Development of a Novel Eco-Friendly Thermoelectric Air-Conditioning System

Tiasha Joardar © Tiasha Joardar ©

5124 Water Haven Lane Plano, TX 75093

Acknowledgements First I would like to thank Ms. Deanna Shea for her help and guidance with the entire science fair process. I would like to thank my father for purchasing all the items required for this project, for his help with constructing some of the apparatus, for allowing me to use the garage and laundry room, for explaining several electrical and thermal concepts, and for proof reading and help with formatting this report. All experimental work was done in our home in Plano, Texas.

2

Abstract A novel thermoelectric air conditioning system is reported that can remove heat without requiring the use of energy intensive compressors or environmentally harmful chlorofluorocarbons (CFCs). Thermoelectric cooling units (TECs) operate on the Peltier Effect and use electricity to pump heat. A TEC based air conditioning unit is demonstrated which consumes significantly less energy than a vapor compression unit of similar size when used to maintain a scaled model home at 8°C below outside temperature under a heat load of 150 W/m2. In addition, a method for optimization of the currents in each stage of a two-stage cascaded thermoelectric cooling system is developed theoretically and confirmed experimentally for the first time. Since commercially available TECs have low Coefficients of Performance (COP), several innovative steps are taken to overcome this limitation. First, it is recognized from theoretical analysis that the energy used by TECs decreases exponentially with heat load. By using a "divide and conquer" approach where the heat load is shared by multiple TEC modules in a room, the overall energy efficiency of the system is greatly improved. Second, it is also noted from theoretical analysis that the COP of thermoelectric cooling systems improves significantly if its hot side temperature is kept as low as possible. This is accomplished by using a watercooled heat sink which efficiently removes heat from the module, thereby keeping its hot side at a temperature no higher than the outside air temperature. The active heat sink is powered by solar cells, which have no operating cost. Thirdly, theoretical analysis is used to show that energy efficiency of a thermoelectric system can be increased further by using a cascaded dualstage system if the current in each stage is optimized. Several experiments are conducted the results from which support the theoretical findings. A scaled model home is constructed for experiments and fitted with a traditional vapor compression air conditioner unit on one side and a thermoelectric unit on the other. The walls are insulated using Styrofoam insulation. A resistor bank driven by a variable power supply is placed inside the model home. This serves as a controllable heat load. Energy used by the vapor compression and thermoelectric systems as a function of home indoor temperatures is investigated. It is found that under identical heat load and temperature conditions, it is possible to obtain up to 40% savings in energy usage using a dual-stage cascaded thermoelectric system by optimizing the current driven through the TECs. The findings of this project provide an opportunity to reduce energy usage in homes and buildings greatly and open up the possibility of providing air conditioning in homes in less developed areas. Additional efficiency gains are possible by using real time adaptive control on the current flowing through the units. This aspect will be studied in future.

3

Table of Contents 1.

Introduction ………………………………………………………………………………… 7 1.1 Problem Statement………………………………………………………………........

8

1.2 Fundamentals of Solid State Thermoelectric Cooling …………………………….....

8

1.3 Review of Prior Work…………………………………………………….………..... 11

2.

Theory…………………………………………………..……………………………………12 2.1 Single Stage Thermoelectric Cooling ………………….................................................14 2.2 Two-stage Cascaded Thermoelectric Cooling.……………………………………..…..14 2.3 Design Innovations Based on Theoretical Analysis ………………………..………… 18 2.4 Theoretical Calculations and Hypothesis ……………………………… ……………19

3.

Experimental Setup and Procedures …...……………………………………………….. 20 3.1 Variables……………………………………………………..……………………….. 20 3.2 Apparatus Used………………………………………….……………………………. 20 3.3 Measurement Procedures……………………................................................................ 23

4.

Experimental Results……………………………………………………………………… 23 4.1 Energy Consumption of Vapor Compression System………………………………... 23 4.2 Energy Consumption of Single Stage Thermoelectric Systems……

…………......25

4.3 Energy Consumption of Two-Stage Cascaded Thermoelectric Systems………………28

5.

Discussion………………...………………………………………………………………….29

6.

System Cost Comparisons……………….………………………………………………….30

7.

Future Work………………………………………………………………………………... 32

8.

Conclusions…………………………………………………………………………………. 32

9.

References ………………………………………………………………………………….. 33

4

List of Figures Fig. 1

Fig. 2

Simplified diagram of a TEC showing (a) a single junction and (b) an assembled view of a TEC module..

9

Theoretical dependence of optimum COP of a TEC module on its hot side temperature.

11

Theoretically calculated plots of indoor temperature, cooling power, and COP as a function of current passed through a single-stage thermoelectric air conditioning system.

14

Fig. 4

Diagram showing a two-stage thermoelectric cooling system.

14

Fig. 5

Theoretically calculated variation of cold side temperature TC with currents in the upper and lower TECs. (a) is a 3-d surface plot and (b) is a contour plot of the data. 16

Fig. 6

3-d surface plot of COP versus IC and IH for the two-stage system described in this section.

17

Theoretically calculated TC as a function of input power for TH = 37°C and Qgen = 4W.

18

.Diagram showing how an active water cooled heat sink is used to control the temperature of the hot side of the TEC.

19

Diagram showing equipment used to monitor energy usage by different air conditioning units.

21

Fig. 10 (a) Diagram of the single-stage, quad-module thermoelectric cooling system, (b) diagram of the two-stage cascaded thermoelectric system, and (c) photograph of fully assembled cascaded thermoelectric cooling unit used in this project.

22

Fig. 11 Energy consumed as a function of time by the vapor compression air conditioner over a 20 minute duration for three different T settings, each with a heat load of 24W.

24

Fig. 12 Temperature inside home as a function of time for the vapor compression air conditioner for different T values.

24

Fig. 13 Summary of energy used over one hour by the vapor compression system as a function of indoor temperature setting.

25

Fig. 3

Fig. 7

Fig. 8

Fig. 9

5

Fig. 14 Energy consumed as a function of time by the single-stage, single-module, thermoelectric air conditioner over a 20 minute duration for three different T settings, each with a heat load of 24W.

26

Fig. 15 Temperature inside home as a function of time for the single-stage, single-module, thermoelectric air conditioner over a 20 minute duration for different T settings.

26

Fig. 16 Measured variation of indoor temperature versus electrical energy used over one hour by a single-stage single-module thermoelectric system.

27

Fig. 17 Measured variation of indoor temperature (blue) and energy consumed in one hour versus electrical current input for a single-stage quad-module thermoelectric system.

28

Fig. 18 Measured variation of indoor temperature versus (a) electrical current IH, and (b) versus hourly energy use for a dual-stage cascaded thermoelectric system.

29

Fig. 19 Comparison of energy usage of vapor compression and thermoelectric air conditioning systems as a function of T with a heat load of 24W.

30

Fig. 20 Dependence of purchase price of vapor compression air conditioners on cooling power found from a market survey (dots). Dashed line represents estimated dependence of cost per cooling watt for TEC systems based on current market prices.

31

6

1. Introduction From an environmental sustainability viewpoint it has become critical that global energy use be brought under control. One of the largest sources of energy consumption is air conditioning systems. According to the American Council for an Energy-Efficient Economy, home air conditioning consumes 5% of the total electricity produced in the US [1, 2]. Energy use from residential air conditioning is also increasing rapidly in Brazil, China, and India as their economies expand. Reducing this energy usage will not only provide economic relief to residents but, on a larger scale, it will help protect our environment by lowering pollution from fossil fuel based power plants. Further, the refrigerants used in today's air conditioners are hydrochlorofluorocarbons (HCFCs) which are known to be harmful to the environment. Therefore, it is worth investigating alternative energy efficient and eco-friendly ways of heat pumping. One possibility is the use of the thermoelectric effect. Thermoelectric cooling modules (TECs) pump heat from a high temperature region to a low temperature region by passing an electric current through a thermoelectric material. Since this does not require compressors or refrigerants, and since its cooling power can be electrically controlled, it may be possible to develop thermoelectric air conditioners without the drawbacks of traditional vapor compression units. Thermoelectric systems may be most suitable as small room air conditioners that are popular in Asia, Europe, and South America. In this project, a thermoelectric based room air conditioning system is developed. Since the practical use of such a system will depend entirely on how well it compares against a vapor compression air conditioner, this aspect will be the main theme of this project. The basis for comparing a thermoelectric and a vapor compression system will be the energy consumed by each system operating under identical environmental conditions. Energy consumption of an air conditioning system is normally stated in terms of its Coefficient of Performance (COP) which is defined as the ratio of the amount of heat energy pumped by the system to the amount of electric energy required to do so. The higher the COP of a system, the lower its energy consumption under constant heat loads. COP 

Heat _ Pumped . Electricit y _ Used

(1)

7

1.1 Problem Statement The question addressed in this project can be stated quantitatively as follows. Is it possible to demonstrate a thermoelectric air conditioning system, using commercially available parts, that consumes 20% less energy when compared to a similarly sized conventional vapor compression air conditioning system while maintaining a scaled model home at 8°C below outside air temperature, with an internal heat load of 150 W/m2? The choice of an 8°C temperature difference is explained as follows. The industry standard conditions for stating the COP or energy efficiency ratings (EER) of air conditioners require an indoor temperature of 80°F (27°C) and outside temperature of 95°F (35°C), which is a difference of about 8°C. Regarding the target of 20% reduction in energy use, it is noted that if air conditioners would consume about 30% less energy than they do today, it would become practical to operate them off solar panels, a green source of energy [3]. If the objective of this project can be met, then the results could be a useful step towards controlling and sustaining the rapid growth in worldwide energy usage. Lastly, regarding the use of 150 W/m2 as internal heat load, the reasons for this are explained in Section 2.4. 1.2 Fundamentals of Solid State Thermoelectric Cooling Thermoelectric cooling is based on the Peltier effect. This effect, discovered by Jean-Charles Peltier in 1834, states that when an electric current is driven through a junction between two different conductors heat is absorbed or generated, depending on the direction of current flow. The amount of heat transferred depends on the materials used to form the junction [4, 5]. Extensive research has been conducted in the last few decades to develop materials that exhibit thermoelectric properties strong enough to be of practical value [6]. Today's commercially available thermoelectric modules use tightly arranged pellets of bismuth telluride, and are capable of pumping almost 10 watts/cm2. Fig. 1 shows a simplified diagram of such a TEC module. The figure on the left, Fig. 1(a), represents a thermoelectric single junction and shows the flow of electric current and heat. The figure on the right, Fig. 1(b), shows how several of these thermoelectric junctions are assembled to make a TEC module.

8

HEAT

Conductor Semiconductor pellet

p

n

Heat sink HEAT

(a)

(b) Fig. 1. Simplified diagram of a TEC showing (a) a single junction and (b) an assembled view of a TEC module.

The Coefficient of Performance (COP) of TEC modules depends on a "figure of merit" of the thermoelectric material used. This figure is denoted as ZT and is given as [6]: ZT 

  S 2 T K

,

(2)

where  is the electrical conductivity of the thermoelectric material, S is its Seebeck coefficient,

 is its thermal conductivity, and T is the average temperature of the TEC. For commonly used bismuth antimony telluride alloys the highest ZT is about 1. New superlattice materials can have

9

ZTs as high as 3.5 but these are not commercially available [7]. The dependence of maximum achievable COP of a TEC module on ZT is given by [8]:

COPmax

TH   TC  ZT  1  TC   ZT  1  1  TH  TC  

  ,  

(3)

where TH is the hot side temperature and TC is the cold side temperature in Kelvin (degrees Celcius + 273). Using equation (3) it is found that for a hot side temperature of 37°C (outside temperature) and a cold side temperature of 27°C (temperature in home) a maximum COP of about 5 is possible with bismuth telluride TECs (ZT = 1). In comparison, vapor compression air conditioners have COPs between 2 and 3. This may suggest that TEC based air conditioners will almost always consume less energy than vapor compression systems. Unfortunately, there are additional considerations that pose significant challenges when using TEC based systems. Fig. 2 shows a plot of COP as a function of the high side temperature TH of a TEC module, obtained using equation (3). The low side temperature TC is assumed to be 27°C and ZT is assumed to be 1. It can be seen that the COP drops very quickly as TH increases. For TH larger than about 45°C the COP drops below 2, the COP of many vapor compression systems of good quality. In theory, TH of a TEC based air conditioner can be as low as the outside air temperature. In reality, though, it could be several degrees higher. This is because TH is determined by how effectively the heat being pumped by the TEC, plus the heat it generates internally, can be carried away from its hot side. If the heat energy accumulates on the hot side of the TEC, TH will continue to increase and, as a result, its COP will drop. An effective heat sink is therefore critical to the success of a TEC based cooling system. Similarly, due to inefficiencies in heat transfer, TC needs to be a few degrees lower than the desired room temperature. This also leads to lower COP. Finally, the promisingly high COP values calculated from (3) occur at very low cooling powers, meaning a large number of TECs will be required to attain reasonable cooling.

10

10 9 8

COP

7 6 5 4 3 2 1 0 30

32.5

35

37.5

40

42.5

45

47.5

50

TEC Hot Side Temperature, TH (C)

Fig. 2. Theoretical dependence of optimum COP of a TEC module on its hot side temperature. The red dots are COPs reported by the US Department of Energy from tests on a high quality Mitsubishi vapor compression air conditioning system (model FE12NA). Both data sets are at cold side temperature of 27°C [9].

For purposes of comparison, the measured variation of COP with TH of a very high quality Mitsubishi vapor compression system is also shown in Fig. 2 (red dots). In theory, from Fig. 2 it can be assumed that it is feasible to design a thermoelectric air conditioning system that will outperform typical vapor compression systems of similar size on the basis of COP. 1.3 Review of Prior Work Since Goldsmid first demonstrated in 1954 the possibility of producing large Ts using semiconductor thermoelectrics like bismuth telluride, there have been many attempts at developing a commercial thermoelectric unit [10]. Unfortunately these systems were not commercially successful because they consumed large amounts of energy. The general opinion appears to be that until thermoelectric materials with high ZT can be developed, vapor compression systems will be more energy efficient and practical. However, it is also recognized that thermoelectric cooling is likely to be effective (i) under relatively low heat loads, (ii) where the heat load shows large variation, and (iii) where large temperature differences are not required [11]. Since room air conditioning satisfies these conditions there is reason to continue to pursue the development of a thermoelectric system for such use. A review of technical literature revealed only one prior technical report on thermoelectric air conditioning where the installation of a TEC based air conditioner in a studio size room with

11

a heat load of 960 watts is described [12]. A total of 48 TEC modules were required. However, no quantitative comparison with traditional vapor compression systems is provided. A major aspect of this work involves the use of two-stage cascaded thermoelectric modules. Cascaded thermoelectric cooling systems have been studied in the past, but all of the existing literature is theoretical with almost no experimental confirmation of results [13, 14]. In addition, the majority of existing work on cascaded thermoelectric cooling systems is centered on optimizing the geometry ratio of the different stages of the structure. In this project, for the first time, a method for optimization of the currents in each layer of a two-stage cascaded thermoelectric cooling system is developed theoretically and confirmed experimentally.

2. Theory 2.1 Single Stage Thermoelectric Cooling The cooling power at the cold end of a single stage thermoelectric cooling module is given by [10]:

1 L A  Q p  2 N  SITC  I 2  KT  , 2 A L 

(4)

where N is the number of thermoelectric junction pairs, S is the Seebeck coefficient of the thermoelectric material, I is the current through the thermoelectric module, TC is the cold side temperature,  is the resistivity of the thermoelectric material, A is the cross-sectional area of each thermoelectric pellet, K is the thermal conductivity of the thermoelectric material, and T is the temperature difference between the hot and cold sides of the TEC. The first term inside braces is the heat removed by the Peltier effect. The second and third terms represent, respectively, the heat injected into the cold side due to the electrical resistance of the TEC and heat conduction through the TEC from the hot side. These two latter effects tend to reduce the cooling power of the TEC. Under steady temperature conditions the above cooling power is balanced by the heat load, which consists of any internal sources of heat in the room being cooled, plus conduction of heat through the walls and ceiling of the room. This can be expressed as: Qload  Qgen  KW T ,

(5)

12

where Qgen represents heat generated by any internal or external heat sources such as appliances or sunlight coming in through windows. KW is the effective thermal conductivity of the walls and ceiling of the room. From equations (4) and (5) the temperature in the room being cooled can be solved as:

TC 

Qgen  N ( L / A) I 2  {KW  2 NK ( A / L)}TH 2 NSI  {K W  2 NK ( A / L)}

.

(6)

The total rate of energy consumption by the TEC (QTE) is the sum of Qp and the rate at which heat is generated within the TEC due to its electrical resistance. It is given by [10]:

 L  QTE  2 N  I 2  SIT  . A  

(7)

Lastly, the COP is given by the ratio of Qp/QTE. Using equations (4) and (7) the COP is expressed as follows:

COP 

1 2 L A I  KT 2 A L . L SIT  I 2 A

SITC 

(8)

Fig. 3 shows plots of TC, Qp, and COP as a function of current I, as computed from the equations above. The following values, which are typical of commercially available TECs, were used for the various quantities in the equations: S = 2x10-4 V/K, K = 3 W/mK, N = 127,  = 2.5x10-4 ohm-cm. TH was taken to be 37°C (i.e. 310K) and a 4W internal heat source was assumed. It can be seen that the temperature in the room decreases with I and reaches a minimum value at about I = 6.5 amps. At this current the cooling power of the system reaches its maximum value of about 20W. If I increases beyond this value the temperature rises as the TEC begins to lose its ability to pump heat due to increasing resistive heating. The COP is also a function of I, decreasing rapidly with increasing I. Therefore, from this theory, it is found that in order to maximize the efficiency of a thermoelectric air conditioner it should be operated at as low a current as possible. In other words, a high efficiency thermoelectric air conditioner should be designed to achieve a maximum cooling power much larger than the maximum heat load it is expected to handle. In the example shown here, in order to maintain an indoor temperature of 27°C (i.e. T = 10°C), a current of 1.3A is needed. At this current, the system's cooling power is

13

about 2W, which is roughly one tenth of its maximum possible value. The associated COP is about 2. 45 Qp Tc COP

35

8

30

6

25 20

4

COP

Room Temperature, Tc (C) & Cooling Power, Qp (W)

40

15 10

2

5 0

0 0

2

4

6

8

10

12

Current, I (Amp)

Fig. 3. Theoretically calculated plots of indoor temperature (blue line), cooling power (grey, dashed), and COP (black) as a function of current passed through a single-stage thermoelectric air conditioning system. An indoor heat load of 4W and an outdoor temperature of 37°C are assumed.

2.2 Two-stage Cascaded Thermoelectric Cooling In this section a two-stage thermoelectric cooling system is analyzed, and it is shown that such a system can, when operated optimally, can have a much higher COP than a single stage unit. Fig. 4 shows a diagram of a two stage cascaded thermoelectric cooling module. In this section the theoretical behavior of such a module is described.

Qwall

TC

Qgen

TH

IC IH Fig. 4 Diagram showing a two-stage thermoelectric cooling system.

The expressions for the various heat flow rates for the structure shown in Fig. 4 are as follows. Qpc is the heat pumped by the cold side of the upper TEC, Qrc is the heat exiting the

14

bottom surface of the upper TEC, Qph is the heat pumped by the lower TEC, and Qrh is the heat exiting the bottom surface of the lower TEC. 1 L KA   Q pc  2 N  SI C TC  I C2  (TM  TC  , 2 A L  

(9)

1 L KA   Qrc  2 N  SI C TM  I C2  (TM  TC  , 2 A L  

(10)

1 L KA  Q ph  2M  SI H TM  I H2  (TH  TM 2 A L 

 , 

(11)

1 L KA   Qrh  2M  SI H TH  I H2  (TH  TM  , 2 A L  

(12)

where N is the number of thermoelectric pairs in the upper TEC, M is the number of pairs in the lower TEC, TM is the temperature at the boundary between the upper and lower TECs, IC is the current through the upper TEC, and IH is the current through the lower TEC. Other symbols have the same meanings as described in the previous section. By equating Qrc to Qph it is possible to solve for TM. The result is:

L  KA  2 L MTH  NTC   NI C2  MI H 2 A A L  . TM  KA M  N  2S MI H  NI C   2 L

(13)

This expression for TM can be substituted in equation (9) resulting in the following expression for Qpc: Q pc  2 NI C TC  N

L A

I C2 





2 NKR MI H2  NI C2  4 N KA / L  MTH  NTC  . 2S MI H  NI C   2( KA / L)( M  N ) 2

(14)

Then, by equating this expression for Qph to Qload (equation 5), TC can be solved for. The result is:

TC 

N

L A

I C2  2 NKR MI H2  NI C2 / D  4MN KA / L  TH / D  K W TH  Qgen 2

2SNI C  4 N 2 K 2 / D  2 N KA / L   K W

.

(15)

Once TC and TM are found, Qpc and Qrh can be evaluated using equations (9) and (12). Finally the COP of this two-stage cascaded system can be found from: COP 

Q pc Qrh  Q pc

.

(16)

15

Fig. 5 shows a 3-d surface and a contour plot of TC as a function of the currents IC and IH as computed from equation (15). The numerical values of the various quantities were the same as in the previous section. Tc 50.0 45.0

40.0 35.0 30.0 25.0 20.0 15.0

0.0

2.75

4

9

7.75

Ih (A)

6.5

5.25

4

2.75

1.5

0.25

10.0 5.0

6.5

5.25

1.5

0.25

Ic (A)

0.25 1.25 2.25 3.25 4.25 5.25 6.25

0.25

1.5

2.75

Ih (A)

4

5.25

6.5

7.75

9

(a)

40.0-50.0 30.0-40.0

20.0-30.0 10.0-20.0 0.0-10.0

Ic (A)

(b) Fig. 5. Theoretically calculated variation of cold side temperature TC with currents in the upper and lower TECs. (a) is a 3-d surface plot and (b) is a contour plot of the data. Outside air temperature of 37°C and a 4W internal heat source are assumed.

16

As with the single stage case, TC drops as IC and IH increase, reaching a minimum point, beyond which it begins to increase as resistive heat from the TEC begin to dominate. However, what is more interesting is that there are now multiple combinations of IC and IH that can result in the same TC. For example, the blue band on the contour plot of Fig. 5(b) contains all possible combinations of IC and IH that will result in TC between 20°C and 30°C. It is possible that amongst all of these possible combinations, some pairs of IC and IH values will result in significantly higher COP than a single stage system. Fig. 6 shows a 3-d surface plot of the corresponding COP values as a function of currents IC and IH. As for the single-stage system, COP drops rapidly with increasing values of current. A more detailed analysis of COP done using the solver function in Excel shows that setting IC = 0.92 A and IH = 0.78 A results in the highest COP of 3.72 while maintaining the internal temperature at 27°C, i.e. 10°C below outside air temperature. This value is considerably larger than the COP of about 2 that was obtained for a single-stage system under the same conditions. It should be noted that by optimizing the currents such a thermoelectric air conditioning system can be operated with its COP constantly maximized, even under varying values of heat load, TH and TC. Such a control mechanism that adapts with changing thermal conditions is very complex to implement in vapor compression systems.

20.0 18.0 16.0

12.0

10.0

0.5 1 1.25 1.5

2

1.75

Ih (A)

1.5

1.25

1

0.75

0.5

0.25

8.0

1.75

COP

14.0

18.0-20.0

6.0

16.0-18.0

4.0

14.0-16.0

2.0

12.0-14.0

0.0

10.0-12.0

0.25

8.0-10.0

0.75

6.0-8.0

Ic (A)

4.0-6.0 2.0-4.0 0.0-2.0

Fig. 6. 3-d surface plot of COP versus IC and IH for the two-stage system described in this section.

17

2.3 Design Innovations Based on Theoretical Analysis In this section the main takeaways from the theoretical analysis presented above are summarized. First, it is noted that the electrical power consumed by a thermoelectric cooling system to maintain a constant temperature difference, T, is strongly dependent on the heat load, Qload. With decreasing Qload, the power consumption decreases. In reality, the heat load itself cannot be controlled, but one way to reduce the effective heat load on each TEC module is to use multiple modules operating simultaneously. Since each module pumps a fraction of the total heat load, the overall power consumption, and hence the efficiency of the system, is improved, compared to having a single module handle the total heat load. Fig. 7 illustrates this "divide and conquer" approach. In this figure the theoretically calculated variation of TC for with input power is plotted for the single-stage system described in Section 2.1. TH = 37°C and Qgen = 4W are used as before. Results from two cases are shown, one for a single TEC module and a second with four modules connected in series. It is seen that for any value of TC the 4-module system requires much less input power, i.e. has a much higher COP.

Room Temperature, Tc (C)

40 1 TEC Module

35

4 TEC Modules 30 25 20 15 10 5 0 0

10

20

30

40

50

Input Power, Pin (W)

Fig. 7. Theoretically calculated TC as a function of input power for TH = 37C and Qgen = 4W.

A second technique to improve the effective COP of a thermoelectric cooling system is based on that fact that for a given heat load and cold side temperature, its energy consumption depends on the temperature of it hot side. Using a water-cooled active heat sink it should be possible to lower the hot side temperature of the TEC module, thus reducing its energy consumption. Existing thermoelectric cooling units typically use passive heat sinks such as large

18

metal fins with or without a fan attached to them. These approaches are not very effective in conducting away all the heat pumped by the TEC from its hot surface. Fig. 8 shows the approach that is used in this project. It consists of placing a metal block with an inlet and an outlet for water, between the TEC and a finned metal heat sink. The detailed theory of heat conduction by water cooled heat sinks is complicated and not covered in this study. However, it is known that water cooled heat sinks can be ten times more effective than air cooled ones [11].

Fan TEC Water in Water cooled heat sink Water out

Metal fin passive heak sink

Hot side

Fig. 8. Diagram showing how an active water cooled heat sink is used to control the temperature of the hot side of the TEC.

The third technique to improve the effective COP of a thermoelectric cooling system is to use TEC modules in cascade. As shown in Section 2.2, such a system has, in theory, a much higher COP than a single-stage system.

2.4 Theoretical Calculations and Hypothesis In this section a hypothesis is developed from theory for the objective of this project. Typically 15% of a home's floor area is covered with windows. Normal solar intensity can be taken to be 1000 W/m2. Therefore, the peak total solar power expected to enter through windows of a typical home is 150 W/m2 of floor area. The model home used in this study has a floor area of 40 x 40 cm2. Thus, the solar heat load on such a home can be estimated to be 24W. In order to maintain a steady temperature inside the home the air conditioner will have to pump out a total of 24 watts. This assumes that all of the solar energy entering the room is trapped inside the room. In order to come up with a reasonable hypothesis regarding the project objective that the TEC system must have a COP that is 20% higher than that of a similar vapor compression unit, some additional assumptions are needed. A COP of 2 is assumed for a vapor compression air conditioner. This is typical of small commercially available units. The TEC system must then

19

have a COP of 2.4 or higher. The equations of Section 2 were set up in an Excel calculator to help design the kind of thermoelectric cooling system that will be required to pump out 24 watts and the associated COP. It was found that for a single-stage design with one TEC module, it would require 15W of input power to attain a T of 8°C. This implies a COP of 1.6 which does not meet the objective. Next, with a single-stage system consisting of 4 serially connected TEC modules it was found that 8W would be required to attain a T of 8°C, resulting in a COP of 3, which meets the project objective. Lastly, using a two-stage system, the theoretical calculations showed that under optimum current combinations, about 4W of input power would be needed to attain a T of 8°C, which implies a COP of 6. Therefore, based on the theoretical calculations it is hypothesized that the project objective can be met by using a two-stage design, if it is ensured that the TEC's hot side temperature is equal to the outside air temperature. It is noted that the energy consumed by the pump used in the water cooled heat sink is not considered in these calculations because it expected to be operated on solar power.

3. Experimental Setup and Procedures 3.1 Variables The independent variables of this project are the type of air conditioning system used, (thermoelectric / vapor compression). For the thermoelectric system, the number of modules and number of stages are also independent variables. The dependent variable is the energy consumed by each type of air conditioning system under identical heat load conditions. The heat load, the difference maintained by air conditioners between room and outside air temperature, and the volume of space being cooled are kept constant, i.e. these are the control variables.

3.2 Apparatus Used A scale model home with outside dimensions of 60cm x 60cm x 60cm is constructed out of foam board. 10 cm thick insulation is placed on the inside of its walls, floor, and ceiling. A bank of resistors with an equivalent resistance of 80 ohms is placed in the center of the model home and connected to a variable ac power source (variac) to act as the heat load. It is assumed that the electric power consumed by the resistors is equal to the heat load. The heat load can be changed

20

by changing the variac output voltage. For example, when the variac is set to supply 44 volts, a power output of 24.2 watts results from the resistors. The exact relation between the heat load and variac output voltage is given as

Qheat 

Vac2 80

(17)

where Vac is the variac output voltage. Electronic thermometers are placed both inside and outside the house. The house is fitted with the thermoelectric cooling unit on one side, and with a small vapor compression unit on the opposite side. Temperatures inside and outside the house, as well as the energy consumed by the air conditioners are measured and logged electronically. A Velleman PCS-10 logger is used for recording the temperatures and a "Watts Up Pro" unit is used for recording the electric energy used by vapor compression unit. The thermoelectric system is driven by adjustable DC power supplies and its energy usage is found by monitoring the voltage and current supplied by the power supplies. Fig. 9 shows a diagram of the setup with the connections for the thermometers, thermostat, and data loggers. To control the vapor compression unit, a programmable thermostat is used. The temperatures sensed inside and outside the house are inputted to the thermostat, which turns off the system when the inside temperature drops to a preset value below the outside temperature. For the thermoelectric unit, the input electrical power is adjusted manually till the desired room temperature is attained.

V

Thermometer (inside air)

Heater

Variac

AC (TE)

Thermometer (outside air) AC (VC)

Insulated walls, roof, and floor

Data logger (temperature)

Thermostat

Data logger (energy)

Fig. 9. Diagram showing equipment used to monitor energy usage by different air conditioning units.

21

Details of the thermoelectric cooling unit are shown in Fig. 10. Fig. 10(a) is a diagram of the single stage design. Four model TEC modules, each rated at 72 watts maximum cooling power, are connected in series as shown. They are placed on a water cooled heat sink and held in place by two 1/8 inch thick aluminum plates. Nylon sleeves and washers are used with the clamping screws to ensure that the top and bottom plates are thermally insulated from each other. A fan is attached to the top plate to help circulate cold air (not shown in the diagram to keep clarity). Fig. 10(b) is shows the design of the two-stage cascaded thermoelectric unit. A fifth TEC module is used here. It is placed on top of the upper aluminum plate and held in place by the fan. Fig. 10(c) shows a photograph of the final assembled unit. Top Al plate Active heat sink

TEC module Bottom Al plate

Clamping screw

(a)

2nd stage TEC module

(b)

(c)

Fig. 10. (a) Diagram of the single-stage, quad-module thermoelectric cooling system, (b) diagram of the two-stage cascaded thermoelectric system, and (c) photograph of fully assembled cascaded thermoelectric cooling unit used in this project.

22

3.3 Measurement Procedures The heating resistors and air conditioning unit are turned on and the outside and inside temperatures are monitored till they become stable. After that the data loggers are turned on and temperature and energy usage are logged for a period of 20 minutes. After twenty minutes the data from the loggers are transferred to a computer and saved for analysis later.

4. Experimental Results 4.1 Energy Consumption of Vapor Compression System In this set of measurements, the amount of energy consumed by a traditional vapor compression air conditioner to cool the scale model home described previously is measured under different T. Although the project objectives are for T = 8°C, measurements are done for several T values to get a more complete picture of the performance of the system around the target operating conditions. Fig. 11 shows the power and energy consumed over time by the vapor compression system under a 24W internal heat load while maintaining an average T of 10°C, 5.9°C, and 0.7°C. The step-like shape seen in the graph of energy versus time is because the air-conditioner turns off when the home has cooled to the desired temperature and turns back on when the temperature rises above the set point temperature. Whenever the air conditioner is off, energy consumption also stops increasing, resulting in the flat portions of the "steps" in Fig. 11. Fig. 12 shows the temperature variation inside the home as a function of time under the same conditions as described above. It is seen that the temperature varies above and below some average value as the thermostat turns the system on and off. The T values used here are based on the difference between the outside temperature and the average inside temperature over the 20 minute measurement duration (which is the reason why they are somewhat irregular).

23

30 DT = 0.7C

Energy Used (Wh)

25

DT = 5.9C 20

DT = 10C

15 10 5 0 0

2

4

6

8

10

12

14

16

18

20

tim e (m in)

Fig. 11. Energy consumed as a function of time by the vapor compression air conditioner over a 20 minute duration for three different T settings, each with a heat load of 24W.

Temperature (C)

23.0

18.0

DT = 0.7C

13.0

DT = 5.9C DT = 10C 8.0 0

2

4

6

8

10

12

14

16

18

20

Tim e (m in)

Fig. 12. Temperature inside home as a function of time for the vapor compression air conditioner for different T values.

Fig. 13 is a graph that shows a summary of results from all experiments of this set. The graph shows electric energy used as a function of indoor temperature for an internal heat load of 24W and an outdoor temperature of 21°C. For convenience, the results obtained from the 20 minute measurements are multiplied by 3 and presented as energy consumed in an hour. The gray line denotes the best linear fitting to the measured data. As expected, the energy consumed by the system decreases as T decreases. For an indoor temperature of 13°C, corresponding to a T of 8°C as required by the project objective, the vapor compression system consumes 70Wh of energy in one hour.

24

Energy Used in One Hour (Wh)

100.0 90.0 80.0 70.0 60.0 50.0 40.0 30.0 20.0 Qgen = 24W 10.0 0.0 4

6

8

10

12

14

16

18

20

22

Temperature in Room (C)

Fig. 13. Summary of energy used over one hour by the vapor compression system as a function of indoor temperature setting. Outside temperature = 21°C. The gray line is a linear equation fitted to the measured data. The vertical dotted line is at 13°C which is the target indoor temperature for energy usage comparison.

4.2 Energy Consumption of Single Stage Thermoelectric Systems In this set of measurements the energy consumed by single stage thermoelectric air conditioners to cool the scale model home is measured. As with the vapor compression system, measurements are done at several indoor temperatures and each measurement is done over a 20 minute time span. The energy consumed during this period is multiplied by three to estimate the hourly energy usage of the system. Fig. 14 shows the energy consumed over time by a single-stage, single-module thermoelectric system under a 24W internal heat load while maintaining a T of about 10°C, 5°C, and 0°C. Unlike in the vapor compression system an on-off type thermostat is not used in this case. The current through the system is adjusted till the desired temperature is reached inside the home. Once the desired temperature is achieved the system runs on its own while the temperatures and power consumption information are logged. Unlike in the vapor compression system, the graph of energy versus time is smooth, i.e. not step-like. Fig. 15 shows the temperature variation inside the home when cooled by a thermoelectric system. Fig. 16 is a graphical summary of results from all experiments on a single-stage singlemodule system. The graph shows measured indoor temperature as a function energy input. Outdoor temperature is kept constant at 21°C and internal heat load is 24W. The gray lines 25

denote the best linear fitting to the measured data. As expected, as the energy input into the system increases, the indoor temperature decreases. In theory, the inside temperature should increase if input power is increased beyond a point, but the system was not pushed to that limit. [Note: Results from thermoelectric air conditioning systems are plotted differently than those from vapor compression system because of the difference in how the two systems are controlled. While in the vapor compression system the desired indoor temperature is set on the thermostat and the energy usage is observed, in the thermoelectric systems the input current, which is equivalent to the input power, is set to a certain value and the resulting indoor temperature is observed. Thus it is more convenient to plot the input power on the x-axes when dealing with thermoelectric systems.]

Energy Used (Wh)

40 35

DT = 5C

30

DT = 10C

25

DT = 0C

20 15 10 5 0 0

2

4

6

8

10

12

14

16

18

20

Tim e (m in)

Fig. 14. Energy consumed as a function of time by the single-stage, single-module, thermoelectric air conditioner over a 20 minute duration for three different T settings, each with a heat load of 24W.

Temperature (C)

23.0

18.0

DT = 5C

13.0

DT = 10C DT = 0C 8.0 0

2

4

6

8

10

12

14

16

18

20

Tim e (m in)

Fig. 15. Temperature inside home as a function of time for the single-stage, single-module, thermoelectric air conditioner over a 20 minute duration for different T settings.

26

Temperature in Room (C)

22.0 20.0 18.0 16.0 14.0 12.0 10.0 8.0 Qgen = 24W

6.0 4.0 0.0

20.0

40.0

60.0

80.0

100.0

120.0

Energy Used (Wh)

Fig. 16. Measured variation of indoor temperature versus electrical energy used over one hour by a single-stage single-module thermoelectric system. Outside temperature = 21°C. The gray line is a quadratic equation fitted to the measured data. The horizontal dotted line at 13°C represents the target indoor temperature.

It is found from the data of Fig. 16 that about 90Wh of energy is consumed by the singlestage, single-module thermoelectric system to keep the indoor temperature of the model house at 8°C below outside air temperature. This is significantly higher than the 70Wh consumed by the vapor compression system under the same conditions. Obviously, such a system does not meet the objectives of this project. However, it was observed in Section 2.3 that, in theory, by using multiple TEC modules in series, the energy usage of a thermoelectric system can be reduced significantly as it reduces the heat load on each module. This concept was tested next using four TEC modules (quad-module) in series. Fig. 17 shows the summary results obtained from the quad-module, single-stage system shown previously in Fig. 10(a). It is seen that to maintain an indoor temperature of 13°C, about 0.9A of current needs to be passed through the TEC modules. The energy consumed over one hour at this current is about 50W. This meets the project objective of developing a thermoelectric system that uses 20% less energy than a similar vapor compression system under the same operating conditions.

27

90

20

80

18

70

16

60

14

50

12

40

10

30

8

Energy Used (Wh)

Temperature in Room (C)

22

20 Temp in Room

6

Energy Used

4 0

0.2

0.4

0.6

0.8

1

1.2

10 0 1.4

Current, I (A)

Fig. 17. Measured variation of indoor temperature (blue) and energy consumed in one hour (red) versus electrical current input for a single-stage quad-module thermoelectric system. Outside temperature = 21°C. The horizontal dotted line at 13°C represents the target indoor temperature.

4.3 Energy Consumption of Two-Stage Cascaded Thermoelectric Systems Although the project objective was met with a single-stage, quad-module thermoelectric system, additional experiments were done using a dual-stage cascaded system to confirm if additional gains in efficiency would be possible as indicated by theory. The system shown earlier in Fig. 10(b) and 10(c) was assembled and tested under identical conditions of 21°C outside temperature and 24W heat load. Fig. 18(a) shows plots of indoor temperature as a function of current IH in the lower layer of the cascaded arrangement for several values of current IC in the upper layer (IH and IC refer to the currents shown earlier in Fig. 4). As expected from theory, the plots have a parabolic shape; once the current exceeds a critical value, the room temperature begins to rise due in increased resistive losses in the TEC. But, more importantly, it can be seen by following the dotted line that the target temperature of 13°C can be obtained by several possible combinations of IC and IH. Fig. 18(b) shows the associated temperature versus energy consumption data. The shapes of the plots are as expected, and, again, it is seen that the target temperature can be attained by multiple combinations of IC and IH. From the available data, the minimum energy required to obtain the target 13C inside temperature is at IC = 0.5A and IH = 0.8A. At these currents the

28

hourly energy used is about 40Wh. This represents a 40% lower energy usage than the vapor compression system.

Temperature in Room, Tc (C)

22 Ic = 0.25A

20

Ic = 0.5A

18

Ic = 1A

16

Ic = 2A

14 12 10 8 6 4 0

0.5

1

1.5

2

2.5

Current Ih (A)

(a)

Temperature in Room, Tc (C)

22 Ic = 0.25A

20

Ic = 0.5A

18

Ic = 1A

16

Ic = 2A

14 12 10 8 6 4 0

40

80

120

160

200

240

280

320

360

Energy Used (Wh)

(b) Fig. 18. Measured variation of indoor temperature versus (a) electrical current IH, and (b) versus hourly energy use for a dual-stage cascaded thermoelectric system. Outside temperature = 21°C. The horizontal dotted line at 13°C represents the target indoor temperature.

5. Discussion It was seen in the previous section that a dual-stage cascaded thermoelectric air conditioning system can be significantly more energy efficient than a similarly sized vapor compression

29

system. The experimentally observed behaviors showed trends similar to what was predicted by theory, but a closer comparison would be useful in validating the results. Fig. 19 is a comparison of theoretical and experimentally measured values of indoor temperature as a function of energy input for the dual-stage thermoelectric system. The experimental data are the same as shown in Fig. 18 and the theoretical data is obtained from the equations of section 2.2 with the same inputs as used in the experiments. It can be seen that although there is reasonable similarity in trends between experiment and theory, there is a large difference between the numerical values. Two main reasons for this are (a) while the theory is based on 1-d cascade structure and heat flow, in reality it is a 3-d effect, and (b) while the theory assumed that the TEC cold side is at the same temperature as the indoor air, in reality the air temperature is likely to be significantly higher. Temperature in Room (C)

25

Theory (Ic = 0.5A) Expt (Ic = 0.5A)

20

Theory Ic = 1A Expt Ic = 1A

15 10 5 0 0

25

50

75

100

125

150

Energy Used (Wh)

Fig. 19. Comparison of energy usage of vapor compression and thermoelectric air conditioning systems as a function of T with a heat load of 24W.

6. System Cost Comparisons In real-world use, cost is a very important factor. Even though it is found that thermoelectric air conditioners used with IR filtering and active heat sinks can be made to consume less energy and are therefore less expensive to operate, if the cost of purchasing and installing such a system is high then it is unlikely to be accepted widely. Table 2 is a summary of the costs involved in the two types of systems considered in this project. It is seen that the thermoelectric system is about 40% less expensive than the vapor compression system. The system costs given above are for a very small model home. To be of practical value it is important to consider how the costs mentioned above would increase as the system is scaled to 30

a larger size cooling area. This is discussed next. The scope is limited to window mounted room air conditioners. The heat pumping power of commercially available vapor compression window air conditioners range from 1500 W to 7500 W. Purchase prices of several of these units are obtained from a market survey. Fig. 18 shows the dependence of purchase price on the cooling power. It is observed that the price increases roughly linearly at the rate of 8.5 cents per watt. A similar cost per watt of cooling power can be estimated for TEC based systems as follows. At wholesale rates, TEC modules such as the one used in this project cost $3 each. These modules have an optimum heat pumping power of about 20 watts. The cost of the active heat sink used in this project is mainly due to its motor. Since the motor is large enough to cool 100 TEC modules the heat sink cost can be reduced to $0.2 per TEC. From these assumptions it is estimated that at current market prices for TEC modules, practical TEC based air conditioners will cost about 16 to 20 cents per watt of cooling power. Table 2. Comparison on system costs.

Vapor Compression Air Conditioner A/C unit $55.00

TOTAL

Thermoelectric Air Conditioner TEC module $5 x 4 Water cooled heat sink $20 Passive heat sink $2 TOTAL $42.00

$55.00

1200

1000

Cost ($)

800

600

400

200

0 0

1000

2000

3000

4000

5000

6000

7000

8000

Cooling Power Qp (W)

Fig. 20. Dependence of purchase price of vapor compression air conditioners on cooling power found from a market survey (dots). Dashed line represents estimated dependence of cost per cooling watt for TEC systems based on current market prices.

31

7. Future Work The primary advantage of a thermoelectric air conditioning system is that its cooling power can be adapted to the existing heat load and temperatures by controlling the current through the modules. In case of a dual stage cascaded system there is even more flexibility because the current in each stage can be varied separately. One of the main items that will be addressed in future is a control system that can automatically take into consideration the existing heat load and temperatures and set the currents such that energy efficiency of the system is maximized. 8. Conclusions In conclusion, a novel thermoelectric based air conditioning system was demonstrated using commercially available components. The objectives of the project were successfully met. It was found that with a dual-stage cascaded design it is possible to construct a thermoelectric air conditioner that is significantly more energy efficient than a similarly sized vapor compression system. At current market prices the purchase cost of a practical thermoelectric air conditioner is expected to be a little more than a comparable vapor compression unit.

32

9. References 1. American Council for an Energy Efficient Economy Website. Internet: aceee.org/topics/airconditioners, 2012 [Oct. 01, 2012]. 2. US Department Energy, Energy Inform. Administration, "Annual Energy Review 2011," Internet: www.eia.gov/totalenergy/data/annual/pdf/aer.pdf., Sep. 2012 [Oct. 15, 2012]. 3. Marsicek, G., Klein, S., Nellis, G., "Feasibility of Combined Solar/Heat Pump Systems for Net-Zero Buildings," International Building Performance Simulation Association SimBuild Conference. [Online]. Available: http://www.ibpsa.us/simbuild2012/Papers/SB12_TS03a_3_Marsicek.pdf 4. Rowe, D. M., Editor, CRC Handbook of Thermoelectrics, CRC Press, 1995 5. Nolas, G. S., Sharp, J., Goldsmid H. J., Thermoelectrics – Basic Principles and New Materials Developments, Springer, 2001. 6. Li, J., Liu, W., Zhao, L., Zhou, M., "High-performance Nanostructured Thermoelectric Materials," NPG Asia Materials, vol. 2, pp. 152-158, 2010. 7. Harman, T. C., Walsh, M. P., laforge, B. E., Turner, G. W., "Nanostructured Thermoelectric Materials," Journal of Electronic Materials, vol. 34, pp. L19-L22, 2005. 8. Johnson, D. A., Bierschenk, J., "Latest Developments in Thermoelectrically Enhanced Heat Sinks," Electronics Cooling. [Online]. Available: http://www.electronics-cooling.com/2005/08/latestdevelopments-in-thermoelectrically-enhanced-heat-sinks. [Aug. 1, 2005]. 9. Winkler, J., "Laboratory Test Report for Fujitsu 12RLS and Mitsubishi FE12NA Mini-Split Heat Pumps," Internet: www.nrel.gov/docs/fy11osti/52175.pdf, Sep. 2011 [Oct. 01, 2012]. 10. Goldsmid, H. J., Douglas, H. W., "The Use of Semiconductors in Thermoelectric Refrigeration," British Journal of Applied Physics, vol. 5, pp. 386 - 390, 1954. 11. Brown, D. R., Dirks, J. A., Fernandez, N., Stout, T. B., " Prospects of Alternatives to Vapor Compression Technology for Space Cooling and Food Refrigeration Applications," Internet: www.pnl.gov/main/publications/external/technical_reports/pnnl-19259.pdf, Mar. 2010 [Oct. 10, 2012]. 12. A.Melero, D.Astrain, J.G.Viin, L.Aldave, J.Albizua, and C.Costa, "Application of Thermoelectricity and Photovoltaic Energy to Air Conditioning," Proc. 22nd International Conference on Thermoelectrics, pp. 627-630, 2003. 13. H. Lai, Y. Pan, and J. Chen, "Optimum Design on the Performance Parameters of a Two-Stage Combined Semiconductor Thermoelectric Heat Pump," Semiconductor Sci. Technol., vol. 19, pp. 17 – 22, 2004.

33

14. M. Olivares-Robles, F. Vazquez, and C. Ramirez-Lopez, "Optimization of Two-Stage Peltier Modules: Structure and Exergetic Efficiency," Entropy, vol. 14, pp. 1539 – 1552, 2012. 15. R.E. Simons and R. C. Chu, "Application of Thermoelectric Cooling to Electronic Equipment: A Review and Analysis," Proc. 16th IEEE SEMI-THERM Symposium, pp. 1-4, 2000. 16. Karim, O., Creiber, J. C., Gillot, C., Schaffer, C., Mallet, B., Gimet, E., "Heat Transfer Coefficient for Water Cooled Heat Sink: Application for Standard Power Modules Cooling at High Temperature," Proc. Of 32nd Power Electronics Specialists' Conf., pp. 1938 - 1943, 2001. 17. O'Brien, B. J., Wallace, C. S., Landecker, K., "Cascading of Peltier Couples for Thermoelectric Cooling," J. Applied Physics, 27, pp. 820 - 824, 1956.

34