Comparison of Thermocouple Temperature Measurements of Simple ...
5th Asia-Pacific Conference on Combustion, The University of Adelaide, Adelaide, Australia 17-20 July 2005
Comparison of Thermocouple Temperature Measurements of Simple and Precessing Jet Propane Flames 1
1
2
1
E. J. Smith , G. J. Nathan , N. H. Qamar , and B. B. Dally 1
School of Mechanical Engineering The University of Adelaide, S.A. 5005, AUSTRALIA 2
School of Chemical Engineering The University of Adelaide, S.A. 5005, AUSTRALIA While the findings of both Newbold et al. [7] and Qamar et al. [8] are significant, the local instantaneous temperature is also required to advance understanding of the relationship between flame strain, temperature and soot in turbulent flames. Such data has not previously been available.
Abstract An experimental study was conducted to compare the temperature characteristics of simple and precessing jet flames of commercial propane. A fine wire R-type thermocouple is used to measure both axial and radial temperature profiles. The measurements are corrected for effects of radiation, convection and conduction. The results show that the overall maximum temperatures are similar in both the precessing and simple jet flames, but are also affected by the amount of soot. In general, the precessing jet flame has lower mean temperatures. The differences in temperature between the two flames and the influence of flame strain on temperature, through its influence on soot, are discussed.
This current work aims to provide such measurements of temperature in two turbulent diffusion flames, a precessing jet and a simple jet flame using a fine wire R-Type thermocouple.
2 Background on the Precessing Jet Cement kilns require high temperatures of approximately 1450ºC to drive chemical reactions in the bed. This is usually supplied by the combustion of fossil fuels. One method for reducing the NOx emissions and improving overall combustion within rotary kilns firing either gas or solid fuels is the use of the precessing jet technology (Nathan et al. [4]).
1 Introduction It is well understood that temperature and mixing rates play significant roles in pollutant formation and emissions, heat transfer, soot formation and destruction in turbulent flames. In diffusion flames especially, the relationship between flame temperature, soot and radiation are interrelated. An increase in flame temperature enhances soot formation, which in turn increases radiation and lowers the flame temperature. Higher flame temperatures also increase the rate of soot oxidation. At the same time, increased strain rates in a reaction zone act to reduce soot volume fraction.
Jet precession can be generated by a naturally occurring instability within an axisymmetric chamber downstream from a large sudden expansion. A schematic diagram of the precessing jet nozzle is shown in figure 1. Flow enters the precessing jet (PJ) chamber through a large sudden expansion (Nathan et al. [4]). The jet flow deflects asymmetrically to the wall of the chamber, where it reattaches. An unstable pressure field is created in the nozzle chamber causing the attached flow to precess azimuthally as it travels through the nozzle. On exiting the chamber the jet flow is deflected by a small lip and leaves at an angle of 40-60o to the nozzle geometric axis.
To assess the complex relationships between all of these parameters, under conditions of relevance to practical turbulent combustion systems, previous experimental studies have compared three types of turbulent diffusion flames. The selected flames span high to low global mixing rates but are otherwise matched as closely as possible. They are a momentum dominated simple jet flame; buoyancy dominated precessing jet flame and a highly strained bluff body flame (Newbold et al. [7] and Qamar et al. [8]).
Orifice
Cylinder
Centrebody
Lip
Fuel
Precession
Radiant fraction, global residence time, NOx emissions and global temperature (adiabatic flame temperature corrected for radiation) have been measured for each of these flames by Newbold et al. [7]. It has been found that the precessing jet flame has the highest global residence time and the lowest global flame temperature. Recently, local instantaneous soot volume fraction, time averaged soot volume fraction and total soot volume have been measured in the same flames (Qamar et al. [8]). It has been found that the precessing jet has about 2.5 times more total soot volume than the simple jet and about 2 orders of magnitudes more than the bluff body flame. The precessing jet flame also has the highest local instantaneous soot volume fraction.
Figure 1. A schematic diagram of the precessing jet (PJ) nozzle (Newbold et al., [5]) The internal and initial emerging flow from a PJ nozzle is three dimensional and highly unsteady, resulting in initial rates of spread and decay that are much greater than those of simple jet (SJ) flows. Within the initial region large-scale flow structures fold the air and fuel together to produce a low strain, highly radiant flame. The low strain flame promotes the formation of soot, increasing radiant heat transfer and simultaneously reducing NOx emissions (Newbold. [6]).
89
The size of the thermocouple bead will greatly affect the response time of the thermocouple and hence both measurement accuracy and the rate at which data should be collected. To assess its effect on accuracy, temperature measurements were also performed in the simple jet flame using a thermocouple with a wire diameter of 250µm and a bead diameter of 800µm. Figure 3 compares the results obtained with the two thermocouples. The faster response of the fine wire 50µm thermocouple results in measured temperatures whose maximum and minimum values differ significantly from the mean. However the response of the 250µm thermocouple is so poor that there is little difference between the minimum, maximum and mean measured temperatures. Both thermocouples collect similar mean temperatures, giving confidence in the mean data from the 50µm wire.
3 Experimental Technique The burners used in this experimental study are the same as those used by Newbold et al. [7] and Qamar et al. [8]. The long pipe burner has internal diameter d=5mm with a long development length, L/d=100, to ensure fully developed pipe flow at the exit. The precessing jet burner comprises an axisymmetric sudden expansion chamber, with internal diameter D=25mm, attached to the exit of a long pipe burner, with internal diameter of d=5mm. Commercial grade bottled propane, with 98% propane, was fed via pressure and flow rate regulators to the inlet of each burner. The Reynolds number at the outlet of the long pipe burner and inlet to the precessing jet chamber was maintained at 18,500, providing a constant velocity of 16.9m/s (matching the initial conditions of Qamar et al. [8]). (Note that the Reynolds number of 18,500 was observed to provide a flow that was consistently in the precessing jet mode. A PJ without a centrebody (figure 1) requires a Reynolds number of 20,000 to ensure consistent precession (Nathan et al. [4])).
1800 Max Mean Min Max Mean Min
Temperature (Deg_C)
1600
The burners were mounted under the middle of an exhaust hood. The distance between the burner and the hood was kept constant. Temperature measurements were collected in unconfined surroundings; however the presence of a nearby wall caused a bias in the direction of the flames. Each flame tilted slightly towards to wall, but this was more noticeable for the PJ flame since it has the lowest momentum. This bias in flame direction matched those in the investigation of Qamar et al. [8], who used the same exhaust hood for their experiments. The thermocouples were mounted to a traverse and moved radially and axially throughout the flame. The traverse has a measurement range in both axially and radially directions of 1 meter and a position accuracy of 0.025mm.
1400 1200
50um 50um 50um 250um 250um 250um
1000 800 600 400 200 0 0
1
2
3
r/d
4
5
6
7
Figure 3. R-Type thermocouple comparison for 50µm and 250µm diameter wire in a simple Jet flame at x/d=40. (diameter of jet, d=5mm) Since the simple jet and precessing jet both produce turbulent sooty flames some soot build-up can be expected on the thermocouple bead. The larger the thermocouple bead, the more soot is deposited and the greater its effect on temperature measurement. However, for the 50µm thermocouple, the effect of soot was found to be minimal, with no soot deposit observed. This is consistent with Goh et al. [3], who found that, when measuring in highly sooting flames with a fine wire thermocouple (bead diameter < 200 µm) the amount of soot deposited on the bead is minimal and is instantly burnt off.
The temperature was measured using an in-house-made thermocouple of type R (Platinum-Platinum 13% Rhodium) with bead diameter of 110µm and wire diameter of 50µm. The thermocouple configuration is shown in figure 2. The thermocouple was connected to a thermocouple transmitter (electronic cold junction). A data logger collected the measurements with a sampling rate of 1 kHz for 6 minutes at each location, for both flames. The measured temperatures were corrected to account for radiation, convection, and conduction loses using an energy balance equation for the thermocouple bead following Frinstrom et al. [2].
4 Results and Discussion
R-TYPE (Pt/13%Rh-Pt)
φ 6.35 mm
φ 1.59 mm
φ 0.05 mm
400 mm
23 mm
4 mm
Simple jet and precessing jet axial mean temperature decay and integrated soot volume are compared in figure 4. Measurements for soot volume are from Qamar et al. [8]. Axial distance is normalised by the simple jet diameter and the inlet to the PJ nozzle, that is d=5mm.
A Ceramic Insulating rod (kept thin to reduce perturbation of the flow) 5m R-Type lead (connected to T/C transmitter)
The mean centre-line temperature of the PJ peaks at ~11000C in the near field region at x/d=20. The temperature then decays in the far field. The peak PJ integrated soot volume fraction of ~3.2×10-3 mm2 occurs downstream, at x/d=60. As soot volume fraction increases with distance, the mean temperature decreases, consistent with the effect of increased radiant cooling. The mean centre-line temperature of the simple jet peaks in the far field with a temperature of ~12200C at x/d=140. The SJ soot volume also peaks at x/d=140 with ~1.1×10-3 mm2. The rise in soot volume follows the rise in mean temperature.
Weld Bead φ110um
x
r Very Fine Thermocouple wire (φ50um = 0. 05 mm – for fast response time)
Detail at A Not to Scale
PJ Nozzle
Figure 2. Schematic diagram of the fine wire (50µm) RTYPE (Pt/13%Rh-Pt) thermocouple (0-1700oC) configuration.
90
The temperatures in the PJ flame are more influenced by soot due to the higher volume of soot in it. The peak mean temperature is reached close to the base of the flame where the amount of soot is relatively low. The mean temperature then reduces as soot volume increases, and then with greater dilution toward the flame tip. The situation is different for the simple jet flame, which has less total soot. Hence the two peak at around the same location, demonstrating different dynamics.
Temperature (Deg_C)
1200
PJ Soot Volume Peak
SJ Temperature Peak 3.0
1000
2.0
800 600
SJ Soot Volume Peak
400
1.0
200
0
20
40
60
80
100
120
140
160
180
Mean (mid flame)
1400
Mean (flame tip) Geometric Centreline
1200 1000 800 600 400 200 0
0
0
Mean (flame base)
1600 Temperature (Deg_C)
PJ Temperature Peak
1800
Soot volume x10-3 per mm (mm2)
1400
jet (x/d=20) and near to the flame tip for the simple jet (x/d=140). The general shape of the profiles in Figure 7, are similar for the two flames, but subtle differences are also evident. The maximum instantaneous temperatures measured in both jets are at least 400 oC less than the adiabatic fuel temperature for propane, which is 1,994oC (Turns [9]). This is consistent with the flame being cooled by the presence of soot.
-30
-20
-10
200
0
10
20
30
r/d
x/d
Figure 6. PJ Radial profiles of mean temperature at three downstream locations, flame base (x/d=20), mid flame (x/d=60) and flame tip (x/d=120).
Figure 4. Comparison of PJ and SJ mean axial temperature decay and soot volume fraction (soot data from Qamar et al. [8])
1800
Three downstream radial profiles for the simple jet are presented in figure 5, at the flame base (x/d=40, just above lift off), mid flame (x/d=100), and flame tip (x/d=160). The profile at the flame base shows the distinctive double hump structure, showing that reaction occurs at the outer edge of the flame. The middle of the flame profile exhibits a flat region on the axis and a bell profile is found in the flame tip region. The profile at the flame tip shows a bias towards the right; this is due to the presence of the wall.
Temperature (Deg_C)
1600
Temperature (Deg_C)
1000 800 600 400 0
Mean (flame base) Mean (mid flame) Mean (flame tip) Geometric Centreline
1400
1200
200
1800 1600
1400
-30
-20
SJ: Max x/d=140 PJ: Max x/d=60 PJ: Max x/d=20
1200 1000
-10
0 r/d
SJ: Mean x/d=140 PJ: Mean x/d=60 PJ: Mean x/d=20
10
20
30
SJ: Min x/d=140 PJ: Min x/d=60 PJ: Min x/d=20
800
Figure 7. Radial profiles of mean, max and min temperature for the SJ and PJ flames at the axial locations of peak mean temperature (PJ: x/d=20 and SJ: x/d=140) and peak soot volume (PJ: x/d=60 and SJ: x/d=140).
600 400 200 0 -30
-20
-10
0
10
20
The maximum temperatures are chosen as being a representative, although not absolute, measure of the local reaction zone temperature. The spatial location of the local reaction zones, whose thickness is of the order of millimetres, fluctuates with time and space in turbulent diffusion flames (Donbar et al. [1]). These lead to spikes in the instantaneous temperature, although not all local peaks in temperature are necessarily associated with the high temperature reaction zone. To determine the reaction zone temperatures thus requires conditional measurements involving multiple species. However, the absolute maximum of the instantaneous data sets at any given location within a flame must clearly be associated with a reaction zone. Hence this measurement can be used to assess how reaction zone temperatures vary within a flame or between different flames.
30
r/d
Figure 5. SJ Radial profiles of mean temperature at three downstream locations, flame base (x/d=40), mid flame (x/d=100) and flame tip (x/d=160). Three downstream radial profiles for the precessing jet are presented in figure 6, at the flame base (x/d=20), mid flame (x/d=60), and flame tip (x/d=120). All radial profiles are bell shaped. Flame temperature is higher towards the flame base and decreases downstream towards the tip. The mid flame and flame tip profile also show a bias towards the right, due to the wall, previously mentioned. Radial profiles of mean, minimum and maximum temperature are presented in figure 7 at the axial height of peak temperature, that is, near to the flame base for the precessing
For both jets the maximum temperatures are much more uniform across the flame than are the mean values, and the
91
presence of cold ambient fluid is seen to extend well within the mean flame envelope. This suggests that the reduction in mean temperature with radial distance is dominated by fluctuations in the location of the instantaneous flame zone, rather than by significant changes to the local reaction zone. The peak precessing jet temperature near to the flame base is slightly higher than that for the simple jet near to the middle of the flame, consistent with the lower soot there. While for the PJ at x/d=60, where soot is at its maximum (figure 7), the instantaneous maximum temperature is lower than that in the SJ flame due to increased soot and, in turn, increased radiant cooling.
5 Conclusion These experimental results show that temperature is significantly affected by the differences in flame dynamics that is, mixing characteristics and soot formation. The values of maximum instantaneous temperatures are deemed to be representative indicators of trends in the reaction zones. The overall maximum values are similar in both flames, since the same fuel has been used for both jets. However at points where soot volume is higher, the local maximum temperature is reduced due to radiant cooling. Hence for the low strain precessing jet flame, where soot volume is greater, the local instantaneous maximum temperatures are reduced, thus reducing mean temperatures. The PJ flame also exhibits greater variability in temperature, due to a more dynamic flame structure produced by large buoyant structures observed elsewhere.
Throughout the PJ flame the instantaneous minimum temperatures are lower than in the SJ flame, suggesting more dynamic flame movement. This is consistent with the observations of flame dynamics of Newbold et al. [5]. Hence the profiles of minimum and maximum instantaneous temperatures profiles can provide valuable information regarding reaction zones and mixing characteristics.
6 Acknowledgments
Instantaneous temperature PDF’s (probability density function) along the centreline for the simple jet and precessing jet are shown in figures 8 and 9 respectively. The simple jet PDF shows a distinctive peak along the centreline, while for the PJ the peak is much flatter and broader. This shows that the range of temperatures within the PJ flame is much greater, consistent with the increased movement of the reaction zones due to large buoyant structures (Newbold et al. [5]).
The authors would like to acknowledge the following. Stephane Maupetit for his assistance in collecting the experimental measurements; Silvio De leso and George Osborne for assembling the thermocouple equipment; the ARC SPIRT scheme, the Sugar Research Industry (SRI) and Fuel and Combustion Technology (FCT) for their financial support of this work.
7 References [1] Donbar, J. M. and Driscoll, J. F., Reaction Zone Structure in Turbulent Nonpremixed Jet Flames – From CH-OH PLIF Images, Combustion and Flame, 122 (1/2), 2000 [2] Frinstrom, R. M. and Westenberg, A. A., Flame Structure, McGraw-Hill, 1965. [3] Goh, S. F., Kusadome, S. and Gollahalli, S. R., Effects of Jet Dilution and Co-flow on Sooting and Emission Characteristics of Hydrocarbon Fuels, Clean Air, International Journal on Energy for a Clean Environment, 4 (4), 2003. [4] Nathan, G.J., Hill, S.J. and Luxton, R.E., Axisymmetric ‘Fluidic’ Nozzle to Generate Precession, J. Fluid Mech., 370, 1998, 347-380. (9)
An Jet
[5] Newbold, G.J.R., Nathan, G.J. and Luxton, R.E., Largescale dynamics of an unconfined precessing jet flame, Combust. Sci. Tech., 126 (1-6), 53, 1997.
Figure 8. Simple jet temperature PDF along centreline.
[6] Newbold, G.J.R., Mixing and Combustion in Precessing Jet Flows, Ph.D. Thesis, The University of Adelaide, Australia, 1997. [7] Newbold, G.J.R., Nathan, G.J., Nobes, D.S. and Turns, S.R., Measurement and Prediction of NOx Emissions from Unconfined Propane Flames from Turbulent-Jet, Bluff-Body, Swirl, and Precessing Jet Burners, Proceedings of the Combustion Institute, 28, 2000, 481487. [8] Qamar N., Nathan G. J., Alwahabi, Z. T. and King, K. D. The Effect of Global Mixing on Soot Volume Fraction: Measurements in Simple Jet, Proceedings of the Combustion Institute, 30, 2005, 1493-1500. [9] Turns, S. R., An Introduction to combustion: concepts and Applications, Second Edition Mc Graw-Hill Singapore, 2000
Figure 9. Precessing jet temperature PDF along centreline.
92
Comparison of Thermocouple Temperature Measurements of Simple and Precessing Jet Propane Flames 1
1
2
1
E. J. Smith , G. J. Nathan , N. H. Qamar , and B. B. Dally 1
School of Mechanical Engineering The University of Adelaide, S.A. 5005, AUSTRALIA 2
School of Chemical Engineering The University of Adelaide, S.A. 5005, AUSTRALIA While the findings of both Newbold et al. [7] and Qamar et al. [8] are significant, the local instantaneous temperature is also required to advance understanding of the relationship between flame strain, temperature and soot in turbulent flames. Such data has not previously been available.
Abstract An experimental study was conducted to compare the temperature characteristics of simple and precessing jet flames of commercial propane. A fine wire R-type thermocouple is used to measure both axial and radial temperature profiles. The measurements are corrected for effects of radiation, convection and conduction. The results show that the overall maximum temperatures are similar in both the precessing and simple jet flames, but are also affected by the amount of soot. In general, the precessing jet flame has lower mean temperatures. The differences in temperature between the two flames and the influence of flame strain on temperature, through its influence on soot, are discussed.
This current work aims to provide such measurements of temperature in two turbulent diffusion flames, a precessing jet and a simple jet flame using a fine wire R-Type thermocouple.
2 Background on the Precessing Jet Cement kilns require high temperatures of approximately 1450ºC to drive chemical reactions in the bed. This is usually supplied by the combustion of fossil fuels. One method for reducing the NOx emissions and improving overall combustion within rotary kilns firing either gas or solid fuels is the use of the precessing jet technology (Nathan et al. [4]).
1 Introduction It is well understood that temperature and mixing rates play significant roles in pollutant formation and emissions, heat transfer, soot formation and destruction in turbulent flames. In diffusion flames especially, the relationship between flame temperature, soot and radiation are interrelated. An increase in flame temperature enhances soot formation, which in turn increases radiation and lowers the flame temperature. Higher flame temperatures also increase the rate of soot oxidation. At the same time, increased strain rates in a reaction zone act to reduce soot volume fraction.
Jet precession can be generated by a naturally occurring instability within an axisymmetric chamber downstream from a large sudden expansion. A schematic diagram of the precessing jet nozzle is shown in figure 1. Flow enters the precessing jet (PJ) chamber through a large sudden expansion (Nathan et al. [4]). The jet flow deflects asymmetrically to the wall of the chamber, where it reattaches. An unstable pressure field is created in the nozzle chamber causing the attached flow to precess azimuthally as it travels through the nozzle. On exiting the chamber the jet flow is deflected by a small lip and leaves at an angle of 40-60o to the nozzle geometric axis.
To assess the complex relationships between all of these parameters, under conditions of relevance to practical turbulent combustion systems, previous experimental studies have compared three types of turbulent diffusion flames. The selected flames span high to low global mixing rates but are otherwise matched as closely as possible. They are a momentum dominated simple jet flame; buoyancy dominated precessing jet flame and a highly strained bluff body flame (Newbold et al. [7] and Qamar et al. [8]).
Orifice
Cylinder
Centrebody
Lip
Fuel
Precession
Radiant fraction, global residence time, NOx emissions and global temperature (adiabatic flame temperature corrected for radiation) have been measured for each of these flames by Newbold et al. [7]. It has been found that the precessing jet flame has the highest global residence time and the lowest global flame temperature. Recently, local instantaneous soot volume fraction, time averaged soot volume fraction and total soot volume have been measured in the same flames (Qamar et al. [8]). It has been found that the precessing jet has about 2.5 times more total soot volume than the simple jet and about 2 orders of magnitudes more than the bluff body flame. The precessing jet flame also has the highest local instantaneous soot volume fraction.
Figure 1. A schematic diagram of the precessing jet (PJ) nozzle (Newbold et al., [5]) The internal and initial emerging flow from a PJ nozzle is three dimensional and highly unsteady, resulting in initial rates of spread and decay that are much greater than those of simple jet (SJ) flows. Within the initial region large-scale flow structures fold the air and fuel together to produce a low strain, highly radiant flame. The low strain flame promotes the formation of soot, increasing radiant heat transfer and simultaneously reducing NOx emissions (Newbold. [6]).
89
The size of the thermocouple bead will greatly affect the response time of the thermocouple and hence both measurement accuracy and the rate at which data should be collected. To assess its effect on accuracy, temperature measurements were also performed in the simple jet flame using a thermocouple with a wire diameter of 250µm and a bead diameter of 800µm. Figure 3 compares the results obtained with the two thermocouples. The faster response of the fine wire 50µm thermocouple results in measured temperatures whose maximum and minimum values differ significantly from the mean. However the response of the 250µm thermocouple is so poor that there is little difference between the minimum, maximum and mean measured temperatures. Both thermocouples collect similar mean temperatures, giving confidence in the mean data from the 50µm wire.
3 Experimental Technique The burners used in this experimental study are the same as those used by Newbold et al. [7] and Qamar et al. [8]. The long pipe burner has internal diameter d=5mm with a long development length, L/d=100, to ensure fully developed pipe flow at the exit. The precessing jet burner comprises an axisymmetric sudden expansion chamber, with internal diameter D=25mm, attached to the exit of a long pipe burner, with internal diameter of d=5mm. Commercial grade bottled propane, with 98% propane, was fed via pressure and flow rate regulators to the inlet of each burner. The Reynolds number at the outlet of the long pipe burner and inlet to the precessing jet chamber was maintained at 18,500, providing a constant velocity of 16.9m/s (matching the initial conditions of Qamar et al. [8]). (Note that the Reynolds number of 18,500 was observed to provide a flow that was consistently in the precessing jet mode. A PJ without a centrebody (figure 1) requires a Reynolds number of 20,000 to ensure consistent precession (Nathan et al. [4])).
1800 Max Mean Min Max Mean Min
Temperature (Deg_C)
1600
The burners were mounted under the middle of an exhaust hood. The distance between the burner and the hood was kept constant. Temperature measurements were collected in unconfined surroundings; however the presence of a nearby wall caused a bias in the direction of the flames. Each flame tilted slightly towards to wall, but this was more noticeable for the PJ flame since it has the lowest momentum. This bias in flame direction matched those in the investigation of Qamar et al. [8], who used the same exhaust hood for their experiments. The thermocouples were mounted to a traverse and moved radially and axially throughout the flame. The traverse has a measurement range in both axially and radially directions of 1 meter and a position accuracy of 0.025mm.
1400 1200
50um 50um 50um 250um 250um 250um
1000 800 600 400 200 0 0
1
2
3
r/d
4
5
6
7
Figure 3. R-Type thermocouple comparison for 50µm and 250µm diameter wire in a simple Jet flame at x/d=40. (diameter of jet, d=5mm) Since the simple jet and precessing jet both produce turbulent sooty flames some soot build-up can be expected on the thermocouple bead. The larger the thermocouple bead, the more soot is deposited and the greater its effect on temperature measurement. However, for the 50µm thermocouple, the effect of soot was found to be minimal, with no soot deposit observed. This is consistent with Goh et al. [3], who found that, when measuring in highly sooting flames with a fine wire thermocouple (bead diameter < 200 µm) the amount of soot deposited on the bead is minimal and is instantly burnt off.
The temperature was measured using an in-house-made thermocouple of type R (Platinum-Platinum 13% Rhodium) with bead diameter of 110µm and wire diameter of 50µm. The thermocouple configuration is shown in figure 2. The thermocouple was connected to a thermocouple transmitter (electronic cold junction). A data logger collected the measurements with a sampling rate of 1 kHz for 6 minutes at each location, for both flames. The measured temperatures were corrected to account for radiation, convection, and conduction loses using an energy balance equation for the thermocouple bead following Frinstrom et al. [2].
4 Results and Discussion
R-TYPE (Pt/13%Rh-Pt)
φ 6.35 mm
φ 1.59 mm
φ 0.05 mm
400 mm
23 mm
4 mm
Simple jet and precessing jet axial mean temperature decay and integrated soot volume are compared in figure 4. Measurements for soot volume are from Qamar et al. [8]. Axial distance is normalised by the simple jet diameter and the inlet to the PJ nozzle, that is d=5mm.
A Ceramic Insulating rod (kept thin to reduce perturbation of the flow) 5m R-Type lead (connected to T/C transmitter)
The mean centre-line temperature of the PJ peaks at ~11000C in the near field region at x/d=20. The temperature then decays in the far field. The peak PJ integrated soot volume fraction of ~3.2×10-3 mm2 occurs downstream, at x/d=60. As soot volume fraction increases with distance, the mean temperature decreases, consistent with the effect of increased radiant cooling. The mean centre-line temperature of the simple jet peaks in the far field with a temperature of ~12200C at x/d=140. The SJ soot volume also peaks at x/d=140 with ~1.1×10-3 mm2. The rise in soot volume follows the rise in mean temperature.
Weld Bead φ110um
x
r Very Fine Thermocouple wire (φ50um = 0. 05 mm – for fast response time)
Detail at A Not to Scale
PJ Nozzle
Figure 2. Schematic diagram of the fine wire (50µm) RTYPE (Pt/13%Rh-Pt) thermocouple (0-1700oC) configuration.
90
The temperatures in the PJ flame are more influenced by soot due to the higher volume of soot in it. The peak mean temperature is reached close to the base of the flame where the amount of soot is relatively low. The mean temperature then reduces as soot volume increases, and then with greater dilution toward the flame tip. The situation is different for the simple jet flame, which has less total soot. Hence the two peak at around the same location, demonstrating different dynamics.
Temperature (Deg_C)
1200
PJ Soot Volume Peak
SJ Temperature Peak 3.0
1000
2.0
800 600
SJ Soot Volume Peak
400
1.0
200
0
20
40
60
80
100
120
140
160
180
Mean (mid flame)
1400
Mean (flame tip) Geometric Centreline
1200 1000 800 600 400 200 0
0
0
Mean (flame base)
1600 Temperature (Deg_C)
PJ Temperature Peak
1800
Soot volume x10-3 per mm (mm2)
1400
jet (x/d=20) and near to the flame tip for the simple jet (x/d=140). The general shape of the profiles in Figure 7, are similar for the two flames, but subtle differences are also evident. The maximum instantaneous temperatures measured in both jets are at least 400 oC less than the adiabatic fuel temperature for propane, which is 1,994oC (Turns [9]). This is consistent with the flame being cooled by the presence of soot.
-30
-20
-10
200
0
10
20
30
r/d
x/d
Figure 6. PJ Radial profiles of mean temperature at three downstream locations, flame base (x/d=20), mid flame (x/d=60) and flame tip (x/d=120).
Figure 4. Comparison of PJ and SJ mean axial temperature decay and soot volume fraction (soot data from Qamar et al. [8])
1800
Three downstream radial profiles for the simple jet are presented in figure 5, at the flame base (x/d=40, just above lift off), mid flame (x/d=100), and flame tip (x/d=160). The profile at the flame base shows the distinctive double hump structure, showing that reaction occurs at the outer edge of the flame. The middle of the flame profile exhibits a flat region on the axis and a bell profile is found in the flame tip region. The profile at the flame tip shows a bias towards the right; this is due to the presence of the wall.
Temperature (Deg_C)
1600
Temperature (Deg_C)
1000 800 600 400 0
Mean (flame base) Mean (mid flame) Mean (flame tip) Geometric Centreline
1400
1200
200
1800 1600
1400
-30
-20
SJ: Max x/d=140 PJ: Max x/d=60 PJ: Max x/d=20
1200 1000
-10
0 r/d
SJ: Mean x/d=140 PJ: Mean x/d=60 PJ: Mean x/d=20
10
20
30
SJ: Min x/d=140 PJ: Min x/d=60 PJ: Min x/d=20
800
Figure 7. Radial profiles of mean, max and min temperature for the SJ and PJ flames at the axial locations of peak mean temperature (PJ: x/d=20 and SJ: x/d=140) and peak soot volume (PJ: x/d=60 and SJ: x/d=140).
600 400 200 0 -30
-20
-10
0
10
20
The maximum temperatures are chosen as being a representative, although not absolute, measure of the local reaction zone temperature. The spatial location of the local reaction zones, whose thickness is of the order of millimetres, fluctuates with time and space in turbulent diffusion flames (Donbar et al. [1]). These lead to spikes in the instantaneous temperature, although not all local peaks in temperature are necessarily associated with the high temperature reaction zone. To determine the reaction zone temperatures thus requires conditional measurements involving multiple species. However, the absolute maximum of the instantaneous data sets at any given location within a flame must clearly be associated with a reaction zone. Hence this measurement can be used to assess how reaction zone temperatures vary within a flame or between different flames.
30
r/d
Figure 5. SJ Radial profiles of mean temperature at three downstream locations, flame base (x/d=40), mid flame (x/d=100) and flame tip (x/d=160). Three downstream radial profiles for the precessing jet are presented in figure 6, at the flame base (x/d=20), mid flame (x/d=60), and flame tip (x/d=120). All radial profiles are bell shaped. Flame temperature is higher towards the flame base and decreases downstream towards the tip. The mid flame and flame tip profile also show a bias towards the right, due to the wall, previously mentioned. Radial profiles of mean, minimum and maximum temperature are presented in figure 7 at the axial height of peak temperature, that is, near to the flame base for the precessing
For both jets the maximum temperatures are much more uniform across the flame than are the mean values, and the
91
presence of cold ambient fluid is seen to extend well within the mean flame envelope. This suggests that the reduction in mean temperature with radial distance is dominated by fluctuations in the location of the instantaneous flame zone, rather than by significant changes to the local reaction zone. The peak precessing jet temperature near to the flame base is slightly higher than that for the simple jet near to the middle of the flame, consistent with the lower soot there. While for the PJ at x/d=60, where soot is at its maximum (figure 7), the instantaneous maximum temperature is lower than that in the SJ flame due to increased soot and, in turn, increased radiant cooling.
5 Conclusion These experimental results show that temperature is significantly affected by the differences in flame dynamics that is, mixing characteristics and soot formation. The values of maximum instantaneous temperatures are deemed to be representative indicators of trends in the reaction zones. The overall maximum values are similar in both flames, since the same fuel has been used for both jets. However at points where soot volume is higher, the local maximum temperature is reduced due to radiant cooling. Hence for the low strain precessing jet flame, where soot volume is greater, the local instantaneous maximum temperatures are reduced, thus reducing mean temperatures. The PJ flame also exhibits greater variability in temperature, due to a more dynamic flame structure produced by large buoyant structures observed elsewhere.
Throughout the PJ flame the instantaneous minimum temperatures are lower than in the SJ flame, suggesting more dynamic flame movement. This is consistent with the observations of flame dynamics of Newbold et al. [5]. Hence the profiles of minimum and maximum instantaneous temperatures profiles can provide valuable information regarding reaction zones and mixing characteristics.
6 Acknowledgments
Instantaneous temperature PDF’s (probability density function) along the centreline for the simple jet and precessing jet are shown in figures 8 and 9 respectively. The simple jet PDF shows a distinctive peak along the centreline, while for the PJ the peak is much flatter and broader. This shows that the range of temperatures within the PJ flame is much greater, consistent with the increased movement of the reaction zones due to large buoyant structures (Newbold et al. [5]).
The authors would like to acknowledge the following. Stephane Maupetit for his assistance in collecting the experimental measurements; Silvio De leso and George Osborne for assembling the thermocouple equipment; the ARC SPIRT scheme, the Sugar Research Industry (SRI) and Fuel and Combustion Technology (FCT) for their financial support of this work.
7 References [1] Donbar, J. M. and Driscoll, J. F., Reaction Zone Structure in Turbulent Nonpremixed Jet Flames – From CH-OH PLIF Images, Combustion and Flame, 122 (1/2), 2000 [2] Frinstrom, R. M. and Westenberg, A. A., Flame Structure, McGraw-Hill, 1965. [3] Goh, S. F., Kusadome, S. and Gollahalli, S. R., Effects of Jet Dilution and Co-flow on Sooting and Emission Characteristics of Hydrocarbon Fuels, Clean Air, International Journal on Energy for a Clean Environment, 4 (4), 2003. [4] Nathan, G.J., Hill, S.J. and Luxton, R.E., Axisymmetric ‘Fluidic’ Nozzle to Generate Precession, J. Fluid Mech., 370, 1998, 347-380. (9)
An Jet
[5] Newbold, G.J.R., Nathan, G.J. and Luxton, R.E., Largescale dynamics of an unconfined precessing jet flame, Combust. Sci. Tech., 126 (1-6), 53, 1997.
Figure 8. Simple jet temperature PDF along centreline.
[6] Newbold, G.J.R., Mixing and Combustion in Precessing Jet Flows, Ph.D. Thesis, The University of Adelaide, Australia, 1997. [7] Newbold, G.J.R., Nathan, G.J., Nobes, D.S. and Turns, S.R., Measurement and Prediction of NOx Emissions from Unconfined Propane Flames from Turbulent-Jet, Bluff-Body, Swirl, and Precessing Jet Burners, Proceedings of the Combustion Institute, 28, 2000, 481487. [8] Qamar N., Nathan G. J., Alwahabi, Z. T. and King, K. D. The Effect of Global Mixing on Soot Volume Fraction: Measurements in Simple Jet, Proceedings of the Combustion Institute, 30, 2005, 1493-1500. [9] Turns, S. R., An Introduction to combustion: concepts and Applications, Second Edition Mc Graw-Hill Singapore, 2000
Figure 9. Precessing jet temperature PDF along centreline.
92
