隨著 CPU、GPU 及高性能電子元件的運算能力持續提升，單位面積內產生的熱量也快速增加，使散熱設計成為影響元件的可靠度與持續壽命的重要因素。傳統氣冷方式在高熱通量條件下逐漸受到限制，因液冷散熱器因具備較高的熱移除能力，成為高功率電子冷卻技術的重要發展方向。本研究首先以文獻中的雙螺旋液冷散熱器作為基準模型，建立三維共軛熱傳數值模型，並透過 Case 1 與 Case 2 的溫度分布及熱性能趨勢進行驗證。在完成模型驗證後，進一步在相同散熱器尺寸與邊界條件下重新設計三螺旋與四螺旋流道，探討多螺旋流道配置對冷卻性能與溫度均勻性的影響。
本研究使用 SolidWorks 建立流道幾何模型，並以 ANSYS Fluent 進行穩態數值模擬。本文中的模擬過程皆假設流體為不可壓縮的單相流，並考慮固體與流體之間的共軛熱傳現象。工作流體採用去離子水，散熱器材料則設定為純銅。主要操作條件包含入流速度 0.15 m/s、入流溫度 25 °C、出口壓力 0 Pa，並於 CPU 表面施加 10 W/cm² 的均勻熱通量。為評估不同流道設計的散熱表現，本研究以 CPU 表面之最高溫度、平均溫度及表面溫差作為主要比較指標。
結果顯示，重建後的雙螺旋散熱器能合理再現近似參考文獻中的溫度分布與熱傳性能趨勢。在入流速度 0.15 m/s 下，Case 1 表面溫差之誤差約為 0.5%，Case 2 平均溫度之誤差約為 0.35%，顯示本次研究所建立的數值模型具有可接受的可靠性。進一步比較三種流道配置後發現，四螺旋流道在整體冷卻能力與溫度均勻性方面表現最佳。在相同的操作條件下，四螺旋流道的最高溫度為 39.75 °C，平均溫度為 39.07 °C，表面溫差為 1.63 °C，皆優於雙螺旋與三螺旋流道。此結果表示，四螺旋配置能提供較完整的流道覆蓋範圍，使冷卻效果分布較其餘兩者更加均勻。
此外，本研究進一步針對四螺旋流道進行最大允許功率分析。以 CPU 表面最高溫度需小於等於 90 °C 作為限制條件時，四螺旋流道可承受的最大允許熱通量約為 45.15 W/cm²，對應總輸入功率約為 722.44 W。在此條件下，CPU 表面溫差約為 7.73 °C，顯示四螺旋流道即使在高熱輸入的條件下，仍能展現良好的溫度均勻性。最後，將 CPU 表面劃分為 16 個區域進行局部熱源分析，結果顯示熱源位置會明顯影響溫度分布，其中角落區域因鄰近絕熱邊界，熱傳路徑較受限制，較容易形成局部高溫。雖然四螺旋流道具有較佳的整體散熱能力，但角落區域仍是後續流道優化的重要方向。

As the power density of CPUs, GPUs, and other high-performance electronic components continues to increase, thermal management has become a critical factor affecting device reliability, operating stability, and service life. Under high heat flux conditions, conventional air-cooling techniques gradually become insufficient because of their limited heat removal capability. Therefore, liquid-cooled heat sinks have received increasing attention as an effective cooling design solution for compact and high-power electronic devices. In this investigation, a double-spiral heat sink reported in the literature was first reconstructed as the reference model. A three-dimensional numerical model considering conjugate heat transfer was then established and validated by comparing the simulated temperature distributions and thermal performance trends of Case 1 and Case 2 with the reference results. After the validation process, three-spiral and four-spiral liquid-cooled channel designs were developed under the same heat sink dimensions and operating conditions to investigate the effect of multi-spiral channel arrangements on cooling performance and temperature uniformity.
The geometric models were created using SolidWorks, and steady-state numerical simulations were performed by using ANSYS Fluent. The working fluid was assumed to be incompressible, single-phase deionized water, while pure copper was used as the heat sink material. The main boundary conditions included an inlet velocity of 0.15 m/s when inlet temperature is 25 °C, an outlet pressure was set to 0 Pa, and a uniform heat flux of 10 W/cm² applied to the bottom of the heatsink. To predict the thermal performance of different channel configurations, the maximum temperature, average temperature, and surface temperature difference of the CPU surface were selected as the main comparison parameters.
The validation results show that the reconstructed DSHS model was reasonably reproduce the thermal behavior reported in the reference study. When inlet velocity was set to 0.15 m/s, the error of Case 1 surface temperature difference was approximately 0.5%, while the average temperature error of Case 2 was approximately 0.35%. These results indicate that the numerical model established in this study provides acceptable reliability for further channel comparison. Among the three channel configurations, the four-spiral channel exhibited better overall cooling performances. Under the same operating conditions, the four-spiral channel achieved a maximum temperature of 39.75 °C, an average temperature of 39.07 °C, and a surface temperature difference of 1.63 °C, all of which were lower than those of the DSHS and three-spiral channels. This suggests that the four-spiral arrangement provides wider channel coverage and produces a more uniform cooling distribution over the heated surface.
Furthermore, the maximum allowable power of the four-spiral channel was evaluated by using 90 °C as the upper limit of the CPU surface temperature. The results show that the four-spiral channel can stand a maximum heat flux of approximately 45.15 W/cm², corresponding to a total heat input of about 722.44 W. Under this condition, the surface temperature difference was approximately 7.73 °C, indicating that the four-spiral channel can still maintain good temperature uniformity under high-heat-flux input. Finally, CPU surface was divided into 16 regions for localized heat source analysis. The results reveal that the location of the heat source has a significant influence to the temperature distribution. Particularly on the corner regions tend to generate higher local temperatures because the adjacent adiabatic boundaries limit the available heat conduction paths. Although the four-spiral channel demonstrates strong overall cooling capability, the corner regions remain important targets for future channel optimization.

[1] S. M. S. Murshed and C. A. Nieto de Castro, "A critical review of traditional and emerging techniques and fluids for electronics cooling," Renewable and Sustainable Energy Reviews, vol. 78, pp. 821–833, 2017.
[2] I. Mudawar, "Assessment of high-heat-flux thermal management schemes," IEEE Transactions on Components and Packaging Technologies, vol. 24, no. 2, pp. 122–141, 2001.
[3] S. S. Kang, "Advanced Cooling for Power Electronics," in 7th International Conference on Integrated Power Electronics Systems (CIPS), 2012.
[4] R. J. McGlen, R. Jachuck, and S. Lin, "Integrated thermal management techniques for high power electronic devices," Applied Thermal Engineering, vol. 24, pp. 1143–1156, 2004.
[5] P. Wang, P. McCluskey, and A. Bar-Cohen, "Hybrid Solid- and Liquid-Cooling Solution for Isothermalization of Insulated Gate Bipolar Transistor Power Electronic Devices," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 3, no. 4, pp. 601–611, 2013.
[6] F. Chen, C. Liu, X. Wu, S. Yao, W. Liu, and W. Chang, "The flow and heat transfer performance of a novel double-spiral heat sink for processing unit cooling," International Journal of Heat and Fluid Flow, vol. 119, 110299, 2026.
[7] C. Ma, Y. Guo, D. Kong, and X. Liu, "The analysis of flow and heat transfer characteristics in spiral microchannel heat sinks," International Journal of Heat and Fluid Flow, vol. 117, 110106, 2026.
[8] S. Liu et al., "Multi-objective optimization of a bionic microchannel heat sink based on Fibonacci spiral for electronic components," International Journal of Heat and Mass Transfer, vol. 253, 127544, 2025.
[9] H. M. Rashad, Y. M. Najim, and H. H. Ismaeel, "Hydrothermal analysis of archimedean spiral single and dual channel heat sink for CPU cooling," Journal of Thermal Engineering, vol. 10, no. 1, pp. 50–61, 2024.
[10] X. Wei et al., "Design and Evaluation of the Zigzag-fin Multi-parallel Serpentine Heat Sink for IGCTs," IEEE Transactions on Components, Packaging and Manufacturing Technology, early access, 2026.
[11] E. S. Cho, J. W. Choi, J. S. Yoon, and M. S. Kim, "Experimental study on microchannel heat sinks considering mass flow distribution with non-uniform heat flux conditions," International Journal of Heat and Mass Transfer, vol. 53, pp. 2159–2168, 2010.
[12] A. Tikadar and S. Kumar, "Local hotspot thermal management using metal foam integrated heat sink," Applied Thermal Engineering, vol. 221, 119632, 2023.
[13] Z. Xue, Y. Yan, K. Shen, J. You, C. Zhang, and Z. He, "Investigation on thermal-hydraulic performance for local hotspot via variable density design of a micro-jet heat sink," International Communications in Heat and Mass Transfer, vol. 156, 107668, 2024.
[14] B. Cao and Z. Wu, "Microchannel heat sinks for hotspot thermal management: achieving minimal pressure drop and maximal thermal performance," International Journal of Heat and Mass Transfer, vol. 236, 126411, 2025.
[15] X. Li and Y. Xuan, "Embedded Cooling of High-Heat-Flux Hotspots Using Self-Adaptive Microchannel/Pin-Fin Hybrid Heat Sink," IEEE Transactions on Components, Packaging and Manufacturing Technology, vol. 12, no. 10, pp. 1627–1635, 2022.
[16] Intel, “Intel® Core™ i9-11900K Processor specifications,” Intel ARK. [Online]. Available: Intel official website. Accessed: Jun. 28, 2026.