Experimental study of topology optimized, additively manufactured microchannel heat sinks designed using a homogenization approach

Abstract

Topology optimization generates complex geometry heat sink designs having intricate features well-suited for fabrication by additive manufacturing. In particular, a homogenization approach to topology optimization creates microchannel heat sink designs wherein the material porosity distribution corresponds to microstructures of varying dimensions, thereby eliminating the need to penalize the porosities to achieve binarized solid/liquid designs. This approach is inherently able to take into consideration the capabilities and limitations of available additive manufacturing processes as the types of microstructures can be user-defined. The current study is the first to fabricate and experimentally test microchannel heat sink designs that are topology optimized using the homogenization approach. To this end, a multi-objective optimization is performed to generate a series of Pareto optimal designs that minimize pressure drop and thermal resistance. The effect of the grid resolution is investigated. The resulting topology optimized designs are found to have different geometries and performances with different grid cell sizes because this impacts the physical dimensions of the microstructures. A series of pin fin arrays with different grid cell sizes and fin thicknesses are fabricated using Direct Metal Laser Sintering of AlSi10Mg. The additively manufactured surfaces have high roughness which creates connections between pin fins with small gap sizes and creates tortuous flow paths. The smallest pin fin that consistently survived the fabrication process is ∼0.25 mm thick and hence series of topology optimized microchannel heat sinks with a 0.5 mm grid cell size are additively manufactured. The pressure drop and thermal resistance of the fabricated heat sinks are experimentally characterized, and the measured results lie on the Pareto optimality curve predicted by the topology optimization algorithm. This work demonstrates that the homogenization approach to topology optimization generates high-performance microchannel heat sink designs that can be achieved using available additive manufacturing processes.