A hierarchical data reconciliation based on multiple time-delay interval estimation for industrial processes. (October 2020)

• Record Type: Journal Article
• Authors: Xie, Sen; Wang, Huaizhi; Peng, Jianchun; Liu, Xiaoli; Yuan, Xiaofeng
• Is Part Of: ISA transactions. Volume 105(2020)
• Page Start: 198
• Page End: 209
• Publication Date: 2020-10
• Subjects: Data reconciliation -- Long-running industrial process -- Multiple time-delay interval estimation -- Improved discrete state transition algorithm -- Hierarchical modeling; Engineering instruments -- Periodicals; Engineering instruments; Periodicals; Electronic journals; 629.805
• DOI: 10.1016/j.isatra.2020.06.001
• Languages: English
• ISSNs: 0019-0578

Abstract

Abstract: With the increasing demand for energy conversation and high efficiency, data quality is of great important to the operation management and monitoring in industrial applications. Data reconciliation, as a data processing technology, provides great potential to improve quality of process data, and is widely used to reduce measurement error and estimate unmeasured parameters. However, there are reactors connected in series in the long-running industrial processes so that liquid material information is difficult to mark and trace, and the liquid material has different residence times in each reactor due to the differences in the internal structure and operation mode. The time-delay in different reactors may be various and time-varying. In this paper, to solve these problems, a multiple time-delay interval estimation based hierarchical data reconciliation method is put forward. First, the multiple time-delay interval estimation is developed according to the process mechanism analysis and modeling. Then, an improved discrete state transition solution approach is presented to solve the data time-matching with multiple time-delay interval estimation for different reactors. Finally, a hierarchical data reconciliation frame is built by data characteristics. The feasible of the proposed data reconciliation method is verified utilizing the industrial application results. Graphical abstract: Highlights: A hierarchical DR is proposed based on multiple time-delay interval estimation. Multiple time-delay interval estimation is achieved by process mechanism analysis. An improved discrete STA is beneficial for solving data time-matching. The performance of developed method is validated on long-running industrial application.