Artificial intelligence‐based process control in chemical, biochemical, and biomedical engineering. (10th August 2021)

• Record Type: Journal Article
• Authors: Dutta, Debaprasad; Upreti, Simant R.
• Is Part Of: Canadian journal of chemical engineering. Volume 99:Number 11(2021)
• Page Start: 2467
• Page End: 2504
• Publication Date: 2021-08-10
• Subjects: artificial intelligence -- biochemical engineering -- biomedical engineering -- chemical engineering -- process control; Chemical engineering -- Periodicals; Technology -- Periodicals; 660.05
• DOI: 10.1002/cjce.24246
• Languages: English
• ISSNs: 0008-4034

Abstract

Abstract: In the last three decades, artificial intelligence (AI) has been increasingly and vigorously utilized for process control in chemical, biochemical, and biomedical engineering. These disciplines frequently involve fairly sophisticated processes under risks of operational upsets, and thus have an ever increasing demand of superior control strategies. As the deployment of AI in process control pushes the limits in this regard, the research advances, which are multitudinous and varied, need to be assimilated and organized to help promote further utilization and progress in the field. To that end, we examine more than 280 relevant research publications, and systematically collate the information. The AI‐based technologies are classified, and their over‐arching control paradigm is presented. Common AI‐based control technologies are then presented, which are based on expert systems, fuzzy logic, artificial neural networks, nature‐inspired algorithms, and hybrid approaches. Their working principles, types, and implementations are summarized along with advantages, limitations, and comparisons, if available. Important applications in the above disciplines are included with the help of tables that capture important details. A discussion is also provided on advanced and newly emerging AI‐based control technologies with pertinent applications. Overall trends are analyzed, and future prospects are identified based on the survey.