Minimization of Heat Sink Mass Using CFD and Mathematical Optimization

Abstract

This paper describes the use of CFD and mathematical optimization to minimise heat sink mass given a maximum allowable heat sink temperature, a constant cooling fan power and heat source. Heat sink designers have to consider a number of conflicting parameters. Heat transfer is influenced by, amongst others, heat sink properties (such as surface area), airflow bypass and the location of heat sources, whilst size and/or mass of the heat sink needs to be minimized. This multiparameter problem lends itself naturally to optimization techniques. In this study a commercial CFD code, STAR-CD, is linked to the DYNAMIC-Q method of Snyman. Five design variables are considered for three heat source cases. Optimal designs are obtained within six design iterations. The paper illustrates how mathematical optimization can be used to design compact heat sinks for different types of electronic enclosures.