Field Visualization of Pin Fin Arrays Designed by Computationally Efficient Three-dimensional Optimization

Pin fin heat sinks with optimized geometries hold great promise for surpassing existing performance boundaries. This dissertation focuses on the design, fabrication, characterization, and flow field analysis of additively manufactured pin fin heat sinks. First, optimal pin shapes were designed by employing numerical shape optimization methods, including non-gradient-based genetic algorithms and gradient-based adjoint optimization. To assess the physical characteristics of design candidates, the genetic algorithm is coupled to a finite element model, while the adjoint solver is combined with a finite volume model. The optimization problem focuses on a pin fin array within a cooling channel utilizing water as the working fluid, with Reynolds numbers ranging from 500 to 5000. The objective is to minimize both the heated surface temperature and the pressure drop. The optimized fin array was fabricated using the laser powder bed fusion technique, employing aluminum alloy Al6061-RAM2 as the printing material. This alloy boasts exceptional ductility, high strength, and excellent thermal conductivity. An optimal symmetrical pin shape identified through the genetic algorithm exhibits a 44% enhancement in the heat transfer coefficient with a 3% decrease in pressure drop compared to traditional cylindrical pins. Additionally, the optimal non-symmetrical pin shape introduced by the adjoint method demonstrates up to a 48% improvement in the heat transfer coefficient with a 2% boost in pressure drop penalty compared to its cylindrical counterpart. Finally, to elucidate the mechanisms behind the enhanced heat transfer in the additively manufactured pin fin arrays, laser-induced fluorescence (LIF) and particle image velocimetry (PIV) techniques were utilized. These methods reveal the flow patterns, mixing phenomena, and temperature distribution within the fluid.