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 sim... [more]

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... [more]

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 no... [more]

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 sim... [more]

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.

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.

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.

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 perf... [more]

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... [more]

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 correlat... [more]

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 de... [more]

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 positi... [more]

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 configuratio... [more]

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.

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.... [more]

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... [more]

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 w... [more]

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... [more]

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.

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... [more]

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... [more]

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... [more]

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... [more]

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.

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 outco... [more]