• A rectangular microchannel cold plate with alternating independent dual channels is proposed. • Orthogonal experiments identify 20% propylene glycol as the optimum coolant. • Inlet temperature highly affects cooling efficiency with minimal impact on pressure drop. • An optimal parameter set improves performance while reducing temperature and pressure drop. • A highly reliable Nu correlation is established with excellent predictive accuracy. The widespread use of high-power-density energy storage batteries presents notable thermal management challenges. Conventional liquid cooling often faces uneven flow distribution, pronounced local hotspots, and high pressure drop, which collectively impair system efficiency and safety. To address these issues, this paper proposes a cross-flow microchannel cold plate structure. The design features cross-flow channels for uniform coolant distribution and adopts opposite-side inlet/outlet layouts to reduce temperature gradients. Through combined optimization of flow velocity and straight-channel geometry, the structure significantly reduces pressure drop while mitigating localized hot spots. Simulations were conducted to evaluate the effects of coolant type (aqueous solution, 20% ethylene glycol, 50% ethylene glycol, 20% propylene glycol), inlet temperatures (26–32 °C), and inlet flow velocities (0.5–0.8 m/s). A total of 16 orthogonal experiments were designed to determine the optimal operating condition. Results demonstrate that the 20% propylene glycol solution offers the best cooling performance, whereas the aqueous solution yields the lowest pressure drop. Moreover, lower inlet temperatures significantly improve cooling efficiency with negligible effect on pressure loss. Similarly, higher flow rates enhance both cooling capacity and temperature uniformity, though incur increased pressure drop. Consequently, the optimal parameter set (Combination 14: 20% propylene glycol, 26 °C, 0.7 m/s) was selected. Compared with the group exhibiting the lowest pressure drop, this configuration lowered the heat source surface temperature by 2.43 °C, while in comparison with the group having the lowest heat source surface temperature, it reduced the pressure drop by 728.67 Pa (approximately 16.3%).Furthermore, the Nusselt number (Nu) was fitted with a coefficient of determination (R 2 ) of 0.978, accompanied by narrow confidence and prediction intervals, which affirms the accuracy of the simulations. Finally, the application of the proposed Nu fitting formula for computational verification against other models resulted in a maximum error of as low as 13.2%, thereby demonstrating its broad applicability under the appropriate conditions.
Ding et al. (Sat,) studied this question.