Study of the Performance of a Novel Radiator with Three Inlets and One Outlet Based on Topology Optimization

In recent years, in order to obtain a radiator with strong heat exchange capacity, researchers have proposed a lot of heat exchangers to improve heat exchange capacity significantly. However, the cooling abilities of heat exchangers designed by traditional design methods is limited even if the geometric parameters are optimized at the same time. However, using topology optimization to design heat exchangers can overcome this design limitation. Furthermore, researchers have used topology optimization theory to designed one-to-one and many-to-many inlet and outlet heat exchangers because it can effectively increase the heat dissipation rate. In particular, it can further decrease the hot-spot temperature for many-to-many inlet and outlet heat exchangers. Therefore, this article proposes novel heat exchangers with three inlets and one outlet designed by topology optimization to decrease the fluid temperature at the outlet. Subsequently, the effect of the channel depth on the heat exchanger design is also studied. The results show that the type of exchanger varies with the channel depth, and there exists a critical depth value for obtaining the minimum substrate temperature difference. Then, the flow and heat transfer performance of the heat exchangers are numerically investigated. The numerical results show that the heat exchanger derived by topology optimization with the minimum temperature difference as the goal (Model-2) is the best design for flow and heat transfer performance compared to other heat sink designs, including the heat exchanger derived by topology optimization having the average temperature as the goal (Model-1) and conventional straight channels (Model-3). The temperature difference of Model-1 can be reduced by 37.5%, and that of Model-2 can be decreased by 62.5% compared to Model-3. Compared with Model-3, the thermal resistance of Model-1 can be reduced by 21.86%, while that of Model-2 can be decreased by 47.99%. At room temperature, we carried out the forced convention experimental test for Model-2 to measure its physical parameters (temperature, pressure drop) to verify the numerical results. The error of the average wall temperature between experimental results and simulation results is within 2.6 K, while that of the fluid temperature between the experimental and simulation results is within 1.4 K, and the maximum deviation of the measured Nu and simulated Nu was less than 5%. This indicated that the numerical results agreed well with the experimental results.