An adaptive spare parts demand forecasting method based on degradation modeling

Proactive maintenance and spare parts ordering are two fundamental components for prognostics and health management of system. To achieve timely proactive maintenance and spare parts ordering, it is necessary to forecast the demand of spare parts accurately according to the system's operating state. In this paper, a degradation modeling based method is proposed to adaptively forecast the demand of spare parts. Specifically, the system's lifetime is predicted in a probability form by modeling its degradation process as a Wiener process. To achieve adaptive lifetime estimation, the degradation rate parameter is recursively updated by Kalman filtering algorithm and then the lifetime is derived by considering the updated degradation rate parameter. Based on the lifetime distribution, the demand distribution of spare parts in a future time span can be forecasted by evaluating the convolution of the lifetime distribution. Finally, a case study for gyro's data is provided to illustrate the implementation of the proposed method.