As a key component in power electronic systems, the accurate reliability assessment of metallized film capacitors (MFC) is very critical to assure the system reliability and safety. In most operating conditions, the gradual capacitance decrease induced by dielectric film aging and the accumulation of each sudden capacitance loss by random self-healing events under elevated temperature and voltage are the main degradation mechanisms of MFC. However, most investigations tend to directly fit the MFC degradation data by some statistical means, which ignores the specific degradation mechanisms. Such treatments for degradation data often perform a lower modeling accuracy, which may also bring an obvious deviation of reliability assessment result. In this article, we design a degradation mechanisms-based reliability modeling approach for MFC under temperature and voltage stresses. The dielectric film aging is characterized by Wiener process, where the aging rate is dependent with temperature. The compound Poisson process is developed to capture the random self-healing events, where the arrival rate and each sudden capacitance loss are related to the operating voltage as well as the dielectric film aging rate. Finally, an ADT with temperature and voltage acceleration for MFC is conducted to validate the effectiveness of the proposed approach.

作为电力电子系统中的关键部件，金属化薄膜电容器（MFC）的准确可靠性评估对于保证系统的可靠性和安全性至关重要。在大多数操作条件下，介电膜老化引起的电容逐渐降低以及在升高的温度和电压下随机自愈事件导致的每次突然电容损失的累积是 MFC 的主要退化机制。然而，大多数研究倾向于通过一些统计方法直接拟合 MFC 降解数据，而忽略了特定的降解机制。这种对退化数据的处理往往会导致建模精度较低，这也可能带来可靠性评估结果的明显偏差。在本文中，我们设计了一种基于退化机制的 MFC 在温度和电压应力下的可靠性建模方法。介电膜老化以维纳过程为特征，其中老化速率取决于温度。开发了复合泊松过程来捕获随机自愈事件，其中到达率和每次突然的电容损失都与工作电压以及介电膜老化率有关。最后，针对 MFC 进行了具有温度和电压加速的 ADT，以验证所提出方法的有效性。其中到达率和每次突然电容损失与工作电压以及介质膜老化率有关。最后，针对 MFC 进行了具有温度和电压加速的 ADT，以验证所提出方法的有效性。其中到达率和每次突然电容损失与工作电压以及介质膜老化率有关。最后，针对 MFC 进行了具有温度和电压加速的 ADT，以验证所提出方法的有效性。