Active fault diagnosis and tracking control for stochastic distribution control systems with bounded disturbance

Abstract

A new active fault diagnosis (AFD) method for stochastic distribution control (SDC) systems with bounded disturbance is proposed, and tracking control is performed using the information obtained from active fault diagnosis. Considering the actuator fault of the SDC system with non‐Gaussian bounded disturbance and noise, the zonotope is used to describe the set of weight under different fault types. When the weight of the SDC system falls within a specific zonotope, the fault information of the system can be obtained and the tracking control can be carried out in time. Therefore, the primary objective of the AFD is to separate the zonotopes corresponding to different fault types. In this paper, the zonotopic box is introduced, and the distance from the center of the zonotopic box to its vertex is taken as the separation boundary of two or more zonotopes. For single‐input systems, the paper provides an analytical solution for the auxiliary input. In the case of multi‐input systems, the auxiliary input can be determined by solving a convex optimization problem. Finally, the algorithm is applied to a SDC system, and the effectiveness of the method is verified.