Shape optimization of inlet header of microchannel heat sink using surrogate model combined with genetic algorithm

A lot of work has documented the significance fluid flow uniformity on thermal-hydraulic performance microchannel heat sink (MCHS). The purpose this is to optimize shape inlet/outlet headers a MCHS improve distribution characteristics using back-propagation neural networks combined with genetic algorithm(GA-BP) as surrogate model. slanted edge inlet header defined quadratic parabola instead straight line. Meanwhile, optimally designed for different rates. 40 training sample points and 6 testing geometry structures are by Latin hypercube sampling (LHS) method. 3-D Computational Fluid Dynamic (CFD) calculation performed all models. objective function nonuniformity flow. It found that prediction GA-BP capable obtaining values within space. Through optimizations, optimal structure can be reduced 52.43%to 33.17%for velocity 0.05m/s to0.1m/s,respectively, compared original design. results demonstrate parabolic treatment well structural optimization greatly MCHS.