Dynamic Process Optimization Research Articles

On combining the stages of parametric identification and optimization of dynamic processes

In the paper, we propose a method for jointly solving problems of parametric identification of a mathematical model’s parameters and searching for the optimal mode of a dynamic object. As a result of this joint process, we obtain a suboptimal solution to the problem of controlling a dynamic object in the vicinity of the optimal mode. We present computation formulas and algorithms for implementing the proposed approach, as well as the results of computer-based experiments.

Создание и развитие высоконадежных информационно-управляющих систем с элементами искусственного интеллекта для перспективных технологических комплексов

Methods and means of improving complex technological equipment are shown through the operational assessment of the quality of implemented technologies realtime. This is proved by the fact that implementing technological operations on universal equipment each time requires modeling dynamic processes and taking into account a large number of uncertainty factors that affect the geometry generation quality. It is not possible to be a priori aware of these factors. There is a need to create new information technologies with the possibilities of universal application for immediate understanding of various dynamic processes in diagnostic, identification and control systems. Standard computer systems for statistical analysis and optimization of dynamic processes with the possibilities of universal application for various implementations of modern technologies have been introduced. The possibility of using integrative criteria and methods of artificial intelligence for diagnostic systems, identification and control of advanced technological complexes is shown. The implementation of information systems for the management of complex objects of various technological purposes is presented. The proposed modeling methods and approaches have been tested at various machine-building enterprises when processing parts on turning, milling and grinding machines, both universal and CNC. The research results made it possible to implement new principles of automated control and optimal adjustment of technological processes in real time and create an automated system for evaluating their quality, which allows increasing the efficiency and reliability of management decisions by conducting optimization directly on operating equipment. Based on the methods and approaches described above, new results have been obtained in the implementation of plasma technologies for the modification of geometrically complex surfaces of mechanical engineering products aimed at increasing wear resistance, hardness and other technical characteristics of the working surfaces of precision engineering products. A fairly complete approbation of methods, approaches, procedures and decision-making criteria for various technologies allows them to be recommended for universal applicability.

Dynamic grey relational analysis-based optimization of electrochemical machining processes

Dynamic grey relational analysis-based optimization of electrochemical machining processes

Towards a digital twin framework in additive manufacturing: Machine learning and bayesian optimization for time series process optimization

Laser directed-energy deposition (DED) offers notable advantages in additive manufacturing (AM) for producing intricate geometries and facilitating material functional grading. However, inherent challenges such as material property inconsistencies and part variability persist, predominantly due to its layer-wise fabrication approach. Critical to these challenges is heat accumulation during DED, influencing the resultant material microstructure and properties. Although closed-loop control methods for managing heat accumulation and temperature regulation are prevalent in DED literature, few approaches integrate real-time monitoring, physics-based modeling, and control simultaneously in a cohesive framework. To address this, we present a digital twin (DT) framework for real-time model predictive control of process parameters of the DED for achieving a specific process design objective. To enable its implementation, we detail the development of a surrogate model utilizing Long Short-Term Memory (LSTM)-based machine learning which uses Bayesian Inference to predict temperatures across various spatial locations of the DED-built part. This model offers real-time predictions of future temperature states. In addition, we introduce a Bayesian Optimization (BO) method for Time Series Process Optimization (BOTSPO). Its foundational principles align with traditional BO, and its novelty lies in our unique time series process profile generator with a reduced dimensional representation. BOTSPO is used for dynamic process optimization in which we deploy BOTSPO to determine the optimal laser power profile, aiming to achieve desired mechanical properties in a DED build. The identified profile establishes a process trajectory that online process optimizations aim to match or exceed in performance. This paper elucidates components of the digital twin framework, advocating its prospective consolidation into a comprehensive digital twin system for AM.

Thermodynamic cycle analysis of the fuel precooled multi-mode turbine engine mode transition process: Why? When? How?

The fluctuation of engine thrust during the mode transition process is a critical concern that poses a significant risk to aircraft safety. In order to address this issue, a thermodynamic model for mode transition in a dual fuel precooled multi-mode turbine engine is established, and three key issues in the mode transition process are comprehensively examined from the perspectives of steady-state analysis and dynamic process optimization: Why? When? How? Firstly, the basis for mode transition under high Mach number conditions is proposed for the first time from exergy efficiency. The engine exergy efficiency and combustion exergy efficiency in the turbofan mode decrease as the Mach number increases, while the turbojet mode exhibits superior combustion exergy efficiency. To achieve efficient operation, it is imperative to switch from the turbofan mode to the turbojet mode. Subsequently, a novel approach is proposed for determining the mode transition range based on bidirectional cyclic correction, which effectively addresses the limitation of large thrust fluctuations at the mode transition point determined by the conventional maximum state thrust continuous method. The mode transition envelope that can achieve a smooth transition between thrust and airflow rate is determined. Finally, a mode transition method is newly put forward by integrating directional adjustment and offline optimization. Throughout the entire mode transition envelope, the maximum thrust fluctuation during the process remains below 1 %, while the maximum airflow rate change stays within 3 %, which exhibits excellent efficacy in achieving smooth transitions.

A Mixed-Integer Dynamic and Stochastic Algae Process Optimization

With increased energy demand as it gets scarcer, a great deal of research is being carried out into alternatives to non - renewable energy resources. One of the promising studies is the biofuel production from micro algae. Microalgae are photosynthetic organisms and capture carbon dioxide, reducing emissions and providing valuable products (fuel, fertilizer, etc.). Thus, efficiency in the design and optimization of process related units are important. In this study, the optimal experimental conditions for Nannochloropsis Oculata were calculated under the constraints of the model equations and other process related constraints through simultaneous optimization approach. The economic evaluation of the process is also handled by introducing the uncertainty in the economic measures sampled from normal distribution to maximize the average profit. Unlike traditional approaches, the MINLP formulation, which is solved stochastically, dynamically, and simultaneously, provides more robust and reliable results, flexibility, improved decision making, reduced risks to be taken and a better understanding of risk factors.

Process optimization of microbial fermentation with parameter uncertainties via distributionally robust discrete control

There are some uncertain kinetic parameters in microbial fermentation system because of the unclear intracellular metabolic mechanisms. Considering the affection of these uncertain parameters on system performance, dynamic process optimization of the fermentation system can be modeled as a distributionally robust discrete control problem under moment uncertainty, which aims to maximize the mean productivity by optimizing the discrete-valued dilution rate function. Based on duality theory, the established min–max discrete optimal control problem is first transformed into a single level minimization problem, which is then discretized into a large-scale parameter optimization problem with semi-infinite constraint via time transformation and control parameterization. A new two-step estimation of distribution algorithm is developed to solve the obtained large-scale optimization problem. Numerical results show the feasibility and effectiveness of the proposed solution approach together with the superiority of the obtained control strategy considering parameter uncertainties.

Online Dynamic Optimization of Multi-Rate Processes with the Case of a Fluid Catalytic Cracking Unit

Due to operational limitations in the industrial field, the operating variables of fluid catalytic cracking units (FCCU) are of multiple operating frequencies, which are CO combustion promoter amount, recycle slurry flow rate, combustion air flow rate, heat escape, and reaction temperature, from low frequency to high frequency. There are usually two schemes for operation optimization of FCCU. The former is called single-rate, single-window optimization, whose operating variables are optimized only once in the whole operation cycle, which is easy to achieve, but the optimization effect is poor. The latter is called single-rate multi-window optimization, whose operating variables are optimized repeatedly and whose operation cycle is discretized into multiple optimization periods with the same frequency, which costs a heavy calculation burden and cannot adapt to the optimization variables with multiple operating frequencies. So, a multi-rate, variable-window online dynamic optimization method is proposed. In an operation cycle, the high-frequency operating variable is optimized in a short optimization period, and the low-frequency operating variable is optimized in a long optimization period; each optimization period has integer multiples to the minimum optimization period. Each optimized result for each optimization period is put into use online immediately. The optimization model involves the time domain differential equations, integral cost objective function, and measured disturbances. The experimental results show that compared with the single-rate, single-window optimization method and single-rate multi-window optimization method, the optimization effect of multi-rate, variable-window online dynamic optimization is better than single-rate, single-window optimization but worse than single-rate multi-window optimization. However, the optimization results are consistent with the operation frequency of each optimization variable, which can be implemented in complex chemical processes and increase certain economic benefits.

Overcoming the modeling bottleneck: A methodology for dynamic gray-box modeling with optimized training data

For model-based control and real-time optimization, the existence of sufficiently reliable process models is a prerequisite. The development of such models often is the bottleneck in developing advanced solutions and particularly challenging in cases where the underlying physico-chemical phenomena are not fully understood. When process data is available, machine learning (ML) approaches can be used to build data-based models but the validity of such models is limited to the conditions for which sufficiently rich training data was available. To combine the flexibility of ML-models and the reliability of purely mechanistic models, gray-box models can be applied. In this work we consider gray-box models in which ML-submodels are embedded into differential–algebraic equations that have been derived from first principles. To identify the ML-submodels is challenging as no explicit input/output data for the submodels is available. Therefore, suitable training data must first be estimated from the available data. We propose a systematic methodology for developing gray-box models in a sequence of steps, employing specific regularizations and numerical techniques. We apply the methodology to two case studies: a distillation column with a reaction occurring in the rectifying section, and a fermentation process of a sporulating bacterium. The value of the fermentation model is demonstrated by its use in dynamic process optimization.

A DeepAR based hybrid probabilistic prediction model for production bottleneck of flexible shop-floor in Industry 4.0

Predicting production bottleneck plays an important role in dynamic production process optimization and throughput improvement of flexible shop-floor. However, great challenges aroused by the high uncertainty of shop-floor environment and the interactive effects from many influencing factors still remains on production bottleneck prediction. Therefore, a Deep Auto-Regressive Neural Network with hybrid model (H-DeepAR) is developed to provide production bottleneck probabilistic prediction. In this proposed model, convolution neural network (CNN) and multi-head attention mechanism (MHA) are constructed to capture high-level representations from the multivariate input data. Then, a gated recurrent unit (GRU) is leveraged to capture temporal correlations of production bottleneck. After that, fully connection layers (FCN) are designed to make production bottleneck probability prediction. Finally, a case study based on aircraft overhaul shop-floor is illustrated to validate the effectiveness and superiority of the proposed method. The results indicate that the proposed model yields better performance than the benchmark models.

Key influencing factors for the success of external innovation strategies in the biotechnology industry

PurposeNumerous biotechnology and pharmaceutical firms have undergone considerable changes and adapted to the challenge of developing sustainable products and services. However, few studies have explored the factors that contribute to the success of external innovation and value co-creation strategies adopted by biotechnology and pharmaceutical firms. The purpose of this study is to examine how biotechnology and pharmaceutical industries use value co-creation strategies to obtain external resources.Design/methodology/approachThis study developed a conceptual framework based on the relevant literature. The study applied a resource-based approach, dynamic capability theory and a qualitative multiple-case study design to investigate several research questions; semi-structured interviews were conducted with representatives from 11 biotechnology/pharmaceutical firms in Taiwan, and the data extracted from the interview content were axially coded.FindingsThis study revealed that factors such as dynamic marketing capabilities and process optimization contributed to the success of the aforementioned strategies; several propositions were also developed on the basis of the literature review and coded data, thereby providing insights regarding the relative efficacy and propriety of various external innovation and value co-creation strategies and models in various situations and contexts. Firms and technology providers might enter a technology licensing agreement, establish a joint venture company; participate in a merger/acquisition depending on their size, research and development capabilities; or goals and time- and cost-related factors.Originality/valueThe main original contributions of this study are the proposed conceptual framework and the insights provided regarding the relative efficacy and propriety of different external innovation and value co-creation strategies and models in different situations and contexts.

Understanding the effect of temperature downshift on CHO cell growth, antibody titer and product quality by intracellular metabolite profiling and in vivo monitoring of redox state.

The strategy of temperature downshift has been widely used in the biopharmaceutical industry to improve antibody production and cell-specific production rate (qp ) with Chinese hamster ovary cells (CHO). However, the mechanism of temperature-induced metabolic rearrangement, especially important intracellular metabolic events, remains poorly understood. In this work, in order to explore the mechanisms of temperature-induced cell metabolism, we systematically assessed the differences in cell growth, antibody expression, and antibody quality between high-producing (HP) and low-producing (LP) CHO cell lines under both constant temperature (37°C) and temperature downshift (37°C→33°C) settings during fed-batch culture. Although the results showed that low-temperature culture during the late phase of exponential cell growth significantly reduced the maximum viable cell density (p < 0.05) and induced cell cycle arrest in the G0/G1 phase, this temperature downshift led to a higher cellular viability and increased antibody titer by 48% and 28% in HP and LP CHO cell cultures, respectively (p < 0.001), and favored antibody quality reflected in reduced charge heterogeneity and molecular size heterogeneity. Combined extra- and intra-cellular metabolomics analyses revealed that temperature downshift significantly downregulated intracellular glycolytic and lipid metabolic pathways while upregulated tricarboxylic acid (TCA)cycle, and particularly featured upregulated glutathione metabolic pathways. Interestingly, all these metabolic pathways were closely associated with the maintenance of intracellular redox state and oxidative stress-alleviating strategies. To experimentally address this, we developed two high-performance fluorescent biosensors, denoted SoNar and iNap1, for real-time monitoring of intracellular nicotinamide adenine dinucleotide/nicotinamide adenine dinucleotide + hydrogen (NAD+ /NADH) ratio and nicotinamide adenine dinucleotide phosphate (NADPH) amount, respectively. Consistent with such metabolic rearrangements, the results showed that temperature downshift decreased the intracellular NAD+ /NADH ratio, which might be ascribed to the re-consumption of lactate, and increased the intracellular NADPH amount (p < 0.01) to scavenge intracellular reactive oxygen species (ROS) induced by the increased metabolic requirements for high-level expression of antibody. Collectively, this study provides a metabolic map of cellular metabolic rearrangement induced by temperature downshift and demonstrates the feasibility of real-time fluorescent biosensors for biological processes, thus potentially providing a new strategy for dynamic optimization of antibody production processes.

Digital twin driven production progress prediction for discrete manufacturing workshop

• This article focuses on a novel synchronous update method for digital twin workshop model with integrating learning approach. • The representation model is established to describe the performance degradation rules of discrete manufacturing workshop. • An Adaboost-DNN-LSTM based synchronous update method is proposed to guarantee the consistency between DTM and physical workshop. • A competitive election mechanism is designed to select base learner among DNN and LSTM between historical and dynamic sample. In make-to-order manufacturing enterprises, accurate production progress (PP) prediction is an important basis for dynamic production process optimization and on-time delivery of orders. Digital twin technology offers an enabling tool for PP analysis. Although the production process can be observed, analyzed, and controlled in real-time by digital twin model (DTM), there exist some uncertain events, degradation of manufacturing elements, and abnormal disturbance in physical workshop (PW), which would cause the deviation between DTM and PW performance and affect the prediction accuracy of PP. Synchronous evolution of DTM for precision holding to ensure the consistency between DTM and the performance of PW, and guarantee the accuracy of DTM is still a challenging issue, especially when dealing with new dynamic samples for complex production environment of discrete manufacturing workshop (DMW). This article focuses on how to effectively construct DTM synchronous update methods based on dynamic sample data for DMW. This study proposes a representation model of performance degradation and an Adaboost-DNN-LSTM based synchronous update model with competitive election mechanism to enhance the accuracy of PP prediction with time in industrial environment. The experiment is conducted in the realistic production dataset, which demonstrates that the proposed synchronous evolution model has good performance for realizing the synchronization of the performance of physical workshop in industrial environment, and can greatly improve the prediction ability for PP.

Generation of parallel colocation difference schemes using the integro-interpolation approach

The issues of development of stable parallel difference schemes of a given order of accuracy, focused on the discretization of differential equations or their systems with concentrated and distributed parameters using an integro-interpolation approach to the generation of calculation coefficients, are considered. The proposed difference schemes improve the efficiency of parallel simulation and optimization of dynamic processes described by evolution equations with distributed parameters. Such equations are characterized by the absence of analytical solutions. The introduction of higher-order derivatives into difference schemes improves the accuracy of the obtained results and equalizes the approximation orders for all calculation points in the block.The proposed approach is universal for obtaining difference equations of various types. The computational formulas obtained on the basis of this approach for multipoint difference equations have less computational complexity and are oriented to implementation in computing systems with a parallel architecture.&#x0D; Keywords: integro-interpolation approach, difference schemes, colocation block methods, parallel computing&#x0D;

Control-oriented dynamic process optimization of solid oxide electrolysis cell system with the gas characteristic regarding oxygen electrode delamination

Fast response, strong reliability, and high efficiency as the critical objectives of a solid oxide electrolysis cell (SOEC) system, which requires the system to operate under optimal operating conditions and switching manners. Therefore, various research has been carried out on its characteristic security. Most of them focus on the thermal characteristics security without further considering the security of gas characteristics regarding oxygen electrode delamination. However, this is one of the most critical degradations of SOEC. Herein, a novel SOEC system dynamic model with the gas characteristic regarding oxygen electrode delamination is proposed for the first time to optimize operation. Specially, the advantages are discussed through the optimization analysis of both the secure operation range regarding certain operating conditions and the steady-state operating points of the system. Subsequently, the dynamic process constraints of oxygen electrode delamination, fuel starvation, and power fast tracking during dynamic switching are investigated. The feasibilities of multi-step, parabola, and ramp schemes in solving the problems above are also evaluated. The outcomes show that the ramp scheme is preferred. Furthermore, the priority issue between oxygen electrode delamination and fuel starvation is studied, and corresponding dynamic process optimization strategies are developed for high and low initial power with a two-step switching method. After that, the dynamic process optimizations of two representative cases are investigated in detail. The optimization results show that the system efficiency is increased, and the safe ramp duration under the low initial power is significantly shortened. The findings of this study could provide a basis for developing control strategies for SOEC systems with fluctuating energy input.

Dynamic Bayesian network-based disassembly sequencing optimization for electric vehicle battery

The sharply increasing end-of-life (EOF) battery volume in the global complex energy market has created significant challenges for its recycling and reuse, to reduce environmental pollution and resource waste, and efforts have been focused on the disassembly process considering the uncertainty of electric vehicle (EV) battery pack categories and quality. Compared with traditional disassembly, the EV battery disassembly process needs to consider more uncertainty factors for each EOF battery pack to represent its disassembly structure, which significantly reduces disassembly production efficiency. Even though sequence optimization methods for the disassembly process have been developed to solve these problems, there are still two important challenges that remain: uncertain disassembly structure representation and optimal disassembly sequence selection. To address these challenges, this paper proposes a dynamic disassembly Bayesian network approach based on an EV battery disassembly graph model. This method offers dynamic process optimization to manufacturers to deduce the optimal disassembly sequences using the forward–backward algorithm and the Viterbi decoding algorithm. To validate the proposed method, an EOF battery is used to demonstrate the disassembly sequence selection, which indicates the possibility of massive EV battery disassembly prediction.

Optimal control and the Pontryagin's principle in chemical engineering: History, theory, and challenges

AbstractIn the mid‐1950s, Pontryagin et al. published a principle that became a fundamental concept in optimal control (OC) theory. The principle provides theoretical and practical methods to find the solution of OC problems, in particular, open‐loop control problems. In chemical engineering, the principle has played an important role as a decision making framework for more than 60 years. This study gathers the main contributions on the application of the Pontryagin's principle to the dynamic optimization of chemical processes. A concise overview of the optimality conditions for a wide class of constrained OC problems is provided. Numerical methods to solve the necessary conditions and strategies to address inequality constraints are summarized. The information and illustrative case study presented in this work can be used as a guide to implement the principle in different settings. Opportunities for further application of the principle in relevant chemical engineering problems are also discussed.

Robust dynamic optimization for nonlinear chemical processes under measurable and unmeasurable uncertainties

AbstractWe formulate an integrated framework for the robust dynamic optimization of nonlinear chemical processes under measurable and unmeasurable uncertainties. An affine decision rule is proposed to approximate the causal dependence of the wait‐and‐see decision variables on the gradually revealed measurable uncertainties. To overcome the computational intractability of the proposed model, a linearization technique based on the first‐order Taylor expansion is introduced around the nominal values of uncertainties to derive the robust dynamic counterpart, which can be discretized to a large‐scale nonlinear programming (NLP) formulation. Effects of first discretizing the dynamic models or introducing the affine decision rule are investigated. The proposed framework is also compared with the state‐of‐the‐art re‐optimization and traditional robust optimization approaches. An illustrative example and an industrial semi‐batch 2‐mercaptobenzothiazole production case are involved to demonstrate the advantages and applicability of the proposed framework.

Dynamic simulation and optimization of anaerobic digestion processes using MATLAB

Time series-based modeling provides a fundamental understanding of process fluctuations in an anaerobic digestion process. However, such models are scarce in literature. In this work, a dynamic model was developed based on modified Hill’s model using MATLAB, which can predict biomethane production with time series. This model can predict the biomethane production for both batch and continuous process, across substrates and at diverse conditions such as total solids, loading rate, and days of operation. The deviation between literature and the developed model was less than ± 7.6%, which shows the accuracy and robustness of this model. Moreover, statistical analysis showed there was no significant difference between literature and simulation, verifying the null hypothesis. Finding a steady and optimized loading rate was necessary to an industrial perspective, which usually requires extensive experimental data. With the developed model, a stable and optimal methane yield generating loading rate could be identified at minimal input.

An investigation for integration of deep learning and digital twins towards Construction 4.0

PurposeThe purpose of this paper is to investigate the potential integration of deep learning (DL) and digital twins (DT), referred to as (DDT), to facilitate Construction 4.0 through an exploratory analysis.Design/methodology/approachA mixed approach involving qualitative and quantitative analysis was applied to collect data from global industry experts via interviews, focus groups and a questionnaire survey, with an emphasis on the practicality and interoperability of DDT with decision-support capabilities for process optimization.FindingsBased on the analysis of results, a conceptual model of the framework has been developed. The research findings validate that DL integrated DT model facilitating Construction 4.0 will incorporate cognitive abilities to detect complex and unpredictable actions and reasoning about dynamic process optimization strategies to support decision-making.Practical implicationsThe DL integrated DT model will establish an interoperable functionality and develop typologies of models described for autonomous real-time interpretation and decision-making support of complex building systems development based on cognitive capabilities of DT.Originality/valueThe research explores how the technologies work collaboratively to integrate data from different environments in real-time through the interplay of the optimization and simulation during planning and construction. The framework model is a step for the next level of DT involving process automation and control towards Construction 4.0 to be implemented for different phases of the project lifecycle (design–planning–construction).