Ordered clustering method and degradation trend analysis for performance degradation of tantalum capacitor

Abstract

The performance degradation trend of a tantalum capacitor has different characteristics in different degradation regimes with different temperatures and humidity. Clustering method and degradation trend analysis can be useful for early warning of life. In this paper, according to the degradation characteristics of the capacitance and the dissipation factor with different temperature and humidity, the experiment is designed, and the failure mechanisms of different degeneration regimes are analyzed. After the ordered clustering method is introduced, using the natural property and the change rate α(k) of the slope β(k) of the error function δ[p(m,k)], the determination method of the optimization clustering number is put forward. The Davies‐Bouldin Index is proposed as the evaluation method of the ordered clustering result, and the validation of the evaluation method was verified by the degradation data with other stresses. For the stable degradation regime, the degradation trend analysis is completed. The error result shows that the degradation fitting functions can be a good fit for the degradation trend with different stress. According to the degradation fitting analysis, the degradation prediction method is described to find the relationship between prediction life and different temperature and humidity to realize the failure time prediction with different stresses using the early degradation data. © 2019 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.