icef2016-9349€¦ · emissions standards for new engines and rely less on advanced in-cylinder...

Proceedings of the ASME 2016 Internal Combustion Engine Division Fall Technical Conference

ICEF2016

October 9-12, 2016, Greenville, South Carolina, USA

1 Copyright © 2016 by ASME

ICEF2016-9349

EFFICACY OF ADD-ON HYDROUS ETHANOL DUAL FUEL SYSTEMS TO REDUCE NOX

EMISSIONS FROM DIESEL ENGINES

Jeffrey T. Hwang, Alex J. Nord, William F. Northrop*

University of Minnesota

Department of Mechanical Engineering

Minneapolis, MN 55455

ABSTRACT

Aftermarket dual-fuel injection systems using a variety of

different fumigants have been proposed as alternatives to

expensive after-treatment to control NOX emissions from

legacy diesel engines. However, our previous work has shown

that available add-on systems using hydrous ethanol as the

fumigant achieve only minor benefits in emissions without re-

calibration of the diesel fuel injection strategy. This study

experimentally re-evaluates a novel aftermarket dual-fuel port

fuel injection (PFI) system used in our previous work, with the

addition of higher flow injectors to increase the fumigant

energy fraction (FEF), defined as the ratio of energy provided

by the hydrous ethanol on a lower heating value (LHV) basis

to overall fuel energy. Results of this study confirm our earlier

findings that as FEF increases, NO emissions decrease, while

NO2 and unburned ethanol emissions increase, leading to no

change in overall NOX. Peak cylinder pressure and apparent

rates of heat release are not strongly dependent on FEF,

indicating that in-cylinder NO formation rates by the

Zel’dovich mechanism remains the same. Through single zone

modeling, we show the feasibility of in-cylinder NO

conversion to NO2 aided by unburned ethanol. The modeling

results indicate that NO to NO2 conversion occurs during the

early expansion stroke where bulk gases have temperature in

the range of 1150 -1250 K. This work conclusively proves that

aftermarket dual fuel systems for fixed calibration diesel

engines cannot reduce NOX emissions without lowering peak

temperature during diffusive combustion responsible for

forming NO in the first place.

INTRODUCTION

Diesel engines are known for reliability, durability, low

manufacturing cost and high power density. Given their

longevity, legacy diesels regulated to older emissions levels

will continue to be used in practice for decades to come. New

diesel engines have ~27% lower NOX emissions than engines

of a decade ago [1] in part due to selective catalytic reduction

(SCR) aftertreatment systems. Although aftertreatment is an

effective method for reducing emissions, in-cylinder

techniques are also attractive to reduce SCR urea dosing rate

requirements or to possibly eliminate the need for NOX

aftertreatment altogether.

In-cylinder NO is primarily formed during combustion

through a combination of chemical pathways including the

extended Zel’dovich, prompt (Fenimore) and N2O

mechanisms [2–4]. NO in diesel engines mainly arises through

the thermally controlled Zel’dovich mechanism in lean to

stoichiometric regions found near the periphery of the

diffusive flame front. NO is oxidized to NO2 and

concentrations “freeze” short of thermodynamic equilibrium

soon after the end of injection and mixing of burned gases [5]

in the expansion stroke.

Low temperature combustion modes like dual fuel

reactivity controlled compression ignition (RCCI) have been

shown to simultaneously limit in-cylinder NOX and soot

production over a wide speed and load range [6–8]. RCCI uses

fumigation of a low reactivity fuel, like gasoline into the

intake manifold and early direct injection of a high reactivity

fuel like diesel to avoid high temperature NOX formation

regions found in conventional diesel combustion. Various

fumigants have been investigated for RCCI including

hydrogen, gasoline, hydrous ethanol, and natural gas [9–14]

and all have similar impacts on avoiding in-cylinder NOX

formation.

Although RCCI is an attractive method for in-cylinder

emissions reduction, it must be implemented in new engines

due to the requirement for significant modifications to engine

hardware and software. To date, manufacturers have chosen to

employ NOX aftertreatment like SCR to meet stringent

emissions standards for new engines and rely less on advanced

in-cylinder techniques like dual fuel RCCI.

For legacy diesel engines regulated to older emissions

standards, add-on SCR and lean NOX trap aftertreatment

systems have been marketed to meet new in-use NOX

regulations [15]. Dual fuel retrofit kits are available that also

claim to reduce NOX emissions without aftertreatment while

also substituting diesel fuel for lower carbon fuels like

compressed natural gas or partially renewable ethanol [16,17].

These aftermarket systems incorporate a separate fuel system

and fumigate the secondary fuel directly into the intake


plumbing, retaining the stock engine calibration and hardware.

Therefore, these systems cannot achieve the emissions

reductions possible with RCCI since they do not change the

diesel fuel injection strategy. Fumigation for diesel engines

has a long history [18]. Previous work has examined

advancing diesel injection timing of a mechanically injected

engine with fumigation of hydrous ethanol [19] to increase the

diesel replacement quantity but this strategy is not applicable

to modern electronically controlled diesel engines.

Our previous work investigated using hydrous ethanol as

the fumigant in an add-on configuration both with an existing

commercial fumigation system [16] and with a novel port-

injection system [20]. This work along with other published

literature on aftermarket dual fuel systems [9,13,21,22] show

that NOX is not appreciably reduced with increased fumigant

energy fraction (FEF), defined as the ratio of fumigant lower

heating value to overall input fuel lower heating value. Others

have shown that when the water content of hydrous ethanol

exceeds 50%, NOX can be mitigated through intake charge

cooling, lowering peak combustion temperatures in the

diffusive flame [2,23].

Although overall NOX emissions do not significantly

decrease for add-on dual fuel systems, our work and research

by others have shown that the NO2 to NOX fraction increases

with increasing FEF [2,14,16,20]. This suggests that NO

formation during combustion remains unchanged for dual fuel

operation but that NO oxidation to NO2 occurs with more

fumigation.

Increased unburned hydrocarbon emissions also increase

with FEF, which have been implicated in the NO to NO2

conversion process. HO2 radicals formed within the cylinder

from the oxidation of intermediate species such as CH3CHOH

have been shown to be responsible for the conversion of NO

to NO2 [3,4]. At high temperatures, HO2 is unstable and

unable to react, however as temperatures decrease, the HO2

radical becomes stable and begins to promote the reaction of

NO to NO2. Hori et al. illustrated a kinetic mechanism by

which hydrocarbons facilitate the conversion of NO to NO2 at

temperatures between 600 and 1200 K [24]. As in-cylinder

temperatures are generically higher than engine exhaust

temperatures, it can be concluded that the NO to NO2

conversion via unburned hydrocarbons occurs within the

cylinder during the expansion stroke. Though hydrocarbon

assisted NO conversion chemistry has been studied with light

hydrocarbons like methane and ethane, ethanol has not been

investigated.

The work presented here provides a thorough set of

performance and emissions data for an add-on hydrous ethanol

port injection dual fuel diesel engine covering a larger range

of FEF and hydrous ethanol water content than our previous

work. It also investigates the effect of unburned hydrocarbons

on in-cylinder NO to NO2 conversion through comparison of

experimental data to a single zone kinetic model.

EXPERIMENTAL

The objective of the experimental work was to investigate

a hydrous ethanol dual fuel PFI system over a large range of

engine operation using varying hydrous ethanol water content

and FEF. A John Deere 4045HF475 Tier 2 diesel engine was

used in the experiments. The specifications of the engine and

PFI system are shown in Table 1.

The same custom PFI fuel rail from our previous work

was used in this study [20]. The PFI rail was integrated into

the existing intake manifold, and incorporated two

automotive-grade fuel injectors aligned to spray directly in

between the intake ports of the cylinders.

Table 1: Engine Specifications

Manufacturer/Model John Deere 4045HF475

Engine Type 4-Stroke DI Diesel

Cylinders 4, in-line

Displacement (L) 4.5

Bore x Stroke (mm) 106 x 127

Compression Ratio 17.0:1

Maximum Power

(kW/rpm) 129/2400

Aspiration Turbocharged & After

Cooled

Diesel Injection System Common Rail

Ethanol Injection System Port Fuel Injection

Ethanol Heating System None

Emissions Certification EPA Tier 2 (Off-Highway)

IVO (CAD ATDCF) 339

EVC (CAD ATDCF) 380

Hydrous ethanol injections were controlled using the

signal from the manufacturer-installed camshaft sensor.

Analysis of the signal provided engine speed and the location

of cylinder one top dead center (TDC) to a National

Instruments (NI) cRIO controller. Hydrous ethanol injection

pulse width and timing were then output by the cRIO to each

injector. Each injector injected twice per two engine rotations

starting at 360 CAD ATDCF, 21 CAD after IVO, to partially

mitigate fuel bypass from positive valve overlap inherent to

this engine. The hydrous ethanol was stored in a secondary

container and pumped to the PFI rail at constant flow. A

digital scale was used to determine the time rate change of

mass during a given testing duration for hydrous ethanol while

diesel fuel flow was measured using a CUB5 series

mechanical fuel flow meter.

A laminar flow element (LFE) was used to measure

intake airflow rate, and after-cooler outlet temperature was

maintained between 40 and 50 °C using an air-water heat

exchanger. Heated intake air was required to ensure complete

combustion of the charge due to ethanol’s high latent heat of

vaporization.

Gaseous emissions were measured using an AVL Fourier

Transform Infrared Spectrometer (FTIR), while soot emissions

were measured using an AVL Micro-Soot Sensor (MSS).

Engine exhaust was first diluted at a ratio of 5-7 in a residence

chamber with compressed air before being measured by the

MSS. The FTIR sampled both raw exhaust and diluted

exhaust, where the ratio of CO2 emissions before and after


dilution was used as a measurement of dilution ratio at every

testing point. The experimental setup is given in Figure 1.

In addition to engine performance and emissions data,

high-speed in-cylinder pressure data was collected at each

testing condition. Kistler Type 6065A pressure transducers

were mounted in custom Kistler 6542Q128 glow plug adapters

for all four engine cylinders. A 0.1 CAD resolution BEI H25

incremental optical encoder was mounted to the engine crank

to trigger data acquisition using a NI BNC-2110 and PXI-6123

DAQ. A NI LabVIEW interface was programmed to sample

100 cycles for a total of 720,000 data points per cylinder.

Apparent heat release rate was calculated using a custom post-

processing code and a first law analysis as outlined in ref.

[25].

Figure 1: Diagram of engine test setup

Hydrous ethanol with 10% (180 Proof) and 20% (160

Proof) water by volume were used in this study. Both proofs

of hydrous ethanol have a lower distillation energy to LHV

ratio during the refining stages and do not require dehydration,

thereby significantly increasing the renewability of the fuel

[26]. Hydrous ethanol blends were mixed by volume with lab

grade non-denatured anhydrous ethanol and distilled water.

The primary direct injected fuel used was non-oxygenated #2

ultra-low sulfur diesel (ULSD). The experimental testing plan

consisted of operating the engine over a modified type C1 off-

road vehicle ISO 8178 eight point testing cycle with and

without hydrous ethanol PFI [27]. The testing modes are

shown in Table 2.

Table 2: Modified ISO 8178 engine operation conditions

Mode Engine Speed

[rpm]

Engine Load

[N-m]

BMEP

[bar]

1 2400 450 12.6

2 2400 350 9.77

3 2400 250 6.98

4 2400 50 1.40

5 1400 450 12.6

6 1400 350 9.77

7 1400 250 6.98

8 1000 0 (idle) 0.00

At each testing mode, the engine was first allowed to

reach steady state diesel fuel only combustion. The PFI system

was then toggled “on” for PFI of hydrous ethanol and data was

taken once emissions, temperature, and pressure data reached

steady state. The stock engine ECU was not modified in any

way; all diesel injection parameters followed the OEM

calibration. Data was collected at intervals of two minutes at

steady state operation and then averaged for reported results.

Ethanol injector pulse width was varied to increase FEF,

while engine load was held constant by varying the engine

pedal position, effectively decreasing diesel fuel flow to

accommodate increased load during PFI operation. After data

collection for the selected testing mode was completed, the

PFI system was toggled “off” and cycled to the next testing

mode under conventional diesel combustion (CDC).

Experiments were conducted over a four-day period, where

the eight-point test cycle was conducted in its entirety on each

day for 160 proof, 180 proof, selected repeats (160 and 180

proof), and CDC respectively. FEF was determined using the

time rate change of mass of hydrous ethanol and measured

diesel fuel flow rate at each testing condition in conjunction

with respective LHV values. The ratio of ethanol energy input

over total energy input was then calculated as the FEF, where

maximum FEF corresponds to the maximum pulse width

achieved at each testing condition.

An uncertainty analysis was conducted using standard

deviations of measurements during steady state and between

repeated data sets. Error bars on result figures are based on the

root mean square value of two times the standard deviation,

representing the 95% confidence interval. Propagation of error

calculations were estimated using the numerical sequential

perturbation approach [28]. Systemic error was small

compared to standard deviation error and was only used for

hydrous ethanol time rate change of mass.

RESULTS AND DISCUSSION

Performance and Emissions

The diesel engine equipped with the PFI hydrous ethanol

dual fuel system was operated over a range of injector pulse

widths for each operating mode with 160 proof, 180 proof,

and diesel-only modes. Performance results for the max FEF

achieved at each condition are given in Table 3. All values

were calculated on a diesel equivalent basis. Combustion

efficiency (CE) and air/fuel ratio (AFR) decreased very

slightly with increasing FEF for all conditions. The decrease

in CE can be attributed to charge cooling effects from the

latent heat of hydrous ethanol vaporization and the increased

amount of water being introduced to the engine. BTE

decreased with increasing FEF for most cases, but increased

slightly for a few 180 proof cases. This behavior is coupled

with the brake specific fuel consumption (BSFC), which

increased with increasing FEF for most cases, but decreased

for the same 180 proof cases with increased BTE.

All testing modes were stability limited, defined as when

an increase in injector pulse width resulted in unstable

combustion or audible engine knock. 160 proof hydrous

ethanol testing modes were able to reach higher injector pulse

widths as compared to 180 proof, before the onset of knock.


The higher knock tolerance of 160 proof ethanol can be

explained by the increased charge cooling from higher water

content. A significant barrier for diesel fuel replacement at

low engine speeds exists as illustrated in Table 3 especially at

high load (Modes 5 and 6) where the stability limit was

reached between an FEF of 23-25%.

Figure 2 gives the in-cylinder pressure and calculated

apparent rate of heat release (RoHR) for Mode 3 (2400 rpm, 7

bar BMEP) for selected FEF and 180 proof hydrous ethanol.

As FEF increased, the premixed heat release event and peak

pressure increased in magnitude. Combustion phasing, as

measured by CA50 advanced slightly with increasing FEF,

while burn duration, defined as CA90 – CA05 decreased with

increasing FEF. For example, for the cases shown in Figure 2,

CA50 advanced from 16.4 DATDC to 13.5 DATDC, while

burn duration decreased from 27.5 CAD to 20.0 CAD. These

trends are primarily due to the increased premixed portion of

combustion with increased ethanol. The bimodal shape of the

RoHR reflects amplified premixed and diffusion portions of

combustion. This has been previously explained by the

combustion of directly injected diesel fuel providing ignition

energy for the combustion of the premixed fumigant [29].

Unlike with RCCI that uses advanced diesel injection timing

to generate a primarily premixed combustion event, the

mixing controlled mode of heat release remained present with

increasing FEF.

Table 3: Engine performance parameters at maximum

FEF achieved

Mode Operation

Max

FEF

[%]

BSFC

[g/kW-hr]

BTE

[%]

CE

[%]

A/F

Ratio

1

160 Proof 41.7 252 34.0 99.6 28.5

180 Proof 39.1 210 40.8 99.8 32.6

Diesel 0 220 39.0 99.9 32.2

2

160 Proof 61.8 274 31.2 99.3 28.5

180 Proof 60.0 220 39.0 99.5 34.9

Diesel 0 226 38.0 99.9 35.2

3

160 Proof 51.6 288 29.8 99.3 32.1

180 Proof 49.6 229 37.4 99.5 39.9

Diesel 0 234 36.6 99.9 39.0

4

160 Proof 41.2 630 13.6 98.9 51.3

180 Proof 46.0 590 14.5 98.8 55.7

Diesel 0 435 19.7 99.9 71.0

5

160 Proof 23.8 214 40.1 99.9 22.0

180 Proof 21.6 201 42.6 99.9 23.1

Diesel 0 206 41.6 99.9 23.4

6

160 Proof 27.6 223 38.5 99.9 23.2

180 Proof 28.7 205 41.8 99.9 25.0

Diesel 0 209 41.0 99.9 25.6

7

160 Proof 33.4 230 37.2 99.8 27.4

180 Proof 26.0 215 39.9 99.9 29.5

Diesel 0 213 40.3 99.9 31.1

8

160 Proof 48.9 980 8.74 99.1 80.1

180 Proof 53.0 848 10.1 98.9 90.5

Diesel 0 604 14.2 99.9 131

Figure 2: In-cylinder pressure traces and apparent RoHR

for Mode 3 and 180 proof hydrous ethanol

Figure 3: Brake specific CO emissions as a function of FEF

for 160 and 180 proof hydrous ethanol

Brake specific CO emissions are shown in Figure 3 for

160 and 180 proof hydrous ethanol injection over a range of

FEF for each testing mode. The dotted lines depicted on

emission plots represents diesel only combustion. CO

emissions increased uniformly with increasing FEF, and were

largely independent of ethanol proof. Incomplete combustion

is the primary cause of increased CO emissions, while low in-


cylinder temperatures prevent the oxidation of CO to CO2.

The trend also shows CO emissions reaching a horizontal

asymptote at high FEF where the increased A/F ratio prevents

the formation of CO.

Increases in CO have been shown to directly correlate

with an increase in HC emissions for combustion with

alcohols [13]. Further, excess HC emissions in dual fuel

modes are known to arise from the fumigant and not the

directly injected diesel fuel [9]. Figure 4 shows the light HC

distribution for 160 and 180 proof hydrous ethanol at Mode 3

for a range of FEF as measured by FTIR. Methane (CH4) and

ethylene (C2H4) emissions increase with FEF indicating

incomplete combustion becomes more significant. Using 160

proof hydrous ethanol leads to higher light HC emissions

overall. Increased water injection at the same FEF level leads

to additional cylinder charge cooling that lowers in-cylinder

temperature and leads to more incomplete combustion. This is

reflected in the lower CE values for higher ethanol proof given

in Table 3.

Figure 4: Selected HC emissions on a brake-specific basis

as a function of FEF for 160 and 180 proof hydrous

ethanol at Mode 3

Although injected hydrous ethanol proof had a significant

impact on light HC emissions, it did not change brake specific

ethanol emissions as shown in Figure 5. Similar to our

previous work, high engine load cases exhibit the lowest

amount of unburned ethanol due to higher engine temperatures

allowing sufficient heating to combust the ethanol completely.

Unburned ethanol emissions increase with FEF mainly

because it arises from areas in the combustion chamber

uninfluenced by the diffusive combustion event. These areas

include the squish and crevice regions. With increasing FEF,

the concentration of ethanol is greater in these regions leading

to higher emissions.

Further, at low engine load conditions, unburned ethanol

emissions increase much more rapidly as a function of FEF

due to fewer sufficiently hot regions in the combustion

chamber. Unburned ethanol concentration is lower for the

1400 rpm modes because low engine speed conditions operate

at higher temperatures than high engine speed conditions for a

given load. Fang et al. have also shown that ethanol delays

ignition and combustion phasing, also resulting in increased

unburned ethanol in the exhaust at higher engine speed [9]. In

addition, the injected ethanol is premixed with intake air,

causing any overlap between exhaust valve close (EVC) and

intake valve open (IVO) events to increase unburned ethanol

emissions through short-circuiting [20]. Although the PFI

injection strategy used in this study injected 21 CAD after

IVO to mitigate this effect, some short-circuiting is still

expected. In addition, gases from crevice volumes are known

to increase unburned ethanol emissions.

Figure 5: Brake specific unburned ethanol emissions as a

function of FEF for 160 and 180 proof hydrous ethanol

Figure 6 shows soot concentration in mg/m3 of exhaust

for all tested conditions. There was no uniform trend in soot

emissions as a function of FEF. High load modes (Modes 1, 2,

5, 6) showed increasing soot emissions with FEF. Of those

that achieved high FEF (Modes 1 and 2), the soot eventually

decreased. Initial increases in soot could be due to higher

temperature diffusion combustion and richer mixture around

the flame zones due to premixed ethanol. At high FEF

however, significantly less diesel is injected and combustion is

shifted more to a premixed mode.


Figure 6: Soot concentration as a function of FEF for 160

and 180 proof hydrous ethanol

For lighter load modes at high speed (Modes 2 and 3),

premixed ethanol had a more immediate effect in decreasing

soot emissions. At these conditions, shorter residence time and

diesel injection duration allowed soot to decrease more rapidly

with FEF. At idle, soot emissions were negligible and

independent of FEF due to primarily premixed low

temperature combustion. Ethanol proof also had no impact on

soot emissions for any mode indicating that any benefits

gained were due to ethanol replacement of diesel and not

through water influence on soot formation.

Ethanol may also chemically play a role in decreasing soot

formation at higher FEF. Previous work has shown that dual-

fuel combustion reduces soot concentrations, especially when

the secondary fuel is ethanol where increased OH radicals lead

to greater post-combustion soot oxidation [2,30,31]. Ethanol

(C2H5OH) consists of C-H, C-C, C-O, and O-H bonds. During

combustion, the C-C and C-O bonds can be readily broken due

to their lower bond energies. This chemical reaction causes an

increased concentration of OH radicals within the combustion

chamber [14]. If the theory is accurate, higher OH radical

concentration at higher FEF may result in decreasing exhaust

soot concentration.

NOx Emissions Results

Figure 7 depicts the brake specific NO emissions for 160

and 180 proof hydrous ethanol. For every operating mode,

NO emissions decreased with increasing FEF, as might be

expected due to the increase in charge cooling lowering in-

cylinder temperatures. However, there was no discernable

change in NO emissions between the use of 160 and 180 proof

hydrous ethanol at a given FEF indicating that NO formation

is independent of water content in the fuel. This discrepancy is

evidence that other factors besides charge influence NO

emissions for dual fuel combustion, such as the propensity of

unburned hydrocarbons facilitating the conversion of NO to

NO2 during the expansion stroke. The chemical kinetics and

mechanisms responsible for this conversion are discussed in

the single zone combustion modeling presented later in this

work.

Figure 7: Brake specific NO emissions as a function of FEF

for 160 and 180 proof hydrous ethanol

As in previous work, increasing FEF had very little

impact on overall NOX emissions as shown in Figure 8. This

indicates that fumigation with hydrous ethanol for all cases,

except for the very highest FEF conditions, does not have an

impact on NO formation during diffusive diesel combustion

and that NO is converted to NO2 at some point during the

closed cycle. Only at very high FEF does charge cooling play

a role in mitigating formation via the thermal Zel’dovich

mechanism, especially noticeable during Mode 4.

The experimental data imply that NO is converted to NO2

while overall NOX concentration nearly constant. As a

measure of the conversion process, Figure 9 shows the

NO2/NOX ratio for all experimental data points collected in the

study as a function of unburned ethanol measured in the

exhaust normalized by the diesel-only NOX emissions per

mode respectively. The NO to NO2 conversion process occurs

rapidly as a function of unburned ethanol until reaching a


horizontal asymptote around 0.72. At this point, no further

conversion of NO occurs with increasing FEF. Reasons for

this trend are not apparent from the experimental data and

require additional kinetics modeling to investigate.

Figure 8: Brake specific NOX emissions as a function of

FEF for 160 and 180 proof hydrous ethanol

Figure 9: NO2/NOX ratio as a function of unburned

ethanol for all modes and ethanol proofs

Single Zone Combustion Modeling

Our experimental results indicate that NO to NO2

conversion increases with increasing unburned ethanol in the

exhaust. In previous research, Hori et al. [24] used an

adiabatic constant pressure single zone reactor model and a

constant temperature quartz flow experimental reactor to

illustrate the mechanism by which hydrocarbons influence the

NO to NO2 conversion at different temperatures, with ethylene

and propane being very effective as compared to methane and

ethane. However, work conducted in Hori et al. did not

consider ethanol.

A single zone constant pressure reactor model was created

in the open-source thermochemistry and kinetics code Cantera

to compare ethanol’s effectiveness to other hydrocarbons for

converting NO to NO2. The C1-C4 Hydrocarbon with NO

Addition mechanism from Lawrence Livermore National

Laboratory (LLNL) was used in the model [24]. Initial

conditions for the model were the same as for the Hori et al

paper including atmospheric pressure and concentrations 20

PPM NO in N2, 50 PPM of HC in N2, and the remainder air.

The constant pressure model was run over a sweep of

operating temperature between 600 and 1200 K at 50 K

increments, where the gases were controlled to a 1.5 second

residence time within the reactor. Ethanol, ethylene and

methane were used as hydrocarbons because these had the

highest concentrations in the measured engine exhaust.

Concentrations of NO and NO2 were exported and the ratio of

final NO2/NOX ratio for each operating temperature can be

seen in Figure 10.

Figure 10: NO2/NOX ratio as a function of reaction

temperature for selected HC’s using a constant pressure

reactor

Figure 10 shows that ethanol has a peak NO2/NOX ratio

similar to ethylene but over a smaller temperature window.

Hydrocarbons predominantly oxidize NO to NO2 through the

NO + HO2 ↔ NO2 + OH mechanism. Hori et al. have shown

that the effectiveness of a hydrocarbon at converting NO to

NO2 is dependent on its ability to simultaneously produce

radicals like OH to sustain fuel oxidation and HO2 for NO to


NO2 conversion [24]. According to the model, methane does

not readily promote the conversion of NO to NO2. Because the

oxidation of methane is relatively slow, there is a limited

amount of HO2 produced, in addition to methyl radicals

reducing NO2 to NO via CH3 + NO2 ↔ CH3O + NO [24].

Ethanol reacts with OH radicals to make CH3CHOH, and

CH3CH2O, which can be oxidized to produce HO2 in the

following reaction scheme.

𝐶𝐻3𝐶𝐻𝑂𝐻 + 𝑂2 ↔ 𝐶𝐻3𝐻𝐶𝑂 + 𝐻𝑂2 (1)

𝐶𝐻3𝐻𝐶𝑂 + 𝑂 ↔ 𝐶𝐻3𝐶𝑂 + 𝑂𝐻 (2)

Reaction 1 forms HO2 radicals to promote NO to NO2

conversion, while reaction 2 produces OH radicals to feed the

ethanol consumption reaction. The model shows that ethanol

has a high tendency to convert NO to NO2 at temperatures

between 800 - 1200 K. This temperature range is greater than

engine out exhaust temperatures; however temperatures during

the expansion stroke would fall within this temperature range.

To more closely compare the model to experimental

results, a variable pressure single zone model was created.

This model uses the recorded in-cylinder pressure data versus

time to model NO conversion kinetics occurring during the

expansion stroke. The model assumes that NO formation is

complete and the burned gases are mixed by the crank angle

location of 90% gross heat release (CA90). Therefore, CA90

was chosen as the initial condition for the variable pressure

model. The pressure data starting at CA90 was used in

conjunction with a range of initial local in-cylinder

temperatures and a sweep of unburned ethanol concentrations

to predict NO to NO2 conversion. A range of temperatures at

CA90 was used because local in-cylinder temperatures were

unknown and can vary within the cylinder. The in-cylinder

temperatures were calculated using the polytropic relations for

each initial CA90 temperature and measured pressure data.

Figure 11 illustrates the conditions under which the model

was run. The symbols on the plot are for clarity between the

trends, and are not indicative of actual data. The TCA90 surface

represents the isentropic temperature curve fits for initial

temperatures from 1000 to 2000 K at 50 K increments. A

larger lower range of temperatures was chosen because the

introduction of water is well known to decrease in-cylinder

temperatures. The mean in-cylinder temperature curve starting

at CA90 was calculated to validate the TCA90 range used for

the model. The initial exhaust composition points were taken

from a CDC mode (1400 rpm, 250 N-m), and a range of

ethanol (500 - 4000 PPM) was added to the mixture at the start

of the model. Similar results were obtained when running the

model at different engine testing modes, and were not

included for brevity.

Figure 11: Apparent RoHR, in-cylinder pressure, mean in-

cylinder temperature and CA90 temperature range as a

function of CAD

The NO conversion trajectory as a function of crank angle

for two different CA90 temperatures is shown in Figure 12. At

lower temperatures, NO is readily converted to NO2 early in

the expansion stroke, and then “freezes” when the temperature

becomes too low to promote conversion. At higher

temperatures, NO has a sudden decrease followed by an

increase before settling as in-cylinder temperatures do not

enter the right temperature range to promote conversion. The

decrease in NO at high temperatures is caused by the

oxidation of ethanol, while the increase right after is because

the reversion of NO2 to NO is more favored at high

temperatures. The model suggests that unburned ethanol will

not be evident in regions of high in-cylinder temperatures,

preventing the conversion of NO to NO2. Instead, unburned

ethanol evident in low local in-cylinder temperature regions

will readily convert NO to NO2.

Figure 12: NO conversion trajectory as a function of CAD

starting at CA90


Figure 13: NO2/NOX contour as a function of local in-

cylinder temperatures at CA90 and normalized unburned

ethanol concentration

Figure 13 shows the NO2/NOX ratio at the end of

expansion, plotted as a contour against the initial temperature

at CA90 and the ratio of unburned ethanol concentration

normalized against initial NOX concentration. A clear island

can be seen for TCA90 between 1150 and 1250 K where

unburned ethanol in the exhaust promotes the conversion of

NO to NO2. At local in-cylinder temperatures greater than

1400 K, complete ethanol conversion is predicted, preventing

the conversion. In addition, higher local temperatures favor

the reduction of NO2 to NO, while the consumption of ethanol

yields daughter radicals. Hori et al. have shown that daughter

radicals resistant to oxidation by O2 will reduce NO2 to NO by

the reaction R + NO2 ↔ NO + RO. In contrast, at local in-

cylinder temperatures lower than 1150 K, limited OH radicals

reduces the production of HO2, slowing down the conversion

of NO to NO2. It is also important to note that this model only

takes into account unburned ethanol in the expansion stroke.

As seen in Figure 4, other unburned hydrocarbons such as

ethylene (C2H4), and methane (CH4) increase with increasing

FEF, promoting additional NO to NO2 conversion.

CONCLUSIONS

In this study, a comprehensive dataset was presented

characterizing an add-on dual fuel PFI system using hydrous

ethanol as the secondary fuel. Data was collected over a range

of FEF for each point along a modified ISO 8178 eight point

testing plan. The results show that 160 proof hydrous ethanol

can achieve up to 61.8% FEF, while 180 proof reaches up to

60% FEF. CO, THC, and unburned ethanol emissions all

increase with increasing FEF, while NOX emissions initially

show no change, but begin to decrease at high FEF where

significant charge cooling lowers the diesel combustion

temperature. It was also found that both 160 and 180 proof

hydrous ethanol follow the same emissions trends, with few

significant different between similar FEF values.

Single-zone reactor models were created using Cantera to

investigate the conversion of NO to NO2 due to unburned

hydrocarbons in the exhaust. The first model found that

different hydrocarbons have different propensities for

promoting the NO to NO2 conversion, a finding in validation

with work conducted by Hori et al. [23] Results from this

model indicated that unburned ethanol concentrations promote

the conversion of NO to NO2 at a temperature range from 800

– 1200 K, temperatures likely to be encountered during the

expansion stroke of an engine. The second single zone model

used high-speed in-cylinder pressure traces and CA90 as

initial conditions to estimate NO to NO2 conversion during the

expansion stroke. The model found that in-cylinder

temperature regions between 1150 and 1250 K at CA90 have a

high NO to NO2 conversion rate and that conversion mostly

occurs near the beginning of the expansion stroke.

Overall, our findings indicate that aftermarket dual fuel

systems that directly introduce the secondary fuel into a fixed

calibration engine cannot achieve the emissions reductions

possible with low temperature RCCI combustion modes. To

effectively use aftermarket dual fuel strategies for NOX

emissions reduction, mitigation of NO formation in the diesel

diffusion flame must be achieved by lowering combustion

temperature. Strategies such as exhaust gas recirculation or

hydrous ethanol reformation could be used to increase the heat

capacity of the unburned gas and reduce overall combustion

temperature while maintaining high FEF and thermal

efficiency.

ACKNOWLEDGMENTS

This research was conducted with funding from the

Minnesota Corn Growers Association, The Agricultural

Utilization Research Institute and the University of Minnesota

Institute for Renewable Energy and Environment under grant

AIC209. We wish to acknowledge our colleagues at the

Thomas E. Murphy Engine Research Laboratory at the

University of Minnesota, especially Darrick Zarling for

technical guidance, Andrew Kotz for assistance with

developing the high speed in-cylinder pressure trace data

logging system, and Wei Fang for high speed data processing

assistance.

NOMENCLATURE

AFR – Air/Fuel Ratio

ATDCF – After Top Dead Center Firing

BMEP – Brake Mean Effective Pressure

BSFC – Brake Specific Fuel Consumption

BTE – Brake Thermal Efficiency

CA90 – Crank Angle location of 90% gross heat release

CAD – Crank Angle Degree

CDC – Conventional Diesel Combustion

CE – Combustion Efficiency

CI – Compression Ignition

DATDC – Degrees After Top Dead Center

ECU – Electronic Control Unit

EGR – Exhaust Gas Recirculation

EVC – Exhaust Valve Close

FEF – Fumigant Energy Fraction

FTIR – Fourier Transform Infrared Spectrometer

GHG – Greenhouse Gas

HC – Hydrocarbon

IVO – Intake Valve Open


LFE – Laminar Flow Element

LHV – Lower Heating Value

LLNL – Lawrence Livermore National Laboratory

MSS – Micro-Soot Sensor

NI – National Instruments

PM – Particulate Matter

PFI – Port Fuel Injection

RCCI – Reactivity Controlled Compression Ignition

RoHR – Rate of Heat Release

SCR – Selective Catalytic Reduction

TDC – Top Dead Center

THC – Total Hydro Carbons

ULSD – Ultra-Low Sulfur Diesel

REFERENCES

[1] EPA, 2013, “Nonroad Compression-Ignition Engines -

- Exhaust Emissions Standards” [Online]. Available:

http://www.epa.gov/otaq/standards/nonroad/nonroadci

.htm.

[2] Olson, A. L., 2010, “The Effect of Ethanol-Water

Fumigation on the Performance and Emissions from a

Direct-Injection Diesel Engine,” (September).

[3] Bowman, C. T., 1992, “Control of combustion-

generated nitrogen oxide emissions: Technology

driven by regulation,” Symp. Combust., 24(1), pp.

859–878.

[4] Lilik, G. K., Zhang, H., Herreros, J. M., Haworth, D.

C., and Boehman, A. L., 2010, “Hydrogen assisted

diesel combustion,” Int. J. Hydrogen Energy, 35(9),

pp. 4382–4398.

[5] Lavoie, G. A., Heywood, J. B., and Keck, J. C., 1970,

“Experimental and Theoretical Study of Nitric Oxide

Formation in Internal Combustion Engines,” Combust.

Sci. Technol., 1(4), pp. 313–326.

[6] Dempsey, A. B., Walker, N. R., Splitter, D., Wissink,

M., and Reitz, R. D., 2012, “Characterization of

Reactivity Controlled Compression Ignition ( RCCI )

Using Premixed Hydrated Ethanol and Direct

Injection Diesel in Heavy-Duty and Light-Duty

Engines.”

[7] Splitter, D., Hanson, R., Kokjohn, S. L., and Reitz, R.

D., 2011, “Reactivity Controlled Compression Ignition

(RCCI) Heavy-Duty Engine Operation at Mid-and

High-Loads with Conventional and Alternative Fuels,”

SAE Tech. Pap. Ser., 1(Ci), p. 0363.

[8] Reitz, R. D., and Duraisamy, G., 2015, “Review of

high efficiency and clean reactivity controlled

compression ignition (RCCI) combustion in internal

combustion engines,” Prog. Energy Combust. Sci., 46,

pp. 12–71.

[9] Fang, W., Huang, B., Kittelson, D. B., and Northrop,

W. F., 2012, “Dual-Fuel Diesel Engine Combustion

with Hydrogen, Gasoline, and Ethanol Fumigants:

Effect of Diesel Injection Timing,” ASME Intern.

Combust. Engine Div., pp. 1–9.

[10] El-hagar, M. M. E., 2014, “Exhaust Emissions of a

Single Cylinder Diesel Engine with Addition of

Ethanol,” 3(1), pp. 74–81.

[11] Rahman, M. M., Stevanovic, S., Brown, R. J., and

Ristovski, Z., 2013, “Influence of Different

Alternative Fuels on Particle Emission from a

Turbocharged Common-Rail Diesel Engine,” Procedia

Eng., 56, pp. 381–386.

[12] Cheng, C. H., Cheung, C. S., Chan, T. L., Lee, S. C.,

and Yao, C. D., 2008, “Experimental Investigation on

the Performance, Gaseous and Particulate Emissions

of a Methanol Fumigated Diesel Engine.,” Sci. Total

Environ., 389(1), pp. 115–24.

[13] Imran, A., Varman, M., Masjuki, H. H., and Kalam,

M. a., 2013, “Review on Alcohol Fumigation on

Diesel Engine: A Viable Alternative Dual Fuel

Technology for Satisfactory Engine Performance and

Reduction of Environment Concerning Emission,”

Renew. Sustain. Energy Rev., 26, pp. 739–751.

[14] Surawski, N. C., Ristovski, Z. D., Brown, R. J., and

Situ, R., 2012, “Gaseous and Particle Emissions from

an Ethanol Fumigated Compression Ignition Engine,”

Energy Convers. Manag., 54(1), pp. 145–151.

[15] Code of Regulations, C., and ARB, Regulation to

Reduce Emissions of Diesel Particulate Matter,

Oxides of Nitrogen and Other Criteria Pollutants from

In-use Heavy-Duty Diesel-Fueled Vehicles.

[16] Hwang, J. T., and Northrop, W. F., 2014, “Gas and

Particle Emissions from a Diesel Engine Operating in

a Dual-Fuel Mode using High Water Content Hydrous

Ethanol,” ASME Intern. Combust. Engine Div., pp. 1–

9.

[17] Papagiannakis, R. G., and Hountalas, D. T., 2004,

“Combustion and exhaust emission characteristics of a

dual fuel compression ignition engine operated with

pilot diesel fuel and natural gas,” Energy Convers.

Manag., 45(18-19), pp. 2971–2987.

[18] Alperstein, M., Swim, W. B., and Schweitzer, P. H.,

1957, “Fumigation Kills Smoke Improves Diesel.”

[19] Chaplin, J., and Binjanius, R., 1987, “Ethanol

fumigation of a compression-ignition engine using

advanced injection of diesel fuel,” Trans. Asae, 30(3),

pp. 610–614.

[20] Nord, A. J., Hwang, J. T., and Northrop, W. F., 2015,

“Emissions from a Diesel Engine Operating in a Dual-

Fuel Mode Using Port-Fuel Injection of Heated

Hydrous Ethanol,” ASME Intern. Combust. Engine

Div., pp. 1–9.

[21] Soloiu, V., Muinos, M., and Harp, S., 2015,

“Investigation of Dual Fuel PCCI (PFI of n-Butanol

and DI-ULSD) Compared with DI of Binary Mixtures

of the Same Fuels in an Omnivorous Diesel Engine,”

SAE Int., (X).

[22] Morsy, M. H., 2015, “Assessment of a direct injection

diesel engine fumigated with ethanol/water mixtures,”

Energy Convers. Manag., 94, pp. 406–414.

[23] Greeves, G., Khan, I. M., and Onion, G., 1977,

“Effects of water introduction on diesel engine

combustion and emissions,” Symp. Combust., 16(1),

pp. 321–336.

[24] Hori, M., Matsunaga, N., Marinov, N., William, P.,

and Charles, W., 1998, “An experimental and kinetic

calculation of the promotion effect of hydrocarbons on


the NO-NO2 conversion in a flow reactor,” Symp.

Combust., 27(1), pp. 389–396.

[25] Heywood, J. B., Internal Combustion Engine

Fundamentals.

[26] Saffy, H. A., Northrop, W. F., Kittelson, D. B., and

Boies, A. M., 2015, “Energy, carbon dioxide and

water use implications of hydrous ethanol

production,” Energy Convers. Manag., 105, pp. 900–

907.

[27] 1997, “ISO 8178 Emission Test Cycle,” DieselNet

[Online]. Available:

https://www.dieselnet.com/standards/cycles/iso8178.p

hp.

[28] Figliola, R., and Beasley, D., 2000, Theory and

Design for Mechanical Measurements, John Wiley

and Sons, New York.

[29] Northrop, W. F., Fang, W., and Huang, B., 2013,

“Combustion Phasing Effect on Cycle Efficiency of a

Diesel Engine Using Advanced Gasoline Fumigation,”

J. Eng. Gas Turbines Power, 135(3), p. 032801.

[30] Bika, A. S., Franklin, L., Olson, A. L., Watts, W., and

Kittelson, D., 2009, “Ethanol Ultilization in a Diesel

Engine.”

[31] Can, Ö., Çelikten, İ., and Usta, N., 2004, “Effects of

Ethanol Addition on Performance and Emissions of a

Turbocharged Indirect Injection Diesel Engine

Running at Different Injection Pressures,” Energy

Convers. Manag., 45(15-16), pp. 2429–2440.

icef2016-9349€¦ · emissions standards for new engines and rely less on advanced in-cylinder...

Documents