| Popis: |
With the miniaturized and high-powered development of power equipment and electronic devices, the microchannel heat sink is becoming an important device widely used in the modern cooling field. The heat transfer performance of the microchannel heat sink is affected by many influence factors, and amongst them there exists a complex coupling effect. So far, the mechanism of the microchannel heat sink has not been clearly elucidated. To figure out the dominant characteristics that affect the heat transfer performance of the microchannel heat sink and the interaction mechanism between relevant characteristics, this work carried out numerical simulations and parameter sensitivity analysis to investigate the key factors of the fluid flow and heat transfer performance in rectangular microchannels, and considered variables included key physical parameters of the fluids and dimensional parameters of the microchannels. To satisfy the enormous amount requirement of sample size in the parameter sensitivity analysis, a machine learning model, Gradient Boosting Tree (GBT), was constructed based on computational fluid dynamics (CFD) simulation results; then, the Sobel sensitivity analysis algorithm was coupled in to analyze the multi-parameter contribution of each input parameter. Results show that the number of channels N and Reynolds number Re have a greater impact on heat transfer performance under low Re conditions, and the cross-sectional area A and aspect ratio α have a much higher effect on the pumping power rather than Nusselt numbers Nu; N and A have a positive second-order interaction on Nu. When moving to the conditions with a higher Re, the number of channels N became the dominant factor that influenced both heat transfer and flow performance of the microchannel heat sink. This work also provides effective guidance for the design and optimization of high-efficiency microchannel heat sinks. |
Full Text from ScienceDirect