Integration of nonlinear model predictive fault tolerant control and fault detection and diagnosis

This paper presents an integration of fault detection and diagnosis and control reconfiguration approach to form a comprehensive active fault tolerant control system for class of nonlinear systems with input constraints. The FDD is designed based on multiple model method. The bank of extended kalman filter is used to detect the predefined lock in place actuator fault and estimate the unknown parameter of actuator position. Delay on FDD decision may lead to instability. This is because of mismatch between the process and the model. The fault information is sent to the Fault Tolerant Control block in order to reconfigure the controller. The reconfigurable controller is designed base on stabilizing Nonlinear Model Predictive Control scheme. The speed of FDD on decision making process and the ability of the NMPC approach to guaranty stability are important factors in the fault tolerant control system. These factors are investigated on a chemical process. It is shown that this scheme can provide the system stability when a fault occurs.

Keywords:

fault detection and diagnosis , extended kalman filter , nonlinear model predictive control