Upload
api-3844809
View
100
Download
2
Embed Size (px)
Citation preview
Thermo-Acoustic Characterization of a Gas Turbine Burner on a Single Burner Test Rig
A project report submitted
To
Manipal University
For Partial Fulfilment of the Requirement for theAward of the Degree
Bachelor of Engineering
In
Mechanical
Manipal Institute of Technology, India
(Constituent Institute of Manipal University)
By
Akhouri Piyush Raj,
Reg no. 040909190
Under the Guidance of
DEPARTMENT OF MECHANICAL ENGINEERING
MANIPAL INSTITUTE OF TECHNOLOGY(A constituent Institute of MANIPAL UNIVERSITY)
MANIPAL - 576 104, KARNATAKA, INDIA
June 2008
M.Sc. Panduranga Reddy AlemelaScientific Co-worker
Chair of Thermodynamics, Technical University of Munich.
External guide
Dr. N. Y. SharmaHead of Department,
Department of Mechanical and Mfgg. Engineering,Manipal Institute of Technology.
Internal guide
ACKNOWLEDGEMENTS
I would like to express my gratitude to all those who gave me the opportunity to complete this
project successfully. I want to thank Mechanical and Manufacturing Dept. of Manipal
Institute of Technology, India, for giving me the permission to pursue my Project in
Lehrstuhl of Thermodynamics, Technical University of Munich, Germany. I would like to
thank Deptt. Of Thermodynamics, TUM, Munich for giving me the opportunity to do my
project as a part of completing my Bachelor of Engineering degree.
My special thanks to Prof. Dr.-Eng. Thomas Settelmayer; Head, Department of
Thermodynamics for giving me the opportunity to do the internships in one of his ongoing
supervised projects in the chair of thermodynamics.
I am deeply indebted to my guide and mentor at TU Munich, Msc. Panduranga Reddy
Alemela, whose help, stimulating suggestions and encouragement helped me in all the time of
this project and writing of this Report. I would also like to thank Dipl.Eng Urban Neunert
and Dipl.Eng Dan Fanaca for providing valuable suggestions and information during the
course.
I would like to express my gratitude to all my colleagues at the Chair of Thermodynamics, for
their support, help and hints.
I would like to express my sincere thanks to all the members of local committee of IAESTE
IndiaMIT and LC Munich, for encouraging and helping me time to time to do my project in
this esteemed organisation.
I will be obliged to my internal guide and HOD, Deptt of Mech. and Manufacturing Engg,
MIT Manipal, Dr. N. Yagnesh sharma, for all his support and cooperation during the course
of this project.
Thanking You,
Yours Sincerely,Akhouri Piyush Raj
MANIPAL INSTITUTE OF TECHNOLOGY(A constituent Institute of MANIPAL UNIVERSITY
MANIPAL - 576 104 .
DEPARTMENT OF MECHANICAL AND
MANUFACTURING ENGINEERING
CERTIFICATE
This is to certify that the practice school titled “Thermo-Acoustic Characterization of a Gas
Turbine Burner on a Single Burner Test Rig” is a bonafide work of Akhouri Piyush Raj
from 07.01.2008 to 23.06.2008 carried out in partial fulfillment of the requirements for
awarding the degree of Bachelor of Engineering in Mechanical discipline in Manipal Institute
of Technology under MANIPAL University, Manipal during the academic year 2007-
2008.The contents of this report, in full or in part, have not been submitted to any other
Institute or University for the award of any Degree or Diploma.
External guide:
M.Sc. Mr. Panduranga Reddy Alemela,
Scientific Co-worker,
Lehrstuhl für Thermodynamik,
Technische Universität München, Signature
Bavaria, Germany-81547 Date:
Internal guide:
Dr. N. Yagnesh Sharma,
Head of Department,
Deptt. Of Mechanical and Mfg. Engineering, Signature
Manipal, Karnataka-576104 Date:
Annexure – 2
TABLE OF CONTENTS
Page No.
CHAPTER 1_INTRODUCTION 2
1.1 Need for the project 3 1.2. Objective 4
CHAPTER 2_LITERATURE REVIEW2.1 Concept of thermo acoustics 52.2 Physics of Combustion Instabilities 52.3 Growth of Thermo acoustical research 7 2.4 Research Work at TU Munich. 11 2.5 Nomenclature 12
CHAPTER 3_EXPERIMENTAL SETUP 14 3.1 Understanding the test Rig 153.2 Test rig working conditions 16 3.3 Various Components of test rig
CHAPTER 4_MICROPHONE CALIBRATION234.1 Essentials of Microphone calibration 234.2 Procedure 244.3 Results 334.4 Graph Plots 344.5 Conclusion 42
For Final practice school report…..CHAPTER 5_Obtaining the Burner Transfer Matrix 445.1 Essencials of Experiment5.2 Concept5.3 Test Rig Preparation5.4 Procedure5.5 Data collection and post processing5.6 Results and discussion
REFERENCES 50
ABSTRACT
KEYWORDS: Combustion instability; Oscillation Premixed flame ; Thermo acoustics ;
Stability ; Combustion chamber ; Gas turbine ;
In the following thesis,
CHAPTER 1: INTRODUCTION
Continuous combustion processes are encountered in many applications related to power
generation, heating, as well as domestic and industrial burners. The need for high thermal
efficiency and low levels of pollutant emissions requires continuous research and
development in the field of power generation. To comply with today’s stringent regulations,
the concept of lean premixed combustion has been adopted as it offers a certain number of
advantages in controlling the emissions level. The reliability and flexibility of the Modern
cycle power gas turbine has imposed this technology as standard energy supply source in the
present and in the foreseeable future.
These processes exhibit a wide range of dynamics, thereby promoting the resonant coupling
between the unsteady parts of the heat release and the acoustic field, leading to sustained large
amplitude pressure oscillations, known as Thermo acoustic Instabilities (Rayleigh criteria).
These instabilities are more likely to occur in combustion systems near the extinction limit
(increased flame receptivity). Modern, ultra-low NOx gas turbines running in a lean,
premixed mode meet all these conditions and avoiding instability remains a challenging task
for the design of power plants.
Industrial practice has proven this technology to be sensitive to the development of self
sustained thermo-acoustic combustion instabilities resulting in high levels of pressure
pulsations in the combustion chamber. These disturb the normal operation and can produce
extensive hardware damage to the system. The driving mechanism for this undesired
phenomenon is mainly the feedback loop with a positive growth rate linking the unsteady heat
release of the turbulent premixed flame and the acoustic field of the combustion system. [1]
Active control to suppress thermo acoustic instabilities and consequently to enhance stable
heat release close to the lean extinction limit offers the advantage to increase the efficiency
and lifetime of the engine, keeping the pollutant emissions low, in particular NOx and CO.
1.1 Need for the project:
It is clear that it is very difficult to model the thermo acoustic interaction mechanisms in a
flame stabilized by such a mechanism. Nevertheless, it will be demonstrated here that
surprisingly simple models suffice to describe the thermo acoustic behaviour of the flame.
However, some of the parameters in these models need experimental input.
Research is done in this field to develop analysis tools to predict instabilities in the early
design stages of a combustor. Thus important design countermeasures could be applied and
future costly interventions during the nominal operations avoided. The preferred approach,
due to robustness and reliability, is the stability analysis based on the acoustic network
modelling method. For this however, the dynamic behaviour of every hardware component in
terms of the transfer matrices must be apriority known. The transfer matrix (Eqn.1) describes
the change in amplitude and phase of the acoustic parameters pressure and velocity upstream
(u) and downstream (d) of the test element. For the burner and flame these are, because of
complexity, still derived from experimental investigations or numerical simulations.
Eqn.1. Definition equation of a transfer matrix linking the acoustic parameters Pressure (p) and velocity (u) across the element
In this project the dynamic behaviour of burners and flames in single burner configuration is
experimentally investigated and the transfer matrices are determined. A special attention is
dedicated to the comparison between the two cases; theoretical approach and practical
approach, when the similarities and differences are assessed.
The research methodology adopted in this project goes in accordance to the Combustion
Instabilities methodology proposed by Dr. Wolfgang Polifke (2004). The Data has been acquired
using LABVIEW (V5) Software. The modeling and simulations have been performed using
MATLAB (2007B).The design modifications of the test rig have been planned on CATIA V14
Software. The author basically assists MSC Panduranga Reddy Alemela in his experiments and
conducts microphone Calibration and Obtaining the Burner Transfer Matrix as his final semester
practice school internship project.
To measure the flame transfer matrices, firstly the operating domain of the thermo acoustic
experiment with flame is performed.
1.2 Objective:
1.2.1 To perform the Microphone Calibrations for the thermo acoustic test on the single
burner test Rig and obtain microphone coefficients..
1.2.2 Thermo-Acoustic Characterization of a Gas Turbine Burner on a Single Burner Test
Rig by performing cold test run of the test rig and post-process the data to study flow
characteristics of the burner.
Following this introduction, some of literature related to the evolution, concept and present
status of research going on thermo acoustic instabilities are presented outlining the research
area and the work at TU Munich.
Figure 1. Single burner test Rig facility
CHAPTER 2:
2.1 Concept of Thermo Acoustics
A sound wave in a gas is usually regarded as consisting of coupled pressure and motion
oscillations, but temperature oscillations are always present, too. When the sound travels in all
channels, oscillating heat also flows to and fro inside the channel walls. The combination of
all such oscillations produces a rich variety of “thermo acoustic” effects. [2]
Research in thermo acoustics began with simple curiosity about the oscillating heat transfer
between gas sound waves and solid boundaries. These interactions are too small to be obvious
in the sound in air with which we communicate every day. However, in intense sound waves
in pressurized gases, thermo acoustics can be harnessed to produce powerful engines,
pulsating combustion, heat pumps, refrigerators, and mixture separators. Hence, much current
thermo acoustics research is motivated by the desire to create new technology for the energy
industry that is as simple and reliable as sound waves themselves.
2.2 Physics of Combustion Instabilities
Consider mass conservation across a premix flame (equation 2) in steady state,
Eqn.2. Mass conservation across a premix flame
With the density of the combustion products (index h for ‘hot’) being lower than the
density of the fresh fuel/air mixture (index c for ‘cold’), the velocity u or the volume flux
must increase across the flame. Now, if the heat release rate of the flame fluctuates, the
volume ”produced” by the flame will also fluctuate, and this will generate sound; just like a
loudspeaker box with its oscillating membrane.[3]
The heat release rate may be perturbed by turbulent fluctuations of the velocity field upstream
of the flame front. This gives rise to combustion noise, e.g. a camping burner or a blow torch
which ‘hisses’ or ‘roars’. Combustion noise often exhibits a broad band frequency
distribution, which derives from the size distribution of the turbulent eddies perturbing the
flame. Combustion noise may also be generated by fairly large scale, vortical
coherent structures, originating from hydrodynamic instability of the
base flow (e.g. a shear layer or swirling flow).
In any case, if one speaks of combustion noise, it is usually implied that there is no significant
feedback from the sound emitted back to the flow fluctuations which perturbed the heat
release in the first place.
However, if the flame is enclosed in a combustion chamber, sound may be reflected back to
the flame such that a feedback loop is established. If the phase between the sound field
established in the chamber and the fluctuations of heat release is just right, a self-excited
combustion instability may occur, where small (infinitesimal) perturbations are amplified ever
more, until eventually some kind of saturation mechanism kicks in. For saturated thermo-
acoustic combustion instabilities, limit cycle velocity fluctuations often exceed the mean flow
velocities; amplitudes of pressure fluctuations can reach more than 120 dB in atmospheric
flames, and several atmospheres in rocket engines. Damage to the combustion equipment can
result very quickly due to excessive mechanical or heat loads, emissions of noise or pollutants
like oxides of nitrogen or carbon monoxide are often intolerable. This is why combustion
instabilities are not merely a fascinating phenomenon, but of great technical importance in
aerospace, energy and process engineering. [4]
2.3 History/Growth of Thermo Acoustics
The rich history of thermo acoustics has many roots, branches, and trunks intricately
interwoven, supporting and cross-fertilizing each other. It is a complicated history because in
some cases invention and technology development, outside of the discipline of acoustics, have
preceded fundamental understanding; at other times fundamental science has come first.
Dr. N. Rott in 1969 took the meaning of the word “thermo acoustics” to be self-evident; a
combination of thermal (heat) effects and sounds [7-9]. He developed the mathematics
describing acoustic oscillations in a gas in a channel with an axial temperature gradient, with
lateral channel dimensions of the order of the gas thermal penetration depth δ κ (typically of
the order of 1 mm), this being much shorter than the wavelength (typically of the order of 1
m). The problem had been investigated by Rayleigh and by Kirchhoff, but without
quantitative success. In Rott’s time, motivation to understand the problem arose largely from
the cryogenic phenomenon known as Taconis oscillations–when a gas-filled tube reaches
from ambient temperature to a cryogenic temperature, the gas sometimes oscillates
spontaneously, with large heat transport from ambient to the cryogenic environment. Dr. T.
Yazaki [10] later demonstrated most convincingly that Rott’s analysis of the Taconis
oscillation was quantitatively accurate.
A century earlier, Lord Rayleigh [11] understood the qualitative features of such heat-driven
oscillations: “If heat be given to the air at the moment of greatest condensation [i.e., greatest
density] or be taken from it at the moment of greatest rarefaction, the vibration is
encouraged.” He had investigated Sondhauss oscillations [12] the glassblowers’ precursor to
Taconis oscillations. Raleigh’s criterion was also understood to apply to Rijke’s oscillations.
[13] Similar oscillations can also occur when combustion takes place in a cavity [14] The
oscillations occur spontaneously if the combustion progresses more rapidly or efficiently
during the compression phase of the pressure oscillation than during the rarefaction phase.
Such oscillations must be suppressed in rockets to prevent catastrophic damage, but they are
deliberately encouraged in some gas-fired residential furnaces and hot-water heaters to
improve their efficiency.
To us, the word “thermo acoustics” represents one unifying analytical and conceptual
approach to all of these devices and phenomena. [27] The thermo acoustic approach begins
with the assumptions that the oscillations of pressure p, temperature T , density ρ, velocity u,
and entropy s can be thought of as “small” and that they are adequately represented as
sinusoidal functions of time. Results of engineering interest are obtained as time-averaged
products of the oscillating variables: heat fluxes are proportional to the product of T and u,
work to the product of p and u, mass fluxes to the product of ρ and u, etc. Surprisingly,
despite the assumption that the oscillations must be small and mono-frequencied, the results
of the thermo acoustic approach are usefully accurate even for large oscillations with
substantial harmonic content.
This summary highlights only some of the interesting inventions, discoveries, insights, and
fundamental demonstrations of thermo acoustics in the past half century.
2.4 Research at Technical University of Munich
Modern design of the low emission combustor is characterized by swirling air in the
combustor dome coupled with distributed fuel injection to maximize mixing. This design
results in efficient combustion with extremely low emissions. The burner on which we are
experimenting is known as EV5 burner (Provided by ALSTOM) and has the unique property
of flame instabilization in free space near the burner outlet utilizing the sudden breakdown of
a swirling flow, called vortex breakdown. The swirler is of exceptionally simple design,
consisting of two halves of a cone, which are shifted to form two air slots of constant width.
Gaseous fuels are injected into the combustion air by means of air distribution tubes
comprising of two rows of small holes perpendicular to the inlet ports of the swirler.
Complete mixing of fuel and air is obtained shortly after injection.
The characteristic of combustion stabilization by vortex breakdown are controlled by the flow
dynamics associated with this particular flow phenomenon. Vortex breakdown is defined as a
flame instability that is characterized by the formation of an internal stagnation point on the
vortex axis, followed by reverse flow. Upstream of the vortex breakdown location, the
velocity profile is highly jet-like with a peak velocity almost three times greater than the mean
velocity. Very close to the down Stream of the breakdown, the flow in the core may
completely stagnate and then change to a wake-like flow. Downstream of the breakdown
turbulence increases, axial velocities are substantially lower and reverse flow is possible.
In this project the dynamic behaviour of burners and flames in both single and annular
configurations are experimentally and theoretically investigated and the transfer matrices
determined. A special attention is dedicated to the comparison between the two cases, when
the similarities and differences are assessed. The experimental method consists mainly in
recording the acoustic response of the combustion system under external excitation for
determining the above mentioned quantities. The test object is a lean premixed burner from
the industrial partner ALSTOM Power Generation.
2.5 Nomenclatures
BTM Burner Transfer Matrix
FTM Flame Transfer Matrix
FTF Flame Transfer Function
C Speed of sound (m/s)
F Complex amplitude of upstream wave(m/s)
Ĝ Complex amplitude of downstream wave (m/s)
γ ratio of specific heats
k= ω / c Acoustical wave number (1/m)
M Mach Number(-)
N Interaction Index (-)
p' Acoustic Pressure (Pa)
φ Phase Angle (rad)
Q Heat Release (W)
Ρ Density(kg/m3)
Σ Time Delay Distribution (s)
Τ Temperature(K)
Τ Convective Time delay(s)
U Mean velocity(m/s)
Ú Acoustic Velocity (m/s)
Ω Angular Frequency
Ø Diameter
Chapter 3: Understanding the Single Burner Test Rig
3.1 Experimental Setup
The single burner test rig with its main components is shown in figure 2. The compressed air
enters through a mass flow controller providing the desired air mass flow rate. An electric
Preheater (32 KW) is used to preheat the air up to 500° C. Primary Fuel Injection into the
main air stream using small venturi nozzles is positioned well upstream of the combustor and
provides a perfectly premixes fresh gas.
The fuel used for this investigation is Natural Gas with 98.04 Mol.% methane. Speed Control
Sirens are placed up and downstream of the burner to provide two source forcing with
sufficient amplitude. As the upstream siren directly modulates the fresh gas mixture, a bypass
valve is used to control the level of acoustic excitation. The thermally insulated and
segmented Ø 120 mm plenum of 1500 mm length is equipped with microphone ports to
measure the upstream acoustic field.
The EV Burner located at the end of the plenum consists of two slightly shifted half cones
providing slots for the mixture to enter and induce swirl to the flow. Combustion occurs in an
air cooled 90 mm square combustion chamber of 700 mm length also equipped with multiple
micro phone ports. Quartz glass windows on the two opposite sides provide optical access.
Finally, a downstream exhaust system with a flexible acoustic boundary condition completes
the setup. A low reflection or even anechoic boundary condition is desirable for making
precise acoustic measurements. For this a perforated end plate is required. Up to four water
cooled condenser microphones(G.R.A.S. 40 BP-¼-inch) used with preamplifier(¼-inch,type
26 AL), with a large dynamic range and frequency response are mounted both up and
downstream of the burner. The dynamic pressure is measured using a sample and hold data
acquisition system from a multi-channel input-output Board (333 kHz, 12 bit) at a sampling
rate of 10 kHz. In parallel the heat release fluctuations are measured with a OH* -
chemiluminescence which peaks at a wavelength of 308 nm. For Optical Characterization of
the flame, i.e. flame structure and flame length, a high speed camera (APX Intensified) has
been used. A sequence of five hundred images is used to get the ensemble averaged image to
characterize each static operating point.
3.2 Test Rig working Conditions
The operating range of the test rig was chosen to closely meet the start up conditions in real
aero engines. The test rig can be operated within the following boundary conditions:
Air mass flow .
Thermal power of the flame
Static combustor pressure p: atmospheric (open combustor exit) up to 2 bar (choked
operation with outlet nozzle)
Optional air preheating
The fact that rumble is observed at low combustor pressures (1-4 bar abs.) is favourable for
the investigation at atmospheric pressure conditions. This permits maximum optical access
into the primary combustor zone and the high degree of adaptability to the requirements of
Exhaust
Downstream Siren Unit
Water Cooled Microphones
Secondary Fuel supply
Upstream Siren Unit
Bypass Valve
Primary Fuel supply
EV5 Burner
Air Supply12 Bar
Cooling Air
m
P
Air Flow Controller
Combustion Chamber
Flexible acoustic Boundary Condition
Plenum
Diagram 1. Schematic of the Test Rig
measurement techniques employed. A modular test rig design was realized, which provides
the option of tuning the test rig acoustics.
Some interesting features of the test rig facility are:
The mixing air can be optionally switched off.
Thermocouples are located in the bypass duct and in the vicinity of the injector inlet to
capture the air temperatures at the combustor front panel and the mixing air.
A third thermocouple located at the pre-heater exit monitors the pre heaters
performance and prevents overheating.
The mean and oscillating pressure losses over the injector are detected by pair of
pressure taps and microphone ports close to the burner inlet and outlet.
Optional microphone access is provided in the supply tube and in the combustion
chamber for measuring the axial distribution of acoustic pressures.
3.3 Various Components of test rig
3.3.1 Air/ fuel supply
The air is supplied from the centralized pressurized air supply unit, TUM at a pressure of 12
bars. The air used for combustion is filtered and the air mass flow is controlled by a controller
of Type Bronkhorst In flow F206BI-FBD-99V.
Figure 2. Air supply with air mass flow controller
3.3.2 Pre-heater
An electrical pre-heater (32 kW) is placed upstream of the siren valve and it is capable of
providing maximum air temperatures of 800 K. Its temperature is always under supervision
by sensors and is also provided with an integral pre-heater emergency air cooling facility.
Figure 3. Air pre-heater
3.3.3 Fuel Supply
Fuel supply is from the central infrastructure of the lab facility. Natural gas at 5 bars is
supplied to the test rig through a electronic mass flow controller that can be controlled by the
control computer. Fuel can be supplied either at the plenum or the combustor entry, or can be
mixed in piloting according, as per choice and demand of the experiment. This arrangement is
provided through swagelok valves.
Figure 4. fuel injection in the plenum
3.3.4 Siren:
The test rig design features a siren to accomplish the acoustic excitation. Compared to
speakers, which are also commonly used in thermo acoustic research, e. g. (PGW98), the
attainable forcing amplitudes of sirens are particularly higher at low frequencies. This is
relevant with respect to the intended experiments in the low frequency regime including
atomisization.
The siren excites an acoustic field by deterministically modulating the air supply of the
combustor front panel containing the air blast injector.
The excitation unit of a siren consists of a rotor stator combination. Following the principle of
hole sirens, the waveform of the generated acoustic perturbation is assumed to correlate with
the open area of the two shapes passing each other at a specified velocity. The velocity
determines the excitation frequency obtained and can be set by the rotation speed of the siren
shaft carrying the rotor blade. The siren features six rectangular orifices, being equally
distributed on the rotor, and six double sine shapes in the stator plate. The shapes have been
adapted to account for the finite reference radii.(r=30 mm) of both the stator and the rotor
disk. They are shown in figure 3.4. The axial clearance between the rotor and the stator is
approximately 0.3 mm. This configuration has shown to provide high signal quality with
dominating ground move and low noise level.
Figure 5. Upstream siren for wave generation in plenum
Figure 6. Second siren for downstream wave triggering
3.3.5 Plenum/ Supply tube
The siren is connected to the combustor by a supply tube. The supply tube aims at
establishing a well-defined one-dimensional acoustic state upstream of the burner inlet by
homogenizing the jet flow exiting the siren orifices. Furthermore, the supply tube can be used
as a resonator to achieve higher excitation amplitudes.
The supply tube consists of a cylindrical duct with a inner diameter of 124 mm. In order to
achieve an axial one dimensional acoustic field inside the duct, the transverse modes must be
suppressed. Three exchangeable segments, allow the flexible adaptation of supply tube length.
The overall length of supply tube; thus, can be varied from 850 mm to 1450 mm. Ten equally
spaced microphone ports are welded to the tube in the far field of the siren and can used to
determine the local acoustic pressures. The distance between the adjacent ports is 60 mm.
Figure 7. Plenum/ supply tube with microphone slots
3.3.6 Combustion Zone
The model combustor accommodates the injector, the front panel and the combustion
chamber. It has a squared cross section and measures 90mm x 90mm x 90mm, comparable to
dimensions of a single sector in an annular aero engine combustor. The construction consists
of a fixed frame accommodating exchangeable plates, which allows the adaptation to the
requirements of the measurement techniques applied. The primary combustion zone and the
upstream section of the dilution are optically accessible by exchangeable quartz glass
windows, The viewed area measures 150 mm x 80 mm. Pressure transducers and dynamic
temperature probes can be located along the entire length of the combustor. Ignition is
achieved using a high performance spark plug of type Champion G54V.
Figure 8. Combustion chamber
3.3.7 Burners
The burner on which we are experimenting is known as EV5 burner (Provided by ALSTOM)
and has the unique property of flame instabilization in free space near the burner outlet
utilizing the sudden breakdown of a swirling flow, called vortex breakdown. The swirler is of
exceptionally simple design, consisting of two halves of a cone, which are shifted to form two
air slots of constant width. Gaseous fuels are injected into the combustion air by means of air
distribution tubes comprising of two rows of small holes perpendicular to the inlet ports of the
swirler. Complete mixing of fuel and air is obtained shortly after injection.
Figure9. The EV5 burner (not in use )
3.3.8 Exhaust
Figure 11. The exhaust system
When natural gas fuel is combusted in the test rig, the hot combustion product gases that are
formed are called flue gases. Those gases are generally exhausted to the ambient outside air
through chimneys or industrial flue gas stacks (sometimes referred to as smokestacks).
The combustion flue gases inside the chimneys or stacks are much hotter than the ambient
outside air and therefore less dense than the ambient air. That causes the bottom of the vertical
column of hot flue gas to have a lower pressure than the pressure at the bottom of a corres-
ponding column of outside air. That higher pressure outside the chimney is the driving force
that moves the required combustion air into the combustion zone and also moves the flue gas
up and out of the chimney. That movement or flow of combustion air and flue gas is called
"natural draught/draft", "natural ventilation", "chimney effect", or "stack effect". The taller the
stack, the more draught or draft is created. To further fasten this process, an exhaust fan is fit -
ted midway that further pushes out the gas. Care is taken that the area of exhaust gas release
is safely away and disposes the gas into free air efficiently and carefully. The Chimney fan
can be shut On/off as from the touch panel near the control PC.
3.3.9 Water Cooling
To prevent the microphones from the excessive heat released in the combustion chamber and
to maintain the temperature of the combustion chamber, a water flow is supplied to the
microphone holding tube on the top side of the secondary combustor zone. This receives
water supply from the central lab infrastructure facility and is capable of keeping microphones
around 65 degree Celsius that is the safe operating range for a microphone; still when the
combustor temperature may be reaching 1700 degree Celsius.
Figure12. Water cooling
3.3.10 Air cooling
To cool down the combustor during the experiment, so that the temperature does not rises
above operating limits of combustor zone, jets of air are gushed over the primary and the
secondary combustor zone to convectively carry the heat from the outside of the combustor
zone; and still not disturbing the flame. There is a facility of 10 such air jets, but in the
experiment; 8 are actually utilizes, 5 to the primary combustor (2 from top,2 for both sides
and 1 from below) and 3 to the secondary combustor(2 for both sides and one from below.)
Figure 13. Air cooling supply knobs Figure 14.Air cooling on the secondary combustor
CHAPTER 4: Calibration of Microphones
4.1 Microphones usage in thermo acoustics
Acoustic measurements involve the measurement of sound pressure level or the generation of
a sound field, or both. Typical examples are noise measurements, loudspeaker measurements,
microphone measurements and measurements on systems like hearing aids and mobile
phones.
A sound level meter, for example, is supposed to directly display the sound pressure level in
dB SPL (dB referred to 20 μPa sound pressure). If an audio analyzer like UPV, which
measures voltage, is used for this purpose, the sensitivity of the microphone is given by
Eqn.3.Sensitivity of a microphone
has to be determined, wherein is the r.m.s value of the microphone output voltage and
is the r.m.s value of the sound pressure which produced the output voltage. The value of the
sound pressure is obtained by dividing the measured voltage by this sensitivity value.
The measurement with which the value of the microphone sensitivity is obtained is commonly
called microphone calibration. The sound pressure level for the microphone calibration in the
multi microphone method is generated with a so-called acoustic calibrator (In the Experiment,
we use sine wave signal with defined frequency range (from 0 to 600 hertz with steps of 10
Hz in the region 0-240 Hz and steps of 20 in the range 260-620 Hz) and defined sound
pressure (usually10 Pa/ mV).
As measurement microphones are small and have a well-defined mechanical structure, the
sensitivity is frequency-independent within a certain frequency range. Therefore calibration at
a single frequency at a time is sufficient.
Certain measurements require the generation of a defined sound pressure at a certain point. In
order to be able to set a desired sound pressure level in the generator of the UPV, the
sensitivity of the loudspeaker
Eqn.4. Sensitivity of a loudspeaker
has to be determined, wherein is the sound pressure generated by the loudspeaker at the
pre-defined measurement point, and is the generator output voltage which produced the
sound pressure.
Figure 15. A microphone with the adaptor jacket and cooling pipes
4.2 Essentials of Microphone Calibration
4.2.1 Waveform Generator:
An arbitrary waveform generator (AWG) is a piece of electronic test equipment used to
generate electrical waveforms. These waveforms can be either repetitive or single-shot (once
only) in which case some kind of triggering source is required (internal or external). The res-
ulting waveforms can be injected into a device under test and analyzed as they progress
through the device, confirming the proper operation of the device or pinpointing a fault in the
device. Unlike function generators, AWGs can generate any arbitrarily defined wave-shape as
their output. The waveform is usually defined as a series of "waypoints" (specific voltage tar-
gets occurring at specific times along the waveform) and the AWG can either jump to those
levels or use any of several methods to interpolate between those levels.
For example, a 50% duty cycle square wave is easily obtained by defining just two points: At
t0, set the output voltage to 100% and at t50%, set the output voltage back to 0. Set the AWG to
jump (not interpolate) between these values and the result is the desired square wave. By
comparison, a triangle wave could be produced from the same data simply by setting the
AWG to linearly interpolate between these two points.
Figure 16. Arbitrary Waveform generator
Because AWGs synthesize the waveforms using digital signal processing techniques, their
maximum frequency is usually limited to no more than a few megahertz. The output con-
nector from the device is usually a BNC connector and requires a 50 or 75 ohm termination.
AWGs have various means of modulating the output waveform, and often contain the ability
to automatically and repetitively "sweep" the frequency of the output waveform (by means of
a voltage-controlled oscillator) between two operator-determined limits. This capability
makes it very easy to evaluate the frequency response of a given electronic circuit. Some
AWGs also operate as conventional function generators.
The waveform generator used in the process is from Agilent model number 33220A LXI and
has a capacity to generate sine, square, ramp, pulse, noise and arbitrary signal waveforms up
to a maximum frequency of 20MHz.
4.2.2 Microphones
The Piezoelectric effect is an effect in which energy is converted between mechanical and
electrical forms. It was discovered in the 1880's by the Curie brothers. Specifically, when a
pressure (piezo means pressure in Greek) is applied to a polarized crystal, the resulting
mechanical deformation results in an electrical charge. Piezoelectric microphones serve as a
good example of this phenomenon. Microphones turn an acoustical pressure into a voltage.
Alternatively, when an electrical charge is applied to a polarized crystal, the crystal undergoes
a mechanical deformation which can in turn create an acoustical pressure. An example of this
can be seen in piezoelectric speakers in nowadays PC’s.
Figure 17. Microphone with cooling adaptor showing the microphone at the tip, receiver
at the middle and bronze cooling adapter.
Electrets are solids which have a permanent electrical polarization. (These are basically the
electrical analogs of magnets, which exhibit a permanent magnetic polarization). Figure (cite
figure number) shows a diagram of the internal structure of an electret. In general, the
alignment of the internal electric dipoles would result in a charge which would be observable
on the surface of the solid. In practice, this small charge is quickly dissipated by free charges
from the surrounding atmosphere which are attracted by the surface charges. Electrets are
commonly used in microphones.
Figure 18: Internal Structure of an electret Figure 19: A sensor based on the piezoelectric effect
Permanent polarization as in the case of the electrets is also observed in crystals. In these
structures, each cell of the crystal has an electric dipole, and the cells are oriented such that
the electric dipoles are aligned. Again, this results in excess surface charge which attracts free
charges from the surrounding atmosphere making the crystal electrically neutral. If a
sufficient force is applied to the piezoelectric crystal, a deformation will take place. This
deformation disrupts the orientation of the electrical dipoles and creates a situation in which
the charge is not completely cancelled. This results in a temporary excess of surface charge,
which subsequently is manifested as a voltage which is developed across the crystal.
Therefore, piezoelectric crystals act as transducers which turn force, or mechanical stress into
electrical charge which in turn can be converted into a voltage. Alternatively, if one was to
apply a voltage to the plates of the system described above, the resultant electric field would
cause the internal electric dipoles to re-align which would cause a deformation of the material
4.2.3 Microphone Cooling
Water cooling is a method of heat removal from microphones and to keep them at a
reasonable operating temperature while in use at high temperatures.. As opposed to air
cooling, water is used as the heat transmitter. The main mechanism for water cooling is
convective heat transfer.
In the microphone cooling system, water flows from the tank ;forced by the pump to run
through the tubes in and out the microphone cover tube where it whirls around the
microphone, absorbs the heat of the microphone and then carries it to the water tank. A
thermostat has been provided to maintain the temperature of water cooling to a constant value
so as to have uniform cooling. The microphone cooling system is capable to handle 8
microphones at a time and has been custom built by TUM, Munich for thermo-acoustic
research purpose. Water cooling should be started at least 15 minutes before the experiment to
bring all the microphones at a common temperature.
Figure 20. Microphone water cooling system
4.2.4 Microphone amplifier
Generally, an amplifier is any device that changes, usually increases, the amplitude of a
signal received from the microphones. The model being used in the experiment is a
nexus model from Brüel and Kjaer mfg. The amplification of the microphone response
from the amplifier was manually set to 10V/Pa and the response coefficients of the
microphones were fed in the amplifier manually with its GUI LCD screen.The amplifier
is pretty good in noise reduction and signal conditioning, that is clear from its technical
specifications:
Input signal range: +/-10 V peak
-3dB bandwidth: 100 kHz @ 30 dB gain
Gain accuracy: +/-1 dB
S/N (typical): 110 dB (20 Hz to 30 kHz at 9.9V)
Noise floor: 9.2 μV rms
THD (typical): < 0.002% (1 kHz tone, +/-7 V peak)
Spectral variation: 3 dB (20 Hz to 30 kHz, 45 dB gain)
Input impedance: 600 ohms
Output signal range: +/-10 V peak
Output impedance: 5 ohms
Bias voltage: 10 V, 150 mA max (XLR input)
Power and communication: zBus required
In popular use, the term today usually refers to an electronic amplifier, often as in audio
applications. The relationship of the input to the output of an amplifier — usually expressed
as a function of the input frequency — is called the transfer function of the amplifier, and the
magnitude of the transfer function is termed the gain. A related device that emphasizes
conversion of signals of one type to another (for example, a light signal in photons to a DC
signal in amperes) is a transducer, or a sensor. However, a transducer does not amplify power.
The Microphone Amplifier is a two-channel high gain, low noise preamplifier with both
phone and balanced XLR microphone inputs for optimum impedance and noise
characteristics. The Microphone Amplifier features variable gain from 10 dB to 55 dB in 5 dB
steps, a toggle switch providing 20 dB of additional gain (maximum amplification of 5600x),
and a bias switch for microphones requiring a bias voltage. Four BNC outputs provide easy
connection to any TDT System 3 device.
Figure21. Microphone amplifier
4.2.5 Acoustic calibrator/ Sound generator
Like a loud speaker, this device receives signal inputs from the waveform generator and
converts them into sound wave signals precisely inside its closed end tube. Microphones are
fitted by firmly gripping them in the slots of the calibration tube which is at the symmetric
locations. An acoustic calibration device for intensity measuring systems comprises for four
pressure microphones to be calibrated without using an anechoic chamber. The calibration
device comprises four cavities interconnected by means of one or several acoustic resistance
elements. One of the pressure microphones is to be inserted into one of the cavities and the
other three microphones are to be inserted into the other three cavities symmetrically located.
A sound source is connected to one of the cavities. In connection with the subsequent cavity
the acoustic resistance thus forms an acoustic RC-link providing a phase shift proportional to
the frequency corresponding to the conditions in the free field. By a suitable dimensioning of
the RC-link a phase shift corresponds to the phase shift over a distance of e.g. 50 mm in the
free field. The sound source is able to generate either white noise or pink noise depending on
whether measurements are performed over fixed frequency intervals or relative frequency
intervals.
Figure 22. Acoustic calibrator
4.2.6 Wave Monitoring Device
The digital storage oscilloscope, or DSO for short, is the preferred type of oscilloscope for
most industrial applications. It replaces the unreliable storage method used in analogue
storage scopes with digital memory, which can store data as long as required without
degradation. It also allows complex processing of the signal by high-speed digital signal
processing circuits.
The vertical input, instead of driving the vertical amplifier, is digitised by an analog to digital
converter to create a data set that is stored in the memory of a microprocessor. The data set is
processed and then sent to the display, an LCD flat panel. DSOs with color LCD displays are
common. The data set can be sent over a LAN or a WAN for processing or archiving. The
screen image can be directly recorded on paper by means of an attached printer or plotter,
without the need for an oscilloscope camera. The scope's own signal analysis software can
extract many useful time-domain features (e.g. rise time, pulse width, amplitude), frequency
spectra, histograms and statistics, persistence maps, and a large number of parameters
meaningful to engineers in specialized fields such as telecommunications, disk drive analysis
and power electronics.
Digital storage also makes possible another unique type of oscilloscope, the equivalent-time
sample scope. Instead of taking consecutive samples after the trigger event, only one sample
is taken. However, the oscilloscope is able to vary its timebase to precisely time its sample,
thus building up the picture of the signal over the subsequent repeats of the signal. This
requires that either a clock or repeating pattern be provided. This type of scope is frequently
used for very high speed communication because it allows for a very high "sample rate" and
low amplitude noise compared to traditional real-time scopes.
Figure 23. Digital storage electroscope
4.2.7 Processing Software/ Program
Labview
A Labview program continuously monitors the response of the microphones under calibra-
tion, collects the on line data and stores them as *.fft files per frequency step. It contains all
the basic functions to monitor and control almost every part of the experiment being carried
out. The version what we use in the microphone calibration is version 5 and pretty old com-
pared to the version available nowadays. But still, it efficiently controls the process and gath-
ers Data.
Figure 24. Labview GUI on the process control computer.
4.3 Procedure
As a loudspeaker has a far more complex mechanical structure than a measuring microphone,
and radiation effects additionally influence the loudspeaker sensitivity depending on the
frequency and on the location of the measurement point, the sensitivity of a loudspeaker
usually is frequency dependent. For this reason the calibration of a loudspeaker consists of
two steps:
1. Measurement of the absolute sensitivity at one frequency
2. Measurement of the frequency response relative to this frequency.
In order to generate a defined sound pressure at the measurement point, the generator output
voltage has to be set to the desired sound pressure divided by the loudspeaker sensitivity, and
to be corrected by the inverse frequency response (equalization).
It has to be observed that the loudspeaker calibration is only valid for the point of the
calibration, due to the propagation properties of the sound wave.
Note that an equalization in the sense of a frequency dependent amplitude correction is only
possible with signals which are defined in the frequency domain, like (swept) sine-wave and
multi-sine signals. For the equalization of a complex signal defined in the time domain like
speech, a filter is required. For numerical function results, the microphone sensitivity can be
entered in the Function Configuration Panel as Reference Value.
4.3.1 Calibration:
The microphone calibration is available for each of the two input channels. “Microphone
Calibration” starts the calibration routine. Once the calibration routine has successfully
finished, a file can be specified to store the results. This file name is simultaneously entered
into the “Calibration File” text box which specifies the file from which the calibration value to
be entered into the UPV setup is taken.
To calibrate a microphone, the arbitrary wave generator is connected to the acoustic
calibrator. The microphones are fitted in the symmetrically located slots of the calibrator with
one microphone being the reference value for calibration. The output of the microphones is
connected to the microphone amplifier. This microphone amplifier finally sends the data to
the computer via BNC connectors. A digital oscilloscope monitors the waveforms, both from
the arbitrary wave generator and the one of the microphones. The Labview program
continuously monitors the functioning of the whole experiment.
Step1. The microphone cooling is turned on and kept to stabilize temperature diversities for
10-15 minutes.
Step2. The microphones are checked for any loosened contacts and a firm grip with the slots.
Care is taken to see if they are properly inserted.
Step3. Microphone coefficients and sensitivity are fed into the microphone amplifier
manually.
Figure25. Section showing microphone amplifications manually fed in the GUI
Step 4. An initial test run is activated by feeding a signal output from the arbitrary wave
generator to the calibrator and receiving microphone responses.
Step 5. Working directory, microphone amplifications, file name, output data type, number of
loops and required frequency as same in AWV is fed to the LABVIEW Program.
Figure 25. Section to feed in file name, file type and working directory
Step 6. All the 4 BNC connector ports are checked on the oscilloscope for smooth responses.
Step 7.The loops are recorded by pressing the microphone recording button on Labview. The
data is saved as *.prc , *.stp and *.fft file formats.
Step 8. After 5 loops, the output file saved in the working directory is checked. If reasonable,
manual feeding of data starting from 10 Hz is started on the system and file names are
accordingly changed instead of test readings.
Figure 26. Section showing no. of loops currently processed.
Step 9. After 10-240 in steps of 10 Hz, and 240-260 in steps of 20 hz, the total data aquired is
transferred to the main processing computer with MATLAB installed for further post
Processing.
4.3.2 Appending the Data
The data obtained from the Labview Program is in the form of a set of *.fft files generated by
the lab view software because of the conversion of output records into Fourier transform files.
This enables files to be easily read by the data post processing software MATLAB.
The set of the files as named from ‘10’ to ‘620’ with a suffix .fft are appended into two .fft
files; one with a step of 10 hertz difference between data ‘n’ and ‘n-1’ from 10 to 240 Hertz,
and the other with a frequency step of 20 hertz for the values from 260 till 620.
The following MATLAB Program was used to append the files. The example used here is to
append files of the frequency 260 hertz to 620 Hertz into a common file known as
‘test_260to620hz.fft’.The same program is later customized accordingly for other data
readings.
To append the obtained .fft files into one common File
%This Routine appends the binary data obained for each single frequency, into a common file and saves as a IEEE standard format, to able read in Mathematica routines
clcclear all %*********************************************************** % id_freq= [10:10:240];% id_freq= [20:20:40]; id_freq= [260:20:620];% id_freq= [260:20:620];% id_freq= [10:10:600];
nb_freqs=size(id_freq,2); % TUM\ICLEAC_test_rig\Ev5\measurements\mic_calibration\Mic_cali_121207\set3_CIKL\'; %working directoryworkdir= '\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Mic_calibn_230408\Calibrn_G_I_J_E\'; % to feed in the working directoryappdata=[]; for k=1:nb_freqs %frequency loop (if desired) filenamein = strcat(workdir,int2str(id_freq(k)),'.fft'); fidin=fopen(filenamein, 'rb', 'b'); %big endian 32 bit-float! if fidin > 0 [data,in] = fread(fidin, inf, 'float32'); appdata=[appdata data']; end end filenameout = [workdir,'test_260to620Hz.fft'];
fidout = fopen(filenameout,'w'); out = fwrite(fidout, appdata, 'float32'); %to finally write the output in a IEEE standard format, consistent with the mathematica! fclose all;
After appending the data values into one common file, this file is fed to another MATLAB
Program as a input where it is post processed to get the out put graphs of Absolute Amplitude,
Absolute phase angle, Relative amplitude and relative phase difference Respectively. The
following program locates the working directory and the working file,generates a output
directory, plots the graph and saves them into the output directory along with the a .txt
numerically writing the data values in a text file type.
%Program to get the calibration coefficients from the standard .fft file generated from labview of the ICLEAC labour.The routine also plots the relative amplitudes and phases of each mic with respect to the selected refernace mic!%06:03:08 Last modified by Piyush Raj clcclose allclear all %***********************************************************%User Input
workdir= '\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Mic_calibn_230408\Calibrn_G_F_H_N\'; %working directory
filenamein = [workdir,'test_260to620Hz.fft'];
path_out=[workdir,'test_260to620HZ_output\']; if isdir(path_out)==0 mkdir(path_out);end scan= 10000; %scan frequency (per channel)number_of_samples=10000;chan= 200000; %chan frequency (between channels)seq=[2,3,4,5,0,0,0,0,0,1];nb_mics=4; %no. of Mics used ref=1; % select the reference MIC%id_freq= [10:10:240]; %excitation frequencies id_freq= [260:20:620]; %excitation frequencies nb_freqs=size(id_freq,2); freqstep=id_freq(2)-id_freq(1);pic_save = 'y' ;phasecorr=1;filterfft=0; %flag whether .fft data should be filtered with respect to the reference signalfiltercta=0;down=2.0; %the ref signal voltage lies in between up and down, i.e., 1.6 to 3V for the case of up stream forcing up=3; Refmic_pistophone_factor=1; appdata=[];calibcoef=[]; nb_cols_fft=13; fidfft=fopen(filenamein, 'r'); %for reading the data obtained from parallel board, using HP-VEE / Fischer data/IEEE standard[fftdata,lfftdata]=fread(fidfft, [nb_cols_fft inf], 'float32'); %Signals FFTfftdata=fftdata'; collphases=zeros(nb_freqs,nb_cols_fft); %initialising.....collamps=zeros(size(collphases));[collamps,collphases]= getfftsubsmeantoref_EV5(filenamein,chan,seq,scan,id_freq,filtercta,filterfft,down,up); refamp=ref+3;refphase=ref+3; for k=1:nb_freqs %frequency loop (if desired) calibcoef(k,1)=id_freq(k);% collamps(k,refamp)=collamps(k,refamp)/Refmic_pistophone_factor; for i=2:1:nb_mics+1calibcoef(k,2*i-2)=collamps(k,i+2)/collamps(k,refamp); % Amplitude factor in alternative columns
calibcoef(k,2*i-1)=collphases(k,i+2)-collphases(k,refphase); % Phase factor in alternative columns if phasecorr if calibcoef(k,2*i-1)> pi fac1=ceil((calibcoef(k,2*i-1)+pi)/(2*pi))-1; %How many times must (2*pi) be substacted in order to match the range calibcoef1(k,2*i-1)=calibcoef(k,2*i-1)-2*fac1*pi; elseif calibcoef(k,2*i-1)< - pi fac2=floor((calibcoef(k,2*i-1)-pi)/(-2*pi)); %How many times must (2*pi) be added in order to match the range calibcoef(k,2*i-1)=calibcoef(k,2*i-1)+2*fac2*pi; end end end endfor i=1:nb_mics for j=1:nb_freqs Abs_Amp(i,j) = collamps(j,3+i); Phases(i,j)= collphases(j,3+i); Rel_Amp(i,j) = calibcoef(j,2*i); Rel_Phase(i,j) = calibcoef(j,(2*i)+1); end end %Plot absolute Amplitudesfigure(1) plot(collamps(:,2),Abs_Amp) xmin = min(id_freq); xmax = max(id_freq); ymin = 0; ymax = 720; axis([xmin xmax ymin ymax]) title('Absolute Amplitude'); xlabel('Frequency [Hz]'); ylabel('Amplitude [mbar]'); h = legend('Mikrofon {G REF}','Mikrofon {F}','Mikrofon {H}','Mikrofon {N}',1);% if pic_save == 'y' saveas(gcf,[path_out,'Abs_amp'],'tif') end%Plot Phase angles figure(2) plot(collamps(:,2),Phases) xmin = min(id_freq); xmax = max(id_freq); ymin = -pi; ymax = pi; axis([xmin xmax ymin ymax]) title('Phases'); xlabel('Frequency [Hz]'); ylabel('Phase angle [rad]'); h = legend('Mikrofon {G REF}','Mikrofon {F}','Mikrofon {H}','Mikrofon {N}',1); if pic_save == 'y' saveas(gcf,[path_out,'Phases'],'tif') end%Plot relative Amplitudes
figure(3)plot(collamps(:,2),Rel_Amp) xmin = min(id_freq); xmax = max(id_freq); ymin = 0.2; ymax = 1.5; axis([xmin xmax ymin ymax]) title('Relative Amplitudes'); xlabel('Frequency [Hz]'); ylabel('Relative amplitude [-]'); h = legend('Mikrofon {G REF}','Mikrofon {F}','Mikrofon {H}','Mikrofon {N}',1); if pic_save == 'y' saveas(gcf,[path_out,'Relative Amplitudes'],'tif') end
%plot the relative Amplitudesfigure(4) plot(collamps(:,2),Rel_Phase) xmin = min(id_freq); xmax = max(id_freq); ymin = -1.5; ymax = 1.5; axis([xmin xmax ymin ymax]) title('Relative Phases'); xlabel('Frequency [Hz]'); ylabel('Phase angle [rad]'); h = legend('Mikrofon {G REF}','Mikrofon {F}','Mikrofon {H}','Mikrofon {N}',1); if pic_save == 'y' saveas(gcf,[path_out,'Relative Phases'],'tif') end dlmwrite([path_out,'_relative_amplitudes_and_phases_to_mic',sprintf('%.1d',ref),'.txt'],calibcoef,'delimiter','\t','precision','%4f');
Due to nonlinearities of loudspeakers there may be deviations in sensitivity and frequency
response between calibrations executed at different levels. It is also possible to perform
calibrations for a set of different levels and store them under different file names. After the
level has been confirmed, the calibration starts with first determining the absolute sensitivity
at 1 kHz, then measuring the relative frequency response. In a second run the obtained
equalization is checked and corrected before a third sweep is done to verify the result of the
calibration. After completion of the measurement, a file name has to be specified for storing
the calibration values. Subsequently the UPV is returned to local control and the calibration
tool window is minimized in order to show the sweeps:
4.4 Results
After obtaining the graphs for each of the four appended .fft files, we finally obtain the graphs
for all the seven microphones named G, I, J, E, F, H, N with G being the reference
microphone for subsets GIJE and GFHN. Following are the results of Microphone
calibrations with first set showing absolute amplitudes, second set showing absolute phase
angles, third set showing relative amplitudes and fourth set showing relative phase differences
between the microphones.
Subset 1.1: Reference Microphone is G; others are I, J and E. Frequency Range is 0-240 Hz.
Subset 1.2: Reference Microphone is G; others are I, J and E. Frequency Range is 260-620 Hz.
Subset 2.1: Reference Microphone is G; others are F, H and N. Frequency Range is 0-240 Hz.
Subset 2.2: Reference Microphone is G; others are F, H and N. Frequency Range is 260-620 Hz.
Graph Plots 1-4 respond to subset 1.1; 5-8 respond to subset 1.2, 9-12 respond to Subset 2.1, and 13-16 respond to 2.2 respectively.
Graph Plots 17-20 are regenerated plots for subset GIJE as the experiment were repeated to
investigate faulty behaviour of Microphone I in the range 0-160 Hertz. After re-doing the
experiment in the zone 0-160 Hz for GIJE, the behaviour of I comes out to be normal
indicating manual error in the previous reading.
Plot 1 Absolute Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 2. Absolute Phase Plot for subset GIJE in the range 0-240 Hz.
Plot 3. Relative Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 4 .Relative Phase Plot for subset GIJE in the range 0-240 Hz.
Plot 5. Absolute Amplitude Plot for subset GIJE in the range 260-620 Hz.
Plot 6. Absolute Phase angle Plot for subset GIJE in the range 260-620 Hz.
Plot 7 . Relative Amplitude Plot for subset GIJE in the range 260-620 Hz.
Plot 8. Relative Phase Plot for subset GIJE in the range 260-620 Hz.
Plot 9. Absolute Amplitude Plot for subset GFHN in the range 0-240 Hz
Plot 10. Absolute Phase Angle Plot for subset GFHN in the range 0-240 Hz
Plot 11. Relative Amplitude Plot for subset GFHN in the range 0-240 Hz
Plot 12. Relative Phase Angle Plot for subset GFHN in the range 0-240 Hz
Plot 13. Absolute Amplitude Plot for subset GFHN in the range 260-620 Hz
Plot 14. Absolute Phase Angle Plot for subset GFHN in the range 260-620 Hz
Plot 15. Relative Amplitude Plot for subset GFHN in the range 260-620 Hz
Plot 16. Relative Phase Angle Plot for subset GFHN in the range 260-620 Hz
Plot 17. Absolute Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 18. Absolute Phase Angle Plot for subset GIJE in the range 0-240 Hz.
Plot 19. Relative Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 20. Relative Phase Angle Plot for subset GIJE in the range 0-240 Hz.
4.5 Conclusion
From the graph plots 1-4 for GIJE subset, we see that there is an abnormal behaviour of
microphone I in the range of 0-160 Hz. As the abnormal behaviour suddenly vanishes after
160 Hz, it means there must be some manual or other error and gives a hint a hint that may be
microphone I was not inserted properly in the calibration tube slots.
The calibration is repeated again in the zone 0-160 Hz, and the data is replaced to plot new
graphs, here the microphone ‘I’ behaves normally indicating that our assumption about it was
correct.
Due to lack of repeatability, there is a small deviation of phase and amplitudes of all the
microphones from the reference microphone in the Graphs 17-20.
The Microphones work exceptionally well in the range 260-620 Hz, for GIJE indicated by the
Graphs 5-9.
During post processing of the results, it was found that the .fft file for subset GFHN was not
properly recorded for the frequency step 170 Hz. This lead to inability of obtaining further
results in the region 0-240 Hz.
Experiment was re- conducted, for GFHN, this time only for 170 Hz; the data was obtained
and replaced in the original database. After plotting it for 0-240 Hz; we see that there is a
small yet small irregularity peak in the graph 9 in the set graph plots 9-12 for the subset
GFHN in the range 0-240 Hz. This indicates lack of good repeatability of microphones during
calibration process.
The Microphones work exceptionally well in the range 260-620 Hz, for GFHN indicated by
the Graphs 5-9.
CHAPTER 5: Obtaining the Operating Domain of the Test Rig
(Still working on)
4.1 Experiment Objective
Obtaining the Operating Domain of the Test Rig by obtaining the microphone response on
flame instability comparing the internal and internal piloting the Gas into the Combustor.
4.2 Research Methodology
4.3 Data acquisition
4.3 Results
4.4 Discussions
4.5 Interpretations
REFERENCES
1. Wolfgang Polifke, Combustion Instabilities, Lehrstuhl für Thermodynamik, TU München, Germany, March 18th, 2004 (Corrected printing Oct. 15th 2007)
2. Dr. R. I. Sujith, ‘Acoustic Instabilities in Aerospace Propulsion’, Associate Professor, Department of Aerospace Engineering, Indian Institute of Technology Madras, Compiled and Collated by Vipluv Aga, April 2003
3. Alemela, P.R., Fanaca, D., Ettner, F., C. Settelmayer, T., Schuermans, B., ‘Flame Transfer Matrices of a premixed flame and a global check with modelling and experi-ments’, GT 2008-50111,Submitted for Proc. ASME Turbo Expo 2008,June 9-13, 2008, 2008, Berlin, Germany.
4. Dr.-Eng. Markus Raffel,Dr. Christian E. Willert,Dr. Jürgan,’Particle Image Veloci-metry’, ISBN 3-540-63683-8, (1998).
5. National Instruments Labview 8.0, ‘Labview User Guide’, 2007.
6. MATLAB R2007b,’MATLAB User Guide’, Version 7.0.5.042, June 15, 2007.
7. N. Rott. Damped and thermally driven acoustic oscillations in wide and narrow tubes. Z. Angew. Math. Phys., 20:230—243, 1969.
8. N. Rott. Thermally driven acoustic oscillations, part III: Second-order heat flux. Z. An-gew. Math. Phys., 26:43—49, 1975.
9. N. Rott. Thermoacoustics. Adv. Appl. Mech., 20:135—175, 1980.
10. T. Yazaki, A. Tominaga, and Y. Narahara. Experiments on thermally driven acoustic oscillations of gaseous helium. J. Low Temp. Phys., 41:45—60, 1980.
11. Lord Rayleigh. The explanation of certain acoustical phenomena. Nature, 18:319—321, 1878.
12. K. T. Feldman. Review of the literature on Sondhauss thermoacoustic phenomena. J. Sound Vib., 7:71—82, 1968.
13. K. T. Feldman. Review of the literature on Rijke thermoacoustic phenomena. J. Sound Vib., 7:83—89, 1968.
14. B. Zinn. Pulsating combustion. In F. J. Weinberg, editor, Advanced Combustion Meth-ods, pages 113—181. Academic, London, 1986.
15. T. Hofler, J. C. Wheatley, G. W. Swift, and A. Migliori. Acoustic cooling engine, 1988. US Patent No. 4,722,201.
16. T. J. Hofler. Thermoacoustic refrigerator design and performance. PhD thesis, Physics department, University of California, San Diego, 1986.
17. W. E. Gifford and R. C. Longsworth. Pulse tube refrigeration progress. Adv. Cryogenic Eng.,10B:69—79, 1965.
18. E. L. Mikulin, A. A. Tarasov, and M. P. Shkrebyonock. Low-temperature expansion pulse tubes. Adv. Cryogenic Eng., 29:629—637, 1984.
19. R. Radebaugh. A review of pulse tube refrigeration. Adv. Cryogenic Eng., 35:1191—1205, 1990.
20. G. Walker. Stirling Engines. Clarendon, Oxford, 1960.
21. J. Organ. Thermodynamics and Gas Dynamics of the Stirling Cycle Machine. Cambridge University Press, 1992.
22. Urieli and D. M. Berchowitz. Stirling Cycle Engine Analysis. Adam Hilger, Bristol UK, 1984.
23. P. H. Ceperley. A pistonless Stirling engine–The traveling wave heat engine. J. Acoust. Soc. Am., 66:1508—1513, 1979.
24. P. H. Ceperley. Gain and efficiency of a short traveling wave heat engine. J. Acoust. Soc. Am., 77:1239—1244, 1985.
25. T. Yazaki, A. Iwata, T. Maekawa, and A. Tominaga. Traveling wave thermoacoustic engine in a looped tube. Phys. Rev. Lett., 81:3128—3131, 1998.
26. C. M. de Blok. Thermo acoustic system, 1998. Dutch Patent: International Application Number PCT/NL98/00515. US Patent 6,314,740, November 13, 2001.
27. S. Backhaus and G. W. Swift. A thermo acoustic-Stirling heat engine. Nature, 399:335—338, 1999.
28. G. W. Swift. Thermo acoustics: A Unifying Perspective for some Engines and Refrigerat-ors. Acoustical Society of America Publications, Sewickley PA, 2002.
29. J. C. Wheatley and A. Cox. Natural engines. Physics Today, 38:50—57, August 1985.
30. J. R. Olson and G. W. Swift. Acoustic streaming in pulse tube refrigerators: Tapered pulse tubes. Cryogenics, 37:769—776, 1997.
31. N. Rott. The influence of heat conduction on acoustic streaming. Z. Angew. Math. Phys., 25:417-421, 1974.
32. P. S. Spoor and G. W. Swift. Thermo acoustic separation of a He—Ar mixture. Phys. Rev. Lett., 85:1646—1649, 2000.
33. D. A. Geller and G. W. Swift. Thermo acoustic enrichment of the isotopes of neon. J. Acoust. Soc. Am., 115:2059—2070, 2004.
CHAPTER 5_Operating Domain of Single Burner test Rig
Fig: A photograph of flame during combustion in the single burner test rig
5.1 Introduction
In this section, we conduct experiments to determine the operating limits of a lean combustion
mixture to prevent flame blow out. Combustion of the lean fuel air mixture is observed by
varying the ‘lambda ratio’ and ‘piloting’ percentage for same power produced from
combustion.
It is pointed out, in this connection, that the λ-ratio is generally
used for describing the air/fuel ratio in combustion operations.
The λ-ratio is a measure of the ratio of the air quantity introduced
into the combustion space to the air quantity required
theoretically for complete combustion. Gas Turbine Burners of
the type used in the present invention are operated, as a rule, in a
range of 1.5 λ to 1.8 λ.
Piloting is the process of injecting fuel into the combustion chamber. It can be done in the two
ways, Internal Piloting, External Piloting; or the mixture of both. In internal Piloting, the fuel
is mixed into the compressed air right at the entry of Plenum, so that it mixes thoroughly in
the plenum. In external Piloting, the fuel is mixed into the combustion chamber right into the
burner; so it gets time to mix in the air only for a short while. But due to turbulent behaviour
of combustors and high mass flow of compressed air swirl, its efficiently mixes just before
combustion.
In the experiment, the piloting percent (ratio of external piloting to internal piloting) was
varied and the flame experiments were carried out.
There were 3 microphones at the first trial; two at the upstream side and one in the
downstream side.
The experiment was carried out at approximately 50 kW power and lambda ratio was varied
manually 1.5 to 1.8 to preventing flame blow off. Flame pictures were taken using CCD Cam-
era and microphone responses were recorded for a time period of 2 seconds at a sampling rate
of 10000 Hz. For this experiment, commercially available UV flame detector CCD camera is
used to monitor the status (flame on or off) of a flame. It continuously monitors the flame and
amidst experiment, it records 100 pictures continuously and takes an average picture based on
average pixel value of 100 images.
5.2 Concept of lean blow out
Stringent environmental emission regulations have motivated changes in the design and oper-
ation of combustion processes, in particular gas combustion systems. Many developers of gas
combustion systems, such as stationary gas turbines, use some form of lean-premix combus-
tion (LPM). In LPM , fuel is mixed with air upstream of the combustion zone at deliberately
fuel-lean conditions. A significant reduction of thermal NO x formation is achieved using
LPM system. To meet the target of reducing NO x levels to under 10ppm, modern premix tur-
bine combustors must operate with a finely controlled fuel/air ratio, (equivalence ratio) near
the lean extinction limit. In practice, changes in flow splits caused by manufacturing toler-
ances or engine wear can compromise emissions performance.
Control of the combustion process at the burner can be performed by metering the flows of
fuel and oxidant, through appropriately regulated valves (electrically or pneumatically driven)
that are controlled by a programmable controller (PC). The ratio of oxidant to fuel flow is pre-
determined using the chemical composition of the fuel and of the oxidant. To be effective, the
flow measurements for the fuel and oxidant must be very accurate and readjusted on a regular
basis. Typically this situation often leads the operator to use a large excess of air to avoid the
formation of CO. In small combustion chambers, often instabilities of flames lead to flame
blow off.
Serious problems can result when flames reach an extinction limit, or blow-out. Operation
near the lean extinction limit is desired to reduce NO x emissions; however this desire must be
balanced by the risks of encountering a sudden flame extinction, or lean blow-out event. Cur-
rently, there are no commercial methods to sense when lean blow-out may be incipient. Thus,
we manually obtain the operating domain of the single burner test rig by controlling the flame
at certain operating then manually, manipulating it by decreasing and decreasing it by some
amount, keeping in mind the flame blow off and refraining from it to occur.
5.3 Test Rig Preparation
Step 1: The microphones are arranged over the test rig at required positions.
Step 2: Microphone cooling is set on and is left to stabilize its temperature for 15-20 minutes.
Step 3: The air cooling and water cooling supply is installed over the combustion chamber for
counteracting high heat build ups.
Step 4: Microphones are connected to the signal Amplifiers and BNC connections are
checked on the Oscilloscope. This is done by sending some air flow in the plenum and then
comparing the signal responses of the microphones. A crude and rough; but continuous and
less fluctuating response is acceptable.
Step 5: The CCD camera is adjusted in front of the transparent quartz glass screen of the
combustion chamber, with its view frame plane exactly perpendicular. This is a hectic task,
but needs attention and delicate care to set the camera focussed on a a test calibration plate
that is a planer rectangular tin plate, with exact fit to the combustion chamber height as its
weight and length as glass window, with markings in the X and Y direction. The camera is
said to be set if it captures the images of all the marking on the calibration plate clearly, with
the origin marked in the centre of the camera frame. Once set, the camera must not move from
that position through out the experiment. The CCD camera recording software is started and
test runs are conducted to check if everything is running smoothly and data is getting
efficiently recorded as .bmp images.
Step 6: Some air flow (around 25gps) is sent through the plenum to ensure that there are no
wage and unnecessary particles in the air flow during experiment.
Step 7: The CCD camera setup is activated for a data capture any time. This is done by
activating the camera control and Data acquisition software. The software we are using for
our purpose is Streak Star Data capturing Software. Later the data is processed on IrfanView
software as .bmp Files.
Step8: The Spark Plug is fitted in the base of the combustion chamber and is connected to the
igniting switch.
5.4 Experimental Procedure
Step1: Some air flow is provided to the test rig. Initially starting from a less value (approx 10
gps), the air flow is gradually increased in steps of 5-10 gps to a value close to 30 gps.
Step 2: Air Cooling and Combustion chamber cooling are activated along with the emergency
pre-heater cooling system. Ear Plugs are worn from Noise protection.
Step 3. The spark plug is activated and the spark is ignited at the combustion chamber. This is
monitored continuously by the process observation camera focussed at the combustion
chamber and real time video display to the process controller.
Step 4. The fuel supply is gradually added to the air in the plenum in very small increments.
This process is continued till the combustion starts. The theoretical power predictor in the
Labview Program indicates the Theoretical Power Being produced from the Combustion.
Lambda ratio is increased to an extent, that the Theoretical power
attained by combustion reaches 50 kW.
Step 5. Keeping the power produced to be fixed (approximately 50
kW), for a particular λ-ratio , the piloting is gradually varied from
0 percent (no external piloting) to higher values in some calculated
guess variations; to ensure that there is no flame blow off.
Step 6. For a particular λ-ratio and piloting percentage, (what we
call ‘operating points’), the microphone response for a particular
time domain is recorded. Because of the self excitation of the
microphones due to flame, they reach a particular peak frequency
in a sample time period. In the experiment, the sampling time
being 2 seconds and sampling rate being 10000 Hz, the data
collected is in the form of .mi* files with the * being the suffix as
per the number assigned to the microphone. This is basically the
continuous recording of microphones in the time domain. For
each operating point, The number of time domain output files are
equal to the microphones involved in the experiment; i.e. 4 for this
experiment along with the *.prc and *.stp files.
Step 7. When the Microphones are taking the recording for a
particular Operating point, the CCD Camera is activated and a
continuous set of 100 pictures are taken of the flame through the
transparent glass screen. These 100 pictures, later are merged into
one *.bmp file by averaging all the pixel values of the 100 pictures
taken. This gives a smooth average image for each operating point
due to less fluctuation. The data is stored accordingly and
renamed later as per the name of the operating point parameters
for further post processing to obtain flame length for each
reading.
Step 8. These steps are repeated for each lambda value and
piloting percentage is varied in steps. The external piloting is
incremented only to a certain value, because as the piloting
increases, the flame tries to rush away from the Burner end and
this is undesired. Piloting is done till the verge of lean blow out. of
fuel. Higher piloting may lead to a Lean blow off and should be
prevented. Similar readings are taken for other lambda values
with increasing piloting.
Step 9. For all the operating points, the data is collected. The
*.bmp files are saved in the same order as the time domain
recordings of operating points and are later renamed to the
operating point parameters. The general format followed here is
ex: ‘50kW_150Lambda_000Preheat_10pct.bmp’.Here the first
fragment represents power, 2nd represents lambda ratio, 3rd
represents preheating and the 4 fragment represents the piloting
percentage.
5.5 Post processing
5.5.1 Getting the Flame length
The obtained *.bmp file for each bmp file is saved in a separate
folder for classification as internal and external piloting. For each
of these folders; separately, an XL sheet of the data is prepared
with the parameters of each operating points in the same order
the data was created. This includes power, lambda ratio; preheat
temperature and piloting percentage respectively.
The program we later use to post process the flame images to
obtain flame length is as follows:
clcclearclose all
%DATA FILE INFORMATIONoffset=41; % for streak star camera
% in pixels: the side plate which sits between the front plate of the burner
% (flushes with the glass) and the side of the picture % this offset has to be eliminatedBurnerFlangeOffset=0;
%this is the offset added on the flame length, as the true burner exit plane is 5 mm inside the burner flange (which cannot be seen in the flame images.
poly_order=7;plotoption=3; working_dir='\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Preheat_080508\flame pictures\Internal\';
path_in_protokol_file='\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Preheat_080508\flame pictures\Internal\protokol_file_080508_piyush_internal.xls'; output='flame_length\'; integr_plots='integr_plots\';derivI_plots='derivI_plots\';derivII_plots='derivII_plots\';intensity_plots='intensity_plots\'; delete([working_dir,output,sprintf('%.2d',poly_order),'peaks.xls']);
delete([working_dir,output,sprintf('%.2d',poly_order),'peaks.txt']); mkdir([working_dir], output);mkdir([working_dir, output],integr_plots);mkdir([working_dir, output],derivI_plots);mkdir([working_dir, output],derivII_plots);mkdir([working_dir, output],intensity_plots); [parameters, text]=xlsread(path_in_protokol_file, 'for_matlab');Leistung=parameters(:,1);Lambda=parameters(:,2);Temperature=parameters(:,3);SFFinpct=parameters(:,4);
%resolution of the pictures rows X columns in Px% resolution=[1024,512];
%the coordonation in pixel of the window of the combustion cham-ber
X=[1 384];Y=[12 264];Height_Glass_pixels=Y(2)-Y(1); length_cc_pix=X(2)-X(1)+1;[length_cc_mm,faktor]=pixelstomillimeters(length_cc_pix,Height_Glass_pixels);length_cc_mm_vector=X(1)*faktor:1*faktor:X(2)*faktor;length_cc_pix_vector=X(1):1:X(2); peak_mm_total=[];peak_mm=[];for i=1:1:size(Leistung,1) clear file_in; clear scale; clear col; clear sum_col; clear medium_col; clear signal_integr; clear int;
clear heat; clear heat_sum; clear A; clear map; clear derivII_roots; file_in=[working_dir,sprintf('%.2d',Leistung(i)),'kW_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperat-ure(i)),'CPH_',sprintf('%.2d',SFFinpct(i)),'.bmp'];
%file_in=[working_dir,'50kW_176_300PH_2.bmp']; 45kw_176_30pct_270cph_mean_image.bmp %file_in=[working_dir,sprintf('%.2d',Leistung(i)),'kW_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperature(i)),'_mean_image.bmp'];
[A,map]=imread(file_in); for x=X(1):1:X(2) col=A(Y(1):Y(2),x); sum_col(x)=sum(col); medium_col(x)=sum_col(x)/(Y(2)-Y(1)+1); end flip_sum_col=fliplr(sum_col); for y=X(1):1:X(2) signal_integr(y)=sum(flip_sum_col(1:y)); %Integration des eigentlichen Signals end
%integration of the sum plot %roots_interval_pix=[70 1000];% for APX images
roots_interval_pix=[80 384]; % for streak star images p_signal=polyfit(X(1):X(2),sum_col,3); f_signal=polyval(p_signal,X(1):X(2)); p_signal_integr=polyfit(X(1):X(2),signal_integr,poly_order); f_signal_integr=polyval(p_signal_integr,X(1):X(2)); p_derivI=polyder(p_signal_integr); f_derivI=polyval(p_derivI,X(1):X(2)); p_derivII=polyder(p_derivI); f_derivII=polyval(p_derivII,X(1):X(2)); derivII_roots=roots(p_derivII); for k=1:max(size(derivII_roots)) if angle(derivII_roots(k))==0 if (derivII_roots(k)>roots_interval_pix(1)) & (derivII_roots(k)<roots_interval_pix(2)) peak_pix=derivII_roots(k); end end end peak_mm=((peak_pix-offset)*faktor)+BurnerFlangeOffset; %PLOTS if plotoption %paths of the plots
header_plot=[sprintf('%.2d',Leistung(i)),'kW Lambda: ',sprintf('%.3d',Lambda(i)),' Temperature: ',sprintf('%.3d',Temperature(i)),'°C'];
path_out_plot_integr=[working_dir,output,integr_plots,sprintf('%.2d',Leistung(i)),'_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperature(i)),sprintf('%.2d',SFFinpct(i)),'pct.tif']; path_out_plot_derivI=[working_dir,output,derivI_plots,sprintf('%.2d',Leistung(i)),'_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperature(i)),sprintf('%.2d',SFFinpct(i)),'pct.tif']; path_out_plot_derivII=[working_dir,output,derivII_plots,sprintf('%.2d',Leistung(i)),'_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperature(i)),sprintf('%.2d',SFFinpct(i)),'pct.tif']; path_out_plot_intensity=[working_dir,output,intensity_plots,sprintf('%.2d',Leistung(i)),'_',sprintf('%.3d',Lambda(i)),'_',sprintf('%.3d',Temperature(i)),sprintf('%.2d',SFFinpct(i)),'pct.tif']; %integration plot plot(length_cc_mm_vector,signal_integr,length_cc_mm_vector,f_signal_integr,'-r','LineWidth',2); grid on; h_plot_integr=findobj('Type','figure'); xlabel('length of the combustion chamber [mm]'); ylabel('[-]'); title(header_plot); saveas(h_plot_integr, path_out_plot_integr); close(h_plot_integr); %derivative I plot plot(length_cc_mm_vector,f_derivI,'LineWidth',2); grid on; h_plot_derivI=findobj('Type','figure'); xlabel('length of the combustion chamber [mm]'); ylabel('derivI [-]'); title(header_plot); saveas(h_plot_derivI, path_out_plot_derivI); close(h_plot_derivI); %derivative II plot plot(length_cc_mm_vector,f_derivII,'LineWidth',2); grid on; h_plot_derivII=findobj('Type','figure'); xlabel('length of the combustion chamber [mm]'); ylabel('derivII [-]'); title(header_plot); saveas(h_plot_derivII, path_out_plot_derivII); close(h_plot_derivII); %intensity plot plot(length_cc_mm_vector,fliplr(sum_col)); grid on; h_plot_intensity=findobj('Type','figure'); xlabel('length of the combustion chamber [mm]'); title(header_plot); saveas(h_plot_intensity, path_out_plot_intensity); close(h_plot_intensity); end fid=fopen([working_dir,output,sprintf('%.2d',poly_order),'peaks.txt'],'a+'); count=fprintf(fid,'%.3d %.3d %.3d %.2d %6.2f\n', Leistung(i), Lambda(i), Temperature(i),SFFinpct(i),peak_mm); fclose(fid); peak_mm_total=[peak_mm_total peak_mm];
end path_out=[working_dir,output]; dlmwrite([path_out,'flamelengths.xls'],peak_mm_total','delimiter','\t','precision','%4f')fclose all;
Program summery:
Figure: A sample Flame image *.bmp File
(50kW_160lambda_300CPH_0pct.bmp)
Initially the Programs reads the working Directory, Defines an
output directory for sub folders separately for the Integration
plots of Flame length, the first Derivative of pixel values, the
second derivative of pixel values and the intensities for each of the
operating point *.bmp image. To overcome calculations error, it
offsets the calculation from just the tip of the burner exit. It reads
the XL sheet to get the wanted files from the data pool by
comparing XL sheet parameters with that of the *.bmp image file
name. It further reads the *.bmp file and divides the image into a
m n pixel matrix. Now, row wise, it initially integrates the pixel
value for every horizontal pixel row. The length of the combustion
chamber is 120 mm. So graph is plot as per intensity versus
combustion chamber length.
Figure: A sample column integral plot vs. combustor length for
50kW_160lambda_300CPH_0pct.bmp
In the next step, all of the integrated rows are differentiated with
respect to the corresponding column to obtain the maxima and
minima. We notice in this plot that the maxima comes somewhere
in 45 mms from the tip of swirler exit.
Figure: A sample first derivative plot vs. combustor length for
50kW_160lambda_300CPH_0pct.bmp
Later, it is double differentiated to acquire the trend of value
change. We see that at near 45 mms, the value starts to decrease.
It shows that the maxima was reaches somewhere around 45 mms.
Figure: A sample second derivative plot vs. combustor length for
50kW_160lambda_300CPH_0pct.bmp
Finally the maximum intensities are plot. The maximum
intensities plot for each operating point are appended and saved
as final output in a XL sheet in the same order as the files were
read.
Figure: A sample intensity plot vs. combustor length for
50kW_160lambda_300CPH_0pct.bmp
As per post processed, for the sample operating point image file
‘50kw_160lambda_300preheat_0pct.bmp’, the flame length was
obtained to be 47.798047 mm. This value is close but more accurate
then if MATLAB Program wouldn’t have been used and flame
length would have been guessed manually.
Post processing is repeated for both internal and external piloting
and finally comparison of Internal Versus external piloting is
done. This is obtained manually by plotting graph of internal
flame length vs. External flame length over increasing piloting
percentage. The trends are later studied.
Time Domain response post-processing:
The time domain response is simply the response of the
microphones over the sampling time (two seconds) .The sampling
rate is 10000 Hz so it makes the sampled data to be 20000
responses recorded over one microphone. Each recording is saved
as a #.mi* file with * being the microphone number. For each
operating point we get the number of #.mi* files as equal to the
active microphone recordings. The data received is basically the
voltage recording of the microphones and thus contains a
continuous detailed summery of fluctuations. If plot straight, it is
the voltage response of microphones over the time interval. We
are interested in noticing the maximum amplitude and maximum
frequency response of each microphone over the time domain. For
this, the wanted graph is amplitude response over frequency
response for all the microphones for a particular operating point.
The response of these microphones is post-processed using the
following program on MATLAB:
%Program to check the time domain signals from the ICLEAC test rig upto 10 channels in total.This is modified to plot the frequency spectrum of only few chosen channels on 26.04.08
%%%%%%%%%%%%%%%%%%%%%
clear allclc;
%user inputpath_in='\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Preheat_080508\Internal_pilot\';path_out='\\Andromeda\allhomes\akhouri\piyush raj ka matlab waala kaam\Preheat_080508\Internal_pilot\Out-put\'; if isdir(path_out)==0 mkdir(path_out);end mic_calib_factors=[1.049803,0.919499,0.927001,0.859807,0.618178,0.9878,0.85990]; sampling_frequency=10000; no_samples=20000; total_time=no_samples/sampling_frequency;f=1/sampling_frequency;time=0:f:total_time-f;time_plot=time(1:length(time)/2);faktor_fft=sampling_frequency/no_samples;low_limit=0; frequency_array=[0:faktor_fft:sampling_frequency-faktor_fft];frequency_array_plot=frequency_array(low_limit+1:length(frequency_array)/2); plots=1; %flag whether to plot ?no_ch=5;ampl_factor=1; %to convert back the pascals into a voltage from a mic signal and
have a consistancy with the CTA and UV signal filename='50kW_178_300CPH_80-000Hz-001';commontitle='Rig response to self excitation';pref=2*10^-5; %reference pressure amp_phases_total=[];peaks=[]; count=0;
% for j=1:1:size(Leistung,1) for i=1:1:no_ch file=strcat(path_in,filename,'.mi',int2str(i)); fid(i)=fopen(file); if fid(i)>0
%if i==1||i==5 % choosent to plot only mics 1(upstream located in the plenum) and 5 (downstream located in the Combn. chamber)
count=count+1; signal=getbindata(file)/ampl_factor;
fft_mic=fft(signal); amplitudes_mic=abs(fft_mic); phases_mic=angle(fft_mic); amplitudes_mic(1)=0; amplitudes_mic_plot(:,i)=amplitudes_mic(low_limit+1:no_samples/2)/(no_samples/2); [mic_prime(count),index(i)]=max(amplitudes_mic_plot(:,i)); phases_mic_plot(:,i)=phases_mic(low_limit+1:no_samples/2); freq_prime(count)= frequency_array(index(i)); end end %end for the i loop % PLOTS colourcode=['k','b','g','r','c','m','y','g:','k:','r:']; if plots %Plot the frequency Vs amplitudes fig_title=strcat(commontitle,' amplitude spectrum'); path_out_plot_bmp=strcat(path_out,filename,'_amplitudes.tif'); path_out_plot_fig=strcat(path_out,filename,'_amplitudes.fig');
%title([fig_title]) for i=1:1:no_ch if fid(i)>0 %if i==1||i==5 subplot(5,1,i) plot(frequency_array_plot,amplitudes_mic_plot(:,i), colourcode(i)),grid on, ylabel(strcat('amp[Pa] mi: ',int2str(i))), xlim([0 1000]) xlabel('frequency [Hz]') title([fig_title]) path_out_plot_bmp=strcat(path_out,filename,'mic',int2str(i),'_amplitudes.tif'); h_plot=findobj('Type','figure'); hold on; end end saveas(h_plot,path_out_plot_bmp); close(h_plot); endamplitudes_Pa_dB(1,:)=mic_prime;amplitudes_Pa_dB(2,:)=freq_prime'; dlmwrite([path_out,filename,'.xls'],amplitudes_Pa_dB,'delimiter','\t','precision','%4f'); fclose all;
Program summery:
The program first reads the working directory and creates an output folder to store output
results. It reads the sampling rate, microphone calibration coefficient, frequency, total time
and from that it deduces the Fourier transform factors.
Later a frequency array is defined from this #.mi* file, the file is converted into a *.fft file.
This *.fft file is opened, processed to give a plot of maximum amplitude versus maximum
frequency recorded by each microphone in the time domain.
Later the subplot option of MATLAB appends all these separate plots in one common image
file. These files are simplified spectrum plots of amplitude responses with respect to peak
frequency responses generated because of the self excitation of the test rig due to the flame.
Plot: Rig response to self excitation amplitude spectrum for Microphones, A , B and C for operating point
50kW_160_000Preheat_0Piloting
The peak frequency and peak amplitude recorded for each
microphone for each operating point are recorded in a XL sheet.
Later all these data are appended manually into one common XL
Sheet. From this XL sheet responses are categorized in separate
sheets as per microphones for external and external piloting.
Firstly, all the microphones for a particular piloting type (internal
or external) are plotted to a set of operating points having same
power and lambda ratio. An example of such a plot is shown for
power 50 kW and lambda ratio 1.6. Here the piloting percentage
is varied from 40 percent to 80 percent.
Plot: Peak Frequency of Microphones A, B, C vs. Pilot gas percentage over operating points 50_160_40-80
5.5.3 Final Post processing:
The data recorded for each operating point (peak frequency and peak amplitude) is manually
appended for each microphone. Later it is sub classified for each microphone for a particular
power value, particular lambda value and increasing piloting value. Separately done for each
microphone for every operating point with internal and external piloting, the internal and
external peak amplitudes are compared for internal and external piloting with increasing
piloting ratio.
This is done for each microphone and final output are three graphs for Microphone A, B and
C where internal and external piloting trend is studied with Increasing Piloting ratio.
5.6 Results
The flame length comparison plot for Internal and External piloting for microphone A, B and
C for increasing piloting are shown below. The corresponding tables show the peak
amplitudes reached due to the self excitation of the test rig due to flame instabilities.
Microphone A:
Power Lambda Temperature pilotgas A(peak_internal) pilotgas A(peak_external)
[kW] 150 [C] [%] Amplitude [%] Amplitude
50 150 0 0 57 0 56
50 150 0 10 109.5 20 117.5
50 150 0 20 134.5 20*(repeat) 143
50 150 0 20*(repeat) 126 30 142
50 150 0 30 15.5 40 156.5
50 150 0 40 134 50 125.5
50 150 0 40*(repeat) 144 60 142
50 150 0 50 150.5 70 145.5
50 150 0 60 144.5 80 171
50 150 0 70 144.5 90 126
50 150 0 80 138.5 100 133.5
50 150 0 90 146
Recordings were repeated for operating points 50kW_150lambda_000pct_20internal,
50kW_150lambda_000pct_20external, and 50kW_150lambda_000pct_40internal.Following
was the graph plot showing the external and Internal piloting trend together for 50kw and 1.5
lambda ratio.
Microphone B:
Power Lambda Temperature pilotgas B(peak_internal) pilotgas B(peak_external)
[kW] 150 [C] [%] Amplitude [%] Amplitude
50 150 0 0 57 0 56
50 150 0 10 109.5 20 97
50 150 0 20 55.5 20.5 85.5
50 150 0 20.5 61 30 142
50 150 0 30 99.5 40 156.5
50 150 0 40 134 50 171.5
50 150 0 40.5 134.5 60 179
50 150 0 50 151 70 181.5
50 150 0 50.5 184 80 186
50 150 0 60 203.5 90 187.5
50 150 0 60.5 156.5 100 193
50 150 0 70 144.5
50 150 0 80 124
50 150 0 90 40
50 150 0 100
Recordings were repeated for operating points 50kW_150lambda_000pct_20internal,
50kW_150lambda_000pct_20external,50kW_150lambda_000pct_50internal,50kW_150lamb
da_000pct_60internal and 50kW_150lambda_000pct_40internal.Following was the graph plot
showing the external and Internal piloting trend together for 50kw and 1.5 lambda ratio.
Microphone C:
Power Lambda Temperature pilotgas C(peak_internal) pilotgas C(peak_external)[kW] 150 [C] [%] Amplitude [%] Amplitude
50 150 0 0 57 56 050 150 0 10 109.5 97 2050 150 0 20 55.5 85.5 20.550 150 0 20.5 61 105 3050 150 0 30 11.5 156.5 4050 150 0 40 134 844 5050 150 0 40.5 20.5 179 6050 150 0 50 151 220.5 7050 150 0 50.5 184 186 8050 150 0 60 203.5 187.5 9050 150 0 60.5 86 133.5 10050 150 0 70 70.550 150 0 80 117
50 150 0 90 2650 150 0
Recordings were repeated for operating points 50kW_150lambda_000pct_20internal,
50kW_150lambda_000pct_20external,50kW_150lambda_000pct_50internal,50kW_150lamb
da_000pct_60internal and 50kW_150lambda_000pct_40internal.Following was the graph plot
showing the external and Internal piloting trend together for 50kw and 1.5 lambda ratio.
Flame length Comparison:Flame length comparison of External Versus Internal Piloting led to following data after post
processing:
Leistung Lambda Temperature pilotgasInternal pi-
lotingExternal pi-
loting[kW] [-] [C] [%] [mm] [mm]50 150 0 10 69.385241 70.40829950 150 0 20 49.59637 34.50849950 150 0 30 42.253789 35.46503150 150 0 40 41.864582 33.95512650 150 0 50 41.55036 30.32781550 150 0 60 42.329884 28.95925750 150 0 70 42.387926 24.72322350 150 0 80 43.910127 20.92911250 150 0 90 46.48504 18.4461650 150 0 100 16.25079
The internal and external piloting are compared after being plot versus increasing piloting
percentage.
In the experiment 19 total operating points (including repetition of some)(for power 50 kW,
lambda ratio 1.5) were recorded in which 9 were for Internal piloting and 10 for External
Piloting
5.7ConclusionsDuring the Experiment , following were the parameters where flame was about to shut off:
Based on the following operating points, the operating domain of the single burner test rig can be sketched as follows:
Plot: Operating Domain of burner Test Rig without preheat
5.8 Problems encountered during the experiment
1. First, clear optical access is necessary which requires positioning of a viewing port in
a strategic location with respect to the flame for collecting the flame light emission.
Second, the environment is difficult because of excessive heat being produced by the
burner.
2. Furthermore, unexpected changes in fuel composition, or momentary changes in fuel
delivery can lead to problems with flame anchoring.
3. The flame tends to leave the burner exit and tends to push away from the swirler exit.
This should be prevented as it lead to false flame length measurements and even flame
blow offs.
4. Typically the high temperature-operating environment of the burners necessitates the
need for water or gas-cooled probes for use either in or near the burner.
5. The environment may be dusty which is not favourable for the use of optical
equipment except with special precautions, such as gas purging over the optical
components.
6. Safety of operation is an essential characteristic expected from all industrial
combustion systems. Automated control of the presence of the flame in the
combustion can be used to stop the flow of oxidant when the fuel flow is suddenly
interrupted.
7. The experiment should be conducted when the test rig is heated up and has obtained
temperature saturation and stability. Abrupt variations in temperatures of test rig may
induce temperature stresses and may harm costly gadgets around.
List of Figures
Figure: A photograph of flame during combustion in the single burner test rig
Figure: A sample Flame image *.bmp File (50kW_160lambda_300CPH_0pct.bmp)
Plot: A sample column integral plot vs. combustor length for 50kW_160lambda_300CPH_0pct.bmp
Plot: A sample first derivative plot vs. combustor length for 50kW_160lambda_300CPH_0pct.bmp
Plot: A sample second derivative plot vs. combustor length for 50kW_160lambda_300CPH_0pct.bmp
Plot: A sample intensity plot vs. combustor length for 50kW_160lambda_300CPH_0pct.bmp
Plot: Rig response to self excitation amplitude spectrum for Microphones, A , B and C for operating point 50kW_160_000Preheat_0Piloting
Plot: Peak Frequency of Microphones A, B, C vs. Pilot gas percentage over operating point’s 50kW_160lambda_300CPH_40-80pct
Plot: Microphone A peak amplitude response (Internal and External piloting) versus piloting percentage
Plot: Microphone B peak amplitude response (Internal and External piloting) versus piloting percentage
Plot: Microphone C peak amplitude response (Internal and External piloting) versus piloting percentage
Plot: Flame Comparison (External and Internal piloting) versus increasing piloting percentagePlot: Operating Domain of burner Test Rig without preheat
Plot 1. Absolute Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 2. Absolute Phase Plot for subset GIJE in the range 0-240 Hz.
Plot 3. Relative Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 4 . Relative Phase Plot for subset GIJE in the range 0-240 Hz.
Plot 5. Absolute Amplitude Plot for subset GIJE in the range 260-620 Hz.
Plot 6. Absolute Phase angle Plot for subset GIJE in the range 260-620 Hz.
Plot 7 . Relative Amplitude Plot for subset GIJE in the range 260-620 Hz.
Plot 8. Relative Phase Plot for subset GIJE in the range 260-620 Hz.
Plot 9. Absolute Amplitude Plot for subset GFHN in the range 0-240 Hz
Plot 10. Absolute Phase Angle Plot for subset GFHN in the range 0-240 Hz
Plot 11. Relative Amplitude Plot for subset GFHN in the range 0-240 Hz
Plot 12. Relative Phase Angle Plot for subset GFHN in the range 0-240 Hz
Plot 13. Absolute Amplitude Plot for subset GFHN in the range 260-620 Hz
Plot 14. Absolute Phase Angle Plot for subset GFHN in the range 260-620 Hz
Plot 15. Relative Amplitude Plot for subset GFHN in the range 260-620 Hz
Plot 16. Relative Phase Angle Plot for subset GFHN in the range 260-620 Hz
Plot 17. Absolute Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 18. Absolute Phase Angle Plot for subset GIJE in the range 0-240 Hz.
Plot 19. Relative Amplitude Plot for subset GIJE in the range 0-240 Hz.
Plot 20. Relative Phase Angle Plot for subset GIJE in the range 0-240 Hz.
Figure 1. Single burner test Rig facility
Eqn.1. Definition equation of a transfer matrix linking the acoustic parameters Pressure (p) and velocity (u) across the element
Eqn.2. Mass conservation across a premix flame
Diagram 1. Schematic of the Test RigFigure 2. Air supply with air mass flow controller
Figure 3. Air pre-heater
Figure 4. fuel injection in the plenum
Figure 5. Upstream siren for wave generation in plenum
Figure 6. Second siren for downstream wave triggering
Figure 7. Plenum/ supply tube with microphone slots
Figure 8. Combustion chamber
Figure9. The EV5 burner (not in use )
Figure 11. The exhaust system
Figure12. Water cooling
Figure 13. Air cooling supply knobs
Figure 14.Air cooling on the secondary combustor
Eqn.3.Sensitivity of a microphoneEqn.4. Sensitivity of a loudspeaker
Figure 15. A microphone with the adaptor jacket and cooling pipes
Figure 16. Arbitrary Waveform generator
Figure 17. Microphone with cooling adaptor showing the microphone at the tip, receiver
at the middle and bronze cooling adapter.
Figure 18: Internal Structure of an electret
Figure 19: A sensor based on the piezoelectric effect
Figure 20. Microphone water cooling system
Figure21. Microphone amplifier
Figure 22. Acoustic calibrator
Figure 23. Digital storage electroscope
Figure 24. Labview GUI on the process control computer
Figure25. Section showing microphone amplifications manually fed in the GUIFigure 25. Section to feed in file name, file type and working directory
Figure 26. Section showing no. of loops currently processed.