An Online Remaining Life Prediction Method Based on Nonlinear Degradation Model of Equipment Electronic System

Electronic equipment will be affected by various types of fault mechanisms under various pressures, and its reliability needs to be strictly guaranteed in industrial applications. It is urgent to develop robust prediction models to detect abnormal behaviors and predict the life of electronic equipment. Therefore, this paper proposes an online remaining life prediction method based on a nonlinear degradation model. Taking the BUCK circuit as the research object, aiming at the nonlinear degradation characteristics of the product, a nonlinear degradation model based on the diffusion process is established, and the nonlinear drift parameters in the degradation model are defined as state variables, and the monitored performance degradation is used as the observation variable to establish the state space model. The Kalman filter is used to update the state variables, and the degradation data is divided into 0-50h, 0-70h, 0-90h, and 0-110h as the specific values of the actual operation, and the predicted values of the nonlinear drift parameters are obtained as 0.1722, 0.1553, 0.1420, 0.1272, And then from the remaining life distribution function to obtain the online average remaining life of 160.6048h, 149.0391h, 137.3590h, 128.1768h. Compared with the linear degradation model, the result of the nonlinear degradation model is 23.63% lower than the result of the linear degradation model on average.