|  41 
 
Numerical analysis of oil combustion  
in a small combustion device 
Numerična analiza oljnega gorilnika v mali 
kurilni napravi 
Matej Zadravec
1,
*, Boštjan Rajh
2
, Niko Samec
1
, Matjaž Hriberšek
1
 
1
 Faculty of Mechanical Engineering, University of Maribor, Smetanova 17,  
2000 Maribor, Slovenia 
2
 EnerB d.o.o., Hmeljarska ulica 7, 3310 Žalec, Slovenia 
E-Mails: matej.zadravec@um.si ; bostjan.rajh@enerbee.eu ; niko.samec@um.si ; 
matjaz.hribersek@um.si    
* Corresponding author  
Izvirni znanstveni članek 
TEHNIKA - 
mehanika,  
numerično modeliranje  
Datum prejema: 
17. marec 2014 
ANALI PAZU 
4/ 2014/ 1: 41-50 
www.anali-pazu.si 
 
Abstract: This paper focuses on the numerical analysis of the combustion of the liquid 
fuel Extra Light Fuel Oil (ELFO) in a small combustion device. The calculation was 
performed by ANSYS CFX computer code for experimental optimal air-fuel ratio (λ
opt
) and 
the middle power of the oil burner Weishaupt WL5/1-A (37 kW). The goal of the numerical 
analysis was to establish, whether the experimentally determined λopt ensured the conditions 
for a complete combustion in each point of the combustion chamber. The comparison of 
numerical results and measured values confirm the suitability of the chosen numerical model.  
Key words: oil burner, oil combustion, multicomponent reactive flow, computational 
fluid dynamics (CFD), model validation. 
Povzetek: Prispevek obravnava numerično analizo zgorevanja v oljnem gorilniku ekstra 
lahkega kurilnega olja (ELKO)  v mali kurilni napravi. Numerična analiza je bila izvedena v 
programskem paketu ANSYS CFX pri eksperimentalno določenem optimalnem razmerju 
zraka in olja (λ
opt
) in srednji moči oljnega gorilnika Weishaupt WL5/1-A (37 kW). Namen 
numerične analize je bil preveriti ali eksperimentalno določeno optimalno razmerje (λ
opt
) 
zagotavlja popolno zgorevanje v vsakem območju kotla male kurilne naprave. Ujemanje med 
eksperimentalnimi podatki in izbranimi parametri za numerično analizo kažejo dobro 
ujemanje. 
Ključne besede: oljni gorilnik, zgorevanje olja, večkomponentni reaktivni tok, 
računalniška dinamika fluidov (RDF), validacija modela. 

42  |  
ANALI PAZU, 4/ 2014/ 1, str. 41-50                                                                                                                  Matej ZADRAVEC et al. 
1. Introduction 
A detailed analysis of the combustion process under 
actual operating conditions of a specific combustion 
device requires considering of all diffusion and reaction 
processes of reacting flow at the same time. Next to the 
chemical kinetics, the impact of convection, diffusion 
and other physical processes have to be taken into 
account. Monitoring of all stated processes during the 
experiment is very difficult or even impossible. 
Therefore, Computational Fluid Dynamics (CFD) is 
being increasingly used in industry for the purposes of 
consideration and more detailed analysis of combustion 
processes as well as for optimization of combustion 
devices. The greatest issue in modeling the turbulent 
reacting flow is the fact that the majority of numerical 
models are based on empirical presuppositions. 
Therefore, an experiment has to be carried out in order 
to determine the missing parameters that are necessary 
for an appropriate numerical model. 
In the field of ELFO combustion, only a few 
literature sources [1] focusing on determination of wall 
heat fluxes in the large scale test combustors can be 
found. Some other sources [2] focus on the combustion 
of light residual fuel oils. Many authors have been 
studied CFD modeling of heavy fuel oil combustion [3-
5]. In general it is not possible to find any sources 
dealing with the CFD numerical model as a whole used 
for ELFO combustion in a small combustion device to 
analyse the achievement of complete combustion 
conditions in the whole space of combustion chamber. 
2. Combustion in a small combustion device 
Oil burners are appliances that heat the home using 
fuel oil. Oil is delivered to the home and pumped into a 
storage tank. Oil from the storage tank is drawn into the 
oil burner. The oil burner atomizes the oil (turns it into a 
mist of tiny oil droplets), mixes it with air and ignites it 
with a spark. The resulting flame shoots out of the blast 
tube and into the oil burner chamber (see fig. 1). 
Combustion gasses move through the heat exchanger 
where heat is transferred to the household air stream. 
The combustion gasses then make their way up the flue 
and out of the house. 
All practical combustion processes are turbulent in 
nature. Turbulent non-premixed flames are encountered 
in a large number of industrial systems for two main 
reasons. The first, compared to premixed flames, non-
premixed burners are simpler to design and to build 
because a perfect reactant mixing, in given proportions, 
is not required. Moreover, non-premixed flames are also 
safer to operate as they do not exhibit propagation 
speeds and cannot auto-ignite in undesired locations [6]. 
On the basis of experimental measurements of flue 
gas emissions, carried out by a flue gas analyzer for 
different air-fuel ratios (λ=1.15, λ=1.2 and λ=1.3), the 
optimal air-fuel ratio λopt = 1.2 was determined. It 
enables well operation of a combustion device under 
different conditions (permanent changes of air pressure, 
of air temperature and of pressure in the furnace as well 
as other weighting factors) within one service interval. 
In the scope of the experiment, specific data necessary 
for creation of the numerical model were obtained. This 
was performed by determining the λopt representing 
most optimal operating conditions, by creating the 
precise geometry of the furnace and the oil burner, and 
by collecting data concerning the oil nozzle. After that, 
we attempted to define an appropriate numerical model 
by applying the CFD. This model makes it possible to 
examine, whether the appropriate λopt was determined 
during the experiment, and if conditions for complete 
combustion are provided in every point of combustion 
chamber. 
        (a)                                                               (b)                   
Figure 1. (a) Experimental furnace Weishaupt. (b) Mixing device of the oil burner Weishaupt WL5/1-A. 

|  43 
NUMERICAL ANALYSIS OF OIL COMBUSTION IN A SMALL COMBUSTION DEVICE 
Liquid ELFO evaporation and combustion process 
are described by a numerous of evaporation and 
combustion models which are available to address many 
different applications. These applications are 
characterized by the type of mixing (e.g. non-premixed, 
premixed or partially premixed) and reaction chemistry 
(e.g. finite-rate chemistry or fast chemistry). The 
applied model contains the liquid Jet-A fuel C
12
H
23
 
(average composition of ELFO) as a pure substance 
with specified adequate substance properties (ρ = 840 
kg/m
3
, η = 0.00336 kg/ms, λ = 0.14 W/mK, latent heat = 
280 kJ/kg at 70 °C, surface tension coefficient of oil = 
0.026 N/m). The Liquid Jet-A Fuel C
12
H
23
 evaporates 
into its equivalent gas phase. 
Combustion primarily takes place in a diffusion-
limited mode (non-premixed) where fuel and oxidant 
are brought into contact via mixing and can then react. 
In these flames, it can generally be assumed that the 
turbulent mixing rate is much slower than the chemical 
kinetics rates (fast chemistry), and hence, will be the 
rate-limiting step. The current study uses the Eddy-
Dissipation Combustion (EDC) model which assumes 
that the reaction rate is controlled by the turbulent 
mixing rate. Hence, the turbulent mixing rate in 
conjunction with a global reaction mechanism is used to 
predict local temperatures and species profiles. This 
model solves the conservation equations describing 
convection, diffusion, and reaction sources for each 
component species. The gas phase is a part of 
changeable gas mixture that burns using a single-step 
reaction of the two-step reactions mechanism. At the 
first stage, the reaction specifies the gas phase and O
2
 as 
reactants, while CO and H
2
O are determined as 
combustion products (eq. 1). At the second stage, 
however, CO and O
2
 are specified as reactants, while 
CO
2
 is a combustion product (eq. 2):  
Additionally, the formation of NOX (thermal NO) 
was included into the model on the basis of Zeldovich 
mechanism (NO Formation with Temperature PDF). 
The selection of suitable reaction rates for chemical 
reactions is a very difficult task due to a shortage of data 
as well as ambiguity and inconsistency in scientific 
resources. After careful review the data of Westbrook 
and Dryer [7] have been used for C
10
H
22
 scaled by 
molar mass. The reaction rates for NO formation were 
taken from CH
4
 combustion mechanism. 
3. Numerical model 
3.1. Governing Equations 
The features of the turbulent flame depend mainly 
on the nature of the reacting flow. At sufficiently high 
Reynolds Numbers the flow becomes turbulent. The 
turbulent flow is unstable; it moves in irregular patterns 
and contains vortices of different sizes. These eddies 
move with different rotation and translation motion 
through the combustion chamber. The contour of the 
turbulent flow is not steady, but is changing all the time. 
Also unsteady is the flame front; its boarder surface is 
extremely curved. Exactly such type of flame is typical 
of ELFO combustion in a small combustion device. 
According to many authors [8-9], it is very difficult to 
establish a precise definition of the turbulent reacting 
flow. However, it is known that such flow has features 
like great instability, high value of Reynolds Number, 
3D fluctuation of eddies, their dissipation and turbulent 
diffusion that causes fast mixing of reactants, and 
increases the relation between momentum and the mass 
and heat transfer. An adequate description of the 
movement of incompressible viscous two-phase 
turbulent reacting flow is given by using the Reynolds-
averaged Navier-Stokes (RANS) equations for mass 
(eq. 3), momentum (eq. 4) and energy (eq. 5) 
conservation: 
,   ,    are the mean values of density, enthalpy 
and fluid velocity,     is the sum of all volume forces, p  
is the pressure,    is the combustion and    is the 
radiation source/sink term,     and    are the viscous 
stress tensor and the heat flux, respectively. 
By applying RANS equations, some additional 
terms have been produced in the momentum (eq. 4) and 
energy conservation equation (eq. 5), named Reynolds 
stresses            and fluxes       . These fluxes are 
modeled by the introduction of the turbulent viscosity 
η
t
, which can be determined by applying appropriate 
turbulent models to complete the system of RANS 
equations. The system of conservation RANS equations 
regarding the issue presented was solved by applying 
the two-equation RNG k-ε turbulent model developed 
using Re-Normalization Group (RNG) methods [10]. 
The RNG k–ε turbulence model is robust, economic and 
reasonably accurate over a wide range of turbulent flow. 
This model solves transport equations for kinetic energy 
(k) and its dissipation rate (ε). It assumes that the flow is 
fully turbulent and the effects of molecular viscosity are 
negligible, and is valid only for fully turbulent flows. 
The system of RANS equations, maintaining the 
mass, energy and momentum conservation equation, has 
 
(1) 
 
(2) 
                                                         (3) 
 
 
                                                                                                (4) 
 
 
                                                                                                (5) 
 

44  |  
ANALI PAZU, 4/ 2014/ 1, str. 41-50                                                                                                                  Matej ZADRAVEC et al. 
to be supplemented by the convection-diffusion 
equation (sometimes called transport equation) of mass 
species (ξ
n
) of the reacting flow component n (eq. 6), 
which has, due to Reynolds averaging, an additional 
term called the turbulent mass species flux: 
where      is turbulent Scmidt number and     is 
molecular diffusion coefficient of the component n. 
The source terms of the energy (    ) and convection-
diffusion equation (    ) are computed by the following 
two equations where      is computed by the turbulent 
combustion model: 
where           is the standard formation heat and      is the 
molecular mass of the component n. In the equations 
(eq. 7) and (eq. 8) the     stands for the formation or 
consumption rate of component n and is defined by:  
which is written for the general form of the chemical 
reaction: 
where    and    are the stoichiometric coefficients of 
component for the reactants and products, respectively. 
The chemical reaction rate      in eq. (eq. 9) is calculated 
using appropriate combustion models. 
In order to describe the oil droplet movement, the 
Lagrangian Particle Tracking Model was used, where it 
was necessary to consider that an oil droplet moves 
through a continuous component of the gas phase. 
Forces that act on every individual oil droplet, and 
influence its acceleration appear on account of the 
differences between the velocity of the oil droplet and 
the gas phase and also because of the gas phase 
displacement. We considered that every oil droplet is 
influenced by the buoyancy force F
B
 and the drag force 
F
D
. Other forces, however, were not included, since the 
density of oil droplets is much larger than that of the gas 
phase: 
 
 
 
At sufficiently small value of particle Reynolds 
numbers (the viscous regime), fluid particles behave in 
the same manner as solid spherical particles. In this case 
the drag coefficient is well approximated by the Schiller
-Neumann correlation which is defined as: 
At larger value of particle Reynolds numbers, the 
inertial or distorted particle regime and the surface 
tension effects become important. The fluid particles 
become, at first, nearly ellipsoidal in shape, and finally, 
spherical cap-shaped [11]. The drag coefficient of oil 
droplets was modeled using Ishii-Zuber Drag Model 
where the drag coefficient is approximately constant, 
independent of Reynolds number, but dependent on the 
particle shape through the dimensionless group known 
as the Eotvos number, which measures the ratio 
between gravitational and surface tension forces:  
where     stands for the density difference between the 
phases,   is the gravitational acceleration and   is the 
surface tension coefficient. 
The Ishii-Zuber correlation is given as [12]: 
The liquid evaporation model is a model for 
particles with heat transfer and one component of mass 
transfer, and in which the continuous gas phase is at a 
higher temperature than the particles. The model uses 
two-mass transfer correlations depending on whether 
the droplet is above or below the boiling point. This is 
determined through the Antoine equation: 
where          is the pressure scale used to scale the units 
for the vapor pressure, A is the Antoine Reference State 
Constant, B is the Antoine Enthalpic Coefficient and C 
is the Antoine Temperature Coefficient. 
The particle is boiling if the vapor pressure      is 
greater than the gaseous pressure. When the particle is 
above the boiling point, the mass transfer is determined 
by the convective heat transfer: 
                               
(6) 
 
 
(7) 
 
(8) 
 
(9) 
 
(10) 
 
(12) 
 
(13) 
(11) 
 
(14) 
 
(15) 
 
(16) 

|  45 
NUMERICAL ANALYSIS OF OIL COMBUSTION IN A SMALL COMBUSTION DEVICE 
When the particle is below the boiling point, the 
mass transfer is given by: 
where      and      are the molecular weights of the vapor 
and the mixture in the continuous phase, while    and              
are the molar fractions in the drop and gas phase. In 
either case, the rate of mass transfer is set to zero when 
all of the non-base substances in the particle evaporated. 
The EDC model has been applied based on an 
assumption that chemical reactions are faster in 
comparison to diffusion processes. Besides, the 
numerical model uses the EDC model for the 
combustion of the volatile gases in the gas phase. The 
rate of progress of elementary reaction n in the EDC 
model is determined by the two following expressions 
for reactant (eq. 18) and product limiter (eq. 19): 
 
 
 
where       is the molar concentration of component n 
and including only the components of the reactants, A is 
a constant and its value varies from case to case (in our 
case            ). 
 
 
 
where P includes all product components in the 
elementary reaction n (eq. 10) and B is additional model 
constant (in our case               ). 
Because the radiation is also a physical important 
transport phenomena during the combustion the P1 
radiation model has been chosen, since it proved as 
most appropriate [13-14]. The P1 radiation model is 
also a simplification of the Radiation Transport 
Equation which assumes that the radiation intensity is 
isotropic or direction independent at a given location in 
space. 
3.2. Geometry of the Model 
Before developing a numerical model, the geometry 
of the model (see fig. 2) has been created, including the 
combustion chamber in its original size, the oil burner 
torch together with the baffle plate for the combustion 
air supply, and the oil nozzle for the fuel supply. Special 
attention was paid to creating the oil burner torch, the 
baffle plate and the fuel supply, since the effect of the 
air flow swirling caused by the baffle plate of the oil 
burner has been considered. Specific details of the 
actual combustion chamber causing just a minor effect 
on the behaviors in the combustion chamber have been 
simplified and disregarded. 
Great attention was given to mesh selection. It is 
important to adjust and adapt the required size of mesh 
control volumes according to the size of the oil nozzle 
and of the combustion chamber. In the expected flame 
front area, the mesh had to be appropriately discretized 
for sufficient inclusion of all complex special processes 
of the turbulent combustion. In the opposite case, the 
results would not have been precise enough, providing a 
distorted picture of actual behaviors. Selection of the 
mesh has a direct influence on a number of control 
volumes, while a higher number of control volumes 
prolongs the calculation time. From the aspect of 
calculation time, it is therefore reasonable to choose the 
number of volumes that is as small as possible and that 
still ensures qualitative results at appropriate calculation 
times. However, it is also necessary to consider the size 
and the form of the control volumes at critical spots that 
have to be appropriately discretized. It can easily 
happen that an inappropriate form and size of the mesh 
at these spots have an impact on the convergence of 
numerical results. Considering mentioned facts 705 881 
control volumes have been applied.  
 
(17) 
 
(18) 
 
(19) 
Figure 2. Discretized geometry and detailed view illustrating the blast tube and the torch of the oil burner.  

46  |  
ANALI PAZU, 4/ 2014/ 1, str. 41-50                                                                                                                  Matej ZADRAVEC et al. 
3.3. Boundary Conditions 
The calculation was based on the transient analysis 
with the time limit of 1 second, during which the 
periodical flow field should set up. In this study the 
High Resolution was used for the advection scheme 
which is more accurate than the Upwind Scheme [11] 
and First Order for turbulence because turbulence is a 
fairly diffusive process. The calculation of conservation 
equations was performed in time steps of 0.004 seconds. 
During each time step it was, however, necessary to 
specify eight additional non-linear iterations to ensure 
the conditions of complete convergence. In our case, the 
convergence criterion 10
-4
 was selected, which indicates 
that the calculation accuracy is satisfactory. In the 
calculation domain, two points were also selected, at 
which the velocity value was examined during the 
calculation. The velocity oscillation at these two points 
has stabilized after a certain time or, as the case may be, 
the oscillation has become periodical having the 
minimal amplitude. This indicates steady-state solution 
of the flow field at specific points. 
Two-phase system of vaporization-combustion of 
the oil with air has been used. The required boundary 
conditions were determined at the following points: the 
INLET (mass flow of fuel, mass flow of air), the 
OUTLET (pressure), and the SURFACE OF THE 
BOILER JACKET (determined boundary condition on 
the wall – heat transfer (deducted heat flux), no-slip 
boundary condition – v=0, aggregation oil droplets on 
the wall). It is very difficult to determine the intensity of 
the turbulent movement of the air passing through the 
baffle plate into the combustion chamber. Therefore the 
medium intensity (5%) has been chosen. Specific 
boundary conditions, like the inlet mass flow of fuel, the 
mass flow rate of the air supply and the Sauter mean 
diameter of an oil droplet, had to be estimated and 
changed several times. Besides, the appropriate 
numerical models have been chosen to provide similar 
conditions under which the experiment was carried out, 
since not all volume measurements of quantities 
appearing during the combustion process and being 
possible to measure in the experiment were available. 
Some model parameters have been collected in tab. 1. 
Table 1. Model parameters. 
Air mass flow [kg/s]: 0.014952 
Air temperature [K]: 297 
Air mass fraction: 0.232 (O
2
), 0.768 (N
2
) 
Turbulence Intensity: Medium (5%) 
Fuel mass flow [kg/s]: 0.000861 
Fuel temperature [K]: 343 
Liquid evaporation method: Light oil modification 
Coefficients of Antoine Equation: A = 23.3, B = 5600 [K], C = 25 [K], p
scale 
= 1 [Pa] 
Coefficient of EDC Combustion model: A= 4, B = -1 
Fluid Pair: 
 
Heat Transfer Model:  Fluid dependent 
Heat Transfer Model for Gas Mixture: Thermal Energy 
Heat Transfer Model for Extra Light Fuel Oil: Particle Temperature 
Particle Coupling: Fully Coupled 
InterPhase Heat Transfer: Ranz Marshall 
InterPhase Thermal Radiation Transfer: Opaque 
Buoyancy Model: Full Buoyancy Model (Density Difference) 
Buoyancy Turbulence: Production and Dissipation 
Wall temperature [K]: 363 
Outlet Static Pressure [Pa]: -3 
Reference Pressure [atm]: 1 

|  47 
NUMERICAL ANALYSIS OF OIL COMBUSTION IN A SMALL COMBUSTION DEVICE 
Actual distribution of oil droplets, which is the 
function of the diffusion oil nozzle and the diffusion 
parameters, was defined by applying the function of 
Rosin-Rammler distribution:  
 
 
 
with the diameter                      and the spread parameter 
n, which is a measure of dispersion of particle sizes 
where lower values indicate wider dispersion,          . 
The dispersion has been determined in the shape of a 
hollow cone with the angle 60° and the dispersion angle 
5°, the injection velocity magnitude 28 m/s, and the 
injection oil droplet temperature 70 °C. 
In fig. 3 it is demonstrated that the biggest oil 
droplets appear at the beginning of dispersion within the 
size range 100 μm. After that, oil droplets disperse into 
smaller ones and vaporize. As a result, conditions for 
dispersion fineness are fulfilled. Oil droplets of such 
size class represent conditions similar to those in the 
experiment. From combustion process modeling 
viewpoint, droplets should have various diameters that 
have different combustion times, since that is the only 
way to provide formation of a stable flame. 
4. Results and Discussion 
4.1. Numerical results 
Flame is a typical feature of combustion, and 
includes a narrow area where basic physical-chemical 
processes of oxidation take place, and broadens through 
space filled with reactants. In most cases, flame can be 
visible, which is the consequence of the presence of 
different temperatures and molecules of intermediate 
combustion products. 
High temperatures arise in the flame zone. This 
means that the temperature field in the highest 
temperature range at the same time also presents the 
flame position (see fig. 4a). It is possible to conclude 
that in the flame front, high temperature gradients and 
changes in the reacting flow composition are present. 
The reacting flow is characterized by maximal 
combustion temperatures where their values depend on 
λ and the heat exchange rate with the environment. On 
the basis of the velocity field (see fig. 4b and 4c) it can 
be established that the highest velocities appear in the 
centre of the flame. This is understandable, since there is 
the greatest impact of the temperature gradients on the 
density, and consequently through the momentum 
equation also on the velocity components. The field of 
consumption of the air or O
2
 concentrations is located 
within the combustion zone where fast consumption is 
typical (see fig. 4d). The excess of the O
2
 mass 
concentration is distributed relatively equally outside the 
flame through the whole combustion chamber. After O
2
 
enters the combustion chamber, the concentration of 
H
2
O (water vapor) (see fig. 4e) and of CO
2
 (see fig. 4f) 
noticeably increases in the flame zone. An intensive 
chemical reaction of combustion namely takes place in 
the centre of the flame where the highest temperature is 
present. The concentration begins to distribute equally 
through the whole combustion chamber, what is also to 
be expected, since CO
2
 as well as H
2
O (water vapor) are 
products of complete combustion. The indicator of 
incomplete combustion is the presence of CO in 
combustion products (see fig. 4g) or in flue gases. The 
highest concentrations of CO appear immediately in 
front of the flame zone on the side of the fuel supply. 
Further, the combustion is complete, which indicates 
that there is enough air, and that conditions for complete 
combustion are ensured.  
 
(20) 
Figure 3. Distribution and size of oil droplets. 

48  |  
ANALI PAZU, 4/ 2014/ 1, str. 41-50                                                                                                                  Matej ZADRAVEC et al. 
 
Figure 4. Numerical results. 

|  49 
NUMERICAL ANALYSIS OF OIL COMBUSTION IN A SMALL COMBUSTION DEVICE 
Thermal NO represents 95% of all NO
X
, which arise 
through the combustion of ELFO (see fig. 4h). The 
highest concentrations of NO appear in the flame area 
where the highest temperatures and enough O
2
 for 
oxidation of N
2
 are present. The reasons for formation 
of thermal NO are in particular high temperatures near 
the flame front. For their appearance sufficient quantity 
of O
2
 is needed. O
2
 reacts then with N
2
 (see fig. 4i), 
which leads to creation of thermal NO. The temperature 
load of the combustion device wall (see fig. 4j) shows 
that the back wall is most exposed to the temperature on 
account of the changeable length of the flame. 2D 
surface streamlines (see fig. 4k) and 3D streamlines (see 
fig. 4l) shows exact information about behaviors within 
the combustion chamber. The findings in our case 
indicate that the flow is extremely turbulent and 
completely unsteady. The numerical results confirm that 
an appropriate λ
opt
 was determined in the scope of the 
experiment, and consequently the conditions for 
complete combustion were ensured. 
4.2. CFD Validation with Experimental Data 
Numerical results had to be properly verified in 
order to obtain information about the suitability of the 
numerical model. The comparison between different 
results is in Table 2: numerical results, results of 
experimental measurements, analytical calculation 
results based on solving the empirical stoichiometric 
equations for liquid fuels, and results taken from 
literature. Comparison in tab. 2 shows that these results 
are in good agreement with each other (deviation is less 
than 5%). 
In the scope of the experiment has been possible to 
record the appearance of the turbulent flame, and then 
compare it with the CFD numerical model. The findings 
indicate that the area with the highest temperatures is 
present in the flame center where the injection of the 
liquid fuel is also performed (see fig. 5).  
Deviations between numerical and experimental 
results appear because the numerical analysis does not 
consider the effect of atomization of liquid fuel into oil 
droplets. In carrying out the experiment, however, the 
following factors should be taken into account: 
permanent air pressure changes, permanent air 
temperature changes, permanent pressure changes in the 
furnace, and other weighting factors. Despite these 
simplifications, the numerical model provides 
sufficiently accurate description of the combustion 
physical behavior. 
Table 2. Comparison of results. 
  
λ = 1,2 
Numerical 
[vol. %] 
Experimental 
[vol. %] 
Analytical 
[vol. %] 
Literature [15] 
[vol. %] 
Literature [16] 
[vol. %] 
O
2,wet
 3.22 3.16 3.32 3.4 3.28 
O
2,dry
 3.61 3.5 3.71 3.77 3.69 
CO
2,dry
 12.72 12.8 12.62 12.6 12.74 
Figure 5. Comparison of the turbulent flame in the experiment and in the CFD analysis. 

50  |  
ANALI PAZU, 4/ 2014/ 1, str. 41-50                                                                                                                  Matej ZADRAVEC et al. 
5. Conclusion 
To sum up, the paper illustrates that our description 
of the turbulent combustion process of liquid fuels and 
physical behaviors inside the combustion chamber is 
sufficiently accurate. This description is namely based 
on the selected numerical models in the scope of the 
CFD as well as on assumptions and experimental 
presuppositions. The findings were confirmed with the 
presented case of ELFO combustion in a small 
combustion device. In practice oil burner always need to 
operate with some excess of air (generally 15-20% of 
the required stoichiometric  amount) to achieve a 
complete combustion. The suggested optimal air-fuel 
ratio for the presented case is λ
opt
 = 1.2, which is based 
on detailed analysis of the experimental measurement 
results of flue gas emissions. The most sensitive 
indicator of the incomplete combustion is concentration 
of CO, which is formed because of insufficient amount 
of the oxygen. Another reason for the incomplete 
combustion can also be an improper mixing of the 
combustion air with the fuel. Therefore the CO can be 
regarded as a good indicator of the combustion quality. 
Incomplete combustion means that less energy is 
released and combustion efficiency is reduced. 
Appropriate settings of the oil burner can enable fuel 
savings up to 10%, sometimes even more. That is why 
combustion control and determination of an appropriate 
optimal air-fuel ratio λ
opt
 are of great importance. By 
applying the appropriate λ
opt
, it can be established, 
whether the optimal combustion is achieved, since only 
1% of CO in flue gases would be enough to cause an 
increase of the furnace losses of 4 to 5%. 
Literature 
1. J. Broukal, J. Hajek, Wall Heat Fluxes in Swirling 
Combustion of Extra Light Fuel Oil in Large Scale 
Test Combustor: Experiment and Modelling Using 
Eddy Dissipation Model. 13th Conference on 
Process Integration, Modelling and Optimisation 
for Energy Saving and Pollution Reduction, Book 
Series: Chemical Engineering Transactions, 
Volume 21, 1111--1116, 2010. 
2. FK Jager, H. Kohne, CFD modelling in process 
burner development: Combustion of light residual 
fuel oils. Computational Fluid and Solid 
Mechanics, Vols 1 and 2, Proceedings, 966--970, 
2003. 
3. A. Saario, A. Rebola, PJ Coelho, M. Costa, A. 
Oksanen, Heavy fuel oil combustion in a 
cylindrical laboratory furnace: measurements and 
modeling. Fuel, Volume 84, 359--369, 2005. 
4. DR Schneider, Z. Bogdan, Modelling of SO3 
formation in the flame of a heavy-oil fired furnace. 
Chemical and Biological Engineering Quarterly, 
Volume 17, 175--181, 2003. 
5. M. Takei, R. Weber, T. Niioka, Mathematical 
modeling of industrial furnaces considering 
detailed oil spray characteristics. Combustion 
Science and Technology, Volume 175, 1237--1262, 
2003. 
6. T. Poinsot, D. Veynante, Theoretical and 
Numerical Combustion, Second Edition, 2005. 
7. C. K. Westbrok, F. L. Dryer, Simplified Reaction 
Mechanisms for the Oxidation of Hydrocarbon 
Fuels in Flames. Combustion Science and 
Technology, 31--43, 1981. 
8. N. Peters, Turbulent combustion, Cambridge, 2004. 
9. J. Warnatz, U. Maas, R.W. Dibble, Combustion: 
Physical and Chemical Fundamentals, Modelling 
and Simulation, Experiments, Pollutant Formation, 
Springer Verlag, Berlin Heidelberg, 1996. 
10. Yakhot, V., Orszag, S.A., Thangam, S., Gatski, 
T.B. & Speziale, C.G., Development of turbulence 
models for shear flows by a double expansion 
technique. Physics of Fluids A, Vol. 4, No. 7, 
pp1510-1520, 1992. 
11. Ansys, Ansys – CFX Documentation, Release 13, 
November 2010. 
12. Ishii M. and Zuber N., Drag coefficient and relative 
velocity in bubbly, droplet or particulate flows. 
AIChE Journal, Volume25, Issue 5, 843--855, 
1979. 
13. A. Habibi, B. Merci, G. J. Heynderickx, Impact of 
radiation models in CFD simulations of steam 
cracking furnaces. Computer & Chemical 
Engineering, 1389--1406, Volume 31, Issue 11, 
2007. 
14. B. Rajh, Experimental – numerical analysis of oil 
combustion in a small burning device, diploma 
work, Maribor, 2009. 
15. Weishaupt, Technische Arbeitsblätter, 2001. 
16. J. Oman, A. Senegačnik, A. Mirandola: Air, fuels 
and flue gases: physical properties and combustion 
constants. Padova: Progetto, 106--109, 2006.