Towards Prognosis on Interconnected Distributed Closed-Loop Control System: A Data-Driven Perspective

Through distributed monitoring and control, interconnected distributed systems have extensively promoted the efficiency of industrial production. Remaining useful life (RUL) prediction is a critical aspect of system prognosis. The existing RUL prediction approaches are mainly for conventional centralized systems. How to accomplish the RUL prediction and corresponding faulttolerant control on interconnected distributed systems needs further work. Consequently, from a data-driven perspective, this article integrates the state space model and the stochastic degradation process model. A datadriven particle filter approach is proposed for accurate hidden degradation state estimation and RUL prediction for interconnected distributed closed-loop control systems. Moreover, considering the degradation of system performance, this paper designs a fault-tolerant control strategy for the system, which can alleviate the degradation of the entire system to a certain extent. In this paper, an inertial platform control system is adopted to verify the proposed approach. Experiment shows that the proposed approach can accurately predict and extend the RUL.