A Study on Biogas-Fueled SI Engines: Effects of Fuel

Composition on Emissions and Catalyst Performance

by

Robert Abader

A thesis submitted in conformity with the requirements

for the degree of Master of Applied Science
Graduate Department of Mechanical and Industrial Engineering
University of Toronto

Copyright c 2014 by Robert Abader

 Abstract

A Study on Biogas-Fueled SI Engines: E↵ects of Fuel Composition on Emissions and

Catalyst Performance

Robert Abader

Master of Applied Science

Graduate Department of Mechanical and Industrial Engineering

University of Toronto

2014

Biogas as a fuel is attractive from a greenhouse standpoint, since biogas is carbon neutral.

To be used as such, increasingly stringent emission standards must be met. Current

low-emission technologies meet said standards by precisely controlling the air-fuel ratio.

Biogas composition can vary substantially, making air-fuel ratio control difficult. This

research was conducted as part of a larger project to develop a sensor that accurately

measures biogas composition. Biogas was simulated by fuel mixtures consisting of natural

gas and CO2 ; the e↵ects that fuel composition has on emissions and catalyst performance

were investigated. Engine-out THC and NOx increased and decreased, respectively, with

increasing CO2 in the fuel mixture. Doubling the catalyst residence time doubled the

conversion of THC and CO emissions. The e↵ectiveness of the catalyst at converting

THC emissions was found to be dependent on the relative proportions of engine-out

THC, NOx and CO emissions.

ii
 Acknowledgements

First and foremost I would like to thank Professor James Wallace for providing the

opportunity for me to work on this project, and for all the support he’s provided over the

last couple of years. I never left his office without a sense of direction, and his guidance

will not be forgotten.

I would also like to thank my lab mates Charles, Dan, Manuel, Mark, Phil, Silvio and

Vahid for always being there to lend a helping hand, o↵er their thoughts and just overall

making my experience here enjoyable. A special thank you goes out to Osmond Sargeant

and Terry Zak for all the help they’ve provided, and for not letting me take myself too

seriously. I would not have been able to complete this project without their expertise.

Last but definitely not least, I’d like to thank my family and friends for all their love,

support and unending questions about when I’d finally finish.

iii
Contents

1 Introduction 1

2 Background and Literature Review 3

2.1 Current Energy Needs . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2.2 Biogas As a Promising Alternative . . . . . . . . . . . . . . . . . . . . . 4

2.3 Biogas Composition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

2.4 Previous Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

2.4.1 Stoichiometric and Lean-Burn Hydrogen Addition Experiments . 9

2.4.2 Hydrogen Addition with EGR Systems . . . . . . . . . . . . . . . 12

2.5 Summary and Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . 15

3 Test Equipment and Setup 18

3.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

3.2 Engine Specifications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.3 Catalyst Formulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.4 Stamford Electrical Generator . . . . . . . . . . . . . . . . . . . . . . . . 21

3.5 Measurement Apparatus . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.6 Throttle and Air-Fuel Ratio Control . . . . . . . . . . . . . . . . . . . . 25

3.6.1 Speed Control . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.6.2 Air-Fuel Ratio Control . . . . . . . . . . . . . . . . . . . . . . . . 29

iv
4 Testing 31

4.1 Preliminary Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

4.1.1 Natural Gas Justification . . . . . . . . . . . . . . . . . . . . . . . 31

4.1.2 Determination of Optimal Spark Timing and Equivalence Ratio . 34

4.2 Brake Specific Energy Consumption . . . . . . . . . . . . . . . . . . . . . 37

4.3 Emissions Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.3.1 Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40

4.4 Brake Specific Emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . 44

4.5 Catalyst Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.5.1 Emission Reductions . . . . . . . . . . . . . . . . . . . . . . . . . 48

4.5.2 Catalyst Performance vs. Relative Emission Ratios . . . . . . . . 54

5 Conclusions 58

6 Recommendations for Future Work 61

Appendix A Engine Tuning 67

A.1 Spark Timing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67

A.2 Optimal Equivalence Ratio for Catalyst Performance . . . . . . . . . . . 74

v
List of Tables

2.1 Theoretical gas yields and methane content [33]. . . . . . . . . . . . . . . 6

2.2 Fuel composition (in volume %) of di↵erent gaseous fuels [30]. . . . . . . 6

3.1 Kubota DF-750 specifications. . . . . . . . . . . . . . . . . . . . . . . . . 20

3.2 Catalyst specifications [31]. . . . . . . . . . . . . . . . . . . . . . . . . . . 21

3.3 Calibration gases used for their respective measured emissions. . . . . . . 23

4.1 Optimal spark timing for emissions at di↵erent fuel compositions. . . . . 34

A.1 Optimal spark timing for emissions at di↵erent fuel compositions. . . . . 73

vi
List of Figures

3.1 Schematic of engine setup used for testing. . . . . . . . . . . . . . . . . . 19

3.2 Generator Efficiency Curve. . . . . . . . . . . . . . . . . . . . . . . . . . 22

3.3 LabView Program Front Panel. . . . . . . . . . . . . . . . . . . . . . . . 24

3.4 ITB and Trim Valve wiring. Letter ”a” indicates a connection to the ITB

and letter ”b” to the Trim Valve. . . . . . . . . . . . . . . . . . . . . . . 26

3.5 The ’Overview’ tab displayed when monitoring the speed controller through

the L-Series Service Tool. . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3.6 The ’Overview’ tab displayed when monitoring the air-fuel ratio controller

through the L-Series Service Tool. . . . . . . . . . . . . . . . . . . . . . . 28

4.1 Comparison of engine-out emissions when the engine is fueled by NG vs.

CH4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

4.2 Comparison of emission reductions across a catalyst with a space velocity

of 24 457 hr-1 (1 catalyst) when the engine is fueled by NG vs. CH4 . . . . 33

4.3 Comparison of emission reductions across a catalyst with a space velocity

of 12 228 hr-1 (2 catalysts in series) when the engine is fueled by NG vs.

CH4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4.4 Reduction of THC emissions across the catalyst with a mean value of

1.013 and with oscillations about the mean as displayed on the x-axis.

Tests conducted with a constant load of 6 kW with NG as the fuel. . . . 35

vii
4.5 Reduction of NOx emissions across the catalyst with a mean value of

1.013 and with oscillations about the mean as displayed on the x-axis.

Tests conducted with a constant load of 6 kW with NG as the fuel. . . . 36

4.6 Reduction of CO emissions across the catalyst with a mean value of

1.013 and with oscillations about the mean as displayed on the x-axis.

Tests conducted with a constant load of 6 kW with NG as the fuel. . . . 36

4.7 Brake specific energy consumption at di↵erent fuel compositions with a 6

kW engine load. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

4.8 Brake specific energy consumption at di↵erent fuel compositions and at

two di↵erent load conditions. . . . . . . . . . . . . . . . . . . . . . . . . . 38

4.9 THC emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 40

4.10 NOx emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 41

4.11 CO emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 41

4.12 THC emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 42

4.13 NOx emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 43

4.14 CO emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions. . . . . . . . . . . . . . . . . 43

4.15 A comparison of engine-out brake specific THC emissions achieved with

the biogas setup, and previous testing running on gasoline. . . . . . . . . 45

4.16 A comparison of engine-out brake specific NOx emissions achieved with

the biogas setup, and previous testing running on gasoline. . . . . . . . . 45

viii
4.17 Engine-out brake specific CO emissions achieved with the biogas setup, at

varying fuel compositions. . . . . . . . . . . . . . . . . . . . . . . . . . . 46

4.18 A comparison of post-catalyst (space velocity = 12 228 hr-1 ) brake spe-

cific THC emissions achieved with the biogas setup, and previous testing

running on gasoline (engine-out). . . . . . . . . . . . . . . . . . . . . . . 47

4.19 A comparison of post-catalyst (space velocity = 12 228 hr-1 ) brake spe-

cific NOx emissions achieved with the biogas setup, and previous testing

running on gasoline (engine-out). . . . . . . . . . . . . . . . . . . . . . . 47

4.20 Post-catalyst (space velocity = 12 228 hr-1 ) brake specific CO emissions

achieved with the biogas setup, at varying fuel compositions. . . . . . . . 48

4.21 THC emissions reductions with di↵erent fuel compositions. Operating

with a mean of 1.013 and an engine load of 6 kW. . . . . . . . . . . . . 50

4.22 NOx emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 6 kW. . . . . . . . . . . . . . . 51

4.23 CO emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 6 kW. . . . . . . . . . . . . . . 51

4.24 THC emissions reductions with di↵erent fuel compositions. Operating

with a mean of 1.013 and an engine load of 8 kW. . . . . . . . . . . . . 52

4.25 NOx emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 8 kW. . . . . . . . . . . . . . . 53

4.26 CO emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 8 kW. . . . . . . . . . . . . . . 53

4.27 THC emissions reductions against NOx /THC ratio at catalyst entry. Op-

erating with a mean of 1.013 and at two di↵erent load conditions. . . . 56

4.28 Variation in the engine-out NOx /THC concentration ratio with changing

fuel composition. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56

4.29 Further analysis clarifies changes in the catalyst THC conversion efficiency. 57

ix
4.30 Increase in CO2 concentration in the exhaust gas stream as it flows through

the catalysts, confirming that the NO reaction with CO is occurring. . . 57

A.1 Spark timing sweep, showing the e↵ect that spark advance has on fuel

flow. Test engine fuelled by NG (7 kW load). . . . . . . . . . . . . . . . . 69

A.2 Spark timing sweep, showing the e↵ect that spark advance has on exhaust

temperature. Test engine fuelled by NG (7 kW load). . . . . . . . . . . . 70

A.3 Spark timing sweep, showing the e↵ect that spark advance has on THC

emissions. Test engine fuelled by NG (7 kW load). . . . . . . . . . . . . . 70

A.4 Spark timing sweep, showing the e↵ect that spark advance has on NOx

emissions. Test engine fuelled by NG (7 kW load). . . . . . . . . . . . . . 71

A.5 Spark timing sweep, showing the e↵ect that spark advance has on fuel

flow. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW load). 71

A.6 Spark timing sweep, showing the e↵ect that spark advance has on exhaust

temperature. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6

kW load). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

A.7 Spark timing sweep, showing the e↵ect that spark advance has on THC

emissions. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW

load). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

A.8 Spark timing sweep, showing the e↵ect that spark advance has on NOx

emissions. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6

kW load). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73

A.9 Reduction of THC emissions across the catalyst at di↵erent values.

Sweep conducted with a constant load of 7 kW and with NG as the fuel. 74

A.10 Reduction of NOx emissions across the catalyst at di↵erent values.

Sweep conducted with a constant load of 7 kW and with NG as the fuel. 75

A.11 Reduction of CO emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 7 kW and with NG as the fuel. . . . . 75

x
A.12 Reduction of THC emissions across the catalyst at di↵erent values.

Sweep conducted with a constant load of 6 kW and with a mixture con-

sisting of 60% NG and 40% CO2 as the fuel. . . . . . . . . . . . . . . . . 76

A.13 Reduction of NOx emissions across the catalyst at di↵erent values.

Sweep conducted with a constant load of 6 kW and with a mixture con-

sisting of 60% NG and 40% CO2 as the fuel. . . . . . . . . . . . . . . . . 76

A.14 Reduction of CO emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 6 kW and with a mixture consisting of

60% NG and 40% CO2 as the fuel. . . . . . . . . . . . . . . . . . . . . . 77

xi
Nomenclature

Excess air ratio

Fuel/air equivalence ratio

bmep Brake mean e↵ective pressure

bsec Brake specific energy consumption

BT C Before top centre

CAD Crank angle degrees

CN G Compressed natural gas

COVimep Coefficient of variation of indicated mean e↵ective pressure

DAQ Data acquisition system

EGR Exhaust gas recirculation

GHG Greenhouse gas

Gtoe Gigatonne of oil equivalent

HC Hydrocarbon

ICE Internal combustion engine

LF G Landfill gas

xii
LP G Liquefied petroleum gas

M BT Maximum brake torque

NG Natural gas

N Ox Oxides of Nitrogen

SI Spark ignition

SLM Standard litres per minute

T DC Top dead centre

T HC Total hydrocarbon

W EGO Wide range exhaust gas oxygen

xiii
Chapter 1

Introduction

Utilization of biogas as a fuel is very attractive from a greenhouse standpoint since biogas

is carbon neutral. There are various forms of biogas, including digester gas produced by

an anaerobic digester. Landfill gas is similar in composition. The common constituent

is methane, with varying amounts of carbon dioxide and nitrogen as well as a variety of

other gases in small amounts. Methane is the main source of fuel energy in biogas.

Biogas can be used to fuel either stationary engines or vehicles. Gas storage and

fueling technology developed for compressed natural gas (CNG) fueled vehicles would be

employed for vehicular uses of biogas. Despite the environmental attractiveness of biogas,

new challenges have arisen in the form of stringent engine exhaust emission standards for

both vehicles and most recently for stationary engines. Low emission engine technology

requires precise control of the air-fuel ratio, which in turn requires knowledge of the

fuel’s chemical composition. However, biogas composition varies substantially with time

(and location, in the application where vehicles refill in di↵erent locations). Thus, there

is a need for a gas composition sensor that can characterize the biogas fuel for the

engine control system. A gas composition sensor would also facilitate seamless multi-

fuel capability, i.e. the ability to run on natural gas or LPG or gasoline if biogas is not

available.

1
Chapter 1. Introduction 2

Low cost sensors that provide unambiguous gaseous fuel composition information

do not currently exist. This engine research is part of a larger project focused on the

development of a fiber optic based gas composition sensor. The larger project developed

a laboratory prototype sensor system using an absorption cell with multipoint sensing

on a single fiber. Absorption sensors measure the spectral absorption features of gas

species methane (CH4 ) and carbon dioxide (CO2 ), the main components of biogas from

an anaerobic digester or a landfill site. Reliable gas concentrations for both CH4 and

CO2 can be obtained for gas mixtures in the 30 to 60% CH4 concentration range with a

maximum error of 4%. Further work is needed to achieve the same accuracy at higher

CH4 concentrations.

The focus of this engine study was on the e↵ect of biogas, especially with high CO2

concentrations, on the operation of a standard 3 way catalytic converter. Methane is

particularly difficult to oxidize and specialized catalysts have been developed for natural

gas. Use of biogas as a fuel changes both the chemical composition of the exhaust gas

and the temperature of the exhaust gas, both of which have an impact on catalytic

converter operation. The overall objective of the engine study was to evaluate the e↵ect

of biogas on the operation of a standard 3-way catalytic converter. Specific questions to

be answered include:

• Does the fuel composition a↵ect the engine spark timing required for optimal cat-

alyst operation?

• What is the optimal air-fuel ratio to maximize conversion and does this optimal

air-fuel ratio change with fuel composition?

• Does the fuel composition have an impact on the catalyst size requirement?

• What are the minimum emission levels achievable?

• Does the fuel composition have an impact on engine efficiency, as characterized by

brake specific energy consumption?

Chapter 2

Background and Literature Review

2.1 Current Energy Needs

Worldwide energy consumption has had an increasing trend over the past several decades,

and was recently estimated to be at 11.8 Gtoe (2009), 23% greater than what it was only

a decade earlier in 1999 [1]. Fossil fuels have been primarily used to meet this demand, as

80.9% of the total energy consumption was comprised of fossil fuel consumption in 2009

(varying by only 1% since 1990). If this trend of increasing energy demand continues, as

is expected, then a rise in fossil fuel consumption would also result [33]. However, with a

new worldwide commitment (proceeding from the 2011 United Nations Climate Change

Conference) to reducing greenhouse gas emissions that includes the major fossil fuel

consuming nations of China, India (both of which did not have to cut emissions under

the first round of the Kyoto protocol) and the U.S. (which did not ratify the Kyoto

protocol), it is clear that, if the goals outlined are to be achieved, fossil fuel consumption

must be reduced. Alternative energy sources will have to provide clean energy and will

therefore play a large role in the future.

3
Chapter 2. Background and Literature Review 4

2.2 Biogas As a Promising Alternative

Biogas is an interesting alternative fuel, which can be used as a replacement of fossil fuels

in heat and power production, or as an engine fuel [33]. It is produced through anaerobic

digestion, which is an after-treatment of biomass that is applied in landfill sites and

also used on farms. Biogas provides the benefit of displacing the need for fossil fuels,

along with the additional advantage of not allowing methane (the gas with the highest

content in the mixture) [8], which is a greenhouse gas (GHG), to be released directly to

the atmosphere as it would through ordinary biological decay. By using this fuel in an

internal combustion engine (ICE), the methane content is combusted, releasing CO2 to

the atmosphere instead, which has a global warming potential that is 21 times less than

that of methane [12][28].

Several other benefits also exist [20], which have been quantified by means of economic

analyses. The economic analysis conducted by White et al. [34] showed that for small-

scale Ontario dairy and beef farms (with digesters capable of producing below 250 kW),

savings of CAN$7.70 and CAN$5.32 per animal per year, respectively, can be realized.

When taking into account the Ontario Feed-in-Tari↵ (FIT) program (which guarantees to

pay approximately 5 times the wholesale market price of electricity for 20 years to farms

that generate electricity by means of an anaerobic digester system), it was ultimately

determined that approximately 4000 dairy farms in Ontario have favourable profitability

(ROI greater than 10% in a payback time of less than 10 years) and that 5000 beef farms

in the province do also. If the Ontario FIT program were to be extended Canada wide,

these numbers become 20,000 and 30,000, respectively. Yiridoe et al. [38] conducted a

more comprehensive analysis which took into account other benefits of on-farm biogas

energy production such as odour reduction, toxicity and pathogen reduction, reductions

in weed seed germination, the organic manure produced as a by-product of the spent

digestate, heat energy production and GHG reduction. They showed that with a medium

level of governmental support (30% cost sharing), the digester/ICE power systems of all
Chapter 2. Background and Literature Review 5

swine farms in a large size range (with between 50 and 800 sows, inclusive) would have a

net present value greater than zero, and likewise for dairy farms with 100-800 cows. It is

worth noting that these papers focused on the economic feasibility of small-scale biogas

plants only, since large-scale plants have already been shown to be profitable [34][38].

Since such small-scale installations have already been shown to be economically viable,

improvements, such as in the e↵ectiveness of engines fueled by the biogas and used to

power generators to produce electricity, would only make them more so.

2.3 Biogas Composition

Weiland [33] has shown that biogas production through anaerobic digestion has been

steadily increasing at a rate of 20% per year, with 6 Mtoe of energy having been produced

in 2007. An issue with the end use of the gas however, is that the gas content varies

significantly based on the feedstock, the process technology used and the retention time

of the waste in the digester. In the agriculture sector, biogas plants digest manure from

pigs, cows, and chickens along with additional co-substrates (which increase the content of

organic material to achieve higher gas yields). These co-substrates usually include harvest

residues, such as organic wastes from agriculture-related industries, collected municipal

bio- waste from households, energy crops, etc. The e↵ect of di↵erent feedstocks is shown

in Table 2.1, which displays the theoretical biogas yields per ton of total solids (TS) and

the methane content for di↵erent biomass substrates [33].

Research is currently being undertaken on the e↵ects that di↵erent digester processes

have on gas yields and compositions. Processes can be classified into two types: wet and

dry, which operate with a total solids concentration in the fermenter of less than 10% and

between 15-30%, respectively. Wet fermentation processes are the most widely used in

the agriculture sector. In a study of 61 plant farms, Weiland [33] showed that two-stage

digester systems, consisting of a high-loaded main fermenter and a low-loaded secondary

Chapter 2. Background and Literature Review 6

Table 2.1: Theoretical gas yields and methane content [33].

Substrate Biogas (Nm3 /t TS) CH4 (%) CO2 (%)

Carbohydrates 790-800 50 50

Raw protein 700 70-71 29-30

Raw fat 1200-1250 67-68 32-33

fermenter in series, resulted in higher gas yields and a reduced residual methane potential

of the digestate. However, due to the complexity and the availability of parameters that

can be measured online, these processes are difficult to control. Research is currently

underway as to what indicators would be the most efficient and e↵ective for control,

and on how they may be obtained [33]. As a result of less than optimal control, the

composition of biogas varies even if it is produced using the same process and at the

same location. When taking into account all of the di↵erent processes that can be used

and the variability in feedstock, the composition of biogas varies significantly. The range

of biogas compositions typically found is shown in Table 2.2, along with the composition

of natural gas for comparison.

Table 2.2: Fuel composition (in volume %) of di↵erent gaseous fuels [30].

CH4 CO2 N2 H2 O2 H2 S NH3

Biogas 55-65 35-45 0-3 0-1 0-2 0-1 0-1

Landfill Gas 45-60 40-60 2-5 0-0.2 0.1-1 0-1 0.1-1

Natural Gas 70-90 0-8 0-5 0 0-0.2 0-5 0

A higher CH4 content is preferred, as it gives the gas a higher energy content (since

it and hydrogen are the flammable gases), whereas carbon dioxide and nitrogen are inert

and therefore have no energy content. Hydrogen sulfide and ammonia, which are toxic
Chapter 2. Background and Literature Review 7

and require treatment to be removed, are present only as minor impurities [24]. An

investigation on the e↵ect of the methane concentration in biogas when used as a fuel

for a spark ignition engine by Porpatham et al. [25], showed that as CO2 levels were

reduced, the brake power output and thermal efficiency increased for the entire range of

equivalence ratios tested (0.6 to 1.2). E↵ects of these composition changes on emissions

were also researched. It was seen that as the concentration of CO2 decreased, there was

a drop in hydrocarbon emissions, because an increase in combustion temperature leads

to better oxidation. An inverse relationship was found to exist between NO emissions

and the CO2 concentration, for the same reason. Near stoichiometric equivalence ratios,

CO and CO2 emissions behaved similarly. The researchers determined that a reduction

of CO2 concentration to 30% of the fuel by volume was optimal for hydrocarbon and

NO emissions reduction, though at the expense of brake torque and thermal efficiency

[25]. Since biogas is often used as the fuel in engines designed to be fuelled by natural

gas, research has been conducted in order to alter biogas combustion properties to more

closely resemble those of natural gas, either at the production stage, or by using additives.

Chandra et al. [7] compared the evaluated parameters of power developed, brake specific

gas consumption and brake thermal efficiency for an engine running on three di↵erent

fuels: CNG, methane enriched biogas (95% CH4 , 3% CO2 , 2% other gases) and raw biogas

(65%CH4 , 32% CO2 , 2% other gases). They reported that methane-enriched biogas

was able to produce 1.43 times as much power as raw biogas under similar operating

conditions. Brake specific gas consumption (BSGC) values were also measured, and it

was found that the BSGC of raw biogas was relatively high as compared to the other

two fuels, which had comparable BSGCs. E↵ects on the thermal efficiency were also

investigated, and for the methane enriched biogas, the maximum brake thermal efficiency

was 26.2%, at a brake load of 59%, whereas for the raw biogas, it was 23.3%, at a brake

load of 53.5% [7], which is in agreement with the findings of Porpatham et al. [25]. For

this study, these brake loads correspond to the respective maximum brake power outputs
Chapter 2. Background and Literature Review 8

for the di↵erent gaseous fuels.

It should be noted that several factors significantly a↵ect the performance and emis-

sions of an engine, which must be taken into account when using biogas as the fuel for an

engine not designed to run on it. Compression ratio and spark timing are usually altered

and have been for many of the engines used in this area of research, and the altered

values can significantly a↵ect the data. One example is in the research by Bade Shrestha

and Narayanan [28], where it was determined that for the same natural gas engine with a

compression ratio of 8.5 running on pure methane, comparable performance was achieved

by running on landfill gas with a compression ratio of 12 (the engine used had variable

compression ratios). Although the same engine could not be used at a compression ratio

greater than 10 when fueled by methane due to knock issues, it was able to run at a

higher compression ratio with landfill gas because the diluents (CO2 and N2 ) in the gas

act as heat sinks and knock suppressant agents. Along with increasing the compression

ratio on these engines to achieve similar performance when landfill gas is used as the fuel,

spark advance also had to be increased from 12 to 25 BTC [28].

Nges et al. [23] did a study on the feasibility of supplementing an industrial waste

anaerobic digester feedstock, consisting of wastes such as pig manure, slaughterhouse

waste, food processing and poultry waste, with carbohydrate-rich energy crops. They

found that the methane yield per ton of feedstock improved by 27% by adding energy

crops such that the maximum total solids content, 12% for the plant studied, of a con-

tinuous stirred tank reactor was reached.

The e↵ect of hydrogen addition to engine fuels, including biogas, has been the subject

of several recent studies [24][3][6][14][16][18][39]. Some of hydrogens advantages as a fuel

include its excellent combustion properties (a laminar flame speed of 1.6 m/s, compared

to 0.38 m/s for CH4 ), the absence of carbon emissions since the only product from its

combustion with O2 is water, it is highly ignitable, and has a high heating value on a

mass basis. Some disadvantages are that the storage of hydrogen is problematic, because
Chapter 2. Background and Literature Review 9

its heating value on a volume basis is very low. Therefore, instead of using it alone as a

fuel, hydrogen is being considered as a fuel additive to enhance combustion [28].

2.4 Previous Work

2.4.1 Stoichiometric and Lean-Burn Hydrogen Addition Exper-

iments

Bade Shrestha and Narayanan [28], simulating landfill gas with a test fuel composition

of 53% CH4 , 42% CO2 and 5% N2 , studied the e↵ect that hydrogen addition has on com-

bustion. Using hydrogen as an additive in landfill gas, it was found that H2 increased the

thermal efficiency (only with a H2 /(LFG+H2 ) ratio up to 5%, above which it decreased),

increased the peak cylinder pressure significantly in all operating conditions, significantly

reduced the coefficient of variation of peak cylinder pressure, and significantly reduced

ignition lag and combustion duration. There were particularly significant improvements

in thermal efficiency and power at lean conditions. It was also reported that the coeffi-

cient of variance for the indicated power was very high when the equivalence ratio was

greater than unity and low otherwise, suggesting that operating at stoichiometry would

require very tight control on the air-fuel ratio in order for results to be consistent and

repeatable [28].

Bysveen’s [6] research used a mix of CNG and hydrogen with a much higher hydrogen

content (29-vol %) and with excess air ratios ( ) ranging from 1.1 to 1.8. She showed that

the amount of unburned hydrocarbons increased with increasing , and, at a given , was

lower when running on the mix with hydrogen as opposed to just on CNG. NOx emissions

were found to be higher for the mix for the same , because with H2 , combustion speeds

and temperatures increase. However, since running lean generally reduces NOx emissions

significantly, and the addition of H2 extends the lean burn limit, the counter e↵ect gener-
Chapter 2. Background and Literature Review 10

ally keeps the levels quite low. Also, the brake thermal efficiency was considerably higher

when hydrogen was added for the same given power (31% compared to 27.5% at 24 kW),

and the brake thermal efficiency advantage hydrogen provided increased at higher engine

speeds and higher excess air ratios. This was attributed to the fact that H2 addition

allows for higher flame speeds and therefore shorter combustion duration, and less cyclic

variability, which are issues that exist with lean-burn engines fueled by NG [6].

Kahraman et al. [16] investigated the e↵ect that adding hydrogen to methane has

on engine performance, cylinder pressures, brake thermal efficiencies and emissions. The

experiments were performed at wide-open throttle, at several engine speeds and varying

of 1, 1.05, 1.1, 1.15 and 1.2. CH4 /H2 volume fraction ratios of 100/0, 90/10, 80/20 and

70/30 were used. They reported that peak cylinder pressures increased as the volume

fraction of H2 increased for the entire range of that was tested. It was also observed

that the e↵ect that has on peak pressure timing relative to top dead center (TDC) is

dependent on the fuel blend used, although the relationship is not necessarily linearly

dependent on the H2 volume fraction. For a constant , the relationship between H2 con-

tent and peak pressure timing showed that as the volume fraction of H2 is increased, less

spark advance is required for peak pressure to occur at the same location, as the engine

speed is increased up until about 2000 rpm, after which changes become insignificant.

Carbon monoxide emissions were shown to decrease with an increasing , and generally

settled to negligible levels as the was increased past a certain limit. At higher engine

speeds, the CO emissions settle to these negligible levels at lower values. Another

trend observed was that as the H2 fraction in the fuel increased, CO emissions decreased,

though it should be noted that the research of Zhunqing and Zhang [39] concluded that

CO emissions increase as the hydrogen fraction is increased. Carbon dioxide emissions in-

crease with increasing up until values of approximately 1.05-1.11. Ratios higher than

this lead to a decrease in CO2 emissions. As the H2 fraction increases, CO2 emissions de-

crease. Across the entire range of tested, the e↵ect of increasing H2 by a certain amount
Chapter 2. Background and Literature Review 11

on CO2 emissions was fairly consistent, except for values approaching 1.35, where the

e↵ect of adding H2 decreases. As the hydrogen fraction increases, hydrocarbon emissions

decrease. The lowest emission values for all the di↵erent fuel blends are observed when

is between 1.1 and 1.2, and when the value is above or below this range, the HC emissions

increase. This is due to the fact that as the excess air ratio is increased, generally HC

emissions are reduced, but after a certain increase, the flame propagation slows enough

that there is insufficient time to burn all of the fuel in the combustion chamber, resulting

in a net increase in HC emissions. As the hydrogen fraction is increased, flame speeds

increase, and as such, fuels with higher hydrogen content were found to have decreasing

HC emissions over a wider range of . The brake thermal efficiency was also found to

increase when the fuel used had a higher hydrogen content. Also, with increasing ,

⌘ th increases, though its e↵ect varies based on the engine speed, which also exhibited a

proportional influence on ⌘ th [16].

Since prior research provided mixed results, Park et al. [24] aimed to more clearly

demonstrate the e↵ect of H2 addition on the thermal efficiency of a biogas engine. A mix

of only natural gas and N2 was used to simulate biogas, as the researchers stated that

the overall trends of N2 dilution are similar to those of CO2 dilution. Using di↵erent

gas compositions, ranging from 0% to 80% N2 as fuel and at an equivalence ratio of

unity the thermal efficiency at each composition was calculated. The maximum thermal

efficiency observed was 31.1% at 80% N2 . It was also found that as the inert gas (N2 )

concentration was increased, the MBT spark timing was advanced. However, contrary

to the result that this would have on NOx and THC emissions when using higher density

fuels, NOx emissions significantly decreased with rising N2 concentrations, while THC

emissions increased. The e↵ect of H2 addition was also investigated, ranging from 5%

to 30% addition. Using biogas with the lowest heating value that could be reasonably

expected (80% N2 ), and running at stoichiometric conditions (since combustion stability

would not allow for lean operation at low hydrogen concentrations) they observed a
Chapter 2. Background and Literature Review 12

maximum thermal efficiency of 32%, at 5% hydrogen addition, and a MBT timing of 16

BTC. At H2 concentrations greater than 5%, cooling losses became so high that thermal

efficiency decreased, as opposed to the proportional relationship that exists between the

concentration of H2 and thermal efficiency in the 0% to 5% range. The e↵ect of hydrogen

addition at lean operating conditions was also studied, and the peak thermal efficiency

for this test was found at 30% excess air and 10% hydrogen addition [24].

2.4.2 Hydrogen Addition with EGR Systems

Though hydrogen addition allows for more favourable combustion characteristics in bio-

gas, the e↵ects it has on reducing emissions are not significant, and for example, in the

case of NOx emissions, hydrogen addition actually increases emissions. Therefore, other

means of reducing emissions are considered, one of which is the use of an exhaust gas

recirculation (EGR) system.

Research on the e↵ects that di↵erent volume fractions and exhaust gas recirculation

rates would have on the heat release rate, flame development duration, rapid combustion

duration and the coefficient of variance of the indicated mean e↵ective pressure, has also

been done. Hu et al. [14] used natural gas-hydrogen fuel blends in which the volumetric

fractions of H2 were 0%, 10%, 20%, 30% and 40% along with varying EGR rates. They

reported that as the EGR rate increases, the optimum ignition timing occurs at a greater

advance for all compositions, and that the opposite is observed with increasing H2 con-

centrations. Also, as the EGR rate increases, the beginning of heat release is postponed

(increasing the combustion duration) and the mass fraction burned increases at a slower

rate due to the decrease of propagating flame speed. With an increase in the H2 concen-

tration, the flame duration decreases, since the early stage of flame development is highly

related to ignition delay, which depends on the mixture concentration and temperature.

Results showed that hydrogen addition has a greater e↵ect on the flame development

duration as compared to its e↵ect on the rapid combustion portion of the combustion.
Chapter 2. Background and Literature Review 13

The coefficient of variation of the indicated mean e↵ective pressure (CoVimep ) was also

analyzed, and was observed to be highly dependent on engine speed. When running at

2000 rpm, it was determined that a very slight negligible di↵erence exists when using

an EGR rate between 0-10%. However, at an EGR rate greater than this, a significant

spike was seen. When using 15% EGR, the CoVimep jumped from 4% to 32% when using

a blend with 0% H2 , but up to 5% from 2% when using the 40% H2 blend. The e↵ect

was greater at 20% EGR for the higher hydrogen fraction fuel blend, as a jump from 5%

to about 42% in the CoVimep was observed. When running at 3000 rpm, the change in

the EGR rate from 0-25% had a negligible e↵ect on the CoVimep , though a very slight

increasing trend was seen. When using 30% EGR, the CoVimep increased to 22% for the

0% H2 fuel blend, but no significant change was seen among the other blends at this

EGR rate. At 35% EGR, a significant increase was seen for the 0%, 10% and 20% H2

blends, and at 40% EGR, increases of 12% and 9% in the CoVimep were observed for the

30% and 40% H2 blends, respectively [14].

The e↵ect on generating efficiency and emissions of hydrogen addition along with the

use of an EGR system were studied by Lee et al. [18]. Their research investigated the

relationships between generating efficiency, spark timing, the EGR rate, and emissions

(NOx and CO2 ) in an engine fueled by a biogas-hydrogen blend. These relationships were

investigated alongside those for natural gas and a simulated biogas without a hydrogen

additive as well. The composition of gas used to simulate biogas consisted of 60% CH4

and 40% CO2 . For the experiments with a hydrogen additive, 5% of the total heating

value of the fuel was added (approximately 9% of the total gas flow volume), as their

studies showed this to be the most e↵ective composition for improving the generating

efficiency. Previous studies also showed that using a value of 1.2 was the most e↵ective

for improving this efficiency, and so all their experiments were run at such conditions.

They reported that MBT spark timing is more retarded with hydrogen addition (from

16 to 13 BTC). However, as EGR rates increased to 15%, spark timing advanced, and
Chapter 2. Background and Literature Review 14

EGR rates higher than this did not a↵ect the spark timing. Efficiency was also shown to

decrease from 30.15% to 29.02% as the EGR rate was increased from 0% to 15%. When

studying the emissions, they showed that the biogas with the H2 additive generally had

higher NOx concentrations than the plain biogas for the same EGR rate, though it was

found that this di↵erence decreased significantly at an EGR rate greater than or equal to

15%. CO2 emissions in the H2 addition test were also found to be lower than those of the

plain biogas. With an EGR rate of 15%, and optimal spark timing, efficiency increased

from 29.5% to 31% when hydrogen was added [18].

In a separate paper, Hu and Huang [13] researched the optimization of the EGR ratio

for a spark ignition engine fueled by natural gas-hydrogen blends. They showed that at

lower EGR rates, the thermal efficiency increases with increasing EGR primarily due to

the fact that combustion is improved for the blend because the intake charge temperature

is increased with increasing EGR rates (since the recirculated gas was cooled by a fixed

amount of cooling water). At higher EGR rates, the decrease in the e↵ective thermal

efficiency as the EGR rate was increased was due to the possible occurrence of bulk

mixture quenching and partial burning. Flame development duration increased with an

increase in the EGR rate, since the dilution of EGR increases the ignition delay and

slows down the early stage of flame development. With advanced timing, the duration

increased, because flame development is defined as the crank angle interval between the

start of ignition and the point at which 10% of the mass fraction is burned. Since the

in-cylinder temperatures are lower when the spark timing is advanced, flame speed is

decreased, thus increasing the flame development duration. Rapid combustion duration

was also investigated, and the total duration was found to increase with an increase of the

EGR rate and to decrease with the advance of spark timing. The concentration of NOx

was found to decrease with increased EGR rate, which is due to mixture dilution of large

specific heat capacity gases (CO2 and H2 O), which absorb the released heat, reducing

the combustion temperature. Also, an increase in the EGR rate means a decrease in
Chapter 2. Background and Literature Review 15

fresh air and fuel, which decreases the amount of heat released and the combustion

temperature. NOx emissions increased with the advance of spark timing, due to the

higher in-cylinder pressures and temperatures that it causes. Hydrocarbon emissions

were shown to increase with an increase in the EGR rate, due to mixture bulk quenching

at low flame velocities and large cycle-by-cycle variations. Also, by decreasing the spark

advance HC emissions decreased, as emissions temperatures rise and the combustion of

HC emissions in the expansion stroke is enhanced. They reported only slight variations

were found in CO and CO2 emissions with changes in the EGR rate and spark timing, and

emissions were relatively constant, with about 0.5% and 10% concentrations observed for

each, respectively. Though the relation between the e↵ective thermal efficiency and NOx

emissions (when using di↵erent EGR rates) provided scattered results, it was determined

that the best results for low NOx emissions and high efficiency were when the EGR rate

was set at 15%, and MBT timing was used [13].

Along with these developments, come advances in catalyst technology as well. Kwak

et al. [17] report on the NOx storage behavior of Pt/BaO/CeO2 ceria-based catalysts in

the presence of SO2 . They found that the catalyst has a high sulfur tolerance (maintaining

the time of complete NOx uptake even upon being exposed to approximately 3 g/L of

SO2 ).

2.5 Summary and Conclusion

Meeting stringent emissions standards yet realizing as much of a power output as possible

is the goal with internal combustion engines. However, as attested to in the literature,

there is often a trade-o↵ between the two. Lean operation has been a method used to

reduce emissions. Consistency exists in the findings of Bysveen et al. [6] and Kahraman et

al. [16], which have shown that generally as increases, unburned HC emissions decrease,

except for when operating at high excess air ratios. Furthermore, reductions in CO
Chapter 2. Background and Literature Review 16

emissions [16] and NOx emissions [6] display the motivation behind moving towards lean

operation. However, additional unfavourable results of increasing include significant

reductions in the brake thermal efficiency and power output of an engine [6].

To alleviate these adverse e↵ects of running lean, hydrogen addition has been pro-

posed, with a consensus in the literature that has shown experimentally that the thermal

efficiency and power output of an engine can be increased significantly when hydrogen

is added to the fuel mix [28][24][6][16]. It is interesting to note that the findings of

Bade et al. [28] and Park et al. [24] agree that the maximum thermal efficiency that

can be achieved occurs with a 5% hydrogen addition. It is also interesting to note that

research by the latter was conducted when running stoichiometric, meaning that the

additional favourable characteristics that hydrogen addition provides could be coupled

with stoichiometric operating conditions (to maximize efficiency and power output).

Hydrogen addition provides these advantages along with decreased CO [16], CO2 [16]

[18] and unburned HC [16] emissions, although it also increases NOx emissions [6][18].

To counteract the increase in NOx emissions, EGR systems can be put into e↵ect, which

cause significant decreases in NOx emissions [18][13]. A very e↵ective means of emissions

control, used widely today in cars to meet emissions standards is the three-way catalyst.

With the findings of Rasi et al. [26], which have shown that sulfur can be e↵ectively

removed from biogas, along with those of Kwak et al. [17] on catalysts with high sulfur

tolerances, the conclusions of Roubaud et al. [27] may not apply, and three-way catalysts

could possibly be an e↵ective means of reducing emissions.

The major challenge in being able to do so is due to the operating principles of three-

way catalysts, which require stoichiometric conditions. With the varying composition

of biogas, this is difficult. A solution would be to have a fuel composition sensor that

can determine within a reasonable amount of time, the composition of the fuel. Such a

sensing system would need to be capable of determining the portion of the fuel that is

made up of CH4 , CO2 , N2 and H2 , and thus capable of sensing these gases within the
Chapter 2. Background and Literature Review 17

ranges specified in Table 2.2 (though it should be capable of sensing H2 from the range

of 0% to 5% of the overall fuel composition, since 5% hydrogen addition gives favourable

performance). Ideally it would also be compact and inexpensive, making it suitable for

automotive applications.
Chapter 3

Test Equipment and Setup

3.1 Overview

The equipment used for measurement and testing is described in this chapter. A schematic

of the equipment setup used for testing is shown in Figure 3.1. Several components are

primarily unchanged from the setup detailed by Tadrous [31]. For the convenience of the

reader, a brief review is provided. Several of the components, including the engine, the

system by which the engine is loaded, the measurement apparatus, the speed control and

air-fuel ratio control, will be discussed in more detail.

18
 Generator
Voltage and
Solenoid Valve Ball Valve Flow Controller Current

NG/
CH4
Service Tool Exhaust
Generator WEGO
Data Temperature
Sensor
and Pressure Exhaust

ITB Catalyst Catalyst

Engine

Manual Valve Flow Controller Trim Valve

Air Emissions
Temperature Analyzers
CO2 and Pressure Encoder

Air Plenum
Chapter 3. Test Equipment and Setup

Spark Exhaust

Air Flow

IMAP
Intake Air and
EMAP

Flame Arrester Solenoid Valve Manual Valve Flow Controller

H2

Figure 3.1: Schematic of engine setup used for testing.

THPS 400

DAQ
19
Chapter 3. Test Equipment and Setup 20

3.2 Engine Specifications

For this research, a naturally aspirated 3-cylinder Kubota DF-750 four-stroke SI engine

was used. The engine specifications are provided in Table 3.1. Although this engine

was designed to run on either gasoline or LPG, only minor modifications were made for

it to run on biogas and natural gas. An MSD digital programmable 6 AL-2 ignition

system was used to control spark timing, as the optimal timing varies with varying fuel

composition, and the engine originally had fixed ignition timing.

Table 3.1: Kubota DF-750 specifications.

Type Vertical

Number of Cylinders 3

Bore ⇥ Stroke 68 ⇥ 68 mm

Total Displacement 740 cc

Brake Power SAE Net (Continuous) 13.4 kW/3600 rpm

Brake Power SAE Net (Intermittent) 16.4 kW/3600 rpm

Compression Ratio 9.0:1

3.3 Catalyst Formulation

Tadrous [31] showed that a catalyst formulation as that detailed in Table 3.2 can ad-

equately reduce THC, NOx and CO emissions, and so the same one was used for all

catalyst tests. The space velocity (defined by Equation 4.2) of one catalyst is 24 457 hr-1 .

To show the e↵ect of space velocity on emissions, a second catalyst was placed in series

with the first (making the e↵ective space velocity equal to 12 228 hr-1 ), and emissions

were measured before the catalysts, after 1 catalyst, and after 2 catalysts.
Chapter 3. Test Equipment and Setup 21

Table 3.2: Catalyst specifications [31].

Part No. PGM Loading PGM Loading Loading Ratio Wash-Coat Feature

(g/ft3 ) (g/litre) (Pt/Pd/Rh)

Z60-0290 20 0.706 5/0/1 ZX ceria

3.4 Stamford Electrical Generator

A Stamford model BCI-162D electrical generator, directly coupled to the engine, and

a resistive load bank were used to apply a load to the engine. It is a 2-pole generator

configured to deliver 240V AC at 60 Hz. Ten 240 V AC electric heaters, which could be

switched on and o↵ in di↵erent combinations by relays, were used to control the load on

the engine. A more comprehensive description of this setup has been written by Foster

[9]. The generator efficiency curve is shown in Figure 3.2.

3.5 Measurement Apparatus

As seen in Figure 3.1, several measurements were made to aid with analysis. This in-

cludes several thermocouples (K-type thermocouples were used) to measure the intake

air temperature, exhaust temperature (both pre- and post-catalyst). For safety, and to

ensure that the engine was operating within specifications, coolant temperature was also

measured, though not recorded. Also, absolute pressure transducers were used to mea-

sure the intake air pressure along with pressure in the intake and exhaust manifolds. A

wide range exhaust gas oxygen (WEGO) sensor was used to measure exhaust oxygen

for air-fuel ratio control (discussed further in section 3.6). To measure the electrical

power generated, the voltage and current of the generator were measured. A transformer

reduced voltage such that it could be measured by the data acquisition system (DAQ)

and a current transducer converted the current to a linear voltage output. In order to
Chapter 3. Test Equipment and Setup 22

75

70

65

60
Generator Efficiency (%)

55

50

45 y=−0.00007279x4+0.00243x3−0.03196x2+0.192+0.2631

40

35

30

25
0 1 2 3 4 5 6 7 8 9 10
Electrical Power Output (kW)

Figure 3.2: Generator Efficiency Curve.

[31]
Chapter 3. Test Equipment and Setup 23

simulate biogas, flow controllers were used to mix di↵erent pure gases, and the flows of

each gas, measured in standard litres per minute (SLM), were measured and monitored

to confirm that the biogas composition conformed with that required for the testing in

progress.

Table 3.3: Calibration gases used for their respective measured emissions.

Emission Range Calibration Gas

THC 0-5000 ppm 2000 ppm methane, N2 balance

NOx 0-100 ppm 90 ppm NOx , N2 balance

0-1000 ppm 980 ppm NOx , N2 balance

0-10 000 ppm 4500 ppm NOx , N2 balance

CO 0-20 000 ppm 15 000 ppm CO, N2 balance

CO2 0-14% 13.5% CO2 , N2 balance

O2 0-1.1% 0.9% O2 , N2 balance

Emissions measurements were made using CAI instruments, including a heated flame

ionization detector (HFID) to measure THC emissions, a heated chemiluminescent de-

tector (HCLD) for NOx emissions and non-dispersive infrared detectors (NDIRs) for CO,

CO2 and O2 emissions. A heated line was used to maintain sample temperature at

191 C to prevent condensation as it flowed to the emissions rack, where the sample is

split between the HFID, HCLD and a chilling system to remove water. A flow manifold

downstream of the chilling system further split o↵ the sample into the analyzers requiring

dry samples, that is, the CO, CO2 and O2 analyzers. Table 3.3 shows the calibration

gases used for each measurement.

LabView was used to interface with the DAQ and the measured values. The front

panel of the program used is shown in Figure 3.3.

Chapter 3. Test Equipment and Setup 24

Throttle Position (%) Throttle CH4 CO2 H2

0.00 4.5 5 5.5
Stop 40 60
Current Amps 4 6 100
3.5 6.5
20 80 0.00 3 7
90
0 100 2.5 7.5
Electrical Power(KW) Fuel Flow
Recording Intake Air Pressure (kPa)
2 8 80
Trim Valve Position Trim 0.00 1.5 8.5
0.0 Elapsed (s) 0
0.00 1 Power kW 9 70
40 60 rms value Voltage VAC
60 Target (s) In. Man. Press (kPa) 0.5 9.5
20 80 0.00 0.00 0 0 60
Output File Path 0 10
0 100
50

Air Flow CH4 CO2 H2

60 40
50 70 1450 115 115 20

40 Manifold 80 30
18
2500 Phi to Trim
Pressure 100 100
60 30 90 1200 2000 3000
16 0.000 20
40 80 90 90

20 100
Intake Air Temp 20 100
14
1500 RPM 3500
80 80 10
1000 1250 3750
0 120 11
10 110 0
70 70 12 1000 4000
120 0
0 800 750 4250
kPa 60 07:00:00 PM 07:00:10 PM 07:00:20 PM 07:00:30 PM 07:00:45 PM
60 10 4500
250 0.8019
Before After 50 50
600 8 0 5000
800 40 0.8018
40
6 Engine must be turning,
600 400 30 30 or errors will occur. 0.8017
Exhaust Temp 4
400 20 20
0.8016
200
200 10 2
10
Inc 2 Dec H2 % Req'd H2 Flow Required 0.8015
0 0 0 0 0 0 0
19:00:00 19:01:16 0.8014
0.00 0.000 0.000 0.000 CO2 % Req'd CO2 Flow Required
Inc 3 Dec
0 0 0.8013

Phi Offset
Exhaust THC 0 ppm 0.8012
% Off Tolerance

0.000 1000.000 2000.000 3000.000 4000.000 5000.000 0.3 0.8011

2500
2750 Offset
0 ppm 2000 0.8010
Exhaust NO X
1750 RPM_anal 3250
0.000
0.000 1000.000 2000.000 3000.000 4000.000 5000.000 1500 3500 0.8009
0.00 0 10 20 30 40 50 60 70 80 90 100 107
1250 3750
Exhaust CO 2 0.00 % 1
1000 4000

0.000 2.500 5.000 7.500 10.000 12.500 15.000 17.500 20.000 750 4250 0.5
500 4500
Exhaust O 2 0.00 % 250 0
0 5000
0.000 5.000 10.000 15.000 20.000 25.000
-0.5
Exhaust CO - High Range 0.00 ppm RPM_analog
0
Reset Overspeed Overspeed? -1
0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500 5.000
OFF 0 5 10 15 20 25 30 35 40 45 51

Exhaust CO - Low Range 0.00 ppm

0.000 0.500 1.000 1.500 2.000 2.500 3.000 3.500 4.000 4.500 5.000

Figure 3.3: LabView Program Front Panel.

Chapter 3. Test Equipment and Setup 25

3.6 Throttle and Air-Fuel Ratio Control

Originally, the engine’s stock configuration used a carburetor to control gas mixing, and

a governor to activate the throttle in response to engine speed. The stock carburetor

(designed for LPG fuel) was replaced with a venturi-type natural gas carburetor to pro-

vide better mixing characteristics, and the governor was replaced with a servo motor to

control the throttle. However, these changes still left much to be desired with regards

to stability. Since one of the goals outlined from the onset of this project was to look at

emissions and performance parameters when running at constant power, with a constant

load applied (as the system had already been set up), it was deduced that the desired

stability could be realized if the engine speed could be well controlled. Cyclic variations

would also be reduced if the air-fuel ratio could be better controlled.

3.6.1 Speed Control

In order to control engine speed, a Woodward L-series integrated throttle body (ITB)

controller was installed. The ITB is composed of a venturi mixer (to mix fuel with air)

and an electric actuator. The actuator accepts a speed signal from an optical encoder

(the signal provided to the actuator gives 36 pulses per revolution), compares it to the

configured speed setpoint, and drives a 0-60 output throttle shaft to the commanded

position based on an internal shaft position sensor [35]. Software to configure the ITB

was available on the Woodward website (L-Series Service Tool Version 2.6 was used).

Wiring for both the ITB and Trim Valve (air-fuel ratio controller) is shown in Figure 3.4.

Also, Figures 3.5 and 3.6 show the respective ’Overview’ tabs for the L-Series Service

Tool when connected to the ITB and Trim Valve, respectively.

Both the ITB and Trim Valve controllers are connected by means of 12 pin Deutsch

connectors, as stated in their respective manuals [35][37]. A brief explanation of the

wiring is stated here. In Figure 3.4, a number preceded by the letter ’a’ denotes that
Chapter 3. Test Equipment and Setup 26

Power On LED
+12 VDC 6 A fuse
a5

560 Ω  
a1

b1

a5

b5

+5 VDC
Analog Mode LED

a5

Mode
180 Ω  
Selector a4 (incr)

a6 (decr)

10k pot

a8

Auto LED
a5

180 Ω   b4
Φ+ Signal b8
10k pot
Φ GND Signal b3
b6
Manual LED
MPU + Signal a11
a5

MPU GND Signal a3

180 Ω  

Throttle Position a2
DAQ DAQ
Output: Dev 2: pin 21 a10 Inputs: Trim Valve Position b2

Figure 3.4: ITB and Trim Valve wiring. Letter ”a” indicates a connection to the ITB

and letter ”b” to the Trim Valve.

Chapter 3. Test Equipment and Setup 27

Figure 3.5: The ’Overview’ tab displayed when monitoring the speed controller through

the L-Series Service Tool.

Chapter 3. Test Equipment and Setup 28

Figure 3.6: The ’Overview’ tab displayed when monitoring the air-fuel ratio controller

through the L-Series Service Tool.

Chapter 3. Test Equipment and Setup 29

it is the respective pin number on the ITB, whereas a number preceded by the letter

’b’ is that on the Trim Valve. When pins a4 and a6 are both activated, the ITB is in

analog mode (analog mode can be switched on and o↵ via the ’Mode Selector’ switch).

The MPU+ signal is the A signal from the encoder, divided by 100 (resulting in a TTL

signal that gives 36 pulses per revolution). Pin 2 (a2) on the ITB provides a 0-5V signal

(referenced from the common ground) which indicates throttle position information and

is monitored through the DAQ. Pin 2 (b2) on the Trim Valve provides a similar signal to

indicate the valve position. When a10 is activated (a +5V signal output from the control

box), the control switches from the idle speed to the rated speed (3600 rpm). The +

signal is an analog signal from the WEGO sensor, and when the AFR ’Mode Selector’

switch is not set to ’Auto’, then it is in the preconfigured position for open-loop control.

Safety considerations included the use of an overspeed trip. Upon sensing an over-

speed condition, the system was set up (on an external circuit) to cut o↵ the fuel supply.

Throughout the setup, especially when tuning the throttle PID, high amplitude spikes

were seen in engine speed, demonstrating the importance of the trip. These spikes would

cause misreads in the encoder counter signal, giving instantaneous values much higher

than the overspeed condition (greater than 4000 rpm), which rendered the use of this

signal as an input to the overspeed trip ine↵ective. Using a digital-to-analog circuit,

the frequency was converted to a 0-5 V signal which gave more accurate engine speed

information during high acceleration transitions. This analog signal was more reliable

and therefore used as the input to the overspeed trip.

3.6.2 Air-Fuel Ratio Control

When fuelling the engine with simulated biogas (NG diluted in CO2 ), the flow require-

ments increased such that the recommended upstream fuel line setup for the speed and

air-fuel ratio controllers was not adequate. A work-around involved removing the zero-

pressure regulator that was directly upstream of the trim valve. However, due to the
Chapter 3. Test Equipment and Setup 30

increased fuel pressure at the trim valve, slight changes in trim valve position resulted

in large changes in , and thus poor air-fuel ratio control. To counteract this, the pro-

portional gain on the trim valve controller was significantly reduced (by a factor of 10).

Since the tests were run at steady state, the slow response that this resulted in was

not a problem. In order to get the engine to the required steady state operating condi-

tions however, a larger proportional gain was needed, and so the system was set up with

two gain settings. To quickly bring the engine near the desired set point with regards

to air-fuel ratio, the ’dynamics setting’ with the larger proportional gain was used and

then the controller was switched to operate with the lower gain setting by means of the

potentiometer on the control box that controls the voltage seen by pin b6.
Chapter 4

Testing

The test program was conducted in two parts. The first part consisted of preliminary

tests to validate the substitution of natural gas for pure methane and to determine

the optimal spark timing and optimal fuel-air equivalence ratio for maximum emissions

reduction. The second part of the tests evaluated the impact of carbon dioxide content

(reduced methane content) of the simulated biogas fuels.

4.1 Preliminary Tests

4.1.1 Natural Gas Justification

In all ensuing tests, natural gas was used as a substitute for methane, as it was much

more accessible and much less expensive. Natural gas provided through the Enbridge

distribution system consistently has a methane content of about 95%. In order to justify

this substitution, tests were conducted to compare results with pure methane and pure

natural gas. These tests were conducted with a constant of 1.013 (see Section 4.1.2)

and a load of 6 kW. Figure 4.1 shows the di↵erence in engine-out emissions when using

the two di↵erent fuels, and Figures 4.2 and 4.3 show the di↵erences in the e↵ectiveness

of the catalyst in reducing emissions when using NG versus CH4 as the fuel. Engine-

31
Chapter 4. Testing 32

out THC and NOx emissions are similar in both cases (within 5% of each other). The

reduction of both of these emissions is also consistent at the two catalyst space velocities

tested. Carbon monoxide emissions are slightly higher when using NG versus CH4 , which

could explain the slightly lower reductions across the catalyst. Further analysis showed

that although the nominal value was 1.013 for both of these tests, the mean measured

values for the NG and CH4 data points were 1.0124 and 1.0127, respectively. The slightly

richer conditions during the NG tests may have contributed to the discrepancy seen with

regards to CO emissions. However, CO is almost completely reduced across the catalyst

with a space velocity of 12 228 hr-1 , for either fuel. Since the emissions results are very

close, natural gas was considered a valid substitute for pure methane.

3500
3261.8
3021.0
3000
Engine Out Emission Concentration (ppm)

2500

2000
NG
CH
4
1500

1000 871.8
830.2
603.6 616.1
500

0
THC NOx CO
Emission

Figure 4.1: Comparison of engine-out emissions when the engine is fueled by NG vs.

CH4 .
Chapter 4. Testing 33

80

Emission Reduction (Catalyst Space Velocity = 24 457 hr−1) (%)

71.3 70.0
70

60

51.2
50 47.8

NG
40 CH4

30

20.1 20.9
20

10

0
THC NOx CO
Emission

Figure 4.2: Comparison of emission reductions across a catalyst with a space velocity of

24 457 hr-1 (1 catalyst) when the engine is fueled by NG vs. CH4 .

99.3 99.8 98.3

100
96.1
Emission Reduction (Catalyst Space Velocity = 12 228 hr−1) (%)

90

80

70

60

49.6 49.8 NG
50 CH4

40

30

20

10

0
THC NOx CO
Emission

Figure 4.3: Comparison of emission reductions across a catalyst with a space velocity of

12 228 hr-1 (2 catalysts in series) when the engine is fueled by NG vs. CH4 .
Chapter 4. Testing 34

4.1.2 Determination of Optimal Spark Timing and Equivalence

Ratio

In order to determine the optimal spark timing at each fuel composition, spark timing

sweeps were conducted, as detailed in Appendix A.1. For the convenience of the reader,

the optimal spark timing for each fuel composition is repeated in Table 4.1 (also shown in

Appendix A.1). Although these spark timings were optimized for minimum catalyst-out

emissions and were not MBT timings, they do agree with the findings of Hu et al. [14],

Hu and Huang [13], and Park et al. [24], which all state a requirement for greater spark

advance with increased dilution. Likewise, sweeps were conducted with the maximum

methane-content fuel (NG) and the minimum methane-content fuel (a mixture consisting

of 60% NG and 40% CO2 ), in order to determine the optimal for catalyst performance

(see Appendix A.2). It was found that the optimal value of 1.013 was the same with both

fuel compositions, and was therefore used as the optimal value for all fuel compositions

tested. This is close, although slightly leaner than the optimal value (1.015) determined

by Tadrous [31].

Table 4.1: Optimal spark timing for emissions at di↵erent fuel compositions.

% NG % CO2 Spark Advance (CAD BTC)

100 0 25

80 20 35

75 25 40

70 30 45

65 35 45

60 40 50

Both the mean value and the amplitude of the fluctuations about the mean value of

the fuel-air equivalence ratio ( ) are important for catalyst performance. Due to the
Chapter 4. Testing 35

difficulties experienced with regards to air-fuel ratio control, mentioned in Section 3.6.2,

a set of tests was conducted in which the equivalence ratio was held constant but the

amplitude of the oscillations about the mean was increased. The e↵ect that changing this

has on THC, NOx and CO emissions is shown in Figures 4.4, 4.5 and 4.6, respectively.

From these figures, the only significant e↵ect seen with changes in the oscillation window

is the significant increase in the variability in the reduction of NOx as the oscillation

about the mean value increases above ± 0.005. This is fairly consistent with the

findings of Heywood [11] which showed that the width of the window in which all three

emissions are removed with high efficiency is about 0.007 equivalence ratio units. With

the system set up as it was, this could be achieved for all fuel compositions, and as such,

only data sets in which the standard deviation of was equal to or less than 0.005 were

used for analysis throughout this thesis.

65
Catalyst Space Velocity = 12 228 hr−1

60
THC Emissions Reduction (%)

55

50

45

40

35

30
−0.0010 0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080
Φ Oscillation Window

Figure 4.4: Reduction of THC emissions across the catalyst with a mean value of 1.013

and with oscillations about the mean as displayed on the x-axis. Tests conducted with a

constant load of 6 kW with NG as the fuel.

Chapter 4. Testing 36

120
Catalyst Space Velocity = 12 228 hr−1

110

NOx Emissions Reduction (%) 100

90

80

70

60

50
−0.0010 0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080
Φ Oscillation Window

Figure 4.5: Reduction of NOx emissions across the catalyst with a mean value of 1.013

and with oscillations about the mean as displayed on the x-axis. Tests conducted with a

constant load of 6 kW with NG as the fuel.

100
Catalyst Space Velocity = 12 228 hr−1

98
CO Emissions Reduction (%)

96

94

92

90

88
−0.0010 0.0000 0.0010 0.0020 0.0030 0.0040 0.0050 0.0060 0.0070 0.0080
Φ Oscillation Window

Figure 4.6: Reduction of CO emissions across the catalyst with a mean value of 1.013

and with oscillations about the mean as displayed on the x-axis. Tests conducted with a

constant load of 6 kW with NG as the fuel.

Chapter 4. Testing 37

4.2 Brake Specific Energy Consumption

With increasing CO2 content in the fuel mixture, the engine fuel requirement at the same

load increases. As such, brake specific fuel consumption would not be a useful measure

of the e↵ect that fuel composition has on engine efficiency. Instead, the brake specific

energy consumption (bsec), defined in Equation 4.1 was used, where Ė is the energy flow

rate and Pb is the brake power. The energy flow rate is the product of the fuel mass

flow rate and its lower heating value. Figure 4.7 shows that at the lower load condition,

with increasing CO2 content, the bsec increases, which is in agreement with the findings

of Tadrous [31]. Although the bsec at the higher load condition is greater at the fuel

compositions that include CO2 , compared to the NG test point, the trend here is not

clear. Figure 4.8 shows that the e↵ect of load is much greater than that of changing fuel

composition.

Ė
bsec = (4.1)
Pb

9000
Load=6.0 kW

8900

8800
bsec (kJ/kWhr)

8700

8600

8500

8400
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.7: Brake specific energy consumption at di↵erent fuel compositions with a 6 kW

engine load.
Chapter 4. Testing 38

4
x 10
9
Load=6.0 kW
Load=8.0 kW
8

6
bsec (kJ/kWhr)

0
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.8: Brake specific energy consumption at di↵erent fuel compositions and at two

di↵erent load conditions.

4.3 Emissions Tests

Engine-out emissions were measured along with post-catalyst emissions (after 1 and after

2 catalysts), in order to show the e↵ect of space velocity, or, inversely, residence time.

Space velocity, s, is defined in equation 4.2, where Vcatalyst is the catalyst volume in ft3

and V̇air is the volumetric flow rate of air into the engine in standard ft3 /hr [31].

V̇air
s= (4.2)
Vcatalyst

These test points e↵ectively showed emissions reductions across catalysts with respective

space velocities of 24 457 hr-1 (one catalyst) and 12 228 hr-1 (two catalysts in series).

Engine-out emissions, as seen in Figures 4.9, 4.10, 4.12 and 4.13 show trends in

THC and NOx emissions with regards to fuel composition. At both 6 and 8 kW loads,

THC emissions increase with increasing CO2 fuel content. This is in agreement with the

findings of Propatham et al. [25] and Tadrous [31]. A study by Park et al. [24] used

N2 as opposed to CO2 as a diluent with methane (stating that the overall trends of N2
Chapter 4. Testing 39

dilution are similar to those of CO2 dilution), and also found a similar trend with regards

to THC emissions and fuel methane content. Hu and Huang [13] also conducted a study

in which NG was diluted with EGR and similarly showed an increase in THC emissions

with increasing dilution. Hydrocarbon emissions form through several mechanisms in SI

engines, but the increase with increasing dilution is due mainly to incomplete combustion.

With increasing dilution, unburned gas temperature and pressure during combustion

decrease, resulting in lower combustion quality. Typically, a substantial amount of the

unburned hydrocarbons that escape the primary combustion process are oxidized during

the latter segment of the expansion stroke, and during the exhaust process. However,

with increasing dilution, exhaust temperatures are lower, resulting in worse oxidation and

exacerbating the situation where already relatively high THC emissions have resulted

from incomplete combustion [11].

These same figures also show that at both load conditions, engine-out NOx emissions

decrease with increasing CO2 content in the fuel. Studies by Lee et al. [18] along with Hu

and Huang [13], where EGR was used to dilute the fuel mixture, showed similar trends.

Park’s study with N2 dilution and Porpatham et al.’s research with biogas showed the

same trend [24][25]. Nitrogen oxide engine emissions are predominantly comprised of

NO, the formation rates of which have strong temperature dependence. Therefore, for

the same reasons that THC emissions increase with increasing CO2 in the fuel mixture,

namely, decreasing temperatures, NOx emissions decrease [11].

Carbon monoxide emissions do not seem to change with increasing CO2 content (see

Figures 4.11 and 4.14). Since the equivalence ratio was held constant at 1.013 for all test

points, this was expected, as CO emissions are dependent on [11].

Chapter 4. Testing 40

4.3.1 Emissions

6 kW Load

1200
Engine−Out
−1
After Catalyst (Space Velocity = 24 457 hr )
1100
After Catalyst (Space Velocity = 12 228 hr−1)

1000

900
THC Emissions (ppm)

800

700

600

500

400

300
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.9: THC emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

Chapter 4. Testing 41

700
Engine−Out
After Catalyst (Space Velocity = 24 457 hr−1)
600 After Catalyst (Space Velocity = 12 228 hr−1)

500
NOx Emissions (ppm)

400

300

200

100

−100
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.10: NOx emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

4000
Engine−Out
After Catalyst (Space Velocity = 24 457 hr−1)
After Catalyst (Space Velocity = 12 228 hr−1)
3500

3000
CO Emissions (ppm)

2500

2000

1500

1000

500

0
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.11: CO emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

Chapter 4. Testing 42

8 kW Load

1300
Engine−Out
−1
1200 After Catalyst (Space Velocity = 24 457 hr )
After Catalyst (Space Velocity = 12 228 hr−1)
1100

1000
THC Emissions (ppm)

900

800

700

600

500

400

300
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.12: THC emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

Chapter 4. Testing 43

1200
Engine−Out
−1
After Catalyst (Space Velocity = 24 457 hr )
−1
After Catalyst (Space Velocity = 12 228 hr )
1000

800
NOx Emissions (ppm)

600

400

200

−200
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.13: NOx emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

4000
Engine−Out
After Catalyst (Space Velocity = 24 457 hr−1)
After Catalyst (Space Velocity = 12 228 hr−1)
3500

3000
CO Emissions (ppm)

2500

2000

1500

1000

500

0
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.14: CO emissions with a mean value of 1.013 as measured at 3 di↵erent

locations and with various fuel compositions.

Chapter 4. Testing 44

4.4 Brake Specific Emissions

Previous research conducted on this engine by Foster [9] included tests with the engine

running on gasoline. A comparison of engine-out brake specific emissions (normalized to

engine brake power) is presented here. Figures 4.15, 4.16 and 4.17 display brake specific

emissions at two di↵erent load conditions and against varying fuel compositions. The

experimental setup detailed by Foster [9] uses the same generator as that used for the

biogas testing, and as such, the generator efficiency curve shows that 6 kW and 8 kW

loads correspond to brake power outputs of approximately 8 kW and 10 kW, respectively.

Foster’s tests were conducted at engine power outputs ranging from 1 to 13 kW, but the

test points at 8 and 10 kW were used as comparison points to the tests conducted for

this research. Specific emissions at these two load points are plotted alongside biogas

emission data in the succeeding plots. It should be noted that the CO plots do not

provide this data, as the brake specific CO emissions in Foster were much higher than

the emissions seen when running on the various fuel mixtures used to simulate biogas.

Whereas the highest engine-out CO emissions value seen in Figure 4.17 is 12 g/kWh,

the gasoline testing conducted by Foster showed values of approximately 180 and 110

g/kWh at the 8 and 10 kW engine power points, respectively. The engine-out THC

emissions were much higher with the NG and biogas tests than with gasoline. This

can be attributed to the recurrent theme of difficulties with regards to the oxidation of

methane. It can also be seen from Figure 4.15 that with increasing dilution, brake specific

THC emissions increase. Engine-out NOx emissions are very similar between NG and

gasoline, and, the e↵ect of engine load is much more significant than it is on brake specific

THC emissions, even though these emissions are already normalized to the engine power,

a finding that Foster also noted. Increasing fuel CO2 content clearly results in decreasing

NOx emissions. At the lower load condition, specific CO emissions are higher, and the

trend seems to show that these increase with greater dilution, although the trend is not

consistent.
Chapter 4. Testing 45

4
6 kW Load
8 kW Load
Foster 8 kW Brake Power
3.5 Foster 10 kW Brake Power

3
bsTHC (g/kWh)

2.5

1.5

0.5
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.15: A comparison of engine-out brake specific THC emissions achieved with the

biogas setup, and previous testing running on gasoline.

5
6 kW Load
8 kW Load
4.5 Foster 8 kW Brake Power
Foster 10 kW Brake Power

3.5
bsNOx (g/kWh)

2.5

1.5

1
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.16: A comparison of engine-out brake specific NOx emissions achieved with the

biogas setup, and previous testing running on gasoline.

Chapter 4. Testing 46

9.5
6 kW Load
8 kW Load

bsCO (g/kWh) 8.5

7.5

6.5
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.17: Engine-out brake specific CO emissions achieved with the biogas setup, at

varying fuel compositions.

In order to demonstrate the value of a catalyst, these same engine-out brake specific

emissions from the gasoline tests were further compared to post-catalyst brake specific

emissions from the biogas tests in Figures 4.18, 4.19 and 4.20. Here, although there is a

significant decrease, specific THC emissions from the biogas tests are still much higher

than those achieved with gasoline fuel. Also the trend with regards to the e↵ect of CO2

dilution still holds. A significant reduction is seen in specific NOx emissions. Specific CO

emissions, already much lower than their counterparts from the gasoline tests, are even

further reduced.
Chapter 4. Testing 47

2
6 kW Load
8 kW Load
Foster 8 kW Brake Power
1.8 Foster 10 kW Brake Power

bsTHC (g/kWh) 1.6

1.4

1.2

0.8

−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.18: A comparison of post-catalyst (space velocity = 12 228 hr-1 ) brake specific

THC emissions achieved with the biogas setup, and previous testing running on gasoline

(engine-out).

3
6 kW Load
8 kW Load
Foster 8 kW Brake Power
2.5 Foster 10 kW Brake Power

2
bsNOx (g/kWh)

1.5

0.5

0
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.19: A comparison of post-catalyst (space velocity = 12 228 hr-1 ) brake specific

NOx emissions achieved with the biogas setup, and previous testing running on gasoline

(engine-out).
Chapter 4. Testing 48

1
6 kW Load
8 kW Load
0.9

0.8

bsCO (g/kWh) 0.7

0.6

0.5

0.4

0.3

0.2

0.1
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.20: Post-catalyst (space velocity = 12 228 hr-1 ) brake specific CO emissions

achieved with the biogas setup, at varying fuel compositions.

4.5 Catalyst Performance

4.5.1 Emission Reductions

Next, catalyst performance was evaluated. Figures 4.21, 4.22, 4.23, 4.24, 4.25 and 4.26

show the e↵ectiveness of the catalyst at reducing THC, NOx and CO emissions at two

load conditions and with two di↵erent catalyst space velocities.

There does not seem to be a consistent trend with regards to any e↵ect that the fuel

composition may have on the e↵ectiveness of the catalyst at reducing THC emissions, at

either load condition. However, with two catalysts instead of one, approximately twice as

much reduction is seen across all fuel compositions, at both load conditions. This suggests

that with a decrease in space velocity, full hydrocarbon reduction can be achieved. It

is worth noting that low THC emissions reduction was expected, as methane is difficult

to oxidize. These results are in agreement with previous research conducted by Tadrous

[31].

The e↵ectiveness of the catalyst at reducing NOx emissions decreases with increasing
Chapter 4. Testing 49

CO2 , at both load conditions. Also, whereas halving the space velocity approximately

doubles the reduction of THC and CO emissions, the same does not hold true for NOx

emissions, even though more than 50% (mean values) of emissions are reduced across

only 1 catalyst under both load conditions. This suggests that there is another factor

limiting the full reduction of NOx , separate from the catalyst space velocity.

Under both load conditions, CO emissions are reduced by approximately 50% when

passed through the catalyst with a space velocity of 24 457 hr-1 (1 catalyst) and are

approximately fully reduced when the space velocity is halved (2 catalysts).

Chapter 4. Testing 50

6 kW Load

80
Catalyst Space Velocity = 24 257 hr−1
−1
70 Catalyst Space Velocity = 12 228 hr

60
THC Emissions Reduction (%)

50

40

30

20

10

−10
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.21: THC emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 6 kW.

Chapter 4. Testing 51

120
Catalyst Space Velocity = 24 257 hr−1
−1
110 Catalyst Space Velocity = 12 228 hr

NOx Emissions Reduction (%) 100

90

80

70

60

50

40

30
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.22: NOx emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 6 kW.

110
Catalyst Space Velocity = 24 257 hr−1
Catalyst Space Velocity = 12 228 hr−1
100

90
CO Emissions Reduction (%)

80

70

60

50

40

30
−5 0 5 10 15 20 25 30 35 40 45
CO2 % in Fuel

Figure 4.23: CO emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 6 kW.

Chapter 4. Testing 52

8 kW Load

70
Catalyst Space Velocity = 24 257 hr−1
Catalyst Space Velocity = 12 228 hr−1
THC Emissions Reduction (%) 60

50

40

30

20

10

0
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.24: THC emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 8 kW.

Chapter 4. Testing 53

120
Catalyst Space Velocity = 24 257 hr−1
−1
110 Catalyst Space Velocity = 12 228 hr

NOx Emissions Reduction (%) 100

90

80

70

60

50

40

30

20
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.25: NOx emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 8 kW.

100

90

80
CO Emissions Reduction (%)

70

60

50

40

30
Catalyst Space Velocity = 24 257 hr−1
Catalyst Space Velocity = 12 228 hr−1
20
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.26: CO emissions reductions with di↵erent fuel compositions. Operating with

a mean of 1.013 and an engine load of 8 kW.

Chapter 4. Testing 54

4.5.2 Catalyst Performance vs. Relative Emission Ratios

To provide further insight on the factors a↵ecting catalyst performance, the reduction

of THC and NOx emissions versus the relative proportions of engine-out emissions was

explored. Typically, three way catalysts operate best at a stoichiometric equivalence

ratio, as fluctuations about a stoichiometric mean allow the catalyst to undergo reduction-

oxidation reactions. When the catalyst is in an oxidized state, CO and THC emissions

can be oxidized, reducing the catalyst component, which can then react with NO under

lean conditions (leaving the catalyst component in an oxidized state, to repeat the cycle)

[11]. Since slightly rich equivalence ratios were used for these tests however, less O2 was

available to oxidize CO and hydrocarbons. NO catalysis, which involves the reduction of

NO by CO, hydrocarbons and H2 in the exhaust, is actually more e↵ective under these

rich conditions as there are much more reducing species than oxidizing species present.

Under these reducing conditions, possible NO reactions are [11]:

1
N O + CO ! N2 + CO2 (4.3)
2

2N O + 5CO + 3H2 O ! 2N H3 + 5CO2 (4.4)

Figure 4.27 illustrates the indirect e↵ect that NOx emissions have on the ability of the

catalyst to oxidize hydrocarbon emissions. An interesting observation from this figure

is that THC and NOx are competing. As the NOx /THC ratio increases, the conversion

efficiency with regards to THC emissions decreases. Although the equivalence ratio is

slightly rich, CO is oxidized through Equation 4.3, but there is less oxidation of THC

emissions. However, as the NOx /THC ratio continues to increase, there comes a point

where THC oxidation increases. Further analysis showed that the NOx /THC ratio was

dependent on the operating conditions, namely, the fuel composition. With increasing

dilution, the NOx /THC ratio decreases, as seen in Figure 4.28. It is also known that

with increasing dilution, there are two major temperature e↵ects that significantly a↵ect

the formation of NOx and the oxidation of hydrocarbons. The increased heat capacity of
Chapter 4. Testing 55

the charge in the cylinder that is associated with increased dilution, corresponds to lower

peak combustion temperatures, and thus, a decrease in NO formation (as NO formation

rates are highly temperature dependant). Although lower peak temperatures result, there

are also higher exhaust temperatures, which are due to increased combustion durations,

and subsequently, higher temperatures at the exhaust valve opening event. These trends

are shown in Figure 4.29, and can be used to explain the changes in THC oxidation

across the catalyst. The decreasing exhaust temperatures that coincide with increasing

NOx /THC ratios explain the decreasing THC conversion rates, as less oxidation occurs

in the catalyst at lower temperatures. The results of these temperature e↵ects can also

explain the eventual increase in THC conversion at high enough NOx /THC ratios.

While looking at Figure 4.29 keep in mind that approximately full CO oxidation is

seen through both catalysts. It is believed that unburned hydrocarbons are competing

with CO for oxidation via the catalyst oxidation sites. As CO is much more reactive

than hydrocarbons, and in particular, methane, CO oxidation suppresses hydrocarbon

oxidation. However, with high enough NOx levels, CO oxidation via the NO reaction

decreases the CO concentrations in the exhaust, freeing up catalyst oxidation sites for

THC oxidation. Evidence that this reaction is indeed occurring can be found in Figure

4.30, which shows that, at all fuel compositions tested, the CO2 concentration in the

exhaust gas stream increases as it flows through the catalysts. As CO2 (along with N2 )

is a product of the NO reduction reaction, this explanation is plausible.

Chapter 4. Testing 56

60
Space Velocity = 24 457 hr−1, 6 kW Load
55 Space Velocity = 12 228 hr−1, 6 kW Load
Space Velocity = 24 457 hr−1, 8 kW Load
50 Space Velocity = 12 228 hr−1, 8 kW Load
THC Emissions Reduction (%)

45

40

35

30

25

20

15

10
0 0.2 0.4 0.6 0.8 1 1.2 1.4
NOx/THC

Figure 4.27: THC emissions reductions against NOx /THC ratio at catalyst entry. Oper-

ating with a mean of 1.013 and at two di↵erent load conditions.

1.05
Load=8.0 kW
1

0.95

0.9

0.85
NOx/THC

0.8

0.75

0.7

0.65

0.6

0.55
−5 0 5 10 15 20 25 30 35
CO % in Fuel
2

Figure 4.28: Variation in the engine-out NOx /THC concentration ratio with changing

fuel composition.
Chapter 4. Testing 57

60
Space Velocity = 24 457 hr−1, 6 kW Load
55 Space Velocity = 12 228 hr−1, 6 kW Load
Space Velocity = 24 457 hr−1, 8 kW Load
THC Emissions Reduction (%) 50 Space Velocity = 12 228 hr−1, 8 kW Load

45

40

35 Increasing Peak In−cylinder Temperatures

30

25
Decreasing Exhaust Temperatures
20

15

10
0 0.2 0.4 0.6 0.8 1 1.2 1.4
NOx/THC

Figure 4.29: Further analysis clarifies changes in the catalyst THC conversion efficiency.

18
Engine−Out
After Catalyst (Space Velocity = 24 457 hr−1)
17 After Catalyst (Space Velocity = 12 228 hr−1)

16
CO Emissions (%)

15

14
2

13

12

11
−5 0 5 10 15 20 25 30 35
CO2 % in Fuel

Figure 4.30: Increase in CO2 concentration in the exhaust gas stream as it flows through

the catalysts, confirming that the NO reaction with CO is occurring.

Chapter 5

Conclusions

Fuel mixtures containing NG and CO2 in varying proportions were used to fuel a 3-

cylinder Kubota SI engine, and the e↵ects that the fuel composition had on emissions

and catalyst performance were studied.

In order for subsequent tests to be of greater value, the optimal spark timing for

emissions reduction was first determined. This set of tests revealed that as the spark

advanced, THC and NOx emissions increase, and so the most retarded timing at which

the engine would operate before stability began to deteriorate was selected as optimal

from an emissions standpoint, at each fuel composition. The trend seen here was that as

CO2 content in the fuel increases, greater spark advance is required, which is consistent

with the literature.

Once engine-out emissions had been minimized by means of spark timing, catalyst

performance was investigated. Catalyst performance is dependent mainly on temper-

ature (light-o↵ must be achieved) and the fuel-air equivalence ratio. With the spark

timings determined to be optimal at each fuel composition, catalyst temperatures were

sufficiently high, and so the equivalence ratio at which superior performance was ob-

tained was investigated. It was found that fuel composition does not change the optimal

equivalence ratio for catalyst performance, which was found to be 1.013 with the catalyst

58
Chapter 5. Conclusions 59

formulation used for testing.

Having completed these preliminary tests, primary testing on the e↵ect of fuel com-

position on engine-out emissions and catalyst performance was conducted, using the

optimal spark timing and equivalence ratio for each test point. With increasing CO2

content, engine out THC emissions increased, and NOx emissions decreased, at both

load conditions tested. As the load was increased, THC emissions rose slightly and NOx

emissions increased significantly. CO emissions did not show a trend with increasing

CO2 content, and slightly decreased as the load was increased. Brake specific emissions

were also plotted versus the fuel composition, and a comparison was made between the

engine-out emissions achieved with NG and those achieved in previous research with

gasoline as the fuel. These tests showed comparable specific NOx emissions. Due to

oxidization issues, higher specific hydrocarbon emissions were associated with NG as

opposed to gasoline, but much lower CO emissions were also a result. Furthermore, a

comparison of the engine-out specific emissions from the gasoline tests to post-catalyst

specific emissions from biogas tests, revealed that NOx emissions could be reduced to

levels significantly lower than those achieved with gasoline and that THC emissions also

could be reduced such that they approach the levels of engine-out gasoline emissions.

CO emissions, already much lower with NG than with gasoline, were also significantly

reduced with the catalyst.

Emission reductions across the catalyst were conducted at two di↵erent space veloci-

ties, e↵ectively achieved by using either one catalyst or two catalysts in series. At both

load conditions, THC emissions were reductions were similar. With 2 catalysts, at the

higher load condition, the lowest THC emissions concentration achieved was approxi-

mately 500 ppm (about 400 ppm at lower load), with pure NG as the fuel. Although

no clear trend was seen with regards to the e↵ect of fuel composition on catalyst perfor-

mance, the tailpipe (or catalyst-out) emissions under both load conditions showed higher

hydrocarbon emissions with increasing CO2 content in the fuel. Doubling the catalyst
Chapter 5. Conclusions 60

residence time resulted in approximately doubling the reduction, suggesting that with

a large enough catalyst, full THC reduction can be achieved. Tailpipe NOx emissions

increased with increasing CO2 fuel content, despite an opposite trend in engine-out emis-

sions, suggesting that catalyst performance decreases with increasing CO2 dilution. The

majority of NOx emissions are reduced across the first catalyst, and the second cata-

lyst provides further reduction, though at a lower efficiency. Tailpipe NOx emissions

were below 200 ppm at the higher load condition and below 100 ppm at the lower load.

Carbon monoxide emissions were almost fully reduced at both load conditions, with ap-

proximately 50% reduction through the first catalyst and the rest through the second

catalyst. If a larger catalyst is used (for better THC reduction), then CO emissions are

not a concern as they are removed the most e↵ectively of all three. No trend with regards

to catalyst CO reduction and fuel composition was observed.

Plots were made showing the e↵ect that the relative proportions of THC and NOx

engine-out emissions had on catalyst performance. It was found that as the THC/NOx

ratio increased, catalyst performance with regards to the oxidation of THC emissions

was hindered, due to reactions involving both NOx and CO, resulting in CO emissions

being oxidized while hydrocarbon emissions were not. Under rich conditions, and with

methane being particularly difficult to oxidize, this e↵ect was amplified. This suggests

that an emissions control strategy for biogas-fueled engines would require tuning that is

specific to a particular catalytic converter formulation.

The results show that low emissions can be achieved from a spark ignition engine

fueled by biogas through the use of conventional three-way catalytic converter technology.
Chapter 6

Recommendations for Future Work

The results discussed in the latter parts of Chapters 4.5.2 and 5 suggest that the catalyst

performance pertaining to its reduction of THC emissions is dependent on the NOx /CO

ratio. However, changing exhaust temperatures make this conclusion arguable. Testing

using a steady-state test bench rather than an engine (as suggested previously by Tadrous

[31]) could provide further insight towards what factors have dominant e↵ects on the

outcomes.

61
Bibliography

[1] International energy agency, 2011.

[2] International energy outlook 2011. Technical report, U.S. Energy Information Ad-

ministration, September 2011.

[3] S. Orhan Akansu, Zafer Dulger, Nafiz Kahraman, and T. Nejat Veziroğlu. “Internal

combustion engines fueled by natural gas-hydrogen mixtures”. International Journal

of Hydrogen Energy, 29(14):1527–1539, 2004.

[4] A.B. Baspinar, M. Turker, A. Hocalar, and I. Ozturk. “Biogas desulphurization at

technical scale by lithotrophic denitrification: Integration of sulphide and nitrogen

removal”. Process Biochemistry, 46(4):916–922, 2011.

[5] Iván D. Bedoya, Samveg Saxena, Francisco J. Cadavid, Robert W. Dibble, and

Martin Wissink. “Experimental study of biogas combustion in an HCCI engine

for power generation with high indicated efficiency and ultra-low NOx emissions”.

Energy Conversion and Management, 53(1):154–162, 2011.

[6] Marie Bysveen. “Engine characteristics of emissions and performance using mixtures

of natural gas and hydrogen”. Energy, 32(4):482–489, 2007.

[7] R. Chandra, V.K. Vijay, P. M.V. Subbarao, and T. K. Khura. “Performance evalu-

ation of a constant speed IC engine on CNG, methane enriched biogas and biogas”.

Applied Energy, 88(11):3969–3977, 2011.

62
Bibliography 63

[8] M. Fatih Demirbas and Mehmet Balat. “Progress and recent trends in biogas pro-

cessing”. International Journal of Green Energy, 6(2):117–142, 2009.

[9] Matthew Foster. “Design of a Hydrogen Injection System for a Prechamber

Hydrogen-Fueled Internal Combustion Engine”. Master’s thesis, University of

Toronto, 2009.

[10] Brian Eric Hallgren. Impact of Retarded Spark Timing on Engine Combustion,

Hydrocarbon Emissions, and Fast Catalyst Light-O↵. PhD thesis, MIT, 1976.

[11] John B. Heywood. Internal Combustion Engine Fundamentals. McGraw-Hill, 1988.

[12] John Theodore Houghton, Bert Bolin, and L. Gylvan Mera Filho. Climate Change

1995: The Science of Climate Change. Number 22. Cambridge University Press,

1996.

[13] Erjiang Hu and Zuohua Huang. “Optimization on ignition timing and EGR ratio of

a spark-ignition engine fueled with natural gas-hydrogen blends”. Number 2011-01-

0918, Detroit, 2011. SAE 2011 World Congress and Exhibition, SAE.

[14] Erjiang Hu, Zuohua Huang, Bing Liu, Jianjun Zheng, and Xiaolei Gu. “Experi-

mental study on combustion characteristics of a spark ignition engine fueled with

natural gas-hydrogen blends combining with EGR”. International Journal of Hy-

drogen Energy, 34(2):1035–1044, 2009.

[15] Peter V.K. Funk Jr., Daniel J. Bauer, and Duane Morris. “The promise and realities

of biogas produced from landfills and farm waste”. Cogeneration and Distributed

Generation Journal, 25(2):44–51, 2010.

[16] Nafiz Kahraman, Bilge Cepera, S. Orhan Akansu, and Kadir Aydin. “Investigation of

combustion characteristics and emissions in a spark-ignition engine fueled with nat-

Bibliography 64

ural gas-hydrogen blends”. International Journal of Hydrogen Energy, 34(2):1026–

1034, 2009.

[17] Ja Hun Kwak, Do Heui Kim, Janos Szanyu, and Charles H.F. Peden. “Excellent

sulfur resistance of Pt/BaO/CeO2 lean NOx trap catalysts”. Applied Catalysts B:

Environmental, 84(3-4):545–551, 2008.

[18] Kyungtaek Lee, Taesoo Kim, Hyoseok Cha, Soonho Song, and Kwang Min Chin.

“Generating efficiency and NOx emissions of a gas engine generator fueled with a

biogas–hydrogen blend and using an exhaust gas recirculation system”. International

Journal of Hydrogen Energy, 35(11):5723–5730, 2010.

[19] Stefan Luck, Gabriele Büge, Holger Plettenberg, and Martin Ho↵mann. “Near-

infrared spectroscopy for process control and optimization of biogas plants”. Engi-

neering in Life Sciences, 10(6):537–543, 2010.

[20] D. I. Massé, G. Talbot, and Y. Gilbert. “On farm biogas production: A method to

reduce GHG emissions and develop more sustainable livestock operations”. Animal

Feed Science and Technology, 166-167:436–445, 2011.

[21] D. Mescia, S. P. Hernández, A. Conoci, and N. Russo. “MSW landfill biogas desul-

furization”. International Journal of Hydrogen Energy, 36(13):7884–7890, 2011.

[22] Nirendra N. Mustafi and Robert R. Raine. “Application of a spark ignition engine

simulation tool for alternative fuels”. Journal of Engineering for Gas Turbines and

Power, 130(1), January 2008.

[23] Ivo Achu Nges, Federico Escobar, Fu Xinmei, and Lovisa Björnsson. “Benefits of

supplementing an industrial waste anaerobic digester with energy crops for increased

biogas production”. Waste Management, 32(1):53–59, 2011.

Bibliography 65

[24] Cheolwoong Park, Seunghyun Park, Yonggyu Lee, Changgi Kim, Sunyoup Lee, and

Yasuo Moriyoshi. “Performance and emission characteristics of a SI engine fueled

by low calorific biogas blended with hydrogen”. International Journal of Hydrogen

Energy, 36(16):10080–10088, 2011.

[25] E. Porpatham, A. Ramesh, and A. Nagalingam. “Investigation on the e↵ect of

concentration of methane in biogas when used as a fuel for a spark ignition engine”.

Fuel, 87(8-9):1651–1659, 2007.

[26] S. Rasi, J. Läntelä, and J. Rintala. “Trace compounds a↵ecting biogas energy

utilisation - A review”. Energy Conversion and Management, 52(12):3369–3375,

July 2011.

[27] Anne Roubaud, Roger Röthlisberger, and Daniel Favrat. “Lean-burn cogeneration

biogas engine with unscavenged combustion prechamber: comparison with natural

gas”. International Journal of Thermodynamics, 5(4):169–175, 2002.

[28] S.O. Bade Shrestha and G. Narayanan. “Landfill gas with hydrogen addition - A

fuel for SI engines”. Fuel, 87(17-18):3616–3626, 2008.

[29] Gabriela Soreanu, Patricia Faletta, Michel Béland, Kara Edmonson, Bert Ventresca,

and Peter Seto. “Empirical modelling and dual-performance optimisation of a hy-

drogen sulphide removal process for biogas treatment”. Bioresource Technology,

101(23):9387–9390, 2010.

[30] James G. Speight. Synthetic Fuels Handbook: Properties, Process, and Performance.

McGraw-Hill, United States, 2008.

[31] Mark Tadrous. “Achieving Low Emissions from a Biogas Fuelled SI Engine Using a

Catalytic Converter”. Master’s thesis, University of Toronto, 2012.

Bibliography 66

[32] N. Tippayawong and P. Thanompongchart. “Biogas quality upgrade by simultaneous

removal of CO2 and H2 S in a packed column reactor”. Energy, 35(12):4531–4535,

2010.

[33] Peter Weiland. “Biogas production: current state and perspectives”. Applied Mi-

crobiology and Biotechnology, 85(4):849–860, 2009.

[34] Andrew J. White, Donald W. Kirk, and John W. Graydon. “Analysis of small-scale

biogas utilization systems on Ontario cattle farms”. Renewable Energy, 36(3):1019–

1025, 2011.

[35] Woodward, 1000 East Drake Road, Fort Collins CO 80525, USA. L-Series Integrated

Speed Control, 26250 edition.

[36] Woodward, 1000 East Drake Road, Fort Collins CO 80525, USA. L-Series ITB

LC-50, 26249 edition.

[37] Woodward, 1000 East Drake Road, Fort Collins CO 80525, USA. L-Series Process

Controller, 26251 edition.

[38] Emmanuel K. Yiridoe, Robert Gordon, and Bettina B. Brown. “Nonmarket coben-

efits and economic feasibility of on-farm biogas energy production”. Energy Policy,

37(3):1170–1179, 2008.

[39] Hu Zhunqing and Xin Zhang. “Experimental study on performance and emissions of

engine fueled with lower heat value gas-hydrogen mixtures”. International Journal

of Hydrogen Energy, 37(1):1080–1083, 2011.

Appendix A

Engine Tuning

A.1 Spark Timing

Spark timing can have a significant e↵ect on engine emissions. Tests are typically con-

ducted with the spark timing set such that minimum timing for best torque, or MBT

is achieved, although that is not optimal from an emissions reduction standpoint. To

illustrate this, Figure A.1 shows the e↵ect that spark advance has on fuel flow. With

the engine producing constant power and operating at a constant equivalence ratio, it

follows that the value of spark advance at which fuel consumption is lowest, corresponds

with the operating condition where the throttle is least open, and thus where volumetric

efficiency is maximized; this is also the MBT spark advance. In Figure A.1, it is seen

that MBT spark advance lies between 30 and 40 CAD BTC. However, the engine’s stock

spark advance is listed as 18 CAD BTC.

In order to determine the engine operating conditions to be used for testing, spark tim-

ing sweeps were conducted with the engine fuelled by both NG (the maximum methane-

content fuel used in testing), and a 60% CH4 -40% CO2 fuel mixture (he lowest methane-

content fuel used in testing). Seven kilowatt and 6 kW loads were applied to the engine,

for both fuel cases, respectively, as these were the maximum loads at which the engine

67
Appendix A. Engine Tuning 68

would run in a stable manner for the respective operating conditions and for all spark

timing settings. The spark advance was set to the maximum possible advance with the

given setup, and retarded in increments of 10 CAD BTC. When stability deteriorated,

the advance was advanced by 5CAD BTC from the setting in place to provide further

information as to the optimal spark timing to be used for the composition being tested.

Spark timing sweeps with the e↵ects on fuel consumption, THC emissions, NOx emissions

(engine-out emissions, in both cases) and the exhaust temperature at the catalyst are

shown in Figures A.1, A.2, A.3, A.4, A.5, A.6, A.7 and A.8.

In the maximum-methane content fuel case, it can be seen that the minimum fuel

consumption occurs with the spark timing set somewhere between 30 and 40 CAD BTC,

and in the minimum-methane content fuel case, it seems that minimum fuel consumption

occurs with the spark advance set to greater than the maximum setting of 60 CAD BTC

that the system is capable of. The exhaust temperatures in both cases are greater than

500 C, which is well above the temperature required to activate the catalysts, which is

important. Exhaust temperature seems to level o↵ when the spark advance is set to 50

CAD BTC or more when the engine is fuelled by NG. When fueled by NG, the engine

stability deteriorated with the spark advance set to 20 CAD BTC or less. When fueled

by the minimum methane-content fuel, engine stability deteriorated with the advance

set to 45 CAD BTC or less. In both cases, both THC and NOx emissions decreased

with less spark advance. Hydrocarbon emissions decrease as spark timing is retarded

(relative to MBT) because the peak in-cylinder pressure is reduced, resulting in less

unburned mixture being forced into crevice regions and exiting the cylinder unburned.

Also, as seen in Figures A.2 and A.6, the exhaust temperatures that are a result of

lower efficiencies associated with not using MBT timing, contribute to better oxidation

rates, further decreasing engine-out THC emissions [10]. NOx emissions decrease with

spark retard mainly because of the lower in-cylinder peak pressures, and thus, peak

temperatures, which the formation of NOx is highly a function of [11].

Appendix A. Engine Tuning 69

As such, it was determined that the optimal spark timing for emissions was 25 CAD

BTC with the engine fueled by NG, and 50 CAD BTC when fueled by the minimum

methane-content fuel. Since these spark timing settings are di↵erent, it was necessary to

conduct spark timing sweeps at all compositions to be used in testing. These sweeps were

performed in the same manner as that described for these two fuel compositions. The

trends with regards to THC and NOx emissions were consistent for all fuel compositions

in between, and as such, the minimum spark advance at which the engine would operate

with stability was selected as the optimal timing for each composition. The result of

these tests is shown in Table A.1.

39
Φ=1.00

38.5

38
Total Fuel Flow (SLM)

37.5

37

36.5

36

35.5

35
15 20 25 30 35 40 45 50 55 60 65
Spark Advance (CAD BTC)

Figure A.1: Spark timing sweep, showing the e↵ect that spark advance has on fuel flow.

Test engine fuelled by NG (7 kW load).

Appendix A. Engine Tuning 70

590
Φ=1.00

580

570
Exhaust Temperature (°C)

560

550

540

530

520

510
15 20 25 30 35 40 45 50 55 60 65
Spark Advance (CAD BTC)

Figure A.2: Spark timing sweep, showing the e↵ect that spark advance has on exhaust

temperature. Test engine fuelled by NG (7 kW load).

1400
Φ=1.00

1300

1200
THC Emissions (ppm)

1100

1000

900

800

700

600
15 20 25 30 35 40 45 50 55 60 65
Spark Advance (CAD BTC)

Figure A.3: Spark timing sweep, showing the e↵ect that spark advance has on THC

emissions. Test engine fuelled by NG (7 kW load).

Appendix A. Engine Tuning 71

2800
Φ=1.00
2600

2400

2200
NO Emissions (ppm)

2000

1800

1600
x

1400

1200

1000

800
15 20 25 30 35 40 45 50 55 60 65
Spark Advance (CAD BTC)

Figure A.4: Spark timing sweep, showing the e↵ect that spark advance has on NOx

emissions. Test engine fuelled by NG (7 kW load).

58
Φ=1.00

57.5
Total Fuel Flow (SLM)

57

56.5

56

55.5
44 46 48 50 52 54 56 58 60 62
Spark Advance (CAD BTC)

Figure A.5: Spark timing sweep, showing the e↵ect that spark advance has on fuel flow.

Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW load).
Appendix A. Engine Tuning 72

575
Φ=1.00

570
Exhaust Temperature (°C)

565

560

555

550

545
44 46 48 50 52 54 56 58 60 62
Spark Advance (CAD BTC)

Figure A.6: Spark timing sweep, showing the e↵ect that spark advance has on exhaust

temperature. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW load).

1200
Φ=1.00

1150

1100
THC Emissions (ppm)

1050

1000

950

900

850
44 46 48 50 52 54 56 58 60 62
Spark Advance (CAD BTC)

Figure A.7: Spark timing sweep, showing the e↵ect that spark advance has on THC

emissions. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW load).
Appendix A. Engine Tuning 73

280
Φ=1.00

260

240
NOx Emissions (ppm)

220

200

180

160
44 46 48 50 52 54 56 58 60 62
Spark Advance (CAD BTC)

Figure A.8: Spark timing sweep, showing the e↵ect that spark advance has on NOx

emissions. Test engine fuelled by a 60% CH4 -40% CO2 fuel mixture (6 kW load).

Table A.1: Optimal spark timing for emissions at di↵erent fuel compositions.

% NG % CO2 Spark Advance (CAD BTC)

100 0 25

80 20 35

75 25 40

70 30 45

65 35 45

60 40 50
Appendix A. Engine Tuning 74

A.2 Optimal Equivalence Ratio for Catalyst Perfor-

mance

As discussed in 4.1, sweeps were conducted with fuel compositions containing the maxi-

mum and minimum methane contents, as with the spark timing sweeps. These tests were

conducted with a constant catalyst space velocity of 12 228 hr-1 , in order to determine the

value of at which the catalyst converted THC, NOx and CO emissions most e↵ectively,

at each di↵erent fuel composition. The results, shown in Figures A.9, A.10, A.11, A.12,

A.13 and A.14, show that a value of 1.013 is optimal for catalyst performance. Since

the compositions tested represent the maximum and minimum methane content fuels

used for all tests, it was inferred that this value was optimal for all fuel compositions

used in other tests.

55
Catalyst Space Velocity = 12 228 hr−1
50

45
THC Emissions Reduction (%)

40

35

30

25

20

15

10
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.9: Reduction of THC emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 7 kW and with NG as the fuel.

Appendix A. Engine Tuning 75

100
Catalyst Space Velocity = 12 228 hr−1
90

NOx Emissions Reduction (%) 80

70

60

50

40

30

20

10

0
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.10: Reduction of NOx emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 7 kW and with NG as the fuel.

100
Catalyst Space Velocity = 12 228 hr−1
95

90
CO Emissions Reduction (%)

85

80

75

70

65

60

55

50
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.11: Reduction of CO emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 7 kW and with NG as the fuel.

Appendix A. Engine Tuning 76

80
−1
Catalyst Space Velocity = 12 228 hr

70

THC Emissions Reduction (%)

60

50

40

30

20

10
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.12: Reduction of THC emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 6 kW and with a mixture consisting of 60% NG and

40% CO2 as the fuel.

120
Catalyst Space Velocity = 12 228 hr−1

100
NOx Emissions Reduction (%)

80

60

40

20

−20
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.13: Reduction of NOx emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 6 kW and with a mixture consisting of 60% NG and

40% CO2 as the fuel.

Appendix A. Engine Tuning 77

105
Catalyst Space Velocity = 12 228 hr−1
100

95
CO Emissions Reduction (%)

90

85

80

75

70

65

60
0.9950 1.0000 1.0050 1.0100 1.0150 1.0200 1.0250
Φ

Figure A.14: Reduction of CO emissions across the catalyst at di↵erent values. Sweep

conducted with a constant load of 6 kW and with a mixture consisting of 60% NG and

40% CO2 as the fuel.