Estimating the Probability Density Function of Remaining Useful Life for Wiener Degradation Process with Uncertain Parameters

The effective prediction of remaining useful life is essential to realize system failure diagnosis and health management. The existing researches often assume that the degradation model is constant or the degradation process is measurable. The accurate degradation model, however, usually can not be established, and the parametric variation and measurement error of the degradation process are unavoidable, which makes it hard to obtain the exact value for predicting the remaining useful life. Regarding this problem, on basis of the concept of first failure time, a real-time probability density function is derived for the Wiener degradation process with the uncertainty of parameters, the stochasticity of degradation process and the randomness of measurement error. The main steps are as follows: firstly, the degradation model with three kinds of uncertainties is established, and then the stochastic degradation state and the parameters of the uncertainty model are estimated by fusion Kalman/UFIR filter; then, the analytical expression of the probability density function of remaining useful life is deduced. Finally, the correctness and effectiveness of the proposed method are verified by a group of comparison experiments and Monte Carlo simulations.