Digital Chemical Engineering最新文献_第6页

Toward predicting CO2 loading capacity in monoethanolamine (MEA) aqueous solutions using deep belief network 基于深度信念网络的单乙醇胺（MEA）水溶液CO2承载能力预测

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-04-01 DOI: 10.1016/j.dche.2025.100235

Mahdi Abdi-Khanghah , Fahimeh Hadavimoghaddam , Saeid Atashrouz , Elnaz Nasirzadeh , Meftah Ali Abuswer , Mehdi Ostadhassan , Ahmad Mohaddespour , Abdolhossein Hemmati-Sarapardeh

The viability of CO₂ capture projects, particularly through absorption with monoethanolamine (MEA) and other commercial absorbents, strongly depends on the CO₂ loading capacity. Therefore, comprehending the impact of variables on the CO₂ loading capacity of MEA is crucial in designing CO₂ capture units, which can be further optimized through multi-objective optimization. To this end, four machine learning models—Bagging Regression (BR), Categorical Boosting (CatBoost), Deep Belief Network (DBN), and Gaussian Process Regression with Rational Quadratic kernel function (GPR-RQ)—were utilized to predict the CO₂ loading capacity of MEA aqueous solutions. Temperature, partial pressure of CO₂, and MEA concentration were inputted into the intelligent network to calculate the CO₂ loading capacity. The binary values of R² and standard deviation (SD), which were 0.9889 and 0.0628 for Bagging Regression, 0.9932 and 0.06586 for CatBoost, 0.9957 and 0.0588 for GPR-RQ, and 0.9971 and 0.0329 for DBN, confirm that DBN has the highest accuracy in statistical analysis, followed by GPR-RQ, CatBoost, and Bagging Regression. Additionally, graphical methods like scattered plots and relative deviation plots corroborate the superior performance of the DBN model over all other intelligent techniques. By conducting a relevancy factor analysis on DBN outcomes, sensitivity analysis demonstrates that pressure has the most significant influence among the inputs. Furthermore, the Leverage technique affirms that the DBN model has a substantial degree of validity in forecasting the CO₂ loading capacity of MEA. Finally, 3-D image plots were systematically examined to analyze the binary interactive effect of (temperature, CO₂ partial pressure), (temperature, MEA concentration), and (CO₂ partial pressure, MEA concentration) on the carbon absorption efficiency, which is essential to reach the net-zero emission purpose.

二氧化碳捕获项目的可行性，特别是通过单乙醇胺和其他商业吸收剂的吸收，很大程度上取决于二氧化碳的装载能力。因此，在设计CO2捕集装置时，了解各变量对MEA CO2承载能力的影响至关重要，可通过多目标优化进一步优化。为此，采用bagging Regression （BR）、Categorical Boosting （CatBoost）、Deep Belief Network （DBN）和Gaussian Process Regression with有理二次核函数（GPR-RQ）四种机器学习模型对MEA水溶液的CO2承载能力进行了预测。将温度、CO2分压和MEA浓度输入到智能网络中，计算CO2的负荷能力。R2和标准差（SD）的二值分别为Bagging Regression的0.9889和0.0628，CatBoost的0.9932和0.06586，GPR-RQ的0.9957和0.0588，DBN的0.9971和0.0329，证实了DBN在统计分析中的准确性最高，其次是GPR-RQ， CatBoost和Bagging Regression。此外，散点图和相对偏差图等图形方法证实了DBN模型优于所有其他智能技术的性能。通过对DBN结果进行相关性因子分析，敏感性分析表明，在输入因素中，压力的影响最为显著。此外，杠杆技术证实了DBN模型在预测MEA的CO2负荷能力方面具有相当程度的有效性。最后，对三维图像图进行了系统分析，分析了（温度，CO2分压）、（温度，MEA浓度）和（CO2分压，MEA浓度）对碳吸收效率的二元交互作用，这是实现净零排放目标所必需的。

{"title":"Toward predicting CO2 loading capacity in monoethanolamine (MEA) aqueous solutions using deep belief network","authors":"Mahdi Abdi-Khanghah , Fahimeh Hadavimoghaddam , Saeid Atashrouz , Elnaz Nasirzadeh , Meftah Ali Abuswer , Mehdi Ostadhassan , Ahmad Mohaddespour , Abdolhossein Hemmati-Sarapardeh","doi":"10.1016/j.dche.2025.100235","DOIUrl":"10.1016/j.dche.2025.100235","url":null,"abstract":"<div><div>The viability of CO<sub>2</sub> capture projects, particularly through absorption with monoethanolamine (MEA) and other commercial absorbents, strongly depends on the CO<sub>2</sub> loading capacity. Therefore, comprehending the impact of variables on the CO<sub>2</sub> loading capacity of MEA is crucial in designing CO<sub>2</sub> capture units, which can be further optimized through multi-objective optimization. To this end, four machine learning models—Bagging Regression (BR), Categorical Boosting (CatBoost), Deep Belief Network (DBN), and Gaussian Process Regression with Rational Quadratic kernel function (GPR-RQ)—were utilized to predict the CO<sub>2</sub> loading capacity of MEA aqueous solutions. Temperature, partial pressure of CO<sub>2</sub>, and MEA concentration were inputted into the intelligent network to calculate the CO<sub>2</sub> loading capacity. The binary values of R<sup>2</sup> and standard deviation (SD), which were 0.9889 and 0.0628 for Bagging Regression, 0.9932 and 0.06586 for CatBoost, 0.9957 and 0.0588 for GPR-RQ, and 0.9971 and 0.0329 for DBN, confirm that DBN has the highest accuracy in statistical analysis, followed by GPR-RQ, CatBoost, and Bagging Regression. Additionally, graphical methods like scattered plots and relative deviation plots corroborate the superior performance of the DBN model over all other intelligent techniques. By conducting a relevancy factor analysis on DBN outcomes, sensitivity analysis demonstrates that pressure has the most significant influence among the inputs. Furthermore, the Leverage technique affirms that the DBN model has a substantial degree of validity in forecasting the CO<sub>2</sub> loading capacity of MEA. Finally, 3-D image plots were systematically examined to analyze the binary interactive effect of (temperature, CO<sub>2</sub> partial pressure), (temperature, MEA concentration), and (CO<sub>2</sub> partial pressure, MEA concentration) on the carbon absorption efficiency, which is essential to reach the net-zero emission purpose.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100235"},"PeriodicalIF":3.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143839775","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Enhancing cybersecurity of nonlinear processes via a two-layer control architecture 通过双层控制架构加强非线性过程的网络安全

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-04-01 DOI: 10.1016/j.dche.2025.100233

Arthur Khodaverdian , Dhruv Gohil , Panagiotis D. Christofides

This work proposes a novel two-layer multi-key control architecture to enhance the resilience of nonlinear chemical processes to cyberattacks. The architecture consists of an upper-layer nonlinear controller and a lower-layer of encrypted linear controllers. The nonlinear controllers process unencrypted sensor data to determine optimal control actions, which are then used to estimate the closed-loop state trajectory using a first-principle model of the plant. This trajectory is sampled and mapped to a valid subset before encryption, which can lead to minor inaccuracies. The resulting encrypted state-space data samples are used as set-points for the lower-layer controllers, which can be implemented using encrypted signals, allowing for obfuscation of the computation and transmission of the applied control inputs, thereby enhancing cybersecurity. This study further improves security by taking advantage of the Single-Input-Single-Output nature of some linear control methods to allocate a unique encryption key to each linear controller and its respective sensor data. Two nonlinear chemical process applications, including a benchmark chemical reactor example and one application modeled through the use of Aspen Dynamics, are used to demonstrate the application of the proposed two-layer architecture.

本文提出了一种新的两层多键控制体系结构，以增强非线性化学过程对网络攻击的弹性。该结构由上层非线性控制器和下层加密线性控制器组成。非线性控制器处理未加密的传感器数据以确定最优控制动作，然后使用植物的第一性原理模型来估计闭环状态轨迹。在加密之前对该轨迹进行采样并映射到一个有效子集，这可能导致较小的不准确性。由此产生的加密状态空间数据样本用作下层控制器的设定点，这可以使用加密信号实现，允许应用控制输入的计算和传输混淆，从而增强网络安全。本研究利用某些线性控制方法的单输入-单输出特性，为每个线性控制器及其各自的传感器数据分配唯一的加密密钥，进一步提高了安全性。两个非线性化学过程应用，包括一个基准化学反应器示例和一个通过使用Aspen Dynamics建模的应用，用于演示所提出的两层架构的应用。

{"title":"Enhancing cybersecurity of nonlinear processes via a two-layer control architecture","authors":"Arthur Khodaverdian , Dhruv Gohil , Panagiotis D. Christofides","doi":"10.1016/j.dche.2025.100233","DOIUrl":"10.1016/j.dche.2025.100233","url":null,"abstract":"<div><div>This work proposes a novel two-layer multi-key control architecture to enhance the resilience of nonlinear chemical processes to cyberattacks. The architecture consists of an upper-layer nonlinear controller and a lower-layer of encrypted linear controllers. The nonlinear controllers process unencrypted sensor data to determine optimal control actions, which are then used to estimate the closed-loop state trajectory using a first-principle model of the plant. This trajectory is sampled and mapped to a valid subset before encryption, which can lead to minor inaccuracies. The resulting encrypted state-space data samples are used as set-points for the lower-layer controllers, which can be implemented using encrypted signals, allowing for obfuscation of the computation and transmission of the applied control inputs, thereby enhancing cybersecurity. This study further improves security by taking advantage of the Single-Input-Single-Output nature of some linear control methods to allocate a unique encryption key to each linear controller and its respective sensor data. Two nonlinear chemical process applications, including a benchmark chemical reactor example and one application modeled through the use of Aspen Dynamics, are used to demonstrate the application of the proposed two-layer architecture.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100233"},"PeriodicalIF":3.0,"publicationDate":"2025-04-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143759018","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Green hydrogen extraction from natural gas transmission grids using hybrid membrane and PSA processes optimized via bayesian techniques 通过贝叶斯技术优化的混合膜和PSA工艺从天然气输电网中提取绿色氢气

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-31 DOI: 10.1016/j.dche.2025.100234

Homa Hamedi, Torsten Brinkmann

Green hydrogen (H₂) is a leading enabler for the decarbonization of hard-to-abate industries where electrification is either uneconomical or infeasible. Establishing an adequate and cost-effective infrastructure for hydrogen distribution remains one of the primary barriers to its widespread adoption. A promising short-term solution to this challenge involves H₂ storage and co-transportation via existing gas grids. For H₂ extraction from distribution gas grids, standalone pressure swing adsorption systems are considered the most viable option, whereas a hybrid process is suggested in the literature for transmission gas networks. This article presents a comprehensive techno-economic model for the proposed hybrid process, developed using an integrated platform based on Aspen Adsorption and Aspen Custom Modeler. The system consists of a single-stage hollow fiber Matrimid membrane module, followed by a 4-bed adsorption process operating in 8 sequential steps to meet H₂ market purity requirements with an acceptable recovery rate. Since the performances of these two separation modules, as an integrated system, significantly influence each other, the study identifies a unique opportunity to minimize separation costs through process optimization. To reduce computational time, a cyclic steady-state approach was employed to simulate the PSA process. Bayesian optimization, facilitated by the integration of Python with Aspen Adsorption, was used to efficiently identify the optimal solution with a minimal number of objective function evaluations. The levelized cost of H₂ separation (99.0 % purity at 10 bar) from natural gas containing 10 % H₂ at pressures of 35 bar and 60 bar is estimated to be 2.7310 and, $2.5116/kg-H₂, respectively. These estimates correspond to a scenario with 10 identical trains, each handling a feed flowrate of 200 kmol/hr. Increasing the number of trains keeps the cost contribution of PSA constant; however, the total cost decreases as the compression fixed cost is distributed across more trains.

绿色氢（H₂）是电气化不经济或不可行的难以减少的行业脱碳的主要推动者。建立一个足够的和具有成本效益的氢气分配基础设施仍然是其广泛采用的主要障碍之一。解决这一挑战的短期解决方案是通过现有的天然气网进行氢储存和联合运输。对于从配气网中提取H，独立变压吸附系统被认为是最可行的选择，而文献中建议在输气网络中采用混合过程。本文提出了一个综合的技术经济模型，该模型是利用基于杨木吸附和杨木定制建模器的集成平台开发的。该系统由单级中空纤维基质膜模块组成，然后是4层吸附工艺，分8个顺序步骤操作，以满足市场对h2纯度的要求，回收率可接受。由于这两个分离模块作为一个集成系统的性能会显著地相互影响，因此本研究确定了通过流程优化来最小化分离成本的独特机会。为了减少计算时间，采用循环稳态方法模拟PSA过程。利用Python和Aspen吸附相结合的贝叶斯优化方法，以最少的目标函数评价次数有效地识别出最优解。在35 bar和60 bar的压力下，从含有10% H2的天然气中分离H2（纯度为99.0%,10 bar）的平均成本估计分别为2.7310和2.5116美元/kg-H2。这些估计对应于10个相同列车的场景，每个列车处理200 kmol/hr的进料流量。增加列车数量保持PSA的成本贡献不变；然而，当压缩固定成本分布在更多的列车上时，总成本会降低。

{"title":"Green hydrogen extraction from natural gas transmission grids using hybrid membrane and PSA processes optimized via bayesian techniques","authors":"Homa Hamedi, Torsten Brinkmann","doi":"10.1016/j.dche.2025.100234","DOIUrl":"10.1016/j.dche.2025.100234","url":null,"abstract":"<div><div>Green hydrogen (H₂) is a leading enabler for the decarbonization of hard-to-abate industries where electrification is either uneconomical or infeasible. Establishing an adequate and cost-effective infrastructure for hydrogen distribution remains one of the primary barriers to its widespread adoption. A promising short-term solution to this challenge involves H₂ storage and co-transportation via existing gas grids. For H₂ extraction from distribution gas grids, standalone pressure swing adsorption systems are considered the most viable option, whereas a hybrid process is suggested in the literature for transmission gas networks. This article presents a comprehensive techno-economic model for the proposed hybrid process, developed using an integrated platform based on Aspen Adsorption and Aspen Custom Modeler. The system consists of a single-stage hollow fiber Matrimid membrane module, followed by a 4-bed adsorption process operating in 8 sequential steps to meet H₂ market purity requirements with an acceptable recovery rate. Since the performances of these two separation modules, as an integrated system, significantly influence each other, the study identifies a unique opportunity to minimize separation costs through process optimization. To reduce computational time, a cyclic steady-state approach was employed to simulate the PSA process. Bayesian optimization, facilitated by the integration of Python with Aspen Adsorption, was used to efficiently identify the optimal solution with a minimal number of objective function evaluations. The levelized cost of H₂ separation (99.0 % purity at 10 bar) from natural gas containing 10 % H<sub>2</sub> at pressures of 35 bar and 60 bar is estimated to be 2.7310 and, $2.5116/kg-H<sub>2</sub>, respectively. These estimates correspond to a scenario with 10 identical trains, each handling a feed flowrate of 200 kmol/hr. Increasing the number of trains keeps the cost contribution of PSA constant; however, the total cost decreases as the compression fixed cost is distributed across more trains.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100234"},"PeriodicalIF":3.0,"publicationDate":"2025-03-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143776826","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A tutorial review of policy iteration methods in reinforcement learning for nonlinear optimal control 非线性最优控制强化学习中的策略迭代方法教程综述

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-27 DOI: 10.1016/j.dche.2025.100231

Yujia Wang , Xinji Zhu , Zhe Wu

Reinforcement learning (RL) has been a powerful framework for designing optimal controllers for nonlinear systems. This tutorial review provides a comprehensive exploration of RL techniques, with a particular focus on policy iteration methods for the development of optimal controllers. We discuss key theoretical aspects, including closed-loop stability and convergence analysis of learning algorithms. Additionally, the review addresses practical challenges encountered in real-world applications, such as the development of accurate process models, incorporating safety guarantees during learning, leveraging physics-informed machine learning and transfer learning techniques to overcome learning difficulties, managing model uncertainties, and enabling scalability through distributed RL. To demonstrate the effectiveness of these approaches, a simulation example of a chemical reactor is presented, with open-source code made available on GitHub. The review concludes with a discussion of open research questions and future directions in RL-based control of nonlinear systems.

强化学习（RL）已成为设计非线性系统最优控制器的有力框架。本教程综述提供了对强化学习技术的全面探索，特别关注用于开发最优控制器的策略迭代方法。我们讨论了关键的理论方面，包括闭环稳定性和收敛分析的学习算法。此外，该综述还解决了在实际应用中遇到的实际挑战，例如开发准确的过程模型，在学习过程中结合安全保证，利用物理信息的机器学习和迁移学习技术来克服学习困难，管理模型不确定性，并通过分布式强化学习实现可扩展性。为了演示这些方法的有效性，本文给出了一个化学反应器的模拟示例，并在GitHub上提供了开源代码。最后，讨论了基于rl的非线性系统控制的开放性研究问题和未来发展方向。

引用次数: 0

Study on the Switching Model Predictive Control Algorithm in Batch Polymerization Process 间歇聚合过程中切换模型预测控制算法的研究

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-20 DOI: 10.1016/j.dche.2025.100232

Jong Nam Kim , Chun Bae Ma , Hyok Jo , Un Chol Han , Hyon-Tae Pak , Son Il Hong , Ri Myong Kim

In the batch polymerization process, temperature control is generally a challenging task. In this paper, a new switching model predictive control algorithm that can be effectively used for the temperature control of batch polymerization process is developed and its effectiveness is verified by introducing it to industrial batch polyvinyl chloride polymerization process. Firstly, a general analysis of the polymerization process is conducted, and based on this, the reaction starting point is determined. Secondly, a switching model identification method considering the reaction starting point and the reaction heat generated after the reaction starts is proposed. Finally, a switching model predictive control algorithm that determines the optimal manipulated value based on the on-line updated step response model is constructed, and a cascade control system using this algorithm is introduced to the temperature control of batch polyvinyl chloride suspension polymerization process. The results show that the proposed control system can significantly improve temperature control performance (overshoot: 0.2%, root mean square error: 0.3) compared to before introduction (overshoot: 1.1%, root mean square error: 1.2ྟC) .

在间歇聚合过程中，温度控制通常是一个具有挑战性的任务。本文提出了一种可有效用于间歇聚合过程温度控制的切换模型预测控制算法，并将其应用于工业间歇聚氯乙烯聚合过程中，验证了该算法的有效性。首先对聚合过程进行总体分析，在此基础上确定反应起始点。其次，提出了考虑反应起始点和反应开始后产生的反应热的切换模型辨识方法。最后，构建了基于在线更新阶跃响应模型确定最优操纵值的切换模型预测控制算法，并将该算法引入到间歇聚氯乙烯悬浮聚合过程的温度控制中。结果表明，与引入前（超调量：1.1%，均方根误差：1.2）相比，该控制系统能显著提高温度控制性能（超调量：0.2%，均方根误差：0.3）。

{"title":"Study on the Switching Model Predictive Control Algorithm in Batch Polymerization Process","authors":"Jong Nam Kim , Chun Bae Ma , Hyok Jo , Un Chol Han , Hyon-Tae Pak , Son Il Hong , Ri Myong Kim","doi":"10.1016/j.dche.2025.100232","DOIUrl":"10.1016/j.dche.2025.100232","url":null,"abstract":"<div><div>In the batch polymerization process, temperature control is generally a challenging task. In this paper, a new switching model predictive control algorithm that can be effectively used for the temperature control of batch polymerization process is developed and its effectiveness is verified by introducing it to industrial batch polyvinyl chloride polymerization process. Firstly, a general analysis of the polymerization process is conducted, and based on this, the reaction starting point is determined. Secondly, a switching model identification method considering the reaction starting point and the reaction heat generated after the reaction starts is proposed. Finally, a switching model predictive control algorithm that determines the optimal manipulated value based on the on-line updated step response model is constructed, and a cascade control system using this algorithm is introduced to the temperature control of batch polyvinyl chloride suspension polymerization process. The results show that the proposed control system can significantly improve temperature control performance (overshoot: 0.2%, root mean square error: 0.3) compared to before introduction (overshoot: 1.1%, root mean square error: 1.2ྟC) .</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100232"},"PeriodicalIF":3.0,"publicationDate":"2025-03-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143738419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Real-time process safety and systems decision-making toward safe and smart chemical manufacturing 实时过程安全和系统决策，实现安全和智能化工制造

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-12 DOI: 10.1016/j.dche.2025.100227

Austin Braniff , Sahithi Srijana Akundi , Yuanxing Liu , Beatriz Dantas , Shayan S. Niknezhad , Faisal Khan , Efstratios N. Pistikopoulos , Yuhe Tian

The ongoing digital transformation has created new opportunities for chemical manufacturing with increasing plant interconnectivity and data accessibility. This paper reviews state-of-the-art research developments which offer the potential for real-time process safety and systems decision-making in the digital era. An overview is first presented on online process safety management approaches, including dynamic risk analysis and fault diagnosis/prognosis. Advanced operability and control methods are then discussed to achieve safely optimal operations under uncertainty (e.g., flexibility analysis, safety-aware control, fault-tolerant control). We highlight the connections between systems-based operation and process safety management to achieve operational excellence while proactively reducing potential safety losses. We also review the developments and showcases of digital twins paving the way to actual cyber–physical integration. Outstanding challenges and opportunities are identified such as safe data-driven control, integrated operability, safety and control, cyber–physical demonstration, etc. Toward this direction, we present our ongoing developments of the REal-Time Risk-based Optimization (RETRO) framework for safe and smart process operations.

随着工厂互联性和数据可访问性的提高，正在进行的数字化转型为化工制造业创造了新的机遇。本文回顾了最新的研究进展，为数字时代的实时过程安全和系统决策提供了潜力。首先概述了在线过程安全管理方法，包括动态风险分析和故障诊断/预测。然后讨论了在不确定条件下实现安全最优运行的先进可操作性和控制方法（如柔性分析、安全感知控制、容错控制）。我们强调以系统为基础的操作和过程安全管理之间的联系，以实现卓越的操作，同时主动减少潜在的安全损失。我们还回顾了数字孪生的发展和展示，为实际的网络物理集成铺平了道路。指出了安全数据驱动控制、综合可操作性、安全和控制、网络物理演示等方面的突出挑战和机遇。朝着这个方向，我们展示了我们正在开发的基于实时风险的优化（RETRO）框架，用于安全和智能的过程操作。

{"title":"Real-time process safety and systems decision-making toward safe and smart chemical manufacturing","authors":"Austin Braniff , Sahithi Srijana Akundi , Yuanxing Liu , Beatriz Dantas , Shayan S. Niknezhad , Faisal Khan , Efstratios N. Pistikopoulos , Yuhe Tian","doi":"10.1016/j.dche.2025.100227","DOIUrl":"10.1016/j.dche.2025.100227","url":null,"abstract":"<div><div>The ongoing digital transformation has created new opportunities for chemical manufacturing with increasing plant interconnectivity and data accessibility. This paper reviews state-of-the-art research developments which offer the potential for real-time process safety and systems decision-making in the digital era. An overview is first presented on online process safety management approaches, including dynamic risk analysis and fault diagnosis/prognosis. Advanced operability and control methods are then discussed to achieve safely optimal operations under uncertainty (e.g., flexibility analysis, safety-aware control, fault-tolerant control). We highlight the connections between systems-based operation and process safety management to achieve operational excellence while proactively reducing potential safety losses. We also review the developments and showcases of digital twins paving the way to actual cyber–physical integration. Outstanding challenges and opportunities are identified such as safe data-driven control, integrated operability, safety and control, cyber–physical demonstration, etc. Toward this direction, we present our ongoing developments of the REal-Time Risk-based Optimization (RETRO) framework for safe and smart process operations.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100227"},"PeriodicalIF":3.0,"publicationDate":"2025-03-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143628822","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Surrogate-based flowsheet model maintenance for Digital Twins 基于代用流程图的数字孪生模型维护

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-12 DOI: 10.1016/j.dche.2025.100228

Balázs Palotai , Gábor Kis , János Abonyi , Ágnes Bárkányi

Digital Twins (DTs) are transforming industrial processes by providing virtual models that mirror physical systems, enabling real-time monitoring and optimization. A major challenge in DTs in process industry, is maintaining the accuracy of flowsheet simulation models due to changes like equipment degradation and operational shifts. This study proposes a novel surrogate-based approach for the automated calibration of these models, which reduces reliance on manual adjustments and adapts to changes in the physical system. This study leverages surrogate models and particle swarm optimization to incorporate modeling considerations and measurement uncertainties, thereby automating model calibration and reducing manual interventions. In a refinery case study, our approach reduced calibration time for the sour water stripper Hysys model by 80% while maintaining the desired accuracy. These results highlight the method’s potential to enhance flowsheet model accuracy in digital twin systems and to support more robust and adaptable DT applications.

数字孪生（dt）通过提供反映物理系统的虚拟模型，实现实时监控和优化，正在改变工业流程。在过程工业中，DTs面临的一个主要挑战是，由于设备退化和操作转换等变化，保持流程图仿真模型的准确性。本研究提出了一种新的基于代理的模型自动校准方法，减少了对人工调整的依赖，并适应了物理系统的变化。本研究利用替代模型和粒子群优化来结合建模考虑和测量不确定性，从而自动化模型校准并减少人工干预。在一个炼油厂的案例研究中，我们的方法在保持预期精度的同时，将酸水提提器Hysys模型的校准时间减少了80%。这些结果突出了该方法在提高数字孪生系统中的流程模型精度以及支持更健壮和适应性更强的DT应用方面的潜力。

引用次数: 0

Assessment of forward and forward–backward Bayesian filters 前向和后向贝叶斯滤波器的评估

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-10 DOI: 10.1016/j.dche.2025.100224

Daniel Martins Silva, Argimiro Resende Secchi

This paper investigates a forward–backward filtering approach comprised of forward filters and backward smoothers assimilating estimations of a moving horizon estimation. Those evaluations were carried out for extended, unscented, and cubature combinations of the Kalman filters, besides a particle filter, an ensemble Kalman filter, and a moving horizon estimation. Three simulation scenarios were defined for two nonlinear case studies with different complexity to evaluate the estimation accuracy and computational time under different uncertainty conditions. The backward smoothing was found to degenerate for longer horizons; however, it improved the estimation accuracy with smaller horizons in most simulation scenarios in comparison to the respective filters alone. In addition, the method successfully reduced steady-state estimation bias under model mismatch with a small increase in computational time. The performance of the forward–backward filtering was found to be sensitive to active constraint; however, this drawback does not outweigh the meaningful performance improvements found in this study.

研究了一种由前向滤波和后向平滑同化估计组成的前向后向滤波方法。除了粒子滤波器、集合卡尔曼滤波器和移动视界估计之外，还对卡尔曼滤波器的扩展、无气味和培养组合进行了评估。针对两种不同复杂程度的非线性案例，定义了三种模拟情景，以评估不同不确定性条件下的估计精度和计算时间。当视界较长时，后向平滑会退化；然而，在大多数模拟场景中，与单独使用各自的滤波器相比，它提高了较小视界的估计精度。此外，该方法成功地减少了模型失配下的稳态估计偏差，且计算时间增加较少。发现前向后向滤波的性能对主动约束很敏感；然而，这个缺点并没有超过本研究中发现的有意义的性能改进。

引用次数: 0

Polynomial Neural Networks for improved AI transparency: An analysis of their inherent explainability (operational rationale) capabilities 提高人工智能透明度的多项式神经网络：对其内在可解释性（操作原理）能力的分析

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-10 DOI: 10.1016/j.dche.2025.100230

Donovan Chaffart , Yue Yuan

The demand for reliable Artificial Intelligence (AI) models within critical domains such as Chemical Engineering has garnered significant attention towards the use and development of transparent AI methodologies. Nevertheless, the field of AI transparency has received an uneven level of attention, such that crucial aspects like explainability (i.e., the transparency of the AI's operational rationales) have remained understudied. To address this challenge, this study investigates the inherent explainability capabilities of Polynomial Neural Networks (PNNs) for applications within Chemical Engineering. PNNs, which implement higher-order polynomials in lieu of linear expressions within their hidden layer neurons, are inherently nonlinear, and thus do not require an activation function to accurately capture the behavior of a system. Accordingly, these neural networks provide continuous, closed-form algebraic expressions that can be used to ascertain the contributions of individual features in the AI architecture towards the network operational behavior. In order to study this behavior, the PNN method was adopted in this work to capture the relationships of noiseless and noisy data derived according to simple mathematical expressions. The PNN polynomials were then extracted and examined to highlight the insights they provide regarding the system operational rationales. The PNN method was furthermore applied to capture the behavior of a circulating fluidized bed reactor to fully showcase the explainative capability of this method within a Chemical Engineering application. These studies highlight the intrinsic explainability capabilities of PNNs and demonstrated their potential for reliable AI implementations for applications in Chemical Engineering.

在化学工程等关键领域，对可靠的人工智能（AI）模型的需求已经引起了人们对透明人工智能方法的使用和开发的极大关注。然而，人工智能透明度领域受到的关注程度参差不齐，诸如可解释性（即人工智能操作原理的透明度）等关键方面仍未得到充分研究。为了解决这一挑战，本研究探讨了多项式神经网络（PNNs）在化学工程应用中的固有可解释性能力。pnn在其隐藏层神经元中实现高阶多项式代替线性表达式，其本质上是非线性的，因此不需要激活函数来准确捕获系统的行为。因此，这些神经网络提供连续的、封闭形式的代数表达式，可用于确定人工智能架构中各个特征对网络运行行为的贡献。为了研究这种行为，本文采用PNN方法捕捉由简单数学表达式推导出的无噪声和有噪声数据之间的关系。然后提取并检查PNN多项式，以突出它们提供的关于系统运行原理的见解。PNN方法进一步应用于捕获循环流化床反应器的行为，以充分展示该方法在化学工程应用中的解释能力。这些研究强调了pnn内在的可解释性能力，并展示了它们在化学工程应用中可靠的人工智能实现的潜力。

{"title":"Polynomial Neural Networks for improved AI transparency: An analysis of their inherent explainability (operational rationale) capabilities","authors":"Donovan Chaffart , Yue Yuan","doi":"10.1016/j.dche.2025.100230","DOIUrl":"10.1016/j.dche.2025.100230","url":null,"abstract":"<div><div>The demand for reliable Artificial Intelligence (AI) models within critical domains such as Chemical Engineering has garnered significant attention towards the use and development of transparent AI methodologies. Nevertheless, the field of AI transparency has received an uneven level of attention, such that crucial aspects like <em>explainability</em> (i.e., the transparency of the AI's operational rationales) have remained understudied. To address this challenge, this study investigates the inherent <em>explainability</em> capabilities of Polynomial Neural Networks (PNNs) for applications within Chemical Engineering. PNNs, which implement higher-order polynomials in lieu of linear expressions within their hidden layer neurons, are inherently nonlinear, and thus do not require an activation function to accurately capture the behavior of a system. Accordingly, these neural networks provide continuous, closed-form algebraic expressions that can be used to ascertain the contributions of individual features in the AI architecture towards the network operational behavior. In order to study this behavior, the PNN method was adopted in this work to capture the relationships of noiseless and noisy data derived according to simple mathematical expressions. The PNN polynomials were then extracted and examined to highlight the insights they provide regarding the system operational rationales. The PNN method was furthermore applied to capture the behavior of a circulating fluidized bed reactor to fully showcase the <em>explainative</em> capability of this method within a Chemical Engineering application. These studies highlight the intrinsic <em>explainability</em> capabilities of PNNs and demonstrated their potential for reliable AI implementations for applications in Chemical Engineering.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"15 ","pages":"Article 100230"},"PeriodicalIF":3.0,"publicationDate":"2025-03-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143684687","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Operability for process flowsheet analysis 工艺流程分析的可操作性

IF 3 Q2 ENGINEERING, CHEMICAL

Digital Chemical Engineering

Pub Date : 2025-03-06 DOI: 10.1016/j.dche.2025.100229

Ulysses Guilherme Ferreira , Sérgio Mauro da Silva Neiro , Luís Cláudio Oliveira-Lopes , Thiago Vaz da Costa , Heleno Bispo , Fernando Vines Lima

Operability establishes the relationship between available input and achievable output sets through a system's mathematical representation. This work aims to develop a Flowsheet Operability analysis for a chemical process using rigorous models in a process simulator. The analysis focuses on a typical Air Separation Unit (ASU) in UniSim® Design (Honeywell) and integrates the simulator with the open-source Python operability tool (Opyrability) developed at West Virginia University. The performed assessment incrementally adds the output space of the process flowsheet units and examines how one group of units output space affects downstream units. The results underscore the importance of Flowsheet Operability analysis and the inclusion of inter-unit operability spaces for efficiently identifying unfavorable operating conditions that traditional Plantwide Operability analysis might overlook.

可操作性通过系统的数学表示建立了可用输入集和可实现输出集之间的关系。本工作旨在利用过程模拟器中的严格模型为化学过程开发流程图可操作性分析。分析的重点是UniSim®Design（霍尼韦尔）中典型的空气分离单元（ASU），并将模拟器与西弗吉尼亚大学开发的开源Python可操作工具（Opyrability）集成在一起。执行的评估以增量方式添加流程流程图单元的输出空间，并检查一组单元输出空间如何影响下游单元。研究结果强调了流程可操作性分析和单元间可操作性空间的重要性，以有效识别传统的全厂可操作性分析可能忽略的不利操作条件。

引用次数: 0