Multi-objective shape optimisation of natural convection cooled plate-fin heat sinks with localised heat sources

Plate-fin heat sinks are used as a passive cooling technique to dissipate the heat generated by high power electronic components such as power amplifiers in wireless telecommunication hardware. There has been an increasing demand for cooling performance enhancement, and this could be achieved using numerical shape optimisation. Here, computational fluid dynamics (CFD) simulations are combined with a Multi-Objective Genetic Algorithm to optimise natural convection from parallel plate-fin heat sinks, with the objective of minimising the thermal resistance and heat sink mass within a given box volume. Two heat sinks are studied: with one and two localised heat sources. The performance enhancement includes the optimisation of the fin spacing, the position and length of secondary fins. Firstly, a quantitative comparison is presented between two well-established CFD packages, reporting their strengths and weaknesses for this specific application. Secondly, the heat sink performance is verified against data available in the literature, followed by multi-objective optimisation. The findings lead to a best practice procedure for determining the optimum heat sink configuration, resulting in a more effective cooling technology to improve electronic component performance and reliability.