0

Experimental validation of the optimum design of automotive air-to-air

thermoelectric air conditioner (TEAC)

Alaa Attar [1, 2, 3]

HoSung Lee [1]

Sean Weera [1]

1. Department of Mechanical and Aeronautical Engineering, Western Michigan University, Kalamazoo, MI

49008-5343, USA. 2. Department of Mechanical Engineering, King Abdulaziz University, Rabigh, Saudi

Arabia. 3. —email: alaa.m.attar@wmich.edu

1

Abstract

The optimization of thermoelectric air conditioners (TEAC) has been a challenging topic due to the

multitude of variables needed to be considered. The present work discusses the experimental validation of

the optimum design for automotive air-to-air TEAC. The TEAC optimum design was obtained by using a

new optimal design method with dimensional analysis that has been recently developed. The design

constraints were obtained through a previous analytically study on the same topic. In order to simplify the

problem, a unit cell that represents the entire TEAC system was analytically simulated and was

experimentally tested. Moreover, commercial TEC modules and heat sinks were selected and tested based

on the analytical optimum design results.

Key words: Thermoelectric air conditioner; automotive thermoelectric cooling; thermoelectric automobile

application.

2

Nomenclature

Ae cross-sectional area of thermoelement (mm2) T∞c cold fluid temperature (°C)

Ab total base area of thermoelectric air conditioner

(mm2)

T∞h hot fluid temperature (°C)

Ac total fin surface area cold side heat sink (mm2) ∆T thermoelectric temperature different (∆°C)

Ah total fin surface area hot side heat sink (mm2) ∆Tcooling cold air temperature inlet – outlet (∆°C)

AUC unit cell base area (mm2) tc cold side air fin thickness

COP the coefficient of performance th hot side air fin thickness

cp specific heat (J/kg.K) Vh hot fluid volume flow rate (L/min for liquid) or

(CFM for air) 𝐺𝑒 thermocouple geometric ratio

H total height of thermoelectric air conditioner

(mm)

W total width of thermoelectric air conditioner

(mm)

h heat transfer coefficient of the fluid (W/m2K) Z the figure of merit (1/K) = α2/ρk

I electric current (A) zc fin spacing for the cold side air (mm)

L total length of thermoelectric air conditioner

(mm)

zh fin spacing for the hot side air (mm)

Lc unit cell cold side length (mm)

Le length of thermoelement (mm) Greek symbols

Lh unit cell hot side length (mm) α Seebeck coefficient (V/K), α = αp − αn

k thermoelement thermal conductivity (W/m K),

k = kp + kn

ρ electrical resistivity (Ω cm), ρ = ρp + ρn

𝜑 alimunum block thermal resistance (K/W)

n the number of thermocouples ηc fin efficiency of cold side heat sink

nc number of fins for the cold side heat sink ηh fin efficiency of hot side heat sink

nh number of fins for the cold side heat sink

Nk dimensionless thermal conductance, Nk =

n(Aek/Le)/ηhhhAh

Subscripts

Nh dimensionless convection, Nh = ηchcAc/ηhhhAh c cold

NI dimensionless current, NI = αI/(Aek/Le) e thermoelement

Qc total cooling power from thermoelectric air

conditioner (W)

h hot

Pin total input power for from thermoelectric air

conditioner (W)

p p-type element

PD power density (W/cm2) n n-type element

R electrical resistance of a thermocouple (Ω) m measured

Re fluid Reynolds’s number opt. optimal quantity

Tc cold junction temperature (°C) UC unit cell

Th hot junction temperature (°C) ∗ dimensionless

3

Introduction

Most automobile air conditioners use refrigerant R-134a, which does not have ozone-depleting

properties like Freon, but is nevertheless a terrible greenhouse gas [1]. Soon enough, this refrigerant will

be banned which means alternative air-conditioning technology is needed [1]. In 2009, the U.S. Department

of Energy (DOE) and the California Energy Commission funded a project to improve the air conditioning

(AC) system of vehicle by developing thermoelectric heating ventilation and air conditioner (TE HVAC)

system which would replace the current conventional AC system [2]. Using a thermoelectric air

conditioning (TEAC) system instead of a conventional AC system has two main benefits: it will eliminate

the need of R-134a and it will provide the ability to cool selected zones instead of the entire cabin which in

turn will reduce fuel consumption [3]. On average, 73% of a vehicle’s mileage occurs when the driver is

the only occupant and it is estimated that the total cooling power required to cool the zone of a single

occupant is around 630 W while 3.5 to 4.5 kW is needed to cool the entire cabin [2]. The goal of the DOE

project was to build a TEAC device that could provide the needed 630 W of cooling power for a single

occupant with a coefficient of performance (COP) of 1.3 or higher [4].

Several studies have been conducted on automotive TEAC systems in order to test their

performance and feasibility. Junior et al. [5] compared a thermoelectric liquid-gas heat exchanger system

used for steady state and transient simulation models with the conventional auto HVAC system. For

ambient temperatures of 25°C and 30°C, the conventional auto HVAC system has cooling capacity of five

times higher than the thermoelectric HVAC system at the same input power [5]. Wang et al. [6] designed

and analyzed an air-to-liquid thermoelectric HVAC system for a passenger vehicle using a numerical

model. They also constructed an experiment to validate their model that was able to reach a COP of 1.55

at a cooling power of 1.55 kW with the same air and liquid inlet temperatures of 30°C [6]. The use of the

thermal isolation method allowed COP improvement and determination of the fluid and junction

temperatures [6]. Raut and Walke [7] constructed and tested a thermoelectric cooler system on a small

passenger vehicle where the goal was to remove 222 W of heat from the cabin. They used six TEC modules

4

(each module has a maximum cooling power of 48 W) connected electrically in series and sandwiched

between two heat sinks. Their results showed that the system was able to reduce the cabin temperature to

as low as 4°C [7]. Hsu et al. [8] studied and tested an air-to-liquid TEAC. After optimizing the heat sink,

they analyzed the effects of the figure of merit and thermoelement thickness on the system COP and cooling

performance. They stated that the change of the element thickness can improve the cooling performance

but not the COP. They built and tested an air-to-liquid TEAC installed in a vehicle (Honda Civic Exi). The

experimental results showed that the TEAC performance curves and cooling power trends followed the

results from the basic equations.

The Ford Motor Company, in collaboration with Gentherm, presented their design of a TEAC along

with a performance curve in the 2012 Directions in Engine-Efficiency and Emissions Research (DEER)

Conference [9]. They tested an air-to-liquid TEAC system where it was able to reach a COP of 1.3 at an

input power of 400 W using a cold air flow rate of 60 CFM. They decided on using liquid instead of air

for the hot (waste) side fluid because of many advantages such as higher heat transfer coefficients, higher

power density, and lower operating noise [10]. However, there are significant considerations to account

for such as air-to-liquid coolant leakage as well as the complexity of an additional radiator system [10].

Table I highlights some of the pros and cons for air-to-air vs. air-to-liquid TEAC systems. Attar et al. [11]

applied an optimal design method developed by Lee [12] to design air-to-liquid TEAC based on Gentherm

design and they were able to obtain a COP of 1.68 at the same input power. This optimal design method

used the dimensional analysis in order to optimize current supplied and the geometric ratio of

thermoelement (or number of thermoelement couples) simultaneously for a given set of fixed parameters.

More details of this method are discussed in the next section. Moreover, with use of this optimal design

method, Attar et al. [11] also designed air-to-air TEAC system that has a COP of 1.3 at the same input

power (400W). This performance was obtained from designing a unit cell that can simulate the whole

system in order to simplify the problem. Even though this work is showing the optimized design, it is still

considered to be analytical work and many uncertainties may occur. Therefore, the present work discusses

the experimental validation of this unit cell of air-to-air TEAC design.

5

Table I Comparison between air-to-air TEAC and air-to-liquid TEAC [10]

Air Waste Stream Liquid Waste Stream

Pros Cons Pros Cons

Low weight Poor heat transfer Higher power density More weight

No risk of coolant leaks Lower power density Less noise Risk of leaks

Difficult to vent the

waste heat

Requires an additional

radiator

Noise at higher flow rate

Waste side temp tied to

ambient

Background

In the previous work [11], the air-to-air TEAC model was built and analytically optimized. The

cold air heat sink is sandwiched between two layers of the thermoelectric modules while two layers of hot

air heat sinks are attached to the hot sides of the thermoelectric modules as shown in Fig. 1a. The optimum

design model was built based on a unit cell (Fig. 1b) located at the center of the TEAC system. The unit

cell ambient cold and hot fluids temperatures, T∞c and T∞h, are calculated by assuming linear change of

the cold and hot temperatures along the TEAC system. The design of the heat sinks were optimized (using

optimization method found in Lee [13]) to give the minimum thermal resistances at a provided flow.

Moreover, the Nusselt number correlation devolved by [14] is used to calculate the heat transfer coefficients

for both cold and hot flows.

6

Fig. 1. (a) Schematic of the air-to-air TEAC and (b) unit cell schematic.

7

It is found that in order to obtain the optimum design for the thermoelectric cooling system to

maximize the cooling power �̇�𝑐, the electrical current I and the thermocouple geometric ratio (𝐺𝑒 = 𝐴𝑒/𝐿𝑒)

must be optimized simultaneously. Therefore, the optimum design method using the dimensional analysis

technique that was developed by Lee [12] is used. In this method, dimensionless numbers were defined

under the assumption that the electrical and thermal contact resistances in the TEC are negligible, the

material properties are independent of temperature, the TEC is perfectly insulated, and the p-type and n-

type element dimensions are identical. This method converts the four basic heat balance equations (Eqns.

1 to Eqn. 4) into two non-dimensional equations (Eqns. 6 and 7). Fig. 2 (a) and (b) show schematics of a

thermoelectric module with heat sinks and thermoelectric couple, respectively. Moreover, the two

aluminum blocks, which are sandwiched between the cold and hot sides of the TEC and their respective

heat sinks, needed to be considered in the analysis where 𝜑 is thermal resistance of each block and it is

equal to 0.149 K/W.

8

Fig. 2 (a) thermoelectric cooler module (TEC) with two heat sinks, (b) schematic of thermoelectric

couple.

�̇�𝑐 = 𝜂𝑐ℎ𝑐𝐴𝑐[𝑇∞𝑐 − (𝜑𝑄𝑐 − 𝑇𝑐)] (1)

�̇�𝑐 = 𝑛 [𝛼𝐼𝑇𝑐 −1

2𝐼2𝑅 +

𝐴𝑒

𝐿𝑒𝑘(𝑇𝑐 − 𝑇ℎ)] (2)

�̇�ℎ = 𝑛 [𝛼𝐼𝑇ℎ +1

2𝐼2𝑅 +

𝐴𝑒

𝐿𝑒𝑘(𝑇𝑐 − 𝑇ℎ)] (3)

�̇�ℎ = 𝜂ℎℎℎ𝐴ℎ[(𝑇ℎ − 𝜑𝑄ℎ) − 𝑇∞ℎ] (4)

𝑃𝑖𝑛 = �̇�ℎ − �̇�𝑐 (5)

9

where 𝜂𝑐 is the fin efficiency, ℎ𝑐 is the convection coefficient, and 𝐴𝑐 is the total surface area of the cold

heat sink.

𝑁ℎ(𝑇∞

∗ − 𝑇𝑐∗)

𝑁𝑘= (𝜂𝑐ℎ𝑐𝐴𝑐𝜑 + 1) [𝑁𝐼𝑇𝑐

∗ −𝑁𝐼

2

2𝑍𝑇∞ℎ+ (𝑇𝑐

∗ − 𝑇ℎ∗)] (6)

(𝑇ℎ

∗ − 1)

𝑁𝑘= (𝜂ℎℎℎ𝐴ℎ𝜑 + 1) [𝑁𝐼𝑇ℎ

∗ −𝑁𝐼

2

2𝑍𝑇∞ℎ+ (𝑇𝑐

∗ − 𝑇ℎ∗)] (7)

ZT∞h, Nh , Nk , and NI are defined as the dimensionless figure of merit, convection ratio, the ratio

of thermal conductance to the convection conductance, and dimensionless current, respectively, and can be

defined as

𝑍𝑇∞ℎ =𝛼2

𝜌𝑘𝑇∞ℎ (8)

𝑁ℎ = 𝜂𝑐ℎ𝑐𝐴𝑐

𝜂ℎℎℎ𝐴ℎ (9)

𝑁𝑘 =𝑛(𝐴𝑒𝑘/𝐿𝑒)

𝜂ℎℎℎ𝐴ℎ (10)

𝑁𝐼 =𝛼𝐼

𝐴𝑒𝑘/𝐿𝑒 (11)

Tc∗, Th

∗ and T∞∗ are the dimensionless cold junction temperature, the dimensionless hot junction temperature,

and the fluid temperature ratio, respectively, and are defined as

𝑇𝑐∗ =

𝑇𝑐

𝑇∞ℎ (12)

𝑇ℎ

∗ =𝑇ℎ

𝑇∞ℎ

(13)

10

𝑇∞

∗ =𝑇∞𝑐

𝑇∞ℎ

(14)

the dimensionless temperatures are then a function of five independent dimensionless parameters as

𝑇𝑐∗ = 𝑓(𝑁𝑘 , 𝑁ℎ , 𝑁𝐼 , 𝑇∞

∗ , 𝑍𝑇∞ℎ) (15)

𝑇ℎ∗ = 𝑓(𝑁𝑘 , 𝑁ℎ , 𝑁𝐼 , 𝑇∞

∗ , 𝑍𝑇∞ℎ) (16)

setting 𝑍𝑇∞ℎ, 𝑇∞∗ and 𝑁ℎ to be the inputs, the dimensionless parameters 𝑁𝑘 and 𝑁𝐼 can be optimized to

solve Eqns. 6 and 7 for the maximum cooling power. The design requirements were to have an input power

of 400W and COP of 1.3 (or input power of 4.5W and COP of 1.3 for the unit cell). Therefore, for the

current experimental analysis, the input power is fixed at 4.5W and 𝐶𝑂𝑃 can be maximized.

Experimental Setup

In order to conduct an experiment that can simulate the unit cell of air-to-air TEAC system, a

commercial thermoelectric cooler module is sandwiched between two heat sinks (for hot and cold air). The

selection of the TEC and the heat sink were based on the optimized analytical design that was built

previously. Since the optimum heat sinks are not commercially available, closer heat sinks were selected.

Therefore, two heat sinks ALPAH UB30-20B and ALPAH UB30-25B were used for the cold and hot sides,

respectively. Fig. 3a shows the overall experimental setup while Fig. 3b shows the details of the test section.

Moreover, two aluminum blocks (30 × 30 × 19.1 𝑚𝑚3) were fabricated and inserted between the TEC

module and each of the heat sink. Two parallel (5mm apart) K-type thermocouples were drilled to the

center of each block where the average hot and cold blocks temperatures occurred. Moreover, the cold and

hot airs were driven by variable speed centrifugal blowers where a temperature bath controller and a heater

are used to control the inlet cold and hot air temperatures, respectively. The air speed was obtained by

measuring the dynamic pressure (the difference between the total pressure and static pressure) using a pitot

tube that was connected to a manometer. The blowers were set to have mass flow rates for cold and hot air

to be 3.21CFM and 6.1CFM, respectively. E-type thermocouples were installed at the air inlet and exit for

11

both cold and hot air in order to measure the air temperatures (𝑇∞𝑐,𝑖𝑛, 𝑇∞𝑐,𝑜𝑢𝑡, 𝑇∞ℎ,𝑖𝑛, and 𝑇∞ℎ,𝑜𝑢𝑡) so that

the average ambient temperature at the heat sinks could be averaged such that 𝑇∞𝑐 =1

2(𝑇∞𝑐,𝑖𝑛 + 𝑇∞𝑐,𝑜𝑢𝑡)

and 𝑇∞ℎ =1

2(𝑇∞ℎ,𝑖𝑛 + 𝑇∞ℎ,𝑜𝑢𝑡). On the other hand, the TEC input power was controlled by a variable

DC power supply which allowed controlling the input voltage.

From the analytical model at the required unit cell input power (𝑃𝑖𝑛,𝑈𝐶 = 𝐼𝑚 × 𝑉𝑖𝑛 = 4.5𝑊), the

average ambient cold and hot temperatures were required to be at 21.6 oC and 33.6 oC, respectively.

Therefore, the TEC supplied voltage, the cold air inlet temperature 𝑇∞𝑐,𝑖𝑛, and the hot air inlet

temperature 𝑇∞ℎ,𝑖𝑛 were adjusted accordingly until the input power and the average ambient temperatures

match the above values. After that, the measurements were taken under steady state conditions for each

input voltage (with increment of 1V) until reaching maximum voltage, 𝑉𝑚𝑎𝑥 provided by the manufacturer

as shown in Fig. 4.

The objective was to measure the cooling power 𝑄𝑐, the heat rejection 𝑄ℎ, the cold junction

temperature 𝑇𝑐 and the hot junction temperatures 𝑇ℎ. 𝑇𝑐 and 𝑇ℎ can be obtained by extrapolating the two

measured temperatures of each block (𝑇𝑐1& 𝑇𝑐2 for cold side and 𝑇ℎ1 & 𝑇ℎ2 for hot side) assuming the

temperatures change linearly across the aluminum blocks. For 𝑄𝑐 and 𝑄ℎ, the thermoelectric ideal Eqs.

2 and 3 were used where the electrical current and junction temperatures are experimentally obtained and

thermoelectric effective material properties are used for the values of 𝛼, 𝜌, and 𝑘 [15]. The effective

material properties technique is based on the maximum thermoelectric module parameters (typically

measurements), which were provided by the manufacturer, in order to calculate the material properties.

This technique enables to reduce the errors associated with the assumption of neglecting the contact

resistances and to provide more practical module properties.

The test was done for three TEC modules (module 1: Tellurex C2-30-1503, module 2: Tellurex C2-30-

0904, and module 3: Marlow RC12-4) in order to explore the effect of 𝑁𝑘 on the TEAC performance. All

three modules had the same base area (30 × 30 𝑚𝑚2) but different number of couples and/or

thermoelement geometric ratio, 𝐺𝑒.

12

Fig. 3 (a) Experimental setup, (b) test section

13

Fig. 4 Flowchart of the experimental procedure

Results & Discussion

A comparison has been made between experimental and analytical junction temperatures where the

cold and hot ambient temperatures, electrical current, and air flow rates were the inputs. Then, the

experimental junction temperatures were obtained by extrapolating the temperature readings from the

aluminum blocks while the analytical junction temperatures were obtained by using the four basic heat

14

balance equations (Eqs. 1, 2, 3, and 4). The results show very good agreements as shown in Fig. 5. From

these junction temperatures, the TEAC experimental performances were compared with the predicted

results for the three modules. Fig. 6 also shows a good agreement between experimental and analytical

𝐶𝑂𝑃 for the three tested modules. These results shown in Figs 5 and 6 indicate that the basic heat balance

equations with the effective material properties predict very well the measurements.

(a) (b) (c)

Fig. 5 Comparison between experimental and analytical junction temperatures vs. input current for

(a) module 1, (b) module 2, and (c) module 3

(a) (b) (c)

15

Fig. 6 Comparison between experimental and analytical COP vs. input power for (a) module 1, (b)

module 2, and (c) module 3

The results from the modified analytical model show the maximum possible 𝐶𝑂𝑃, at given

conditions, and it is equal to 1.09. This result came from the optimization of the two dimensionless values

𝑁𝑘 and 𝑁𝐼 simultaneously. Moreover, the values of 𝑁𝑘 and 𝑁𝐼 for the TEC modules number 1, 2, and 3

are obtained by using Eqs. 10, and 11 where the effective material properties is also been applied. Table II

shows a comparison between the three modules’ performances at the same input power (𝑃𝑖𝑛 = 4.5𝑊).

Among the three modules, module number 2 has the closest values of 𝑁𝑘 and 𝑁𝐼 to the optimum analytical

model to give the maximum COP at a given conditions. The table also shows that the performances of the

modules are degraded by the presence of the aluminum blocks. In addition, if the optimum heat sinks

(found in [11]) are used instead of the commercial ones, the system will converge to the optimum design

at the given conditions as shown in the last row of the table.

Table II Comparison between the three tested modules, optimized design without the aluminum

blocks, and the optimum design using the optimized heat sink

Module 𝑵𝑰 𝑰

(Amp) 𝑵𝒌

𝒏 × 𝑮𝒆

(𝒄𝒎) 𝑷𝒊𝒏 (W) COP

1 0.162 0.92 0.258 9.22 4.5 1.044

2 0.199 1.42 0.18 7.08 4.5 1.085

3 0.155 0.96 0.28 11.42 4.5 1.025

Analytical model 0.204 1.46 0.172 6.77 4.5 1.09

w/o blocks 0.219 2.9 0.154 11.21 4.5 1.16

w/o blocks & w/

optimal heat sinks 0.189 2.98 0.119 17.94 4.5 1.38

16

One of the experiment goals was to study the effect of the thermoelement geometric ratio (or

thermoelement number of couples) at the optimum input electrical current. This can be analyzed by

studying the ratio of thermal conductance to the convection conductance, Nk, and its relationship with the

optimum design. Testing three different modules validated the goal and gave the ability to view the closest

module to the optimum design. Fig. 1 shows a prediction and comparison of COP vs. 𝑁𝑘 between the three

modules at fixed input power (𝑃𝑖𝑛 = 4.5𝑊). The predicted curve was obtained by fixing 𝑃𝑖𝑛 and assuming

a constant temperature difference at the junctions (for a short range) and then resolved for 𝑁𝐼 to be only as

a function of 𝑁𝑘. This assumption allows expressing COP independently from 𝑁𝐼 for the range where all

of the three tested modules can be included. It can be seen from the figure that module 2 is the best module

among three at the given conditions because it has the closest values of 𝑁𝑘 and 𝑁𝐼 to the analytical optimal

design.

Fig. 7 COP vs. Nk at Pin = 4.5W

Conclusion

17

The present work discusses the experimental validation of the optimum design for automotive air-

to-air TEAC. The TEAC optimum design was obtained in a previous work [11] by using a new optimal

design method with dimensional analysis that has been recently developed by Lee [12]. The dimensional

analysis method obtains the maximum cooling power by simultaneously determining the dimensionless

current supplied NI and the ratio of the thermal conductance to the convection conductance Nk for a given

set of fixed parameters. In order to simplify the problem, a unit cell of the TEAC system was used instead,

which is considered to be located at the center of the whole TEAC system.

The experimental set up was designed to obtain the TEC module junction temperatures by

extrapolating two temperatures from thermocouples inserted into an aluminum block sandwiched between

the heat sink and the module. The results from the experiments showed a good agreement with the

analytical model that uses the four thermoelectric basic equations. These basic equations were solved at

the same input parameters from the experiment and by using the effective material properties [15] instead

of the intrinsic properties. Moreover, three TEC modules were tested in order to study the optimum

𝑁𝑘 value in addition to the optimized current. Therefore, module number 2 was the optimum module for

the given conditions.

Even though the use of the aluminum blocks was a necessity, they add extra thermal resistances

which act negatively on the TEAC performance. The use of commercial heat sinks instead of the optimized

designs also limits the reach of the optimum design of the air-to-air TEAC. Once these two constrains are

removed, the optimum design can be obtained with 20% improvement from module 2 performance.

18

REFERENCES

[1] C. B. Vining, "An inconvenient truth about thermoelectrics," nature materials, vol. 8, no. 2, pp. 83-85, 2009.

[2] J. Fairbanks, "Vehicular thermoelectric applications session DEER 2009," U.S. Department of Energy, 05 08

2009. [Online]. Available:

http://www1.eere.energy.gov/vehiclesandfuels/pdfs/deer_2009/session7/deer09_fairbanks.pdf. [Accessed 06

11 2013].

[3] R. Gonzalez, "Automotive thermoelectric HVAC development and demonstration project," California Energy

Commission, 10 2010. [Online]. Available: http://www.energy.ca.gov/2010publications/CEC-500-2010-

FS/CEC-500-2010-FS-018.PDF. [Accessed 06 11 2013].

[4] C. W. Maranville, "Thermoelectric HVAC and Thermal Comfort Enablers for Light-Duty Vehicle

Applications," United States Department of Energy, 18 05 2012. [Online]. Available:

http://www1.eere.energy.gov/vehiclesandfuels/pdfs/merit_review_2012/adv_combustion/ace047_maranville

_2012_o.pdf. [Accessed 06 11 2013].

[5] N. C. S. N. C. L. J. K. C. S. Junior, "Modeling a Thermoelectric HVAC System for Automobiles," Journal of

Electronic Materials, vol. 38, no. 7, pp. 1093-1097, 2009.

[6] D. C. D. a. L. J. Wang, "Design and Analysis of a Thermoelectric HVAC System for Passenger Vehicles,"

SAE International, Vols. 2010-01-0807, 2010.

[7] M. S. Raut and D. V. Walke, "Thermoelectric Air Cooling For Cars," International Journal of Engineering

Science and Technology (IJEST), vol. 4, no. 5, pp. 2381-2394, 2012.

[8] C.-Y. Hsu, S.-L. Li, C.-K. Liu, R.-M. Tain, H.-C. Chien, S.-F. Hsu, C.-M. Tzeng, M.-J. Dai, H.-S. Chu and

J.-D. Hwang, "Non-refrigerant thermoelectric air conditioning technique on vehicles," in Microsystems,

Packaging, Assembly and Circuits Technology Conference (IMPACT), 2011 6th International, Taipei, 2011.

[9] J. S. L. C. T. B. J. H. C.W. Maranville, "Improving efficiency of a vehicle HVAC system with comfort

modeling, zonal design, and thermoelectric devices," United States Department of Energy, 18 10 2012.

[Online]. Available:

http://www1.eere.energy.gov/vehiclesandfuels/pdfs/deer_2012/thursday/presentations/deer12_maranville.pd

f. [Accessed 06 11 2013].

[10] D. T. G. R. T. Barnhart, "Development of a Thermoelectric Device for an Automotive Zonal HVAC System,"

United States Department of Energy, 20 03 2011. [Online]. Available:

https://www1.eere.energy.gov/vehiclesandfuels/pdfs/thermoelectrics_app_2012/tuesday/barnhart.pdf.

[Accessed 06 11 2013].

[11] A. Attar, H. Lee and S. Weera, "Optimal Design of Automotive Thermoelectric Air Conditioner (TEAC),"

Journal of ELECTRONIC MATERIALS, vol. 43, no. 6, pp. 2179-2187, 2014.

19

[12] H. Lee, "Optimum Design," Applied Energy, pp. 115-123, 2013.

[13] H. Lee, Thermal Design: Heat Sinks, Thermoelectrics, Heat Pipes, Compact Heat Exchangers, and Solar

Cells, Hoboken: John Wiley & Sons, Inc., 2010.

[14] C. F. W. Zhimin, "The optimum thermal design of microchannel heat sinks," in Electronic Packaging

Technology Conference, 1997.

[15] S. L. L. Weera and H. Lee, "analytical performance evaluation of thermoelectric modules using effective

material properties," (Master's thesis) Western Michigan University, Kalamazoo, Michigan, 2014.