Health Indices Construction for Stochastically Deteriorating Feedback Control Systems

Health index is a key input for the predictive maintenance or control reconfiguration of systems. This paper deals with the degradation and failure of a feedback control system using proportional-integral-derivative controller. Actually, the controller allows system fault tolerance to fulfil certain required missions, but also hides the effect of damages inherent in the feedback control system, and hence renders more difficult system health monitoring. Faced with this problem, we propose in this paper an efficient approach to specify indices of the system’s degradation and failure. The originality resides mainly in the assumption that the system input-output are the only monitoring information available for the feedback control system, and that no knowledge about the system physics is required. The monitoring data are first employed to construct the transfer function of the system. On the basis of this model, we next build two degradation indices for the system: (i) one based on the system poles, and (ii) another based on the maximum gain of the transfer function. Associated with each degradation index, a threshold is proposed to define the system failure. For illustration, we apply the proposed approach to a stochastically deteriorating stabilization loop control device in an inertial platform to assess its health state.