A Wiener Process Model With Kernel-based Time Transformation For Nonlinear Degradation Data

Hanyu Wang,Zhen Chen,Tangbin Xia,Di Zhou,Ershun Pan
DOI: https://doi.org/10.1109/phm-nanjing52125.2021.9613044
2021-01-01
Abstract:Degradation processes of complex systems generally have two key characteristics: nonlinearity and stochasticity. The stochastic process degradation models with nonlinear transformation functions are widely used in degradation modeling, such as Wiener process model with an exponential term. However, the exponential terms for time transformation are not suitable for the nonlinear degradation processes which caused by complex failure mechanisms. To capture the nonlinear characteristics more accurately, this study proposes a nonlinear Wiener process model with kernel-based time transformation for nonlinear degradation analysis. A Bayesian estimation method is developed for parameter estimation. Compared with the experimental results of linear Wiener model, exponential Wiener model and RVM model, the effectiveness and universality of the method proposed in this paper is verified among different degradation curves of several products.
What problem does this paper attempt to address?