文献类型：期刊文献

英文题名：Representation evaluation block-based teacher-student network for the industrial quality-relevant performance modeling and monitoring

作者：Yang, Dan[1]; Peng, Xin[1]; Lu, Yusheng[1]; Huang, Haojie[1]; Zhong, Weimin[1,2]

机构：[1] Key Laboratory of Advanced Control and Optimization for Chemical Processes, Ministry of Education, East China University of Science and Technology, Shanghai, 200237, China; [2] Shanghai Institute of Intelligent Science and Technology, Tongji University, Shanghai, 200092, China

年份：2021

外文期刊名：arXiv

收录：EI(收录号：20210044875)

语种：英文

外文关键词：Fault detection - Industrial water treatment - Learning systems - Process control - Process monitoring - Students - Uncertainty analysis

摘要：Quality-relevant fault detection plays an important role in industrial processes, while the current quality-related fault detection methods based on neural networks main concentrate on process-relevant variables and ignore quality-relevant variables, which restrict the application of process monitoring. Therefore, in this paper, a fault detection scheme based on the improved teacher-student network is proposed for quality-relevant fault detection. In the traditional teacher-student network, as the features differences between the teacher network and the student network will cause performance degradation on the student network, representation evaluation block (REB) is proposed to quantify the features differences between the teacher and the student networks, and uncertainty modeling is used to add this difference in modeling process, which are beneficial to reduce the features differences and improve the performance of the student network. Accordingly, REB and uncertainty modeling is applied in the teacher-student network named as uncertainty modeling teacher-student uncertainty autoencoder (TSUAE). Then, the proposed TSUAE is applied to process monitoring, which can effectively detect faults in the process-relevant subspace and quality-relevant subspace simultaneously. The proposed TSUAE-based fault detection method is verified in two simulation experiments illustrating that it has satisfactory fault detection performance compared to other fault detection methods. Copyright ? 2021, The Authors. All rights reserved.