A Residual Storage Life Prediction Approach for Systems With Operation State Switches

This paper concerns the problem of predicting residual storage life for a class of highly critical systems with operation state switches between the working state and storage state. A success of estimating the residual storage life for such systems depends heavily on incorporating their two main characteristics: 1) system operation process could experience a number of state transitions between the working state and storage state; and 2) system's degradation depends on its operation states. Toward this end, we present a novel degradation model to account for the dependency of the degradation process on the system's operation states, where a two-state continuous-time homogeneous Markov process is used to approximate the switches between the working state and storage state. Using the monitored degradation data during the working state and the available system operation information, the parameters in the presented model can be estimated/updated under Bayesian paradigm. Then, the posterior probabilistic law of the number of state transitions and their transition times are derived, and further, the formulation for the predicted residual storage life distribution is established by considering the possible state transitions in the future. To be solvable, a numerical solution algorithm is provided to calculate the distribution of the predicted residual storage life. Finally, we demonstrate the proposed approach by a case study for gyroscopes.