[American Institute of Aeronautics and Astronautics 41st AIAA/ASME/SAE/ASEE Joint Propulsion...

American Institute of Aeronautics and Astronautics

1

ON THE APPLICABILITY OF CHEMILUMINESCENCE TO THE ESTIMATION OF UNSTEADY

HEAT-RELEASE DURING UNSTABLE COMBUSTION IN LEAN PREMIXED COMBUSTOR

Jong G. Lee, Esteban Gonzalez and Domenic A. Santavicca Department of Mechanical and Nuclear Engineering

The Pennsylvania State University University Park, PA 16802

ABSTRACT This paper presents results of experimental study of flame chemiluminescence from an atmospheric, swirl-stabilized, turbulent (Re=10.8·103 – 57.0·103) lean premixed flame. A detailed spectroscopic measurement of chemiluminescence over the wavelength of 395-495 nm is made for various flow conditions and used to determine the background chemiluminescence intensity from CO2* in CH-chemiluminescence band around 431 nm. Simultaneous CH- and CO2-chemiluminescence from the whole flame are collected over the wavelength of 430±5 and 470±5 nm, respectively and the CH-chemiluminescence is corrected for the background CO2-chemiluminescence. For a fixed equivalence ratio, the background-corrected CH-chemiluminescence intensity normalized by the fuel mass flow rate is found to be constant with respect to the combustor inlet velocity and irrespective of flame stability, while either CO2-chemiluminescence or uncorrected CH-chemiluminescence intensity normalized by the fuel mass flow rate increases as the mean flow velocity increases. Consequently the background-corrected overall CH- chemiluminescence intensity can be used as a measure of overall heat release of flame under unstable as well as stable combustion where there is no equivalence ratio fluctuation.

INTRODUCTION Lean premixed (LP) combustion is widely

regarded as the most promising strategy for meeting current and future NOx emission regulations for industrial gas turbines. Unfortunately, LP combustion systems are known to be very susceptible to thermo-acoustic instability which is the result of closed-loop coupling between pressure fluctuation and heat release oscillation. Therefore, the measurement of heat release rate is of great importance in the study of these instabilities. For its relative simplicity, the chemiluminescence emission has

been used as an indicator of heat release in such flames and extensively used in combustion instability studies [1-5].

Chemiluminescence is the emission of photons which occurs when electronically excited molecules, formed by chemical reactions, return to their ground state. Alternatively, these excited molecules may return to their ground state through collisions with walls or other molecules. Consequently, chemiluminescence measurements of these species will be in direct proportion to their production rates, and can offer a better measure of burning rate than a non-rate based measure. Figure 1 shows a typical spectrum of chemiluminescence emission from a turbulent lean premixed flame at an atmospheric pressure. The strong chemiluminescence emissions in lean hydrocarbon flames are primarily from CH*, OH* and CO2* (the asterisk indicates an excited species). In these flames, the chemiluminescence emission from CH* (431 nm) and OH* (309 nm) occurs at distinctly different and relatively narrow wavelength intervals, while the CO2- chemiluminescence lies over a broad wavelength interval (300-600 nm) and overlaps the CH- and OH-chemiluminescence spectra. It can be noted that the CO2-chemiluminescence can contribute significantly to the measurement of either CH- and OH-chemiluminescence if a filter with a broad bandwidth is used to collect either one of these species.

The base for using chemiluminescence as an indicator of the location of reaction zone and to infer local and overall heat release rates relies on early experimental studies by Clark [6] and Hurle et al. [7] of jet premixed laminar and weakly turbulent hydrocarbon-air flames near stoichiometric conditions. In these studies it was found that for a fixed equivalence ratio the chemiluminescence captured from the whole flame, hereafter referred to as the overall chemiluminescence, increases linearly with the fuel flow rate, being the slope function of the equivalence ratio.

41st AIAA/ASME/SAE/ASEE Joint Propulsion Conference & Exhibit10 - 13 July 2005, Tucson, Arizona

AIAA 2005-3575

Copyright © 2005 by the American Institute of Aeronautics and Astronautics, Inc. All rights reserved.


2

Numerical studies of strained laminar methane-air flames with detailed chemical kinetics [8-9] have shown that the local rate of heat release and the local chemiluminescence emission increase exponentially with temperature, and that they are affected by unsteady strain and flame curvature. This leads to a power law relationship between the local chemiluminescence (per unit flame area) and the local rate of heat release (per unit flame area), that is,

Ilocal ∝ (HRlocal)α

where the exponent α is a positive number and depends on the flame temperature (as determined by the equivalence ratio, unburned gas temperature, dilution and radiation losses) and on the effects of unsteady strain and flame curvature.

In order to determine the relationship between the overall chemiluminescence and the overall rate of heat release, one must integrate the local values over the flame area, i.e., Ioverall = ∫A Ilocal dAflame and HRoverall = ∫A HRlocal dAflame If the flame temperature and the effects of strain and curvature are constant over the flame area, this leads to the result that the overall chemiluminescence emission and the overall rate of heat release are proportional, that is,

Ioverall = C HRoverall

where the constant C depends on the flame temperature and the effects of strain and

curvature. This result is consistent with the experimental results discussed previously, where for fixed equivalence ratio and inlet temperature the overall chemiluminescence emission increased linearly with the overall rate of heat release and that the slope depends on the equivalence ratio.

The relationship between the overall chemiluminescence emission and flow parameters becomes more complicated if the flow is turbulent. The effect of turbulence on chemiluminescence is not clearly known and there exists some uncertainty of whether chemiluminescence will be an adequate measure of heat release in highly turbulent flames where the chemistry of the flame can be change by the high turbulence. For example, Clark [6] finds that turbulence (Re=8000-14000) reduces the intensity of the overall chemiluminescence in rich and lean propane-air and ethane-air flames. The numerical studies of strained laminar flames by Samaniego et al. [8] and Najm et al. [9] show that the use of CO2* as a quantitative measure of heat release is adequate only for well-controlled situations such as two-dimensional and axisymmetric flow conditions, where unsteady strain-rate is the only significant parameter. They suggest that CO2* be used with caution as a qualitative measure of heat release in turbulent premixed flames, unless the effects of curvature and three-dimensionality on CO2* are further studied. From their experimental results, Najm et al. [9] conclude that "images of CH*, OH* and C2* are not generally reliable measures of turbulent flame front topology, much less of burning and heat release rates". However, the study by Higgins et al. [10-11] using a counter-flow high-pressure laminar premixed methane-air flame at lean conditions finds no significant dependence of CH* on the relatively small strain rate (200-700 s-1). However, they suggest that for highly turbulent flows it is expected that flame curvature will affect CH*.

Nonetheless, chemiluminescence imaging has been used extensively in many studies of turbulent flames as a measure of heat release and/or equivalence ratio as well as a visualization method of flame structure. Most experimental studies have used narrow bandpass filters (typically with the full width at half maximum of 10 nm) to measure the chemiluminescence from CH* and OH*. However, as pointed out previously, if the background chemiluminescence from CO2* is not properly accounted for, the effect of flow parameters on chemiluminescence and hence the relationship between chemiluminescence intensity and heat

Fig 1. Typical chemiluminescence emission spectrum from the lean premixed dump combustor operating at 100 kPa on natural gas at an equivalence ratio of 0.8 with an inlet temperature of 673K [5].

Wavelength (nm)300 350 400 450 500 550 600

Inte

nsity

(a.u

.)

0.0

2.0

4.0

6.0

8.0

10.0CH*

OH*

CO2*

Wavelength (nm)300 350 400 450 500 550 600

Inte

nsity

(a.u

.)

0.0

2.0

4.0

6.0

8.0

10.0CH*

OH*

CO2*

Fig 1. Typical chemiluminescence emission spectrum from the lean premixed dump combustor operating at 100 kPa on natural gas at an equivalence ratio of 0.8 with an inlet temperature of 673K [5].

Wavelength (nm)300 350 400 450 500 550 600

Inte

nsity

(a.u

.)

0.0

2.0

4.0

6.0

8.0

10.0CH*

OH*

CO2*

Wavelength (nm)300 350 400 450 500 550 600

Inte

nsity

(a.u

.)

0.0

2.0

4.0

6.0

8.0

10.0CH*

OH*

CO2*


3

release and/or equivalence ratio may not be correctly addressed.

The objective of the present work is firstly to experimentally study the effect of flow parameters such as combustor velocity, equivalence ratio and fuel-air premixing on the overall CH- and CO2- chemiluminescence from a swirl-stabilized turbulent lean premixed combustor operating stably at an atmospheric pressure. Especially, the CH-chemiluminescence collected using a narrow bandpass filter is corrected for the background chemiluminescence from CO2*. Secondly, with an established relationship between the overall chemiluminescence and flow parameters under stable flames, it will be determined whether the chemiluminescence can be used to estimate fluctuating heat release during unstable combustion.

EXPERIMENTAL APPARATUS AND PROCEDURE

Coaxial dump combustor

The coaxial dump combustor used in this study is depicted in Fig. 2. It consists of an annular mixing section and combustor section. The combustor consists of an optically accessible fused-silica section. Two different length of combustors are used: 280 mm-long combustor is used to achieve stable combustion without any combustion oscillation over all of the operating conditions listed in Table 1 and 635 mm-long combustor to achieve unstable combustion over some of the operating conditions.

The flame is stabilized on a bluff centerbody which is centered in the annular mixing section using swirl vanes mounted at an angle of 30o with respect to the flow direction. A choke plate at the entrance of the mixing section ensures that the inlet flow is choked at all operating conditions. Natural gas (96 % methane) is used as a fuel and injected well upstream of combustor, providing a well premixed mixture at the inlet of combustor.

Air is heated by an electric heater (30-kW). The combustor inlet temperature is monitored by a K-type thermocouple located 40 mm upstream of the dump plane and regulated by a temperature controller. The air and fuel flow rates are measured with linear mass flow meters.

The apparatus is provisioned with two different fuel injection locations. In one of the locations, the fuel is injected radially outward from the centerbody into the mixing section through 16 equally spaced holes (0.635 mm diameter) located

25.4 mm upstream of the dump plane, providing a partially-premixed fuel-air mixture at the inlet to the combustor. The other injection location is located well upstream of the chocked inlet to the mixing section, producing a well mixed fuel-air mixture in the combustor.

Experimental approach

Commonly, CH-chemiluminescence is

collected using a narrow bandpass filter over the wavelength of 430±5 nm. In order to correct for the background CO2-chemiluminescence in the light collected using the filter, the amount of background level should be determined. This requires a detailed spectrum of flame emission over the wavelength range. Once the level of background CO2-chemiluminescence is determined, it should be accounted for using CO2-chemiluminescence collected over different wavelength range.

A spectroscopy system was used to obtain emission spectrum over the wavelength of 395-495 nm. Based on the measured emission spectrum, two PMT detection systems (i.e. one over the wavelength of 430±5 nm and the other

Choked inlet

Fused silica section combustor

(110 mm dia. x 275 mm)

Swirl vanes

Choked inlet


(110 mm dia. x 275 mm)

Swirl vanes

Fig. 2 Schematic drawing of the coaxial dump combustor

Choked inlet


(110 mm dia. x 275 mm)

Swirl vanes

Choked inlet


(110 mm dia. x 275 mm)

Swirl vanes

Fig. 2 Schematic drawing of the coaxial dump combustor

Table 1. Operating conditions

10.8x103

- 57.0x103Reynolds number

0.50-0.70Equivalence ratios

200 and 350Mixture inlet temperature, Ti (oC)

3-12Mean velocity in the combustor, Uc (m/s)

36-120Mean velocity in the mixing section, Um (m/s)

10.8x103

- 57.0x103Reynolds number

0.50-0.70Equivalence ratios

200 and 350Mixture inlet temperature, Ti (oC)

3-12Mean velocity in the combustor, Uc (m/s)

36-120Mean velocity in the mixing section, Um (m/s)


4

over the wavelength of 470±5 nm) were used to collect emission simultaneously over the two different wavelength ranges and account for the background CO2-chemiluminescence level. Chemiluminescence spectrum measurement

Chemiluminescence from the whole flame over the wavelength range of 395-495 nm was collected using a lens and coupled into an optical fiber. Collected light, I(λ), is converted into a voltage output in the detector through the detection system efficiency, S(λ) which depends on the collection efficiency of optics train, including optical fiber, lenses and spectrometer as well as the sensitivity of each camera pixels to the wavelength of light: i.e.

In other words, if the detection system efficiency does not depend on wavelength, the intensity of light collected by the detection system at different wavelengths will be represented by the same relative quantity as the voltage output at each wavelength. However, it is hardly the case.

Shown in Fig. 3 is the schematic drawing of the setup to determine S(λ). A tungsten lamp calibrated at 650 nm is used as the reference light source since its spectral irradiance is known at the calibrated wavelength. Light from the whole tungsten filament is collected using a lens (an achromatic microscope objective, N.A.=0.65) and coupled into a fused-silica optical fiber (N.A.=0.22). The light output from the other end of optical fiber is delivered onto an imaging spectrometer (SPEX 1000M: 1-m, f/8, 1200 grooves/mm) and the dispersed light is detected using an intensified CCD camera. The spectrometer can capture light spectrum over 10 nm per grating orientation. More detailed procedure can be found in the appendix-A.

Once S(λ) is determined, the detected chemiluminescence intensity from flame is

corrected for by S(λ). Also, the measured chemiluminescence intensity is corrected for the gradual transmittance change of fused silica tube by monitoring chemiluminescence intensity at a fixed reference condition before and after each run.

Overall chemiluminescence measurement

Overall chemiluminescence is measured for both stable and unstable operating conditions only with well-premixed flame where there is no equivalence ratio fluctuation. Chemiluminescence from the whole flame is collected at the downstream of combustor to minimize the effect of gradual transmittance change of fused-silica tube with respect to the operating conditions. The collected light is split into two streams of light using a beam splitter. One stream of light is filtered using a narrow band pass filter centered at 430 nm with a full width at half height (FWHM) of 10.3 nm (hereafter it is called CH-filter) and the other is filtered using a narrow band pass filter centered at 470.5 nm with 10 nm of FWHM (hereafter it is called CO2-filter). The same type of PMT’s are used to detect each light stream. It is ensured that the whole flame is captured and its image is kept well inside of the PMT’s detector area (6 mm by 24 mm) within 6-mm in diameter for all operating conditions.

In order to correct for the background CO2-chemiluminescence from the light measured using the CH-filter, the high voltages of each PMT are set in such a way that the light intensity measured using the CO2-filter is the same as the background CO2-chemiluminescence intensity. The combinations of PMT-voltage setting are obtained using the tungsten lamp. Then, by subtracting the light intensity measured using CO2-filter from that measured using CH-filter, the chemiluminescence only from CH* can be determined.

RESULTS AND DISCUSSIONS

Chemiluminescence spectrum

Measured chemiluminescence is corrected for the detection system efficiency according to the procedure presented in the previous section. Fig. 4 shows the resulting chemiluminescence spectrum for the premixed flames at φ=0.7 and Ti=400 oC. For a fixed equivalence ratio and temperature, the overall chemiluminescence

L1 L2ML

SPOF

FH FHTL ICCD

L1 L2ML

SP

SP

OFFH FH

TL ICCD

ICCD

Fig. 3 Schematic drawing of the setup for spectroscopic measurement (TL: Tungsten lamp, ML: Microscope objective lens, FH: Fiber holder, OF: Optical fiber, L1 & L2: Lenses, SP: Spectrometer, ICCD: Intensified CCD camera)

L1 L2ML

SPOF

FH FHTL ICCD

L1 L2ML

SP

SP

OFFH FH

TL ICCD

ICCD

L1 L2ML

SPSPOF

FH FHTL ICCDICCD

L1 L2ML

SP

SP

OFFH FH

TL ICCD

ICCD

Fig. 3 Schematic drawing of the setup for spectroscopic measurement (TL: Tungsten lamp, ML: Microscope objective lens, FH: Fiber holder, OF: Optical fiber, L1 & L2: Lenses, SP: Spectrometer, ICCD: Intensified CCD camera)

∫ •= λλλ dSIVOUT )()(


5

intensity is shown to increase as the mean mixture velocity in combustor and hence mass flow rate of fuel increases. This indicates that the overall chemiluminescence intensity is proportional to the overall heat release from flame. The background CO2-chemiluminescence level is estimated from a curve-fitting of the spectrum over the wavelengths of 405-420 and 440-495 nm where there exists negligible CH*. An interesting finding is that the

ratio of background-CO2* embedded in the chemiluminescence over the wavelength of 420-440 nm to the chemiluminescence over the wavelengths of 460-480 nm is 1.20 for all the velocity cases. And this is found not to change for all operating conditions run in this study, strongly suggesting that the relative shape of the broad CO2-chemiluminescence spectrum does not change with respect to the operating condition.

Also it shows that the chemiluminescence from CH* accounts for only about 30% of the chemiluminescence intensity collected over the wavelengths of 420-440 nm. However, as will be seen, the background-subtracted CH- chemiluminescence is found to be more sensitive to the equivalence ratio and premixedness of fuel-air mixture.

The effect of equivalence ratio on the chemiluminescence spectrum for Uc=10 m/sec and Ti=400 oC is shown in Fig. 5. For the change of equivalence ratio from 0.55 to 0.70, the CO2- and background-subtracted CH-chemiluminescence intensities increase as much as 200% and 470%, respectively.

Fig. 6 shows the effect of fuel-air premixing on the chemiluminescence spectrum at Uc=10 m/sec and Ti=400 oC. The change of fuel-air premixedness results in the change of the shape of chemiluminescence spectrum as well as its magnitude: partially premixed fuel-air mixture results in greater overall chemiluminescence intensity than well-premixed mixture. This is because the partially premixed mixture results in a locally higher temperature region in the flame. Total chemiluminescence increases about 8% for CO2-chemiluminescence integrated over 460-480 nm and 67% for the background-subtracted CH- chemiluminescence. Overall chemiluminescence measurement for premixed flames under stable and unstable combustion Fig. 7 shows the overall chemiluminescence intensity collected using CH- and CO2-filter (hereafter they will be called as the overall CH- and CO2-chemiluminescence, respectively) as a function of the mean combustor inlet velocity (Uc) for different equivalence ratios (φ’s). Uc is varied from 3 to 8 m/sec and φ from 0.5 to 0.65. The linear curve fits for each equivalence ratio are also shown in the figure. These results show that the overall CH- and CO2- chemiluminescence intensity

Fig. 4 Chemiluminescence spectrum for well-premixed flames at φ=0.70 and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0

2e+6

4e+6

6e+6

8e+6

1e+7

Uc= 4 m/sec

Uc= 6 m/sec

Uc= 8 m/sec

Uc=10 m/sec

Fig. 4 Chemiluminescence spectrum for well-premixed flames at φ=0.70 and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0

2e+6

4e+6

6e+6

8e+6

1e+7

Uc= 4 m/sec

Uc= 6 m/sec

Uc= 8 m/sec

Uc=10 m/sec

Fig. 5 Chemiluminescence spectrum for well-premixed flames at φ=0.55 and 0.7 with Uc=10 m/sec and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0

2e+6

4e+6

6e+6

8e+6

1e+7

φ=0.70

φ=0.55

Fig. 5 Chemiluminescence spectrum for well-premixed flames at φ=0.55 and 0.7 with Uc=10 m/sec and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0

2e+6

4e+6

6e+6

8e+6

1e+7

φ=0.70

φ=0.55

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0

2e+6

4e+6

6e+6

8e+6

1e+7

φ=0.70

φ=0.55

Fig. 6 Chemiluminescence spectrum for well-premixed and partially-premixed flames atUc=10 m/sec, φ=0.70 and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0.0

2.0e+6

4.0e+6

6.0e+6

8.0e+6

1.0e+7

1.2e+7

1.4e+7

Well-premixed

Partially-premixed

Fig. 6 Chemiluminescence spectrum for well-premixed and partially-premixed flames atUc=10 m/sec, φ=0.70 and Ti=400oC

Wavelength (nm)

400 420 440 460 480 500Che

milu

min

esce

nce

inte

nsity

(a.u

.)

0.0

2.0e+6

4.0e+6

6.0e+6

8.0e+6

1.0e+7

1.2e+7

1.4e+7

Well-premixed

Partially-premixed


6

is a linear function of the mean combustor inlet velocity (hence fuel flow rate) and the slope of this function increases with increasing equivalence ratio.

Since the overall heat release increases linearly with respect to the increase of fuel flow rate, the CH- and/or CO2-chemiluminescence intensity normalized by the combustor inlet velocity should not change with respect to the change of the combustor inlet velocity for a fixed φ if the overall CH- and/or CO2-chemiluminescence intensity is proportional to the overall heat release. Fig. 8 (a) and (b) show that for a fixed φ the overall CH- and CO2-chemiluminescence intensity normalized by the combustor inlet velocity increase with respect to the combustor inlet velocity. These results indicate that either CH- or CO2-chemiluminescence is not a good measure of overall heat release.

Fig. 9 shows the CO2-background subtracted overall CH-chemiluminescence intensity

normalized by the combustor inlet velocity for different equivalence ratios. The results show that

Mean velocity in the combustor (m/sec)2 3 4 5 6 7 8 9

CH

-Che

milu

min

esce

nce

(a.u

.)

0.0

0.2

0.4

0.6

0.8

1.0φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e (a

.u.)

0.0

0.1

0.2

0.3

0.4

0.5φ=0.50φ=0.55φ=0.60φ=0.65

(a)

(b)

Fig. 7 Overall (a) CH- and (b) CO2-chemiluminescence as a function of the mean combustor inlet velocity for different equivalence ratio


CH

-Che

milu

min

esce

nce/

Uc (

a.u.

)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e/U

c (a.

u.)

0.00

0.02

0.04

0.06

0.08

φ=0.50φ=0.55φ=0.60φ=0.65

Fig. 8 Overall (a) CH- and (b) CO2-chemiluminescence intensity normalized by the combustor inlet velocity for different equivalence ratios

(a)

(b)


CH

-Che

milu

min

esce

nce

(a.u

.)

0.0

0.2

0.4

0.6

0.8

1.0φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e (a

.u.)

0.0

0.1

0.2

0.3

0.4

0.5φ=0.50φ=0.55φ=0.60φ=0.65

(a)

(b)



CH

-Che

milu

min

esce

nce

(a.u

.)

0.0

0.2

0.4

0.6

0.8

1.0φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e (a

.u.)

0.0

0.1

0.2

0.3

0.4

0.5φ=0.50φ=0.55φ=0.60φ=0.65

(a)

(b)



CH

-Che

milu

min

esce

nce/

Uc (

a.u.

)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e/U

c (a.

u.)

0.00

0.02

0.04

0.06

0.08

φ=0.50φ=0.55φ=0.60φ=0.65


(a)

(b)


CH

-Che

milu

min

esce

nce/

Uc (

a.u.

)

0.00

0.02

0.04

0.06

0.08

0.10

0.12

φ=0.50φ=0.55φ=0.60φ=0.65


CO

2-C

hem

ilum

ines

cenc

e/U

c (a.

u.)

0.00

0.02

0.04

0.06

0.08

φ=0.50φ=0.55φ=0.60φ=0.65


(a)

(b)


(Bac

kgro

und-

subt

ract

edC

H-C

hem

ilum

ines

cenc

e)/U

c (a.

u.)

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08φ=0.50φ=0.55φ=0.60φ=0.65

Fig. 9 CO2-background subtracted overall CH-chemiluminescence intensity normalized by the combustor inlet velocity for different equivalence ratios


(Bac

kgro

und-

subt

ract

edC

H-C

hem

ilum

ines

cenc

e)/U

c (a.

u.)

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08φ=0.50φ=0.55φ=0.60φ=0.65

Fig. 9 CO2-background subtracted overall CH-chemiluminescence intensity normalized by the combustor inlet velocity for different equivalence ratios


7

CO2-background subtracted overall CH-chemiluminescence intensity is constant with respect to the change of inlet combustor velocity after it is normalized by the combustor inlet velocity, suggesting there exists no apparent effect of turbulence on the relationship between CO2-background subtracted overall CH-chemiluminescence, and heat release and it is a good measure of overall heat release for stable premixed flames over the velocity and equivalence ratio ranges run in this study.

As mentioned previously, by changing the length of the combustor unstable combustion is achieved over a part of the operating conditions.

Fig. 10 shows a typical time traces of combustor pressure, CH- and CO2-chemiluminescence intensity during an unstable combustion. Fig. 11 shows the plot of the CO2-background subtracted mean CH-chemiluminescence intensity normalized by Uc versus equivalence ratio for both stable and unstable flames. It can be noted that for a fixed equivalence ratio the relationship between the normalized mean value of the CO2-background subtracted CH-chemiluminescence intensity and equivalence ratio does not change with either the amount of fuel or the flame stability. This means that, though CH-chemiluminescence corrected for the background CO2-chemiluminescence fluctuates during unstable combustion, its mean value does not change compared to that of stable combustion. In other words, the background corrected CH-chemiluminescence is a good measure of instantaneous overall heat release under unstable combustion where there is no equivalence ratio fluctuation.

Fig. 12 is a plot of the RMS fluctuation of CO2-background subtracted CH-chemiluminescence intensity normalized by its mean value (as a measure of normalized heat release) vs. RMS fluctuation of combustor pressure normalized by the combustor inlet velocity for all unstable combustion cases. Using the linear acoustics approximation of acoustic velocity fluctuation (i.e. urms= Prms/ρc), the velocity fluctuation level can be estimated and the normalized velocity fluctuation corresponding to the given pressure fluctuation is shown in the top axis. It shows that the normalized heat release increases linearly as the pressure fluctuation increases until it becomes saturated at high enough pressure fluctuation for the normalized velocity fluctuation to be around 17 %.

SUMMARY AND CONCLUSIONS

The effect of flow parameters such as mean mixture velocity, equivalence ratio and fuel-air premixing on overall CH- and CO2- chemiluminescence is studied experimentally for a swirl-stabilized turbulent lean premixed combustor. A detailed spectroscopic measurement of chemiluminescence is made for each flow condition over the wavelength of 405-495 nm. Using the results from chemiluminescence spectrum measurements, the CH-chemiluminescence collected using a narrow bandpass filter is corrected for the background chemiluminescence from CO2*. The change of

0.00 0.01 0.02 0.03 0.04

Pc'

(psi

)

-0.8

-0.4

0.0

0.4

0.8

0.00 0.01 0.02 0.03 0.04CH

-Che

milu

min

esce

nce

(a.u

.)

0.0

0.2

0.4

0.6

0.8

Time (sec)0.00 0.01 0.02 0.03 0.04C

O2-C

hem

ilum

ines

cenc

e(a

.u.)

0.0

0.1

0.2

0.3

0.4

Fig. 10 Time traces of combustor pressure (Pc’), CH- and CO2-chemiluminescence intensity during unstable combustion

0.00 0.01 0.02 0.03 0.04

Pc'

(psi

)

-0.8

-0.4

0.0

0.4

0.8

0.00 0.01 0.02 0.03 0.04CH

-Che

milu

min

esce

nce

(a.u

.)

0.0

0.2

0.4

0.6

0.8

Time (sec)0.00 0.01 0.02 0.03 0.04C

O2-C

hem

ilum

ines

cenc

e(a

.u.)

0.0

0.1

0.2

0.3

0.4

Fig. 10 Time traces of combustor pressure (Pc’), CH- and CO2-chemiluminescence intensity during unstable combustion

Fig. 11 CO2-background subtracted overall CH-chemiluminescence intensity normalized by combustor inlet velocity versus equivalence ratio for both stable and unstable flames

Equivalence ratio0.48 0.50 0.52 0.54 0.56 0.58 0.60 0.62 0.64 0.66

LOG

[(CH

*-C

O2*

)/Uc]

-1.9

-1.8

-1.7

-1.6

-1.5

-1.4

-1.3

Stable combustionUnstable combustion

Fig. 11 CO2-background subtracted overall CH-chemiluminescence intensity normalized by combustor inlet velocity versus equivalence ratio for both stable and unstable flames

Equivalence ratio0.48 0.50 0.52 0.54 0.56 0.58 0.60 0.62 0.64 0.66

LOG

[(CH

*-C

O2*

)/Uc]

-1.9

-1.8

-1.7

-1.6

-1.5

-1.4

-1.3

Stable combustionUnstable combustion


8

fuel-air premixedness results in the change of the shape of chemiluminescence spectrum as well as its magnitude. The background-CO2 subtracted CH- chemiluminescence is found to be more sensitive to the premixedness of fuel-air mixture than CO2- chemiluminescence.

For a fixed equivalence ratio, the relationship between the normalized mean value of the CO2-background subtracted overall CH-chemiluminescence intensity and equivalence ratio does not change with respect to either the combustor inlet velocity or the flame stability. It means that there exists no apparent effect of turbulence and unsteadiness associated with unstable combustion on the relationship between CO2-background subtracted overall CH-chemiluminescence and heat release, and the background corrected CH-chemiluminescence is a good measure of instantaneous overall heat release from flames under unstable as well as stable combustion where there is no equivalence ratio fluctuation.

.

ACKNOWLEDGMENTS

The Authors are grateful for the support provided by University Turbine Systems Research Program of the Department of Energy under DOE Award

DE-FC26-02NT41431. The authors would also like to acknowledge the support of Mr. Esteban Gonzalez-Juez by a NSF Graduate Fellowship (NSF grant number DGE-0338240).

Appendix A - Spectroscopy system calibration

In order to determine the detection system

efficiency, a light source with a known spectral irradiance is required. A tungsten lamp calibrated at 650 nm was used for this purpose. Since the lamp was calibrated at 650 nm, the relationship between the color temperature and current setting is known only at the wavelength of 650 nm. Therefore, it requires a conversion process to determine the spectral irradiance at different wavelengths for a given current to the lamp.

The spectral irradiance (Eλ) from an object at the temperature of T is given by the Planck's law as:

where ελ is the emissivity of tungsten filament which is a function of temperature (T) and wavelength (λ) [12], C1=3.7418x10-16 Wm2, C2=1.4388x10-2 mK and Ta is the color

Prms/Uc, psi/(m/sec)0.000 0.025 0.050 0.075 0.100 0.125 0.150

(CH

-CO

2)rm

s/(C

H-C

O2)

mea

n

0.0

0.2

0.4

0.6

0.8

urms/Um (%)

0 5 10 15 20 25 30Uc=4 m/sec, φ=0.550Uc=5 m/sec, φ=0.550Uc=6 m/sec, φ=0.550Uc=7 m/sec, φ=0.550Uc=8 m/sec, φ=0.550Uc=3 m/sec, φ=0.575Uc=3 m/sec, φ=0.575Uc=4 m/sec, φ=0.575Uc=5 m/sec, φ=0.575Uc=6 m/sec, φ=0.575Uc=7 m/sec, φ=0.575Uc=8 m/sec, φ=0.575Uc=3 m/sec, φ=0.600Uc=4 m/sec, φ=0.600Uc=5 m/sec, φ=0.600

Fig. 12 Standard deviation of CO2-background subtracted overall CH-chemiluminescence intensity normalized by its mean value vs. RMS fluctuation of combustor pressure normalized by the combustor inlet velocity for all unstable combustion cases

Prms/Uc, psi/(m/sec)0.000 0.025 0.050 0.075 0.100 0.125 0.150

(CH

-CO

2)rm

s/(C

H-C

O2)

mea

n

0.0

0.2

0.4

0.6

0.8

urms/Um (%)

0 5 10 15 20 25 30Uc=4 m/sec, φ=0.550Uc=5 m/sec, φ=0.550Uc=6 m/sec, φ=0.550Uc=7 m/sec, φ=0.550Uc=8 m/sec, φ=0.550Uc=3 m/sec, φ=0.575Uc=3 m/sec, φ=0.575Uc=4 m/sec, φ=0.575Uc=5 m/sec, φ=0.575Uc=6 m/sec, φ=0.575Uc=7 m/sec, φ=0.575Uc=8 m/sec, φ=0.575Uc=3 m/sec, φ=0.600Uc=4 m/sec, φ=0.600Uc=5 m/sec, φ=0.600

Fig. 12 Standard deviation of CO2-background subtracted overall CH-chemiluminescence intensity normalized by its mean value vs. RMS fluctuation of combustor pressure normalized by the combustor inlet velocity for all unstable combustion cases

)1()1( 22 51

51

, −=

−⋅=⋅=

aTCTCb eC

eCEE λλλλλλ λλ

εε


9

temperature. Then, the actual temperature of the tungsten filament (T) can be calculated as:

From this, the color temperature at a different wavelength for the same current setting can be calculated using the emissivity data of tungsten at the filament temperature and the wavelength as:

Figure A-1 shows the calculated color temperature per current applied across the tungsten filament of the lamp for wavelengths of 430 and 470 nm as well as for the calibrated wavelength of 650 nm. The wavelength of 430 nm is chosen because it is the center wavelength of a narrow bandpass filter which is commonly used to measure the overall CH* chemiluminescence. Also, the wavelength of 470 nm is chosen as the center wavelength of a narrow bandpass for CO2* chemiluminescence

measurement. From the relationship between the color temperature and current, the spectral irradiance of tungsten lamp at a given current setting can be calculated using the Planck's law. This is shown in Fig. A-2.

Table A-1 summarizes these results for two lamp-current settings. It shows that the color temperature decreases as the wavelength increases for a given current setting. Also, for a given lamp-current setting the color temperatures at 430 nm or 470 nm are quite different from that calibrated at the wavelength of 650 nm. This suggests that the spectral irradiance curve determined from the calibration at 650 nm will not correctly represent the actual spectral irradiance from the tungsten lamp. This is illustrated in Table A-1 by comparing the ratios of spectral irradiance (Eλ=470=nm /Eλ=430=nm) at 430 nm and 470 nm: one based on the calibrated relationship between color

[ ]1)1(ln 650,6502

650650

2

+−⋅=

aTCeCT λ

λελ

⎥⎥⎦

⎤

⎢⎢⎣

⎡+−⋅

=

1)1(1ln 2

2,

TCnew

newa

new

new

e

CTλ

λελ

* Based on the relationship between colortemperature and lamp-current calculatedat each wavelength

** Based on the calibrated relationship betweencolor temperature and lamp-current at 650 nm

Table A-1. Color temperature and the ratio of irradiance for the lamp-current settings of 20.0 and 25.0 A

2.953.98(Eλ=470= nm /Eλ=430=nm)2**

2.593.50(Eλ=470= nm /Eλ=430=nm)1*

1866.01560.0(Ta)@ 650 nm (K)

1909.51588.9(Ta)@ 470 nm (K)

1919.21595.5(Ta)@ 430 nm (K)

2007.61655.6T (K)

25.020.0Lamp current (A)

2.953.98(Eλ=470= nm /Eλ=430=nm)2**

2.593.50(Eλ=470= nm /Eλ=430=nm)1*

1866.01560.0(Ta)@ 650 nm (K)

1909.51588.9(Ta)@ 470 nm (K)

1919.21595.5(Ta)@ 430 nm (K)

2007.61655.6T (K)

25.020.0Lamp current (A)

Fig. A-2 Calculated spectral irradiance for different lamp-current settings

Wavelength (nm)

400 410 420 430 440 450 460 470 480 490 500

Spec

tral

irra

dian

ce (a

.u.)

Current = 25.0 ACurrent = 22.5 A

Current = 20.0 A

Fig. A-2 Calculated spectral irradiance for different lamp-current settings

Wavelength (nm)

400 410 420 430 440 450 460 470 480 490 500

Spec

tral

irra

dian

ce (a

.u.)

Current = 25.0 ACurrent = 22.5 A

Current = 20.0 A

Fig. A-1 Color temperature vs. lamp-current settings for different wavelengths

Current (A)

18 19 20 21 22 23 24 25 26 27 28

Col

or te

mpe

ratu

re (K

)

1500

1600

1700

1800

1900

2000

λ= 430 nmλ= 470 nm

λ= 650 nm

Fig. A-1 Color temperature vs. lamp-current settings for different wavelengths

Current (A)

18 19 20 21 22 23 24 25 26 27 28

Col

or te

mpe

ratu

re (K

)

1500

1600

1700

1800

1900

2000

λ= 430 nmλ= 470 nm

λ= 650 nm Fig. A-3 Detection system efficiency over the wavelength between 395 and 405 nm

Col 1 vs Col 2

Wavelength (nm)

395 396 397 398 399 400 401 402 403 404 405

Mea

sure

d in

tens

ity (a

.u.)

0

10000

20000

30000

40000

50000

60000

Rel

ativ

e irr

adia

nce

from

lam

p (a

.u.)

0.0

5.0e-11

1.0e-10

1.5e-10

2.0e-10

2.5e-10

Rel

ativ

e sy

stem

effi

cien

cy (a

.u.)

0

2

4

6

8

10

Measured intensity

Irradiance from tungsten lamp

Relative system efficiency

Fig. A-3 Detection system efficiency over the wavelength between 395 and 405 nm

Col 1 vs Col 2

Wavelength (nm)

395 396 397 398 399 400 401 402 403 404 405

Mea

sure

d in

tens

ity (a

.u.)

0

10000

20000

30000

40000

50000

60000

Rel

ativ

e irr

adia

nce

from

lam

p (a

.u.)

0.0

5.0e-11

1.0e-10

1.5e-10

2.0e-10

2.5e-10

Rel

ativ

e sy

stem

effi

cien

cy (a

.u.)

0

2

4

6

8

10

Measured intensity

Irradiance from tungsten lamp

Relative system efficiency


10

temperature and lamp-current at 650 nm and that calculated at each wavelength.

The detection system efficiency is determined from the measured light intensity and the calculated irradiance for a fixed lamp-current setting. Typical detection system efficiency over the wavelength between 395 nm and 405 nm is shown in Fig. A-3. It should be noted that the system efficiency is not constant but changes with respect to the wavelength of light.

REFERENCES

[1] W.-P. Shih, J.G. Lee and D.A. Santavicca,

"Stability and emissions characteristics of a lean premixed gas turbine combustor," Proceedings of the Combustion Institute, The Combustion Institute, Vol. 26, pp. 2771-2778, 1996.

[2] J.R. Roby, J.E. Reaney and E.L. Johnsson, "Detection of temperature and equivalence ratio in turbulent premixed flames using chemiluminescence", FACT-Vol.22, 1998 International Joint Power Generation Conference, Volume 1, ASME 1998.

[3] J.G. Lee, K. Kim and D.A. Santavicca, "Measurement of equivalence ratio fluctuation and its effect in heat release during unstable combustion," Proceedings of the Combustion Institute, Volume 28, 2000, pp. 415-421.

[4] T.M. Muruganandam, B. Kim, R. Olsen, M. Patel, B. Romig and J.M. Seitzman, "Chemiluminescence based sensors for turbine engines," AIAA 2003-4490, 39th AIAA/ASME/SAE/ASEE Joint Propulsion Conference and Exhibit, July 20-23, 2003, Huntsville, Alabama.

[5] S.A. Miller, "Development of a flame chemiluminescence probe for determination of primary zone equivalence ratio in gas turbine combustors," M.S. Thesis, Dept. of Mechanical and Nuclear Engineering, The Pennsylvania State University, University Park, PA, August 1999.

[6] T.P. Clark, "Studies of OH, CO, CH, and C2 radiation from laminar and turbulent propane-air and ethylene-air flames," NACA Technical Note 4266, 1958.

[7] I.R. Hurle, R.B. Price, T.M. Sugden and A. Thomas, "Sound emission from open turbulent premixed flames," Proceedings of the Royal Society of London Series A., Vol. 303: 409-427, 1968.

[8] J.M. Samaniego, F.N. Egolfopoulos and C.T. Bowman, "CO2* chemiluminescence in premixed flames," Combustion Science and Technology, Vol.109 (1-6): 183-203 1995.

[9] H.N. Najm, P.H. Paul, C.J. Mueller and P.S. Wyckoff, "On the adequacy of certain experimental observables as measurements of flame burning rate," Combustion and Flame, Vol. 113 (3): 312-332, May 1998.

[10] B. Higgins, M.Q. McQuay, F. Lacas, J.C. Rolon, N. Darabiha and S. Candel, "Systematic measurements of OH chemiluminescence for fuel-lean, high-pressure, premixed, laminar flames," Fuel, Vol. 80 (1): 67-74, Jan 2001.

[11] B. Higgins, M.Q. McQuay, F. Lacas and S. Candel, "An experimental study on the effect of pressure and strain rate on CH chemiluminescence of premixed fuel-lean methane/air flames," Fuel, Vol.80 (11): 1583-1591, Sep 2001.

[12] Larrabee, R., "Spectral emissivity of tungsten," J. of the Optical Society of America, 1959.

[American Institute of Aeronautics and Astronautics 41st AIAA/ASME/SAE/ASEE Joint Propulsion...

Documents

Transcript of [American Institute of Aeronautics and Astronautics 41st AIAA/ASME/SAE/ASEE Joint Propulsion...