干旱区生态环境知识资源中心(Arid): Dynamic risk-based fault prediction of chemical processes using online sparse model identification

Arid

	Dynamic risk-based fault prediction of chemical processes using online sparse model identification
	Bhadriraju, Bhavana; Kwon, Joseph Sang-Il; Khan, Faisal
通讯作者	Bhadriraju, B (corresponding author), Texas A&M Univ, Artie McFerrin Dept Chem Engn, College Stn, TX 77845 USA.
会议名称	American Control Conference (ACC)
会议日期	MAY 25-28, 2021
会议地点	ELECTR NETWORK
英文摘要	Fault prognosis has emerged as an essential monitoring technique that predicts the occurrence of a fault and provides a guide for taking an appropriate action before the fault propagates into a failure. This is especially useful for industrial systems with growing process complexities. Recently, data-based prognosis methods have become popular because of their ease of implementation and availability of abundant data. However, at times when the process conditions vary, the offline trained models are not sufficient to predict the changing dynamics. In such situations, it is useful to have an adaptive model that can readily accommodate the dynamical changes. Hence, we propose a systematic approach that combines operable adaptive sparse identification of systems (OASIS) and dynamic risk assessment for fault prediction. At every sampling instance, the OASIS algorithm is used to adaptively predict the future dynamics for a certain period of time-steps. This multi-step prediction approach is utilized to estimate the process risk associated with a fault in an online manner. A fault is expected to occur if the maximum risk value from the prediction window exceeds a pre-determined threshold. For demonstration, we apply the proposed method to predict the occurrence of a process fault in a chemical reactor earlier than it actually affects the process.
来源出版物	2021 AMERICAN CONTROL CONFERENCE (ACC)
ISSN	0743-1619
EISSN	2378-5861
出版年	2021
ISBN	978-1-6654-4197-1
出版者	IEEE
类型	Proceedings Paper
语种	英语
收录类别	CPCI-S
WOS记录号	WOS:000702263305009
WOS关键词	PROGNOSIS ; SYSTEMS
WOS类目	Automation & Control Systems ; Engineering, Electrical & Electronic
WOS研究方向	Automation & Control Systems ; Engineering
资源类型	会议论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/365608
作者单位	[Bhadriraju, Bhavana; Kwon, Joseph Sang-Il; Khan, Faisal] Texas A&M Univ, Artie McFerrin Dept Chem Engn, College Stn, TX 77845 USA
推荐引用方式 GB/T 7714	Bhadriraju, Bhavana,Kwon, Joseph Sang-Il,Khan, Faisal. Dynamic risk-based fault prediction of chemical processes using online sparse model identification[C]:IEEE,2021.