Vaccines are typically produced in large facilities to take advantage of economies of scale. However disease outbreaks are often local in nature and require flexible, small-scale production, especially in regions with poor infrastructure. In this work, mobile on-demand vaccine production is explored as a solution to future outbreaks. An mRNA vaccine process is scaled down to the size of two 20-foot shipping containers, so that 10,000 vaccine doses can be produced in one batch in less than 16 hours. The container is self-sufficient except for the regular resupply of water and electricity being able to produce 100 batches without resupply raw materials and consumables. The final cost per dose is estimated to be 25 e with a likely range between 4 to 45 e depending on dose size, raw material prices, and other underlying assumptions. The practicality of a container-based facility at the presented scale is demonstrated by two case studies.

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 industr... [more]

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 represen... [more]

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.

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.

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 simu... [more]

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,... [more]

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-st... [more]

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... [more]

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 ap-plied 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 be-yond a certain cycle number. This model-based approach provides a more robust strategy for agglomeration control compared to conventional methods, offering valuable insights for indus... [more]

This model uses standard CHEMCAD unit operations and thermodynamic models to simulate the separation of ethanol and water from a fermenter broth.
It is used as a template in the course Green Processes at Berlin University of Applied Science (BHT), where students use it to regress measured data from lab experiments and to design an optimal process.

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]

Traditional two-dimensional dynamic fault detection methods describe nonlinear dynamics by constructing a two-dimensional sliding window in the batch and time directions. However, determining the shape of a two-dimensional sliding window for different phases can be challenging. Samples in the two-dimensional sliding windows are assigned equal importance before being utilized for feature engineering and statistical control. This will inevitably lead to redundancy in the input, complicating fault detection. This paper proposes a novel method named attention-based two-dimensional dynamic-scale graph autoencoder (2D-ADSGAE). Firstly, a new approach is introduced to construct a graph based on a predefined sliding window, taking into account the differences in importance and redundancy. Secondly, to address the training difficulties and adapt to the inherent heterogeneity typically present in the dynamics of a batch across both its time and batch directions, we devise a method to determine t... [more]

Batch processing is a widely utilized technique in the manufacturing of high-value products. Traditional methods for quality assessment in batch processes often lead to productivity and yield losses because of offline measurement of quality variables. The use of soft sensors enhances product quality and increases production efficiency. However, due to the uneven batch data, the variation in processing times presents a significant challenge for building effective soft sensor models. Moreover, sensor failures, exacerbated by the manufacturing environment, complicate the accurate modeling of process variables. Existing soft sensor approaches inadequately address sensor malfunctions, resulting in significant prediction inaccuracies. This study proposes a fault-tolerant soft sensor algorithm that integrates two Long Short-Term Memory (LSTM) networks. The algorithm focuses on modeling process variables and compensating for sensor failures using historical batch quality data. It introduces a... [more]

Batch processes in the biopharmaceutical and chemical manufacturing industries often develop new products to meet changing market demands. When the dynamic models of these new products are trained, dynamic modeling with limited data for each product can lead to inaccurate results. One solution is to extract useful knowledge from past historical production data that can be applied to the product of a new grade. In this way, the model can be built quickly without having to wait for additional modeling data. In this study, a subspace identification combined common feature learning scheme is proposed to quickly learn a model of a new grade. The proposed modified state-space model contains common and special parameter matrices. Past batch data can be used to train common parameter matrices. Then, the parameters can be directly transferred into a new SID model for a new grade of the product. The new SID model can be quickly well trained even though there is a limited batch of data. The effec... [more]

Support vector data description (SVDD) has been widely applied to batch process fault detection. However, it often performs poorly, especially when incipient faults occur, because it only considers the shallow data feature and omits the probabilistic information of features. In order to provide better monitoring performance on incipient faults in batch processes, an improved SVDD method, called deep probabilistic SVDD (DPSVDD), is proposed in this work by integrating the convolutional autoencoder and the probability-related monitoring indices. For mining the hidden data features effectively, a deep convolutional features extraction network is designed by a convolutional autoencoder, where the encoder outputs and the reconstruction errors are used as the monitor features. Furthermore, the probability distribution changes of these features are evaluated by the Kullback-Leibler (KL) divergence so that the probability-related monitoring indices are developed for indicating the process stat... [more]

In this paper, a two-dimensional, two-layer quality regression model is established to monitor multi-phase, multi-mode batch processes. Firstly, aiming at the multi-phase problem and the multi-mode problem simultaneously, the relations among modes and phases are captured through the analysis between process variables and quality variables by establishing a two-dimensional, two-layer regression partial least squares (PLS) model. The two-dimensional regression traces the intra-batch and inter-batch characteristics, while the two-layer structure establishes the relationship between the target process and historical modes and phases. Consequently, online monitoring is carried out for multi-phase, multi-mode batch processes based on quality prediction. In addition, the online quality prediction and monitoring results based on the proposed method and those based on the traditional phase mean PLS method are compared to prove the effectiveness of the proposed method.

Safe and stable operation plays an important role in the chemical industry. Fault detection and diagnosis (FDD) make it possible to identify abnormal process deviations early and assist operators in taking proper action against fault propagation. After decades of development, data-driven process monitoring technologies have gradually attracted attention from process industries. Although many promising FDD methods have been proposed from both academia and industry, challenges remain due to the complex characteristics of industrial data. In this work, classical and recent research on data-driven process monitoring methods is reviewed from the perspective of characterizing and mining industrial data. The implementation framework of data-driven process monitoring methods is first introduced. State of art of process monitoring methods corresponding to common industrial data characteristics are then reviewed. Finally, the challenges and possible solutions for actual industrial applications a... [more]

Biosorption has great potential in removing toxic effluents from wastewater, especially heavy metal ions such as cobalt, lead, copper, mercury, cadmium, nickel and other ions. Mathematically modeling of biosorption process is essential for the economical and robust design of equipment employing the bioadsorption process. However, biosorption is a complex physicochemical process involving various transport and equilibrium processes, such as absorption, adsorption, ion exchange and surface and interfacial phenomena. The biosorption process becomes even more complex in cases of multicomponent systems and needs an extensive parametric analysis to develop a mathematical model in order to quantify metal ion recovery and the performance of the process. The biosorption process involves various process parameters, such as concentration, contact time, pH, charge, porosity, pore size, available sites, velocity and coefficients, related to activity, diffusion and dispersion. In this review paper,... [more]

To obtain quality variables that cannot be measured in real time during the production process but reflect information on the quality of the final product, the batch production process has the characteristics of a strong time-varying nature, non-Gaussian data distribution and high nonlinearity. A locally weighted partial least squares regression quality prediction model (KICA-JITL-LWPLS), based on wavelet kernel function independent meta-analysis with immediate learning, is proposed. The model first measures the similarity between the normalized input data and the historical data and assigns the input data to the group of historical data with high similarity to it, based on the posterior probability of the Bayesian classifier; subsequently, wavelet kernel functions are selected and kernel learning methods are introduced into the independent meta-analysis algorithm. An independent meta-analysis, based on the wavelet kernel function, is performed on the classified input data to obtain pr... [more]

By-products of laying hens represent a promising raw material source with a high collagen content, which is currently not adequately used. The aim of the paper is to prepare gelatins from laying hen paws. The purified collagen raw material was processed by a biotechnological process using the food endoprotease Protamex®. After cleavage of the cross-links in the collagen structure, the gelatin was extracted by a batch process with a stirrer in two extraction steps. The influence of the extraction process on the yield of gelatins and on selected qualitative parameters of gelatins was monitored by two-level factor experiments with three selected process factors. The studied factors were: enzyme dosage (0.2−0.8%), enzyme processing time (24−72 h) and gelatin extraction time (30−120 min). After the first extraction step at 75 °C, gelatin was extracted with a yield of 8.2−21.4% and a gel strength of 275−380 Bloom. In the second extraction step at 80−100 °C, it is possible to obtain another p... [more]

The problem of scheduling multiproduct and multipurpose batch processes has been studied for more than 30 years using math programming and heuristics. In most formulations, the manufacturing recipes are represented by simplified models using state task network (STN) or resource task network (RTN), transfers of materials are assumed to be instantaneous, constraints due to shared utilities are often ignored, and scheduling horizons are kept small due to the limits on the problem size that can be handled by the solvers. These limitations often result in schedules that are not actionable. A simulation model, on the other hand, can represent a manufacturing recipe to the smallest level of detail. In addition, a simulator can provide a variety of built-in capabilities that model the assignment decisions, coordination logic and plant operation rules. The simulation based schedules are more realistic, verifiable, easy to adapt for changing plant conditions and can be generated in a short perio... [more]

A phenomenological model was developed to describe the dynamic evolution of the batch suspension polymerization of methyl methacrylate in terms of reactor temperature, pressure, concentrations and molecular properties of the final polymer. Then, the phenomenological model was used as a process constraint in dynamic data reconciliation procedures, which allowed for the successful monitoring of reaction variables in real-time and on-line. The obtained results indicate that heat transfer coefficients change significantly during the reaction time and from batch to batch, exerting a tremendous impact on the process operation. Obtained results also indicate that it can be difficult to attain thermodynamic equilibrium conditions in this system, because of the continuous condensation of evaporated monomer and the large mass transfer resistance offered by the viscous suspended droplets.