FINAL TECHNICAL REPORT

hammercoupleMécanique

22 févr. 2014 (il y a 3 années et 1 mois)

153 vue(s)


i

FINAL TECHNICAL REPORT

IEC Grant No.: 06
-
05a

Title:
NOx Emissions From Biodiesel Burned in Utility Generators


Modeling

Principal Investigator: Song
-
Charng Kong

Grantee Organization:
Mechanical Engineering Department, Iowa State University


______________
____________________

Signature of PI


Public Abstract


This study performed computer modeling of diesel/biodiesel combustion in large stationary
engines used for utility generation. This work is part of the IEC project awarded to Iowa
Association of Munici
pal Utilities on burning biodiesel in utility generators. The present
modeling study used stack test results to calibrate computer models that have been developed for
diesel combustion and emissions simulation. The purpose of this study is to use the model

as a
tool to explore NOx reduction strategies when biodiesel blends are used in large diesel engines.


Two engines were modeled in this study, including a Caterpillar engine

(2 MW)

in Winterset
facilities and a Fairbanks Morse engine

(2 MW)

in Story City
facilities
. After

model validation
,
parametric studies were performed to investigate effects of fuel injection timing and exhaust gas
recirculation on NOx emissions

reduction
.
Model results indicate that NOx can be reduced by
retarding the injection timing

or increasing exhaust gas recirculation rates.
On average, a 20 ~
25%
reduction in NOx
emissions can be achieved
by retarding injection timing
for

5
crank angle
degrees.
On the other hand, a

25 ~
35%

reduction in
NOx can also be achieved
per

10% exhaust
g
as recirculation rates.

Model results indicate that both strategies can be effective.
However,

exhaust gas recirculation requires extensive hardware modifications
, in particular for large diesel
engines. Therefore, retarding injection timing
appears to

be

a more feasible means to reduce
NOx emissions from burning biodiesel blends in large utility generators.


ii

Table of Conten
ts



1.

Introduction…………………………………………………………
……
……


1

1.1 NOx Emissions from Biodiesel Combustion ……………………


2

2.

Project Objective

……………………………………
………
…….
…………


2

3.

Computer Models ………………………………………
…….
………………


3

4.

Modeling Caterpillar Engine at

Winterset

Facilities ………………
..
……


5

4.1

Engine Geometry and Conditions ……………………….………


5

4.2

Model Calibration
…………………..

…………………………….

6

4.3

Parametric Study ………………………………
…………………..

8

5.

Modeling Caterpillar Engine at Winterset…………
…..
……………………


11

5.1

Engine Geometry and Conditions ……………………….………


11

5.2

Model Calibration …………………
……………………
………….

11

5.3

Parametric Study …………………………………………………..

14

6.

Summary ………………………………..…………..…………………………


16

7.

References……………………………………………
…………………
...




1
7


1

1.
Introduction


This work is part of the IEC project awarded to Iowa Association of Municipal Utilities

(IAMU)

on burning biodiesel in utility generators.
The present work is a modeling study to complement
the engine test that would be carried by IAMU at both Winterset and Story City municipal
facilities.


Particulate matters (PM) and nitrogen oxides (NOx) are the two major pollutants of diesel
engines and can cause adverse effects on the environment and hum
an health.
S
oot from diesel
engines
can cause

as much as a quarter of all global warming by reducing the ability of snow and
ice to reflect sunlight (
Sato et al., 2003).

On the other hand,
NOx is a major pollutant of concern
as a component of smog and acid

rain, contributing to the formation of ground
-
level ozone and
contributing to the greenhouse effect. When biodiesel is used in the engine, i
t is known that
biodiesel combustion produces less
PM

emissions but more NOx emissions.
Previously
published litera
ture has been focusing on laboratory engines or transit buses. Modeling study of
large diesel engines for utility generation is not available and is the subject of study in this work.


Biodiesel can displace
petroleum
-
based fuels and contribute to bioecono
my. However, the
increase in NOx emissions due to biodiesel combustion can limit its application.
In Iowa, NOx is
the limiting pollutant for stationary diesel generators. While generator permit types vary, a
typical Iowa Department of Natural Resources (ID
NR) air permit (Iowa Administrative Code,
2005), for example, allows a generator to operate up to a certain number of hours per year, or to
burn a certain number of gallons of fuel, based on the expected emissions of NOx predicted from
emissions factors (A
P
-
42 factors, EPA 2005) for specific fuels combined with manufacturer data
about engine performance. There are no emissions factors calculated for biodiesel blends and
there have not until recently been any emissions tests on large stationary engines used
for
electricity generation. The lack of emissions data from large diesel generators has been a
significant barrier to utilities wanting to burn biodiesel.


The present modeling study
is intended to help determine strategies to reduce NOx emissions
from uti
lity generators burning biodiesel blends. The present computer code consists of various

2

physical and chemistry models to describe the in
-
cylinder spray combustion process. Multi
-
dimensional simulation will be performed including the simulation of flow turb
ulence, spray
dynamics, vaporization, mixing and chemical reactions within the engine cylinder. This work
will use

stack test results to calibrate computer models that have been developed for diesel
combustion and emissions simulation.


1.1
NOx Emissions

from Biodiesel Combustion


It has been suggested that the increase in NOx is due to injection timing differences caused by
the low compressibility of biodiesel. Research that used spray chamber testing showed one crank
angle degree shift in using B100, i.e
., actual start of injection is earlier (Szybist and Boehman,
2003). The shift in injection timing resulted in an earlier ignition by 4 crank angle degrees that
caused a higher combustion temperature in the cylinder and produced more NOx emissions.
Other r
esearch indicates that the increase in NOx emissions is due to the lack of soot radiation
that causes a higher flame temperature in the cylinder when oxygenated fuel is used, such as
biodiesel (Mueller, 2005). In any case, various NOx reduction strategies
have been proposed
including retarding the injection timing setting, intake charge cooling, fuel additive and blending,
and the use of exhaust gas recirculation (EGR) to lower the combustion temperature (Yoshimoto
and Tamaki, 2001; McCormick et al., 2002;
Szybist et al., 2003). For instance, a recent study by
Minnesota Center for Diesel Research demonstrated that charge
-
air cooling (from 90 to 40
degree C) was very effective at reducing NOx emissions (by 25%) in a utility engine (Zarling et
al., 2004).


The

updated engine performance model will be used to suggest NOx reduction methods for the
test engines. The suggested methods will be examined to consider the feasibility of the engine
hardware.


2.
Project Objective


The goal
of the present modeling study
is to use the test data collected from the stack tests
,
performed by IAMU,

to calibrate the engine model. The c
alibrated model will be used to p
redict

3

effects of injection timing and related engine operating parameters on NOx reduction.

Potentially
the mod
el can also be used to predict emissions from other engine geometries in the future.


It is also hoped that the model results, together with stack test data, will

pr
ovide information to
utilities
and regulators
to

better evaluate their options for controll
ing NOx emissions.
We

disseminate the results from this project
through

research papers and technical meetings.


3.
Computer Models


The computer code consists of various models to describe the in
-
cylinder fluid dynamics
and
chemical reaction processes. T
he models have been developed

over the pasts few years

and have
been validated mainly for on
-
high vehicle engines

(Kong et al., 1995; Kong et al., 1999; Kong et
al., 2002)
. This study will further validate the model predictions using large diesel engines i
n
power plants that have different geometries and operating conditions.


The computer code was designed to simulate engine combustion that is a transient (due to piston
motion), multiphase (due to liquid spray and gas phases) and turbulent (due to high pis
ton speed)
process with chemical reactions.
The major model
s

include the spray atomization, drop
-
wall
impingement, wall heat transfer, piston
-
ring crevice flow,
combustion,
and
NOx and
soot
formation and oxidation models
(Kong et al., 1995; Kong et al., 19
99)
. The RNG
k
-


turbulence
model was used for in
-
cylinder flow simulations using the standard values for turbulence
parameters a
s those derived originally.


The combustion model uses multi
-
step chemical kinetics to describe the low
-
temperature
autoignition chemistry and

high
-
temperature flame chemistry that are both important to diesel
engine.
The
reaction
chemistry and flow solutions were coupled

by using a characteristics time
approach
.
The turbulence also affects the combustion by property transport, wall heat flux, e
tc.
Details of the model can be found in the original literature

(Kong et al., 1995)
.


The extended Zeldovich NO mechanism was used to simulate the NO emission process. The
effect of OH radical on NO formation is also considered through the reactions, as
listed below
.


4

By invoking the steady
-
state assumption for N radical and using equilibrium approach for O, H
and OH radicals, the formation rate of NO can be obtained as in the following.

O + N2 NO + N

N + O2 NO + O

N + OH NO + H

d
dt
NO
k
O
N
NO
K
O
N
k
NO
k
O
k
OH
f
b
f
f




R
S
|
T
|
U
V
|
W
|
2
1
1
1
2
2
12
2
2
1
2
2
3
/
/
d
i
O + N2 NO + N

N + O2 NO + O

N + OH NO + H

d
dt
NO
k
O
N
NO
K
O
N
k
NO
k
O
k
OH
f
b
f
f




R
S
|
T
|
U
V
|
W
|
2
1
1
1
2
2
12
2
2
1
2
2
3
/
/
d
i




(1)


Soot emissions
are

predicted
using

a phenomenological soot
model
.
Two competing processes
are

considered in this model,
namely
soot formation and oxidation. The rate of
change of
soot
mass
s
M

within a computational cell is determined from the soot formation rate
sf
M


and soot

oxidation rate
so
M

.


dt
dM
dt
dM
dt
dM
so
sf
s











(
2
)


The formation rate uses an Arrhenius expression and the oxidation rate is based on a carbon
oxidation
mode
l, described as




0.5
exp
sf
f fv
dM
A M P E RT
dt
 







(3
)


6
so c
s Total
s s
dM Mw
M R
dt D


.









(4
)


Based on the current combustion model,

fuel was used as the s
oot inception species in Eq. (3
).
The pre
-
exponential constant
A
f

was adjusted accordingly for the present implementation and
also to account for the fuel effects
, i.e., r
egular diesel vs. biodiesel blends
. On the other hand, the
s
oot oxidation rate is determined by the Nagle
-
Strickland
-
Constable

model that considers
carbon
oxidation by two
reaction pathways

whose rates depend on surface chemistry
of

two different
reactive

sites, as in Eq. (4).


5


The combustion and emissions chemistry in the present model
was

calibrated

based on the test
data to account for the characteristics of biodiesel

blends
. The
calibrated was then

used to
perform parametric study

for NOx reduction.


4.

Modeling Caterpillar Engine at Winterset Facilities


4.1
Engine Geometry and Conditions


The engine is a Caterpillar 3516B engine. The specification is listed in Table 1.
The baseline
condition has a start
-
of
-
injection (SOI) timing of

10

ATDC (after top
-
dead
-
center). The
parametric study includes sweeps of EGR and SOI. The engine wa
s run at full load conditions
with 1
,
825 kW output. The fuel injector has 8 nozzle holes such that the computational
domain

only use
s

1/8 of the entire cylinder to
take advanta
ge of symmetry, as indicated in Fig. 1.

The
computational domain contains one fuel spray plume as will be seen in the following figures.



Table 1 C
aterpillar 3516B engine specifications

Bore × Stroke

170 mm × 190 mm

Compression Ratio

14:1

Displacement

6
9

L

for 16 cylinders

Connecting Rod Length

392.5

mm

Engine

power and

speed

185 kW at
1800 rpm

BSFC

204.7 g/kW
-
hr

Modeling cases
:

(1) Baseline

(2
) 0% EGR

(3
) SOI=

5 ATDC


0% EGR,
SOI=

10

ATDC

SOI=

10,

5, 0, 5 ATDC

EGR=0, 5, 10, 15%




6

Computational
domain
Cylinder head
Piston surface
Computational
domain
Cylinder head
Piston surface

Figure 1
Com
bustion chamber and the c
omputational domain
in this study.


4.2
Model Calibrations

for Baseline Case


The present three
-
dimensional computational fluid dynamics (CFD) model is capable of
simulating fuel spray combustion and emissions inside the engine cyl
inder. Figure 2 shows the
distributions of liquid fuel spray and combustion temperature on a cross section in the cylinder.
The fuel spray experiences atomization and vaporization, and the fuel vapor mixes with air such
that combustion occurs in the cylind
er. It is seen that the fuel spray is surrounded by the
combustion flame as expected in the real engine combustion process.

The predicted spray and
combustion processes seem very realistic as compared with existing literature.



Figure 2

Model results of
in
-
cylinder fuel spray and combustion temperature
distributions on a cross section inside the combustion chamber. The fuel spray is
surrounded by the high temperature combustion flame.




7

Soot
NOx
TDC
20 ATDC
50 ATDC
Soot
NOx
TDC
20 ATDC
50 ATDC

Figure 3

Model results of soot and NOx distributions on a cross sec
tion for a sequence
of times.
The cylinder axis is located at the left boundary of each plot and o
nly half of
the
cylinder cross section is shown.


The present model also includes the chemical kinetics for NOx and soot formation. Figure 3
shows the distrib
utions of soot and NOx on the cross section for a sequence of timing. It is seen
that the NOx forms at the region where there is high combustion temperature since NOx is
sensitive to temperature. Soot is formed in the leading edge of the spray plume where
the fuel
-
air
mixture is rich. The soot cloud is then moved by the in
-
cylinder flow in a counterclockwise
direction to the squish region later in the engine cycle. Part of the in
-
cylinder soot will oxidize
later in the engine cycle, as will be seen in Fig.
4.


Since the experimental emissions data are available for B0, the model was further calibrated for
this baseline condition. Figure 4 shows the history of in
-
cylinder soot and NOx emissions in
lb/MMBTU to be consistent with the engine data. The predicted
emissions at exhausts valve

8

open (approximately 120 ATDC) are compared with the engine
-
out stack measurements. In the
present model formulation, the kinetics rate constant for soot formation was adjusted such that
the engine
-
out PM emissions was matched. O
n the other hand, no model constants were adjusted
in the NOx model while the NOx emissions was already well predicted. More engine conditions
using biodiesel blends will be modeled once the experimental data are available.


0
1
2
3
4
5
-100
-50
0
50
100
NOx (lb/MMBTU)
Crank Angle (ATDC)
Engine Data
Model Prediction

0
0.01
0.02
0.03
0.04
0.05
0.06
-100
-50
0
50
100
PM (lb/MMBTU)
Crank Angl e (ATDC)
Engi ne Data
Model Prediction

Figure 4

History of in
-
cyl
inder soot and NOx emissions. The predicted emissions at
exhausts valve open are compared with the engine
-
out stack measurements.


4.3
Parametric Study


Parametric study was performed to investigate effects of EGR and SOI
on NOx and soot
emissions. Note th
at using EGR or retarding SOI are the most common and effective methods to
reduce NOx emissions.

This is because the

two
methods can reduce the combustion temperature
inside the cylinder leading to

lower

NOx emissions.



Figure 5

shows the predicted NOx a
nd soot emissions, cylinder pressure history and peak
cylinder pressure for various SOI conditions. It is seen that, as injection timing is retarded, the
cylinder pressure d
ecreases resulting in lower NOx emissions but higher soot emissions due to
lower co
mbustion temperatures.
An average of 20% reduction in NOx emissions can be achieved
by retarding injection timing for 5 crank angle degrees.



9


-70
-60
-50
-40
-30
-20
-10
0
10
-10
-5
0
5
10
NOx Variation (%)
SOI (ATDC)

-50
0
50
100
150
200
-10
-5
0
5
10
Soot Variation (%)
SOI (ATDC)

Cylinder Pressure and Heat Release Rate vs. Crank Angle
CAT 3500 Full Load 1800 rpm
0% EGR
0
5
10
15
-60
-40
-20
0
20
40
60
Crank Angle (dATDC)
Cylinder Pressure (Mpa)
0
250
500
750
1000
1250
1500
Heat Release Rate (J/deg)
SOI 5 EGR 0%
SOI 0 EGR 0%
SOI-5 EGR 0%
SOI-10 EGR 0%

Peak Cylinder Pressure vs. SOI
CAT 3500 Full Load 1800 rpm
0% EGR
0.0
5.0
10.0
15.0
-10.0
-5.0
0.0
5.0
10.0
SOI (dATDC)
Peak Cylinder Pressure (Mpa)

Figure 5

Predicted NOx and soot emissions, cylinder pressure history and peak cylinder
pressure for 0%
EGR with various SOI.


On the other hand, EGR is another effective means of reducing NOx emissions. The EGR rate is
defined as the ratio of re
-
inducted exhaust gas to the total intake charge.

intake
[%] 100
EGR
m
EGR
m
 








(5)

It is an indication of ho
w much exhaust gas is re
-
inducted to the intake. If exhaust gas is not re
-
inducted into the intake, the EGR rate will be zero. It is known that the specific heat (C
p

and C
v
)
of H2O and CO2 (exhaust) are higher than O2 and N2 (fresh air). If part of the exh
aust gas is
directed to the intake and displacement some of the fresh air, the overall specific heat of the in
-
cylinder mixture will be higher. Since the energy release from combustion is the same, a higher
specific heat will result in lower gas temperatur
e rise during combustion. Therefore, NOx
production can be reduced due to low combustion temperature.



10


-60
-50
-40
-30
-20
-10
0
10
0
5
10
15
20
NOx Variation (%)
EGR (%)

-10
0
10
20
30
40
50
0
5
10
15
20
Soot Variation (%)
EGR (%)

Cylinder Pressure and Heat Release Rate vs. Crank Angle
CAT 3500 Full Load 1800 rpm
SOI -5 dATDC
0
5
10
15
-60
-40
-20
0
20
40
60
Crank Angle (dATDC)
Cylinder Pressure (Mpa)
0
250
500
750
1000
1250
1500
Heat Release Rate (J/deg)
SOI-5 EGR 0%
SOI-5 EGR 5%
SOI-5 EGR 10%
SOI-5 EGR 15%

Peak Cylinder Pressure vs. EGR
CAT 3500 Full Load 1800 rpm
SOI -5 dATDC
0.0
2.0
4.0
6.0
8.0
10.0
12.0
0.0
5.0
10.0
15.0
20.0
EGR (%)
Peak Cylinder Pressure (MPa)

Figure 6

Predicted NOx and soot emissions, cylinder pressure history and peak cylinder
pressure for SOI=

5 ATDC with various EGR.


Figure 6 s
hows the predicted engine data of using different levels of EGR. Results are consistent
with the understanding of EGR that NOx emissions decrease with increased EGR.
Approximately a 35% reduction in NOx by increasing EGR by 10% was predicted. It is of
inte
rest to note that NOx emissions decrease significantly while the engine cylinder pressure
does not decrease noticeably. Also note that the increase in soot emissions by using EGR is not
as significant as retarding the injection timing.


Note that the stack

test data for B20 and B100 are recently available and the modeling is
currently being performed for the two biodiesel blends. Modeling results will be presented in the
final report as the parametric study data are available for B20 and B100.



11

5.
Modeling
Fairbanks Morse Engine at Story City Facilities


5.1
Engine Geometry and Conditions


The present computer model was also used to simulate the combustion and emission process of
the Fairbanks Morse engine in Story City facility. The engine geometry and oper
ating conditions
are provided by Fairbanks Morse Engine. The engine is a two
-
stroke engine and each cylinder
has two
opposed
piston
s compressing against each other, as shown in Fig. 7. The combustion
chamber is at the middle section of the cylinder where t
he fuel injectors are located. Each
cylinder has two fuel injector

that produce

hollow
-
cone sprays. The
entire
fuel injection process
was modeled as well as the combustion process.


Top piston
Bottom piston
Injector #1
Injector #2
Top piston
Bottom piston
Injector #1
Injector #2

Figure 7 Geometry and computational mesh of the Fairbanks Morse engine.


5.2
Model Calibrations


The model was calibrated for the baseline case (
720
rpm, 0% EGR).

A sequence of combustion
images with in
-
cylinder details (fuel spray distribution and combustion temperature) are shown
in Fig. 8. It is seen that the two injectors

issue fuel spray from the side the cylinder wall toward
the center. Some of the liquid drops impinge on the piston surface while pistons are compressing
against each other. Auto
-
ignition occurs at the leading edge of the spray plumes and temperature

12

rises

causing liquid fuel continuing to vaporize.
Soon after ignition, the entire fuel spray is
engulfed in the flame. As time goes by, two flames merge and combustion spreads across the
majority of the combustion chamber. The model predicts that combustion is
still confined in the
center of the combustion chamber with the present two
-
injector setup,
as indicated by the high
temperature contours.


Figure 9 shows the predicted in
-
cylinder pressure and temperature histories. The ignition occurs
near top
-
dead
-
cente
r and the peak cylinder pressure occurs at the early stage of piston expansion.
The predicted NOx and soot histories are also shown in Fig. 9. The model was calibrated using
the stack test results. Comparisons of measured and predicted NOx and soot emissio
ns for B0,
B20 and B100 are shown in Fig. 10. The soot formation rate constant in Eq. (3) is adjusted for
different fuel blends.



13





(No c
ombustion yet)



(Combustion has started)










(Combustion of two fuel spray merges and
spread over t
he combustion chamber.)


Figure 8 Predicted fuel spray and combustion temperature in the cylinder at different
views for a sequence of timing.



14


0
1
2
3
4
5
6
0
500
1000
1500
2000
-100
-50
0
50
100
FM Engine
Pressure
Temp
Cylinder Pressure (MPa)
Gas Temperature (K)
Crank Angl e (TDC)

0
5
10
15
20
25
30
35
40
0
10
20
30
40
50
60
70
-40
-20
0
20
40
60
80
100
FM Engine
NOx
Soot
NOx (g/kg-f)
Soot (g/kg-f)
Crank Angl e (TDC)

Figure 9

Histories

of in
-
cylinder
pressure, average combustion

temperature
, NOx and
soot emissions
.


0
0.5
1
1.5
2
2.5
3
0
20
100
FM Engine
NOx - Test
NOx - Model
Emissions (lb/MMBTU)
Biodi esel Bl end (%)

0
0.05
0.1
0.15
0.2
0.25
0.3
0
20
100
FM Engine
Soot - Test
Soot - Model
Emissions (lb/MMBTU)
Biodi esel Bl end (%)

Figure 10
Comparisons of
test data and model prediction for

NOx and soot emissions
after model calibration
.


5.3
Parametric Study


After model calibration for different fuel blends, p
arametric study was performed to investigate
effects of EGR and SOI o
n NOx and soot emissions

for B20 and B100
.
Results are shown in Fig.
11
.


By dela
ying the fuel injection timing from

5 ATDC to 0 ATDC, approximately 25% reduction
in NOx can be obtained for both B20 and B100. Further retarding injection timing can benefit


15

NOx reduction for B20 but no further NOx reduction was predicted for B100. On the other hand,
for using EGR,
an average of 25% reduction in NOx

per 10% EGR can be achieved. Such effects
of EGR on the reduction in NOx are similar for both B20 and B100.


Al
though EGR seems to be more effective and consistent in reducing NOx emissions for B20
and B100, the increase in soot emissions is more significant for EGR cases. In addition,
considering the requirement in hardware modification to implement the EGR system
, retarding
the injection timing by 5 crank angle degrees seems to be a more feasible means for NOx
reduction for burning biodiesel blends without significant penalty of soot emissions.


-60
-50
-40
-30
-20
-10
0
10
20
-5
0
5
10
B20
Emission Variation (%)
SOI (ATDC)
Soot increase
NOx reducti on

-40
-20
0
20
40
60
-5
0
5
10
B100
Emission Variation (%)
SOI (ATDC)
Soot increase
NOx reducti on

-100
-50
0
50
100
150
0
5
10
15
20
25
30
35
B20
Emission Variation (%)
EGR (%)
Soot increase
NOx reducti on

-100
-50
0
50
100
150
0
5
10
15
20
25
30
35
B100
Emission Variation (%)
EGR (%)
Soot increase
NOx reducti on


Figure 11 Predicted NOx and soot emissions, cylinder press
ure history and peak
cylinder pressure for SOI=

5 ATDC with various EGR.



16

6.
Summary


The present computer

model
s were validated

by emissions test data

for a Caterpillar engine and a
Fairbanks Morse engine used in municipal facilities.

Trends of soot and
NOx emissions for
different biodiesel blends are predicted by the model. P
arametric studies were performed to
investigate effects of fuel injection timing and exhaust gas recirculation on NOx emissions
reduction.


Model results indicate that NOx can be re
duced by retarding the injection timing or increasing
exhaust gas recirculation rates
because both methods can

create a lower combustion temperature
environment

that can reduce NOx formation
. For the Caterpillar engine studied, an average of
20% reduction
in NOx emissions can be achieved by retarding injection timing for 5 crank angle
degrees. On the other hand, a 35% reduction in NOx can also be achieved per 10% exhaust gas
recirculation rates. For the Fairbanks Morse engine, by delaying the fuel injection

timing from

5
ATDC to 0 ATDC, approximately 25% reduction in NOx can be obtained for both B20 and
B100. In using exhaust gas recirculation, an average of 25% reduction in NOx per 10% EGR can
be achieved but with significant penalty in soot emissions.


M
odel results indicate that both NOx reduction strategies can be effective.
However,

exhaust gas
recirculation requires extensive hardware modifications to the intake and exhaust manifolds, the
cost may be significant, in particular for large diesel engines
. Therefore, it is concluded that
retarding injection timing would be a more feasible means to reduce NOx emissions from
burning biodiesel blends in large utility generators.


17

7.
References


EPA, Research Triangle Park, North Carolina, Compilation of Air P
ollutant Emissions Factors,
AP
-
42 Fifth edition, Volume 1, stationary point and and area sources, January 2005.
Website:
http://www.epa.gov/ttn/chief/ap42/index.html

Iowa Administrative Code, Se
ction 567, Chapters 20
-
32.
Iowa Administrative Rules for Air
Quality, 2005.

Kong
,
S.C.
,

Han
, Z. and
Reitz,

R.D.

“The Development and Application of a Diesel Ignition and
Combustion Model for Multidimensional Engine Simulations,”
Journal of Engines
, Vol. 10
4, pp.
502
-
518 (1995).


Kong
, S.C.

and
Reitz,

R.D.

“Multidimensional Modeling of Diesel Ignition and Combustion
Using a Multistep Kinetics Model,”
J. Engng Gas Turbines Power
, Vol.101, pp.781
-
789 (1993).


Kong
, S.C.
,

Senecal
, P.K.

and
Reitz,

R.D.

“Developm
ents in Spray Modeling in Diesel and
Direct
-
Injection Gasoline Engines,”
Oil & Gas Science and Technology


Rev. IFP
, Vol. 54, No.
2, pp. 197
-
204 (1999).


McCormick, R.L., Alvarez, J.R., Graboski, M.S., Tyson, K.S. and Vertin, K. “Fuel Additive and
Blendin
g Approaches to Reducing NOx Emissions from Biodiesel,” SAE 2002
-
01
-
1658, 2002.


Mueller, C., Private Communication, Combustion Research Facilities, Sandia National
Laboratory, Livermore, CA, 2005.

Sato, Mki., J. Hansen, D. Koch, A. Lacis, R. Ruedy, O. Dub
ovik, B. Holben, M. Chin, and T.
Novakov.
Global atmospheric black carbon inferred from AERONET
.
Proc. Natl. Acad. Sci.

100
, 6319
-
6324, doi:10.1073/pnas.0731897100, 2003.

Szybist, J.P
. and A. L. Boehman “Behavior of a Diesel Injection System with Biodiesel Fuel,”
SAE 2003
-
01
-
1039, 2003.


18


Szybist, J.P., Simmons, J., Druckenmiller, M., Al
-
Qurashi, K., Boehman, A.L. and Scaroni, A.
“Potential Methods for NOx Reduction from Biodiesel,” SAE 2003
-
01
-
3205, 2003.


Yoshimoto, Y. and Tamaki, H. “Reduction of NOx and Smoke Emissions in a Diesel Engin
e
Fueled by Biodiesel Emulsion Combined with EGR,” SAE 2001
-
01
-
0649, 2001.


Zarling, D.D., Bickel, K.L., Waytulonis, R.W. and Sweeney, J.R. “Improving Air Quality by
Using Biodiesel in Generators,” SAE 2004
-
01
-
3032, 2004.