Vehicle Fuel Economy

Improvement through
Thermoelectric Waste Heat
Recovery
John LaGrandeur & Doug Crane, BSST LLC
Andreas Eder, BMW of North America, LLC

DEER Conference, Chicago, IL

August 25, 2005

marlow industries
 BSST Thermoelectrics Program

Program funded by the DOE- Freedom Car Office

One of four teams working to achieve 10% fuel

economy improvement through conversion of waste
heat to electric energy using thermoelectric power
generation

BSST has started the second of four phases and will

have a bench system operational in Q4 2006

BSST is developing a system that will be installed in a

BMW series 5 platform

2005 DEER Conference 2

 System Architecture

Base Radiator Aux Radiator

DC/DC
Pump

to valves
to pumps

Thermoelectric sensors Controller

Generator Module
(TGM)

Pump Field
Pump Controller

Primary
Catalytic Rear Exhaust
Engine Heat
Converter
Exchanger with Muffler
•Accessory
Drive Loads
Alternator •Existing and
New Electric
Motor Loads
Engine Coolant Flow
Visteon Responsibility
BMW Responsibility Control Signal Flow
BSST Responsibility High Electrical Power
Primary Coolant Flow Accessory Torque

Visteon

2005 DEER Conference 3

 Vehicle / Engine Selection
The selected vehicle is a
state-of-the-art BMW sedan
with a 3 liter displacement
engine (BMW 530i, MY 2006,
Selected vehicle platform
(BMW 530i, MY2006)
automatic transmission)

The engine is the newest

generation of highly efficient,
in-line, 6-cylinder engines
with characteristics
Selected engine platform (Inline representative of engines in
6 cylinder, 3.0L displacement) the 2010 to 2015 timeframe

2005 DEER Conference 4

 Exhaust Gas Temperature

Ref: Eder, Bertram, Liebl: Visions of Thermoelectrics in Vehicle Applications, DoEThermoelectricityWorkshop, San Diego, 2004

Exhaust gas temperatures for different engine speeds

(1000, 3000, 6000 rpm) measured in front of (P1) and
behind (P2) the catalytic converter
2005 DEER Conference 5
 Available Thermal Power in
Exhaust for FTP-75 Drive Cycle

2005 DEER Conference 6

 Performance Targets
Minimize the increase in exhaust gas pressure

No degradation of engine power or torque

Minimize weight and volume increase

Average Thermoelectric Generator Module (TGM)
electric output 700-1000W
• Electric output should be “clean” DC without noise

Positive influence on engine heat-up time (fuel

economy gain/emissions reduction) due to heat
transfer from the exhaust gas to the coolant

2005 DEER Conference 7

 Capturing Exhaust Gas
Waste Energy
Primary heat exchanger (PHx) designed to recover waste
heat from the exhaust gas (downstream from the catalytic
converter) and provide high quality thermal energy to the
TGM
PHx optimized to manage wide variations in exhaust gas
mass flow and temperatures
Backpressure created by the PHx is traded-off Vs heat
exchange effectiveness (subsequent analysis will take into
account muffler backpressure)
Preliminary design and modeling based on existing heat
exchanger
• Performance models validated
• Candidate designs identified for prototyping
2005 DEER Conference 8
 Primary Heat Exchanger
He/Xe working fluid
Shell & tube heat exchanger transports thermal
for exhaust gas heat transfer energy to TGM

Cat converter
Muffler

Exhaust gas bypass flow

2005 DEER Conference 9

 TGM Configuration
with Primary Control Loop
Improves efficiency
• Improves thermal impedance match with exhaust
gas
• Enables heat flux control
• Allows thermodynamic cycle optimization
Contains TE materials within a separate hermetically
sealed package
• Allows easier recycling of TE materials
• Compact design improves ruggedness
• Reduces TGM size, weight, and cost
Enables electrical load matching

2005 DEER Conference 10

 Current TE Materials & Segmented
TE Couple

P-type TE material N-type TE material

975 K
p-C eFe3R uS b12
n-C oS b3

p-TA G S
n-P bTe
300 K p-B iS bTe n-B i2Te3

Load
Ref: Modified from - http://www.its.caltech.edu/~jsnyder/thermoelectrics/ 11
2005 DEER Conference
 TE Power Generator Concept

1st generation TGM will

incorporate:
• Segmented
thermoelectric
elements
• Advanced TE
materials
• Effective ZT of 0.85

Hot fluid

Hot side heat exchanger

Hot side thermal and electrical links
Thermally conductive material Electrical path
Cold side thermal and electrical links
Cold side heat exchanger

Cold fluid
Thermoelectric element

2005 DEER Conference 12

 ADVISOR Development
Baseline BMW vehicle information implemented into
ADVISOR
Each individual subsystem validated and integrated into
ADVISOR
• PHX
– Heat transfer effectiveness is calculated for each time step
– Exhaust backpressure not included in model
• Power Converter
– Dynamic operation implemented
• Alternator
– Recalibrated ADVISOR to accept dynamic performance of
Visteon alternator
• TGM
– Extensive lookup table was created from in-depth BSST-
developed TGM model for faster code operation with equivalent
accuracy for each time step
2005 DEER Conference 13
 Model Validations
ADVISOR, originally developed by NREL and
commercialized by AVL, has been validated in a number of
independent tests and studies
Fuel economy results were validated against BMW-provided
data to within 2%
Exhaust temperatures validated against BMW-provided data
to within 5% with the trends captured
Alternator model derived directly from Visteon experimental
data
PHX model based on standard heat transfer correlations and
Visteon modeling experience
TGM model uses similar concepts to those for TE heating
and cooling models developed by BSST that have been
validated to accuracies of 5%
2005 DEER Conference 14
 Model Validation for TE Heating
and Cooling Devices
25 Stack Performance

4.0 Data: dTc=10 dTh=5

Tcin=Tcih=30
3.5
Data: dTc=10 dTh=20
3.0 Tcin=Tcih=40
Data: dTc=25 dTh=12.5
2.5 Tcin=Tcih=40
COPC

Data: dTc=25 dTh=30

2.0
Tcin=Tcih=35
1.5 Model: dTc=10 dTh=5
Tcin=Tcih=30
1.0 Model: dTc=10 dTh=20
Tcin=Tcih=40
0.5
Model: dTc=25 dTh=12.5
0.0 Tcin=Tcih=40
0.0 0.1 0.2 0.3 0.4 0.5 0.6 Model: dTc=25 dTh=30
Tcin=Tcih=35
ε

2005 DEER Conference 15

 ADVISOR Selection Screen

2005 DEER Conference 16

 Drive Cycle Simulation Results
Present (2005) System Capability Projected for Dyno Test, 2008 Target for Dyno Test, 2008

Drive cycle FTP-75 HWFET combined (1) FTP-75 HWFET combined (1) FTP-75 HWFET combined (1)

Average alternator load (W) 1000 1000 1000 2000 (2) 2000 (2) 2000 (2) 2000 (2) 2000 (2) 2000 (2)

Average ZT 0.85 0.85 0.85 1.00 1.00 1.00 1.25 1.25 1.25

% improvement - mpg 8.36 8.25 8.28 9.60 10.50 10.03 11.64 12.61 12.10

% change - HC (3) -1.67 0.19 -1.03 -2.19 0.58 -1.26 -2.77 0.58 -1.65

% change - CO (3) -1.86 -1.75 -1.82 -2.07 -2.65 -2.27 -2.53 -3.16 -2.75

% change - NOx (3) -2.99 -1.50 -2.53 -3.77 -1.48 -3.09 -4.25 -2.22 -3.64

(1) Combined drive cycle weighted 60% FTP-75 and 40% HWFET
(2) Increase in average alternator load is due to the estimated increase in electrification of vehicles by the year 2012
(3) Emissions results do NOT included significant reduction in emissions due to faster coolant warm-up

2005 DEER Conference 17

 Factors Not Currently Captured in
Simulation Results
Opportunities

• Effect of downsizing muffler

• Effect of downsizing the alternator
• Ability to run system without the use of power
conversion
• Further optimization of primary control loop flow for
maximized thermal management
Risks

• Exhaust backpressure effects

• Additional pump power losses
• Effect of potential unmatched load resistance
• Certain weight and drag coefficient changes
2005 DEER Conference 18
 Diesel vs. Gasoline
Engine Systems

3L BMW diesel and gasoline engines

2005 DEER Conference 19
 Engine Efficiency Comparison
At part load
operating modes
(0.2 kJ/liter, 2000
rpm), effective
engine efficiencies
are almost identical

Diesel engines
have higher heat
losses to the walls
of the cylinders at
this FTP-75
representative load
condition
Ref: Eder, Bertram, Liebl: Visions of Thermoelectrics in Vehicle Applications,
DoE Thermoelectricity Workshop, San Diego, 2004

17% higher fuel density in diesel fuel compared to gasoline

accounts for much of the difference in fuel consumption
between the two engine types
2005 DEER Conference 20
 Temperature Gradients in Different
Exhaust Systems
Diesel engine
exhaust
temperatures are
approximately
100C lower than
gasoline engine
exhaust at part
load and
approximately
200C lower at full
load.

Ref: R. Richter, BMW Group, Germany

2005 DEER Conference 21
 Potential in Diesel Engine Systems
A lower fraction of heat is rejected to the exhaust in
diesel compared to gasoline engine vehicles (22 – 35%
compared to 34 – 45%)1
A higher fraction of heat is rejected to the coolant in
diesel compared to gasoline engine vehicles (16 – 35%
compared to 17 – 26%)1

With lower temperatures and potentially less heat in the

exhaust system and more heat in the coolant system for
diesel engine vehicles, TE waste heat recovery from
these engines would be more focused on the following:
• Increased ZT for more low (100C – 250C) to medium (250C
– 500C) temperature TE materials
• Extracting waste heat from more components than just the
exhaust system (i.e. cooling system)

1. J. B. Heywood, Internal Combustion Engine Fundamentals, McGraw-Hill, Inc., 1988

2005 DEER Conference 22
 Commercialization
Vehicle-level
We estimate that TE power generation in gasoline
engine vehicles can be ready for production introduction
around CY2012 (MY2013).

• Implementation discipline of automotive OEMs requires a

~4-5-year lead-time prior to Start-of Production (SOP).

• Prototype vehicle performance results in the CY2010/11

timeframe will help determine the rate of expansion of
application beyond initial introduction.

2005 DEER Conference 23

 Conclusion
Around 10% improvement in vehicle fuel economy
through thermoelectric waste heat recovery in gasoline
engine vehicles in the next few years is achievable.

Although technical development may be slightly

different than that for gasoline engine vehicles,
significant fuel economy improvement through
thermoelectric waste heat recovery in diesel engine
vehicles is also achievable.

In Phase 2 (which has been funded with work

concurrently beginning), the team will iterate the model,
design, build and test a bench system to validate
analytically derived predictions in preparation for Phase
3, system integration, and Phase 4, engine level
integration and test.

2005 DEER Conference 24