A fault detection and diagnosis system based on input and output residual generation scheme for a CSTR benchmark process

Abstract

Aim of this study is to propose fault detection and diagnosis (FDD) algorithm based on input and output residuals that consider both sensor and actuator faults separately. The existing methods which have capability of fault diagnosis and its magnitude estimation suffer from great computational complexity, so they would not be suitable for the real-time applications. The proposed method in this paper has the advantage of simple structure and straightforward computations but at the cost of losing precision. The introduced approach incorporates an auxiliary-PI controller in a feedback configuration with an extended Kalman filter (EKF) algorithm to constitute an actuator input-output residual generator (AIORG) unit. Similarly, a sensor output residual generator (SORG) unit is realized with an EKF-based algorithm to cover for simultaneous sensor possible faults. The generated residuals are then fed to a FDD unit to extract diagnostic and fault estimation results using a threshold-based inference mechanism. A set of test scenarios is conducted to demonstrate the performance capabilities of the proposed FDD methodology in a simulated continuous stirred tank reactor (CSTR) benchmark against sensor and actuator faults.