Detailed Measurements in An Idealized and A Practi
Detailed Measurements in An Idealized and A Practi
Detailed Measurements in An Idealized and A Practi
net/publication/242196880
Article
CITATIONS READS
0 98
12 authors, including:
Howard Levinsky
University of Groningen
95 PUBLICATIONS 1,469 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Graham Hargrave on 02 March 2014.
ABSTRACT
The results of an experimental study of the combustion processes in an idealized boiler are presented. This
experimental study was part of a project supported by the European Community (Brite-Euram III program)
which aimed at developing tools for the design of clean and efficient household boilers. The idealized boiler
was designed with an optically accessible combustion chamber and independent controls of power input,
primary and secondary aerations. Combustion processes were investigated by combined measurements of
temperature, velocity, mean species concentrations and radical concentrations using probe sampling and
non-intrusive optical techniques. Measurements were obtained for five different flame cases. Results
focused on understanding the pollutant formation mechanisms in these five flames. The work reported here
represents one of the first comprehensive studies of combustion processes in a practical boiler.
RESUME
Cette publication présente les résultats de l’étude expérimentale des processus de combustion dans une
chaudière domestique modèle. Cette étude expérimentale représente l’une des tâches d’un projet co-
financé par la Communauté Européenne (programme Brite-Euram III) dont l’objectif est de développer
des outils d’aide à la conception de chaudières domestiques à haut rendement et faible émission de
polluants. La chaudière modèle a été conçue avec une chambre de combustion optiquement accessible
et un contrôle indépendant de la puissance et des taux d’aération primaires et secondaires. Les
processus de combustion ont été étudiés par mesures combinées de température, de concentration
d’espèces chimiques stables et radicalaires, et de vitesses grace à des sondes de prélèvement et par des
techniques optiques. Ces mesures ont été effectuées sur 5 cas-tests différents. Les résultats portent
sur la compréhension des mécanismes de formation de polluants dans ces 5 cas-tests. Ces travaux
représentent l’une des premières études des processus de combustion dans une chaudière domestique.
Laminar premixed combustion has been widely studied by others as a basic combustion
research problem (1-4). Nevertheless, up to now, few studies were devoted to laminar partially
premixed flames with natural gas and operating conditions pertaining to practical devices (5,6). This
lack of basic knowledge is restraining the understanding of stabilization and pollutant emission
mechanisms in household boilers. This basic knowledge is required today to meet the increasingly
stringent regulations on NOx emissions limits. For instance, although NOx emissions for most of the
natural gas powered household boilers sold in Europe are currently around 120 mg/kWh (7), the
European standards are moving towards lower limits, i.e., 50 mg/kWh.
Over the years, burner manufacturers have developed rules of thumb for choosing
combination of design parameters, such as primary fuel-air ratio, port loading, and port spacing, to
meet the approved specifications. However, this approach is a global one, where all measurements
are made at the input and output of the appliance. This often results in unwanted effects such as the
increase in CO emissions or decrease in flame stability which occur as NOx emissions are reduced.
Since the formation of pollutants in combustion systems depends upon the details of the combustion
process itself, insights into these details are required for proper design of the appliance.
In a project supported by the European Community in the framework of the Brite-Euram III
program, detailed measurements in the combustion chamber of an idealized household boiler were
obtained. The objective was to understand, using advanced measurement techniques, how burner
operation affects combustion performance and how this understanding can be used for improving
boilers and burners. It was also to identify and develop measurement tools most adapted to boiler
design aids. To that effect probe measurements were compared with measurements obtained using
non-intrusive techniques. Finally, these detailed measurements were also used to provide an
accurate and reliable database for validating numerical models also developed in the framework of
the project.
The idealized boiler was developed with the same characteristics as that of a commercial
boiler but with a simplified 2-dimensional geometry. Global measurements in the exhaust of the
boiler were obtained and compared with measurements performed in a commercial boiler. Five
different flame regimes were also selected. For the five regimes temperature, velocity, chemical
species concentrations and flame emission measurements were obtained in the combustion chamber
of the idealized boiler. Preliminary results are presented.
EXPERIMENTAL SETUP
meters.
The combustion chamber (150 x Figure 1. Schematic of the idealized boiler.
115 x 150 mm) is protected from
entrainment of surrounding air by 4 glass (Vycor) windows, held in place by 4 stainless steel
mounts. A heat exchanger made of copper fins is placed on the top edges of the windows. This heat
exchanger was designed to support a maximum load of 11 kW (Gross Calorific Value) with
minimum pressure drop and to have an efficiency similar to that of commercial boilers. It was
operated with an inlet water temperature of 60°C and an outlet water temperature of 65°C for all
flame conditions investigated. A chimney is placed on top of the heat exchanger for flue gases
exhaust. The composition of the flue gases was measured with a probe consisting of a copper tube
located at the top of the chimney.
The combustion chamber is characterized by the X, Y, and Z axis (see Figure 1). The X axis
is the horizontal axis which runs across the burner slots. The location at X = 0 mm corresponds to
the center of the central slot. The slots being 16 mm apart, their centers are located at X = -48, -32,
-16, 0, +16, +32, and +48 mm. The Y axis is the vertical axis (Y = 0 mm corresponds to the surface
of the burner slot, and Y = 150 mm corresponds to the bottom of the heat exchanger). Last, the Z
axis is the horizontal axis that runs along the burner slots (Z = 0 mm corresponds to the center of the
slot).
In order to study the combustion mechanisms in the idealized boiler, two levels of
measurements were performed: global measurements in the flue gases and detailed measurements
in the combustion chamber. The overall measurements consisted of varying the primary aeration
ratio, the secondary aeration ratio, and the power input independently, and measuring the NOx and
CO emissions at the flue gas probe. For the detailed measurements in the combustion chamber, five
operating conditions referred to as « test cases » were defined. They are shown in Table 1. In all
cases selected, the total aeration ratio was set to 1.4 and the power input was set to 10 kW. Case 1
and 2 have the same primary and secondary aeration ratio but a different fuel composition: natural
gas like blend vs. pure methane, respectively. Case 3 uses pure methane with no secondary air (i.e.,
Table 1. Test cases selected as operating conditions for the idealized boiler.
Test Case Fuel Composition Fuel GCV Flame Primary Second.
3
(kWh/m (n)) Type Aeration Aeration
CH4 C2H6 C3H8
1 0.9 0.06 0.04 11.04 Bunsen 0.6 0.8
2 1 0 0 9.87 Bunsen 0.6 0.8
3 1 0 0 9.87 Bunsen 1.4 0
4 1 0 0 9.87 Bunsen 1.2 0.2
5 1 0 0 9.87 V shaped 1.2 0.2
Commercial Boiler
The combustion performances of the idealized boiler were compared with those of a
modified commercial boiler. It was a floor standing boiler with a cast iron combustion chamber,
designed for a nominal thermal input of 8.8 kW. Water was produced with a temperature of 70°C.
This boiler had partially premixed bar burners with atmospheric injection and free entrainment of
secondary air, thus creating 3-dimensional flames. This contrasts with the idealized boiler where
the primary premixing is controllable and the burner slots give 2-dimensional flames. The primary
aeration ratio was varied between 1.12 and 1.45, giving a total aeration ratio between 1.45 and 1.54.
The combustion performance of the boiler were investigated in terms of flue gas temperature, total
air factor and NOx, and CO emissions.
Measurement Techniques
Direct visualization of the total UV/visible emission and OH, CH and C2 emissions was
achieved using an ICCD camera (Princeton Instruments) with a spatial resolution of 0.2 mm per
pixel, and a camera lens (UV-Nikkor, 105 mm, f 4.5). Interference filters were used in order to
select the emission wavelength from different radicals. The camera was connected to a controller
(ST138, Princeton Instruments) for image acquisition and a pulse generator (DG535, Stanford
Research) to synchronize the shutter and the exposure time. The camera was placed along the Z
direction (see Figure 1). The exposure time varied for each flame condition and for each radical
investigated. The digitized images were stored on a PC for image processing and analysis.
Fast Fourier Transform Infrared spectroscopy (FTIR) measurements of CO2, H2O, CO and
NO and Laser Induced Fluorescence (LIF) spectroscopy measurements of OH, CH, and CN were
performed in the idealized boiler for the five selected test cases. Due to space limitations, the results
of these measurements will not be presented in this paper but will be discussed during the poster
session of the present symposium.
Mean species (i.e., CO, CO2, O2, CH4, NO and NOx) concentrations were also measured in
the combustion chamber of the idealized boiler using Sonic Nozzle Probe Sampling (SNPS). The
SNPS (Pyro-Controle Chauvin Arnoux, licensed by Gaz de France) uses a suction probe with an
outer diameter of 4 mm. A water cooled sonic nozzle is located at the probe tip. This nozzle is
designed to quench chemical reactions in the sampled gases by fast reduction of temperature (down
to about 60°C) and pressure (down to about 100 mbar of absolute pressure). The sampled gases are
channeled through a heated tube maintained at 60°C, to a permeation membrane for water removal.
These gases are then sent to a series of analyzers. The volumetric concentrations of CO2 and CO are
measured by infrared spectroscopy Ultramat 22P, (Siemens), CH4 by infrared spectroscopy (Uras
4, Hartmann & Braun), O2 by paramagnetism (Oxymat 5E, Siemens), and NO and NO2 by
Overall Measurements
Overall measurements were performed in the flue gases of the idealized boiler as a function
of primary and secondary aeration ratio, and power input. The objective of these measurements was
to determine the performances of the idealized boiler in terms of NOx and CO emissions and to
compare its behavior with that of commercially available boilers.
0.3
100
300
0.5
250 80
NOx (mg/kWh)
CO (mg/kWh)
200
60
1.0
0.3
150
40 0.5 Secondary
aeration ratio =
100
1.0
Secondary
20 0.0
50 aeration ratio = 0.0
0 0
0.6 0.8 1 1.2 1.4 1.6 1 1.2 1.4 1.6 1.8 2
Primary aeration ratio Total aeration ratio
Figure 2 shows the NOx concentration measured by probe sampling in the flue gases of the
idealized boiler as a function of primary aeration ratio for different secondary aeration ratios. The
concentrations were obtained when the burner was operated at 10 kW. As one can see, the NOx
concentration decreases sharply when the primary aeration ratio increases. This decrease is
somewhat less sharp as the secondary aeration ratio decreases. Note that very low NOx emissions,
below 50 mg/kWh) are observed with no secondary aeration, i.e., for fully premixed flames. Lean
premixed flames are known to favor NOx reduction. In these flames, the concentration of C2, CH
and CH2 radicals are reduced, hence the attack of N2 molecules by these radicals in the oxidation
zones which is the first step for prompt NO formation (Fenimore pathway) is limited. In addition,
lean premixed flames have lower temperatures, also resulting in smaller NO formation through the
thermal channel (Zeldovich pathway). Figure 3 shows CO concentrations measured in the idealized
boiler as a function of total aeration ratio for different secondary aeration ratios. Total aeration was
selected as the X axis in this case, to stress the weak effect of secondary aeration. As one can see
in Figure 3, CO emissions decrease with total aeration ratio. This decrease is almost independent of
secondary aeration ratio. This shows that in the combustion chamber with enough residence time
for burnout between the burner and the heat exchanger, the oxidation of CO is governed by the total
availability of molecular oxygen, irrespective of whether it is provided in the primary or secondary
air.
Measurements of NOx and CO as a function of primary aeration at constant power was also
obtained using a commercially available floor standing boiler. These measurements are reported in
Figure 4. The strong decrease in NOx emission with increasing primary aeration ratio as observed
in the idealized boiler is evident. However, in contrast with the idealized boiler, the CO emission
in the practical boiler increases sharply with primary aeration. This difference in behavior is
traceable to the fact that whereas the flame in the idealized boiler blows off suddenly at a given
primary aeration ratio, practical burners begin to lift long before finally blowing off. These lifted
flames leak a fuel-air mixture at the edges of the burner, that is then converted to CO en route to the
heat exchanger (5). These effects of primary aeration are responsible for the existence of only a
narrow window in terms of operating parameters for acceptable NOx and CO emissions.
Temperature Measurements
The temperatures were measured in the combustion chamber by FWCT and by CARS.
Figure 6 compares the temperature profiles measured by both techniques along the centerline of the
central burner slot for Case 3. As one can see, the two profiles are in good agreement. The
agreement is also good for Case 4 and 5, but only fair for Case 1 and 2. Since space limitations
preclude a full comparison of the results obtained with both techniques, the following discussion will
focus on FWCT measurements.
The temperature maps measured by FWCT are shown below Y = 40 mm in Figure 7 for the
five cases investigated. As a reminder, the FWCT measurements were obtained by introducing a
thermocouple on the negative side of the X axis through a slot machined in the glass windows. Since
this slot caused some flow disturbances up to about X = 0 mm, the measurements shown were
performed above the slot located at X = +16 mm.
2
2
Temperature (K)
14
Temperature (K)
14 O2 O2
12 1600 12 1600
10 NO x10
3 10 3
NO x10
CO2 CO2
8 1200 8 1200
6 6
3 3
4 NO 2 x10 800 4 NO2 x10 800
2 2
CO CH4 CO
CH4 0 400
0 400
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Y (mm) Y (mm)
Figure 8. Axial profiles for flame case 1. Figure 9. Axial profiles for flame case 2.
20 2400 20 2400
18 18 Temp.
16 Temp. 2000 16 2000
O2
2
2
Temperature (K)
Temperature (K)
14 14
O2
12 NO 2 x10
4 1600 12 1600
10 4 10
NO x10 CO 2
CO2
8 1200 8 1200
6 6
4 800 4 CH4 3 800
NO x10
2 CH4 2 3
CO CO NO2 x10
0 400 0 400
0 10 20 30 40 50 60 0 10 20 30 40 50 60
Y (mm) Y (mm)
Figure 10. Axial profiles for flame case 3. Figure 11. Axial profiles for flame case 4.
20 2400
18
Temp.
16 2000
Mole Fraction (x10 )
2
14 Temperature (K)
O2
12 1600
10
CO 2
8 1200
6
CH4
4 3 800
NO x10
2 3
NO 2 x10
CO
0 400
0 10 20 30 40 50 60
Y (mm)
The mean species concentration profiles are similar for Case 1 and 2 as expected from
temperature and emission spectroscopy measurements. In both cases, the maximum CO
concentration measured corresponds to the location where the temperature gradient is maximum,
which is assigned to the location of the inner premixed flame. The NO concentration profiles follow
the same trend and can be divided in three stages: a stage of sharp concentration increase is observed
from the burner mouth up to the location of the CO concentration peak (i.e., the location of the inner
flame); a stage of gradual concentration increase follows, up to the location where the temperature
is maximum (i.e., Y = 24 mm); finally above this location, the NO concentration remains almost
constant up to Y = +60 mm. At Y = +60 mm, the NO and NO2 concentrations are 90 and 30 ppm,
respectively, which corresponds to their respective concentrations in the flue gases. The stage of
sharp concentration increase can be assigned to NO formed through the prompt (Fenimore) pathway.
At the end of this stage, the NOx concentration (taken as [NO]+[NO2]) represents half of the total
NOx produced (i.e., the NOx concentration is 60 ppm at Y = 8 mm). This, however, does not take
into account the dilution with secondary air. The stage of gradual increase, which starts around
CONCLUSIONS
ACKNOWLEDGMENTS
The results of this project are published by kind permission of Companies, Universities and
Organizations which are involved in TOPDEC, and of the European Commission which financial
contribution to this project is gratefully acknowledge.
1. Nguyen, Q.V., Dibble, R., Carter, C.D., Fiechtner, G.J. and Barlow R.S. (1996). Raman-LIF
Measurements of Temperature, Major Species, OH, and NO in a Methane-Air Bunsen
Flame. Combustion and Flame, 105:499-510.
2. Van Oostendorp, D.L., Borghols, W. T. A., and Levinsky, H. B. (1991). The influence of
Ambient Air Entrainment on Partially Premixed Burner Flames: LIF Imaging of CO and OH.
Combustion and Flame, 79:195-206.
3. Arrigotti, S., Bernstein, S., Levinsky, H.B., Van Oostendorp, D.L., and van der Meij, C.E.
(1992). Numerical and Experimental Investigation of a Laminar Bunsen Flame. Proceedings
1992 International Gas Research Conference, Orlando, FL, 1753-1762.
4. Hasko, S.M., Fairweather, M., Bachman, J.S., Imbach, J., van der Meij, C.E., Mokhov, A.V.,
Jacobs, R.A.A.M. and Levinsky, H.B. (1995). Optical-Diagnostic and Numerical Study of
an Experimental Slot Burner. Proceedings 1995 International Gas Research Conference,
Orlando, FL, 33-42.
5. Van der Meij, C.E., Mokhov, A.V., Jacobs, R.A.A.M. and Levinsky, H.B. (1994). On the
Effect of Fuel Leakage on CO Production from Household Burners as Revealed by LIF and
CARS. Twenty-Fifth Symposium (International) on Combustion, The Combustion Institute
243-250.
6. Levinsky, H.B., van Oostendorp, D.L., van der Meij, C.E., Jacobs, R.A.A.M. and Mokhov,
A.V. (1992). Pollutant Formation in Household Burners: LIF Imaging of NO. Proceedings
1992 International Gas Research Conference, Orlando, FL, 1578-1586.
7. Joos, L. (1995). State of the Art in Pollution Control for Gas Appliances. Proceedings 6th
International Gas Congress (IRGUC), Prague, Czech Republic.
8. Neveu, F., Corbin, F., Perrin, M. and Trinité, M. (1995). Characterization of the couplings
of aerodynamics and combustion in a nonpremixed turbulent flame. Rev. Gén. Therm.,
34:305-314.
9. Shandross, R.A., Longwell, J.P. and Howard, J.B. (1991). Noncatalytic Thermocouple
Coating for Low-Pressure Flames. Combustion and Flame, 85:282-284.
10. Kent, J.H. (1970). A Noncatalytic Coating for Platinum-Rhodium Thermocouples.
Combustion and Flame, 14: 279-282.
11. Muzio, L.J. and Quartucy, G.C. (1997). Implementing NOx Control: Research to
Application. Prog. Energy Combust. Sci., 23:233-266.