Modeling and Application of Catalytic Ignition in Internal ... … · Modeling and Application of Catalytic ... Two-Dimensional Propane-Air Mixture Mixed Mode Heat Transfer ... of

Modeling and Application of Catalytic Ignition in Internal Combustion Engines

FINAL REPORT FEBRUARY 2004

Budget Number KLK312

N04-03

Prepared for

OFFICE OF UNIVERSITY RESEARCH AND EDUCATION U.S. DEPARTMENT OF TRANSPORTATION

Prepared by

National Institute for Advanced Transportation Technology University of Idaho

Judi Steciak and Steve Beyerlein

TABLE OF CONTENTS

TABLE OF FIGURES............................................................................................................. iii

LIST OF TABLES.................................................................................................................... v

LIST OF TABLES.................................................................................................................... v

EXECUTIVE SUMMARY ...................................................................................................... 1

PART I. STEADY STATE DYNAMOMETER TESTING TO COMPARE OPERATION

ON GASOLINE AND AQUEOUS ETHANOL IN A PASSENGER VAN ........................... 3

I.A INTRODUCTION ........................................................................................................ 3

I.B DESCRIPTION OF PROBLEM................................................................................... 4

I.C APPROACH AND METHODOLOGY ....................................................................... 5

Benefits of Aqueous Fuel Combustion ......................................................................... 5

Catalytic Igniter Technology ........................................................................................ 6

Fuel Handling System................................................................................................... 7

Fuel Injection System ................................................................................................... 8

Test Protocol ............................................................................................................... 11

Dynamometer Testing................................................................................................. 13

I.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS......................................... 16

Preliminary Results..................................................................................................... 16

Conclusions................................................................................................................. 17

Future Work ................................................................................................................ 18

Acknowledgements..................................................................................................... 18

PART II. MODELING CATALYTIC IGNITION CONDITIONS OF PROPANE/AIR

MIXTURES OVER PLATINUM WIRES ............................................................................. 19

II.A INTRODUCTION ...................................................................................................... 19

II.B DESCRIPTION OF PROBLEM................................................................................. 20

II.C APPROACH AND METHODOLOGY ..................................................................... 21

Pressurizable Plug--Flow Reactor............................................................................... 22

Experimental Determination of Surface Reaction Temperature................................. 24

Two-Dimensional Propane-Air Mixture Mixed Mode Heat Transfer Model ............ 24

Sensitivity Study (Fluid Flow).................................................................................... 26

Modeling and Application of Catalytic Ignition in Internal Combustion Engines i

Sensitivity Study (Catalyst Geometry) ....................................................................... 26

Sensitivity Study (Power Input).................................................................................. 27

Three-Dimensional FEA Model ................................................................................. 27

I.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS......................................... 29

Comparisons of Two-Dimensional and Three-Dimensional Results with the

Experimental Results .................................................................................................. 29

Ongoing Research....................................................................................................... 30

Acknowledgments....................................................................................................... 30

PART III. THEORETICAL STUDY OF AQUEOUS ETHANOL-AIR COMBUSTION

IN PLUG FLOW..................................................................................................................... 31

III.A INTRODUCTION .................................................................................................. 31

III.B DESCRIPTION OF PROBLEM............................................................................. 32

III.C APPROACH AND METHODOLOGY ..................................................................... 33

Hydrodynamics, Combustion, and Transport (HCT) Code ........................................ 34

Theoretical Impact of Water on Gas-Phase Ethanol Combustion .............................. 35

III.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS......................................... 47

Future Plans ................................................................................................................ 47

Acknowledgements..................................................................................................... 48

REFERENCES ....................................................................................................................... 49

Modeling and Application of Catalytic Ignition in Internal Combustion Engines ii

TABLE OF FIGURES

FIGURE 1 Dual-fuel demonstration vehicle. ....................................................................... 5

FIGURE 2 Plan view of the van, showing physical location of components..................... 10

FIGURE 3 FTP Urban driving cycle trace. ......................................................................... 11

FIGURE 4 Highway driving cycle trace............................................................................. 12

FIGURE 5 Dynamometer inputs and desired outcomes..................................................... 14

FIGURE 6 MAX fuel metering system. ............................................................................. 15

FIGURE 7 Dual-fuel vehicle test platform schematic. ....................................................... 15

FIGURE 8 Sections of the flow reactor. ............................................................................. 22

FIGURE 9 Prototype mixing nozzle mounted downstream of the evaporator. .................. 23

FIGURE 10 A two-dimensional propane air mixture FEA model of platinum temperature

distribution. ..................................................................................................................... 25

FIGURE 11 A three-dimensional propane air mixture FEA model of platinum temperature

distribution. ..................................................................................................................... 28

FIGURE 12 Catalyst temperatures versus changes in power supply. ................................ 29

FIGURE 13 Logic flow and file interaction for the MKCDAT, HCT, and HCTPLT

programs ......................................................................................................................... 35

FIGURE 14 Species consumption path analysis for ethanol oxidation.............................. 36

FIGURE 15 Comparison between numerical calculations and experimental data for flow

reactor studies of ethanol oxidation at φ = 0.81.............................................................. 37

FIGURE 16 C2H5OH vs. residence time at φ = 0.81.......................................................... 38

FIGURE 17 Temperature vs. residence time at φ = 0.81. .................................................. 39

FIGURE 18 O2 vs. residence time at φ = 0.81.................................................................... 39

FIGURE 19 CO vs. residence time at φ=0.81. ................................................................... 40

FIGURE 20 CO2 vs. residence time at φ = 0.81. ................................................................ 41

FIGURE 21 C2H6 vs. residence time at φ = 0.81................................................................ 42



FIGURE 24 CH3CHO vs. residence time at φ = 0.81......................................................... 44

Modeling and Application of Catalytic Ignition in Internal Combustion Engines iii

FIGURE 25 Water vs. residence time at φ = 0.81. ............................................................. 45

FIGURE 26 Water vs. residence time at φ = 0.81. ............................................................. 45

FIGURE 27 CH4 vs. residence time at φ =0.81. ................................................................. 46

Modeling and Application of Catalytic Ignition in Internal Combustion Engines iv

LIST OF TABLES

TABLE 1 Description of Six-Mode Points......................................................................... 12

TABLE 2 Five-Gas Emissions and Fuel Consumption at Six Modal Points for Gasoline. 16

TABLE 3 Torque and Phrottle Position.............................................................................. 17

TABLE 4 Design Features.................................................................................................. 17

TABLE 5 Sensitivity to Mesh Density ............................................................................... 24

TABLE 6 Effects of Fluid Velocity Changes on Catalyst Temperature ............................ 26

TABLE 7 Effects of Wire Diameter Changes on Average Catalyst Temperature. ............ 26

TABLE 8 Effects of Power Supply Changes on Average Catalyst Temperature............... 27

Modeling and Application of Catalytic Ignition in Internal Combustion Engines v

EXECUTIVE SUMMARY

We progressed towards our ultimate goal of developing catalytic igniters for aqueous

ethanol as a transportation fuel with three different actions: 1) developing a test matrix to

compare the performance of a passenger van operating with gasoline with the same van

operating with aqueous ethanol; 2) determining the average temperature when surface

reactions occur on a heated platinum wire catalyst; and 3) theoretically modeling the

impact of water on the gas-phase oxidation of ethanol.

We used a steady-state chassis dynamometer approximating urban and rural driving

cycles to develop our test matrix. We evaluated the matrix using a van fueled with

gasoline. The van was converted for dual-fuel use with a programmable dual fuel

computer and injectors replacing the carburetor, and catalytic igniters instead of

sparkplugs.

Ignition of fuel/oxygen/nitrogen mixtures over a platinum wire was studied using

microcalorimetry. Experimental results were compared with predictions from a steady-

flow finite element (FEA) model. The average catalyst wire temperatures obtained from a

three-dimensional FEA model over predicted the experimentally obtained temperatures

by only threepercent. The FEA analysis indicated that axial conduction dominated heat

losses from the Pt catalyst in comparison with radial convection.

The thermal decomposition and combustion kinetics of gas-phase ethanol oxidation were

modeled with a computer code. The output was compared with flow reactor data

available in the literature. Reasonable agreement was found between the model and data

for 100 percent ethanol oxidation.

The vehicle test matrix developed in this work makes it possible for us to compare

vehicle performance on gasoline and ethanol-water. Comparisons of alternative fuels

with conventional fuels fueling the same vehicle platform are required to quantify

differences in performance and emissions.

Modeling and Application of Catalytic Ignition in Internal Combustion Engines 1

The FEA modeling indicates that thermal breaks and/or substrates with lower thermal

conductivity will reduce heat losses from catalytic igniters, a concern during cold start.

Our new ability to model detailed combustion kinetics permits us to determine the

optimal ignition and combustion conditions needed to reduce the formation of toxins and

environmental contaminants from renewable transportation fuels.


PART I. STEADY STATE DYNAMOMETER TESTING TO COMPARE OPERATION ON GASOLINE AND AQUEOUS ETHANOL IN A PASSENGER VAN

I.A INTRODUCTION

Comparisons of alternative fuels with conventional fuels fueling the same vehicle

platform are required to quantify differences in performance and emissions. To make

comparison tests possible, we developed a test matrix using a steady-state chassis

dynamometer approximating urban and rural driving cycles. The matrix was evaluated

using a passenger van operating with gasoline. The van was converted for dual-fuel use

with a programmable dual fuel computer and injectors replacing the carburetor, and

catalytic igniters instead of sparkplugs. The van conversion provided us with a robust,

reliable vehicle platform for evaluating alternative fuel handling system that shows no

sign of corrosion. Furthermore, gasoline fuel economy and emissions are far improved

over the original carbureted configuration with a 95 percent reduction in NOx and a 67

percent reduction in unburned hydrocarbons. The test matrix developed in this work will

be used to compare vehicle performance on gasoline and ethanol-water.

In the following sections, we summarize the progress of our research in this area.

Part I. Steady State Dynamometer Testing to Compare Operation on Gasoline and Aqueous Ethanol in a Passenger Van 3

I.B DESCRIPTION OF PROBLEM

Previous research on catalytic igniters and aqueous fueled engines showed potential for

lowering emissions and increasing engine efficiency over conventional engine

configurations. To quantify these improvements in a vehicle platform, we converted a

transit van owned by Valley Transit of Lewiston, Idaho, to operate on both gasoline and

aqueous fuels, with changeover possible in less than one hour.

Back to back comparisons of baseline engine and converted engine performance are

critical for documenting benefits of alternative fuel operation on the same vehicle. To

facilitate these comparisons, we developed a test matrix using a steady-state chassis

dynamometer approximating urban and rural driving cycles. Preliminary results for the

converted transit van using gasoline, which agree closely with driving cycle data from the

original vehicle, are presented in this report.


I.C APPROACH AND METHODOLOGY

The purpose of this project was to create infrastructure for research on alternative vehicle

performance involving catalytic igniter technology. This resulted in the dual-fuel van

(Fig. 1) as well as test protocols involving a steady-state dynamometer that can be used to

rigorously evaluate vehicle performance under different operating conditions. Future

phases of this work will lead to Federal Test Protocol (FTP) driving cycle tests on

gasoline and alternative fuels. Ultimately, this dual fueled vehicle will be studied in over-

the-road tests as part of a local transit system.

FIGURE 1 Dual-fuel demonstration vehicle.

Benefits of Aqueous Fuel Combustion

Ethanol is a renewable fuel that is primarily made from agriculture crops [Nadkarni,

2000]. Currently, blends of 85 percent ethanol and 15 percent gasoline (E85) are

commercially available [Wyman, 1996]. Much is published on the emission benefits of

ethanol, and many have taken advantage of ethanol’s inherent attraction to water by

mixing small quantities of water into the fuel [Lee and Geffers, 1977]. Water in the

combustion chamber significantly reduces flame temperatures and thus reduces NOx

formation. In most cases reported in the literature, the engine is not capable of cold

starting on ethanol and must be started and warmed on a pilot fuel [Jehlik, et al., 1999].

These cases are also limited to small amounts of water, as conventional ignition sources


are unable to ignite blends of greater than 15 percent water. In all studies with spark

ignition, significant misfire or non-fire occurs above 10 percent water present in the fuel.

The University of Idaho and Automotive Resources, Inc., of Sandpoint, Idaho, have been

working with Aquanol fueled engines since 1996. Aquanol is a mix of 65 percent ethanol

and 35 percent water by volume. While this mixture has not been optimized, it provides a

good balance between NOx minimization and complete combustion of in-cylinder

hydrocarbons. Aquanol-fueled engines have run on mixtures up to 50 percent ethanol and

50 percent water, and shown cold starting capability [Morton, et al., 1999].

Catalytic Igniter Technology

The primary drawback in using Aquanol is the difficulty of initiating combustion.

Traditional means of spark ignition are insufficient to initiate flame propagation of

Aquanol-air mixtures [Cherry, et al., 1992]. An ignition source using a catalytic reaction

in a pre-chamber provides a high-power torch ignition that has proven successful at

igniting mixtures previously un-ignitable by spark or compression ignition [Gottschalk,

1995]. This ignition source is elementary when converting an engine to operate on

Aquanol.

Automotive Resources Inc. (ARI) has held the patent on catalytic ignition in a pre-

chamber since 1990 [Cherry, 1990]. Since then ARI has made many improvements in the

robustness and ignition control of the catalytic igniter. They have applied this technology

to improve performance and emissions of rotary engines, two-stroke engines and to

combust unconventional fuels in reciprocating internal combustion (IC) engines.

The University of Idaho began its first research project in 1966, converting aYanmar

three-cylinder, direct-injection, compression-ignition (CI) engine to run on Aquanol

[Morton, 2000]. Direct diesel injectors were replaced with catalytic igniters, and a

manifold fuel injection system was put in place for fuel delivery. This engine had no

throttle plate, so load was controlled only by the amount of fuel delivered to the engine.


Initial research showed that it was possible to burn high water content with ethanol in an

IC engine using catalytic ignition. Further testing was done to acquire full brake specific

fuel consumption and brake specific emissions maps on both a stock Yanmar diesel

engine and the Yanmar converted to Aquanol fuel [Clarke, 2001].

As expected, the emissions of NOx were reduced significantly, with slight increases in

CO and HC—both of which are easily reduced with exhaust after treatment. Net

indicated thermal efficiency increased from 35 percent on diesel to 42 percent on

Aquanol [Cordon, et al., 2002]. With the addition of an intake air pre-heater, the Yanmar

engine demonstrated cold-start and smooth idle capabilities on a 35 percent water/65

percent ethanol fuel, operating over air/fuel ratios between 8:1 and 45:1.

Fuel Handling System

One of the major systems unique to a flexible-fuel vehicle is fuel handling. This research

platform was designed to keep two different fuels onboard by installing individual

storage for gasoline and Aquanol. Nearly all fuel system components were designed to be

compatible with gasoline, since gasoline is the predominant fuel used in the United

States. Alcohol fuels are more corrosive than gasoline and the addition of water makes

them even more corrosive. Fuel handling components designed for gasoline are

susceptible to rapid corrosion when used with Aquanol.

The first Aquanol storage tank used in this research was made of steel with the interior

surface coated with Teflon. The Teflon was intended to act as a barrier against corrosion.

After two years of use, examination of the fuel tank showed the coating had failed and

particles were clogging the remainder of the fuel system. In addition, the fuel pickup and

fuel lines were not Aquanol compatible. Materials compatible with Aquanol are:

polypropylene, stainless steel, hard anodized aluminum, brass, Teflon, and most synthetic

rubbers.

That first handling system was removed and replaced with a pair of polypropylene fuel

tanks with anodized fuel pickups. The two Aquanol tanks combined have a capacity of 46


gallons, which yields a similar vehicle range as a 25-gallon tank of gasoline. The flexible

fuel lines, Earls Auto-Flex hose constructed of HTE synthetic rubber bonded to a braided

stainless steel shell, use hard-anodized Ano-Tuff hose ends. All the rigid fuel lines and

fittings are made of 304 stainless steel, with the exception of anodized aluminum pre-

filters.

Because of space limitations, only a single fuel injection system will fit on the engine.

Both fuels share a common fuel pump, high-pressure lines, injectors and regulator. To

accommodate this, a set of stainless valves is used to control the flow path of fuel. In fuel

injection systems the fuel tank has two lines. The suction line is connected to the fuel

pump where a constant flow and pressure is attained. Any fuel not used in the engine is

returned back to the fuel tank via a return line. Dual fuel tanks are common on large fuel

injected vehicles, and tank selector valves are available that switch both the suction and

return lines at once. These valves are not compatible with Aquanol. More importantly,

using such a valve will allow the fuel enclosed in the loop between the inlet and return

valves to contaminate the other fuel when switched. The set of fuel selector valves

designed for the van consists of a single T-ball valve that switches between the two

different fuel sources, and a valve body of three T-valves to handle flow of the return

line. All four valves are modular and made of 316 stainless with a Teflon seat. The three

valves are joined by half-inch stainless pipefittings coated with anti-seize compound to

prevent galling.

Fuel Injection System

Two principal design constraints for the fuel system were corrosion resistance and a wide

fuel metering capability. Conscientious material selection provides a solution to the first

constraint. Aquanol has 36 percent the energy per unit mass of gasoline (15.9 MJ/kg vs.

44.0 MJ/kg). Even though engines running Aquanol and catalytic igniters typically show

an increase in net indicated thermal efficiency, Aquanol still requires over twice the

volume flow rate of gasoline to maintain comparable performance. A careful balance of

shared components was necessary to ensure successful operation on both fuels.


The fuel injection system is responsible for precise metering of the fuel delivered to the

engine. CI engines control engine load by varying only the amount of fuel delivered to

the engine, and thus operate over a wide range of air/fuel ratios. This is the case for diesel

engines, and the Yanmar Aquanol conversion. Spark ignition (SI) engines control load by

throttling—restricting airflow to the engine. Ideally, SI engines will have a constant

air/fuel ratio, but the amount of air/fuel mixture will vary greatly throughout operating

conditions. The job of the fuel injection system is to meter the fuel flow to maintain a

consistent air/fuel ratio. This is done with a central electronic control module (ECM) that

sends and receives information from various engine components.

Using a programmable ECM allows injection components to be shared between both

fuels. This provides easy changing between fuels without the need for replacing fuel

hardware as would be necessary with carburetion. When switching fuels, a new fuel map

is uploaded to the ECM. The Haltec ECM interfaces with a laptop computer for making

changes to fuel parameters. When connected, the laptop is used to display sensor input

and output data. This data recorded on the laptop is particularly useful for diagnostics and

gathering information about engine load.

Fuel injection requires fuel pressures between 30 and 60 psi. Fuel flow rates must also be

considered when selecting a pump. At full speed and load, the fuel pump must be capable

of over 300 lbf/hour of Aquanol at 40 psi. Another major consideration for the fuel pump

is corrosion. Bosch is the only manufacturer that made a high-pressure automotive fuel

pump fully compatible with Aquanol. Unfortunately, due to low demand, that pump has

been discontinued. Aeromotive, Inc., makes a high quality, high flow pump that is

compatible with our requirements. The internal surfaces are stainless, brass, copper, and

bronze with an aluminum housing. The pump is capable of 450 lbf/hour at 45 psi. One of

the distinguishing features of this pump is that it can be easily disassembled for

inspection or maintenance. If a vehicle were to sit idle for more than a month when setup

for Aquanol use, it is recommended that the fuel loop be purged with gasoline.


Fuel injectors are simply inductance-activated solenoids with a needle and seat to control

fuel flow. The engine has eight injectors so each injector must be capable of 40 lbf/hour

flow. MSD makes a series of injectors that are fully stainless where in contact with the

fuel. Flow ratings were given at 35 psi, and injectors with a max flow rate of 38 lbf/hour

were selected. Running a fuel pressure of 40 psi provides adequate flow for peak

requirements on Aquanol. Larger injectors are available, but should not be used if dual

fuel capability is desired. An injector larger than 40 lbf/hour will not be capable of

accurately metering low gasoline flows at idle conditions.

FIGURE 2 Plan view of the van, showing physical location of components.

The various components, including includes the fuel handling and fuel injection systems

and the exhaust after treatment system, are arranged on the transit van as shown in Fig. 2.


Test Protocol

A primary thrust of this research was to create a test protocol that would allow local

testing of the vehicle that would mimic the FTP driving cycles. This allowed preliminary

testing and comparisons of fuel economy and emissions between the two fuels.

Approximating a FTP driving cycle locally allows the fuel mapping and exhaust after

treatment to be evaluated and modified for best possible vehicle emissions and

performance.

FIGURE 3 FTP Urban driving cycle trace.

The FTP driving cycle is a speed-time trace that a vehicle must follow while a transient

chassis dynamometer mimics road power requirements. The FTP-72 trace imitates city

driving and is shown in Fig. 3, while the Highway Fuel Economy Test (HWFET) is

shown in Fig. 4. Since no transient dynamometer is available for use in the Northwest, an

approximation using the steady-state chassis dynamometer was used.


FIGURE 4 Highway driving cycle trace.

We created a six-mode test that collects data at four steady state points and two mock-

acceleration points (see Table 1). Emissions are reported in percent and parts-per-million

(ppm) depending on the species, and fuel consumption reported in both kg/hour and miles

per gallon (mpg). To estimate driving cycle performance, weighting factors were applied

to each data point representing percent time of each point in the FTP-72 and HWFET

driving cycles. Many countries still use weighted steady-state modal tests for new vehicle

certification. Until 1996, Japan used a six-mode certification for vehicles carrying more

TABLE 1 Description of Six-Mode Points Mode Speed Load Weighting Factor

1 Idle --- 0.05

2 25 mph Road Load 0.35

3 20 mph 50% throttle 0.15

4 45 mph Road Load 0.25

5 40 mph 50% throttle 0.10 6 60 mph Road Load 0.10

than ten passengers, which has been updated to a ten-mode test. Emissions in their six-

mode test are expressed in percent composition, but their loading parameters and

weighting factors differ significantly from the above protocol.


When preparing for chassis dynamometer testing, data was gathered about road load for

three points. The Haltec Halwin software and radar speed sensor was used for initial data

collection. As with roll-down testing, the data was collected on a flat, straight road and

verified multiple times in each direction. Before measuring road load, the vehicle

speedometer was first calibrated. The radar speed sensor was used to determine the

vehicle’s speedometer reading for each speed, and can vary with tire size and slightly

with tire pressure and wear. To measure road load, we drove the vehicle at a steady-state

speed, using the software to collect data on the engine speed, percent throttle, manifold

pressure, engine and air temperature, and injector duty cycle.

The data gathered on the above road testing was used to recreate the same conditions on

the chassis dynamometer. Once the vehicle is set up on the chassis dynamometer, we

began with some light driving to bring the engine to operating temperature. Once the

dynamometer control was stabilized, we adjusted the throttle to achieve the desired

manifold pressure associated with the given road load speed. After verifying that the

injector duty cycle matches the road load data, we collected data on emissions, fuel

consumption, engine and air temperature, and vehicle speed.

Dynamometer Testing

Chassis Dynamometer testing was conducted to establish baseline performance on

gasoline and to calibrate equipment. Figure 5 shows a diagram of the dynamometer

variables measured, identifying user inputs and desired outcomes.

Since performance measurements were dependent on the inputs of dynamometer speed

control and vehicle throttle position, a procedure was developed to calibrate wheel

speeds. First, a road test was performed. Throttle position readings were taken from an

onboard programmable control module while vehicle speed was calibrated as compared

to speed read from a Garmin E-trex GPS with an accuracy of +/- 0.1 mph. The van was

then run on the chassis dynamometer under no load conditions. Six calibrated speeds


FIGURE 5 Dynamometer inputs and desired outcomes.

taken from the road tests were compared to the dyno speed-readings, which gave us a

correction factor on the dynamometer wheel speed. The corrected dyno wheel speed

along with throttle positions determined from the road tests became our input. The

dynamometer’s speed control was set to the six modal points and the matching throttle

positions held constant. When steady state was reached, a torque reading measured at the

wheels was recorded.

In addition to the torque, parallel measurements were taken to record fuel flow and

emissions. The MAX Machinery 710 series positive displacement fuel-metering system

shown in Fig. 6 was used to monitor fuel consumption. Also shown is the remote fuel

tank and fuel-metering box that have the ability to measure both the feed and the return

lines from the engine.


FIGURE 6 MAX fuel metering system.

Exhaust gas species were recorded using an EMS model 5001 five-gas analyzer.

Emissions were recorded about 15 inches from the output of the tailpipe. Although

transient response of this unit is slow, achieving steady state conditions is constrained by

the dynamometers speed control not the five-gas analyzer. The five-gas analyzer has the

ability to accurately measure CO2, CO, NOx, O2, and HC’s. Future work will address

aldehyde emissions with data collected from a Radian FTIR analyzer. Figure 7 represents

key components of the vehicle test platform.

FIGURE 7 Dual-fuel vehicle test platform schematic.


I.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS

Preliminary Results

These results represent road load data and the six-mode simulation performed on the

chassis dynamometer. At this time chassis dynamometer data has been collected on

gasoline and extensive baseline fuel mapping has been performed. A Haltec E6A

electronic control module allows adjustment of air-fuel ratios to near stoichiometric

conditions. Table 2 presents emissions data as well as fuel consumption for the six modal

points.

TABLE 2 Five-Gas Emissions and Fuel Consumption at Six Modal Points for Gasoline

MODE 1 2 3 4 5 6

CO2 (%) 10.1 10.5 10.5 10.4 -- 10.6

CO (%) 0.61 0.46 -- 1.11 -- 0.83

NOx (ppm) 15.5 14.3 18.8 13.7 -- 13.2

HC (ppm) 143 86 -- 52 22 53

Fuel Rate (Kg/hr) 3.35 5.4 22.3 8.3 22.7 12.2

Compared with earlier data taken at similar load points prior to installation of catalytic

converters and map optimization, these results indicate a 95 percent decrease in NOx

emissions and a 67 percent decrease in HC’s. Table 3 shows torques for the throttle

positions at the corresponding modal points.


TABLE 3 Torque and Phrottle Position

Mode Throttle Position (%)

Dyno Speed (mph)

Torque (ft*lb) Horsepower

1 0 0 0 0

2 3 25 155 7

3 50 20 1823 62

4 7 44 128 10

5 50 39 842 58

6 11 58 216 23

Conclusions

This work has produced a robust, reliable vehicle platform for evaluating alternative fuel

handling system that shows no sign of corrosion, and with gasoline fuel economy and

emissions far improved over the original carbureted configuration. Data logging

capabilities have been added that will facilitate analysis of vehicle performance under

controlled test conditions. Design features of the converted transit van are listed below.

TABLE 4 Design Features

Item Feature

Programmable fuel injection fuels Allows control of fuel metering for multiple

Manual fuel switching valves Permits purging to prevent fuel contamination

Total Seal piston rings contact Reduces blow-by and eliminates oil-fuel

On-board data logging values Tracks and records ECM input and output

Fuel system components Resistant to highly corrosive fuels

Stand-alone ignition system source Provides reliable, easily disabled ignition

Next generation catalytic igniter fuels Supports ignition & combustion of alternative

Pressure transducer port Collects in-cylinder data from one cylinder

Duel fuel design Allows fuel switchover in less than one hour


Future Work

Research is currently underway to optimize fuel maps for 65 percent ethanol and 35

percent water blended fuel. Once this is complete, the test protocol developed in this

work will be used to compare vehicle performance on gasoline and ethanol-water. We

expect results to be similar to previous engine conversions at the University of Idaho

where NOx and CO were greatly reduced compared to gasoline combustion. An FTIR

system is currently being installed in our vehicle test cell to support unburned

hydrocarbon speciation monitoring during driving cycle simulations.

Acknowledgements This research is sponsored by the National Institute of Advanced Transportation

Technology (NIATT), the U.S. Department of Transportation, and the Idaho

Transportation Department. Special thanks also to Valley Transit for use of the commuter

van as a vehicle test platform.


PART II. MODELING CATALYTIC IGNITION CONDITIONS OF PROPANE/AIR MIXTURES OVER PLATINUM WIRES

II.A INTRODUCTION

Ignition of fuel/oxygen/nitrogen mixtures over a platinum wire was studied using

microcalorimetry. Experimental results were compared with predictions from a steady-

flow finite element (FEA) model. The average catalyst wire temperatures obtained from

both two- and three-dimensional FEA models over predicted the experimentally obtained

temperatures by 26 and 3 percent, respectively. Modeling assumptions (prescribed end

temperatures of the wire and negligible thermal radiation) and uncertainty in the

convective heat transfer coefficient contributed to the differences between model and

data.

The FEA analysis indicated that axial conduction dominated heat losses from the Pt

catalyst in comparison with radial convection. This finding suggested that thermal breaks

and/or substrates with lower thermal conductivity will reduce heat losses from catalytic

igniters. Future work will determine the threshold-heating rate of the Pt wire at which the

ignition of aqueous ethanol fuel-air mixtures occurs. This work is expected to support

design improvements for catalytic igniters to help alleviate cold starting problems.

Part II. Modeling Catalytic Ignition Conditions of Propane/Air Mixtures over Platinum Wires 19

II.B DESCRIPTION OF PROBLEM

The ignition of gas phase combustion by a heated catalytically active surface involves

coupling of transport processes and chemical kinetics. Cold starting is one of the least

understood aspects of igniter operation and has motivated the construction of a catalytic

plug flow reactor at the University of Idaho. A platinum catalyst inserted downstream of

a specially designed mixing nozzle preheats the fuel/air mixture and is a source of

chemical activation in a well-defined control volume. Ignition of propane/oxygen/

nitrogen mixtures over platinum is currently being studied using microcalorimetry.

Of particular interest is the temperature distribution along the length of the platinum

catalyst. In this work, experimental results are compared with predictions by a steady-

flow finite element model. Relative significance of fuel/air surface heating, axial

conduction to the catalyst supports, surface heat generation, supplemental electrical

heating, radiative losses, and convective losses have been documented from the finite

element studies and preliminary experiments.

Also of interest is the determination of minimum external heating required to achieve

catalytic ignition conditions. This work is expected to support design improvements for

catalytic igniters as well as the catalytic flow reactor, alleviating cold starting conditions.


II.C APPROACH AND METHODOLOGY

Lean mixtures in internal combustion engines are not readily ignited and are

characterized by a low flame propagation velocity. One effective method to improve

flame initiation is catalytic assistance of ignition and combustion. The transition state of

the molecule catalyst complex in this low energy configuration provides an easier route

for the reactants to proceed in subsequent reactions to form the final products. The fact

that the catalyst can have strong influence on the reaction mechanism, in addition to

lowering the overall activation energy, is suggested by the work of Patterson and Kembal

[1963], who obtained a negative reaction order in fuel for the oxidation of olefins on

platinum films. Similar findings have been reported by Moro-Ok, et al. [1967], Schwartz

et al. [1971], and Cardoso and Luss [1969].

Previous experimental research was conducted in a flow reactor to determine the axial

position at the end of a mixing nozzle for insertion of a platinum catalyst wire. For this

purpose, a hot-wire anemometer was used to obtain velocity profiles at various radial

positions and at different axial positions. The experimental results indicate that the

platinum catalyst can be inserted in a plug-flow region between 5.625 diameters and

11.625 diameters downstream of the mixing nozzle.

The primary purpose of the present investigation was to study the temperature

distribution along the length of the platinum catalyst experimentally and compare it with

the finite element (FEA) model developed using the software ALGOR. A finite element

sensitivity study was conducted to study the temperature variation with the changes in the

power supply. Research was performed to investigate the changes in wire temperature

with corresponding changes in the catalyst geometry and fluid (propane-air mixture) flow

conditions. Further, the ALGOR studies determined theoretically the minimum external

heating required to achieve catalytic ignition conditions; the analysis will be helpful for

determining the amount of heat power require to be supplied to the electrical system in

order to achieve a desired temperature. This ALGOR data was used for conducting the

microcalorimetry experiments.


In conventional design engines, excess fuel is supplied during cold starting. Studies in the

specially designed flow reactor aim at understanding the problems associated with cold

starting without the need to supply excess fuel.

Pressurizable Plug--Flow Reactor

A literature search for flow reactor systems provides information about standard designs

features of reactors for combustion studies. Mitchell, et al. [1992] used the flow reactor

described by Hoffman, et al. to compare the combustion of methanol in the reactor with

the combustion of methanol in a diesel engine [Hoffman, et al., 1991]. Another key point

from the literature involving safety, as discussed by Karim, et al. [1994] is that

experiments conducted in a highly lean environment (equivalence ratio of 0.05) offer less

potential for problems [Karim, et al., 1994]. Karim, Dryer [1972], Hoffman, Koert

[1990], Mao and Barat [1996], and Vermeercsh, et al. [1971] also discuss surface

reactions of catalysts and modeling the reactor system. The literature search provided

specifications for the University of Idaho reactor design, operating parameters and

introduced reactor hardware concepts used for previous research.

Propane)

F

Part II. Mo

Fuel Inlet Mixing Flow Combustion

Exhaust Analysis(FTIR)

Syringe Pump O2

N2

Pressure Vent

Catalyst Wire

Exhaust Analysis (FTIR)

IGURE 8 Sections of the flow reactor.

deling Catalytic Ignition Conditions of Propane/Air Mixtures over Platinum Wires 22

First, the fuel—propane— is introduced into the fuel inlet section of the catalytic plug

flow reactor. Next, nitrogen carries the fuel from the inlet section to the mixing section,

where the fuel mixes with oxygen. A turbulence grid injection system is used as a mixing

nozzle. The mixing section is designed in such a manner that it minimizes the mixing

times and ensures complete mixing. In the mixing nozzle, the nitrogen-fuel mixture

tapers into a tube and is vented outward through six small holes, similar to a design

developed at Drexel University [Hoffman, et al., 1991]. Oxygen enters the middle ring

where it fills a large cavity and is mixed with N2 via 18 small holes. A sleeve creates the

outside wall of the cavity, forming a plenum, where oxygen enters.

A prototype-mixing nozzle was machined from polymethyl methacrylate and tested for

fit with the evaporator section (Fig. 9). The gas-mixing nozzle design aims to preserve a

step-change when the concentration of one gas stream—the fuel-nitrogen stream—is

suddenly changed. Anemometer experiments verified the plug flow region. Experiments

using two CO2-N2 gas streams of different concentrations verified complete mixing.

FIGURE 9 Prototype mixing nozzle mounted downstream of the evaporator.

The combustion section of the reactor will contain catalysts in different arrangements as

needed for the experiments. The ongoing research will make use of platinum catalyst

wire for igniting the fuel-air mixture. Instrumentation and calculations related to the wire

will be used to determine the transition temperature from kinetic to diffusion control of

the ignition process, based on the experimental method of Schwartz, et al. [1979].


Experimental Determination of Surface Reaction Temperature

The electrical resistance of platinum varies as

R = Ro [1 + α (T --To) ],

where R is the electrical resistance at temperature T, Ro is the electrical resistance at

temperature To, and α is the coefficient of resistance for Pt, a property of the metal.

With the Pt wire connected to a variable power supply and with voltage monitored

separately, power to the wire was slowly increased as a propane-air mixture was passed

over it. When surface reactions began, the temperature of the wire increased abruptly as

indicated by a sudden change in voltage. The corresponding change in resistance was

calculated and the average wire temperature found from the equation above. The average

temperature obtained at the power level where surface reactions began was used for

comparison with the flow finite element (FEA) model results.

Two-Dimensional Propane-Air Mixture Mixed Mode Heat Transfer Model

A two-dimensional FEA model was generated to study the effects of conduction and

convection on the wire temperature. Initially, a lower mesh density, four-by-four nodes,

was selected, and gradually increased until the temperature remained constant. Table 4

shows the mesh sensitivity analysis.

TABLE 5 Sensitivity to Mesh Density

Mesh Size ALGOR Temperature (K)

4 x 4 467 10 x 10 482 20 x 20 490


Input parameters: We considered a platinum catalyst (rectangle) of 508 micron by

0.015 m in the Y-Z plane. The thickness was specified to be 508 microns. The top

surface was kept at room temperature, 25oC. The temperature profile is symmetrical; the

bottom surface of the FEA model corresponds to the intersection of the catalytic wire and

the axial centerline of the reactor tube. The left surface was exposed to convection

cooling by the fluid flow. The convective heat transfer coefficient calculated in the FEA

model is 53.06 W/m2K. Figure 10 shows the temperature distribution along the length of

the platinum catalyst for a two-dimensional FEA model.

FIGURE 10 A two-dimensional propane air mixture FEA model of platinum temperature distribution.


Sensitivity Study (Fluid Flow)

The fluid (propane-air mixture) velocity for our research application is 1.96 m/s. We are

interested in studying the changes in temperature with respect to the changes in fluid

velocity. The other parameters of mesh size, heat flux, and wire geometry are kept

constant. The results show that the temperature of the catalyst increases as the fluid

velocity decreases and decreases as the fluid velocity increases (Table 6)

When the fluid velocity decreases, heat losses due to convection decrease. Hence, the

temperature of the platinum catalyst increases. Likewise, the convective heat transfer

coefficient changes significantly with changes in the fluid velocity.

TABLE 6 Effects of Fluid Velocity Changes on Catalyst Temperature

Fluid Velocity (m/s) Temperature (K)

0.196 507 1.96 491 19.6 453

Sensitivity Study (Catalyst Geometry)

Studies were conducted to investigate the changes in the average wire temperature with

corresponding changes in the wire diameter (Table 7). The results indicate that higher

average catalyst temperatures were obtained for thicker wires and the catalyst

temperature decreases as we used thinner wires. This is due to the fact that axial

conduction dominates radial convection.

TABLE 7 Effects of Wire Diameter Changes on Average Catalyst Temperature.

Wire Diameter (micro m) Temperature (K)

254 472 508 491 762 496


Although this case study gives the expected results, one parameter, the heat transfer

coefficient, still puzzles us. The heat transfer coefficient remained the same for the three

wire diameters. According to convective heat transfer correlations for cross flow over

heated rods, we expect the heat transfer coefficient to change as wire diameter changes.

Further calculations were conducted by changing the wire diameter by a factor of 10, but

the same value of the heat transfer coefficient was obtained.

Sensitivity Study (Power Input)

Calculations were carried out to study the average temperature variation with the changes

in the power supply. See Table 8 for temperature results at various power inputs.

TABLE 8 Effects of Power Supply Changes on Average Catalyst Temperature.

Power (W) r Temperature (K)

0.79 4620.85 4760.93 4911.06 5191.09 5241.16 5401.24 5561.32 605

The results show an increase in average catalyst temperature with the changes in the

power supply. As the heat input to the system increases, the average catalyst temperature

increases. In the laboratory experiments we observed a similar trend. The average

temperature obtained at the power level where surface reactions began using this two-

dimensional analysis is 605 K, 26 percent higher than the experimental average

temperature of 480 K.

Three-Dimensional FEA Model

A three-dimensional FEA model was generated to study the effects of conduction and

convection on the wire temperature. For our input parameters, we considered a platinum


catalyst (cylinder) of 508-micron diameter and 0.015. The top surface was kept at room

temperature, 25oC. The temperature profile is symmetrical; the bottom surface of the

FEA model corresponds to the intersection of the catalytic wire and the axial centerline of

the reactor tube. The average temperature obtained at the power level where surface

reactions began using this three-dimensional analysis is 495 K, only three percent higher

than the experimental average temperature of 480 K.

FIGURE 11 A three-dimensional propane air mixture FEA model of platinum temperature distribution.

A three-dimensional model allows specifying the boundary conditions more uniformly

throughout the entire surface area, whereas in a two-dimensional model, we can specify

the boundary conditions on only one of the four surfaces. Hence, a three-dimensional

analysis predicts the catalyst temperature more accurately than a two-dimensional

analysis. Figure 11 shows the temperature distribution along the length of the platinum

catalyst for a three-dimensional FEA model.


I.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS Comparisons of Two-Dimensional and Three-Dimensional Results with the Experimental Results

Figure 12 compares the average catalyst wire temperature obtained experimentally with

that calculated by the two-dimensional and three-dimensional FEA models. Differences

between experiment and model may be due to modeling assumptions (the end

temperature of the wire is prescribed and thermal radiation is neglected) and uncertainty

in the convective heat transfer coefficient.

0

100

200

300

400

500

600

700

0.79 0.85 0.93 1.06 1.09 1.16 1.24 1.32

POWER (WATTS)

TEM

PER

ATU

RE

(K)

2-D ALGOR RESULTS

EXPERIMENTAL RESULTS

3-D ALGOR RESULTS

FIGURE 12 Catalyst temperatures versus changes in power supply.

The three-dimensional boundary conditions predict the catalyst temperature more

accurately than the two-dimensional boundary conditions. By conducting the FEA

propane-air mixture analysis, we learned that the axial conduction dominates radial

convection on the platinum catalyst. Because both two-dimensional and three-

dimensional results differ from the experimental results, it indicates that physical

assumptions such as a prescribed top surface temperature and negligible thermal


radiation, and uncertainties in the convective heat transfer coefficient are critical for

matching experimental conditions.

Ongoing Research

The next step in the research is to include radiation losses in the three-dimensional model

that we have already generated, perform parametric studies for model sensitivity to a

prescribed end temperature and convective heat transfer coefficient, and compare with

the experimental results. Further efforts will be made to conduct transient heat transfer

analysis for this model using ALGOR. The basic aim will be to find a correlation

between the experimental wire temperatures achieved and FEA wire temperatures

calculated for the same power supply. The FEA propane air mixture analysis will be

helpful to determine the amount of power required to be supplied to the electrical system

in order to achieve a desired temperature. Further, the results of the FEA sensitivity

analysis will provide an insight into accommodating system modifications including

changing catalyst geometry, power supply and wall heating.

A platinum catalyst will be inserted into plug flow region, between 5.625 diameters and

11.625 diameters, identified by hot-wire analysis.

Future work will also include testing with other fuels including aqueous ethanol-air

mixtures. Data will be collected in order to determine the temperature of the platinum

catalyst required for ignition of a particular volumetric content of water in ethanol at a

specified equivalence ratio. Further investigation will be carried out in order to

understand the changes in ignition temperature for different volumetric concentrations of

aqueous ethanol at a constant excess air coefficient. This research aims at determining the

threshold point at which the ignition of aqueous ethanol fuel-air mixtures occur.

Acknowledgments In addition to a University Transportation Center grant from the US Department of

Transportation through NIATT, this research is sponsored by a DOD-EPSCOR grant.


PART III. THEORETICAL STUDY OF AQUEOUS ETHANOL-AIR COMBUSTION IN PLUG FLOW

III.A INTRODUCTION

The thermal decomposition and combustion kinetics of gas-phase ethanol oxidation were

modeled with a computer code. The output was compared with flow reactor data available

in the literature. Additional calculations were performed with different percentages of

water added to the fuel. Temperature, the consumption of ethanol, and the mole percent of

the major products carbon dioxide, water and excess oxygen (CO2, H2O and O2) and

intermediate species carbon monoxide, methane, acetylene, ethene, ethane, hydrogen, and

acetaldehyde (CO, CH4, C2H2, C2H4, C2H6, H2 and CH3CHO) were plotted as a function of

time. Reasonable agreement was found between the model and data for 100 percent ethanol

oxidation. Model results showed that combustion temperature dropped as water content

increased. Because dilution masked the impact of water on decomposition and combustion

kinetics, an analysis is underway to “dry” model results. The computer code is being

modified to include the kinetics of catalytic surface reactions.

Part III: Theoretical Study of Aqueous Ethanol-Air Combustion in Plug Flow 31

III.B DESCRIPTION OF PROBLEM

A detailed chemical kinetic mechanism developed for homogeneous ethanol oxidation was

examined using the HCT (Hydrodynamics, Combustion, and Transport) code developed at

Lawrence Livermore National Laboratory. The output was compared with combustion data

from a flow reactor available in the literature. Additional calculations were performed with

different percentages of water added to the fuel. Results show that combustion temperature

dropped as water content increased, with corresponding decreases in the rate of production

of stable intermediary species. The consumption of ethanol and the mole percent of major

products CO2, H2O and O2 and intermediate species CO, CH4, C2H2, C2H4, C2H6, H2 and

CH3CHO are plotted as a function of residence time. These calculations are a first step in

developing an understanding of cold starting and emissions from ethanol-water combustion

in air after catalytic ignition over platinum.


III.C APPROACH AND METHODOLOGY

A catalytic igniter [Cherry, et al., 1990; Cherry, 1992; Cherry, et al., 1992] permits stable

combustion at fuel lean conditions in internal combustion engines [Morton, 2000]. The

igniter also supports combustion of high water content fuel, for example 70 percent

ethanol/30 percent water blends, and permits the combustion of heavy fuels in light

engines.

Water in fuel offers several advantages in lowering harmful combustion emissions. High

water content lowers combustion temperature, thus impeding the formation of thermal NO.

Due to the thermodynamics of CO oxidation, water also encourages the oxidation of CO to

CO2. Soot formation also appears to be inhibited in the presence of water [Hall-Roberts, et

al., 2000]. However, increased acetaldehyde emissions may be a problem during cold

starting since this species is a stable intermediary of the dominant path for gas-phase

ethanol combustion [Marinov, 1998; Egolfopoulos, et al., 1992; Norton and Dryer, 1992].

Our approach to assist in the development of the igniter takes three parallel paths:

a) Understanding the chemistry of catalytic ignition through detailed modeling and

plug-flow reactor studies

b) Working with stand-alone engines to improve igniter durability and

manufacturability, and modeling key characteristics for predicting future designs;

c) Creating demonstration vehicles for over-the-road testing and public outreach.

The work presented here describes our progress in modeling the impact of water on gas-

phase ethanol combustion. These calculations are a first step in developing an

understanding of ethanol water combustion after catalytic ignition over platinum. From

detailed modeling, a simplified model will be derived that improves a catalytic ignition-

timing model being developed to assist igniter designs for new engine applications.


Hydrodynamics, Combustion, and Transport (HCT) Code

Our goal is to model the detailed chemical kinetics of platinum catalyzed ethanol-water

ignition. We are using the Hydrodynamics, Combustion, and Transport (HCT) code

developed at Lawrence Livermore National Laboratory. HCT is a finite-difference code

that calculates one-dimensional time-dependent problems involving gas hydrodynamics,

transport, and detailed chemical kinetics. It can calculate ignition occurrence, laminar

flame propagation, and species mole fractions. The code solves the coupled equations for

conservation of mass, momentum, energy, and conservation of each chemical species. The

inherently stiff set of equations is solved implicitly using the numerical method, LU

decomposition.


FIGURE 13 Logic flow and file interaction for the MKCDAT, HCT, and HCTPLT programs

The HCT codes include MKCDAT, HCT and HCTPLT. Figure 13 shows the HCT

flowchart and the changes we made for plotting. HCT contains the main calculation

program. The input file, HCT.INP describes the combustion model, including reactions and

the species involved, initial conditions and boundary conditions, the model execution

method, and output files includes hsp0, defile0, and plot.xls. Among these output files,

hsp0 is an ASCII file that contains the simulation process, species mole fraction changes

and other information; dfile0 is a binary file for plotting with the HCTPLT program.

The MKCDAT code produces binary and ASCII chemistry data files, reads in.cdat and a

possible changes file (e.g. changes.195) and writes cdat (binary file required by the HCT

code) and out.cdat, which we rename in.cdat (for changes.195) for use in the next version.

The HCTPLT code is a post-processor plotter and requires NCAR and GPS graphics

library with input command file INPUT. Types of plots made include spatial and time.

Variables plotted are mole fractions, temperature, zone width, velocity, and flame position

and speed.

To satisfy more flexible and powerful plotting requirements, we made some changes in

HCT. When running HCT, the file plot.xls is produced so that with MS Excel or other data

analysis and graphics tools, we can plot powerfully and easily.

Theoretical Impact of Water on Gas-Phase Ethanol Combustion

For the flow reactor combustion of ethanol at atmospheric pressure, initial temperatures

near 1100 K and equivalence ratio φ = 0.61, 0.81 and 1.24, experimental profiles of stable

species mole fractions and temperature are reported by Norton and Dryer [1991].

The stable species CO2, O2, CO, C2H6, CH4, H2O, C2H5OH, H2, CH3CHO, C2H4, C2H2, and

temperature are plotted.


CH3CH2OCH2CH2OH CH3CHOH

C2H5OH

CH2OC2H4 CH3CHO

HO2

C2H3

CH3COHCO

CH3

CH2O

C2H6 CH2O CH3OCH4 CH3OH

FIGURE 14 Species consumption path analysis for ethanol oxidation [Egolfopoulos, 1992; Norton and Dryer, 1991].

The species decomposition paths for ethanol/air oxidation are shown in Fig. 14. We

developed these paths by integrating all reactions in the oxidation process and determining

the fraction of each species consumed by a particular reaction. C2H5OH is mainly attacked

by H, OH, O, CH3, C2H5 and HO2, which abstract H atom and produce three C2H5O

isomers CH3CHOH, CH2CH2OH and CH3CH2O:

C2H5OH + X -> {CH2CH2OH, CH3CHOH, CH3CH2O} +

XH {X = OH, H, CH3, HO2, C2H5} R1

The ethanol mechanism calculates the three distinct sites of hydrogen abstraction from the

ethanol molecule; subsequently the mechanism considers the reactions involving these

three isomers. CH3CHOH reacts with O, O2 and OH to produce acetaldehyde (CH3CHO).

CH2CH2OH decomposes to C2H4, which is the main step to produce ethane. CH3CH2O

thermally decomposes to CH3 with simultaneous product CH2O. The major products of

these three isomers reaction series and their subsequent reactions are expected to be

important in the mechanism.


Detailed chemical kinetics model developed since 1990 have greatly improved the degree

of understanding of ethanol oxidation. The modeling studies of Borisov, et al. [1992] and

Norton and Dryer [1991] addressed the issue of three distinct H-atom abstraction sites in

ethanol as described above, and the resulting temperature dependent product distribution in

combustion. Egolfopoulos, et al. [1992] published the detailed kinetic scheme of the study

of ethanol in laminar-premixed flames, flow reactors, and shock tubes, and developed an

ethanol oxidation mechanism based on his methanol kinetics model. These models had

reasonable agreement with experimental data (Fig. 15).

FIGURE 15 Comparison between numerical calculations and experimental data for flow reactor studies of ethanol oxidation at φ = 0.81 [Egolfopoulos, 1992; Norton and Dryer, 1991].

More recently, Marinov [1998] studied ethanol oxidation at high temperature and

developed a detailed and improved model. The ethanol combustion mechanism reported by

Marinov [1998] was in particular agreement with experiment data in laminar flames, shock

tubes and jet-stirred reactors. Unfortunately his model was unable to reproduce the ethanol

consumption profile for φ = 0.61 in the flow reactor even considering the uncertainty in the

experimental induction time. Certainly for equivalence ratio φ = 0.81 and 1.24, numerical

results were time-shifted to match the experimental data from Norton and Dryer [1991]

correctly.


Curran and Pitz [1997] modeled dimethyl ether CH3OCH3 (DME) combustion. This DME

kinetics model was rigorously tested through comparisons with experimental data. At high

temperature, the fuel consumption pathway for DME is similar to some degree to ethanol in

this model, since some dimethyl ether changes to ethanol at first, then ethanol reacts along

the reaction channels described by Norton and Dryer [1991], Egolfopoulos [1992],

Brorisov [1992], and Marinov [1998] (Fig. 13), while remaining DME reacts along its

decomposition channels. Hence, to some extent, this DME model can simulate ethanol

oxidation mechanism. However, the DME model is particularly developed for DME; using

it to predict ethanol oxidation requires additional effort. We modified the DME model with

more detailed ethanol decomposition reactions developed by Marinov [1998] and formed a

modified DME model. This model was used to simulate ethanol-water oxidation. The

results presented here show reasonably good agreement in comparison with plug-flow data

reported by Norton and Dryer [1991].

In our research about catalytic ethanol-water air oxidation, we first consider homogenous

oxidation at atmospheric pressure. The initial reactant mixture includes ethanol and air. We

examine the effect of increasing fuel-water content by varying the initial fuel from 100

percent ethanol and 0 percent water to 70 percent ethanol and 30 percent water in a plug

flow reactor. The numerical results for equivalence ratio φ = 0.81 were time-shifted by 15

msec.

0 20 40 60 80 1000.0

0.1

0.2

0.3

0.4

0.5

0.6

30% h2o

20% h2o

10% h2o

0% h2o

C2H5OH @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 16 C2H5OH vs. residence time at φ = 0.81.


Although both the HCT simulation and the experimental data show steady consumption of

ethanol (Fig. 16), the data indicate a faster combustion rate than calculated by the

simulation. As water content increases, the rate of ethanol consumption decreases because

of lowered combustion temperature, as shown in Fig. 17.

0 10 20 30 40 50 60 70 80 90 100

1100

1125

1150

1175

1200

1225

1250

1275

1300

1325

30% h2o

20% h2o

10% h2o0% h2o

Temperature @T=1090, φ=0.81

Tem

pera

ture

(K)

Time (msec)

FIGURE 17 Temperature vs. residence time at φ = 0.81.

0 20 40 60 80 1000.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

2.2

2.4

30% h2o

20% h2o

10% h2o

0% h2o

O2 @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 18 O2 vs. residence time at φ = 0.81.

Figure 18 shows that the HCT prediction is in reasonable agreement with experimental

data. As the amount of water increases in the fuel, the consumption rate of O2 becomes

lower in concert with corresponding decreases in the rate of ethanol decomposition and

temperature. Closer comparison between Fig. 17 and 18 shows that once the temperature

starts to increase abruptly, the consumption of O2 increases quickly.


0 20 40 60 80 1000.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

30% h2o

20% h2o

10% h2o

0% h2o

CO @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 19 CO vs. residence time at φ=0.81.

CO mole percent is plotted as a function of residence time in Fig. 19. With 10 percent and

more water, the rate of CO formation slows in comparison with the 0 percent water

condition and the CO peaks are not so abrupt as in 0 percent water. The peak production of

CO is coincident with the rapid temperature increase.

The CO oxidation reactions include the following:

2 2CO +O CO + O R2→

2

2

O + H2O OH + OH R3CO + OH CO + H R4H + O OH + O R5

→→

→

Glassman [1987] indicated that with the peroxy radical HO2 present, another route for CO

oxidation is possible:

2 2CO + HO CO + HO R6→

In Fig. 19, HCT shows that CO2 mole faction with 0 percent water in initial condition is in

reasonable agreement with the experimental data. The abrupt increase of production of CO2

is in good conformity to the temperature curve in Fig. 16, for 0 percent water, 10 percent

water, 20 percent water and 30 percent water. The CO2 formation rate increases abruptly in

concert with the abrupt temperature change. And with the increase of water, the final CO2

fraction decreases because of dilution due to the large amount of water present.


0 20 40 60 80 1000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

30% h2o

20% h2o10% h2o

0% h2o

CO2 @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 20 CO2 vs. residence time at φ = 0.81.

Reactions R2-R6 shows the mechanism between CO and CO2. Reaction 2 is the initiator of

the chain sequence; the fastest and the dominant CO oxidation mechanism is R4 [Turns,

1996]. As water content increases, thermodynamic equilibrium considerations argue that

the reaction R4 is pushed further towards reactants. However, chemical kinetics

calculations show that with increase in water, the reaction temperature decreases

considerable in Fig. 20, which causes the severe decrease in the CO2 production rate.

In Fig. 21, the formation reaction R7 of C2H6, methyl radical recombination

3 3 2 6CH + CH C H R7→

is the major contributor to C2H6 [Egolfopoulos, 1992], following the ethanol decomposition

path in Fig. 13. The key reactions involving C2H6 include

2 6 2 5 2

2 6 2 5

2 6 2 5 2

C H + H C H + H R8C H + O C H + OH R9C H + OH C H +H O R10

→→→

Figure 20 shows the general tendency for the addition of H2O to lower the reaction speed.

HCT simulation also correctly describes the fast decrease of C2H6 after the mole fraction

apex, but significantly underestimates the amount of ethane produced.


0 20 40 60 80 1000

20

40

60

80

100

120

140

160

180

200

220

240

30% h2o20% h2o

10% h2o

0% h2o

C2H6 @T=1090, φ=0.81

Mol

e P

PM

Time (msec)

FIGURE 21 C2H6 vs. residence time at φ = 0.81.

0 20 40 60 80 1000

200

400

600

800

1000

1200

1400

30% h2o20% h2o

10% h2o

0% h2o

C2H4 @T=1090, φ=0.81

Mol

e P

PM

Time (msec)


The main reactions involved in C2H4 and C2H2 include

2 4 2 2 2

2 4 2 3

2 4 2 4 2 3 2 5

C H + M C H + H + M R11C H + M C H + H + M R12

C H + C H C H +C H R13

→→

→

and

2 3 2 2 2

2 3 2

2 3 2 2 2 2

2 3 2

C H + H C H + H R14 C H + O H CO + H R15C H + O C H + HO R16C H + M C H

C→→→→ 2 + H + M R17


In Fig. 22, the HCT simulation shows that C2H4 production mainly follows the

decomposition of CH2CH2OH in Fig. 13. Ethene’s consumption follows R11-R17 and the

path about C2H4 in Fig. 13, and quickly disappears with the abrupt increase of temperature.

Figure 23 shows significant difference between the simulated C2H2 mole fraction and the

experimental data for the case with pure ethanol (0 percent water). One reason for the

difference is the difficulty reading the exact experimental data from Norton and Dryer

[1992] due to the graphics scale. As the temperature decreases with increasing water

content, the production rate of acetylene slows. At a fixed temperature, the rate of C2H2

production decreases abruptly. The dominant reactions here are

2 2 2

2 2 2

2 2 2

C H + M C H + H + M R18C H + O HCCO + OH R19C H + O HCO + HCO R20

→→→

It is clear that C2H2 production is closely connected with C2H4 directly in R11through R17

and through the intermediary C2H3.

0 20 40 60 80 1000

10

20

30

40

30% h2o

20% h2o

10% h2o

0% h2o

C2H2 @T=1090, φ=0.81

Mol

e P

PM

Time (msec)



0 20 40 60 80 1000

200

400

600

800

1000

1200

30% h2o

20% h2o

10% h2o0% h2o

CH3CHO @T=1090, φ=0.81

Mol

e P

PM

Time (msec)

FIGURE 24 CH3CHO vs. residence time at φ = 0.81.

The reactions that directly affect the production and consumption of CH3CHO are

3 3

2 2

3 3

CH CHOH + X -> CH CHO + XH R21 {X=O , O, H, OH and HO }CH CHOH + M -> CH CHO + H + M R22

and

3 3

2 2

3 2

CH CHO + X -> CH CO + XH R23 {X=O , O, H, OH and HO }CH CHO + X -> CH CHO + XH R24

3 2

{X=CH , O, H, OH and HO }

Egolfopoulos, et al. [1992] developed a CH3CHO reaction channel and Marinov [1998]

further refined this channel. The production reactions R21 and R22 of CH3CHO, following

the production of CH3CHOH, directly result from the ethanol decomposition in Fig. 16.

The CH3CHO consumption reactions R23 and R24 are sensitive to temperature. Hence

with 10 percent, 20 percent, and 30 percent H2O, the rate of decrease of acetaldehyde

production is steep. In Fig. 24, HCT shows that the modified ethanol mechanism is in

reasonable agreement with the experimental data for CH3CHO.

The intermediate H2 reaction mechanism connects with many reactions and intermediate

and stable species in the modified DME mechanism because of the importance of H, O and


OH chemistry. The rate of H2 formation increases steadily before the temperature reaches

its sharp increase (Fig. 25). Higher H2O content slows the rate of H2 formation. The

experiment data and HCT simulation are in reasonable agreement at the 0 percent fuel

water case.

0 20 40 60 80 1000.00

0.05

0.10

0.15

0.20

0.25

0.30

30% h2o

20% h2o

10% h2o0% h2o

H2 @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 25 Water vs. residence time at φ = 0.81.

0 20 40 60 80 1000.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

2.0

0% h2o

H2O @T=1090, φ=0.81

Mol

e P

erce

nt

Time (msec)

FIGURE 26 Water vs. residence time at φ = 0.81.

The water routes in the oxidation mechanism are connected to many other reactions and are

not included here. The experiment data and HCT simulation are in reasonable agreement at

the 0 percent fuel water case for the production of water in combustion (Fig. 26). The plot


does not include simulation with increased water in the fuel because the fuel-water

overwhelms the plot and obscures slight changes in mole percent with time.

As shown in Fig. 27 for CH4, its production mainly follows the reaction R25:

3 4CH + H + M -> CH + M R25

Methane consumption mainly reverses to the production of CH3 again. Once the abrupt

temperature increase starts, most of CH3 is consumed, which leads to an abrupt drop in CH4

production.

0 20 40 60 80 1000

200

400

600

800

1000

1200

1400

30% h2o

20% h2o

10% h2o

0% h2o

CH4 @T=1090, φ=0.81

Mol

e P

PM

Time (msec)

FIGURE 27 CH4 vs. residence time at φ =0.81.


III.D FINDINGS; CONCLUSIONS; RECOMMENDATIONS

The development of a catalytic igniter that permits stable combustion of high water-

content ethanol in internal combustion engines has promoted a need for fundamental

understanding of Pt-catalytic ethanol water ignition chemistry. As a first step in

developing a heterogeneous model, the HCT code was used with an ethanol oxidation

mechanism to model gas-phase ethanol-water-air combustion.

A modified ethanol mechanism was based on an existing dimethyl ether mechanism,

and included the species consumption path analysis for ethanol oxidation.To gain

confidence in the ethanol oxidation mechanism, simulated gas-phase ethanol-air

combustion was compared with experimental data from a plug flow reactor.

Due to the uncertainty in experimental induction time, the simulation results were

“time-shifted.” The calculations are in reasonable agreement with experimental data

obtained with 100 percent ethanol, with the exception of intermediate species C2H6,

(overperdicted) and C2H2 (underpredicted), and slower C2H5OH decomposition. The

simulation shows that as more water is added to ethanol, the reaction temperature

decreases and the rate of the oxidation process decreases. The addition of water results

in lower peak valued of intermediate species because of dilution.

Future Plans

1. Refine the ethanol-air combustion mechanism as new information and insights

become available to improve agreement with data at φ=0.81. Compare an improved

model with data available at φ=0.61 and 1.24.

2. Search for and evaluate hydrocarbon-Pt surface reaction data, especially that for

initial ethanol-Pt decomposition reactions. Find characteristics of Pt-hydrocarbon

oxidation that will be useful in developing a Pt-ethanol-water oxidation mechanism.

3. Develop a 1-D model for catalyst surface reactions through a literature search and

theoretical analysis of surface chemistry.


4. Modify the HCT code to model the coupled homogeneous-heterogeneous

catalytically assisted ignition of ethanol-water fuel.

5. Compare the result with ethanol-water-air plug flow experiments in a reactor that is

being developed.

6. Derive simplifications that can be used to improve a catalytic ignition timing model

being developed to assist igniter designs for new engine applications

Acknowledgements

Drs. C. Westbrook, W. Pitz and L. Chase of Lawrence Livermore National Laboratory

provided valuable assistance understanding the operation of HCT and its application.


REFERENCES

Borisov, A. A., V. M. Zamanskii, A. A., Konnov, V. V., Lisyanskii, S. A., Rusakov, and G. G. I. Skachkov, “A mechanism of high-temperature ethanol ignition,” Sov. J.Chem.Phys. 9:2527-2537, 1992.

Cardoso, M., and D. Luss, Chem. Eng. Sci. 24:1699-1710 (1969).

Cherry, M. A., “Catalytic-compression timed ignition,” US Patent 5,109,817, May 5, 1992.

Cherry, M., R. Morrisset, and N. Beck, “Extending lean limit with mass-timed compression ignition using a plasma torch,” SAE Paper 921556, 1992.

Cherry, M., Catalytic-Compression Timed Ignition, US Patent 5 109 817, December 18, 1990.

Cherry, M. A. and Elmore, C. L., “Timing chamber ignition method and apparatus,” US Patent 4,977, 873, December 18, 1990.

Clarke, E., “Characterization of Aqueous Ethanol Homogeneous Charge Catalytic Compression Ignition,” Masters Thesis, University of Idaho, 2001.

Cordon, D., E. Clarke, S. Beyerlein, J. Steciak, and M. Cherry, “Catalytic igniter to support combustion of ethanol-water/air mixtures in internal combustion engines,” SAE Paper 02FFL-46, 2002.

Curran, H.J. and Pitz W. J., “A wide range modeling study of dimethyl ether oxidation,” Int. J. Chem. Kin., Vol. 30, pp. 229-241, 1998.

Dryer, F. L., “High temperature oxidation of carbon monoxide and methane in a turbulent flow reactor,” AMS Report T-1034, 1972.

Egolfopolous, F. N., D. X. Du, and C. K. Law, “A study on ethanol oxidation kinetics in laminar premixed flames, flow reactors, and shock tubes,” Twenty-Fourth Symposium (International) on Combustion, The Combustion Institute, pp. 833-841, 1992.

Glassman, I., Combustion. 2nd Ed. Orlando: Academic Press, 1987.

Gottschalk, Mark A., “Catalytic ignition replaces spark plugs,” Design News, May 22, 1995.

Hall-Roberts, V. J., Hayhurst, A. N., Knight, D. E., Taylor, S. G., “Origin of soot in flames: is the nucleus an ion?” Combustion and Flame, vol. 120, No. 4, pp. 578-584, March, 2000.


Hoffman, J. S., W. Lee, A. Litzinger, D. A. Santaciaa, and W. J. Pitz, “Oxidation of propane at elevated pressures: experiments and modeling,” Combustion Science and Technology, vol. 77, pg.95-125, 1991.

Jehlik, F., M. M. Jones, P. Shepherd, J. Norbeck, K. Johnson, and M. McClanahan, “Development of a low-emission, dedicated ethanol-fuel vehicle with cold-start distillation system,” SAE Paper 1999-01-0611, 1999.

Karim, H., L. D. Pfefferle, P. Markatou and M.Smooke, “Modeling heterogeneous/ homogeneous reactions and transport coupling of catalytic combustion of methyl chloride over a Mn-based catalyst,” Twenty-Fifth International Symposium on Combustion, pg.299-306, 1994.

Koert D. N., “Effects of pressure on hydrocarbon oxidation chemistry.” Ph.D. Dissertation, Drexel University, 1990.

Konnov, A.A., “Development and validation of a detailed reaction mechanism for the combustion of small hydrocarbons,” 28th Symposium (Int.) on Combustion, Edinburgh, Abstr. Symp. p. 317, 2000.

Lee, W., and Geffers, W., “Engine performance and exhaust emissions characteristics of spark ignition engines burning methanol and methanol mixtures,” A.I.Ch.E. Symposium Series #165, Vol. 73, 1977.

Lund, C. M., “HCT: A general computer program for calculating time-dependent phenomena involving one-dimensional hydrodynamics, transport, and detailed chemical kinetics,” revised by L. Chase, Lawrence Livermore National Laboratory, UCRL-52504, 1995.

Mao, F., and R. B. Barat, “Experimental and modeling studies of staged combustion using a reactor engineering approach,” Chemical Engineering Communications, vol.145, pg. 1-21, 1996.

Marinov, N. M., “A detailed chemical kinetic model for high temperature ethanol oxidation,” U.S. Department of Energy / Lawrence Livermore National Lab Contract W-7405-ENG-48, 1998.

Mitchell, W. L., T. A. Litzinger, and W.Lee, “Effects of in-cylinder catalysts on combustion and emissions of a D.I. diesel engine fueled on methanol,” Society of Automotive Engineers, Paper # 920688, 1992.

Morton, A. T., “Homogeneous charge combustion of aqueous ethanol in a high compression catalytic engine,” MS Thesis, University of Idaho, 2000.

Morton, A., M. Genoveva, S. Beyerlein, J. Steciak, and M. Cherry, “Aqueous ethanol-fueled catalytic ignition engine,” SAE Paper 1999-01-3267, 1999.


Nadkarni, R. A., Guide to ASTM Test Methods for the Analysis of Petroleum Products and Lubricants. West Conshohocken, PA: ASTM Publishing, 2000.

Norton, T. S., and Dryer, F. L., “An experimental and modeling study of ethanol oxidation kinetics in an atmospheric pressure flow reactor,” International Journal of Chemical Kinetics, Vol. 24, pp.319-343, 1992.

Patterson, W.R., and Kemball, C., J. Catalysis 2:465-478 (1963).

Schwartz, A., L.L.Holbrook, H.Wise, “Catalytic oxidation studies with platinum and palladium,” Journal of Catalysis, Vol. 21, pg.199-207, 1971.

Turns, S. R., An Introduction to Combustion. New York: McGraw-Hill, Inc., 1996.

Vermeersch, M. L., T. J. Held, Y. Stein, and F. I. Dryer, “Auto-ignition chemistry studies of n-butane in a variable pressure flow reactor,” Society of Automotive Engineers, Paper # 912316, 1991.

Westbrook, C. K., and F. L. Dryer, “Chemical kinetic modeling of hydrocarbon combustion,” Prog. Energy Combust. Sci., Vol. 10, pp. 1-57, 1984

Wyman, C. E., Handbook on Bioethanol, Production and Utilization. National Renewable Energy Laboratory, NICH Report 24267 (424), 1996.


Modeling and Application of Catalytic Ignition in Internal ... … · Modeling and Application of Catalytic ... Two-Dimensional Propane-Air Mixture Mixed Mode Heat Transfer ... of

Documents