Optimal Process Control Research Articles

Research on Torque Current Error Control of PMSM Based on Model Predictive Control

To solve the problem that the torque pulsation is serious due to the large current error in the period when the duty cycle is calculated by the deadbeat method in the duty cycle dual-vector model predictive current control method, a vector switching method is proposed. The control objective of the proposed method is not to equate the predicted torque current with the reference value under the dead-beat idea, but to reduce the torque current control error by changing the switching time of the voltage vector and ensuring the symmetry of the positive and negative current errors. By deducing the duty cycle, the maximum current errors in a single control cycle before and after vector switching is compared and analyzed. The vector switching method is verified on the Simulink simulation platform and compared with the duty cycle double vector method. The results show that the vector switching method can effectively reduce the intra-cycle torque current tracking error and realize the optimal process control in each cycle.

Case Studies and Best Practices in Cloud-Based Big Data Analytics for Process Control

In this research paper, case studies and exemplars and lessons learnt in cloud-based big data analytics for process control are reviewed. The paper presents big data, cloud computing and industrial process control system with prospects of enhancing effectiveness, increasing production rates, and effective decision making in the industries. The research in this paper involves a comprehensive literature review of the research topic, and an extension of the analysis to four specific business industries as well as a discussion of architectural elements for cloud-based big data solutions for process control business. It also presents various crucial issues such as data protection, adherence to legal requirements, and compatibility with other systems, giving solutions. In addition, the research compares the effectiveness of cloud-based solutions with on-premise ones and discuss other novelties, including edge computing and artificial intelligence as the tendencies potentially influencing process control. Consequently, the findings of this research can be helpful for both industry practitioners and researchers who aim to optimize process control and organization operation with the help of cloud-based big data analytics

SimDFBA: A framework for bioprocess simulation and development

Many efforts have been made to model bio-processes and investigate the main effective parameters using conventional known softwares. The main shortcoming of such models is that the bioreactor is considered a black box. Such simplification causes uncertainty in following different metabolic pathways of microorganisms. This study presents a new method for simulating bio-processes by integrating available conventional simulation softwares such as ASPEN PLUS and MATLAB. The former is used to simulate common process equipment, and the latter is employed to model Bio-reactor. Besides, the Bio-reactor is modeled mathematically with Dynamic Flux Balance Analysis (DFBA) concept using the metabolic network reconstruction model. This framework will simultaneously provide a rigorous basic model of both upstream and downstream equipment. This approach can also be used as a toolbox to implement process optimization and modern control strategies on the bio-processes. A syngas fermentation bioprocess case study was chosen to implement the new framework.

Adaptive Stochastic ADMM for Decentralized Reinforcement Learning in Edge IoT

Edge computing provides a promising paradigm to support the implementation of Internet of Things (IoT) by offloading tasks to nearby edge nodes. Meanwhile, the increasing network size makes it impractical for centralized data processing due to limited bandwidth, and consequently a decentralized learning scheme is preferable. Reinforcement learning (RL) has been widely investigated and shown to be a promising solution for decision-making and optimal control processes. For RL in a decentralized setup, edge nodes (agents) connected through a communication network aim to work collaboratively to find a policy to optimize the global reward as the sum of local rewards. However, communication costs, scalability, and adaptation in complex environments with heterogeneous agents may significantly limit the performance of decentralized RL. Alternating direction method of multipliers (ADMM) has a structure that allows for decentralized implementation and has shown faster convergence than gradient descent-based methods. Therefore, we propose an adaptive stochastic incremental ADMM (asI-ADMM) algorithm and apply the asI-ADMM to decentralized RL with edge-computing-empowered IoT networks. We provide convergence properties for the proposed algorithms by designing a Lyapunov function and prove that the asI-ADMM has <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$\mathcal {O}(1/k) + \mathcal {O}(1/M)$ </tex-math></inline-formula> convergence rate, where <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$k$ </tex-math></inline-formula> and <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$M$ </tex-math></inline-formula> are the number of iterations and batch samples, respectively. Then, we test our algorithm with two supervised learning problems. For performance evaluation, we simulate two applications in decentralized RL settings with homogeneous and heterogeneous agents. The experimental results show that our proposed algorithms outperform the state of the art in terms of communication costs and scalability and can well adapt to complex IoT environments.

Deep Learning-Based Predictive Control of Injection Velocity in Injection Molding Machines

Rapid and reliable optimal control of injection molding machines (IMMs) is critical for the effective production of injection-molded goods, especially in the situation of restricted computer resources of embedded equipment in IMMs. In this paper, an optimal tracking injection velocity control problem arising in a typical IMM is studied. An effective hybrid intelligent control approach with less computing resources for real-time implementation based on the deep learning (DL) method to mimic the classical model predictive control rule is developed to deal with the tracking control of the injection speed. The proposed method utilizes the gated recurrent unit neural network to learn and predict the optimal time series control process data produced by the traditional model predictive controller. The benefits of this approach over the conventional optimization method are illustrated through simulation results, which show that the convergent DL-based controller can effectively avoid the complex calculation in the control process of IMMs and meet the requirements of more robustness and resist environmental uncertainty to a certain level and can be potentially implemented in embedded hardware much more efficiently and conveniently with a smaller memory footprint and faster computation time.

Data-driven modeling of power generation for a coal power plant under cycling

Increased penetration of renewables for power generation has negatively impacted the dynamics of conventional fossil fuel-based power plants. The power plants operating on the base load are forced to cycle, to adjust to the fluctuating power demands. This results in an inefficient operation of the coal power plants, which leads up to higher operating losses. To overcome such operational challenge associated with cycling and to develop an optimal process control, this work analyzes a set of models for predicting power generation. Moreover, the power generation is intrinsically affected by the state of the power plant components, and therefore our model development also incorporates additional power plant process variables while forecasting the power generation. We present and compare multiple state-of-the-art forecasting data-driven methods for power generation to determine the most adequate and accurate model. We also develop an interpretable attention-based transformer model to explain the importance of process variables during training and forecasting. The trained deep neural network (DNN) LSTM model has good accuracy in predicting gross power generation under various prediction horizons with/without cycling events and outperforms the other models for long-term forecasting. The DNN memory-based models show significant superiority over other state-of-the-art machine learning models for short, medium and long range predictions. The transformer-based model with attention enhances the selection of historical data for multi-horizon forecasting, and also allows to interpret the significance of internal power plant components on the power generation. This newly gained insights can be used by operation engineers to anticipate and monitor the health of power plant equipment during high cycling periods.

Machine learning modelling of a membrane capacitive deionization (MCDI) system for prediction of long-term system performance and optimization of process control parameters in remote brackish water desalination

Membrane Capacitive Deionization (MCDI) is a promising electrochemical technique for water desalination. Previous studies have confirrmed the effectiveness of MCDI in removing contaminants from brackish groundwaters, especially in remote areas where electricity is scarce. However, as with other water treatment technologies, performance deterioration of the MCDI system still occurs, hindering the stability of long-term operation. Herein, a machine learning (ML) modelling framework and various ML models were developed to (i) investigate the performance deterioration due particularly to insufficient charging/discharging of the electrode caused by accumulation of ions and electrode scaling and (ii) optimise MCDI operating parameters such that the impacts of these deleterious effects on unit performance were minimized. The ML models developed in this work exhibited a prediction accuracy of cycle time with average mean absolute percentage error (MAPE) values of 16.82% and 16.09% after 30-fold cross validation for Random Forest (RF) and Multilayer Perceptron (MLP) models respectively. The pre-trained ML model predicted different declining trends of water production for two different operating conditions and provided corresponding recommendations on frequencies of chemical cleaning. A case study on the adjustment of operating parameters using the results suggested by the optimization ML model was conducted. The model validation results showed that the overall water production and water recovery of the system using the cycle-based optimized process control parameters (SCN 1) exceeds the MCDI system performance under three fixed parameter settings that were used at each stage of SCN 1 by 1.78% to 4.48% and 2.95% to 9.46%, respectively. Permutation-based and Shapley additive explanation (SHAP) coefficients were also employed for variable importance (VIMP) analysis to uncover the “black-box” nature of the ML models and to better understand the various features’ contributions to overall MCDI system performance.

Fast and Ultrasensitive Glycoform Analysis by Supercritical Fluid Chromatography-Tandem Mass Spectrometry.

The glycoform of a therapeutic monoclonal antibody (mAb) has a significant impact on its effector function as well as its safety and pharmacokinetics. Glycoform heterogeneity is influenced by various factors, including the producing cells and cell culture processes. Therefore, accurate glycoform characterization is essential for drug design, process optimization, manufacturing, and quality control of therapeutic mAbs. In this study, we developed a fast, quantitative, and highly sensitive analytical platform for glycan profiling by supercritical fluid chromatography-tandem mass spectrometry (SFC-MS/MS) and applied the technique to the glycan structural analysis of mAbs. To achieve both the highest sensitivity and the most comprehensive glycan profiling, we integrated our energy-resolved oxonium ion monitoring (Erexim) method with SFC-MS to construct a new SFC-Erexim technology. An 8 min analysis of bevacizumab, nivolumab, ramucirumab, rituximab, and trastuzumab by SFC-Erexim detected a total of 102 glycoforms, with a detection limit of 5 attomoles. The dynamic range of glycan abundance was over 6 orders of magnitude for bevacizumab analysis by SFC-Erexim compared to 3 orders of magnitude for conventional fluorescence HPLC analysis. This method revealed the glycan profile characteristics and lot-to-lot heterogeneity of various therapeutic mAbs. We were also able to detect a series of structural variations in pharmacologically important glycan structures. The SFC-MS-based glycoform profiling method will provide an ideal platform for the in-depth analysis of precise glycan structure and abundance.

RESEARCH AND IMPROVEMENT OF OIL FIELD DEWATERING

A steady increase in the water content of the extracted oil, an increase in the volume of heavy grades of oil with stable emulsions, an increase in the proportion of thick oil in the total volume of refined oil due to the imperfection of the oil dewatering process is far from a complete list of factors determining the importance and relevance of the issue of optimal automatic control of the oil dewatering process in the fields today. Although this problem is not new to the oil industry, its solution is not yet completed. This circumstance is due to insufficient knowledge of the processes of oil preparation due to the emerging objective difficulties and the lack of adequate models that could be used to solve the problems of stabilization and optimization of the oil dewatering process. The paper presents the research on the process of breaking highly-stable heavy oil emulsions by water evaporation. An equation has been derived and relevant curves have been plotted to calculate the required heating temperature. This study shows the dependence of hydrocarbon distillate yield on heating temperature, boiling point oil composition and the initial water cut of the emulsion. The technology of oil dehydration from the field has been tested.

Design and optimization of an inherently safe and sustainable process for the separation of anisole

Sustainable process design problems are multi-faceted approaches that pose challenges in terms of the evaluation of indicators to optimize inherent safety, economics, and control. In recent years, chemical engineering has taken advantage of recent advances in process systems engineering more notably considering intensification to generate green and more sustainable processes. This work presents the design and optimization of the anisole separation process under three intensified and one conventional schemes using a systematized methodology to find optimal designs of purification processes where the interactions between the different sustainability indicators are balanced to obtain optimal configurations. The indeces chosen in the course of process optimization (inherent safety, economic, and control) aim at generating green and more sustainable process alternatives. The results indicate that the most intensified processes, as is the case of the Dividing Wall Column (DWC), is the one with the best sustainability indicators, with better inherent safety and better operability despite it being an intensified process with fewer operational degrees of freedom. In addition, the environmental impact after the optimization of each of the processes were evaluated. Separation through DWC resulted in 2.43% savings in economic terms, 0.31% in inherent safety, improved controllability performance, and a 3.35% reduction in environmental impact, in comparison with the conventional sequence used as baseline. This indicates that in the case of anisole separation, process intensification can yield a more sustainable process design.

Mechanism for the induction and enhancement of inclusions on crack source and simulation analysis for hot tearing tendency of aluminum alloy

Inclusions are typical discontinuous defects, those defects significantly reduce the local mechanical properties of castings during solidification, and play a strong role in inducing and enhancing hot tearing in alloys. Based on the coupling characterization of micro solidification morphology and macro inclusions, this paper proposed a construction method of heterogeneous model containing multi-scale information, and carried out dynamic analysis on the evolution process of stress field at the end of solidification, so as to investigate the influence mechanism of inclusions on the initiation and propagation of hot tearing. The results showed that inclusions and their local content significantly changed the stress field at the end of solidification, which easily resulted in local stress concentration and induced crack source in defects area. During the crack propagation, nearby inclusions greatly reduced the propagation resistance and promoted the formation of hot tearing, which had significant impact on the propagation path of hot tearing. Introducing inclusions defect into stress analysis and comprehensively considering the induction and enhancement of inclusions on hot tearing can better improves the prediction accuracy of hot tearing, which has engineering guiding significance for casting process optimization and product quality control.

A double-layer model predictive control algorithm for valve position control systems

Valve position control systems have the characteristics of combining static and dynamic, as well as fast and slow, which enables the system’s quick response while maintaining some economic performance. However, there is further potential to optimize the economic performance of the valve position control system under conventional control algorithms. This article proposes an improved double-layer model predictive control algorithm for the valve position control system to exploit the potential economic performance while satisfying control goals. In the steady-state optimization layer of the improved double-layer structural model predictive control, the ideal residence point of the key manipulated variable is calculated from the set point of the controlled variable and the steady-state model to characterize the coupling relationship between the primary and secondary manipulated variables in the valve position control system, which realizes the successive actuation feature between the manipulated variables. In order to improve the response speed of the system, the dynamic control layer adopts a two-mode switching strategy, where the mode switching symbol is the time at which the controlled variable first approaches a given steady-state threshold value. Based on the feedback from the dynamic control layer, the steady-state optimization layer optimizes set points of controlled and manipulated variables that can handle disturbances entered into the process at any time to improve the economic performance of valve position control systems. This article implements classical control methods via advanced process strategies to consider the optimal control of processes from a superior perspective.

Algorithm for Solving the Problem of Optimal Control of a Chemical-Technological Process with Terminal Constraints

Introduction. The problem of determining the optimal mode parameters during the mathematical modeling of chemical and technological processes is the most important. Numerical methods and algorithms for the solution provide the basis for developing software packages to calculate processes and their digital twins. The mathematical model of the chemical-technological process can be described by a system of differential equations, highlighting the phase variables that determine the state of the process, and the control parameters, which can be changed and thereby affect the course of the process. The aim of the work is to develop a numerical algorithm for solving the problem of optimal control of a chemical-technological process in the presence of terminal constraints and the constraints on the control parameter. Materials and Methods. There was formulated the problem of optimal control in general terms. To solve it, the penalty method and method of artificial immune systems were applied. There was described a method for including constraints in the penalty function and for choosing a sequence of coefficients with which the penalty is taken. To overcome local extrema, a random choice of initial values of control parameters was used. Results. The article presents a step-by-step numerical algorithm for solving the problem of optimal control of a chemical-technological process with terminal constraints. A computational experiment was carried out for a model example, as a result of which the structure of the optimal process control and the corresponding optimal trajectories of phase variables are determined. It is shown that the calculated solution of the optimal control problem consists with the solution obtained by the needle linearization method. Discussion and Conclusion. The developed algorithm allows finding a numerical solution to the problem of optimal control of a chemical-technological process with terminal constraints. The solution does not depend on the choice of the initial approximation.

Multi-objective optimal control of bioconversion process considering system sensitivity and control variation

This paper aims to determine a time-varying dilution rate of the feed medium which maximizes mean productivity while minimizing both system sensitivity and control cost in glycerol continuous culture. A multi-objective optimal control problem subjected to a nonlinear control system and its auxiliary system is proposed and transformed into a finite-dimensional large-scale multi-objective optimization problem via discretization and time-scaling transformation. A multi-objective competitive swarm optimization algorithm is constructed to solve this transformed problem based on pairwise competition, mutation operation and environment selection. The Pareto front obtained shows the distribution of the optimal target value which indicates the rationality and feasibility of the control system under the approximate optimal strategy. Numerical simulations verify the robustness of the optimal control system and the effectiveness of our numerical solution method. The value of IGD shows the good accuracy and distribution of the proposed multi-objective optimization algorithm.

Applications of machine learning in thermochemical conversion of biomass-A review

Thermochemical conversion of biomass has been considered a promising technique to produce alternative renewable fuel sources for future energy supply. However, these processes are often complex, labor-intensive, and time-consuming. Significant efforts have been made in developing strategies for modeling thermochemical conversion processes to maximize their performance and productivity. Among these strategies, machine learning (ML) has attracted substantial interest in recent years in thermochemical conversion process optimization, yield prediction, real-time monitoring, and process control. This study presents a comprehensive review of the research and development in state-of-the-art ML applications in pyrolysis, torrefaction, hydrothermal treatment, gasification, and combustion. Artificial neural networks have been widely employed due to their ability to learn extremely non-linear input–output correlations. Furthermore, the hybrid ML models outperformed the traditional ML models in modeling and optimization tasks. The comparison between various ML methods for different applications, and insights about where the current research is heading, is highlighted. Finally, based on the critical analysis, existing research knowledge gaps are identified, and future recommendations are presented.

Data-driven constrained reinforcement learning for optimal control of a multistage evaporation process

It is challenging for reinforcement learning to solve the optimal control problem of industrial processes with constraints under uncertain operating conditions. In this context, this paper proposes a novel data-driven constrained reinforcement learning algorithm for the optimal control of a multistage evaporation process consisting of multiple evaporators in series with coupled liquid levels. We first formulate the optimal control problem as a constrained Markov decision process. Then, with the cumulative tracking error of the outlet liquor density taken as the cumulative constraint, a Lagrangian-based constrained policy optimization is developed. The fast setpoint tracking is achieved by gradient iteration of the policy and the dual variable. An action correction layer based on the online sequential version of random vector functional-link networks is built on the output of the policy network to address the instantaneous constraints of the liquid levels. The infeasible action is corrected in real-time so as to keep the liquid levels in each evaporator within operating range. Finally, we utilize both on-policy and off-policy data generated by the interaction between the constrained policy and the evaporation environment to update our algorithm, which is more data-efficient. Experiments have been carried out on a multistage evaporation system, and the results validate the effectiveness of the proposed algorithm.

Single and Multiobjective Optimal Control of the Wastewater Treatment Process

Single and Multiobjective Optimal Control of the Wastewater Treatment Process

Direct prediction of steam cracking products from naphtha bulk properties: Application of the two sub-networks ANN

Steam cracking of naphtha is an important process for the production of olefins. Applying artificial intelligence helps achieve high-frequency real-time optimization strategy and process control. This work employs an artificial neural network (ANN) model with two sub-networks to simulate the naphtha steam cracking process. In the first feedstock composition ANN, the detailed feedstock compositions are determined from the limited naphtha bulk properties. In the second reactor ANN, the cracking product yields are predicted from the feedstock compositions and operating conditions. The combination of these two sub-networks has the ability to accurately and rapidly predict the product yields directly from naphtha bulk properties. Two different feedstock composition ANN strategies are proposed and compared. The results show that with the special design of dividing the output layer into five groups of PIONA, the prediction accuracy of product yields is significantly improved. The mean absolute error of 11 cracking products is 0.53wt% for 472 test sets. The comparison results show that this indirect feedstock composition ANN has lower product prediction errors, not just the reduction of the total error of the feedstock composition. The critical factor is ensuring that PIONA contents are equal to the actual values. The use of an indirect feedstock composition strategy is a means that can effectively improve the prediction accuracy of the whole ANN model.

Model Based Optimal Control of the Photosynthetic Growth of Microalgae in a Batch Photobioreactor

The present paper investigates the optimal control of the photosynthetic growth process in an artificial light photobioreactor operated in batch mode, the objective being to find an optimal incident light intensity for which the consumption of light energy, for any amount of newly formed biomass, is minimal. By using a simple and reliable model for the photosynthetic growth of microalgae of microalgae, predictions can be made on the quantity of produced biomass and on the amount of light consumed, whose ratio gives the biomass yield on light energy. This variable is unimodal on the allowed range of incident light intensities and has been used as objective function. An improved objective function is proposed by using the specific growth rate and a weighing factor that allows obtaining the desired amount of biomass while the light energy consumption is optimal. A closed-loop control structure has been designed based on the developed optimization algorithm. The optimal controller has been validated in simulation, comparing different lengths of the optimization horizon and the sampling period. It was found that a bigger sampling period, for the cases where there is no online information on the biomass concentration, does not significantly affect the productivity. The optimization algorithm can be used either online or offline, being useful for various experimental setups.

Intelligent Optimal Control of Sewage Treatment Based on Multiobjective Evolutionary Algorithm

In order to solve the problem of optimal control of the sewage treatment process based on a multiobjective evolutionary algorithm, an intelligent optimal control of sewage treatment based on a multiobjective evolutionary algorithm is proposed in this paper. In this paper, the decomposition based multiobjective evolutionary algorithm (MOEA/D) is improved, and it is expected that the uniformly distributed approximate Pareto frontier can be obtained with fewer evolution times. For each new solution generated by the MOEA/D algorithm, the improved algorithm in this paper finds the most suitable subproblem for the new solution from all subproblems and replaces the population in its neighborhood. On the basis of the original subproblem, it carries out secondary optimization to improve the utilization rate of the children and then finds the approximate Pareto frontier in the optimization problem with fewer iterations. The experimental results show that AE, PE, and EC Based on SS–MOEA/D optimal control method are reduced by 6.91%, 1.54%, and 5.58%, respectively. Conclusion. The algorithm significantly reduces the number of steps to find the Pareto frontier, significantly improves the performance of the MOEA/D algorithm, and achieves the optimization goal in the optimization of the sewage treatment process.