Numerical investigation of single-phase manifold microchannel heat exchanger for use in electric vehicle battery cooling/heating

Abstract

Three microchannel geometries were numerically studied to investigate its effects on a Manifold Microchannel Heat Exchanger system using a full conjugate heat transfer model. Fundamental theories and recent advancements in the field were first discussed, followed by the numerical tools and methodology employed in the study. The geometries tested are rectangular microchannels, circular re-entrant cavity microchannels and sinusoidal wavy microchannels. The pressure drop and heat transfer coefficient data are presented graphically against Re, whereby the investigated Re in the microchannel ranged from 166 to 1659. Their overall performance is evaluated using two performance evaluation criteria where considerations of both pressure drop and heat transfer performance are made simultaneously. Insights were then taken from the simulation data for possible geometry optimization to further boost performance. The sinusoidal wavy microchannel enables modest heat transfer coefficient improvements (18.48% on average) over the rectangular microchannel but at a cost of significant increase in pressure drop that rises exponentially with Re. The re-entrant cavity improves the heat transfer coefficient of the rectangular microchannel by 7.21% on average while simultaneously decreasing pressure drop by 4.8% on average within the investigated Re. Subsequently, the numerical model was validated with experimental measurements, and it was found that the model matches very closely to the physical measurements (within 5%). It was found that the inaccuracy in the 3D printing used to manufacture the microchannel heat sink has a significant effect on the hydrodynamic performance of the MMHX system, and the model was able to duplicate it. The inaccuracy in 3D printing has actually reduced pressure drop by about 37% on average in the simulated Re. Comparing the heat transfer coefficient curves of a one-pass system and a four-pass system has led to the conclusion that the one-pass system models the behavior of the four-pass system very closely and can be used to explore different geometric designs at reduced computational power.