熱擴散介質應用於電子冷卻上之熱傳特性研究

Thermal Behavior for Heat Spreaders in Electronics Cooling Applications

林大衛(Ta-Wei Lin)

指導教授 ： 洪英輝 ； 傅建中

國立清華大學/工學院/動力機械工程學系/博士(2008年)

摘要

本論文研究乃針對含單一均勻熱源或單一非對稱熱源之熱擴散介質在無限制空間或限制空間下應用不同冷卻方法作一系列的熱流實驗與數值探討。探討之冷卻方法有自然冷卻與風扇噴流衝擊冷卻兩種。在研究中有關影響自然冷卻及風扇噴流衝擊冷卻之熱流性能的相關參數則分別列之如下：（一）自然冷卻－格拉雪夫數(GrT或GrH)、限制平板間距與特徵長度比（S/LH）、熱源面積與熱擴散介質面積比（r）、熱源位置及熱擴散介質種類。這些參數的探討範圍是GrH = 8.15×106 – 3.75×107， S/LH=0.077、 與 r=0.217-1。（二）風扇噴流衝擊冷卻－格拉雪夫數(GrT或GrH)、雷諾數（Re）、熱源面積與熱擴散介質面積比（r）、熱源位置及熱擴散介質種類。這些參數的探討範圍是GrH = 1.39×108 – 3.42×108，Re = 4344-7125與 r = 0.217-1。
為了準確估算各種熱擴散介質之有效熱傳導係數，本研究提出一個結合熱網絡數值計算及自然冷卻實驗量測的半經驗方法，可以省時且有效地求得熱擴散介質之有效熱傳導係數。研究中針對十一種熱擴散介質提出了三條溫度函數的有效熱傳導係數之經驗公式。另外，本研究也成功地利用類似的半經驗方法探討了含單一均勻熱源或單一非對稱熱源之熱擴散介質在自然冷卻與風扇衝擊噴流冷卻下之局部與平均熱傳特性。研究中不僅針對相關的實驗熱傳特性作有系統的參數探討外，同時也應用變異數分析F檢測法而針對相關影響參數作統計敏感度分析，其結果與結論與實驗皆相當吻合。此外，針對含單一均勻熱源或單一非對稱熱源之熱擴散介質在自然冷卻與風扇衝擊噴流冷卻下之平均紐賽數，本研究提出五條新的通用經驗公式。為了探討含非對稱熱源之熱擴散介質表面混合對流平均熱傳特性，在本研究中提出了兩條新的平均紐賽數組合經驗公式，此兩條公式為r及GrT/Re2 (或GrH/Re2.5)之函數，它們可準確地預測自然對流與風扇噴流衝擊交互影響之混合對流的平均熱傳特性。另外，本研究更進一步地針對外部熱阻與整體熱阻提出了新的經驗公式。
最後，針對含單一均勻熱源或單一非對稱熱源之熱擴散介質在自然冷卻與風扇衝擊噴流冷卻下之平均熱傳特性，本研究使用反應曲面法搭配實驗計劃法求得迴歸曲面模型，可以準確地預測熱擴散介質在自然冷卻與風扇衝擊噴流冷卻下之平均紐賽數與整體熱阻。此統計模型結果與實驗數據比對，其最大及平均誤差分別小於10.84%與5.89%。而與經驗公式比對，其最大及平均誤差則分別小於9.98%與5.73%。

關鍵字

熱擴散介質 ； 電子冷卻 ； 自然對流 ； 風扇衝擊噴流 ； 熱傳導係數 ； 雷諾數

並列摘要

A series of experimental and numerical studies on the heat transfer behavior for unconfined/confined heat spreaders having a uniform or an asymmetric discrete heating load with different cooling methods have been performed. Two types of cooling method such as natural cooling and fan jet impingement cooling are employed in the present study. The relevant parameters influencing heat transfer performance on the studies with natural cooling and fan jet impingement cooling are: (1) natural cooling －Grashof number(GrT or GrH), ratio of confinement spacing to characteristic length (S/LH), ratio of heater area to spreader area (r), heater installation location and type of spreader material. The ranges of these parameters studied are GrH = 8.15×106 – 3.75×107; S/LH=0.077 and ; and r=0.217-1. (2) fan jet impingement cooling － Grashof number (GrT or GrH), Reynolds number (Re), ratio of heater area to spreader area (r), heater installation location and type of spreader material. The ranges of these parameters explored are GrH = 1.39×108 – 3.42×108, Re = 4344-7125 and r = 0.217-1. Their effects on heat transfer behavior with natural cooling and with fan jet impingement cooling have been systematically explored.
For precisely evaluating the effective thermal conductivities of various types of heat spreader, an effective and time-saving semi-empirical method, by using both a numerical thermal-network method and an experimental measurement together with a statistical least-square technique, has been proposed. Three new correlations of effective thermal conductivity in terms of temperature for eleven types of heat spreader have been presented.
For investigating the local and average thermal behavior of heat spreaders having a uniform or an asymmetric discrete heating load with natural cooling or with fan jet impingement cooling, an effective semi-empirical method integrated by the experimental measurement and thermal network analysis has been successfully developed. The parametric studies on relevant thermal performances are conducted; and the results and conclusions are also verified by a statistical sensitivity analysis of ANOVA F-test. In addition, five new generalized correlations of average Nusselt numbers for heat spreaders having a uniform or an asymmetric discrete heating load with natural cooling or with fan jet impingement cooling are presented with satisfactory agreements. Furthermore, two new composite correlations of average Nusselt number for a confined heat spreader having an asymmetric discrete heating load in mixed convection due to fan jet impingement and buoyancy in terms of r and GrT/Re2 (or GrH/Re2.5) have been proposed; and another new correlations for evaluating external and total thermal resistances for all the cases have also been presented.
Finally, the empirical response formulas with a quadratic RSM model for evaluating all relevant thermal performances such as average Nusselt number and total thermal resistance for various types of heat spreader having a uniform or an asymmetric heating load with either natural cooling or fan jet impingement cooling have been presented. Comparisons of the predictions with such empirical response formulas with the experimental data and relevant experimental correlations are made. The maximum and average deviations compared with the experimental data are less than 10.84% and 5.89%, respectively; those compared with the experimental correlations are less than 9.98% and 5.73%, respectively.

並列關鍵字

Heat Spreader ； Electronics Cooling ； Natural Cooling ； Fan Jet Impingement ； Thermal Conductivity ； Re

參考文獻

[1] Schneider, G. E., Yovanovich, M. M., and Cane, R. L. D., 1980, “Thermal Resistance of a Convectively Cooled Plate with Non-uniform Applied Flux,” Journal of Spacecraft and Rockets, 17(4), pp.372-376.

[2] Bobco, R. P., and Starkovs, R. P., 1985, “Rectangular Thermal Doublers of Unoform Thickness,” AIAA Journal, 23(12), pp.1970-1977.

[3] Venkataraman, N. S., Cardoso, H. P., and Filho, O. B. D., 1982, “Thermal Resistance of Convectively Cooled Plates with Applied Heat Flux and Internal Heat Generation,” Acta Astronautica, 9(10), pp.613-620.

[5] Welch, J. W., and Lam, T. T., 1997, “Minimal Thermal Resistance of a Convectively Cooled Isotropic Plate with Non-uniform Heat Flux,” Advances in Electronic Packaging, 2, pp.1933-1939.

[8] Yan, L., Haji-Sheikh, A., and Beck, J. V., 1993, “Thermal Characteristics of Two-Layered Bodies with Embedded Thin-Film Heat Source,” ASME Journal of Electronic Packaging, 115(3), pp.276-283.