Sources and mitigation of CO and UHC emissions in low-temperature diesel

combustion regimes: Insights obtained via

homogeneous reactor modeling �

Paul Miles CRFSandia National Laboratories

Acknowledgements:

� Will Colban� (Sandia National Laboratories) CRF

� Seungmook Oh� (Korea Institute of Machinery and Materials)

� Sungwook Park, Rolf Reitz� (University of Wisconsin - ERC) ��

�� ��

�� �������

� ��

���� �� ��

����

�� �� ��

��

��

���� ��

���� ��

����

��

��

�� �� ��

��

����

��

��

��

KIMM

� Gurpreet Singh� (DOE OFCVT Program Manager)

DEER 2007 13th Diesel Engine Efficiency and Emissions Research Conference

August 13-16, Detroit, Michigan

Objectives

To assemble previous knowledge, fundamental underpinnings, and recent research results into a framework promoting clarity of thought

•� Review conventional wisdom regarding CO and UHC sources, clarifying

� � � � -� differences between heavy- and light-duty engines

� � � � -� the influence of dilution

•� Employ the results of homogeneous reactor simulations to clarify the impact of T, �, P and dilution on CO and UHC emissions

•� Review the tools / diagnostics available to clarify dominant sources of emissions, and summarize studies that have identified important sources under various operating conditions

For conventional HD diesel combustion, CO emissions stem from both fuel-rich and fuel-lean regions

At high load, CO is dominated by under- CO @ constant � & T , P = 60 bar, �t=2 ms , 21% O2

Soot/NOx contours from Kitamura, et al., JER 3, 2002 mixed fuel 4

15%

10%

Soot

5%

1%

500 ppm

NOx

5000 ppm

(fuel-rich regions)

� •� CO tracks soot

1 10

3

Equ

ival

ence

Rat

io

� •� CO decreased with

increased mixing � � - Increased Pinj � � - Optimized swirl

At light load, CO can be dominated by

CO

[% o

f fuel carb

on

]

50

40 2

30

20

over-mixed fuel (fuel-lean regions)

0� •� CO correlates with 600 1000 1400 1800 2200 2600 φig, inversely with Temperature [K] Tad, max

� •� Can be increased see also Park and Reitz (Comb. Sci and Tech, 2007) and Golovitchev,

with increased Pinj et al. (ICE2007) for detailed multi-dimensional calculations

60

0

21% O2

12% O2

With dilution, sources of CO change little

At low T, rich CO contours 1% by volume (roughly 200 g/Kg-fuel), P = 60 bar, t = 2.0 ms

Soot/NOx contours from Kitamura, et al., JER 3, 2002 4 mixtures pro­vide low CO:

15% SoFo

ot rmation

10%

TypLTC

ical path

5%

1%

500 ppm

CO

NOx

5000 ppm

Very little difference in oxidation behavior for T>1000 with changing dilution (EGR)

Conclusion:

Fuel mixed with

No combustion

(at a fixed, low temperature rich mixtures with EGR produce more CO) E

quiv

alen

ce R

atio

3

2

1

sufficient O2,0600 1000 1400 1800 2200 2600 CO production in early enough to

Temperature [K] the low temperature allow 2 ms at oxidation phase is Behavior of detailed T>1600K, will impeded by EGR simulations is well-captured burn (major species only) (Park & Reitz, Golovitchev et al.) completely

CO – Summary of recent LTC studies

Like conventional diesel combustion, CO emissions from low-temperature combustion systems can be dominated by either rich pockets (under-mixed fuel) or by lean pockets (over-mixed fuel) depending on load, EGR rate, SOI, etc. (2005-01-3837, 2006-01-0197, 2007-01-0193)

•� When under-mixed fuel dominates CO emissions, emissions can be reduced by

� -� Increased injection pressure (see also MTZ Sept 2003)

� -� Appropriate injection timing/targeting

� -� Formation of beneficial late-cycle flow structures

� -� Increasing boost (both � and Δ are increased) (see also 2006-01-3412, � � Colban, et al. SAE 2008)

•� When over-mixed fuel dominates CO emissions, emissions can be reduced by

� -� Increasing boost (increased reaction rates dominate)

� -� Reducing Pinj (2006-01-0076)

� �� Dilution has little influence on the completion of CO oxidation

Potential UHC sources include poorly mixed, fuel-rich regions, fuel-lean regions, and low temperature regions

Poor mixture formation

� -� Low injection velocity UHC @ constant � & T, P = 60 bar, �t=2 ms, 21% O2 Soot/NOx contours from Kitamura, et al., JER 3, 2002 � -� Large droplets 4

15%

10%

Soot

5%

1%

500 ppm

NOx

5000 ppm

� � (may delay mixing)

3 80

� -� Under- or over- UH

C [%

of fu

el carbo

n]

penetration

(e.g. Greeves, Khan, etc.)

Cold quench layers on

Equ

ival

ence

Rat

io

60

2 40

20 combustion chamber walls

1

0600 1000 1400 1800 2200 2600

Temperature [K] Over-lean mixtures

(formed during φig)

100

0

15%

10%

5%

1%

Soot

500 ppm

5000 ppm

NOx

�

�

4

3

Equ

ival

ence

Rat

io

dribble

For well-optimized HD engines, two UHC sources are generally thought to dominate

UH

C [%

of fu

el carbo

n]

Sac volume / hole

Lean mixtures formed during ignition delay

1000 1400 1800 2200 2600 Temperature [K]

600 0

100

80

60

2 40

20 1

0

Conventional strategies to reduce UHC:

� � Reduce sac volume

� Decrease φig � � � -� Increase compression ratio � � � -� Increase coolant temperature � � � -� Increase Tintake � � � -� Increase displacement

� � Optimize mixing � � � -� Nozzle hole number � � � -� Nozzle hole area � � � -� Swirl � � � -� Bowl design � � � -� Micro-holes - Not! (SAE 2005-01­� 0914)

� � Increase boost

� Increase cetane number (SAE 2004-01­�

? 1868)

Undesirable for LTC or high power density applications

15%

10%

5%

1%

Soot

500 ppm

5000 ppm

NOx

For light-duty engines, additional UHC sources become important

Increased importance due to smaller displacement

Liquid films

� -� Survive multiple Complete mixing more difficult due engine cycles to lower global � (high power density)

� -� Delay vaporization

4

100

2 40

�

1000 600

Equ

ival

ence

Rat

io

1

0

-� More significant for

� � re-entrant bowls

� •� Post-injection (not well understood) remain important � •� Close pilot injections at light load � •� Reduce spray asymmetry

� •� Optimize mixing (swirl, bowl design, targeting)

and mixing

UH

C [%

of fu

el carbo

n]

3 80 Additional strategies to Squish-volume 60

quench layers reduce light-duty UHC: •� Increase bowl diameter 20

high-squish,

Lean mixtures

0 •� Disrupt film formation 1400 1800 2200 2600 � (may impact high-load Temperature [K]

performance)

What happens when we add significant EGR?

Like CO, UHC yield is not strongly influenced by dilution

EGR promotes HC contours 1% of fuel carbon, P = 60 bar, t = 2.0 ms low temperature formation of non-fuel UHC - � Enhanced CO

production, too

EGR may reduce UHC emissions -� "re-burning"

recirculated UHC - � Enhanced low

temperature oxidation (see also JSAE 20030230)

4

3

600

Equ

ival

ence

Rat

io

2

1

0

15%

10%

5%

1%

Soot Formation

500 ppm

5000 ppm

NOx

Typical LTC path

21% O212% O2

non-fuel UHC

Soot/NOx contours from Kitamura, et al., JER 3, 2002

Low HC yield is observed from moderately rich mixtures

In the region of most interest, EGR has little influence

1000 1400 1800 2200 2600 (a small advantage

Temperature [K] to EGR is seen at For �<1 and moderate temper- the higher T, �)ature, EGR inhibits HC oxidation

Tools available to identify UHC sources

•� Cycle-resolved UHC measurements obtained via fast FID (2007-01-1837)

•� Load / parameter sweeps to identify controlling processes

•� Optical measurements

� -� Liquid film imaging (2007-01-1836)

� -� Imaging of in-cylinder UHC (2007-01-0907; Proc. Comb Inst. 31, 2963-2970)

•� Modeling predictions

� -� Homogeneous reactors

� -� Multi-dimensional simulations using detailed kinetics (Park & Reitz, CST 2007)

� -� Liquid impingement calculations

•� UHC emission behavior contrasted with CO emissions

•� Speciated UHC measurements

•� Fuel effects

Homogeneous reactor modeling studies help identify situations which result in high CO but low HC

HC=1% fuel carbon, CO = 1% by volume, P = 60 bar, 12% O2, Yield at 2 msSoot/NOx contours from Kitamura, et al., JER 3, 2002 4

15% SoFo

ot rmation

10% LTC elem

fuel ent 5%

path a 1%

Path b

Total HC 500 ppm

CO X NOx

5000 ppm

a)�Combustion near �=1 (high load or EGR rate)

� -� Even with high late-cycle mixing rates, there may be insufficient O2 available locally to achieve mixtures with � < 1

Equ

ival

ence

Rat

io

3

2

12% O2, P = 60 bar, � = 0.4,T=1400 K 1 0.03

0.00 UHC

CO

Mol

e Fr

actio

n [ ]

0.02

0600 1000 1400 1800 2200 2600

Temperature [K] 0.01

0 0.2 0.4 0.6 0.8 1b)� Quenching of lean mixtures during expansion Time [ms]

� - HCs can be consumed well before CO oxidation is complete

0

Conversely, high HC but low CO may be observed

Cold crevices regions can result in high HC with very

HC=1% fuel carbon, CO = 1% by volume, P = 60 bar, 12% O2, Yield at 2 msSoot/NOx contours from Kitamura, et al., JER 3, 2002 4

15%

10%

SoFo

ot rmation

5%

1%

Total HC 500 ppm

CO

NOx

5000 ppm

little CO...

...so can liquid films! 3

Equ

ival

ence

Rat

io

35 30 25 20 15 10 5H

C, C

O/1

0 [g

/kg-

fuel

]

CO

HC Boosted

Naturally aspirated

8 10 12 14 16 18 Intake O2 Mole Fraction

2

1

0Cool–to–moderate 600 1000 1400 1800 2200 2600 temperature, lean regions Temperature [K] can also lead to HC emissions with little CO

Speciation of HC emissions can provide information on the sources of UHC

nC7H16 [% fuel carbon]

CH2O [% fuel carbon]

CH4 [% fuel carbon]

Non-fuel UHC [% fuel carbon]

6.0

4.0

2.0

0.0

100

80

60

40

20

0

40

30

20

10

0

12

10

8

6

4

2

0

UHC summary

•� Like CO emissions, UHC can stem from either rich pockets (under-mixed fuel) or by lean pockets (over-mixed fuel). Direct in-cylinder imaging studies have shown either source can dominate, depending on conditions and engine geometry (2007-01-0907; Proc. Comb Inst. 31, 2963-2970; 2007-01-1836)

•� Unlike CO, UHC emissions can also stem from cold boundary layer or crevice regions. This source may also dominate under some conditions.

•� Homogeneous reactor simulations suggest that dilution has only a small effect on UHC emissions at a fixed T and �.

•� EGR may reduce UHC emissions stemming from cold regions (at the expense of increased CO emissions). There is little influence of dilution on UHC emissions in bulk gases.

•� Boost increases both chemical rates and mixing rates, and can reduce UHC stemming from over-lean as well as over-rich regions (Careful! There may be a soot penalty) (Colban, et al., SAE 2008)

Closure

•� There is no simple answer to the question "What is the dominant source of CO & UHC emissions in low-temperature diesel combustion systems. Under-mixed fuel, over-mixed fuel, and crevice/wall layers can all play important roles

•� Understanding and controlling the mixture formation process including

� � � � � - � Sufficiently rapid mixing of fuel-rich regions

� � � � � -� Avoidance of over-mixing

� � � � � -� Avoidance of HC in crevices and quench layers

� � � � � -� Avoidance of liquid films

� is key to the successful development of low-temperature diesel combustion systems