Time Series Chain Graph for Reliability Covariate Modelling

In realistic applications, based on periodic or continuous monitoring, knowledge about system operating state and degree of degradation measured by some diagnostic variables as well as operating conditions have effect on the estimation of system remaining useful life (RUL) and planning of maintenance activities. On the other hand, the time series chain graph extends the directed graph by relating variables to each other over adjacent time steps. In this paper, by considering multiple external factors and multi-sensor based monitoring, the system degradation is modelled based on time series chain graph and the RUL is estimated using the proposed method.