Remaining Useful Life Estimation of Stochastically Deteriorating Feedback Control Systems with a Random Environment and Impact of Prognostic Result on the Maintenance Process

Abstract

The objective and originality of this work are twofold. On one hand, it considers the degradation modeling and Remaining Useful Life (RUL) estimation for the closed-loop dynamic systems, which have not been addressed extensively in the literature. On the other hand, the paper examines how the prognosis result impacts the maintenance process. Indeed, due to their natural ageing and/or non desired effects of the operating condition, actuators deal with the loss of effectiveness which is a source of performance degradation of closedloop system. In this paper, we consider a control system with classical Proportional-Integral-Derivative controller and stochastically deteriorating actuator. It is assumed that the actuators are subject to shocks that occur randomly in time.An integrated model is proposed which jointly describes the states of the controlled process and the actuators degradation. The RUL can be estimated by a probabilistic approach which consists of two steps. First, the system state regarding the available information is estimated online by Particle Filtering method. Then, the RUL of the system is estimated by Monte Carlo simulation. To illustrate the approach and highlight the impact of the prognosis result on the maintenance process, a well-known simulated tank level control system is used. The maintenance decision rule is based on the quantiles of RUL histogram. In order to evaluate the performance of the maintenance policy, a cost model is developed.