ROYAL INSTITUTE OF TECHNOLOGY

Study of gas fuel jet burning in low oxygen

content and high temperature oxidizer

Magnus Mörtberg

Doctoral Dissertation

Stockholm 2005

Royal Institute of Technology School of Industrial Engineering and Management Department of Materials Science and Engineering

Division of Energy and Furnace Technology SE–100 44 Stockholm

Sweden

Akademisk avhandling som med tillstånd av Kungliga Tekniska Högskolan i Stockholm framlägges för offentlig granskning för avläggande av teknologie doktorsexamen fredagen den 27 maj 2005, kl. 14.00 i sal D3, Lindstedtsvägen 5, Kungliga Tekniska Högskolan, Stockholm.

ISRN KTH/MSE--05/26--SE+ENERGY/AVH

ISBN 91-7178-043-2

ii

Magnus Mörtberg. Study of gas fuel jet burning in low oxygen content and high temperature oxidizer

Royal Institute of Technology School of Industrial Engineering and Management Department of Materials Science and Engineering Division of Energy and Furnace Technology SE–100 44 Stockholm Sweden ISRN KTH/MSE--05/26--SE+ENERGY/AVH ISBN 91-7178-043-2 © the author

iii

Abstract

iv

v

During the past decade, new advanced combustion systems that share the same basic concept of using a substantially diluted and high-temperature oxidizer in the reaction volume have gained a great deal of interest regarding their application in industrial and power systems. These novel combustion technologies have proved to offer significant benefits compared to traditional combustion techniques. These benefits include reductions in pollutant emissions and energy consumption, as well as a higher and more uniformly distributed heat flux. This entails the potential to, for example, reduce the size of equipment in industrial units or increase production rates while fuel consumption and the subsequent CO2 emissions are decreased or maintained at the same level.

Although the development of these new combustion technologies has occurred fairly recently, it has gained worldwide recognition. During the past few years the technique has been used commercially with several different types of burners. Despite its widespread use, the basic understanding of the chemical-physical phenomena involved is limited, and a better understanding of the combustion phenomena is required for more effective utilization of the technology.

The objectives of this work have been to obtain fuel-jet characteristics in combustion under high-temperature, low-oxygen conditions and to develop some theoretical considerations of the phenomena. The effect of the preheat temperature of the combustion air, combustion stoichiometry and the fuel-jet calorific value on flame behavior was investigated. Temperature and heat-flux distribution were also studied using a semi-industrial test furnace to see if similar flame features would be found for the small- and large-scale experiments.

Particle Image Velocimetry (PIV) was used for the first time to obtain information on the flow dynamics of a fuel jet injected into a crossflow of oxidizer at either a normal temperature or a very high temperature. Light emission spectroscopy was used to collect information on time-averaged radical distributions in the combustion jet.

Jet turbulence, time-averaged velocity distribution, fuel-jet mixing, the distribution of radicals such as CH, OH and C2, and flame photographs were investigated. The results showed delayed mixing and combustion under high-temperature low-oxygen-concentration conditions. The combustion air preheat temperature and oxygen concentration were found to have a significant effect on the burning fuel-jet behavior. The results of the semi-industrial-scale tests also showed the features of even flame temperature and heat flux.

vi

vii

Acknowledgements

viii

ix

Many thanks to my supervisor Professor W. Blasiak for providing me with the opportunity to complete significant parts of this thesis abroad and his scientific support and guidance throughout the years as a student in his group.

Professor A.K. Gupta is acknowledged for his excellent supervision in the combustion science. Scientific support and guidance, always seeing the possible when I saw the impossible, thus pushing me to strive further for a successful accomplishment of data collection and interpretation. Further also for accepting me as a visiting graduate student in his laboratory at the University of Maryland.

Doctor Yang Weihong is great fully acknowledged for many interesting discussion, both in the sciences and in many other topics as well.

My colleagues at the energy and furnace division are acknowledged for the scientific and cultural exchange that is taking place all the time.

The financial support to the project from the Swedish Energy Agency, STEM, NFK and Jernkontoret are great fully acknowledged.

Mr Göran Andersson and Mrs Birgitta Lindblad at Jernkontoret are acknowledged for providing financial support to the project and for supervision during my time working for TO51 and Jernkontoret.

The colleagues and friends in the combustion lab at University of Maryland with who I drunk countless numbers of beers during the Friday nights at the Cornerstone.

All my friends at the department of materials science and engineering are acknowledged, for many great moments of fun. No one mentioned, hopefully no one forgotten.

My friends outside the academia brought a lot of fun to me when science was not as exciting.

Finally but mostly I would like to express my sincere gratitude to my love and my best friend, my wife Helene who has always been on my side supporting me through happy moments and trough though periods.

x

xi

Papers included in the thesis

xii

xiii

Paper 1: “Combustion of Low Calorific Fuels in High Temperature Oxygen Deficient Environment” Mortberg M., Blasiak W. and Gupta A.K.

Published in J. Industrial Heating, November 2004 (In Japanese) ISRN KTH/MSE—05/27—SE+ENERGY/AVH

Paper 2: “Experimental Investigation of Flow Phenomena of a Single Fuel Jet in Cross-Flow during Highly Preheated Air Combustion Conditions” Mortberg M., Blasiak W., Kitagawa K. and Gupta A.K.

Submitted to ASME J. Eng. for Gas Turbines and Power for publication December 2004 ISRN KTH/MSE—05/28—SE+ENERGY/AVH

Paper 3: “Combustion of Low Calorific Fuels in High Temperature Oxygen Deficient Environment” Mortberg M., Blasiak W. and Gupta A.K.

Accepted for publication, Combustion Science and Technology, March 2005 ISRN KTH/MSE—05/29—SE+ENERGY/AVH

Paper 4: “Lateral Fuel Jet Injection into High Temperature Oxygen Deficient Conditions” Mortberg M., Gupta A.K. and Blasiak W.

Presented at Joint International Conference on Sustainable Energy and environment (SEE) December 1-3, 2004, Hua-Hin, Thailand. ISRN KTH/MSE—05/30—SE+ENERGY/AVH

Paper 5: “Measurements of temperature, heat flux and flue gas composition in HTAC flame” Mortberg M., Rafidi N., and Blasiak W. Presented at IOM, challenges in reheating furnaces conference, October, 2002, London, United Kingdom. ISRN KTH/MSE—05/31—SE+ENERGY/AVH

xiv

xv

Papers not included in the thesis

xvi

xvii

Mörtberg, M., Blasiak, W. and Gupta, A.K., Flameless combustion of methane fuel jet in transverse air flow, Proceedings of the 3rd International Energy Conservation Engineering Conference (ICECE), San Francisco, CA, USA, 2005. Mörtberg, M., Blasiak, W. and Gupta, A.K., Effect of fuel property on flame characteristics under high temperature air combustion conditions, Proceedings of the 24th International Conference on Incineration and Thermal Treatment Technologies (IT3), Galveston, TX, USA, 2005. Blasiak W., Mörtberg M., Rafidi N., Krishnamurthy N., Yang W., Kalisz S., Szewczyk D., Study and application of high temperature air combustion for gas fuels and wastes, J. Industrial Heating, (in Japanese), Vol 41, No 5, 2004. Blasiak W., Szewczyk D., Mortberg M., Rafidi N., Yang W., Combustion tests in a test furnace equipped with high temperature air combustion mode, J. Industrial Heating (in Japanese), Vol. 41 No. 2, 2004. Yang W., Mortberg M. and Blasiak W., Influences of flame configurations on flame properties and NO emissions in combustion with high-temperature air, Scandinavian Journal of Metallurgy, vol 33, 2004. Mortberg, M., Blasiak, W. and Gupta A. K. Combustion of Low Calorific Fuels in High Temperature Oxygen Deficient Environment, Proceedings of the 23rd International Conference on Incineration and Thermal Treatment Technologies, Phoenix, AZ, U.S.A., 2004. Blasiak W., Mörtberg M., Rafidi N., Krishnamurthy N., Yang W., Kalisz S., Szewczyk D., Study and application of High Temperature Air Combustion for gas fuels and wastes, Proceedings of 14th IFRF Members’ Conference, May, NH Leeuwenhorst Conference Centre, Noordwijkerhout, The Netherlands, 2004. Blasiak W., Szewczyk D., Rafidi N., Mörtberg M., Experimental study of High Temperature Air Combustion (HiTAC) in test furnace equipped with radiant tube High-cycle Regenerative System, Topic Oriented Technical Meetings – TOTeM 26, Ijmuiden Netherlands, 2003.

xviii

Mörtberg M., Blasiak W. and Gupta A. K., Experimental Investigation of Physical Properties of a Single Fuel Jet in Cross-Flow During Highly Preheated Air Combustion Conditions, Topic Oriented Technical Meetings TOTeM 25, Stockholm, Sweden, 2003. Szewczyk D., Mörtberg M., Rafidi E. N., Dobski T., and Blasiak W., Measurements of temperature and heat flux in HiTAC flame for unsteady state condition, 5th International Symposium on High Temperature Air Combustion and Gasification, Yokohama, Japan, 2002. Lille S., Blasiak W., Mörtberg M., Dobski T., Yang W., Heat flux evaluation in a test furnace equipped with high temperature air combustion (HTAC) technology, Proceedings of IJPGC.02, 2002 International Joint Power Generation Conference, Phoenix, AZ, USA, 2002.

xix

1. INTRODUCTION 1 2. STATE OF THE ART 5 2.1. Objectives of Work 11 3. EXPERIMENTAL STUDY 13 3.1. Single jet flame experimental facility 15 3.2. Semi industrial experimental facility 18 4. MEASUREMENT & ANALYSIS 21 4.1. Particle Image Velocimetry 23 4.2. Light Emission Spectroscopy 24 4.3. Direct Flame Photographs 24 4.4. Suction pyrometer and heat flux probe 24 5. EXPERIMENTAL CONDITIONS 27 5.1. Single fuel jet experiments 29 5.2. Semi-industrial experiments 32 6. THEORETICAL ANALYSIS 33 6.1. Fuel jet trajectory 35 6.2. Jet entrainment 36 6.3. Combustion noise 37 6.4. Jet size and shape 39 6.5. Flame temperature 44 7. RESULTS AND DISCUSSION 47 7.1. Jet flame trajectory and jet width 49 7.2. Jet flame entrainment 51 7.3. Combustion noise 52 7.4. Jet flame size and appearance 53 7.5. Distribution of combustion radicals 55 7.6. Jet flame aerodynamics 61 7.7. Superimposition of experimental data 75 7.8. Flame temperature 78 7.9. Semi-industrial HiTAC jet flame study 82 8. CONCLUSIONS 83 9. REFERENCES 87

xx

1

1. INTRODUCTION

2

3

The combustion of fossil fuels provides the vast majority of the world's energy requirements, and is thus also a major source of air-pollutant emissions and greenhouse gases. In 1992 a first initiative was taken by the United Nations to preserve the global environment. This initiative led to the Kyoto Protocol in 1997 which stipulated that developed countries should reduce their total emission of greenhouse gases by at least 5% of 1990 emission levels by the year 2012. The increased concern regarding environmentally harmful pollutants has led to greater emphasis on the development of combustion systems marked by low emission of undesirable pollutants and high energy efficiency. For industrial furnaces, fuel consumption represents a major operational cost and results in large amounts of greenhouse gases being emitted into the atmosphere. Industrial furnaces consume a significant amount of the total energy consumption of fossil fuels and are at the same time characterized by much lower efficiency than, for example, boilers. As a consequence of fossil fuel consumption, the industrial furnace also emits large amounts of CO2, which is one of the important greenhouse gases. Clearly there is an urgent need to reduce furnace fuel consumption, and with that, CO2 emissions, if we are to reduce costs for CO2-trading systems and the environmental impact of the emissions.

Several different combustion technologies have been developed to meet new and more stringent demands on energy consumption, environmental impact and production expenditure. These technologies have been developed by the initiatives of different groups. The underlying concept of all of the technologies is separate injection of the oxidizer and fuel at very high injection velocities which results in a substantial entrainment of reaction products back to the reaction zone, thus leading to a reduced oxygen concentration in the reaction zone and consequently a reduced reaction rate. These novel combustion technologies use combustion with the flame not attached to the burner face, which is opposed to traditional techniques. Moreover, the reaction volume and distribution are controlled by the injection velocity, which is an important factor since it has been proved that it is possible to use pure oxygen as an oxidizer due to the strong entrainment of combustion products back to the reaction zone which renders a very low local oxygen concentration within the reaction volume.

One of the most distinctive features of these new technologies is combustion with a high oxidizer temperature and low oxygen concentration inside the reaction volume. Combustion at the low oxygen concentration forms flames which have different properties compared to conventional air-combustion flames. In practical applications such combustion phenomena are achieved by separated injection of combustion air and fuel into the reaction volume. The regenerative burners, partly studied in this work, inject the combustion air and fuel into the

4

combustion region at very high velocities, causing strong entrainment and thus lowering the oxygen concentration in the combustion region. This has been denoted “oxygen-deficient combustion” as it is characterized by a lower excess oxygen ratio. This yields a uniform and a relatively low flame temperature in the combustion region.

Utilization of these new combustion technologies offers several environmental and economical benefits/advantages compared to traditional combustion technologies [1]:

- reduced fuel consumption and consequently reduced CO2 emissions

- reduced NOX emissions at very low CO levels

- reduced combustion noise

- enhanced temperature uniformity

- enhanced heat transfer

- increased production rates or downsizing of production equipment as a consequence of the above-mentioned benefits

These benefits offer significant economic and environmental incentives when retrofitting or rebuilding industrial heating facilities, which also has resulted in a large number of the systems being implemented in facilities throughout the world. Energy savings in the area of about 30% coupled with CO2-emission reductions have been reported by many industrial organizations that have implemented these new types of combustion systems [2]. These technologies have been adopted mainly in heating furnaces, heat-treatment furnaces, melting furnaces and ladle heaters. With further development, they can also be applied to steam reformers, chemical process heaters, gas turbines, waste incinerators and boilers.

13

3. EXPERIMENTAL STUDY

14

15

In order to obtain information on the phenomena arising from a combusting fuel jet when preheating the oxidizer to high temperatures and reducing its oxygen content, two types of experimental apparatuses were employed equipped with various measurement and analysis instruments.

The single fuel jet was injected into a crossflow of oxidizer composed of air diluted with nitrogen. The oxidizer preheat temperature and oxygen concentration was varied. Several non-intrusive experimental techniques were employed to investigate the phenomena arising under the conditions of high-temperature and low oxygen concentration. Non-intrusive techniques are preferred in experimental investigations because no disturbance is introduced into the investigated system. The oxidizer used in the single-fuel-jet experiments was regular air diluted with nitrogen in order to obtain the correct oxygen concentration for the experiments.

The aim of the work was to complete experimental investigations of a transverse fuel jet under different combustion stoichiometry, oxidizer temperatures and fuel calorific values in order to obtain information on the phenomena which arise when oxidizer temperature and oxygen concentration change. The experimental set-up for the single jet flame allowed for well-controlled conditions and convenient optical access to the studied fuel jet.

Furthermore, semi-industrial-scale jet-flame experiments were performed where temperature distribution and heat flux was studied. The purpose of the semi-industrial study was to determine if the found benefits of the jet-flame in a high-temperature and oxygen-deficient environment also applied on a larger scale, and if the phenomena were similar to what was seen with the small-scale experiments. In studying the most important aspects of the occurring phenomena, a simple analytical approach was taken in order to obtain more widely applicable results from the empirical investigation.

3.1. Single jet flame experimental facility The single-flame experimental apparatus was constructed to permit investigation of the interaction between the fuel jet and the high-temperature and low-oxygen-concentration crossflow under steady-state conditions throughout the experiment. This permitted determining the structure and distribution of flames under high-temperature and low-oxygen conditions.

The single flame experimental apparatus, shown with a photograph in Figure 1 and schematically in Figure 2, is a high-temperature air combustion test apparatus. It consists of two burners, with

16

firing alternating between two different furnace sections (A and B). The furnace sections are equipped with ceramic honeycomb regenerators for thermal storage of the thermal energy before the exhaust gases are released to the environment. Before discharge to the atmosphere, almost all of the thermal energy of the exhaust gas is extracted and used to preheat the oxidizer.

The apparatus generates a continuous mass flow of a highly preheated oxidizer (up to 1400 K) in the test section by passing the oxidizer through ceramic honeycomb regenerators where heat is exchanged to the oxidizer. The composition of the oxidizer, especially the oxygen concentration, is controlled by the addition of nitrogen to the oxidizer. By adding different amounts of nitrogen, any oxygen concentration could be maintained in the test section. The air at ambient temperature is preheated during its passage through the thermal-storage honeycomb bed. The ceramic honeycombs are heated by alternate firing of the burners in the A and B furnace sections. The inlet flow was conditioned by a honeycomb element placed immediately below the test section of the apparatus to ensure a laminar flow was entering the test section.

Every 20 seconds the flow direction was automatically switched, but during the switching sequence no tests were run, and consequently the experiments were performed in steady state. Throughout the experiment there was a temperature drop of 2 K per second in the experimental section.

The flow rates of the oxidizer and fuel were monitored via a flow control panel throughout the experiments. At the panel the oxidizer and the fuel flow through a plug valve into a glass tubular flow meter. The pressure in the flow meter was measured by a pressure gauge. From the panel the flows were directed in to the single jet experimental apparatus. Calibration data for all of the gauges is provided by the manufacturer.

Figure 1. Photograph of the single jet experimental facility

17

Test fuel

Heat up fuel

Combustion air

Flue gases

Dilutant (N2)

4-way switching valve

Pressurized gas

Fluidized bed

Test section

TC

TC

TC

TC

TC

Gas analyzer

Figure 2. Schematic layout of the High Temperature Air Combustion facility

Prior to running an experiment, the air was preheated to a specific temperature. The heating process is illustrated in Figure 3.

In cycle A when burner A is firing, the fuel valve for burner A is opened and fuel is supplied to the burner. At the same time, combustion air flows in through the 4-way switching valve and regenerator B where the combustion air is heated from the heat stored in regenerator B. The combustion in chamber A creates hot furnace gases that flow directly through regenerator A, which stores the heat, and the gases are then released to the ambient air at lower temperatures.

After 20 seconds the system is switched to cycle B. Now burner B is firing and the combustion air flows in through regenerator A. While the air passes through regenerator A, heat is exchanged. Now regenerator A releases heat to the air. Upon exiting regenerator A, the air can have reached temperatures up to 50K below the flame temperature of the fuel gas. The furnace gases flow directly to chamber B and through regenerator B, where heat is exchanged once more. Now the regenerator again stores the heat from the gases and they are released to the ambient air at lower temperatures.

18

By repeating this process it is possible to preheat the incoming air to very high temperatures. A fraction of the combustion air flowing through the burner system bypasses and flows through the test section.

The dimensions of the test section were 200 x 260 mm (in the X and Y directions, respectively). The dimensions of the flow inlet were 95 x 58 mm (in the X and Z directions, respectively). The test section was equipped with quartz glass windows of 4 mm thickness to allow for optical observation of the flame on three sides. Alumina-oxide insulation was used to minimize heat loss. The oxygen concentration and gas temperature were determined shortly before the flow entered the test section. The flow rate of the oxidizer and the fuel was also monitored.

To form a jet-in-crossflow arrangement, the fuel jet to be examined was introduced into the test section normal to the preheated oxidizer via a nozzle 1 mm in diameter. The fuel jet pressure was maintained at 40 kPa throughout the experiment. The different fuel jets used were methane, propane and a simulated low-calorific-value fuel (methane 10 % diluted with nitrogen). The fuel jet was at room temperature while the combustion air in the crossflow was either 298 K, 1173 K or 1373 K.

Cycle A Cycle B

Figure 3. Illustration of test facility cycle

3.2. Semi industrial experimental facility Investigation of a single HiTAC jet flame in semi-industrial conditions was carried out. The semi-industrial test apparatus was equipped with a high-cycle regenerative system (HRS). The burner system was equipped with “honeycomb” regenerators. The burner was able to operate in both

19

conventional flame mode and in HiTAC mode, so comparisons could be made between the conventional jet flame and the flame in HiTAC mode.

The inner dimensions of the furnace body were 3.5x2.2x2.2 m. The test facility was built for a maximum firing rate of 500 kW. The furnace body was insulated with a 0.3 m thick layer of ceramic fiber material.

The inner volume of the combustion chamber was 6.7 m3. There were also a number of openings in the furnace body for observation and obtaining measurements from inside the combustion chamber. The top view of the test equipment with observation and test openings and the experimental burner can be seen in Figure 4.

The heat was removed from the furnace with four horizontal cooling tubes made of a special temperature-resistant kanthal alloy. The tubes were air cooled and enabled a large amount of heat to be removed.

The test furnace was also equipped with two different flue gas channels: one with a water-cooled heat exchanger for hot flue gases and one for cold flue gases from the heat regenerators. The flue gas channel for hot gases allowed the furnace to be run in conventional mode and was also used when the furnace was operated in one-burner mode, when 20% of the flue gases had to be extracted through the hot flue gas channel.

The burner system was a so-called “one-flame system”, which provided a single HiTAC flame in the combustion chamber. The switching time of the burner was constant and set to 10 seconds. The firing capacity of the one-burner system was 200 kW. The burner system was fired with LPG as fuel.

Figure 4. Top view of semi industrial test furnace with traversing system and burner system

20

21

4. MEASUREMENT & ANALYSIS

22

23

In order to attain well-controlled conditions during the experiments, all decisive parameters were monitored. The oxidizer temperature was recorded before and after the thermal storage honeycombs throughout the experimental apparatus and directly below the test section. The final measurement of the oxidizer temperature was carried out directly before the preheated oxidizer entered the test section. In the same cross-section of the apparatus, before the flow entered the test section, the composition of the oxidizer was also determined with a Horiba PG-250 gas analyzer. The gas analyzer was a portable stack gas analyzer that can simultaneously measure up to five separate gas components. Composition data was obtained on the concentration of oxygen, carbon oxide, carbon dioxide and nitrous oxides. This facilitated adjustment of the flow rates of the oxidizer and nitrogen diluent in order to achieve the desired temperature, flow rate and oxygen concentration of the oxidizer. The mass flow rates of the combustion air, the preheat fuel and the test fuel were monitored with flow meters located before the flow entered the experimental apparatus in order to achieve stable and reproducible conditions for the experiments conducted. Thermocouples were placed on both sides of the regenerators to allow monitoring of oxidizer temperature in the test apparatus.

After obtaining stable experimental conditions, several different experimental techniques were employed to investigate the high-temperature air combustion phenomena briefly described below.

Measurement and analysis details are presented in appendix A.

4.1. Particle Image Velocimetry Particle Image Velocimetry (PIV) is one of the most important achievements in experimental fluid dynamics as it allows non-intrusive flow analysis. Particle Image Velocimetry (PIV) is a technique providing an instantaneous flow field image in a cross section of a flow.

The basic principle of PIV is simple. The fluid motion is traced by the addition of micron-size particles to the flow which is then illuminated by a laser light sheet with two instantaneous pulses. The displacement of the particles is calculated and the velocity field determined.

Compared to other techniques for flow-field determination such as pitot tubes, hot wire anemometry, or LDV, PIV offers the benefits outlined below.

24

• Non-intrusive velocity measurement. No probe is present in the flow, thus the flow remains undisturbed. PIV also enables indirect determination of fluid motion since the velocity of the tracer particles is measured.

• PIV is a whole-field technique which allows for simultaneous determination of the velocities at a large number of points, while other techniques only permit measurement at a single point.

• Statistical determination of the velocity at each point as several hundreds of images are collected and gives a statistical correlation of the velocity distribution.

These advantages mean more reliable data in comparison to velocity data obtained by intrusive methods.

4.2. Light Emission Spectroscopy Light emission spectroscopy was used to analyze the spontaneous emission of intermediate species generated in the flame from the combustion. With light emission spectroscopy it was possible to determine the spatial distribution of different flame radicals, such as OH, CH and C2. This was used to provide information on the distribution in the reaction zone and heat release within the flame.

The investigated area was 130.0 * 156.0 mm (X * Y) for the high-temperature oxygen-deficient conditions (at 1173 K and 1373 K) and 130.0 * 78.0 mm (X * Y) for the normal-temperature (298 K) combustion air conditions.

4.3. Direct Flame Photographs Direct flame photographs were captured with a digital camera to obtain time-averaged flame characteristics such as flame color, size, and volume. The digital camera was equipped with a CMOS sensor with an active pixel array of 3072 x 2048 pixels. Several different exposure times were used during the experiments in order to obtain flame images of suitable quality. The exposure time under the direct flame photograph experiments was 8 seconds for the LCV fuel jet under high-temperature conditions and 1/8 second for the methane and propane fuel jets. The different exposure times were chosen with respect to the different luminosity of the different flames under the changed conditions. The time-averaged photographs provide an estimation of the flame structure over time, although the photographs give no information about the dynamic features of the flame which is why PIV was used in this study.

4.4. Suction pyrometer and heat flux probe Water-cooled measurement probes were inserted at various points inside the furnace through 13 openings in the furnace roof as shown in Error! Reference source not found.. Each probe was mounted at the top of the furnace and moved by a computer-controlled traversing system.

For measurement of the flue-gas temperature, a suction pyrometer was used. The suction pyrometer was made of regular stainless steel with a ceramic tip shielding from thermal radiation. Inside the shield, a regular thermocouple of type S was mounted. The thermocouple was placed 150 mm from the tip of the ceramic shield. The flue gases were sucked through the ceramic shield at a velocity of 140 m/s. The high velocity coupled with the radiation shields surrounding

25

the thermocouple ensure convective heat transfer from the hot furnace gas to the thermocouple tip and results in highly accurate temperature measurement. [62]

The total heat flux was recorded with a water-cooled probe with a steel plug in its tip. The temperature gradient was measured between the surface of the steel plug and the water-cooled surface and the total heat flux was calculated [63].

26

27

5. EXPERIMENTAL CONDITIONS

28

29

5.1. Single fuel jet experiments The experiments were carried out under several different conditions in order to investigate the combustion phenomena arising under high-temperature and low-oxygen conditions. Investigations were conducted in order to investigate the influence of the:

• oxidizer preheat temperature [varying from 298 K to 1373 K]

• oxidizer oxygen concentration [varying from 21 % O2 to 2 % O2]

• fuel-jet calorific value [varying from 1023 to 21800 kcal/m3]

The experiments were performed at ambient temperature, 298 K, under either reacting or non-reacting conditions, which allowed for basic investigation of the influence of the combustion process on a normal fuel jet. The elevated temperature levels used in the experiments were 1173 and 1373 K with reacting fuel jets. These temperatures facilitated investigation of combustion phenomena under high-temperature and low-oxygen-concentration conditions.

The influence of the oxidizer oxygen concentration was also considered for the elevated temperatures. The oxygen concentration was controlled by mixing a diluent with the combustion air. The diluent utilized throughout the experiments was nitrogen. The investigated cases used oxidizer oxygen concentrations of 21 %, 15 %, 10 %, 5 % and 2 %.

Several different gaseous fuels were investigated. These were propane (commercial grade >99 %), methane (commercial grade >99 %) and a low calorific value fuel, which consisted of 11.9% methane diluted with nitrogen, used in order to simulate the calorific value of a low-grade fuel gas from pyrolysis processes of waste, biolfuels etc. A compilation of the different experimental cases can be seen in Table 1.

The momentum flux ratio between the gas jet and crossflow was maintained constant when examining different gaseous fuel jets and crossflows, thus providing similarity in mixing for different test conditions. When the momentum flux ratio is maintained constant throughout the experiment, the mixing and flow patterns between the jet fluid and crossflow fluid should be the same. Differences in velocity, turbulence and other measured quantities can then be related to the phenomena originating from the combustion process itself and the variation in parameters in the different experiments.

30

The momentum flux ratio, r, was calculated by:

2

2

∞∞ ⋅

⋅=

VU

r jj

ρρ

Equation 1

where, Uj and V∞ represent the inlet fuel-jet velocity and crossflow velocity, respectively, and ρj and ρ∞ the density of the gas jet and the crossflow, respectively. In addition, the velocity ratio between the gas jet and crossflow is also important. The velocity ratio,λ, sometimes also denoted the Patrick number (which is the inverse of momentum flux ratio number), was calculated as:

2

22

jj UV∞∞ ⋅=

ρρλ Equation 2

The equivalence ratio is calculated by dividing the stoichiometric oxygen concentration of the combustion air by its actual oxygen concentration:

actual

tricstoichiome

OOratioeEquivalenc

%%

2

2= Equation 3

To calculate the stoichiometric oxygen concentration, the reaction between propane and oxygen is used.

31

Table 1. Experimental cases

Oxidizer temperature

[K] Fuel jet caseEquivalance

ratiojet Flowrate

[l/h]Jet Velocity

U0 [m/s]

Oxidier flowrate [m3/s]

oxygen concentration

[%]

Cross flow Velocity V0

[m/s]Momentum flux

ratioPatrick

number λ2

298 C3H8 non-reacting - 42,0 14,9 16,0 21,0 0,92 19,7 0,0025298 C3H8 reacting 0,079 42,0 14,9 16,0 21,0 0,92 19,7 0,0025

1173 C3H8 reacting 0,079 25,0 8,8 16,0 21,0 0,92 19,8 0,00251173 C3H8 reacting 0,11 25,0 8,8 16,0 15,1 0,92 19,8 0,00251173 C3H8 reacting 0,21 25,0 8,8 16,0 9,8 0,92 19,8 0,00251173 C3H8 reacting 0,33 25,0 8,8 16,0 5,0 0,92 19,8 0,00251173 C3H8 reacting 0,83 25,0 8,8 16,0 2,0 0,92 19,8 0,0025

1373 C3H8 reacting 0,079 20,0 7,1 16,0 21,0 0,92 20,2 0,00251373 C3H8 reacting 0,11 20,0 7,1 16,0 15,1 0,92 20,2 0,00251373 C3H8 reacting 0,21 20,0 7,1 16,0 9,8 0,92 20,2 0,00251373 C3H8 reacting 0,33 20,0 7,1 16,0 5,0 0,92 20,2 0,00251373 C3H8 reacting 0,83 20,0 7,1 16,0 2,0 0,92 20,2 0,0025

298 CH4 non-reacting - 70,0 24,8 16,0 21,0 0,92 20,1 0,0025298 CH4 reacting 0,03 70,0 24,8 16,0 21,0 0,92 20,1 0,0025

1173 CH4 reacting 0,03 42,0 14,9 16,0 21,0 0,92 20,4 0,00241173 CH4 reacting 0,04 42,0 14,9 16,0 15,1 0,92 20,4 0,00241173 CH4 reacting 0,07 42,0 14,9 16,0 9,8 0,92 20,4 0,00241173 CH4 reacting 0,13 42,0 14,9 16,0 5,0 0,92 20,4 0,00241173 CH4 reacting 0,3 42,0 14,9 16,0 2,0 0,92 20,4 0,0024

1373 CH4 reacting 0,03 33,0 11,7 16,0 21,0 0,92 20,4 0,00241373 CH4 reacting 0,04 33,0 11,7 16,0 15,1 0,92 20,4 0,00241373 CH4 reacting 0,07 33,0 11,7 16,0 9,8 0,92 20,4 0,00241373 CH4 reacting 0,13 33,0 11,7 16,0 5,0 0,92 20,4 0,00241373 CH4 reacting 0,3 33,0 11,7 16,0 2,0 0,92 20,4 0,0024

298 LCV non-reacting 58,0 20,5 16,6 8,1 1,00 22,4 0,00201173 LCV reacting 32,5 11,5 16,0 8,1 0,92 22,0 0,0021

32

5.2. Semi-industrial experiments All semi-industrial-scale measurements were performed under two different furnace operational conditions. The two compared cases were conventional combustion and High-Temperature Air Combustion. A one-flame regenerative burner system was used.

When the furnace was operated in one-flame regenerative burner mode, 20 % of the flue gas had to be extracted through the hot-flue-gas channel. In the conventional mode, 100 % of the flue gas exited through this channel.

The furnace temperature during the experiments was 1273 K.

Table 2 Configuration of semi industrial test stand

Fuel flow rate

(Nm3/h)

Fuel type

Combustion air flow rate

(Nm3/h)

Furnace Temperature (K)

Combustion air temp (K)

Conventional 7.7 LPG 210 1273 293

One-flame HTAC 7.7 LPG 210 1273 1213

33

6. THEORETICAL ANALYSIS

34

35

The changed properties of the flame under high-temperature and low-oxygen-concentration combustion conditions indicate that oxidizer temperature, oxygen concentration and fuel type have a significant impact on the fluid dynamics, which affects the fuel and oxidizer mixing. In order to enhance our fundamental understanding of the phenomena, theoretical assessments had to be done. This enabled prediction of experimental results. Some of the most important parameters evaluated are covered below. The theoretical considerations also explain some of the behavior of high-temperature low-oxygen combustion phenomena.

6.1. Fuel jet trajectory The transverse jet has been extensively investigated over the years, due to the well-documented positive effects on jet mixing when the jet is injected into a transverse air stream. Pratte and Baines [56] reported that jets injected into transverse air streams easily could be normalized with the length scale rd and the trajectory for jets with different momentum flux ratios could be described with the following equation:

B

rdxA

rdy

⎟⎠⎞

⎜⎝⎛=

Equation 4

In this equation x is the radial distance from the jet nozzle in the jet exit direction, and y is the axial distance in the crossflow direction from the fuel-jet nozzle and d is the fuel-jet diameter. A and B are experimental constants. Several different values for A and B have been reported, generally ranging between 1.5-2.5 for A and 0.25-0.38 for B [57]. In this study, A was found to be 2.1 and B to be 0.38. The constants are affected by several different factors such as fuel-jet and crossflow properties and turbulence.

36

6.2. Jet entrainment The noticeable change in the flame appearance, reported by several other researchers, indicates a slower mixing of the reactants and a reduced reaction rate. With this it is obvious that entrainment of the crossflow fluid into the fuel jet is of significant importance. Hasselbrink and Mungal [58] derived a correlation of the mass entrained by a transverse jet by using the centerline trajectory information. Their work was based upon calculation of the momentum of the crossflow required to bend the fuel jet injected laterally into the crossflow. The time-averaged conservation principles for the mass and the momentum in the x and y directions are:

mmm j &&& =+ ∞ Equation 5

xumum && =∞∞ Equation 6

yjj umum && = Equation 7

In using momentum equations, the pressure term is to be neglected when the momentum flux ratio is much greater than one. In this work’s experimental cases the momentum flux ratio was twenty. Furthermore, the effect of buoyancy was neglected in the calculation of the momentum flux ratio term, which means that the equation is only valid for large Froude numbers. In these experimental cases the Froude number is much larger than one.

At a large distance from the jet nozzle, where y>rd one can assume that the jet width is proportional to its height from the jet exit level, yc1=δ , the mass flux can then be expressed by:

2214

ycum πρ∞=& Equation 8

By combining the equations, the mass flux problem can be described with:

j

jj

jjj uu

mm

ycd

mm

uu

mm ∞

∞

∞ =⎥⎥

⎦

⎤

⎢⎢

⎣

⎡

⎟⎟⎠

⎞⎜⎜⎝

⎛−⎟

⎟⎠

⎞⎜⎜⎝

⎛−

2

221

2

&

&

&

&

&

&

ρρ

Equation 9

When the trajectory equation is coupled with the mass flux equation it can be written:

x

yBBB

uu

rdAyAB

rdxAB

dxdy

=⎟⎠⎞

⎜⎝⎛=⎟

⎠⎞

⎜⎝⎛=

−−

11

Equation 10

37

which is the ratio of the velocity components. According to Hasselbrink and Mungal [58], the mass flux, thus the mass entrainment, can be written as:

B

jj

ej

j rdx

ABr

mmm

mm −

∞ ⎟⎠⎞

⎜⎝⎛+=

+=

1

1ρρ

&

&&

&

& Equation 11

where jm& is the initial fuel-jet mass-flow rate and em& the mass-flow rate entrained from the crossflow stream.

6.3. Combustion noise The issue of combustion noise has become of greater importance over the years. The reduced turbulence in the high-temperature flames suggests reduced levels of combustion-generated noise. The noise is associated with pressure pulses generated by fluctuation and movement of the flame front. According to a study by Gaydon and Wolfhard [59], the intensity, I, of sound is expressed as a function of the pressure variation ΔP:

cPIρ

2Δ

= Equation 12

where, ρ is the density and c the sound velocity. The pressure difference is generated from expansion of the gas due to heat release from the combustion process. For theoretical consideration of the issue, a simplified approach was chosen and the sound intensity was calculated at distance d from a single point in the flow field. The pressure pulsation originating from the combustion process was then calculated as:

⎟⎠⎞

⎜⎝⎛⋅=Δ

dtdV

dtd

dP

πρ

4 Equation 13

where dV/dt is the rate of change of the volume in the considered volume element. The rate of change of the volume is proportional to the combustion intensity. Assuming spherical expansion, the rate of change of the source volume is defined by:

⎟⎟⎠

⎞⎜⎜⎝

⎛ −⋅=

V

V

EE

dtdrr

dtdV 1

4 2π Equation 14

where EV is the volumetric expansion ratio of the combusted gas compared to the volume of the non-combusted reactants. The pressure fluctuation can be calculated by:

38

( )⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

⋅+⎟⎠⎞

⎜⎝⎛−

=

=⎪⎭

⎪⎬⎫

⎪⎩

⎪⎨⎧

⎟⎟⎠

⎞⎜⎜⎝

⎛ −⋅⋅=Δ

2

22

2

2

21

144

dtrdr

dtdrr

dEE

EE

dtdrr

dtdr

dP

V

V

V

V

ρ

ππρ

Equation 15

During steady burning, the flame travels at a uniform speed Vu ESdtdr = , were Su is the burning velocity. For a steady-state combustion process, the burning velocity becomes more or less constant and the second derivative, 022 =dtrd , describing fluctuation of the radius of the expanding volume element approaches and becomes zero. Building on these assumptions, Equation 11 becomes:

( ) 212uVV rSEE

dP −⎟

⎠⎞

⎜⎝⎛=Δ

ρ

Equation 16

To estimate the pressure fluctuation, the burning velocity has to be estimated. Chomiak [60] presented a study of the burning velocity. Analysis of the burning velocity was based on the magnitude of the thickness of the combustion region δl and the characteristic time τ of the reaction. δl is estimated by dimensional analysis by:

τδ al = Equation 17

where a is the thermal diffusivity of the mixture, calculated by:

ρλ

pCa =

Equation 18

where, λ is the thermal conductivity, Cp the specific heat and ρ the density. The characteristic time for a reaction is related to the reaction rate w, i.e.,

wρτ =

Equation 19

The reaction rate, w, is based on the changes in concentration of the different species in the combustion reaction by:

39

2

2OHC

OHC CCkdtC

dtC

wyx

yx ⋅⋅=−

=−

= Equation 20

where yxHCC and

2OC denote the concentration of propane and oxygen, respectively, and k is

the reaction rate constant. The reaction rate constant is calculated using the Arrhenius formula:

⎟⎠⎞⎜

⎝⎛−⋅= RT

EAk aa exp Equation 21

where Aa is a constant, Ea is the activation energy, R the universal gas constant and T the temperature. In evaluating the reaction-rate calculation, the results show the reaction rate to increase significantly with an increase in temperature T. A reduced oxygen concentration reduces the reaction rate. With the calculated information, the burning velocity Su can be calculated assuming that during a short time τ, the combustion will propagate the distance δl , i.e.,

wC

~Sp

lu

λρτ

δ 1=

Equation 22

Now the pressure fluctuation and the theoretical sound intensity can be determined. Regarding the burning velocity, it can be seen to be dependent on the combustion air temperature and oxygen concentration. The formulas indicate an increase in burning velocity with an increase in temperature at a maintained oxygen concentration and a decrease in burning velocity with a decrease in oxygen concentration in the oxidizer. Similar results were reported by Gaydon and Wolfhard [59].

6.4. Jet size and shape The empirical PIV and light emission spectroscopy data , as well as the flame photographs, reveal changed fluid-dynamics properties under the high-temperature and low-oxygen conditions. The jet appears to be wider when compared to the normal-temperature cases. In an attempt to theoretically quantify the jet width, some properties of the fluid dynamics of a jet in crossflow had to be determined.

40

Figure 5. Scheme of jet in cross flow.

The jet deflection results from the balance between the jet momentum and the crossflow momentum. First it is assumed that the fuel-jet momentum is much larger than one. The jet centerline is defined as the maximum-velocity line along the jet trajectory. The jet border is the point where the velocity in the jet flow direction rises above the surrounding flow velocity. The force balance, for a separate mass element of the jet can be expressed as:

0=+ cp dFdF Equation 23

A centrifugal force which occurs when a mass element, of area Aj, moves at the velocity Uj along a curve linear trajectory of changing radius Rj, with the angle αj. The centrifugal force is written as:

dsARU

dF jjj

jc α

2

= Equation 24

where

js

jAj

j AddsUA

Uj

∫∫=1

Equation 25

41

The other force in the force balance, dFp, may be expressed as proportional to the dynamic pressure, with Cj as the drag coefficient, ρc, the crossflow density, Uc the crossflow velocity and bj the width of the jet, which can be determined by means of the velocity component normal to the jet axis.

dsbU

CdF jjCC

jP ⋅= αρ 2

2

sin2

Equation 26

Incorporating Equations (24), (25) and (26) into (23) renders:

0sin2

22

2

=+⋅ jj

jjj

CCj A

RU

bU

Cρ

αρ

Equation 27

Considering the momentum conservation principles, it is reasonable to assume that the jet momentum is constant along the y-axis.

42

Figure 6. Scheme of jet trajectory

00000 sinsin jjjjjjjjjj UUAUUA αραρ = Equation 28

To comply with the above condition, and incorporating the following dimensionless relationships,

( )B

db

RdR

U

U

j

j

j

j

jj

CC === ;;0200

2

ρ

ρ Equation 29

the equation is transformed into the following:

01sinsin2 3

0 =+jj

j RBC

αγαπ

Equation 30

43

This uses the fact that:

'tan ydxdy

== α Equation 31

Some more relationships are introduced

( )2

322

1sin1

yy

yyR j

′+

′′=

′′′+

= α Equation 32

The incorporation of these into the equation results in a differential equation to be solved:

ByC

yj

j 30sin

2′⋅

−−=′′ γ

απ Equation 33

or, when solving assuming y’=t

BtC

tj

j 30sin

2⋅−=′ γ

απ Equation 34

In order to solve the differential equation it is necessary to define the relationship between the jet width and the coordinates of the jet axis. To be able to successfully solve the differential equation it should be transformed so it represents changes as a function k which should be inverse in its relation to t ( jtk α1tan1 −== k=1/t=tan-1αj). Furthermore, the k function should not follow

the x-y coordinates, but the jet axis. Using the trigonometric relationships:

dsdx

dxdk

dsdk

kk

dydx

j =+

== ;1

cosα Equation 35

transforms the differential equation into

20 1sin2

kBC

dsdk

j

j

+= γ

απ Equation 36

44

With the boundary conditions s=0, k=k0= 010 tan,0 jkks α−=== , it is the differential

equation of the jet axis. Changes in jet width (bj)N are expressed by adopting an equation of a jet flowing into a stationary medium. The equation was corrected for the influence of the crossflow by an empirical correlation. The equation is:

( ) ⎟⎟⎠

⎞⎜⎜⎝

⎛

+

−+=

−

0

10 cot

1SS

kbB jn

Nj

ααγ Equation 37

With the following parameters defined as follows: ( ) ( )02 SSCbNj += ;

jdsS = ;

jdSS 0

0 = ;

⎟⎠⎞

⎜⎝⎛=

2tan βC

6.5. Flame temperature Estimation of the flame temperature in the experimental jet is possible by measuring the C2 vibrational intensity and using the relationship between the two peaks at 470 nm and 515 nm. The spectral intensity for the electronic transition, or the integrated intensity including all of the related rotational and vibrational transitions, changes depending on the flame temperature T and the number of combustion radicals N. The intensity can be expressed as [33, 34, 35]:

⎟⎟⎠

⎞⎜⎜⎝

⎛ −=

)(/expTQkTENgAhI υ Equation 38

where h is the Plank’s constant, υ the frequency, N the number density, g the statistical weight of the excitation level, A the transition probability for spontaneous emission, E the upper-level energy, k the Boltzmann’s constant, and Q(T) the partition function.

Flame temperature was determined by taking the ratio of C2 emission intensities obtained at two wavelengths (470 and 515 nm). The profiles for the net emission intensities of C2 species were estimated by subtracting the background intensity level (at 490 nm) from the C2 emission intensity level (at 470 or 515 nm). The intensity ratio I1/I2 of the two vibrational bands was taken from [35]. The instrument constant, p, was determined as an average value. The calibration was based on the temperatures measured at two spatial points in the flame by means of an Rh-Pt thermocouple.

∑

∑

⎟⎟⎠

⎞⎜⎜⎝

⎛ −

⎟⎟⎠

⎞⎜⎜⎝

⎛ −

=

j

jjj

i

iii

TQkTE

Av

TQkTEAv

pII

)(/

exp

)(/exp

2

1 Equation 39

45

Using equations (38) and (39) we can derive a simple expression for the intensity ratio I1/I2 in terms of flame temperature T, i.e.:

kTEEKpII

CC/)exp( 515

2470

22

1 −= Equation 40

where K is the ratio of the term νgA/Q(T) (this ratio is almost constant at temperatures below 2500K).

46

47

7. RESULTS AND DISCUSSION

48

49

7.1. Jet flame trajectory and jet width Figure 7 shows information on the trajectories. By using the data extracted from the velocity-field plots it can be seen that preheat temperature, oxygen concentration and fuel calorific value have a minor effect on the jet trajectory. This shows that there is an influence on the jet trajectory of parameters such as oxidizer temperature, oxygen concentration and the fuel jet calorific value and that the trajectory is affected by the momentum flux ratio between the fuel jet and the crossflow. The trajectory is expected to be the same as the momentum flux ratio, thus the force balance between the fuel jet and the crossflow fluid is the same. The PIV data shows that with an ambient-temperature oxidizer, combustion of the gaseous fuel jet causes the jet to penetrate further into the crossflow. This was also reported by Hasselbrink and Mungal [36]. Furthermore, the high-temperature and low-oxygen-concentration conditions that were investigated showed similar tendencies with a stronger penetration of the fuel jet into the crossflow when compared to ambient non-burning conditions. The figure shows that the fluid dynamics, thus the mixing, do not change significantly between the fuel jet and the crossflow under high-temperature low-oxygen conditions when compared to cases of normal temperature and oxygen concentration. The results were as expected as the trajectory of the fuel jet is controlled by the momentum flux ratio according to Equation 2. The effect of oxidizer temperature, thus the oxidizer density, can be seen in Figure 8. The plot reveals that the jet width increases when the oxidizer temperature is reduced, i.e. the oxidizer density is lowered.

50

Figure 7. Fuel jet trajectories at varying oxidizer preheat temperatures and oxygen concentration

Figure 8. Jet width for different fuels and preheat temperatures

51

7.2. Jet flame entrainment The entrainment is illustrated in Figure 9, where the effect of oxidizer temperature and oxygen concentration on the entrainment can be seen. The data shows that within the first 40 nozzle diameters from the fuel jet orifice, the entrainment is similar in all experimental cases. However, it can be clearly seen that when an oxidizer of 21 % oxygen is maintained and the preheat temperature is increased, the entrained mass is smaller. This clearly shows the effect of the reduced flame volume and reaction zone as the ignition is faster and the reaction more intense at high preheat temperatures with an oxygen concentration of 21 %. The reaction is also more intense under high oxidizer temperatures with 21% oxygen in the oxidizer. The intense combustion reaction causes strong thermal expansion of the reaction volume which hampers entrainment of the surrounding crossflow fluid. When the oxygen concentration is reduced and a high preheat temperature maintained, the mass entrainment is increased. This is due to the increased reaction volume that the high-temperature low-oxygen-concentration conditions render, which is created by the less intense reaction in the flame volume. The increased entrainment of the crossflow fluid is also necessary in order to complete the combustion reaction due to the low oxygen concentration. From the theoretical considerations it can be concluded that the temperature difference between the oxidizer and the fuel-jet fluid has a significant influence on the entrainment as the density of the different fluids is affected by the temperature. A higher temperature difference renders stronger entrainment of the crossflow fluid into the fuel jet. However, the theory does not explain why the entrainment again increases when the oxidizer oxygen concentration is reduced. It can be concluded from the results that the oxygen concentration has a stronger influence on the entrainment than the temperature does, and that the reduced amount of available oxygen hampers the combustion reaction which in turn leads to the stronger entrainment.

52

Figure 9. Entrainment of cross flow oxidizer into a propane fuel jet

7.3. Combustion noise The relative sound pressure levels are illustrated in Figure 10. Here the effects of the high-temperature low-oxygen-concentration conditions can easily be seen. When the oxygen concentration is reduced and the preheat temperature is increased, the relative sound pressure level is significantly lowered. The reduced combustion noise is a result of the decreased turbulence in the flame under high-temperature oxygen-deficient conditions. The PIV data plotted in Figure 19 and Figure 20 show reduced turbulence levels under these conditions. The reduced turbulence levels are an effect of the laminarization of the flow that occurs when combustion is taking place. The laminarization is primarily an effect of the heat release from the combustion process. The data in Figure 10 presents the relative sound pressure levels for several different experimental cases. It can be seen that the oxygen concentration and oxidizer temperature have a strong hampering effect on the relative sound pressure levels. The high oxidizer preheat temperatures reduce the relative sound pressure levels by 50 % when the oxygen concentration is maintained at 21 %. Although the more intense combustion reaction which could be seen from the flame photographs in Figure 11 and Figure 12 and in the spatial data of in-flame generated radicals in Figure 13 and Figure 14, the laminarization of the flow has a strong effect on the relative sound pressure levels, which are significantly reduced. Moreover, when the oxygen concentration is reduced to 5 %, the relative sound pressure levels drop further. This is an effect of the lower combustion intensity as the reaction extends over a larger volume.

When considering the burning velocity (Equation 22), it can be seen to be dependent on combustion air temperature and oxygen concentration, which explains why the relative sound pressure levels are reduced. The formulas indicate an increase in burning velocity with an increase

53

in temperature at a maintained oxygen concentration and a decrease in burning velocity with a decrease in the oxygen concentration of the oxidizer. Similar results were reported by Gaydon and Wolfhard [59].

0

1

2

3

4

5

6

7

8

Rel

ativ

e so

und

pres

sure

leve

ls, S

PL

(dB

)

a b c d eExperimental case

Figure 10. Relative sound pressure levels (in dB) from the jet flames: (a) 21% oxygen

concentration at 298K; (b) 21% oxygen concentration at 1173K; (c) 5% oxygen concentration at 1173K; (d) 21% oxygen concentration at 1373K; and (e) 5% oxygen concentration at 1373K

7.4. Jet flame size and appearance The direct flame photographs can be seen in Figure 11 and Figure 12. The photographs present the flame appearance under different conditions and for different fuel jets. They also illustrate the effect of oxidizer preheat temperature on flame size and color.

The flame-color information is of interest when studying the combustion chemistry. Flame color has been reported to be significantly dependent on the combustion chemistry and the formation of radicals, which is much more prevalent under high-temperature low-oxygen conditions.

When the oxidizer is preheated, the flame appears to be reduced and more intensive. The photographs indicate that when the preheat temperature is increased from 298 K to a high temperature, the flame shrinks and takes on a yellow-white bright appearance. When the oxygen concentration in the oxidizer is reduced, the flame size increases and the flame color turns to blue. At very low oxygen concentrations, the flame is invisible. The changed flame color shows that a different and less intense combustion reaction takes place under the high-temperature and low-oxygen-concentration conditions. Flame color has been reported to clearly correlate with the generation of radicals during the combustion process in the flame, which in the present case is much less intense. Flame color is highly dependent on the diluent in the oxidizer used to lower the oxygen concentration. Several researchers have reported a green flame under high-

54

temperature and low-oxygen conditions. The green flame color was not observed in the experiments conducted in this work. Nitrogen was used as the diluent in this investigation, which explains the blue flame color at low oxygen concentrations. Other researchers have used carbon dioxide or flue gas to lower the oxygen concentration, which affects the combustion chemistry slightly, thus the changed flame color.

The images obtained with light emission spectroscopy can be seen in Figure 11, Figure 13 and Figure 14. Further results are also provided in Paper 1 and Paper 3. The intensity plots show the time-averaged spatial distribution of flame-generated radicals. The data presented gives information on the distribution and intensity of light emission from the combustion-generated species. The spatial intensity plots illustrate the relative intensity of the emitted light from the radicals, where red indicates the strongest intensity and black indicates a zero-emission intensity on the current wavelength. The spatial distribution of radicals gives information on time-averaged changes in the radicals light emission in the jet flame as the air preheat temperature and oxygen content in the combustion air changes. The selected species included in the determined spatial distributions were OH, CH and C2 which corresponds to wavelengths of 307 nm, 340 nm and 470 nm, respectively. The OH radical serves as the flame marker and provides an estimation of the flame borders. The CH radical indicates the intensity of the time-averaged heat release and the C2 radical provides information on the intensity of the combustion process.

Strong dependence on combustion air temperature and oxygen concentration of the radical species can be seen in the spatial distributions for the different experimental cases. When the oxygen concentration is reduced and the combustion-air preheat temperature is increased from 298 K to 1173 K and 1373 K, there is a dramatic change in the distribution of the different radicals. The data shows a significantly enlarged reaction zone under the high-temperature low-oxygen conditions. This can be explained by the low oxygen concentration of the oxidizer, which reduces the combustion rate. Due to the limited availability of oxygen molecules in the reaction volume, different flame radicals are generated as the combustion reaction can not be completed immediately. This explains the elevated levels of radicals that can be detected in the high-temperature and low-oxygen-concentration flames. The enlarged reaction volume is also one of the characteristic features of high-temperature and low-oxygen combustion.

The ignition delay time is the time that elapses from when an element of the fuel-jet fluid emerges from the fuel nozzle to the time the element is ignited. The ignition delay time in this work’s experiments was estimated from the distance between the fuel jet and the ignition point, which can be seen in the photos. Knowing the fuel-jet velocity, the time for the fuel to be ignited can easily be estimated. With an increase in air preheat temperature, the ignition delay is reduced. This is due to the high temperature that the fuel jet is heated to immediately after exiting the fuel nozzle. At high temperatures the fuel is ignited faster when less energy is needed from the combustion process itself to heat the fuel to a temperature above the auto-ignition temperature. Evidence of this are the elevated levels of OH closer to the nozzle along the trajectory, which render a shorter flame stand-off distance at elevated temperatures. The flame stand-off distance has been determined as the distance between the fuel-jet nozzle and the point where the flame is ignited and radicals were detected by light emission spectroscopy. The data clearly indicates a shorter ignition delay time at a higher air preheat temperature. At an oxygen concentration of 21 % in the combustion air, the difference between the different air preheat temperatures was small. At 5% oxygen, the ignition delay was found to be 3 μs at an air preheat temperature of 1173 K.

The ignition delay time was found to be significantly greater for the LCV fuel jet when compared to the methane and propane fuel jets. This can be explained by the lower calorific value of the LCV fuel jet. Having a lower calorific value, the fuel has to be heated to a very high temperature due to the low heating value of the fuel and a high auto-ignition temperature. The LCV fuel jet takes longer time to mix with the oxygen in order to be able to ignite and combust. However, the

55

LCV fuel jet, once ignited, can be considered a well-stirred reactor because of the complete mixing of the reactants which is illustrated in the homogenous reaction zone in the intensity plots.

7.5. Distribution of combustion radicals When studying the C2 distribution within the flame in Figure 13 and the reaction-zone distribution was found to change with the oxygen concentration and air preheat temperature of the combustion air. The reaction zone is defined as the region in the C2 plots where the intensity from the emission is larger than for the background radiation at the current wavelength.

The C2 distribution serves as a reaction-zone marker where the combustion intensity is highest, i.e. where oxidation of the fuel takes place. The different plots of the C2 distribution reveal a significantly larger, but less intense reaction zone in the case of high-temperature air combustion compared to the normal-temperature air case. The maximum reaction intensity location moves further downstream in the jet trajectory under conditions of high air temperature, revealing the ignition delay.

Under high-temperature oxygen-deficient conditions, the reaction zone is significantly enlarged. The strongest intensity of the C2 species clearly shows that combustion reaction is delayed, i.e. it is found at a position further downstream in the trajectory. Furthermore, a more evenly distributed reaction zone can be distinguished as the C2 distribution was found to be significantly larger at the 5 % and 2 % oxygen contents and at a 1373K air preheat temperature. This is opposite the C2 distribution at 21 % for which a small and intense distribution can be observed. This was captured with a short exposure time and a long exposure time was used for high-temperature low-oxygen conditions.

The wider and extended C2 distribution is a characteristic of the high-temperature air combustion. The enlarged C2 distribution indicates a larger reaction zone. This has been noted by several other researchers who have reported spatial distribution data on radicals or direct flame photographs which illustrated significantly enlarged reaction zones. This indicates a reduced reaction rate of the combustion process due to the lowered oxygen content in the oxidizer. This also complies with

wC

Sp

lu

λρτ

δ 1≅≅ Equation 22

that indicates an increase in burning velocity with an increase in temperature at a maintained oxygen concentration, and a decrease in burning velocity with a decrease in oxidizer oxygen concentration. The reduced reaction rate is also what some researchers denote “mild combustion”.

The intensity of the spatial emissions of CH and OH in Figure 13 and Figure 14, is higher in the ignition region of the fuel jet as compared to the intensity further downstream the fuel-jet trajectory. An extended combustion zone can be seen downstream along the trajectory for the cases with high-temperature preheat air. This also indicates that the reaction rate of the fuel jet is reduced because the combustion is occurring over a larger region under high-temperature oxygen-deficient conditions. This is one of the main characteristics of high-temperature air combustion.

56

The CH-radical distribution indicates the intensity of the time-averaged heat release from the flame. In comparing the different cases, it can be seen that not only the reaction zone increases, but also the heat release. The novel combustion technology using a low oxygen concentration in a large reaction volume is well known for its even and high heat release. The effect of the evenly distributed heat release is a consequence of a reduced reaction rate which results in a larger reaction zone. Moreover, the even heat release has a positive effect on the entrainment, as a strong and sudden heat release has a hampering effect on the entrainment when the thermal expansion is substantial due to an intense combustion process.

A secondary or extended reaction zone can be observed in Figure 13 and Figure 14 in the high-temperature low-oxygen environment. The enlarged reaction zone shows that two-stage burning is taking place when the oxygen concentration is reduced. An enlarged reaction zone can be seen when the oxygen concentration is reduced, and a reduced reaction zone when the oxidizer temperature is increased. The extending of the flame stems from the lower oxygen concentration which results in a reduced reaction rate. The extended flame can also be coupled with the higher axial strain rates in the high-temperature low-oxygen environment (detected in the PIV analyses). This can be observed in Figure 17 and Figure 18. The elevated axial strain rates indicate more entrainment of the crossflow fluid into the fuel jet. The stronger entrainment is desirable in the high-temperature and low-oxygen-concentration cases as it facilitates providing the fuel with enough oxygen to complete the combustion process and counterbalances the lower oxygen concentration in the crossflow oxidizer.

The propane fuel jet in Figure 13 exhibits a stronger tendency towards two-stage burning, or a significantly enlarged reaction zone when compared to the methane fuel jet in Figure 14, which can be seen in the C2 distribution. This indicates slower oxidation of the larger propane molecules which take several steps to be broken down and to completely react. This is a characteristic of the high-temperature oxygen-deficient oxidizer and explains the enlarged flames and the more even heat flux distribution which is a characteristic feature of the high temperature and low oxygen concentration combustion.

57

OH

C2

CH

C3H8 CH4

Flame photograph

LCV 1173 K LCV 1273 K

Figure 11. Flame photographs of C3H8 and CH4 at 298 K, 21 % O2 and LCV fuel at 1173 K and 1273 K at 8 % O2 and intensity plots of radicals generated in the flame

58

2% O2

1373K1173K1373K1173K

5% O2

10% O2

15% O2

21% O2

Combustion air diluted with N2

CH4

C3H8

Figure 12. Flame fotographs under 1173 and 1373 K oxidizer preheat temperature and 21 - 2 % oxygen concentration

59

CH OH

C2

1373K1173K1373K 1173K1373K1173K

2% O2

5% O2

10% O2

15% O2

21% O2

C3H8, Combustion air diluted with N2

Figure 13. Light Emission spectroscopy data showing C2, CH and OH specie at 1173 and 1373 K oxidizer preheat temperature and 21 - 2 % oxygen concentration in C3H8 fuel jet flames

60

2% O2

CH OH

C2

1373K1173K1373K 1173K1373K1173K

5% O2

10% O2

15% O2

21% O2

CH4, Combustion air diluted with N2

Figure 14. Light Emission spectroscopy data showing C2, CH and OH specie at 1173 and 1373 K oxidizer preheat temperature and 21 - 2 % oxygen concentration in CH4 fuel jet flames

61

7.6. Jet flame aerodynamics Some examples of the flow-field results for the non-burning and burning cases using propane and methane as fuels are shown in Figure 15 to Figure 22. The color-coded bar above each image gives information on the magnitude of the particular velocity component and the vectors indicate flow direction. The data that the figures are based on was normalized by the jet exit velocity in order to normalize the results with respect to the jet exit velocity. Also, the same momentum flux ratio was used in order to attain similarity in the mixing and jet and crossflow mixing, and thus be able to identify the effects of combustion air preheat temperature and oxygen concentration.

The average velocity distribution, turbulence characteristics, axial strain rate and total strain rate were calculated from the obtained data. The axial strain rate in the direction of the crossflow is calculated by:

dydvS y = Equation 42

where Sy is the axial strain rate and v is the axial velocity in the y direction of the flow. The rotation of the fluid is represented by the vorticity, which is defined as the curl of the velocity vector u:

dydu

dxdv

z −=×∇=Ω u Equation 43

where v is the velocity in the y direction and u is the velocity in the x direction. The vorticity also provides a measure of where the mixing takes place between the jet and the surrounding fluid. The average velocity distribution is defined by the average of the u and v velocities, i.e.:

22AVG vuV +== V Equation 44

The average velocity distribution normalized by the jet exit velocity is presented in Figure 15 and Figure 16. Some results are also presented in Paper 1. The plots give information on the effects of combustion-air preheat temperature and oxygen content on the mean average velocity distribution. The average velocity distribution for a burning jet injected into a crossflow of ambient temperature air of 21% oxygen concentration is presented. The effect of the combustion can be directly seen from the results. The distribution of the jet is wider for the burning case compared to the non-burning case using a normal air temperature. (This is discussed in greater detail in Papers 1 and 2.) The heat release from the combustion process causes thermal expansion of the jet. This can be clearly seen when comparing the burning and the non-burning jets in ambient conditions. When the combustion-air preheat temperature is increased, higher velocities are measured in the jet. This is due to the direct effect of combustion-air properties on fuel-jet penetration. At higher air preheat temperatures, the fuel jet penetrates slightly further into the crossflow. Under high-temperature combustion conditions, the fuel jet shows a stronger expansion compared to normal temperature conditions. This indicates a strong entertainment

62

during the high-temperature air combustion and a stronger expansion as a function of the more intense heat release.

At lower oxygen concentrations, greater amounts of combustion air must be entrained into the fuel jet before complete combustion. At high temperatures combined with a combustion-air oxygen concentration of 21%, the jets with a 1373K preheat air are distinctly smaller than those with a preheat air of 1173K. This correlates well with the flame photographs, showing a smaller flame size with an increase in air preheat temperature for normal air. The smaller expansion is also due to the lower flow rate of fuel, as the momentum flux ratio between the fuel and the crossflow was held constant. The mixing pattern of the different cases indicated similar behavior due to the constant momentum flux ratio used. As the oxygen concentration was changed from 21% to 5% at a constant temperature, the averaged flow field indicates dramatic expansion of the flame. This also correlates well with the flame photographs that illustrated a larger flame volume at a lower oxygen concentration in the combustion air. The effect of oxygen concentration on jet expansion was also expected to have an influence on the flow since a sudden reaction renders faster mixing of the fuel and the crossflow. When the reaction rate is slower, this jet expansion occurs over a larger volume, in both axial and radial directions. The results also imply, on the other hand, a smaller jet expansion for the methane fuel jets, which is due to the less complex fuel molecules which react faster.

The CH intensity distribution plots in Figure 13 and Figure 14 can be coupled with the larger expansion of the average velocity pattern resulting from the fuel jet. This is because of the less intense heat release in the high-temperature and low-oxygen cases which allows a stronger entrainment and thus a more widely distributed jet flow field.

63

(a) (b)

(c) (d)

(e)

Figure 15. Average velocity distribution normalized by jet exit velocity at (a): 298K, 21% O2combustion air, C3H8 burning; (b): 1173K, 21% O2 combustion air, C3H8 burning; (c): 1173K, 5% O2 combustion air, C3H8 burning; (d): 1373K, 21% O2 combustion air, C3H8 burning; (e): 1373K, 5% O2 combustion air, C3H8 burning

64

(a) (b)

(c) (d)

(e)

Figure 16. Average velocity distribution normalized by jet exit velocity at (a): 298K, 21% O2combustion air, CH4 burning; (b): 1173K, 21% O2 combustion air, CH4 burning; (c): 1173K, 5% O2 combustion air, CH4 burning; (d): 1373K, 21% O2 combustion air, CH4 burning; (e): 1373K, 5% O2 combustion air, CH4 burning

65

The axial strain rate distributions normalized by the jet exit velocity for the different cases are illustrated in Figure 17 and figure 18. The axial strain rate is correlated to the acceleration or deceleration of a finite volume element. The acceleration of the flow in the axial (v or y) direction can also be determined (in the case of crossflow) from the entrainment of crossflow fluid into the fuel jet. The deceleration of the flow upstream of the fuel jet can be interpreted as a wake immediately upstream of the fuel-jet path where a portion of the crossflow is entrained back into the fuel jet.

The experimental results presented in Figure 17 show that for the ambient temperature cases, the combustion process is strongly dependent on the entrainment. It can be seen that the combustion process increases the entrainment of the crossflow into the jet path. The magnitude of the axial strain rate during high-temperature conditions was higher compared to the normal-temperature air cases. This also indicates dependency on the combustion air temperature. For a 21% oxygen combustion air at higher air preheat temperatures, the reaction is more intense from the stronger entrainment and axial strain rate. The extended reaction zone, which is a distinct feature of the low oxygen concentration, can be clearly distinguished from the axial strain rate distribution data when the oxygen content is reduced to 5 %. This trend can be observed for all different fuels examined in this work. The results indicate an extended zone of elevated axial strain rates. This shows the influence of the oxygen content, as larger volumes of the crossflow have to be entrained into the fuel-jet region before the fuel can react completely. The axial strain rate distribution shows similar dependency on the temperature of the combustion air. The axial strain rate levels are higher at higher temperatures, although the entrainment is distributed quite similarly when the oxygen content is kept constant and the air preheat temperature is changed. The oxygen content shows a larger influence on the entrainment. At a low oxygen concentration, the reaction zone is enlarged. Thus a stronger entrainment of combustion air into the reaction zone is required in order to combust the fuel jet completely. The influence of the fuel jet fluid on the axial strain rate can also be seen in Figure 17 and Figure 18. The propane fuel jet shows a more widespread axial strain rate distribution, which indicates that more of the crossflow fluid is entrained in order to complete the combustion process.

The results provided in Paper 3 are for the LCV and methane fuel jets. For the normal air case, higher strain rates were obtained for the LCV fuel gas jet compared to the methane fuel gas jet. Also, a more coherent zone of elevated strain rate levels was found when the LCV gas was used. This suggests a slower mixing of the LCV gas fuel jet. The zones of elevated strain rates can be identified in the zone below the jet trajectory where the jet and the crossflow interact. The slower mixing might in this particular case depend on the higher density of the LCV gas (1.134 kg/m3) compared to the methane gas (0.669 kg/m3), which results in a higher inertia of the gas jet, thus reducing entrainment and mixing in the case of the LCV gas jet.

The plots for the high-temperature cases indicate better defined zones of both positive and negative axial strain rates below and above the jet trajectory, respectively. Furthermore the axial strain rate levels were found to be higher for the low-calorific-fuel gas jet. In the high-temperature case, the changed fluid-dynamics properties of the combustion air would be expected to have a large impact on the axial strain rate distribution, which also can be seen from the results.

This can also be seen in the direct flame photographs, where a flame region can be identified at a longer distance downstream along the jet trajectory during the LCV fuel-jet combustion case compared to the methane fuel-jet case, thus indicating an ignition delay as the methane fuel jet renders faster ignition of the fuel.

66

(a) (b)

(c) (d)

(e)

Figure 17. Axial strain rate distribution normalized by jet exit velocity at (a): 298K, 21% O2 combustion air, C3H8 burning; (b): 1173K, 21% O2 combustion air, C3H8 burning; (c): 1173K, 5% O2 combustion air, C3H8 burning; (d): 1373K, 21% O2 combustion air, C3H8 burning; (e): 1373K, 5% O2 combustion air, C3H8 burning

67

(a) (b)

(c) (d)

(e)

Figure 18. Axial strain rate distribution normalized by jet exit velocity at (a): 298K, 21% O2 combustion air, CH4 burning; (b): 1173K, 21% O2 combustion air, CH4 burning; (c): 1173K, 5% O2 combustion air, CH4 burning; (d): 1373K, 21% O2 combustion air, CH4 burning; (e): 1373K, 5% O2 combustion air, CH4 burning

68

The results show that combustion decreases the degree of turbulence in the region where the jet begins to deflect into the crossflow, denoted laminarization. Figure 19 and Figure 20 present the distribution of the intensity of radial turbulence levels normalized by jet exit velocity resulting from the entrained jet. The diagrams show a similar turbulence distribution in the nozzle region. This is as expected since the mixing is primarily governed by the momentum ratio between the jets, which was maintained constant for all the cases presented here. Reduced turbulence levels can be observed for the high-temperature air combustion. The results shown here suggest that radial turbulence intensity is lower under conditions of high air preheat temperature and low oxygen concentration. The areas of elevated turbulence in the reaction region can be seen where the axial strain rates are elevated also. This region provides information on the local mixing of the fuel jet with the crossflow.

The heat release generated from the combustion process has a hampering effect on the turbulence levels. The results showed that combustion decreases the degree of turbulence in the region where the jet begins to deflect into the crossflow. Paper 4 presents results on the LCV and methane fuel jets in normal temperature air. They show that the LCV fuel jet produces similar turbulence levels as the methane fuel jet in close proximity to the nozzle. This was expected as the mixing is primarily governed by the momentum ratio between the jets and this ratio was held constant for the two cases. The experimental findings suggest that high-temperature cases reduce the radial turbulence intensity significantly. Comparison of the results from the non-burning and combustion cases for LCV and methane shows that the turbulence levels are low for the high-temperature air case compared to the normal-temperature air case. A very small local region of higher turbulence can be seen in the zone of high axial strain rate. This region provides local mixing of the fuel jet with the crossflow. Furthermore, it can be seen that the turbulence is even further reduced when using methane as the fuel as compared to the low-calorific-value fuel (methane diluted with nitrogen). This indicates that the combustion process reduces the turbulence in the flow path. Also, the stronger heat release of the methane and propane fuel jets, compared to the LCV fuel jet, has a stronger effect on dampening the turbulence along the trajectory of the fuel jet. Similar results were reported by Eickhoff et al.[38] from studies on stabilization mechanisms of jet diffusion flames. This information suggests the clear role played by combustion in the flow dynamics associated with normal- and high-temperature air combustion conditions.

69

Figure 19. RMS radial velocity distribution normalized by the jet exit velocity at at (a): 298K, 21% O2 combustion air, C3H8 burning; (b): 1173K, 21% O2 combustion air, C3H8 burning; (c): 1173K, 5% O2 combustion air, C3H8 burning; (d): 1373K, 21% O2 combustion air, C3H8 burning; (e): 1373K, 5% O2 combustion air, C3H8 burning

(a) (b)

(c) (d)

(e)

70

Figure 20. RMS radial velocity distribution normalized by the jet exit velocity (a): 298K, 21% O2 combustion air, CH4 burning; (b): 1173K, 21% O2 combustion air, CH4 burning; (c): 1173K, 5% O2 combustion air, CH4 burning; (d): 1373K, 21% O2 combustion air, CH4 burning; (e): 1373K, 5% O2 combustion air, CH4 burning

(a) (b)

(c) (d)

(e)

71

Some results of the vorticity normalized by jet exit velocity are presented in Figure 21 and Figure 22. The vorticity is generated as soon as the jet starts to deflect into the crossflow, leading to rotation of the jet. A generally stronger vorticity for the high-temperature air case as compared to the normal-temperature air case can be noted. This indicates that strong rotation is initiated in the fuel jet/crossflow interaction area. In examining the different plots, it can be seen that in the regions of strong vorticity there are also elevated levels of axial strain. An explanation for this is that as soon as the main flow begins to entrain into the jet, rotation of the flow occurs. The slower mixing under high-temperature conditions for the low-calorific-value case would have the consequence of stronger rotation which would lead to delayed mixing. For the normal-temperature case, the findings indicate that the vorticity is distributed in a similar manner along the jet trajectory for both cases, giving an indication that similar mixing of the jet and crossflow takes place under non-burning conditions.

For the high-temperature case, a larger vorticity area can be seen for the combustion of LCV fuel jets compared to the methane fuel-jet case. The heat release from chemical reactions affects vorticity negatively, reducing vortex formation and destroying local vortex tubes [31]. This can be observed where the vorticity is concentrated to a smaller area for the methane and propane cases, which entail a more immediate heat release, and is distributed over a larger area for the LCV fuel for which combustion is slower.

72

(a) (b)

(c) (d)

(e)

Figure 21. Vorticity distribution normalized by the jet exit velocity (a): 298K, 21% O2 combustionair, CH4 burning; (b): 1173K, 21% O2 combustion air, CH4 burning; (c): 1173K, 5% O2combustion air, CH4 burning; (d): 1373K, 21% O2 combustion air, CH4 burning; (e):1373K, 5% O2 combustion air, CH4 burning

73

Figure 22. Vorticity distribution normalized by the jet exit velocity (a): 298K, 21% O2 combustion air, CH4 burning; (b): 1173K, 21% O2 combustion air, CH4 burning; (c): 1173K, 5% O2 combustion air, CH4 burning; (d): 1373K, 21% O2 combustion air, CH4 burning; (e): 1373K, 5% O2 combustion air, CH4 burning

(a) (b)

(c) (d)

(e)

74

7.7. Superimposition of experimental data In order to be able to couple the distinctive characteristics of the flames in the high-temperature low-oxygen environment, the flame photographs were superimposed with the PIV plots and intensity plots from the light emission spectroscopy analyses (Figure 23 and Figure 24). In Appendix 3 the results of some other experimental cases are presented. The Appendix shows the results for the LCV fuel jet under high-temperature conditions and for methane fuel jets both under high-temperature low-oxygen conditions and ambient conditions and finally, for a propane fuel jet under ambient conditions. In Figure 23a, the flow field is superimposed on a flame photograph of a C3H8 fuel jet in 1173K and with a crossflow oxygen concentration of 5 %. The figure illustrates the flow-field distribution and correlation of the flow field to the visible flame. The jet expansion and ignition of the fuel jet can be seen. The flow field follows the flame contour. Figure 23b, with the superimposed radial RMS distribution, clearly shows how the crossflow interacts with the fuel jet, and in the interaction zone where the mixing is highest, the turbulence levels are also highest. Figure 23c presents the vorticity distribution superimposed on the flame photograph. Two different crossflow-fluid rotation zones can be identified above and below the fuel jet. This is as expected, in line with the fluid dynamics of a fuel jet in crossflow. In Figure 23d, the superimposed axial strain distribution can be seen. The axial strain indicates acceleration or deceleration of the crossflow fluid, and it can be seen that the highest strain rates lie below the fuel jet. This indicates strong entrainment of the crossflow fluid into the fuel jet and explains the extended reaction zones found in Figure 24.

The reaction zone clearly shows the extended flame that occurs in the environment. A comparison with the results from the other experimental cases in Appendix 3 shows a significantly larger reaction zone with a secondary burning zone in the high-temperature oxygen-deficient cases.

75

(a) (b)

(c) (d)

Figure 23. Superimposed data of flame photographs and; (a) average velocity field, (b) radial RMS values, (c) vorticity distribution and (d) axial strain rate distribution. The fuel jet was C3H8

at 1173 K and 5 % oxygen concentration

76

(a) (b)

(c)

Figure 24. Superimposed data of flame photographs and light emission spectroscopy data; (a) C2-distribution, (b) CH-distribution, (c) OH-distribution. The fuel jet was C3H8 at 1173 K and 5 % oxygen concentration in the oxidizer

77

7.8. Flame temperature The results of the flame temperature calculations are illustrated in Figure 25, Figure 26, figure 27, Figure 28 and Figure 29. The influence of the oxidizer oxygen concentration and temperature can be observed, as well as the influence of fuel type.

Calculated temperatures for the vibrational temperatures of the C2 species are shown. The temperature was calculated using Equation 39. The temperature data used for the plots pertain only to within the reaction volume. This is because the temperature was calculated as a function of the C2 species that is only present within the reaction zone. Thus, there are no C2 radicals outside the reaction zone and calculated temperatures for that area would be meaningless.

In considering the effect of oxidizer oxygen concentration, a significant difference in the temperature field can be seen when the oxygen concentration is changed between 21 % and 5 %. At 21 % oxygen, the temperature plot reveals a small region of high temperatures; when the oxygen concentration drops to 5%, the flame is significantly larger. This is due to the extended reaction zone that results from the slower reaction rate due to insufficient oxygen in the reaction volume. This enlargement of the reaction zone can be seen in both the propane and methane fuel-jet cases.

Preheat temperature was found to have an effect on the temperature distribution as well. When the oxidizer temperature is increased, the zone of high temperatures tends to shrink. Thus, a more intensive reaction takes place. This is due to faster ignition, and a greater amount of energy being available for more rapid heating of the fuel. The ignition is determined for the position where the temperature in the trajectory is first elevated. More rapid ignition of the fuel jet can also be noted when the temperature is 1373 K versus 1173 K. However, an enlarged reaction zone can be seen for the higher oxidizer temperature and oxygen content of 5 %. This shows that oxidizer oxygen concentration has a stronger influence on flame volume than does oxidizer temperature.

A change in fuel calorific value had a minor impact on the temperature distribution within the flame regarding comparison of the propane and methane fuel jets. The methane fuel-jet results indicated faster ignition when considering the point where the first elevated temperatures can be detected along the trajectories. The behavior of the LCV fuel jet significantly differed from that of the methane and propane fuel jets. The temperature plot of the LCV fuel jet shows a delayed ignition and a very evenly distributed temperature within the reaction volume. This implies that the LCV fuel jet has a slow ignition, as it has to be heated up to a high temperature and be well mixed with the oxidizer in order to react. This shows that it is possible to combust fuels with low heating values, but the mixing and the oxidizer temperature are of great importance in ignition of the fuel jet.

78

Figure 26. Temperature plots based on the C2 vibrational temperature calculation forC3H8 gas fuel jets at 1173 K (left) and 1373 K (right) oxidizer preheat and 5% oxygen concentration

Figure 25. Temperature plots based on the C2 vibrational temperature calculation forC3H8 gas fuel jets at 1173 K (left) and 1373 K (right) oxidizer preheat and 21% oxygen concentration

79

Figure 27. Temperature plots based on the C2 vibrational temperature calculation forCH4 gas fuel jets at 1173 K (left) and 1373 K (right) oxidizer preheat and 21% oxygen concentration

Figure 28. Temperature plots based on the C2 vibrational temperature calculation forCH4 gas fuel jets at 1173 K (left) and 1373 K (right) oxidizer preheat and 5 %oxygen concentration

80

[°C

Figure 29 Temperature plots based on the C2 vibrational temperature calculation for LCV gas fuel jet at 1173 K oxidizer preheat and 8.1 % oxygen concentration

81

7.9 Semi-industrial HiTAC jet flame study In order to compare the experimental results from the small-scale test apparatus and to determine if the characteristics found for the single flame were similar to those for a larger flame, experiments where conducted using a semi-industrial experimental apparatus equipped with a burner that could be operated in both regenerative and conventional mode. The experiments were performed with the burner running in conventional combustion mode and with a burner running in regenerative combustion mode. Flame temperature and heat flux were determined using probes that were inserted from the furnace ceiling. The experimental results showed a more compact zone close to the burner face with high temperatures in the conventional mode, and in the regenerative combustion mode the peak temperatures were lower and the temperature more evenly distributed within the entire combustion chamber. The results of the temperature and total-heat-flux measurements are illustrated in Figure 30 and Figure 31, respectively. The results clearly show that the same features of even flame-temperature distribution and even heat-flux distribution were obtained in both the large-scale and small-scale experiments.

The temperature plots in Figure 30 show, for conventional mode, a peak flame temperature at the burner face. In regenerative burner mode, the temperature peak is absent. This is due to the high injection velocity which generates a strong entrainment and in turn leads to slower ignition as the flame is not attached to the burner face. Thus, a larger volume to react exists once the fuel is ignited, leading to a more even temperature distribution. The enlarged reaction volume can also be seen in Figure 32which shows the CO distribution for both regenerative burner mode and for conventional mode. A larger reaction zone can be seen in the case of the regenerative burner mode. The enlarged reaction zone is an effect of enhanced mixing due to the higher injection velocities that result in a more diluted and low-oxygen-concentration reaction zone.

The results indicated a significantly enlarged reaction zone, which is coherent with the results of the radical determinations in the single-flame test apparatus. This also shows that the reaction zone was found to be enlarged in the semi-industrial-scale results and that the flame features observed were similar to those found for the single-jet experiments.

82

Figure 31. Heat flux profiles at the burner face level in the semi-industrial test furnace for HiTAC mode (left) and for conventional mode (right)

Figure 32. CO concentrations along a vertical cut in the semi-industrial test furnace for HiTAC mode (left) and for conventional mode (right)

Figure 30. Temperature profiles at the burner face level in the semi-industrial test furnace forHiTAC mode (right) and for conventional mode (left)

83

8. CONCLUSIONS

84

85

An investigation on the effects of oxidizer temperature, oxidizer oxygen concentration and fuel-jet calorific value on the combustion process and flame features has been carried out using the configuration of a single fuel jet in crossflow. Experiments were also performed using a semi-industrial test apparatus where conventional and HiTAC flames from an experimental burner were compared. The fuels used in the investigation were propane, methane and partially a low-calorific-value fuel.

Particle Image Velocimetry (PIV) was employed for the first time to investigate the flow field of a gaseous single fuel jet in a high-temperature oxygen-deficient environment. Light emission spectroscopy and direct flame photographs were also used to investigate the experimental single fuel jet.

The study’s results showed the oxidizer preheat temperature to have a significant effect on the single gas jet flame. The flame shrunk in volume when the preheat temperature was increased. When the oxygen concentration in the oxidizer was reduced, the reaction volume increased significantly. This was due to a reduced reaction rate as not enough oxygen was present to complete the combustion process in one step, leading to two-stage burning. This can clearly be seen in the LES plots.

The fuel calorific value was found to have an effect on ignition delay. The LES results showed significant ignition delay for the LCV fuel.

The mass entrainment decreased when the oxygen concentration was increased due to faster and more intense reaction, which in turn rendered a reduced reaction volume. The mass entrainment under high-temperature low-oxygen conditions increased as a consequence of an enlarged reaction volume.

The jet expansion was larger for high-temperature low-oxygen conditions due to the larger entrainment required for complete combustion of the gaseous fuel.

Combustion noise was shown to be reduced for jet flames in the high-temperature low-oxygen environment compared to normal-temperature jet flames. Moreover, the turbulence levels in the flames were found to be significantly lower under high-temperature low-oxygen conditions.

Finally, the same flame features observed in the results from the small-scale experiments were found for the semi-industrial-scale experiments.

86

87

9. REFERENCES

88

89

1. Tanaka, R., New progress of energy saving technology toward the 21st century; Frontier of combustion & heat transfer technology; Proceedings of 11th IFRF Members Conference; 10-12th May 1995.

2. Yasuda, T. Ueno C., Dissemination project of industrial furnace revamped with HTAC; 2nd International High Temperature Air Combustion (HTAC) Symposium; January 17-18, 2000.

3. Hardesty D. R. and Weinberg F. J. Burners Producing Large Excess Enthalpies. Combust. Sci. Technol., 8, pp 201-214, 1974.

4. Cornforth, J. R.. Combustion Engineering and Gas Utilization, 3rd edition, 1992.

5. Suzakawa Y., Sugiyama S., Mori I., Heat transfer improvement and NOx reduction in an industrial furnace by regenerative combustion system, Proc. IECEC 96., 2 (1996) 804 809, IEEE (96360), Washington DC.

6. Rafidi N and Blasiak W, Experimental Study: Thermal Performance of Ceramic Regenerative Heat Exchangers Used in HiTAC Regenerative burning Systems`, IFRF Online Combustion Journal October 2003, Stockholm, Sweden.

7. Hasegawa T., Mochida S. and Gupta, A.K., Development of Advanced Industrial Furnace Using Highly Preheated Combustion Air, Journal of Propulsion and Power, vol. 18, no 2, 2002.

8. Wünning J.A., and Wünning J.G.; Burners for flameless oxidation with low NO formation even at maximum air preheat; GASWÄRME Int. 41, 10; October 1992.

9. Cavaliere A and Joannon, M, Mild combustion, Progress in Energy and Combustion Science 30, pp 329-366, 2004.

10. Mochida S., Hasegawa T., Tanaka R.;Characteristics of High Temperature Air Combustion; RAN’95, 1995.

11. Amagai, K., Arai, M., Propane flame appeared in a hot air heated over self-ignition temperature, RAN95, International Symposium on Advanced Energy Conversion Systems and Related Technologies, Nagoya, Japan, pp. 703-710, 4- 6 December, 1995.

90

12. Gupta A.K., Li Z.; Effect of fuel property on the structure of highly preheated air flames; International Joint Power Generation Conference, IJPGC 1997, ASME EC-vol 5, Volume 1. 1997.

13. Bolz, S., Gupta, A.K. . Effect of air preheat temperature and oxygen concentration on flame structure and emission, FACT-Vol. 22, International Joint Power Generation Conference, Volume 1, pp 193 – 205, ASME 1998.

14. Gupta A.K., Bolz S, Hasegawa T., Effect of air preheat temperature and oxygen concentration on flame structure and emission, Journal of Energy Resources Technology, Vol. 121, pp 209-216, September 1999.

15. Hasegawa T, Tanaka R, Niioka T, Combustion with high temperature low oxygen air in regenerative burners, The First Asia-Pacific Conference on Combustion, Osaka, Japan, pp. 290 – 293, May 12-15, 1997.

16. Kishimoto, K., Watanabe, Y., Kasahara, M., Hasegawa, T., Tanaka, R., Observational study of chemiluminescence from flames with preheated and low oxygen air, The First Asia-Pacific Conference on Combustion, Osaka, Japan, pp. 468-471, May 12-15, 1997.

17. Kitagawa, K., Konishi, N., Arai, N., Gupta, A.K., Two-dimensional distribution of flame fluctuation during highly preheated air combustion, FACT-Vol. 22, International Joint Power Generation Conference, Volume 1, pp 239 – 242, ASME 1998.

18. Blasiak W., Szewczyk D., Dobski T., Influence of N2 addition on combustion of single jet of methane in highly preheated air, Proceedings of IJPGC’01, paper FACT-19048, New Orleans, USA, 4-7 June 2001.

19. Lille S, Dobski T, Blasiak W, Visualisation of fuel jet in conditions of highly preheated air combustion, AIAA Journal of Propulsion and Power, Vol. 16, Number 4, pp 595-600, 2000.

20. Ito Y., Gupta A. K., Yoshikawa K. and Shimo N., Combustion Characteristics of Low Calorific Value Gas with High Temperature and Low-Oxygen Concentration Air, Proc. 5th High Temperature Air Combustion and Gasification (5th HTACG), Yokohama, Japan, October 28-31, 2002.

21. Lille, S., Experimental study of single fuel jet combustion and high-cycle regenerative systems, doctoral thesis, Royal Institute of Technology, Stockholm, Sweden, May 2002.

22. Blasiak, W., Yang, W.H. and Rafidi N, Physical properties of a LPG flame with high-temperature air on a regenerative burner, Combustion and Flame, 136, pp 567-569, 2004.

23. Yang, W. and Blasiak W., Numerical study of fuel temperature influence on single gas jet combustion with preheated and oxygen deficient air, ENERGY, article in press, 2004.

24. Yang, W. and Blasiak, W. Chemical Flame Length and Volume in Liqufied Propane Gas Combustion Using High-Temperature and Low-Oxygen-Concentration Oxidizer Energy & Fuels, vol 18, no 5, pp 1329-1335, 2004.

25. Dally, B. B., Riesmeier, E. and Peters, N., Effect of fuel mixture on moderate and intense low oxygen dilution combustion, Combustion and Flame, vol. 137, issue 4, pp 418-431, June, 2004.

26. Yuan, J. and Naruse, I., Effects of Air dilution on Highly preheated air combustion in a regenerative furnace, Energy & Fuels, vol 13, no 1, pp 99-104, January 1999.

27. Margason, R. J, The path of a jet directed at large angles to a subsonic free stream. Technical Report TN D-4919, NASA., 1968.

91

28. Margason, R. J. Fifty years of jet in crossflow research. In AGARD Jet in Cross Flow Symposium, Volume AGARD CP–534, Winchester, United Kingdom, 1993.

29. Broadwell, J. E. and Breidenthal, R. E., Structure and Mixing of a Transverse Jet in Incompressible Flow, Combustion and Flame, Vol. 114, No. 3/4, pp. 319-335, 1984.

30. Choudhuri, A. R. and Gollahalli, S. R., Effects of Ambient Pressure and Burner Scaling on the Flame Geometry and Structure of Hydrogen Jet Flames in Cross-flow, International Journal of Hydrogen Energy 25, pp. 1107-1118, 2000.

31. Donghee, H. and Mungal, M. G: Simultaneous Measurements of Velocity and CH Distribution. Part II: Deflected Jet Flames, Combustion and Flame vol. 133, pp. 1 – 17, 2003.

32. Hasselbrink, E. F. and Mungal M. G., Transverse Jets and Jet Flames. Part I. Scaling Laws For Strong Transverse Jets, J. Fluid Mech., vol. 443, pp 1-25, 2001.

33. Konishi, N., Kitagawa, K., Arai, N. and Gupta, A. K.: Two-Dimensional Spectroscopic Analysis of Spontaneous Emission from a Flame using Highly Preheated Air Combustion, J. Propulsion and Power, Vol. 18, No. 1, January-February, pp. 199-204, 2002.

34. Kitagawa, K., Shinoda, S., Arai, N., Ishiguro, T., Tsuge, S., Hasegawa, T. and Tanaka, R.: Spectroscopic Observation of Homogenization and Stabilization during Combustion of Hydrocarbons with Pre-Heated Air, Proc. 1st Asia-Pacific Conference on Combustion, Osaka, Japan, 1997.

35. Ishiguro, T., Tsuge, S., Furuhata, T., Kitagawa, K. Arai, N., Hasegawa, T., Tanaka, R. and Gupta, A. K.: Homogenization and Stabilization During Combustion of Hydrocarbons with Preheated Air, Proc. 27th Symposium (Intl.) on Combustion, The Combustion Institute, Pittsburgh, PA, pp. 3205-3213, 1998.

36. Hasselbrink, E. F. and Mungal M. G., Transverse Jets and Jet Flames. Part II. Velocity and OH Field Imaging, J. Fluid Mech., vol. 443, pp 27-68, 2001.

37. Lilley, D. G., Lateral Injection Into Swirling Crossflow, RAN2001, 3rd International Symposium on Advanced Energy Conversion Systems and Related Technologies, Nagoya, Japan, December 15-17, 2001.

38. Eickhoff, H., Lenze, B. and Leuckel, W., Experimental Investigation on the Stabilization Mechanism of Jet Diffusion Flames, Proceedings of the 20th Symposium (Intl.) on Combustion, The Combustion Institute, Pittsburgh, PA, pp 311-318, 1984.

39. Simon, M.A., Baird, B. D., and Gollahalli, S. R., Characteristics Of A Laminar Diffusion Flame In A Cross-Flow Of Combustion Products, Proceedings of PWR2005, Chicago, Illinois U.S.A, April 5-7, 2005.

40. Weber R., Verlaan V.L., Orsino S., Lallemant N., Flow and mixing in gas-fired furnaces operated with highly preheated combustion air, Proceedings of 2nd Int. High Temperature Air Combustion symposium, Taiwan January, 1999.

41. Lille, S., Blasiak, W., Mortberg, M., Dobski, T., Weihong, Y., Heat Flux Evaluation in a Test Furnace Equipped with High Temperature Air Combustion (HTAC) Technique, 2002 International Joint Power Generation Conference, Phoenix, AZ, USA, June 2002.

42. Quinqueneau A, Miquel P.F., Wills B., Experimental studies in an industrial staged burner with preheated air, 12th IFRF Members Conference, Nooordwijkerhout, The Netherlands, 5-8 May, 1998.

92

43. Quinqueneau A, Touzet A, Oger M., Experimental studies on regenerative industrial burners operating in the flameless oxidation mode, 13th IFRF Members Conference, Ijmuiden, The Netherlands, May, 2001.

44. Pipers O., On combustion using separated air and natural gas injection systems in various furnace atmospheres and internal furnace geometry, 17th International Symposium on Combustion Processes, Poznan University of Technology, Poznan, September 24-27, pp 66-74, 2001.

45. Rafidi N., Blasiak W., Measurements of Flue Gases Compositions and Flame Volume in High Temperature Air Combustion (HiTAC) Furnace, Finnish - Swedish Flame Days 2002, Vaasa, Finland, September 24-25, 2002.

46. Hasselbrink, E. F., Transverse Jets and Jet Flames: Structure, Scaling, and Effects of Heat Release, Technical report TSD-121, Thermosciences Division, Department of Mechanical Engineering, Stanford University, Stanford, California 94302-3032, U.S.A.

47. Adrian R. J., Particle-imaging techniques for experimental fluid mechanics, Ann. Rev. Fluid 261. Mech. 23, 1991.

48. Raffel M., Willert C. and Kompenhans J., U Particle Image Velocimetry – a practical guideU, Springer-Verlag, Heidelberg, 1998.

49. Melling, Tracer particles and seeding for Particle Image Velocimetry, Meas. Sci. Technol. 8, 1406, 1997.

50. Kitagawa N., Arai, N., and Gupta, A. K., Two-Dimensional Spectroscopic Analysis of Spontaneous Emission from a Flame using Highly Preheated Air Combustion, J. Propulsion and Power, Vol. 18, No. 1, pp. 199-204, January-February, 2002.

51. Kitagawa, K., Shinoda, S., Arai, N., Ishiguro, T., Tsuge, S., Hasegawa, T. and Tanaka, R.: Spectroscopic Observation of Homogenization and Stabilization during Combustion of Hydrocarbons with Pre-Heated Air, Proc. 1st Asia-Pacific Conference on Combustion, Osaka, Japan, 1997.

52. Ishiguro, T., Tsuge, S., Furuhata, T., Kitagawa, K. Arai, N., Hasegawa, T., Tanaka, R. and Gupta, A. K.: Homogenization and Stabilization During Combustion of Hydrocarbons with Preheated Air, Proc. 27th Symposium (Intl.) on Combustion, The Combustion Institute, Pittsburgh, PA, pp. 3205-3213, 1998.

53. Hughes, P.M. and Parameswaran, T., Emission Spectroscopy for Flame Characterization, 14th IFRF Members Conference, Noordwijkerhout, The Netherlands, May, 2004.

54. Gaydon, A. G., The Spectroscopy of Flames, John Wiley and Sons, New York, pp. 1-9, pp 159-163, 1974.

55. Gaydon, A. G., Spectroscopy and Combustion Theory, Chapman and Hall, London, pp. 1-18, 42-58, 1942.

56. Pratte, B. D. and Baines, W. D., Profiles of Round Turbulent Jet in a Crossflow, Journal of Hydronautics, vol. 92, pp. 53-64, 1967.

57. Han, D., Orozco, V. and Mungal, M. G., Gross-Entrainment Behavior of Turbulent Jets Injected Obliquely into a Uniform Crossflow, AIAA Journal, vol. 38, No. 9, pp 1643-1649, September 2000.

58. Hasselbrink, E. F., and Mungal M. G., An Analysis of the Time-Averaged Properties of the Far Field of the Transverse Jet, AIAA Paper 96-0201, January 1996.

93

59. Gaydon, A. G., and Wolfhard, H. G., Flames, Their structure, radiation and temperature, John Wiley & Sons, Inc., New York, USA, 1978.

60. Chomiak, J., Combustion, A study in theory, fact and application, Abacus press/Gordon & Breach science publishers, Eds. A. K. Gupta and D. G. Lilley, New York, USA, 1989.

61. Bis, Z., Blasiak, W., Collin, T. and Vaclavinek J., Non-isothermal turbulent Jet in a Cross-flow, First Baltic Heat Transfer Conference, Gothenburg, Sweden, 1991.

62. Szewczyk, D., Mörtberg, M., Rafidi, E. N., Dobski, T., and Blasiak, W., Measurements of temperature and heat flux in HiTAC flame for unsteady state condition, 5th International Symposium on High Temperature Air Combustion and Gasification, Yokohama, Japan, October 28-30, 2002.

63. Chedaille, J. and Braud, Y. in Beér, J. M. and Thring, M. W. (ed)., Measurements in Flames, International Flame Research Foundation, 1972.

64. Zhang, T., A study of thermal counterflow in HE II using particle image velocimetry (PIV) technique, doctoral thesis, The Florida State University, Tallahassee, FL, U.S.A., 2004.

94