Diluted combustion in a aerodynamically staged swirled burner

fueled by diesel oil

M. Torresi1, S.M. Camporeale1, B. Fortunato1, S. Ranaldo1, M. Mincuzzi1,
A. Saponaro2
1. DIMeG -Dipartimento di Ingegneria Meccanica e Gestionale – Politecnico di Bari
2. CCA – Combustion & Environment Research Centre – Ansaldo Caldaie

1. Abstract
Among new combustion technologies, MILD (Moderate and Intense Low Oxygen Diluted)
combustion is particularly attractive since it is characterized by: flame temperature peaks
lower than registered in conventional systems, reducing NOx formation; a stable oxidation
without a flame front even with lean fuels, limiting vibrations and noise due to flame front; an
increase in combustion efficiency, and a lower sensitivity to fuel characteristics. Such a kind
of combustion mode can be obtained by means of the dilution of reactants with inert gas and
the increase of their temperature. For this reason, an aerodynamically staged swirled burner
fuelled by diesel oil has been experimentally tested and numerically simulated under diluted
and highly preheated inlet flow conditions. The staging is obtained through a double coaxial
air inlet with the same swirl orientation. The diesel oil is injected through a central atomizing
nozzle characterized by a very high rangeability. The numerical results, in agreement with
experiments, confirm that the burner, under properly operating conditions, is able to
completely burn the fuel without a flame front and with a very uniform temperature field.

Keyword: MILD combustion, CFD, aerodynamically staged combustion, diesel oil.

2. Introduction
Diluted combustion is nowadays a well established combustion technology. Actually, a subset
of this technology, the MILD combustion, is holding more and more the attention of the
scientific community [1]. According to Cavaliere and de Joannon, it is possible to talk about
MILD combustion when the oxidant inlet temperature Tin and the increase of temperature
inside the combustion chamber ∆T are respectively higher than and lower than the self
ignition temperature of the fuel Tign [2]. In the MILD combustion regime, the high presence of
diluents, which don’t directly participate to the reaction mechanism, moves the mixture
composition outside the flammability limits so that a deflagrative flame propagation is not
allowed. To make the fuel oxidation possible, preheating of the oxidant beyond the fuel self
ignition temperature is needed. A great interaction between the turbulent mixing and the
chemical kinetic (approximately unitary Damköhler number) has been recognized to be the
principal characteristic in this kind of combustion [3]. Particularly, thermochemical
fluctuations are responsible of delocalizing the reaction, distributing the heat release. Radiant
heat transfer plays an important role too, since the recirculating exhaust gases optically absorb
a wide amount of energy from the reaction zone, due to the high concentrations of H2O and
CO2, thus contributing to a redistribution of temperatures. In the MILD combustion regime, in
the proximity of the reactant inlets, a high level of uniformity is reached, which is extended to
the entire volume of reaction, with reduced temperature and species concentration gradients

ISBN 978–88–88104–11-9 / doi : 10.4405/ptse2010.I11

1
Processes and Technologies for a Sustainable Energy

[4]. In this way, a combustion characterized by no visible flame (flameless oxidation), a

reduction of pollutant emissions and a reduction of acoustic energy release (noiseless
combustion) are favoured, as shown by Wünning and Wünning [5]. They have investigated
the fundamentals of flameless oxidation (FLOX), defining a schematic diagram of the
stability limits. Most of the initial research activities have been focused on gaseous fuels, such
as methane, propane, LPG, process gases [6], [7], [8]. The diluted combustion technology has
been mainly applied in furnaces and boilers for industrial applications, especially in the
metallurgy field [9], [10]. This R&D effort allowed to achieve significant results especially in
terms of energy saving (up to about 60%), pollutant reduction (including NOx and greenhouse
gases, such as CO2), and downsizing of the equipments. In principle, diluted combustion can
be applied in almost all kinds of combustion systems. These results have encouraged the
scientific community to conduct R&D on diluted combustion and currently many projects are
already in progress.
In the present paper the attention is devoted to the MILD combustion of automotive Diesel
oil. The numerical simulations were performed both in flame and flameless configuration and
they correspond to two experimental tests carried out at the combustion laboratory placed at
the Department of Mechanical and Management Engineering (DIMeG) of the Politecnico di
Bari University. Thanks to the use of high temperature diluted air and above all to the
aerodynamic characteristic of the burner, stable MILD combustion regime was reached even
though the temperature of combustion chamber walls could be considered low (773 K).

3. Plant description
The experimental test rig (Fig. 1) was designed in order to obtain dilution and temperature
condition of the oxidant suited for MILD combustion. The main parts of the plant are: a
Forced Draft fan (FD fan) and an Induced Draft fan (ID fan), which allow the control both of
the flow rate and the pressure inside the combustion chamber; the flow rates are measured by
means of two turbine volumetric flowmeters type SM-RI-X-G 400 ITI-Instromet. The
flowmeter is characterized by a rangeability of 32-650m3/h and incorporates a high frequency
blade passage counter, which at the highest flow rate returns a frequency value equal to 160
Hz. The output signal is 1…3 mA according to the NAMUR standard, specifically used for
intrinsically safe applications. In order to acquire this signal on a Field Point unit, an opto-
isolated proximity signal converter with an open collector output is used.
The test rig has an auxiliary burner parallel to the air piping, diluting the oxidant; a mixing
system; a cooling system based on a double-tube water heat-exchanger; a fully instrumented
experimental 100 kW combustor; a quench cooler with atomized water injection for the
exhaust gas cooling in order to protect the ID fan. The exhaust gases coming from the
auxiliary furnace are mixed with fresh air in order to obtain a diluted oxidant for the MILD
combustion; whereas, the auxiliary furnace is excluded by the experimental plant circuit when
flame combustion is desired. The rig is remotely controlled by PC. It is possible to control
flow rates, oxygen concentration, fuel-air ratio and temperature at the combustor inlet. The
control system is programmed in LabVIEW environment. In order to follow the combustion
process, two Bosch LSU 4.2 wideband oxygen sensors, that uses a flat ceramic zirconia
element, in conjunction with their AWS2 Signal Conditioning Units, are installed upstream
and downstream the experimental combustor. They provide a precise indication of the exact
air/fuel ratio over a broad range of mixtures. The Bosch LSU 4.2 wideband oxygen sensor has
a response time of less than 100 milliseconds to changes in the air/fuel mixture, and reaches
operating temperature of 700°C to 800°C within 20 seconds or less using its internal heater.

2
 Ischia, June, 27-30 - 2010

To feed the two burners connected to the two furnaces an ad hoc diesel oil circuit was
designed in compliance with regulations (Directive 97/23/EC - PED 94/9/EC - ATEX.).
The atomizing nozzle, for the experimental burner under investigation, is the Variflo P/N
33769-9 by Delavan, that provides good atomization over a wide range of flow rates and wide
spray angles even at low flow rate. This type of nozzle is based on the principle of the bypass.
Hence, in order to evaluate the fuel mass flow rate, two flow meters of the VS1-GPO12V-
12D11 / 3 12VDC type by VSE are installed both on the supply line and on the bypass line.
These are volume rate measuring sensors based on the meshing gear principle and are
designed for use with liquids. Two precisely matched gear wheels are enclosed in a very
accurately machined housing. Gear rotation is sensed by a non-contacting signal pick-up
system. They are characterized by a measuring range between 0.002 l/min and 2 l/min, with
an accuracy of ± 0.3% repeatability and ± 0.05% of measured value.

Fig. 1 Experimental plant.

3.1 Burner and Combustion Chamber

The aerodynamically staged swirled burner is represented in Fig. 2.

Fig. 2 Burner head details in flame (on the left) and flameless (on the right) configuration.

3
Processes and Technologies for a Sustainable Energy

There are two swirled oxidant flows (primary and secondary air) entering the combustion
chamber. Their mass flow rates are controlled by the relative positions that the fuel injection
gun and the primary air pipe can assume. In particular, MILD combustion configuration is
characterized by a maximum secondary flow crossing section, because the immediate contact
between fuel and oxidant is not desired since it would originate a flame front. In Fig. 2 on the
left and on the right the flame and flameless combustion configurations are respectively
represented. The combustion chamber is cylindrical with a 0.6 m diameter and 1.2 m length.

4. Model description
In order to numerical predict the burner performance in both flame and flameless combustion
conditions, the 3D thermo-fluid dynamic simulation have been carried out by means of a
commercial CFD code (FLUENT). The computational domain (Fig. 3) has been discretized
by means of a 3D hybrid multiblock grid (unstructured mesh are used where the geometries of
the elements are more complex). Taking into account the geometrical symmetry of the
domain, under the hypothesis of thermo-fluid dynamic symmetry, only an half of the entire
domain has been considered. In order to accurately simulate the swirl induced on the primary
and secondary air flows, the computational domain starts from the wind box inlet and
includes the entire region across the burner and the combustion chamber.

Fig. 3 Computational domain.

The two grids for flame and flameless combustion are very fine (2,555,852 and 2,298,570
cells respectively) and differ only for the relative positions of the fuel injection gun and of the
primary air pipe, as in the corresponding experimental configurations.
The steady incompressible three-dimensional RANS equations are discretized by means of a
finite volume approach. Additional transport equations (e.g. mixture fraction and its variance,
radiation intensity, etc.) are also included in the model. In this specific case the equations are
solved one at time using the segregated approach. The pressure velocity coupling is achieved
by means of the SIMPLE algorithm (Semi-Implicit Method for Pressure-Linked Equations).
The convection terms are discretized by means of a second order upwind scheme, while the
pressure and viscous terms are discretized by a centred second order scheme. The Boussinesq
approach is applied to relate the Reynolds stresses to the mean velocity gradient. Closure for
turbulence effects in the RANS equations has been achieved by means of the standard k-ε
model (Launder & Spalding,1972). The link between the solution variables at the near-wall
cells and the corresponding quantities on the wall has been performed by means of standard
wall functions (Launder & Spalding,1974). The P-1 radiation model (used in this work) is
based on the expansion of the radiation intensity I into an orthogonal series of spherical

4
 Ischia, June, 27-30 - 2010

harmonics. The combustion chemistry is modelled using a non-premixed combustion

approach. This involves the solution of transport equations for the mixture fraction and its
variance, whereas equations for individual species are not solved. Species concentrations are
then derived from the predicted mixture fraction field. The thermochemistry calculations are
preprocessed and then tabulated for look-up. Interaction of turbulence and chemistry is
accounted for with an assumed-shape Probability Density Function (PDF). During the present
work, the Diesel oil was assimilated to C16H29 and other 29 chemical species (C, H, N2, O2,
C<s>, H2, CH4, CO, H2O, CO2, OH, O, C2H2, CH3, CN, C2N2, HO2, C2H4, N, C2H6, C4H2,
C2N, C3H2, HNO, H2CCCH, C2H, CH2, H2CCCCH, C2H3) have been considered.
In order to consider the liquid fuel injection, the Discrete Phase Model (DPM) implemented
in the FLUENT code is used. The DPM performs Lagrangian trajectory calculations, as well
as heat and mass transfer, for the dispersed phase (particles, droplets, or bubbles). The code
can also include the interaction of the discrete phase with the continuous phase by means of
source terms added to the different transport equations. In the case of fuel droplets, the code
considers both the particle heating and its evaporation.
In order to simulate the fuel injection a pressure-swirl atomizer model called Linearized
Instability Sheet Atomization (LISA), which determines an hollow cone distribution of the
spray, is used [11].

5. Results and Discussion

In Tab. 1 the main inlet boundary conditions for the two simulations are reported [12]. In the
MILD combustion configuration, the oxidant is not fresh air but a mixture of fresh air diluted
by the exhaust gases coming from the auxiliary furnace. The oxidant temperature has been
experimentally measured, as well as the combustion chamber wall temperature, whereas the
chemical composition has been analytically computed.
In order to perform the numerical simulations, the following procedure has been adopted: at
first, the stabilization of the “cold” flow (i.e. the flow field of the oxidant, without the fuel
injection) has been reached; and then the fuel injection and the combustion simulation have
been performed. Furthermore, in the MILD combustion simulation, prior to the introduction
of the diluted oxidant, a traditional flame configuration with fresh air has been reached.

Tab. 1 Inlet boundary conditions.

Combustion regime a) Flame combustion b) MILD combustion
Fuel Diesel oil (C16H29) Diesel oil (C16H29)
12.59% O2 ,74.48% N2 ,
Oxidant fresh air
9.43% CO2 , 3.50% H2O
Oxidant temperature [K] 302 673
Oxidant flow rate [kg/s] 0.04 0.06746
Fuel flow rate [kg/s] 0.00228 0.00244
Fuel inlet temperature [K] 300 300
Combustion chamber
[K] 650 773
wall temperature
Turbulence intensity [-] 5% 5%
Turbulence length scale [m] 0.005 0.005

Fig. 4 represent the structures of the flow fields by means of the axial velocity contours in the
flame and flameless conditions respectively. In the diluted configuration, due to the higher
inlet temperatures and the higher mass flow rate, the volume flow rate is larger than in

5
Processes and Technologies for a Sustainable Energy

traditional combustion, so that velocities have higher picks especially in the restricted
sections. In both simulations two main recirculation zones can be identified: an internal
recirculation zone closed to the burner head and an annular recirculation zone near the side
walls of the combustion chamber. In the case of MILD combustion, the recirculation zone is
more intense, hence further diluting the evaporating fuel and the oxidant and improving the
delocalization of the combustion reactions.

Fig. 4 Contousr of axial velocity for flame (a) and flameless (b) conditions.

Fig. 5 shows temperature fields for both flame and flameless combustions. The flameless
combustion regime presents very smooth temperature distribution (Tmax = 1400K) compared
to the flame regime, in which high temperature peaks are evident (Tmax = 1800K). Due to the
“mixed is burned” assumption in both simulations the combustion begins near the burner
head; however, its intensity is much lower in the case of flameless combustion with respect to
flame combustion. In the same figure, images captured during the two experimental tests are
reported. Since no measurements in flame have been carried out, only a qualitative
comparison can be performed. A good agreement in the shape of the flame front can be
recognized. Furthermore, in the MILD combustion configuration, no flame front is visible.
Actually, in order to take the picture under flameless condition, the diaphragm needed to be
opened to its maximum. The MILD temperatures make reasonably think to a lower NOx
production, because the chemical reactions of main NOx production mechanism (thermal
NOx) are inhibited under temperatures of 1800 K.
The contours of CO mass fraction (Fig. 6) show that in both cases the combustion process
could be considered complete. When the flame condition is considered, a peak of CO mass
fraction is evident in the core of the flame, whereas in the flameless condition, due to dilution
and reaction delocalization, CO concentrations are smaller.

6
 Ischia, June, 27-30 - 2010

Fig. 5 Contours of temperature for flame (a) and flameless (b) conditions with the relative ex-
perimental images.

Fig. 6 Contours of CO mass fraction for flame (a) and flameless (b) conditions.

6. Conclusions
The present work has been devoted to the numerical simulations of both flame and flameless
regimes obtained in the experimental facility available at DIMeG combustion laboratory.
During the preliminary experimental activity only global parameters have been registered.
This was the first time that in the DIMeG combustion Lab a stable flameless combustion was
obtained. An important characteristic of these tests is the use of a liquid fuel.
Under a qualitatively point of view, the results of CFD combustion analysis can be considered
compatible with the experimental results.
Of particular interest are the temperature fields, which clearly confirm the different thermo
fluid dynamic behaviour when the burner is operated in either flame and flameless conditions.

7
Processes and Technologies for a Sustainable Energy

In both cases, the absence of unburned gases at the combustion chamber exit confirm that the
fuel oxidation is complete, demonstrating the validity of the proposed burner design.

References
1. De Joannon M., Langella G., Beretta F., Cavaliere A., Noviello C., "Mild Combustion:
Process Features and Technological Constrains Combustion Science and Technology",
Vol.153, (2000), pp.33-50.
2. Cavaliere A., de Joannon M., Ragucci R., “Physical and chemical aspects of the MILD
combustion. Examples related to gas turbine combustion”, Proc. of the 4th Int. Symposium
on HTACG, Rome, Nov 26-30 2001.
3. Plessing T., Peters N., Wünning J.G., “Laser optical investigation of highly preheated
combustion with strong exhaust gas recirculation”, Proc. of the 27th Int. Symposium on
combustion, the combustion institute, Pittsburgh, 1998, pp 3197-3204.
4. Amoresano A., de Joannon M., “Oxydation regimes in MILD combustion of low and high
molecular weight paraffins”, 59° Congresso Annuale ATI, Genova 2004.
5. Wünning J.A., Wünning J.G., “Flameless Oxidation to reduce thermal NO-formation”,
Progress in Energy and Combustion Science, Vol.23, pp. 81-94, 1997.
6. Dally B.B., Riesmeier E., Peters N., "Effect of fuel mixture on moderate and intense low
oxygen dilution combustion", Combustion and Flame, Vol. 137, (2004), pp 418–431.
7. Özdemir I.B., Peters N., “Characteristics of the reaction zone in a combustor operating at
mild combustion”, Experiments in Fluidsm Vol. 30, (2001) pp.683-695.
8. Murer S., Pesenti B., Lybaert P., "Characterization of flameless combustion of natural gas
in a laboratory scale furnace", ECM (European Combustion meeting) 2005, Louvain-la-
Neuve (Belgium), Apr 2005.
9. Aguilé F., "Overview of the Gaz the France R&D activities on flameless oxidation applied
to high temperature processes", 23rd World Gas Conference, Amsterdam, 2006.
10. Touzet A., Lhomme P.J., Quinqueneau A., “New Efficient technologies with very low
NOx emissions available for the industry: two recent examples in the French metallurgy
field”, Proc. of the 4th International Symposium on HTACG (High Temperature Air Com-
bustion and Gassification), Rome, Nov 26-30 2001.
11. ANSYS FLUENT “User’s guide”.
12. Mincuzzi M., “Analisi fluidodinamica di un bruciatore industriale alimentato con
combustibili liquidi”, Tesi di Laurea anno accademico 2008-2009, Politecnico di Bari.