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 material... [more]

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.

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... [more]

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. S... [more]

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 simul... [more]

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 th... [more]

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... [more]

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.

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 per-formance of graph-based machine learning models, specifically Graph Convolutional Networks (GCN), Graph At-tention Networks (GAT), and a graph-transformer model, PharmHGT, for predicting CMC values. We aim to de-termine 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.

Supplementary Material for the contribution "Synthesis of Distillation Flowsheets with Reinforcement Learning using Transformer Blocks" by Niklas Slager and Meik Franke (UTwente) for ESCAPE 35

The document is the digital supplementary material for the article titled "Application of Artificial Intelligence in process simulation tool", submitted to the ESCAPE 35 conference. It contains additional information and figures.

Overview presenation by Prof. Alexander Dowling at the Public Affairs & Information Committee (PAIC) Town Hall at AIChE Annual Meeting in 2024.

For mitigating global warming, polymer electrolyte membrane fuel cells have become promising, clean, and sustainable alternatives to existing energy sources. To increase the energy density and efficiency of polymer electrolyte membrane fuel cells (PEMFC), a comprehensive numerical modeling approach that can adequately predict the multiphysics and performance relative to the actual test such as an acceptable depiction of the electrochemistry, mass/species transfer, thermal management, and water generation/transportation is required. However, existing models suffer from reliability issues due to their dependency on several assumptions made for the sake of modeling simplification, as well as poor choices and approximations in material characterization and electrochemical parameters. In this regard, data-driven machine learning models could provide the missing and more appropriate parameters in conventional computational fluid dynamics models. The purpose of the present overview is to expl... [more]

This research delves into the impact of Twist Winglets−Cross-Section Twist Tape (TWs-CSTT) structures within heat exchangers on thermal performance. Utilizing Computational Fluid Dynamics (CFD) and machine learning methodologies, optimal geometrical parameters for the TWs-CSTT configuration were examined. The outcomes demonstrate that fluid undergoing a rotational motion within tubes featuring this structure leads to more effective secondary flows, intensified mixing, and improved thermal boundary layer disturbance. Moreover, by integrating machine learning with multi-objective optimization techniques, the performance of heat exchangers can be accurately predicted and optimized, facilitating enhanced heat exchanger design. Through the application of the multi-objective optimization algorithm Non-dominated Sorting Genetic Algorithm II (NSGA-II), the ideal configurations for TWs-CSTT were ascertained: L1 is the cross-sectional length of the Twisted Wings, L2 is the radius of the Central... [more]

Physicochemical properties of ionic liquids (ILs) are essential in solvent screening and process design. However, due to their vast diversity, acquiring IL properties through experimentation alone is both time-consuming and costly. For this reason, the creation of prediction models that can accurately forecast the characteristics of IL and its mixtures is crucial to their application. This study proposes a model for predicting the three important parameters of the IL-organic solvent−organic solvent ternary system: density, viscosity, and heat capacity. The model incorporates group contribution (GC) and machine learning (ML) methods. A link between variables such as temperature, pressure, and molecular structure is established by the model. We gathered 2775 viscosity, 6515 density, and 1057 heat capacity data points to compare the prediction accuracy of three machine learning methods, namely, artificial neural networks (ANNs), extreme gradient boosting (XGBoost), and light gradient boos... [more]

The present work focuses on predicting the steel melt temperature following the vacuum treatment step in a vacuum tank degasser (VTD). The primary objective is to establish a comprehensive methodology for developing and validating machine learning (ML) models within this context. Another objective is to evaluate the model by analyzing the alignment of the SHAP values with metallurgical domain expectations, thereby validating the model’s predictions from a metallurgical perspective. The proposed methodology employs a Random Forest model, incorporating a grid search with domain-informed variables grouped into batches, and a robust model-selection criterion that ensures optimal predictive performance, while keeping the model as simple and stable as possible. Furthermore, the Shapley Additive Explanations (SHAP) algorithm is employed to interpret the model’s predictions. The selected model achieved a mean adjusted R2 of 0.631 and a hit ratio of 75.3% for a prediction error within ±5 °C. De... [more]

Some educators place an emphasis on the commonalities between engineering mathematics with process control, among others and this helps students see the bigger picture of what is being taught. Traditionally, some of the concepts such as diffusion and heat transfer are taught with a mathematical point of view. Now-a-days, Machine Learning (ML) has emerged as topic of greater interest to both educators and learners and new and disparate modules are sometimes introduced to teach the same. With the emergence of these new topics, some students (falsely) believe that ML is a new field that is somehow different and not linked to engineering mathematics. In this work, we show the link between the different topics from engineering mathematics, that are traditionally taught in UG education, with ML. We hope that educators and learners will appreciate the treatise and think differently, and we further hope that this will further increase the interest to improve ML models.

Critical minerals are essential to the future of clean energy, especially energy storage, electric vehicles, and advanced electronics. In this paper, we argue that process systems engineering (PSE) paradigms provide essential frameworks for enhancing the sustainability and efficiency of critical mineral processing pathways. As a concrete example, we review challenges and opportunities across material-to-infrastructure scales for process intensification (PI) with membranes. Within critical mineral processing, there is a need to reduce environmental impact, especially concerning chemical reagent usage. Feed concentrations and product demand variability require flexible, intensified processes. Further, unique feedstocks require unique processes (i.e., no one-size-fits-all recycling or refining system exists). Membrane materials span a vast design space that allows significant optimization. Therefore, there is a need to rapidly identify the best opportunities for membrane implementation, t... [more]

Initial design stages are inherently complex and often lack comprehensive information, posing challenges in evaluating sustainability metrics. Machine Learning (ML) emerges as a valuable solution to address these challenges. ML algorithms, particularly effective in predicting environmental impacts of new chemicals with limited data, enable more informed decisions in sustainable design. This study focuses on employing ML for predicting the environmental impacts related to human health, ecosystem quality, climate change, and resource utilization to aid in early-stage environmental impact assessment of chemical processes. The effectiveness of the ML algorithm, eXtreme Gradient Boosting (XGBoost) tested using a dataset of 350 points, divided into training, testing, and validation sets. The study also includes a practical application of the model in a cradle-to-cradle LCA of N-Methylpyrrolidone (NMP), demonstrating its utility in sustainable chemical process design. This approach signifies... [more]

Most integrated energy system (IES) optimization frameworks employ the price-taker approximation, which ignores important interactions with the market and can result in overestimated economic values. In this work, we propose a machine learning surrogate-assisted optimization framework to quantify IES/market interactions and thus go beyond price-taker. We use time series clustering to generate representative IES operation profiles for the optimization problem and use machine learning surrogate models to predict the IES/market interaction. We quantify the accuracy of the time series clustering and surrogate models in a case study to optimally retrofit a nuclear power plant with a polymer electrolyte membrane electrolyzer to co-produce electricity and hydrogen.

In this work, we present a follow-up work of reinforcement learning (RL)-driven process design using the Institute for Design of Advanced Energy Systems Process Systems Engineering (IDAES-PSE) Framework. Herein, process designs are generated as stream inlet-outlet matrices and optimized using the IDAES platform, the objective function value of which is the reward to RL agent. Deep Q-Network is employed as the RL agent including a series of convolutional neural network layers and fully connected layers to compute the actions of adding or removing any stream connections, thus creating a new process design. The process design is then informed back to the RL agent to refine its learning. The iteration continues until the maximum number of steps is reached with feasible process designs generated. To further expedite the RL search of the design space which can comprise the selection of any candidate unit(s) with arbitrary stream connections, we investigate the role of RL reward function and... [more]

Machine learning (ML) has become a powerful tool to analyze complex relationships between multiple variables and to unravel valuable information from big datasets. However, an open research question lies in how ML can accelerate the design and optimization of processes in the early experimental development stages with limited data. In this work, we investigate the ML-aided process design of a microwave reactor for ammonia production with exceedingly little experimental data. We propose an integrated approach of synthetic minority oversampling technique (SMOTE) regression combined with neural networks to quantitatively design and optimize the microwave reactor. To address the limited data challenge, SMOTE is applied to generate synthetic data based on experimental data at different reaction conditions. Neural network has been demonstrated to effectively capture the nonlinear relationships between input features and target outputs. The softplus activation function is used for a smoother... [more]

We present a string representation for hybrid extraction and distillation using symbols representing phenomena building blocks. Unlike the conventional equipment-based string representation, the proposed representation captures the design details of liquid-liquid extraction and distillation. We generate a set of samples through the procedure of input parameter sampling and superstructure optimization that minimizes separation cost. We convert these generated samples into a set of string representations based on pre-defined rules. We use these string representations as descriptors and connect them with conditional variational encoder. The trained conditional variational encoder shows good prediction accuracy. We further use the trained conditional variational encoder to screen designs of hybrid extraction and distillation with desired cost investment.

Machine learning presents opportunities to improve the scale-specific accuracy of mechanistic models in a data-driven manner. Here we demonstrate the use of a machine learning technique called Sparse Identification of Nonlinear Dynamics (SINDy) to improve a simple mechanistic model of algal growth. Time-series measurements of the microalga Chlorella Vulgaris were generated under controlled photobioreactor conditions at the University of Technology Sydney. A simple mechanistic growth model based on intensity of light and temperature was integrated over time and compared to the time-series data. While the mechanistic model broadly captured the overall growth trend, discrepancies remained between the model and data due to the model's simplicity and non-ideal behavior of real-world measurement. SINDy was applied to model the residual error by identifying an error derivative correction term. Addition of this SINDy-informed error dynamics term shows improvement to model accuracy while maint... [more]

Water chemistry plays a critical role in the design and operation of water treatment processes. Detailed chemistry modeling tools use a combination of advanced thermodynamic models and extensive databases to predict phase equilibria and reaction phenomena. The complexity and formulation of these models preclude their direct integration in equation-oriented modeling platforms, making it difficult to use their capabilities for rigorous water treatment process optimization. Neural networks (NN) can provide a pathway for integrating the predictive capability of chemistry software into equation-oriented models and enable optimization of complex water treatment processes across a broad range of conditions and process designs. Herein, we assess how NN architecture and training data impact their accuracy and use in equation-oriented water treatment models. We generate training data using PhreeqC software and determine how data generation and sample size impact the accuracy of trained NNs. The... [more]