Enhanced moving horizon Bayesian-based fault diagnosis for multisampling rate data in a plantwide process

A plantwide process typically involves multisampling rates. However, traditional fault diagnosis methods usually assume that all monitored variables are obtained simultaneously. Therefore, an enhanced moving horizon Bayesian-based fault diagnosis for a plantwide process with multisampling rate data is proposed in this study. The focus of this method is to estimate the possible realization for the missing part in multisampling rate system and to calculate the corresponding possible fault reason, by using the likelihood probability obtained from the historical incomplete and complete dataset as well as the online moving horizon. Innovation of this study is that the method proposed considers the process state changes and the occurrence of incipient faults. The experiments demonstrate that the proposed method achieves considerable improvement. (C) 2020 Published by Elsevier Ltd.