Page 1
The Effects of EGR and Its Constituents on the Autoignition of Single- and Two-Stage Fuels
John Dec, Magnus Sjöberg, and Wontae Hwang Sandia National Laboratories
13th Diesel Engine-Efficiency and Emissions Research Conference August 13 – 16, 2007, Detroit, MI
Sponsors: U.S. DOE, Office of FreedomCAR & Vehicle Technologies
Program Manager: Gurpreet Singh
Saudi Aramco Oil Company Program Manager: Ahmar Ghauri
Page 2
Introduction
z Advanced LTC diesel and HCCI engines can provide both high efficiencies and very low emissions of NOX and PM.
z Nearly all LTC and HCCI strategies use high levels of EGR/residuals. 1.Reduce peak combustion temperatures to control NOX.
2.Delay the onset of ignition. ⇒Controlling combustion phasing. ⇒For HCCI-like diesel, it allows more time for premixing before ignition.
3. For gasoline HCCI, hot residuals are used to promote autoignition, and combined with cooled EGR, to control combustion phasing and NOX.
z Although EGR is widely used, its thermodynamic and chemical effects on autoignition are only understood in general terms.
z Objective: Provide a fundamental understanding of how EGR and its constituents affect the autoignition of single- and two-stage ignition fuels. – Conduct well-controlled experiments in an HCCI engine.
Page 3
HCCI Engine and Subsystems
Only premixed fueling is used for this study.
Cummins B 0.98 liter / cyl.
CR = 14 piston
EGR Systems z Real EGR z Simulated EGR &
EGR constituents
Page 4
Baseline Points40
330 335 340 345 350 355 360 365 370 Crank Angle [°CA]
iso-Octane Gasoline PRF80 PRF60, 29% CSP
CA50 = 368°CA
PRF60, 29% CSP PRF80 Gasoline iso-Octane
Baseline Points
EGR
C/F = 37.8 (φ = 0.40) CA50 = 368°CA
Single-stage Ignition Two-stage
Ignition (LTHR)
No EGR ≈29% CSP
Gasoline
TDC
z Establish baseline operating points for all four fuels.Î CA50 = 368°CA (8° aTDC)
35
30
25
20
15z Study the retarding effects of EGR 10
and its constituents. 1200
1100
z Iso-octane and gasoline exhibit single-stage autoig. at these conds.
1000
900
800
z PRF80 and PRF60 have two-stage 700
autoignition with LTHR. – Similar to diesel fuel (CN = 21 & 30).
180
z Baseline conditions have no EGR except for PRF60 ⇒ 29% CSP.
Operating Conditions Pin = 100 kPa (naturally asp.)
1200 rpm. C/F-mass ratio = 37.8, (φ = 0.40 without EGR.)
160
140
120
100
80
60
40
Pres
sure
[bar
]M
ass-
Ave
rage
d Te
mp.
[K]
Req
uire
d In
take
Tem
p. [°
C]
50 60 70 80 90 100PRF Number, Anti-knock Index
Page 5
Composition of EGR (CSP) z Consider complete combustion of iso-octane in "air".
(Argon and atmospheric CO2 lumped with N2.)
z C8H18 + 12.5 (O2 + 3.773 N2) ⇒ 8 CO2 + 9 H2O + 47.16 N2.
z For complete combustion with φ = 1, the gas composition (excl. fuel) changes from air: 21% O2 & 79% N2 (mole basis).
z to: 12.5% CO2, 14.0% H2O and 73.5% N2 for wet exhaust (CSP).
z to: 14.5% CO2 and 85.5% N2 for dry exhaust (Dry CSP).
z These gas compositions will be referred to as: Complete Stoichiometric Products Î CSP & Dry CSP.
Page 6
Thermodynamic Effects of EGR (CSP)
z EGR influences the compressed-gas temperature. – Heat-capacities of CSP & individual components are different from that of air.
z Start with motored operation. ⇒ Examine effects of CSP & its components on the mass-averaged temperature near TDC.
z Displace air while maintaining Pin = 100 kPa. 900
890
z Trends can be explained by 880
changes in specific heat (Cv). Te
mpe
ratu
re @
350
°CA
[K]
870
– CO2 has highest Cv (mole bs.). 860
– H2O high Cv. 850
– N2 ⇒ Cv slightly less than air. 840
– CSP between air and H2O. 830
– Dry CSP slightly lower Cv 820
than CSP. 1112131415161718192021
Intake O2 Mole Fraction [%]
N2 Dry CSP CSP H20 CO2
Motored Data
Increasing Diluent
Page 7
900
Autoignition Retard for iso-Octane
z Investigate the effect of these gases on combustion phasing.
z The autoignition retard is highly dependent on the type of the diluent.
z The retarding effect of the various gases is ordered consistently with their "cooling capacity" - thermodynamic effect.
z However, N2 addition increases the compression temperature, but the phasing is retarded, so there must be a weak chemical [O2] effect.
1112131415161718192021 Intake O2 Mole Fraction [%]
N2 Dry CSP CSP H20 CO2
Motored Data
10%
Bur
ned
[°C
A]
373
372
371
370
369
368
367
1919.2 19.4 19.6 19.8 2020.2 20.4 20.6 20.8 21 Intake O2 Mole Fraction [%]
CO2 H2O CSP Dry CSP N2
Tin = 175°C
Tem
pera
ture
@ 3
50°C
A [K
] 890
880
870
860
366
365 820
830
840
850
Page 8
365
366
367
368
369
370
371
372
373
16.51717.51818.51919.52020.521Intake O2 Mole Fraction [%]
10%
Bur
ned
[°C
A]
CO2H2OCSPDry CSPN2
iso-OctaneTin = 175°C
373
Temperature Traces for iso-Octane
1050
Air, Baseline N2 CO2 CA10
≈ 370.5°CA
CA10 ≈ 365.7°CA
CO2
N2
Air
z Retarding effect of lower [O2] becomes more significant at higher levels of N2 addition.– Clearly, reduction of [O2] has to
lead to misfire at some point.
z Temperature traces illustrate
1040
1030
1020
1010
1000 990 980 970 960
10%
Bur
ned
[°C
A]
Mas
s-A
vera
ged
Tem
p. [K
]
different mechanisms for retard. 950 350 355 360 365 370 375
Crank Angle [°CA]
16.5 1717.5 1818.5 1919.5 2020.5 21 Intake O2 Mole Fraction [%]
CO2 H2O CSP Dry CSP N2
iso-Octane Tin = 175°C
373
372
371
370
369
368
372
1919.2 19.4 19.6 19.8 2020.2 20.4 20.6 20.8 21 Intake O2 Mole Fraction [%]
CO2 H2O CSP Dry CSP N2
Tin = 175°C
10%
Bur
ned
[°C
A]
371
370
369
367
366
365 365
366
367
368
Page 9
900
Autoignition Retard for PRF80 (2-Stage Ignition)
z For PRF80, the retarding effects are not consistent with the "cooling capacity" of the added gases (except for CO2).
z N2 addition (reduced [O2]) strongly retards combustion phasing.
10%
Bur
ned
[°C
A]
Tem
p. R
ise
335
- 352
°CA
[K] 147
146
145
144
143
142
z PRF80's high sensitivity to [O2] occurs for two reasons:– First, a reduced [O2] reduces the
LTHR.
CO2 H2O N2 Dry CSP CSP
Tin = 72°C
PRF80, N2 2nd Order Curve Fit
890
1112131415161718192021 Intake O2 Mole Fraction [%]
N2 Dry CSP CSP H20 CO2
Motored Data
373
374
Tem
pera
ture
@ 3
50°C
A [K
]
880 372
371870
860 370
369
368
367
366 21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19
820
Intake O2 Mole Fraction [%]
830
840
850
Page 10
900
Autoignition Retard for PRF80 - [O2]z Second, a reduced [O2] also affects hot ignition.
1060
1040
1020
1000
980
960
940
920
900
880
Tem
pera
ture
@ 3
50°C
A [K
]M
ass-
Ave
rage
d Te
mp.
[K]
N2CO2
[O2] = 20.7%
[O2] = 19.3%
Post-LTHR temperature is the same.
10%
Bur
ned
[°C
A]
Tem
p. R
ise
335
- 352
°CA
[K] 147
146
145
144
143
142
CO2 H2O N2 Dry CSP CSP
Tin = 72°C
PRF80, N2 2nd Order Curve Fit
350 355 360 365 370 375Crank Angle [°CA]
890
1112131415161718192021 Intake O2 Mole Fraction [%]
N2 Dry CSP CSP H20 CO2
Motored Data
373
374
880 372
371870
860 370
369
368
367
366
21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19820
Intake O2 Mole Fraction [%]
830
840
850
Page 11
820
830
840
850
860
870
880
890
900
1112131415161718192021Intake O2 Mole Fraction [%]
Tem
pera
ture
@ 3
50°C
A[K
]
N2Dry CSPCSPH20CO2
Motored Data
400
500
600
700
800
900
1000
260 280 300 320 340 360 380Crank Angle [°CA]
Mas
s-A
vera
ged
Tem
p.[K
]
PRF80, N2
PRF80, H2O
880
900
920
940
960
980
1000
350 355 360 365 370 375Crank Angle [°CA]
Mas
s-A
vera
ged
Tem
p.[K
]
N2H2O
CA10 ≈ 371°CA
CA
10
All Constituents for PRF80 - H2O
z Thermodynamic "cooling" should add to [O2] effect for all constituents.
z Why is the retarding effect of H2O equal to N2?
z H2O enhances the intermediate chemistry, so thermal run-away occurs at lower temperature. – Almost perfectly counteracts the cooling effect of H2O!
z Also explains why CSP is less retarded than dry CSP.
z H2O does not enhance the intermediate chemistry for iso-Octane. 374
373
372
371
370
369
368
367
366
10%
Bur
ned
[°C
A]
880
900
920
940
960
980
1000
350 355 360 365 370 375 Crank Angle [°CA]
Mas
s-A
vera
ged
Tem
p. [K
]
iso-Octane, CO2 iso-Octane, H2O PRF80, N2 PRF80, H2O
PRF80 CA10 ≈ 371°CA
iso-Octane CA10 ≈ 370.4°CA
1919.2 19.4 19.6 19.8 2020.2 20.4 20.6 20.8 21 Intake O2 Mole Fraction [%]
CO2 H2O N2 Dry CSP CSP
Tin = 72°C
Page 12
366
367
368
369
370
371
372
373
131415161718192021Intake O2 Mole Fraction [%]
10%
Bur
ned
[°C
A]
N2 addition
366
367
368
369
370
371
372
373
131415161718192021Intake O2 Mole Fraction [%]
10%
Bur
ned
[°C
A]
N2 addition
366
367
368
369
370
371
372
373
131415161718192021Intake O2 Mole Fraction [%]
10%
Bur
ned
[°C
A]
N2 addition
Comparing [O2] Sensitivities z Have now identified three mechanisms by which EGR affects phasing. 1. Cv effect (thermodynamic – retarding). 2. O2-concentration effect (chemical – retarding). 3. H2O effect (chemical – enhancing).
z Compare side by side for all fuels.
z N2 addition gives most reduction in [O2] for least change in Cv.
373
372 z The two-stage ignition fuels
PRF80 and PRF60 are 371
much more sensitive. 370
369 z Lower [O2] both reduces
LTHR and counteracts 368
the hot ignition. 367
366 z Gasoline is initially very
insensitive to [O2].
10%
Bur
ned
[°C
A]
15.5 1616.5 1717.5 1818.5 1919.5 2020.5 21 Intake O2 Mole Fraction [%]
PRF60 PRF80 iso-Octane Gasoline Gas., 3rd order
N2 addition
Page 13
Comparing Thermal Sensitivities
1. Cv effect (thermodynamic – retarding). 2. O2-concentration effect (chemical – retarding).
3. H2O effect (chemical – enhancing).
z CO2 addition gives most thermodynamic cooling, with 373
least change in [O2].
20.2 20.3 20.4 20.5 20.6 20.7 20.8 20.9 21 Intake O2 Mole Fraction [%]
Gasoline iso-Octane PRF80 PRF60
CO2 addition 372
371
10%
Bur
ned
[°C
A]
z The single-stage ignition fuels,iso-octane and gasoline, aremuch more sensitive to thecooling effect of CO2.
370
369
368
367z The reason for this explained in:
Proceedings of the Combustion 366
Institute, Vol. 31, pp. 2895–2902, 2007.
Page 14
Comparing CSP Effects
1. Cv effect (thermodynamic – retarding).
2. O2-concentration effect (chemical – retarding).
3. H2O effect (chemical – enhancing).
z The overall effect of CSP results from the combination of these three mechanisms.
374 z CSP gives similar amounts of 373
retard for all four fuels. 372
z But underlying mechanisms 37110
% B
urn
Poin
t [°C
A]
are different. 370
369
z Trace species in real EGR 368 add to these effects. 367
366 19.2 19.4 19.6 19.8 2020.2 20.4 20.6 20.8 21
Intake O2 Mole Fraction [%]
iso-Octane PRF60, Shifted 6.1% PRF80 Gasoline
CSP addition
Page 15
Real EGR Effects
z Real EGR contains trace species 373 372
that also affect the autoignition. 371
– Unburned fuel, partially oxidized 370 369
fuel, formaldehyde, CO, and other 368
species. 367 366
z For iso-Octane and Gasoline, 365
372 real EGR retards less than CSP. 371 370
z For PRF80, real EGR has a 369
retards slightly more than CSP. 368 367 366
z The net effect of these species is 365
to advance the ignition for iso- 372 371
octane and gasoline... 370 369
z ...but to retard it for PRF80. 368 367 366
10%
Bur
n Po
int [
°CA
]CSP Real EGR
PRF80 Tin = 72°C
CSP Real EGR
Gasoline Tin = 149°C
CSP Real EGR
iso-Octane Tin = 175°C
21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19 Intake O2 Mole Fraction [%]
Page 16
Iso-Octane Gasoline PRF80 PRF60
± Effect: Single-stage Two-stage with LTHR
Cv effect (thermodynamic) Strong Weak Retarding
[O2] effect (chemical) Weak Strong
Enhancing H2O effect (chemical) None to weak Strong
Trace Species Moderate enhancing Moderate retarding
Iso-Octane Gasoline PRF80 PRF60
± Effect: Single-stage Two-stage with LTHR
Cv effect (thermodynamic) Strong Weak Retarding
[O2] effect (chemical) Weak Strong
Enhancing H2O effect (chemical) None to weak Strong
Trace Species Moderate enhancing Moderate retarding
Summary / Conclusions z EGR is very effective for suppressing autoignition reactivity.
– Can be used beneficially for controlling combustion phasing across load and speed ranges.
Iso-Octane Gasoline PRF80 PRF60
± Effect: Single-stage Two-stage with LTHR
Retarding Cv effect (thermodynamic) Strong Weak
[O2] effect (chemical) Weak Strong
Enhancing H2O effect (chemical) None to weak Strong
Trace Species Moderate enhancing Moderate retarding
z The net result of the different thermal, O2, and H2O sensitivities is a fairly similar effect of CSP for all fuels.
z Trace species (unburned and partially oxidized fuel, and CO) influence the effect of real EGR.
Details in: SAE 2007-01-0207
Page 18
900
Autoignition Retard for PRF80 - [O2] z Lower [O2] also affects hot ignition, in addition to reducing the LTHR.
Mas
s-A
vera
ged
Tem
p. [K
]
N2
CO2
[O2] = 20.7%
[O2] = 19.3%
1040
1000
960
920
880
840
800
10%
Bur
ned
[°C
A]
Tem
p. R
ise
335
- 352
°CA
[K] 147
146
145
144
143
142
CO2 H2O N2 Dry CSP CSP
Tin = 72°C
PRF80, N2 2nd Order Curve Fit
760
335 340 345 350 355 360 365 370 375Crank Angle [°CA]
890
1112131415161718192021 Intake O2 Mole Fraction [%]
N2 Dry CSP CSP H20 CO2
Motored Data
373
374
Tem
pera
ture
@ 3
50°C
A [K
]
880 372
371870
860 370
369
368
367
366
21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19820
Intake O2 Mole Fraction [%]
830
840
850
Page 19
820
830
840
850
860
870
880
890
900
1112131415161718192021Intake O2 Mole Fraction [%]
Tem
pera
ture
@ 3
50°C
A[K
]
N2Dry CSPCSPH20CO2
Motored Data
Autoignition Retard for Gasoline
z N2 addition shows that the chemical [O2] effect is very weak.
z Retarding effect of H2O is weaker than expected based on its high Cv.
z CSP (with water) has a slightly weaker retarding effect than dry CSP.
z Temperature traces confirm that H2O has an enhancing effect on autoignition for gasoline.
z Effect is stronger than for iso-octane, but much weaker than for PRF80.
350 355 360 365 370 375 Crank Angle [°CA]
CO2 H2O
Gasoline CA10 ≈ 370.3°CA
373 1020
CO2 H2O Dry CSP CSP N2
Gasoline Tin = 149°C
Mas
s-A
vera
ged
Tem
p. [K
] 372
371
370
1000
10%
Bur
ned
[°C
A]
980
369960
368
367
366
365
21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19 Intake O2 Mole Fraction [%]
900
920
940
Page 20
Real EGR Effects (2)z Intake concentrations increase with
EGR level and retard. 373 372
z These effects of Real EGR are also 371 370 important for explaining the effects of 369
retained residuals. 368 367
z Discussed in Proceedings of the 366 Combustion Institute, Vol. 31, pp. 365
CSP Real EGR
PRF80 Tin = 72°C
CSP Real EGR
Gasoline Tin = 149°C
CSP Real EGR
iso-Octane Tin = 175°C
2895–2902, 2007.
10%
Bur
n Po
int [
°CA
] 372
371
370
369
368
367
366
PRF80 iso-Octane Gasoline
Intake CO
iso-Octane PRF80 Gasoline
Intake HC
300
250
200
150
100
Inta
ke C
O -
[ppm
]
50 365 0 372
1200
1000
Inta
ke H
C -
[ppm
] 371
370
369
368
400 367 200 366
21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19 21 20.8 20.6 20.4 20.2 20 19.8 19.6 19.4 19.2 19 Intake O2 Mole Fraction [%]
Intake O2 Mole Fraction [%]
0
600
800
Page 21
Compare with NIST Database z All air displacement except N2 reduce the compression temperature. z Can be explained by changes of the specific heat capacity. z CO2 has highest specific heat capacity on a mole basis. z H2O has high specific heat capacity. z N2 has slightly lower specific heat capacity compared to air. z CSP falls between Air and H2O. z Dry CSP has slightly lower Cv compared to CSP.
Tem
pera
ture
@ 3
50°C
A [K
]
15
20
25
30
35
40
45
50
Hea
t Cap
acity
- C
v [J
/K*m
ol]
CO2 H2O CSP Dry CSP Air N2
NIST Chemistry WebBook
900
890
880
870
860
850
840
Motored Data
N2 Dry CSP CSP H20 CO2 830
820 300 400 500 600 700 800 900 1000 1100 1200 21 20 19 18 17 16 15 14 13 12 11
Temperature [K] Intake O2 Mole Fraction [%]