Printer Friendly

Coefficient of performance improvement in small thermoelectric refrigerators.


Thermoelectric coolers operate based on the Peltier effect which was discovered in 1834 (Goldsmid 2010). Since then, the subject has received a fair amount of attention by researchers. Peltier coolers are comprised of small-size thermoelectric modules. When a current is applied to the thermoelectric circuit, heat is removed from the conditioned space and rejected to the other end of the module. No moving parts exist anywhere in the system to generate operating noise which in turn makes the system appropriate for certain applications. Another advantage of Peltier coolers is in their lack of need for maintenance. The main disadvantage is their quite low coefficient of performance (COP) which limits their use to small- scale applications only. It is established knowledge that the COP of a thermoelectric cooler highly depends on the heat rejection rate of the system to the high-temperature sink (Khonsue 2012). Astrain et al. (2003) studied ways of increasing the COP of a thermoelectric refrigerator and suggested that by using phase change heat transfer in thermosyphon mode, the COP can be improved by as much as 32% compared to the system that utilizes commercial fin dissipator. Goldsmid (1999) reviewed and commented regarding various ways the performance of thermoelectric refrigerators can be improved. The efficiency of a thermoelectric cooling system can be improved either via improving the material characteristics of the module, or improving the heat transfer characteristics of the whole system which includes heat removal and heat rejection characteristics of the cooler. Forced air has been used on the heat rejection side due to convenience and portability (Sofrata 1996, Yu et al. 2012). For better heat rejection from the hot side, forced convection by water flow in a heat exchanger attached to the hot side has also been used (Lertsatitthanakom et al. 2008). The method tested in the current research makes use of an open stream of water as the cooling fluid on the hot side of the thermoelectric refrigerator. The use of an open stream of water adds latent heat transfer through evaporation to the other heat transfer mechanisms. Since the heat transfer coefficient during evaporation is considerably higher than that of sensible heat transfer alone, the COP of the thermoelectric refrigeration system is expected to improve. Attey (1998) proposed using a pumped water cooling unit on both sides of a thermoelectric refrigerator and also discussed that the COP of the whole system depends highly on its ability to reject heat. The second issue the current study will address is the transient behavior of a thermoelectric refrigerator. Few studies have been published on this topic to this date. Ahiska et al (2012) proposed a new method for measuring the time constant of real thermoelectric modules. Naji et al. (2003) studied the transient behavior of thermoelectric generators and refrigerators and presented plots of the COP variations with time. The third issue that the current study will tackle is the load dependency of the performance of a thermoelectric refrigerator. A specific literature search on this issue did not produce any result.


Four temperature measurements were taken in order to calculate the COP at various times during the experiment. The time interval used during the experiment was 5 seconds, and the temperatures were recorded with the use of a custom made LabVIEW[TM] module. The ambient air temperature, Tout, along with cold side heat sink surface temperature, [T.sub.Suface], refrigerated space temperature, [], and can temperature, [T.sub.can], were used to calculate the COP. In order to determine the overall and transient COP, the properties of the materials used during the experiment were obtained at the ambient temperature. A thermal conductivity of 0.1 W/m-K (0.0578 Btu/h-ft-R) was used for the plastic wall of the container. The mass of each soda can was taken as 0.375 kg (0.826 lbm). The total thermal resistance was calculated using the electrical resistance analogy to be 18.01 W/K (34.15 Btu/h-R). In order to calculate the transient behavior of the thermoelectric modules and determine various heat transfer values, it was necessary to determine the temperature decrease increment over the length of the time step as shown in Equation (1):

[DELTA]T = [T.sub.n-5s] - [T.sub.n] (1)

Using this approach, all heat transfer and COP calculations started at zero for time T = 0 s. After 5 seconds, new temperature measurements were taken, and heat transfer and COP were calculated. From this, [Q.sub.gain], [Q.sub.can], [Q.sub.air], and [] were calculated for one time interval by using Equations 2-5, respectively:

[Q.sub.gain,n] = [[T.sub.out,n] - [,n]/[summation][]] (2)

[Q.sub.can,n] = [m.sub.can][c.sub.p,can] ([T.sub.can,n-5s] - [T.sub.can,n]) (3)

[Q.sub.air,n] = [m.sub.air][c.sub.p,air] ([,n-5s] - [,n]) (4)

[,n] = [,n-5s] + [Q.sub.gain,n] + [Q.sub.can,n] + [Q.sub.air,n] (5)

The total work was then determined form the steady state power input times the length of the experiment at time n. The work input was consisted of the electrical power input to the modules, and the power input to the fan for the forced air case, or the power required to generate the water flow for the water cooling case. The DC electric power input to the fan and the modules were calculated by multiplying the supplied voltage by the measured current. The instantaneous COP was calculated as in Equation (6):

[COP.sub.inst] = [,n]/[,n] (6)


A portable beverage container 10.8x7.9x7 in (27x20x18 cm) inside space volume was retrofitted into a thermoelectric refrigeration system. The thermoelectric system consisted of 4 individual thermoelectric modules 1.57"x1.57" (40x40 mm) each. The modules were arranged in a 2 by 2 square matrix form with aluminum fins on the cold side. For the case of forced air cooling, the hot side was also equipped with aluminum fins for increased heat transfer area. However, for the case of water cooling, the hot side was equipped with a flat aluminum plate oriented vertically on which the cooling water can form a thin film and flow downward due to gravity. Two adjustable voltage power supplies were used to power the four modules. The supplied voltage to the modules was fixed for the duration of the experiment. The voltage varied from 4 to 12 V for different test configurations. In order to decrease the contact thermal resistance between the ceramic plates of the modules and the base plate of the aluminum fins, a thin layer of thermal paste was applied. The modules were sandwiched between the fins with bolts while the tightening torque was adjusted to ensure good thermal contact between the surfaces. The cooling package consisting of the four modules the inner fins, the outer fins (or the flat aluminum plate) were installed on the larger wall of the container through a rectangular hole that was cut into the wall. In order to model thermal load, 1 to 6 unopened regular-size soda cans were placed inside the container, depending on the test configuration. A K-type surface thermocouple was attached to the wall of one of the cans to measure its temperature variations. Three other thermocouples were used to measure the air temperature inside the container, the ambient air temperature, and the surface temperature of the fins located inside the cooled space. Figure 1 shows the experimental setup.


The four thermocouples were connected to a data acquisition card and a LabVIEW[TM] program was compiled to monitor and record the data at 5-s intervals for a period of approximately 2.5 hours which assured that a steady state was prevailed. The relatively small time step size was chosen because the transient behavior of the system was of interest. A total of 17 configurations were tested as shown in Table 1. For forced air heat rejection a 12-V brushless DC fan was used. For water flow heat rejection, an array of four 1/4" copper tubes sprayed water on the aluminum played at a constant rate to ensure a thin film of water existed on the plate and no dry spots existed. A pump was used to circulate water from the drip pan back to the spray system. Before each test, the cans were placed in the container with the container lid open in the lab for at least 24 hours so that every component of the system was at thermal equilibrium with the lab ambient air. The fan or the water spray pump was turned on. The container lid was closed. The thermoelectric modules were powered on and immediately the data acquisition program was started.


Regardless of the test configuration, the COP showed a similar transient behavior as depicted in Figure 2 for an air- cooled single-can test at a supply voltage of 6 V. During the first few seconds there was a spike in the instantaneous value of COP followed by a drop and then gradual increase until the COP of the system reached its steady value after about 1 hour for the case shown in Figure 2. In a numerical study performed by Naji et al. (2003) a similar transient behavior was observed and reported.

The effect of the supply voltage on the steady-state COP of the refrigerator is depicted in Figure 3. As can be seen from the Figure, for the case of a single can as the thermal load, increasing the supply voltage will decrease the COP of the refrigerator. The reason is due to the size of the thermal load. It seems that the lowest power applied to the modules can overcome the small heat load and further increasing the applied power will have no effect other than heat generation in the module hence increased inefficiency of the refrigerator. Furthermore, it can be seen from Figure 3 that in the case of single can tests, water cooling has an insignificant effect on the COP except for the lowest power rating.



In a refrigeration system, the achievable cold surface temperature is also an important characteristic of the system. The effect of supply power on the minimum temperature of the inner surface of the cold-side fins is presented in Figure 4. It is evident from the Figure that the surface temperature has a minimum which is different for each cooling mode. In the case of forced air heat rejection, the minimum temperature occurs at 8 V which is higher than the case of water-stream heat rejection. This can be attributed to the fact that with water cooling on the heat rejection side, better heat rejection can be achieved at lower power ratings.


Considering the results shown in Figurte 4, it was determined that the optimum supply voltage is in the vicinity of 4 to 6 V. Figure 5 was then plotted ffrom the results of the tests at a supply voltage of 6 V for varying thermal loads ranging from 1 to 6 cans. The plots show that increasing the thermal load will increase the COP of the refrigerator while heat rejection with water assures better performance of the system.


Finally, an attempt was made to clarify the load dependency of the COP of the system with water heat rejection. The result is presented in Table 2. The COP of the water-cooled refrigerated decreased from 1.23 to 0.6 as the supply voltage increased from 4 to 6 V, while the COP increased again when the thermal load was increased to 6 cans instead of 4 cans. Attey (1998) reported an improved COP of 0.82 for the case where forced-circulation water cooling was used on both sides of the refrigerator, over the traditional finned forced-air heat rejection. The range of the COP values attained in the current study is in good agreement with Attey (1998). The results reported by Lertsatitthanakom et al. (2008) is also in the same range as they reported a COP of 0.47 to 0.51 for water-cooled heat rejection system.


A small beverage cooler equipped with thermoelectric cooling modules was investigated with regards to its coefficient of performance. Several configurations were tested at different supply voltage rates, number of cans, at two different heat rejection modes namely, forced air, and liquid cooling using an open stream of water. It was found that at constant thermal load, increasing the supply voltage to the thermoelectric refrigerator will decrease the coefficient of performance of the refrigerator. This is true regardless of the method of heat rejection. The lowest cold surface temperature corresponds to a certain supply voltage which is a function of the thermal load. The water-cooled heat rejection system provided better COP values compared to the traditional forced-air systems. The maximum steady-state COP of 1.23 was achieved at a supply voltage of 4 V when there are 4 soda cans in the refrigerator.


COP Coefficient of performance

[c.sub.p]--Specific heat

GPH Gallons per hour

k--Thermal conductivity


n--Current time

[Q.sub.air]--Heat removed from the air inside the container

[Q.sub.can]--Heat removed from the cans

[Q.sub.gain]--Heat transfer to the container through the walls

[]--The sum of heat removed from all the components of the system for the time period n

[]--Temperature inside the cooler

[T.sub.out]--Air temperature outside the cooler

[summation][]--Total thermal resistance of the cooler


Ahiska, R., Dislitas, S., Omer, G. 2012. A new method and computer-controlled system for measuring the time constant of real thermoelectric modules. Energy Conversion and Management, 53: 314-321.

Astrain, D., Vian, J.G., Dominguez, M. 2003. Increase of COP in the thermoelectric refrigeration by the optimization of heat dissipation. Applied Thermal Engineering, 23: 2183-2200.

Attey, G.S. 1998. Enhanced thermoelectric refrigeration system COP through low thermal impedance liquid heat transfer system. 17th International Conference on Thermoelectrics, Nagoya, Japan.

Goldsmid, H.J. 1999. Possibilities of improvement in thermoelectric refrigeration. 18th International Conference on Thermoelectrics, Baltimore, MD.

Goldsmid, H.J. 2010. Introduction to Thermoelectricity. Springer-Verlag, Berlin.

Khonsue, O. 2012. Experimental on the liquid cooling with thermoelectric for personal computer. Heat and Mass Transfer, 48: 1767-1771.

Lertsatitthanakom, C., Tipsaenprom, W., Srisuwan, W., Atthajariyakul, S. 2008. Study on the cooling performance and thermal comfort of a thermoelectric ceiling cooling panel syatem. Indoor Built Environment, 17: 525-534.

Naji, M., Alata, M., Al-Nimr, M.A. 2003. Transient behaviour of a thermoelectric device. Proceedings of the Institution of Mechanical Engineers, Part A: Journal of Power and Energy, 217: 615-621.

Sofrata, H. 1996. Heat rejection alternatives for thermoelectric refrigerators, Energy Conversion and Management, 37: 269-280.

Yu, H., Chen, Y., Yu, L., Lu, Y., Zhang, D. 2012. The design of enhancing thermoelectric cooler system based on forced air cooling. 2012 International Conference on Systems and Informatics, Yantai, China.

Hessam Taherian, PhD


William L. Adams Jr.

Author A is an Assistant Professor in the Department of Mechanical Engineering, the University of Alabama at Birmingham, Birmingham, AL.

Author B is a Mechanical Engineer at Southern Company, Birmingham, AL.
Table 1. Test Configurations

Configuration     Supply      Number    Cooling medium
number          voltage (V)   of cans

1                    4           1            air
2                    6           1            air
3                    8           1            air
4                   10           1            air
5                   12           1            air
6                    4           1      water at 20 GPH
7                    6           1      water at 20 GPH
8                    8           1      water at 20 GPH
9                   10           1      water at 20 GPH
10                  12           1      water at 20 GPH
11                   6           1      water at 40 GPH
12                   4           4            air
13                   6           4            air
14                   4           4      water at 20 GPH
15                   6           4      water at 20 GPH
16                   6           6            air
17                   6           6      water at 20 GPH

Table 2. The COP of the Refrigerator at
Different Configurations

Supply voltage   Number of cans   COP

4                      4          1.23
6                      4          0.6
6                      6          0.75
COPYRIGHT 2013 American Society of Heating, Refrigerating, and Air-Conditioning Engineers, Inc.
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2013 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Author:Taherian, Hessam; Adams, William L., Jr.
Publication:ASHRAE Transactions
Article Type:Report
Geographic Code:1USA
Date:Jul 1, 2013
Previous Article:Experimental and analytical study of the transient process of infiltration/exfiltration in walk-in coolers.
Next Article:Modeling and testing of an R-23/R-134a mixed refrigerant system for low temperature refrigeration.

Terms of use | Privacy policy | Copyright © 2020 Farlex, Inc. | Feedback | For webmasters