文献类型：期刊文献

英文题名：VAE-Based Interpretable Latent Variable Model for Process Monitoring

作者：Pan, Zhuofu[1,2,3];Wang, Yalin[3];Cao, Yue[4];Gui, Weihua[3]

机构：[1]Hunan Univ Technol & Business, Xiangjiang Lab, Changsha 410205, Peoples R China;[2]Hunan Univ Technol & Business, Sch Microelect & Phys, Changsha 410205, Peoples R China;[3]Cent South Univ, Sch Automat, Changsha 410083, Peoples R China;[4]East China Univ Sci & Technol, Minist Educ, Key Lab Smart Mfg Energy Chem Proc, Shanghai 200237, Peoples R China

年份：2024

卷号：35

期号：5

起止页码：6075

外文期刊名：IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS

收录：;EI(收录号：20232614288413);WOS:【SCI-EXPANDED(收录号:WOS:001013601800001)】；

基金：This work was supported in part by the National Natural Science Foundation of China under Grant 92267205 and Grant U1911401, in part by the Science and Technology Innovation Program of Hunan Province under Grant 2021RC4054, in part by the Changsha Social Laboratory of Artificial Intelligence, in part by the Open Project of Xiangjiang Laboratory under Grant 22XJ03019, and in part by the Scientific Research Fund of Hunan Provincial Education Department under Grant 22A0459.

语种：英文

外文关键词：Fault detection; Analytical models; Principal component analysis; Mathematical models; Taylor series; Process monitoring; Kernel; Activation function selection; interpretable structural design; Index Terms; process monitoring (PM); Taylor expansion; variational autoencoder (VAE)-based latent variable model (LVM)

摘要：Latent variable-based process monitoring (PM) models have been generously developed by shallow learning approaches, such as multivariate statistical analysis and kernel techniques. Owing to their explicit projection objectives, the extracted latent variables are usually meaningful and easily interpretable in mathematical terms. Recently, deep learning (DL) has been introduced to PM and has exhibited excellent performance because of its powerful presentation capability. However, its complex nonlinearity prevents it from being interpreted as human-friendly. It is a mystery how to design a proper network structure to achieve satisfactory PM performance for DL-based latent variable models (LVMs). In this article, a variational autoencoder-based interpretable LVM (VAE-ILVM) is developed for PM. Based on Taylor expansions, two propositions are proposed to guide the design of appropriate activation functions for VAE-ILVM, allowing nondisappearing fault impact terms contained in the generated monitoring metrics (MMs). During threshold learning, the sequence of counting that test statistics exceed the threshold is considered a martingale, a representative of weakly dependent stochastic processes. A de la Pena inequality is then adopted to learn a suitable threshold. Finally, two chemical examples verify the effectiveness of the proposed method. The use of de la Pena inequality significantly reduces the minimum required sample size for modeling.