Pub Date : 2025-02-15DOI: 10.1016/j.dche.2025.100221
Ali Nik-Khorasani , Thanh Tung Khuat , Bogdan Gabrys
This paper addresses the challenges of predicting bioprocess performance, particularly in monoclonal antibody (mAb) production, where conventional statistical methods often fall short due to time-series data’s complexity and high dimensionality. We propose a novel Hyperbox Mixture Regression (HMR) model that employs hyperbox-based input space partitioning to enhance predictive accuracy while managing uncertainty inherent in bioprocess data. The HMR model is designed to dynamically generate hyperboxes for input samples in a single-pass process, thereby improving learning speed and reducing computational complexity. Our experimental study utilizes a dataset that contains 106 bioreactors. This study evaluates the model’s performance in predicting critical quality attributes in monoclonal antibody manufacturing over a 15-day cultivation period. The results demonstrate that the HMR model outperforms comparable approximators in accuracy and learning speed and maintains interpretability and robustness under uncertain conditions. These findings underscore the potential of HMR as a powerful tool for enhancing predictive analytics in bioprocessing applications.
{"title":"Hyperbox Mixture Regression for process performance prediction in antibody production","authors":"Ali Nik-Khorasani , Thanh Tung Khuat , Bogdan Gabrys","doi":"10.1016/j.dche.2025.100221","DOIUrl":"10.1016/j.dche.2025.100221","url":null,"abstract":"<div><div>This paper addresses the challenges of predicting bioprocess performance, particularly in monoclonal antibody (mAb) production, where conventional statistical methods often fall short due to time-series data’s complexity and high dimensionality. We propose a novel Hyperbox Mixture Regression (HMR) model that employs hyperbox-based input space partitioning to enhance predictive accuracy while managing uncertainty inherent in bioprocess data. The HMR model is designed to dynamically generate hyperboxes for input samples in a single-pass process, thereby improving learning speed and reducing computational complexity. Our experimental study utilizes a dataset that contains 106 bioreactors. This study evaluates the model’s performance in predicting critical quality attributes in monoclonal antibody manufacturing over a 15-day cultivation period. The results demonstrate that the HMR model outperforms comparable approximators in accuracy and learning speed and maintains interpretability and robustness under uncertain conditions. These findings underscore the potential of HMR as a powerful tool for enhancing predictive analytics in bioprocessing applications.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100221"},"PeriodicalIF":3.0,"publicationDate":"2025-02-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143436459","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}
In the oil industry, gas lift is essential for facilitating fluid flow toward the production unit. However, the challenge lies in balancing gas availability constraints to achieve maximum efficiency in an oil field. This study utilizes the integrated production modeling (IPM) software to simulate an oil field operation in Iran. To this end, 154 data points constructed by a central composite design (CCD) experiment were utilized to develop neural network models. Therefore, four robust models, including multilayer perceptron (MLP), radial basis function (RBF), general regression neural network (GRNN), and cascade forward neural network (CFNN), were implemented for modeling. In addition, the net present value (NPV) serves as the objective function. To optimize the selected input variables, including tubing inside diameter, gas injection rate, and separator pressure, various optimization algorithms such as particle swarm optimization (PSO), ant colony optimization (ACO), genetic algorithm (GA), and a Novel optimization algorithm in a gas-lift study called grey wolf optimization (GWO), were utilized considering the constraint of the limited available gas. A penalty function was used to incorporate this constraint into the optimization procedure. There has previously been much research in the area of gas lift optimization. However, robust neural networks (GRNN and CFNN) have not been used for integrated production system modeling, nor have GWO algorithms been used to maximize the production or NPV in gas lift operations until now. The results for model errors were found to be %2.09, %2.99, %10.68, and %1.75 for MLP, RBF, GRNN, and CFNN, respectively. These findings imply that the CFNN model is more efficient. Also, comparing the GWO approach to other algorithms, the largest NPV ($788,512,038$) was yielded with less sensitivity of its adjustable parameters. Thereupon, NPV and cumulated oil production indicate a significant increase compared to ordinary NPV and oil production with values of 351,087,876.4 $ and 14,308 STB, respectively. High NPV effectively captures the overall added value of the project and, as a benchmark, helps to make informed decisions about investment and resource allocation, ultimately driving economic growth and increasing competitiveness in using this method.
{"title":"Integration of artificial intelligence and advanced optimization techniques for continuous gas lift under restricted gas supply: A case study","authors":"Leila Zeinolabedini , Forough Ameli , Abdolhossein Hemmati-Sarapardeh","doi":"10.1016/j.dche.2025.100220","DOIUrl":"10.1016/j.dche.2025.100220","url":null,"abstract":"<div><div>In the oil industry, gas lift is essential for facilitating fluid flow toward the production unit. However, the challenge lies in balancing gas availability constraints to achieve maximum efficiency in an oil field. This study utilizes the integrated production modeling (IPM) software to simulate an oil field operation in Iran. To this end, 154 data points constructed by a central composite design (CCD) experiment were utilized to develop neural network models. Therefore, four robust models, including multilayer perceptron (MLP), radial basis function (RBF), general regression neural network (GRNN), and cascade forward neural network (CFNN), were implemented for modeling. In addition, the net present value (NPV) serves as the objective function. To optimize the selected input variables, including tubing inside diameter, gas injection rate, and separator pressure, various optimization algorithms such as particle swarm optimization (PSO), ant colony optimization (ACO), genetic algorithm (GA), and a Novel optimization algorithm in a gas-lift study called grey wolf optimization (GWO), were utilized considering the constraint of the limited available gas. A penalty function was used to incorporate this constraint into the optimization procedure. There has previously been much research in the area of gas lift optimization. However, robust neural networks (GRNN and CFNN) have not been used for integrated production system modeling, nor have GWO algorithms been used to maximize the production or NPV in gas lift operations until now. The results for model errors were found to be %2.09, %2.99, %10.68, and %1.75 for MLP, RBF, GRNN, and CFNN, respectively. These findings imply that the CFNN model is more efficient. Also, comparing the GWO approach to other algorithms, the largest NPV ($788,512,038$) was yielded with less sensitivity of its adjustable parameters. Thereupon, NPV and cumulated oil production indicate a significant increase compared to ordinary NPV and oil production with values of 351,087,876.4 $ and 14,308 STB, respectively. High NPV effectively captures the overall added value of the project and, as a benchmark, helps to make informed decisions about investment and resource allocation, ultimately driving economic growth and increasing competitiveness in using this method.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100220"},"PeriodicalIF":3.0,"publicationDate":"2025-01-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143198410","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}
Pub Date : 2025-01-17DOI: 10.1016/j.dche.2025.100218
Hasan Nikkhah , Zahir Aghayev , Amir Shahbazi , Vassilis M. Charitopoulos , Styliani Avraamidou , Burcu Beykal
Planning and scheduling are crucial components of enterprise-wide optimization (EWO). For the successful execution of EWO, it is vital to view the enterprise operations as a holistic decision-making problem, composed of different interconnected elements or layers, to make the most efficient use of resources in process industries. Among different layers of the operating decisions, planning and scheduling are often treated sequentially, leading to impractical solutions. To tackle this problem, integrated approaches, such as bi-level programming are utilized to optimize these two layers simultaneously. Nonetheless, the bi-level optimization of such interdependent and holistic formulations is still difficult, particularly when dealing with mixed-integer nonlinear programming (MINLP) problems, due to a lack of effective algorithms. In this study, we employ the Data-driven Optimization of bi-level Mixed-Integer NOnlinear problems (DOMINO) framework, a data-driven algorithm developed to handle single-leader single-follower bi-level mixed-integer problems, to solve single-leader multi-follower planning and scheduling problems subject to MINLP scheduling formulations. We apply DOMINO to the continuous production of multi-product methyl methacrylate polymerization process formulated as a Traveling Salesman Problem and demonstrate its capability in achieving near-optimal guaranteed feasible solutions. Building on this foundation, we extend this strategy to solve a high-dimensional and highly constrained nonlinear crude oil refinery operation problem that has not been previously tackled in this context. Our study further evaluates the efficacy of using local, NOMAD (Nonlinear Optimization by Mesh Adaptive Direct Search), and a global data-driven optimizer, ARGONAUT (AlgoRithms for Global Optimization of coNstrAined grey-box compUTational), within the DOMINO framework and characterize their performance both in terms of solution quality and computational expense. The results indicate that DOMINO-NOMAD consistently achieves superior performance compared to DOMINO-ARGONAUT by identifying lower planning costs and generating more feasible solutions across multiple runs. Overall, this study demonstrates DOMINO’s ability to optimize production targets, meet market demands, and address large-scale EWO problems.
计划和调度是企业范围优化（EWO）的关键组成部分。为了成功地实施EWO，至关重要的是将企业运营视为一个整体决策问题，由不同的相互关联的要素或层次组成，以最有效地利用过程工业中的资源。在操作决策的不同层次中，计划和调度通常是顺序处理的，从而导致不切实际的解决方案。为了解决这一问题，利用双层编程等综合方法同时优化这两层。然而，由于缺乏有效的算法，这种相互依赖的整体公式的双层优化仍然很困难，特别是在处理混合整数非线性规划（MINLP）问题时。本文采用数据驱动优化双级混合整数非线性问题（Data-driven Optimization of bi-level Mixed-Integer NOnlinear problems， DOMINO）框架来解决单领导者单追随者双级混合整数问题，这是一种用于处理单领导者单追随者双级别混合整数问题的数据驱动算法。我们将多米诺应用于多产品甲基丙烯酸甲酯聚合过程的连续生产中，该过程被表述为一个旅行推销员问题，并证明了它在获得近最优保证可行解决方案方面的能力。在此基础上，我们将该策略扩展到解决高维、高约束的非线性原油炼油厂运行问题，这是以前在此背景下尚未解决的问题。我们的研究进一步评估了在DOMINO框架中使用局部NOMAD（通过网格自适应直接搜索的非线性优化）和全局数据驱动优化器ARGONAUT（约束灰盒计算的全局优化算法）的效果，并从解决方案质量和计算费用两方面描述了它们的性能。结果表明，与DOMINO-ARGONAUT相比，DOMINO-NOMAD通过确定更低的规划成本并在多次运行中生成更可行的解决方案，始终具有更优越的性能。总体而言，本研究证明了DOMINO优化生产目标、满足市场需求和解决大规模EWO问题的能力。
{"title":"Bi-level data-driven enterprise-wide optimization with mixed-integer nonlinear scheduling problems","authors":"Hasan Nikkhah , Zahir Aghayev , Amir Shahbazi , Vassilis M. Charitopoulos , Styliani Avraamidou , Burcu Beykal","doi":"10.1016/j.dche.2025.100218","DOIUrl":"10.1016/j.dche.2025.100218","url":null,"abstract":"<div><div>Planning and scheduling are crucial components of enterprise-wide optimization (EWO). For the successful execution of EWO, it is vital to view the enterprise operations as a holistic decision-making problem, composed of different interconnected elements or layers, to make the most efficient use of resources in process industries. Among different layers of the operating decisions, planning and scheduling are often treated sequentially, leading to impractical solutions. To tackle this problem, integrated approaches, such as bi-level programming are utilized to optimize these two layers simultaneously. Nonetheless, the bi-level optimization of such interdependent and holistic formulations is still difficult, particularly when dealing with mixed-integer nonlinear programming (MINLP) problems, due to a lack of effective algorithms. In this study, we employ the Data-driven Optimization of bi-level Mixed-Integer NOnlinear problems (DOMINO) framework, a data-driven algorithm developed to handle single-leader single-follower bi-level mixed-integer problems, to solve single-leader multi-follower planning and scheduling problems subject to MINLP scheduling formulations. We apply DOMINO to the continuous production of multi-product methyl methacrylate polymerization process formulated as a Traveling Salesman Problem and demonstrate its capability in achieving near-optimal guaranteed feasible solutions. Building on this foundation, we extend this strategy to solve a high-dimensional and highly constrained nonlinear crude oil refinery operation problem that has not been previously tackled in this context. Our study further evaluates the efficacy of using local, NOMAD (Nonlinear Optimization by Mesh Adaptive Direct Search), and a global data-driven optimizer, ARGONAUT (AlgoRithms for Global Optimization of coNstrAined grey-box compUTational), within the DOMINO framework and characterize their performance both in terms of solution quality and computational expense. The results indicate that DOMINO-NOMAD consistently achieves superior performance compared to DOMINO-ARGONAUT by identifying lower planning costs and generating more feasible solutions across multiple runs. Overall, this study demonstrates DOMINO’s ability to optimize production targets, meet market demands, and address large-scale EWO problems.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100218"},"PeriodicalIF":3.0,"publicationDate":"2025-01-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159674","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}
Pub Date : 2025-01-16DOI: 10.1016/j.dche.2025.100219
Xiaojie Li , Mingxue Yan , Xuewen Zhang , Minghao Han , Adrian Wing-Keung Law , Xunyuan Yin
Model predictive control (MPC) has become a key tool for optimizing real-time operations in industrial systems and processes, particularly to enhance performance, safety, and resilience. However, the growing complexity and nonlinearity of modern industrial systems present significant challenges for both first-principles modeling and real-time implementation of typical non-convex optimization associated with conventional MPC designs based on nonlinear models. In this review, we aim to provide an overview of current data-driven predictive control methods that have attributes of being computationally efficient as well as having the distinctive potential to address the above two challenges simultaneously. We focus particularly on two promising frameworks: (1) Koopman-based model predictive control, and (2) data-enabled predictive control, both of which are capable of formulating the optimization problem into a convex form even in the presence of strong nonlinearity in the underlying system. Additionally, we provide an outlook on the potential applications of these methods and briefly discuss their future directions across various industrial sectors.
{"title":"Efficient data-driven predictive control of nonlinear systems: A review and perspectives","authors":"Xiaojie Li , Mingxue Yan , Xuewen Zhang , Minghao Han , Adrian Wing-Keung Law , Xunyuan Yin","doi":"10.1016/j.dche.2025.100219","DOIUrl":"10.1016/j.dche.2025.100219","url":null,"abstract":"<div><div>Model predictive control (MPC) has become a key tool for optimizing real-time operations in industrial systems and processes, particularly to enhance performance, safety, and resilience. However, the growing complexity and nonlinearity of modern industrial systems present significant challenges for both first-principles modeling and real-time implementation of typical non-convex optimization associated with conventional MPC designs based on nonlinear models. In this review, we aim to provide an overview of current data-driven predictive control methods that have attributes of being computationally efficient as well as having the distinctive potential to address the above two challenges simultaneously. We focus particularly on two promising frameworks: (1) Koopman-based model predictive control, and (2) data-enabled predictive control, both of which are capable of formulating the optimization problem into a convex form even in the presence of strong nonlinearity in the underlying system. Additionally, we provide an outlook on the potential applications of these methods and briefly discuss their future directions across various industrial sectors.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100219"},"PeriodicalIF":3.0,"publicationDate":"2025-01-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159036","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}
Water pollution, worsened by rapid industrialization, poses severe challenges to global water management, particularly in developing countries like India. Conventional water quality monitoring methods, which rely on manual sampling and laboratory analysis are, inadequate for handling the dynamic and real-time nature of industrial wastewater contamination. To address this issue, this research article presents the state-of-the-art IoT-based autonomous real-time monitoring system (APAH), a scalable and frugal solution for industrial wastewater management. APAH integrates multi-parameter sensors to continuously monitor critical water quality parameters such as pH, dissolved oxygen (DO), electrical conductivity (EC), total dissolved solids (TDS), turbidity, and temperature. The system's layered architecture, comprising a sensing layer, edge layer, and application layer, enables data acquisition, processing, and remote access via APAH i.e. developed Android mobile application, respectively. APAH utilizes advanced technologies including, the Internet of Things (IoT) and Machine learning (ML) to provide real-time monitoring and control of wastewater treatment processes. Automated valve controls and real-time alerts enable timely intervention, preventing contamination and ensuring compliance with environmental standards. The system's performance was validated through field tests at four industrial wastewater treatment plants in Maharashtra, India particularly directed towards textile, dairy, and greywater effluents, demonstrating significant improvements in water quality post-treatment. The APAH system offers a promising solution for enhancing industrial wastewater treatment efficiency and ensuring sustainable water resource management. By integrating IoT technologies, real-time monitoring, and predictive analytics, APAH can contribute to addressing the urgent need for effective water quality management in industrial environments, particularly in regions facing acute water scarcity and pollution challenges.
{"title":"APAH: An autonomous IoT driven real-time monitoring system for Industrial wastewater","authors":"Nishant Chavhan , Resham Bhattad , Suyash Khot , Shubham Patil , Aditya Pawar , Tejasvi Pawar , Palomi Gawli","doi":"10.1016/j.dche.2025.100217","DOIUrl":"10.1016/j.dche.2025.100217","url":null,"abstract":"<div><div>Water pollution, worsened by rapid industrialization, poses severe challenges to global water management, particularly in developing countries like India. Conventional water quality monitoring methods, which rely on manual sampling and laboratory analysis are, inadequate for handling the dynamic and real-time nature of industrial wastewater contamination. To address this issue, this research article presents the state-of-the-art IoT-based autonomous real-time monitoring system (APAH), a scalable and frugal solution for industrial wastewater management. APAH integrates multi-parameter sensors to continuously monitor critical water quality parameters such as pH, dissolved oxygen (DO), electrical conductivity (EC), total dissolved solids (TDS), turbidity, and temperature. The system's layered architecture, comprising a sensing layer, edge layer, and application layer, enables data acquisition, processing, and remote access via APAH i.e. developed Android mobile application, respectively. APAH utilizes advanced technologies including, the Internet of Things (IoT) and Machine learning (ML) to provide real-time monitoring and control of wastewater treatment processes. Automated valve controls and real-time alerts enable timely intervention, preventing contamination and ensuring compliance with environmental standards. The system's performance was validated through field tests at four industrial wastewater treatment plants in Maharashtra, India particularly directed towards textile, dairy, and greywater effluents, demonstrating significant improvements in water quality post-treatment. The APAH system offers a promising solution for enhancing industrial wastewater treatment efficiency and ensuring sustainable water resource management. By integrating IoT technologies, real-time monitoring, and predictive analytics, APAH can contribute to addressing the urgent need for effective water quality management in industrial environments, particularly in regions facing acute water scarcity and pollution challenges.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100217"},"PeriodicalIF":3.0,"publicationDate":"2025-01-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159062","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}
Pub Date : 2025-01-02DOI: 10.1016/j.dche.2024.100216
Yiming Lang , Michelle Xin Yi Ng , Kai Xiang Yu , Binghui Chen , Peng Chee Tan , Khang Wei Tan , Weng Hoong Lam , Parthiban Siwayanan , Kek Seong Kim , Thomas Shean Yaw Choong , Joon Yoon Ten , Zhen Hong Ban
Illegal practices like open electronic waste incineration release hazardous pollutants, endangering the environment and human health. The Internet of Things (IoT) enables online real-time gas concentrations, but its capability to predict leak sources accurately remains a challenge. A large amount of historical data is required to train the source localization model, as gas dispersion is affected by wind speed and wind direction. Furthermore, sensor placement critically affects precise detection and prediction. This study introduces an innovative approach integrating Computational Fluid Dynamics (CFD), Mixed-Integer Linear Programming (MILP), and Artificial Neural Network modeling (ANN). CFD was utilized for machine learning model training. The MILP was used to optimize sensor placement, while the ANN model was used to optimize sensor number. The source localization model was again realized by the ANN model with optimized sensors data. The trained model was able to identify the unknown illegal electronic waste treatment locations with 97.22 % accuracy in this study. This method allows for the rapid detection of gas sources, as well as the execution of an emergency response, in line with Sustainable Development Goal Target 3.9.
{"title":"A novel CFD-MILP-ANN approach for optimizing sensor placement, number, and source localization in large-scale gas dispersion from unknown locations","authors":"Yiming Lang , Michelle Xin Yi Ng , Kai Xiang Yu , Binghui Chen , Peng Chee Tan , Khang Wei Tan , Weng Hoong Lam , Parthiban Siwayanan , Kek Seong Kim , Thomas Shean Yaw Choong , Joon Yoon Ten , Zhen Hong Ban","doi":"10.1016/j.dche.2024.100216","DOIUrl":"10.1016/j.dche.2024.100216","url":null,"abstract":"<div><div>Illegal practices like open electronic waste incineration release hazardous pollutants, endangering the environment and human health. The Internet of Things (IoT) enables online real-time gas concentrations, but its capability to predict leak sources accurately remains a challenge. A large amount of historical data is required to train the source localization model, as gas dispersion is affected by wind speed and wind direction. Furthermore, sensor placement critically affects precise detection and prediction. This study introduces an innovative approach integrating Computational Fluid Dynamics (CFD), Mixed-Integer Linear Programming (MILP), and Artificial Neural Network modeling (ANN). CFD was utilized for machine learning model training. The MILP was used to optimize sensor placement, while the ANN model was used to optimize sensor number. The source localization model was again realized by the ANN model with optimized sensors data. The trained model was able to identify the unknown illegal electronic waste treatment locations with 97.22 % accuracy in this study. This method allows for the rapid detection of gas sources, as well as the execution of an emergency response, in line with Sustainable Development Goal Target 3.9.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100216"},"PeriodicalIF":3.0,"publicationDate":"2025-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143160287","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}
Pub Date : 2025-01-02DOI: 10.1016/j.dche.2024.100211
Bastian Oetomo , Ling Luo , Yiran Qu , Michele Discepola , Sandra E. Kentish , Sally L. Gras
With the rapid growth of the biopharmaceutical sector in recent years, in conjunction with many recent successful developments in machine learning and artificial intelligence, the demand for the sector to shift to Industry 4.0 has emerged. Process Analytical Technology (PAT) makes it possible to monitor and control the manufacturing processes of monoclonal antibodies (mAbs), both in upstream and downstream processing. Despite downstream processing being responsible for approximately 60% of the cost of biological drug production, most of the recent developments focus on its upstream counterpart. This paper investigates existing literature on the application of machine learning and/or process control in downstream processing, with an emphasis on ultrafiltration/diafiltration (UF/DF) via tangential flow filtration (TFF). Literature on the intersection between control systems and machine learning will also be explored.
{"title":"Controlling tangential flow filtration in biomanufacturing processes via machine learning: A literature review","authors":"Bastian Oetomo , Ling Luo , Yiran Qu , Michele Discepola , Sandra E. Kentish , Sally L. Gras","doi":"10.1016/j.dche.2024.100211","DOIUrl":"10.1016/j.dche.2024.100211","url":null,"abstract":"<div><div>With the rapid growth of the biopharmaceutical sector in recent years, in conjunction with many recent successful developments in machine learning and artificial intelligence, the demand for the sector to shift to Industry 4.0 has emerged. Process Analytical Technology (PAT) makes it possible to monitor and control the manufacturing processes of monoclonal antibodies (mAbs), both in upstream and downstream processing. Despite downstream processing being responsible for approximately 60% of the cost of biological drug production, most of the recent developments focus on its upstream counterpart. This paper investigates existing literature on the application of machine learning and/or process control in downstream processing, with an emphasis on ultrafiltration/diafiltration (UF/DF) via tangential flow filtration (TFF). Literature on the intersection between control systems and machine learning will also be explored.</div></div>","PeriodicalId":72815,"journal":{"name":"Digital Chemical Engineering","volume":"14 ","pages":"Article 100211"},"PeriodicalIF":3.0,"publicationDate":"2025-01-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143159037","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}
扫码关注我们
求助内容：
应助结果提醒方式：
京公网安备 11010802042870号