Soft sensor design and fault detection using Bayesian network and probabilistic principal component analysis

Abstract

Bayesian network (BN) and probabilistic principal component analysis (PPCA) are powerful tools in artificial intelligence. In this study, two intelligent soft sensors, designed based on BN and PPCA, were proposed for fault detection and data prediction in chemical processes. A gas sweetening process was considered as a case study in which the existing data for H 2 S concentration in sweet gas stream were incomplete. In order to detect faults during the operation, a soft sensor was designed using BN and PPCA for predicting H 2 S concentration in the sweet gas stream. Then, another soft sensor was developed based on the BN for fault detection, considering the Gaussian mixture model with hidden nodes. The Tennessee-Eastman challenge process was used to assess the efficiency of the fault detection method. Results showed that the predicted values of the soft sensors are very close to real data. The fault occurring in this process was detected in early stages, demonstrating the good performance of the proposed fault detection system. It was also shown that the performance of the BN is better than the PPCA in prediction of incomplete data. Moreover, the confidence interval can be evaluated for the predicted values when using the BN. An uncertainty analysis was performed to assess the quality of predicted data, and it was observed that the error magnitude of predicted data is smaller when using the BN compared with the PPCA. In particular, the results showed that the BN model is capable of estimating H 2 S concentration with nearly 96.1% accuracy, whereas the accuracy of the PCA-based method was 93.2%. K E Y W O R D SBayesian network, fault detection, gas sweetening process, probabilistic principal component analysis, soft sensor, Tennessee-Eastman challenge process

Adaptive inference for Bayesian network soft‐sensor in the presence of process and sensor drift

“…[2][3][4] Recently, Bayesian networks (BNs), which are a family of probabilistic graphical models, have shown significant advantages in their application to soft-sensor development compared to conventional approaches. [1,5,6] Due to their ability to incorporate prior process knowledge and ease of handling uncertainty and missing data, BNs possess several advantages compared to conventional approaches. The BNs soft-sensors are typically identified using historical data, and the identified model is used for carrying out predictions in real-time.…”

Welcome to the second year of the Journal of Advanced Manufacturing and Processing

“…Cybersecurity and the digital transformation of informational and operational technologies in the manufacturing environment requires new ways of how to protect systems and processes from unintended and unauthorized consequences. This year, we expanded on the topic of data driven manufacturing, advanced informatics and machine learning which was introduced to the journal in the paper by Mohammadi et al [17] The intersection of cyber and physical systems for manufacturing will continue to be an important element of advanced manufacturing, and we hope to see further articles elaborating the important connections between them. We will also have a special issue on process intensification amplifying this theme.…”