以熱電發電器進行廢熱(冷)回收之分析研究

An Investigation of the Waste Heat/Cold Recovery in Use of Thermoelectric Power Generators

翁健洲(Chien-Chou Weng)

指導教授 ： 黃美嬌

國立臺灣大學/工學院/機械工程學研究所/博士(2015年)

https://doi.org/10.6342/NTU.2015.00611

摘要

本論文針對三種不同的熱電發電器應用實例，分別提出適合的分析方法來進行探究，並對回收系統建立熱阻模型，對模型近似解的準確性進行探討。
第一個應用實例是汽車廢熱回收。我們建立一個從汽車排氣管抽取廢熱的熱電發電器廢熱回收系統數值模型，並以套裝軟體FLUENT模擬研究熱電偶數量及熱電偶覆蓋率對此廢熱回收系統效率的影響。我們發現較多的熱電偶並不一定能產生較多的電力，因為貼附在下游熱電偶的熱沉會從上游的熱交換器抽取熱能，導致上游熱電偶的發電力下降。此外，當熱電偶數量固定的時候，下游未貼附熱電偶的熱交換器有助於提升發電力。這是因為未貼附熱電偶的區域有較高的溫度，熱會藉由牆壁從下游傳至上游的熱電偶，進而增加上游熱電偶的溫差及發電功率。
第二個應用實例是從桌燈廢熱回收。我們將熱電模組與桌燈結合並使用自然對流鰭片進行散熱，同時進行模擬及實驗研究。在模擬中使用Compact heat sink 模型取代真實鰭片的模擬。為了節省計算量且正確評估發電力功率，我們先模擬開迴路系統得開路溫差，再提出熱電發電器熱阻模型預測閉迴路發電功率；另一方面為求驗證，我們也直接進行同時考慮熱電效應之閉迴路系統模擬。結果顯示，前者預測出來的結果與後者的結果相近，但兩者與量測結果相比卻有較高的最大發電力率及較低的最佳匹配負載。我們將此歸因於電路上寄生電阻的影響，而這同時也是為何將熱電模組並聯所得到的電力會比串聯低的原因。最後，從分析結果可知，燈管與熱電模組熱端間的高熱阻是低廢熱回收率的主要原因之一。
最後一個應用實例是低溫氮系統廢冷回收。針對低溫氮排放系統，我們研究如何利用此低溫氮與室溫之溫差來發電。我們先使用熱電發電器熱阻模型來預測，再以實驗來進行驗證。模型預測及實驗量測結果皆顯示，使用堆疊式熱電模組可利用更大的溫差且能獲得較多的電力；當使用四片雙層熱電模組時，我們可以從3.6g/s之低溫氮中得到0.93W的電力。然而，實驗得到的電力卻低於模型預測的結果，未考慮系統表面結冰之影響是造成預測失準的原因。另一方面，我們也探討了熱電模組數量對發電力的影響，且在實驗中發現了最佳熱電模組數量的存在。
最後，熱電應用常使用的熱電發電器數學模型常需要搭配數值方法才能求解；為求方便，我們推導此模型的最大發電力及最佳匹配負載的的近似解，並探討它們的適用性及準確性。結果指出，此近似解對具有較高冷熱端熱阻或相近冷熱端熱阻的熱電發電器能源回收系統，以及高操作溫度和低溫差的應用有較好的準確性。

關鍵字

熱電發電器 ； 廢熱回收 ； 廢冷回收 ； 熱阻模型 ； 汽車 ； 桌燈

並列摘要

In this thesis, the proper analysis methods were proposed for three thermoelectric power generator (TEG) applications; relevant thermal-resistance model was carefully built and the accuracy of the approximate solution for this model was discussed.
The first TEG application is associated with the automotive waste heat recovery. The influences of the number and the coverage rate on the surface of the heat-exchanger of the thermoelectric couples (TE couples) were explored via FLUENT simulation. It was found that implementing more TE couples does not necessarily generate more power because the heat sinks attached to the downstream TE couples loot heat from the upstream hotter wall, resulting in a performance degradation of the upstream TE couples. Furthermore, for a given number of TE couples, the performance is better when a portion of the heat exchanger is uncovered with TE couples at the downstream side, because the uncovered part becomes hotter and consequently transfers heat from the downstream wall to the upstream TE couples.
The second application is to recover waste heat from a table lamp. The table lamp integrated with TEG modules and cooled by a natural convection heat sink was investigated experimentally as well as numerically. In the simulation, the heat sink was not truly simulated but modeled by the so-called compact heat sink model. To reduce the computational amount and still accurately estimate the power generation, the open-circuit system was first simulated and a TEG thermal-resistance model was built for predicting the power generation rate based on the open-circuit temperature difference. For verifying, closed-circuit simulation which properly takes Peltier and Joule’s heats into consideration was also performed. The investigation shows the prediction from the TEG thermal-resistance model is in a good agreement with the closed-circuit simulation result; the maximum power generation rate is slightly larger and the optimal electric load is slightly smaller than the experimental measurements however. We attribute these differences to the effect of the parasitic electric resistances in the system. A same reason is employed to explain a lower maximum power generation rate observed in the parallel network than in the series network. Finally, it was concluded that the low hot-side thermal conductance is the main reason for the low power generation efficiency.
The final application is about the waste cold recovery. A TEG waste-cold recovery system was studied for the cryogenic-nitrogen exhaust system. A TEG thermal-resistance model was constructed for predicting the power generation rate and experiments were performed for verification. Both the model analysis and the experiment show that using cascade TEG modules can access more temperature difference and thus generate more power; a power generation rate as high as 0.93W was obtained by the present system when four two-layer cascade TEG modules were employed at a mass flow rate of cryogenic nitrogen of 3.6 g/s. However, the measured power generation rates are less than the predictions. The ice frozen over the thermal spreader, which is not taken into consideration in the model, must take the responsibility. On the other hand, the influence of the TEG-module number was also explored and the existence of the optimal number of TEG modules was observed in the experiment.
Finally, the analytical solution of a common TEG thermal-resistance model is usually unavailable. For convenience, the approximate (empirical) solutions of the maximum power generation rate and the optimal electric load for the TEG thermal-resistance model were derived and investigated in various situations. The investigation shows that the approximate solution has a better accuracy when in the TEG recovery system the hot-side and cold-side thermal conductances are low or nearly equal or when the system is operated at high operating temperatures or with a low temperature difference.

並列關鍵字

Thermoelectric Power Generator (TEG) ； Waste Heat Recovery ； Waste Cold Recovery ； Thermal-Resistance Model ； Automobile ； Table Lamp

參考文獻

50.S. Lal, S. Loo, D. Y. Chung, T. Kyratsi, M. G. Kanatzidis, C. Cauchy, and T. P. Hogan, Thermoelectric module for low temperature applications, Materials Research Society Symposia Proceedings, vol. 691, pp. G6.2.1-G6.2.9, 2002.

51.S. R. Harutyunyan, V. H. Vardanyan, A. S. Kuzanyan, V. R. Nikoghosyan, S. Kunii, K. Winzer, K. S. Wood, and A. M. Gulian, Laser-powered thermoelectric generators operating at cryogenic temperatures, Applied Physics Letters, vol. 87, pp. 194114(1-3), 2005.

67.M. C. Lin, C. Wang, M. H. Tsai, F. T. Chung, M. S. Yeh, T. T. Yang, M. H. Chang, and L. H. Chang, Suppression of thermal oscillation induced by liquid-nitrogen in the CESR-type SRF module, IEEE Transactions on Applied Superconductivity, vol. 19, pp. 1427-1431, 2009.

84.C. S. Chang, J. K. Liu, C. H. Wang, and H. S. Pei, LED lamp cooling apparatus with pulsating heat pipe, US patent 7547124B2, 2009.

83.H. Ye, S. Koh, H. van Zeijl, A. W. J. Gielen, and G. Q. Zhang, A review of passive thermal management of LED module, Journal of Semiconductors, vol. 32, pp. 014008(1-4), 2011.