Non-Parametric Optimization of Heat Sinks for Power Dense Motor Controllers 2022-26-0009

With the future of mobility moving towards electrification, there is an ever-increasing demand in both aerospace and automotive industry for achieving higher power density in inverters, controllers, etc. This has made thermal management a challenging task and warrants a need for exploring innovative cooling techniques to manage the dissipated heat. This paper focuses on a liquid cooled thermal management system for power dense applications. The heatsink design presented here is a pin fin arrangement staggered to induce swirling flow, which has been proven to enhance heat transfer. The traditional heatsink optimization involves creating a design of experiments (DoE) with parameters like fin diameter, spacing and height and performing thermal simulations to arrive at a design with enhanced heat transfer characteristics. The unique idea in this work is that, in addition to a traditional DoE based optimization, the topology of these circular pin fins has also been optimized using Ansys adjoint solver, resulting in lower junction temperature of the electronics. This gradient based topology optimization solver takes the flow simulation data and uses it to morph the pin fins to create unique wavy shaped pin fins that have enhanced heat transfer characteristics with negligible increase in pressure drop across the heatsink. Furthermore, the authors have also explored flow impingement localized to electronics mounting locations to further increase the heat transfer characteristics and reduce electronics junction temperature. These studies will help in increasing the power density of the electronics system and at the same time improve system reliability.