Computers & Chemical Engineering最新文献_第2页

Adaptive fault detection via machine unlearning 基于机器学习的自适应故障检测

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-12 DOI: 10.1016/j.compchemeng.2025.109394

Amal Anto , Deepak Kumar , Hariprasad Kodamana , Manojkumar Ramteke

{"title":"Adaptive fault detection via machine unlearning","authors":"Amal Anto , Deepak Kumar , Hariprasad Kodamana , Manojkumar Ramteke","doi":"10.1016/j.compchemeng.2025.109394","DOIUrl":"10.1016/j.compchemeng.2025.109394","url":null,"abstract":"<div><div>Data-driven models are widely relied upon for process fault detection. However, their performance is susceptible to the quality of training data. Anomalous data in training or online updates degrade detection models, increasing false alarms or missed detections, and retraining models on corrected datasets is impractical for real-time fault detection. To address this problem, we propose a machine unlearning based adaptive fault detection that updates the model parameters to selectively remove the influence of faulty data from trained models without retraining and compromising the accuracy on normal data. We implement blindspot unlearning on four deep learning-based fault detection models: Autoencoder (AE), Variational Autoencoder (VAE), LSTM-AE, and LSTM-VAE, and evaluate their performance on two benchmark datasets: the Tennessee Eastman Process (TEP) and the Wastewater Treatment Plant (WWTP). We evaluate the models’ fault detection performance before and after unlearning. Our findings demonstrate that unlearning improves fault detection performance while significantly reducing computational overhead. Compared to the original models, unlearned models showcased improved fault detection rate, achieving a 44% increase on the TEP dataset and reaching 90% on the WWTP dataset. Unlearned models achieved fault detection performance comparable to retrained models, reducing computational time by up to 46% on TEP and 33% on WWTP. This validates the effectiveness of machine unlearning for adaptive fault detection.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109394"},"PeriodicalIF":3.9,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145118169","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Synchronous optimization framework for the integrated hydrogen network and heat exchanging system: A decomposition optimization strategy of the system based on the structure generation method 氢网换热一体化系统同步优化框架：基于结构生成法的系统分解优化策略

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-12 DOI: 10.1016/j.compchemeng.2025.109364

Shizhao Chen , Dan Yang , Xin Peng , Weichao Ding , Shuai Tan , Weimin Zhong

引用次数: 0

Uniting neural network-based control and model predictive control: Application to a large-scale nonlinear process 结合神经网络控制与模型预测控制：在大规模非线性过程中的应用

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-12 DOI: 10.1016/j.compchemeng.2025.109396

Arthur Khodaverdian , Dhruv Gohil , Panagiotis D. Christofides

{"title":"Uniting neural network-based control and model predictive control: Application to a large-scale nonlinear process","authors":"Arthur Khodaverdian , Dhruv Gohil , Panagiotis D. Christofides","doi":"10.1016/j.compchemeng.2025.109396","DOIUrl":"10.1016/j.compchemeng.2025.109396","url":null,"abstract":"<div><div>This work proposes a method to overcome the issue of nonlinear model predictive control (MPC) requiring practically infeasible computation times for large-scale systems. In particular, the use of Neural Networks (NN) to approximate nonlinear MPC calculated control actions in a real-time closed-loop implementation with externally enforced stability guarantees is explored. Using Lyapunov-based stability constraints, the reduced computational complexity of NNs paired with the ability to train using MPC that would be infeasible to apply in real-time systems (due to the use of a large prediction horizon to ensure good closed-loop performance) enables the training of an NN-based approximate control policy that directly substitutes MPC. With a stabilizing fallback controller available, this NN controller enables real-time stabilizing control of high-dimensional nonlinear systems. To demonstrate this, Aspen Plus Dynamics, a dynamic chemical process simulation software, is used to create a large-scale nonlinear chemical process example. Using an NN trained off of an offline MPC using a first-principles model and a large prediction horizon, a comprehensive study of the resulting closed-loop behavior is carried out to evaluate the closed-loop stability, performance, and robustness properties of the approach.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109396"},"PeriodicalIF":3.9,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145044734","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Safe reinforcement learning for optimization of batch processes with uncertainties: A Bayesian predictive exploration approach 不确定批处理优化的安全强化学习：贝叶斯预测探索方法

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-12 DOI: 10.1016/j.compchemeng.2025.109391

Jingsheng Qin , Lingjian Ye , Xiaofeng Yuan

引用次数: 0

Nonlinear predictive regulation of an integrated green hydrogen and ammonia production system under time-varying renewable energy supply 时变可再生能源供应下绿色氢氨一体化生产系统的非线性预测调节

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-11 DOI: 10.1016/j.compchemeng.2025.109376

Thiago Oliveira Cabral , Davood B. Pourkargar

{"title":"Nonlinear predictive regulation of an integrated green hydrogen and ammonia production system under time-varying renewable energy supply","authors":"Thiago Oliveira Cabral , Davood B. Pourkargar","doi":"10.1016/j.compchemeng.2025.109376","DOIUrl":"10.1016/j.compchemeng.2025.109376","url":null,"abstract":"<div><div>This work presents a comprehensive model for a modular system that integrates green hydrogen and ammonia production with renewable energy generation. The chemical module comprises a high-temperature water electrolyzer for hydrogen production and an ammonia synthesis reactor. When solving the models over time, the system exhibits complex yet predictable dynamics, with the chemical module having a much faster response than other components. Under typical weather conditions, the renewable energy module generates over 50 kW for most of the day, partially meeting the chemical module’s energy demands. Nonlinear model predictive control (NMPC) is employed to manage the operation of the chemical module in response to variable renewable energy availability. The proposed NMPC framework determines the optimal supplemental energy required from the conventional energy grid to sustain the process. When renewable energy availability is high, the controller minimizes grid energy usage, maintaining the chemical module near its desired operating conditions with minimal reliance on external sources. Conversely, during low renewable energy availability periods, the controller increases grid energy acquisition to ensure stable system operation, demonstrating a greater dependence on external energy supplies.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109376"},"PeriodicalIF":3.9,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145105881","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Application of unsupervised machine learning methods to actor–critic structures in reinforcement learning for training and online implementation 无监督机器学习方法在训练和在线实施强化学习中的行为-批评结构中的应用

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-11 DOI: 10.1016/j.compchemeng.2025.109392

Daniel Beahr, Debangsu Bhattacharyya

{"title":"Application of unsupervised machine learning methods to actor–critic structures in reinforcement learning for training and online implementation","authors":"Daniel Beahr, Debangsu Bhattacharyya","doi":"10.1016/j.compchemeng.2025.109392","DOIUrl":"10.1016/j.compchemeng.2025.109392","url":null,"abstract":"<div><div>A fundamental obstacle to the implementation of reinforcement learning (RL) to continuous systems is the large amount of data and training that must take place to achieve a satisfactory control policy. This is exacerbated when the focus is an online implementation. It is the goal of this work to investigate the use of unsupervised learning to make more efficient decisions with the data available, both for learning and exploration in the typical RL algorithm. Gaussian mixture models (GMMs) are used to form a probabilistic prediction for the outcome of proposed actions during exploration, while high-performing data points are subsequently over-sampled to accelerate learning and convergence. With respect to the exploration policy, GMMs are used to predict the outcomes for given actions and used for preventing undesired exploratory actions that can lead to significant loss in control performance and/or violation of safety or other operational constraints. The proposed approach was integrated within a Deep Deterministic Policy Gradient algorithm and was applied to the control of a selective catalytic reduction unit. It was found that a satisfactory policy was found faster and with less overall performance degradation than the standard RL approaches.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109392"},"PeriodicalIF":3.9,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145105345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Process modeling and sensitivity analysis for integrated carbon capture and conversion with ionic liquids 离子液体碳捕获与转化过程建模及敏感性分析

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-11 DOI: 10.1016/j.compchemeng.2025.109390

Adhika P. Retnanto , Mark A. Stadtherr , Michael Baldea

{"title":"Process modeling and sensitivity analysis for integrated carbon capture and conversion with ionic liquids","authors":"Adhika P. Retnanto , Mark A. Stadtherr , Michael Baldea","doi":"10.1016/j.compchemeng.2025.109390","DOIUrl":"10.1016/j.compchemeng.2025.109390","url":null,"abstract":"<div><div>Integrated carbon capture and conversion (ICCC) processes convert captured CO<sub>2</sub> directly in the capture medium, avoiding the solvent regeneration, separation, and compression required in conventional capture–conversion (CCC) schemes. We present a prototype ionic liquid (IL)-based ICCC process integrated with an ethylene plant, producing methanol via <figure><img></figure> hydrogenation. A sensitivity analysis evaluates process energy use with respect to (1) CO<sub>2</sub> removal rate and (2) <figure><img></figure> conversion, under both mass-transfer- and reaction-rate-limited regimes. Results show that energy demand is strongly linked to IL recirculation and vapor recycle rates. Importantly, a reaction-rate-limited regime is advantageous: fast CO<sub>2</sub> absorption (even with slower hydrogenation) allows unreacted CO<sub>2</sub> in the vapor recycle to be reabsorbed, lowering flowrates and energy use. Because purging in the conversion loop affects net CO<sub>2</sub> removal, it is critical to design ICCC processes holistically, integrating capture and conversion to optimize overall carbon and energy efficiency.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109390"},"PeriodicalIF":3.9,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145105882","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Modeling and multi-criteria assessment of polyamine-based CO2 capture in rotating packed beds using artificial intelligence 基于人工智能的旋转填料床多胺CO2捕集建模与多准则评估

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-10 DOI: 10.1016/j.compchemeng.2025.109400

Theofilos Xenitopoulos , Athanasios I. Papadopoulos , Panos Seferlis

{"title":"Modeling and multi-criteria assessment of polyamine-based CO2 capture in rotating packed beds using artificial intelligence","authors":"Theofilos Xenitopoulos , Athanasios I. Papadopoulos , Panos Seferlis","doi":"10.1016/j.compchemeng.2025.109400","DOIUrl":"10.1016/j.compchemeng.2025.109400","url":null,"abstract":"<div><div>Rotating Packed Beds (RPB) are receiving wide attention as a CO<sub>2</sub> capture technology that intensifies mass transfer and enables substantial equipment size reduction compared to conventional packed columns. This study employs a data-driven approach to model CO<sub>2</sub> absorption in a RPB system through experimental literature data for five polyamines across various liquid flow rates, rotational speeds and concentration. Polyamines are promising solvents as the combination of multiple amine groups in the same molecule enables high absorption rate, kinetics and CO<sub>2</sub> solubility, among others. The Artificial Intelligence (AI) algorithms used are Partial Least Squares (PLS), Random Forest Regression (RFR), Light (LightGBM) and Extreme Gradient Boosting Machine (XGBoost), Categorical Boosting (CatBoost), Support Vector Regressor (SVR) and Multilayer Perceptron (MLP). Shapley Additive Explanations (SHAP) is used to analyze the combined influence of key parameters on absorption efficiency. The LightGBM model achieved the highest predictive accuracy in carbon capture rate. It was then used for factorial design of simulations, and calculation of RPB motor power requirement enabling a multi objective assessment. Results revealed that ethylenediamine (EDA) offers superior trade-offs between carbon capture and energy requirement. The work underscores the potential of using directly process-level experimental data to model and investigate the performance in polyamine-based capture systems where sufficient data are not available to develop first-principles models.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109400"},"PeriodicalIF":3.9,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145105344","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Real-world implementation of offline reinforcement learning for process control in industrial dividing wall column 离线强化学习在工业分界墙柱过程控制中的实际实现

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-09 DOI: 10.1016/j.compchemeng.2025.109383

Joonsoo Park , Wonhyeok Choi , Dong Il Kim , Ha El Park , Jong Min Lee

{"title":"Real-world implementation of offline reinforcement learning for process control in industrial dividing wall column","authors":"Joonsoo Park , Wonhyeok Choi , Dong Il Kim , Ha El Park , Jong Min Lee","doi":"10.1016/j.compchemeng.2025.109383","DOIUrl":"10.1016/j.compchemeng.2025.109383","url":null,"abstract":"<div><div>Reinforcement Learning (RL) has emerged as a promising approach for automating industrial process control, particularly in handling stochastic disturbances and complex dynamics. However, conventional RL methods pose significant safety concerns in real-world applications due to their reliance on extensive real-time interactions with the environment. Offline RL, which derives an optimal policy solely from historical operational data, provides a safer alternative but remains underexplored in industrial chemical processes. In this study, we apply Calibrated Q-Learning (Cal-QL), an offline-to-online RL algorithm, to temperature control of an industrial dividing wall column (DWC). We propose a practical procedure for deploying offline RL in chemical plants, integrating a Long Short-Term Memory (LSTM) network with a Deep Q-Network (DQN) to effectively process time series data structure and discrete action distributions commonly encountered in plant operations. Extensive simulation studies and real-world experiments on an industrial DWC demonstrate the suitability of the proposed framework. We also highlight the critical role of reward function design in balancing short- and long-term objectives, significantly influencing control performance. Our best performing configuration achieved stable temperature control with a high automation ratio of 93.11%, underscoring the feasibility and practical effectiveness of offline RL for complex industrial plant operations.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109383"},"PeriodicalIF":3.9,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145044804","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimum design under uncertainty of CO2 capture, utilization, mineralization and sequestration networks using rotating packed beds 旋转填料床CO2捕集、利用、矿化和封存网络不确定性优化设计

IF 3.9 2区工程技术

Computers & Chemical Engineering Pub Date : 2025-09-08 DOI: 10.1016/j.compchemeng.2025.109398

Thomas Prousalis , Babis Kantouros , Athanasios I. Papadopoulos , Panos Seferlis

{"title":"Optimum design under uncertainty of CO2 capture, utilization, mineralization and sequestration networks using rotating packed beds","authors":"Thomas Prousalis , Babis Kantouros , Athanasios I. Papadopoulos , Panos Seferlis","doi":"10.1016/j.compchemeng.2025.109398","DOIUrl":"10.1016/j.compchemeng.2025.109398","url":null,"abstract":"<div><div>This study presents a comprehensive framework for the optimal design of integrated CCUS networks within industrial clusters. The framework incorporates rotating packed bed (RPB) reactors for solvent-based CO<sub>2</sub> capture, CO<sub>2</sub> capture by mineralization and CO<sub>2</sub> utilization for precipitated calcium carbonate (PCC) nanoparticles production, as well as SNG production, pipeline transportation, and geological sequestration. Rigorous models for these processes are used to derive linear or piece-wise linear surrogate models, through a systematic approach that ensures accurate process predictions. The derived models are integrated into a mixed-integer linear programming (MILP) model for efficient network optimization. The framework is applied to a case study involving an industrial cluster, comprising of 5 emitters (power generation, refinery, quicklime, cement and paper plants), 3 sequestration sites, and one minerals deposit site. The framework is tested under both deterministic and uncertain conditions, accounting for fluctuations in CO<sub>2</sub> capture costs, CO<sub>2</sub> utilization raw material and product market prices, and carbon permit price. The CO<sub>2</sub> capture cost is the most influential factor in the operation of the entire network, with mineralization becoming favorable beyond a specific threshold. SNG synthesis becomes viable with modest raw material price decrease and product market price increase, whereas PCC production is consistently selected as the most favorable utilization route. Under high carbon permit prices, partial CO<sub>2</sub> capture may be economically optimal, with the remainder offset through purchasing carbon permits. The findings provide strategic insights into economic viability of alternative CO<sub>2</sub> pathways and support informed decision-making for future CCUS infrastructure deployment under real-world complexities.</div></div>","PeriodicalId":286,"journal":{"name":"Computers & Chemical Engineering","volume":"204 ","pages":"Article 109398"},"PeriodicalIF":3.9,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145044732","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":2,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0