Multi objective Design Optimization of Branching, Multifloor, Counterflow Microheat Exchangers

Abstract

Heat removal capacity, coolant pumping power requirement, and surface temperature nonuniformity are three major challenges facing singlephase flow microchannel compact heat exchangers. In this paper multiobjective optimization has been performed to increase heat removal capacity, and decrease pumping power and temperature nonuniformity in complex networks of microchannels. Threedimensional (3D) fourfloor configurations of counterflow branching networks of microchannels were optimized to increase heat removal capacity from surrounding silicon substrate (15 أ— 15 أ— 2 mm). Each floor has four different branching subnetworks with opposite flow direction with respect to the next one. Each branching subnetwork has four inlets and one outlet. Branching patterns of each of these subnetworks could be different from the others. Quasi3D conjugate heat transfer analysis has been performed by developing a software package which uses quasi1D thermofluid analysis and a 3D steady heat conduction analysis. These two solvers were coupled through their common boundaries representing surfaces of the cooling microchannels. Using quasi3D conjugate analysis was found to require one order of magnitude less computing time than a fully 3D conjugate heat transfer analysis while offering comparable accuracy for these types of application. The analysis package is capable of generating 3D branching networks with random topologies. Multiobjective optimization using modeFRONTIER software was performed using response surface approximation and genetic algorithm. Diameters and branching pattern of each subnetwork and coolant flow direction on each floor were design variables of multiobjective optimization. Maximizing heat removal capacity, while minimizing coolant pumping power requirement and temperature nonuniformity on the hot surface, were three simultaneous objectives of the optimization. Paretooptimal solutions demonstrate that thermal loads of up to 500 W/cm2 can be managed with fourfloor microchannel cooling networks. A fully 3D thermofluid analysis was performed for one of the optimal designs to confirm the accuracy of results obtained by the quasi3D simulation package used in this paper.