Genetic Optimization of Pin Fin Heat Sinks Using a Finite Difference Formulation

The demand for electronics cooling is increasing with growing power density in modern electronic devices. High heat transfer capabilities and a small pressure drop thereby characterize optimal heat sink designs. In the development of efficient heat sinks, additive manufacturing allow more geometrical degrees of freedom. However, conjugate convective heat transfer analyses of entire heat sinks are computationally expensive within an optimization process. The presented method is based on a finite difference implementation approximating the thermal performance of the heat sink by its base plate. Heat sources and the characteristics of a priori optimized single pin fins, are coupled into the system via artificial boundary conditions. Based on a 2D heat sink setup, the methodology is compared to CFD models and the arrangement of the pin fins is optimized using a genetic algorithm. Although the FD model is imprecise in determining absolute values, it provides an appropriate estimation of the relative quantities and can serve as an initial geometry setup for topology optimizations.