Converter reliability optimization based on mission profile

This paper presents the solution to an optimization problem that involves two objective functions to minimize thermal resistance and weight of a heatsink. Several constraints are taken into account, in such a way that the overall failure rate is reduced, thus yielding an enhanced reliability. A Genetic Algorithm was applied to optimize the design of the heatsink structure. The optimized heatsink is numerically simulated using computational fluid dynamics (CFD) software. The maximum temperature difference is 15.4 °C, and the transferred power of heat flux is 3.7 W, so the thermal resistance θ_JA is 4.1627 °C/W, which agrees well with the analysis result of thermal resistance network model. The optimization yielded an improvement of the converter failure rate, from 315.43 FITs, to 97.73 FITs.