Remaining Useful Life Prediction of Aluminum Electrolytic Capacitor with a Strain-based Health Indicator

This paper proposes a strain-based remaining useful life (RUL) prediction method for aluminum electrolytic capacitors (AECs). A strain sensing-based health indicator of AECs has recently been reported with the advantages of being non-invasive high sensitivity to degradation, and less susceptible to interference. This discovery gives a new opportunity for the penitential of AECs RUL prediction. In this paper, the degradation data of strain sensing of AECs are collected by an accelerated aging platform, and the degradation model of force signals is defined. Coefficient estimation, iterative updating, and RUL prediction of the degradation model are performed. The analyzed results show that the method can effectively fit the strain-based degradation characteristics of AECs and accurately predict the RUL.