Comparative Study on Remaining Useful Life Prediction of Domestic and Imported Electrolytic Capacitors Based on a Power-Function Nonlinear Wiener Process

In practical engineering applications of domestic electrolytic capacitors, standard degradation indicators such as capacitance are often difficult to obtain continuously, whereas temperature monitoring data are more readily available. To address this issue, this paper investigates a comparative remaining useful life prediction method for domestic and imported electrolytic capacitors based on a power-function nonlinear Wiener process. For the domestic capacitor, a temperature sampling sequence is used to construct a degradation representation through max-min normalization, moving-average smoothing, and baseline-shift processing, and then a nonlinear Wiener process under a power-function time scale is established for lifetime prediction. For the imported capacitor, the targetsample degradation data are digitally reconstructed from published degradation curves, and remaining useful life prediction is performed under the same modeling framework. The results show that the fitted time-scale parameter of the domestic capacitor is 1.3630 with an RMSE of 27.5384, while that of the imported capacitor is 2.4049 with an RMSE of 3.7540. Since both time-scale parameters are greater than 1, the degradation processes of both capacitors exhibit clear nonlinear accelerating characteristics. The results indicate that the power-function nonlinear Wiener process can effectively characterize the degradation evolution of both domestic and imported electrolytic capacitors under the current data condition, and demonstrate the feasibility of remaining useful life prediction using temperature-derived degradation representation for domestic electrolytic capacitors.