Performance Assessment for Process Monitoring and Fault Detection Methods

The objective of Kai Zhang and his research is to assess the existing process monitoring and fault detection (PM-FD) methods. His aim is to provide suggestions and guidance for choosing appropriate PM-FD methods, because the performance assessment study for PM-FD methods has become an area of interest in both academics and industry. The author first compares basic FD statistics, and then assesses different PM-FD methods to monitor the key performance indicators of static processes, steady-state dynamic processes and general dynamic processes including transient states. He validates the theoretical developments using both benchmark and real industrial processes.

Assessing the performance of T2 and Q fault detection statistics.- Proposing a new performance evaluation index called expected detection delay (EDD).- Assessing the performance of different PM-FD methods using EDD when applied to detecting different types of faults.- Assessing the state-space-based PM-FD methods when applied to a real hot strip mill process.

Kai Zhang has just finished his PhD defense. His research area covers multivariate statistical process monitoring (PM) methods, data-driven fault detection (FD) methods and performance evaluation for PM-FD methods.