The mathematical modeling of enzymatic reactors for esterification of fatty acids with sugars in the production of biosurfactants has been a useful tool for studying and optimizing the process. In particular, artificial neural networks and fuzzy systems emerge as promising methods for developing models for those processes. In this work, regarding artificial neural networks application, coupling of networks to reactor mass balances was considered in hybrid models to infer reactant concentrations over time. Computationally, an algorithm was constructed incorporating material balances, neural reaction rates, and step-by-step numerical integration (employing the classical Runge-Kutta method). Besides, based on an available set of experimental data, fuzzy logic was applied for modeling and optimization of the conversion of esterification as a function of operational process parameters (such as time, temperature and molar ratio of substrates). All computational development was carried out using the Matlab software. The neural networks were able to predict the kinetic behavior of xylose esterification process by applying them to reactor mass balances, obtaining R2 values above 0.94, indicating a good predictive capacity. The trained fuzzy models were able to simulate the relationships between input variables and the output variable, enabling the construction of surfaces and estimating the optimal operational condition at 60 h of reaction, 55°C, molar ratio of substrates of 5:1 and enzyme loading of 37.5 U/g. The same condition was obtained when applying the particle swarm optimization algorithm. Thus, this work presented a complete tool based on computational intelligence for modeling, simulation, and optimization of biosurfactant synthesis.

Mixing is a critical operation in numerous industrial processes, traditionally performed in agitated tanks to ensure homogenization. Despite its importance, the design of tanks and impellers is often neglected during agitation system selection, resulting in excessive energy consumption and inefficient mixing. To mitigate these challenges, Computational Fluid Dynamics (CFD) serves as a powerful tool for analyzing tank hydrodynamics and quantifying mixing times. CFD employs mathematical models to simulate mass, heat, and momentum transport phenomena within fluid systems. Among the latest advancements in modeling stirred tank hydrodynamics is Smoothed Particle Hydrodynamics (SPH), a mesh-free Lagrangian approach that tracks individual particles characterized by properties such as mass, position, velocity, and pressure. SPH provides significant advantages over traditional mesh-based methods by accurately capturing fluid behavior through particle interactions. In this study, the performance of three impellers—double ribbon, paravisc, and hybrid—was evaluated based on hydrodynamics and mixing times during the homogenization of water and ethanol in a 0.5 L stirred tank. The tank and impellers were meticulously designed, operating at 70% capacity, with the fluids exhibiting the following rheological properties: ?1 = 1000 kg/m³, ?2 = 789 kg/m³, µ1 = 1E-6 m²/s, and µ2 = 1.52E-6 m²/s. The simulations were conducted under turbulent flow conditions (Reynolds number of 10,000) for a duration of 2 minutes using the DualSPHysics software. The stirring speed was set at 34 rpm, and the initial particle spacing was configured to 1 mm, generating 270,232 fluid particles and 187,512 boundary particles representing the tank and agitator. The analysis included velocity profiles, flow patterns, vorticity, divergence, and density fields to assess mixing performance. The Q-criterion was employed to discern the dominance of rotational or deformational motion and to identify stagnation zones. The results revealed that the double ribbon impeller exhibited superior performance, achieving 88.28% mixing within approximately 100 seconds, while the paravisc and hybrid impellers reached mixing efficiencies of 12.36% and 11.8%, respectively. These findings underscore the potential of SPH as a robust computational approach for linking hydrodynamic behavior with mixing efficiency and identifying key parameters to optimize mixing processes.

Optimizing twin-screw extruder (TSE) performance is critical in the plastics industry but is often resource-intensive. This study introduces a novel surrogate modeling approach using a Recurrent Deep Embedding Network (RDEN) that integrates deep autoencoders with recurrent neural networks to capture sequential dependencies and physical relationships in TSE processes. Leveraging Progressive Latin Hypercube Sampling (PLHS), the RDEN achieves robust predictions of key process variable, like mean residence time. Results demonstrate the model’s accuracy, generalization capabilities, and potential for automated screw design optimization.

This article studies four dimple shapes: spherical, smoothed-spherical, normal distribution, and error distribution and how they enhance heat transfer on a plate within a plate heat exchanger using computational fluid dynamics. The dimple that showed the greatest efficiency of heat transfer was the normal distribution dimple, giving a temperature increase of 7.5 times of the smoothed-spherical and 15% more than the error distribution dimple shape. This was primarily due to the large increase in the turbulent kinetic energy caused by the eddies created upon the flow over the normal distribution shape. With the normal distribution shape being found to be the most effective in enhancing heat transfer, a layout of multiple normal distribution dimples based on the stage of flow development was also studied. It was found that a fully developed flow resulted in 9.5% more efficiency than half developed flow and 31% more efficient than placing dimples directly next to each other.

This study introduces the novel application of a Multi-Task Gaussian Process (MTGP) model to predict biogas production and critical anaerobic digestion (AD) performance indicators (soluble COD, volatile fatty acids (VFAs)), addressing feedstock variability and dynamic process behavior. We compare the MTGP against the widely used mechanistic AM2 model to evaluate its accuracy and applicability for probabilistic modeling in AD systems. The MTGP framework leverages multi-output correlations and uncertainty quantification, trained on experimental data, achieving superior predictive performance over AM2in this study, with lower RMSE (SCOD: 0.32 g/L; VFAs: 0.87 mmol/L; biogas: 0.15 L/day) and higher R² values (SCOD:0.91, VFAs:0.94, biogas :0.88) under the conditions tested. While AM2 provides biochemical insights, its reliance on unvalidated assumptions may limits robustness. The flexibility of MTGP and precision suggest its potential for real-world applications such as Bayesian Optimization and Design of Experiments, enabling data-driven process enhancement without mechanistic constraints. This work establishes MTGP as a pioneering tool for AD optimization, bridging data-driven efficiency with practical bioenergy challenges.

The synthesis of membrane networks to recover components from liquid mixture is challenging due to an extensive array of feasible network configurations and the added complexity of modeling membrane permeators caused by nonidealities in liquid mixtures. We present a mixed-integer nonlinear programming (MINLP) framework for synthesizing membrane networks to recover multiple components from liquid mixtures. First, we develop a physics-based nonlinear surrogate model to accurately describe crossflow membrane permeation. Second, we propose a richly connected superstructure to represent numerous potential network configurations. Third, the two aforementioned elements are integrated into an MINLP model to determine the optimal network configuration. Finally, the effectiveness of the proposed approach is demonstrated through a range of applications.

Data-driven soft sensors help in process operations by providing real-time estimates of otherwise hard to measure process quantities, e.g., viscosities or product concentrations. Currently, soft sensors need to be developed individually per plant. Using transfer learning, machine learning based soft sensors could be re-used and fine-tuned across plants and applications. However, transferring data-driven soft sensor models is in practice often not possible, because the fixed input structure of standard soft sensor models prohibits transfer if, e.g., the sensor information is not identical in all plants. We propose a topology-aware graph neural network approach for transfer learning of soft sensor models across multiple plants. In our method, plants are modeled as graphs: Unit operations are nodes, streams are edges, and sensors are embedded as attributes. Our approach brings two advantages for transfer learning: First, we not only include sensor data but also crucial information on the plant topology. Second, the graph neural network algorithm is flexible with respect to its sensor inputs. This allows us to model data from different plants with different sensor networks. We test the transfer learning capabilities of our modeling approach on ammonia synthesis loops with different process topologies [1]. We build a soft sensor predicting the ammonia concentration in the product. After training on data from one process, we successfully transfer our soft sensor model to a previously unseen process with a different topology. Our approach promises to extend the data-driven soft sensors to cases to leverage data from multiple plants.

New ways of reducing environmental impact of solid waste are constantly developed. Thermochemical conversion with focus on material or energy recovery is one of the viable options. To make the feedstock properties more suitable for such a process, refuse derived fuel (RDF) is created. Although several studies have focused on thermochemical conversion in recent years, only few have comprehensively compared the main aspects of incineration, gasification, and pyrolysis processes from multiple aspects. This study focuses on mathematical modeling of these three processes in the Aspen Plus environment. Comparison from economic, safety, and environmental viewpoints was performed. As a base for the calculations, 10 t/h of RDF was selected. All three processes demonstrated the suitability to be used for energy recovery. Pyrolysis showed the greatest potential for material recovery. Payback period was used as a parameter of economic comparison with pyrolysis being the most profitable process. Based on the calculated Dow Fire and Explosion Index (F&EI) and Chemical Exposure Index (CEI) gasification was identified as the most hazardous process. In summary, based on this extensive techno-economic analysis, the ranking of processes from an economic standpoint is as follows: pyrolysis, incineration, gasification. From an environmental perspective, the ranking is: gasification, pyrolysis, incineration. In terms of safety, the order is: incineration, pyrolysis, gasification.

In the upcoming decades, the scale-up of hydrogen production will play a crucial role for the integration of renewable energy into energy system. One scale-up strategy is the numbering-up of standardized electrolysis units in modular plant concepts. The use of modular plants can support the integration of different technologies into heterogeneous electrolyzer plants to leverage technology-specific advantages and counteract disadvantages. This work focuses on the analysis of technical operability of large-scale modular electrolyzer plants in heterogeneous plant layouts using co-simulation. Developed process models of low-temperature electrolysis components are combined in Simulink as shared environment. Strategies to control process parameters, like temperatures, pressures and flowrates in the subsystems and the overall plant, are developed and presented. An operability analysis is carried out to verify the functionality of the presented plant layout and control strategies. The dynamic progression of all parameters is presented for different operative states, like start-up, continuous operation, load change and hot-standby behavior. It is observed that the exemplary plant is operational, as all relevant process parameter can be held within the allowed operating range during all operative states. Some limitations, regarding the possible operating range of individual technologies, are introduced and solution approaches are conceptualized. Additionally, relevant metrics for efficiency, such as the specific energy consumption and flexibility are calculated to prove the feasibility and show the advantages of heterogeneous electrolyzer plant layouts, such as a heightened operational flexibility without mayor reductions in efficiency.

One of the current challenges for hydrogen-related technologies is its storage and transportation. The low volumetric density and low boiling point require high-pressure and low-temperature conditions for effective transport and storage. A potential solution to these challenges involves storing hydrogen in chemical compounds that can be easily transported and stored, with hydrogen being released through decomposition processes. Ammonia stands out as a promising hydrogen carrier due to its high hydrogen content (17.8% by weight), relatively mild liquefaction conditions (~10 bar at 25°C), and the availability of a well-established storage and transportation infrastructure. The objective of this study was to develop a mathematical model to analyze and design a membrane fixed-bed reactor (MFBR) for large-scale ammonia decomposition. The kinetic model for the Ru-K/CaO catalyst was obtained from the literature and validated using the experimental data reported in the original study. This catalyst was selected due to its effective performance under high-pressure conditions, which increases the drive force for hydrogen permeation in the membrane reactor. The model was developed in Aspen Custom Modeler (ACM) using a 1D pseudo-homogeneous approach. The governing equations for mass, energy, and momentum conservation were discretized via a first-order backward finite difference method. The effectiveness factor was incorporated to account for intraparticle mass transfer limitations, which are prevalent with the large particle sizes typically employed in industrial applications. The study further investigated the influence of sweep gas ratio, temperature, relative pressure, and space velocity on ammonia conversion and hydrogen recovery, employing response surface methodology generated through an ACM-Python interface. The proposed multi-tubular membrane reactor achieved approximately 94% ammonia conversion and 90% hydrogen recovery, operating at an inlet temperature of 380°C and a pressure of 40 bar. Under comparable conditions of temperature, pressure, and WHSV, the membrane reactor demonstrated an approximately 41.7% higher ammonia conversion compared to a conventional fixed-bed reactor. Furthermore, the developed model is easily transferable to Aspen Plus, facilitating subsequent process conceptual design and economic analyses.

Power-to-Ammonia technology offers sustainable pathways for energy storage and chemical production, with fixed-bed reactors being critical components for efficient synthesis. Understanding reactor dynamics under varying conditions is essential for optimizing these systems, particularly when integrated with intermittent renewable energy sources. This study aims to develop and validate a 2D axisymmetric CFD model for analysing the dynamic response of a ruthenium-catalysed ammonia synthesis reactor to thermal perturbations. The model incorporates detailed reaction kinetics, multicomponent mass transport, and heat transfer mechanisms to predict system behaviour under transient conditions. Results reveal that a step increase in wall temperature from 400°C to 430°C enhances NH3 concentration by 136% (from 2.2 to 5.1 vol.%), with rapid system stabilization achieved within 0.5 seconds. The thermals response maintains consistent heat transfer patterns, exhibiting ~400K differentials between inlet and maximum temperature zones

This study compares the probability density function (PDF) of the power generated by a wind farm obtained analytically with the PDF considering the wake effect between wind turbines, a phenomenon that reduces the power generation capacity of wind farms. Instead of considering the wake effect in the analytical method, which is complex and difficult to solve, it has been proposed to use kernel estimators to obtain the PDF. To calculate it, a wind farm power output data set has been used. This data set was generated using historical wind speed and direction data and the Katic multiple wake model. Discrepancies between the analytical PDF and PDF fitted with the kernel estimators, can lead to an overstatement of the annual available energy by 4 an 9 %, depending on the complexity of the wind farm layout. These inconsistencies can have significant implications for production planning, wind farm design, and integration of wind power into the grid. Therefore, this analysis underscores the necessity of incorporating the wake effect in wind farm modelling, to guarantee more precise projections of generation and energy availability.

Pillow-plate heat exchangers (PPHEs) represent a viable alternative to conventional shell-and-tube and plate heat exchangers. The waviness of their channels intensifies fluid mixing in the boundary layers and facilitates heat transfer. Applying secondary surface structuring can further enhance the overall thermo-hydraulic performance of PPHEs, thus increasing their competitiveness against conventional heat exchangers. In this work, streamlined secondary structures applied on the PPHE surface were studied numerically to explore their potential in enhancing near-wall fluid mixing. Computational fluid dynamics (CFD) simulations of single-phase turbulent flow in the inner PPHE channel were performed and pressure drop, heat transfer coefficients, and overall thermo-hydraulic efficiency were determined. The simulation results clearly demonstrate a favourable influence of secondary structuring on the heat transfer performance of PPHEs.

The three systems, water, energy and food, are intertwined since the effect of any of these systems can affect others. This study proposes a mathematical model incorporating uncertain parameters in the biomass energy-water-food nexus system. The novel aspects of this work include formulating and solving the problem as a mixed-integer linear program and addressing the presence of uncertain parameters through a two-stage stochastic mathematical programming approach. Taking maximising economic benefit as an objective function, this work compares the results of the deterministic model with the results computed by incorporating uncertainty in the model parameters. The results indicate that incorporation of uncertainty gives rise to reduced profitability, but increased greenhouse gas emission (GHG) as compared to the deterministic model. On the other hand, when minimisation of GHG emission is considered as an objective function, a significantly greater reduction in the profitability is observed for both, stochastic and deterministic, models. The model results are used for allocating optimal resources, reducing carbon footprint, increasing economic potential and managing resources sustainably in the whole system.

This work presents an improved thermodynamic model, an equilibrium model, and a unified kinetic model for ammonia synthesis. The thermodynamic model accurately describes the non-ideality of the reaction system up to 1000 bar using a modified Soave-Redlich-Kwong Equation-of-State. The developed Langmuir-Hinshelwood kinetic model accurately describes ammonia synthesis on iron-based catalysts by incorporating N* and H* surface species, whereas H* species are mainly relevant below 400°C. The model fits an extensive dataset across diverse conditions (251-550°C, 1-324 bar, H2/N2 ratios 0.33-8.5, and space velocities of 1-1800 Nm3 kg-cat-1 h-1) and accounts for catalyst activity variations through a Relative Catalytic Activity factor.

This study presents the application of a Q-learning algorithm to optimize the selection of chemical reactions for methane reforming processes. Starting with a set of 11 candidate reactions, the algorithm identified three key reactions. These reactions effectively represent the experimental data while aligning with the underlying physics of the process and previously reported findings. The algorithm employed an epsilon-greedy policy to balance exploration and exploitation during the training process. Furthermore, simulations based on the identified reactions revealed trends consistent with experimental data. This work highlights the efficiency and adaptability of Q-learning in modeling complex catalytic systems and provides a framework for further exploration and optimization of methane reforming processes.

Fluid catalytic cracking is a crucial process in the refining industry, capable of converting lower-quality feedstocks into higher-value products. Due to the variability in feedstock properties and fluctuations in product market prices, timely adjustment and optimization of the FCC unit are essential. In this context, data-driven models have garnered increasing attention for their capacity to handle the complex, nonlinear reactions involved in the FCC process. However, on account of the limited operating range of the plants and the black-box nature of data-driven models, relying solely on these models for optimization may lead to contradictory decisions in optimization processes. To address these challenges, we integrate gradient information of product yields with respect to key variables derived from the mechanistic model Petro-SIM, into the training process of data-driven models. To mitigate the high computational demands of the Petro-SIM model, we propose the use of active learning methods for efficient sampling and thereby constructing a surrogate model. The results demonstrate that the active learning approach reduces the required sampling size by 25%. More importantly, the data-driven model trained with gradient information improves the accuracy of trend direction prediction by 34.6%, significantly enhancing its effectiveness in supporting the optimization process. The code will be available at https://github.com/xwl514/fcc-hybrid-loss.

Sustainability is a thriving global topic of concern and following the advancement of technological progress and increased standards of living, the demands for energy, fuels, chemicals and other requirements have increased significantly. Methanol is one such chemical which has seen increases in demand due to its importance as a precursor in the development of widely used chemicals such as formaldehyde. In order to gain insight into the reaction mechanisms driving the process, it is beneficial to develop kinetic models that accurately describe the system for several reasons: (i) to develop process understanding; (ii) to facilitate control and optimisation; (iii) to reduce experimental burdens; and (iv) to expedite scale up and scale down of processes. Two commonly used kinetic reaction rate models are the power law and Langmuir-Hinshelwood expressions, however the strong assumptions made when developing such models may limit their predictive performance through the introduction of inductive bias (i.e. model structural uncertainty). A solution to counter these drawbacks is known as hybrid modelling where, the inauguration of a data-driven component within the kinetic modelling framework allows for any complex, less understood kinetics to be instead learnt from historical data by a machine learning model. In order to identify the pros and cons associated with each kinetic and hybrid modelling strategy for chemical reaction network modelling, a thorough comparison was made using syngas methanol production as a case study. It was shown that hybrid models offered increased predictive accuracy, robust uncertainty quantifications, and improved generalisability under limited data availability.

Planning and scheduling are integral components of process supply chains. The presence of data correlation, particularly multivariate demand data dependency, can pose significant challenges to the decision-making process. This necessitates the consideration of dependency structures inherent in the underlying data to generate good-quality, feasible solutions to optimisation problems such as planning and scheduling. This work proposes a chance-constrained optimisation framework integrated with copulas, a non-parametric data estimation technique to forecast uncertain demand levels in accordance with specified risk thresholds in the context of a planning and scheduling problem. We focus on the integrated planning and scheduling problem following a bi-level optimisation formulation. The estimated demand forecasts are subsequently utilised within the Data-driven Optimisation of bi-level Mixed-Integer NOnlinear problems (DOMINO) framework to solve the integrated optimisation problem, and derive decisions with guaranteed demand satisfaction rates. Computational experiments demonstrate that our proposed copula-based chance-constrained optimisation framework can incorporate demand correlation and achieve higher joint demand satisfaction rate, lower total costs with higher efficiency.

Process monitoring is essential in industrial settings to ensure system functionality, necessitating the identification and understanding of fault causes. While a substantial body of research focuses on fault detection, fault diagnosis has received significantly less attention. Typically, faults originate either from abnormal instrument behavior, indicating the need for calibration or replacement, or from process faults, signaling a malfunction within the system. A primary objective of this study is to apply the proposed fault diagnosis methodology to a benchmark that closely mirrors real-world conditions. Specifically, we introduce a fault diagnosis framework for a wastewater treatment plant (WWTP) that effectively addresses the challenges posed by imbalanced big data commonly encountered in large-scale systems. In our study, four distinct fault scenarios were investigated: fault-free conditions, process faults only, sensor faults only, and simultaneous sensor and process faults. To enhance our research, we explored various techniques and approaches to optimally address the fault diagnosis problem. Implementing an undersampling technique based on clustering and determining appropriate hyperparameters proved challenging, alongside deciding the most effective approach for undersampling. Additionally, advanced time-series models, including Long Short-Term Memory (LSTM) and Bidirectional LSTM (BiLSTM), were employed, with particular emphasis on identifying the correct regularization factor and number of epochs being critical. When optimal model parameters are selected, each fault scenario achieves an accuracy of at least 90%, and the maximum simulation time does not exceed 15 minutes. Furthermore, we developed a Python-based GUI for this study to enable users to fully explore the capabilities of the proposed fault diagnosis algorithm.

This study models a Heat Recovery System (HRS) within a petrochemical plant, assessing its economic and environmental viability. The system integrates four combustion processes and a condensing steam turbine combined heat and power (ST-CHP) generation unit, along with waste heat recovery technologies to reduce the plant’s energy use. The developed system-based approach extends a previous methodology, initially focused on reducing energy consumption in production processes, to encompass energy supply systems (in which CHP is included) as well. Simulation models were developed for two improvement scenarios regarding the integration of the ST-CHP into the HRS: preheating either the combustion air stream or the inlet water of the ST-CHP’s boiler. The latter demonstrated greater potential for reducing energy-related operational costs, thus an NLP optimisation model was developed based on that scenario. Both simulation and optimisation models were created resorting to the capabilities of the ThermWatt computational tool (comprising Modelica and Python languages). A payback period of about 2 years and 3 months and a reduction in equivalent carbon dioxide emissions of 5.14 kt CO2,eq/year were estimated for the final system, which are significantly reasonable in comparison to benchmark values.

Hydrogen-powered heavy-duty trucks have a high potential to significantly reduce CO2 emissions in the transportation sector. Therefore, efficient hydrogen storage onboard vehicles is a key enabler for sustainable transportation, as achieving high storage densities and extended driving ranges is essential for the competitiveness of hydrogen-powered trucks. Cryo-compressed hydrogen (CcH2), stored at cryogenic temperatures and high pressures, emerges as a promising solution. This study presents a comprehensive dynamic thermodynamic model that is capable of simulating the tank system across all operating conditions and, therefore, enables thermodynamic analysis of drive cycles. The core of the model is a differential-algebraic equation system that describes the thermodynamic state of the hydrogen in the tank. Additionally, surrogate models based on artificial neural networks are applied to efficiently describe quasi-steady-state heat exchangers integrated into the tank system. Several use cases are explored to demonstrate the model’s ability to simulate the thermodynamic behavior and to find optimal operating strategies. The optimal hydrogen density, when to stop driving and refuel the tank to maximize overall driving ranges, is investigated both in ideal and real operation, taking into account the limited availability of refueling stations in early market applications. Further, driving range and venting losses are considered for longer periods of dormancy. These results provide insights into how operational strategies can be tailored to maximize driving range, minimize hydrogen losses, and improve overall system efficiency, ultimately supporting the adoption of hydrogen in long-haul transportation.

Dynamic modelling of wind turbines and their simulation is a very useful tool for studying their behaviour. One of the key elements concerning the physical models of wind turbines is the power coefficient Cp, which acts as an efficiency in the extraction of power from the wind. Unfortunately, this coefficient is often unknown a priori, as it does not usually appear in the information provided by manufacturers. This paper first describes a methodology for obtaining the power coefficient parameters of a commercial wind turbine model using the power curve provided by the manufacturer, which indicates the theoretical power that the wind turbine can produce at each wind speed. To achieve this, a parameter estimation problem is formulated and solved to determine the power coefficient parameters. Nevertheless, this information is often insufficient, requiring additional knowledge, such as operational data, to improve the fit. Finally, a new parameter estimation is performed using only real data measured from the process, validating the proposed methodology. The operational dataset was obtained from the SMARTEOLE project in France and corresponds to the pitch-controlled variable-speed Senvion MM82/2050 wind turbine.

This study evaluates the potential for computer-aided real-time monitoring and decision-making in pilot-scale ion-exchange chromatography operations using only historical data from the pilot-scale facility. Historical data of flow and conductivity were utilized from students running pilot-scale ion exchanges that resemble industrial ion exchange processes. A Lumped Rate Model (LRM) with a Steric Mass Action (SMA) isotherm was implemented and parameterized to characterize the fixed-bed column. The Discontinuous Galerkin Spectral Element Method (DGSEM), implemented in CADET-Julia, enabled efficient simulation and parameter estimation. Using DGSEM, the LRM with SMA was solved in less time than the sensor measurement frequency. This development allows for the prediction of batch evolution in real time for operators of the ion-exchange column. Despite challenges related to data preprocessing and manual operation inconsistencies, the results demonstrate the feasibility of integrating real-time analysis into pilot-scale operations.

Clinker is the main constituent of cement, produced in the pyroprocessing section of the cement plant. This comprises some high temperature and carbon intensive processes, which are responsible for the vast majority of the CO2 emissions associated with cement production. This paper presents first-principles mathematical models for the simulation of the pyroprocess section; more specifically the preheating cyclones, the calciner and the rotary kiln. The models incorporate material and energy balances, the major heat and mass transport phenomena, reaction kinetics and thermodynamic property estimation models. These mathematical formulations are implemented in the gPROMS® Advanced Process Modelling Environment and the resulting index-1 DAE (Differential Algebraic Equation) system can be numerically solved for various reactor geometries and operating conditions. The process models developed for each unit are then used to build a cement pyroprocess flowsheet model. The flowsheet model is validated against published data, demonstrating the ability to predict accurately operating temperatures, degree of calcination, gas and solid compositions, fuel consumption and overall CO2 emissions. The substitution of conventional coal with more sustainable fuels is also investigated, to evaluate the potential for avoiding CO2 emissions by replacing part of the fossil-based coal fuel (used as a reference case). Trade-offs between different process KPIs (f.e. calcination efficiency, specific CO2 emissions per tonne of clinker) are identified and evaluated for each fuel utilization scenario.

This study presents an economic analysis of a blue H2-N2 production system, using a novel intensified reformer system with a hydrogen production efficiency of 80%. The system’s ability to produce both high-purity H2 and N2 creates opportunities for small-scale blue H2 and distributed ammonia production. The system consists of three identical, optimized fixed-bed reforming reactors, a heat recovery system, and shift reactors. A dynamic model was developed to simulate three small-scale H2 production systems: 2.8 tpd, 7.1 tpd, and 17.1 tpd, enabling an evaluation of their economic viability. The results indicate that the cost of H2 production ranges from 2.7 to 3.1 USD/kgH2. Sensitivity analysis reveals that natural gas and CO2 transportation costs have a significant impact on the variability of H2 price. This research provides valuable insights into the economic feasibility of small-scale blue hydrogen production, offering a pathway to support the broader adoption of hydrogen technologies in advancing a cleaner energy future.

A Digital Twin (DT) is a purposeful digital representation of a physical entity that employs data, algorithms, and software to enhance operations, making it possible to e.g., forecast failures, or evaluate new designs through the simulation of real-world scenarios. DTs are enablers for real-time monitoring, simulation, and optimization. However, traditional simulation DTs often rely on complex, non-linear mechanistic models with high computational demands, complex structures, and a large number of specific parameters and thus pose quite a challenge to maintainability. Surrogate models, on the other hand, are simplified approximations of more complex, higher-order models. These approximations are typically built using data-driven approaches, such as Random Forest Regression, facilitating faster simulations, simpler adaptation, and quicker deployment. This study analyzes the complexity of mechanistic and surrogate modeling approaches in the context of DTs to aid model selection. A model with reduced complexity enhances computational efficiency, simplifies implementation, and supports real-time monitoring and predictive maintenance. Complexity analysis evaluates metrics such as analytical, structural, space, behavioral, training, and prediction complexity, resulting in an overall complexity score for model selection. However, the decision involves trade-offs, such as balancing high fidelity with low complexity or prioritizing high explainability over structural simplicity. Addressing these trade-offs is essential in selecting a model that balances the accuracy, usability, and efficiency of DTs. Using a stirred tank reactor as a use case, the mechanistic model is compared to a surrogate model to quantify complexity scores and select a less complex model for DT development.

This study evaluated the air/steam co-gasification of crude glycerol (CG) and linear low density polyethylene (LLDPE). It was demonstrated that operating the process using air or a mixture of air and steam has significant implications for carbon conversion efficiency (CCE), cold gas efficiency (CGE), lower heating value (LHV) gasifier output temperature and syngas concentration. The CCE reached a maximum value of 100% at equivalence ratio (ER) of 0.3 for 25% LLDPE and an ER of 0.35 for 75% LLDPE when air was used. When steam was introduced in the gasifier at a fixed rate (SFR =0.5), the CCE of 100% was maximised at ER of 0.25 for 25% LLDPE and 0.3 for 75% LLDPE content. An increase in the steam to feedstock ratio (SFR) did not alter the CCE for 25% LLDPE at a constant ER, but for that of 75% LLDPE, a CCE was maximized at an SFR of 0.25. In the case of CGE, a maximum value of 79.24% and 78.12% was reached at ER of 0.3 and 0.35 for 25% LLDPE and 75% LLDPE respectively when pure air was used. Keeping the SFR constant, the CGE slightly increased to 81.33 and 81.45% at ER of 0.25 and 0.3, respectively for 25% LLDPE and 75% LLDPE. The LLDPE content did not significantly alter the syngas LHV at a fixed ER or SFR. The gasifier temperature increased with an increase in ER but decreased when steam was fed into the gasifier. Furthermore, the syngas ratio (H2:CO) achieved in this study can be useful for the synthesis of methanol and dimethyl ether.

Nucleophilic aromatic substitutions (SNAr) are key chemical transformations in pharmaceutical and agrochemical synthesis, yet their complex mechanisms (concerted or two-step) complicate kinetic model identification. Accurate kinetic models for SNAr are essential for scale-up, optimization, and control of the reaction process, but conventional methods struggle with mechanism uncertainty driven by substrates, nucleophiles, and reaction conditions, with data collection being difficult due to its source-intensive nature. We address this using DoE-SINDy, a data-driven framework for generative modelling without complete theoretical understanding. A benchmark study on the SNAr reaction of 2,4-difluoronitrobenzene with morpholine in ethanol was conducted, incorporating parallel and consecutive side-product formation. Ground-truth kinetic models validated in prior studies were used to generate in-silico data under varying noise levels and sampling intervals. DoE-SINDy successfully identified the true kinetic model with minimal runs, quantifying the impact of key design factors such as inlet concentrations, residence time, sample size and experimental budget on model identification.

This study evaluates the introduction of Carbon Capture and Utilization (CCU) process in two Colombian refineries, focusing on their potential to reduce CO2 emissions and their associated impacts under a scenario aligned with the Net Zero Emissions by 2050 Scenario defined in the 2023 IEA report. The work uses a MILP programming tool (Linny-R) to model the operational processes of refinery sites, incorporating a net total cost calculation to optimize process performance over five-year intervals. This optimization was constrained by the maximum allowable CO2 emissions. The methodology includes the calculation of surplus refinery off-gas availability, the selection of products and CCU technologies, and the systematic collection of data from refinery operations, as well as scientific and industrial publications. The results indicate that integrating surplus refinery fuel gas (originally used for combustion processes) and HTL bio-crude off-gas (as a source of biogenic CO2) can significantly lower scope 1 and 2 CO2 emissions, aligning with long-term decarbonization goals. However, these advantages carry additional costs due to significant increases in utility demands. In the high-complexity refinery, electricity consumption increases by a factor of 10, and steam demand and water usage each increase by a factor of 2.5 and 3, respectively. Similarly, in the medium-complexity refinery, electricity consumption rises by a factor of 11, steam demand by a factor of 4, and water usage by a factor of 6. The renewable energy requirements for water electrolyzers and CO2 capture units primarily drive these increases. Furthermore, despite achieving CO2 neutrality in scope 1 and 2 emissions by 2050, scope 3 emissions increase due to additional CO2-based methanol production. Economic analyses highlight profit opportunities in the long term. The production costs of CO2-based methanol are lower than the forecasted cost of production of fossil-based methanol, enhancing their economic viability in the long term. The study emphasizes the critical influence of refinery complexity levels on the scale and timeline for implementing these technologies to achieve short- and long-term CO2 reduction targets. However, further evaluation is necessary to align these results with national electrical grid capacity, water supply availability, and expansion plans.

Pressure-swing distillation (PSD) is a frequently applied method to separate pressure-sensitive azeotropic mixtures; however, its energy demand is very high. In continuous mode, PSD is performed in a system consisting of a high- and a low-pressure column. If the composition of the feed is between the azeotropic compositions at the two pressures, it can be introduced into any of the columns, leading to two possible column sequences. Depending on the feed composition, one of the sequences is optimal whether in terms of energy demand or total annual cost (TAC). In the present work, surrogate model-based optimization is applied to determine the optimal TAC value as a function of the feed composition between the azeotropic ones. As a first step, the column sequence with feeding into the high-pressure column is studied here. The mixture to be separated consists of water and ethylenediamine, which form a maximum-boiling azeotrope. The columns are modeled separately and a large number of simulations are performed in points generated by Latin hypercube sampling, then, algebraic surrogate models are fitted to the simulation results. The resulting surrogate optimization problem is solved for different feed composition values and the accuracy of the models is verified by rigorous simulations, as well. The energy cost and TAC has a maximum as a function of the composition. The error of the estimation of TAC remains under 6 % except at the lowest water concentration studied.

Reject brine discharge and high CO2 emissions from desalination plants are major contributors to environmental pollution. Managing reject brine involves significant costs, mainly due to the energy-intensive processes required for brine dilution and disposal. In this context, Solvay process represents a mitigation scheme that can effectively reduce reject brine salinity and sequestering CO2 while producing sodium bicarbonates simultaneously. The Solvay process represents a combined approach that can effectively manage reject brine and CO2 in a single reaction while producing an economically feasible product. Therefore, this study reports a systematic techno-economics assessment of conventional and modified Solvay processes, while incorporating brine and carbon tax. The model evaluates the significance of implementing a brine and CO2 tax on the economics of conventional and Ca(OH)2 modified Solvay compared to industries expenditures on brine dilution and treatment before discharge to the sea. The results show that the conventional Solvay process becomes profitable after applying a brine tax of 1 dollar per meter cube of brine and a CO2 tax of 42 dollars per tonne CO2 —both figures lower than the current costs associated with brine treatment and existing carbon taxes. Moreover, the profitability of the Ca(OH)2-modified Solvay process increases even further with minimal brine and CO2 taxes. The findings highlight the significance of adopting modified Solvay process as an integrated solution for sustainable brine management and carbon capture.

This study explores multiphase flow dynamics with a focus on the annular flow regime using Computational Fluid Dynamics (CFD) simulations. The methodology included defining the physical model, generating the computational mesh, and analyzing flow patterns. The Volume of Fluid (VOF) model captured fluid interactions, while the k-? SST turbulence model ensured accurate flow predictions. Simulations examined mixture density behavior and identified optimal configurations. A dataset was generated and analyzed using k-means clustering to classify flow patterns effectively. The results demonstrate the reliability of this approach for improving multiphase flow systems, with applications in oil-water processes.

A new methodology for the assessment of the performance loss in catalytic reactors due to deactivation was developed and applied to fixed- and fluidized-bed CO methanation, with catalyst subject to coking. The methodology is based on the solution of heat and mass balances, by decoupling the reactor and deactivation dynamics. This is possible by using consecutive 1D, steady-state calculations for the characterization of the reactor performance. In this way, the progressively lower values of catalyst activity along the time on stream are computed with the integration of a dedicated dynamic model. This method has shown promising results in the characterization of the loss of performance of the reactor over time. The model correctly describes a progressive deactivation of the catalyst in fixed-bed reactors, while it shows that the decrease in activity is sudden for the whole reactor volume in fluidized bed reactors and occurs after a critical time-on-stream. Besides, it was observed that the fluidized bed reactor is in general more resistant to deactivation by CO, with the time needed to reach 25% conversion being 1.5 to 2 orders of magnitude higher than for a fixed bed in similar conditions – for example around 250 hours against 50 for a fixed bed at 300°C. This will allow optimizing the reactor with respect to deactivation, hinting the modifications possible in the conditions or in the reactor geometry to maximise the time-on-stream prior to loss in the product quality.

This study presents a numerical simulation of the liquid-vapor (L-V) equilibrium stage in a sieve plate distillation column using the Smoothed Particle Hydrodynamics (SPH) method. To simulate the equilibrium stage, periodic temperature boundary conditions were applied. The column design was carried out in Aspen One, considering an equimolar benzene-toluene mixture and an operating pressure ensuring a condenser cooling water temperature of 120°F. The Chao-Seader thermodynamic model was employed for property calculations. Key outputs included liquid and vapor velocities per stage, mixture viscosity and density, operating pressure, and column diameter. The geometry of the distillation column stage and sieve plate was developed using SolidWorks, and Computational Fluid Dynamics (CFD) simulations were performed using the DualSPHysics code. The results demonstrate the influence of sieve plate design on velocity and temperature distributions within the stage, providing insights for enhancing stage design and operational efficiency.

Dual function materials (DFMs) integrate CO2 capture and conversion, offering a streamlined approach to Power-to-Gas (PtG) processes. This study develops a cyclic steady-state model for the DFM-based methanation of CO2 using the finite difference method. The model captures the adsorption, purge, and methanation stages and incorporates a semi-implicit numerical scheme for stability and accuracy. Bayesian optimisation is used to explore operational and design parameters to maximise methane productivity, CO2 conversion, and product purity. Multi-objective optimisation reveals key trade-offs among these metrics, while the impact of pressure, hydrogen concentration, DFM weight, geometry and cycle times is systematically evaluated. Results reveal that lower flow rates enhance recovery and purity, while higher flow rates improve productivity. Extended adsorption times favour purity, whereas longer methanation times significantly benefit recovery and productivity. Multi-objective optimisation, presented through Pareto fronts, highlight trade-offs among performance metrics. Variations in hydrogen pressure, DFM mass, and H2 concentration further emphasise their impact on the system’s efficiency. This work provides a scalable, reproducible framework for optimising DFM-based PtG systems, advancing their potential for sustainable energy storage and decarbonisation.

Dynamic process modeling in process industries has been extensively studied, especially with the development of deep learning techniques. Recurrent neural networks (RNN) and stacked autoencoders (SAE) are two powerful tools for dynamic modeling and data processing. However, most existing research primarily focuses on extracting features from process input data, often neglecting the temporal autocorrelation of output variables. In this work, a hierarchical model based on time-series RNN structure is proposed. The upper layer employs a long short-term memory (LSTM) network to extract temporal features from process input data. The lower layer uses a gated recurrent unit (GRU) to model the temporal dependencies of output variables across samples. These two parts are concatenated to form the model. Additionally, SAE is utilized to perform dimensionality reduction and reconstruction of process input, seamlessly integrating the reconstruction process with the RNN into a unified framework, termed the AR-SAE-RNN model. Distributed training is employed to effectively learn the model parameters. The proposed AR-SAE-RNN model has been applied to an industrial distillation column. The results demonstrate the model’s effectiveness in capturing the dynamic behavior of the system.

Real-time carbon accounting is crucial for advancing policies that effectively meet sustainability objectives. This work introduces a carbon tracking tool specifically designed for the European electricity grid. The tool collects hourly data on electricity consumption and generation, cross-border power exchanges, and weather information to assess the real-time environmental effects of electricity use, employing locally-specific emission factors for the generation sources. It utilizes weather data from various stations across Europe to produce week-ahead forecasts of carbon intensity in the grid. Predictions are created using a random forest regressor, integrated within the optimal controller of an operational industrial batch process. This prediction-based optimizer seeks to reduce total emissions tied to the process schedule’s electricity consumption by implementing a rolling horizon strategy. By leveraging enhanced energy flexibility, the controller provides significant opportunities for load shifting and emission reductions, though at a higher cost. An industrial batch process demonstration highlights the potential of this approach, showing environmental savings ranging from 4% in grids with low renewable energy to 32% in nations that incorporate a substantial proportion of renewable energy in their energy mix.

The pulp and paper industry has significant potential to reduce its carbon footprint by optimizing energy and water usage, contributing to global efforts toward net-zero emissions. A critical unit operation in this industry is pulp washing using twin roll press washers, which enhance pulp quality. The filtrate from this unit operation can be concentrated and combusted to generate a significant amount of energy for the plant and nearby industries. However, these washers are prone to blockages that disrupt production and decrease energy efficiency. Despite their importance, models for managing blockages in twin roll press washers are lacking. This study addresses this gap by developing empirical models to predict roll speed based on key process parameters. These models performed well on a case study of un-sanitized data from a real pulp and paper plant, achieving R2 of the order of 0.7. The models can potentially be used by operators to ensure uninterrupted production, and optimize resource usage, supporting the industry’s sustainability and decarbonization goals.

An essential problem in precision agriculture is to accurately model and predict root-zone (top 1 m of soil) soil moisture profile given soil properties and precipitation and evapotranspiration information. This is typically achieved by solving agro-hydrological models. Nowadays, most of these models are based on the standard Richards equation (RE), a highly nonlinear, degenerate elliptic-parabolic partial differential equation that describes irrigation, precipitation, evapotranspiration, runoff, and drainage through soils. Recently, the standard RE has been generalized to time-fractional RE with any fractional order between 0 and 2. Such generalization allows the characterization of anomalous soil exhibiting non-Boltzmann behavior due to the presence of preferential flow. In this work, we focus on inverse modeling of time-fractional RE; that is, how to accurately estimate the fractional order and soil property parameters of the fractional RE given soil moisture content measurements. Specifically, we introduce a novel Bayesian variational autoencoder (BVAE) framework that synergistically integrates our in-house developed fractional RE solver and adaptive Fourier decomposition (AFD) to accurately estimate the parameters of time-fractional RE. Our proposed AFD-enhanced BVAE framework consists of a probabilistic encoder, latent-to-kernel neural networks and convolutional neural networks. The BVAE framework is theoretically explainable and enhanced by the AFD theory, a novel signal processing technique that achieves superior computationally efficiency. Through illustrative examples, we demonstrate the efficiency and reliability of our AFD-enhanced BVAE framework.

Decarbonizing the watch industry and improving its sustainability requires enhancing the energy efficiency, intensifying the waste heat recovery and integrating renewable energy resources to significantly reduce the energy consumption and reduce the overall process emissions. The watchmaking industry operates across different periods, each one characterized by own levels of average power demand, plant activity, and duration. While working days (Work D) experience the highest energy demand; other periods, such as weekend (Wend D1, Wend D2 and Wend N), exhibit a substantial drop in energy consumption. It requires a suitable selection of utility systems and an energy planning that can meet the energy requirements regardless of the different operation regimes, reducing the consumption of fossil resources. To this end, this work leverages a systematic framework to perform the techno-economic analysis and explore the feasibility of incorporating alternative energy conversion (e.g. electrification) and storage solutions (for heat) in the watchmaking industry, accounting for the monthly variations of energy demand and environmental conditions.

Liquid Organic Hydrogen Carriers (LOHCs) represent a promising solution for the efficient transport and storage of hydrogen, addressing critical challenges associated with its low volumetric density and safety concerns in gaseous and liquefied forms. LOHCs are oil-like substance, capable of reversibly binding and releasing hydrogen through catalytic hydrogenation and dehydrogenation. Hydrogenation is performed where renewable energy is extensively available: at the loading terminal, green H2 is produced and is chemically bonded to the LOHC molecule. In this way, the hydrogenated molecule is transported safely under ambient conditions using existing liquid fuels infrastructures. At the delivery site, the dehydrogenation process releases high-purity hydrogen for industrial or mobility applications, with the regenerated LOHC carrier being recycled back to the H2 production site for reuse. In view of highlighting critical issues associated to LOHCs implementation at large scale, this paper explores green H2 transport through toluene and dibenzyltoluene. Considering the overall H2 transport value chain, the performances of these two alternatives are assessed by means of equivalent H2 methodology. As a result, insights are provided on their role in enabling a sustainable and efficient hydrogen economy.

As initiatives to decarbonize societies increase, industry is also being considered for policies to encourage its sustainability. Ammonia (NH3) industry relies entirely on Haber-Bosch (HB) process, consuming fossil fuels for hydrogen production and energy purposes, accounting for more than 1 % of anthropogenic carbon dioxide (CO2) emissions. Emerging technologies such as the electrochemical synthesis of NH3 promise sustainable production from water, air, and renewable energies, but low TRLs are still reported. The electrification of the HB process opens a more viable path for sustainable NH3 production in the near future, where hydrogen (H2) is produced by electrolysis of water, powered from renewable energy sources. Many studies have focused on the production of 100 % green NH3 using only electric HB. In this work, a different approach is presented, which consists of studying the gradual incorporation of green H2 into a conventional NH3 plant. An Aspen Plus® V14 model of the methane-fed HB process was developed, and the limitations of the process were identified by sequentially incorporating 0 to 10 % green H2. To avoid overheating the equipment in the natural gas reforming stage, a new configuration is suggested for this step. With this modification, the study is extended to incorporating higher percentages of green H2, up to a limit of around 61 %. Finally, the impact of the reforming temperature tolerance on the green H2 incorporation limits for the suggested configuration was studied.

Simulation plays a crucial role in the design and optimization of gasifiers by providing a detailed understanding of the involved physical processes and complex chemical reactions without the need for extensive trial-and-error experiments. It can also serve as a valuable tool for identifying potential technical issues in experimental devices that operate below expected performance. This study presents a comprehensive simulation of biomass gasification using the open-source software DWSIM. The simulated results were compared with experimental data from a pilot-scale spouted bed reactor, featuring a square-based design with a 20 kWth capacity, using pruning of apple tress as feedstock. Experimental results revealed that the reactor operated effectively at temperatures exceeding 850°C, maintaining stable conditions across a wide range of equivalence ratios. However, the distribution of products—particularly hydrogen (H2)—did not match expected results based on both literature and simulations. A joint analysis of experimental data and expected behavior from simulation helped identify the observed inefficiencies and optimize the reactor’s performance.

In offshore environments, where space, weight, and energy efficiency are critical constraints, the effective removal of water from turbine exhaust gases is essential to enhance gas treatment processes. In this context, replacing conventional methods, such as molecular sieves, with supersonic separators (SSRs) emerges as a promising alternative. This study aims to determine the most suitable operating conditions and design parameters for water removal via supersonic separation (SS) in turbine exhaust gases (TxGs) on offshore platforms. Simulations were performed in Aspen HYSYS using a unit operation extension, based on typical TxGs compositions from offshore platforms. Key parameters, including operating conditions, separator dimensions, and shock Mach number, were evaluated to maximize efficiency while minimizing equipment footprint. The results indicated a water capture efficiency of 99.45%, demonstrating that SS technology is not only compact but also a viable and efficient alternative for water removal from exhaust gases, particularly in space-constrained environments.

The increasing global population has resulted in the scramble for more energy. Hydrogen offers a new revolution to energy systems worldwide. Considering its numerous uses, research interest has grown to seek sustainable production methods. However, hydrogen production must satisfy three factors, i.e., energy security, energy equity, and environmental sustainability, referred to as the energy trilemma. Therefore, this study seeks to investigate the sustainability of hydrogen production pathways through the use of a Multi-Criteria Decision- Making model. In particular, a modified Simultaneous Evaluation of Criteria and Alternatives (SECA) model was employed for the prioritization of 19 options for hydrogen production. This model simultaneously determines the overall performance scores of the 19 options and the objective weights for the energy trilemma in a South African context. The results obtained from this study showed that environmental sustainability has a higher objective weight value of 0.37, followed by energy security with a value of 0.32 and energy equity with the least at 0.31. Of the 19 options selected, steam reforming of methane with carbon capture and storage was found to have the highest overall performance score, considering the trade-offs in the energy trilemma. This was followed by steam reforming of methane without carbon capture and storage and the autothermal reforming of methane with carbon capture and storage. The results obtained in this study will potentially pave the way for optimally producing hydrogen from different feedstocks while considering the energy trilemma and serve as a basis for further research in sustainable process engineering.

Sustainable aviation fuel (SAF) is a proven alternative to reduce CO2 emissions in the aviation sector, supporting sustainable growth. However, SAF processes remain economically uncompetitive with fossil-derived jet fuel, prompting interest in strategies to address these challenges. In 2022, Carrasco-Suárez et al. explored process intensification in the SAF separation zone of a biorefinery using waste cooking oil (WCO), achieving a 3.07% reduction in CO2 emissions and lower operational costs for steam and cooling water. Despite these gains, the WCO biorefinery remains economically unviable with high energy demands. This work presents the energy integration of the entire WCO biorefinery addressed from the pinch point methodology, combined with separation zones intensification (EI-PI-S), using the principles of sections movement for distillation columns; these energy efficiency strategies were applied on the biorefinery in Aspen Plus V.10.0 in order to improve the scheme. Key indicators—total annual cost (TAC), energy investment per product energy (EI-P), energy investment per main product mass (EI-MP), and CO2 emissions per main product mass (CO2-MP)—were used to compare the conventional scheme (CS) and the intensified scheme before energy integration (PI-S). The EI-PI-S scheme achieved the best performance, reducing steam and cooling water requirements by 14.34% and 31.06%, respectively, and CO2 emissions by 13.85% and 14.13% compared to CS and PI-S. However, TAC for EI-PI-S was 0.89% higher than PI-S. Despite this, the integrated and intensified WCO biorefinery emerges as a feasible option for SAF production, adhering to energy minimisation principles and improving economic performance.

Dynamic or Programmable Catalysis is an innovative strategy to improve heterogeneous catalysis processes by modulating the binding energies (BE) of adsorbates on a catalytic surface. The technique enables the periodic favoring of different reaction steps, overcoming limitations imposed by the Sabatier Principle and allowing for higher overall reaction rates, otherwise unattainable. Previously, we implemented a simultaneous simulation approach using the algebraic modeling language Pyomo and the solver IPOPT to obtain cyclic steady state results for a unimolecular reactive system with up to four-order of magnitude increases in computational performance compared to the previously reported sequential approach. The flexibility of the method allowed for the investigation of the influence of forcing signal parameters on system behavior and provided a framework for waveform design. In this study, we use a hybrid framework that combines the sequential and the simultaneous simulation approaches to find the cyclic steady state of a more complex system, of ammonia synthesis, comprising 19 reversible elementary reaction steps. The framework allowed us to investigate sine wave parameters with approximately 220 times less computational effort compared to the sequential approach alone. With the parameters studied, our findings indicate that frequencies exceeding 1000 Hz and compressive strains greater than 2% can negatively impact the system performance. Future work will focus on expanding the model to include lateral interactions between molecules, using other waveform as forcing signals, and integrating systematic mathematical optimization approaches. These advancements pave the way to establishing a general framework for identifying optimal waveforms across diverse dynamic catalysis systems.

Natural gas production is expected to increase, leading to the exploitation of low-quality reserves that contain high levels of acid gases, such as carbon dioxide. The aim of this work is to compare various innovative and conventional technologies for the removal of CO2 from natural gas, considered as a binary mixture of methane and carbon dioxide, with CO2 contents ranging from 10 to 70 mol%. The processes are simulated using Aspen Plus® V9.0 and compared in terms of energy consumption, which is evaluated through the net equivalent methane method. The results show that novel low-temperature and hybrid technologies, which combine distillation and physical absorption, are the most energy-efficient for CO2 removal from natural gas with high acid gas contents, while conventional physical absorption processes are optimal for natural gas with low to moderate acid gas contents.

Temperature regulating in liquid tanks is critical in the chemical industry and conventionally relies on sensor feedback. However, due to the complex thermo-hydrodynamics, unsensed local temperatures can deviate from desired thresholds, underscoring the need for improved tank temperature modeling. The absence of internal thermal or flow data, however, poses significant challenges for the development and validation of effective control strategies. In this study, a virtual model for regulating liquid tank temperature was developed using computational fluid dynamics (CFD). Adaptions were made mainly by involving (1) a simple on-off mechanism of feeding based on a virtual sensor to achieve temperature within the acceptable range and (2) the imposition of unfavorable temperatures on the walls representing ambient influences. Leveraging this virtual system, several new cases were simulated. The simulation results highlighted pronounced temperature non-uniformity, with discrepancies exceeding 2°C between sensor readings and interior points, even leading to inefficiencies of regulatory mechanism. Additionally, the buoyancy was found to significantly influence the internal thermal distributions, and proper outlet position can help alleviate the temperature non-uniformity. Thus, the proposed CFD model serves as a reliable platform for generating comprehensive datasets to support temperature regulation studies in liquid tanks.

Mathematical modelling and simulation are pivotal components in process systems engineering. Focusing on polymerization process systems, identifying microscopic properties of polymers is highly sought after for advancing kinetic comprehension and facilitating industrial applications. Among various computational methods predicting polymeric properties microscopically, kinetic Monte Carlo (kMC) offers a stochastic framework to characterize individual polymer chains and track dynamic system evolution, providing mechanistic insights into complex polymerization kinetics. In this study, an accurately accelerated Superbasin-aided kMC model is developed for enhancing the kinetic understanding of the advanced photo-iniferter RAFT (PI-RAFT) polymerization. The contribution is twofold, presenting advancements in both the mathematical modelling techniques for complex dynamic process systems and the mechanistic understanding of photo-induced polymerizations. Leveraging the increased computational performance and the gained kinetic understanding, batch blending is examined as a viable approach to produce polymers with desired microscopic properties. This work bridges stochastic molecular simulations with scalable polymerization processes, laying groundwork for multiscale modelling in polymer engineering, and also the foundation for process design, control, and optimization.

This study presents the design and evaluation of a plantwide control (PWC) system for Formic acid (FA) production under unsteady green Hydrogen supply. Starting from a steady-state foundation in Aspen Plus V12, the system was prepared to handle variable inputs and was subsequently transitioned into Aspen Dynamics for real-time responsiveness. The two-level design methodology to build a PWC scheme, which is comprised of equipment-specific and plantwide controllers, effectively managed fluctuations in feed rates ranging up to ±20%, maintaining FA purity and production rate targets. Gradual SRAMP (sinusoidal ramp) adjustments of 1% per hour provided optimal stability. These results confirm the PWC system’s effectiveness in maintaining production goals under the variability of throughput.

This work presents a macroscopic, high-level representation of the interconnected nexus system, utilizing a resource allocation model to capture the interactions between the power and water subsystems. The model is employed to assess the system’s performance under various external stressor impact scenarios, determining the thresholds at which the system can no longer maintain a continuous supply of functional services (i.e. power and water), which reveal the system’s vulnerabilities. Resilience metrics are incorporated to interpret these results and characterize the nexus performance. The proposed methodology is generalizable, and its capabilities will be demonstrated through a case study on the energy-water nexus in the Gulf Cooperation Council region.

The increasing need for Direct Air Capture (DAC) technologies is driven by the urgent global need to mitigate rising CO2 levels due to climate change. In humid climates, DAC systems face challenges as high humidity increases the energy required for regeneration. This study introduces a novel DAC system integrated within an Air Handling Unit (AHU) that includes a silica gel wheel for air dehumidification before CO2 capture, significantly enhancing physisorbent performance by optimizing conditions for CO2 adsorption. This system, tailored for the AHU of Doha Tower, involves dehumidifying return air, subsequently cooling it for effective CO2 capture. The introduction of the silica gel wheel notably reduced the energy requirements by 81.5% for NbOFFIVE compared to configurations without dehumidification, and the thermal energy cost for NbOFFIVE when integrated with HVAC and silica gel is 70 USD/ tonCO2, compared to 160 USD/ tonCO2 for SBA-15 + TEPA used alone. Additionally, the thermal energy cost for the HVAC system was reduced from 430 USD/tonCO2 to 303 USD/ tonCO2. A technoeconomic analysis highlights these improvements as crucial for reducing operational costs and cutting capital expenditures by leveraging existing AHU infrastructure. Despite the higher initial cost of NbOFFIVE, the levelized cost for DAC using NbOFFIVE is estimated at 184 USD per ton of CO2, making it economically advantageous compared to 208 USD per ton for SBA-15.

Demographic shifts and increased automation in chemical plants are reducing the experience and skill levels of plant operators. Therefore, BASF has implemented Operator Training Simulators (OTS) to allow operators to practice and improve their skills in this safe and controlled environment. The OTS consists of a dynamic model of the process, a control system and safety logics. This paper describes the learnings from using OTS at BASF, where they are used to train operators in process understanding, optimization, procedural training, and disturbance handling. Benefits include reduced training costs, minimized risks and improved efficiency. Also organizational guidelines are provided to ensure that the mentioned benefits are realized in industrial practice. Additionally, high-accuracy OTS models support HAZOP, debottlenecking, and optimization studies.

The urgent need for sustainable energy drives the exploration of biomass and plastic waste co-gasification, a promising route for producing clean fuels and chemicals, reducing greenhouse gas emissions, and minimizing fossil fuel dependence. Modeling and simulation are vital for optimizing this process, particularly syngas yield, yet comparative studies on Aspen Plus modeling techniques for steam co-gasification are limited. This research addresses this gap by comparing three Aspen Plus strategies: thermodynamic equilibrium modeling (TEM), restricted thermodynamic modeling (RTM), and kinetic modeling (KM), for simulating the co-gasification of pine sawdust and polyethene (PE) with steam in bubbling fluidized bed gasifier (BFBG). The primary objective is to evaluate the effectiveness of each strategy in predicting the syngas composition under varying conditions. Three models were developed in Aspen Plus on the basis of each strategy, and their predicted syngas compositions were compared with published experimental data. A detailed sensitivity analysis was performed within the RTM framework to obtain optimized values of the approach temperatures to enhance predictive accuracy. The results showed that RTM provided the highest precision, with an average root mean square error (RMSE) of 0.0296, while TEM showed the lowest precision with RMSE of 0.1234. KM performed moderately with an RMSE of 0.0929. These findings demonstrate that RTM is superior for predicting the syngas composition for this co-gasification process, and its optimized solution presents a viable alternative when detailed kinetic data are lacking, thus reducing computational expense.

Proton Exchange Membrane (PEM) fuel cells are gaining traction in automotive applications due to their efficiency and environmental benefits, but they face challenges such as high costs, degradation rates, and limited hydrogen availability. To address these issues, novel operational methods have been developed, focusing on customized designs rather than traditional uniform configurations. These advancements include the variable temperature flow field, which maintains high relative humidity without external humidification by leveraging internally generated water and heat, and graded catalyst loading, which enhances current density distribution. Additionally, complex flow fields have been designed using 3D metal printing to mitigate liquid water accumulation. These innovations have shown significant performance improvements, particularly when combined, demonstrating a 260% increase in current density at 0.6 V. These advancements hold promise for overcoming the limitations of conventional PEM fuel cell designs and enhancing their viability for commercial applications.

Cell culture processes play a central role in the production of various therapeutic compounds. These processes are multiscale and highly complex, making them challenging to describe comprehensively using fully mechanistic models. In this study, we employ an integrated hybrid machine learning and first principles model to predict the viable cell density, product titer, and metabolite concentration profiles. We employ the concept of degree of hybridization, where we create a family of hybrid models each with increasing degree of process knowledge. Predictions from the feasible hybrid architecture were integrated with a genome scale metabolic model to evaluate the flux distribution of reactions related to the central carbon metabolism of the cell throughout the process duration. We demonstrate that the current approach not only reasonably predicts the bioprocess profile but also provides biologically relevant information that can uncover dynamics of intracellular metabolism which can open opportunities for new optimization strategies.

Process simulation is essential for optimizing chemical processes, offering a cost-effective alternative to the experimental approach. This study presents a co-current hollow fibre membrane model for CO2 separation, integrated into Aspen HYSYS® using the CAPE-OPEN standard. A one-dimensional boundary value problem (BVP) is solved through the shooting method, ensuring accuracy for complex gas separation processes. The unit is implemented in C++, facilitating interoperability, error handling, and optimization of key performance indicators like energy consumption and separation efficiency. Appropriate output variables are employed in the Aspen HYSYS® environment to enable direct sensitivity analysis and optimization within the process simulator. Results Sensitivity analysis results demonstrate that the co-current hollow fiber membrane unit improves methane recovery compared to a pressure swing water absorption (PSWA) column for biogas upgrading to biomethane. While membrane technology showed a 24% reduction in energy consumption for biomethane production, PSWA exhibits higher efficiency for biomethanol production. This study not only demonstrates the value of CAPE-OPEN standards in implementing custom unit operations but also lays the groundwork for future developments in process simulation using advanced mathematical modelling and optimization techniques.

The co-gasification of biomass and plastic waste offers a promising pathway for sustainable syngas production, necessitating precise prediction of its composition to optimize efficiency. This study compares the performance of Aspen Plus models, including the thermodynamic equilibrium model (TEM) and restricted thermodynamic equilibrium model (RTM), with machine learning (ML) techniques, focusing on the support vector regression (SVR) for syngas prediction during steam and air co-gasification. Aspen Plus simulations provided valuable mechanistic insights, while the ML model demonstrated superior predictive accuracy. The SVR, enhanced by principal component analysis (PCA), significantly improved performance, achieved R² values of 0.879 for H2, 0.856 for CO, 0.859 for CO2, and 0.744 for CH4 on the testing dataset. It also outperformed other models in terms of RMSE, achieving exceptional precision for CH4 (0.0087), CO (0.0193), and H2 (0.0194). In contrast, RTM exhibited moderate accuracy with minimal deviations, while TEM showed the highest RMSE values across all predictions, indicating its limited reliability for real-world applications. This study advances gasification technologies by demonstrating the advantages of ML models, particularly SVR, in predicting syngas composition and supporting the global transition toward sustainable energy systems.

This work presents a comprehensive multiscale model for Solid Oxide Fuel Cells (SOFCs), integrating microscale and macroscale simulations to analyze internal reforming and its impact on overall cell performance. The microscale model [1], [2] captures the intricate mass and charge transport phenomena at the pore scale of porous electrodes, resolving electrochemical reactions at the triple-phase boundaries and modeling chemical reactions at pore spaces. Simultaneously, the macroscale model provides a broader view of the entire cell’s behavior by solving the same transport equations on a coarser computational mesh. The multiscale approach is particularly useful for addressing the challenges posed by simultaneous chemical and electrochemical reactions at the anode, which complicate the modeling of internal reforming. To overcome these challenges, a novel approach is introduced [3], spatially separating the regions of chemical and electrochemical activity in the pore scale domain by taking the electrochemical active layer thickness into consideration. The integrated multiscale model is applied to a complete internal reforming SOFC to explore how electrode morphology, particularly the use of gradient microstructures, influences cell performance.

This paper evaluates the thermodynamic efficiency of a newly synthesized large-scale pre-combustion CO2 capture process using a novel ionic liquid (IL) 1-octyl-2,3-methylimidazolium thiocyanate [OMMIM][SCN] for blue H2 production. In addition, the potential eco-toxicity of the selected IL was assessed using the ADMETlab 2.0 web tool. The results of these analyses were compared to those of an established IL 1-butyl-2,3-dimethylimidazolium bis(trifluoromethyl sulfonyl)imide [BMMIM][TF2N]. The eco-toxicity assessment confirmed that [OMMIM][SCN] is less environmentally toxic than [BMMIM][TF2N]. Thermodynamic analysis of the novel system shows the COOLER unit accounts for the highest energy demand; however, the [OMMIM][SCN] system demonstrates a 7.45% reduction in energy consumption in the COOLER unit compared to [BMMIM][TF2N]. The system experienced the highest exergy losses (irreversibilities) in the COOLER unit for [BMMIM][TF2N] (12982 kW) and in the flash separator unit for [OMMIM][SCN] (8256 kW). The thermodynamic efficiency was analyzed within a specified IL inlet temperature range (25oC – 75oC) and absorption pressure range (7 – 16 bar). The results show that the [OMMIM][SCN] system outperforms [BMMIM][TF2N] with higher IL inlet temperature but decreases at higher absorption pressure, while [BMMIM][TF2N] maintains constant energy efficiency. The exergy efficiency of the [OMMIM][SCN] system shows better performance across the studied conditions but decreases beyond the peak temperature range of 50–60°C. These findings highlight the potential of [OMMIM][SCN] as an excellent alternative solvent for pre-combustion CO2 capture applications.

Accurately predicting biogas production for real-time applications remains a challenge in anaerobic digestion (AD) due to the process’s complexity and dynamic nature. While mechanistic models are essential for understanding and modelling AD processes, however they are highly non-linear and depend on detailed feedstock characterisation and parameter calibration. In contrast, data-driven models do not rely on predefined equations and rather use process data to capture the system’s underlying dynamics. This study compares mechanistic and data-driven models for biogas prediction using lab-scale data. A state estimation framework with a rolling window was used for the mechanistic model, based on biomass and substrate concentrations with Haldane kinetics, achieved an accuracy of (R² = 0.91). A Long Short-Term Memory (LSTM) model with Bayesian Optimisation for hyperparameter optimisation, trained on the same data showed superior performance (R² = 0.93–0.98) and captured temporal dependencies inherent to the AD process. The LSTM model was further applied to industrial data, maintaining high accuracy (R² = 0.95–0.97) and demonstrating its scalability. Its strong predictive capabilities, combined with practicality for real-time applications, make it a promising tool for optimising operations in large-scale AD plants.

In the present paper, we come up with application of machine learning using data for flow visualization as a method for predicting unsteady flow patterns in oscillatory baffled reactors (OBRs). Application of the proper orthogonal decomposition (POD) is investigated for dynamic analysis of spatio-temporal data acquired by particle image velocimetry (PIV) to determine inputs and outputs for neural network model. It has demonstrated that three sets of modes and time-varying mode coefficients extracted by the POD could be useful for dynamic analysis and prediction of time-variant flow patterns in OBR. Also it is shown that decomposition of the time-series data for the mode coefficients by Fourier series expansion was effective for deriving reduced order model.

Predicting the critical micelle concentration (CMC) of surfactants is essential for optimizing their applications in various industries, including pharmaceuticals, detergents, and emulsions. In this study, we investigate the performance of graph-based machine learning models, specifically Graph Convolutional Networks (GCN), Graph Attention Networks (GAT), and a graph-transformer model, PharmHGT, for predicting CMC values. We aim to determine the most effective model for capturing the structural and physicochemical properties of surfactants. Our results provide insights into the relative strengths of each approach, highlighting the potential advantages of transformer-based architectures like PharmHGT in handling molecular graph representations compared to traditional graph neural networks. This comparative study serves as a step towards enhancing the accuracy of CMC predictions, contributing to the efficient design of surfactants for targeted applications.

The escalating urgency to address climate change has driven carbon capture (CC) technologies into the spotlight, particularly for large-scale emitters, which benefit from economies of scale. However, small-scale emitters account for a significant share of CO2 emissions, yet such applications remain largely overlooked in the literature. While CC cost is often used as a key performance indicator (KPI) for CC technologies, the lack of standardized cost estimation methods leads to inconsistencies, complicating comparisons, and hindering the deployment of CC systems. This study addresses these challenges by developing flexible short-cut correlations for selected CC technologies, providing estimates of the total equipment cost (TEC) and energy consumption specific to small-scale applications across various CO2 inlet concentrations (mol%) and capture scales (10 – 100 kt/y). The flexibility of the correlations enables the integration of various cost estimation methods available in the literature and case-specific assumptions (e.g., utility prices), enhancing their consistency and applicability across various scenarios. This approach provides decision-makers with a practical yet simple method for assessing the technical and economic viability of CC systems without conducting extensive simulations or detailed techno-economic assessments (TEAs). Ultimately, this work aims to improve the accessibility of CC technologies for small-scale industries, facilitating their broader application and contributing to the overall goal of reducing greenhouse gas emissions across various sectors.

Integrated Carbon Capture and Conversion (ICCC) technologies offer an efficient alternative to conventional Carbon Capture, Utilization, and Storage (CCUS) methods by simultaneously capturing and converting CO2 into value-added chemicals. Dual-function materials (DFMs) are particularly promising due to their capability to integrate adsorption and catalysis in a single step, thereby reducing both energy consumption and associated costs. This study models the dynamic behavior of CO2 adsorption within a laboratory-scale packed-bed reactor employing DFMs. The mathematical model incorporates momentum, mass, and heat transfer equations, implemented using COMSOL Multiphysics v5.6, and evaluates various axial dispersion models (ADMs) and mass transfer coefficients (MTCs). The results indicate that the Rastegar-Gu ADM, combined with an MTC of 8.3 × 10-2 s-1, provides the most accurate representation of breakthrough and saturation times, as well as the total quantity adsorbed. Furthermore, relative errors associated with this model are consistently lower than those observed in alternative models. These findings offer valuable insights for the scaling-up of ICCC technologies, underscoring the significance of precise Computational Fluid Dynamics (CFD) simulations in advancing sustainable CO2 capture and conversion solutions.

As demand for long and accurate molecular simulations increases so too does the computation demand. Beyond using new, enterprise scale processor developments – such as the ARM neoverse chips – or performing simulations leveraging Graphics Processing Unit compute, there exists a potentially faster and more power efficient option in the form of custom hardware. Using hardware description languages it is possible to transform existing algorithms into custom, high performance hardware layouts. This can lead to faster and more efficient simulations but compromises on the required development time and flexibility. In order to take the greatest advantage of the potential performance gains, the focus should be on transforming the most computationally expensive parts of the algorithms. When performing molecular dynamics simulations in a polar solvent like water, non-bonded electrostatic calculations dominate each simulation step, as the interactions between the solvent and the molecular structure are calculated. However, simply developing a non-bonded electrostatics co-processor may not be enough, as transferring data between the host program and the Field Gate Programable Arrays itself incurs a significant time delay. For any changes to be made, competitive to existing calculation solutions, the number of data transfers must be reduced. This could be achieved by simulating multiple time-steps between memory transfers, which may impact accuracy, or performing more calculations in the custom hardware.

The deployment of CO2 capture technologies at a large scale will largely benefit from the knowledge acquired during pilot testing. A mobile CO2 capture pilot unit is currently being designed at the University of Liège. Here, the pilot plant is introduced, and the column sizing results are presented. The sizing was performed with a process model built in Aspen Plus. Overall, the pilot installation is expected to serve for process model validation, data collection and technology de-risking while assisting Belgian industries in their transition towards carbon neutrality.

Process Systems Engineering (PSE) provides the advanced conceptual framework and software tools to formulate and optimise well-considered integrated solutions that could accelerate the sustainability transition within the industrial sector. The landscape of advanced PSE is poised to undertake a considerable transformation with the rise in popularity of open-source and script-based software platforms with predictive modelling capabilities based on modern mathematical optimization techniques. This paper highlights three leading equation-based platforms—IDAES, Modelica, and GEKKO — that are increasingly utilised for the modelling, simulation, and optimisation of complex systems within the advanced PSE domain, alongside the strengths and limitations of each approach. Following this, we present a framework through which emerging techniques within the domain of Software Engineering could be leveraged to address these limitations, with a vision of improving the accessibility and flexibility of complex modelling tools for industrial partners.

This study focuses on enhancing heat and shaft power integration within existing nitric acid production processes to optimize waste heat recovery and identify opportunities to improve process efficiency. A digital twin of the operational plant is utilized, which features a dual-stage pressure nitric acid production process with a capacity of 50 tons/h of HNO3 (100% equivalent). The authors conducted a simultaneous analysis of the thermal energy potential and the expansion capacity of tail gases to effectively fulfil the primary process’s heating, cooling, and power requirements while increasing steam generation through waste heat recovery, all without compromising plant throughput. The proposed process modifications lead to a 23.8% reduction in cooling water usage and a 35.6% decrease in CO2 equivalent emissions while achieving a 13.1% increase in steam generation. These utility savings culminate in a 10.2% enhancement in plant throughput.

In this research, a chemical looping scheme is combined with dry reforming of natural gas in a novel microwave reactor to produce industrial quantity of methanol. Simulation results show that the chemical looping scheme can produce all the hydrogen required by the methanol reactor as well as a significant portion of the carbon dioxide required for the syngas reactor. A heat exchanger network is developed to substantially reduce the hot and cold utility usage. A technoeconomic analysis indicates a significant positive net present value along with a substantial reduction in carbon dioxide emissions as well as a reduction in energy consumption.

In this research, a novel process is developed that utilizes low density polyethylene from plastic waste to produce ethylene. In this process, waste polyethylene is reacted in a microwave reactor to produce ethylene. A conceptual flowsheet based on this reactor is developed in the ASPEN Plus environment. Heat integration tools are utilized to reduce the hot and cold utilities used in this process. This novel design is compared with the conventional process of making ethylene from ethane via cracking. A technoeconomic analysis is conducted to demonstrate the economic feasibility of this process.

Bipolar membrane electrodialysis are receiving the attention of the research community in the last years because they can help the electrification and the spread of direct air capture systems. In this work, a mathematical model of a bipolar membrane electrodialysis cell for carbon dioxide recovery is carried out in order to find the most significant parameters on efficiency through a global sensitivity analysis. The electrochemical cell can be integrated into an absorption column capturing carbon dioxide from the air. Results show that the most important parameter over all investigated figures of merit (specific energy consumption, costs, carbon dioxide desorption efficiency, potassium transport number, removal ratio of potassium cation and carbon) is the potassium cation concentration in the rich solution feeding the cell. A trade-off between energy efficiency, process speed and operational cost is suggested. Future research should be conducted in order to apply the global sensitivity analysis to the overall integrated system considering the whole carbon cycle between the bipolar membrane electrodialysis and absorption column.

An intensified approach to ?-valerolactone (GVL) production is achieved using a reactive distillation column. Conventional methods require multiple units, leading to high energy consumption, costs, and limited scalability. The proposed technology integrates reaction and separation into a single unit, enhancing process efficiency for biomass-derived chemicals. A multiobjective optimization framework balances economic, environmental, and operational goals, reducing total annual cost (TAC) by 43% and environmental impact (EI99) by 45% compared to conventional processes. Additionally, energy consumption drops by 63%, while GVL production increases by 25%, highlighting the potential of reactive distillation for improved efficiency and sustainability.

Oxidation of arabinose to arabinoic acid is an innovative way to valorize local biomass to a high add value product. Previously done experiments on oxidation of arabinose to arabinoic acid with molecular oxygen were used to determine the optimum reaction conditions, scale-up the process and analyse the techno-economic aspects. These results were utilized to analyse the environmental impact of the scaled-up process during its lifetime using the life cycle assessment (LCA) methodology. SimaPro software combined with the impact assessment method IMPACT 2002+ were applied. The results revealed that heating seems to be the largest contributor to the environmental impact even if the reaction is performed under rather mild conditions (70oC). This highlights the importance of reducing the energy consumption via efficient heat integration.

Steel mill off gas fermentation presents a promising green alternative to petrochemical isopropyl alcohol (isopropanol, IPA) and acetone production while potentially reducing greenhouse gas emissions. A pilot-scale study stated negative global warming potential (GWP) at 85% gas conversion and 90% product selectivity. However, industrial-scale plant design including detailed techno-economic assessment (TEA) and life cycle assessment (LCA) remain undescribed. Therefore, this study modelled a heat-integrated 47.5 kton/ year gas fermentation process to IPA and acetone, based on pilot-scale data. The downstream processing was designed using vacuum distillation and heat-pump integrated (extractive) distillation to purify the 50 gproduct/ L broth with biomass and acetate as byproducts, to obtain 41.8 kton/ year of 99.6 wt. % IPA and 5.64 kton/ year of 99.0 wt. % acetone. Notably, no steam is consumed and 2.6 MWh of electricity is generated by utilising the energy from the steel mill off gas. The estimated unit production cost (UPC; 0.57 $/kgproduct) is significantly below market prices (1.65 $/ kgIPA and 1.45 $/ kgacetone). Moreover, the cradle-to-gate LCA gave GWPs of -1.42 kgCO2-eq/ kgIPA and -1.25 kgCO2-eq/ kgacetone, mainly due to avoided steel mill emissions. Other environmental impacts studied were also lower than for petrochemical production. Only the freshwater use is higher (70 L/ kgIPA and 62 L/ kgacetone) compared to petrochemical production (16 L/ kgIPA and 28 L/ kgacetone). Future research should study the impact of IPA and acetone selectivity and titer on the economic and environmental sustainability of steel mill off gas fermentation to IPA and acetone.

Acrylonitrile is a critical commodity chemical used to produce a variety of industrial polymers, such as carbon fibers, plastics, etc. Currently 90% of the global acrylonitrile production is based on propylene ammoxidation. However, there is no literature reporting the whole process holistically in detail, and which also looks into the energy utilization of the whole process including the reaction heat as well as the energy demands of the downstream separation. This original study provides a rigorous process design of the full process from a holistic viewpoint, covering 7 sections of acrylonitrile production (reaction, acid quenching, absorption-desorption, hydrogen cyanide recovery, acrolein recovery, acrylonitrile-acetonitrile-water separation, acetonitrile recovery sections). In order to further improve the energy efficiency, three energy integration strategies are proposed (1) Energy integrated downstream processing; (2) Systematic heat integration utilizing the heat of reaction; (3) Power generation by process surplus heat. Design 1 employs a direct heat exchanger network, recovering 30.1 MW heat through multiple heat exchanges, saving 86.91% fuel gas, 17.38% low pressure steam and 43.96% cooling utilities. Design 2 and 3 recovers 58.39 MW heat, and 12.20 MW low pressure steam is output as waste heat recovery in design 2, whereas 11.06 MW low pressure steam and 7.8 MW electricity are generated in design 3. The pro/con each options are also discussed for further guidance. As the first comprehensive description of the design of the entire acrylonitrile production process, this work highlights the potential for improved energy efficiency in the acrylonitrile production.

Plastic recycling is prevalently mechanical, which is inefficient at removing contaminants and produces low-grade materials. Solvent-based polymer dissolution and precipitation is emerging as a low-energy alternative to mechanical recycling when tackling highly contaminated plastic waste streams. We present a computer-aided molecular and process design (CAMPD) formulation for the selection of optimal solvents and process temperatures for polymer recycling via a dissolution and precipitation process. A mixed-integer nonlinear programming (MINLP) model is proposed to minimize the energy requirement for the dissolution of commercial low-density polyethylene, a ubiquitous polymer in industrial and municipal plastic waste, while minimizing the solvent viscosity and toxicity through multiobjective optimization. We integrate the SAFT-??Mie group-contribution equation of state in the optimization framework to predict key thermodynamic properties and to ensure that the desired phase behaviour is achieved. A ranked list of solvents and the corresponding process temperatures is obtained using integer cuts. Two polyethylene molecular weights are considered. In each case, we construct a Pareto front to quantify the trade-offs between the energy requirements of the process and the selected solvent properties. Furthermore, we investigate the simultaneous design of the dissolution and cooling precipitation to analyse the effect of varying the precipitation temperature on the energy of dissolution. The current study provides valuable insights into selecting optimal solvents for polyethylene dissolution, advancing the design of more efficient plastic recycling processes.

Biomass conversion to chemical derivatives and essential intermediates is regarded as a long-term strategy for the chemical sector. Among the numerous valuable chemicals obtained from biomass, 5-hydroxymethylfurfural (HMF) is considered an industrially relevant compound due to its capacity to be converted into a variety of value-added chemicals. Compared to conventional catalytic synthesis, bio-catalysis has emerged as a potential greener substitute for HMF conversion to value-added compounds. HMF conversion through bio-catalysis, although more sustainable, seldom leads to the production of a single derivative. Thus, the development of efficient purification and separation processes of several products are crucial to scalability. The downstream process for the novel enzymatic conversion of HMF to high value-added chemicals (i.e., 1-phenylethylamine, 2,5-bis(hydroxymethyl)furan, 1-phenylethylalcohol, and 5-(aminomethyl)-2-furanmethanol) was designed by means of rigorous simulations in Aspen Plus®. The proposed process resulted in product streams with 97 wt% or larger for their respective component, and at least 95% product recovery is attained for each value-added HMF derivative. The global greenhouse gas emissions for the manufacture of a kg of product is 2.26 kgCO2e/ kgproduct, while 0.086 m3wastewater/ kgproduct is generated. The proposed downstream sequence offers an overview of the feasibility for the scalability of the enzymatic process, crucial in developing sustainable bioprocesses.

Engineering can be made simple and more impactful by observing and understanding how organisms in nature solve eminent problems. For example, scientists around the world have observed green plants thriving without organic food inputs using the complex photosynthesis process to kick-start a self-sustaining biochemical food chain. In this study, two biological analogues for advanced water treatment, i.e., visible-light photocatalysis using porphyrin-Bi12O17Cl2 and BiOIO3 compounds and interfacial solar desalination by a by Reduced Graphene Oxide-Black TiO2 (rGO-Black TiO2) were investigated. For the visible-light photocatalytic process for dye decolouration, a porphyrin@Bi12O17Cl2 system was applied to successfully degrade Rhodamine B dye in batch experiments, achieving up to 79% degradation within 240 minutes. These results show that more advances and more efficient engineered systems can be achieved by observing nature and how these systems have survived over billions of years. The rGO-Black TiO2 system achieved increases in temperature for 25oC to 60oC in less than 5 seconds. Based on these observations, the research group from the Water Utilisation Group at the University of Pretoria have studied and developed fundamental processes for the degradation and remediation of unwanted compounds such as disinfection byproducts (DBPs), volatile organic compounds (VOCs) and pharmaceutical products from water.

Climate and environmental problems caused by increasing CO2 concentration in the atmosphere make the direct air capture (DAC) technology having great prospects for development. As the widely used solvent in carbon capture based on chemical absorption processes, MEA still fails to address the issues of high energy consumption and high costs when used in DAC process. In this study, piperazine (PZ) was used as the new solvent for DAC process. The new configuration was simulated in Aspen Plus® V11 and the model was validated through experimental data and model comparison. It is followed by investigation of the potential for energy efficiency and cost reduction. The standard DAC-PZ configuration could reduce the reboiler duty from 10.7 GJ/tCO2 to 8.9 GJ/tCO2 for DAC-MEA process. Economic analysis will be carried out through Aspen Process Economic Analyzer®. Further analysis (e.g. sensitivity analysis for different parameters and optimisation) will be performed to further reduce the energy consumption and cost for the DAC-PZ process with different configurations.

By-product fuel gases from refinery operations are a major heat source in fossil refineries and their availability poses a challenge to the deployment of low-carbon heat sources. This study evaluates the valorization of refinery fuel gases (RFG) into low-carbon methanol via co-processing with residual biogenic gas streams from biomass thermochemical conversion. Results from techno-economic analysis indicate that up to 44 wt.% of biogenic blend is possible without the need for external hydrogen supply, while electricity and heat requirements per tonne of methanol change by -4 % and + 80% respectively. Nevertheless, at the 44 wt.% blend, the estimated methanol cost increases only by 2.4 % (0.43 EUR/kg), while the reduction in methanol carbon intensity is approximately 40 %. This highlights promising benefits that can contribute to the integration of bio-oils producing technologies within fossil refineries.

Triethylene glycol (TEG) or mono-ethylene glycol (MEG) absorption are the commercial technologies for natural gas dehydration processes. Nevertheless, the necessity of regenerating solvents under high temperatures results in environmental footprint and complex operation. Membrane with advantages in small footprint and high feasibility operation in hostile conditions is considered as promising technology for natural gas dehydration processes. In this work, system scale design and mesoscale modelling are synchronously adopted to optimize natural dehydration process design. Aspen HYSYS with ChemBrane extension is used for natural gas dehydration process. Taking pressure ration, membrane area and sweep gas flowrate as decision variables for minimizing specific process cost is optimized through NSGA-II algorithms. The minimum specific cost of < 3.06×10-2 $/m3 natural gas is estimated to achieve the separation requirement of <100 ppm. Then, the module length, and membrane thickness of the hollow fiber membrane design is investigated using Computational fluid dynamics (CFD), which better configures the simulation results and concentration, velocity profile to analyze the simulation results. The system scale engineering design and mesoscale modelling provide an in-depth insight into natural gas dehydration process.

The pharmaceutical and specialty chemical industries are challenged with the requirement of faster time-to-process to meet market demands. Here, Modular Plants made up of predesigned process equipment assemblies (PEAs) are advantageous. Moreover, equipment-based Digital Twins of these modules can further reduce the time-to-process when combined with methods such as Quality by Design (QbD) and model-based design of experiments (MBDoE) to reduce uncertainty. This paper presents a lab scale-based workflow using an equipment-based Digital Twin, which applies QbD and MbDoE methods to identify the Design Space in the lab scale which can be transferred to production scale equipment as part of a Digital Twin based workflow for scale-up in Modular Plants.

Acetic acid is an important bulk chemical and one of the major downstream products of methanol. However, it has received less attention from an environmental sustainability perspective. Here, we evaluate the absolute sustainability of several acetic acid synthesis routes, considering both fossil and renewable feedstocks. More specifically, we studied the business-as-usual (BAU) methanol carbonylation and the novel, low technology readiness level (TRL) methane carboxylation and semi-artificial photosynthesis routes. Using process simulation and life cycle assessment (LCA), our results reveal that the alternative routes have the potential to outperform the fossil BAU in at least 14 out of the 16 evaluated impact categories. However, despite the overall improvements, their performance in SDGs 3, 6, 13, 14 and 15 remains poor in any of the studied scenarios, which could potentially be addressed by hybridizing fossil and renewable feedstocks. All in all, our analysis underscores the importance of absolute sustainability to contextualize the environmental footprint of alternative chemical synthesis pathways, going beyond climate change to embrace a broader sense of categories relative to absolute thresholds.

This work investigated synergies for improved energy efficiency between integrated first- (1G) and second-generation (2G) sugarcane ethanol distillation, an energy-intensive unit operation, especially for stand-alone 2G ethanol. For this investigation, integrated and separated 1G2G distillation simulations were conducted using Aspen Plus v.10 assuming a dilute 2G fermentation beer with titer varying from 5 to 40 g/L. The results were then assessed in heating energy demand savings for distillation, and it was measured the potential of saved bagasse (boiler fuel) for valorization in either electricity or 2G ethanol. A life cycle assessment was also performed for a consequential approach to carbon emission reductions from energy savings. As our main result, distillation integration can maintain the heat demand of a stand-alone 1G mill, regardless of the 2G ethanol beer titer. This means energy savings between 9 and 15% in total ethanol heat demand, and between 46 and 92% in 2G ethanol heat demand. The savings are due to the reflux ratio associated with 2G ethanol distillation becoming close to the host 1G distillation process. However, although the energy savings may result in economic gains, not assessed in this study, by using saved bagasse for either electricity (between 0.14 and 0.23 MJElectricity/kgEthanol) or 2G ethanol production (between 6.8 and 4.1 g2G Ethanol/kgEthanol), the boiler fuel savings do not significantly decrease carbon emissions of ethanol production (less than 1%). Nonetheless, the lower energy demand has the potential to debottleneck the operational cost challenge of distillation for 2G ethanol production without associated trade-offs.

Emerging technologies require sophisticated design and optimization due to their rapid advancement and potential to alter material flows and life cycles. However, their future development remains uncertain due to sociotechnical factors such as regulations, infrastructure, and market dynamics. Multiple technologies are often considered simultaneously, but their interactions and synergies are not systematically evaluated. This study addresses pre-emptive life cycle design in social challenges by integrating emerging technologies into superstructures, which help visualize alternative candidates for design problems. Case studies on battery chemistry, plastics, and regional resource circulation demonstrate this approach. For battery technology, nickel-manganese-cobalt lithium batteries have dominated over lithium iron phosphate alternatives. Superstructures were developed to assess recycling technologies and were refined through communication with managers of Japanese national battery projects. In plastics, system designs integrating biomass-derived and recycled carbon sources were explored, and scenario planning with life cycle assessment (LCA) was shared with stakeholders. Regional resource circulation was studied through technological integration and local stakeholder engagement, leading to demonstration tests. Findings suggest that superstructures and technology assessments enhance understanding of applicable technologies and their trade-offs. Given the critical role of social acceptance in technology implementation, computer-aided process engineering (CAPE) tools should facilitate discussions on socio-technical and socio-economic aspects of process systems engineering.

This study explores an energy transition strategy that leverages reversible solid oxide cells (rSOC), power-to-gas (PtG) conversion, and CO2 management to enhance the efficiency and sustainability of secondary aluminum production. A comparative analysis between conventional and integrated energy scenarios highlights the benefits of multi-technology integration. The results indicate that the integrated system increases total energy demand by 27% due to additional energy conversion steps, but eliminates natural gas consumption, reducing dependency on fossil fuels. Additionally, net CO2 emissions are reduced more than fivefold, demonstrating the potential of carbon capture and utilization strategies. The seasonal storage of synthetic natural gas (SNG) and biogenic CO2 further enhances system flexibility, allowing excess renewable electricity to be converted into storable fuels for winter use. Despite higher capital expenditures, the operational costs of the integrated system are 11% lower than the conventional approach due to improved energy efficiency and CO2 taxation savings. These findings underscore the importance of process integration, energy storage, and renewable electricity utilization in achieving a low-carbon, resilient aluminum industry and urban heating networks.

The growing concern over climate change and rising carbon dioxide (CO2) emissions have spurred the development of strategies to upcycle greenhouse gases. One promising solution is the synthesis of green methanol via catalytic hydrogenation of captured CO2 using renewable hydrogen (H2). This provides a versatile chemical feedstock for fuels and industrial processes while reducing CO2 levels. Recent advancements in CO2 capture technologies achieve purities ranging from 83% to 98% (v/v), enabling a sustainable integration with green hydrogen for methanol production. While research has largely focused on CO2 purities above 96%, such models overlook the variability and lower purities typical of industrial carbon capture streams. Addressing this gap, this study examines the economic impacts of CO2 purity on methanol synthesis. Using Aspen Hysys V14, the hydrogenation process is simulated to assess the effects of varying CO2 purities on operational costs, yield, and profitability, providing a realistic evaluation under actual industrial conditions. The results reveal that even slight changes in CO2 purity markedly influence both costs and profitability, highlighting the importance of studying CO2 capture technologies for methanol production. This study quantifies the relationship between CO2 purity and operational cost, contributing critical insights for process operation. By emphasizing the role of CO2 purity in enhancing both sustainability and economic feasibility, it advances the understanding of green methanol production as a viable strategy for climate change mitigation.

This study presents a gate-to-gate Life Cycle Assessment (LCA) evaluating the sustainability and environmental impacts of utilising CO2 for Enhanced Oil Recovery (EOR) in Dukhan Field. The assessment employs a detailed model that encompasses CO2 capturing, transportation, injection, and oil production processes. Utilising Gabi software, the study assesses CO2 emissions across different stages of the EOR process and evaluates the environmental efficiency using two functional units: ‘1 kg of CO2 captured’ and ‘1 kg of oil produced’. Results indicate that Post-Combustion Capture (PCC) contributes the highest emissions, accounting for 76% of the total Global Warming Potential (GWP), while transportation pipelines and separators contribute only 2% and 4%, respectively. By Year 21, emissions drop by over 98%, with a corresponding GWP reduction from 4.73 billion kgCO2e in Year 1 to 94.97 million kgCO2e. Emission rates for CO2 capture and oil production also decrease significantly, reaching 0.01 kgCO2e per kg of CO2 captured and 0.03 kgCO2e per kg of oil produced. These findings highlight the transformative potential of closed-loop CO2 management, where recycling reduces reliance on external CO2 sources and enhances the sustainability of EOR operations. This study provides actionable insights for optimising CO2-EOR systems, contributing to both economic and environmental sustainability goals.

This study investigates the design of a two-column distillation (TCD) process to separate a dilute ternary Ethyl Acetate (EtAc)-Methanol (MeOH)-water waste solvent into nearly pure components. The separation is complicated by the presence of a homogeneous EtAc-MeOH azeotrope and a heterogeneous EtAc-water azeotrope, creating a distillation boundary that divides the ternary composition space into two distinct regions. To address this, the proposed flowsheet incorporates liquid-liquid phase separation to cross the distillation boundary, enabling feasible separation. Additionally, the pressure sensitivity of the distillation boundary is exploited to reduce the recycle rate, enhancing energy efficiency. The basic TCD flowsheet consists of a decanter, a high-pressure simple column, and a low-pressure divided-wall column (DWC). Heat integration (HI) is achieved using external process-to-process heat exchangers and vapor recompression (VR)-driven reboilers. The resulting energy-efficient HIVR-TCD configuration is more cost-effective and energy-efficient than existing designs in the literature. Specifically, the total annualized cost (TAC) of the proposed HIVR-TCD process is 15.4% lower than the literature best design, with reductions in energy consumption and CO2 emissions of 34.3% and 31.4%, respectively, representing a significant improvement.

This work presents a robust flowsheet design for the recovery and purification of waste solvent streams containing ethyl acetate (EtAc), methanol (MeOH), and water. Separation of this mixture is challenging due to the presence of two azeotropes: a homogeneous EtAc-MeOH azeotrope and a heterogeneous EtAc-water azeotrope. These azeotropes create a distillation boundary that divides the ternary composition space into two distinct regions, making separation via conventional distillation difficult. Additionally, the wide variability in waste solvent compositions requires a versatile design, as flowsheets optimized for dilute mixtures may not be feasible for concentrated ones. The key to this design is using a liquid-liquid extractor (LLX) with recycled water as the solvent, ensuring the mixture remains within the liquid-liquid equilibrium (LLE) split region, which facilitates spontaneous separation across the distillation boundary and promotes energy-efficient separation. The raffinate composition lies near the EtAc-water edge in the EtAc-rich region and is stripped to recover pure EtAc, while the extract phase is processed to yield pure MeOH and water. Converged designs are developed for equimolar, EtAc-rich, MeOH-rich, and water-rich compositions, with the proposed robust structure validated for all four compositions. This study represents a significant advancement in sustainable solvent recovery by eliminating the need for separate designs tailored to specific compositions, offering a unified, robust solution for separating various solvent feed mixtures.

Trichlorosilane (TCS) is a platform chemical used in the manufacture of silicon metals, silicones, and functional silanes. Despite this, very little information is available on the environmental impact (EI) associated with its manufacture. This work addresses this gap by developing estimates for the EI of reagent grade TCS (RG-TCS) based on a combination of process modelling & optimisation and life cycle assessment (LCA). Two production methods are considered: 1) direct chlorination (DC) producing RG-TCS as a main product, and 2) the Siemens process (SP) producing RG-TCS as a co-product. Results of a bi-objective process optimization suggest that the DC approach provides consistently better pareto-optimal (PO) trade-offs between the global warming potential (GWP) of RG-TCS and process profit; predicted GWPs are 3.2 to 3.3 kgCO2-eq/kg for DC-derived RG-TCS and 3.8 to 4.9 kgCO2-eq/kg for SP-derived PO designs. This suggests that processing history is important when considering the EI of RG-TCS. Nonetheless, further discussions of model uncertainty stress that these GWP estimates should be considered as preliminary, with improvements to modelling being required to refine EI predictions of RG-TCS in the future.

As anthropogenic CO2 emissions continue to drive global warming, innovative approaches to repurpose CO2 into valuable products emerge as pivotal solutions to mitigate its environmental impact. CO2 utilization encompasses a range of technologies, including its conversion into fuels, chemicals, and materials, leveraging CO2 as a resource rather than treating it solely as a waste. This shift not only reduces greenhouse gas emissions but also supports the circular economy by integrating industrial processes with carbon capture and storage technologies. Specifically, in the Waste-to-Energy (WtE) context, sodium bicarbonate production can be an attractive solution, considering that it is required in the plant for SOx and acidic gases abatement. In this work, the carbon dioxide utilization to give sodium bicarbonate in a WtE context is analyzed. With reference to an existing waste-to-energy plant in Italy, the potential of this CO2 utilization method is highlighted by means of process simulation. A preliminary feasibility study is presented, based on heat and material balances retrieved from Aspen Plus® V14, properly set for the accurate modeling of the system’s behaviour. Process performances are discussed, in view of the pilot plant engineering and construction.

A simple mathematical model of the co-precipitation of Ni-Mn-Co hydroxides is developed and applied to investigate the effect of pH, initial concentration of ammonia in the solution, concentration of the ammonia feed, nucleation rate constant and exponent, growth rate constant and growth exponent over the model output. The model is shown to produce a correct representation of the precipitation variables, and the general trends obtained for different sets of parameters are found in agreement with results presented elsewhere. A sensitivity analysis is carried out and the sensitivity indices are calculated. It is found that pH, initial concentration of ammonia and growth rate constant are the input parameters with the most relevant effect over the model input.

The transition to a circular economy (CE) requires agents in circular supply chain (SC) networks to take a variety of different initiatives, many of which are dynamic in nature. We use a system dynamics (SD)-based approach to develop a generic framework for dynamic modeling of CE networks and propose a prototypical circular SC network by combining dynamic models for five actors: a manufacturer, consumer, material recovery facility (MRF), recycling facility, and the Earth. We apply this framework to the supply chain for Polyethylene Terephthalate (PET) plastic packaging by considering different scenarios over a 65-year time horizon in the US. We include both “slow-down-the-loop” initiatives (i.e., those that extend product use time through demand reduction or reuse) and “close-the-loop” initiatives (i.e., those that reintroduce product to the supply chain through recycling) by the consumer, as well as sorting and recycling capacity expansion. We find that, given the current recycling infrastructure in the U.S., “slow-down-the-loop” initiatives are more effective than “close-the-loop” initiatives for improving circularity and minimizing environmental impact. However, combining the two initiatives eliminates the need for capacity expansion and leads to the highest circularity in the shortest amount of time.

L-lactic acid (L-LA), a key monomer in biodegradable plastics, is a sustainable alternative that can be derived from LCB. The L-LA production process typically involves various technologies such as fermentation, filtration, and distillation. In the L-LA production process, large amounts of buffers are used to maintain proper pH during fermentation, so conventional buffers (e.g., CaCO3) are often selected because of their low cost. However, these buffers cannot be recycled efficiently, and the potential for alternative buffers remains uncertain. In this work, we aim to develop and evaluate novel processes for sustainable L-LA production using the alternative buffer (i.e., KOH). The processes involve a series of different unit operations such as pretreatment, fermentation, extraction, and electrolysis. An efficient buffer regeneration process using membrane electrolysis is implemented to recycle the buffer with minimal energy input. Then, we evaluated the viability of the proposed processes compared to the conventional process based on minimum selling price (MSP), and global warming potential (GWP). The MSP for L-LA was evaluated to be 0.83 USD /kg L-LA, and the GWP was assessed to be 2.93 kg CO2-eq/kg L-LA. These results represent a 25% reduction in MSP and a 32% reduction in GWP compared to the conventional process. Additionally, a sensitivity analysis was performed to identify the major cost drivers of MSP, such as interest rate, and the price of biomass. The proposed process, as a cost-effective and eco-friendly process, promotes biotechnology practices for sustainable production of L-LA.

Hydrogen (H2), as the promising alternative to fossil fuel-based energy carriers, faces the critical challenge of diversifying its sources and lowering production costs. Biogas, produced from organic waste, offers a renewable and carbon-neutral option for H2 production, but its high CO2 content requires a pre-separation process of CO2 from CH4 or specialized catalysts for use in existing reforming processes. Chemical looping reforming (CLR), as an advanced H2 production process, uses an oxygen carrier (OC) as the oxidant, allowing raw biogas to be used directly in the reforming process. Recently, numerous studies on CLR design and analysis have demonstrated their growing economic feasibility. However, deploying the CLR process in the biogas treatment industry requires further research to analyze its technical, economic, and environmental performance under target capacities and H2 purity. This study proposes biogas-based CLR processes and analyzes the capability of the processes from techno-economic and environmental perspectives: ?) chemical looping partial oxidization (CLPO), ?) chemical looping steam reforming (CLSR), ?) chemical looping water splitting (CLWS), and ?) chemical looping CO2 splitting (CLCS). Evaluation metrics include energy efficiency, exergy efficiency, unit production cost (UPC), and net CO2 equivalent emission (NCE) to compare technical, economic, and environmental performance, respectively. Each process is simulated in Aspen Plus to obtain mass and energy balance data. The heat exchanger network (HEN) is optimized to maximize heat recovery and achieve autothermal conditions. As a result, we comparatively analyze the economic and environmental capability of the proposed processes by identifying the major cost drivers and CO2 emission contributors.

Electrified steam methane reforming has emerged as a promising technology for electrifying the hydrogen production process industries. Unlike conventional fossil fuel-based steam methane reforming, the electrified steam methane reforming process relies exclusively on electrical heating, eliminating the need for fossil fuel combustion. Beyond that, however, significant amounts of electricity required for the electrified process should be imported from the renewable energy-based system rather than fossil fuel-based grid electricity to have an environmental advantage over the conventional process. This study suggests a framework for integrating renewable energy systems into the electrified process for decarbonized hydrogen production. Considering the variability of renewable energy, wind and solar power are supplemented by battery storage, to facilitate a stable electricity supply to the electrified hydrogen production process. A Mixed-Integer Linear Programming (MILP) model is developed to optimally size and operate both the renewable system and potential grid imports. Case studies under various carbon tax scenarios, using historical weather data from a region in Germany, are conducted, followed by a techno-economic assessment to estimate the Cost of Hydrogen (COH). The results show that higher carbon taxes and reduced capital costs for wind, solar, and storage technologies significantly increase the share of renewable-based electricity. These findings highlight the importance of more stringent carbon taxation and improvements in the technology readiness level (TRL) of renewable energy are critical for accelerating large-scale, clean hydrogen production and industrial decarbonization.

As greenhouse gas emissions continue to increase worldwide, the growing energy demand must be met using low-carbon technologies. Renewable energy and carbon capture and storage are the two important technologies that can mitigate CO2 emissions. The two technologies have been primarily developed independently. However, their hybridization can offer complementary benefits and lower the costs of greenhouse gas abatement. Accordingly, in this article, we develop a novel carbon-neutral process that combines concentrated solar power (CSP) and fuel-based combustor with redox-based thermochemical energy storage (TCES) materials. The TCES materials are used for energy storage and as a source of oxygen (O2) for combusting fuel. We optimize the process’ economic performance considering variability in solar irradiance by developing a two-stage stochastic programming model. We illustrate that compared to the CSP-TCES process employing the Mn2O3/Mn3O4 TCES system, the proposed hybrid process has a 25% lower levelized cost of electricity (LCOE) for an 80% capacity factor. We also compare the economic performance of our process with another hybrid configuration containing CSP-TCES, boiler for fuel combustion, and air separation unit (ASU) to produce O2, which is then used to combust fuel. We conclude that the LCOE of the proposed process is 11.5% lower than the LCOE of the process with an ASU.

Biomass gasification is a promising technology for sustainable energy production. To date, extensive research has been conducted on biomass gasification, particularly focusing on the reaction models of the process. However, existing models are too complex to apply to the control system or to optimize the process operating conditions effectively, limiting their practical use for industrial applications. To address this, a simple reaction model for biomass gasification was developed and validated. A steady state simulation of the biomass gasification process is conducted to analyze gasifier behavior and provide insights into reaction dynamic. The findings in this study align well with existing literature, confirming the reliability of the approach. This simulation serves as a foundation for further study in process control and optimization. Future work will include experimental validation to enhance model accuracy and applicability.

In this contribution, a thermodynamic model-based approach for the optimal design of a solid-state hydrogen storage and release system utilizing the reversible iron oxide/iron thermochemical redox mechanism is presented. Existing storage processes using this mechanism face significant limitations, including low hydrogen conversion, high energy input requirements, limited storage density, and slow charging/discharging kinetics. To address these challenges, a custom thermodynamic model using NIST thermochemistry data is developed, enabling an in-depth analysis of redox reaction equilibria under different conditions. Unlike previous studies, this approach integrates a multi-objective optimization framework that explicitly balances competing objectives: maximizing hydrogen yield while minimizing thermal energy demand. By systematically identifying optimal trade-offs, the study provides new insights into improving process efficiency and reactor design for thermochemical hydrogen storage. These findings contribute to advancing energy-efficient and scalable hydrogen storage technologies.

This study investigates the development of a probabilistic design space (DS) for a tableting process, focusing on the uncertainty in critical model parameters. A an empirical model is used to assess the impact of critical process parameters (CPPs), including lubrication extent and porosity, on tablet tensile strength (CQA). By incorporating Monte Carlo and Bayesian techniques, the uncertainty of five model parameters is propagated, allowing the estimation of feasibility probabilities for achieving CQAs with a reliability greater than 0.95. The resulting probabilistic DS provides manufacturers with a tool to assess the likelihood of meeting CQAs under varying production conditions. The findings indicate that specific combinations of lubrication rate and porosity define a robust DS within the acceptable operating region, ensuring consistent tableting performance even in the presence of uncertainties. This approach emphasizes the importance of probabilistic DS in optimizing manufacturing processes and delivering built-in quality assurance.

Multi-component batch distillation, wherein multi-component mixtures are separated using a single column, is a crucial separation technique in the chemical industry. Traditionally, the components are separated in the descending order of volatility (direct sequence). Similar to continuous distillation, a specific separation sequence can optimize batch distillation. This work aims to generate such optimal sequence for a batch distillation in a computationally efficient manner. Specifically, the proposed approach extends the marginal vapor rate method, which is used for sequencing continuous distillation to multi-cut batch separation. The approach addresses challenges arising due to dynamic nature of batch distillation. The proposed methodology is validated using simulation case studies.

Traditional steel processes are energy-intensive and rely heavily on fossil fuels, contributing to significant greenhouse gas emissions. By adopting electrification technologies, such as electric boilers and compressors, particularly when powered by renewable energy, steel plants can reduce their carbon footprint, enhance process flexibility, and lower long-term operational costs. This transition also aligns with increasing regulatory pressures and market demand for greener practices, positioning companies for a more competitive and sustainable future. This work investigates the potential of replacing conventional steam crackers in a steel plant that relies on the use of fossil fuels, with electrically driven heating systems powered by renewable energy sources. The overall aim was to significantly lower greenhouse gas emissions by integrating electric furnaces and heat pumps into the steel production process. This study evaluates the potential carbon savings from the integration of solar energy in a steel plant with a production capacity of 300,000 tons per month. The solar field required for this integration was found to span an area ranging from 122,142 to 362,360 m², while the wind field area required is estimated at 550.373 m2. By incorporating solar power into the plant’s energy mix, the analysis reveals a significant reduction in carbon emissions, with an estimated saving of 707,597 tons of CO2 per year.

This work aims to develop an energy storage system that allows fluctuating energy inputs (i.e. from process sections driven by renewable sources) to power two process units that are operated continuously at different temperatures. The system consists of two vessels storing diathermal mediums: one for the hotter- and the other for the colder-energy fluxes. The investigated solutions include sensible-heat-, latent-heat-, and thermochemical-TES (thermal energy storage). Organic Rankine cycles (ORCs) with lithium-ion batteries and thermoelectric generators were also assessed. Indeed, all these technologies allow the exploitation of low-temperature thermal energy to supply the high-temperature unit during periods of energy scarcity. Both vessels aim for total self-sufficiency; however, the option to rely on external utilities has been included to meet the energy demand of both units when sufficient process-side power is unavailable. Two energy profiles were investigated to assess the proposed storage systems’ performance: one showing only solar energy inputs and the other showing only wind energy inputs. Finally, an optimization problem was formulated to estimate the optimal size of both storage vessels. Indeed, increasing their capacities would lead to higher capital expenses (CAPEX) and lower operating expenses (OPEX). As a result, the assessed process-integrated configurations (i.e. those featuring any energy storage system) proved competitive with the non-integrated solution (i.e. the one fully relying on external sourcing of electricity) only if referring to high electric energy prices (300 USD/MWh). Conversely, when considering an electricity price of 100 USD/MWh, all the assessed process-integrated configurations proved more expensive than the non-integrated solution.

This study presents a comparative analysis of various configurations for sustainable olefins production via chemical recycling of plastic/biomass wastes, integrating CO2 capture, storage and management technologies. The co-gasification, methanol synthesis and methanol-to-olefins process models were developed on the Aspen Plus® software. Optimization of processing conditions is achieved through the OSMOSE Lua platform, for minimizing the total cost of operation while accounting for seasonal variability in the electricity prices. CO2 valorization processes have been shown to increase carbon efficiency from 55% up to 97% compared to steam naphtha cracking, making chemical recycling of plastics an appealing alternative. In addition, direct CO2 emissions can be fully eliminated, resulting in up to 70% lower net CO2 emissions even when fossil-based plastic waste is used as feedstock. Seasonal CO2 storage can extend the economic benefits by acting as a buffer against high electricity costs and serving as a feedstock for CO2 valorization processes when excess renewable electricity is available. The benefits of the proposed configurations are expected to become increasingly significant with the decarbonization of the electricity generation, relying more on renewable resources.

In this study, we evaluate the economic and environmental impacts of integrating waste polyethylene (PE) pyrolysis with naphtha-based steam cracking for 660 Mt/y ethylene production. We compare six integration scenarios to both business-as-usual (BAU) steam cracking and greenfield waste PE pyrolysis plant. We perform process simulations and equipment design in Aspen Plus® V12, followed by a techno-economic analysis (TEA) and a life-cycle assessment (LCA. The integration capacity we considered corresponds to one full-capacity PE pyrolysis furnace, reducing naphtha feed by 7% in BAU steam cracking. Through the TEA, we identify the most cost-effective scenario by merging the PE pyrolysis gas with the steam cracker furnace outlet after preheating the PE feed. This integration reduces production costs by 6.46MM€/y, improving costs a 0.3% compared to BAU and 30% compared to the pyrolysis greenfield design. LCA results show that the greenfield pyrolysis plant achieves the lowest global warming potential (GWP), reducing emissions by 14% compared to BAU. Among the integration scenarios, lowest GWP occurs when PE pyrolysis gases, also with feed pre-heating, merge before the main compressor train, obtaining a 0.12% GWP reduction compared to BAU. This analysis highlights the potential of introducing circular technologies in existing plants to progressively shift from fossil fuel usage.

The Life Cycle Assessment (LCA) of a solid oxide electrolyser (SOE) has been performed using publicly available data. The system for producing 1 kg of hydrogen at 25bar and 99.9% purity is represented by a modular structure, which includes the 20-kW solid oxide stack manufacturing, balance of plant equipment, operation consumables, and end-of-life processes. A parametrized life cycle inventory modeling approach was developed. The results illustrate that SOE performs better than steam methane reforming only if supplied by electricity from renewable or nuclear sources. The operation consumables have been identified as the most contributive life stage (67%-89% of potential impacts), followed by equipment manufacturing (7%-22%) and stack manufacturing (4%-11%). Considering the predominant contribution of electricity supply in the consumables, no compromise should be made on ensuring clean electricity sourcing and on the stack energy conversion efficiency. The lifetime of the stack and the balance of plant equipment, as well as the heat mix have been identified as sensitive parameters to minimize the environmental impact of the hydrogen technology. These LCA results have been used to produce a tailored eco-design process for hydrogen projects: (i) organization of an eco-design workshop to present LCA results & environmental hotspots and related key parameters where to leverage eco-design innovations through an open discussion (brainstorm); (ii) provide an eco-design tool obtained from a simplified version of the LCA model parametrized around a limited number key life cycle inventory parameters, enabling the designers/developers to independently test the environmental performance of their innovations.

A generally applicable framework for the evaluation of the sustainability of distillation processes is proposed by aligning indicators directly to selected sustainable development goals (SDGs) created by the United Nations. The indicators are related to the goals good health and well-being (SDG 3), clear water and sanitation (SDG 6), affordable and clean energy (SDG 7), decent work and economic growth (SDG 8), industry, innovation and infrastructure (SDG 9), responsible consumption and production (SDG 12), climate action (SDG 13) and life below water (SDG 14). A total of 12 sustainability indicators, including human toxicity potential, wastewater generation, water consumption, renewable energy share, energy demand, material footprint, profit, waste generation, recycling ratio of waste, greenhouse gas emission, eutrophication potential and acidification potential are assigned to selected SDGs. The application of the indicators is illustrated by two case studies: a batch (BD) and a continuous (pressure-swing) distillation process. Case BD6, where the CO2 emission was minimized, proves to be the most sustainable BD process. It has the lowest greenhouse gas emission, acidification potential and waste generation. Although it has the lowest profit, the variations across the BD cases are minimal. By the pressure-swing distillation, L4 (feeding into low-pressure column, heat integration with optimization) is the best case, with the highest profit and the lowest energy demand, material footprint and greenhouse gas emission values.

Direct Air Capture (DAC) technology has gained significant attention as a promising solution for mitigating CO2 emissions and meeting climate goals. However, the current challenges of high energy demand, capital costs, and scalability present critical challenges to the widespread deployment of DAC systems. Integrating DAC with Heating, Ventilation, and Air Conditioning (HVAC) systems in buildings offers a potential solution by enhancing indoor air quality while capturing CO2, thus lowering energy consumption and capital investment compared to standalone DAC systems. This study evaluates the techno-economic performance of an integrated DAC-HVAC system against a standalone DAC system. This analysis combines thermodynamic estimation of CO2 and H2O loadings and energy requirements with an economic evaluation of capital and operating costs to calculate the levelized cost of CO2 capture (LCOD) for both DAC-HVAC and DAC-standalone. A sensitivity analysis explores the effects of varying climatic conditions, specifically temperature and humidity, on the economic performance of the DAC-standalone system. Further sensitivity analysis examines the impact of discount rates and electricity prices on both systems’ economic viability. Results show that the DAC-HVAC integration achieves lower energy consumption and capital costs, with an LCOD of approximately 219 $/ton CO2, compared to the DAC-standalone system, which ranges from 349–429 $/ton CO2 depending on climatic conditions. In addition to offering a lower levelized cost, the DAC-HVAC system also provides resilience against climate and cost fluctuations. This study demonstrates the significant economic benefits and scalability potential of integrating DAC technologies within existing building infrastructures.

The rising levels of carbon dioxide (CO2) in the atmosphere significantly contribute to climate change, highlighting the need for effective CO2 mitigation strategies. While capturing and storing CO2 is important, converting it into useful products offers additional environmental and economic benefits. One promising method is the reverse water gas shift (RWGS) reaction, which transforms CO2 into carbon monoxide (CO). Membrane reactors (MR), which integrate selective membranes with equilibrium limited chemical reactions, have the potential to intensify processes based on the RWGS reaction. In such reactors, by-products like water are removed in-situ from the reaction zone, effectively shifting the reaction equilibrium to favor higher CO2 conversion. This study develops a comprehensive multi-scale mathematical model for RWGS membrane reactors. We integrate the microscale permeance model (for LTA-4A membrane) with the RWGS MR unit scale and the system’s scale models. The effectiveness of applying sweep gas has been suggested, and the optimal H2 perm-selectivity is found in the range from 30 to 50. A detailed membrane permeance model for LTA-4A membranes is integrated in the MR model to assess how species permeance changes with varying reactor conditions. We propose a sweep recycle configuration and compare it with the conventional sweeping configuration commonly assumed in most literature. A thermal efficiency analysis is conducted to evaluate the proposed recycling configuration, providing insights into the practical feasibility of membrane reactors for CO2 utilization.

This work develops process options using a novel protonic membrane reformer (PMR) and liquefaction-based CO2 capture process for low-carbon hydrogen production from natural gas. Several hybrid concepts of the PMR and liquefaction process are suggested based on the strategies to handle the residual gas from the reformer. The process intensification and optimization results indicate that the hybrid system with a water-gas-shift reactor and off-gas recycling guarantees high H2 and CO2 recovery rates for the PMR operating at relatively low H2 recovery. The hybrid concept also has 74% energy conversion efficiency, which is higher than a conventional steam-methane reforming (SMR)-based H2 production with chemical absorption CO2 capture.

Despite the potential of electrification in transportation, diesel will remain one of the main fuels for decades. The replacement of diesel with biodiesel is one of the solutions to decrease the net emissions of diesel engines. However, biodiesel has limited performance in cold weather and requires fuel additives. In this context, choosing additives from non-edible, inexpensive, renewable sources is important. Ethyl levulinate, an ester derived from levulinic acid that can be produced from sugarcane, is a promising option because it improves the cold-flow properties of fuels and reduces soot emissions. In this work, a reactive distillation column was designed to produce ethyl levulinate. Because of the volatility order of the components involved in this reaction, levulinic acid was chosen as the excess reactant. Production cost was calculated based on ethanol price, capital cost, and operating expenses for several scenarios. The results showed that the optimized reactive distillation column using excess levulinic acid provides a production cost 29% lower than that of a similar esterification process with a continuous stirred tank reactor followed by distillation.

In this work, a Claus reaction furnace was analyzed in a sulfur recovery unit (SRU) of the Abadan Oil Refinery where the combustion operating temperature is important since it ensures optimal performance in the reactor, this study focused on temperature of control of 1400, 1500 and 1600 K and excess air of 10, 20 and 30% to improve the reaction yield and H2S conversion. The CFD simulation was carried out in Ansys Fluent in transitory state and in 3 dimensions, considering turbulence model ? -e standard, energy model with transport by convention and mass transport with chemical reaction using the Arrhenius Finite – rate/Eddy dissipation model for a Kinetic model of destruction of acid gases H2S and CO2, obtaining a good approximation with experimental results of industrial process of the Abadan Oil Refinery, Iran. The percentage difference between experimental and simulated results varies between 0.5 to 5 % depending on species. The temperature of 1600 K and with excess air of 30% was the best, with one fraction mol of 0.054 of S2 at the outlet and with conversion of the acid gas (H2S) of 97.64%, which is quite good compared to the experimental one.

The integration of carbon capture and storage (CCS) into coal and biomass co-firing systems (CBCCS) offers a promising solution for reducing carbon emissions in electricity generation. This study evaluates hypothetical scenarios in Malaysia and Indonesia, focusing on techno-economic-environmental transparency. The analysis shows a negligible change in plant net efficiency (~1%) across biomass co-firing ratios of 5-20% in both countries. The capture penalty increases at higher biomass ratios, particularly at 20% co-firing, due to higher auxiliary power demands and steam extraction. As biomass share increases, net CO2 emissions decrease by an average of 43% in Malaysia and 34% in Indonesia. Economic evaluations show a positive revenue increase for Malaysia at a 20% co-firing ratio, while Indonesia faces a revenue deficit (0.6%) under the same condition, mainly due to an unattractive carbon price and feed-in tariff from 2027 onward. Malaysia faces a higher risk of stranded assets due to earlier retirement plans, compared to Indonesia. As the co-firing ratio increases, the risk of stranded assets decreases by 3% in Malaysia and 0.6% in Indonesia. The present work showcases the intricate and synergistic effects of energy and commodity market dynamics from a micro-sector perspective towards plant revenue. This allows both Indonesia and Malaysia to learn from each other in their quest to achieve energy independence and security. Collaboration in tapping renewable sources is one way the two countries can secure energy for their domestic needs while addressing the challenges of carbon lock-in and the impact of stranded assets.

The development of Carbon Capture, Transport, and Storage (CCTS) and hydrogen pipeline networks is crucial for achieving deep decarbonisation in industrial sectors. However, existing network design models often assume perfect foresight, limiting their applicability to real-world infrastructure planning, which is inherently uncertain and iterative. This study introduces a novel rolling-horizon methodology for pipeline network expansion, leveraging a genetic algorithm-based approach that allows for adaptive routing and incremental infrastructure development. By comparing rolling-horizon designs to 2050-optimised networks in a case study of the Humber region in the UK, the analysis highlights the trade-offs between adaptability and cost efficiency. Results indicate that while rolling-horizon approaches better reflect real-world decision-making, they also introduce inefficiencies, increasing capital expenditures by approximately 8% for both hydrogen and CCTS infrastructure. Additionally, the study examines the economic risks associated with transitioning from blue to green hydrogen production, revealing a 20-fold increase in CCTS costs per tonne of CO2 when hydrogen production shifts to green. The findings underscore the importance of integrating uncertainty-aware strategies in infrastructure planning to mitigate inefficiencies while maintaining adaptability to technological and policy changes.

This study introduces a mathematical programming approach to optimize biomass-to-hydrocarbon supply chain design and planning, aiming to balance economic and environmental outcomes. The model incorporates a range of residual biomass types from forestry, sawmills, and the pulp and paper industry, with the option to establish various processing facilities and technologies over a multi-period planning horizon. The analysis involves selecting forest areas, identifying biomass sources, and determining the optimal locations, technologies, and capacities for facilities converting wood-based residues into methanol and pyrolysis oil, which can be further refined into biodiesel and drop-in fuels. Using Life Cycle Assessment (LCA) in a gate-to-gate analysis, forest supply chain carbon emissions are estimated and integrated into the optimization model, extending previous research. A multi-objective framework is employed to minimize CO2-equivalent emissions while minimizing present costs, with efficient Pareto points evaluated in a case study focused on the Argentine forest industry.

Traditional scheduling approaches often rely on simplified process representations to reduce computational complexity, failing to capture the real-world dynamics where tasks often overlap, and their timing depends on finer operational steps. To address these limitations, this paper proposes a novel process representation that breaks down production tasks into smaller, more primitive steps called operations. Unlike traditional methods, this approach provides a more granular depiction of task timing and resource dependencies. Operations can define the start or end of other tasks, utilize shared resources, and incorporate recipe constraints that mandate task sequencing. The proposed representation is utilized to develop two MILP models to address the makespan and the cycle time minimization problems. Finally, the efficiency and practical applicability of the developed models are showcased with a help of a case study from the pharmaceutical industry.

This study focuses on optimizing production scheduling in multi-product plants with shared resources and costly changeover operations. Specifically, two main challenges are addressed, the unknown changeover behavior of new products and the need for rapid schedule generation after unforeseen events. An innovative framework integrating Machine Learning (ML) techniques with Mixed-Integer Linear Programming (MILP) is proposed for single-stage production processes. Initially, a regression model predicts unknown changeover times based on key product attributes. Then, a representation where distances correlate with changeover times is compiled through multidimensional scaling, allowing constrained clustering to group production orders according to available packing lines. Ultimately, the MILP model generates the production schedule within a constrained solution space, utilizing optimal product-to-line allocation from cluster segmentation. A case study inspired by a Greek construction materials plant is used to validate the proposed approach. The results showed that this framework improves scheduling efficiency by providing rapid solutions, reducing downtime and facilitating the introduction of new products. Overall, a novel scheduling solution is proposed for manufacturing industries that face unknown production data and need quick schedule alternations.

The growing demand for sustainable energy has driven research into renewable methane production to reduce greenhouse gas emissions and reliance on fossil fuels. Promising feedstocks include lignocellulosic dry residues, wet waste, and captured CO2, converted via gasification, anaerobic digestion, and synthetic processes with renewable hydrogen. This study uses a multiscale approach to compare these sources, incorporating a techno-economic evaluation to identify key performance indicators (KPI) for facilities and renewable energy sources. A facility location pro- blem (FLP) determines plant locations and production capacities, considering material availability and transportation costs. The analysis focuses on the decentralised use of wastes and CO2 from point and diluted sources across Spain, employing an MILP model to optimise waste and CO2 utilisation alongside solar and wind energy systems. Results highlight lignocellulosic dry waste and CO2 captured with MEA from point sources as the most promising options. Sensitivity analysis predicts methane prices between 13.028 €/MWh and 47.216 €/MWh through 2050, requiring substantial investment for full methane self-sufficiency. With carbon taxes, the price could drop to 10.735 €/MWh by 2050, competitive with current natural gas prices.

Data-driven decision-making is crucial in the transition to a low-carbon economy, especially as global industries strive to meet stringent sustainability goals. Traditional life cycle assessments often rely on static emission factors, overlooking the dynamic nature of the energy grid. As renewable energy penetration increases, grid carbon intensity fluctuates significantly across time and regions, due to the inherent intermittency of renewable sources like wind and solar. This variability introduces discrepancies in emission estimations if time-averaged factors are applied, leading to sub-optimal process operations and unintended environmental consequences. To this end, we present a real-time emission-aware optimization framework, which is implemented through a mixed-integer linear programming formulation that can determine optimal design configurations and operation schedules while simultaneously mitigating emissions by utilizing electricity price forecasts, time-varying emission factors, sporadic weather data, and supply and demand variability. The optimization model integrates life cycle assessment criteria to evaluate various environmental inputs, categorized into direct on-site emissions (Scope 1), indirect emissions from energy use (Scope 2), and upstream, downstream, and construction-related emissions associated with the process system (Scope 3). The framework, demonstrated through a detailed hydrogen production case study, showcases the effect of time-varying emissions by generating a set of optimal solutions that balance the trade-offs between environmentally sustainable and economical operational strategies.

Green hydrogen will play a key role in the decarbonization of the steel sector via the direct reduction path [1]. To meet the demand side, both a highly efficient numbering-up based scaling strategy for water electrolysis is needed as well as flexible operation strategies that follow the fluctuating electricity load. This paper presents a modularization approach for electrolysis systems that addresses both aspects by combining different electrolysis technologies into one heterogeneous electrolysis system. We present a modular design of such a heterogeneous electrolysis system that can be scaled for large-scale applications. The impact of different degrees of technological and production capacity-related heterogeneity is investigated using system co-simulation to find an optimal solution for the goal-conflict, that the direct reduction of iron for green steel production requires a constant stream of hydrogen while the renewable electricity profile is fluctuating. For this use-case the distribution of technology and production capacity in the heterogeneous plant layout is optimized regarding overall system efficiency and the ability to follow flexible electricity profiles. The simulation results are compared with conventional homogenous electrolyser plant layouts. First results underline the benefits of combining different technologies and production capacities of individual systems in a large-scale heterogeneous electrolyser plant. It is shown that the presented modular electrolysis systems can follow the current flow without losing significant efficiency. However, for a downstream system that requires a constant hydrogen mass flow, supplementary hydrogen storage is required.

The growing size and complexity of energy system optimization models, driven by high-resolution spatial data, pose significant computational challenges. This study introduces methods to reduce model’s size and improve computational efficiency while preserving solution accuracy. First, a composite-curve-based approach is proposed to aggregate granular data into larger resolutions without averaging out specific properties. Second, a general clustering method groups geographically proximate fields, replacing multiple transportation arcs with a single arc to reduce transportation-related variables. Lastly, a two-step algorithm that decomposes the supply chain design problems into two smaller, more manageable subproblems is introduced. These methods are applied to a case study of switchgrass-to-biofuels network design in eight U.S. Midwest states, demonstrating their effectiveness with realistic and detailed spatial data.

The ongoing energy transition involves decarbonization across different sectors. Amongst these, the transportation sector contributes significantly owing to its reliance on traditional fossil fuels as feedstock. Attaining decarbonization goals requires the adoption of novel sustainable technologies such as electric vehicles (EVs), and hydrogen fuel cell vehicles (HFCVs), amongst others. The feedstock transition towards electricity and dense energy carriers is challenged by the requirement for additional infrastructure to manage intermittency, power generation, and grid expansion which requires both materials and capital investment. By evaluating and redirecting the role of carbon value vector from fossil fuel production towards the production of carbon-based materials such as polymers to empower the energy transition, we can optimize resource allocation and maintain economic viability, all while reducing environmental impact. In this work, we propose an integrated framework to systematically address energy-materials-mobility transition nexus challenges. The proposed multiscale framework utilizes a resource-task-network (RTN) representation and a life cycle assessment (LCA) step and considers future material demand and production capacities. Its capabilities are demonstrated through a case study on the transition from gasoline-fueled vehicles to EVs, analyzing (i) the role of carbon value vectors in resources and materials production, and (ii) electricity generation, storage, and dispatch using intermittent renewables. The study reveals the interactions between energy, material, and mobility value chains while providing configurations for exploiting such synergies.

The turbulence in energy markets poses risks to energy suppliers, impacting profitability. Whilst risk mitigation is crucial for new projects, adapting existing infrastructure to evolving conditions incurs additional costs. For natural gas dependent economies, the natural gas industry faces exogenous uncertainties represented by demand and price fluctuations, and endogenous risks arising from inadequate proactive planning. This study evaluates the resilience of optimised Qatar’s natural gas monetisation infrastructure under different cases by examining the network’s ability to meet production targets amid process disruptions and market volatility. The analysed network includes 6 direct and indirect utilisation routes, represented by liquefaction, Haber-Bosch, methanol, gas-to-liquids, MTBE and urea processes to produce 9 products. First, process simulations and market assessments were used to obtain operational and market input data. Second, a mixed-integer linear programming model was used to optimise the natural gas allocation to processes under 9 case studies for a planning horizon of 10 year (2024-2033) with an objective of profitability maximisation. Third, the resilience% of processes and the overall network was estimated as per the model’s suggested allocation strategy to evaluate the infrastructure utilisation efficiency. Process resilience below 100% indicates vulnerability to disruptions. Incorporating operational and market data increased network resilience from 74% to at least 82%, highlighting the value of flexibility in planning. Furthermore, allowing dynamic natural gas allocation improved profitability by at least 27% compared to a baseline scenario lacking operational and market considerations. The findings highlight the importance of adaptive strategies in ensuring long-term economic sustainability.

A dual-stack fuel cell model was developed to simulate the hydrogen consumption a fuel cell-powered vehicle for a specific drive cycle. Two fuel cell stacks, each consisting of 65 parallel cells at different aging status and thus with different efficiency profiles (i.e., low and high) were considered. A constrained optimization for power distribution between individual stacks was performed where the objective function was to minimize the hydrogen consumption while meeting the total demand. For proper power management each stack has its own power controller which manipulates the stack current to control the stack power at its desired-set point. Computed optimal power values constitute the desired set-points for the local power PID controllers of the individual stacks. Closed-loop simulations were performed by simulating the developed mechanistic model together with optimization and PID controllers in SIMULINK platform. The closed loop simulations demonstrate how well the power demand of the drive cycle is tracked and the hydrogen consumption is minimized. To assess the impact of optimization strategy used, hydrogen consumption is compared to that of equal power sharing and a daisy-chain strategies. The results demonstrate that the optimization strategy effectively reduces the total hydrogen consumption.

With increasing emphasis on energy efficiency, more researchers are focusing on energy-efficient flexible job shop scheduling problems. Mathematical programming is a commonly used optimization method for such scheduling challenges, offering the advantages of achieving global optima and serving as a foundation for other approaches. However, current mathematical programming formulations face several challenges, including insufficient consideration of various forms of energy consumption and low efficiency, particularly in handling large-scale instances, which struggle to converge. In this study, we propose a novel global sequence-based approach with high computational efficiency. In this model, immediate precedence relationships are identified using constraints, enabling the precise determination of idle durations within any idle slots. The proposed formulation achieves a significant reduction in energy consumption by up to 20% relative to other formulations. Furthermore, it successfully reaches feasible solutions in challenging cases.

Transition from fossil fuels to clean energy technologies (CETs) is critical, but material shortages threaten to hinder progress. This study analyzes the potential deficits in 14 key materials – such as lithium, nickel, and cadmium – based on capacity projections for CETs by eight Integrated Assessment Models (IAMs) for 2020-2050. It focuses on technologies including battery storage, concentrated solar power (CSP), electrolyzers, solar photovoltaics (PV), and wind turbines. Our findings show that these materials could face shortages of up to 97% by 2050. To meet rising demand, material production rates must increase sharply, with some materials like cadmium, selenium, and tellurium requiring about 31% increases, peaking in this decade. Immediate actions are needed to accelerate production and improve recycling efforts. However, recycling targets, such as 325% for lithium, seem highly challenging to achieve. Without these measures, material shortages could delay CET deployment, risking Paris Agreement goals. To prevent these disruptions, we call for: (i) revising CET capacity projections, (ii) urgent planning for mining expansion, and (iii) a significant increase in recycling infrastructure to ensure a sustainable energy transition.

Commodity chemicals and energy supply chains are an essential part of the hydrocarbon industry in several countries. As these supply chains are susceptible to disruptions caused by various risks, the economies of countries that depend on the hydrocarbon sector as a major source of income might be negatively affected. One major risk is the price fluctuations of the resources used in the multiple stages of the supply chains. Investment decisions in this sector aim to secure the investment portfolio’s financial returns against the risk of price fluctuations. This work introduces an adaptation of a portfolio optimization technique, the modern portfolio theory (MPT) to the case of commodity chemicals and energy supply chain investments by considering all supply chain stages in formulating the MPT framework. A case study considering four chemical commodities and three potential importing countries is presented with a sensitivity analysis that studies the impact of changing the costs associated with the several stages of the supply chain on the investment portfolio. Results show the importance of integrating both the chemical production and the shipping stages of the supply chain in developing a portfolio optimization framework and its impact on the optimal investment portfolios. The developed model guides investment planners to target the desired financial returns at a minimum risk.

The unit commitment problem determines the optimal strategy to meet the electricity demand at minimum cost by committing power generation units at each point of time. Solving the unit commitment problem gives rise to a challenging optimization problem due to its combinatorial complexity and potentially long solution time requirements. Our proposed solution approach utilizes a decomposition method in conjunction with alternative models from the EGRET library. Results of this decomposition approach tested against four benchmarking systems show that significant computational speed ups are achieved.

A key technology for integrating fluctuating renewable energy into the process industry is the production of green hydrogen through water electrolysis plants. Scaling up electrolysis plant capacity remains a significant challenge for the renewable energy transition. System simulation of large-scale electrolysis plants can support process design, monitoring, optimization, and maintenance scheduling. Hybrid modeling methods are promising for improving simulation reliability by combining process knowledge with process data, addressing gaps in understanding of the underlying processes. These hybrid, semi-parametric models have shown improved accuracy than purely mechanistic models. This study develops a hybrid, semi-parametric model for an anion exchange membrane electrolysis (AEMEL) stack unit. Parameters such as heat loss and heat transfer, which cannot be directly measured, are estimated using real process data. Sensors provide data on lye tank temperature, outlet temperature, and flow rate, enabling estimation of heat transfer coefficients and losses. The hybrid model is validated against operational data from different load settings of the AEMEL stack unit. To test its scalability, a large-scale electrolysis plant configuration is simulated, comprising multiple AEMEL stack units, a water supply module. Performance accuracy and efficiency of the hybrid model are compared with the mechanistic model. This hybrid model lays the foundation for future use in efficient system simulations with surrogate models, potentially enhancing large-scale electrolysis plant performance and renewable energy integration.

With an increasing share of non-dispatchable renewable energy sources in the European grid, energy flexibility will be key for the industrial sector to support the green transition. The EU-project Flex4Fact aims at finding solutions for energy and process flexibility for industry, using SINTEF’s open-source energy system model EnergyModelsX to quantify the potential benefits. This work presents some extensions done in EnergyModelsX, denoted as EnergyModelsFlex, to accommodate energy and industrial flexibility, adding new functionalities to assist with industrial flexibility potential. The extended EnergyModelsX model is described and demonstrated through two case studies in the plastic and polymeric products manufacturing sector to evaluate their potential for increasing renewable generation and flexibility. The first use case, being energy intensive, consumes both natural gas and electricity. This site enables the use of heat recovery and utilization, hydrogen blending, on-site hydrogen production, which can reduce CO2 emissions. The second use case relies solely on electricity consumption, and the considered flexibility is energy shifting by electric batteries and production flexibility. The focus of this case study is on the interplay between energy storage, on-site energy production and process flexibility to increase the degree of self-produced renewable energy in the energy mix. Together, the two case studies demonstrate how the extended EnergyModelsX framework can be used to explore process and energy flexibility in the industry to aid the transition from a fossil-based society to a renewable based society.

Superstructure based approaches have long been employed for optimal process synthesis problems. Due to the difficulties of using rigorous process models and simultaneous solutions, shortcut calculations have been the preferred means of modeling unit operations within larger process network problems. However, even with the use of shortcut equations to model the behaviour of unit operations, the resulting mixed-integer programs can be challenging to solve. Furthermore, generating the problem superstructure has often been done manually, presenting issues for scaling to larger problems. We demonstrate the use of an algorithmic approach to generate network superstructures for synthesis problems coupled with equation reformulations to yield an MIQCP (Mixed-Integer Quadratically Constrained Program) for networks of thermally coupled distillation columns. The combination of rapid problem generation with the ability to leverage recent advances in the performance of QCP (Quadratically Constrained Program) solvers enables efficient and scalable solutions for these process synthesis applications.

Olefins are essential precursors in producing a wide range of chemical products, including plastics, detergents, adhesives, rubber, and food packaging. Ethylene and propylene are the most ubiquitous olefin components and are predominantly produced through steam cracking. However, steam cracking is highly energy- and carbon-intensive, making its decarbonization a priority as the energy sector shifts toward clean, renewable electricity. Electrifying the steam cracking process is a promising pathway to reduce carbon emissions. However, this is challenged by the intrinsic conflict between the continuous operational nature of ethylene plants and the intermittent nature of renewable energy sources (e.g., solar and wind) in modern power systems. Massive energy storage systems or full plant reconfigurations to meet the power demand of electrified crackers are shown to be economically and practically infeasible. Thus, a more viable solution is to pursue a gradual electrification pathway and operate an ethylene plant as a microgrid that adopts diverse energy sources. To optimize the operational strategy of such a microgrid considering uncertainties in renewable energy generation and market prices, in this work, we introduce a stochastic multi-objective optimization approach that minimizes operating costs and carbon emissions of steam cracking. Results from a case study not only elucidate the trade-offs between economic and environmental objectives, but also provide insights into the optimal operating scheme in sustainable ethylene production.

In response to global efforts to reduce carbon emissions, the oil refining sector, a major source of industrial emissions, has set ambitious decarbonization targets. This study analyzes trade-offs between minimizing CO2 emissions and costs through the use of Pareto optimal solutions. A superstructure optimization framework evaluates various technological pathways and timelines, employing a bi-criterion optimization approach using the ?-constraint method. Results show that cost-effective, higher-emission solutions often involve natural gas-based technologies with carbon capture, while expensive, low-emission solutions favor electric power-based technologies. The analysis incorporates detailed assumptions about grid carbon intensity of varying degrees and accounts for varying national policies. Comparative case studies across locations highlight how grid carbon profiles influence optimal strategies, providing insights to inform local policies and incentivize technologies.

Green hydrogen is widely recognized as a key player in the decarbonization of the energy system. To transport it efficiently, hydrogen must be converted into a carrier, such as liquefied hydrogen or ammonia, to increase its volumetric density. The supply chain of these carriers includes hydrogen conversion into the carrier, overseas transport, and carrier reconversion back to hydrogen. A case study involving hydrogen transportation across the Mediterranean Sea is used to evaluate the carrier efficiency. The processes involved in the supply chain are simulated in Aspen Plus® V11 to determine material and energy balances, and the “net equivalent hydrogen” method is applied to calculate the equivalent amount of hydrogen needed to supply thermal or electric power. The efficiency, defined as the ratio of net hydrogen delivered (after accounting for consumption and boil-off losses) to the initial hydrogen input, is higher for ammonia than for liquefied hydrogen (73% vs 60%, respectively). This advantage lies in the lower overall hydrogen consumption for ammonia synthesis and cracking compared to liquefaction.

The urgent need to mitigate climate change and reduce reliance on fossil fuels highlights green hydrogen as a key component of the global energy transition. This study assesses the feasibility of producing hydrogen offshore using wind energy, focusing on economic and environmental aspects. Offshore wind energy offers several advantages: access to water for electrolysis, potentially lower hydrogen export costs compared to electricity, and storage systems that stabilize wind energy output. However, significant challenges remain, including the high costs of storage solutions, capital expenditures (CAPEX), and operational costs (OPEX). A Mixed-Integer Linear Programming (MILP) model optimizes the production units, storage, and distribution processes. A case study in southern France examines hydrogen production from a 150 MW floating wind farm. While hydrogen produced from offshore wind ranks among the most environmentally friendly, its costs remain high, and production volumes fall short of meeting regional demand. The findings emphasize the need to optimize power capacities and explore hybrid renewable systems to better balance economic and environmental trade-offs, enhancing offshore wind’s competitiveness in hydrogen production and its role in the global energy transition.

Deriving accurate cost projections associated with producing hydrogen within the context of an energy-export paradigm is a challenging feat due to non-deterministic nature of weather systems. Many research efforts employ deterministic models to estimate costs, which could be biased by the innate ability of these models to ‘see the future’. To this end we present the findings of a multistage stochastic model of hydrogen production for energy export (using liquid hydrogen or ammonia as energy vectors), the findings of which are compared to that of a deterministic programme. Our modelling found that the deterministic model consistently underestimated the price relative to the non-deterministic approach by $ 0.08 – 0.10 kg-1(H2) (when exposed to the exact same amount of weather data) and saw a standard deviation 40% higher when modelling the same time horizon. In addition to comparing modelling paradigms, different grid-operating strategies were explored in their ability to mitigate three critical co-sensitive factors of the production facility: high-cost hydrogen storage, uncertainty in weather forecasting and sluggish production processes. We found that a ‘grid-wheeling’ strategy substantially reduces the production cost for a solar system (by 16% and 21% for LH2 and NH3, respectively) due to its ability to guarantee the return of energy borrowed overnight during the day, but was not effective for the wind system, due to the non-periodic nature of aeolian weather patterns.

Fossil fuel reliance in the transportation sector remains a leading contributor to global greenhouse gas emissions, underscoring the urgent need for renewable alternatives like biodiesel. Derived from renewable feedstocks, biodiesel can reduce emissions, enhance energy independence, and support rural economies. However, its production faces challenges such as low energy efficiency, process optimization barriers, and the limited utilization of byproducts like glycerol, which elevate costs and hinder large-scale adoption. This study addresses these challenges by developing an integrated framework for biodiesel production and byproduct valorization, supporting the long-term decarbonization of biofuel production. Three key feedstocks—refined palm oil, rapeseed oil, and soybean oil—are evaluated for biodiesel yield. The single-step transesterification process is enhanced through a two-stage approach, optimizing fatty acid methyl ester conversion under varying methanol and NaOH catalyst split ratios. Glycerol valorization strategies, including combustion, purification, and anaerobic digestion, are analyzed to improve economic and environmental sustainability. Using mixed-integer linear programming (MILP), the study minimizes total costs by balancing operational and capital expenditures while incorporating process integration constraints. The findings provide actionable insights into cost-effective and sustainable biodiesel production pathways, promoting the broader adoption of biofuels in the transition toward net-zero emissions.

This study aims to develop an optimization-based approach to design the carbon capture, utilization, and storage (CCUS) supply chain and analyze the optimal configuration and investment strategies. To achieve this goal, we develop an optimization model that determines the logistic decision-making to maximize the net present value (NPV) and minimize the net CO2 emissions (NCE) of the strategies of the CCUS supply chain under logical and practical constraints. We estimate the technical (production scale and energy consumption), economic (capital and operating expenditure), and carbon-related (CO2 emissions) parameters based on the literature. By adjusting major cost drivers and economic bottlenecks, we determined major decision-making problems in the CCUS framework, such as sequestration vs. utilization. As a real case study, the future CCUS system of South Korea was evaluated, which includes three major CO2 emitting industries in South Korea (power plants, steel, and chemicals), as well as transportation modes and sequestration sites. As a result, we analyzed different design and operation strategies based on various objectives. We provided a strategic solution for national-level planning of the CCUS supply chain. The major finding of this study can support industry stakeholders and government policymakers by providing a practical guideline for investing in the deployment of CCUS.

Perishable consumer products like food, cosmetics, and household chemicals face challenges in supply chain management due to limited shelf life and uncertainties in demand and transportation. To address some of these issues, this work proposes a robust optimization framework for jointly optimizing facility allocation and inventory management. The framework determines optimal locations for distribution centers and their assigned customers, as well as inventory policies that minimize the total costs related to transportation, distribution, and storage under uncertain demand in a robust setting. Specifically, we develop a two-stage mixed-integer linear programming (MILP) model is that incorporates First-In-First-Out (FIFO) inventory policy to reduce spoilage. The bilinear FIFO constraints are linearized to improve computational efficiency. Social equity is integrated by defining a fairness index and incorporating it in facility allocation. Demand uncertainty is tackled using a robust optimization approach with affine demand functions to handle multiple scenarios. The model is solved using row and column generation techniques for scalability. Overall, this robust optimization framework is expected to enhance supply chain resilience, reduce waste, and improve cost-effectiveness in managing perishable products.

This work proposes a two-level model, combining process and supply chain models, and an optimization framework for an integrated biorefinery system to convert agricultural residues into renewable urea via gasification routes. The process model of the gasification, ammonia and urea synthesis was developed in Aspen Plus® to identify key performance indicators such as energy consumption and relative yields for urea for different biomasses and operating conditions; then, these key process data were used in a mixed-integer linear programming (MILP) model, designed to identify the optimal combination of energy source, technological route of urea production and plant location that maximizes the net present value of the system. The model was applied to the whole Brazilian territory, divided into 5569 cities and 558 micro-regions. Each region’s agricultural production was evaluated to estimate biomass supply and urea demand. The Assis microregion, in close proximity with sugarcane and soybean crops, was found to be an optimal location for the plant, with an NPV of 110 MMUSD. Urea demand was found to be the main limiting factor, with more than 30 microregions needed to consume the plant’s production. The results indicate the feasibility of producing renewable urea from agricultural waste and the importance of considering a two-level approach to assess the economic performance of the entire system.

Production scheduling in the process industry is an area of significant activity in research and of great practical importance for the performance of industrial companies. In the vast majority of research papers, the scheduling problem is formulated as an off-line problem where a number of jobs is scheduled on a number of resources and the efficiency of the formulation and the solution algorithms is discussed. In reality, however, scheduling is a continuous activity that has to react to the arrival of new orders, to variations in processing times, breakdowns, lack of resources etc. This is termed real-time (or online) scheduling. Available commercial solutions usually provide solutions with relatively long update intervals due to the necessary computation times and a delayed flow of information from the manufacturing execution systems where the data on the current state of the production is collected. Thus the computed schedules are outdated quickly, if not already at the point in time when they become available, and new orders and all kinds of variations and disturbances are handled by the operators and plant supervisors. There has been little concern for systematic solutions for real-time scheduling that continuously react to the changing situation in the plant. In this contribution we present a simulation-optimization approach to real-time scheduling where a detailed discrete-event simulation of the plant is coupled with an evolutionary algorithm. The simulation model is continuously updated with the latest plant data, and the optimization algorithm continuously improves the schedule based on the currently available information. We validate our approach using an example of a multiproduct, multistage batch plant in the pharmaceutical industry from the literature, demonstrating that the proposed approach can generate high-quality solutions quickly. We compare the results with those of an ideal (clairvoyant) scheduler, demonstrating that our method effectively manages unforeseen disturbances.

By 2050 around 12% of cumulative emissions reductions will come from Carbon Capture, Utilisation and Storage (CCUS) making it an essential component in the path towards net zero [1]. Focus will initially be on the retrofitting of fossil fuel power plants, which will shift to hard-to-decarbonise industries such as iron, steel, and concrete [1]. Such industries are often grouped together in industrial clusters. Comprising both large and small point sources concentrated over a defined geographical area, industrial clusters offer an opportunity to maximise the impact of CCUS whilst also improving economic feasibility [2]. The North Sea Port (NSP) cluster an example of this. Within the NSP cluster an initial set of five emitters are to join a capture, conditioning, and transport network by 2030. From there other emitters within the area will be able to join incrementally to 2050 [3]. However, the emitters who join and the timing of their connection will have a significant effect on the evolution the network. The pipeline network design will balance design requirements for initial emitters in a backbone network, with requirements for encouraging and enabling expansion. This study builds on scenarios defined between 2030 and 2050 [3] and applies a multi-period evolutionary-based approach (Steiner tree with Obstacles Genetic Algorithm (StObGA)) to predict pipeline year-on-year network growth in the NSP cluster. This provides a novel approach to the problem. The results are used in an examination of the potential growth of the pipeline network and an investigation of trade-offs necessary in the infrastructure design.

Renewable carbon sources, such as biomass and waste, are being explored as alternatives for sustainable plastic production. However, the significant uncertainties surrounding the environmental impact of biomass supply processes raise questions about whether these plastics positively contribute to society. Furthermore, the lack of systematic knowledge about plastics and incomplete understanding among stakeholders pose challenges to conducting comprehensive assessments and designing effective plastic life cycle systems. This study aims to clarify the carbon flow within the life cycle of biomass- and recycled-derived plastics and to design a plastic life cycle that enables the introduction of renewable carbon sources. To this end, the study analyzed the structure of plastics containing renewable carbon and conducted a flow analysis of packaging plastics in Japan. The flow analysis was conducted in the form of an optimization problem. Greenhouse gas (GHG) emissions and the proportion of renewable carbon were selected as assessment indicators and objective functions. Constraints for optimization included the availability of biomass supply, the demand for product resins, and the processing capacity of various End-of-Life (EoL) methods. As a result, it was revealed that the priority of resin types to be produced and recycled shifts in response to changes in constraints such as biomass input and recycling capacity. Furthermore, a trade-off between reducing environmental impact and increasing the renewability of carbon was identified, highlighting the necessity of designing systems that balance these factors. This research provides valuable insights for designing sustainable plastic life cycles in the future.

To address the dynamic operational demands of the gas processing sector, which is continuously evolving due to gas market opening, increase in natural gas production, and the growing challenge of upstream-midstream integration in a competitive environment, this work presents the Integrated-Gas-Scheduling-System, IntegraGAS. The proposed methodology innovates by using first principles rigorous process simulation coupled with a scheduling tool for short/medium/long-term, enabling gas plants to swiftly adapt to varying operational conditions and meet the requirements of this new market. IntegraGAS was implemented in Petrobras and has significantly enhanced scheduling efficiency, reducing execution time by up to 99.2% and avoiding approx. US$ 2.3 million in annual labor costs, optimizing resource utilization. By integrating Excel for the frontend, Aspen HYSYS for process simulation, VBA for automation, and Microsoft PowerBI for real-time data visualization, IntegraGAS improves decision-making, regulatory compliance, and operational agility. Its key functionalities include alerts for operational limit violations, automated mass and energy balance calculations, optimized gas allocation, integration with maintenance shutdown plans and KPI monitoring. With an intuitive interface and a robust architecture driven by digital transformation, IntegraGAS eliminates manual inefficiencies, ensures seamless coordination among stakeholders, and enables rapid responses to market dynamics.

The battle against climate change and the search for innovative solutions to mitigate its effects have become the focus of the researchers’ attention. One potential approach to reduce the impacts of the global warming could be the design of a Carbon Capture and Storage Supply Chain (CCS SC). However, the high complexity of the model requires exploring alternative ways to optimise it. In this work, a CCS multi-period supply chain for Europe is designed. Data on CO2 emissions have been sourced from the EDGAR database, which includes information spanning the last 50 years. Since this problem involves optimising the cost and operation decisions over a 10-year time horizon, it would be advisable to forecast carbon dioxide emissions to enhance the reliability of the data used. For this purpose, a neural network-based model is implemented for forecasting N-Beats. Furthermore, a temporal decomposition scheme is used to address the intractability issues of the model. The selected method is Lagrangean decomposition, which has been employed in other high complexity works, demonstrating strong performance and significant computational savings.

Despite the potential benefits of decision-level integration for process and energy supply chains (SCs), these systems are traditionally assessed and optimised by incorporating simplified unit operation models within a spatially distributed network. The desired organisational-level integration cannot be achieved without leveraging complex computational tools and concepts. This work proposes a multi-scale agent-based model to facilitate the transition from traditional practices to coordinated SCs. The proposed multi-agent system framework incorporates different enterprise dimensions of the process and energy SCs, including raw material suppliers, rigorous processing plants, and consumers. The behaviour of each agent type and its interactions are implemented, and their impact on the overall system is investigated. This approach allows for the simultaneous assessment and optimisation of process and SC decisions. By integrating each decision level into the operation, the devised framework goes beyond existing studies in which the impacts of lower decision levels are neglected. A biofuel SC example comprising farmers, biorefineries, and end-users is presented to demonstrate the application of the proposed multi-agent system. The involved actors seek to increase their payoffs given their interdependencies, intra-organisational variables, and the underlying dynamics of the network. The aggregated payoff of the supply network is optimised under different scenarios, and fractions of capacity allocated to biofuel production and consumption are obtained. The results indicate that integrated decision-making significantly influences SC performance. The proposed research expounds a more realistic view of multi-scale coordination schemes in process and energy SCs.

In the last decades, the growing awareness of climate change and the high political sensitivity of critical resources such as energy have emphasized a need for local, renewable and optimized energy mixes at various scales. Several studies have therefore aimed to optimize renewable energy technologies and plant locations to develop more renewable and efficient Energy Mixes. Following this trend, this paper applies and compares Goal Programming, Branch-and-Cut and NSGA-II to a multi-objective combinatorial optimization problem focused on the energy mix of Oise, France. Results show more optimality for Goal Programming and Branch-and-Cut, accompanied by a high sensitivity to constraints, while NSGA-II provides more technological diversity in the computed solutions.

Real-world flowshop problems which are very common in the chemical industry are often difficult to solve in a reasonable time with allocation, sequencing, and lot-sizing decisions. Although great progress has been made in the last 20 years regarding MILP model formulations and solution algorithms, realistically-sized flowshop problems with resource and buffer constraints are still difficult to solve. On the other hand, discrete-event simulation (DES) allows for very detailed modelling of process plants, but lacking of optimization capabilities. Simulation Optimization (SO) combines the high-detail DES with mathematical optimization. We show that is possible to integrate MILP and DES using Benders decomposition. We explain the Benders-DES (BDES) approach with a small motivation example with makespan minimization objective and apply it to a real-world case study of a formulation plant with seven formulation and filling lines with sequencing, allocation, and lot-sizing decisions. We show that the BDES approach performs comparably to the original monolithic-sequential MILP-DES approach. However, the key advantage of this method is the ability to leverage MILP optimization power without requiring a detailed model description.

Sustainable industrial development depends on optimizing resource and energy integration within Eco-industrial parks (EIPs), combined with stringent carbon emissions reduction policies. The main challenge is ensuring transparency, accountability, and data privacy while optimizing the conversion of raw materials and energy into valuable products and controlling emissions within EIPs. This research introduces an innovative framework to design optimized EIPs and deploy a blockchain-enabled trading platform for resources and emissions management, tackling these key issues. The proposed framework integrates EIPs with emission control policies, supported by two distinct smart contracts: one dedicated to blockchain-based resource trading and another handling financial transactions related to emission control policies, including other regulations such as income tax. The resource trading platform fosters transparency, enabling accurate tracking of material and energy flows. Furthermore, the framework integrates an off-chain Mixed-Integer linear Programming model (MILP) to optimize EIP design and operations, which is seamlessly integrated with smart contracts on the Ethereum blockchain (BC) to ensure data privacy and traceability among processes to meet environmental targets. The model also determines emission reductions and investments in carbon capture technology, promoting operational efficiency. By incorporating identity verification and external entities for compliance, the framework ensures secure and regulated operations. Offering a powerful tool to decision-makers and authorities, this framework enhances comprehension of resource and emissions tracking, paving the way for the development of innovative policies and fostering regulatory compliance. This development promotes sustainable industrial activities and supports environmental goals.

This work presents an approach to implement and assess fractional controllers in an extractive distillation system. The experimental dynamic data for an extractive distillation column is used as a case study. A strategy is developed to fit the operation data to fractional-order transfer functions. Then, the fractional controllers are designed in the Simulink environment in Matlab, tuning the controllers through a hybrid optimization approach. First, the approach uses a genetic algorithm to find an initial point, and then the solution is improved through the fmincon algorithm. According to the results of the design of fractional controllers, the sum of the square of errors is below 2.9×10-6 for perturbations in heat duty, and 1.2×10-5 for perturbations in the reflux ratio. Moreover, after controller tuning, a minimal value for ISE of 1,278.12 is obtained, which is approximately 8% lower than the value obtained for an integer-order controller.

This paper focuses on the control and dynamics of an oil refinery process on an intermediate level – the flows, masses and compositions of and between units within the refining operation. It aims to elucidate optimal strategies for the routing of streams during upset events imposed on the process. A general flowsheet simulation technique including tunable controllers for flows, compositions, levels and reaction extents is incorporated in a Linear Programming model. A standard node represents a mixed receiving tank, with exit streams which can be split, converted and separated. These nodes can be inter-connected arbitrarily in the flowsheet. The method is demonstrated for the case of a planned 3-day shutdown of the catalytic cracker.

The optimization of shell-and-tube heat exchangers (STHEs) is critical for improving energy efficiency, reducing operational costs, and mitigating environmental impacts in industrial applications. This study evaluates the performance of the Success-Based Optimization Algorithm (SBOA), a novel metaheuristic strategy inspired by behavioral patterns in success perception, against seven established algorithms—Cuckoo Search, Differential Evolution (DE), Grey Wolf Optimization (GWO), Jaya Algorithm, Particle Swarm Optimization, Teaching-Learning Based Optimization, and Whale Optimization Algorithm—for minimizing the total annual cost (TAC) of STHE designs. Using the Bell-Delaware method, the optimization framework incorporates eleven decision variables, including geometric and operational parameters, subject to thermo-hydraulic constraints. A penalty function method enforces feasibility by dynamically adjusting constraint weights. Statistical analysis of 30 independent runs reveals that DE achieves the lowest mean TAC (6,090 USD/year) with minimal variability (SD = 22.57), while GWO attains the best median TAC (6,076 USD/year). Although SBOA identifies competitive solutions (minimum TAC = 6,074 USD/year), its high standard deviation (270.66) underscores inconsistency in convergence.

The motivation of this work is an application of a production scheduling based on Real-Time Optimization and Zone Control Nonlinear Model Predictive Controller on a liquid recovery unit of an LPG production plant. In this unit, the scheduling-relevant disturbances occur on a time scale relevant to the system dynamics; thus, we propose a novel combination of a well-known control strategies leading to a hierarchical two-layered strategy, where the lower layer employs a zone control nonlinear model predictive controller (NMPC) to define inventory setpoints while the upper layer uses real-time optimization (RTO) to determine optimal plant-wide flow rates from an economic perspective. Unlike a traditional RTO, the proposed upper-layer problem is parameterized by product demands, with a distinct optimization problem formulated for each demand scenario. Our novel approach allows for proactive mitigation of potential inventory issues by dynamically recalculating the distribution of plant products. This approach addresses the challenges of storage bottlenecks and inventory fluctuations, and it is more effective than the typical operator decisions from an economic perspective in the system of interest.

Model-based control and optimisation strategies can play a key role in improving energy efficiency and reducing emissions into produced water re-injection facilities. However, building a model that adequately describes the plant is challenging and can also be used in online decision-making procedures. This work proposes a simple model based on a real water re-injection facility operating on the Norwegian continental shelf. The results demonstrate the model’s flexibility, which could be fitted to different plant operating points while being fast to solve when embedded in optimisation problems. The developed model is expected to aid the implementation of strategies like self-optimising control and real-time optimisation on produced water re-injection facilities.

Process systems are negatively impacted by manufacturing uncertainties, and increasingly by unknown-unknown disruptive events. To this effect, systems need to be designed with the inherent flexibility and resilience to overcome the impacts of uncertainties and disruptions respectively as it is more challenging to retrofit existing systems with such capabilities. To this end, we propose a methodology based on flexibility analysis to systematically explore the feasibility of design alternatives under parameter uncertainty and discrete disruption scenarios simultaneously. Multi-parametric programming is utilized to generate explicit relationships between design decisions and the resulting system’s ability to maintain feasible operations under uncertainty and disruptive events. We capture this ability by introducing the Combined Flexibility-Resilience Index (CFRI), which describes the likelihood that the system is feasible under the relevant uncertainty and disruption sets. With explicit functions for the CFRI, the flexibility and resilience objectives can be incorporated into a general superstructure design optimization problem alongside other objectives, allowing for the exploration of trade-offs among flexibility, cost and resilience. Our methodology provides the foundation of a holistic approach for design optimization for both flexibility and resilience. A case study involving a process plant with multiple sources of uncertainties and disruptive events is used to illustrate the application of the proposed methodology.

One market scope of bioindustries is the production of recombinant proteins for its application in serotherapy. However, its process’s monitoring and optimization present limitations. There are different approaches to optimize bioprocess performance; one is using model-based control strategies such as Model Predictive Control (MPC). Another strategy is learning-based control, such as Reinforcement Learning (RL). In this work, an RL approach was applied to maximize the production of recombinant proteins in E. coli at the induction phase using as a control variable the substrate feed flow rate (Fin). The RL model was trained using the actor-critic Twin-Delayed Deep Deterministic (TD3) Policy Gradient agent. The reward corresponded to the maximum value of protein productivity. The environment was represented with a dynamic hybrid model. The optimization was evaluated by stages of two hours to check the protein productivity performance. Afterwards, the results were compared with an MPC approach. Finally, the control approaches were trained considering temperature disturbances. The results elucidate that the RL approach could be implemented as a control strategy, reaching values from 0.014 mg/h to 0.079 mg/h through all the optimization stages previously demonstrated to be the optimal ones. Despite exhibiting temperature disturbances, the RL approach demonstrated its robustness by adapting the control action to maintain similar protein productivity values.

Safe Bayesian Optimization (Safe BO) has demonstrated significant promise in enhancing data-driven optimization strategies in safety-critical settings, where model discrepancies, noisy measurements, and unknown safety constraints are prevalent. Despite these advancements, there still remains a limited understanding on the effectiveness and applicability of these Safe BO methods, particularly within process system engineering. Specifically, this study adapts and examines Safe Exploration for Optimization with Gaussian Processes (SafeOpt), Goal-oriented Safe Exploration (GoOSE), Gaussian Processes with Trust Region (GPs-TR) and Adversarially Robust Gaussian Processes (StableOpt). Methods such as SafeOpt and GoOSE face challenges in managing continuous systems due to their reliance on system discretization and together with StableOpt, lack the capability to manage multiple safety constraints. Thus, this work presents a comprehensive evaluation of state-of-the-art safe BO methods, with our own addition and enhancements, focusing on their performance in process system engineering. The performances of methods are assessed using a case study on the William Otto Reactor, focusing on convergence speed, unknown constraints mitigations, practicality, and robustness against adversarial perturbations. Overall, this study underscores the unique strengths and limitations of safe BO methods in process system engineering, advancing the field on data-driven approaches in safety-critical processes while also identifies areas where further improvements are necessary.

Successful implementation of optimization-driven control techniques, such as model predictive control (MPC), is highly dependent on an accurate and detailed model of the process. As complexity in the system increases, linear approximation used in MPC may result in poor performance since a critical operating point is valid in only a small neighborhood of operation. To address this problem, this paper proposes a collaborative approach that combines linear and data-based models to predict state variables individually. The outputs of these models, along with constraints, are then incorporated into the MPC algorithm. For data-based process model, a multi-layered feed-forward network is used. Additionally, the offset-free technique is applied to eliminate steady-state errors resulting from model-process mismatch. To demonstrate the results, a binary distillation column process which is multivariable and inherently nonlinear is chosen as testbed. We compare the performance of the proposed method to MPC using the full nonlinear model and also to single-model MPC methods for both the linear model and neural network model. We show that the proposed method is only slightly suboptimal with respect to the best available performance and greatly improves over individual methods. In addition, the computational load is reduced when compared to the full nonlinear MPC.

Liquid-liquid extraction (LLX) is an essential technique for separating heat-sensitive, highly diluted, or azeotropic mixtures. However, the design and optimization of LLX processes can be challenging due to mass transfer limitations and complex fluid dynamics. While distillation can often be modeled using equilibrium-based (EQ-based) approaches with (constant) height equivalent to theoretical stage (HETS) values, these kinetic effects can limit the applicability of EQ-based LLX models for conceptual design. Non-equilibrium (NEQ) or rate-based modeling can account for detailed mass transfer and fluid dynamics but further increases the nonlinearity and complexity of the respective optimization problems, which should account for closed-loop solvent recovery. To successfully address these complexities, we propose an integrated methodology combining NEQ-based simulation with EQ-based superstructure optimization to design a hybrid extraction-distillation process. An NEQ model is first used to derive operation-dependent HETS correlations, which are then incorporated into an EQ-based superstructure model for techno-economic optimization targeting total annualized cost. This approach balances model fidelity and computational efficiency, providing more reliable solutions by capturing the solvent-specific mass transfer behavior. We illustrate the methodology for a dilute acetone-water system with different solvents.

Due to the intermittent nature of most renewable energy sources, developing good and flexible control structures for green electrolysis systems is crucial for maintaining efficient and safe plant operation. This work uses the “top-down” section of Skogestad’s plantwide control procedure to propose a suitable control architecture for PEM electrolysis systems based on advanced regulatory control. Advanced regulatory control structures, such as active constraint control, may offer several advantages over MPC and AI-based control methods as they are computationally less expensive, less affected by model accuracy, easier to scale, and offer fast disturbance rejection. In our approach, we first mapped the constraint regions for the system. Then, we reduce the complexity by reformulating the optimization problem slightly, to remove some constraint regions to obtain a simpler solution structure that gives a negligible loss. Finally, we propose an active constraint control architecture using PI controllers and a selector for switching between the active constraint regions. We found that the system’s fully constrained solution is close to the optimal solution, implying that our active constraint control approach may be used with little loss in efficiency. The proposed control structure was shown to perform well in dynamic simulations.

Microfluidic mixing has gained popularity in the Pharmaceutical Industry due to its application in the field of Nano-based Drug Delivery Systems (DDS). The flow conditions in Microfluidic mixers enable very efficient mixing conditions, which are crucial for the production of Nanoparticles by Flash Nanoprecipitation (FNP), as it enables reproducible production of particles with low-size variability. Mixer geometry is one of the most determinant factors, as it largely determines the flow patterns and the degree of contact between the two mixing streams. In this paper, a shape optimization methodology using Computational Fluid Dynamics (CFD) and Bayesian optimization is applied to the toroidal micromixer design, considering three different operating conditions. It consists of first defining a geometry solution space and then using Multi-Objective Bayesian optimization to explore the different designs. Mixer performance is evaluated with CFD simulations and two objective functions are considered: mixing time and pressure drop. Approximations of the Pareto front for each case study are obtained and the analysis of the best geometries enabled to derive some general geometric features of optimal mixer designs.

Recent advancements in machine learning and time series analysis have opened new avenues for improving predictive control in complex systems such as mineral flotation. Techniques leveraging multivariate predictive control in mineral flotation have seen significant progress in recent years. However, challenges in developing an accurate dynamic model that encapsulates both the pulp and froth phases have hindered further advancements. Now, with a readily available model containing equations that describe the physics of flotation froths, an opportunity for novel control strategies presents itself. In this study, a Gaussian Process (GP) Model Predictive Control (MPC) strategy is proposed to integrate uncertainty quantification directly into the control framework. By leveraging the probabilistic nature of GP models, this approach captures process variability and adapts dynamically to new data, ensuring continuous refinement of the GP model within the MPC strategy. Unlike previous implementations where model parameters remained static, this methodology updates the GP model in real time, allowing for improved decision-making in the face of process uncertainty. The GP model was trained, optimized with JAX, evaluated using relevant metrics, and implemented as a surrogate within the MPC framework. The results demonstrate the capability of the GP model to accurately represent process dynamics while minimising prediction errors. Moreover, incorporating uncertainty reduction through standard deviation minimisation in the objective function enhances both control performance and system robustness. This paper sets the basis for the potential of using GP-MPC to enhance both the accuracy and robustness of mineral froth flotation control.

Recently, the focus on increasing process efficiency to reduce energy consumption has driven the adoption of alternative systems, such as Petlyuk distillation columns. It has been proven that, when compared to conventional distillation columns, these systems offer significant energy and cost savings. From an economic standpoint, achieving high-purity products alone does not ensure the feasibility of a process. Instead, balancing the trade-off between product purity and cost necessitates multi-objective optimization. While conventional optimization methods are effective, novel strategies like Bayesian optimization offer distinct advantages for handling complex systems. Bayesian optimization requires no explicit mathematical model and can efficiently optimize even when starting from a single initial point. However, as a black-box method, it demands a detailed analysis of hyperparameters, such as the acquisition function and the number of initial points, to ensure optimal performance. This work presents a case study of a Petlyuk distillation column, focusing on the influence of key hyperparameters in Bayesian optimization. Additionally, the role of uncertainty in the optimization process is explored, as it is a critical factor in real-world applications. By simulating uncertainty, this study provides insights into its impact on the optimization process.

Asserting the phase stability of multi-component mixtures is an important task in computer-aided mixture/blend design (CAMbD), but it is often hindered by the lack of reliable and tractable models. In this paper, we propose a group-contribution machine-learning (GC-ML) method to predict phase coexistence for a large set of ternary mixtures consisting of two solvents and one (fixed) solute. Each solvent is represented by a vector of functional group numbers, encoded by integer values. The solvent vectors are combined with mixture composition and temperature to form the input features to a GC-ML surrogate classifier, which distinguishes between four types of stable phase configurations as possible outputs: liquid (L), solid-liquid (SL), liquid-liquid (LL) or solid-liquid-liquid (SLL). To explore the performance of the trained GC-ML multi-classifier, it is embedded as a surrogate phase-stability constraint in the optimisation of an ibuprofen crystallisation process. A two-step solution strategy is proposed, iterating between a surrogate-based subproblem and a rigorous UNIFAC-based subproblem, to design binary solvent mixtures that improve the yield of ibuprofen. A high classification accuracy score of over 0.96 is achieved in identifying the correct phase configurations with the surrogate model, making it possible to accurately predict whole phase diagrams for numerous mixtures. Moreover, reliable solutions to the crystallisation design problem are generated with the two-step strategy, yielding objective function values close to those with the UNIFAC-based model at a lower computational cost. These results reveal the value of the proposed surrogate model in guiding the search for better solvent mixtures in CAMbD applications.

Process optimisation and quality control are crucial in process industries for minimising product waste and improving plant economics. Identifying robust operational regions that ensure both product quality and performance is particularly valued in industries. However, this task is complicated by operational uncertainties, which can lead to violations of product quality constraints and significant batch discards. We propose a novel robust optimisation strategy that integrates advanced machine learning and process systems engineering to systematically identify optimal operational regions under uncertainty. Our approach begins by using a process model to screen a broad operational space across various uncertainty scenarios, pinpointing promising control trajectories to satisfy process constraints and product quality. Machine learning is then employed to cluster these trajectories into sub-regions. Finally, a two-layer dynamic optimisation framework is employed to determine the optimal control trajectory and corresponding operable space within each promising sub-region. To demonstrate the efficiency of our approach, we used a case study focusing on the quality control of a dynamic batch process for formulation product manufacturing. The resulting operational regions were shown to meet product quality demands and offer a significant improvement in optimality over the current operation, highlighting the advantage and industrial potential of our strategy.

The Reverse Osmosis (RO) system has the potential as a vibrant technology to generate high-quality water from brackish water sources. Nevertheless, the progressive growth in water and electricity demands necessitates the development of a sustainable desalination technology. This can be achieved by reducing the specific energy consumption of the process, which would also reduce the environmental footprint. This study proposes the concept of reducing the overall energy consumption of a multistage multi-pass RO system of Arab Potash Company (APC) in Jordan via heating the feed brackish water. The utilisation of waste heat generated from different units of production plant of APC such as steam condensate supplied to a heat exchanger is a feasible technique to heat brackish water entering the RO system. To systematically assess the contribution of water temperature on the performance metrics including specific energy use, a generic model of RO system is developed. Model based simulation is used to evaluate the effect of water temperature. The results indicate a clear enhancement of specific energy consumption while using water temperatures close to the maximum recommended temperature of the manufacture. It has been noticed that an increase in water temperature from 25 ºC to 40 ºC can result an overall energy saving of more than 27%.

Seawater desalination is considered as the only available solution that can cope with the increasing demand for freshwater around the world. Improving the desalination techniques may help to cut off the cost and increase sustainability. In this paper, a mathematical model describing the MED process is developed within gPROMs software. The model includes all the necessary mass and energy balance equations together with thermodynamic and physical properties equations. The model predictions are validated against the actual plant data before using the model for optimizing the process to achieve minimum heat transfer area. For two different operating conditions (summer and winter) and a fixed production demand, the heat transfer area is minimised while optimising different parameters of the MED process. The results showed that a 10.4% reduction in the heat transfer area can be achieved under summer operating conditions and around 26% decrease in the heat transfer area can be met under winter operating conditions.

Enhanced weathering (EW) of alkaline minerals in chemical reactors with a controlled environment is recognized as a promising approach for gigaton-level carbon dioxide removal. However, reactor configuration and operating conditions must be optimized to balance the interfacial areas between gas, liquid and solid phases prior to industrial application. We developed a physics-based and data-driven hybrid modelling approach, coupled with multi-objective optimization, to study and compare three typical chemical reactors, i.e., trickle bed, packed bubbling columns, and stirred slurry reactors, and the optimal design to improve CO2 capture rate and reduce energy and water consumptions. Then an adaptive optimization is proposed to dynamically adjust the operating of the reactors in response to intermittent CO2 emission and renewable energy supply. Results indicated that forced stirring enhances CO2 capture rates by accelerating mass transport but increases energy consumption. Trickle bed reactor achieves the highest capture rate and lowest energy and water consumption. The AI-enabled, multi-objective, and adaptive optimization developed in response to intermittent CO2 and energy supplies shows considerable potential for industrial applications, advancing progress towards achieving SDG #13.

Maintaining operational efficiency while ensuring safety is a longstanding challenge in industrial process control, particularly in high-risk environments. This paper presents a novel Dynamic Risk-Informed Explicit Model Predictive Control (R-eMPC) framework that integrates safety and operational objectives using probabilistic constraints and real-time risk assessments. Unlike traditional approaches, this framework dynamically adjusts safety thresholds based on Bayesian updates, ensuring a balanced trade-off between reliability and efficiency. The validation of this approach is illustrated through a case study on tank level control, a safety-critical process where maintaining the liquid level within predefined safety limits is paramount. The results demonstrate the framework’s capability to optimize performance while maintaining robust safety margins. By emphasizing adaptability and computational efficiency, this research provides a scalable solution for integrating safety into real-time control strategies for similar process systems.

The evaluation of control properties in industrial processes is essential to achieve sustainability, a very relevant topic today. This study emphasizes the importance of control studies to ensure that processes are efficient, operable and safe. While strategies such as process intensification can reduce the size, cost, and consumption of energy, it can present challenges in control and operability. This work focuses on the evaluation of the control properties of schemes with different degrees of intensification for the purification of levulinic acid, with the aim of identifying designs with the best control properties and the best economic and environmental indicators. The schemes were designed under a systematic synthesis strategy and optimized using the hybrid method of differential evolution with a tabu list, considering the total annual cost and Eco-indicator 99. An open-loop study analyzed the relationship between manipulable variables and output variables using total condition number, sensitivity index, and relative gain matrix analysis. The dynamic behavior in a closed loop was subsequently analyzed using the minimization of the absolute error integral as a criterion. The results showed that the design, which includes a liquid-liquid extraction column, three distillation columns, and thermal coupling, presented the best dynamic performance. This design had a low total condition number, a below-average sensitivity index, a stable control structure, and low values of the absolute error integral. In addition, it stood out for its excellent cost and environmental impact indicators, which makes it the most favorable option among the proposed designs.

Distillation is not only a widely-used but also an energy-intensive separation process, in which internals such as structured packings play an important role. Increasing mass transfer efficiency by designing improved structured packings in order to provide a large interfacial area while enabling low pressure drop is one promising approach to quickly reduce the energy requirements of vacuum distillation where low pressure drop is important for separation efficiency and thermal stability of the processed media. The current work presents an innovative method to optimize structured packings by means of constrained shape optimization on the basis of computational fluid dynamics simulations to minimize the pressure drop while maintaining a constant specific surface area. To solve the fluid dynamic optimization problem, a gradient-based local optimization algorithm in a continuous adjoint formulation is utilized. The shape optimization is applied for a commonly used Rombobak packing, and tested as a refinement method for an initial structure derived by topology optimization. The results demonstrate a successful reduction of the pressure drop in both cases, which can be attributed to several factors, including the refinement of the edges and the mending of the dead zones.

Circular economy is recognized as one of the most effective strategies for promoting plastic sustainability. However, its implementation requires to enhance consumer engagement, which remains a primary target of regulatory initiatives designed to promote plastic circular economy. To ensure sustained consumer participation, it is essential to evaluate and optimize various incentives, including regulatory policies, voluntary programs, and market-related mechanisms. This study applies Stackelberg Game Approach to quantitatively capture the strategic interactions between the authorities (as the leader) and consumers (as followers). The model incorporates key consumer behaviors, i.e., “use less,” “use longer,” and “recycling”, to reflect their role in advancing plastic circular economy goals. By integrating factors such as governmental utility (gains of benefits), consumer utility (welfare), and plastic waste reduction, the model identifies the optimal intensities of various public initiatives, which represent the quantitative levels of policy intervention necessary to maximize desired outcomes. A case study based on EU-27 data demonstrates the model’s applicability, revealing optimal intensities of 0.91, 0.41 and 0.8 for regulatory, voluntary, and market-related initiatives. At these optimal intensities, regulatory measures achieve the highest governmental utility (benefits) and plastic waste reduction, while market-related mechanisms yield more favorable outcomes for consumer welfare. The findings highlight the need for a balanced policy mix that effectively aligns government objectives with consumer incentives.

Anthropogenic carbon-dioxide emissions, exceeding 51 billion tons annually, are a major driver of global climate impacts. Aqueous amine scrubbing offers an effective carbon-capture solution, but the energy-intensive thermal regeneration step of the process significantly increases costs, limiting large-scale adoption. To address these challenges, computational optimization of process and molecular design is promising but often too resource-intensive, emphasizing the need for efficient surrogate models. Specifically, we develop a surrogate model based on an artificial neural network (ANN) that is employed to replace rigorous phase-equilibrium computations performed with the SAFT-? Mie group contribution method within a steady-state aqueous amine carbon-capture process model. Our ANN is trained on 32,768 vapour–liquid equilibrium data points of a quaternary mixture of water, monoethanolamine, carbon dioxide, and nitrogen over industrially relevant temperature, pressure, and composition ranges, achieving mean-relative errors below 2% for most variables. The ANN is integrated into a computer aided process design (CAPD) framework and evaluated via mathematical optimization, with the total annualized cost (TAC) as the objective function. Compared to the rigorous process model, the model with the ANN surrogate exhibits reduced computational times of up to 81% while converging to near-identical optima within 1% of the TAC. With a robustness analysis across the lean-loading and desorber-pressure design space we show that the ANN effectively captures 90% of the investigated design space, vs. 77% for the rigorous model. In a multi-start optimization ensuring global optimality, the surrogate converges to the same point as the rigorous model, taking 10 h less. This demonstrates that carefully trained surrogate models can significantly accelerate process optimization and enable larger-scale applications, paving the way for future solvent and process co-design efforts.

Intensified schemes, such as reactive distillation, have been proposed to produce silane (SiH4). Several studies have been carried out around this intensified scheme focusing directly on its improvement in energy or economic criteria. However, these mentioned criteria do not ensure that the scheme is also optimal from the control point of view. There is a direct compromise between the economic criterion and the control criterion. Thus, the best controllable scheme is not necessarily the most economical and vice versa. Analyses have been proposed to evaluate the controllability of steady-state processes using open-loop with Singular Value Decomposition (SVD) under quantitative a criterion such as A? + ?sm with simplified first-order transfer functions. This work considers four feasible designs with multiple reactive zones and evaluates their controllability from their open-loop dynamic responses obtained from Aspen Dynamics® by calculating the condition number for different frequency ranges analyzing qualitative aspects and the A? + ?sm criterion evaluating the controllability of each scheme and the conditions of the simplification of their dynamic responses. For this, simplifications of first and second order transfer function models are compared with the models that robustly predict their behavior. The results of this study show that first order functions, despite their simplicity, can predict well the complex model of these distillation schemes. However, for low frequencies (i.e. ? < 10 rad/h) second order functions are better options because they capture the overshooting (damping) effect. The proposed methodology can be applied in global optimization with controllability as an additional criterion in the objective function.

Solar Thermal Power Plants (STPPs) use solar energy for large-scale electricity production but face significant operational challenges. These include variations in solar radiation, cloud cover, electricity demand fluctuations, and the need for frequent shutdowns if energy storage is inadequate. Deciding an optimal STPP operating conditions is challenging due to these factors. While revenue maximization has been used as an objective in existing literature, current models are often static and fail to capture the dynamic nature of STPPs. In contrast, this work proposes a dynamic model-based revenue optimization approach that accounts for plant dynamics and operational constraints, such as solar radiation variability and changing electricity demand. The objective function is designed to maximize revenue while considering power generation and fluctuating electricity prices. A simulation model of 1 MWe hybrid solar thermal power plant in Gurgaon, India, featuring two solar fields—Parabolic Trough Collector (PTC) and Linear Fresnel Reflector (LFR), is chosen as a case study. The optimization focuses on the oil mass flow rate through the PTC, which influences steam production. The mass flow rate of oil side is optimized at discrete time intervals, forming the decision variables in a nonlinear programming problem (NLP). The results demonstrate the effectiveness of this approach in significantly enhancing the STPP’s revenue.

Wastewater treatment and resources recovery from large wastewater flowrates of the municipalities and circular bio-based economy ask for efficient control solutions. The paper presents solutions for operating the wastewater treatment plant, based on advanced process control methods aimed to merge the benefits of the cooperation between the lower-level regulatory control loops and the upper-level model predictive control strategy. The lower-level is designed to regulate the nitrification in the aerated bioreactors by controlling the Dissolved Oxygen or the ammonia concentration and to control the denitrification in the anoxic reactor by controlling the nitrates concentration. The model predictive controller either sets the setpoints of the regulatory layer or directly manipulates the air and nitrate recycle flow rates. The plant performance results obtained using the regulatory Proportional and Integral control are compared to the direct or the supervisory model predictive control outcomes. Both feedback and feedforward control configurations were considered, with the influent ammonia concentration as a representative disturbance. The WWTP control results were assessed using the Pumping and Aeration Energy, Effluent Quality, and Green House Gases emissions performance indices. They showed the advantage of the two-level control system for ammonia and nitrates concentration control with setpoints driven by the constrained model predictive controller, in association with the feedforward control.

The optimization of the power conversion system, responsible for thermal-to-electrical energy conversion, for a pulsed fusion power plant is presented. A spherical tokamak is modelled as three heat sources, all pulsed, with different stream temperatures and available amounts of heat. A thermal energy storage system is considered in the design to compensate for the lack of thermal power during a dwell. Thermal storage enables continued power generation during a dwell and can avoid thermal transients in sensitive components like turbomachines. Multiple lower grade heat sources are integrated into the process through parallel preheating trains. The evaluation of a dynamic model of the power conversion system is used to define an objective function with multiple criteria. A bi-objective optimization problem is defined to investigate the trade-off between the size of the thermal energy storage system and the variability in turbine power output during a dwell. The set of non-dominated designs demonstrates the trade-off, with designs covering the range from steady state turbine power output to a design that reduces the turbine power output by 46.3% during a dwell. The dynamics of three non-dominated designs demonstrate the power conversion system can maintain relatively stable conditions at the turbine inlet under different operating conditions, but molten salt temperature control is an issue.

In the steam power system for the centralized heat supply in an industrial park, heat demands of all consumers are satisfied by the energy station, leading to the high steam delivery costs caused by the several distant enterprises. Additionally, the number of steam levels is limited due to the trade-off between distance-related costs and heat cascaded utilization, thus some consumers are supplied with heat at higher temperature than that of required, resulting in the low energy efficiency. To deal with the above problems, this work proposes an optimization model for steam power systems (SPSs) in industrial parks, which incorporates the distributed heat supply and auxiliary steam turbines (ASTs). Field erected boilers (FEBs) can independently supply heat to consumers, thereby avoiding the excessive pipeline costs. ASTs are used for the re-depressurization of steam received by consumers, which can increase the electricity generation capacity and improve the temperature matching of heat supply and demand. A mathematical model is established for the problem, in which the influence of different numbers of steam levels is considered, and the SPSs are optimized by the objective function of minimizing the total annual cost (TAC). A case study illustrates that the optimal number of steam levels, and the selective installation of FEBs and ASTs for consumers, can effectively reduce the cost of pipeline network, increase the income of electricity generation, and minimize the TAC of the system.

This work presents a plantwide model of a Haber-Bosch ammonia synthesis loop (HB-loop) in a PtA plant, consisting of heat exchangers, compressors, steam turbines, flash separators and catalytic reactor beds. The total electrical power utility of the HB-loop is a combination of compressor power, refrigeration power, and steam turbine power. We optimise the HB-loop operating parameters, subject to constraints for maximum reactor temperatures, compressor choke and stall, minimum steam temperature, and maximum loop pressure. The loop features six degrees of freedom (DOFs) for the optimisation: three reactor temperatures, reactor N2/H2-ratio, separator temperature, and loop pressure. The optimisation minimises the total loop power utility for a given hydrogen make-up feed flow, with the PtA load varied by ranging the hydrogen make-up feed flow from 10 % to 120 % of the nominal. Across this load range, different constraints become active, with the compressor surge limit being particularly critical at low loads, significantly increasing HB-loop power consumption. To address this, we investigate configurations with two, three, and four compressor trains operating in parallel. Reductions in total power of 55%, 74%, and 84% are achieved at reduced plant loads, with two, three, and four parallel trains, respectively. In terms of total compressor capital and operating cost, we demonstrated savings of 8.06%, 8.29%, and 7.11% in total cost after ten years of operation with two, three, and four parallel compressor trains compared to a single train configuration.

Efficient temperature control in greenhouses is essential for optimal plant growth, especially in arid regions where the harsh environment poses significant challenges to maintaining a stable microclimate. Maintaining the optimum temperature range directly influences healthy plant development and overall agricultural productivity, impacting crop yields and financial outcomes. However, the greenhouse in the present case study fails to maintain the optimum temperature as it operates based on predefined settings, limiting its ability to adapt to dynamic climate conditions. To maintain an ideal temperature range within the greenhouse while dynamically adapting to fluctuating external conditions, this study introduces a control framework using Deep Deterministic Policy Gradient, a model-free deep reinforcement learning algorithm, to optimize temperature control in the closed greenhouse. A deep neural network is trained using historical data collected from the greenhouse to accurately represent the nonlinear behavior of the greenhouse system under varying conditions. The deep deterministic policy gradient algorithm learns optimal control policy by interacting with a simulated greenhouse environment, continuously adapting without needing an explicit system dynamics model. Results from the study demonstrated that, over a five-day simulation period, the deep deterministic policy gradient control system outperformed the existing greenhouse climate system in temperature regulation. It achieved a mean squared error of 0.01°C and a mean absolute error of 0.13°C. Additionally, the deep deterministic policy gradient algorithm demonstrated a significant improvement in energy efficiency, reducing total energy consumption by 6.80% compared to the current system.

This work explores Aotearoa-New Zealand’s potential to fully electrify and source industrial process heat demands from renewable energy for 286 industrial sites while exploring the feasibility of green methanol production using excess electricity. Most energy models rely on spatially aggregated supply and demand, which limits the accurate representation of energy value chains. To address this limitation, the model incorporates industrial sites with varied temperature profiles, enabling the use of diverse heating technologies such as heat pumps, electrode boilers, bubbling fluidised bed reactors and biomass boilers. The proposed Mixed-Integer Linear Programming energy model uses the Accelerated Branch-and-Bound (ABB) algorithm, which is implemented within the P-graph framework to optimise the system. The model considers different energy transportation modes, including road transport for biomass and grid infrastructure for electricity. The multi-period design determines optimal heating technology capacities for each unique industrial site while accounting for renewable energy variability, spot electricity prices, and sectoral energy demand fluctuations. The optimisation results reveal the most effective configurations of heat technologies for industrial sites and projected 1.03 Mt/y of green methanol production in New Zealand when the selling price exceeds NZD1050 /ton (EUR 569 /ton). The findings demonstrate a 41.3% reduction in total process energy supply from 25.9 TWh/year to 18.8 TWh/year, driven by the higher coefficients of performance (COP) and efficiencies of electrification.

The oil production in wells using electric submersible pumps (ESPs) demands precise control of parameters within safety and efficiency constraints to minimise failures, extend equipment lifespan, and reduce costs. This study proposes a non-linear model predictive control (NMPC) system designed for ESP-lifted wells, leveraging pump frequency and choke valve adjustments to maximise production while adhering to operational limits. Tested on a simulated pilot plant using a first-principles model to predict key variables like flow and liquid column height, the NMPC demonstrated offset-free performance, effective disturbance rejection, and ensured stable, safe, and optimised operations, addressing challenges in nonlinear, constraint-intensive environments.

In chemical process control, decision-making often involves balancing multiple conflicting objectives, such as maximizing production, minimizing energy consumption, and ensuring process safety. Traditional approaches for multi-objective optimization, such as linear programming and evolutionary algorithms, have proven effective but struggle to adapt in real-time to the dynamic and nonlinear nature of chemical processes. In this paper, we propose a framework that combines Reinforcement Learning (RL) with Multi-Objective Evolutionary Algorithms (MOEAs) to address these challenges. Specifically, we utilize MOEAs, such as NSGA-II, to optimize the parameter space of policy neural networks, resulting in a Pareto front of policies. This Pareto front provides a diverse set of policies that enable operators to dynamically switch control strategies based on real-time system conditions and prioritized objectives. Our proposed methodology is applied to a Controlled Stirred Tank Reactor (CSTR) case study, focusing on balancing the objectives of minimizing set point error and operational costs. The case studies highlight the adaptability of the method in dynamic environments, including scenarios involving fouling, showcasing its resilience and flexibility in complex, nonlinear systems. This work demonstrates the potential of combining RL and MOEAs to advance decision-making in chemical process control, providing a robust and adaptable framework for navigating environments with changing objectives. The methodology can also be extended to other industrial applications with similarly challenging multi-objective requirements.

The dividing wall column (DWC) has significant energy saving potential compared to conventional column sequences. However, to reach these savings in practice, it is essential that the control structures can track the optimal operation point despite inevitable changes in feed properties, performance characteristics and other uncertainties. Otherwise, the energy consumption may rise significantly or, more commonly, the DWC becomes unable to produce pure products even at its maximum reboiler duty. Extremum seeking control (ESC) is a model-free optimisation technique that may mitigate off-optimal operation in this environment. By active perturbation of selected manipulative variables, the algorithm infers gradient properties of the measured cost function and, by that, enables tracking of a moving optimum. Extremum seeking control can be used also in combination with other approaches, e.g. self-optimising control. Applied to the DWC, the presented perturb-and-observe algorithm, which can be classified as ESC, appears to be able to track changes in the optimal values of the manipulated variables and setpoints.

By capturing carbon dioxide from biomass flue gases, energy processes with negative carbon footprint are achieved. Among carbon capture methods, the fluidized temperature swing adsorption (TSA) column is a promising low-pressure alternative, but it has been developed on small scales. This work aims to model, simulate and optimize a fluidized TSA multi-stage equilibrium system to obtain a cost estimate and a conceptual design for future process scale up. A mathematical model described adsorption in multiple stages, each with a heat exchanger, coupled to the desorption operation. The model was based on elementary macroscopic molar and energy balances, coupled to pressure drops in a fluidized bed designed to operate close to the minimum fluidization velocity, and coupled to thermodynamics of adsorption equilibrium of a mixture of carbon dioxide and nitrogen in solid sorbents (the Toth equilibrium isotherm was used). The complete fluidized TSA process has been optimized to minimize costs, considering equilibrium in each stage. The optimal configuration for heat exchangers was determined and a unit cost for carbon dioxide capture was estimated. It was found that a small number of stages is required to meet captured carbon specification at 95% molar purity, while process cost per metric ton of CO2 was within range of other capture technologies. These findings show that the methodology developed here is useful for guiding the conceptual design of fluidized TSA process for carbon capture.

This work presents a Lyapunov-based framework for safe reinforcement learning (RL) applied to the control of an unstable reactor. The proposed method imposes stability constraints on the value and Q-functions through a Lyapunov candidate function defined as the negative of these functions, L(s)=-V(s) and L(s,a)=-Q(s,a). Constraints enforce positivity of the Lyapunov candidate function and non-positive time derivatives, promoting monotonic behavior aligned with Lyapunov stability conditions. The framework was tested on both on-policy (PPO) and off-policy (SAC, TD3, and DDPG) RL algorithms, with performance evaluated against their baseline versions and a nonlinear Model Predictive Controller (NMPC). Results showed that stability constraints significantly improved control performance across all tested algorithms, yielding consistently higher cumulative rewards, reduced overshoot, and decreased variability. Derivative-based constraints successfully mitigated abrupt changes and oscillatory behavior in the Q and value functions, especially evident in PPO, DDPG, and TD3. This work demonstrates that Lyapunov-based constraints are an effective tool for improving the safety and stability of RL algorithms in safety-critical applications. The proposed approach avoids the complexity of auxiliary optimization and offers a computationally efficient solution for enhancing the safety and stability of RL algorithms in safety-critical control applications.

Industrial decarbonization is considered one of the key objectives in mitigating global climate change. To achieve a net-zero industry requires actively transitioning from fossil fuel-based energy sources to renewable alternatives. However, the intermittent nature of renewable energy sources poses challenges to a reliable and robust supply of energy for industrial sites. Therefore, the integration of renewable energy systems with existing industrial processes, subject to energy storage solutions and main grid interconnections, is essential to enhance operational reliability and overall energy resilience. This study proposes a novel framework for the design and optimization of industrial utility systems integrated with renewable energy sources. A monthly-based analysis is adopted to consider variable demand and non-constant availability in renewable energy supply. Moreover, carbon capture is considered in this work as a viable decarbonization measure, which can be strategically combined with renewable-based electrification. The proposed optimization model evaluates the economic trade-offs of integrating carbon capture, renewable energy, and energy storage. By applying this approach, systematic design guidelines are developed for the transition of a conventional steady-state utility system toward renewable energy integration, ensuring economically viable and sustainable energy management in process industries.

This work addresses a practical challenge in batch process optimization: the need for sample efficient learning methods due to the high cost and time-intensive nature of running physical batch processes. While reinforcement learning (RL) offers a promising framework for optimizing batch processes, traditional approaches require numerous experimental runs to converge to optimal policies. A novel sample efficient RL method that leverages Gaussian Processes (GPs) to accelerate learning from limited batch data is proposed. However, the direct application of GPs becomes computationally intractable as data accumulates batch-to-batch, and their performance degrades when training distributions shift during policy improvement. To address these challenges, an integrated framework that combines Q-learning with GPs was developed and a strategic subset selection mechanism using determinantal point processes is introduced to maintain computational efficiency while preserving diverse, high-performing samples. The method exploits problem structure and backward induction to further maximize sample efficiency and incorporates both aleatoric and epistemic uncertainty for robust policy improvements. The approach is demonstrated on a non-isothermal semi-batch reactor case study, showing significantly improved learning efficiency compared to exact GP strategies while maintaining computational tractability. The results highlight the method’s practical applicability to industrial batch process optimization where experimental data is limited and costly to obtain.

Middle vessel batch distillation (MVBD) is an alternative configuration of the conventional batch distillation with improved sustainability index. This article presents a comparison of model-based control approaches for MVBD column. Specifically, two control approaches – sequential (open-loop optimization followed by closed-loop control) and simultaneous (closed-loop optimization and control) are pursued. These two approaches are compared in terms of their effectiveness, overall performance, and robustness to plant-model mismatch. The effectiveness of these control strategies is illustrated using a simulation case study of a ternary mixture separation consisting of benzene, toluene and o-xylene.

Real-time optimization (RTO) is a model-based technique that drives plants to optimal operating conditions. Modifier Adaptation (MA) is a class of methods that adjusts the optimization problem using gradient information. This enables the plant to reach the optimum operating point or batch trajectory without the need of a precise model which reduces the necessary modeling efforts. However, computing the gradients of the cost function or of the plant outputs with respect to the inputs online is a challenging task. Modifier Adaptation with Quadratic Approximation (MAWQA) integrates MA with Quadratic Approximation (QA), which helps mitigate the challenges of estimating gradients from noisy measurements by utilizing historical operating data. However, the distribution of these past operating points significantly affects the effectiveness of the MAWQA strategy. To address this issue in this contribution, new methods to compute probing points which lead to fast convergence to the optimum are proposed. The effectiveness of these methods is validated using the Williams-Otto reactor benchmark.

Optimal energy scheduling for sector-coupled multi-energy systems is becoming increasingly important as renewable energies such as wind and photovoltaics continue to expand. They are very volatile and difficult to predict. This creates a deviation between generation and demand that can be compensated for by energy storage technologies. For these, rule-based control is well established in industry, and mixed-integer model predictive control (MPC) is an area of research that promises the best results, usually regarding minimal costs. Drawbacks of MPC include the need for an adequate system model, often associated with high modeling effort, high computational effort for larger prediction horizons, and complications with stochastic variables. In this work, Reinforcement Learning is used in an attempt to overcome these difficulties without applying elaborate mixed-integer linear programming. The self-learning algorithm, which requires no explicit knowledge of the system behavior, can learn a control policy and uncertainties of the variables just by interaction with the (simulated) system model. It is demonstrated that Reinforcement Learning (exchange factor = 36.8 %) can learn complex system behavior with comparable quality to model predictive control (ex. = 32.4 %) and outperforms rule-based control (ex. = 41.8 %). This is done in a case study with the goal of minimizing the exchange of energy with the grid, with a battery and hydrogen system providing storage flexibility. These results were achieved in the context that the Reinforcement Learning agent only has instantaneous rather than predictive information, i.e., a very limited state of information compared to the MPC. The trained policy is then deployed while significantly decreasing the computational effort.

Predicting processes’ future behavior based on past data is vital for automatic control and dynamic optimization in engineering. Recent advances in deep learning, particularly Artificial Neural Networks, have improved predictions in various engineering fields. Recurrent Neural Networks (RNNs) are well-suited for time series data, as they naturally evolve through dynamic systems with recurrent updates. Despite their high predictive power, RNNs may underperform if their training ignores the model’s future application. In Model Predictive Control, for example, the model evolves over time using only current information, relying on its own predictions at later steps. A model trained for one-step-ahead predictions may fail when tasked with multi-step-ahead forecasting in autoregressive mode. This study explores deep recurrent neural network models for predicting critical operational time series of a large-scale Electric Submersible Pump system. We present an innovative training approach, framing the task as a multi-step-ahead prediction problem. Results show that aligning model training with its future use is crucial to ensure real-time performance.

Vacuum (pressure) swing adsorption (V(P)SA) has received considerable attention in the past decades. Existing studies typically estimate vacuum pump energy consumption using an approximate constant energy efficiency or an empirical energy efficiency correlation, leading to inaccurate representation of realistic vacuum pump performance. In this paper an enhanced computational approach is proposed for simulation and optimisation of V(P)SA through simultaneous integration of realistic vacuum pump data and adsorption bed fluidisation limits. The computational results show that the developed prediction models accurately represent the actual performance curves of the vacuum pump. Incorporation of the vacuum pump prediction models and fluidisation constraints in V(P)SA optimisation leads to significantly different optimal solutions compared to when these factors are not considered.

This study presents a deterministic optimization methodology for the design of shell-and-tube heat exchangers with twisted tape turbulence promoters, focusing on minimizing the total annualized cost (TAC) while balancing thermal performance and energy consumption. A sensitivity analysis was carried out as Case I (Methanol-Water), it reveals that increasing the twist ratio (TR) reduces flow turbulence, resulting in lower fluid velocity, pressure drop (?Pi), and overall heat transfer coefficient (U). Among the turbulence promoters evaluated, twisted tapes with V-cuts achieved a 21.1% increase in U with a 52.27% increase in pressure drop, demonstrating an optimal balance between thermal enhancement and energy cost. In contrast, promoters with circular rings and multiple perforations showed the highest U improvements (26.7% and 25.8%, respectively) but incurred significant pressure drops (93.5% and 97.9%). The optimization problem has been stated as a nonlinear programming (NLP) problem and solved with the CONOPT solver. Results from two case studies emphasize the importance of operating with similar mass flows, directing the smaller flow through the tubes to maximize promoter effectiveness and achieve balanced heat transfer and pressure drop. For Case II (Residue-Water), the TAC was estimated at 11,467 USD/year, with the equipment cost as the major contributor. Case III (Water-Water) demonstrated improved balance in convective coefficients due to lower thermal effectiveness and similar mass flow rates, highlighting the positive impact of turbulence promoters on performance. This work demonstrates the potential of turbulence promoters for enhancing heat exchangers while emphasizing the importance of careful design to optimize performance and cost.

Reaction processes require optimization to enhance key performance indicators (KPIs) such as yield, conversion, and selectivity. Techniques like Bayesian Optimization (BO), Model-Based Design of Experiments (MBDoE), and Goal-Oriented Optimal Experimental Design (GOOED) play pivotal roles in achieving these objectives. BO efficiently explores the design space to identify optimal conditions, while MBDoE maximizes the information gain by reducing kinetic model uncertainty. In contrast, GOOED focuses solely on maximizing the KPIs without considering the system uncertainty, identifying reactor conditions in the design space guaranteeing optimal performance. This study compares BO, MBDoE, and GOOED in optimizing methane oxidation in an automated flow reactor. Performance is assessed based on optimal methane conversion, reduced system uncertainty and minimal experimental efforts to achieve maximum conversion. BO quickly identifies high-conversion conditions, MBDoE minimizes experimental runs while providing insights into parameter sensitivities, and GOOED prioritizes conversion efficiency. The findings highlight trade-offs between convergence speed, robustness, and information gain, providing valuable insights for designing data-driven, physics-informed experiments..

The definition of strategies for operation of process networks is a key research focus in process systems engineering. This challenge is commonly formulated as a numerical constraint satisfaction problem, where most practical algorithms are limited to identifying inner approximations to the feasible operational envelope. Sampling-based approaches so far have only been developed for formulations that required coordinated operation of the units within the network. We propose a decomposition approach that enables decentralized operation for acyclic muti-unit processes by sampling. Our methodology leverages problem structure to decompose unit-wise and deploys surrogate models to couple the resultant subproblems. We demonstrate it on a serial, batch chemical reactor network. In future research, we will extend this framework to consider the presence of uncertain unit parameters robustly.

Hydrogen isotope separation is a critical component of the fusion fuel cycle, particularly for achieving the desired purity levels of deuterium and tritium while minimising tritium inventory. This study investigates the cryogenic distillation of hydrogen isotopes, with a focus on the effects of isotopic equilibrium reactions at reduced temperatures and different system configurations. A one-column architecture was analysed to evaluate the impact of feed and side stream equilibrator temperatures and flowrates on separation performance and tritium inventory. Additionally, a two-column architecture was studied, incorporating multiple isotopic equilibrators in interconnecting streams, to further reduce unwanted heteronuclear isotopologues and improve system efficiency. Comparative analysis of the proposed configurations highlights significant operational advantages of optimising equilibrator temperatures, including reduced tritium contamination and inventory. Results indicate that reducing equilibrator temperatures to 77.4 K not only improves separation efficiency but also enables reductions in the number of theoretical stages required, minimising tritium inventory. A case study of the ITER Isotope Separation System demonstrates that reducing equilibrator temperatures achieves comparable performance with fewer stages in key columns, validating the benefits of optimised equilibrator operation for practical application. These findings contribute to the development of more efficient, cost-effective designs for fusion fuel cycle isotope separation systems.

Implementing Model Predictive Control (MPC) for large-scale nonlinear systems is often computationally challenging due to the intensive online optimization required. To address this, various reduced-order linearization techniques have been developed. The Koopman operator linearizes a nonlinear system by mapping it into an infinite-dimensional space of observables, enabling the application of linear control strategies. While Artificial Neural Networks (ANNs) can approximate the Koopman operator in a data-driven manner, training these networks becomes computationally intensive for high-dimensional systems as the lifting into a higher-dimensional observable space significantly increases data size and complexity. In this work, we propose a technique, combining Proper Orthogonal Decomposition (POD) with an efficient ANN structure to reduce the training time of ANN for large order systems. By first applying POD, we obtain a low order projection of the system. Subsequently, we train the ANN with an efficient structure to approximate the Koopman operator, significantly decreasing training time without sacrificing accuracy.

The application of Real-time Optimization (RTO) to dynamic operations is challenging due to the complexity of the nonlinear problems involved, making it difficult to achieve robust solutions. The literature on RTO in Pressure Swing Adsorption (PSA) units relies on Model Predictive Control (MPC) and Economic Model Predictive Control (EMPC), which rely heavily on an accurate model representation of the industrial plant. Given the importance of PSA systems on multiple separation operations, establishing alternatives for control and optimization in real-time is in order. With that in mind, this work aimed to explore alternative model-free RTO techniques that depend on simple control elements, as is the case of Extremum Seeking Control (ESC).The chosen case study was Syngas Upgrading. Extremum Seeking Control successfully optimized the CO2 productivity in PSA units for syngas upgrading/H2 purification. The results demonstrate that ESC can be a valuable tool in optimizing and controlling PSA processes and does not require the unit to reach a Cyclic Steady State to adjust the operation.

Monitoring chemical composition is key in several industrial-scale chemical processes. However, traditional composition sensors usually convey drawbacks, including high costs, short lifetimes, and frequent calibration requirements. As an alternative, software (soft) sensors have gained attention in recent years due to their accuracy, ease of training, and potential of integrating widely known machine learning techniques. This study presents the methodology followed to train a soft sensor for hydrogen sulfide monitoring in the gas treatment section of an industrial facility in Italy. In particular, this methodology includes a novel approach for steady-state determination from historical plant data in the presence of several steady states and noise. Unfortunately, only four steady states were found in the plant data, which was insufficient for accurate soft sensor training. As an alternative, these steady states were used to develop and validate a rigorous Aspen HYSYS process simulation. The relative error of the simulation compared to the steady states found was around 1%. This simulation was used to extract additional steady state data, which was sufficient for soft sensor training. The soft sensor, trained following a Kriging Gaussian Process regression, showed a relative error with respect to the rigorous simulation of around 0.7%. The high accuracy of the sensor suggests it is a quick, low-cost, and accurate alternative to estimate the hydrogen sulfide composition and identify potential operating conditions that may lead to off-specification conditions. Future work will focus on field implementation and evaluation of the trained sensor.

This article aims to tutorial a few important extremum seeking control approaches that can be used for the model-free optimization of industrial processes in various fields. The application of several methods is illustrated with a simple case study related to the production of algal biomass in photobioreactors. Other methods and applications are briefly reviewed.

This paper introduces the Updated Absolute Expected Value Solution, U-AEEV, a heuristic for solving multi-stage stochastic programming (MSSP) problems with type 2 endogenous uncertainty. U-AEEV is an evolution of the Absolute Expected Value Solution, AEEV [1]. This paper aims to show how U-AEEV overcomes the drawbacks of AEEV and performs better than AEEV. To demonstrate the performance of U-AEEV, we solve 6 MSSP problems with type 2 endogenous uncertainty and compare the solutions and computational resource requirements.

Addressing challenges in process design and optimisation, especially with complex models and data uncertainties, requires effective tools for model development, selection, and identification. Techniques such as Model-based Design of Experiments (MBDoE) help support this task by screening and discriminating between models and, eventually, calibrating them. Open-source and user-friendly Python packages have implemented some model identification techniques. However, the need for a tool that can couple with various model simulators and account for the steps of model identification as well as physical constraints of systems in design of experiments remains unmet. In that light, we present the python package MIDDOE (Model-(based) Identification, Discrimination, and Design of Experiments) to address this gap. It integrates rival models screening, parameter estimation, uncertainty analysis, and MBDoE techniques, while adapting to various process constraints. These functionalities are demonstrated via an in-silico study for a semi-batch fermentation reactor model identification.

This work approaches a detailed characterization of the aspects inherent to the innovative paradigm of Water and Energy Integration Systems (WEIS). These consists in conceptual physical systems which consider all potential energy-using and water-using processes in a site, all potential recirculation of material and energy streams between these and the integration of several categories of state-of-the-art technologies. The WEIS have the ultimate aim to promote the sustainability character associated to existing installations (through the reduction of energy and water input and contaminants output). The specific characteristics of WEIS are compared to existing similar process integration methodologies and a set of performance indicators are determined, having as a basis two previous case-studies approached for the Engineering project of WEIS. The performed analysis in this work revealed that the innovative paradigm is able to constitute Engineering projects with associated sustainability promotion potential, with the possibility to be modelled with existing computational tools and not being associated to impediments in terms of the maturity and availability of constituting technologies.

In process systems engineering (PSE), obtaining accurate process models for optimization can be expensive and time-consuming. Black-box Bayesian Optimization (BO) with Gaussian process (GP) surrogates offers a promising approach. However, full black-box optimization neglects valuable prior knowledge, which could otherwise improve the optimization process. This work explores methods of integrating prior knowledge in the form of low-fidelity data into BO by evaluating these methods on synthetic multi-fidelity test functions. Our results highlight possibilities for improved convergence of the BO optimization. However, our work further highlights potential pitfalls of these multi-fidelity models, such as bias, convergence to local optima, and overfitting on low-fidelity data. Hence, leveraging low-fidelity data in multi-fidelity models can improve BO convergence, but there are instances where the algorithms are more susceptible to failure.

Job Safety Analysis (JSA) is critical for proactively identifying workplace hazards, assessing their potential consequences, and implementing effective control measures. However, traditional JSA methods can be inefficient and prone to errors, particularly in complex industrial environments. This paper introduces AutoJSA, a knowledge-enhanced framework that leverages large language models (LLMs) to automate and optimize the JSA process. We collected 73 high-quality JSA reports from a chemical engineering company and divided the JSA workflow into three key tasks: hazard identification, consequence identification, and control measure generation. Two approaches — fine-tuning and retrieval-augmented generation (RAG) — were employed on a base LLM (GLM-4-9B-Chat) to adapt it for these domain-specific tasks. Experimental results demonstrate that both fine-tuning and RAG significantly improve task performance relative to the unmodified model, with fine-tuning generally providing larger gains. While RAG offers the advantages of dynamic knowledge retrieval and simpler implementation, it yields more modest improvements compared to fine-tuning. Moreover, a direct combination of fine-tuning and RAG did not show additional benefits under the current approach. Overall, AutoJSA effectively addresses the limitations of traditional JSA by increasing the accuracy and efficiency of safety analysis, laying the groundwork for fully automated JSA report generation. The findings underscore the potential of advanced artificial intelligence and natural language processing techniques to enhance workplace safety management in complex and rapidly evolving industrial settings.

Mechanical compression heat pumps are indispensable to facilitate the transition from thermally driven processes to renewable energy by electrification, upgrading low-temperature waste heat to recycle it at a higher temperature level. However, the implementation of such heat pumps up to date encounters limitations, due to equipment limitations and a lack of tools for the design of process concepts for the application of high-temperature heat pumps. The optimal design of heat pumps relies heavily on the selection of an appropriate refrigerant, as the thermodynamic properties significantly affect the heat pump cycle design and performance. While existing methods are capable of identifying thermodynamically beneficial refrigerants, they do not directly account for practical constraints such as limitations on the compressor discharge temperature, compression ratio, and vacuum operation. The current study proposes a fast-screening approach for arbitrary heat pump applications, considering a large set of established refrigerants. The method automatically assesses the performance of the refrigerants for a specified set of heat sink and source, adjusting the heat pump design with an optional internal heat exchanger in case of necessary superheating prior to compression. The approach is illustrated for the evaluation of heat pump-assisted distillation processes.

Electricity and carbon trading towards carbon reduction are highly coupled. The research on joint trading is essential for helping companies identify optimal strategies and enabling policymakers to detect potential policy loopholes. This study presents a novel game theory optimization model involving both power generation companies (GenCos) and factories to explore optimal operation strategies under electricity-carbon joint trading. By fully capturing the operational characteristics of power generation units and the technical energy consumption of electricity-consuming enterprises, it describes the relationship between renewable energy, fossil fuels, electricity, and carbon emissions detailedly. Considering the correlation between production volume and price of the same product, the case actually encompasses three trading systems: electricity, carbon, and commodities. Transforming this nonlinear model into a mixed-integer linear form through piecewise linearization and discretization, this study examines, through a virtual case with three GenCos and four factories, the impact of various emission reduction targets, comparisons of different carbon allocation mechanisms, and the influence of allowing zero-emission companies into carbon trading. Results reveal that since consumers may cut production rather than implement low-carbon technologies to lower emissions, driving up product prices to maintain profits, high electricity, and carbon prices become unsustainable for GenCos due to reduced electricity demand. Moreover, while intensity mechanisms can incentivize production, overall system profits decrease, which is undesirable for policymakers. Lastly, under strict carbon reduction targets, zero-emission companies may transform the carbon market into a seller’s market by purchasing carbon to raise carbon prices, thereby reducing electricity prices and lowering their operating costs.

Bayesian optimization is known to be able to search for the optimal conditions based on a small number of experiments. However, these experiments are insufficient to understand the experimental condition space. In contrast, we report the development of an algorithm that combines a low-confounding definitive screening design with Bayesian optimization, allowing for rapid optimization and ensuring sufficient experiments to understand the experimental condition space with a low confounding.

This study presents an integrated approach combining the Analytic Hierarchy Process (AHP) with a Mixed-Integer Multi-Objective Linear Programming (MOMILP) model to evaluate sustainable power generation alternatives for Pamplona, Colombia. The MOMILP model includes solar, wind, biomass, and diesel technologies, aiming to minimize costs (net present value) and CO2 emissions while considering design, operational, and budget constraints. The AHP method evaluates multiple criteria such as social acceptance, job creation, technological maturity, and environmental impact. The results show that solar panels are prioritized, with small diesel plants added due to resource limitations. The most sustainable option is a hybrid system with 49% solar, 29% wind, 14% biomass and 8% diesel, generating a net present value of 121,360 USD and 94,720 kg of CO2 emissions. The proposed methodology can be applied to assess and select the most feasible alternative within a wide range of new projects for the integration of renewable energy systems for the extension of existing power grids.

Uncertainty plays a crucial role in strategic supply chain design. In this study, we explore robust approaches to model uncertainty when the non-deterministic parameters are placed in the independent term, on the right-hand side (RHS) of the constraints. We consider the “disjunctive adjustable column-wise robust optimization” (DACWRO), a disjunctive formulation introduced previously in our group, and compare it with the adjustable column-wise robust optimization (ACWRO) formulation, a specific technique for solving robust optimization problems when the original robust optimization approach may assume too-conservative results. Given that the proposed method is based on the generalized disjunctive programming (GDP) technique, it is a higher lever modelling approach that represents the discrete nature of the decision process. In addition, it provides alternative MILP representations that can be further tested and compared. The analysis assesses the computational performance and reformulation tightness of both approaches in the context of the forest supply chain design problem.

Process systems engineering optimization problems may be challenging. These problems often exhibit nonlinearity, non-convexity, discontinuity, and uncertainty, and often only the values of objective and constraint functions are accessible. Additionally, some problems may be computationally expensive. In such scenarios, black-box optimization methods may be appropriate to tackle such problems. A general-purpose multi-agent framework for optimization has been developed to automate the configuration and use of hybrid optimization, allowing for multiple optimization solvers, including different instances of the same solver. Solvers can share solutions, leading to better outcomes with the same computational effort. Alongside cooperation, competition is introduced by dynamically allocating more computational resource to solvers best suited to the problem. Each solver is assigned a priority that adapts to the evolution of the search. The scheduler is priority-based and uses similar algorithms to those in multi-tasking operating systems. The impact on the number of priority levels is investigated. The framework allows for the use of both metaheuristic and direct search methods. Metaheuristics explore the full search space while direct search methods are good at exploiting solutions. The framework has been implemented in Julia, making full use of multiprocessing. A multiobjective case study on the design of a micro-analytic system is presented. The case study demonstrates the benefits of a multi-solver hybrid optimization approach with both cooperation and competition. The framework adapts to the evolving requirements of the search. Often, a metaheuristic method is allocated more computational resource at the beginning of the search while direct search methods are prioritized later.

A key step in developing high-performance industrial products lies in the design of their constituent molecules. Computer-aided molecular design (CAMD) has garnered significant attention for its potential to accelerate and improve the design process. The mainstream method involves using property prediction models to predict the properties of potential molecules and selecting the best candidates based on these predictions. However, prediction errors are inevitable, introducing unreliability into the design. To address this issue, this paper proposes a novel component property modeling framework based on a molecular similarity coefficient. By calculating the similarity between a target molecule and those in an existing database, the framework selects the most similar molecules to form a tailored training dataset. The similarity coefficient also quantifies the reliability of the property predictions. In tests across various properties, this framework not only provides a quantifiable evaluation of prediction reliability but also improves the prediction accuracy of molecules with high reliability, which has the potential to enhance the integrity of molecular design.

Rapid and robust simulation of chemical production processes is critical to address core scientific questions related to process design, optimization, and sustainability. Efficiently solving a chemical process, however, remains a challenge due to their highly coupled and nonlinear nature. Graph abstractions of the underlying physical phenomena within unit operations may help identify potential avenues to systematically reformulate the network of equations and enable more robust convergence of flowsheets. To this end, we further refined a flowsheet graph-theoretic abstraction that consists of a mesh of interconnected variable nodes and equation nodes. The new network of equations is formulated at the phenomenological level—agnostic to the thermodynamic property package—by extending equation formulations widely used to solve multistage equilibrium columns. Decomposition of the graph by phenomena linearizes material and energy balances across the flowsheet by decoupling phenomenological nonlinearities (e.g., phase equilibrium, chemical reactions). Additionally, we further improved a preliminary simulation algorithm which employs this phenomena-based decomposition and demonstrated that it can perform more rapid and robust simulation of large, highly-coupled systems than sequential modular simulation.

The green transition accelerates innovations and developments targeting the integration of green hydrogen in the chemical industry. However, all new hydrogen pathways and process designs must be tested on operability and safety. A big challenge is the typical fluctuating characteristic of green hydrogen supply that contrasts the steady-state operation of most conventional chemical processes. Therefore, to adequately assess control and monitoring techniques, a benchmark model tailored to the relevant aspects of the hydrogen economy is required. We introduce a benchmark model based on the production of green ammonia using the Haber-Bosch process that remains operable when coupled to a fluctuating hydrogen supply from water electrolysis. The main section of the process model is an adiabatic indirect cooled reactor system that provides realistic modeling of industrial applications. Like the ammonia reactor, all process units and the underlying control structure are precisely dimensioned to ensure feasible operation across various hydrogen supply rates. Eventually, the flexible operating ammonia benchmark model can serve as a new benchmark for analyzing process safety and control aspects.

Graph neural networks (GNNs) have proven state-of-the-art performance in molecular property prediction tasks. However, a significant challenge with GNNs is the reliability of their predictions, particularly in critical domains where quantifying model confidence is essential. Therefore, assessing uncertainty in GNN predictions is crucial to improving their robustness. Existing uncertainty quantification methods, such as Deep ensembles and Monte Carlo Dropout, have been applied to GNNs with some success, but these methods are limited to approximate the full posterior distribution. In this work, we propose a novel approach for scalable uncertainty quantification in molecular property prediction using Stochastic Gradient Hamiltonian Monte Carlo (SGHMC). Additionally, we utilize a cyclical learning rate to facilitate sampling from multiple posterior modes which improves posterior exploration within a single training round. Moreover, we compare the proposed methods with Monte Carlo Dropout and Deep ensembles, focusing on error analysis, calibration, and sharpness, considering both epistemic and aleatoric uncertainties. Our experimental results demonstrate that the proposed parallel-SGHMC approach significantly outperforms Monte Carlo Dropout and Deep ensembles in terms of calibration and sharpness. Specifically, parallel-SGHMC reduces the sum of squared errors by 99.4% and 75%, respectively, when compared to Monte Carlo Dropout and Deep Ensembles. These findings suggest that parallel-SGHMC is a promising method for uncertainty quantification in GNN-based molecular property prediction.

Developing piping and instrumentation diagrams (P&IDs) is a fundamental task in process engineering. For designing complex installations, such as petroleum plants, multiple departments across several companies are involved in refining and updating these diagrams, creating significant challenges in data exchange between different software platforms from various vendors. The primary challenge in this context is interoperability, which refers to the seamless exchange and interpretation of information to collectively pursue shared objectives. To enhance the P&ID creation process, a unified, machine-readable data format for P&ID data is essential. A promising candidate is the Data Exchange in the Process Industry (DEXPI) standard. We present pyDEXPI, an open-source implementation of the DEXPI format for P&IDs in Python. pyDEXPI makes P&ID data more efficient to handle, more flexible, and more interoperable. We envision that, with further development, pyDEXPI will act as a central scientific computing library for the domain of digital process engineering, facilitating interoperability and the application of data analytics and generative artificial intelligence on P&IDs. We provide the pyDEXPI package with the documentation on GitHub at https://github.com/process-intelligence-research/pyDEXPI.

Surrogate models are widely used to improve the tractability of process optimization. Some commonly used surrogate models are obtained via machine learning such as multi-layer perceptrons (MLPs), Gaussian processes, and decision trees. Recently, a new class of machine learning models named Kolmogorov Arnold Networks (KANs) have been proposed. Broadly, KANs are similar to MLPs, yet they are based on the Kolmogorov representation theorem instead of the universal approximation theorem for the MLPs. Compared to MLPs, it was reported that KANs require significantly fewer parameters to approximate a given input/output relationship. One of the bottlenecks preventing the embedding of MLPs into optimization formulations is that MLPs with a high number of parameters (larger width or depth) are more challenging to globally optimize. We investigate whether the parameter efficiency of KANs relative to MLPs can be translated to computational benefits when embedding them into optimization problems and solving them to global optimality. We propose a Mixed-Integer Nonlinear Programming (MINLP) formulation of a KAN and test its effectiveness using a case study on the optimization of an auto-thermal reforming process. We observe that KANs are a promising alternative as surrogate models particularly for cases with low number of inputs or outputs. For surrogate models with higher inputs or outputs, stronger formulations must be developed to improve global optimization of models with KANs embedded.

Liquid-liquid extraction stages often deviate from equilibrium due to factors like insufficient mixing, making accurate efficiency modeling essential for process simulation. This study addresses the limitations of Aspen Plus (AP), which distorts equilibrium calculations by directly multiplying efficiency with the distribution coefficient. A modified Murphree efficiency definition, more suitable for liquid-liquid systems but absent in AP’s Extraction Column module, was implemented using Aspen Custom Modeler (ACM). The custom multi-stage extraction column model replaces mole fractions with mole flows to better represent mass transfer and phase interactions, enhancing simulation accuracy when imported into AP. Two test cases validated the custom model’s effectiveness. Test Case I, utilizing the UNIQ-RK thermodynamic model, compared the ACM model to AP’s built-in module, revealing that the ACM model provides a more realistic representation of extraction processes under varying stage efficiencies. It captured gradual efficiency changes and aligned with expected mass transfer behavior, while the built-in module showed reduced sensitivity and accuracy. Test Case II, using the UNIF-LL thermodynamic model, demonstrated the model’s robustness across different thermodynamic frameworks. It highlighted how varying Murphree efficiency from 0 to 1 directly affects separation performance, with higher efficiencies leading to better separation and increased extract phase composition. These findings confirm that the modified Murphree efficiency is a reliable and effective approach, enabling engineers to optimize liquid-liquid extraction processes with greater accuracy and efficiency.

Artificial Photosynthetic Bio-Electrochemical Systems (AP-BES) offer a promising approach for converting CO2 to valuable bioproducts, addressing carbon mitigation and sustainable production. This study employs a stochastic techno-economic assessment (TEA) to estimate the viability of rhodopsin driven AP-BES, from carbon capture to product purification. Unlike traditional deterministic TEAs, this approach uses Monte Carlo simulations to model uncertainties in key technoeconomic parameters, including energy consumption, CO2 conversion efficiency, and bioproduct market prices. The analysis generates probability distributions for economic metrics such as Operational Expenditure (OPEX), Capital Expenditure (CAPEX), and profit. Enhancements in light-harvesting efficiency and advancements in reactor materials were predicted to reduce the payback period to just one year, thereby making large-scale deployment a feasible option.

Industrial process monitoring can benefit from utilizing historical data, providing insights for decision-making and operational efficiency. This study develops a soft-sensor-based approach leveraging multi-fidelity modeling to correct discrepancies between online sensors and laboratory analyses. A Gaussian process-based strategy is used to predict deviations between high-frequency low-fidelity sensor data and less frequent high-fidelity laboratory measurements. By exploring static and dynamic modeling frameworks, we assess their suitability for capturing process dynamics and addressing time-dependent variability. The multi-fidelity soft sensor noticeably improves predictive accuracy, outperforming high-fidelity and low-fidelity methods. This approach demonstrates applicability across various industrial settings where integrating diverse data sources enhances real-time process control and monitoring, reducing reliance on costly laboratory sampling.

The performance of chemical plants is inevitably related to knowledge about the current state of the system. However, both process and equipment drifts may distort state information. Deviations of process values caused by equipment malfunction may be misinterpreted as process drifts and vice versa. Determining the cause of the drift is further complicated by the fact that equipment drifts typically occur in combination with process drifts. This paper presents a method that uses available additional equipment data to reliably detect and decouple combined equipment and process drifts in chemical plants by combining statistical methods with model-based approaches. The utility of additional equipment information is assessed based on its effect on the decoupling of process and equipment drifts. First results demonstrate the feasibility of the approach in a real plant.

In this work, we introduce a quantum-enhanced reinforcement learning (RL) framework for process design synthesis. RL-driven methods for generating process designs have gained momentum due to their ability to intelligently identify optimal configurations without requiring pre-defined superstructures or flowsheet configurations. This eliminates reliance on prior expert knowledge, offering a comprehensive and robust design strategy. However, navigating the vast combinatorial design space poses computational challenges. To address this, a novel approach integrating RL with quantum machine learning (QML) is proposed. QML leverages theoretical advantages over classical methods to accelerate searches in large spaces. Built upon our prior work, the approach begins with a maximum set of available unit operations, represented in a flowsheet structure using an input-output stream matrix as RL observations. A Deep Q-Network (DQN) algorithm trains a parameterized quantum circuit (PQC) in place of a classical neural network (NN). The design structures generated by the RL agent are optimized using the IDAES Process Systems Engineering Framework, with the optimization objectives used as rewards to RL (e.g., cost, productivity). This quantum-enhanced algorithm is performed on a hydrodealkylation process case study, showcasing its efficiency and improved performance in navigating complex design spaces.

New research on complex intensified distillation schemes has popularized the use of several commercial process simulation software. The interfaces between process simulation and optimization-oriented software have allowed the use of rigorous and robust models. This type of optimization is mentioned in the literature as “Black Box Optimization”, since successive evaluations exploits the information from the simulator without altering the model that represents the given process. Among process simulation software, Aspen Plus® has become popular due to their rigorous calculations, model customization, and results reliability. This work proposes a comparative study for Aspen Plus software and Microsoft Excel VBA®, Python® and MATLAB® interfaces. Five distillation schemes were analyzed: conventional column, reactive column, extractive column, column with side rectifier and a Petlyuk column. The optimization of the ?????? (Total Annual Cost) was carried out by a modified Simulated Annealing Algorithm (m-SAA). The evaluation criteria are the time per iteration (????) and ?????? values. The results indicate that the best option to carry out the optimization was by using the VBA interface, however the one carried out with Python did not differ radically (12%).

Bilevel optimization is an active area of research within the operations research community due to its ability to capture the interdependencies between two levels of decisions. This study introduces a metamodeling approach for addressing mixed-integer bilevel optimization problems, exploiting the approximation capabilities of neural networks. The proposed methodology employs neural network embeddings to approximate the optimal follower’s response, bypassing the inner optimization problem by parametrizing it with continuous leader’s decisions. The use of Rectified Linear Unit (ReLU) activations allows the forward pass of the neural network to be represented as a set of mixed-integer linear constraints. Thereby, the bilevel structure is simplified into a single-level optimization model. A case study based on a two-echelon supply chain demonstrates the effectiveness of the approach, with solutions comparable to traditional bilevel optimization methods. The results suggest that neural network embeddings offer a promising alternative for tackling complex bilevel problems even when discrete decisions are involved and unveil the trade-off between prediction accuracy and computational demands. This methodology provides a versatile and efficient framework for solving mixed-integer bilevel optimization problems across diverse domains.

Waste lubricant oil (WLO) is a hazardous residual that requires proper management, being WLO regeneration the preferred approach. However, regeneration is only viable if the WLO does not coagulate in the equipment. Otherwise, the process needs to be shut down for cleaning and maintenance. To mitigate this risk, a laboratory test is currently used to assess the WLO coagulation potential before it enters the process. This laboratory test is, however, time-consuming, presents several safety risks, and is subjective. To expedite decision-making, process analytics technology (PAT) and machine learning were used to develop a model to classify WLOs according to their coagulation potential. Three approaches were followed, spanning linear and non-linear models. The first approach (benchmark) uses partial least squares for discriminant analysis (PLS-DA) and interval PLS combined with standard chemometric preprocessing techniques (27 model variants). The second approach uses wavelet transforms to decompose the spectra and PLS-DA for feature selection (10 model variants). Finally, the third approach uses convolutional neural networks (CNN) to estimate the optimal filter for feature extraction (1 model variant). These models were trained on real industrial data. The results show that the three modelling approaches can attain high accuracy, with an average of 91%. The spectral filtering using wavelet transforms proved to be a viable option to reduce the number of models to explore compared to the benchmark approach. The CNN was also able to streamline the preprocessing selection by implicitly estimating the optimal filter. The models also provided physical insight into the WLO coagulation phenomenon.

Random signals are crucial in chemical and process engineering, where industrial plants generate big data that can be used for process understanding and decision-making. This makes it necessary to unveil the underlying probability histograms from these signals with a finite number of bins. However, the search for the optimal number of bins is still based on empirical optimisation and general rules of thumb. In this work, we introduce an alternative and general method to unveil probability histograms. Our method employs a novel variable-order QMOM, which adapts automatically based on the relevance of the information contained in the random data. The number of bins used to recover the underlying histogram is found to be proportional to the information entropy, where a search algorithm is developed that generates bins and assigns probabilities to them. The algorithm terminates when no more significant information is available for assignment to the newly created nodes, up to a user-defined threshold. In conclusion, our method is a universal histogram reconstruction technique that only requires enough numbers of moments to work. The method has been validated intensively using synthetic random signals and real-life problems.

In applications involving fine powders and small particles, the accuracy of numerical simulations, particularly those employing the Discrete Element Method (DEM) to predict granular material behavior, can be significantly affected by uncertainties in critical parameters. These uncertainties include the coefficients of restitution for particle-particle and particle-wall collisions, viscous damping coefficients, and other related factors. In this study, we use stochastic expansions based on point-collocation non-intrusive polynomial chaos to perform a sensitivity analysis of a fluidized bed system. We treat four key parameters as random variables; each assigned a specific probability distribution over a designated range. This uncertainty is propagated through high-fidelity Large Eddy Simulation (LES)-DEM simulations to statistically quantify its impact on the results. To effectively explore the four-dimensional parameter space, we analyze a comprehensive database comprising over 1,200 simulations. Notably, our findings reveal that variations in the particle and wall Coulomb friction coefficients have a more pronounced influence on streamwise particle velocity than variations in the particle and wall normal restitution coefficients.

Artificial neural networks are widely used as data-driven models for capturing complex, nonlinear systems. However, suboptimal training remains a significant challenge due to the nonlinearity of activation functions and the reliance on local solvers, which makes achieving global solutions difficult. One solution involves reformulating activation functions as piecewise linear approximations to convexify the problem, though this approach often requires substantial CPU time. This study demonstrates that a tailored branch-and-bound algorithm can effectively address these challenges by efficiently navigating the solution space using linear relaxations. The proposed method achieves minimal training error, offering a robust solution to the training bottleneck. Unlike traditional mixed-integer programming approaches, which often struggle to converge within reasonable CPU times, the SOSX algorithm shows superior scalability, with computational demand growing almost linearly rather than exponentially as problem size increases. Experiments on datasets with the same dimensionality confirm that the algorithm’s efficiency is not specific to a particular dataset, as it demonstrates consistent CPU time trends across multiple datasets.

In the chemical industry, soft sensors are deployed to reduce equipment cost or allow for a continuous measurement of process variables. Soft sensors monitor parameters not via physical sensors but infer them from other process variables. On the one hand, the precision of a soft sensors is affected by its architecture, the choice of parametric equations like balances and thermodynamic or kinetic dependencies in the soft sensor model. On the other hand, uncertainty that is inherent to the input variable values propagates through the soft sensor model and impacts the output uncertainty. The latter is affected by the configuration of physical sensors in the chemical process. This paper proposes an approach for the combined optimization of soft sensor architecture and physical sensor configuration. For this purpose, the method combines an automatic extraction of all possible soft sensor architectures from a set of system equations with an uncertainty-based evaluation of sensor configurations. Applied to an in-situ Continuous Stirred Tank Reactor (CSTR), the method is capable of automatically identifying the optimal combination of soft sensor model and physical sensor configuration. It is further discussed, how the explainability of the results can be improved via variance-based uncertainty decomposition, by contributing soft sensor prediction uncertainty to individual inputs or operations along the calculation path.

Process design deals with the problem of finding the best process set-up, subject to a set of constraints defining the design space (DSp). This selection is guided primarily by economic considerations. Flexibility may also play an important factor in process design, since it embodies “how far” from the design space’s bounds are the candidate optimal designs, which in some cases may lead to off-spec products. This work proposes a novel approach for flexibility assessment. In design problems where the design space is constrained by a set of affine bounds, flexibility may be expressed either as the minimum or the maximum distance with respect to the feasible (design) space bounds. For any point in the DSp, the minimum distance provides a good indicator on the minimum flexibility, as the direction that represents the highest risk of violating the constraints. An analogous conclusion can be drawn between the maximum distance and maximum flexibility. These distances can be computed exactly via geometrical approaches, enabling the calculation of a minimum-based and maximum-based flexibility metrics for all points in the DSp. This class of problems are in fact multiparametric programming problems as the goal is to obtain comprehensive flexibility maps, rather than investigating unique points in the DSp. In the case of minimum flexibility, their solutions comprise: (i) a set of critical regions defining a convex hull within the DSp (each associated with a unique nearest bound of the feasible space), (ii) the corresponding optimizer functions (projection at nearest bound), and (iii) objective functions (minimum distance). A novel framework is under development for this class of problems. It enables a new paradigm for flexibility assessment, which can be applied to design problems of any dimension. Complexity is significantly reduced in comparison with the classic multiparametric programming approaches, since only a limited number of active sets need to be considered during solution calculation.

Sustainability is increasingly recognized as a critical global issue. Multi-objective optimization is an important approach for sustainable decision-making, but problems with four or more objectives are hard to interpret due to its high dimensions. In our group’s previous work, an algorithm capable of systematically reducing objective dimensionality for (mixed integer) linear Problem has been developed. In this work, we will extend the algorithm to tackle nonlinear many-objective problems. An outer approximation-like method is employed to systematically replace nonlinear objectives and constraints. After converting the original nonlinear problem to linear one, previous linear algorithm can be applied to reduce the dimensionality. The benchmark DTLZ5(I, M) problem set is used to evaluate the effectiveness of this approach. Our algorithm demonstrates the ability to identify appropriate objective groupings on benchmark problems of up to 20 objectives when algorithm hyperparameters are appropriately chosen. We also conduct extensive testing on the hyperparameters to determine their optimal settings. Additionally, we analyze the computation time required for different components of the algorithm, ensuring efficiency and practical applicability.

A new method for efficient process development is the direct scale-up from laboratory scale to production scale using mechanistic models [1]. The integration of the Quality by Design approach into this scale-up concept may prove beneficial for a variety of product- and process-related aspects. This paper presents a workflow for the digital twin-supported direct scale-up of processes and process plants, which integrates elements of the Quality by Design methodology. To illustrate the concept, the workflow is implemented for the example of an esterification reaction in a stirred tank reactor. Finally, benefits of the implementation of Quality by Design in the direct scale-up using digital twins regarding the product quality and the process development are discussed as well as its limitations.

Heuristic decomposition schemes like moving horizon schemes are a common approach to approximately solve large-scale mixed-integer programs. The authors propose Bayesian optimization as a methodological approach to systematically tune parameters of decomposition schemes for mixed-integer programs. This paper discusses detailed results of three studies of the Bayesian optimization-based approach using hoist scheduling as a case study: Firstly, two objectives of the tuning problem are examined considering sequences of incumbent solutions found by the Bayesian optimization. Secondly, the Bayesian optimization is applied to a set of test instances of the hoist scheduling problem using four types of acquisition functions; they are compared with respect to the convergence of the tuning problem solutions. Thirdly, the scaling behaviour of the Bayesian optimization is studied with respect to the dimension of the space of tuning parameters. The results of the three studies show that the solutions found by the Bayesian optimization converge quickly in smaller and larger tuning parameter spaces using different types of acquisition functions.

One of the most successful continuous high-performance liquid chromatography (HPLC) processes for drug manufacturing is the Simulated Moving Bed (SMB). SMB is a multi-column, continuous, chromatographic process that can handle much higher throughputs than regular batch chromatographic processes. The process is initially transient, but eventually arrives at a cyclic steady state, which makes optimization very challenging, even more so when superstructure optimization is involved. To simplify the optimization problem, many researchers fixed the SMB structure, optimizing only the continuous variables, so they cannot be considered superstructure optimization. In this work, an SMB superstructure that can simultaneously optimize column structure and operation is proposed. The results showed that the superstructure proposed is reliable, and it is more efficient compared to current optimization approaches if the optimal column structure has to be identified.

Anomaly detection is a key technique for maintaining process suitability and safety; however, the quality of process data often deteriorates due to missing or noisy values caused by sensor malfunctions. Such data imperfections may obscure real faults. If anomaly detection models are too sensitive to such abnormal data, they may cause false positives resulting in unnecessary alarms, which may obstruct detection of true process faults. Thus, deterioration of the quality of process data may affect process performance and safety. We propose a new anomaly detection method that utilizes graph Laplacian regularization as a loss function considering data-specific temporal relationships. Graph Laplacian regularization is a mathematical tool used in image processing and denoising to smooth data. We assume that successive process data temporally close to each other have similar values and maintain temporal dependencies among variables. In this study, Laplacian regularization imposes significant penalties when the outputs of neighboring samples lose smoothness, under the assumption that neighboring samples keep similar relationships. Such temporal dependencies can be expressed as a graph structure and extracted with the Nearest Correlation (NC) method. To demonstrate the usefulness of the proposed anomaly detection method, we applied it to an anomaly detection problem in a vinyl acetate monomer (VAM) process. The results show that the model with graph Laplacian regularization achieved higher performance than without graph Laplacian regularization in some fault scenarios. It was confirmed that the proposed method is effective for anomaly detection.

In the domain of process systems engineering, developing software embedded with advanced computational methods is in great demand to enhance the kinetic comprehension and facilitate industrial applications. Polymer production, characterized by complex reaction mechanisms, represents a particularly intricate process industry. In this work, a scientific software is developed for polymerization modelling and simulation with insight on microscopic resolution. From a software architecture perspective, the software is built on a self-developed process modelling platform that allows flexible user customization. A specific design for polymer species with microscopic chain structure information is conducted. From an algorithm perspective, the software offers high-performance solution strategies for polymerization process modelling by utilizing advanced computation approaches. A Ziegler-Natta copolymerization is presented to demonstrate the software’s capability in capturing the microscopic structural information of polymers.

Second-order optimization algorithms that leverage the exact Hessian or its approximation have been proven to achieve a faster convergence rate than first-order methods. However, their applications on training deep neural networks models, partial differential equations-based optimization problems, and large-scale non-convex problems, are hindered due to high computational cost associated with the Hessian evaluation, Hessian inversion to find the search direction, and ensuring its positive-definiteness. Accordingly, we propose a new search direction based on an interval Hessian and incorporate it into a line-search framework. We apply our algorithm to a set of 210 problems and show that it converges to a local minimum for 70% of the problems. We also compare our algorithm with other approaches. We illustrate that our algorithm is competitive to other methods in finding a local minimum using a smaller number of O(n3) operations.

This study introduces a novel methodology that combines Gaussian process regression (GPR) with alpha shape design space reconstruction to visualize multi-dimensional design spaces. The proposed GPR surrogate approach incorporates a kernel optimization step, employing a greedy tree search strategy to identify the optimal combinatorial kernel from a selection of base kernels. This approach efficiently evaluates design spaces around specific points of interest, enabling alpha shape reconstruction. The methodology’s adaptability is demonstrated through its application to both lower-dimensional (2D and 3D) cases and more complex, higher-dimensional systems (up to 7D), showcasing its scalability and versatility. Its effectiveness is further validated by its ability to generate accurate surrogate models from limited data. Overall, this study presents a robust framework that leverages GPR surrogate modeling and alpha shape reconstruction to facilitate design space evaluation in complex, multidimensional engineering problems.

Recent advancements in material design have made adsorption a more energy-efficient alternative to traditional thermally driven separation processes. Accurate modelling of adsorption thermodynamics is crucial for designing and operating equilibrium-limited adsorption systems. High-quality open-source packages like PyIAST, PyGAPsare available for processing adsorption data in Python. They provide a robust set of features for processing and analysing isotherms. However, they have no support for automatic differentiation and are not targeted for performance. Langmuir.jl addresses these limitations by leveraging Julia’s composable and differentiable programming ecosystem. Langmuir.jl includes tools for processing adsorption thermodynamics data—loading data, fitting isotherms with most often used models, predictive multicomponent adsorption through Ideal Adsorption Solution Theory (IAST) — and, importantly, enabling accurate derivative calculations through Julia’s automatic differentiation libraries, without requiring extensive code adjustments. Furthermore, Langmuir.jl achieves two orders of magnitude faster IAST calculations compared to its Python counterpart, PyIAST.

Solvent selection is a critical decision-making process that balances economic, environmental, and societal factors. The vast chemical space makes evaluating all potential solvents impractical, necessitating pre-selection strategies to identify promising candidates. Predictive thermodynamic models, such as the UNIFAC model, are commonly used for this purpose. Recent advancements in deep learning have led to models like the Gibbs-Helmholtz Graph Neural Network (GH-GNN), which overall offers higher accuracy in predicting infinite dilution activity coefficients over a broader chemical space than UNIFAC. This study presents a systematic comparison of solvent pre-selection using GH-GNN and UNIFAC-IL in the context of liquid-liquid extraction. The original GH-GNN model is extended to simultaneously predict organic and ionic systems. This extended GH-GNN model predicts more than 92 % of the logarithmic IDACs with an absolute error of less than 0.3. By comparison, UNIFAC-based models only achieve such accuracy for less than 65 %. A case study is used involving the ionic liquid ethyl-3-methylimidazolium tetrafluoroborate ([EMIM][BF4]) and caprolactam, relevant for the solvolytic depolymerization of polyamide 6. Results indicate a significant correlation in solvent rankings across both methods, with a Spearman’s coefficient of 0.62, suggesting that deep learning-based models like GH-GNN are viable alternatives for solvent pre-selection. Additionally, chemical similarity metrics, such as Tanimoto similarity, can assess confidence in solvent rankings, allowing users to determine acceptable risk levels in predictions across a vast chemical space.

Rubber mixing (RM) is a vital batch process producing high-quality composites, which serve as input material for manufacturing different types of final products, such as tires. Due to its complexity, this process faces two main challenges regarding the final quality: i) lack of online measurement and ii) limited comprehension of the influence of the different factors involved in the process. While data-driven and machine learning (ML) based soft-sensing methods have been widely applied to address the first challenge, the second challenge, to the best of the author’s knowledge, has not yet been addressed in the rubber industry. This work presents a data-driven method for extracting knowledge and providing explainability in the quality prediction in RM processes. The method centers on an XGBoost model while leveraging high-dimensional data collected over extended time periods from one of Michelin’s complex mixing processes. First, a recursive feature elimination-based procedure is used for selecting relevant features, which reduces the number of input features used for building the ML model by 82% while improving its predictive performance by 17%. Secondly, SHapley Additive exPlanations (SHAP) techniques are employed to explain the ML model’s predictions through global and local analyses of feature interactions. The selected quality-related variables can be leveraged to improve process control and supervision. Finally, an uncertainty quantification (UQ) module, based on Split Conformal Prediction (SCP), is combined with the ML model, providing confidence intervals with 90% coverage and empirically verified theoretical guarantees. This module ensures prediction reliability and robustness in real applications.

Processes increasingly need to consider electricity markets, which shifts the traditional demand side management scope towards a more dynamic nature. Instead of only focusing on day-ahead energy trading, demand-side management scope should be broadened towards being able to support the power grid stability during frequency events. This paper studies an artificial example process, similar to the melt-shop process from the steel industry, highlighting the challenges and opportunities of an energy intensive process. We show the potential benefits of having a battery energy storage system on-site, as well as opportunities in lowering the electricity cost by participating in the bidding process of various ancillary products.

Optimization under uncertainty is inherent to many PSE applications ranging from process design to RTO. Reaching process true optima often involves learning from experimentation, but actual experiments involve a cost (economic, resources, time) and a budget limit usually exists. Finding the best trade-off on cumulative process performance and experimental cost over a finite budget is a Partially Observable Markov Decision Process (POMDP), known to be computationally intractable. This paper follows the nonmyopic Bayesian optimization (BO) approximation to POMDPs developed by the machine-learning community, that naturally enables the use of hybrid plant surrogate models formed by fundamental laws and Gaussian processes (GP). Although nonmyopic BO using GPs may look more tractable, evaluating multi-step decision trees to find the best first-stage candidate action to apply is still expensive with evolutionary or NLP optimizers. Hence, we propose modelling the value function of the first-stage decision also with a GP, whose data will correspond to virtual evaluations of second-stage decision trees build upon myopic rollouts. Thus, the nonmyopic initial decision can be efficiently optimized via BO and the virtually learned value function. Effectiveness of the approach is demonstrated in a wide benchmark with synthetically generated functions as well as to optimize small batch production with a chemical reactor.

The growing demand for sustainable alternatives to petroleum-based products drives the development of biomanufacturing using agriculture-based sugars. However, agricultural sugar production faces significant challenges due to limited production capacity and potential negative environmental impacts. This research examines chemical sugar synthesis as an alternative, assessing its environmental impact with conventional agricultural production methods through life cycle assessment. As formaldehyde serves as a primary substrate in chemical synthesis, four production cases were evaluated—comprising two pathways (conventional methods and CO2 capture and utilization (CCU) technologies), each implemented with either fossil fuels or renewable energy sources. The analysis revealed that semi-batch reactors in chemical synthesis substantially reduce environmental impacts compared to batch reactors. Chemical sugar synthesis demonstrated marked advantages in reducing eutrophication, land use change, and water resources consumption across all formaldehyde production methods evaluated. However, the formaldehyde production process was identified as the determining factor in the overall environmental profile. While chemical synthesis offers environmental advantages in several categories, implementing CCU technologies with renewable energy integration remains necessary to reduce climate change impacts and resource consumption. This work demonstrates the potential of optimized chemical sugar synthesis as an alternative to agricultural sugar production and provides direction for future process development.

This study presents a simple tool to provide decision-makers data that will facilitate informed decisions in selecting utilization for small- to medium-scale utilization of stranded natural gas resources that would otherwise be flared. The methodology involves the simulation of different natural gas utilization technologies on Aspen Plus simulation software and utilizing the results to develop a tool on python that enables the user to assess recoverable valuable products from different natural gas profiles. Ten utilization technologies were implemented and six different natural gas profiles (rich and lean) were used as case studies to ascertain the capabilities of the tool. The results provide the user with the Net Present Values (NPV) of different technologies and the most profitable or infeasible utilization technology. The results also show the potentials of utilizing the gas over flaring. For very small volumes of gas the results favored the compressed natural gas (CNG) with positive NPV, and next electricity production. Small to medium volumes of lean gas favoured the turquoise methanol. Mini- Liquefied Natural Gas (LNG) were the least profitable especially for small volumes. This tool can facilitate the evaluation of several gas profiles to assess the impact of utilizing natural gas over flaring in the sustainable future paradigm.

Batch chemical processes face significant challenges due to frequent operational shifts and varying conditions, requiring models to be retrained for each new scenario. This high retraining demand limits the scalability of traditional process monitoring systems, making them unsuitable for dynamic batch operations. To address this, we propose a transfer learning-based framework that enhances adaptability by reusing learned features across different batch conditions, reducing the need for extensive retraining. Proposed method integrates Temporal Convolutional Networks (TCNs) for capturing temporal dependencies in batch data and predicting Quality-Indicative Variables (QIVs) to identify deviations. The core innovation lies in transfer learning, enabling the model to adapt to new process variations with minimal updates. This approach ensures robust, accurate monitoring even under evolving conditions. This framework is validated using the IndPenSim penicillin fermentation dataset, which simulates real-world batch process variability. Results show that the transfer learning-enhanced model effectively predicts QIVs and detects deviations across varying control strategies, demonstrating improved adaptability and reduced retraining requirements. This study highlights the potential of transfer learning to revolutionize batch process monitoring by addressing core challenges in dynamic chemical operations.

Prospective life-cycle design of emerging technologies is important in discussions of decarbonization and resource circulation strategies. This study demonstrates the role of computer-aided process engineering in reflecting technology information with appropriate granularity and accuracy using air conditioning as a case study. Process simulations involving heat exchangers (indoor/outdoor units), compressors, and expansion valves were developed to model air conditioners to quantify changes in performance and heat exchanger size as existing and alternative refrigerants are introduced. The process simulation results were incorporated into a material flow analysis and life cycle assessment to quantify the change in life cycle greenhouse gas (GHG) emissions through 2050 for each refrigerant installed. The results show that operational emissions dominate the life cycle GHG emissions of air conditioners, that decarbonization of electricity can significantly reduce life cycle GHG emissions, with similar results regardless of which refrigerant is used, and that differences in design policies when alternative refrigerants are introduced strongly affect the breakdown of emissions in 2050.

Currently, state-of-the-art approaches to solving complex optimization problems have focused solely on methods requiring high computational time and unable to find the global optimal solution. In this work, a methodology based on quantum computing is presented to overcome these drawbacks. The novelty of this framework stems from the quantum computer’s architecture and taking into consideration the quantum phenomena that take place to solve optimization problems with specific structure. The proposed methodology includes steps for the transformation of the initial optimization problem into an unconstrainted optimization problem with binary variables and its embedding onto a quantum device. Moreover, different resolution levels for the transformation step and different architectures for the embedding process are utilized. To illustrate the procedure, a case study based on Haverly’s pooling and blending problem is examined while demonstrating the potential of the proposed approach. The results indicate that the succinct formulation exhibited higher success rate during the embedding procedure for the different examined architectures, and the quantum annealing solver exhibited the best performance among the various solvers investigated. This highlights the potential of the approach for solving this type of problems with the rapid development and improvement of quantum hardware and expanding it to more complex chemical engineering optimization systems.

Bayesian classification (BC) is a powerful supervised machine learning method for modelling the relationship between a set of continuous variables and a set of discrete variables that represent classes. BC has been successful in engineering and medical applications, including feasibility analysis and clinical diagnosis. Gaussian process (GP) models are widely used in BC methods to model the probability of assigning a class to an input point, typically through an indirect approach: a GP predicts a continuous function value based on Bayesian inference, which is then transformed into class probabilities using a nonlinear function like a sigmoid. The final class labels are assigned based on these probabilities. In this commonly used workflow, the uncertainty associated with the class prediction is usually evaluated as the uncertainty in the GP function values. A disadvantage of this approach is that it does not consider the uncertainty directly associated with the decision-making. In this work, we propagate the uncertainty from the space of GP function values to the class probability space and use this to quantify the uncertainty directly associated with the decision-making process. Additionally, we employ the propagated uncertainty as the objective function in an active learning (AL) method to generate new informative data points for the GP classifier training. We compare the proposed AL method to existing state-of-the-art methods to evaluate its performance.

Individual-based modeling (IbM) has emerged as a powerful approach for studying microbial populations, offering a bottom-up framework to simulate cellular behaviors and their interactions. Unlike continuum-based models, IbM explicitly captures the heterogeneity and emergent dynamics of microbial communities, making it invaluable for studying spatially structured phenomena such as nutrient competition, biofilm formation, and colony interactions. However, IbM faces significant computational challenges, particularly in resolving spatial overlaps during simulations of large microbial populations. Traditional approaches, such as pairwise comparisons or kd-trees, are computationally expensive and scale poorly with population size. The Discretized Overlap Resolution Algorithm (DORA) introduces a novel grid-based solution to overcome these limitations. By encoding spatial information into an occupancy matrix, DORA achieves a time complexity of O(N), enabling efficient resolution of overlaps while maintaining spatial fidelity. To validate its capabilities, we applied DORA to a case study of microbial colony merging. Two scenarios were tested: (1) colonies inoculated 40 µm apart merged at the center due to nutrient depletion, and (2) colonies inoculated 120 µm apart remained separate, with a no-growth zone forming between them. In both cases, DORA replicated observed phenomena with high accuracy and demonstrated linear scalability, significantly reducing simulation time compared to traditional methods. DORA’s computational efficiency and scalability position it as a robust tool for large-scale IbM simulations, advancing our ability to study complex microbial systems in diverse ecological and industrial contexts.

Pressure swing adsorption (PSA) is an energy-efficient technology for gas separation, while the multi-objective optimization of PSA is a challenging task. To tackle this, we propose a hybrid optimization framework, which integrates three steps. In the first step, we establish surrogate models for the constraints using Gaussian processes (GPs) and employ multi-objective Bayesian optimization to search for feasible points that satisfy the constraints. In the second step, we establish surrogate models for the objective function and constraints using GPs and utilize constrained multi-objective Bayesian optimization to search for an approximate Pareto front. In the third step, we perform a local search based on the approximate Pareto front. By employing the trust region filter method, we construct quadratic models for each constraint and objective function and refine the Pareto front to achieve local optimality. This framework demonstrates the efficiency of Bayesian optimization and the local optimality of the trust region method. A comparison with the popular evolutionary algorithm, Nondominated Sorting Genetic Algorithm II (NSGA-II), showed that this framework had a higher hypervolume of the Pareto front while halving the runtime and reducing the number of simulations by a factor of 20.

This work addresses the process design of Power-to-Ammonia plants (i.e. ammonia from renewable-powered electrolysis) by a novel methodology based on the multi-objective optimization of the “Three pillars of sustainability”: economic, environmental, and social. Specifically, we developed a tool estimating the installed capacities of every main process section typically featured by Power-to-Ammonia facilities (e.g., the renewable power plant, the electrolyzer, energy and hydrogen storage systems, etc.) to maximize the plant’s “Global Sustainability Score”.

Robust optimisation is a powerful approach for addressing uncertainty ensuring constraint satisfaction for all uncertain parameter realisations. While convex robust optimisation problems are effectively tackled using robust reformulations and cutting plane methods, extending these techniques to non-convex problems remains largely unexplored. In this work we propose a method that is based on a parallel robustness and optimality search. We introduce a novel spatial Branch-and-Bound algorithm integrated with robust cutting-planes for solving non-convex robust optimisation problems. The algorithm systematically incorporates global and robust optimisation techniques, leveraging McCormick relaxations. The proposed algorithm is evaluated on benchmark pooling problems with uncertain feed quality, demonstrating algorithm stability and solution robustness. The computational time for the examined case studies is within the same order of magnitude as state-of-the-art. The findings of this work highlight the potential of the proposed algorithm in tackling quadratic non-convex robust optimisation problems, particularly in chemical engineering applications.

Large Language Models have demonstrated potential in accelerating scientific discovery, but they face challenges when making inferences in rapidly evolving and niche domains like Process Systems Engineering (PSE). To address this, we propose a Graph-based Retrieval-Augmented Generation (RAG) pipeline specifically designed for PSE papers. Our pipeline includes custom document parsing, knowledge graph construction, and refinement to enhance retrieval accuracy. We evaluate the effectiveness of our approach using an automatically generated benchmark consisting entirely of PSE-related questions. The results show that our pipeline outperforms both non-RAG and vanilla RAG implementations in terms of relevant document retrieval and overall answer quality. Additionally, our implementation is fully customizable, allowing users to select the papers most relevant to their specific tasks. This framework is openly available, providing a flexible solution for those working in PSE or similar domains.

This paper presents a novel fault detection method based on key variable forecasting models. The approach integrates future forecasts of key variables into a time window, allowing for early fault detection without modifying the offline training phase of the existing fault detection model. By incorporating predicted data into the detection process, the proposed method significantly improves fault detection rates and reduces detection delays. Experiments using the Continuous Stirred Tank Heater (CSTH) system demonstrate the superiority of our method over traditional approaches, showing the advantages of forecasting in enhancing detection performance. However, our results also highlight the dependency of the method’s effectiveness on the quality of the forecasting model, suggesting the need for more advanced time-series forecasting techniques. Additionally, the current point forecasting method may not be sufficient in real-world applications, where probabilistic modeling of key variables would provide more reliable detection results. Future research directions will focus on refining forecasting models and exploring probabilistic approaches for more accurate fault detection.

The electrification of industrial heat is a promising pathway for decarbonization, yet challenges persist in balancing capital costs, operating costs, and emissions reduction. While previous studies have assessed electrification through heat integration and graphical methods, these approaches do not inherently determine the optimal hybrid technology configuration. This study introduces an optimization-based framework that systematically evaluates the cost-optimal allocation of electrified and conventional heating technologies. Formulated as a Mixed-Integer Linear Programming (MILP) model and implemented in Gurobi, the framework minimizes Total Annualized Cost (TAC) while satisfying heat demand, technology constraints, and emissions targets. Applied to an industrial case study, the model compares three scenarios: a fully conventional system relying on steam boilers and fired heaters, a fully electrified system utilizing high-temperature heat pumps, electrode boilers, and electric heaters, and an optimized hybrid system that strategically integrates both conventional and electrified technologies. The results demonstrate that the optimized hybrid system achieves a 75% electrification rate, reducing TAC by 7% compared to previous graphical electrification assessments. The model selects a combination of a high-temperature heat pump and a steam boiler, highlighting the importance of accounting for heat pump cost savings in economic evaluations. Additionally, a comparison with Carbon Capture and Storage (CCS) reveals that partial electrification results in a lower CO2 abatement cost of 131 $/tCO2, compared to 140 $/tCO2 for CCS, reinforcing the economic viability of industrial heat electrification. This study provides a systematic optimization framework for guiding industrial electrification decisions and emphasizes the role of carbon taxation, electricity pricing, and renewable energy incentives in shaping cost-effectiveness. The findings underscore the importance of integrating process optimization with techno-economic analysis to accelerate industrial decarbonization.

While bilevel optimization is gaining prominence across various domains, the field lacks standardized tools for generating test problems that can effectively evaluate and guide the development of efficient solution algorithms. We define the term “trivial bilevel optimization problem” as a bilevel problem whose high-point relaxation solution is also feasible. These easy-to-solve problems frequently arise in naïve implementations of random bilevel optimization problem generators, significantly impacting the evaluation of bilevel solution algorithms. However, this problem has not been addressed in the literature to our best knowledge. This work introduces a non-trivial mixed-integer bilevel optimization problem generator, NT-BMIPGen, and a problem library, designed to eliminate the generation of trivial bilevel problems. Through analysis of the bilevel problem structure, we identify key factors contributing to problem triviality. Particularly, the upper to lower variable ratios and the number of decoupled lower-level continuous variables increasing two-level competition significantly impact problem triviality. The proposed generator allows users to customize problem structures while ensuring non-triviality by checking the bilevel feasibility of highpoint relaxation. Additionally, we present a library of 105 non-trivial problems to support algorithm development and testing. This work advances the understanding of bilevel problem complexity and provides essential tools for the systematic evaluation of bilevel optimization algorithms.

Process synthesis is a fundamental step in process design. The aim is to determine the optimal configuration of unit operations and stream flows to enhance key performance metrics. Traditional methods provide just one optimal solution and are strongly dependent on user-defined technologies, stream connections, and initial guesses for unknown variables. Usually, a single solution is not sufficient for adequate decision-making, especially, when properties such as flexibility or reliability are considered in addition to the process economics. Wastewater Treatment network synthesis and design is a complex problem that demands innovative approaches in design, retrofits, and maintenance strategies. Considering this, an enhanced framework for improving reliability in wastewater transportation networks based on graph theory and machine learning is presented. Machine learning models were developed to predict failure probability, where the XGBoost model provided the best predictions. To select the appropriate solution, a trade-off between cost and reliability metrics is presented which is implemented by analyzing the results from the non-dominated solutions obtained for the case study demonstrated in this paper.

Detailed assessment of fuel production processes at an early stage of a project is crucial to identify potential technical challenges, optimize efficiency and minimize costs and environmental impact. While process simulations often are either very rigid and accurate or very flexible and unprecise, informed decision making can only be maintained by establishing a detailed process model as early as possible within the project lifecycle while keeping relevant aspects of the process flexible enough. In this work, we present the development of a framework based on a dynamic interface between AspenPlus® process simulations and Python, enabling enhanced flexibility and automation for process modeling and optimization. This integration leverages the powerful simulation capabilities of AspenPlus® with the versatility of Python for data analysis and optimization, delivering significant improvements in workflow efficiency and process control. By utilizing the dynamic simulation data exchange with Python, extensive parameter studies can be conducted and post-processed for techno-economic and environmental analyses. Furthermore, the interface allows the implementation of complex kinetic models or optimization routines for single process units. An additional extension for heat integration ensures the technical viability of the process route for reliable comparisons of different routes and process configurations. The functionalities are applied to a biomass- and power-based methanol production process including various process designs and operating conditions. To keep the level of detail at a high level, additional Python scripts are implemented securing a proper scaling of process units such as the methanol synthesis reactor system. The process configurations are assessed technically, economically and environmentally.

Artificial neural networks (ANNs) have gained popularity in the last years as tools to develop data-driven models of chemical process units. However, representing a system only with such artificial intelligence models may lead to a loss in the comprehension of the occurring phenomena. Hybrid models allow combining the predictive capabilities of ANNs with the foundational knowledge of rigorous models. This study explores the impact of hybridization in designing and optimizing shell-and-tube heat exchangers, comparing a full ANN-based model with a hybrid model. The hybrid model incorporates ANN predictions for highly nonlinear components, such as heat transfer coefficients, while other calculations are performed using the rigorous Bell-Delaware model. To generate the necessary data, the rigorous model is solved under randomly selected conditions. Using Python, one ANN predicts the exchanger’s cost, while another predicts the heat transfer coefficients. Both models are optimized using the differential evolution algorithm. The results indicate that the hybrid model produces designs with an approximately 67% lower standard deviation compared to the full ANN-based model when compared to the rigorous model’s cost predictions. This highlights the hybrid model’s ability to balance computational efficiency and accuracy, offering a promising approach for process design applications.

Chemical process models often involve high non-linearity due to thermodynamic and kinetic relationships, with non-convex bilinear terms adding complexity to process optimization. Recently, data-driven models, particularly artificial neural networks (ANNs), have gained traction for representing chemical processing units. The predictive accuracy of ANNs depends on data quality, variable interactions, and network architecture, the latter being an optimization challenge itself. This study proposes and evaluates two strategies to optimize ANN architecture for modeling a propylene glycol production process from glycerol. The process includes a reactor and two distillation columns, with training data generated through simulation in Aspen Plus by varying design and operating variables. Two approaches are compared: random ANN structure generation and architecture optimization using the ant colony algorithm, a method suitable for discrete problems. Decision variables include the number of hidden layers and neurons per layer, with the objective of minimizing mean squared error. Both methods yield ANN predictions with high accuracy (R² > 99.9%), closely matching simulation data. However, the ant colony algorithm achieves a better fit despite slower convergence, demonstrating its effectiveness in fine-tuning ANN architecture for chemical process modeling.

As the demand for resources continues to grow, pipelines have become critical for transporting water, fossil fuels, and chemicals. Monitoring pipeline systems is essential, as leaks can lead to severe environmental damage and safety hazards. This study aims to develop a pipeline leakage detection system based on digital twin technology and Physics-Informed Neural Networks (PINNs). By embedding physical principles, such as the continuity and momentum equations derived from the Navier-Stokes equation, into the neural network’s loss function, the model can predict pressure and flow dynamics with high accuracy while adhering to physical constraints. PINNs are particularly advantageous as they require minimal data, maintain physical consistency, and provide reliable interpretations, making them well-suited for addressing pipeline safety challenges. The model is designed to simulate fluid dynamics under normal operating conditions, with deviations in prediction errors signaling potential leaks. Statistical analysis of these errors is used to define control limits, establish rejection regions, and create control charts for leak detection. The detection system is further validated using field data to ensure reliability. By combining physical modeling and neural networks, this approach enhances the accuracy and interpretability of leakage detection, laying a solid foundation for an efficient pipeline monitoring solution.

A piping and instrumentation diagram (P&ID) is a central reference document in chemical process engineering. Currently, chemical engineers manually review P&IDs through visual inspection to find and rectify errors. However, engineering projects can involve hundreds to thousands of P&ID pages, creating a significant revision workload. This study proposes a rule-based method to support engineers with error detection and correction in P&IDs. The method is based on a graph representation of P&IDs, enabling automated error detection and correction, i.e., autocorrection, through rule graphs. We use our pyDEXPI Python package to generate P&ID graphs from DEXPI-standard P&IDs. In this study, we developed 33 rules based on chemical engineering knowledge and heuristics, with five selected rules demonstrated as examples. A case study on an illustrative P&ID validates the reliability and effectiveness of the rule-based autocorrection method in revising P&IDs.

Phase transition enthalpies, such as those for fusion, vaporization, and sublimation, are vital for understanding thermodynamic properties and aiding early-stage process design. However, measuring these properties is often time-consuming and costly, leading to increased interest in computational methods for fast and accurate predictions. Graph neural networks (GNNs), known for their ability to learn complex molecular representations, have emerged as state-of-the-art tools for predicting various thermophysical properties. Despite their success, GNNs do not inherently obey thermodynamic laws. In this study, we present a multitask GNN designed to predict vaporization, fusion, and sublimation enthalpies of organic compounds. We modified the loss function of the GNN, accounting for the thermodynamic cycle of the three phase transition enthalpies. To train the model, we digitized the extensive Chickos and Acree compendium, which encompasses 32,023 experimental measurements. Two approaches were explored: soft constraints, which guide the model toward thermodynamic consistency, and hard constraints, which enforce fully consistent predictions. The GNN achieved root mean squared errors (RMSEs) of 19.9 kJ/mol for sublimation, 11.0 kJ/mol for fusion, and 16.5 kJ/mol for vaporization enthalpies on the test set. Soft constraints were found to provide a good balance between accuracy and thermodynamic consistency, whereas hard constraints prioritized fidelity at the expense of predictive performance. When compared to the conventional Joback group contribution method the GNN demonstrated an improved accuracy and applicability range. This work underscores the potential of thermodynamics-informed GNNs for predicting thermodynamic properties accurately while maintaining consistency, paving the way for more reliable and efficient computational approaches.

This paper presents state-of-the-art object detection and object identification algorithms for digitizing and interpreting chemical engineering diagrams, including, Block Flow Diagrams (BFDs), Process Flow Diagrams (PFDs), and Piping and Instrumentation Diagrams (P&IDs), using computer vision techniques. These diagrams are essential for visualizing plant processes and equipment but are often stored as image-based PDFs, making manual digitization/interpretation labor-intensive and error-prone. The proposed algorithm automates tasks such as detecting unit operations and identifying them using a set of rule-based and predefined approaches including edge and contour-based rules, spatial arrangement rules, and geometric rules. This method avoids data requirements and computational requirements of deep learning approaches, offering a scalable and efficient solution for preliminary extraction of complex process information.

We propose a methodology that allows communication with Piping and Instrumentation Diagrams (P&IDs) using natural language. In particular, we represent P&IDs through the DEXPI data model as labeled property graphs and integrate them with Large Language Models (LLMs). The approach consists of three main parts: 1) P&IDs are cast into a graph representation from the DEXPI format using our pyDEXPI Python package. 2) A tool for generating P&ID knowledge graphs from pyDEXPI. 3) Integration of the P&ID knowledge graph to LLMs using graph-based retrieval augmented generation (graph-RAG). This approach allows users to communicate with P&IDs using natural language. It extends LLM’s ability to retrieve contextual data from P&IDs and mitigate hallucinations. Leveraging the LLM’s large corpus, the model is also able to interpret process information in P&IDs, which could help engineers in their daily tasks. In the future, this work will also open up opportunities in the context of other generative Artificial Intelligence (genAI) solutions on P&IDs, and AI-assisted HAZOP studies.

This study focuses on biomanufacturing case study, i.e. Lovastatin production, employing a hybrid modeling framework that combines mechanistic and data-driven approaches. A time-series dataset was generated using the KT-Biologics I (KTB1) plantwide model, a dynamic simulation of continuous biomanufacturing. The dataset captures critical parameters such as nutrient concentrations and API production. The AI-DARWIN framework was used to develop interpretable machine learning models with constrained functional forms, ensuring both accuracy and clarity. The resulting polynomial-based models reveal key relationships between process variables and system performance, bridging mechanistic insights with data-driven predictions. The models demonstrated reasonable accuracy showing minimal difference between the training and testing errors, highlighting their strong generalization. This work advances hybrid modeling in biomanufacturing by integrating plant-wide mechanistic simulations with interpretable machine learning. The approach ensures both accuracy and transparency while enabling robust process monitoring and control at a ‘plant-wide’ level, contributing to the broader adoption of hybrid modeling in biomanufacturing.

Many small and medium enterprises (SME) often fail to fully utilize the data they collect due to a lack of technical expertise. The ecoKI platform, a low-code solution that simplifies machine learning application for SMEs, showed a promising answer to the challenge. This study explores the application of ecoKI platform to design process monitoring tools for waste sorting plants. NIR separator data were processed through ecoKI’s building blocks to train two neural network architectures—MLP and LSTM—for predicting NIR separation efficiency. The results showed that the models accurately predicted NIR output and effectively identified regions where NIR separation performance declined, demonstrating the potential of data-driven approaches for real-time performance monitoring. This work highlights how SMEs can leverage existing data for operational efficiency and decision-making, offering an accessible solution for industries with limited machine learning expertise. The approach is adaptable to various industrial contexts, paving the way for future advancements in automated, data-driven optimization of equipment performance.

To address the limitations of traditional sensing instrumentation in industrial processes, this work explores the use of industrial text data. Given that current instrumentation often fails to capture the full scope of process-related information, text data resulting from operation of industrial settings (for example: maintenance, inspection and incident reports) can provide valuable insights. This study focuses on accessing the effectiveness of natural language processing (NLP) techniques in retrieving critical information from industrial text data. To achieve this, the classification of Process Safety and Containment Events (PSCE) was used as case study. Overall, we found NLP methods are effective in information retrieval from industrial text data. However, the integration of the embeddings into machine learning (ML) approaches poses some challenges. The complexity of the information encoded in the embeddings makes them too disparate and unique samples of a larger domain, making challenging the training of a ML model.

In an era defined by economic competitiveness and environmental awareness, engineering solutions must maximize profitability, efficiency and sustainability, underscoring the relevance of process optimization and the societal impact any contribution in this research field would bring. In chemical reactor engineering, optimization tasks pose significant challenges due to the highly non-linear and non-convex nature of reactor models, often involving differential equations. While conventional approaches have proven to be reliable strategies for solving these complex problems, their application becomes impractical as problem size and complexity increase. This work introduces a novel application of Physics-Informed Neural Networks (PINNs) to address constrained optimization problems in reactor engineering and demonstrates the proposed methodology through two illustrative case studies in chemical reactor design and control. In doing so, we highlight the capability of PINNs to efficiently learn optimal reactor patterns and analyse the strengths and limitations of this novel methodology. These findings pave the way for broader adoption of PINN-based techniques in process engineering optimization.

As large language models have shown remarkable capabilities in conversing via natural language, the question arises in which way LLMs could potentially assist chemical engineers in research and industry with domain-specific tasks. We generate dynamic chemical reactor models in Modelica code format from textual descriptions as user input. We fine-tune Llama 3.1 8B Instruct on synthetically generated Modelica code for different reactor scenarios. We compare the performance of our fine-tuned model to the baseline Llama 3.1 8B Instruct model as well as GPT4o. We manually assess the models’ predictions regarding the syntactic and semantic accuracy of the generated dynamic models. We find that considerable improvements are achieved by the fine-tuned model with respect to both the semantic and the syntactic accuracy of the Modelica models. However, the fine-tuned model lacks a satisfactory ability to generalize to unseen scenarios compared to GPT4o.

Fault diagnosis is critical for ensuring the safety and efficiency of chemical processes, as undetected faults can lead to catastrophic consequences. While deep learning-based methods have shown promise in this field, they often require manual hyperparameter tuning, which is not efficient since they heavily rely on expert knowledge and need iterative trial-and-error. This work introduces a novel approach combining a Multiscale Convolutional Neural Network (MSCNN) with Tree-Structured Parzen Estimator (TPE) for automated hyperparameter optimization to enhance the performance of fault diagnosis for chemical processes. The Multi-Scale Module is to capture complex nonlinear features from the fault data, while the TPE efficiently searches for optimal hyperparameters for MSCNN. An experimental study was carried out on the Tennessee Eastman Process (TE Process) dataset, where the proposed method was benchmarked against state-of-the-art models. The results indicate that the MSCNN-TPE method demonstrated improved performance in terms of precision and recall, achieving 5.26% and 5.63% higher values, respectively, compared to the CNN model. Comparisons of the MSCNN with default hyperparameters further confirmed the effectiveness of these techniques in improving fault diagnosis performance. Additionally, model ensemble technique was explored to further enhance the performance of the model and provide uncertainty estimations. In conclusion, this approach offers a robust and reliable solution for fault diagnosis in the chemical industry, enhancing process safety and efficiency.

Neural networks (NNs) are widely used as surrogate models but they do not guarantee physically consistent predictions thereby preventing adoption in various applications. We propose a method that can enforce NNs to satisfy physical laws that are nonlinear in nature such as enthalpy balances. Our approach, inspired by Picard’s successive approximations method, aims to enforce multiplicatively separable constraints by sequentially freezing and projecting a set of the participating variables. We demonstrate our Picard-KKT-hPINN for surrogate modeling of a catalytic packed bed reactor for methanol synthesis. Our results show that the method efficiently enforces nonlinear enthalpy and linear atomic balances at machine-level precision. Additionally, we show that enforcing conservation laws can improve accuracy in data-scarce conditions compared to vanilla multilayer perceptron.

Accurate mechanistic models provide valuable physical insight and are crucial for efficient process scale-up and optimisation, but their identification requires lengthy experimental data collection, model construction, validation and discrimination. Traditional black-box machine learning transfer methods leverage prior knowledge but lack interpretability and physical insights. To address this, we propose a novel approach using artificial neural network feature attribution to automatically locate corrections and symbolic regression to make structural modifications to an inaccurate or low-fidelity mechanistic model. In a comprehensive in-silico case study, the framework adapted a kinetic model from one biochemical system to a different but related one, enhancing predictive accuracy. Integrated within an iterative model-based design of experiments routine, it minimised the number of new experiments required. The study also discusses the impact of the inductive bias trade-off and alternative ways of incorporating prior knowledge to improve performance, highlighting the framework’s flexibility and potential for advancing automated knowledge discovery and high-fidelity predictive digital twin design for novel (bio)chemical processes.

This study tackles the challenge of accurate yield prediction in fluid catalytic cracking (FCC) units by comparing conventional supervised regression with time series forecasting methods using industrial data collected from the distributed control system (DCS) of an FCC plant. We introduce a shifted forecast paradigm that preserves temporal relationships between predictors and targets. Our preprocessing pipeline, which employs trimmed mean smoothing, addresses common industrial data challenges. Results demonstrate that the forecasting approach significantly outperforms supervised regression, achieving a mean absolute percentage error (MAPE) of 1.56% for 3-hour shifted predictions compared to 6.20% for supervised regression. The model maintains robust performance even with extended shifts during predictions, showing an MAPE of 3.55% for 14-day forecasts. This research provides valuable insights for implementing predictive analytics in industrial FCC operations, demonstrating the superiority of forecasting methods over traditional supervised regression approaches for process yield prediction.

Accurate yet interpretable prediction of Global Warming Potential (GWP) is essential for the sustainable design of chemical products and processes. However, existing studies that utilize molecular structure and physicochemical properties for GWP prediction often suffer from low interpretability, relying on black-box models that obscure the underlying relationships between molecular descriptors and environmental impact. To address this limitation, this study employs a Kolmogorov–Arnold Network (KAN) to derive symbolic equations that establish explicit relationships between molecular properties and GWP. By extracting interpretable mathematical expressions, our approach provides a transparent foundation for decision-making in chemical processes and reaction development. Our comparative analysis of machine learning models—including Random Forest, XGBoost, Deep Neural Networks (DNN), and KAN—reveals that Mordred descriptors outperform MACCS keys in GWP prediction, emphasizing the importance of physicochemical properties. The proposed KAN model achieves predictive accuracy comparable to conventional deep learning methods while maintaining interpretability, facilitating data-driven and transparent sustainability assessments in the chemical industry.

The rapid advancement of computational drug discovery has enabled the generation of vast virtual libraries of promising drug candidates. However, evaluating the synthetic accessibility (SA) of these compounds remains a critical bottleneck. While computer-aided synthesis planning (CASP) tools can provide synthesis routes to the candidate, their computational demands make them impractical for large-scale screening. Existing rapid SA scoring methods, struggle to generalize to out-of-distribution molecules and do not account for economic viability. To address these challenges, we present MolPrice, an accurate and reliable price prediction tool. By introducing a novel self-supervised learning approach, MolPrice achieves robust generalization to diverse molecular structures of various complexities. Our comprehensive analysis of model architectures and molecular representations reveals that substructure-based features strongly correlate with market prices, supporting the relationship between synthetic complexity and economic value. MolPrice performs well on the standard literature SA benchmark, showcasing its ability for SA estimation. MolPrice thus serves as both an accurate molecular price predictor and a rapid synthetic accessibility assessment tool, enhancing the efficiency of modern drug discovery pipelines.

The increasing complexity of industrial biotechnology demands advanced modeling techniques capable of capturing the intricate dynamics of bioreactors. Traditional regression-based and empirical methods often fall short when confronted with the highly nonlinear behavior and limited experimental data characteristic of bioprocesses. Addressing these challenges requires a more intelligent approach—one that leverages domain knowledge to model complex bioprocess dynamics effectively, even with sparse data, while maintaining interpretability and robustness. In this study, we introduce a process-informed, data-driven methodology for modeling the dynamics of industrial bioreactors, leveraging the capabilities of the rising field of Scientific Machine Learning (SciML). Our approach leverages Physics-Informed Neural Networks (PINNs) to seamlessly integrate domain knowledge encoded in physical laws with sparse experimental data and deep learning techniques, enabling precise simulation and modeling of bioreactor operations. The framework is validated using real-world experimental data, demonstrating its capability for state and space estimation, essential for optimizing biomanufacturing processes. Our findings underscore the potential of Physics Informed Neural Networks to address the challenges of dynamic modeling in bioprocesses, paving the way for the development of intelligent, autonomous, and data-driven solutions in industrial biotechnology. This fusion of data-driven and mechanistic modeling represents a transformative step toward enhancing the efficiency, scalability, and sustainability of modern biomanufacturing practices.

The industrialization of catalytic processes requires reliable kinetic models for design, optimization, and control. While white box models are preferred for their interpretability, they demand considerable time and expertise for their construction. This research enhances the ADoK-S framework by embedding prior expert knowledge using mathematical constraints and integrating uncertainty quantification. The improved methodology consists of: (I) a genetic programming algorithm with constraints to produce physically coherent candidate models, (II) a sequential optimization algorithm for parameter estimation, (III) model selection based on the Akaike information criterion (AIC), and (IV) uncertainty quantification of the chosen model’s predictions. The refined approach not only requires less data for discovering kinetic models but also ensures physically sound proposals. With the inclusion of uncertainty quantification, the method bolsters prediction reliability, and aids in safer system developments – crucial for decision-making and risk management. These improvements enhance data efficiency, model reliability, and position automated knowledge discovery as a real alternative to traditional kinetic modeling techniques.

Understanding data with complex patterns is a significant part of the journey toward accurate data prediction and interpretation. The relationships between input and output variables can unlock diverse advancement opportunities across various processes. However, most AI models attempting to uncover these patterns are not explainable or remain opaque, offering little interpretation. This paper explores an approach in explainable AI by introducing a multi-agent system (MaSR) for extracting equations between features using data. We developed a novel approach to perform symbolic regression by discovering mathematical functions using a multi-agent system of LLMs. This system addresses the traditional challenges of genetic optimization, such as random seed generation, complexity, and the explainability of the final equation. We utilize the in-context learning capabilities of LLMs trained on vast amounts of data to generate accurate equations more quickly. This study presents research on expanding the reasoning capacities of large language models alongside their mathematical understanding. The paper serves as a benchmark in understanding the capabilities of LLMs in mathematical reasoning and can be a starting point for solving numerous complex tasks using LLMs. The MaSR framework can be applied in various areas where the reasoning capabilities of LLMs are tested for complex and sequential tasks. MaSR can explain the predictions of black-box models, develop data-driven models, identify complex relationships within the data, assist in feature engineering and feature selection, and generate synthetic data equations to address data scarcity, which are explored as further directions for future research in this paper.

Process engineers in the Chemical and Oil & Gas industries extensively use process simulation for the design, development, analysis, and optimization of complex systems. This study investigates the integration of Artificial Intelligence (AI) with AVEVATM Process Simulation (APS), a next-generation commercial simulation tool. We propose a framework for a custom chatbot application designed to assist engineers in developing and troubleshooting simulations. This chatbot application utilizes a custom-trained model to transform engineer prompts into standardized queries, facilitating access to essential information from APS. The chatbot extracts critical data regarding solvers and thermodynamic models directly from APS to help engineers develop and troubleshoot process simulations. Furthermore, we compare the performance of our custom model against OpenAI technology. Our findings indicate that this integration significantly enhances the usability of process simulation tools, promoting more innovative and cost-effective engineering solutions.

Hybrid modelling is widely employed in chemical engineering to generate highly accurate predictions. Such an approach merges first-principle modelling with machine learning techniques to identify and model the epistemic uncertainty from experimental data. Despite its advantages, this still requires cross-domain competencies that are difficult to find in the chemical industry and high human involvement. The possibility of automating the identification and training model would be significantly beneficial for the widespread adoption of hybrid modelling methodology within the chemical industry. This work presents a novel algorithm for the automatic identification of hybrid models (HMs) starting from the first-principle representation of the system, described by differential equation sets. The methodology formulates the problem as mixed-integer programming, identifying the equation running under uncertainty, identifying the machine learning model hyperparameters, and training the latter. The Differential Evolution algorithm drives the identification and training tasks. The methodology is validated in three cases, namely a dynamic reaction system, a dynamic bioreactor and a Lotka-Volterra oscillator deviated with polynomial or MRF equation on different levels, generating 14 validation cases. On all of them, the model correctly identifies the position of the uncertainty and the functional form to approximate it. The methodology returns automatically trained HMs with a mean absolute percentage error in the range of 10%, which is in line with the experimental error of the data. The methodology presented in this work presents a step toward the automatic generation of HMs for dynamic systems and the widespread of this technology in the chemical industry.

Modeling process systems by use of partial differential equations is often complex and computationally expensive, especially for inverse problems such as optimization, state identification, or parameter estimation. Data-driven methods typically provide efficient alternatives with lower computational cost. One such method is Graph Neural Networks (GNNs), which can be used to model dynamical systems as graphs. However, dynamic GNNs often face challenges with extrapolation and representability. Integrating mechanistic insights in surrogate models can improve both prediction accuracy and interpretability. This study compares different strategies for embedding physics-based insights into GNNs to model the dynamic behavior of a catalytic CO2 methanation reactor. The hybrid integration of physics-informed GNNs aims to improve the predictive ability and interpretability while reducing the model development time, thereby facilitating faster deployment. Results show that penalizing the predicted change in reactor states from one time step to the next during training reduces the Frobenius error norm of GNN predictions from about 2.2 % to 1.4 %. In contrast, using a traditional physics-based loss function during training results in a Frobenius norm as high as 8.7 %. This highlights the importance of comparing different strategies for embedding physical laws into data-driven models.

The analysis of large chemical reaction space sheds light on reaction patterns between molecules and can inform subsequent reaction pathway planning. With the aim to discover more sustainable production systems, it became worthwhile to explicitly model the reaction space reachable from biobased feedstocks. In particular, the space that reactions in integrated biorefineries span for optimised biorefinery planning is of interest. In this work we show a network-theoretical analysis of biorefinery reaction data. We utilise the directed all-to-all mapping between reactants and products to compare the reaction space obtained from biorefineries with the entire network of organic chemistry (NOC). In our results, we find that despite having 1000 times fewer molecules, the constructed network resembles the NOC in terms of its scale-free nature and shares similarities regarding its small-world property. Additionally, we analyse the coverage rate of the biorefinery reaction data and find that many common second and third-level intermediates are present, while the querying strategy leads to a lack of data on value-added refinery end-products and on feedstock molecules, requiring more advanced and combined data querying strategies in the future.

Photochemical monolith reactors offer advantages over microreactors by providing high mixing efficiency and surface area to volume ratio while being scalable. However, optimizing monolith design parameters like channel number, shape, and stacking is critical to maximizing light usage and reactor efficiency. This work proposes using Bayesian optimization and COMSOL Multiphysics simulations to automatically design translucent monoliths for photochemical reactions. The goal is to maximize both photochemical space-time yield and space-time yield. Ray tracing simulations were performed while evaluating five different channel geometries (circular, elliptical, triangular, square, and pentagonal) and optimizing parameters, including channel diameter, vertical stacking, shape rotation, and ellipse axis ratio. Results showed a clear trade-off between Space-Time Yield (STY) and Photochemical Space-Time Yield (PSTY), with optimized elliptical channels achieving up to 15.3% improvement in STY with minimal PSTY increase and 9.0% enhancement in PSTY with negligible STY decrease.

Piping and Instrumentation Diagrams (P&IDs) play a crucial role in the chemical industry, yet they are often stored as scanned images or computer-aided design (CAD) drawings, limiting their seamless integration into modern digital workflows. Consequently, the task of automating P&ID digitization has attracted significant attention from the computer-aided design (CAD) research community. Traditional approaches, which typically rely on conventional object detection techniques, often demand extensive manual annotations to accurately identify and classify the various symbols in P&IDs. In this paper, we proposed a novel framework for automating the recognition of P&IDs. Our method first extracts key features of geometric primitives in CAD drawings through an automated process. Subsequently, a Transformer-based model is employed to predict the layer assignment of these primitives. Following this, an unsupervised clustering method, guided by predefined rules and geometric distances, is applied within each layer type to facilitate the automatic identification of devices, pipes, instrumentation, valves, insulation, and utilities in P&IDs. Extensive evaluation on an annotated dataset comprising 336 authentic P&IDs demonstrates that the proposed semantic annotation method achieves an overall accuracy exceeding 92%. The results of the unsupervised clustering are visualized, demonstrating promising performance and supporting the effectiveness of the proposed method. This study provides valuable insights into the intelligent digitization of P&IDs and lays the groundwork for further advancements in intelligent chemical process design.

Maintenance is critical for industrial plants to ensure operational reliability and worker safety. In process industries, fouling, the accumulation of solid residues in equipment, poses a significant challenge, causing inefficiencies and productivity losses. Effective modeling of fouling evolution over time is essential for maintenance planning to prevent equipment from operating under suboptimal conditions. Traditional approaches to fouling prediction include equation-based models, which offer high precision but may struggle with continuously changing process boundaries, and machine learning techniques, which are more adaptable but less effective at capturing rapidly evolving trends driven by complex underlying physics. This study introduces an innovative hybrid machine learning approach for predictive maintenance, combining the strengths of both methods. Pressure differential is modeled using an equation-based approach that links pressure data with fouling thickness, while the fouling growth rate is estimated through Gaussian Process Regression (GPR). This hybrid model generates pressure drop profiles with vertical asymptotes that closely mimic real-world behavior. Applied to a vacuum distillation column processing fouling-prone used oil, the proposed predictive method achieves a significant reduction of over 60% in suboptimal operating time compared to the current maintenance strategy, based on a fixed threshold.

The accumulation of heavy polynuclear aromatics (HPNAs) in hydrocracking units (HCUs) poses significant challenges to catalyst performance and process efficiency. This study proposes an integrated machine learning framework that combines ridge regression, K-means, and long short-term memory (LSTM) neural networks to predict HPNA formation, enabling proactive process management. For the training phase, weighted average bed temperature (WABT), catalyst deactivation phase—clustered using unsupervised K-means clustering—and hydrocracker feed (HCU feed) parameters obtained from laboratory analyses are utilized to capture the complex nonlinear relationships influencing HPNA formation. In the simulation phase, forecasted WABT values are generated using a ridge regression model, and future HCU feed changes are derived from planned crude oil blend data provided by the planning department. These forecasted WABT values, predicted catalyst deactivation phases, and anticipated HCU feed parameters serve as inputs to the LSTM model for predicting future HPNA levels. This approach allows us to simulate various operational scenarios and assess their impact on HPNA accumulation before they manifest in the actual process. By identifying critical process parameters and their influence on HPNA formation, the model enhances process engineers’ understanding of the hydrocracking operation. The ability to predict HPNA levels in advance empowers engineers to implement corrective actions proactively, such as adjusting feed compositions or operating conditions, thereby mitigating HPNA formation and extending catalyst life. The integrated framework demonstrates high predictive accuracy and robustness, underscoring its potential as a valuable tool for optimizing HCU operations through advanced predictive analytics and informed decision-making.

This work presents a model for optimizing shell-and-tube heat exchanger design using Q-learning, a reinforcement learning technique. An agent is trained to interact with a simulated environment of a heat exchange model, iteratively refining design configurations to maximize a reward function. This reward function balances heat exchanger effectiveness and pressure drop, emphasizing designs that minimize pressure drop. Results showed that simpler configurations consistently achieved higher rewards, despite complex designs offering better heat transfer efficiency.

Operational decision-making in Process Systems Engineering (PSE) has achieved high proficiency at specific levels, such as supply chain optimization and unit-operation optimization. However, a critical challenge remains: integrating these layers of optimization into a cohesive, hierarchical decision-making framework that enables sustainable and automated operations. Addressing this challenge requires systems capable of coordinating multi-level decisions while maintaining interpretability and adaptability. Multi-agent frameworks based on Large Language Models (LLMs) have demonstrated significant promise in other domains, successfully simulating traditional human decision-making tasks and tackling complex, multi-stage problems. This paper explores their potential application within operational decision-making for PSE, focusing on sustainability-driven objectives. A realistic Gas-Oil Separation Plant (GOSP) network is used as a case study, mimicking a hierarchical workflow that spans from initial back-of-the-envelope multi-objective optimization for cost-emissions trade-offs to a negotiation phase reflecting upper management decision-making, and culminating in high-fidelity simulations to validate operational setpoints at the plant level. This workflow serves as a canvas to assess the benefits of multi-agent LLMs, including their ability to integrate multi-layered decisions, enhance the explainability of strategies, and streamline automation in PSE workflows. The results demonstrate the potential of multi-agent LLMs to address the integration challenge in PSE, supporting sustainable and efficient operational decisions. Beyond GOSPs, this research highlights promising applications of multi-agent LLMs across process engineering, contributing to the vision of hierarchical, automated decision-making for the ‘plant of the future,’ where diverse models and tools operate within an intelligent, unified framework.

Balancing economic and environmental goals in industrial applications is critical amid challenges like climate change. Multi-objective optimization (MOO) and multi-criteria decision analysis (MCDA) are key tools for addressing conflicting objectives. MOO generates viable solutions, while MCDA selects the optimal option based on key performance indicators such as profitability, environmental impact, safety, and efficiency. However, large datasets pose a challenge in selecting the preferred solution during the MCDA process This study introduces a novel machine learning-enhanced MCDA framework and applies the method to analyze decarbonization solutions for a European refinery. A stage-wise dimensionality reduction method, combining AutoEncoders and Principal Component Analysis (PCA), is applied to simplify high-dimensional datasets while preserving key spatial features. Geometric analysis techniques, including Intrinsic Shape Signatures (ISS), are employed to refine the identification of typical configurations for baseline evaluations. Once typical configurations are identified, Large Language Models (LLMs) are utilized to enhance decision-making by providing contextual problem explanations and generating weight proposals for the weighted sum method, ensuring alignment with decision criteria. This framework is designed to support stakeholders in making informed, transparent decisions in complex, uncertain environments.

Physics-Informed Symbolic Regression (PISR) offers a pathway to discover human-interpretable solutions to partial differential equations (PDEs). This work investigates three fitness metrics within a PISR framework: PDE fitness, Bayesian Information Criterion (BIC), and a fitness metric proportional to the probability of a model given the data. Through experiments with Laplace’s equation, Burgers’ equation, and a nonlinear wave equation, we demonstrate that incorporating information theoretic criteria like BIC can yield higher fidelity models while maintaining interpretability. Our results show that BIC-based PISR achieved the best performance, identifying an exact solution to Laplace’s equation and finding solutions with R2-values of 0.998 for Burgers’ equation and 0.957 for the nonlinear wave equation. The inclusion of the Bayes D-optimality criterion in estimating model probability strongly constrained solution complexity, limiting models to 3-4 parameters and reducing accuracy. These findings suggest that a two-stage approach-using simpler complexity metrics during initial solution discovery followed by a post-hoc identifiability analysis may be optimal for discovering interpretable and mathematically identifiable PDE solutions.

As industries embrace the digitalization of Industry 4.0, the abundance of process data creates new opportunities to optimize industrial control systems. Traditional Proportional-Integral-Derivative (PID) controllers often require manual tuning to address changing conditions. This paper introduces an automated, adaptive PID tuning method using historical data and machine learning for a continuously evolving, data-driven approach. The method centers on training a surrogate model using historical process data to replicate real system behavior under various conditions. This enables safe exploration of control strategies without disrupting live operations. An RL (Reinforcement Learning) agent interacts with the surrogate model to learn optimal control policies, dynamically responding to the plant’s state, defined by variables like operational conditions and measured disturbances. The agent adjusts PID parameters in real-time, optimizing metrics such as stability, response time, and energy efficiency. After training, the RL agent is deployed online to monitor and adjust PID controllers in response to real-time deviations. The system continuously integrates new data to refine the surrogate model and RL agent, ensuring adaptability to long-term process changes. This continuous learning enhances resilience and scalability, maintaining optimal performance in dynamic environments. By combining data-driven modeling with RL, this method automates PID tuning, maximizing process data utility and aligning with Industry 4.0 principles. It reduces manual oversight while improving efficiency, reliability, and sustainability, addressing the challenges of increasingly complex and data-rich industrial systems.

Compartmental models are widely used to simplify the analysis of complex fluid dynamics systems, yet subjective compartment definitions and computational constraints often limit their applicability. The CompArt algorithm introduces an AI-driven framework that automates compartmentalization in Computational Fluid Dynamics (CFD) simulations, optimizing both accuracy and efficiency. By leveraging unsupervised clustering techniques such as Agglomerative Clustering, CompArt identifies coherent flow regions based on velocity and turbulent kinetic energy dissipation rate, ensuring a data-driven, physically consistent segmentation. The methodology integrates a connectivity-based clustering strategy, where compartments are dynamically optimized using the Silhouette score and adjacency matrix. This approach enables the reduction of high-resolution 3D CFD simulations into a network of interconnected sub-systems, significantly lowering computational costs while preserving system heterogeneity. The framework was tested on turbulent stirred tank simulations at 200 and 400 RPM, demonstrating its ability to capture both large-scale flow structures and fine-scale turbulence features. At higher RPM, increased mixing leads to a more fragmented dissipation pattern, which the model successfully adapts to. With scalability up to industrial scale meshes (~1M cells) and full automation of hyperparameter selection, CompArt enhances the applicability of CFD compartmental models in crystallization, multiphase flow, and process optimization. By bridging AI-driven clustering with physical modelling, this novel framework streamlines CFD simulations, making real-time industrial applications feasible without sacrificing predictive accuracy.

The efficient separation of carbon dioxide (CO2) and methane (CH4) is crucial for chemical processes, including biogas upgrading and natural gas purification. Metal-organic frameworks (MOFs) have gained significant attention as promising adsorbents for these processes due to their high porosity and tunable structures. Estimating the adsorption capacity of MOFs is essential for screening high performing adsorbents. While molecular simulations are commonly used to estimate the adsorption capacities, their computational intensity acts as a bottleneck in screening MOF adsorbents. In this study, we propose a machine learning (ML)-based framework for the high-throughput prediction of adsorption isotherms for CO2 and CH4 in MOFs. A graph neural network (GNN) model was developed to predict adsorption capacities, effectively replacing the time-consuming molecular simulations. The GNN model processes the structural graphs of MOFs, capturing their spatial configurations, such as surface structure and pore characteristics, which are closely related to adsorption performance. Based on the adsorption capacities predicted by the GNN model, an isotherm model was derived to characterize the adsorption behavior of the MOFs. This framework enables the rapid and reliable screening of MOFs with superior adsorption performance for CO2/CH4 separation, offering a computationally efficient approach that paves the way for broader applications of MOFs in carbon dioxide separation technologies.

This research addresses the challenges associated with data-driven soft sensors in industrial applications, where successful implementations remain limited. The scarcity of practical applications can be attributed to variable operating conditions and frequent disturbances in real-time processes. Industrial data are often nonlinear, dynamic, and highly unbalanced, complicating efforts to capture the essential characteristics of underlying processes. To tackle these issues, we propose a comprehensive solution for industrial application, that encompasses feature selection, feature extraction, and model updating. Feature selection aims to pinpoint the independent variables that have a substantial impact on key performance indicators, including quality, safety, efficiency, reliability, and sustainability. By doing so, it simplifies the model and boosts its predictive accuracy. The process begins with screening variables based on process knowledge, followed by a thorough analysis of correlation and redundancy to eliminate redundant information, which can burden computational resources and degrade prediction accuracy. We propose a mutual information-based algorithm for feature selection that assesses the relevance and redundancy among process variables through a comprehensive correlation function. This algorithm ranks variables by their importance using a Greedy search method to identify the optimal set of variables. After selecting the optimal variables, feature extraction is carried out to derive internal features from this set and establish a relationship between these latent features and the output variables. Given the intricate nature of industrial processes, we employ deep learning techniques, specifically Long Short-Term Memory (LSTM) networks, which are a type of Recurrent Neural Network (RNN) well-suited for capturing long-term dependencies in sequential data. LSTMs excel at modeling temporal correlations due to their ability to maintain memory states that allow for learning from sequential data over extended periods. To address short-term nonstationary features resulting from process disturbances, we incorporate a differential unit into the latent layer of the LSTM network. Once trained, the model is updated during online applications to incorporate gradual changes in equipment and reaction agents. Quality-related data, although typically available only post-measurement, can be leveraged to fine-tune model parameters, ensuring sustained predictive accuracy over time. To validate our approach, we present a case study on a delayed coker unit, yielding promising long-term predictions for tube metal temperature and showcasing the potential of our methodology for industrial applications.

Catalysis design and reaction condition optimization are considered the heart of many chemical and petrochemical processes and industries; however, there are still significant challenges in these fields. Advances in machine learning (ML) have provided researchers with new tools to address some of these obstacles, offering the ability to predict catalyst behaviour, optimal reaction conditions, and product distributions without the need for extensive laboratory experimentation. In this contribution, the potential applications of ML in heterogeneous catalysis and photocatalysis are explored by analysing datasets from different reactions, including Fischer-Tropsch synthesis and photocatalytic pollutant degradation. First, datasets were collected from literature. After cleaning and preparing the datasets, they were employed to train and test several models. The best model for each dataset was selected and applied for optimization.

Deep learning faces significant challenges in efficiently training large-scale models. These issues are closely linked, as efficient training often depends on precise and computationally feasible gradient calculations. This work introduces innovative methodologies to improve deep learning network (DLN) training in complex systems. A novel approach to DLN training is proposed by adapting the block coordinate descent (BCD) method, which optimizes individual layers sequentially. This is combined with traditional batch-based training to create a hybrid method that harnesses the strengths of both techniques. Additionally, the study explores Iterated Control Random Search (ICRS) for initializing parameters and applies quasi-Newton methods like L-BFGS with restricted iterations to enhance optimization. By tackling DLN training efficiency, this contribution offers a comprehensive framework to address key challenges in modern machine learning. The proposed methods improve scalability and effectiveness, especially for handling complex real-world problems. Examples from Process Systems Engineering illustrate how these advancements can directly enhance the training of large-scale systems.

This paper introduces a prototype for constructing an edge AI system utilizing the contemporary Machine Learning Operations (MLOps) concept. By employing microcontrollers such as the Raspberry Pi as hardware, our methodology includes data scrubbing and machine learning model deployment on edge devices. Crucially, the MLOps pipeline is fully developed within the ecoKI platform, a research platform for ML/AI applications. In this study, we thoroughly investigate the performance of our ecoKI platform by comparing it with the established Edge Impulse platform. We deployed the ML model with different weight quantization methods, such as FP32 and INT8, to compare accuracy variations and inference speed between these two platforms and quantization strategies on edge devices. In our experiments, we identified that the average accuracy performance of the ecoKI platform is 3.61% better than the edge impulse. Moreover, real-time AI processing on edge devices enables microcontrollers, even those with limited resources, to effectively handle tasks in areas such as predictive maintenance, process optimization, quality assurance, and supply chain management.

Developing ultra- or high-performance liquid chromatography (HPLC) methods for analysis or purification requires significant amounts of material and manpower, and typically involves time-consuming iterative lab-based workflows. This work demonstrates in two case studies that an autonomous HPLC platform coupled with a mechanistic model that self-corrects itself by performing parameter estimation can efficiently develop an optimized HPLC method with minimal experiments (i.e., reduced experimental costs and burden) and manual intervention (i.e., reduced manpower). At the same time, this HPLC platform, referred to as Smart HPLC Robot, can deliver a calibrated mechanistic model that provides valuable insights into method robustness.

Surface tension is a fundamental property at the liquid/gas interface, influencing phenomena such as capillary action, droplet formation, and interfacial behavior in chemical engineering processes. Despite its significance, experimental determination of surface tension is time-intensive and impractical for in silico-designed compounds. Predictive models are essential for bridging this gap. This study expands on Gaudin’s COSMO-RS-based model, which assumes uniform molecular orientation at the surface, by testing its predictive capability across broader temperatures (5–50°C) and developing a hybrid model combining first-principle and machine learning insights to improve Gaudin’s model predictions. The HM employs a serial configuration where COSMO-RS predictions serve as inputs alongside molecular descriptors, derived using the Mordred library. SHAP analysis guides feature selection, enhancing model interpretability. An artificial neural network refines predictions, optimized via Bayesian optimization for architecture and hyperparameters. Using a diverse dataset of 85 organic molecules, the HM demonstrates improved accuracy, reducing various metrics compared to standalone first-principle predictions.

Sustainable development goals have spurred advancements in bioprocess design, driven by improved process monitoring, data storage, and computational power. High-fidelity models are essential for advanced process system engineering, yet accurate parametric models for bioprocessing remain challenging due to overparameterization, often resulting in poor predictive accuracy. Hybrid modeling, combining parametric and non-parametric methods, offers a promising solution by enhancing accuracy while maintaining interpretability. This study explores hybrid models for succinic acid fermentation by Escherichia coli, a critical process for sustainable bio-based chemical production. The research presents a structured exploration of hybrid model architectures and their robustness under varying conditions. Experimental data were preprocessed to remove noise and outliers, and hybrid model structures were developed with differing levels of hybridization (from one to all reaction rates). Kinetic parameters were recalibrated and compared against original values. Machine learning algorithms, including Artificial Neural Networks, Support Vector Machines, and Gaussian Processes, were tested, with tuning strategies applied to original or recalibrated parameters. Due to a considerable variability in individual model performance for validation, an ensemble learning approach was proposed to enhance robustness. Results demonstrate that despite not solving the overparametrization issues, all hybrid models outperform the original parametric model, with the best-performing hybrid model achieving a 52.3% lower RMSE of validation, avoiding overfitting. Ensemble approaches further improved predictions, reducing RMSE by up to 62.3% compared to individual parametric models. This highlights hybrid modeling’s potential to enhance bioprocess prediction accuracy, even with limited data, supporting future advancements in bioprocess scale-up, digitalization, and sustainable biorefinery implementations.

A reinforcement learning framework is developed for the synthesis of distillation trains. The rigorous Naphtali-Sandholm algorithm for equilibrium separation modeling was implemented in JAX and coupled with the benchmarking Jumanji RL library. The vanilla actor-critic agent was successfully trained to build distillation trains for a seven-component hydrocarbon mixture. A transformer encoder structure was used to apply self-attention over the agent’s observation. The agent was trained on minimal data representation containing quantitative component flows and relative volatility parameters between present components. Training sessions involving 5·104 episodes (3·105 column designs) were typically run in under 60 minutes. While training was fast and reliable with appropriate tuning of the hyperparameters, further improvements are needed in the generalizability performance for similar separation problems.

Optimizing the operational conditions of electrochemical separation systems to achieve higher separation efficiency remains a complex challenge due to their nonlinear and dynamic nature. In this work, we proposed a Reinforcement Learning (RL)-based control framework to address this challenge. By applying various RL algorithms, we trained an RL-based controller that adapts to different system configurations and conditions. Also, the trained model learns the optimality between the removal efficiency and energy consumption. Overall, this approach autonomously learns the optimal operational parameters, significantly improving ion removal efficiency. The proposed RL-based control system enhances the performance of electrochemical system, providing a versatile and adaptive solution for optimizing separation across multiple electrochemical technologies. This work demonstrates the potential of RL in advancing the design and control of sustainable water purification systems.

The transition to renewable energy sources, such as biogas, requires purification processes to separate methane from carbon dioxide, with adsorption-based methods being widely employed. Accurate simulations of these systems, governed by coupled PDEs, ODEs, and algebraic equations, critically depend on precise parameter determination. While traditional approaches often result in significant errors or complex procedures, optimization algorithms provide a more efficient and reliable means of parameter estimation, simplifying the process, improving simulation accuracy, and enhancing the understanding of these systems. This work introduces an Artificial Intelligence-based methodology for estimating the isotherm parameters of a mathematical phenomenological model for fixed-bed experiments. The separation of CO2 and CH4 is used as case study. This work develops an algorithm for parameter estimation for the system’s mathematical model. The results show that the validated model has a close fit with experimental results.

Space exploration demands the integration of multiple scientific and engineering disciplines, with chemical engineering and process systems engineering playing pivotal roles. This paper examines their critical contributions to propulsion systems, life support mechanisms, and advanced materials essential for space missions. Recent advancements in chemical propellants and rocket fuels, illustrated by SpaceX and NASA missions, have significantly improved propulsion efficiency and safety. Chemical engineering is vital in developing air purification, water recycling, and bioregenerative life support systems, ensuring astronaut survival and mission sustainability. Additionally, creating heat-resistant, lightweight materials enhances spacecraft durability under extreme space conditions. Process systems engineering (PSE) complements these efforts by integrating, simulating, and controlling complex systems. PSE ensures reliable subsystem integration and uses predictive analytics and advanced modeling for mission planning and risk mitigation. Automation and control systems are essential for maintaining operations with minimal human intervention. The synergy between these fields is evident in in-situ resource utilization (ISRU) technologies, which extract and process local resources on extraterrestrial bodies, reducing reliance on Earth supplies and enhancing mission viability. Despite significant progress, challenges remain. Addressing harsh space environments, ensuring long-duration mission sustainability, and advancing energy sources and materials are ongoing research areas. This presentation underscores the indispensable roles of chemical and process systems engineering in overcoming space exploration challenges.

The separation of 1,3-butadiene from C4 hydrocarbon mixtures is a crucial step in the production of synthetic rubbers and plastics. Conventional extractive distillation methods using solvents, like N,N-dimethylformamide (DMF), have proven effective but presents significant health and environmental challenges. This study explores the feasibility of using propylene carbonate (PC) as a green solvent alternative for butadiene extractive distillation, leveraging its environmentally friendly properties and industrial compatibility. Simulations were conducted using Aspen Plus®, employing the Non-Random Two-Liquid (NRTL) model coupled with the Redlich-Kwong equation of state to describe phase equilibrium. Results indicate that PC integrates seamlessly into existing processes, achieving comparable operational stability and butadiene separation efficiency with minimal modifications. A significant design improvement was the elimination of the methylacetylene separation column in the PC process, which simplified the overall configuration and reduced cooling water requirements by approximately 43 tonnes per hour. This change resulted in a 0.4% reduction in utility costs. While PC’s higher boiling point requires operational adjustments, such as the use of higher medium-pressure steam, the overall energy consumption was 3.7% lower than with DMF. These findings demonstrate that PC is a viable and sustainable alternative for butadiene recovery processes, offering economic, operational, and environmental benefits.

The electrification of chemical processes and carbon capture and utilisation represent two promising approaches to improve efficiency and decrease carbon emissions of the process industry. The development of electrolyzers has gathered momentum in the last decades, allowing for the possible introduction of renewable electrons into carbon dioxide-based chemicals manufacture. While the performance of the electrolyzers is subject to improvements driven by the experimental community, the generation of waste heat is unavoidable due to the electrical resistances and process inefficiencies within the electrochemical cells. The possibility of re-using this waste heat has not been investigated within the realm of carbon dioxide electrolyzers. Here we show the potential of upgrading this waste heat by means of a heat pump, for its utilisation in the downstream processing of formic acid obtained from carbon dioxide electroreduction. We found that the waste heat represents roughly 62% of the power input to the electrochemical cells and it can be upgraded from 35 °C to 112 °C to drive the extractive distillation of formic acid and water. For a system producing 25 kt/y of formic acid, this results in the electrification of 60% of the separation energy duty, yielding a decrease in carbon dioxide emissions of 39-58% for the downstream processing, while the addition of a heat pump has a payback period of ~4.4 years in the base case.

In this work, we propose a mixed-integer linear programming (MILP) model that optimizes economic and environmental objectives by retrofitting fuel stations for the case study of Spain. The model contains set-covering constraints that ensure that there is at least one retrofitted fuel station within a radius of 20 kilometers of each retrofitted fuel station. The results indicate that by retrofitting fuel stations to allow for electric vehicles, both economic and environmental objectives improve, while showing which power plants would be tasked with the increase in electricity production to satisfy the increased electric demand.

A comprehensive model has been developed to design heat exchanger networks integrated with organic Rankine cycles (ORCs) and heat pumps, aiming to optimize energy efficiency. The model focuses on two key objectives: first, using heat pumps to reduce dependency on external services by enhancing heat recovery within the system; second, utilizing ORCs to recover residual heat or generate additional energy. To achieve optimal performance, the model requires careful selection of fluids for both ORCs and heat pumps, and the determination of optimal operating temperatures for maximum efficiency. The heat exchanger network is designed to be flexible, with non-fixed inlet and outlet temperatures, while simultaneously optimizing the number and operating conditions of ORCs and heat pumps. This approach reduces costs related to external services, electricity, and equipment such as compressors and turbines. Ultimately, the model facilitates the design of a heat exchanger network that efficiently utilizes residual heat and integrates other energy streams, thus improving both operational efficiency and sustainability. It demonstrates the potential for incorporating ORCs into systems that manage various energy streams, extending beyond just residual heat.

Climate change, driven by increasing CO2 emissions, necessitates innovative mitigation strategies, particularly for hard-to-abate industries. Carbon Capture and Utilization technologies offer promising solutions by capturing CO2 from industrial flue gases and converting it into value-added products. Among capture methods, membrane separation stands out for its compact design, energy efficiency, and scalability. Following capture, CO2 can be converted into chemicals like formic acid using electrocatalytic processes, enabling energy storage from renewable sources. This study proposes the design of an industrial demonstrator for a CO2 recycling plant targeting hard-to-abate sectors such as textile and cement industries. The system integrates polymeric membranes for CO2 capture and a 100 cm² electrochemical reactor for CO2 electroreduction into formic acid. Experimental data from both stages are used to develop predictive models based on artificial neural networks (ANN), optimizing system performance. Case studies reveal that CO2 concentration at the capture inlet significantly impacts plant design. For a textile plant with 3.5% CO2 emissions, a four-stage membrane system is required, resulting in higher CAPEX and OPEX. Conversely, a cement plant with 12% CO2 emissions requires only two stages to achieve the target CO2 concentration of >75 %, reducing costs by over 60%. Sensitivity analysis highlights the critical role of inlet CO2 concentration on the membrane area and system stages. The findings underscore the feasibility of modular membrane systems tailored to emission characteristics, paving the way for sustainable CO2 recycling processes adaptable to various industries. This integrated approach offers a pathway to mitigate emissions while generating valuable chemical products.

To achieve long-term greenhouse gas neutrality in aviation, replacing fossil aviation fuels with Sustainable Aviation Fuels (SAF) from renewable sources is essential. A SAF production process from renewable hydrogen and carbon dioxide, was designed using Aveva Process Simulation, followed by comprehensive economical assessments. The designed process leads to an annual production of 37kt of SAF, with 97% of the molecules featuring a carbon chain length between 8 and 16. This output indicates a robust and targeted production capability. With an in-depth optimization of the methanol reactor, it was found that the profitability of the plant aligns with other SAF studies, demonstrating a Minimum Selling Price of Product of $2.46/kg after Heat Integration. In terms of economic profitability, the production of SAF using the methanol pathway appears to be an alternative to other SAF production pathways such as Fischer-Tropsch process but resides dependent on the evolution of H2 production technologies, novel SAF production processes and environmental regulations. Despite this, the price remains higher than the current cost of fossil jet fuels, posing a competitive challenge.

Hydrogen is increasingly recognized as a crucial energy carrier for a low-carbon future. However, most studies on clean hydrogen production devote limited attention to the entire supply chain. This study evaluates the sustainability of 800 combinations of hydrogen production and transportation methods, comparing their environmental impacts against the geophysical limits defined by the Planetary Boundaries framework. Findings reveal that no supply chain alone can make the current economy sustainable, yet powering water electrolysis with bioenergy and carbon capture and storage can meet the CO2-based planetary boundaries. The analysis also underscores the need for decarbonization efforts in the hydrogen transportation sector, as certain options could offset the benefits of clean hydrogen production.

Reduction of greenhouse gas emissions is an important environmental element to actively combat the global warming and climate change. In view of reducing the CO2 concentration from the atmosphere, the Direct Air Capture (DAC) options are promising technologies in delivering negative carbon emissions. The integration of renewable-powered DAC systems with the CO2 utilization technologies can deliver both negative carbon emissions as well as reduced energy and economic penalties of overall decarbonized processes. This work evaluates the innovative energy- and cost-efficient potassium – calcium looping cycle as promising direct air capture technology integrated with various CO2 catalytic transformations into basic chemicals / energy carriers (e.g., synthetic natural gas, methanol etc.). The integrated system will be powered by renewable energy (in terms of both heat and electricity requirements). The investigated DAC concept is set to capture 1 Mt/y CO2 with about 75 % carbon capture rate. Up to 50 % fraction of the captured CO2 stream will be then catalytically converted into the Synthetic Natural Gas (SNG) or methanol using the green hydrogen produced by the water electrolysis (using renewable power), the rest being sent to geological storage. As the results show, the integrated DAC – CO2 utilization system, powered by renewable energy, has promising performances in terms of delivering negative carbon emissions, promoting the utilization of renewables and reducing ancillary plant energy consumptions.

The reduction and recovery of organic fraction of municipal solid waste is a major challenge for contemporary society. It requires the establishment of regional strategies with minimized environmental impact. This study employs life cycle assessment to evaluate the respective environmental performances of the current French system based on incineration, and those of alternative systems including (i) anaerobic digestion with composting and (ii) composting for biowaste treatment under different energy scenarios. The environmental impacts of Parisian biowaste are calculated by considering incineration technologies in the area, the French energy mix in 2022, the average European energy mix in 2022 and the projected French energy mix for 2030. The results show that the proportion of fossil-based sources in the energy mixes significantly influences the environmental performance of waste management systems. Energy mixes based in high-carbon fossil sources dependency tend to favour incineration-based processing systems. This is driven by the significant volume of credits allocated to the alternative energy produced from waste treatment. It is therefore essential for processing system design projects to account for planned changes in the energy mix in order to accurately assess the future impact of waste management strategies.

Energy storage is essential for transitioning to a renewable system based on renewable sources. To meet this challenge, Power-to-X technologies are attracting more attention. This work explores converting the excess of electric energy obtained from wind or solar sources into hydrogen and then into methane leveraging existing natural gas infrastructure for easier storage and transport. The process involves two stages: Firstly, the methane production step using Power-to-X technologies during excess renewable energy periods and, secondly, the electricity generation step during high demand with CO2 capture for reuse in methane synthesis, forming a closed carbon loop. In this way the Power-to-X process is integrated with repurposed combined cycle power plants (CCPPs) creating a Power-to-methane-to-power system. Two approaches are evaluated: oxy-combustion, which simplifies process CO2 purification and air combustion, which needs a more complex CO2 purification, such as amine absorption or PSA systems. The results show that ordinary combustion capturing CO2 by amines is the most profitable process, especially when it is assumed that the oxygen produced by the electrolyzer is sold. For a CCPP with a nominal capacity of 400MW, the cost of electricity production ranges from 450 $/MWh to 490 $/MWh when the income from oxygen sales is neglected. If the oxygen produced can be fully absorbed by the market, the electricity production cost decreases to 250 $/MWh and 300 $/MWh.

The aluminium industry emits approximately 1.1 billion tonnes of CO2-eq annually, contributing about 2% of global industrial emissions. Decarbonization pathways aim to achieve net-zero emissions by 2050, but this requires making decisions today for technologies having lifetimes of 20 – 25 years, based on uncertain economic assumptions, particularly given the volatility of energy prices. Traditional price forecasting models often fail to anticipate major disruptions, such as the 2022 energy crisis. This work applies Monte-Carlo Analysis (MCA) to evaluate the financial stability of decarbonization pathways under energy crisis scenarios and report on the resilience of the alternative solutions. In the modelled secondary aluminium production facility, direct electrification is assumed for lower temperature furnaces of annealing heat treatments or preheating, while the study defines the decarbonization options based on the melter furnace technology, a key bottleneck in terms of load and viable technological alternatives with roof temperatures exceeding 1,000°C. In this study, four decarbonization pathways characterized by unique melting furnace solutions are analyzed, namely: an oxyfuel scenario using natural gas with carbon capture, a biomass-based approach relying on syngas, a hydrogen-based solution utilizing onsite electrolysis, and a plasma-based system with torches for melting. These options are compared to a natural gas baseline. Differences in electrification levels and reliance on imported energy sources result in varied economic resilience under price shocks that can occur during different times of the plant’s life. Results obtained for a set of initial price levels at moderate or crisis energy price profiles revealed a more resilient oxyfuel scenario, followed by biomass utilization, and finally the all-electric scenarios utilizing either hydrogen or plasma. For instance, the oxyfuel option demonstrates a 54% probability of negative incremental Net Present Value (iNPV) with reference to the base case under both crisis and moderate price profiles. Conversely, the biomass scenario was found to have a 95-98% likelihood of economic loss under similar conditions while the all-electric solutions were at 100% likelihood of loss regardless of the initial price or profile. This is because electricity prices start at a higher baseline than combustible fuels and spike dramatically during crisis events. This analysis highlights the importance of considering energy price volatility and the need for diversified energy sources when developing decarbonization strategies for the aluminium industry.

Heat pumps play a crucial role in decarbonizing the chemical industry. The integration and sizing of heat pumps in chemical processes is a challenging task in multi-product chemical processes due to the fluctuating waste heat supply and heat demand. Integrating heat pumps may require a retrofit of the utility system. Mathematical optimization is a useful tool to tackle this challenge by enabling the analysis of correlation between relevant system parameters and equipment sizing. This study demonstrates the utilization of mathematical optimization and parameter studies for utility system equipment sizing addressing fluctuating heat supply and demand profiles.

Hydrothermal liquefaction (HTL) has proven to be an appropriate technology for converting sewage sludge into a valuable resource for renewable energy generation. This study focuses on a prospective analysis of various technological scenarios for sewage sludge-to-fuel pathways via HTL, co-located with a wastewater treatment plant, in support of a circular economy perspective. Four technological foreground scenarios and three prospective background scenarios aligned with the Paris agreement’s climate targets REMIND-SSP2-Base (projecting a 3.5°C temperature rise by the end of the century), PKBudg1150 (aiming to limit the rise to below 2°C), and PKBudg500 (targeting a cap below 1.5°C) are analyzed for sewage sludge-to-fuel conversion in 2030, 2040, and 2050. The superstructure problem of the possible combinations of the developed scenarios is solved using the P-graph studio which is based on the branch and bound approach. The goal of this study is to maximize the objective function (OF) by accounting the credits from avoided GHG emissions, the market value of recovered products, while subtracting operational costs and GHG emission penalties incurred during the biocrude production and upgrading processes. The optimal solution shows a potential OF equal to 858 €/ton of sewage sludge for Pkbudg500 under technological foreground scenario 2 by 2040. The P-graph approach demonstrates that HTL treatment of sewage sludge provides an alternative production pathway within the circular economy concept, capable of identifying optimal and near-optimal solutions for addressing trade-offs between future socio-economic policies and practical implementation for 2030, 2040, and 2050, which are often difficult to monetize.

In this paper, we present a systematic approach to quantify the safe operating window of a proton exchange membrane water electrolysis (PEMWE) system considering energy intermittency and varying hydrogen demand. The PEMWE model has been developed based on first principles, with the polarization curve validated against a lab-scale experimental setup. The impact of key operational variables is investigated which include voltage, inlet temperature, and water flowrate (utilized for both feed and system cooling). Emphasis is given on operating temperature, a safety-critical variable, as its elevation can pose significant hydrogen safety risks within both the electrolyzer cells and the storage system. The impact of temperature on process safety is quantified via a risk index considering the fault probability and consequence severity. Process operability analysis is employed to assess the achievability of a safe and feasible region for design and operations. This analysis provides a comprehensive framework to optimize PEMWE systems for enhanced operational flexibility and robust performance with application to modular hydrogen production using renewable energy sources.

Endpoint indicators provide a concise representation of environmental impacts by aggregating multiple midpoint indicators into a single value. Traditional endpoint weighting systems, however, are often limited by biases introduced through panel reviews and a lack of robustness in scientific process models. Additionally, they typically fail to account for the preferences of key stakeholders, including industry, government, and the public. This work addresses these limitations by developing an endpoint indicator that incorporates stakeholder preferences and minimizes dissatisfaction. A multi-stakeholder optimization framework was formulated to achieve this goal, employing distance, downside risk, and conditional value at risk as objective functions. Stakeholder preferences were derived from emissions data for industry, federal spending on environmental issues for government, and public surveys for societal input. Results highlight regional variations in midpoint indicator weightings across Texas, California, Delaware, and the United States. Furthermore, the influence of public and government preferences reveals a prioritization of human health concerns over environmental impacts. This framework offers a novel approach to creating endpoint indicators that align with stakeholder priorities, promoting more inclusive and collaborative environmental decision-making.

This study optimizes the Sustainable Aviation Fuel Supply Chain Network (SAFSCN) in the Czech Republic, using wheat straw as feedstock. It integrates geospatial data, transportation logistics, and economic feasibility, applying mixed-integer linear programming (MILP) to optimize pyrolysis plant locations and minimize costs. Sensitivity analysis varied wheat production growth by ±0.1% and ±0.2%. Results confirm Sustainable Aviation Fuel (SAF) production is technically and economically viable, with costs projected to decline up to 30.64% and revenues rising 49.07% from 2030 to 2050 due to technological advancements, improved logistics, and economies of scale. The findings underscore the critical role of SAF in achieving EU aviation decarbonization targets and highlight the importance of efficient supply chain planning for scaling SAF production.

Decarbonizing industries could significantly increase electricity demand, necessitating strategic grid expansion. This study evaluates the impact of decarbonizing the Pulp and Paper Sector under four 2050 scenarios: carbon capture, biomass-based, direct electrification, and indirect electrification. A bottom-up approach is employed to estimate 2020 final energy demand by heat grade and subsector. Both final and primary energy demand systems are modeled, accounting for the efficiencies of end-use technologies and primary energy transformation processes. The analysis compares primary renewable energy demand (electricity and biomass) normalized per ton of equivalent CO2 avoided against a business-as-usual scenario. It also considers the requirements for wood residues, organic waste, and CO2 storage. The carbon capture scenario, while low in electricity demand, requires significant organic waste for renewable natural gas production and 2.6 Mt of CO2 storage to offset direct and indirect emissions, making it the least feasible due to uncertainties around carbon storage in Quebec. Among the remaining scenarios, the direct electrification stands out by offering the lowest primary energy demand. It combines heat pumps with electric boilers for steam production and lime kilns are converted to a plasma-based solution. The study also includes a sensitivity analysis highlighting the potential of energy efficiency measures to ease the burden of decarbonization.

Methanol and ammonia are key energy carriers in a decarbonized society. This study assesses their use in power generation via two pathways: direct utilization as green fuels in fuel cells or as hydrogen carriers. Using these chemicals as hydrogen carriers achieves higher efficiencies (around 40%) due to the maturity of hydrogen fuel cells, resulting in electricity costs around 700 €/MWh compared to 1200 €/MWh for direct utilization. While hydrogen offers lower electricity production costs, efficiency advancements in methanol and ammonia fuel cells could enhance their competitiveness. Additionally, for scenarios involving transportation and power generation, methanol and ammonia prove economically viable, particularly for distances exceeding 3000 km. Consequently, both are crucial for addressing hydrogen-related challenges in the new renewable energy systems.

Hydrogen is a key element in the global transition toward a low-carbon economy, with green hydrogen offering significant potential to decarbonize industries and energy systems. This study focuses on designing and optimizing a green hydrogen supply chain (HSC) for the state of Bahia, Brazil, using a deterministic Mixed-Integer Linear Programming (MILP) model. The model evaluates 24 scenarios combining production sites, storage technologies, transportation methods, and energy sources, minimizing the Total Sustainable Cost (TSC). The TSC integrates financial and environmental costs, monetizing CO2 emissions using international carbon pricing. Results indicate that economies of scale play a critical role allowing the minimization of the financial costs while achieving lower greenhouse gas (GHG) emissions compared to other scenarios. The study emphasizes the importance of aligning production strategies with regional renewable energy resources to enhance both cost-effectiveness and sustainability. Future research should incorporate stochastic approaches to address uncertainties in hydrogen demand, energy availability, and market dynamics.

In 2022 the Dutch Energy System used some 2700 PJ of energy. Some 86% of its input was natural gas, crude oil and coal. The other 14% were renewables. A network of power-generation units, refineries and petrochemical complexes converted fossil resources into heat (700 PJ), power (400 PJ), transportation fuels (500 PJ) and chemicals (400 PJ). Some 700 PJ were lost due to conversion and transport. CO2 emissions were 160 Mt in 2022 of which 65 Mt by industry and 30 Mt by mobility. Transformation of this system into a Net Zero CO2 system calls for replacement of fossil resources by renewable heat, power and carbon. Decarbonisation of heat & power for residential and mobility is well underway at the moment. However, decarbonisation of industry and recarbonisation of shipping & aviation fuels, as well as recarbonisation of feeds for chemicals, is hampering progress. This paper concludes that current policies, predominantly based on trading CO2 emission certificates (ETS) is insufficient for industry to justify investment decisions. A carbon tax of 100 €/t CO2 on top of ETS prices (currently some 70 €/t CO2) is still insufficient for viable business cases. This paper shows that a combination of mandated production volumes for renewable fuels supported by opex subsidies, investors’ willingness to accept lower costs of capital, off-takers’ willingness to pay higher prices for green products and producers’ willingness to accept lower returns on their investment, may drive industry towards viable business cases. Financial support comes from government through their return of extra CO2 tax income and income from sales of ETS certificates to producers, off-takers and investors (risk mitigation assurances). This paper shows how QuoMare’s TDES (Transformation of the Dutch Energy System) is deployed to outline a policy framework that provides robust incentives to industry to invest in decarbonisation and recarbonisation projects.

This study designs a green hydrogen supply chain for Portugal, focusing on minimizing both economic costs and water stress. The research uses a multi-objective simulated annealing algorithm to address the trade-offs between the two objectives. The process of producing hydrogen via electrolysis is highly water-intensive, posing a challenge in water-scarce regions like southern Portugal. The study considers Portugal’s uneven water distribution, renewable energy availability and, different hydrological conditions across districts. An aggregate indicator, Water Stress Index (TWSI), quantify the pressure on water resources covering all the Portugal´s districts in a single score. This study explores four scenarios, a baseline scenario, a green hydrogen scenario using only renewable energy, drought conditions, and increased demand in major cities with drought conditions. The quasi pareto front illustrates the trade-offs between supply chain cost and TWSI, enabling decision-makers to select solutions tailored to local priorities. The findings highlight the importance of adaptable and robustness infrastructure to overcome both environmental and demand challenges.

A transition to low-carbon fuels is integral in addressing the challenge of climate change. An essential transformation is underway in the transportation sector, one of the primary sources of global greenhouse gas emissions. The electrofuels that represent methanol synthesis via power-to-fuel technology have the potential to decarbonize the sector. This paper outlines a critical comprehensive life cycle assessment for electrofuels, with this study focusing on the production of synthetic methanol from renewable hydrogen from water electrolysis coupled with carbon from the direct air capture (DAC) process. This study has provided a comparison of the environmental impacts of synthetic methanol produced from grids of five regions (India, the US, China, Switzerland, and the EU) with conventional methanol from coal gasification and natural gas reforming. The results from this impact assessment show a high dependency of environmental scores on the footprint of the grid. Switzerland, with its limited usage of fossil fuels in grid electricity, has shown the least environmental damage in most of the impact categories in comparison to more non-renewable grids like India and China. Interestingly, conventional methanol has proved to be more environmentally friendly when compared with non-renewable grids like India and China in most of the impact categories.

The use of waste heat from electrolysis can significantly increase process efficiency. Alkaline and PEM electrolyzers, the most mature technologies, produce low-temperature waste heat. Most studies focus on using this waste heat for low-temperature applications like district heating. Alternatively, this waste heat can be upgraded to a temperature that can be usable in the chemical industry, e.g., for steam generation. The combination of an alkaline electrolyzer with a heat pump has been recently investigated to supply both hydrogen and medium-temperature heat. Optimizing electrolyzers for both hydrogen and heat production (combined design) has been shown to have advantages over optimizing for hydrogen only and upgrading the waste heat a posteriori (separate design). However, the effects of electrolyzer pressure and hydrogen compression were not considered, and it remains unclear if similar benefits apply to PEM electrolyzers. This work further analyzes the combined system (i.e., electrolyzer with a heat pump) by including hydrogen compression and comparing alkaline and PEM electrolyzers. The results show that designing with waste-heat utilization in mind benefits both alkaline and PEM at low and high pressures. The combined design allows up to 10% cost reduction and up to fourfold reduction in CO2 emissions compared to the separate design. However, this benefit is constrained by the maximum achievable current density, particularly for PEM. Despite this limitation, the combined system can effectively supply hydrogen and heat across various energy price scenarios, making it a promising solution for hydrogen supply and industrial electric heating.

One of the proposed strategies to reach net-zero goals is the diversification of a country’s energy mix and transition to technologies that favour the mitigation of greenhouse gas emissions, while decreasing dependency on conventional fuels. This work presents a mathematical model that describes key production routes for two proposed energy transition vectors, biomethane and hydrogen, expressed as a Mixed-Integer Linear Problem (MILP). The supply chain is optimized with the objective of maximizing the profits from the global supply chain. The problem is formulated as an allocation problem, with production distributed between biomethane and hydrogen markets. The case study focuses on a region in Mexico where second-generation biomass for biogas production is abundant, while hydrogen is produced from biomethane using steam methane reforming. The results highlight the importance of balancing resource allocation in shared supply chains. With a production ratio of 60% biomethane and 40% hydrogen, a Levelized Cost of Hydrogen (LCOH) of 2.07 €/kWh and a Levelized Cost of Biomethane (LCBM) of 0.17 €/kWh are obtained, resulting in total hydrogen and biomethane production of 12,563 GWh/year and 72,374 GWh/year, respectively.

The climate crisis continues to grow as an existential threat. Establishing reliable energy resources that are renewable and zero-carbon emitting is a critical endeavour. Hydrogen has emerged as one such critical resource due to its high gravimetric energy density and near-abundant availability. However, it suffers from low volumetric energy density and is incredibly challenging to store and transport. The metal hydride, a solid-state storage method, provides a viable solution to the current limitations. Storage is achieved through the chemical absorption of hydrogen into a porous metal alloy’s sublattice. But its challenging thermodynamic functionality leaves a gap between the ideal storage capacity that current industry requires and the limited capacity that reusable metal hydrides currently provide. This work used mathematical modelling to determine optimal operating conditions for a metal hydride in order to maximise hydrogen storage capacity. Computational fluid dynamics is used to simulate the coupled heat and mass transfer that occurs during the absorption process into the metal alloy. The finite volume method is used to discretise governing equations, and the alternating direction implicit method is used for numerical solutions as it proved the most stable platform to conduct analyses. An initial investigation into numerical grid sizing is conducted to determine the optimal node allocation. The impact of the hydride bed thickness and supply pressure are varied and optimised. The alloy MmNi4.6Al0.4 is used in the investigation.

Climate change, primarily caused by the excessive emission of greenhouse gases, particularly carbon dioxide (CO2), has intensified global efforts toward achieving carbon neutrality. In this context, renewable energy and CO2 utilization technologies have emerged as key strategies for reducing reliance on fossil fuels and mitigating environmental impacts. In this work, a superstructure model is developed to integrate renewable energy network and chemical production processes. The energy network integrates wind, solar, and biomass energy, complemented by storage systems to enhance reliability and reduce reliance on external power sources. The reaction network features various pathways that utilize CO2 as a raw material to produce high value-added chemicals such as polyglycolic acid (PGA), ethylene-vinyl acetate (EVA), and dimethyl carbonate (DMC), allowing for efficient conversion and resource utilization. A mixed-integer linear programming (MILP) model is formulated to minimize production costs while optimizing energy and reaction pathways. This research supports the green transition of the chemical industry by optimizing the integration of renewable energy and CO2-based chemical processes, promoting more sustainable production practices.

Ensuring access to clean water, preserving water reserves, and meeting energy needs are fundamental for sustainability and a priority for global organizations like the UN and EU. The Mediterranean, particularly Greece, faces severe water imbalances due to rising demand, prolonged droughts, and seasonal tourism pressure. This over-exploitation of water resources threatens agriculture, employment, and regional sustainability. Addressing these challenges, this study analyzes the water-energy nexus in high-stress areas and develops an optimization model for sustainable water resource management. The model integrates sectoral demands, energy consumption, and seasonal variability to improve efficiency while balancing economic and environmental constraints. Additionally, it incorporates demand forecasting to align water use with ecosystem sustainability, reducing environmental impacts. By providing a systematic framework for decision-makers, this research supports the development of long-term, resilient water management strategies, ensuring efficient resource allocation in vulnerable regions. The findings will contribute to sustainable water policies and infrastructure planning.

Hydrogen is increasingly recognized as a key player in future energy systems. However, its production technologies—Steam Methane Reforming (SMR) and electrolysis—present trade-offs. SMR, the dominant method, is cost-effective but has a significant carbon footprint, emitting substantial greenhouse gases (GHGs). In contrast, electrolysis, powered by renewable energy sources, offers a cleaner alternative, albeit at a higher cost. While current hydrogen system optimizations primarily focus on cost reduction and GHG mitigation, they often neglect broader environmental impacts. This paper addresses the challenge of modeling a hydrogen supply chain (HSC) that achieves strong environmental performance at a relatively affordable cost. To this end, a supply chain design optimization is coupled with Life Cycle Assessment (LCA). The novelty of this work lies in the integration of multiple LCA indicators in the supply chain design optimization, rather than focusing solely on costs and Global Warming Potential. The model employs a multi-period (2025–2050) mixed-integer linear programming (MILP) formulation, implemented in the General Algebraic Modeling System (GAMS) environment, using CPLEX as the solver. The Bourgogne-Franche-Comté region is used as a case study to illustrate the framework. Initial optimization results highlight how varying the focus on different parameters can lead to distinct solutions and provide insights into the interactions between indicators. Future research will aim to reduce the dimensionality in multi-objective optimization to improve efficiency while maintaining environmental sustainability in the development of hydrogen networks.

Industrial ecology appears to be a significant means of reducing carbon dioxide emissions from industry. However, beyond flow management, eco-industrial parks can also contribute to a socio-economic transition on a regional scale. Usually, multi-criteria optimization models use economic and environmental criteria in the decision-making process. This article looks at the integration of social criteria in these models, and more broadly at the issues involved in measuring the social impact of an eco-industrial park. The aim of this article is to take a different approach to social indicators, by highlighting the key success factors of an eco-industrial park, such as cooperation within a collective but also between the different scales making up the system. This work is based on bibliographical research and semi-structured interviews with stakeholders in the field. What’s more, the process of developing social indicators, particularly participatory ones, seems to be a strong catalyst in the construction of relationships and their dynamics. Finally, the metric of social indicators also has its limitations, which should be considered if we are to be able to address the various issues in a systemic way and potentially envisage other ways of taking into account the social aspect of an eco-industrial park.

This study evaluates the environmental impact of carbon capture technology in the context of reducing industrial CO2 emissions within Eco-Industrial Parks (EIP). The primary focus is on the post-combustion absorption process, which uses solvents like monoethanolamine (MOA) to capture CO2 before it is released into the atmosphere. The captured CO2 is either stored or utilized to prevent further contribution to climate change. The study employs a Life Cycle Assessment (LCA) methodology to compare the environmental impacts of two scenarios: one with CO2 capture and the other with the direct release of CO2 into the atmosphere. The LCA considers inputs, outputs, energy requirements, and materials needed for the CO2 absorption process. The functional unit of the assessment is 1000 tons of CO2, to standardize comparisons between both scenarios. Results show that the CO2 absorption process significantly reduces the impact on climate change, capturing over 80% of the CO2 from the stream. In terms of climate change, the absorption process has a much lower environmental impact than releasing CO2 into the atmosphere. However, other ecological impacts, such as resource use and ecotoxicity, are higher in the absorption process due to the materials required and the solvents used. In conclusion, while the CO2 absorption process offers clear climate benefits, it introduces additional environmental challenges. Future studies should explore alternative carbon capture technologies and extend the scope to include CO2 storage impacts and a broader range of LCA tools for more detailed assessments.

To achieve carbon neutrality, the installation of variable renewable energy (VRE) has been accelerated. However, the inherent variability of VRE can be mitigated through the installation of energy storage. This study aims to evaluate life cycle impacts of energy storage systems utilizing batteries, hydrogen storage, or thermal energy storage. A model of the energy storage systems with VRE was developed, and the annual energy flow simulations were conducted. The energy storage system targeted in this study assumed that all energy derived from VRE was stored in the energy storage and supplied to consumers. The amount of electricity to-be-sold from the energy storage system, based on the capacity of VRE and the installed energy storage, was calculated. A life cycle assessment was performed to evaluate the greenhouse gas (GHG) emissions, abiotic resource depletion (ARD), and intensity of GHG and ARD as life cycle impacts. The smallest life cycle impacts varied depending on the type and scale of energy storage. This quantitative evaluation of the energy supply capacity and the environmental impacts of the energy storage systems could provide valuable insights for designing efficient energy storage systems integrated with VRE, tailored to regional energy supply requirements.

In recent years, several strategies have been developed to reduce the carbon dioxide (CO2) released into the atmosphere. Carbon Capture, Storage and Utilisation (CCUS) is one of the proposed solutions. However, CCUS systems are expensive to install and operate. Furthermore, most studies in the literature have focused on CO2 utilisation and storage separately, without accounting for the effects of other CO2 emission management strategies. To address this gap, a Mixed Integer Linear Programming (MILP) framework for a supply chain network is developed in this work, incorporating CO2 storage, utilisation, permit trading, and carbon emission taxation. Multi-criteria decision analysis (MCDA) techniques are implemented to select CO2-based products for CO2 utilisation. The MILP framework is set to achieve the maximum environmental and economic performance using a Multi-Objective Optimisation (MOO) approach. This involves using the e-constraint method as a solution procedure to minimise the total cost of the supply chain network while maximising CO2 reduction. A case study of supply chains in Germany is conducted to demonstrate the applicability of the model. The game theoretical Nash approach is successfully applied to determine the fair trade-off between the two objectives.

Industries with low-temperature heat demand, such as laundry and syrup sectors, heavily rely on natural gas-fired boilers, posing challenges to achieving net-zero emissions by 2050. Like hard-to-abate sectors, they must explore energy transition strategies, including heat recovery, fuel substitution, or carbon capture, to reduce CO2 emissions. This paper evaluates the potential of energy transition in these sectors through case studies, using a mixed integer linear programming (MILP) approach. The analysis focuses on three key performance indicators (KPIs): specific energy consumption, CO2 reduction, and variable costs. By 2050, the adoption of heat pumps and waste valorization emerge as the most promising solutions for the syrup and laundry sectors. Specifically, the use of heat pumps reduces energy demand by at least 50%, while on-site biofuel production can fully replace natural gas consumption, thus eliminating dependency on external energy sources. The analysis highlights the importance of sector-specific strategies to meet climate targets, offering a pathway for low-temperature heat industries to reduce emissions while addressing economic and technological constraints.

In 2022, Uruguay launched its green hydrogen roadmap, due to its renewable energy potential, water availability, and favorable logistics to position itself as an exporter of green hydrogen and derivatives. The energy source for water electrolysis is a key factor in both the final cost and the environmental impact of hydrogen production. In this context, this study evaluates the environmental impact of hydrogen production via water electrolysis, powered by 150 MW of new power sources (solar, wind, and hybrid farms) in Uruguay, all connected to the national grid. A life cycle assessment (LCA) was conducted using OpenLCA software and the Ecoinvent database, with 1 kg of hydrogen as the functional unit and the system boundaries included power generation and the electrolysis system. The analysis focused on global warming potential (GWP), considering different scenarios for the grid energy mix and the inclusion or exclusion of surplus energy as carbon credits. The wind plant is the option that generates the highest energy output and surplus while requiring the least energy purchase from the Uruguayan grid, due to its higher capacity factor. Regarding GWP, it is also the alternative with the lowest environmental impact, with annual average emissions ranging from 0.84 to 1.02 kg CO2/kg H2 not considering carbon credits (CC) and between 0.2 and -1.87 kg CO2/kg H2 when CC are considered. These values represent emissions based on grid connection during the years with the highest proportion of renewable and fossil sources, respectively.

Rare earth elements (REE) and many other critical minerals are necessary for the manufacturing of modern everyday technologies, including microchips, batteries and electric motors. Recovery of these materials typically involves aqueous systems which can be modeled as chemical equilibrium problems. One common method for solving these problems is the law of mass action approach (LMA), where a system of non-linear equations involving the equilibrium constants is solved. However, despite being theoretically simple, these problems are in practice very difficult to solve. Currently, the use of iterative heuristics based on saturation indices to decide on which species and reactions to include in the calculations is the state of the art to arrive at a solution. Here, we present an optimization-based alternative to solve chemical equilibria problems involving precipitation/dissolution reactions without the need for such heuristics. Our approach is first validated against the LMA software MINTEQ and PHREEQC for a number of case studies, and then applied to a novel REE recovery process reported in the literature. Overall, our approach was found to have close agreement with MINTEQ and PHREEQC, and we were able to successfully replicate the reported yield and purity for the published REE process.

Sustainable aviation fuels (SAFs) can be pivotal, gradually replacing fossil kerosene and lowering carbon emissions without changing the existing infrastructure. One of the pathways to produce SAFs is through the Fischer-Tropsch synthesis (FTS) process. The present work proposes an integrated process of sustainable aviation fuel production from biogas through a reforming process, Fischer-Tropsch (FT), and a solid oxide electrolysis (SOEC) process. Aspen Plus v14 is used to build an integrated kinetic process model for biogas reforming, FTS and hydrocracking. The technical evaluation is assessed with several key performance indicators, such as carbon efficiency and process efficiency. In addition, two scenarios are investigated in this study for H2 supply from SOEC before and after reforming. The output products consist of kerosene and diesel since the tail gas and naphtha are recycled to the reformer to maximize SAF production. The simulation results show that the carbon efficiency of the integrated process is between 96.3% and 96.5%. Meanwhile, the process efficiency is between 40.1% and 40.9%. The process efficiency can be further improved by investigating the options for heat integration, with subsequent economic and environmental analysis to determine the production cost of SAFs and the CO2 reduction potential compared to conventional aviation fuel. In addition, projection cost analysis is needed to determine the decline in the cost of fuel production in the future.

Sugarcane bioenergy is a reality in Brazil, comprising the production of ethanol and bioelectricity. Sugarcane bioenergy can reduce greenhouse gas (GHG) emissions as compared to fossil fuels. However, there are concerns about the possible implications caused by the expansion of sugarcane production, the displacement of mainly pastureland, but also other croplands, and the potential for indirect land use changes. A promising strategy to enlarge sugarcane bioenergy in Brazil without compromising the cattle industry is to integrate both activities, converting extensive livestock production systems into more intensive ones. The objective of this study is to model and evaluate the socioeconomic impacts and land use change considering the expansion of ethanol production in two scenarios. The first scenario, referred to Business as Usual (BAU), comprises of sugarcane bioenergy and extensive livestock production, without any integration between the two. The second scenario, Integrated Sugarcane-Bioenergy and Livestock (ISBL) in Brazilian agriculture, considers the integration between these activities. To achieve this goal, a computable general equilibrium (CGE) model for the Brazilian economy was implemented. The Brazilian input-output matrix (main data for the model) was estimated for 2021. The closure and the shocks were set to 2031, with an estimated increase of 75% in ethanol production. Additionally, a technological efficiency gain of 5% was considered for the livestock sector within the ISBL system. Our modeling results show that the ISBL scenario is economically more efficient, as a higher gross domestic product and output level were obtained. The aggregate employment level of the economy also increased. Even considering that the integration of sugarcane bioenergy and livestock production took place only in these activities, positive socioeconomic impacts were noticed across other sectors of the economy.

A one-dimensional homogeneous reactor model for a cocurrent flow nonadiabatic catalytic membrane reactor operating water gas shift reaction (WGSR) is developed. The model is used to predict the performance of the reactor and estimate the optimal hydrogen flux profiles required to maximize the CO conversion, and control the temperature rise due to the exothermicity. Under the optimized condition, the secured optimal hydrogen flux is found to be a bang-bang type suggesting constructing reactors of different hydrogen permeabilities. To control the reactor temperature, the activity of the reaction side is diluted by distributing axially certain fractions of inert solid, i.e. 0.35, 0.45 and 0.50. The total volume fraction of the inert solid required to maintain the temperature at 320oC (593.15 K) is 0.50 and the profile is obtained to be a singular-arc type with an observed maximum activity at the reactor inlet.

The production of hydrogen from solar energy has surged in popularity in recent years, driven by global initiatives to combat climate change. The Greater Middle East (GME) region, with its favorable geographical position, offers considerable potential for solar-based hydrogen generation. This study combines Geographic Information System (GIS) spatial analysis and the Analytical Hierarchy Process (AHP) with data-driven optimization models to assess land suitability and hydrogen production potential within the region under various scenarios. Findings highlight that water availability is the primary limiting factor, followed closely by road accessibility in determining land suitability for hydrogen production. According to the AHP analysis, only 3.8% of the GME region is highly suitable for such initiatives. Projections suggest that by 2050, the region could achieve a total hydrogen production capacity of up to 1590 Mt/y, potentially avoiding around 4586 Mt of CO2 emissions if all highly suitable land is utilized effectively. Furthermore, the levelized cost of hydrogen (LCOH) is expected to decrease to a range of $2.56-4.17/kg, while green ammonia and methanol costs are anticipated to fall between $0.56-0.79/kg and $0.67-0.93/kg, respectively. This study offers valuable insights that contribute significantly to the global hydrogen discourse and provides essential guidance for policymakers aiming to develop comprehensive strategies for advancing the hydrogen economy in the GME region.

Active learning is widely recognized as an effective teaching approach that can improve classroom outcomes. This is enabled by providing the time for students to apply new knowledge, make mistakes, correct them, and repeat the process until mastery is achieved. One way to implement active learning is through the flipped classroom paradigm. However, to be effective, active learning depends on providing students with a variety of open-ended problems, ranging in difficulty from introductory to advanced levels. This paper presents four food-themed problems for use in numerical methods and process control courses: 1. Formulating Willy Wonka’s new chocolate bar: An introductory linear programming problem focused on translating verbal descriptions into mathematical models. 2. Optimal production for the Matrix Pizza company: A more advanced mixed-integer linear programming problem involving multiple scheduling scenarios. 3. Optimal frying time for fried ice cream production: A transient heat transfer problem modelled by an initial-value partial differential equation. 4. Control system design for Uncle Kane’s pancake batter machine: A control design problem involving uncertain process models and the selection of an appropriate operating mode and PI controller design. These problems aim to engage students in practical, problem-solving activities central to the active learning process.

Three different implementations of the flipped class paradigm were used to teach Chemical Engineering students at Imperial College London (ICL) in the 2023-24 academic year: (1) The 3rd year elective course Introduction to Numerical Methods (INM) taught in its entirety in flipped format (the “good”); (2) The 2nd year core course on Process Dynamics and Control (PDC), with the first half of the course on process dynamics taught in traditional lecture format, and the second half on process control taught in flipped format (the “bad”); and (3) a one-week workshop on heat integration, taught as part of a 3rd year core course on Process Design (PD), taught in flipped format (the “surprising”). This paper describes these three implementations in detail and presents and analyzes the responses from student surveys intended to ascertain students’ perceptions about the level of their satisfaction with the flipped class approach and the degree to which they achieved mastery of the taught materials. The actual learning outcomes are also analysed, that is, the exam results in the case of the first two courses, and the performance of the students on their design project, for the last course. The paper ends with conclusions and recommendations concerning how the flipped class paradigm should be implemented, depending on the classroom situation for which it is intended.

Digital twins and digital shadows are frequently used terms by industry and academia to describe data-centric models that accurately depict a physical system intended for process monitoring and control. Processes restricted by a low degree of automation rely greatly on operator competencies in key decision-making; a digital shadow can here assist as a guidance tool [1-4]. This work presents a practical implementation of a digital shadow to support operators running a pilot scale-packed batch distillation column at the Technical University of Denmark (DTU) primarily used in education and teaching activities [5]. This operation is selected due to inherent unsteady process dynamics that are controlled by a set of manual valves, which the student operator must continuously balance to meet purity constraints without disrupting the operation. This realisation employ a modular software architecture, separated into four distinct modules compiled into Docker images and independently deployed. These facilitate real-time connection with the equipment via OPC-UA endpoints, process model analysis via HTTP/HTTPS requests, and a dashboard UI presenting the data accessible through any web browser. The final module is realised as a Python package that stores thermodynamic models and is published on the internal GitLab registry for more straightforward utilisation across multiple projects. The advantages of this implementation strategy are flexibility and speed, allowing for continuously updating process models and directly employing them in various projects.

In this paper, we present our recent advances and achievements in automatic control course in the engineering study of cybernetics at the Faculty of Chemical and Food Technology STU in Bratislava. We describe the course elements and procedures used to improve teaching, learning, and administration experience. We discuss on-line learning management system, various teaching aids like e-books with/without solutions to practice examples, computer generated questions, video lectures, choice of computation and simulation tools. The course is provided in the presence form of study for about 20 students, but it relies on on-line tools and methods. Starting from this academic year, flipped design of the course was designed. We describe our experience in the preparation of such a change and some initial feedback from the students. The course concentrates on input/output linear approximation of processes in chemical and food technology and discusses poles/zeros, process dynamics, frequency and time methods, closed-loop system, design and tuning of PID controllers, root-locus analysis. Although the majority of laboratory exercises is performed using simulation tools, we also provide easy-to-use laboratory device that can be given to all students to work either at school or at home.

UCL Chemical Engineering ensures graduates are digitally literate by integrating computational tools like gPROMS, Aspen Plus, and GAMS into the undergraduate curriculum. Students in the first year of undergraduate program use GAMS to solve simple simulation and optimization problems and gPROMS for solving ordinary differential equations (ODEs) in reactor design problems. In the second year, students start using Aspen Plus to simulate more complex chemical process units, interpret and discuss results obtained and justify any differences observed between experimental data and computational results. They use GAMS to simulate and optimize a process flowsheet with considerations of the implications of proper initialization procedures and strategies for obtaining optimal parameters and gPROMS for advanced reactor and separator problems. The computational knowledge acquired in the first two years prepares students for the third-year capstone design project where they use the various tools in designing and simulating their processes during the preliminary, conceptual and detail designs. The computational resources available to students and the additional supports provided by educators help the students succeed on the design project in the third year. This work outlines strategies for delivering the computational tools required for modules in the first two years of the undergraduate program and the application of the knowledge in the third-year design project. It demonstrates how complex process systems engineering (PSE) problems are taught, student support and resources available, and feedback for continuous improvement. It presents some challenges faced due to the use of generative AI tools and how this can be addressed.

Current progress in digitalization has led to a wide interest in learning more from available data. Advanced data analytics can be achieved through commercially available software; however, learning to program allows for more flexibility and, ultimately, more freedom in the potentially tailor-suited investigation. Among other programming languages, Python is one of the most requested, in industry and research alike. To intensify the earlier efforts and create both a pedagogical framework to teach programming to (bio)chemical engineers, and provide students with the opportunity to ask questions, we explore the integration of sPyCE and FermentAI into BioVL, a virtual laboratory for teaching (bio)processes, previously implemented by the authors. sPyCE is an open-source series of Python courses tailored to (bio)chemical engineers, FermentAI is a chatbot trained to answer questions about fermentation processes. The main goal of this work is to enable students to (i) learn (bio)processes and how to model them, and (ii) ask questions to a chatbot to clarify their doubts. We believe that the Python tutorials will enable students with different backgrounds and programming skills to learn how to model (bio)processes. Moreover, the chatbot will enable students to ask for clarification right away, without having to search for an answer or wait for a teacher to reply.

Prompt engineering – improving the command given to a large language model (LLM) – is becoming increasingly useful in order to maximize the performance of the model and therefore the quality of the output. However, in certain instances, the user is not able to enrich the prompt with additional and personalized details, such as the preferred tone and length of generated response. Therefore, it is useful to create models that learn these preferences and implement them directly in the prompt. Current state-of-the-art inductive logic programming (ILP) systems can play an important role in the development and advancement of digitalization strategies. For example, they can be used to learn personal preferences of users without sacrificing human interpretability of the learned outcomes. These systems have recently witnessed the development of data efficient, robust, and human interpretable algorithms and systems for learning predictive models from data and background knowledge. In this paper, one of these systems, ILASP (inductive learning of answer set programs), is used to develop a proof of concept of how personal preferences of groups of students participating in an interview exercise can be learned to tailor and improve the generated response of a LLM used in an educational context.

Process control can be one of the most exciting and engaging chemical engineering undergraduate courses! This paper describes our experience transforming Chemical Process Control into Data Analytics, Optimization, and Control at the University of Notre Dame (second semester required course in the junior year). Our modern course is built around six hands-on experiments in which students practice data-centric modeling and analysis using the Arduino-based Temperature Control Lab (TCLab) hardware. We argue that state-space dynamic modeling and optimization are more critical for educating modern chemical engineers than topics such as frequency domain analysis and controller synthesis emphasized in many classical undergraduate control courses. All the course material is available online at https://ndcbe.github.io/controls.

The integration of mixed reality (MR) and virtual reality (VR) into Chemical, Biochemical, and Biomolecular Engineering (CBB) education presents an opportunity to address one of today’s most pressing pedagogical challenges: sustaining student attention and engagement. Traditional “magistral” approaches often tend to limit the adoption of interactive methodologies. By contrast, MR/VR technologies can heighten immersion and practical intuition, capturing learner focus more effectively than conventional lectures. Yet, if deployed as superficial, isolated demonstrations, these tools may fail to support deep conceptual understanding and risk supplanting core course content. This work proposes a flipped-classroom model that deliberately embeds MR/VR exercises throughout the typical CBB curriculum. The methodology emphasizes a human-in-the-loop concept, whereby the educator strategically orchestrates virtual simulations and real-world problem-solving, reinforcing theoretical concepts through hands-on digital experiences. We address challenges related to scaling this approach, highlighting the importance of aligning content across multiple years and conducting cost-benefit assessments. Preliminary efforts indicate high student acceptance and potential for enhanced learning outcomes. By blending MR/VR with proven instructional strategies, CBB educators can better equip graduates with the skills and engagement needed for success in an evolving industrial landscape.

This work gives a detailed description of the models, methods, and equipment used in a bachelor’s degree lab course. The connections between simulation results and real-world data are highlighted and tools for making the models useful for process design tasks are portrayed. The models cover the production chain for fuel-grade bioethanol, starting from the fermentation of sugar with yeast. In only one semester (14 weeks with 180 minutes per week) the students achieve to produce high-purity ethanol. Some exemplary results of the process designs and their comparison to the realized intermediate and final products are given together with production cost data.

This work explores several examples of how the thermodynamic concept of exergy can be used in the chemical engineering classroom. Examples include using exergy to determine thermodynamic and monetary value of utilities, to identify better heat exchanger network designs, to aid in work-heat integration applications such as heat pumps and organic Rankine cycles, to scope out realistic energy integration cases, and to assess how well chemical potential is being used and managed. The examples are presented in one connected context that makes it easy to see how exergy analyses can be useful across many aspects of chemical and energy industry supply chains.

Carnegie Mellon University (CMU) has a strong tradition and expertise in Chemical Process Systems Engineering. This short article comments on the CMU PSE-related courses and describes in more detail our approach to teaching Chemical Process Design. We discuss (i) our emphasis on proposing processes related to energy and sustainability and (ii) some of the challenges that are currently faced when teaching this course.

The Integrated Project in the Master of Chemical Engineering and Materials Science at the University of Liège (ULiège) aims to consolidate technical knowledge and promote the acquisition of soft skills by integrating various chemical engineering disciplines. The project focus on the design of an industrial process and is divided into five parts: individual work on mass balances and literature reviews, detailed modeling of thermodynamics and key unit operations, sensitivity studies, process integration, and report to a general audience. Key learning outcomes include developing critical thinking, addressing complex multidisciplinary topics, and understanding the role of science and technology in society. Students enhance their soft skills in project management, teamwork, and effective communication in English. Regular interactions with industry and academic experts, along with support from the ULiège Soft Skills Team, ensure comprehensive development. Evaluation includes both technical and soft skills assessments, with individual and group performance contributing to final grades. This project, running for several years, is highly regarded by students for its rigorous and instructive nature, preparing them well for their future careers.

This paper examines the current state of green and digital integration in traditional chemical engineering education, focusing on how artificial intelligence (AI) can enhance learning. A review of curricula shows that sustainability principles, such as green chemistry, circular economy, and resource efficiency, are often confined to electives rather than core courses. Likewise, digital skills are introduced at a basic level, with limited exposure to AI, especially machine learning, and advanced process optimization. The paper emphasizes the need for a structured approach to integrating sustainability and digitalization into core subjects, supported by interdisciplinary learning. It also explores AI’s role in transforming education, particularly in predictive modeling, process optimization, and adaptive learning. The study provides recommendations for redesigning the traditional chemical engineering curriculum to strengthen green and digital transformation.

The field of Process Systems Engineering (PSE) is undergoing a renaissance through the integration of artificial intelligence (AI) and machine learning (ML). This transformation is driven by the vast availability of industrial data and advanced computing power, enabling the practical application of sophisticated ML models. These models enhance PSE capabilities in design, control, optimization, and safety. The progress of ML and ever-present data collection address previously intractable problems, particularly in system integration and life-cycle modeling. ML-powered predictive algorithms are augmenting traditional control systems, showing potential in supply chain optimization and increasing operational resilience. Additionally, ML-driven fault prediction and diagnostics are enhancing process safety systems, allowing for predictive maintenance and minimizing risks of accidents. A case study of the collaboration between the University of West Attica and Helleniq Energy through the MSc program in Oil and Gas Process Systems Engineering showcases the practical application of advanced technologies and knowledge transfer, emphasizing the importance of blending AI with traditional PSE education.

Optimal and autonomous daily use of new technologies isn’t a reality for everyone. In a societal context driven by sociotechnical transitions [1] many people lack access to digital equipment and skills, preventing their participation in digital social life, including energy services. Our exploratory and phenomenological research, guided by European Union directives [2], explores the social appropriation [3] of new technologies during the deployment of smart meters in Wallonia. The study investigates social behaviour of audiences with support during smart meter installation and identifies barriers to technology appropriation. In an exclusively qualitative approach, the field surveys aim to determine to what extent individual participatory forms [4][5] and collective forms [8][9] of support, through active pedagogies like experiential learning [6][7], can include digitally vulnerable users. The central role of field professionals as interfaces [10] is also highlighted within the service relationship [11] that connects, on one hand, the end consumers and, on the other hand, the organization responsible for deploying the smart meters. While our study does not directly utilize CAPE methods and tools, it investigates the current social implications of digitalization.

The paradigm shift into an era of Industry 4.0, also referred to as the fourth Industrial Revolution, has emphasized the need for intelligent networking between process equipment and industrial processes themselves. This has brought on an age of research and framework development for smart manufacturing in the name of Industry 4.0 [1]. While the physical and digital advancements towards smart manufacturing integration are substantial the inclusion of engineers themselves amongst this shift is often less considered [2]. There are educational efforts in Europe to create and implement smart manufacturing curriculum for non-traditional or adult learners already integrated in the workforce, but attention is also needed on a next generation smart manufacturing curriculum for pre-career students [3]. We, the teaching team of CHE 554: Smart Manufacturing at Purdue University, developed and implemented a curriculum geared towards the training of undergraduate, graduate, and non-traditional students in methods of smart manufacturing as they apply directly to industrial scenarios. Through this elective course, taught primarily through the context of chemical engineering but enrolling students from other disciplines, we introduce concepts not covered in its entirety by the core engineering curriculum. Our course includes but is not limited to material on data reconciliation, machine learning, chemometrics, data-driven fault detection, digital twin development, and process optimization. Further, these concepts are executed through the context of open-source Python packages, enabling the accessible and practical application of smart manufacturing in the form of assignments and in the context of professional application in the future [4,5]. The integration of modern tools and Python libraries connects academic solutions to industrial challenges with industrial practice becoming evident for students who would otherwise be unaware of these resources. With the creation and implementation of the Purdue University Smart Manufacturing elective course, we help develop a uniquely prepared and technologically educated generation of engineering students capable of bringing Smart Manufacturing capabilities to industry. This helps catalyze smart manufacturing knowledge transfer where all enrolled students are given the tools to apply their knowledge mastery to their specific engineering discipline and industrial application.

The growing concern over the environmental impacts of the wine industry has driven the search for sustainable technologies to manage its waste, particularly the residual sludge generated during effluent treatment. This sludge, rich in organic matter, represents a significant source of pollution if not properly treated. However, their energy content allows them to turn this environmental liability into an asset through innovative valorization. Hydrothermal liquefaction (HTL) emerges as a promising technology in this context. This process allows the direct conversion of residual sludge into high-energy-value liquid biofuels. Unlike other treatment methods, HTL can process wet biomass without needing prior drying, making it particularly suitable for managing sludge from wine effluents. Thus, this research aims to evaluate the conversion of residual sludge derived from wine effluent treatment into biofuels through a hydrothermal liquefaction simulation, integrating this process into a sustainable biorefinery for levulinic acid and bioethanol production. The methodology considers the determination of the composition and treatment of the wine effluent to define a case study. The biorefinery processes, as well as the HTL, are designed and simulated in Aspen Plus V.14, and also they are evaluated technically and economically. As a result, levulinic acid, sustainable aviation fuel, bioethanol, propylene glycol, and electrical energy are produced. In addition, the biorefinery reduces the effluent’s Chemical Oxygen Demand (COD) by 99%. In conclusion, valorizing suspended solids from wine effluents through HTL is technically and economically feasible.

The 2,3-, 1,4- and 1,3-butanediols (BDOs) are valuable platform chemicals traditionally produced through petrochemical routes. Alternatively, there is growing interest in synthesizing these chemicals through fermentation processes. However, several drawbacks of the fermentation process (e.g. low product concentration, formation of by-products and high-boiling temperatures of BDOs) hinder the downstream process and increase overall production costs. This original research proposes an advanced large-scale (processing capacity of 160 ktonne/y) process design for the purification of different BDOs after fermentation. The initial preconcentration step removes most water and light impurities in heat pump-assisted distillation column. The heart of the developed process is an integrated dividing-wall column that effectively separates high-purity BDO (>99.4 wt% in all cases) from the remaining impurities. Each BDO isomer was purified cost-effectively (0.208 – 0.243 $/kgBDO) and energy-efficiently (1.854 – 2.176 kWthh/kgBDO) using a single process design, offering flexibility in the development of sustainable bioprocesses for BDO production.

Maintaining consistent product quality and yield in bioprocess operations is challenging due to uncertainties inherent in biological systems. Thus, robust strategies are essential to ensure key performance indicators (KPIs), such as product concentration and yield, are consistently met despite the uncertainties. Real-time feedback co Interntrol, though widely used, is often impractical due to its reliance on expensive sensors, rapid data processing, and high-speed control actions. This paper proposes a novel approach to address these challenges by identifying the operational space for control variables, ensuring KPI reliability without requiring real-time control. This operational space serves as a guideline such that, if we operate within this space, the KPIs can be reliably achieved, regardless of the considered uncertainties. Specifically, we reformulate the problem as an optimization task to maximize the operational space, subject to constraints imposed by process dynamics and performance criteria. Using CasADi for symbolic computation and IPOPT for optimization, a stage-wise procedure is implemented to enhance computational efficiency. Unlike surrogate-based methods, this approach directly preserves model fidelity and accommodates path constraints effectively. The method is validated in a case study of fermentation process, involving a time-varying control variable and seven uncertain model parameters. Results demonstrate the ability to define a reliable operational space that enables KPI adherence across parameter variations, proving the method’s efficiency and practical applicability.

Hybrid modelling utilizes advantageous aspects of both mechanistic (white box) and data-driven (black box) modelling. Combining the physical interpretability of kinetic modelling with the power of a data-driven Artificial Neural Network (ANN) yields a hybrid (grey box) model with superior accuracy when compared to a traditional mechanistic model, while requiring less data than a purely data-driven model. This study demonstrates the construction a hybrid model with transfer learning for the predictive modelling and optimization of a high-cell-density microalgal fermentation process for lutein production. Dynamic optimization was conducted to identify a feeding strategy that maximized final lutein production. The results were then experimentally validated. Overall, this work presents a novel digital twin application that can be easily adapted to general bioprocesses for model predictive control and process optimization.

Microalgae cultivation on liquid digestate from the anaerobic co-digestion of agricultural feedstocks is an interesting option for digestate nutrient removal and resource recovery coupled to value-added biomass production. In this paper, a first-principle plant-wide modelling of the process is described. Two well-established models for anaerobic digestion (IWA – ADM1) and algae-based bioremediation processes (ALBA) were considered and modified with necessary equations and extensions to develop a coherent interface between the state variables of the two models. The resulting system is composed by highly non-linear and non-smooth DAEs. Open-loop scenario analysis for different upstream co-digester design and operating conditions was carried out to assess the impacts on the downstream microalgae outputs. It highlighted the importance of a proper biorefinery design and yet a noteworthy robustness of the system performance. The exploitation of the model can facilitate: a more realistic assessment of the process’s technical-economic feasibility and the design/validation of classical and advanced feedback control strategies.

To match the growing demand for biomethane production, anaerobic digestors need an optimal and time-varying adaptation of the input diet. Dynamic co-digestion constitutes a hard challenge for the limited instrumentation and control equipment typically installed aboard full-scale plants. The development of prediction models is foreseen to support process (optimal) design and control. In this work, a rigorous framework was applied to take full-scale applicability into account while dealing with the design and training of both high-fidelity and control-oriented first-principle/grey-box models, to be used for real-time optimization and process control respectively.

Machine learning algorithms have led to the development of numerous vector embeddings for biological entities such as metabolites, proteins, genes, and enzymes. However, these embeddings often lack contextual information due to their specialized focus on individual omics. Disease progression and biosynthesis pathways are increasingly understood through complex, multi-layered networks that integrate diverse omics data and intricate signaling and reaction sequences. Capturing these relationships in a meaningful way requires embeddings that account for both functional and multi-modal dependencies. We propose an embedding approach that unifies these different biological modalities by treating them as directions in a shared space rather than as isolated data types. Similar to how word embeddings in natural language processing reveal meaningful relationships (e.g., Tokyo – Japan + UK = London, indicating a directional representation of capitals), we can model genes and proteins in a way that captures their inherent connections. A gene implies information about the protein it encodes, and vice versa, forming a structured and interpretable representation of biological pathways. Our model, inspired by NLP techniques, breaks down pathway sequences into contextual pairs spanning different omics types. By aligning pathway steps in proximity, the embeddings reflect biologically relevant relationships, enhancing their interpretability and utility. Because these embeddings are generated from pathway sequences, they can be applied to optimize reaction pathways, aiding retrosynthesis in microbiomes, drug development, and even human health interventions.

The European Union is transitioning towards a circular and low-carbon economy, emphasizing renewable biological resources. This study explores the production of high-value compounds like chitosan from fungal biomass and presents a potential design for a sustainable biorefinery process, contributing to the diversification and optimisation of biomass feedstock utilisation. The process simulation includes dedicated sub-models for each unit operation, based on laboratory data and integrated into a comprehensive process flow sheet using COCO-COFE. The productivity of the simulated plant results in 2 500 tons of triglyceride oils and 1 800 tons of chitosan that can be produced from 15 000 tons of Aspergillus niger. On-site acetic acid production meets 45% of the total plant’s demand, significantly reducing the amount of additional acetic acid to be purchased as raw material. Additionally, large-scale enzyme consumption and the substantial heat demand for biomass processing are key economic and environmental factors that need to be central in successful process design.

In recent years, burgeoning interest in products derived from microbial production across various sectors has significantly propelled the evolution of the field of metabolic engineering. As a Gram-negative bacterium, Vibrio natriegens is characterized by its fast growth, robust metabolic capabilities, and a broad substrate spectrum, making it a promising candidate as a standard biological host for the industrial bioproduction of metabolites. Genome-scale metabolic models (GSMMs) are mathematical representations constructed based on genome annotations and gene-protein-reaction (GPR) associations within a cell. These models enable the computational simulation of intracellular reaction flux distributions. In this study, we developed a hybrid method based on metaheuristic algorithms and the GSMM to optimize metabolism for the production of ethanol and 1,3-propanediol (1,3-PDO) as target products in Vibrio natriegens. The modified GSMM used in this study contains 1195 reactions, 1094 metabolites, and 880 genes, with the metaheuristic algorithms employed being Genetic Algorithm (GA) and Particle Swarm Optimization (PSO). The optimization results indicate that the method proposed in this study can achieve the production of ethanol and 1,3-PDO by adjusting the expression levels of just 2-4 genes, with the production of 1,3-PDO reaching its theoretical maximum. The strategies developed in this study can effectively increase the production capacity of specific target metabolites in Vibrio natriegens, serving as a starting point for metabolic engineering and providing guidance for metabolic engineering targets in practical experiments.

Pharmacokinetic and pharmacodynamic (PK/PD) models are used to predict drug transport in the body and to assess treatment efficacy and optimal dosage. The kinetic parameters embedded in the models, which define transport across body compartments or drug efficacy, can be linked to patient-specific characteristics; understanding the parameter space-model output relationship is critical towards linking patient population heterogeneity to the therapeutic outcome variability. Global Sensitivity Analysis (GSA) is a well-established tool used to examine parameter-to-parameter interactions, shedding light on underlying interactions towards enhanced system understanding. Despite its potential and usefulness, GSA performance is dependent to the model complexity; large-scale and nonlinear PK/PD models, which often have large sets of parameters, can render GSA challenging to perform, requiring excessive computational effort. Proposed approaches to reduce GSA complexity, such as segmentation in parameter subsets or the introduction of surrogate metamodels, become less effective as the number of kinetic parameters grows. In this work, we investigate the potential of Machine Learning (ML) to reduce the complexity of PK/PD models by exploring how the level of hybridisation can impact the GSA performance and, critically, whether the use of surrogates affects the resulting model sensitivity to parametric uncertainty. We show that ML-based surrogates can reliably identify parameter interactions and sensitivities while requiring only a limited number of simulations of the reference mechanistic model. Further, surrogates models effectively reduce the computational expenditure of GSA of multi-dimensional nonlinear PK/PD models. The accelerated execution of GSA enables performing patient cohort-specific analysis, with potential applications for optimal study design and for precision medicine.

Incorporating process knowledge from various sources often presents challenges in process development, optimization, and control. To utilize available knowledge, linking existing process mo-dels is a viable approach. This work introduces a methodology using latent variable models, specifically sequential and orthogonalized partial least squares (SO-PLS), to capture and quantify the contribution of first-principles knowledge in process models. Applied to a continuously stirred tank reactor (CSTR) case study, the methodology demonstrates how available knowledge can be quantified and how structural and parametric errors in first-principles are addressed using measured data. The methodology is discussed in relation to root cause analysis in bioprocesses.

Due to the importance of unsaturated fatty acids for human health and the increasing global demand in the food and food crop area, oleaginous yeasts are promising alternative microorganisms for commercial lipid production due to the high volumetric productivity, with Metchnikowia pulcherrima being an underexplored oleaginous yeast with potential as a lipid producer. Critical to achieving high productivity lipid production are nutrient factors. A sensitivity test identified carbon and nitrogen sources as important factors in nitrogen limited broth (NLB) for lipid production in M. pulcherrima i.e. glucose, yeast extract and Ammonium sulphate. Response Surface Methodology (RSM) involving sets of 15 experimental runs of three-factor three-level Box-Behnken Design (BBD) was implemented for exploring the design of the carbon and nitrogen source in the growth media composition. Quadratic surfaces were least-square fitted and used to identify regions of optimal lipid yield. Multiple sets of runs were conducted with the parameter range progressively adapted and repeated until a clear optimum was identified. The highest total lipid production was in the low carbon concentration range (2.27-21.5 g/L) which suggests the process is more productive with the optimised media compared to NLB media (30.4 g/L of carbon). The optimal carbon concentration was 14.8 g/L whereas the dependence on nitrogen was not found to be significant. The yield of optimal point (YP/S) was 2.39 times higher than NLB media after validation because the depletion of glucose at the end of fermentation (72-104 h) contributes the high increase of lipid accumulation (mg Lipid/ 100 ml culture) of the optimal point (94.5%) which was higher than NLB media (34.5%) for 60%.

The extraction of lycopene from tomato waste has been largely evaluated at an experimental level, leading to the creation of polynomial models or response surfaces that allow the representation of the extraction behavior. However, these studies are based on laboratory level and an extraction process has not yet been scaled up. This study evaluates the design and optimization of the lycopene extraction process from tomato waste. The proposed model is solved through a link between Python and Aspen Plus, performing the optimization a genetic algorithm (GA) in Pymoo. The minimum value of TAC is 211,692.2 USD/yr, corresponding to a production of 2.29 g/h of lycopene, starting from 1000 kg/h of tomato waste. This work represents a first approach to the design of a commercial-scale lycopene production process.

Optimizing greenhouse operations in arid regions is essential for sustainable agriculture due to limited water resources and high energy demands for climate control. This paper proposes a multi-objective optimization framework aimed at minimizing both the operational costs and environmental emissions of a climate-controlled greenhouse. The framework determines optimal allocation of growing area among three crops (tomato, cucumber, and bell pepper) throughout the year. These crops were selected for their varying growth conditions, which induce variability in energy and water inputs, providing a comprehensive assessment of the optimization model. The model integrates factors such as temperature, humidity, light intensity, and irrigation requirements specific to each crop. It is solved using a genetic algorithm combined with Pareto front analysis to address the multi-objective nature effectively. This approach facilitates the identification of optimal trade-offs between cost, emissions, and yield, offering a set of efficient solutions for decision-makers. Applied to a greenhouse in Qatar, the model evaluates two scenarios: a hydroponic system and a conventional soil-based system. Results indicate that the multi-objective optimization effectively reduces operational costs and environmental emissions while fulfilling crop demand.The hydroponic scenario demonstrates higher water-use efficiency, resulting in lower overall emissions despite higher initial setup expenses. The soil-based system shows higher emissions due to increased fertilizer use and water consumption. The optimized allocation balances multiple crops simultaneously throughout the year based on seasonal conditions, enhancing overall sustainability. This study underscores the potential of advanced optimization techniques in enhancing the efficiency and sustainability of greenhouse agriculture in challenging environments.

Water lily (Eichhornia crassipes) has been identified as an invasive exotic plant with high proliferation in Mexico, affecting aquatic bodies, such as lakes. After extraction, the water hyacinth biomass can be used as raw material for the production of bioproducts and bioenergy, however, the majority of them not covered the region’s needs, and their economic profitability decreases significantly. Also, few reports present its use as raw material inside a biorefinery scheme. In this work, we propose a local biorefinery scheme to produce power and bioproducts from water lilies, using Aspen Plus V.10.0, per the needs of the Patzcuaro Lake community in Michoacán, Mexico. The scheme has been designed to process the harvested and sun-dried water lily from 197.6 kg/h of total wet harvested biomass, according to the extraction region schedule. The biomass is separated: root (RT) and stems-leaves (SL). The processing scheme involves the RT combustion to produce electric power, and two processes to SL; the first one, anaerobic digestion to produce fertilizer and biogas, while, pyrolysis and gasification as second one. The biogas is converted to electricity, while, the pyrolysis products are separated in bio-oil and biochar; the last one is inserted to gasification reactor to produce syngas, and then, electric power. According with results, this scheme produces 35.6 % of fertilizer, 1 % of bio-oil, and 3.5 kWh of electric power. The scheme requires energy efficiency tools to improve its energetic performance, however, is a potential solution for the use of water lily in the region.

Predicting water quality is essential for effective environmental management and pollution control. Dissolved oxygen (DO), one of key water quality parameters, plays a vital role in biological wastewater treatment [1]. This study aims to forecast DO levels in activated sludge tanks of an oil refinery’s wastewater treatment plant (WWTP). Proper oxygen concentration is critical for microbial activity, as inadequate levels can disrupt the biological breakdown of pollutants. The objective is to develop predictive models to identify operational risks early, enhancing treatment efficiency and optimizing resources like chemicals, bacterial cultures, and aeration systems. Additionally, the study aims to provide early warnings to operators, minimizing reliance on laboratory tests and ensuring optimal conditions for bacteria, leading to better operational performance, cost reduction, and improved water quality ultimately promoting sustainable wastewater treatment. Various deep learning models, including Recurrent Neural Network (RNN), Gated Recurrent Unit (GRU), and Long-Short Term Memory (LSTM), were applied to a two-year real-time dataset with 24 features and 8-hour intervals. Model performance was assessed using process data for training, validation, and testing. The results reveal that GRU-based models effectively predict DO concentration in sludge tanks, achieving an R² of 0.7, MSE of 0.01, and MAE of 0.07. These soft sensors demonstrate strong potential for system control.

Understanding the complexity of human digestion is critical for designing models that serve as valuable research tools for process simulation and prediction. Due to the high cost of medical intervention & recent advancements in in vitro digestion protocols, increased demand for inexpensive in silico solutions emerges. This study aims to develop a mathematical model that simulates the in vitro dynamic Food Digestion SIMulator (FooDSIM) functionalities via a digital twin approach. Ordinary Differential Equations (ODEs) simulate the system as a series of Continuously Stirred Tank Reactors (CSTRs) and describe different regions of human organs (stomach, duodenum, ileum, colon) of the human Gastrointestinal Tract (GIT). Various time horizons were used to investigate the effect of periodic feeding on the dynamic stabilisation of the inherently simulated processes (hydraulics, pH, biochemical interactions between enzymes & substrates, and nutrient absorption). A Polynomial Chaos Expansion (PCE)-based Global Sensitivity Analysis (GSA) was performed to explore causal relationships between input & output variables by performing uncertainty quantification. Results indicate that the system self-stabilizes only after 3-4 feeding cycles. Initially, enzyme concentration increases in the duodenum above the input feed threshold, thus significantly affecting the digestion kinetics in the subsequent reactors. Two cardinal models used for the fitting of enzymatic activity data with pH values indicate satisfactory Mean Squared Error (MSE) scores (<0.0114 for Model 1, <0.0134 for Model 2). Overall, this research highlights the potential of mathematical modeling for simulating digestion and predicting experiment outcomes.

Effective prediction of nutrient demands is crucial for optimising microalgae growth, maximising productivity and minimising the waste of resources. With the increasing amount of data related to microalgae cultivation systems, data mining and machine learning models to extract additional knowledge have gained popularity. In the development of such models, a data preprocessing stage is necessary due to the poor data quality. At this stage, cleaning and outlier removal techniques are employed to eliminate missing data and outliers, respectively. Afterwards, data splitting and cross-validation strategies are employed to ensure that the models are trained and evaluated with representative subsets of the data. Principal component analysis is also applied to simplify complex environmental datasets by reducing the number of features while retaining as much information as possible. To further improve prediction capabilities, ensemble methods are incorporated, leveraging multiple models to achieve a higher overall performance. Stacking, a popular ensemble technique, is used to combine the outputs of individual models into a single meta-model. The application of these machine learning methods has been demonstrated using a dataset acquired from the cultivation of the microalgae Dunaliella in a flat-panel photobioreactor. The results have shown that the data mining workflow, in combination with different machine learning models, was able to describe the nutrients’ requirements enforcing good performance of microalgae Dunaliella ?-carotene production in the carotenogenic phase.

Precision modeling and forecasting of soil moisture are essential for implementing smart irrigation systems and mitigating agricultural drought. Most agro-hydrological models are based on the standard Richards equation, a highly nonlinear, degenerate elliptic-parabolic partial differential equation (PDE) with first order time derivative. However, research has shown that standard Richards equation is unable to model preferential flow in soil with fractal structure. In such a scenario, the soil exhibits anomalous non-Boltzmann scaling behavior. Incorporating the anomalous non-Boltzmann scaling behavior into the Richards equation leads to a generalized, time-fractional Richards equation based on fractional time derivatives. As expected, solving the time-fractional Richards equation for accurate modeling of water flow dynamics in soil faces extensive computational challenges. To target these challenges, we propose a novel numerical method that integrates finite volume method (FVM), adaptive fixed point iteration scheme, and neural network to solve the time-fractional Richards equation. Specifically, we develop an adaptive fixed point iteration scheme to solve the FVM-discretized equation iteratively, which avoids the stability issues when directly solving a stiff and sparse matrix equation. To improve the solution quality which is influenced by numerical errors and computational constraints during actual implementation, we propose to use neural networks that resemble an encoder-decoder architecture to map soil moisture profiles into a latent space and reconstruct them back. Through 1-D examples, we illustrate the accuracy and computational efficiency of our proposed physics-based, data-driven numerical method. Finally, we present a Markov chain Monte Carlo (MCMC) approach to solve the inverse problem to obtain soil-specific parameters given soil moisture solutions.

Miniaturised stirred bioreactors are crucial in high-throughput bioprocesses for their simplicity and cost-effectiveness. To accelerate process optimisation in chemical and bioprocess industries, models that integrate CFD-predicted flow fields with (bio)reaction kinetics are needed. However, conventional two-step coupling methods, which freeze flow fields after solving hydrodynamics and then address (bio)reaction transport, face numerical challenges in miniaturised systems due to unsteady radial flows, recirculation zones, and secondary vortices. These flow fluctuations prevent steady-state hydrodynamic convergence. This study addresses these challenges by time-averaging the RANS solutions of the transitional SST model to achieve statistical hydrodynamic convergence. This method is particularly effective for internal flow problems at low to midrange Reynolds numbers (<10,000), typical of transitional regimes. Following this, photo-bioreaction transport models were solved using these converged fields, considering the bioreactor’s directional illumination and curvature. Applied to a 0.7 L Schott bottle photobioreactor mechanically mixed by a magnetic stirrer (100-500 rpm), the model predictions were thoroughly validated against tracer dye experiments and Rhodopseudomonas palustris biomass growth data. It accurately predicted swirly vortex fields at 500 rpm with a 7% error margin and the biomass growth profiles align to literature datasets. However, parallel computing efficiency did not scale linearly from 16 to 32 processor cores (4G of RAM/core), making time-averaging computationally expensive for simulating scale-up bioreactors. Improved bioreactor mixing influenced cell light/dark cycles and enhanced biomass productivity but stirring speeds above 300 rpm required increased light intensity (>100 W/m²) due to light limitation. This model provides a framework for optimising stirring speeds and refining operational parameters, aiding in the scale-up and scale-down of bioprocesses.

Most industrial fermentations in food and drink use a single, high purity sugar as a substrate. These pure substrates are more expensive and less sustainable than mixed substrates, that can be derived from agricultural byproducts such as straw. However, use of mixed substrates in fermentation leads to challenging modelling and parameter estimation problems, particularly when much academic research, intended to inform industrial applications, uses batch fermentations, while large-scale fermentation is usually continuous, thanks to its cost and productivity advantages. Our findings highlight key challenges in using batch-derived experimental data to inform models of the continuous fermentation processes used at industrial scale. Extrapolating from data obtained in batch to continuous fermentation is risky, as models with near-equivalent data-fit and predictions in a batch context give very different predictions for continuous culture. For continuous fermentations to switch to mixed substrates, we need to improve understanding of dual substrate growth in continuous fermentation, to allow for optimised process design and operation.

Global biomanufacturing is projected to expand rapidly in the coming decade due to advancements in DNA sequencing and manipulation. However, the complexity of cellular behaviour introduces difficulty in modelling and optimizing biomanufacturing processes. Phenomenological models that represent the physics of the system in empirical equations suffer from poor robustness, while their machine learning (ML) counterparts suffer from poor extrapolative capability. On the other hand, hybrid models allow us to leverage both physical constraints and the flexibility of ML. This work describes a new approach for hybrid modeling that integrates the time-variant parameter estimation and ML model training into a singular step. We implement this approach on a proposed scheme for the cell-mediated conversion of a lignin derivative into a bioplastic precursor and show that our integrated hybrid model outperforms the traditional two-step hybrid, phenomenological, and ML model counterparts. Lastly, we demonstrate how to execute an interpretability analysis on the ML component of the integrated hybrid model to reveal new physical insights that are then used to further improve model performance.

The use of metabolic networks is extremely valuable for design and optimisation of bioprocesses as they provide great insight into cellular metabolism. Within bioprocess optimisation, they have enabled better (economic) objective performance through more accurate network-based models. However, one of the drawbacks of using metabolic networks is their underdeterminacy, leading to non-unique flux distributions. Flux Balance Analysis (FBA) reduces this issue by making assumptions on the behaviour of the cell. However, for metabolic networks of higher complexity, can still struggle with underdeterminacy. Metabolic network reduction can remove or greatly reduce this effect but can be difficult, especially when data is limited. Structural analysis of the metabolic network through Elementary Flux Modes (EFM) or Extreme Pathways (EP) can help locate the relevant information within the network. This work presents a metabolic network reduction approach based on the EPs that best explain a small set of available measurements. Many of the reactions will not be active during the process and a significantly smaller network can therefore be constructed. A case study of oxygen-limited Escherichia coli is presented which showcases this approach, enabling accurate prediction of the process with much smaller network-based models. This leads to much lower complexity bioprocess models while keeping the necessary information on cellular metabolism for the given process.

Renewable energy and energy efficiency are increasingly recognised as crucial for creating new economic opportunities and mitigating environmental impacts. Anaerobic digestion (AD) transforms organic materials into a clean, renewable energy source. Co-digestion of various organic wastes and energy crops addresses the disadvantages of single-substrate digestion, increasing production flexibility yet adding process complexity and sensitivity. This study employs a two-pronged approach to optimise biogas production while considering global warming potential: a nonlinear programming (NLP) model for dynamic system economic optimisation with a model predictive control (MPC) strategy for precise temperature regulation within the digester. The NLP model integrates a combined heat and power (CHP) system to leverage dynamic electricity, heat, and gas prices, accounting for physical and economic parameters such as biomethane potential, chemical oxygen demand, and substrate density. A cardinal temperature and pH model ensures accurate depiction of substrate degradation and gas production rates under varying conditions. The MPC scheme, formulated as a system of differential-algebraic equations, offers fine-grained temperature control, capturing real-world complexities like heating/cooling delays, ambient conditions, and multiple feed components with different optimal digestion temperatures. Results demonstrate that this integrated model optimises the interaction between electricity production, biogas generation, and CHP operation for real-time multi-objective optimisation of profit, global warming potential and temperature control. A case study validates the model’s capability for guiding decision-making in biogas facilities, emphasising strategic feedstock management and precise temperature control. Overall, this integrated approach advances the modelling and control of anaerobic co-digestion systems, enhancing both efficiency and profitability in biogas production.

The global reliance on fossil-based solvents has driven the search for sustainable alternatives. This study employs the IBSS® CAMD tool to evaluate building blocks derived, directly or indirectly, from agricultural residues – specifically orange and coffee wastes-, to replace toluene in furfural extraction. A three-stage methodology was implemented: (1) identification of potential building blocks from residues, (2) multi-objective optimization using genetic algorithms and group contribution models for properties calculation, and (3) analysis of the resulting candidates based on performance indicators. A total of 13 families were evaluated, generating millions of candidates. Target properties included minimization of Hansen Solubility Parameters (HSP) distance, boiling point above 250°C, melting point below 10°C, flash point above 61°C, and octanol-water partition coefficient (log(kow)) below 3. The most promising candidates were derivatives of glycerol (performance: 0.9986), limonene (0.9937), and furfural (0.9931), which exhibited low bioaccumulation potential, non-flammability, and liquid phases at typical usage temperatures. Notably, all the top-performing molecules derived from these building blocks outperformed toluene (performance: 0.6724) in terms of the selected properties. The findings highlight the dual benefits of addressing agricultural waste management challenges while advancing sustainable industrial practices. Further studies are required to confirm their stability and scalability for industrial applications.

Soybean is a highly valuable global commodity due to its versatility and numerous derivative products. During harvest, all non-seed materials become “straw”. Currently, this waste is primarily used for low-value purposes such as animal feed, landfilling, and incineration. To address this, the present work proposes a conceptual biorefinery aimed at converting soybean straw into higher-value products. The study began with data collection to identify potential conversion routes. Based on this information, a superstructure was developed, comprising seven conversion routes: four thermochemical routes (pyrolysis, combustion, hydrothermal gasification, and liquefaction), two biological routes (fermentation and anaerobic fermentation), and one chemical route (alkaline extraction). Each process was evaluated based on product yields, conversion times, and associated capital and operating costs. Using this data, an MILP (Mixed-Integer Linear Programming) optimization model was built in Pyomo using CPLEX as the solver. The variables in this problem were the amounts of biomass processed by each conversion process. The objective function was to maximize the Net Present Value (NPV). The optimization considered a maximum raw material supply of 232,500 metric tons per year. As a result, pyrolysis was identified as the most profitable route, yielding US$447 million, producing biochar, bio-oil and gas as products. It was observed that only if the price of the optimum product was 46% lower would the optimum conversion process change. In conclusion, soybean straw offers significant potential for value-added applications in a biorefinery context, with pyrolysis emerging as the most profitable route.

This study provides standardized models for the chemical characterization of complex streams, ensuring the necessary adaptations while considering the differences in biomass types and forms. Several datasets are compiled and examined to establish a valid representation of the mixture, according to industry accepted standards and laboratory protocols. For reliable property estimation, correlations of key biomass properties are obtained from both computational models and experimental measurements. Existing data are used to create datasets for the biomass and the biocrude streams. This model builds upon existing knowledge and data technologies with emphasis on hydrothermal liquefaction (HTL). The proposed approach shows potential as a starting point for the design and modelling of more biorefinery-associated technologies. Sludge and pine wood are used as case studies for biomass feedstocks. Two biocrude samples are employed for biocrude characterization. The performance of the developed optimization model is compared favorably with results obtained using previous works. To ensure a variety of possible and valid solutions, integer cuts are implemented producing solution pools for the analysed streams.

Polyhydroxyalkanoates (PHAs) are microbial bioplastics that are fully biodegradable, biocompatible and can be produced by renewable feedstocks through fermentation. These are all desirable attributes for the replacement of current fossil-based plastics. Strong mathematical models describing bioprocesses are invaluable tools that can be used for enhancing bioprocess understanding as well as optimization. In this study, polyhydroxybutyrate (PHB), by Cupriavidus necator DSM 545 was produced using glycerol and ammonium sulphate (AS) as the sole carbon and nitrogen sources, respectively. In addition, a kinetic bioprocess model was developed. The kinetic parameters of the model were calibrated with five fermentation experiments with different initial conditions (e.g. variable glycerol and AS concentrations) in order to properly establish the inhibition regions and provide a generalized model as much as possible. The model was successfully validated by three independent experiments, two with extreme and one with moderate initial conditions. In addition, the analysis of the data provided valuable metabolic insights for this bioprocess. It was shown that the PHB storage capacity of cells strongly correlates with the initial nitrogen concentration in the medium, meaning that efforts for maximizing biomass production in a two-stage bioprocess strategy should be carefully designed so that the nitrogen concentration does not irreversibly compromise the maximum PHB accumulation potential. Furthermore, a preliminary kinetic analysis of the biodegradation of solvent casted PHB films was performed.

The Fourth Industrial Revolution (Industry 4.0) brings a new chapter at dairy sector. Dairy 4.0 technologies are based on Big Data Analysis, Internet of Things, Robotics and Machine Learning. The usage of smart technologies to processing and analyzing complicated massive data has a significant impact in automation, optimization, functional costs and innovation. Artificial Intelligence tools are applied from dairy farms and production lines – including packaging- to supply chain. The aim of this paper is to demonstrate the most used applications of Machine Learning in dairy production so as to enhance the sustainability and the quality of dairy products. The most significant Machine Learning applications integrate machine vision, smart environmental sensors, activity collars, thermal imaging cameras, and digitized supply chain systems to facilitate inventory management. Challenges like milk adulteration, animal diseases, mastitis, traceability and supply chain losses are also addressed since they increase safety and quality standards in the dairy sector.

The scale-up of processes with complex fluid flow presents significant challenges in process engineering, particularly in fermentation. Computational fluid dynamics (CFD) is a crucial tool for accurately modelling the hydrodynamic environment in bioreactors and understanding the effects of scale-up. This study utilizes Higher Order SVD (HOSVD), which is the multidimensional extension of Singular Value Decomposition (SVD), to identify the dominant structures (modes) of fluid flow in CFD data of fermentation process simulations. Similarly to Proper Orthogonal Decomposition (POD), also based on SVD, this method can be used to identify the dominant structures of fluid flow, and additionally explore the scale parameter space. As a first test case, we examined five scales of a reciprocally shaken flask bioreactor, from 125 mL to 10 L, specified using basic empirical scale-up rules. Results indicate a common set of spatial modes across all scales, thus confirming that the scale-up method assures some dynamic similarity. However, the relative importance of these common spatial modes changes sensibly across process scales. The main spatial mode that represents the large recirculating flow in the flask can account for 19.8% of the data variability for the 125 mL scale but only 11.7% for the 10 L scale case. In this larger scale, one needs a larger number of modes to capture flow dynamics, thus showing an increase in flow complexity. These findings illustrate the impact of scale on fluid dynamics in a particular bioreactor flow but also provide a proof of concept for this methodology.

Extended Dynamic Mode Decomposition (EDMD) and its variant, the pqEDMD, which uses a p-q-quasi norm reduction of polynomial basis functions, are attractive tools to derive linear operators approximating the dynamic behavior of nonlinear systems. This study highlights how this methodology can be applied to data-driven modeling and control of bioprocesses by discussing the selection of several ingredients of the method, such as the polynomial basis, order, data sampling, and preparation for training and testing, and ultimately, the exploitation of the model in linear model predictive control.

Large quantities of used cooking oil (UCO) are produced globally, primarily in densely populated urban centers. Although UCO is highly heterogeneous due to degradation during cooking, it still contains a significant fraction of triacylglycerols (TG) that could be used as raw materials in oleochemical biorefineries. A major challenge in reintegrating this residue into productive cycles is the presence of free fatty acids (FFA), which can affect subsequent catalytic or enzymatic transformations. Conventional processes for FFA removal are energy-intensive, require alkaline feedstocks, and generate problematic residues. To overcome these issues, alcoholic extraction of FFA is considered a promising pretreatment for UCO, enabling the extraction of FFA for subsequent esterification. In this regard, liquid-liquid film contactors (LLFC) have shown potential to intensify FFA extraction because they operate under mild conditions and at laminar flow regime, reducing energy consumption and enhancing productivity. This work presents the modeling and validation of a FFA extraction process suitable for deacidifying UCO using a LLFC with ethanol as a solvent. Experimental data were used to fit model parameters, which were optimized using a genetic algorithm. The validated LLFC model revealed that increasing the UCO mass flow rate and the ethanol-to-UCO ratio enhanced FFA mass transfer. Furthermore, a cascade configuration of three LLFCs was able to reduce the FFA content in UCO to 0.9 wt.%. It is anticipated that this technology will enable the pretreatment of UCO for further transformation into value-added oleochemicals and facilitate the exploitation of recovered FFA via subsequent ethanolic esterification.

This study addresses the challenge of predicting final properties in batch processes with highly uneven durations, using the ibuprofen production process as a case study. Novel methodologies are proposed and compared against traditional regression algorithms, which rely on batch trajectory synchronization as a pre-processing step. The performance of each method is evaluated using established metrics. The data for this study were generated using Aspen Plus V12 simulation software, focused on batch reactors. To handle the unequal-length trajectories in batch processes, this research constructs a dual-transformer deep neural network with multi-head attention and layer normalization mechanism to extract shared information from the high-dimensional, uneven-length manipulated variable profiles into latent space, generating equal-dimensional latent codes. As an alternative strategy for representation learning, a dual-autoencoder framework is also employed to achieve equal-dimensional representations. The representation vectors are then used as inputs for downstream deep learning models to predict the target variables, achieving an accuracy with an R² score exceeding 0.9.

Maintaining sufficient amounts of dissolved oxygen throughout a microbial cultivation is a classic control task in bioprocess engineering to avoid negative effects onto cell physiology and productivity. But traditional PID-based algorithms struggle when faced with pulsed substrate additions and the resulting sudden surge of oxygen uptake. In this work a nonlinear MPC is employed and compared to a PID setup for the cultivation of an E. coli strain exposed to intermittent feeding. Both controllers are tuned for a fast pulse response combined with efficient and robust control action. Their performance was tested in-silico with isolated feed pulses, as well as throughout a full cultivation run. Further, the effects of parameter uncertainty were investigated to assess the impact of a model-plant mismatch. The results showed that the predictive nature of the MPC is well suited for maintaining the dissolved oxygen levels above a threshold and outperforms the PID in almost all investigated simulation scenarios.

This work presents a hybrid-modeling approach to monoclonal antibody (mAb) production processes design using automated bioreactor equipment. Experimental data covering a reasonable yet broad range of cultivation conditions was collected by the equipment. Using the data, a model applicable to a wide range of cultivation conditions was developed. In the modeling, a data-driven model was applied to describe complicated/unknown phenomena that could not be captured by previously proposed mechanistic models. In the hybrid model, while maintaining the mass balance of the mechanistic model, coefficients of the equations were estimated with random forest regression. Overall, the model could describe the dynamic concentration profiles of product mAb and quality-relevant impurities depending on the media/glucose feeding conditions. The model was then applied to determine an optimal condition that maximized product mAb concentration and satisfied the impurity constraints. The work can further support model-based design of cell cultivation processes with considering multi-input & multi-output nature.

This work presents a kinetic model of drug substance synthesis considering slug flow characteristics in Stevens oxidation. The developed model is also applied to determine the feasible range of the process parameters. Flow experiments were conducted to obtain kinetic data, varying the inner diameter, temperature, and residence time. A kinetic model was developed for the change in concentrations of the starting material, products, and catalysis. In the kinetic model, slug flow was considered by including a volumetric mass transfer coefficient during this flow. In the initial experiments, early-stage kinetic data were insufficient, conducting additional experiments at shorter residence times. Furthermore, the initial model could not reproduce the residual of the starting material, introducing the oxidant consumption that inhibits the starting material consumption and improving the initial model. The improved model could reproduce experimental results and demonstrated that, as the inner diameter increases, the efficiency of mass transfer in slug flow decreases with slowing down the reaction. Moreover, the improved model was considered applicable to different scales. The developed model was used to simulate the yields of the desired product, and the dimer, and the process mass intensity, thereby determining the feasible range. As a result, it was shown that when methanol and oxidant concentration was either too high or too low, operating conditions fell outside the feasible range. This kinetic model with flow characteristics will be useful for the process design of drug substance synthesis using liquid-liquid reactions.

Biomacromolecules have intricate crystallisation behaviour due to their size and many interactions in solution and can often only crystallise in narrow ranges of experimental conditions. High solute concentrations are needed for crystal nucleation and growth, exceeding those eluted upstream and therefore preventing the adoption of crystallisation in downstream separation steps. By promoting molecular aggregation and nucleation via a lowered energy barrier, heterogeneous surfaces or templates can relax the supersaturation requirements and widen the crystallisation operating space. Though templates are promising candidates for process optimisation, their experimental testing has generally been limited to small-volume experiments, and quantification of their impact on process intensification and quality metrics at higher volumes remains unexplored. To address the knowledge gap, a model-based investigation of template-induced protein crystallisation systems through evaluation of key metrics is presented. Porous silica nano-particles with three chemical functionalisations (hydroxyl, carboxyl and butyl) are added to batch lysozyme crystallisation experiments at 40ml. Crystallisation population balance models are parametrised with an experimentally-validated parameter estimation methodology and further experiments are simulated. The templates appear to lower the estimated interfacial energy compared to the homogeneous case, leading to nucleation rate profiles which are less dependent on supersaturation. For this reason, the templates can crystallize quicker than the homogeneous system, particularly at lower initial concentrations. The simulation results highlight the ability of heteronucleants to alter nucleation rate profiles, and their potential to be used as process optimisation and intensification tools for biomacromolecule purification.

Choosing suitable solvents for crystallisation processes can be a challenging task when developing new pharmaceuticals, given the vast number of candidates available. To streamline this task, however, process modelling tools can be used to systematically probe the behaviour of different crystallisation setups entirely in-silico. In fact, it is possible to couple thermodynamic models with basic solid-liquid equilibria (SLE) principles to determine the impact of key process variables (e.g., temperature and solvent choice) on process performance, prior to conducting lab-scale experiments. In light of this, in this study we have used thermodynamic computational modelling tools (implemented within MATLAB®) to evaluate the cost and environmental impact of different batch crystallisation processes that may be used to manufacture flurbiprofen – a non-steroidal anti-inflammatory drug (NSAID) that can be used to treat various forms of arthritis. To complete this work, we have used the Apelblat equation to study the solubility of flurbiprofen in four (4) common solvents (n-octane, methyl tert-butyl ether, n-propanol, and isopropyl acetate) over a wide temperature range (283.15-323.15 K). Further to this, we have used green metrics (i.e., E-factor) and established costing methodologies to compare each process candidate; before evaluating their Scope 1 and 2 CO2e emissions by considering their electricity usage and incineration activities.

The increased use of pharmaceuticals globally over the past two decades has contributed to an increase in unused pharmaceuticals and a corresponding surge in pharmaceutical waste. Thus, there is an impetus for the development of processes for the recovery of the active pharmaceutical ingredients (APIs) from these unused drugs. This study introduces a decision framework for solvent selection to enable the recovery of APIs using a general separation train where cooling crystallization is the final step. The framework is designed to base solvent selection not just on the solubilities of the formulation contents but also considers the overall recovery that can be achieved in the process. In addition, the environmental sustainability of the framework is analyzed using the process mass intensity metric (PMI). The effectiveness of this framework is demonstrated by using paracetamol (PA) as a model API in a formulation consisting of five of the excipients commonly found in PA formulations. The results show that selecting a solvent based on the dissolution of pharmaceutical tablets may not be the best possible approach, whereas choosing a solvent that maximizes the theoretical recovery is preferred.

Emerging sectors within the biopharmaceutical industry are undergoing rapid scale-up due to the market boom of gene therapies and vaccine platform technologies. Manufacturers are pressured to orchestrate resources and plan investments under future demand uncertainty and, critically, an early-stage process uncertainty for platforms still under development. In this work, a multi-product multi-stage stochastic optimization problem integrating demand uncertainty is presented and augmented with a worst-case optimization approach with respect to process uncertainty. Results focus on a comparison between fixed equipment facilities and modular technologies, highlighting an inherent flexibility of the latter option due to shorter recourse actions for capacity scale-out. The impact of process uncertainty integration is quantified. With more conservative decisions taken in first-stages of the time horizon, expected costs result lower for modular single-use equipment. This suggests that capacity adjustments also help adapt to varying process performance and reduce the propagation conservative design decisions.

This work introduces a Python-based interface that couples cradle-to-gate Life Cycle Assessment (LCA) with advanced process simulations in continuous biomanufacturing, resulting in dynamic process inventories and thus to dynamic LCA (dLCA). The open-source Brightway2.5 framework is used to dynamically track environmental inventories of the foreground process and LCA indicators (e.g. damage to ecosystems according to ReCiPE 2016) from the v3.10 cut-off ecoinvent database. The framework is applied to KTB1, a dynamic MATLAB–Simulink benchmark model of continuous Lovastatin production. 580 data points are computed across four different 24-hour scenarios. The difference between the hourly and the averaged foreground scenario is between 20-30%; a more pronounced deviation is observed when both background and foreground are averaged. The dLCA framework precisely identifies optimal periods for cleaner electricity usage, enabling future work on direct environmental feedback into process control and optimization for greener high-quality biomanufacturing.

Reactive crystallization (RC) is a chemical process in which the reaction yields a crystalline product. It is used in various industries such as pharmaceutical manufacturing or water purification. In some cases, RC is the only feasible process pathway, such as the precipitation of certain ionic solids from solution. In other cases, a reaction can become a RC by changing the reaction environment to a solvent with low product-solubility. Despite the utility and prevalence of RC, it is not often emphasized in process design software. There are RC models that simulate the inner reactions and dynamics of a RC, but each has limiting assumptions, and are difficult to integrate with the rest of a process-line simulation. This modeling gap complicates RC process design and limits the exploration of the possible benefits to using RC as well as the ability to optimize a system that relies on it. To fill this gap, we built an open-source, customizable model that can be integrated with other unit operations in the Python process simulator package PharmaPy. This model focuses on the reaction-crystallization interactions and dynamics to predict reaction yield and crystal critical quality attributes given inlet streams and reactor conditions. In this way, RC can be integrated with other unit operations to capture the effects RC has on the process overall. The model and assumptions are described in this work. The model space, limitations, and capabilities are explored. Finally, the potential benefits of the RC system are shown using three example cases.

Understanding the complex dynamics of continuous processes in pharmaceutical manufacturing is essential to ensure product quality across the production line. This paper presents a data-driven modeling approach using Sparse Identification of Nonlinear Dynamics with Control (SINDYc) to capture the dynamics of a continuous direct compression (CDC) tableting line. By incorporating delayed control inputs into the candidate function library, the model effectively captures deviations from steady state in response to dynamic changes. The proposed model was developed by finding a balance between accuracy and sparsity, with focus on the ability to generalize to a wide range of operating conditions.

The promising properties of lipid nanoparticles (LNPs) as drug carriers have been attracting significant attention in the field of drug delivery. However, further research is still required for a better understanding of their integration in the pharmaceutical industry. The Quality by Design (QbD) approach aims at ensuring the safety and efficiency in the development of new drugs, through an holistic, risk-based approach that gathers all sources of knowledge available about the system under analysis. One key resource of the QbD framework is the rich toolkit of Design of Experiments (DOE), to deepen the understanding of how the synthesis of LNPs by microfluidics can be effectively conducted and controlled. This study aimed to explore and understand the effectiveness of different DOE strategies, through an in silico study focused on the impact of factors related to the LNPs synthesis, namely the molar ratio of each lipid component in the lipidic mixture and the N/P ratio, while also considering potential economic constraints without disregarding the need for a statistically valid analysis. A simulation model of the LNP synthesis derived from real experiments was adopted as a basis to assess the potential efficacy of estimated models with different levels of complexity, to extract useful insights in future DOEs in these types of systems, given the high cost of each experimental run. The statistical metrics used were the coefficient of determination (R2) and the Root Mean Squared Error (RMSE). With the results obtained, it was possible to verify that different responses from the same system could require quite different model structures, namely, the models developed for potency and for size of the LNPs differed significantly in their complexity. Furthermore, a number of experiments of the order of 30 can be anticipated as necessary for a DOE in a real process, involving similar factors.

The pharmaceutical industry is a highly specialized field where strict quality control and accelerated time-to-market are essential for maintaining competitive advantage. Spherical crystallization has emerged as a promising approach in pharmaceutical manufacturing, offering significant potential to reduce equipment and operating costs, enhancing drug bioavailability, and facilitating compliance with product quality regulations. Emulsions, as an enabling technology for spherical crystallization, present unique advantages. However, the quality of spherical crystallization products derived from emulsions is significantly influenced by the intricate interactions between crystallization phenomena, formulation variables, and solution hydrodynamics. These complexities pose substantial challenges in determining optimal operational conditions to achieve the desired product characteristics. In this study, Bayesian optimization (BO) is employed to refine and optimize the operational conditions for the spherical crystallization of a representative drug, ibuprofen. The primary goal is to improve product flowability, measured by angle of repose, while maintaining the median particle size within a specified range. The optimization process focuses on key variables such as temperature, stirring speed, duration, and BSA concentration. With the help of acquisition functions, BO enables the identification of a high-quality product with fewer experimental trials compared with traditional design of experiments (DoE) methods.

Design space identification (DSId) and flexibility analysis are critical in process systems engineering, enabling efficient design of operating conditions. For bioprocess, these tasks are often hindered by the absence of reliable mechanistic models and limited experimental data. This paper presents an algorithm to address these challenges in bioprocesses. The methodology begins by constructing a Gaussian process (GP) model to predict key performance indicators (KPIs) from process inputs. Leveraging the probabilistic nature of GP predictions, we perform probabilistic design space identification (PDSId), characterizing each input point by its probability of feasibility which is the likelihood that constraints imposed on KPIs are satisfied. To visualize and analyse the feasibility space, contours at varying probability levels are identified using alpha shapes, which define deterministic boundaries corresponding to different confidence levels. This enables the quantification of volumetric process flexibility and operating ranges for each confidence level. The proposed methodology is applied to an antibody-producing Chinese hamster ovary (CHO) cell culture process, optimizing culture temperature and osmolality with respect to product yield and purity. Results are presented through probability heat maps and flexibility metrics, providing both qualitative and quantitative insights into feasibility and operational flexibility, thereby supporting informed decision-making in process design.

This study presents a comprehensive comparison of the three alternative downstream manufacturing technologies for pharmaceuticals: i) Dry Granulation (DG) through roller compaction, ii) Direct Compaction (DC), and iii) Wet Granulation (WG) based on the economic, environmental and product quality performances. Firstly, the integrated dynamic mathematical models of the different downstream (drug product) processes were developed using gPROMS formulated products based on data from the literature or/and our recent experimental work. The process models were developed and simulated to reliably capture the impact of the different design options, process parameters, and material attributes. Uncertainty analysis was conducted using global sensitivity analysis to identify the set of critical process parameters (CPP) and critical material attributes (CMA) that mostly influence the quality and performance of the final pharmaceutical tablets in each case, captured by the critical quality attributes (CQAs). Based on the set of CPP and CMA, the combined design spaces, which guarantee the attainment of the targeted CQA, were identified and compared. Additionally, based on the process simulations results and inventory data, the techno-economic Analysis was performed alongside life cycle assessment (LCA). The LCA provided an in-depth evaluation of the environmental impacts associated with each manufacturing method, considering aspects such as energy consumption, raw material usage, emissions, and waste generation based on a cradle-to-gate approach. By integrating the CQAs and critical emission categories within a Quality and and Sustainability by Digital Design (QSbDD) paradigm, this study offers a holistic analysis that captures both the environmental and product quality performance.

This study aims at developing a holistic approach to establish robust mathematical models of integrated and interactive multistep processes, while systematically identifying the corresponding design space and acceptable operating region (AOR). The overall objective is to reduce the experimentation costs, enhance accuracy of integrated metathetical models, and deliver built-in quality assurance based on a new Quality by Digital Design (QbDD) paradigm. This methodology starts with the construction of a set of model candidates for different unit operations, based on the prior knowledge and inherent assumptions. Several model candidates of the integrated multistep process are considered. A model discrimination based on model prediction performance reveals the best integrated model for the multistep process. In the next step, the estimability analysis and model-based design of experiment (MBDoE) are implemented to deliver information-rich data and systematically refine the integrated model. With the acquisition of the new experimental data, the reliability and robustness of the multistep mathematical model is dramatically enhanced. A blending-tableting process is considered to validate the methodology. The model captures the effects of the blender as well as the composition and porosity of the tablet on the tablet tensile strength. Model discrimination and automated model refinement are then performed to identify and improve the optimal integrated model for this two-step process, and the enhanced model is applied for the design space identification under specified CQA targets and associated bounds.

The environmental impacts of pharmaceutical production underscore the need for comprehensive life cycle assessments (LCAs). Offshoring manufacturing, a common cost-saving strategy in the pharmaceutical industry, increases supply chain complexity and reliance on countries like India and China for active pharmaceutical ingredients (APIs). The COVID-19 pandemic exposed Europe’s vulnerability to global crises, prompting initiatives such as the French government’s re-industrialization plan to relocate the production of fifty critical drugs. Paracetamol production has been prioritized, with recent shortages highlighting the urgency to address supply chain risks while considering environmental impacts. This study uses process engineering to generate life cycle inventory (LCI) data for paracetamol production, offering an eco-design perspective. Aspen Plus was employed to model the API manufacturing process, integrating mass and energy balances to address the scarcity of LCI data. The results highlight significant differences in emissions between India and France. For 1 kg of API, India’s emissions are 0.0826 kg CO2 eq for electricity and 1.3845 kg CO2 eq for heat from steam, compared to France’s significantly lower values of 0.005228 kg CO2 eq and 1.1828 kg CO2 eq, respectively. These differences emphasize the environmental benefits of relocating production to regions with cleaner energy sources. The study demonstrates the value of process modeling for generating robust LCI data, enabling detailed LCAs to support eco-design in pharmaceutical manufacturing. This approach can be extended to other chemicals, facilitating sustainable decision-making in the sector’s re-industrialization.

In 2023, the International Council for Harmonisation (ICH) guideline for the development, implementation, and lifecycle management of pharmaceutical continuous manufacturing (PCM), was implemented in Europe. It promotes quality-by-design (QbD) and quality by control (QbC) strategies as well as the appropriate use of mathematical modelling. This development urges a harmonizing understanding across academia and industry for adoption of interpretable models instead of black-box models for advanced control strategies such as model predictive control (MPC), especially when applied in Good Manufacturing Practice (GMP) regulated areas. To this end, we first propose a comprehensive model development using Dynamic Mode Decomposition with Control (DMDc)to represent complex dynamics in a lower-dimensional space, disambiguating between underlying dynamics and actuation effects. Using data from a digital twin of PCM, our model demonstrates low computational complexity while effectively capturing nonlinear dynamics with significant improvements observed in the performance metrics. Additionally, we demonstrate enhanced uncertainty propagation performance when compared to state-space models obtained with N4SID and Sparse Identification of Nonlinear Dynamical systems with control algorithms. Finally, we develop a closed-loop workflow that seamlessly connects data exchanges between Python (DMDc), GAMS (MPC optimisation) & gPROMS to evaluate the controller performance with setpoint tracking and disturbance rejection studies achieving high accuracy in real-time monitoring and control of granule size. This study offers a novel, interpretable control strategy for PCM. The results demonstrate the potential for real-time release testing, reduced reliance on end-product testing, and improved process control in the pharmaceutical industry.

Crystallization is a relevant process in the pharmaceutical industry for product purification and particle production. An efficient crystallization is characterized by crystals produced with the desired attributes. Therefore, modeling this process is a key point to achieve this goal. In this sense, the objective of this work is to propose a hybrid model to describe paracetamol dissolution in ethanol. The universal differential equations methodology is considered in the development of this model, using a neural network to predict the dissolution rate combined with the population balance equations to calculate the moments of the crystal size distribution (CSD) and the concentration. The model was developed using experimental batches. The dataset is composed of concentration measurements obtained using attenuated total reflectance-Fourier transform infrared (ATR-FTIR). The objective function of the optimization problem is to minimize the relative absolute difference between the experimental and the predicted concentration. The hybrid model efficiently predicted the concentration compared to the experimental measurements. Furthermore, the hybrid approach made predictions of the moments of the CSD similar to the population balance model proposed by Kim et al. [1], being able to successfully calculate batches not considered in the training dataset. Moreover, the performance of the hybrid model was similar to the phenomenological one based on population balance. Therefore, the universal differential equations approach is presented as an efficient methodology for modeling crystallization processes with limited information.

Tableting is a dry granulation process for compacting powder blends into tablets. In this process, a blend of active pharmaceutical ingredients (APIs) and excipients are fed into the hopper of a rotary tablet press via feeders. Inside the tablet press, rotating feed frame paddle wheels fill powder into dies, with tablet mass adjusted by the lower punch position during the die filling process. Pre-compression rolls press air out of the die, while main compression rolls apply the force necessary for compacting the powder into tablets. In this paper, process variables such as feeder screw speeds, feed frame impeller speed, lower punch position during die filling, and punch distance during main compression have been systematically varied. Corresponding responses, including pre-compression force, ejection force, and tablet porosity have been evaluated to optimize the tableting process. After implementing an open platform communications unified architecture (OPC UA) interface, process variables can be monitored in real-time. To enable in-line monitoring of tablet porosity, a novel UV/Vis fiber optic probe has been implemented into the rotary tablet press. To further analyze the overall process, a data-driven modeling approach is adopted. Data-driven modeling is a valuable alternative to modeling real-world processes where, for instance, first principles modeling is difficult or infeasible. Due to the complex nature of the powder compaction process, several model classes need to be explored. To begin with, linear autoregressive models with exogenous inputs (ARX) have been considered. Thereafter, nonlinear autoregressive models with exogenous inputs (NARX) have been considered. Notably, several experiments have been designed to gather the data required for the development of the model.

Mechanistic population balance modeling (PBM) has advanced the design of pharmaceutical crystallization processes, enabling the production of active pharmaceutical ingredient (API) crystals with desired critical quality attributes (CQAs), such as purity and crystal size distribution. However, PBM development can sometimes be resource-intensive, requiring extensive design of experiments (DoE) and high-quality process data, making it impractical under fast-paced industrial development timelines. This study proposes a machine learning (ML)-based workflow for developing ‘fit-for-purpose’ digital twins of crystallization processes, leveraging industrially available DoE data to link operating conditions with CQAs. Validated on industrial data for a commercial API with complex crystallization challenges, the workflow efficiently identifies optimal operating conditions, demonstrating the potential of data-driven digital twins to accelerate the development of pharmaceutical processes.

As pharmaceutical manufacturing is transitioning from traditional batch to continuous manufacturing (CM), there is a lack of tools for CM design and development, which can integrate drug substance and drug product unit operations for overall evaluation. Recently, a Python-based PharmaPy framework was proposed to advance the design, simulation, and analysis of continuous pharmaceutical processes. However, the initial library of models only addressed upstream drug substance processing. In this work, new capabilities, including drug product unit operations such as feeder, blender, and tablet press, have been added to the PharmaPy framework, enabling end-to-end study and optimizing the effects of material properties and process conditions on solid oral dosage products. The platform supports computational efficiency and model accuracy by allowing the development of different mechanistic and semi-mechanistic models. Sensitivity analysis is performed on the integrated end-to-end simulator to identify critical input variables influencing product quality and control strategies. The analysis lowers the complexity of the model by ranking significant input variables. Finally, feasibility studies are conducted on extracted influential input variables to characterize the process design space and achieve desirable output. The enhanced PharmaPy package can now support decision-making from early research and development stages through manufacturing.

Mathematical modeling is essential for the effective control of many chemical engineering processes, including crystallization. However, most existing crystallization models used in industry and academia assume ideal mixing. As a result, the unclear effects of imperfect mixing on crystallization, reported in experimental studies, remain largely unexplained. In this work we aim to address this gap in understanding by examining antisolvent crystallization processes on a general theoretical level, using a novel dimensionless model. To address the impact of mixing on crystallization, we employ the Engulfment model coupled with a population balance, and we nondimensionalize the model equations. Using this model, we explore the dependence of the mean particle size on the homogenization rate, represented by the Damköhler number for crystallization. Moreover, we study the impact of mixing at various values of the model’s kinetic parameters to simulate difference in properties of individual products. We show that we are able to explain the complex interaction between crystallization and mixing, proving our model can serve as a tool for achieving a better understanding of the processes involved. Finally, due to its efficiency and reduced number of parameters, the model is suitable for direct fitting to experimental data.

This study proposes a model-based approach utilizing a hybrid population balance model (PBM) to optimize temperature profiles for minimizing agglomeration and enhancing crystal growth. The PBM incorporates key mechanisms—nucleation, growth, dissolution, agglomeration, and deagglomeration—and is applied to the crystallization of an industrial active pharmaceutical ingredient (API), Compound K. Parameters were estimated through prior design of experiments (DoE) and refined via additional thermocycle experiments. In-silico DoE simulations demonstrate that the hybrid PBM outperforms traditional methods in assessing process performance under agglomeration-prone conditions. Results confirm that thermocycles effectively reduce agglomeration and promote bulk crystal formation, though their efficiency plateaus beyond a certain cycle number. This model-based approach provides a more robust strategy for agglomeration control compared to conventional methods, offering valuable insights for industrial crystallization optimization.

Distillation processes account for a substantial share of the industrial energy demand. Yet, these energy requirements can be reduced by a variety of energy integration methods, including various forms of direct heat integration, multi-effect distillation, thermal coupling and vapor recompression. Consequently, these intensification methods should be evaluated quantitatively in comparison to each other for individual separation tasks, instead of benchmarking single options with conventional sequences or relying on simplified heuristics. In order to overcome the computational burden of a broad assessment of a large number of process alternatives, a computationally-efficient framework for the energetic and economic evaluation of such energy integrated distillation processes is presented, which builds on thermodynamically-sound shortcut models that do not rely on constant relative volatility and constant molar overflow assumptions.