통계적 방법을 이용한 DC-link 필름 커패시터의 수명 예측

In electronic vehicles (EVs) or hybrid electronic vehicles (HEVs), an inverter system has a DC-link capacitor to buffer the high frequency current. A film capacitor has been used as the DC-link capacitor in high level power system. The performance of the film capacitor has deteriorated over operating time. This can be one of critical factors in the EVs and HEVs. For this reason, the lifetime and reliability of the film capacitor are key factors to show the stability of the vehicle inverter system. Therefore, a lot of methods to predict the lifetime of the film capacitor using physical or chemical equations have been researched. However, these researches have a problem about robustness of uncertainty. In this paper, the lifetime and stability of the film capacitor are guaranteed by stochastic methods. A new model combined by random variable deterioration model and gamma process deterioration model to consider natural and sudden loss of the film capacitor is proposed. the parameters of each model are obtained by using curve fitting algorithm and method of moment. Using these models and algorithms, the lifetime of the film capacitor is predicted. Comparing experiment result and the proposed prediction model proves the proposed method has good accuracy.