Fault detection and diagnosis for operational control systems

Abstract

Operational control for complex industrial processes consists of two layers, namely the loop control layer and the operational layer. The former realizes the required loop control for each production unit and the later determines the best set-points to the control loops in the loop control layer. This paper formulates a general framework for the fault detection and diagnosis for such operational control systems, where it is assumed that the faults only take place in the loop control layers. For this purpose, a novel model is established for the representation of the fault detection and diagnosis. This is then followed by the development of fault detection and diagnosis algorithm, where fault diagnosis is realized using an adaptive fault diagnostic observer. A Lyapunov function is used to formulate the fault diagnosis algorithms where the system model uncertainties are also considered. Simulation has been carried out to show the effectiveness of the proposed method.