Particle filter for condition estimation and lifetime prognosis of manufacturing system considering degradation of condition monitoring device

Abstract

Predictive maintenance is a promising maintenance approach for guaranteeing the system performances. In predictive maintenance, condition monitoring plays an important role since it indicates current degradation state of the system, which supports life time prognosis and maintenance decision making. However, the condition monitoring device also deteriorates with time and usage. The degradation of condition monitoring device leads to inaccurate system condition monitoring data and lifetime prognosis. This consequently leads to non-appreciate maintenance decision making. To address this issue, this paper applies Particle filter for estimating the degradation level and predicting Remaining useful life (RUL) of the system based on observed information. This method operates as a Bayesian filter, employing Bayesian theory for state estimation. Several examples are conducted to show the feasibility and effectiveness of the proposed method. The obtain result show that Particle filter can increase the accuracy of system state estimation significantly.