Input Set Design for Active Fault Diagnosis and Control

Abstract

This paper first proposes a novel input set design method for set-based active fault diagnosis and then applies it to implement an integrated diagnosis and control scheme. Particularly, taking multiplicative actuator faults as an example, a separation tendency notion is used to characterize the separability of output sets of a group of healthy and faulty modes. At each time instant, an input set is designed for active fault diagnosis by increasing the separation tendency of output sets at the next time instant. Furthermore, an optimal input out of the input set is designed to minimize the output-tracking error. Based on this logic, integrated active fault diagnosis and control is finally achieved by designing input sets and optimal inputs step by step such that a certain separation tendency is reached for successful diagnosis. At the end, an example is used to illustrate the effectiveness of the proposed method.