International Symposium on Modelling of exhaust-gas after-treatment
Bernd Amon, Herbert Albert, Faurecia
Johann Wurzenberger, Moritz Frobenius, AVL
Development of a SCR System for a
Dual Line Exhaust by Using Two-
Phase Flow CFD Calculations
2MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of the Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
3MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
4MODEGAT, 15. September 2009
Additional DeNOx• LP-EGR• NOx Storage Catalysts• SCR
Europe
NOx[g/km]
USA
0.2 0.3 0.4 0.50.1
JP 2005
JP 2002
JP 2009
PM [g/km]
0.02
0.03
0.04
0.01
0.05
NOx[g/km] 0.4 0.3 0.2 0.10.5
Japan
Tier 2 / Phase In (2004)
Tier II Bin 8 (2004)
Tier II Bin 5 (2009)
EU 3 (2000)
EU 4 (2005)EU 5(2009)
EU 6(2014)
PM [g/km]
0.04
0.03
0.02
0.05
0.01
EMISSION LIMITS – LDV DIESEL
5MODEGAT, 15. September 2009
NH3-SCR
NSC +Tier2 Bin5
Tier2 Bin8/ Euro 6
Euro 5
DeNOx efficiency
NSC
LP-EGR
Vehicle weight
MercedesE320 BT1
MercedesE320 BT2
VW Jetta US Mercedes R/ML/GL320
BMW335d
BMW330d
RenaultEspace 2l
BMW X5
Audi Q7Audi A4
VW Passat
announced
In series
• Primarily heavy vehicleswith big engines will useSCR technology
APPLICATION DeNOx - Technologies
6MODEGAT, 15. September 2009
Diesel
EDC
DPFDOC
SCRcatalyst
SCR SystemDosing system Tank
NOx/NH3Sensor
NOxSensor
Reducing agent
„AdBlue“(NH2) 2CO
Hydrolysiscatalyst
NH3-slipcatalyst
SCR mixer
Slow4 NH3 + 2 NO2 + O2 ⇒ 3 N2 + 6H2O
Fast2 NH3 + NO + NO2 ⇒ 2 N2 + 3 H2O
Standard4 NH3 + 4 NO + O2 ⇒⇒⇒⇒ 4 N2 + 6 H2O
AmmoniakAdBlue (urea/water)NH22HN
O
CNH3 + HO C N
Isocyanic acid
Thermolysis
+ H2ONH3 + CO2
Hydrolysis
SCR SYSTEM
7MODEGAT, 15. September 2009
NH3NOxNOx NH3
• To receive a high reduction rate on NOx by the SCR catalyst a very good distribution of the reducing agent on the catalyst surface is necessary
UNIFORM DISTRIBUTION OF “AD BLUE”
NOx NOx NH3NH3
• complete NOx-reduction (theor.)
• no NH3-slip
• NH3-slip
• lower NOx-reduction
or
8MODEGAT, 15. September 2009
Starting Design• Dual pipe• 2 SCR catalysts• Flex pipe in front of catalyst• No location for injection
DevelopmentConcept• Lean – use simulation/
prediction• Only final validation on
engine test bench
Optimized Packaging• Location for injection optimized• 1 SCR catalyst with optimized
geometry• Optimized outlet cone
Basic design
Find concept to use only oneAdBlue injection
SCR FOR V-ENGINES4 ltr - turbocharged
9MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
10MODEGAT, 15. September 2009
AdBlue injection in one pipe (basic 1)
AdBlue
• With one AdBlue injection valve the ammonia distribution on the catalystsurface is not sufficient (left)
• With one AdBlue injection valve the ammonia distribution on the catalyst surface is notsufficient (left)
• With two AdBlue injection valves the CFD calculation shows a better uniformity of ammonia (right), but at higher costs
AdBlue injection in two pipes (basic 2)
AdBlue
AdB
lue
ANALYSIS OF “BEST vs. WORST CASE”CFD Simulation - Results
11MODEGAT, 15. September 2009
Swirl Cone Design CFD-calculations (single-phase, steady state)
Distribution of passive scalar
Basic Swirl
DESIGN OPTIMIZATION AND VALIDATION
12MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
20MODEGAT, 15. September 2009
CFD-SCR-SIMULATIONFluid Properties and Multi-component evaporation and thermoysis
-1,5
-1,0
-0,5
0,0
0,5
1,0
1,5
0 20 40 60 80 100T (°C)
log
( ηη ηη /
cP)
Urea (Jaeger, 1965)
Water (Jaeger, 1965)
Urea (correlation)
Water (FIRE 8.x)
Temperature and composition dependentmodelling of
� liquid density, viscosity, specific heat
� solution vapour pressure
Properties for gaseous urea
� vapour heat capacity
� vapour dynamic viscosity
Spray / gas phase
� multi-component evaporation
� thermolysis
� hydrolysisH 0 (l)2
H 0 (g)2
H 02
(NH ) CO2 2
(NH ) CO (s or l)2 2
(NH ) CO (g)2 2
NH (g) + HNCO (g)3
(NH ) CO (s or l)2 2
(NH ) CO (g)2 2
21MODEGAT, 15. September 2009
CFD-SCR-SIMULATIONSpray and Wallfilm modelling
Advanced Spray-Wall Modells
� Multicomponent wallfilm evaporationand thermolysis
� Modelling of heat transfer betweendroplets and wall (Wruck-Meingast)
� Wall temperatur dependent splashingmodel (Kuhnke with BirkholdExtension)
� Accounting for heat conduction in solid walls via lateral heat conduction
4/12/1
4/54/3)(
dd
ddd uDK
µσρ=
Sat
Wall
T
TT =*
0 10050 150
0.8
0.9
1.0
1.1
1.2
1.3
1.4
Thermal breakup
Deposition Splash
Rebound
200
D. Kuhnke. Spray Wall Interaction Modelling by Dimensionless Data Analysis. PhD thesis, TU Darmstadt 2004.
N. W. Wruck. Transientes Sieden von Tropfen beim Wandaufprall. PhD thesis, RWTH Aachen, 1998.
22MODEGAT, 15. September 2009
CFD-SCR-SIMULATIONModel validation
Schwarzenberg, M. Untersuchung von Spraykonzepten zur Dosierung von Harnstoff-Wasser-Lösung beim Einsatz eines SCR-Verfahrens. Diplomarbeit, RWTH Aachen, 2005
F. Birkhold, U. Meingast, P. Wassermann, O. Deutschmann. Modeling and simulation of the injection of urea-water-solution for automotive SCR DeNOx-systems. Applied Catalyst B 70 (2007), 119-127
Lichtschnittvisualisierung Simulation
23MODEGAT, 15. September 2009
Heat conductivitySingle wall tube s/λ = 6·10-5 (m²K)/W
Decoupling element s/λ = 7.5·10-2 (m²K)/W
Flange s/λ = 5·10-2 (m²K)/W
Converter cover s/λ = 3.55·10-2 (m²K)/W
Cones s/λ = 8·10-5 (m²K)/W
Heat transfer outside αa = 10 W/(m²K)
Radiation
Emission coeff ε = 0.5 @ 30 °C
Simulation
Thermo camera
CFD-SCR-SIMULATIONEnergy
24MODEGAT, 15. September 2009
2mm wall
EKE
catalyst
1.5mm wall
Wall boundary conditions withheat transfer
Inlet right branch
Inlet leftbranch
Outlet
Spray
block-structured hex-mesh with 600491 cells
SCR-Cat:A=0.0188m²V=0.0045m³400cpsi/4.5milwashcoat 56µm
CFD-SCR-SIMULATIONMeshing
27MODEGAT, 15. September 2009
Basic Design Swirl DesignOP B
CFD-SCR-SIMULATIONResults
28MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
29MODEGAT, 15. September 2009
Swirl
Basis
Engine: Diesel V8, 4.2 ltr, turbochargedMax. power: 240 kW @ 3750 rpmMax torque: 760 Nm @ 1800...2500 rpm
Operating Points:
390160 [Nm]Load
700136[kg/h]Mass Intake Air
300
1700
A
450[°C]T before SCR
3200[rpm]Speed
BOperating point (OP)
ENGINE TEST BENCH INVESTIGATIONS
Dosing strategy
250 ms
108.3 ms
3.3 msOP A
OP B
30MODEGAT, 15. September 2009
Temp
Temp
TempTemp
Gas Pre
Gas Pre
GasPost
FTIRGas divider
Temp
AdBluevalve
AdBluevalve
ENGINE TEST BENCH SET-UP
31MODEGAT, 15. September 2009
Gas divider
cn
ccn
ii
21 1∑
=
−−=γ
Mapping: 48 NOx + NH3 sampling points
NOx (ppm)
550-600500-550450-500400-450350-400300-350250-300200-250150-200100-15050-1000-50
NOx DISTRIBUTION MEASUREMENTS
32MODEGAT, 15. September 2009
Basis Swirl
NOx (ppm)
80-9070-8060-7050-6040-5030-4020-3010-200-10O
P A
NOx (ppm)
550-600500-550450-500400-450350-400300-350250-300200-250150-200100-15050-1000-50
0.760.76
γγγγNOx = 0.69γγγγNOx = 0.69
NOx (ppm)
40-5030-4020-3010-200-10
NOx (ppm)
270-300240-270210-240180-210150-180120-15090-12060-9030-600-30
0.960.96
0.950.95
AdBlue
NOx (ppm)
50-6040-5030-4020-3010-200-10
NOx (ppm)
300-350250-300200-250150-200100-15050-1000-50
Basis 2
OP
B
0.840.84
0.850.85
AdBlue
AdB
lue
AdBlue
NOx DISTRIBUTION MEASUREMENTSResults
33MODEGAT, 15. September 2009
Simulation Measurement
0-75
75-150
150-225
225-300
300-375
375-450
450-525
525-600
NOx [ppm]
7
7x
8
CFD-SCR-SIMULATIONResults basic design– comparison with engine bench
34MODEGAT, 15. September 2009
Simulation
0-37.5
37.5-75
75-112.5
112.5-150
150-187.5
187.5-225
225-262.5
262.5-300
NOx [ppm]
Measurement
7
7x
8
CFD-SCR-SIMULATIONResults swirl design– comparison with engine bench
36MODEGAT, 15. September 2009
� Motivation
� Analysis of the Basic Design of Exhaust System
� Applied CFD Methods
� Test bed Verification
� Conclusions
CONTENT
37MODEGAT, 15. September 2009
� Advanced CFD models for ad-blue sprays and spray-wall interaction
have been applied � accurate prediction of NH3-uniformity in good
correlation with test bed results
� Successful development of a high performance SCR system for a dual
line exhaust system (V8-engine) � Swirl cone design allows a high
performance of the SCR catalytic converter
� In such exhaust system only one AdBlue injection is necessary to
achieve a sufficient ammonia distribution on the SCR catalyst �
Reduction of system costs and complexity of SCR control
� The application of accurate simulation-tools enables a fast and cost-
effective virtual system development
CONCLUSIONS
International Symposium on Modelingof exhaust-gas after-treatment
Development of a SCR System for a
Dual Line Exhaust by Using Two-
Phase Flow CFD Calculations
Bernd Amon, Herbert Albert, Faurecia
Johann Wurzenberger, Moritz Frobenius, AVL