Abstract:

In order to solve the problem that the traditional online prediction method of remaining useful lifetime based on accelerated degradation modeling needs to update the drift coefficient and diffusion coefficient synchronously under the specific conjugate distribution conditions, an online prediction method of equipment residual useful lifetime based on accelerated degradation modeling with proportional relationship is proposed. Firstly, the proportional relationship between diffusion coefficient and drift coefficient is introduced into the traditional Wiener degradation model, which ensure the possibility of synchronous updating of diffusion coefficient and drift coefficient from the modeling perspective. Secondly, a parameter estimation method based on two-step maximum likelihood is proposed to realize the reasonable estimation of model parameters. Then, the degradation data conversion rules are formulated based on the constant principle of acceleration factor, and the degradation status of equipment is updated online by Kalman filter principle. Finally, based on the total probability formula, the probability density function of the residual useful lifetime of the equipment under the condition of constant stress is derived. Taking the accelerated degradation data of an accelerometer as an example, it is proved that the proposed method can effectively improve the accuracy of remaining useful lifetime prediction, and has engineering application value.

Key words: Wiener process, proportion relationship, accelerated degradation model, Kalman filter, remaining useful lifetime prediction