SUBJECT: Ph.D. Proposal Presentation
   
BY: Brian Kelly
   
TIME: Tuesday, October 12, 2021, 4:00 p.m.
   
PLACE: https://bluejeans.com/138243657/2055, Online
   
TITLE: Digital Codesign of a Power Module with Integrated Thermal Management
   
COMMITTEE: Dr. Samuel Graham, Chair (ME)
Dr. Yogendra Joshi (ME)
Dr. Gilbert Moreno (NREL)
Dr. Satish Kumar (ME)
Dr. Shyh-Chiang Shen (ECE)
 

SUMMARY

Power electronics are a necessity for the many types of electrical energy conversions that are required from electricity sources to load devices. Innovation in these devices is typically driven by the need for higher power density, through either higher power or smaller feature sizes. One of the primary bottlenecks to this advancement is failure mechanisms caused by overheating. The packaging of these electronics typically involves several layers of materials with the die material containing the heat source, that is then attached to a convective thermal management solution. Often the effectiveness of a thermal management solution is characterized by both the conductive and convective thermal resistances. Recently, there has been considerable efforts and progress in the development of technologies towards the reduction of the convective thermal resistances (both single- and two-phase cooling). Now there exists an opportunity for innovation in the reduction of the conductive thermal resistances. This can be done by utilizing a direct cooling design that removes several common layers in a typical package such as the encapsulant and direct bonded copper layers, this study seeks to investigate such a design. In order to address some of the design changes and understand their impact both conjugate heat transfer and thermomechanical models are developed. An optimization scheme is implemented to consider the impacts of various design parameters on the thermohydraulic and thermomechanical performance of the device and arrive at a device that has reduced volume with equivalent (or improved) thermohydraulic and thermomechanical performance. Finally, a mission-based reliability analysis of the optimized designs is proposed and the trade-offs of the designs are considered.