Reliability Prediction for Inverters in Hybrid Electrical Vehicles

Due to the increasing importance of electronic components in automobiles, it becomes necessary to consider their reliability. This applies especially to Hybrid Electrical Vehicles (HEV) where a malfunction of the power electronics may force the vehicle to stop immediately. Of paramount importance for the reliability of power electronics is the component temperature and its variation. This paper deals with the development of an advanced simulation tool which is capable of determining the component temperature of a three-phase bridge even for long mission profiles. It determines the expected reliability from the simulation results. At first, losses in the semiconductors and dc-link capacitors are determined. These are fed into a flexible thermal model which computes the component temperatures for the whole mission profile. As basis for the reliability computation, failure rate catalogs, such as Military Handbook 217F or RDF 2000, are used. An approach using simple formulas for lifetime prediction is presented. According to failure rate catalogs, temperature cycles are of particular importance for the reliability of power semiconductors. A novel algorithm, detecting all relevant temperature cycles within the computed temperature curve is developed. Finally, the applicability and significance of the presented reliability prediction methods is assessed.