AI Based Capacitor Aging Diagnosis for DC/DC Converters

As capacitors are under age, their performance degrade, which can lead to reduced reliability of the power converter. In this study, we propose a method for effectively analyzing capacitor’s condition by combining a data preprocessing technique using Lissajous curves with an AI model. The experimental results confirmed that the proposed method accurately predict the aging state of the capacitor by using the proposed Lissajous curves based preprocessing approach, demonstrating performance even under duty conditions that had never been trained on.