See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/225604909

Thermal design of automobile exhaust based thermoelectric generators:

Objectives and challenges

Article  in  International Journal of Automotive Technology · March 2008

DOI: 10.1007/s12239-008-0020-y

CITATIONS READS

107 2,737

3 authors:

Khalid M. Saqr Mohamed Khamis Mansour

Tohoku University Beirut Arab University
123 PUBLICATIONS   1,158 CITATIONS    28 PUBLICATIONS   245 CITATIONS   

SEE PROFILE SEE PROFILE

Md Nor Musa
Universiti Teknologi Malaysia
55 PUBLICATIONS   295 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

DEEP SEA WATER View project

Blood hemodynamics investigations in Cerebral aneurysms and biological relevance View project

All content following this page was uploaded by Khalid M. Saqr on 29 June 2014.

The user has requested enhancement of the downloaded file.

International Journal of Automotive Technology, Vol. 9, No. 2, pp. 155−160 (2008) Copyright © 2008 KSAE
DOI 10.1007/s12239−008−0020−y 1229−9138/2008/039−05

THERMAL DESIGN OF AUTOMOBILE EXHAUST BASED

THERMOELECTRIC GENERATORS: OBJECTIVES AND CHALLENGES
K. M. SAQR , M. K. MANSOUR and M. N. MUSA
1)* 1) 2)

1)
Department of Thermo-Fluids, Faculty of Mechanical Engineering, Universiti Teknologi Malaysia, Malaysia
Research Management Center, Universiti Teknologi Malaysia, 81310 Skudai, Johor Bahru, Malaysia
2)

(Received 15 June 2007; Revised 1 December 2007)

ABSTRACT−The potential for thermoelectric power generation (via waste heat recovery onboard automobiles) to displace
alternators and/or provide additional charging to a vehicle battery pack has increased with recent advances in thermoelectric
material processing. In gasoline fueled vehicles (GFVs), about 40% of fuel energy is wasted in exhaust heat, while a smaller
amount of energy (30%) is ejected through the engine coolant. Therefore, exhaust-based thermoelectric generators (ETEG)
have been a focus for GFV applications since the late 1980s. The conversion efficiency of modern thermoelectric materials
has increased more than three-fold in the last two decades; however, disputes as to the thermal design of ETEG systems has
kept their overall efficiency at limited and insufficient values. There are many challenges in the thermal design of ETEG
systems, such as increasing the efficiency of the heat exchangers (hot box and cold plate), maintaining a sufficient temperature
difference across the thermoelectric modules during different operating conditions, and reducing thermal losses through the
system as a whole. This paper focuses on a review of the main aspects of thermal design of ETEG systems through various
investigations performed over the past twenty years. This paper is organized as follows: first, the construction of a typical
ETEG is described. The heat balance and efficiency of ETEG are then discussed. Then, the third section of this paper
emphasizes the main objectives and challenges for designing efficient ETEG systems. Finally, a review of ETEG research
activities over the last twenty years is presented to focus on methods used by the research community to address such
challenges.
KEY WORDS : Thermoelectric power generation, Automotive waste heat recovery, Energy management

1. INTRODUCTION
Any automobile ETEG typically consists of four compo-
nents: a hot box, thermoelectric modules, cold plate, and
assembly elements. The hot box is the component where
exhaust heat is to be extracted, the cold plate is responsible
for dissipating heat after it passes through the thermoelectric
modules, and the assembly elements are responsible for
applying sufficient compression force to these thermo-
electric modules as well as assembling all other compo-
nents in the desired combination (Vázquez et al., 2002). A
thermoelectric module contains a large number of N-type
and P-type semiconductors, arranged in couples, which can
convert heat directly into electric power (Rowe and Bhandari,
1983). Throughout this paper, the abbreviation “TEMs is Figure 1. Typical construction of ETEG. TEMs are
used to represent thermoelectric modules. Thermal insu- assembled between the hot box and two cold plates using
lation is another important component of any ETEG, and is assembly elements (compression springs and bolts).
used to decrease the thermal losses across the hot box; also,
thermal interface material (i.e., thermal grease) must be direction of working fluids are illustrated in Figure 1.
applied between the hot and cold surfaces and the TEMs to
overcome the thermal contact resistance between the two 2. ETEG HEAT BALANCE AND EFFICIENCY
surfaces and the TEMs. The typical construction and flow
In internal combustion engines, approximately 40% of the
*Corresponding author. e-mail: khaledsaqr@gmail.com fuel energy is wasted in exhaust gas, 30% is dissipated in

155
156 K. M. SAQR, M. K. MANSOUR and M. N. MUSA

on the design of the ETEG. However, a recent case study

suggested that out represents 45% of the total exhaust heat,
Q

while another 45% is represented in , , and together

Q2 Q5 Qo

Figure 2. Heat balance of a typical ETEG structure.
the engine coolant, 5% is lost as radiation and friction, and
25% is reserved for vehicle mobility and accessories (Yang,
(Rowe, 2005).
3. OBJECTIVES AND CHALLENGES
Increasing the overall efficiency is the main objective in
any ETEG design; this increase can be achieved by numer-
ous techniques. The thermal efficiency can be increased by
reducing the heat dissipated due to the thermal contact
resistance ( ), which can be implemented by applying a
Q5

2005). ETEG systems tend to recover heat from the first uniform compression to the TEMs and using highly con-
40%; however, due to many factors, only 5~6% efficiency ductive thermal interface material between the hot/cold
is available for today's ETEG technology (Kushch , et al. sides and the modules. Because the hot box and cold plate
2001). are normally electrically conductive, electrical insulation
The heat balance for the system is explained in Figure 2. should always be applied to the TEMs; ceramic wafers are
Q in and out are the exhaust gas energies at the inlet and exit
Q ideal for fulfilling the thermal interface and electric insu-
of the hot box, respectively, and is the effective heat
Qo lation requirements. A temperature drop of about 10~15 C o

transferred through the TEMs.
The heat losses in the system are expressed as
Equations (1) and (2)), and is the heat lost from the non-
Q1
used sides of the hot box by radiation and convection (i.e.,
heat transferred normal to the drawing plane). is the heat
can therefore be achieved across these wafers (Hi-Z Inc.,
(see QLosses 1996). Other solutions, such as silicon- and metal-based
grease, might not be suitable for ETEG applications
because of the high operating temperature of the modules
Q2 (> 200 C) as well as the cost considerations.
o

lost from the leg-sides of the TEMs by convection and The efficiency of the heat exchanger can be controlled
radiation, and is the heat lost by conduction through the
Q3 by many methods. When the hot box has a relatively large
assembly structure. The heat lost through gaps between the cross-sectional area, the exhaust gas velocity is damped
TEMs is , and is the heat lost by conduction in the
Q4 Q5 when entering the hot box channel. This sudden drop in gas
TEMs due to thermal contact resistance. The heat balance velocity forms a thick thermal boundary layer that causes
of the model in Figure 2 can be expressed as: the overall heat transfer coefficient to be sharply decreased.
Internal fins, turbulators, and corrugated surfaces are most
in− out = + Losses (1)
Q Q Q0 Q
favorable solutions for eliminating the effect of this bound-
Q Losses= + + + +
Q1 Q2 Q3 Q4 Q5 (2) ary layer. However, an important factor that should always
and the overall efficiency of the ETEG can be expressed as: be considered is the free cross-sectional area for the flow,
because it may cause the engine efficiency to decrease. To
η OV = η m × η HX × ρ (3) minimize conduction losses through the assembly, the
where the conversion efficiency of the TEMs is shown as: number of assembly elements contacting the hot box
directly should be minimized as much as possible and free
Generated Power -
η m = -------------------------------------------------------------- (4) surfaces should be strictly insulated.
Heat absorbed by TEMs The ratio (ρ) can also be increased by reducing the heat
loss through air gaps between the TEMs. This can be easily
η m = -----
Po
- achieved by using thermal insulation to prevent convection
Qo
from the hot box surface. Also, spray insulation with
and o is the electric output power from the ETEG.
P conductivity as low as 0.03 W/m.K is available and also
The efficiency of the heat exchanger can be expressed as: sustainable for high temperature applications.
The ability of the cold plate to provide sufficient cooling
Actual heat transferred -
η HX = --------------------------------------------------------------------------------- for TEMs at all operating conditions is another important
Maximum possible heat transfer thermal requirement in ETEG. Generally, the cold plate can
+ be classified to two types according to heat dissipation
η HX = ---------------------- (5)
Q0 Q4

Q in – Q out from the system (Saqr and Musa, 2007):

1. Radiator-based cold plate
ρ = ------------------ (6) 2. Heat sink-based cold plate
Q0

Q + Q 0 4
The percentage of each component of the heat loss, as well
as that of the useful amount of heat passing through TEMs
to the available heat entering the ETEG, are primarily based
In the first type, the cold plate dissipates exhaust heat to the
engine radiator through the engine coolant (i.e., water), while
in the second type the cold plate dissipates heat to the ambient
air through a heat sink; see Figure 3(a) and Figure 3(b).
 THERMAL DESIGN OF AUTOMOBILE EXHAUST BASED THERMOELECTRIC GENERATORS 157

1963 (Neild, 1963), but the most important research in this

area has been performed over the last twenty years. In this
section, three examples of ETEG research are discussed to
explain how the research community has addressed differ-
ent thermal design challenges, and to highlight the most
important developments in ETEG technology.
4.1. 1 KW ETEG
Hi-Z Inc. presented an ETEG for installation on diesel
trucks in 1992 (Bass et al., 1992). A total number of 72
HZ-13 modules were used to produce power in this ETEG.
Each of these modules contained 49 couples of hot pressed
bismuth telluride-based materials. The module size was 5.3
cm square by 0.5 cm thick and 82 g in weight. The hot and
cold side design temperatures were 230oC and 30oC,
respectively (Hi-Z Inc., 1996). The total length of the ETEG
was 48.26 cm with a total diameter of 22.86 cm and a total
weight of 13.6 kg. A hexagonal cross-section hot box was
used with aluminum radiator-based cold plates used as a
heat dissipater. The hot box was fabricated by welding flat
steel plates together to form a hexagonal shape; these steel
plates had a total number of 90 internal longitudinal fins to
enhance the heat transfer from the exhaust gases to the
modules. The TEMs were held against the hot box under a
compressive uniform load of 14 kg/cm2. This compressive
loading was applied using Belleville springs to ensure
minimum thermal contact resistance between TEMs and
heat exchanger surfaces (see Figure 4). The generator was
designed to produce 1 KW of electricity, but when the
experimental tests were performed on a 14 L Cummins
NTC 325 engine, the ETEG produced only 400 W, which
was less than half of the goal. It was concluded from the
temperature profile measurements that the reduction in
generated power was caused by a boundary layer problem
in the gas side, where the exhaust gas velocity decreases
according to the enlargement in the flow cross-section
(Bass et al. , 1995).
In order to resolve this problem, the number of fins was
reduced from 90 to 32 fins, while the length of each fin was

Figure 3. (a) Radiator based ETEG (b) ETEG with a heat

sink cold plate.
In the radiator-based cold plate (Figure 3(a)), the coolant
pump and piping usually require resizing to compensate for
the additional cooling load. Heat sink cold plates can be
efficient if the ETEG is to operate at low temperature
gradients; this technique takes advantage of the movement
of the vehicle to create turbulence in the air flow over the
cold plate.
4. REVIEW OF ETEG RESEARCH
The first ETEG for automobile applications was built in Figure 4. Cross section of Hi-Z ETEG.
158 K. M. SAQR, M. K. MANSOUR and M. N. MUSA

Table 1. Available experimental results from the 1 KW Hi-

Z ETEG.
Hot-side heat exchanger
Material Carbon Steel
Thermal Conductivity 50 W/m.K
Working fluid Exhaust gas
Inlet Temperature −1
Outlet temperature −
Cold-Side heat exchanger
Material Aluminum
Thermal Conductivity 204 W/m.K
Working Fluid Water
Inlet Temperature −
Outlet Temperature −
Heat Transferred from the hot side to −
the cold side of the ETEG ( 0+ Losses)
Q Q

Test engine Commins 14 L Figure 5. Schematic structure of Nissan ETEG.

Maximum power ( o) P 1068 W
Engine operating conditions at 1700 rpm modules were arranged in twelve blocks (M1 to M12), with
maximum ETEG power 300 hp each block of three modules separated by 10 mm; conse-
quently, the proportion of area occupied by modules to the
ETEG overall efficiency at maximum 1.3% faces of the inner shell was only 55%.
power test (ηOV) Measurements were performed under conditions corre-
TEM conversion efficiency (ηm) 4.5% sponding to the 60 km/hr hill climb mode of a 3000cc
Temperature difference across TEM gasoline engine vehicle. A combustor bench was utilized
surfaces at maximum ETEG power 250oC for simulating the engine exhaust at the desired operating
conditions, measuring temperature distribution in the gene-
rator, and evaluating the electric power of the generator.
increased to maintain heat transfer area. A discontinuity The exhaust gas flow produced by the combustor was
was introduced to the fin geometry by adding 0.953 cm divided into two lines. One line was connected to the gene-
gaps at 3.81 cm intervals. In addition, swirl fins were rator, and the other to the bypass line. The temperature and
installed in the hot box to ensure full turbulence of exhaust the flux of exhaust gas from the combustor is controlled by
gas. A summary of the Hi-Z 1 KW available experimental changing the ratio of air to fuel (A/F) and the flow rate of
data is presented in the following table: air blowing into the combustor. The flux of exhaust gas
through the generator is controlled by adjusting the orifice
4.2. 35.6 W ETEG diameter at the inlet of each line, as seen in Figure 6.
In 1998, Nissan Motors Inc published the result of experi- The experimental results are explained in the following
mental testing for a SiGe-based ETEG. The generator table:
contained 72 SiGe-based modules; each module produced
1.2 W at a temperature difference of 563 K between the hot
and cold surfaces of the module (Ikoma , 1998). The
et al.

ETEG overall dimensions were 440 × 180 × 70 mm3 and the

overall weight was 14.5 kg. The modules were mounted
between a hot box made of SUS304 alloy2 and two water-
cooled jackets made of aluminum, as seen in Figure 5. The
inner shell had a rectangular cross-section and two smooth
faces with which the hot side of the module came into
contact. In order to transfer as much heat as possible
through the modules, fins for heat exchange were formed
inside the inner shell in parallel with the gas flow. The
(−) Refers to non-available data
1 Figure 6. Schematic showing the combustor test bench
Chemical composition of SUS304: Fe, <0.08% C, 17.5~20%
2 used to simulate engine exhaust for the Nissan ETEG
Cr, 8~11% Ni, <2% Mn, <1% Si, <0.045% P, <0.03% S testing.
 THERMAL DESIGN OF AUTOMOBILE EXHAUST BASED THERMOELECTRIC GENERATORS 159

Table 2. Experimental results from the 35.6 W ETEG

combustor bench testing.
Hot-side heat exchanger
Material SUS304
Thermal Conductivity 21.5 W/m.K
Working fluid Exhaust gas
Inlet Temperature 592oC
Outlet temperature 527oC
Cold-Side heat exchanger
Material Aluminum
Thermal Conductivity 204 W/m.K
Working Fluid Water
Inlet Temperature 35oC Figure 7. The 300 W ETEG before testing.
Outlet Temperature 39.5oC
Heat Transferred from the hot side to (Tacher , 2006).
et al.

the cold side of the ETEG ( 0+ Losses) 4 KW A PCU (power conditioning unit) was used to adopt the
generated voltage to match the vehicle electrical system at
Q Q

Test engine 3.0 L Gasoline 12 and 24 V. Testing of the ETEG was performed on a V8
Maximum power ( o) P 35.6 W 270 hp gasoline engine from a GM Sierra pickup truck. A
Engine operating conditions at 60 km/hr pre-cooling heat exchanger was used to lower the inlet
maximum ETEG power Hill Climb Mode
ETEG overall efficiency at maximum Table 3. Experimental results from the 300 W ETEG road
power test (ηOV) 0.1% testing at 112.65 km/h.
TEM conversion efficiency (ηm) 1~2% Hot-side heat exchanger
Temperature difference across TEM Material Carbon Steel
surfaces at maximum power 123oC Thermal Conductivity 50 W/m.K
Working fluid Exhaust gas
Inlet Temperature (Case 1) 530.8oC
In their report, Ikoma reported that if η HX × ρ and
et al. Outlet temperature (Case 1) 421.2oC
η m could be increased from 5~10% to 50%, and from Inlet Temperature (Case 2) 617.3oC
1~2% to 5%, respectively, then the same ETEG would Outlet Temperature (Case 2) 484.6oC
produce 950 W at the same testing conditions. They have Cold-Side heat exchanger
emphasized that the key element in enhancing ETEG
efficiency is the thermal effectiveness of the heat ex- Material Aluminum
changers. Thermal Conductivity 204 W/m.K
Working Fluid Water
Inlet Temperature 86.7oC
4.3. 300 W ETEG
(Case 2)

Outlet Temperature (Case 2) 93.9oC

The experimental results from a 300 W ETEG were report- Inlet Temperature (Case 3) 77.6oC
ed in 2004. The project joined Clarkson University and Outlet Temperature (Case 3) 87.8oC
Delphi Systems and was funded by NYSERDA and the
DOE in order to develop a 300 W ETEG to be mounted on Heat Transferred from the hot side to −
a GM Sierra pickup truck. In order to achieve the power the cold side of the ETEG ( 0+ Losses)
Q Q

goal, a total of 16 HZ-20 thermoelectric modules were Test engine

used. Eight modules were mounted on each side of a Maximum power ( o) P 255.1 W
carbon steel hot box, and all modules were connected
electrically in series and thermally in parallel. The HZ-20 Engine operating conditions at 112.65 km/h
bismuth-tullerdie based modules generated 19 W at maximum ETEG power Climb up a grade
minimum if the temperature difference between the two of 7.2%
sides of the modules was 200oC. The module dimensions ETEG overall efficiency at 1.66%
were 75×75×5 mm, and the weight was 115 gm. The maximum power test (ηOV)
TEMs were assembled at a preload pressure of 200 psi TEM conversion efficiency (ηm) 2.8~2.9%
between the hot box and two aluminuim water jackets. The
overall dimensions of the ETEG were 330×273×216 mm Temperature difference across TEM 173.72oC
with a weight of 39.1 kg (Tacher , 2007). See Figure 7
et al.
surfaces at maximum power
 160 K. M. SAQR, M. K. MANSOUR and M. N. MUSA
temperature of the cooling water. The non-effective sides
of the ETEG were strictly insulated. Road testing was
performed at a vehicle speed of 48.28 km/h, 80.47 km/h,
and 112.65 km/h. The highest generated power and ETEG
efficiency were obtained at a vehicle speed of 112.65 km/h,
and the test results are reported in the following table:
In the ETEG test, three different cases were examined,
as explained in Table 3. In Case 1, the ETEG was installed
in the vehicle without the PCHX. In Case 2, the exhaust
pipe between the catalytic converter and the ETEG was
insulated with Therm-Tec™ exhaust wrap and aluminum
backed with high temperature glass fiber insulation. In
Case 3, the exhaust pipe was insulated and the PCHX was
actuated. In flat road testing, the ETEG produced 98.1 W at
Case 1, 130.8 W at Case 2, and 136.5 W at Case 3. A
maximum power of 255.1 W was measured on a 7.2%
graded road test just before PCU failure, which caused the
output power to drop precipitously. These results were all
measured at a vehicle speed of 112.65 km/h. The targeted
300 W power was never achieved by this ETEG during the
road test because of a PCU failure which caused the output
power to degrade.
5. DISCUSSION AND CONCLUSION
A thermal model for automobile ETEG systems was intro-
duced. This model shows the governing factors for enhan-
cement of the thermal efficiency of the ETEG through a
distribution of the available heat in the exhaust. These
governing factors were discussed through a detailed litera-
ture review which describes the most recent and important
research efforts during the last twenty years. There are four
main factors which control the thermal efficiency of any
ETEG:
(1) Heat exchanger geometry
(2) Heat exchanger materials
(3) The installation site of the ETEG
(4) The coolant system of the ETEG
The importance of the hot box geometry can be observed
in the increase of generated power in Hi-Z ETEG when the
number and geometry of fins were changed. Great signifi-
cance for thermal efficiency can be logically predicted if
more thermally conductive materials were used instead of
the low thermal conductivity steel alloys in the hot box
fabrication. When the ETEG is installed at sites closer to
the engine on the exhaust pipe, the possibility of generating
a greater temperature difference across the TEMs increases
significantly. The sensitivity of the ETEG generated power
to the coolant temperature can be observed with the install-
ation of PCHX to assist in reducing the water temperature
after the automobile radiator.
Future research should be focused on implementing new
materials with higher thermal conductivity in ETEG fabri-
cation. Additionally, innovative geometries which offer
more effective heat transfer elements, less outer surface
area, effective assembly elements, and a sufficient exhaust
View publication stats
flow channel should be investigated.
ACKNOWLEDGEMENT–The authors would like to acknow-
ledge the support provided by the Malaysian Ministry of Science,
Technology and Innovation (MOSTI) under ScienceFund VOT
79060.
REFERENCES
Bass, J., Campana, R. J. and Elsner, N. B. (1992). Thermo-
electric generator development for heavy-duty truck ap-
plications. Proc. Annual Automotive Technology Develop-
ment Contractors Coordination Meeting, Dearborn, USA,
743−748.
Bass, J., Elsner, N. B. and Leavitt, A. (1995). Performance
of the 1 kW thermoelectric generator for diesel engines.
Proc. 13th Int. Conf. Thermoelectrics B, Mathiprakisam,
edn., AIP Conf. Proc., New York, 295.
Hi-Z Inc. (1996). Use, Applications, and Testing of Hi-Z
Thermoelectric Modules. Hi-Z Inc: Online Catalog avai-
lable at: http://www.hi-z.com/how-to.htm
Ikoma, K., Munekiyo, M., Furuya, K., Kobayashi, M., Izumi,
T. and Shinohara, K. (1998). Thermoelectric module and
generator for gasoline engine vehicle. Proc. 17th Int.
Conf. Thermoelectrics. IEEE, Nagoya, Japan, 464−467.
Kushch, A. S., Bass, J. C., Ghamaty, S. and Elsner, N. B.
(2001). Thermoelectric development at Hi-Z technology.
Proc. 20th Int. Conf. Thermoelectrics, Beijing, China,
422−430.
Neild, Jr, A. B. (1963). Portable Thermoelectric Generators.
SAE-645A. Society of Automotive Engineers. New York.
Rowe, D. M. and Bhandari, C. M. (1983). Modern Thermo-
electrics. Reston Publishing Inc. Virginia. 93−100.
Rowe, D. M. Edn. (2005). Thermoelectrics Handbook.
Macro to Nano. CRC Press. Florida, 1. Section 52−2.
Saqr, K. M. and Musa, M. N. (2007). State of the art of
thermoelectric generators for waste heat recovery appli-
cations. J. Teknos-2k 7, 2, 22−30.
Thacher, E. F., Helenbrook, B. T., Karri, M. A. and Richter,
C. J. (2006). Progress in Thermoelectrical Energy Re-
covery from a Light Truck Exhaust. Paper Presentation at
the DEER Conf.. Available on: www1.eere.energy.gov/
vehiclesandfuels/pdfs/deer_2006/session6/
2006_deer_thacher.pdf
Thacher, E. F., Helenbrook, B. T., Karri, M. A. and Richter,
C. J. (2007). Testing of an automobile exhaust thermo-
electric generator in a light truck. Proc. I MECH E, Part
D: J. Automobile Engineering 221, 1, 95−107.
Va`zquez, J., Sanz-Bobi, M. A. Palacios, R. and Arenas, A.
(2002). State of the art of thermoelectric generators
based on heat recovered from the exhaust gases of
automobiles. Proc. 7th European Workshop on Thermo-
electrics, Pampelona, Spain.
Yang, J. (2005). Potential applications of thermoelectric
waste heat recovery in the automotive industry. Proc.
24th Int. Conf. Thermoelectrics. Clemson University,
U.S.A: IEEE, 155–160.