MULTIOBJECTIVE OPTIMIZATION OF A SLIT RIB IN A RECTANGULAR COOLING CHANNEL

초록

A slit rib in a rectangular channel is optimized to simultaneously enhance heat transfer and reduce pressure drop using surrogate modeling and a three-dimensional Reynolds-averaged Navier-Stokes analysis. A multiobjective genetic algorithm is used for optimization. The shear stress transport turbulence model is selected as a turbulence closure model among three different turbulence models. Based on the results of a parametric study with the use of geometric parameters of the slit rib, three design variables (rear-to-front slit thickness, front slit thickness-to-rib height, and slit-to-rib height ratios) and two objective functions related to heat transfer and pressure drop in the rectangular channel are selected for optimization. Thirty-two design points are selected by Latin hypercube sampling. A surrogate model using response surface approximation is used to approximate the objective functions. The values of the objective functions at the selected design points are calculated to construct the surrogate models. A Pareto-optimal front that represents a compromise between the two objective functions is obtained using a multiobjective genetic algorithm. Through the optimization, the objective function values are improved by up to 7.9% and 7.2% for the area-averaged Nusselt number and friction factor, respectively, compared to a reference design.

키워드