Optimization of a microchannel heat sink using entropy minimization and genetic aggregation algorithm

A copper heat sink with rectangular U-shaped microchannels for cooling high power laser diode arrays was numerically optimised. The control volume method based commercial software was applied for solving heat transfer and fluid flow in the considered device. The optimisation procedure included over 130 various geometrical configurations with five independent geometric parameters. The Entropy Generation Minimization criterion was applied to find the optimal geometric structure. Some relations between geometrical parameters of the microchannel heat sink were analysed through statistical studies and two most important ones, i.e., the relative height (alpha) and width (beta) of a single microchannel, were selected. Using the genetic aggregation, the entropy generation characteristics were determined as a function of alpha and beta as well as mass flow rate of cooling medium ((G)over dot)). The most optimal from the second law of thermodynamics point of view configuration was the one with alpha(opt) = 0.86, beta(opt) = 0.347 and (G)over dot) = 0.00183 kg/s. Comparing with the initial geometry, over 35% reduction of irreversibility losses for the optimal heat sink had been achieved, i.e., heat sink thermodynamic efficiency increases from 90.05% to 91.8%, while maximum temperature of cooled semiconductor element was reduced from 333.3 K to 324.2 K.