Dissertation

COMBINING A ONE-DIMENSIONAL EMPIRICAL AND
NETWORK SOLVER WITH COMPUTATIONAL FLUID

DYNAMICS TO INVESTIGATE POSSIBLE
MODIFICATIONS TO A COMMERCIAL GAS TURBINE
COMBUSTOR
by
Johannes Jacobus Gouws
Submitted in partial fulfilment

of the requirements for the degree
MASTER OF ENGINEERING
in the Faculty of Mechanical/Aeronautical Engineering

University of Pretoria
Pretoria
Supervisor: Mr R.M. Morris
January 2007
COMBINING A ONE-DIMENSIONAL EMPIRICAL AND
NETWORK SOLVER WITH COMPUTATIONAL FLUID
DYNAMICS TO INVESTIGATE POSSIBLE
MODIFICATIONS TO A COMMERCIAL GAS TURBINE
COMBUSTOR
by
Johannes Jacobus Gouws
Promoter: Mr R.M. Morris

Department: Mechanical and Aeronautical Engineering
Degree: Master of Engineering
SUMMARY
Gas turbine combustion chambers were traditionally designed through trial and error
which was unfortunately a time-consuming and expensive process. The development of
computers, however, contributed a great deal to the development of combustion
chambers, enabling one to model such systems more accurately in less time.
Traditionally, preliminary combustor designs were conducted with the use of one-
dimensional codes to assist in the prediction of flow distributions and pressure losses
across the combustion chamber mainly due to their rapid execution times and ease of use.
The results are generally used as boundary conditions in three- dimensional models to
predict the internal flow field of the combustor. More recent studies solve the entire flow
i
field from prediffuser to combustor exit. This approach is, however, a computationally
expensive procedure and can only be used if adequate computer resources are available.
The purpose of this study is two-fold; (1) to develop a one-dimensional incompressible

code, incorporating an empirical-based combustion model, to assist a one-dimensional
network solver in predicting flow- and temperature distributions, as well as pressure
losses. This is done due to the lack of a combustion model in the network solver that was
used. An incompressible solution of flow splits, pressure losses, and temperature
distributions is also obtained and compared with the compressible solution obtained by
the network solver; (2) to utilise the data, obtained from the network solver, as boundary
conditions to a three-dimensional numerical model to investigate possible modifications
to the dome wall of a standard T56 combustion chamber. A numerical base case model is
validated against experimental exit temperature data, and based upon that comparison,
the remaining numerical models are compared with the numerical base case. The effect of
the modification on the dome wall temperature is therefore apparent when the modified
numerical model is compared with the numerical base case.
A second empirical code was developed to design the geometry of axial straight vane
swirlers with different swirl angles. To maintain overall engine efficiency, the pressure
loss that was determined from the network analysis, of the base case model, is used
during the design of the different swirlers. The pressure loss across the modified
combustion chamber will therefore remain similar to that of the original design. Hence, to
maintain a constant pressure loss across the modified combustion chambers, the network
solver is used to determine how many existing hole features should be closed for the
pressure loss to remain similar. The hole features are closed, virtually, in such a manner
as not to influence the equivalence ratio in each zone significantly, therefore maintaining
combustion performance similar to that of the original design. Although the equivalence
ratios in each combustion zone will be more or less unaffected, the addition of a swirler
will influence the emission levels obtained from the system due to enhanced air-fuel
mixing.
ii
A purely numerical parametric analysis was conducted to investigate the influence of
different swirler geometries on the dome wall temperature while maintaining an
acceptable exit temperature distribution. The data is compared against the data obtained
from an experimentally validated base case model. The investigation concerns the
replacement of the existing splash-cooling devices on the dome wall with that of a single
swirler. A number of swirler parameters such as blade angle, mass flow rate, and number
of blades were varied during the study, investigating its influence on the dome wall
temperature distribution.
Results showed that the swirlers with approximately the same mass flow as the existing
splash-cooling devices had almost no impact on the dome wall temperatures but
maintained the exit temperature profile. An investigation of swirlers with an increased
mass flow rate was also done and results showed that these swirlers had a better impact
on the dome wall temperatures. However, due to the increased mass flow rate, stable
combustion is not guaranteed since the air/fuel ratio in the primary combustion zone was
altered.
The conclusion that was drawn from the study, was that by simply adding an axial air
swirler might reduce high-temperature gradients on the dome but will not guarantee
stable combustion during off-design operating conditions. Therefore, a complete new
hole layout design might be necessary to ensure good combustion performance across a
wide operating range.
iii
ACKNOWLEDGEMENTS
It would not have been possible to complete this work without the help of many people.
However, there are a special few to whom I am grateful that I would like to mention
individually. First of all, I am grateful to my advisor, Mr R.M. Morris, for providing me
the opportunity to conduct my studies on gas turbine combustion chambers. Mr Morris’s
stimulating comments and arguments have been a constant source of inspiration not only
as a study leader but also as a friend. In addition, special thanks to Dr. J.E. van Niekerk
for his valuable inputs into the project.
I would also like to thank Dr. G.P. Greyvenstein and Mr J. van der Merwe from M-Tech
Industrial for providing the network solver, FLOWNEX, and for their support.
Additionally, I would like to thank Armscor for providing the financial assistance for the
this project.
I would like to thank my family members for their support and prayer. First and foremost,
I would like to thank my parents as well as my brother and grandmother for their
continued support. Special thanks to my loving fiancé Nandel, words cannot describe
how much I appreciate all the love and support during this project.
Finally, but certainly not the least, I would like to thank our heavenly Father for
providing me with the strength, endurance and ability to complete this project
successfully.
Philippians 4:13 I can do everything through Him who gives me strength.
iv
TABLE OF CONTENTS
Page
SUMMARY i
ACKNOWLEDGEMENTS iv
NOMENCLATURE viii
CHAPTER 1: INTRODUCTION 1
1.1 Preamble 1
1.2 Overview of the T56 gas turbine engine 3
1.3 Combustor design methodology 6
1.4 Need for this study 13
1.5 Outline of this study 15
CHAPTER 2: COMBUSTOR DESIGN PRINCIPLES:

1-D EMPIRICAL DESIGN 16
2.1 Preamble 16
2.2 Basic combustor layout and design 16
2.3 One-dimensional flow calculations 19
2.4 Heat transfer process 27
2.5 Film-cooling devices 32
2.6 Swirler design 35
2.7 1-D flow and temperature predictions 40
2.8 Summary 43
CHAPTER 3: 1-D NETWORK APPROACH 45

3.1 Preamble 45
3.2 1- D network solver 45
3.3 Prediction of discharge coefficients and jet angles 56
3.4 One-dimensional flow and temperature prediction 57
v
3.5 1-D temperature predictions 60
3.6 Empirical swirler design 62
3.7 Summary 65
CHAPTER 4: NUMERICAL MODEL 66

4.1 Preamble 66
4.2 Turbulence models 66
4.3 Combustion model 70
4.4 Fuel spray model 74
4.5 Grid generation 75
4.6 Boundary conditions 77
4.7 Summary 79
CHAPTER 5: NUMERICAL STUDY 80

5.1 Preamble 80
5.2 Experimental data 80
5.3 Numerical test cases 81
5.4 Numerical analyses of base case 82
5.5 Numerical analysis of swirler 84
5.6 Predicted wall temperature 102
5.7 Summary 104
CHAPTER 6: SUMMARY, CONCLUSION AND RECOMMENDATIONS 105

6.1 Summary 105
6.2 Conclusion and contributions 106
6.3 Recommendations and future work 109
REFERENCES 112
APPENDIX A: MATERIAL PROPERTIES 119
vi
APPENDIX B: GRID INDEPENDENCE STUDY 121
B.1 Grid independence study 121
B.2 Effect of radiation on the liner wall temperature 121
APPENDIX C: EMPIRICAL SWIRLER DESIGN CODE 124

C.1 Empirical swirler design code 124
C.2 Swirler lookup table (Program output) 128
APPENDIX D: 1-D EMPIRICAL CODE 132

D.1 1-D empirical program 132
D.2 Inputs 134
D.3 Flowcheck 138
D.4 Flowsplits 150
D.5 Combustion 160
D.6 Elements 172
D.7 Heat Transfer 175
APPENDIX E: BOUNDARY CONDITIONS 182

E.1 Base case 182
E.2 Base case with blocked splash cooling devices 183
E.3 Case 1 184
E.4 Case 2 185
E.5 Case 3 186
E.6 Case 4 187
E.7 Case 5 188
vii
NOMENCLATURE
A = Area [ m 2 ]
Ar = Aole area ratio
A/F = Air/fuel ratio
C/H = Carbon to hydrogen ratio
Cd = Discharge coefficient
Cp = Specific heat [kJ/kgK]
C1 = Internal convection [W]
C2 = Convection to annulus [W]
C3 = External convection [W]
Dh = Hydraulic diameter [m]
F/A = Fuel/air ratio
Gφ = Tangential momentum flux
Gx = Axial momentum flux
I = Total number of nodes

J = Number of branches associated with a specific node
K = Loss coefficient
Ksw = Swirler blade loss coefficient
K1-2 = Conduction through liner wall [W]
K2-3 = Conduction through casing wall [W]
Lu = Luminosity factor
Lv = Swirler vane length [m]
P = Total pressure [Pa]
Pi = Total pressure upstream of hole [Pa]
Pw = Wetted perimeter [m]
R = Radiation [W]
SN = Swirl number
Sv = Swirler blade pitch [m]
T = Temperature [K]
viii
Ts = Static temperature [K]
Tw,ad = Gas temperature at the combustor liner wall [K]
ΔT = Temperature rise [K]
Q = Volume flow rate [m³/s]
V = Velocity [m/s]
Vn = Normal velocity component
Vt = Tangential velocity component
W = Tangential velocity component
kw = Wall conductivity [W/mK]
lb = Beam length [m]
k = Conductivity [W/mK]
m = Massflow rate [kg/s]
pi = Static pressure upstream of hole [Pa]
pj = Static pressure downstream of hole [Pa]
rt = Tip radius [m]
rh = Hub radius [m]
tw = Liner wall thickness [m]
vnj = Jet velocity component normal to the liner wall
vj = Resultant jet velocity
x = Distance from cooling slot [m]
Symbols
θ = Jet angle [ º]
θv = Swirler blades angle [ º]
φ = Equivalence ratio
ρ = Density [kg/ m 3 ]
ηc = Combustion efficiency
ηf = Film-cooling efficiency
εg = Gas emissivity
εw = Wall emissivity
ix
σ = Stefan Boltzmann constant (5.67 x 10-8 ⎡⎣ W/mK 4 ⎤⎦ )
μ = Viscosity [kg/ms]
α = Mass flow ratio
ϕ = Mass flow ratio/area ratio
Subscripts
a = Air
an = Annulus
c = Casing
g = Gas
h = Hole / Hub
hub = Hub
L = Combustor liner
mr = Maximum circumferential mean value
max = Maximum
out = Zone outlet
q = Fuel/air ratio by mass
sw = Swirler
w1 = Inner liner wall
w2 = Outer liner wall
zone = Recirculation, primary, secondary, dilution zone
ij = Values associated with element eij
nj = Jet normal to flow direction
PZ = Primary zone
RZ = Recirculation zone
1 = Flame side of liner wall
3 = Combustor inlet
4 = Combustor outlet
03 = Total inlet property
x
CHAPTER 1 Introduction
________________________________________________________________________
CHAPTER 1 –
INTRODUCTION
1.1 Preamble
Modern gas turbine combustor design and development is typically a combination of
using empirical correlations, numerical modelling and extensive component testing. The
United States Army Research and Technology Laboratories is one of many organisations
who demonstrated in 1975 the ability of using empirical/analytical methods to
successfully design and develop a small reverse-flow annular combustor. Prior to 1975,
combustion chamber design consisted mainly of expensive and time-consuming trial and
error rig tests. Even though the empirical/analytical method provided valuable design
guidance, the accuracy of the model was limited by sub-models predicting turbulence,
combustion and heat transfer (Mongia et al., 1986).
Due to the ever-increasing computing power, numerical methods such as computational

fluid dynamics (CFD) has become a more attractive and alternative approach that could
be utilised during the design and development of combustors. Numerical modelling
provides the combustor designer with the ability to predict and understand the complex
flow process within the combustion chamber early in the design phase prior to expensive
rig tests. Experimental rig tests will, however, never be totally eliminated, but with the
use of numerical methods, experimental time can be reduced to only a few verification
experiments.
1
________________________________________________________________________
Numerical modelling is a very broad term and often interpreted as the use of CFD in
combustion simulation. Although very powerful, CFD solutions of combustors are very
specialised and time consuming processes. CFD techniques are therefore very seldom
used during the initial sizing of a combustor. Initial sizing of combustors are usually
done through empirical formulations and when extended to a full CFD simulation, the
empirical formulations are used to calculate the boundary conditions for the CFD
simulation.
With the recent developments in compressible network solvers, network solvers now
have the potential of filling the modelling gap between and analytical design and a full
CFD simulation. A network solver has the advantage that it can account accurately for
the geometry and therefore flow splits in the combustor, while also accounting for the
properties of the combustion process.
The purpose of this project is to use numerical modelling to investigate possible

modifications to the dome of a standard Allison T56 combustion chamber due to the
formation of cracks in the area. These cracks are believed to be caused by high-
temperature gradients, which in some cases may even be augmented by blocked splash-
cooling strips on the dome (Figure 1.1). The design of the combustor currently in use is
rather old, compared to modern-day combustion systems, but still widely used.
2
________________________________________________________________________
(a) (b)
Figure 1.1: Defects on combustor dome; (a) Crack formation on dome wall,
(b) Thermal distortion of the splash cooling devices on the inside of the dome
The thermal characteristics of the combustor will be modeled by combining a one-

dimensional analytical model with a commercial network solver, Flownex, to predict
flow distributions, pressure losses, and one-dimensional temperature distributions. The
data obtained will be used in the CFD analysis to investigate possible modifications to the
existing geometry to solve crack formation problems. The numerical model will then
also be used to simulate geometrical changes to the combustor and the effect it will have
on the simulated thermal gradients. Although experimental exit temperature distributions
will be used to validate the numerical data of the base case, the modification analysis will
be purely numerical. Thermal stresses will not form part of this project.
1.2 Overview of the T56 gas turbine engine

The T56 engine is a turboprop gas turbine engine used by several air forces worldwide.
These engines are used on aircraft such as the Lockheed C-130 Hercules, L-188 Electra,
and the P3-Orion, to name a few. The engine is depicted in figure 1.2 and consists of a
14-stage axial compressor, combustion system, and four turbine stages. At take-off the
compressor has a pressure ratio of 9.5:1 and delivers 14.515 kg/s of air at an inlet
temperature and pressure of approximately 566.15 K and 923 897 Pa to six can-annular
throughflow combustion chambers. All of this is achieved at a constant rotational speed
of 13 820 rpm.
3
________________________________________________________________________
Figure 1.2: The T56 gas turbine engine (www.rolls-royce.com)
The combustion system consists of an inner and outer casing that provides the airflow
and mechanical connection between the compressor and turbine. The combustion gases
flow into a four-stage axial turbine, providing it with an acceptable temperature
distribution that promotes the lifespan of the turbine blades. A single combustor unit is
depicted in figure 1.3.
Figure 1.3: Single T56 combustion chamber
The combustion chamber can be divided into three different sections namely; primary,
secondary and dilution zones. The importance of each zone will be discussed in the
subsequent chapter. The primary zone contains a single hole set comprising of seven air
admission holes. The secondary and dilution zone each consists of two hole sets. The first
and second hole sets in the secondary zone consists of three and four holes respectively,
whereas the first and second sets in the dilution zone comprises of four and two holes
4
________________________________________________________________________
respectively. Use is made of eight splash strips on the dome to impart a swirling motion
to the flow within the primary zone and to cool the dome wall. Film-cooling air is
admitted through five wiggle strip sets along the combustor liner as well as a number of
splash-cooling devices. Typical operating conditions during take-off, and used during this
study as the design point, are displayed in table 1.1.
Table 1.1: Operating conditions during take-off

Combustor inlet temperature, T03 [K] 566.15
Combustor inlet pressure, P03 [Pa] 923 897
Total mass flow rate, ma [kg/s] 14.154
Air/fuel ratio 50
Since the T56 engine dates back to the 1950s, minor changes have been made to the
combustion chamber. Due to stringent regulations to reduce aircraft emission levels, such
as NOx, CO, and smoke, some investigations were conducted on the combustion system
to reduce smoke emissions.
Previous work was done by Skidmore (1986) on two different combustion systems, the
T56-A-7 (Series II) and T56-A-15 (Series III), to investigate the discrepancy in smoke
emission characteristics between the two designs. It was found that the Series III
produced approximately 50% more smoke than the Series II engine. It is believed that the
reason for this difference might be due to the design of the primary zone, especially the
amount of air flow through the first set of circumferential holes. The percentage mass
flow rate has decreased from 6.4% for the series II combustor to 3.4% for the series III
combustor. Further investigations were underway to determine the influence of the
primary zone holes on the smoke emission levels (Skidmore, 1986).
A low-smoke modification has been introduced into the Series III Allison T56 engine
combustion system and used by the Royal Australian Air Force (RAAF) in 1990. The
modification consisted of enlarging the primary zone holes from 6.4mm to 11mm to
increase the strength of the main toroidal vortex, thus improving the mixing of the air and
fuel. However, extensive turbine erosion became evident after the modified combustion
5
________________________________________________________________________
system had been employed for two years. After some investigations, it was found that the
erosion was caused by hard carbon deposits which were formed in the fuel-rich region
inside the dome of the combustion chamber (Skidmore et al., 1995).
The fundamental purpose of a gas turbine combustion chamber is to promote stable and
efficient combustion over a wide range of operating conditions while providing an
acceptable exit temperature distribution to the turbine vanes. In addition, the combustor
should have a low total pressure loss to keep the engine efficiency high, low emission
levels that comply with regulations, acceptable liner wall temperatures in order to ensure
structural durability and good re-light capabilities over a wide range of air/fuel ratios.
Designing such a device to comply with all of the above-mentioned requirements,
therefore, becomes a difficult task in which certain compromises have to be made. The
following section will discuss related literature and some of the tools that are used during
a typical preliminary design phase as well as a few important aspects concerning
combustion chamber performance.
1.3 Combustor design methodology

Traditionally, combustor design or modification required a number of experimental rig
tests that were unfortunately a very expensive and time-consuming process (Lawson,
1993; Eccles and Priddin, 1999). Alternative methods were therefore needed to conduct
preliminary design studies to predict total pressure losses and flow distributions across
combustion chambers prior to expensive rig tests. Some of the earliest work was done by
Knight and Walker (1953) who conducted analyses to predict component pressure losses
that could be used to predict isothermal flow distributions of gas turbine combustion
chambers. Losses through swirlers and liner hole features as well as losses due to heat
addition and bends were determined. Certain limitations on the accuracy of these
analyses were due to uncertainty of compressibility and the mixing of the gas streams,
but resulted in pressure loss predictions within 5 percent of measured data. In an attempt
to investigate the influence of a number of inlet parameters on the total pressure loss and
flow distributions of a combustion chamber, Graves and Gronman (1957) presented all
these variables graphically. These graphs were presented in terms of the combustor
6
________________________________________________________________________
reference Mach number, ratio of combustor exit-to-inlet total temperature, fraction of
total airflow passing through the liner, ratio of total hole area in the liner wall to the total
combustor cross-sectional area, and the ratio of the liner cross-sectional area to total
combustor cross-sectional area. The effects of a number of geometries and operating
variables were determined analytically from compressible and incompressible flow
relations. These curves were, however, developed for tubular combustion chambers with
constant annulus and liner cross-sectional area along the combustor axis with flush
circular air admission holes. These results could, however, also be used for can-annular
and annular combustion chambers with constant annulus air velocities, but with less
accuracy.
As the development of computers progressed, it became a viable option for designers to

use it as a tool to conduct one-dimensional analysis that is less time-consuming. Samuel
(1961) described a one-dimensional analytical method that was based on the actual
sequence of processes that the flow experienced as it passed through the combustion
system. The method calculated the pressure distribution and mass flow rates throughout a
gas turbine combustion system when inlet flow conditions, exit temperatures, and the
geometry of the system are known. The method was designed for annular and can-
annular systems with any type of liner flow. Joubert and Hattingh (1982) described a
similar approach in which the pressure drop and flow splits were calculated for a reverse-
flow combustion chamber. The analysis employed the continuity, momentum, and energy
equations in a step-wise manner. The combustor model was divided into a number of
stations and calculations proceeded from one station to the next by using the results from
the previous station as the inputs to the following station. The method proved to be
adequate for combustion chamber design.
The one-dimensional approach described above is, however, not extremely versatile.
When complex geometries need to be analysed this approach becomes difficult to
implement. A one-dimensional network model on the other hand is capable of modelling
complicated geometries effectively without difficulty, while maintaining rapid execution.
Stuttaford and Rubini (1997) described a network model consisting of a number of
7
________________________________________________________________________
independent sub-flows that are linked together to model a certain process. The process
can be modelled by overlaying a network on the system geometry that comprises a
number of elements that are linked together by nodes. The elements define the actual
feature such as orifice and duct sections in the domain of interest. The elements are then
linked together by nodes to form a meaningful overall structure. Semi-empirical
formulations are used to describe the flow through elements, while the overall governing
equations are solved within the nodes. The governing equations that are employed consist
of the continuity equation and pressure-drop flow relations. Pressure drop and flow
distributions can be obtained in this way throughout the region of interest. Heat transfer
can be included in the strategy, by computing conduction, convection, and radiation
effects (Stuttaford, 1997).
A commercially available network solver, FLOWNET (www.flownex.com), was utilised

by Hicks and Wilson (1999) to predict general heat transfer to a combustor liner and to
investigate the influence of cooling devices on the liner wall. The original network code
was developed by a company in South Africa but was adapted for combustor analysis by
Cranfield University. A number of semi-empirical sub-models to represent various
features relevant to a gas turbine combustor were added to the model. Some of the
features included cooling/dilution ports, diffusers, pedestals, and pin-fins. To be able to
predict combustion products at various efficiency levels, a gas property model was
incorporated as well. The heat transfer analysis included in the FLOWNEX model
accounted for conduction, convection, and radiation in the combustor. Combustor film-
cooling effects were included in the convection model. The radiation model initially
made use of a simple semi-empirical relationship based on a luminosity factor as outlined
by Lefebvre (1998). This model proved to be capable of providing quantitive agreement
with experimental results. It is important to note that the accuracy and limitations of
network models are inherent due to the use of empirical formulations that might not be
specifically suited for the problem at hand. Although the network model was capable of
predicting general trends it was unable to predict temperatures within an accuracy less
than 80 K.
8
________________________________________________________________________
With the use of these empirical/analytical models, a number of advanced technology
combustors have been designed over the years. However, even though the models have
proved to be useful in the combustor design process, some limitations persist. These
models cannot predict local hot spots, exit temperature distributions, flow phenomena
inside a combustion chamber, combustor lean-flame stability and ignition characteristics.
A need therefore persists for more accurate modelling techniques that could provide the
engineer with insight regarding the internal flow in a combustor.
Due to the hostile environment inside a combustion chamber, it is difficult to investigate

the internal flow field using experimental measurements. However, numerical methods
can be used to model the internal flow field of a combustor. The numerical solution can
provide detailed information of all flow properties being modelled across the entire flow
field, whereas experimental measurements can only provide flow properties being
measured in the region of the flow field where measurements are taken (Hu and Prociw,
1993). CFD has been developed since the early seventies but due to the requirement of
specialist knowledge and intensive computer power, these tools have remained in the
domain of the specialist numerical scientist. The recent development of relatively
inexpensive but powerful computers has, however, led to the use of CFD as an everyday
designer’s tool. CFD provides the combustor designer with the ability to understand the
complex flow process within the combustion chamber early in the design process, and to
improve the design if necessary, thus requiring fewer expensive rig tests. Rig tests will,
however, never be completely eliminated, but with the use of numerical methods such as
CFD, experimental time can be reduced to only a few verification experiments (Eccles
and Priddin, 1999; Hornsby and Norster, 1997; McGuirk and Palma, 1993).
As described by Sivaramakrishna et al. (2001), the role of CFD in the development of gas
turbine combustion chambers has changed remarkably over the last decade and has
become a valuable part of an overall integrated design system. After validating these
codes with reliable experimental data, it has been utilised for optimisation studies. Such
studies include optimisation of combustion chambers for low emission levels and
9
________________________________________________________________________
adequate exit temperature distributions by determining optimum liner hole size and
location.
Due to more stringent regulations regarding emission levels and its impact on the
environment, it has become a great concern to engine manufacturers and they are
responding to the need. The only feasible changes that can be made to combustion
chambers to reduce emission levels, are changes to fuel injectors and liner hole patterns
and locations. The exhaust gases from a gas turbine engine consist of carbon dioxide,
carbon monoxide, water vapour, unburned hydrocarbons, soot, and excess oxygen and
nitrogen. Carbon dioxide and water vapour are natural products formed during the
combustion process, and the only way to reduce these levels is to add less fuel. Unburned
hydrocarbons, soot, and carbon monoxide are, however, a result of inefficient combustion
(Sturgess et al., 1992). Hornsby and Norster (1997) utilised CFD to predict emission
levels from a can-annular combustor that gave good quantitive results. Others such as
Feitelberg and Lacey (1998) and Liedtke et al (2002) utilised CFD to investigate the
performance of new low emission combustor designs, such as rich-burn quick-quench
lean-burn (RQL) and lean premixed prevapourised (LPP) combustors.
One of the more difficult problems encountered in gas turbine combustion chambers is to
achieve an exit temperature profile that is acceptable to the nozzle guide vanes (NGVs).
The exit temperature profile is affected by; the fuel injector spray characteristics, jet
penetration and mixing, total pressure drop across the liner, as well as the physical
dimensions and shape of the combustor liner. When the exit temperature profile for a
specific combustion chamber is determined experimentally, the tests are usually done at
the maximum pressure because this will be the operating condition with maximum heat
release. The temperature that is most important to the turbine blades is the average radial
temperature profile. These profiles are obtained by adding together the temperature
measurements around each radius and then dividing it by the number of locations at each
radius (Lefebvre, 1998). Another important parameter is the maximum spatial
temperature that controls the stress and erosion on the inlet guide vanes. The maximum
10
________________________________________________________________________
spatial temperature typically occurs at 50 to 70 percent of the radial position and is
normally defined in terms of the pattern factor (Crocker and Smith, 1995).
Snyder et al. (2001) and Crocker and Smith (1995) have proved that CFD analysis could
be utilised to determine optimum exit temperature distributions by studying the position
and size of the dilution holes. Lawson (1993) successfully modelled a high bypass
turbofan aircraft engine combustor in an effort to define liner and dilution modifications
necessary to create inboard peaked and flat exit temperature profiles. Before
modifications were made to the model, it was calibrated first to match existing baseline
data, thereby obtaining qualitative results. The computational modelling of combustors
has shown to be an effective method to predict exit temperature profiles. The process has
also proved to be time- and cost-effective.
The correct evaluation of CFD models is mainly dependent on the specification of

boundary conditions. These boundary conditions are generally obtained from
experimental data, but due to the complexity of combustion chamber geometries,
inappropriate experimental data, and numerical model limitations, simplifications might
be necessary when specifying boundary conditions (McGuirk and Palma, 1993;
Smiljanovski and Brehm, 1999).
A gas turbine combustor is a very complex combustion device. The challenge in such a
device is two-fold, namely (i) the development of models to describe the real-world
component geometries and (ii) an accurate description of the coupled interacting physical
and chemical phenomena (Tangirala et al., 2000). Traditionally CFD analysis of
combustion chambers has been coupled in a weakly manner, with the flow through the air
admission holes being derived from one-dimensional empirical correlations and then used
as boundary conditions to simulate the internal flow of the combustor (Hu and Prociw,
1993; Lawson, 1993; Fuller and Smith, 1993). In addition to flow distribution
predictions, jet angles through liner holes are also predicted through the utilisation of
empirical correlations as shown by Lawson (1993). These methods may, however, be
deficient in at least two ways. No information regarding velocity profile shapes or
11
________________________________________________________________________
turbulence conditions at the port entry locations is provided (McGuirk and Spencer,
2000). The solution obtained from these simulations is therefore strongly dependent on
the accuracy of the description of the boundary conditions. When coupling a CFD model
with a one-dimensional model, certain assumptions are made regarding the specification
of the boundary conditions of the various liner features. Holdeman et al., (1997)
investigated the effect of opposed rows of jets on the mixing with subsonic cross-flow in
rectangular ducts. In their analysis, a uniform flow boundary was assumed for the jets and
mainstream. The discharge coefficients (Cd) for the holes, which are defined as the ratio
between the effective hole area and the geometric hole area, were expected to be less than
unity. A discharge coefficient has therefore been assumed to determine the effective hole
area across which a uniform flow boundary could be specified.
An alternative and more recent approach is to model a combustion system in a fully

coupled manner. This approach consists of modelling the entire flow field from
compressor outlet to turbine inlet (Crocker et al., 1999; McGuirk and Spencer, 2001;
Malecki et al., 2001; McGuirk and Spencer, 2000; Snyder et al., 2001). Two reasons to
implement a fully coupled approach are: (i) flow splits and boundary conditions for the
combustor liner are modelled explicitly and no longer need to be approximated; and (ii)
liner wall temperatures can be predicted when the flow fields on both sides of the liner
walls are modelled in a coupled fashion (Crocker et al., 1999). Although belief in the
importance of fully coupled combustor modelling is evidently growing, insufficient
computer resources may limit the use of coupled calculations. (McGuirk and Spencer,
2001).
The accuracy of the CFD model depends primarily on the accuracy of the geometry being
created and also whether a sufficiently refined grid is used to capture the nature of the
flow (Hornsby and Norster, 1997). Constructing an accurate and realistic CFD model and
generating an acceptable grid are the most time-consuming and expensive phases of CFD
simulations. The two main approaches in generating a grid are either constructing a
structured or an unstructured mesh. When generating a mesh on a complex geometry, a
structured mesh is more difficult and time-consuming to construct and demand a high
12
________________________________________________________________________
degree of user expertise, but allow clustering of the mesh more efficiently in boundary
layer regions. Some geometries can simply not be meshed using a structured approach
unless significant simplifications and assumptions are made to the geometrical features.
The unstructured approach, on the other hand, offers the ability of generating a grid
automatically, and is much easier to learn and use. It also requires fewer simplifications
of the CAD model, and grid refinement near local geometric features is easily controlled.
The unstructured approach, however, requires additional processing time compared to the
structured methodology (Eccles and Priddin,1999; Maleck et al., 2001).
1.4 Need for this study

It is evident from the literature survey that in recent years CFD has become an accepted
method to model gas turbine combustion chambers, assisting designers in predicting the
internal flow and heat transfer of such systems. The results are, however, dependent on
the boundary conditions provided and need to be validated against experimental data.
Although recent computer capabilities have increased providing designers the ability to
model a combustion system in a fully coupled manner from compressor outlet to turbine
inlet, some further computer development is still needed. Modelling a combustion
chamber in an uncoupled manner is, therefore, still the preferred method.
Despite the increased progress made in numerical methods, one-dimensional analysis is

still used for initial sizing of combustion chambers. Consequently, the data obtained from
such an analysis is linked to numerical models, providing the boundary conditions for
three-dimensional numerical analysis. One-dimensional network solvers have the
capability of modelling complex combustor geometries more effectively, providing
simple, fast, and accurate solutions.
The purpose of this study is two-fold; (1) to develop a one-dimensional incompressible

code, incorporating an empirical-based combustion model, to assist a one-dimensional
network solver in predicting flow- and temperature distributions, as well as pressure
losses. This is done due to the lack of a combustion model in the network solver that was
used. An incompressible solution of flow splits, pressure losses, and temperature
13
________________________________________________________________________
distributions is also obtained and compared with the compressible solution obtained by
the network solver. In this way the abilities of a network solver is compared against a
traditional one-dimensional empirical solver; (2) to utilise the data, obtained from the
network solver, as boundary conditions for a three-dimensional numerical model to
investigate modifications to the dome wall of a standard T56 combustion chamber. A
numerical base case model will be validated against experimental exit temperature data,
and based upon that comparison, the remaining numerical models will be compared with
the numerical base case. The effect of the modification can therefore be seen when the
numerical model is compared with the original design.
Seeing that gas temperature will have an effect on the air density, and therefore the flow
distributions and pressure loss, the combustion process has to be accounted for when a
one-dimensional analysis is conducted. A combustion model was however not available
in the network solver, and for this reason, a one-dimensional empirical solver was
developed and combined with the network solver to assist in one-dimensional
combustion predictions. The one-dimensional code that was created is empirically based
and solves for incompressible flow in can- and can-annular type combustion systems.
The model, within the one-dimensional code, is divided into stations so that in moving
from one station to the next the flow distributions and pressure losses can be determined
across the system. In addition, the system model is divided into four zones in which the
total air flow in each zone can be calculated. The adiabatic flame temperature is
determined from general kerosene temperature rise curves, followed by a one-
dimensional heat transfer balance across the combustor liner. These gas temperatures are
then used as inputs to the network solver.
In addition to the one-dimensional code, a second empirical code was developed to

conduct initial sizing of axial swirlers. The input to this code is however the pressure loss
obtained from the network solver. To maintain overall engine performance within the
limits established in the engine cycle, the pressure loss had to be maintained similar to
that of the original design, for all cases. The pressure loss obtained from the network
solver is therefore used within the swirler design code, to generate various geometries
14
________________________________________________________________________
while maintaining a constant pressure loss. The network solver is used to determine
virtually how many hole features should be closed to maintain a constant pressure loss for
a specific swirler design. The hole features were closed in such a manner as not to
influence the equivalence ratio in each combustion zone significantly. Consequently, the
various swirler geometries are analyzed using CFD and its influence on the dome wall
temperatures is investigated.
1.5 Outline of this study

Chapter 2 provides background information regarding general combustor layouts as well
as one-dimensional empirical design methodology. The computational process that was
utilised during the one-dimensional empirical code is described along with the
correlations that were used.
A discussion of one-dimensional models continues in Chapter 3, elaborating more about

network models and the results obtained with such an analysis compared to the empirical
model.
Chapter 4 describes the numerical model and the associated boundary conditions that
were utilised during the study. A short description of the grid generation and numerical
models is provided.
The numerical results obtained during the study are presented in Chapter 5. The original
combustion chamber is solved and used as a base case model with its exit temperature
distribution validated against experimental data. The base case model was validated
against exit temperature distributions obtained experimentally by Skidmore (2004). Six
additional models were developed, one with the original combustion chamber with
blocked splash-cooling devices and five cases investigating the effect of a swirler on the
dome wall temperature and exit temperature distribution.
Finally, Chapter 6 provides a summary of the study, discussing the outcome of the
numerical results as well as some recommendations.
15
CHAPTER 2 Combustion design principles: 1-D empirical design
________________________________________________________________________
CHAPTER 2 –
COMBUSTOR DESIGN PRINCIPLES:
1-D EMPIRICAL DESIGN
2.1 Preamble
The following chapters will discuss the ability of a one-dimensional empirical solver to
predict pressure losses, flow distributions and temperature distributions across gas turbine
combustor chambers. Such an analysis is essential during the preliminary design phase to
reduce costs and development time and has proved to provide comparative results. The
empirical correlations utilised in such an analysis to predict flow distributions and
temperature distributions are presented in this chapter.
2.2 Basic combustor layout and design

Generally, all combustion arrangements incorporate similar components, such as a
perforated liner that is situated within an air casing, diffuser for pressure recovery
purposes, and a fuel injector. The basic combustor features are illustrated in figure 2.1.
A combustion chamber consists of a liner which shelters the flame from the approaching
air and generates an area where stable combustion can take place with the proper addition
of air. The combustion chamber can be divided into three zones, namely primary,
secondary and dilution zone. Air is added to the three different zones through various air
admission holes in the liner. The purpose of the primary zone is to anchor the flame and
to provide adequate amounts of air through the primary holes to sustain the flame. A low
16
________________________________________________________________________
pressure recirculation flow pattern is generally generated in the primary zone due to the
impingement of the primary jets on one another which could, in some cases, even be
enhanced through the utilisation of an air swirler. The recirculation region ensures that
some of the fuel can flow in the upstream direction to mix with the incoming air flow,
thus enhancing air/fuel mixing.
Figure 2.1: Basic combustor features (Mattingly et al., 2002)
The secondary or intermediate zone is downstream of the primary zone and is generally
used to provide enough time at a sufficient temperature for carbon monoxide and
hydrocarbon burn-out to occur, prior to the dilution zone. The temperatures in this zone
should be maintained low enough to avoid dissociation of products in the combustion
gases, and high enough to prevent extreme quenching. The secondary zone’s penetration
and distribution should also be adjusted to improve the temperature distribution
approaching the dilution zone.
The function of the dilution zone is to admit the remaining air after combustion and wall-
cooling requirements have been met, to provide an exit temperature profile with a mean
temperature and temperature distribution that are acceptable to the turbine blades. The
exit temperature profile is affected by the fuel injector spray characteristics, jet
17
________________________________________________________________________
penetration and mixing, total pressure drop across the liner, and the physical dimensions
and shape of the combustor liner. Experimental tests have shown that a suitable exit
temperature profile is dependent on the adequate penetration of the dilution jets into the
main stream, coupled with the correct number of jets to form a sufficient localised mixing
region. Experience has shown that when a large number of small dilution jets are used,
inadequate jet penetration is achieved resulting in a hot core. On the other hand, when a
small number of large dilution jets are used, overpenetration is evident, resulting in a cold
core (Lefebvre and Norster, 1969).
When the exit temperature profile for a specific combustion chamber is determined
experimentally, the tests are usually done at the maximum pressure because this will be
the operating condition with maximum heat release. A dimensionless parameter known as
the pattern factor can be used as an indication of the quality and uniformity of the exit
temperature profile. From an engine performance point of view, the mass-flow-weighted
mean temperature (T4) recorded for all the exit temperatures for a specific liner is of great
concern. For the design of the nozzle guide vanes (NGV), due to its fixed position at the
exit of the combustion chamber, the maximum recorded temperature should be used in
order for the vanes to withstand the temperatures. The pattern factor is normally defined
as (Lefebvre, 1998):
Tmax -T4
Pattern factor = (2.1)
T4 -T3
The profile factor on the other hand is also a dimensionless parameter that is used to
define the radial temperature profile at the combustion chamber exit. The profile factor is
defined as follows (Lefebvre, 1998):
Tmr -T4
Profile factor = (2.2)
T4 -T3
The parameters used in equation (2.1) and (2.2) are explained in figure 2.2.
18
________________________________________________________________________
Figure 2.2: Explanation of exit temperature profile parameters (Lefebvre, 1998)
2.3 One-dimensional flow calculations

For an initial one-dimensional design analysis, the mass flow distributions across the
various features of the combustion liner can be predicted with analytical and empirical
correlations. The flow through specific features is, however, affected by the geometry of
the combustor liner in the vicinity of the hole and not only by its size and pressure drop.
The essential correlations for one-dimensional flow and temperature predictions will be
discussed in the subsequent section, providing the assumption that the flow is
incompressible. The computer code developed to implement these correlations is
presented in Appendix D. The basic empirical equation for incompressible flow through
liner holes, which is derived from the simple Bernoulli equation combined with the
continuity equation, can be expressed as follow:
(
m h =Cd A h 2ρ ( P1 -p j ) )
0.5
(2.3)
where P1 = total pressure upstream of the hole

pj = static pressure downstream of the hole
19
________________________________________________________________________
mh = massflow through hole feature
Cd = hole feature discharge coefficient
Ah = total through flow area of specific hole set
ρ = air density
As described by Adkins and Gueroui (1986), the flow on the outside of a combustion
chamber is normally parallel to the liner walls, and when the flow passes over a liner hole
the air is drawn into it due to the lower static pressure inside the chamber. As the flow is
drawn in, it is both deflected and accelerated and also contracts in cross-sectional area.
This contraction is so rapid that fluid dynamic forces cause it to continue contracting for
some distance after it has passed through the hole. A minimum cross-sectional area is
reached at some distance which is known as the vena contracta. The velocity and static
pressure are uniform over the cross-section due to no streamwise curvature of the jet. The
static pressure will be equal to that of the surrounding gases. The minimum jet area is
normally expressed as a fraction of the geometric hole area known as the discharge
coefficient (Cd), which is generally in the range of 0.6 for plain flush circular holes.
Figure 2.3 is a schematic illustration of this phenomenon.
Figure 2.3: Flow through liner hole (Van Niekerk and Morris, 2001)
20
________________________________________________________________________
Correlations, presented by Norster (1980), can be used to predict the discharge coefficient
through the various liner hole features. These correlations are, however, only valid for
plain and plunged holes. Therefore, initial discharge coefficients of 0.8 can be assumed
for the cooling devices that are aligned with the flow, as described by Dodds and Bahr
(1990).
The correlations for plain and plunged holes are defined by (Norster, 1980):
Plain holes:
1.25(K-1)
Cd = (2.4)
⎡⎣ 4K 2 -K(2-α)2 ⎤⎦
Plunged holes:
1.65(K-1)
Cd = (2.5)
⎡ 4K 2 -K ( 2-α )2 ⎤
⎣ ⎦
In the above-mentioned equations, the hole loss coefficient, K, and the hole bleed
ratio, α , are presented by the following expressions (Norster, 1980):
K=1+0.64 {2ϕ 2 + ⎡⎣ 4ϕ 4 +1.56ϕ 2 (4α-α 2 ) ⎤⎦ } (2.6)
where
.
mh
Mass flow ratio (bleed ratio): α= .
(2.7a)
man
A h,geom
Hole area ratio: Ar = (2.7b)
A an
α
Mass flow ratio/area ratio: ϕ = (2.7c)
Ar
21
________________________________________________________________________
The jet flow angle for plain and plunged holes can also be expressed as a function of the
loss coefficient, K, respectively (Norster, 1980):
Plain holes:
1 (K-1)
sinθ =
1.6Cd K (2.8)
Plunged holes:
1 (K-1)
sinθ = (2.9)
1.2Cd K
With the use of the above-mentioned equations and correlations, the following procedure
can be used to predict pressure drop and flow distributions through liner hole features:
• Manipulate equation (2.3) so that it describe the pressure as a function of the mass
flow rate
• Guess an initial mass flow rate through all the liner holes
• With the initial guess, use the following procedure to calculate the discharge
coefficients (Odgers and Kretchemer, 1980a):
step 1: Calculate α (mass flow ratio) Eq. 2.7 (a) from initial
mass flow
step 2: Calculate Ar (hole area ratio) Eq. 2.7 (b)
step 3: Calculate ϕ Eq. 2.7 (c)
step 4: Calculate K using Eq. 2.6
step 5: Insert K into either Eq. 2.4 or Eq. 2.5 to determine the
appropriate discharge coefficient
• Calculate the area-weighted pressure drop across the liner

• Adapt individual mass flow distributions by correlating the pressure drop
22
________________________________________________________________________
• Restart iteration with new mass flow distributions until pressure drop converges
(assuming that the pressure drop across all the features are the same, therefore
ignoring frictional losses)
The prediction of flow distributions is followed by the prediction of the adiabatic flame
temperature which is a function of the local air/fuel ratio. To facilitate the prediction, the
combustion chamber is divided into a recirculation, primary, secondary and dilution
zones in which the airflow can be assumed for each zone. Following usual practice, as
described by Kretchmer and Odgers (1978), the flow within the recirculation zone
include the sum of all the flow admitted through (a) the swirler, (b) any additional air
admitted through the flare, (c) 2/3 of the air admitted through the first row of holes in the
primary zone, and (d) 1/3 of the air admitted through the second row of primary holes.
None of the air used for film cooling of the primary zone wall is assumed to enter the
recirculation. If there is only one row of holes for primary air admission, then half the air
admitted is assumed to take part in the recirculation. Skidmore (1986) assumed for the
T56 combustion chamber that the primary zone include one third of the air entering
through the first hole set. The remaining air flow then forms part of the air flow for the
secondary zone. In the same way, the secondary zone include half of the air entering
through the third hole set and one third of the air entering through the third cooling
device with the remaining air flow becoming part of the dilution zone. The latter
assumption was utilised during this study.
Chemical reaction rates are generally controlled by the flame temperature which is
usually assumed to be the adiabatic temperature. The adiabatic flame temperature would
be the maximum temperature that can be achieved for given reactants because any heat
transfer from the reacting substance and/or any incomplete combustion would tend to
lower the temperature of the products. The adiabatic temperatures are, however, rarely
achieved due to heat losses from the flame by radiation and convection. Nonetheless, the
adiabatic flame temperature plays a significant role in determining combustion
efficiencies and in heat transfer calculations. However, before the adiabatic flame
23
________________________________________________________________________
temperature can be predicted, some definitions such as stoichiometric air/fuel ratios and
equivalence ratios should be discussed.
Complete combustion of hydrocarbon fuels can be achieved with the proper amount of
air to completely convert the fuel to carbon dioxide and water vapour. Stoichiometric
combustion refers to this complete combustion process. Stoichiometric mixtures contain
sufficient oxygen for complete combustion; therefore, operating at stoichiometric air/fuel
ratios will release all the latent heat of combustion of the fuel (Lefebvre, 1998). The
typical stoichiometric air/fuel ratio used for kerosene is 14.74.
A convenient way of comparing different fuels in terms of their mixing strength is to

make use of the equivalence ratio ( φ ). The equivalence ratio is defined as the actual
fuel/air ratio divided by the stoichiometric fuel/air ratio. It can also be defined as the
stoichiometric air/fuel ratio divided by the actual air/fuel ratio. Therefore, when φ = 1 it
denotes a stoichiometric mixture, when φ < 1 it indicates a fuel-lean mixture, while φ > 1
indicates a fuel-rich mixture.
( F/A )a ( A/F )s
φ= or φ = (2.10)
( F/A )s ( A/F )a
Where “s” and “a” denote stoichiometric and actual mixtures respectively.
The adiabatic flame temperature is influenced by the fuel/air ratio, inlet temperature and
pressure. Since the total amount of air in each of the four combustion zones can be
predicted as discussed in the previous paragraph, the air/fuel ratio and hence the
equivalence ratio in each zone can be predicted. Temperature rise curves for kerosene,
depicted in figure 2.4, can then be utilised to predict the temperature rise ( ΔT ) for a
specific combustion zone. These curves present the undissociated temperature rise due to
combustion at a certain inlet temperature and pressure as a function of the equivalence
ratio. The adiabatic flame temperature can subsequently be obtained using the expression
defined in equation (2.11).
24
________________________________________________________________________
Figure 2.4: Temperature rise curves (Lefebvre, 1998)
Tg =T3 +ΔT (2.11)
With the intention of predicting gas temperatures more accurately, combustion efficiency
has to be accounted for. Gosselin et al. (1999), as well as Odgers and Kretchemer
(1980a), presented empirical correlations for the prediction of combustion efficiency in
the re-circulation, primary, secondary, and dilution zone which in turn predicts the mean
temperatures in the specific zone. The following correlation is used to predict combustion
efficiency in the recirculation zone:
ηRZ =0.56+0.44tanh ⎡⎣1.548×10-3 ( T3 +108lnP3 -1863) ⎤⎦ (2.12)
where,
T3 = inlet temperature [K]
P3 = inlet pressure [Pa]
25
________________________________________________________________________
The temperature distribution in the recirculation zone is considered to be linear between
T3 at the inlet of the injector face and TRZ at the end of the recirculation zone. The
temperature will vary linearly between TRZ and TPZ for the remainder of the primary
zone. Combustion efficiency for the primary zone is presented as follows:
ηPZ =0.71+0.29tanh ⎡⎣1.548×10-3 ( T3 +108lnP3 -1863) ⎤⎦ (2.13)
The correlations presented by Gosselin et al. (1999) for the secondary and dilution zones
indicated lower values of combustion efficiency than that of the primary zone. For this
reason, it was assumed that combustion was completed in the primary combustion zone
and consequently a combustion efficiency of 90 percent was assumed for the secondary
and dilution zones.
Since the combustion efficiencies in the primary zone can be accounted for using
equation (2.13), the mean gas temperature in the zone can be predicted. For the maximum
temperature in the recirculation zone,
Tout,RZ =T3 +ηRZ ΔT (2.14)
According to Odgers and Kretchemer (1980b), this zone is only partially stirred and this
maximum temperature will exist locally only. Therefore, a mean outlet temperature is
assumed,
1 2 (2.15)
Tout,RZ = T3 + TRZ
3 3
The temperatures in the remaining zones can be calculated by assuming a linear

temperature distribution as follows:
Tout,zone =T3 +ηzone ΔTzone (2.16)
26
________________________________________________________________________
2.4 Heat transfer process
Peak gas temperatures in the primary zone of a combustion chamber exceed 2000 K.
Protection of the combustor liner against these high gas temperatures is of great
importance to ensure durability and structural integrity. Combustion chambers are
typically fabricated from a nickel- or cobalt-based alloy such as Hastelloy-X or HS188.
These high-temperature alloys are good for long-term operations at temperatures of up to
1150 K. At temperatures higher than 1150 K, the strength of these materials decreases to
unacceptable levels. At approximately a temperature of 1400 K, rapid oxidation starts
occurring and the liner material starts melting at a temperature range between 1500–
1750 K. The increasing operating conditions of modern-day gas turbine engines, to
improve overall engine efficiency and specific fuel consumption and to reduce emission
levels such as NOx, influence the effective use of film-cooling devices. This leads to the
research and development of more advanced cooling devices and materials that can
withstand higher temperatures (Dodds and Bahr, 1990; Rizk, 1994; Gosselin et al., 1999).
During the present study, the material utilised on the T56 combustion chamber was
analysed with an electron microscope situated at the University of Pretoria, South Africa.
The data obtained from the analysis showed close agreement to Hastelloy-X, and for the
purpose of the study the material properties of Hastelloy-X were used. The data obtained
from the analysis is presented in Appendix A.
A need therefore exists to predict liner wall temperatures prior to more expensive
experimental tests, by conducting a one-dimensional heat balance along the liner wall.
The heat transfer process consists of three modes of heat transfer, namely convection,
radiation, and conduction. The inner liner wall is heated by convection and radiation from
the hot combustion gases and cooled on the outside by the annulus air flow through
convection and radiation from the outer liner surface to the casing wall. Calculations are
conducted for equilibrium conditions where both the internal and external heat fluxes are
equal. Heat loss along the liner wall through conduction is assumed to be very small and
may be neglected. Figure 2.5 is a schematic illustration of such a heat transfer process.
27
________________________________________________________________________
Figure 2.5: Heat transfer process
When steady-state conditions are assumed, the heat flux into a wall segment must equal
the heat loss from the same segment. Considering all three heat transfer modes, the
steady-state heat transfer process can be described by the following expression where the
internal and external liner areas are Aw1 and Aw2 respectively.
(R1 +C1 )A w1 =(R 2 +C2 )A w2 =K1-2 A wm (2.17)
where
R1 = Internal radiation from the hot gas to the liner wall
C1 = Internal convection of the hot gas to the liner wall
R2 = External radiation from the outer liner area to the casing internal wall
C2 = Convection on the external surface of the combustor in the annulus
K1-2 = Conduction through the liner wall
Aw1 = Liner internal heat transfer area
Aw2 = Liner outer heat transfer area
(A w1 +A w2 )
Awm = Mean liner heat transfer area,
2
28
________________________________________________________________________
Gas radiation can be considered to have two components, namely non-luminous and
luminous radiation. Non-luminous radiation originates from certain gases, especially
from carbon dioxide and water vapour. Luminous radiation on the other hand, depends on
the number and size of the solid particles, mainly soot in the flame. The chemistry and
soot formation, atomisation quality, fuel distribution in the combustion zone, and fuel-air
mixing are all effected by the pressure which in turn influences the emissivity of
luminous gases. Increasing the inlet temperature and liner size, owing to a larger beam
length, will noticeably increase the emissivity and radiation. The beam length (lb) is
defined as the radius of gas hemisphere that will radiate to a unit area, at the centre of its
base, the same amount as the average radiation from the actual gas mass. Due to the
complexity in estimating the luminous emissivity, a luminosity factor (Lu) is introduced
into the empirical expression for a non-luminous flame, assuming an isothermal gas
temperature Tg which is bounded by walls at a temperature Tw (Lefebvre, 1998; Rizk,
1994; Lienhard and Lienhard, 2003).
The gas emissivity for luminous gases can be predicted as follows:
ε g =1-e(
-290P3L u (qlb )0.5Tg-1.5 )
(2.18)
where
P = gas pressure, kPa
Tg = gas temperature, K
lb = beam length, m
q = fuel/air ratio by mass
The size and shape of the gas volume determine the beam length ( l b ) which can be
expressed, for most practical purposes, with sufficient accuracy as follows:
volume (2.19)
l b =3.4
surface area
29
________________________________________________________________________
The luminosity factor Lu depends mainly on parameters such as carbon/hydrogen mass
ratios and fuel hydrogen content. A number of correlations exist, but the correlation used
during the present study, and provided by Lefebvre (1998), is presented as follows:
⎛C ⎞
0.75 (2.20)
L u =7.53 ⎜ -5.5 ⎟
⎝H ⎠
The internal radiation flux (R1) from the hot gases into a wall segment is calculated from
the equation proposed by Lefebvre (1998) as follows:
R1 =0.5σ (1+ε w ) ε g Tg1.5 (Tg2.5 -Tw1

2.5
) (2.21)
Assuming a heat transfer relation for a straight pipe, internal convection (C1) between the
gas and the liner wall is given by (Rizk, 1994):
0.8
⎛ . ⎞
k g ⎜ mg
C1 =0.020 0.2
D h ⎜⎜ A Lμ g
⎟
⎟⎟ ( T -T )
g w1 (2.22)
⎝ ⎠
where Dh is defined as the hydraulic diameter,
4A (2.23)
Dh =
Pw
Some difficulties occur when the latter equation is applied to the primary combustion
zone. Due to the flow reversal region in the primary zone, the direction of flow
corresponds, to that of the assumed pipe analogy, only in the region adjacent to the wall.
Another complication can occur when a swirler is used; the local gas velocity close to the
wall will be greater than the downstream component. The question as to whether or not
the bulk gas temperature Tg is appropriate for temperature predictions can also be asked.
To account for this, the value of the constant in equation (2.22) is reduced from 0.020 to
0.017, for the primary zone calculations as described by Lefebvre (1998). Crocker et al.
(1999) presented alternative values of 0.046 and 0.04 for calculations in the primary zone
30
________________________________________________________________________
and remainder of the combustion chamber respectively. However, the former results have
shown to provide better comparative results during this study.
Although heat conduction through the liner wall is considered insignificant it can be
expressed by:
kw
K1-2 = ( Tw1 -Tw2 ) (2.24)
tw
Heat is removed from the combustor liner through radiation to the outer casing and
convection to the annulus air. Using typical values of emissivity for both the casing and
liner walls, the external radiation (R2) can be calculated as follows (Lefebvre 1998):
σ ( Tw2
4
-Tc4 )
R 2Aw = (2.25)
(1-ε w ) + 1
+
(1-ε c )
εw Aw A w Fwc ε c A c
where
Aw = surface area of liner wall
Ac = surface area of casing
Fwc = geometric shape factor between liner and casing
The heat transfer due to radiation is relatively small when compared to the external
convective heat transfer. Its impact on the heat transfer increases with an increase in liner
temperature, but can often be neglected at low temperatures. The geometric shape factor
is also assumed to be unity due to radiation across a long annular space. The net radiation
heat transfer from the liner then reduces to:
σε w ε c ( Tw2
4
-T34 )
R2 = (2.26)
⎛A ⎞
ε c +ε w (1-ε c ) ⎜ w ⎟
⎝ Ac ⎠
31
________________________________________________________________________
The correlation for external convection (C2) is similar to equation (2.22) except that the
properties of the annulus air are used and they are based on the hydraulic diameter of the
annulus.
0.8
⎛ . ⎞
k a ⎜ m an
C2 =0.020 0.2 ⎟ ( Tw2 -T3 ) (2.27)
Dan ⎜ A an μ a ⎟
⎝ ⎠
2.5 Film-cooling devices

A film-cooling device provides a protective film of cooling air between the liner wall and
the hot combustion gases by injecting the air along the inner surface of the liner.
Generally, combustion chambers are designed to incorporate a number of cooling slots
along the length of a combustor liner to overcome the gradual destruction of the film due
to the turbulent mixing with the combustion gases (Rizk, 1994).
Typical film-cooling devices used on conventional combustion chambers include splash-

cooling rings, wiggle strips, stacked rings, machined rings, z-rings, and rolled rings.
These devices are spaced at intervals along the axial direction of the combustion liner to
provide a film-cooled blanket on the inner surface of the combustor liner. Splash-cooling
rings, wiggle strips, stacked rings, and machined rings are fabricated from sheet metal.
The splash-cooling device utilises the static pressure drop across the liner to provide the
desired film cooling. The air jets impinge on a deflector which is attached to the inner
surface of the liner. As the cooling air impinges on the deflector, the air is directed in a
direction parallel to the combustor wall, thus forming a uniform film-cooled blanket. If
the static pressure drop across the liner is too low, a cooling device that utilises the total
pressure drop needs to be used. Wiggle strips and machined rings are examples of
cooling devices that utilise the total pressure drop across the liner. The disadvantage of
utilising the total pressure drop across the combustor liner is the variation of annulus air
velocity due to the availability of the air flow and this may influence the cooling air flow
in downstream locations. The wiggle strip configuration employs a corrugated metal strip
to form a film-cooling slot and provides a stiff structure. The cooling flow area is very
sensitive to variation in material thickness and ring diameter. Due to this it is more
32
________________________________________________________________________
difficult to control the film-cooling air flow through a wiggle strip compared to machined
rings. Machined rings can be constructed of a single piece of metal or by welding several
rings together. A number of orifices, spaced equally apart, are machined into the ring to
provide the required cooling flow. Stacked rings are similar to machined rings but differ
in the construction method (Dodds and Bahr, 1990). Typical film-cooling devices are
depicted in figure 2.6.
Figure 2.6: Film-cooling devices (Dodds and Bahr, 1990)
Another method of protecting the combustor liner is to apply a ceramic coating to the
inside to insulate it from the hot combustion gases. Ceramic tiles are used in industrial
gas turbine engines. The heat transfer coefficient on the cool side can be increased by
using forced convection or by adding fins on the backside. The latter method has the
disadvantage of additional weight and difficulty of manufacturing. An alternative
approach is to merely roughen the backside surface, which will increase the convective
heat transfer coefficient. However, this will also increase the pressure drop across the
combustor liner (Lefebvre, 1998; Dodds and Bahr, 1990).
33
________________________________________________________________________
Ballal and Lefebvre (1986) derived the following expression for film-cooling
effectiveness based on turbulent boundary layer models by utilising available
experimental data:
0.15
0.65
⎛ μa ⎞ ⎛x⎞
-0.2
⎛t⎞
-0.2
ηf =1.10m ⎜⎜ ⎟⎟ ⎜ ⎟ ⎜ ⎟ (2.28)
⎝ μg ⎠ ⎝s⎠ ⎝s⎠
( ρV )a
where m is defined as the mass velocity ratio , μ is the dynamic viscosity, s is the
( ρV )g
cooling slot gap, t is the slot thickness, and x is the distance downstream of the slot. The
subscripts “a” and “g” denote the air and gas streams respectively. The film-cooling
effectiveness is defined as:
Tg -Tw,ad
ηf = (2.29)
Tg -Ta
where
Tg = Predicted gas temperature [K]
Ta = Cooling air temperature [K]
Tw,ad = Gas temperature at the combustor liner wall [K]
A value close to unity for the effectiveness indicates that the wall temperature (Tw,ad)
approaches that of the cooling air temperature (Ta). Equation (2.28) is valid for values of
m ranging between 0.5 and 1.3. For higher values, the first term in equation (2.28) is
replaced by a constant equal to 1.28. To illustrate these variables, figure 2.7 depicts a
typical film- cooling device schematically.
34
________________________________________________________________________
Figure 2.7: Schematic depletion of a film-cooling process
2.6 Swirler design

Turbulent flow in gas turbine combustors, especially in the primary zone, is of great
importance to enhance the mixing of the fuel and air resulting in efficient and complete
combustion. Recirculating flow is generally achieved when the air flow through the
primary jets impinge on one another, forcing some of the air in the upstream direction
and some in the downstream direction. In some cases the jet penetration is so poor that
additional methods must be utilised to enhance the recirculating flow. One method is to
make use of an air swirler. Air is introduced axially into the swirler, which in turn forces
the flow to spiral due to the vanes that are positioned at a fixed predetermined angle.
Such devices are normally designed using empirical correlations as outlined by Lefebvre
(1998). A number of studies have been conducted to determine the effect of a swirler. In
addition to the normal axial swirlers, co-swirlers a well as counter-swirlers were
investigated by a number of authors, including Guoqiang et al. (2004), Gupta and Lewis
35
________________________________________________________________________
(1998), Micklow et al. (1993), and Cia et al. (2002). These two devices consist of two
concentrical swirlers rotating the air flow in the same as well as opposite directions
respectively. Parametric studies were performed in some of the cases to investigate the
influence of flow splits through the swirler on the size of the recirculation zone and
temperature distribution while maintaining a constant equivalence ratio in the combustion
zone. These devices were mainly investigated in an attempt to reduce the emission levels.
It is highlighted by Guoqiang and Gutmarl (2004) that by increasing the recirculation
zone this will enhance the air flow, thus reducing the combustion temperature and hence
thermal NO formation. However, local oxygen levels will also increase as the swirl
increases, thus promoting NO formation.
According to Micklow et al. (1993), the use of an airblast fuel injector will produce a
finer fuel spray, and with the use of a swirler thorough mixing can be achieved, thus
reducing soot and NOx emission levels. Also, such a process will result in a flame of low
luminosity and soot, which may result in cooler liner wall temperatures.
According to Gupta et al. (1984), the induced swirl has an impact on jet growth, flame
size, shape, stability and combustion intensity. The amount of swirl that is being induced,
can be characterised by the dimensional number known as the swirl number (SN). For
axial swirlers, the swirl number can be expressed as a function of the vane angle and is
described by Mattingly et al. (2002) as follows:
Gφ
SN = (2.30)
G 'x rt
Where G φ and G 'x , are the axial fluxes of tangential and axial momentum respectively.
r r
t t
2
Gφ = ∫ ( Wr )ρU ( 2πr ) dr= ( Utanθ v ) U2πρ ∫ r 2 dr= πρU 2 tan θ v ⎡⎣ rt3 -rhub
3
⎤⎦ (2.31)
rhub rhub
3
36
________________________________________________________________________
and
rt rt
G 'x = ∫ ( U )ρU ( 2ρr ) dr= ( U ) 2πρU ∫ rdr=πρU 2 ⎡⎣ rt2 -rhub

2
⎤⎦ (2.32)
rhub rhub
The swirl number is therefore defined as:

⎡ ⎛ r ⎞3 ⎤
⎢ 1- ⎜ hub ⎟ ⎥
2 ⎢ r ⎥
SN = tanθ v ⎢ ⎝ t ⎠2 ⎥ (2.33)
3
⎢1- ⎛ rhub ⎞ ⎥
⎢⎣ ⎜⎝ rt ⎟⎠ ⎥⎦
Strong swirl is characterised by a swirl number greater than or equal to 0.6 (Lefebvre,
1998; Dodds and Bahr, 1990). The recirculation zone diameter and length increase with
an increase in swirl number up to a swirl number of 1.5, thereafter the recirculation zone
diameter continues to increase while the length decreases. If a very strong and long
recirculation zone is induced, it may entrain some of the relatively cool secondary gases,
which may result in stability problems and low combustion efficiency. Due to flow
separation at the inlet to the vane passages, sufficient passage length should be provided
to allow the flow to reattach to ensure stable flow at the preferred swirl angle. A
parameter called the solidity provides an indication to whether or not an appropriate
passage length is available. The solidity is defined at the tip of the swirler vanes and
should be at least unity if sufficient passage length is desired. The solidity is defined as
(Dodds and Bahr, 1990):
Lv (2.34)
Solidity =
Sv
where Lv denotes the swirler vane length and Sv the pitch at the tip.
Lefebvre (1998) describes the mass flow through an axial swirler as a function of the
pressure loss across the swirler, liner area, vane angle, and the type of vanes used:
37
________________________________________________________________________
0.5
⎧ ⎫
⎪ ⎪
. ⎪⎪ 2ρ3ΔPsw ⎪⎪ (2.35)
m sw = ⎨ ⎬
⎪ ⎡⎛ secθ ⎞ 2 1 ⎤ ⎪
⎪ K sw ⎢⎜ A ⎟ - A 2 ⎥ ⎪
v
⎪⎩ ⎢⎣⎝ sw ⎠ L ⎥
⎦ ⎪⎭
Ksw is the blade loss coefficient which is defined by the type of vanes used, either curved
or straight vanes. The value of Ksw is 1.3 for straight vanes and 1.15 for curved vanes.
Curved vanes will result in a slightly better swirl but is more difficult to manufacture.
The flow area of an axial swirler can be calculated as follows:
π 2 ⎛ t ⎞ (2.36)
Asw = ⎡⎣ Dsw -D2hub ⎤⎦ -0.5n ⎜ v ⎟ ( Dsw -Dhub )
4 ⎝ cosθ v ⎠
Where, n represents the number of blades, and tv represents the blade material thickness.
Other types of swirlers that are currently used in combustion chambers include radial
swirlers, counter-rotating swirlers, and converging swirlers as depicted in figure 2.8 and
2.9. The type of swirler used affects the recirculation zone shape and size. Counter-
rotating swirlers with a higher inner swirl number produce a much larger and stronger
recirculation zone than with a co-swirling outer swirler, as described by Dodds and Bahr
(1990). The swirl angle and flow rate through a swirler can be controlled by using a
converging outer swirl cup. Each passage in this arrangement converges from inlet to exit
where the flow is measured at the exit.
38
________________________________________________________________________
Figure 2.8: Swirler passage shapes and vane types (Dodds and Bahr, 1990)
Figure 2.9: Axial and radial swirler (Dodds and Bahr, 1990)
39
________________________________________________________________________
2.7 1-D flow and temperature predictions
By implementing the previous formulations into a computer code, it is possible to predict
one-dimensional incompressible flow and temperature predictions across a combustor
liner. A computer code was thus generated as part of the study and is presented in
Appendix D. With this code it is possible to predict pressure losses across the combustion
chamber, flow distributions as well as temperature distributions across a can or can-
annular combustion system. The model geometry is divided into a number of stations
utilising the formulations described in a step-wise manner at each station.
The data in table 2.1 presents the predicted flow distributions and pressure drop across a
single T56 combustor unit along with the discharge coefficients that were used. The
discharge coefficients for the plain holes were predicted with equation (2.4) whereas the
discharge coefficients for the cooling devices were assumed due to the lack of
appropriate correlations. For the splash-cooling devices, a discharge coefficient of 0.6
was assumed and since the wiggle strip devices are aligned into the flow, utilising the
total pressure loss, a discharge coefficient of 0.8 was assumed. The data in table 2.1 is
described in sequence when moving from the dome, downstream towards the combustor
exit. The hole layout of the combustion chamber is depicted in figure 2.10.
Fig. 2.10: Description of combustor hole layout
40
________________________________________________________________________
Table 2.1: Predicted vs. experimental mass flow distributions
Predicted Experimental Predicted
Hole type Mass fraction [%] Mass fraction [%] Cd [ - ]
1 Injector shroud 0.542 0.74 0.615
2 Dome splash strips 4.639 6.14 0.6
3 Wiggle strip 1 12.78 15.63 0.8
4 Primary hole-set 3.013 4.82 0.582
5 Wiggle strip 2 12.787 10.69 0.8
6 Secondary hole-set 1 2.263 3.51 0.585
7 Splash strip 1 0.876 0.19 0.6
8 Wiggle strip 3 12.78 11.07 0.8
10 Splash strips 2 1.751 0.83 0.6
11 Wiggle strip 4 12.774 9.59 0.8
12 Wiggle strip 5 12.761 9.87 0.8
13 Dilution hole-set 9.511 10.8 0.588
14 Dilution splash strips 2.608 2.62 0.6
15 Single dilution hole 1 3.833 4.36 0.597
Pressure loss 2.16%
Experimental mass flow distributions were obtained from Van Niekerk and Morris
(2001) for a single T56 combustor unit. The data was obtained from isothermal
experimental tests at atmospheric conditions. When comparing the predicted mass flow
predictions with the experimental mass flow distributions, it can be seen that although
there are discrepancies they follow the same trend as shown in figure 2.11. The
discrepancy is due to the inappropriate discharge coefficients assumed for some of the
hole features during the predictions, especially for the cooling devices.
41
________________________________________________________________________
Figure 2.11: Predicted and experimental mass flow comparison
Figure 2.12 depicts the one-dimensional temperature distribution along the combustor
liner. The predictions make use of the correlations previously discussed. The effect of the
cooling devices is clearly depicted in figure 2.12 by the sudden reduction of wall
temperature. Since the effect of the film cooling devices are modelled with film-cooling
efficiency, therefore assuming 100 percent efficiency at the point of injection, a sudden
drop in temperature is evident. The corresponding liner position of the film-cooling
devices is illustrated by the dotted red lines on the temperature graph. The red line in
figure 2.12 depicts the predicted gas temperature distribution along the combustor liner
while the blue line depicts the average liner temperature distribution.
The drawback of the one-dimensional empirical code is the following:

• The effect of swirler flow cannot be predicted
• The flow is considered to be incompressible
42
________________________________________________________________________
Figure 2.12: One-dimensional temperature distribution along combustor liner
Although the analysis cannot predict the three-dimensional influence of a swirler, the
mass flow through a swirler can be predicted. When modifications to a combustor unit
are considered, it is important to maintain the same overall pressure loss as this
influences the engine performance. A swirler needs to be designed with a similar overall
pressure loss, and for this reason accurate pressure drop predictions are important.
2.8 Summary
This chapter started with a general overview of typical gas turbine combustor features
followed by a description of the incompressible empirical formulas to predict mass flow
distributions and pressure losses across combustor features. These correlations are used
43
________________________________________________________________________
within the one-dimensional program presented in Appendix D. In addition, heat transfer
correlations are added to predict combustor liner temperature distributions. Heat transfer
is necessary to be included in such flow calculations, since temperature will influence the
fluid pressure and density, hence mass flow distributions.
44
CHAPTER 3 1-D network approach
________________________________________________________________________
CHAPTER 3 –
1-D NETWORK APPROACH
3.1 Preamble
This chapter describes the network model and its governing equations that were utilised
during the study to predict flow distributions and temperature distributions across the
combustion chamber. The network approach has the ability to account for incompressible
as well as compressible flow effects. These network models have proved to predict flow
distributions and pressure losses with reasonable accuracy across combustion chambers,
which is the reason why these models are generally used during preliminary design
studies.
3.2 1- D network solver

Although combustor design has evolved since the development of numerical methods, it
is still a tedious task requiring several numerical iterations, followed by experimental
verification. Due to its rapid execution time, one-dimensional models are still preferred
during the initial design phase as they are ideal for optimisation studies prior to more
expensive and time-consuming three-dimensional analysis. As part of this investigation, a
one-dimensional empirical model was developed to predict liner temperatures and flow
distributions for the current combustor being considered. This model was described in the
previous chapter. However, due to the limitations of using empirical correlations and
assuming incompressible flow, it was decided to combine the empirical model with a
one-dimensional network solver, FLOWNET. Network models employ the governing
equations of mass, momentum and energy and are therefore suitable for both
45
________________________________________________________________________
compressible and incompressible flow, while parameters such as gas temperatures,
combustion efficiency and gas emissivity can be obtained from the empirical
formulations.
Figure 3.1 illustrates the network layout that was developed for a single combustor liner
to solve the isothermal flow distribution through the different features. Due to the
symmetry, only half of the combustor was modelled. The squares and circles, shown in
more detail for the primary and secondary zone in figure 3.2, denote the nodes and
elements respectively. The type of element that was used varied depending on the type of
geometrical feature modelled. The gas flow inside the combustion chamber and the air
flow in the annulus, were modelled with DG elements (duct with variable area). These
elements account for area changes in a duct as well as frictional losses and momentum
addition. The geometrical flow features, such as primary, secondary, and dilution holes,
as well as film-cooling devices were modelled with RD elements (restrictor with
discharge coefficient). These elements are used when the combustor feature has a known
discharge coefficient. A swirler, on the other hand, can be modelled with an RL element
(restrictor with loss coefficient) which is similar to the RD element but accounts for the
pressure drop through the feature contraction as well as accounting for loss coefficients.
Fig. 3.1: General flow network layout for the combustion chamber
46
________________________________________________________________________
Fig. 3.2: Primary and secondary zone network layout
The initial step in the computational scheme is to assume a pressure at all the nodes and
to calculate the corresponding density. The flow rate can then be calculated from the
initial pressure and density. Thereafter a correction scheme, presented by Greyvenstein
and Laurie (1994), is applied to the pressure, density, and flow rate to satisfy the
governing equations. The overall governing equations constitute the continuity equation,
which is applied at every node in the network, and the pressure drop-flow relationship
which is applied for every element. The steady-state continuity equation (from the
conservation of mass) can be expressed as (Greyvenstein and Laurie, 1994),
J
∑ ρ Q s = -d
j=1
ij ij ij i i=1,2,....,J (3.1)
47
________________________________________________________________________
where,
ρ = density
di = external mass flow into node i
Q = volume flow rate

s = is used to relate the flow direction in the network. If s > 0 the flow is in
one direction and when s < 0, the flow is in the opposite direction.
The pressure drop-flow relationship in each element is typically defined using the Darcy-
Weisbach, Hazen-Williams, Manning or other exponential or empirically determined
correlations. The steady-state pressure drop-flow relationship (from the momentum
conservation) for any branch element of node i (figure 4.3) can be expressed as
(Greyvenstein and Laurie, 1994):
Δpij = pnij -pi = sijHijgijfij (3.2)
where g ij = g ij (ρij ) ,which is a function of ρij ,
f ij = ( )
f ij Qij , which is a function of Qij , and
H ij = Qij / Qij
Note that the mass and energy conservation are applied at the network nodes and
momentum conservation is applied across elements.
48
________________________________________________________________________
Fig. 3.3: Network arrangement (Greyvenstein and Laurie, 1994)
Heat transfer was modelled with convective, radiative, and conductive heat transfer
elements. The heat transfer process is described between two nodes where enthalpy is
balanced within each node, thereby satisfying the energy conservation equation. In the
absence of a film-cooling device, the heat transfer from the combustion gases to the
combustion liner was modelled with convective and radiative elements. These elements
are connected to a conductive element which accounts for the conductive heat transfer
through the combustor liner. External heat transfer is modelled by the connection of a
special heat transfer (HT) element to the conductive element where, for simplicity, a
constant heat transfer coefficient for both convection and radiation is specified. Axial
conduction along the combustor liner was accounted for with a primitive heat transfer
(PCHT) element. Such a process is illustrated in figure 3.4.
49
________________________________________________________________________
Figure 3.4: Heat transfer process network diagram without a film-cooling device
In figure 3.4, only two heat transfer strings are shown due to the complexity of the model
whereas a number of strings are actually used to obtain an adequate resolution. Note in
figure 3.4 the heat transfer string, and flow distribution string are not connected due to
the use of the (HT) element. An HT element is used in order to apply similar boundary
conditions that will compare to its CFD counterpart. The filled nodes denote nodes that
have a fixed property assigned to them, in this case, a fixed temperature. Flame
temperatures and combustion efficiency were predicted with the use of the empirical
correlations presented in the previous chapter.
Modelling of radiation between the combustion gases and the combustor liner, and also
between the liner and inner casing was considered. Radiation could not be modelled
between the internal flow path and combustion liner due to the limitations of the network
model. For this reason, an additional node with fixed temperature, similar to its flow path
counterpart, was added to incorporate radiation from the combustion gases to the liner
wall (figure 3.4). The radiation element was connected between the additional node and
the node connected to the (CHT) element, which describes conduction through the liner
50
________________________________________________________________________
wall. Gas emissivities as well as heat transfer areas were defined within the radiation
element.
Modelling the effect of film-cooling air on the combustor liner temperature is generally
done with the aid of empirical correlations. These correlations make use of the film-
cooling efficiency, as described in the preceding chapter, to predict the temperature
increase of the film-cooling air as it flows along the length of the combustor (Lefebvre,
1998; Stuttaford, 1997). However, these correlations are not available in the network
model. Therefore, a heat transfer process was utilised from the hot combustion gases to
the film-cooling air in a series of heat transfer strings to model the temperature increase
of the air as it flows further downstream from the point of injection. It was assumed that
the film-cooling air will flow from the point of injection up to the following geometrical
feature. Figure 3.5 illustrates the network layout used to define the flow and heat transfer
process through a film-cooling device. A cooling device was modelled with RD and DG
elements to describe the flow through the cooling orifice and the flow path of the cooling
air respectively. For a specific film-cooling heat transfer string, the film-cooling air is
heated through convection from the combustion gases – denoted by element 1 in figure
3.5. The heat transfer of the film-cooling air to the combustion liner is thereafter
modelled with convection but with a heat transfer area similar to that of the cooling
device. This process is depicted by element 2 in figure 3.5. The remaining area with no
film-cooling device is heated by convection and radiation from the combustion gases,
shown as elements 3 and 4 in figure 3.5.
It was assumed that the influence of the deflector wall on the liner temperature was
negligible, due to it being dimensionally very small. In figure 3.5, only a single heat
transfer string is shown to describe the heat transfer layout. However, in modelling the
film cooling devices, several strings were used to obtain an adequate resolution.
51
________________________________________________________________________
Fig. 3.5: Heat transfer and flow layout for a film-cooling device
Figure 3.6 depicts the dome section of the original combustion chamber (base case),
consequently a swirler was not added in this particular model. The initial three heat
transfer strings describe the heat transfer to the dome section where no film-cooling
devices are present. The remaining heat transfer strings describe the heat transfer process
on the dome wall of the combustor when splash-cooling and wiggle strips are present.
The wiggle strip on the dome section was modelled with three heat transfer strings
whereas the wiggle strips along the remaining combustor wall were modelled with six
heat transfer strings individually.
52
________________________________________________________________________
Figure 3.6: Network model of combustor dome section
The green elements in figure 3.6 represent the heat transfer string from the combustion
gases to the combustion liner, whereas the dark blue elements represent the flow path
along the combustor liner. The nodes, with caption ”Tg”, depict the fixed gas
temperatures calculated using the empirical model.
Figure 3.7 presents the complete network layout of the original combustion chamber
(base case). To obtain comparable temperature predictions, four elements were inserted
between each hole feature, which in turn resulted in three heat transfer nodes between
each feature.
With reference to figure 2.9, the network layout of the T56 combustion chamber is shown
in figure 3.8. The final dilution splash-cooling strip, shown in figure 2.9, is however
described as dilution splash strips 1 and 2 respectively.
53
______________________________________________________________________________________________________________________
Section 1 Section 3
Section 2
Figure 3.7: Network layout of T56 combustion chamber
54
________________________________________________________________________
Dome splash Wiggle Wiggle Wiggle

cooling strip 1 strip 2 strip 2
Section 1
Injector Primary hole- Secondary hole-set 1

shroud set
Splash Wiggle Wiggle Splash Wiggle Wiggle

strip 1 strip 3 strip 3 strip 2 strip 4 strip 4
Section 2
Secondary Secondary
hole-set 1 hole-set 2
Wiggle Wiggle Dilution Dilution

strip 5 strip 5 splash strip 1 splash strip 2
Section 3
Dilution Single Dilution Single dilution

hole-set hole 1 hole 2
Figure 3.8: Network liner feature description
55
________________________________________________________________________
3.3 Prediction of discharge coefficients and jet angles
An important parameter that influences the mass flow distributions and pressure losses is
the discharge coefficient of each of the geometrical features. Correlations presented by
Norster (1980) and Adkins and Gueroui (1986) were utilised to predict the discharge
coefficients as well as jet angles, through plain and plunged holes, if inadequate
experimental data were available. These correlations are presented in the preceding
chapter. Jet angles are generally defined as boundary conditions to a numerical model
when coupled with a one-dimensional solver. It is, however, difficult to use these
correlations in the network model since it is based on a number of parameters that are not
available in the network model. An alternative correlation described by Adkins and
Gueroui (1986), provides the prediction of the jet angles as a function of the pressure
drop across a hole feature. The following formulation can therefore be used, with
reference to figure 3.9:
Figure 3.9: Illustration of jet angle
56
________________________________________________________________________
sinθ j =v nj /v j (3.3)
where,
v nj = jet velocity component normal to the liner wall
vj = resultant jet velocity
The velocity component normal to the liner can be expressed in terms of the static
pressure drop across the hole feature, which in turn can be utilised to predict the jet angle
at the vena contracta as shown in equation (3.4).
sinθ j = ( p -p ) / ( P -p )
i j 1 j
(3.4)
3.4 One-dimensional flow and temperature prediction

The data in table 3.1 presents the flow predictions through the various hole features as
well as the overall pressure loss that was obtained with the network analysis. The
predicted data is compared with the experimental isothermal data, obtained from Van
Niekerk and Morris (2001), for a single combustion unit at atmospheric conditions. The
data in table 3.1 is described in sequence when moving from the dome, downstream
towards the combustor exit, similar to that described in table 2.1.
57
________________________________________________________________________
Table 3.1: Comparison of predicted and experimental flow distributions
Predicted Experimental
Hole type Mass fraction [%] Mass fraction [%] Error %
1 Injector shroud 0.7 -- --
3 Wiggle strip 1 12.74 15.63 18.46
5 Wiggle strip 2 12.75 10.69 -19.22
7 Splash strip 1 0.87 0.19 -359.77
8 Wiggle strip 3 12.75 11.07 -15.13
10 Splash strips 2 1.75 0.83 -110.55
11 Wiggle strip 4 12.75 9.59 -32.9
12 Wiggle strip 5 12.75 9.87 -29.15
14 Dilution splash strips 2.62 2.62 -0.11
Pressure loss 2.24%
The prescribed discharge coefficients are the same coefficients as that prescribed in
chapter 2. A comparison between the mass flow distributions for the network model
(table 3.1) and those obtained in the empirical model (table 2.1) shows a close agreement.
The small differences can be attributed to compressibility effects that were accounted for
in the network model. Hence, it can be concluded that the flow can be treated as
incompressible with reasonable accuracy. However, when comparing the data with the
experimental flow distributions some discrepancies are apparent. These discrepancies are
mainly due to the difference in predicted and actual discharge coefficients. The discharge
coefficients assumed for the cooling devices can significantly influence the predictions,
as more than 50 percent of the air supplied is directed through these devices.
Furthermore, the discharge coefficients do not only influence the flow predictions but
also the overall pressure loss. It is important to maintain the overall pressure loss when
considering modifications as this parameter influences the overall engine performance. It
is also important to maintain similar equivalence ratios throughout the combustion
chamber to ensure that combustion stability is not affected.
58
________________________________________________________________________
Since the importance of predicting accurate discharge coefficients is clear, and the
purpose of the present study is to investigate possible modifications to solve practical
problems, it was decided to calibrate the network model using the experimental flow
distributions. This was done by varying the discharge coefficients in the network solver
until the experimental flow distributions for each device were obtained within an error of
approximately one percent. Table 3.2 displays the flow distributions and pressure loss
predictions that were obtained in this manner.
Table 3.2: Comparison of calibration approach with experimental data
Calibration
approach Experimental
Hole type Mass fraction [%] Mass fraction [%] Error %
1 Injector shroud 0.74 -- --
3 Wiggle strip 1 15.51 15.63 0.79
5 Wiggle strip 2 10.61 10.69 0.72
7 Splash strip 1 0.19 0.19 -0.17
8 Wiggle strip 3 10.99 11.07 0.68
10 Splash strips 2 0.82 0.83 1.53
11 Wiggle strip 4 9.52 9.59 0.77
12 Wiggle strip 5 9.81 9.87 0.63
14 Dilution splash strips 2.61 2.62 0.44
Pressure loss 5.10%
A comparison between the empirically determined discharge coefficients and that

obtained from the calibration method is shown in table 3.3. It is evident that the largest
discrepancies between the discharge coefficients occur for the cooling devices. This is
due to the assumptions that were made because of the lack of appropriate correlations.
The difference in pressure loss is also notable, and it is believed that the larger pressure
loss is more representative compared to typical design values. The data obtained from the
59
________________________________________________________________________
network analysis was further utilised as boundary conditions for a refined three-
dimensional model of the combustion chamber.
Table 3.3: Comparison of predicted discharge coefficients
Hole type Assumed Calibration

Cd method Cd
Injector shroud 0.8 0.57
Dome splash strips 0.6 0.532
Wiggle strip 1 0.8 0.655
Primary hole-set 0.583 0.624
Secondary hole-set 1 0.585 0.608
Splash strip 1 0.6 0.087
Secondary hole-set 2 0.586 0.551
Splash strips 2 0.6 0.19
Dilution hole-set 0.588 0.446
Dilution splash strips 0.6 0.4
Single dilution hole 1 0.596 0.452
Single dilution hole 2 0.6 0.549
3.5 1-D temperature predictions

The ability to predict combustor liner temperatures prior to more detailed analysis such as
CFD is of great importance to reduce development time and cost. The network approach
provides the ability to predict temperature distributions that are comparable to three-
dimensional CFD analysis predicting the same basic trend. Even when correlations for
film-cooling efficiencies are not available, a network model can model the actual heat
transfer between the film-cooling air and combustor liner reasonably well.
Figure 3.10 illustrates the predicted one-dimensional network temperature distribution

along the combustor liner. In the dome section of the combustion chamber (distance 0mm
~ 70mm), it is clear that the network model underpredict the wall temperature, since this
is the zone with the highest gas temperatures. This is mainly due to the use of the
predicted empirical gas temperature. The maximum temperature rise is calculated in each
combustion zone through the use of the empirical one-dimensional code and interpolated
60
________________________________________________________________________
between each zone. The temperature for the primary combustion zone will therefore be
interpolated between the combustor inlet temperature and the maximum combustion
temperature rise in the specific zone. Figure 3.11 depicts the predicted gas temperature.
Figure 3.10: Combustor temperature distribution
The gas temperature prediction, presented in figure 3.11, was also used during the
temperature predictions in Chapter 2.
61
________________________________________________________________________
Figure 3.11: Average empirical gas temperature
Comparing the data from figure 2.12 to that of figure 3.10, the trend in temperature
distribution along the combustion liner is very similar. The liner temperature predictions
in figure 2.11 made use of a film-cooling efficiency to model the effect of film-cooling
devices. Conversely, instead of making use of a film-cooling efficiency in figure 3.10, the
effect of the cooling device was modelled with a heat transfer process. In this sense, the
network approach is more versatile in being able to include any type of cooling device.
The network temperature predictions will be compared with CFD predictions in Chapter
5, for validation purposes.
3.6 Empirical swirler design

In addition to the empirical code described in the previous chapter, and the network code
described in the present, an empirical swirler code was developed as well. This code is
used to design a swirler empirically and to investigate its effect on the dome wall
temperatures. For such a design to be conducted, however, an appropriate combustor
pressure loss must be obtained. The network solver, along with the empirical code, can
consequently be used to obtain an appropriate pressure loss as described in section 3.4.
62
________________________________________________________________________
One of the requirements of a successful swirler design is to deliver the required mass
flow rate to the primary zone while at the same time maintaining a strong swirl. The mass
flow rate is dependent on the flow area of the swirler, which in turn is dependent on the
number of blades and the pressure drop across the device. The inner diameter of the
swirler is fixed by the injector diameter, therefore, the only parameter to determine is the
swirler outer diameter for specified flow rates and blade angles. Once this has been
determined, the corresponding swirl number can be calculated. As previously stated,
strong swirl is characterised by a swirl number greater than 0.6.
In the current study, it was decided to investigate five different axial straight vane swirler
designs as depicted in figure 3.12. The mass flow distributions and pressure loss obtained
from the network results were used and the five cases are presented in table 3.4 The
design procedure used can be outlined as:
Step 1: Define initial parameters such as the injector diameter, solidity at the tip,
vane, hub, and outer ring material thickness, percentage pressure drop
across the swirler, inlet total temperature and pressure, total mass flow rate
per combustor and swirler width. Along with the initial parameters a
matrix consisting of percentage mass flow rates through the swirler vs.
blade angle can be calculated.
Step 2: Calculate the flow area of the swirler using equation (2.35)
Step 3: Determine the swirler blade length with predefined swirler width and
blade angle
Step 4: Guess an initial swirler outer diameter
Step 5: By defining the solidity at the tip of the swirler, the number of blades can
be calculated
Step 6: Equation (2.36) should be used to calculate the new swirler flow area
Step 7: Step 5 – 6 is an iterative process adjusting the swirler outer diameter until
the swirler flow area calculated in step 2 and step 6 are equal
Sept 8: The swirl number can now be calculated using equation (2.33)
63
________________________________________________________________________
Figure 3.12: Axial straight vane swirler
The computer code written to perform the above procedure as well as the matrix table
generated, is presented in Appendix C.
The five cases shown in table 3.4 were subsequently investigated in a three-dimensional
numerical model. The influence of the swirler blade angle was investigated in the initial
three models while maintaining an airflow rate similar to that of the flow through the
splash cooling devices on the original combustion chamber. The subsequent two cases
(case 4 and case 5) investigated the effect of increasing the flow rate through the swirler
at a fixed blade angle of 40° while maintaining a constant pressure loss of 5.1 percent.
More detail regarding the swirler designs are presented in Chapter 5.
Table 3.4: Swirler design cases

Case 1 Case 2 Case 3 Case 4 Case 5
Pressure loss [%] 5.1 5.1 5.1 5.1 5.1
Flow split [%] 6.097 6.097 6.097 13.85 13.85
Swirler tip diameter [mm] 40.47 40.92 41.85 47.02 47.02
Blade angle [degrees] 33 40 50 40 40
Number of blades 11 10 8 11 11
Swirl number 0.61 0.78 1.1 0.74 0.74
64
________________________________________________________________________
3.7 Summary
The data from the one-dimensional incompressible empirical code, discussed in the
previous chapter, was successfully used as inputs to a network solver to predict flow
distribution data, pressure losses, and temperature distributions. The distinct advantage of
this approach is its rapid execution time while providing good approximations. The
results showed that discharge coefficients influenced the one-dimensional flow and
pressure drop predictions significantly. Obtaining correlations for the prediction of
discharge coefficients for any hole type is, however, a difficult task. The ability of the
network solver to predict the general temperature distribution along the combustor liner is
also shown, and will be compared with CFD data in the subsequent chapters.
65
CHAPTER 4 Numerical model
________________________________________________________________________
CHAPTER 4 –
NUMERICAL MODEL
4.1 Preamble
Due to the development of computer resources over the past few years, computational
fluid dynamics has become one of the primary tools to investigate the complex flow field
phenomena inside a gas turbine combustion chamber. These analyses are needed to
predict the influence of certain flow parameters on the exit temperature distribution and
wall temperatures. This chapter describes the numerical models and boundary conditions
that were utilised during the study.
4.2 Turbulence models

The equations governing the flow in a gas turbine combustion chamber are the transport
equations for mass, momentum, and energy. The main discussion in this section will be
turbulence and therefore only the former two transport equations will be described.
Turbulence can be described in short as unsteady flow with velocities and flow properties
varying in a random and chaotic way, with various length scales. The transport equation
for mass, assuming steady-state conditions, can be written as follows:
__ __
∂ ρ uj
=0 (4.1)
∂x j
The transport equations for momentum can be represented as:
66
________________________________________________________________________
∂ ∂ ∂p ∂ ⎛ ⎛ ∂u ∂u ⎞ ⎞
( ρu i ) + ( ρu i u j ) = − + ⎜⎜ μ ⎜⎜ j + i ⎟⎟ ⎟⎟ + ρg i + Fi (4.2)
∂t ∂x j ∂x i ∂x j ⎝ ⎝ ∂x i ∂x j ⎠ ⎠
The transport equation for momentum is also known as the Navier-Stokes equations. The
final term (Fi) is a source term defining the effect of body forces.
The above equations can be applied to laminar as well as turbulent flows. To solve
turbulence directly from the above equations will however be a difficult task. To solve all
length scales directly from the time-dependent Navier-Stokes equations are beyond
current-day computer capabilities and is therefore dependent on future computer
hardware development.
In the meantime, adequate information about the turbulence process can be supplied
when utilising the time-averaged properties of the flow.
As an example, figure 4.1 depicts typical measured velocities in turbulent flow. To obtain
__
mean properties of the flow, the velocity, u(t), can be decomposed into a mean value u ,
and the fluctuating component u’(t).
Figure 4.1: Typical point velocity measurement in turbulent flow
67
________________________________________________________________________
__
u= u +u(t)' (4.3)
Replacing ui in equation (4.2) with equation (4.3) yields the following expression known
as the Reynolds average navier-stokes equations (RANS).
⎛ ⎛ __ __ ⎞ ⎞
∂ ⎛ __ ⎞ ∂ ⎛ __ __ ____
⎞ ∂ρ ∂ ∂ u ∂ u
⎜ μ ⎜ j + i ⎟ ⎟ + ρg + __
ρ
⎜ i⎟u + ρ u
⎜ j i u +ρ u ' '
u
j i⎟ = − + Fi (4.4)
∂x i ∂x j ⎜⎜ ⎜⎜ ∂x i ∂x j ⎟⎟ ⎟⎟
i
∂t ⎝ ⎠ ∂x j ⎝ ⎠
⎝ ⎝ ⎠⎠
Comparing equation (4.2) and equation (4.4), it is clear that an additional term has been
added to the navier-stokes equation which introduces additional turbulence stresses
termed the Reynolds stresses. This term can be described as follows when the Boussinesq
hypothesis of μ t is assumed:
⎛ __ __
⎞
____
⎜ ∂ u ∂ u j ⎟ 2
−ρ u u = μ t
' ' i
+ − ρδ ijk (4.5)
i j
⎜⎜ ∂x j ∂x i ⎟⎟ 3
⎝ ⎠
where the turbulent viscosity, μ t , is defined using the turbulent kinetic energy, k, and its
dissipation rate, ε , as follows:

k2
μ t = Cμ ρ (4.6)
ε
In order to obtain adequate information about the turbulence process but without the need
to predict every eddy in the flow, the RANS equations represent transport equations for
the mean flow quantities only, with all the scales of the turbulence being modelled. The
approach of permitting a solution for only the mean flow variables greatly reduces the
computational effort. The turbulence models which are derived from RANS equations
include the standard k-ε model, RNG k-ε model, realizable k-ε model, standard k-ω
model, shear-stress transport (SST) k-ω model, and the Reynolds stress model (RSM).
One of the basic turbulence models is the standard k-ε model. This k-ε model is one of
the most widely used and validated models and presents excellent performance for many
industrial relevant flows. As the name states, it is a two-equation model, one for k and
68
________________________________________________________________________
one for ε which define the velocity and length scales respectively and are valid for fully
turbulent flows. The model, however, performs poorly in certain cases such as
unconfined flows, flows with large extra strains such as swirling flow, rotating flows, and
in fully developed flows in non-circular ducts (Biswas et al., 1997; Versteeg and
Malalasekera, 1995). McGuirk and Palma (1993) showed that although the standard k-ε
model did have some difficulty to predict the vortices created by the primary jets,
acceptable results in other regions of the combustor and the correct prediction of the main
features of the combustor flow were possible. Due to the poor performance of the model
in some cases, improvements were made to the standard k-ε model to enhance its ability
to predict swirling flow in some cases. The renormalisation group (RNG) k-ε model and
realizable k-ε model are two variants of the standard k-ε model. The RNG k-ε model has
more accuracy with low Reynolds number flows compared to the standard k-ε model and
also has enhanced accuracy for swirling flows. A relatively recent development is the
realizable k-ε model which differs from the standard k-ε model in two ways:
• The model contains a new formulation for the turbulent viscosity
• A new transport equation for the dissipation rate, ε, has been derived
An advantage of the realizable k-ε model is to provide improved performance of flows

involving rotation, boundary layers under strong adverse pressure gradients, separation,
and recirculation (Shih et al., 1995).
It should be noted that the above-mentioned models assume μ t to be isotropic which may
lead to inaccurate flow predictions. The transport equations for k and ε respectively are
given by:
∂ ∂ ∂ ⎡⎛ μ ⎞ ∂k ⎤
( ρk ) + ⎛⎜ ρ u i k ⎞⎟ = ⎢⎜ μ+ t
__
⎟ ⎥ + G k − ρε (4.7)
∂t ∂x i ⎝ ⎠ ∂x i ⎣⎝ σ k ⎠ ∂x i ⎦
∂ ∂ ∂ ⎡⎛ μ ⎞ ∂ε ⎤ ε2
( ρε ) + ⎛⎜ ρ u i ε ⎞⎟ = ⎢⎜ μ+ t
__
ε
⎟ ⎥ + C1ε G k -C 2ε ρ (4.8)
∂t ∂x i ⎝ ⎠ ∂x i ⎣⎝ σ ε ⎠ ∂x i ⎦ k k
69
________________________________________________________________________
where Gk is the generation of k:

____
∂u j
G k = -ρ u i' u 'j (4.9)
∂x i
The equations contain five constants; Cμ , σ k ,σ ε , C1ε ,C 2ε . The values of these constants
were determined by comprehensive data fitting for a wide range of turbulent flows and
recommended values are as follows:
Cμ =0.09, σ k =1, σ ε =1.3, C1ε =1.44, C 2ε =1.92 (4.10)
4.3 Combustion model

In addition to turbulence modelling, numerical models to predict the combustion process
should also be considered. Combustion is a process in which fuel reacts with an oxidant
to form products of combustion. The products are not usually formed within a single
chemical reaction but within a series of elementary reactions. When accounting for
combustion through a numerical model, the transport equations must be solved for each
species in addition to the flow equations. The flow field is in turn affected by changes in
temperature and density due to the species. This results in a large number of partial
differential equations (PDEs) to be solved if numerous chemical reactions are to be
considered. Models that consider so many reactions require a huge amount of computer
resources, so simple models that incorporate only a few reactions are often preferred in
numerical combustion procedures used in CFD. The simplest procedure is the simple
chemical reaction system (SCRS). Other modelling approaches to solve turbulent
combustion include the Eddy break-up and laminar flamelet models (Versteeg and
Malalasekera, 1995).
When the SCRS approach is used with the assumption of a one-step, infinitely fast
chemical reaction, a mixture fraction method can be used to determine species
concentrations. A non-dimensional variable f , called the mixture fraction, can be
defined. If the local value of f is 0, the mixture at a point contains only oxidant and if f
70
________________________________________________________________________
equals 1 it contains only fuel. Fluctuating temperatures in the SCRS approach are often
accounted for by incorporating a probability density function (PDF) to calculate mean
properties. In the PDF method, the average value of scalar variables are obtained by
weighting the instantaneous value with a probability density function for mixture
fraction f . The two probability functions that give rise to the best results are the
Gaussian and β - functions (Baron et al., 1994; Versteeg and Malalasekera, 1995;
Smiljanovski et al., 1999; Hoerzer et al., 2002).
A second combustion model is the Eddy break-up model where the rate of consumption
of fuel is specified as a function of local flow properties. The mixing-controlled rate of
k
reactions is expressed in terms of the turbulence time scale , where k is the turbulent
ε
kinetric energy and ε is the rate of dissipation of k. The model considers the dissipation
rates of fuel, oxygen and products, and takes the slowest rate as the reaction rate of fuel.
The Eddy break-up model results in rather good predictions and is simple to implement
but is dependent on the turbulence model used. The combustion model will therefore be
limited if the turbulence model fails to make accurate flow predictions (Versteeg and
Another combustion modelling approach is the laminar flamelet model. The model
allows the addition of experimental information to describe the relationship between the
mixture fraction, mass fraction and temperature. The data is acquired from measurements
in a laminar diffusion flame. A transport equation for the mixture fraction is solved and
the species mass fraction is deduced from laminar flamelet relationships (Versteeg and
The SCRS combustion model that was utilised during the current study shares the basic
assumption, similar to other turbulent combustion models, that the instantaneous scalar
values such as species concentration, temperature, and density are related to a conserved
scalar such as the mixture fraction. The average scalar values are often obtained by
integrating the product of the scalar profile and the probability density function (PDF)
71
________________________________________________________________________
over a mixture fraction space combined with an equilibrium non-adiabatic chemistry
model. A β-function was assumed for the PDF.
The mixture fraction ( f ) of an element can be obtained as follows:
Zi − Z i,o
f = (4.11)
Z i,f − Z i,o
where Zi denotes the elemental mass fraction of element i. The subscripts “o” and “f”
denote the oxidiser and fuel stream inlets respectively. With the assumption of chemical
equilibrium all thermo-chemical scalars, such as temperature, density, and species
fraction, are related to the mixture fraction.
The probability density function P( f ), describes the temporal fluctuation of f in

turbulent flow. Mean scalar values, depending on f , can then be calculated as follows
(Yun et al., 2005).
__ 1
φ = ∫ P( f )φi ( f )df (4.12)
0
__
where φ is the average scalar variable.
Since the shape of the PDF is unknown, it can be modelled as a mathematical function by
assuming its shape. The shape of the function depends exclusively on the mean mixture
_
__
fraction f and its variance f '2 .
The β-function describing the shape of the PDF is given as follows (Yun et al., 2005):
72
________________________________________________________________________
f α −1 (1 − f )
β −1
P( f ) = 1
(4.13)
f α −1 (1 − f )
β −1
∫
0
df
where
⎡ __ ⎛ __ ⎞ ⎤
__ ⎢
f ⎜1 − f ⎟ ⎥
α = f ⎢ ⎝ __ ⎠ − 1⎥ (4.14)
⎢ f '2 ⎥
⎢⎣ ⎥⎦
and
⎡ __ ⎛ __ ⎞ ⎤
f ⎜1 − f ⎟ ⎥
⎛ __ ⎞ ⎢⎢ ⎝ ⎠ − 1⎥
β = ⎜1 − f ⎟ __ (4.15)
⎝ ⎠⎢ f '2 ⎥
⎢⎣ ⎥⎦
_ __
The assumed PDF, P( f ), can therefore be computed as a function of f and f '2 and
used as a weighting function to determine mean values of density, temperature, and
species mass fraction.
A simple one-step non-premixed reaction scheme was utilised during the study, assuming
an infinitely fast reaction. Five basic species were considered in calculating the PDF
look-up tables on the properties of the mixture as a function of density, temperature, and
mass fraction of the species. The five species were; Jet A(g), O2, N2, CO2,H2O. The one-
step reaction scheme is presented as:
⎛ y⎞ y ⎛ y⎞
C x H y + ⎜ x+ ⎟ ( O 2 +nN 2 ) → xCO 2 + H 2 O+n ⎜ x+ ⎟ N 2 (4.16)
⎝ 4⎠ 2 ⎝ 2⎠
73
________________________________________________________________________
4.4 Fuel spray model
The fuel distribution used in the numerical model corresponds to the experimental
measurements obtained by Van Niekerk and Morris (2001). The fuel spray was modelled
as a discrete second phase and the droplets were tracked in a Lagrangian framework. The
interaction of the fuel spray with the gaseous mixture, i.e. the exchange of mass,
momentum and energy between the phases, was taken into account through the particle-
source-in-cell method. The spray cone was modelled as six discrete cones covering the
total included angle of 110º of the real spray. The spray angle of 110º corresponds with
the dual-stage pressure atomisers operating at maximum flow pressure. Equal amounts of
fuel are injected through each of the cones and circumferentially the spray was
discretised into 36 injection points for each cone. The fuel spray data is presented in
tables 4.1 and 4.2 respectively. The fuel that was used during the study was kerosene
(C12H24).
The experimentally determined drop size data that was presented was adapted for the
air/fuel ratio considered. Table 4.1 shows the 12 discretised sizes used in the CFD model.
Table 4.1: Discretised fuel spray data

Size # Mean droplet size Volume fraction Massflow Droplet size [mm]
in group [micron] [kg/s]
1 6.87 0.00226 1.78604E-05 0.00687
2 9.31 0.0058 4.58365E-05 0.00931
3 13.4 0.01669 0.000131899 0.0134
4 18.95 0.03867 0.000305603 0.01895
5 25.7 0.06937 0.000548221 0.0257
6 34.84 0.1066 0.000842444 0.03484
7 47.25 0.14285 0.001128922 0.04725
8 68 0.17171 0.001356998 0.068
9 96.16 0.17766 0.00140402 0.09616
10 130.4 0.14406 0.001138484 0.1304
11 176.8 0.08984 0.000709992 0.1768
12 239.7 0.03448 0.00027249 0.2397
Each cone consisted of equal amounts of fuel. Table 4.2 presents the boundary conditions
for each cone angle when the fuel spray was modelled.
74
________________________________________________________________________
Table 4.2: Fuel cone angles
Cone # Inlet velocity Half angle Fuel flow rate Drop size range
magnitude, [m/s] [degrees] [kg/s] [micron]
1 30 5 0.00790277 6.87 - 239.70
2 30 15 0.00790277 6.87 - 239.71
3 30 25 0.00790277 6.87 - 239.72
4 30 35 0.00790277 6.87 - 239.73
5 30 45 0.00790277 6.87 - 239.74
6 30 55 0.00790277 6.87 - 239.75
4.5 Grid generation

Generally, it is found that approximately two-thirds of the calendar time spent on a CFD
prediction is undertaken on preparing the geometry and obtaining an acceptable grid
(Sivaramakrishna et al., 2001).
For the present combustion chamber and the models being investigated, a single mesh
consisted of approximately 100,000 hexahedral and 900,000 tetrahedral cells
respectively. Several grid sizes, ranging between 500,000 and 1,500,000 computational
cells were solved to verify solution independency on the mesh size. A typical mesh,
depicted upstream of the dome section in figures 4.2 and 4.3, therefore contained
approximately 1,000,000 cells. An unstructured hybrid mesh was used with primarily
hexahedral cells within the core section and film-cooling devices whereas tetrahedral
cells were used close to the walls. The splash-cooling devices on the dome are modelled
as faces adjacent to the wall with a slot height of approximately 2 mm. The swirler inlets
were modelled as faces where unit directional vectors were defined. The normal
component (Vn) at the swirler inlet surface is computed by the relation:
Vn = cosθ v (4.17)
whereas, the tangential (swirling) component Vt, at the swirl plane is computed by the
relation:
Vt = sinθ v (4.18)
75
________________________________________________________________________
Figure 4.2: Grid on outer wall of T56 combustor
Figure 4.3: Grid on dome with modification of swirler
76
________________________________________________________________________
4.6 Boundary conditions
The boundary conditions utilised within the CFD models were obtained from the one-
dimensional models discussed in the previous chapters. The boundary conditions
primarily constitute the air and fuel inlets. The air inlet boundaries were prescribed with a
uniform mass flow rate along with an inlet temperature, presented in table 1.1. The mass
flow rates were obtained from the network results presented in table 3.2 and jet angles
calculated using equation 3.3. Table 4.3 provides a summary for the base case input
parameters at the various inlets. On the outlet plane, a pressure boundary condition was
used. On the outside of the liner wall, a constant heat transfer coefficient of 288
W
was prescribed, calculated from the average Reynolds number within the annulus.
m2K
The effect of radiation was investigated by using the P-1 radiation model, which is a non-
equilibrium diffusion type model, and the variation of the mixture absorption coefficient
was taken into account using the weighted-sum-of-gray-gases model. However, radiation
was not modelled on all models since it is computationally expensive. The grid
independence study and radiation effects on the liner wall temperature are described in
Appendix B.1 and Appendix B.2 respectively.
The data in table 4.3 is described in sequence when moving from the dome, downstream
towards the combustor exit for the base case. The data presents the mass flow splits
through the various hole features as well as the directional vectors that was defined for
the plain holes. The hole layout of the combustion chamber is depicted in figure 4.4.
77
________________________________________________________________________
Figure 4.4: Description of combustor hole layout
Table 4.3: Base case boundary conditions

Flow splits Splits to
Fluent Inlet Flow splits [kg/s] individual Directional vector components Hydraulic
designation [%] 2.37642 holes [kg/s] x y z diameter [mm]
Injector 0.7448178 0.0177 0.0177 normal to inlet boundary 7.4
shell swirl 1 0.08551 normal to inlet boundary 2.4
shell swirl 1+2 6.0969862 0.14489
wigg 1 bottom 0.18428 normal to inlet boundary 3.27
wigg 1 top 0.18428 normal to inlet boundary 3.27
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.96927 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole total 4.7853494 0.11372 6.828
wigg 2 total 10.607973 0.25209
shell3 splash 0.1880981 0.00447 0.00447 normal to inlet boundary 3.63
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027606667 -0.13922 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.0389175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0389175 0.410991 0.899943 0.145567
shell4 hole 3 0.0389175 -0.41099 0.899943 0.145567
shell4 hole 4 0.0389175 -0.64789 -0.7477 0.145567
shell 4 hole total 6.5506097 0.15567 11.254
shell 4 splash 1 0.00977 normal to inlet boundary 3.63
shell 4 splash total 0.8222452 0.01954
78
________________________________________________________________________
wigg 4 total 9.5189403 0.22621
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.063725 0.70639 -0.70639 0.045014
dilute hole 2 0.063725 0.70639 0.70639 0.045014
dilute hole 3 0.063725 -0.70639 0.70639 0.045014
dilute hole 4 0.063725 -0.70639 -0.70639 0.045014
dilute hole total 10.726218 0.2549 16
dilute hole 5 4.3321467 0.10295 0.10295 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2435344 0.07708 0.07708 -0.46946 -0.88293 0.007156 15.86
dilute splash 1.7315963 0.04115 0.04115 normal to inlet boundary 3.63
trans splash 0.8664293 0.02059 0.02059 normal to inlet boundary 3.63
100 2.37642 2.37642
A number of models were generated during the study. The initial two models are of the
original combustion chamber and one where a single set of splash-cooling devices are
blocked on the dome. Another five models were generated to conduct a parametric study
of the swirler and to investigate its influence on dome temperatures and exit temperature
distributions. The boundary conditions for all the models are presented in Appendix E.
4.7 Summary
The accuracy of CFD models depends mainly on the boundary conditions specified as
well as a sufficiently refined grid in areas of high gradients. To maintain adequate
processing time, the grid size was kept to a realistic size while obtaining good resolution.
A model size of 1,000,000 cells was therefore utilised throughout the study. Present-day
computer power is, however, on the increase which may results in more available
computer resources meaning that finer grids may be utilised in future studies.
79
CHAPTER 5 Numerical study
________________________________________________________________________
CHAPTER 5 –
NUMERICAL STUDY
5.1 Preamble
The literature survey has shown that CFD can be used to conduct analyses in combustion
chambers successfully. These analyses are, however, dependent on the physical models
used, as well as the grid size and grid quality which were described in the previous
chapter.
Numerically, five test cases with different swirler designs were investigated. Two
additional cases are investigated showing the original combustion chamber with and
without blocked splash-cooling devices, serving as the base cases. From the results, an
evaluation was made on the dome wall temperatures as well as the exit temperature
profile. The different test cases are discussed and possible reasons for insufficient dome
cooling are also provided. A direct comparison is also made between the experimental
and numerical exit temperature distributions.
5.2 Experimental data

Measured exit temperatures and isothermal flow distributions were obtained from
Skidmore (2004) and Van Niekerk and Morris (2001). Skidmore (2004) obtained
experimental exit temperature distributions at an overall air/fuel ratio of 49.75. Five
readings were taken in the radial direction and measured every 3.15° along the 60° exit
plane of a single T56 combustion chamber. The flow distribution data, from Van Niekerk
and Morris (2001), were obtained at atmospheric conditions. For the purpose of the study,
80
________________________________________________________________________
it was assumed that the flow distributions at atmospheric inlet conditions and operating
inlet conditions were similar.
It was assumed that the measured temperature was the static temperature, since it was not
clear whether or not total or static temperature was measured. When the expression for
the total exit temperature is used (Eq. 5.1), it can be calculated that the dynamic term
only imposes a difference of approximately 4 K when an average exit velocity of 90 m/s
is used.
V2
T0 =Ts + (5.1)
2C p
5.3 Numerical test cases

Initially, the original combustion chamber was modelled to provide a base model against
which the other models can be compared. In addition, a numerical model was generated
describing the original combustor showing the effect of a blocked splash- cooling device
on the dome temperature as well as on the exit temperature profile.
For all the numerical test cases, the pressure drop across the combustion chamber was
maintained similar to the original combustion chamber so as not to compromise overall
engine performance. All the test cases investigated the influence of replacing the existing
splash-cooling devices with an axial swirler to evaluate its influence on the dome wall
temperatures and exit temperature profile. The initial three cases all had exactly the same
mass flow distributions as the original combustion chamber. The swirlers utilised for
these test cases had a mass flow distribution of 6.1 percent which is equivalent to the total
splash-cooling device flow distribution. However, the difference between the initial three
test cases was the number of blades and the blade angle used for each swirler, which in
turn influenced the swirl number and increased the swirler diameter. The inner (hub)
diameter of each swirler was fixed at 32.4 mm due to the fuel injector size.
The remaining two cases investigated an increase in swirler mass flow rate from 6.1
percent to 13.85 percent while maintaining a constant pressure loss of 5.1 percent, which
81
________________________________________________________________________
was determined from the network analysis in the previous chapter. To maintain a constant
pressure loss, 50 percent of the film-cooling air through the first wiggle strip device for
case 4 and the final dilution hole for case 5 was blocked in the numerical model. It
should, however, be noted that the increased mass flow rate in the primary zone will
influence the equivalence ratio, which in turn may influence the combustion stability and
performance. Table 3.4 presented the swirler parameters for the five numerical test cases.
5.4 Numerical analysis of base case

Figure 5.1 depicts the original combustion chamber, at the operating conditions specified
in table 1.1, with dome wall temperature and exit temperature distribution. The numerical
exit temperature distribution was validated against that obtained experimentally by
Skidmore (2004) and is within an error of less than 10 percent.
Figure 5.1: Dome wall temperature and exit temperature distribution of original
combustor (base case)
As discussed in section 2.4, the maximum allowable temperature for Hastelloy-X is

approximately 1150 K. From figure 5.1, it seems that the average dome wall temperature
82
________________________________________________________________________
is within that temperature range, except for a few hot spots. These hot spots are at a
temperature of approximately 1500 K, inducing high-temperature gradients. These
gradients might result in excessive thermal stresses leading to possible crack formations.
It is believed that blocked cooling devices might contribute to the formation of cracks and
therefore an additional analysis was conducted on the original combustion chamber
investigating the effect of a single blocked splash-cooling device on the dome wall. The
results are depicted in figure 5.2. It is apparent that a significant change can be seen in the
temperature distribution on the dome. A high-temperature section is visible on the outer
section of the dome inducing a high-temperature gradient between the inner and outer
dome surface. Alternatively, the exit temperature distribution had an insignificant change
due to the blocked cooling device although the air/fuel ratio in the primary zone has
decreased.
Figure 5.2: Dome wall temperature and exit temperature distribution of original
combustor with blocked splash-cooling strips
83
________________________________________________________________________
Comparing figure 5.2 to figure 1.1, it is clear that the area in which cracks are found is
comparable to that of the area where high-temperature gradients persist.
In addition to the high dome wall temperatures due to the blocked splash-cooling device,
the pressure loss across the combustion chamber will also increase, and consequently
may influence overall engine performance and air-fuel mixing in the primary zone.
5.5 Numerical analysis of swirler

Five swirler modifications, presented in table 3.4, are analysed using CFD. Since there is
a lack of experimental data for these analyses, the data will be compared with the base
case model described in the previous section. The existing splash-cooling devices were
replaced with an axial swirler with a similar air flow distribution as the splash-cooling
devices. For each case, the number of swirler blades, angle, and consequently outer
diameter varied.
Figure 5.3 illustrates the dome wall temperature and exit temperature profile of case 1.
The swirler of case 1 consists of 11 blades positioned at an angle of 33º to the axis of the
combustor. Although the temperature is relatively uniform across most of the dome wall,
it remains too high and may have an effect on the durability of the combustor liner. The
area surrounding the swirler is at a lower temperature, as expected, which may give rise
to crack formation due to the temperature gradient between the inside and outside dome
wall. Also, since the air mass flow rate into the primary combustion zone remained
similar to that of the base case, an insignificant change in the exit temperature profile is
evident.
84
________________________________________________________________________
Figure 5.3: Dome wall temperature and exit temperature distribution of case 1
Figure 5.4 illustrates the dome wall temperature and exit temperature distribution for case
2. This swirler is designed to deliver a similar air flow distribution as the splash-cooling
devices in the base case, consisting of 10 blades positioned at a swirl angle of 40°.
Comparing figure 5.4 and figure 5.3, the effect of increasing the swirler blade angle can
clearly be seen. As the swirl angle increases, the cooling effect on the dome will increase
as well, due to improved impingement of the air on the dome wall. The temperature
contours show clearly that high temperatures still persist on most of the outer surface
inducing high temperature gradients between the inside and outside dome wall.
85
________________________________________________________________________
Finally, figure 5.5 depicts the dome wall temperature and exit temperature distribution for
case 3. The concept makes use of an axial swirler with eight blades positioned at 50º to
the combustor axis, once more, maintaining a similar air mass flow distribution as the
splash-cooling devices on the dome of the base case. A distinct difference can be seen
when comparing the temperature distribution on the dome wall to the previous two cases.
An improvement in temperature distribution is evident due to the larger swirl angles but
the penetration of the cooling air is still not adequate to cool the outer dome wall section.
Crack formation might therefore still be a problem.
86
________________________________________________________________________
According to the three case studies, the swirler has two significant effects on the
combustor. Firstly, the temperature distribution has improved as the swirler blade angle
increased and secondly, the swirl number is enhanced therefore improving air fuel mixing
within the primary zone. The swirl number of case 3 is approximately 1.1 determined
from the one-dimensional swirler analysis. If a swirl angle higher than that of case 3 is
used, a larger swirl number will be induced consequently indicating stronger swirl.
According to Dodds and Bahr (1990), if a very strong and long recirculation zone is
induced it may entrain some of the relatively cool secondary gases, which may result in
stability problems and low combustion efficiency. Considering a maximum swirl number
of 1.1 from case 3, increasing the swirl angle will therefore induce a stronger swirl. It is,
however, evident from the initial three cases that the temperature distribution improves as
the swirl angle increases. However, the air does not penetrate far enough along the dome
wall due to the lack of momentum, resulting in high temperature regions on the outer
dome surface (the hot spot on the dome wall, evident in figure 5.5, is due to a too coarse
grid which might be resolved by refining the mesh in that area). For this reason, two
additional cases were evaluated with a fixed swirler blade angle of 40° and double the
87
________________________________________________________________________
mass flow rate of the previous three cases. The effect of increased swirler mass flow rate
on the dome wall temperature distribution is investigated in case 4 and case 5.
Figure 5.6 illustrates the dome wall temperature and exit temperature distribution of case
4. The swirler utilised in this test case consisted of 11 blades positioned at 40° to the
combustor axis, and has an air mass flow distribution of 13.85 percent compared to the
6.1 percent of the previous cases. To maintain a combustor pressure loss of 5.1 percent
while increasing the swirler air mass flow rate, 50 percent of the first wiggle strip features
were closed in the numerical model. The dome wall temperature has improved
significantly due to the increase of swirler mass flow rate as compared with case 2 with
similar swirler blade angle. Some hot spots still persist, however, on the outer edge of the
dome wall. Comparing the exit temperature profiles, the profile in case 4 has significantly
changed. Using the assumptions of flow distributions into each combustion zone,
described in section 2.3, the equivalence ratio for the primary zone in the base case can
roughly be calculated as 3 whereas the equivalence ratio for primary zone in cases 4 and
5 is reduced to approximately 1.8 due to the additional airflow through the swirler.
Owing to the fact that the equivalence ratio in the primary combustion zone for both case
4 and case 5 is closer to unity, more latent heat is released giving rise to a higher exit
88
________________________________________________________________________
Figure 5.7 illustrates the dome wall temperature and exit temperature distribution for case
5. This case is similar to case 4 but investigates whether the hot spots on the outer dome
surface were due to the reduced film-cooling air. Once again, to facilitate a constant
pressure drop of 5.1 percent some hole features needed to be closed numerically,
therefore the final dilution hole was closed for this case. It was expected that a reduction
in dilution air flow will influence the exit temperature distribution, but was not under
consideration during this investigation. Under consideration was to determine whether
the hot spots on the outer dome area were due to the reduced film-cooling air.
From figure 5.7, it is evident that the addition of 100 percent of the total film-cooling air
does not improve the dome wall temperature. The hot spots on the dome outside wall
might therefore be due to the recirculation air introducing some of the hot gases into that
area. Reducing the dilution air has, however, increased the maximum exit temperature
dramatically.
89
________________________________________________________________________
Table 5.1 depicts the pattern and profile factors for each case. Base case 2 in the table
presents the base case model with a single blocked splash-cooling device. Case 1
correlates well with the experimental determined profile factor. Since all the models are
compared relative to the base case, case 3 is the most satisfactory with respect to profile
factor. When comparing the pattern factors, however, case 4 corresponds closest to the
base case data.
Table 5.1: Profile and pattern factors

Experimental Base Base Case Case Case Case Case
data case case 2 1 2 3 4 5
Profile
0.0523 0.083 0.086 0.059 0.065 0.072 0.15 0.152
factor
Pattern
0.115 0.366 0.373 0.306 0.277 0.264 0.32 0.41
Factor
90
________________________________________________________________________
Figure 5.8 compares all the numerical exit temperature profiles. Notice that the data from
the initial three cases is comparatively good relative to the base case due to the mass flow
distributions that were maintained similar. Although the reduced dilution air played a
significant role in the maximum temperature of case 5, an increase in mass flow
distribution through the swirler results in an increase in the maximum temperature for
case 4 and consequently also influenced case 5.
Figure 5.8: Exit temperature profiles of all numerical models
Figure 5.9 and figure 5.10 depict the numerical and experimental exit temperature
distributions respectively. Figure 5.9 (a) and (b) depict the exit temperature distribution
of the base case and the base case with blocked splash-cooling devices respectively.
When comparing these two figures it is evident that the blocked cooling devices have an
insignificant influence on the exit temperature distribution but will certainly influence the
pressure loss across the combustion chamber. The reduction in cooling air will, however,
91
________________________________________________________________________
influence the equivalence ratio in the primary/secondary zone therefore influencing
combustion performance.
Figure 5.9: Exit temperature distribution of all numerical models
92
________________________________________________________________________
Figure 5.10: Experimental exit temperature distribution
When comparing figure 5.9 (a) and figure 5.10, it is evident that the two figures show the
same trend in exit temperature distribution although the maximum temperature differs by
as much as 200 K. The discrepancies may be due to the numerical model used, such as
the turbulence model, fuel spray and combustion model. The turbulence model may not
describe the recirculation effectively in critical areas whereas the fuel droplet size in the
spray model will influence evaporation and hence heat release from the fuel. In addition,
in the current combustion model only five species were considered with an assumed β
PDF function. Nonetheless, with the assumptions that were made comparative results
were obtained.
Figures 5.9 (c) to (e) represents the temperature distributions of case 1 to 3. The boundary
conditions of these three cases were exactly the same with the only difference being the
outer diameter in the swirler, number of blades used, and the angle of the blades. The
reason for the evaluation of these three cases was to determine the influence of the blade
angle on the exit temperature distribution and dome wall temperature while maintaining a
constant swirler mass flow rate similar to the existing splash-cooling devices. It is
apparent that since the mass flow rate through the swirlers remained constant, the blade
angles had essentially no effect on the exit temperature distribution when comparing the
three cases. However, when comparing the three cases to the base case (figure 5.9 (a)) the
effect of a swirler on the temperature distribution is clearly depicted. It is interesting to
note that the central hot core of figure 5.9 (a) has shifted to the left and that a hot region
appears on the right.
93
________________________________________________________________________
Figure 5.9(f), depicts case 4 where the mass flow rate through the swirler was increased
from 6.1 percent to approximately 13.85 percent. As mentioned earlier, the flow rate
through the first wiggle strip device was reduced by 50 percent in order to maintain a
pressure loss of 5.1 percent and to accommodate for the increase in swirler mass flow
rate. Comparing case 4 to the previous three cases the exit temperature has changed but
the maximum temperature remains at approximately 45 percent of the radial span. An
increase in swirler mass flow rate therefore has a significant effect on the exit
A similar case was conducted with case 5 but with the exception of reducing the flow rate
through the first wiggle strip device, the final dilution hole was also blocked to
accommodate for the increased swirler mass flow rate and to maintain a constant pressure
loss of 5.1 percent. Due to the final dilution hole that was blocked, the exit temperature
increased considerably. Note that in both cases 4 and 5 the equivalence ratio in the
primary combustion zone was altered, compared with the initial three cases, which may
influence combustion stability.
Figure 5.11 depicts, on the center plane, the velocity vectors of the internal flow of the
original combustion chamber (base case). Due to the weak recirculation zone, the air in
the core of the combustion chamber passes through the primary combustion zone without
taking part in the recirculation zone. The recirculation zone is weak and is situated too far
downstream from the dome.
94
________________________________________________________________________
Figure 5.11: Velocity vectors of base case depicted on the center plane
Figure 5.12 depicts path lines of particles being released from the injector shroud. This
figure verifies the statement from the previous figure that most of the core flow does not
take part in a recirculating flow. A swirling motion is, however, only initiated after the
primary combustion zone due to the position of the primary holes.
95
________________________________________________________________________
Figure 5.12: Particle track of particles released from injector shroud – Base case
Figure 5.13 depicts, on the center plane, the velocity vectors of the internal flow field of
case 2. It is apparent that the recirculation zone has increased with the use of a swirler
compared to the base case. Two recirculation zones are, however, present, instead of an
ideal single recirculation zone. The core flow is, however, more active in the recirculation
zone than in the previous case.
It appears that the inner recirculation zone is generated due to the swirler whereas the
second recirculation zone is induced due to the primary air.
96
________________________________________________________________________
Figure 5.13: Velocity vectors of case 2 depicted on the center plane
Figure 5.14 depicts case 2 with particle path lines for particles being released from the
same position as the base case. In case 2, the particles flow outwards around the re-
circulation zone due to the larger swirl angle. This may be due to the low pressure
induced by the swirler. Air-fuel mixing is therefore enhanced by increasing the residence
time.
97
________________________________________________________________________
Figure 5.14: Particle track of particles released from injector shroud – Case 2
Figure 5.15 depicts, on the center plane, the velocity vectors for the internal flow field for
case 4. Two re-circulation zones exist once again but it seems that the recirculation zone
has increased in length. Figure 5.16 depicts the particle path lines of case 4. The flow
field appears to be similar to the flow of case 2.
98
________________________________________________________________________
Figure 5.15: Velocity vectors of case 4 depicted on the center plane
Figure 5.16: Particle track of particles released from injector shroud – Case 4
99
________________________________________________________________________
The following three figures depict the internal temperature contours of the base case, case
2 and case 4 respectively. Due to the core flow that does not participate in the re-
circulating flow, the fuel injection is a long and narrow injection compared to the
following two cases. Due to the lack of a proper recirculation zone, that is evident from
the long fuel injection spray, some of the combustion gases persist near the liner dome
wall.
Figure 5.17: Internal gas temperature contour depicted on the center plane of base
case
The fuel injection for case 2 and 4 does not penetrate as deep as the base case due to the
recirculation flow that is being induced by the swirler. Evidently, the effect of the swirler
mass flow rate can be seen in figures 5.18 and 5.19. The addition of swirler air mass flow
rate (case 4) increased the momentum and consequently improved the penetration of the
cooling air along the combustor liner.
100
________________________________________________________________________
Figure 5.18: Internal gas temperature contour depicted on the center plane
of case 2
Figure 5.19: Internal gas temperature contour depicted on the center plane
of case 4
101
________________________________________________________________________
5.6 Predicted wall temperatures
The average combustor wall temperature that was obtained from CFD analysis is
compared with the predicted one-dimensional network results. Figure 5.20 and figure
5.22 depict the average wall temperatures along the combustor liner. Figure 5.20
represents a comparison between average CFD wall predicted data and network wall
temperature data. In this case, the network wall temperatures were obtained from
empirically predicted gas temperatures shown in figure 3.10. It is evident that the two
graphs show similar trends. Due to the gas temperature that was underpredicted in the
primary zone the wall temperature in the dome section is underpredicted as well. The
underprediction of the empirical gas temperature is discussed in section 3.6.
Figure 5.20: Network vs CFD-predicted average wall temperature using the

empirical gas temperature predictions
102
________________________________________________________________________
The network temperature predictions shown in figure 5.22 were, however, predicted
using the average CFD gas temperatures. The average CFD gas temperature was obtained
from the base case CFD model by determining the average gas temperature on various
cross sectional planes. The position of these planes was associated with the position of
the nodes in the network model that represents the combustor’s central axis.
The average CFD gas temperature is compared with the empirical gas temperature,
presented earlier in figure 3.10, and is depicted in figure 5.21.
Figure 5.21: Average CFD and empirical gas temperatures
From the comparison of gas temperatures, it is apparent that the temperature predictions
in figure 5.22 will be different when compared to the predictions of figure 5.20, but will
consequently have a closer resemblance to the CFD results. Figure 5.22 therefore
illustrates the ability of the network solver to predict comparative CFD results.
103
________________________________________________________________________
Figure 5.22: Network vs CFD-predicted average temperature using the numerically

predicted gas temperatures
In figure 5.20 and figure 5.22 at a distance of approximately 300mm, the network
analysis has overpredicted the wall temperature. This can be ascribed to the fact that the
CFD analysis accounts for the local cooling effect of the hole features along the
combustor wall whereas the network model did not account for such effects. To
incorporate such effects into the network analysis a typical film-cooling efficiency needs
to be included into all hole features. Such data is, however, not available and for the
purposes of this study the results obtained were comparatively good describing the same
trend for the temperatures.
104
________________________________________________________________________
5.7 Summary
The effect of swirler blade angle and mass flow rate on the dome wall temperature was
investigated through the use of a parametric study. The results showed that the addition
of a swirler with a similar flow distribution as the original splash-cooling devices, will
not result in a more uniform dome wall temperature. When the mass flow rate through the
swirler was, however, increased, the possibility of a more uniform dome wall temperature
increased. Case 4 resulted in adequate cooling although hot spots still remained on the
outer dome surface. The initial three tests cases showed that increasing the swirler blade
angle resulted in an improved dome wall temperature distribution. A swirler with 13.85
percent and higher swirl angle in the range of 45° - 50° might therefore improve the
temperature distribution even more. However, due to the fact that a leaner fuel/air ratio is
present in the primary combustion zone, some stability problems may exist. For this
reason, before any swirler as described above can be used, a complete redesign of the
combustor hole sets will be needed.
105
CHAPTER 6 Summary, conclusions and recommendations
________________________________________________________________________
CHAPTER 6 –
SUMMARY, CONCLUSION AND
RECOMMENDATIONS
6.1 Summary
The purpose of this study was two-fold; (1) to develop a one-dimensional incompressible
code, incorporating an empirical combustion model, in conjunction with a network solver
to predict flow distributions, pressure losses, and temperatures across a standard T56
combustion chamber; (2) to use the data obtained form the one-dimensional analysis as
boundary conditions to investigate possible modifications to a T56 combustion chamber,
by replacing the existing dome cooling devices with an axial swirler. The results are
compared against an experimental validated numerical base case model.
The network model accounted for compressibility effects and made modelling complex
geometries much simpler. Due to a lack of empirical correlations in the network model to
model combustion, the model was used in conjunction with the empirical model to obtain
gas temperatures and gas emissivities. The method that was used in the network model to
predict the effect of film-cooling devices proved to be sufficient when compared with the
numerical data. With the network approach, the effect of a film-cooling device can,
however, be modelled sufficiently without the use of empirical models.
Although the network model had to be calibrated with experimental flow distribution
data, critical data such as pressure losses across the combustion chamber was obtained.
106
________________________________________________________________________
The pressure loss was used during the modifications analysis in order to design the
swirler geometries while maintaining overall engine performance. Nonetheless, the one-
dimensional analysis proved to predict sufficient flow distributions. The effect of a
swirler on the dome wall temperature can, however, not be predicted with the use of a
one-dimensional model since the model can only predict the mass flow rate and pressure
loss across a specific device. However, an empirical code was developed and used to
design the swirler geometry with a specified mass flow rate and pressure loss as inputs.
The investigation was conducted with the use of three-dimensional CFD analysis. The
original combustion chamber was analysed and used as a base case model. The numerical
exit temperature distribution of this model was within 10 percent of experimental data.
Since the purpose of the study was to investigate the effect of a swirler on the dome wall
temperatures by combining a one-dimensional empirical and network solver, no
experimental data was available. Therefore, the original combustion chamber was
analysed and validated against experimental data that could be used as a model against
which the test cases can be compared with. Using a swirler with the same mass flow
distribution as the splash-cooling strips, did not offer a sufficient temperature
distribution. Doubling the flow rate through the swirler did, however, have a greater
impact on the wall temperature. Although temperature gradients still persist with the use
of a swirler, lower wall temperatures are apparent which may reduce possible crack
formation are apparent.
6.2 Conclusion and contributions

The following conclusions were made during this study;
• The ability of the one-dimensional empirical code to predict flow distributions,
pressure losses and temperature distributions proved to compare well with
network predictions. The conclusion was made that compressibility had an
insignificant effect on these predictions. It was also concluded that the discharge
coefficients play a great role in determining pressure losses.
107
________________________________________________________________________
• The one-dimensional network solver predicted basically the same trend of
temperature distribution compared to the network solver. The drawback of the
empirical code is that it makes use of a film-cooling efficiency correlation to
predict the cooling effect on the liner wall. Nonetheless, adequate initial data can
be obtained from such an analysis. The network model, on the other hand, used a
heat transfer process to model the cooling effect, since no empirical data was
available on the swirler. Yet, comparative temperature distributions were obtained
compared to CFD data.
• Numerical modelling of the original combustion chamber showed that a poor re-
circulation zone exists. This may result in inadequate air-fuel mixing and
therefore a loss in combustion efficiency and a possibility of high smoke
emissions.
• The original combustion chamber dome wall temperature seems as if the average
temperature is below 1150 K, except for minor hot spots on the outer surface.
These hot spots will cause high temperature gradients on the dome wall, which
may contribute to the formation of cracks. A second numerical analysis proves
that in a case where a splash-cooling device might become blocked, the problem
may be aggravated due to higher-temperature gradients between the outer and
inner surface.
• Particle tracking has shown that the recirculation flow in the primary combustion
zone has improved when a swirler is used. The recirculation zone will improve
combustion efficiency and might even result in lower emission levels, such as
soot and CO2, due to the increased residence time in the region.
• The initial three swirler cases utilised approximately the same mass flow rate as
the splash-cooling devices on the dome. The influence of these swirlers on the
dome wall temperature did not prove to be an adequate solution, since high
temperatures still persisted on the outer wall area. The temperature distribution
108
________________________________________________________________________
improved, however, as the swirler blade angle increased. In all three cases, the
average exit temperature profile stayed virtually unchanged, since the equivalence
ratio in the primary combustion zone was not changed greatly. A drawback of
these three cases is the physical size of these devices. Due to the small amount of
air that it passes (6.1 percent), the dimensions are very small and will therefore be
difficult to manufacture.
• Two additional swirler cases were investigated as well. These cases utilised an
increased mass flow rate, approximately double that of the previous three cases,
while maintaining a similar pressure loss of 5.1 percent. The dome wall
temperature distribution improved significantly, due to the increased momentum,
but resulted in higher average exit temperature distributions. The reason for it
being that the equivalence ratio in the primary combustion zone has been reduced,
thus releasing more latent heat. The swirlers used in case 4 and case 5 utilised 11
blades positioned at a 40º angle relative to the combustor axis. The dome wall
temperature may, however, be improved even more if a larger blade angle is used.
• The addition of a swirler with the same mass flow distribution as the splash-
cooling devices does not result in a more uniform dome wall temperature. In order
for it to be accomplished, a swirler with more than double the mass flow rate
should be used in order for the cooling air to penetrate to the outer dome surface.
The drawback of doubling the swirler mass flow rate is that the equivalence ratio
in the primary zone decreases which may result in stability problems at off-design
operating conditions. Also, NOx emission levels might increase due to the primary
combustion zone being more fuel lean and therefore operating at conditions closer
to the stoichiometric air/fuel ratio. The swirler does, however, enhance the
recirculation flow in the primary combustion zone compared with the original
combustion chamber. An enhanced recirculation zone will ensure more efficient
mixing of the fuel and air which may result in increased combustion efficiency.
109
________________________________________________________________________
From the above discussion, it is evident that by simply replacing the existing splash-
cooling devices with an axial air swirler will not result in an adequate solution. When
the swirler mass flow rate was increased to attempt to lower the temperature
distribution on the dome wall, the equivalence ratio in the primary combustion zone
was altered. Due to the change, stable combustion at other operating conditions
cannot be guaranteed since these analyses were only conducted at a single operating
condition.
To incorporate the swirlers of case 4 and case 5 into the combustion chamber, the
primary zone airflow should be reduced to maintain the pressure loss. An
investigation should also be conducted to ensure combustion performance at other
operating conditions.
6.3 Recommendations for future work

Recommendations for future work, related to this study, can be summarised as follow:
• Investigation to improve the equivalence ratio in the primary combustion zone by
re-designing the size and location of the primary jets. Also, investigating the
influence of reduced film-cooling air along the combustor liner.
• Include radiation in all the numerical models and describe the external heat
transfer coefficient and material conductivity in a user-define-function (UDF).
Constant values were used during this study.
• Increase swirler blade angle to 45º or 50º with mass flow distribution of 13.85
percent and investigate dome wall temperature with improved design.
• Experimental verification of the dome wall temperatures as well as the swirler

designs should be conducted.
• The effect of emission levels for each swirler should be investigated.
110
________________________________________________________________________
• The data utilised for each element in the network model should be generated in a
more generic way since the data had to be inserted individually for every element,
making it a time-consuming task.
111
REFERENCES
Adkins, R.C. and Gueroui, D., 1986, “An Improved Method for Accurate Prediction of
Mass Flows Through Combustor Liner Holes”, Journal of Engineering for Gas Turbine
and Power, Vol. 108, pp 491-497
Ballal, D.R. and Lefebvre, A.H., 1986, “A Proposed Method for Calculating Film-cooled
Wall Temperatures in Gas Turbine Combustion Chambers”, ASME Paper No. 72-
WA/HT-24
Baron, F., Kanniche, M. and Mechitoua, N., 1994, “Influence of Inlet Conditions on Flow
and Combustion Characteristics of a Model Annular Combustor, ASME Paper No. 94-
GT-281
Biswas, D., Kawano, K., Iwasaki, H. and Ishizuka, M., 1997, “Three-Dimensional
Computation of Gas Turbine Combustors and the Validation Studies of Turbulence and
Combustion Models”, ASME Paper No. 97-GT-362
Cia, J., Jeng, S.M. and Tacina, R., 2002, “Multi-Swirler Aerodynamics: Comparison of
Different Configurations”, AMSE Paper No, GT-2002-30464
Crocker, D.S. and Smith, C.E., 1995, “Numerical Investigation of Enhanced Dilution
Zone Mixing in a Reverse Flow Gas Turbine Combustor”, Journal of Engineering for Gas
Turbine and Power, Vol. 177, pp 272-281
Crocker, D.S., Nickolaus, D. and Smith, C.E., 1999, “CFD Modeling of a Gas Turbine
Combustor from Compressor Exit to Turbine Inlet”, Journal of Engineering for Gas
Turbine and Power, Vol. 121, pp. 89-95
112
Dodds, W.J., and Bahr, D.W., 1990, “Combustion System Design”, in A.M. Mellor, ed.,
“Design of Modern Gas Turbine Combustors”, pp 343-476, Academic Press, San Deigo
Eccles, N.C. and Priddin, C.H., 1999,” Accelerated Combustion Design using CFD”,
ISABE Paper No. 99-7094
Feitelberg, A.S. and Lacey, M.A., 1998, “ The GE Rich-Quench-Lean Gas Turbine
Combustor, Journal of Engineering for Gas Turbine and Power, Vol 120, pp. 502-508
Fuller, E.J. and Smith, C.E., 1993, “Integrated CFD Modeling of Gas Turbine
Combustors”, AIAA, A93-50008
Gosselin, P., De Champlain, A., Kretschmer, D., 1999,” Prediction of Wall Heat Transfer
for a Gas Turbine Combustor”, Journal of Power and Energy, Vol. 213, pp. 169-180
Graves, C.C. and Gronman, J.S., 1957, “Theoretical Analysis for Tubular Turbojet
Combustors with Constant Annulus and Liner Cross-Sectional Areas”, NACA Report
1373
Greyvenstein, G.P. and Laurie, D.P., 1994, “A Segregated CFD Approach to Pipe
Network Analysis”, International Journal for Numerical Methods in Engineering, Vol.
37, p3685 – 3705
Guoqiang, L. and Gutmarl, E.J., 2004, “Effects of Swirler Configuration on Flow

Structures and Combustion Characteristics”, ASME Paper No. GT2004-53674
Gupta, A.K. and Lewis, M.J., 1998, “Effect of Swirl on Combustion Characteristics in
Premixed Flames”, Transactions of the ASME, Vol. 120, July, pp. 488-494
Gupta, A.K., Lilley, D.G. and Syred, N., 1984, “Swirl Flow – (Energy and engineering
science series)”, Abacus Press
113
Hicks, R.A. and Wilson, C.W., 1999, “Comparison between measured and predicted wall
temperatures in gas turbine combustors”, ASME Paper No. 99-GT-59
Hoertzer, K., Haswlbacher, H., Griffiths, A.J., Syred, N. and Fraser, T.A., 2002, “CFD-
Analysis of the Combustion in the Primary Stage of a Two-Stage Combustion Chamber”,
ASME Paper No. GT-2002-30610
Holdeman, J.D., Liscinsky, D.S. and Bain, D.B., 1997, “Mixing of Multiple Jets with a
Confined Subsonic Crossflow Part II – Opposed Rows of Orifices in Rectangular Ducts”,
ASME Paper No. 97-GT-439
Hornsby, C., and Norster, E.R., 1997, “Application of CFD to DLN Combustors”, ASME
Paper No. 97-GT-371
Hu, T., and Prociw, L.A., 1993,”Recent CFD Applications in Small Gas Turbine
Combustion Systems Development”, Proc. 81st Symp. Propulsion and Energetic Panel on
Fuel and Combustion Technology for Advanced Aircraft Engines, AGARD Conference,
Rome, Italy
Joubert, F.M. and Hattingh, H.V., 1982, “Rekenaar-Modellering van ‘n Annulere

Teenvloei-verbrandingsruim vir Gasturbienenjins”, Journal of Aeronautical Society, Vol.
3, pp 44-50
Knight, H.A. and Walker, R.B., 1953, “The Component Pressure Losses in Combustion
Chambers” Aeronautical Research Council R and M 2987, England
Kretschmer, D., and Odgers, J., 1978, “A Simple Method for the Prediction of Wall
Temperatures in Gas Turbines”, ASME Paper No. 78-GT-90
Lawson, R.J., 1993, “Computational Modeling of an Aircraft Engine Combustor to

Achieve Target Exit Temperature Profiles”, ASME Paper No. 93-GT-164
114
Lefebvre, A.H., 1998, “Gas Turbine Combustion” 2nd Edition, Taylor & Francis
Lefebvre, A.H. and Norster, E.R., 1969, “The Design of Tubular Gas Turbine
Combustion Chambers for Optimum Mixing Performance”, Proceedings of the Institution
of Mechanical Engineers, Vol. 183, pp. 150-155
Liedtke, O., Schulz, A., and Wittig, S., 2002, “Design Study of a Lean Premixed
Prevaporozed Counter Flow Combustor for a Micro Gas Turbine”, ASME Paper No. GT-
2002-30074
Lienhard IV, J.H. and Lienhard V, J.H., 2003, ”The heat transfer textbook – third
edition”, Phlogiston Press, Cambridge Massachusetts
Malecki, R.E., Rhie, C.M., McKinney, R.G., Quyang, H., Colet, M.B., and Madabhushi,
R.K., 2001, “Application of an Advanced CFD-Based analysis system to the PW6000
Combustor to Optimize Exit Temperature Distribution – Part 1: Description and
Validation of the Analysis Tool”, ASME Paper No. 2001-GT-0062
Mattingly, J.D., Heiser, W.H., Pratt, D.T., 2002, ”Aircraft Engine Design Second
Edition”, AIAA Education Series
McGuirk, J.J. and Palma, J.M.L.M., 1993, “The Flow Inside a Model Gas Turbine
Combustor: Calculations”, Journal of Engineering for Gas Turbines and Power, Vol. 115,
pp 594-602
McGuirk, J.J. and Spencer, A., 2000, “Coupled and Uncoupled CFD Prediction of the
Characteristics of Jets from Combustor Air Admission Ports”, ASME Paper 2000-GT-
0125
115
McGuirk, J.J. and Spencer, A., 2001, “Coupled and Uncoupled CFD Predictions of the
Characteristics of Jets from Combustor Air Admission Ports”, Journal of Engineering for
Gas Turbine and Power, Vol. 123, pp 327-332
Micklow, G.J., Roychoudhury, S., Nguyen, H.L. and Cline, M.C., 1993, “Emission
Reduction by Varying the Swirler Airflow Splits in Advanced Turbine Combustors”,
Journal of Engineering for Gas Turbines and Power, Vol. 115, pp 563-569
Mongia, H.C., Reynolds, R.S., and Srinivasan, R., 1986, “Multidimensional Gas Turbine
Combustor Modeling: Application and Limitations”, AIAA Journal, Vol. 24, pp 890-904
Norster, E.R., 1980, “Combustor Aerodynamics”, in Northern Research and Engineering

Corporation, “The Design and Development of Gas Turbine Combustors”, Woburn,
Massachusetts, Volume 1
Odgers, J. and Kretschmer, D., 1980a “Performance evaluation and the Prediction of
Pollutant Formation”, in Northern Research and Engineering Corporation, “The Design
and Development of Gas Turbine Combustors”, Woburn, Massachusetts, USA, Volume 2
Odgers, J. and Kretschmer, D., 1980b, “Basic Design Method – Can Combustors”, in
Northern Research and Engineering Corporation, “The Design and Development of Gas
Turbine Combustors”, Woburn, Massachusetts, USA, Volume 2
Rizk, N.K., 1994, “A Quasi-3D Calculation Method for Gas Turbine Combustor Liner
Wall Cooling Techniques”, ASME Paper No. 94-GT-316
Samuel, B.P., 1961, “A Jet Engine Combustor Design Analysis Suitable for Electronic
Computers”, ASME Paper No. 61-WA-305
Shih, T.H., Liou, W.W., Shabbir, Yang, Z. and Zhu, J., 1995, “A New k-ε Eddy Viscosity
Model for High Reynolds Number Turbulent Flows - Model Development and
Validation”, Computers Fluids, Volume 24, pp.227-238,
116
Sivaramakrishna, G., Muthuveerappan, N., Vankataraman, S., Sampathkumaran, T.K.,
2001, “CFD Modeling of Aero Gas Turbine Combustors”, ASME Paper No. 2001-GT-
0063
Skidmore, F.W., 1986, “Smoke Emission Tests on Series II and Series III Allison T56
Turboprop Engines”, ARL Aero-Prop-TM-438
Skidmore, F.W., Bennett, J.M., and Glenny, D.E., 1995, “An Investigation into Hard
Carbon Formation in a Modified Gas Turbine Combustor”, ISABE 95-7116
Skidmore, F.W., 2004, Private Communication, Brisbane Australia
Smiljanovski, V. and Brehm N., 1999, “CFD Liquid Spray Combustion of a Single
Annular Gas Turbine Combustor”, ASME Paper No. 99-GT-300
Snyder, T.S., Stewart, J.F., Stoner, M.D., McKinney, R.G., 2001, “Application of an
Advanced CFD-Based Analysis System to the PW6000 Combustor to Optimize Exit
Temperature Distribution – Part II: Comparison of Predictions to Full Annular Rig Test
Data”, ASME Paper 2001-GT-0064
Sturgess, G.J., Sloan, D.G., Lesmerises, A.L., Heneghan, S.P., and Ballal D.R., 1992
“Design and Development of a Research Combustor for Lean Blow-out Series”, Journal
of Engineering for Gas Turbine and Power, Vol. 114, pp. 13-19
Stuttaford, P.J. and Rubini, P.A., 1997, “Preliminary gas turbine combustor design using
a network Approach”, Journal of Engineering for Gas Turbine and Power, Vol. 119, pp.
547-552
Stuttaford, P.J., 1997, “Preliminary Gas Turbine Combustor Design Using a Network
Approach”, PhD Thesis, Cranfield University
117
Tangirala, V.E., Tolpadi, A.K., Danis, AL, M. and Mongia, H.,2000, “Parametric
Modeling Approach to Gas Turbine Combustor Design”, ASME Paper No. 2000-GT-129
Van Niekerk, J.E. and Morris, R.M., 2001, “T56 Isothermal tests and CFD predictions”,
PTE/01/003/1
Versteeg, H.K. and Malalasekera, W., 1995, “An Introduction to Computational Fluid
Dynamics – The Finite Volume Method”, Prentice Hall
Yun, S., Lightstone, M.F. and Thomson, M.J., 2005, “An Evaluation of Beta PDF
Integration using the Density-weighted PDF and the un-weighted PDF”, International
Journal of Thermal Sciences, Vol. 44, pp 421-428
118
APPENDIX A -
Material Properties
APPENDIX A Material Properties
________________________________________________________________________
A1. Material properties of Hastelloy-X
Figure A.1 and table A.1 depicts the results obtained from the electron microscope at the
University of Pretoria in South Africa. Table A.2 present the typical composition of the
material respectively. After comparing the data with various compositions, Hastelloy-X
proved to have the closest agreement with the experimental data. For this reason it was
concluded that the combustion chamber was manufactured from Hastelloy-X. The
material properties for the material, and used during the study, are presented in table A.3.
Although some of the properties were a function of the temperature, mean values were
used during the study
Figure A.1: Electron microscope results of Hastelloy-X
Table A1: Experimental composition of material

Element k-ratio ZAF Atom % Element Wt % Err.
(calc.) Wt % (3-Sigma)
Al-K 0.001 4.91 1.6 0.7 +/- 0.2
Si-K 0.002 3.34 1.4 0.7 +/- 0.1
Cr-K 0.216 0.98 23.6 21.2 +/- 0.4
Mn-K 0.007 0.99 0.7 0.7 +/- 0.2
Fe-K 0.18 1.02 19 18.4 +/- 0.5
Ni-K 0.434 1.1 47.1 47.9 +/- 0.8
Mo-K 0.071 1.16 5 8.2 +/- 2
Co-K 0.012 1.11 1.3 1.3 +/- 0.2
W-M 0.004 2.32 0.3 0.9 +/- 0.4
Total 100 100
119
APPENDIX A Material Properties
________________________________________________________________________
Table A.2: Typical material composition Table A.3: Material properties
Elements Min [%] Max[%] Density [kg/m3] 8820
Molybdenum 8 10 Thermal conductivity [W/m.K] 16.9
Chromium 20.5 23 External emissivity 465
Iron 17 20 Spesific heat Cp [J/kg.K] 465
Tingsten 0.2 1
Cobalt 0.5 2.5
Carbon 0.05 0.015
Silicon -- 1
Manganese -- 1
Boron -- 0.01
Phosphorus -- 0.04
Sulfur -- 0.03
Nickel Remainder
120
APPENDIX B –
Grid independence study

APPENDIX B Grid independence study
________________________________________________________________________
B.1 Grid independence study
A study was conducted on three different mesh sizes consisting of 500 000, 1000 000,
and 1500 000 cells to determine grid independency. Figure B.1 depicts the numerical exit
temperature distribution obtained from each mesh size, and compares it to the
experimental exit temperature distribution obtained by Skidmore (2004). From the study
close agreement between the three cases are evident. Nonetheless, the results obtained
from 1000 000 cells shows closest agreement with experimental data. Due to the
availability of computer recourses and to keep the computational time relatively low, it
was considered that a 1000 000 cells result in comparatively good results and
consequently having grid independence.
Figure B.1: Grid independency study
B.2 Effect of radiation on the liner wall temperatures

Radiation was modeled using the P-1 radiation model, which is a non-equilibrium
diffusion type model, and the variation of the mixture's absorption coefficient was taken
into account using the Weighted-Sum-of-Gray-Gases model. The effect of radiation was
121
________________________________________________________________________
investigated only on the base case model and compared with a similar model without
radiation. Figure B.2.depicts the average liner wall temperature along the combustor liner
with and without radiation and figure B.3 depict the effect of radiation on the exit
temperature profile.
The effect of radiation on the liner wall temperature was found to be a maximum of 10%
on the dilution zone area and not more than 2% on the average exit temperature
distribution.
Figure B.2: Effect of radiation on the combustor average axial wall temperature
122
________________________________________________________________________
Average exit temperature profile
100
90
80
70
60
% Span
50
40
30
20
10
0
0.93 0.98 1.03 1.08 1.13
T-radial/T-average [K/K]
Figure B.3: Effect of radiation on the average outlet temperature distribution
123
APPENDIX C –
Empirical swirler design program

APPENDIX C Empirical swirler design program
________________________________________________________________________
C.1 Empirical swirler design code
---------------------------------------------------------- Inputs -------------------------------------------------------------
THETA=[30 31 32 33 34 35 36 37 38 39 40 45 50 55]; [Swirler blade angles [degrees]]

massf=[5 6 7 8 9 10 11 12 13.67 14 15 16]; [Percentage mass flow rate through
swirler [%]]
[q r]=size(THETA);
[s t]=size(massf);
P=923897; [Combustor inlet pressure [Pa]]
T=566.15;
R=287; [Gas constant ]
dP=5.1; [Percentage pressure drop over
combustor ]
Psw=P*(dP/100); [Swirler pressure loss [Pa]]
Ksw=1.3; [Blade loss coefficient]
L=0.1403; [Average combustor diameter [m]]
AL=(pi*L^2)/4; [Average combustor flow are [m^2]]
ma=2.37642; [Combustor total flow rate [kg/s]]
Injector_diam=0.0324; [Fuel injector diameter [m]]
Hub_and_Ring_th=0.0015; [Swirler hub and outer ring material
thickness [m]]
Vane_th=0.001; [Swirler vane material thickness [m]]
Dhub=Injector_diam+(2*Hub_and_Ring_th);
Solidity=1;
Swirler_width=0.01; [Swirler width [m]]
rho=P/(R*T); [Density [kg/m^3]]

---------------------------------------------------------------------------------------------------------------------------------
CALCULATIONS
ntel =0
for i=1:r
ntel=ntel+1
for j=1:t
m=massf(j); [Percentage swirler mass flow rate]
msw=ma*(m/100); [Swirler mas flow rate [kg/s]]
theta(i)=THETA(i)*(pi/180); [Swirler blade angle [radiant]]
Calculate initial swirler flow area

--------------------------------------------------------------------------------------------------------------------------------
Asw=sqrt((msw^2*Ksw*(sec(theta(i))^2))/((2*rho*Psw)+((msw^2*Ksw)/AL^2)));
---------------------------------------------------------------------------------------------------------------------------------
Lv=Swirler_width/cos(theta(i)); [Swirler blade length]
Dsw=0.02; [Initial swirler blade tip diameter [m]]

ASW=1; [Initial swirler area]
converg=1;
124
________________________________________________________________________
[Determine the number of swirler blades, from which the swirler blade tip diameter will be
calculated]:
---------------------------------------------------------------------------------------------------------------------------------
while abs(converg) > 0.00000001
Number of swirler blades

n=round((pi*Dsw)/(Lv/Solidity));
Swirler flow area

ASW=(pi/4)*(Dsw^2-Dhub^2)-0.5*n*(Vane_th/cos(theta(i)))*(Dsw-Dhub);
converg=Asw-ASW;
Dsw=Dsw+0.0000001;
end
---------------------------------------------------------------------------------------------------------------------------------
[Calculate swirl number and swirler loss coefficient]:

--------------------------------------------------------------------------------------------------------------------------------
DSW(i,j)=Dsw;
N(i,j)=n;
Swirl number
SN(i,j)=(2/3)*((1-(Dhub/DSW(i,j))^3)/(1-(Dhub/DSW(i,j))^2))*tan(theta(i));
Sw_area(i,j)=ASW;
Loss coeff
Loss_coeff(i,j)=(Sw_area(i,j)^2)*(Ksw*(((sec(theta(i))^2)/(Sw_area(i,j)^2)))-1/AL^2); [Loss coeff.]
--------------------------------------------------------------------------------------------------------------------------------
[Calculate hydraulic diameter of single swirler opening]:

--------------------------------------------------------------------------------------------------------------------------------
Blade half thickness
B_half_thick=(Vane_th/cos(theta(i)))/2;
Blade height
Blade_height=(Dsw-Dhub)/2;
Inner arc length of single opening

half_blade_angle=(B_half_thick/(Dhub/2))*2;
arc_angle=(((2*pi)/n)-half_blade_angle);
inner_arc_length=(Dhub/2)*arc_angle;
Outer arc length of single opening

Half_Blade_Angle=(B_half_thick/(Dsw/2))*2;
Arc_Angle=(((2*pi)/n)-Half_Blade_Angle);
Outer_Arc_length=(Dsw/2)*Arc_Angle;
Calculate wettet perimiter

P=(2*Blade_height)+(inner_arc_length+Outer_Arc_length);
Calculate hydroulic diameter

Hd(i,j)=(4*(ASW/n))/P;
end
end
---------------------------------------------------------------------------------------------------------------------------------
125
________________________________________________________________________
[ Write output file]
---------------------------------------------------------------------------------------------------------------------------------
THeta='Vane Angle: ';
Massflow='m_sw [%]';
dswirler='D_sw [mm]';
number_of_vanes='N_vanes';
Swirl_n='SN';
W='Swirler Width [mm]: ';
S_area='Asw [m^2]';
Loss='Loss Coeff';
Hyd='Hydr_D [mm]';
Design_data='******** Design data *******';
Pressure='Total Pressure [Pa] ';
Temp='Temperature [K] ';
dp='Pressure drop [%] ';
massflow='Total Massflow [kg/s] ';
Injector_d='Injector diameter [mm] ';
Blade_th='Blade thickness [mm] ';
Sol='Solidity at tip ';
Hub_ring='Hub & Ring thickness [mm] ';
Swirler_w='Swirler width [mm] ';
final=('******************************************');
fid=fopen('Swirler Data.txt','w');
fprintf(fid,'%6s\n',Design_data);
fprintf(fid,'\n');
fprintf(fid,'%6s',Pressure);
fprintf(fid,'%2.0f\n',P);
fprintf(fid,'%6s',Temp);
fprintf(fid,'%2.2f\n',T);
fprintf(fid,'%6s',dp);
fprintf(fid,'%2.3f\n',dP);
fprintf(fid,'%6s',massflow);
fprintf(fid,'%2.3f\n',ma);
fprintf(fid,'%6s',Injector_d);
fprintf(fid,'%2.3f\n',Injector_diam*1000);
fprintf(fid,'%6s',Blade_th);
fprintf(fid,'%2.3f\n',Vane_th*1000);
fprintf(fid,'%6s',Sol);
fprintf(fid,'%2.2f\n',Solidity);
fprintf(fid,'%6s',Hub_ring);
fprintf(fid,'%2.3f\n',Hub_and_Ring_th*1000);
fprintf(fid,'%6s',Swirler_w);
fprintf(fid,'%2.3f\n',Swirler_width*1000);
fprintf(fid,'%6s\n',final);
fprintf(fid,'\n');
for i=1:r;
th=THETA(i);
fprintf(fid,'%18s',THeta);
fprintf(fid,'%1.1f\n',th);
fprintf(fid,'%18s',W);
fprintf(fid,'%1.2f',Swirler_width*1000);
126
________________________________________________________________________
fprintf(fid,'\n');
fprintf(fid,'\n');
fprintf(fid,'%s',Massflow);
fprintf(fid,'%13s',dswirler);
fprintf(fid,'%12s',number_of_vanes);
fprintf(fid,'%8s',Swirl_n);
fprintf(fid,'%16s',S_area);
fprintf(fid,'%14s',Loss);
fprintf(fid,'%13s\n',Hyd);
for j=1:t;
M=massf(j);
D=DSW(i,j);
NUM=N(i,j);
Sn=SN(i,j);
Loss_C=Loss_coeff(i,j);
Area=Sw_area(i,j);
HD=Hd(i,j);
fprintf(fid,'%7.3f',M);
fprintf(fid,'%15.2f',D*1000);
fprintf(fid,'%9.0f',NUM);
fprintf(fid,'%11.2f',Sn);
fprintf(fid,'%15.8f',Area);
fprintf(fid,'%11.2f',Loss_C);
fprintf(fid,'%14.5f\n',HD*1000);
end
fprintf(fid,'\n');
fprintf(fid,'\n');
end
fclose(fid);
---------------------------------------------------------------------------------------------------------------------------------
127
________________________________________________________________________
C.2 Swirler lookup tables (Program output)
(These specific tables were used to design the swirlers of case 4 & 5)
******** Design data *******
Total Pressure [Pa] 0

Temperature [K] 566.15
Pressure drop [%] 5.100
Total Massflow [kg/s] 2.376
Injector diameter [mm] 32.400
Blade thickness [mm] 1.000
Solidity at tip 1.00
Hub & Ring thickness [mm] 1.500
Swirler width [mm] 10.000
******************************************
Vane Angle: 30.0

Swirler Width [mm]: 10.00
m_sw [%] D_sw [mm] N_vanes SN Asw [m^2] Loss Coeff Hydr_D [mm]
5.000 39.47 11 0.55 0.00021369 1.73 3.35678
6.000 40.23 11 0.54 0.00025641 1.73 3.86495
7.000 40.98 11 0.54 0.00029914 1.73 4.33699
8.000 41.71 11 0.53 0.00034186 1.73 4.77782
9.000 42.49 12 0.53 0.00038458 1.73 5.09273
10.000 43.20 12 0.53 0.00042728 1.73 5.46449
11.000 43.89 12 0.52 0.00046999 1.73 5.81505
12.000 44.57 12 0.52 0.00051268 1.73 6.14679
13.670 45.69 12 0.52 0.00058395 1.73 6.66431
14.000 45.91 12 0.51 0.00059803 1.73 6.76174
15.000 46.65 13 0.51 0.00064070 1.73 6.85171
16.000 47.30 13 0.51 0.00068335 1.73 7.11037
Vane Angle: 31.0

5.000 39.52 11 0.57 0.00021589 1.77 3.38658
6.000 40.29 11 0.57 0.00025906 1.77 3.89789
7.000 41.04 11 0.56 0.00030223 1.77 4.37267
8.000 41.77 11 0.56 0.00034539 1.77 4.81572
9.000 42.50 11 0.55 0.00038855 1.77 5.23127
10.000 43.28 12 0.55 0.00043170 1.77 5.50427
11.000 43.98 12 0.54 0.00047484 1.77 5.85615
12.000 44.67 12 0.54 0.00051797 1.77 6.18907
13.670 45.79 12 0.54 0.00058999 1.77 6.70839
14.000 46.01 12 0.53 0.00060421 1.77 6.80613
15.000 46.77 13 0.53 0.00064732 1.77 6.89434
16.000 47.42 13 0.53 0.00069041 1.77 7.15372
Vane Angle: 32.0

5.000 39.57 11 0.59 0.00021821 1.81 3.41779
6.000 40.34 11 0.59 0.00026185 1.81 3.93243
7.000 41.10 11 0.58 0.00030548 1.81 4.40993
8.000 41.84 11 0.58 0.00034911 1.81 4.85539
9.000 42.57 11 0.57 0.00039273 1.81 5.27301
10.000 43.36 12 0.57 0.00043634 1.81 5.54576
11.000 44.07 12 0.57 0.00047995 1.81 5.89906
12.000 44.77 12 0.56 0.00052355 1.81 6.23329
13.670 45.90 12 0.56 0.00059633 1.81 6.75439
14.000 46.12 12 0.56 0.00061071 1.81 6.85250
15.000 46.78 12 0.55 0.00065428 1.81 7.14082
16.000 47.54 13 0.55 0.00069784 1.81 7.19892
128
________________________________________________________________________
Vane Angle: 33.0

5.000 39.57 10 0.62 0.00022065 1.85 3.48342
6.000 40.40 11 0.61 0.00026478 1.85 3.96854
7.000 41.17 11 0.61 0.00030889 1.85 4.44889
8.000 41.92 11 0.60 0.00035301 1.85 4.89686
9.000 42.65 11 0.60 0.00039712 1.85 5.31660
10.000 43.38 11 0.59 0.00044122 1.85 5.71163
11.000 44.17 12 0.59 0.00048531 1.85 5.94388
12.000 44.87 12 0.58 0.00052940 1.85 6.27936
13.670 46.02 12 0.58 0.00060300 1.85 6.80239
14.000 46.24 12 0.58 0.00061754 1.85 6.90083
15.000 46.91 12 0.57 0.00066160 1.85 7.19014
16.000 47.67 13 0.57 0.00070564 1.85 7.24605
Vane Angle: 34.0

5.000 39.62 10 0.64 0.00022322 1.89 3.51848
6.000 40.46 11 0.63 0.00026786 1.89 4.00633
7.000 41.24 11 0.63 0.00031249 1.89 4.48968
8.000 41.99 11 0.62 0.00035711 1.89 4.94019
9.000 42.74 11 0.62 0.00040173 1.89 5.36214
10.000 43.47 11 0.61 0.00044635 1.89 5.75912
11.000 44.27 12 0.61 0.00049096 1.89 5.99065
12.000 44.98 12 0.61 0.00053555 1.89 6.32746
13.670 46.14 12 0.60 0.00061001 1.89 6.85246
14.000 46.36 12 0.60 0.00062472 1.89 6.95120
15.000 47.04 12 0.59 0.00066929 1.89 7.24154
16.000 47.70 12 0.59 0.00071384 1.89 7.51933
Vane Angle: 35.0

5.000 39.67 10 0.66 0.00022591 1.94 3.55523
6.000 40.47 10 0.66 0.00027108 1.94 4.09772
7.000 41.31 11 0.65 0.00031626 1.94 4.53227
8.000 42.08 11 0.65 0.00036142 1.94 4.98547
9.000 42.83 11 0.64 0.00040658 1.94 5.40972
10.000 43.57 11 0.64 0.00045174 1.94 5.80871
11.000 44.29 11 0.63 0.00049688 1.94 6.18540
12.000 45.10 12 0.63 0.00054201 1.94 6.37757
13.670 46.27 12 0.62 0.00061737 1.94 6.90454
14.000 46.49 12 0.62 0.00063225 1.94 7.00368
15.000 47.17 12 0.62 0.00067736 1.94 7.29507
16.000 47.84 12 0.61 0.00072246 1.94 7.57384
Vane Angle: 36.0

5.000 39.73 10 0.69 0.00022874 1.99 3.59359
6.000 40.53 10 0.68 0.00027448 1.99 4.14032
7.000 41.38 11 0.68 0.00032022 1.99 4.57679
8.000 42.16 11 0.67 0.00036595 1.99 5.03270
9.000 42.92 11 0.66 0.00041167 1.99 5.45932
10.000 43.67 11 0.66 0.00045739 1.99 5.86041
11.000 44.40 11 0.66 0.00050310 1.99 6.23895
12.000 45.12 11 0.65 0.00054880 1.98 6.59755
13.670 46.40 12 0.64 0.00062510 1.98 6.95884
14.000 46.63 12 0.64 0.00064017 1.98 7.05834
15.000 47.32 12 0.64 0.00068585 1.98 7.35080
16.000 47.99 12 0.64 0.00073151 1.98 7.63057
129
________________________________________________________________________
Vane Angle: 37.0

5.000 39.79 10 0.71 0.00023172 2.04 3.63388
6.000 40.60 10 0.71 0.00027805 2.04 4.18486
7.000 41.40 10 0.70 0.00032438 2.04 4.69710
8.000 42.25 11 0.69 0.00037071 2.04 5.08201
9.000 43.02 11 0.69 0.00041703 2.04 5.51113
10.000 43.78 11 0.68 0.00046334 2.04 5.91432
11.000 44.52 11 0.68 0.00050964 2.04 6.29478
12.000 45.25 11 0.67 0.00055594 2.04 6.65515
13.670 46.54 12 0.67 0.00063323 2.04 7.01536
14.000 46.77 12 0.67 0.00064850 2.04 7.11527
15.000 47.47 12 0.66 0.00069476 2.04 7.40879
16.000 48.15 12 0.66 0.00074102 2.04 7.68961
Vane Angle: 38.0

5.000 39.85 10 0.74 0.00023484 2.09 3.67591
6.000 40.68 10 0.73 0.00028180 2.09 4.23139
7.000 41.48 10 0.73 0.00032876 2.09 4.74750
8.000 42.27 10 0.72 0.00037571 2.09 5.22951
9.000 43.13 11 0.71 0.00042265 2.09 5.56513
10.000 43.89 11 0.71 0.00046959 2.09 5.97058
11.000 44.64 11 0.70 0.00051652 2.09 6.35302
12.000 45.38 11 0.70 0.00056344 2.09 6.71511
13.670 46.69 12 0.69 0.00064177 2.09 7.07420
14.000 46.92 12 0.69 0.00065724 2.09 7.17449
15.000 47.63 12 0.69 0.00070413 2.09 7.46915
16.000 48.32 12 0.68 0.00075100 2.09 7.75101
Vane Angle: 39.0

5.000 39.92 10 0.76 0.00023813 2.15 3.72000
6.000 40.75 10 0.76 0.00028574 2.15 4.28011
7.000 41.57 10 0.75 0.00033335 2.15 4.80009
8.000 42.37 10 0.75 0.00038096 2.15 5.28553
9.000 43.24 11 0.74 0.00042856 2.15 5.62147
10.000 44.01 11 0.73 0.00047615 2.15 6.02918
11.000 44.77 11 0.73 0.00052374 2.15 6.41365
12.000 45.52 11 0.72 0.00057131 2.15 6.77756
13.670 46.73 11 0.72 0.00065074 2.15 7.34553
14.000 46.97 11 0.71 0.00066643 2.15 7.45248
15.000 47.79 12 0.71 0.00071397 2.15 7.53193
16.000 48.49 12 0.71 0.00076151 2.15 7.81492
Vane Angle: 40.0

5.000 39.99 10 0.79 0.00024157 2.22 3.76604
6.000 40.83 10 0.78 0.00028988 2.21 4.33099
7.000 41.66 10 0.78 0.00033818 2.21 4.85509
8.000 42.47 10 0.77 0.00038648 2.21 5.34406
9.000 43.27 10 0.77 0.00043477 2.21 5.80244
10.000 44.14 11 0.76 0.00048305 2.21 6.09025
11.000 44.91 11 0.75 0.00053133 2.21 6.47682
12.000 45.66 11 0.75 0.00057959 2.21 6.84261
13.670 46.89 11 0.74 0.00066017 2.21 7.41331
14.000 47.13 11 0.74 0.00067609 2.21 7.52078
15.000 47.97 12 0.73 0.00072432 2.21 7.59723
16.000 48.67 12 0.73 0.00077254 2.21 7.88132
130
________________________________________________________________________
Vane Angle: 45.0
5.000 40.33 9 0.94 0.00026171 2.60 4.07653
6.000 41.23 9 0.93 0.00031405 2.60 4.69055
7.000 42.12 9 0.92 0.00036637 2.60 5.26089
8.000 43.07 10 0.91 0.00041869 2.60 5.67745
9.000 43.93 10 0.91 0.00047101 2.60 6.15276
10.000 44.76 10 0.90 0.00052331 2.60 6.59932
11.000 45.58 10 0.89 0.00057561 2.60 7.02071
12.000 46.38 10 0.89 0.00062790 2.60 7.41988
13.670 47.82 11 0.88 0.00071520 2.60 7.79599
14.000 48.08 11 0.87 0.00073244 2.60 7.90616
15.000 48.84 11 0.87 0.00078469 2.60 8.23004
16.000 49.59 11 0.86 0.00083694 2.60 8.54005
Vane Angle: 50.0

5.000 40.77 8 1.12 0.00028790 3.15 4.47438
6.000 41.75 8 1.10 0.00034547 3.15 5.15000
7.000 42.81 9 1.09 0.00040303 3.15 5.66972
8.000 43.75 9 1.08 0.00046059 3.15 6.22840
9.000 44.67 9 1.07 0.00051814 3.15 6.75122
10.000 45.57 9 1.06 0.00057568 3.14 7.24291
11.000 46.45 9 1.06 0.00063321 3.14 7.70734
12.000 47.46 10 1.05 0.00069073 3.14 7.89994
13.670 48.87 10 1.04 0.00078676 3.14 8.54403
14.000 49.14 10 1.03 0.00080573 3.14 8.66528
15.000 49.97 10 1.03 0.00086322 3.14 9.02196
16.000 50.77 10 1.02 0.00092068 3.14 9.36355
Vane Angle: 55.0

5.000 41.35 7 1.33 0.00032264 3.95 4.99421
6.000 42.54 8 1.31 0.00038715 3.95 5.65067
7.000 43.60 8 1.30 0.00045166 3.95 6.31909
8.000 44.64 8 1.29 0.00051617 3.95 6.94158
9.000 45.64 8 1.27 0.00058066 3.95 7.52459
10.000 46.63 8 1.26 0.00064514 3.95 8.07331
11.000 47.75 9 1.25 0.00070962 3.95 8.32420
12.000 48.70 9 1.24 0.00077408 3.95 8.78348
13.670 50.24 9 1.23 0.00088170 3.95 9.50056
14.000 50.54 9 1.23 0.00090296 3.95 9.63570
15.000 51.43 9 1.22 0.00096738 3.95 10.03330
16.000 52.30 9 1.21 0.00103178 3.95 10.41446
131
APPENDIX D –
Incompressible 1-D code

APPENDIX D Incompressible 1-D code
________________________________________________________________________
D.1 1-D empirical program
The one dimensional empirical program utilise empirical correlations throughout. The
incompressible pressure drop flow relationships, discussed in Chapter 2, were used to
predict pressure losses and flow distributions.
The one-dimensional incompressible code was coded in Matlab 6 and consists of a

number of functions which will be described in the subdequent section. The initial section
consists of an INPUT function where the primary inputs to the program are defined. The
following function called FLOWCHECK is a function that combines the input function
and the remaining functions. The remaining functions are FLOWSPLITS,
COMBUSTION, ELEMENTS, and HEAT TRANSFER. The function called
FLOWSPITS performs the flowsplit and pressure loss predictions and the function,
COMBUSTION, calculates the gas temperature at each node along the combustor axis.
Combustion efficiency can be accounted for during these calculations. ELEMENTS
calculate the relevant properties at each of the nodes that will be used during the HEAT
TRANSFER calculations. HEAT TRANSFER predicts the heat transfer along the
combustion liner at each node. The following diagram depicts the logic behind the
program.
132
________________________________________________________________________
Inputs
Heat Transfer 1
Flowsplits
2
9
3
8
Flowcheck
7 4
6 5
Elements Combustion
10
Finally Results!!
Diagram A.1: Calculation sequence of one-dimensional empirical model
The numbering in the diagram describes the sequence in which calculations are done.
133
________________________________________________________________________
D.2 INPUTS
This file is used to define the inputs to the program
Define function:
---------------------------------------------------------------------------------------------------------------------------------
function[R,Zones,Nodes,Com,T3,P3,ma,S_A_F,Overall_A_F,Rz_boundary,Pz_boundary,Sz_boundary,Tot
al_comb_length,Hole_type,Number_of_holes,...
Hole_diameter,Hole_Position,Discharge_Coeff,cooling_slot_height,slot_height,Rz_flow,Pz_flow,Sz_flow,
Dz_flow,Liner_Diameter,...
Casing_outer_Diameter,Casing_inner_Diameter,Casing_Diameter,Amb_temp,eLi,eLo,eCi,eCo,sigma,k_L,
k_C,t_L,t_C,Radiation_Model,bb,nn,...
d,s,Nb,nCans,CFD,s_d,steek,qq,Num_elements,DSw_out,DSw_in,BladeTh,NBlades,Theta,K_sw,h0,Outsi
de_conv,Outside_rad,mm,qm]=INPUTS(ii);
---------------------------------------------------------------------------------------------------------------------------------
General inputs:
---------------------------------------------------------------------------------------------------------------------------------
R=287.5; [Gas constant]
Zones=4; [Number of combustion zones]
Nodes=25; [Number of nodes in each zone]
---------------------------------------------------------------------------------------------------------------------------------
Select combustor type:

---------------------------------------------------------------------------------------------------------------------------------
Com=2 [Com = 2; Tubular combustor
Com = 1; Can-annular combustor]
if Com==1
nCans=1;
elseif Com ==2
nCans=6; [Define number of combustion chambers in
can-annular combustion system]
end
---------------------------------------------------------------------------------------------------------------------------------
Inlet conditions:
---------------------------------------------------------------------------------------------------------------------------------
T3=566.15;%Temperature in K [Combustor inlet temperature [K]]
P3=923897;%Pressure in Pa [Combustor inlet pressure [Pa]]
ma=14.25859/nCans;%Total massflow [kg/s] [Total flow rate of combustion system [kg/s]]
S_A_F=14.7;%Stoichiometric air/fuel ratio [Stoichiometric air-fuel ratio]
Overall_A_F=49.75;%Overall air/fuel [Operating air-fuel ratio]
Rz_boundary=75; %[mm] rel to dome [Re-circulation zone boundary [mm]]
Pz_boundary = 95;%[mm] rel to dome [Primary zone boundary [mm]]
Sz_boundary = 245;%[mm] rel to dome [Secondary zone boundary [mm]]
Total_comb_length = 548.5;%[mm] rel to dome [Total combustor length [mm]]
---------------------------------------------------------------------------------------------------------------------------------
Hole layout:
---------------------------------------------------------------------------------------------------------------------------------
[Two options: 1, calculate flow splits(mm=1) or 2, specify flow splits (mm=2)]
134
________________________________________________________________________
if mm==2
[Define individual mass flow rates through hole sets [kg/s]:
----------------------------------------------------------------------------------------------------------
qm=[0.0176925 0.085480296 0.059362303 0.3687117 0.1137038 0.1260885 0.1260885
0.0828009 0.0044821 0.13057065 0.13057065 0.155694 0.0195797 0.11311405 0.11311405
0.11641665 0.11641665 0.254772 0.04118814 0.1028524 0.0771393 0.02059407];
----------------------------------------------------------------------------------------------------------
else
qm=0;
end
Define hole type:

------------------------------------------------------------------------------------------------------------------------------
Hole_type=[1 3 3 4 1 4 4 1 3 4 4 1 3 4 4 4 4 1 3 1 1 3]; 1=Plain holes
2=Plunged holes
3=Splash cooling device
4=Wiggle strip
5=Machined ring
6=Swirler
------------------------------------------------------------------------------------------------------------------------------
Define number of holes in each hole set:
------------------------------------------------------------------------------------------------------------------------------
Number_of_holes=[1 32 8 44 7 22 22 3 4 22 22 4 8 22 22 22 22 4 8 1 1 4];
------------------------------------------------------------------------------------------------------------------------------
Define hole diameters of each hole set [mm]:

------------------------------------------------------------------------------------------------------------------------------
%If holetype == 6 , Hole_diameter = flow area of swirler
Hole_diameter=[7.46 3 5 9 11 9 9 9.02 4.8 9 9 11.254 4.8 9 9 9 9 16 4.8 2 1 4.8]
Note:Not DIAMETER of Wiggle strip, but slot width

-----------------------------------------------------------------------------------------------------------------------------
Define hole set position [mm]:

------------------------------------------------------------------------------------------------------------------------------
Hole_Position=[0 11.889 22.626 27.5 67.23 74.5 86.5 113.2 113.5 119.2 131.2 156.4 158.5 165.5 177.5
231.5 242.5 283 314.5 334.4 334.97 434.5];
------------------------------------------------------------------------------------------------------------------------------
Define hole discharge coefficient – if value is 0, discharge coefficient will be calculated (Set only 0
for Plain and Plunged Holes!!) :
------------------------------------------------------------------------------------------------------------------------------
Discharge_Coeff =[0.569 0.532 0.532 0.655 0.624 0.448 0.448 0.609 0.087 0.465 0.465 0.552 0.191
0.403 0.403 0.416 0.417 0.451 0.406 0.461 0.561 0.418];
------------------------------------------------------------------------------------------------------------------------------
Define cooling slot height [mm]:

---------------------------------------------------------------------------------------------------------------------------------
cooling_slot_height=[2 2 2 2 2 2 2 2 2 2 2 2 2 2 2];
slot_height=[0 2 2 2 0 2 2 0 2 2 2 0 2 2 2 2 2 0 2 0 0 2]; [cooling_slot_height and slot_height are the
same, except for the matrix size]
---------------------------------------------------------------------------------------------------------------------------------
135
________________________________________________________________________
Define air mass fraction into each zone:
---------------------------------------------------------------------------------------------------------------------------------
%Fraction of Massflow into zone
Rz_flow=[1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]; [Recirculation zone]
Pz_flow=[0 0 0 0.333 0.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]; [Primary zone]
Sz_flow=[0 0 0 0.666 0.5 1 1 1 1 0.333 0.333 0.5 1 0 0 0 0 0 0 0 0 0]; [Secondary zone]
Dz_flow=[0 0 0 0 0 0 0 0 0 0.666 0.666 0.5 0 1 1 1 1 1 1 1 1 1]; [Dilution zone]
[r s]=size(Hole_type);
Nb=s; [Number of hole sets]
---------------------------------------------------------------------------------------------------------------------------------
Define liner diameter at each hole set position [mm]:

---------------------------------------------------------------------------------------------------------------------------------
Liner_Diameter=[34 75 75 142 138 142 142 138 138 142 142 138 138 142 142 142 142 138 138 138
138 138];
---------------------------------------------------------------------------------------------------------------------------------
Define casing diameter at each hole set position [mm]:

---------------------------------------------------------------------------------------------------------------------------------
if Com ==1 [Tubular combustor]
Casing_Diameter=[191.3287 191.3287 191.3287 191.3287 191.3287 191.3287 191.3287 191.3287
191.3287 191.3287 191.3287 191.3287 191.3287 191.3287 191.3287 191.3287
191.3287 191.3287 191.3287 191.3287 191.3287 191.3287 ];
Casing_outer_Diameter=493*(zeros(1,Nb));
Casing_inner_Diameter=153*(zeros(1,Nb));
elseif Com ==2 [Can-annular

combustor]
Casing_Diameter=zeros(1,Nb);
Casing_outer_Diameter=493*(ones(1,Nb)); %Casing outer diameter at hole
Casing_inner_Diameter=153*(ones(1,Nb)); %Casing inner diameter at hole
end
---------------------------------------------------------------------------------------------------------------------------------
Swirler data:
---------------------------------------------------------------------------------------------------------------------------------
DSw_out=23.09; [Swirler outer diameter
[mm]]
DSw_in=13.4; [Swirler inner diameter
[mm]]
BladeTh=1; [Blade thickness [mm]]
NBlades=12; [Number of blades]
Theta=50; [Blade angle]
K_sw=1.3; [Blade loss coeff.]
---------------------------------------------------------------------------------------------------------------------------------
Heat transfer:
---------------------------------------------------------------------------------------------------------------------------------
Amb_temp=25; [Ambient temp [ºC]]
eLi=0.4; [Liner inner emisivity]
eLo=0.8; [Liner outer emisivity]
eCi=0.8; [Casing inner emisivity]
eCo=0.8; [Casing outer emisivity]
136
________________________________________________________________________
sigma=5.67e-8; [Stefan-Boltzmann
constant]
k_L=16.9; [Liner wall conductivity]
k_C=16.9; [Casing wall
conductivity]
t_L=1; [Liner wall thickness
[mm]]
t_C=1; [Casing wall thickness
[mm]]
[Radiation Model (1- Simple, 2-Accurate, 3-Off)]
Radiation_Model = 2;
[Outside Convection (1- Default convection on, 2- Define outside heat transfer coeff, 3- Outside
convection off)]
Outside_conv = 3; %1- Default convection on , 2- Define outside heat transfer coeff , 3- Outside
convection off
[Outside radiation (1-Outside radiation on, 2- Outside radiation off)]

Outside_rad = 2; % 1- Outside Radiation on , 2- Outside Radiation off
if Outside_conv ==2
h0=input('Define casing outside heat transfer coefficeint : ');
else
h0=0;
end
---------------------------------------------------------------------------------------------------------------------------------
Properties:
---------------------------------------------------------------------------------------------------------------------------------
[Turn on/off Film cooling (1-on, 2-off)]
bb=1;
[Account for combustion efficiency (1-on, 2-off)]

nn=1;%input('Do you want to incorp. combustion eff ? (Y/N)-(1/2) ');
---------------------------------------------------------------------------------------------------------------------------------
Network Solver:
---------------------------------------------------------------------------------------------------------------------------------
[Print network data file (1-yes, 2-no)]
qq=1; %Do you want to print the network data file? (Y/N)-(1/2)
[Define number of elements between nodes (matrix size same as Nb)]

Num_elements=[4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 4 5];
---------------------------------------------------------------------------------------------------------------------------------
**************************************************************************************
---------------------------------------------------------------------------------------------------------------------------------
137
________________________________________________________________________
D.3 FLOWCHECK
Function:
---------------------------------------------------------------------------------------------------------------------------------
[R,Zones,Nodes,Com,T3,P3,ma,S_A_F,Overall_A_F,Rz_boundary,Pz_boundary,Sz_boundary,Total_com
b_length,Hole_type,Number_of_holes,Hole_diameter,Hole_Position,Discharge_Coeff,cooling_slot_height,
slot_height,Rz_flow,Pz_flow,Sz_flow,Dz_flow,Liner_Diameter,Casing_outer_Diameter,Casing_inner_Dia
meter,Casing_Diameter,Amb_temp,eLi,eLo,eCi,eCo,sigma,k_L,k_C,t_L,t_C,Radiation_Model,bb,nn,d,s,N
b,nCans,CFD,s_d,steek,qq,Num_elements,DSw_out,DSw_in,BladeTh,NBlades,Theta,K_sw,h0,Outside_co
nv,Outside_rad,mm,qm]=INPUTS(ii);
---------------------------------------------------------------------------------------------------------------------------------
Convert data to [m]:

---------------------------------------------------------------------------------------------------------------------------------
Rz_boundary=Rz_boundary/1000;
Pz_boundary=Pz_boundary/1000;
Sz_boundary=Sz_boundary/1000;
Total_comb_length=Total_comb_length/1000;
t_L=t_L/1000;
t_C=t_C/1000;
---------------------------------------------------------------------------------------------------------------------------------
Calculate total number of nodes in computational domain:

---------------------------------------------------------------------------------------------------------------------------------
Number_of_nodes=(Zones*Nodes);%-(Zones-1);
---------------------------------------------------------------------------------------------------------------------------------
Convert ambient temperature from ºC to K:

---------------------------------------------------------------------------------------------------------------------------------
Tamb=Amb_temp + 273.15; %Ambient temp in [K]
---------------------------------------------------------------------------------------------------------------------------------
Determine how many hole features are cooling slots:

---------------------------------------------------------------------------------------------------------------------------------
Number_of_cooling_slots=0;
for i=1:Nb
Holetype=Hole_type(i);
if Holetype == 3 %Splash Cooling
Number_of_cooling_slots = Number_of_cooling_slots + 1;
end
if Holetype == 4 % Wiggle Strip
end
if Holetype == 5 % Machined Ring
end
end
if Number_of_cooling_slots == 0
bb=2;
end
--------------------------------------------------------------------------------------------------------------------------------
138
________________________________________________________________________
Calculate annulus air flow area and effective casing diameter for can-annular combustor:
---------------------------------------------------------------------------------------------------------------------------------
if Com == 2
for i=1:Nb
Aflow=((0.25*pi)*(Casing_outer_Diameter(i)^2 - Casing_inner_Diameter(i)^2) –
(nCans*0.25*pi*(Liner_Diameter(i)^2)))/nCans;
Casing_Diameter(i)=sqrt((4/pi)*Aflow + Liner_Diameter(i)^2);
end
end
---------------------------------------------------------------------------------------------------------------------------------
Calculate zone lengths:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Zones
if i==1
Length_zone(i)=Rz_boundary;
end
if i==2
Length_zone(i)=Pz_boundary-Rz_boundary;
end;
if i==3
Length_zone(i)=Sz_boundary-Pz_boundary;
end
if i==4
Length_zone(i)=Total_comb_length-Sz_boundary;
end
end
for i=1:Zones
if i==1
Length(i)=Length_zone(i);
else
Length(i)=Length(i-1)+Length_zone(i);
end
end
--------------------------------------------------------------------------------------------------------------------------------
Initial conditions for solver:

---------------------------------------------------------------------------------------------------------------------------------
T_annulus=(ones(1,Number_of_nodes)*T3); [Initial Annulus air temperature [K]]
T_GAS=(ones(1,Number_of_nodes)*T3); [Initial Gas temperature [K]]
dp_hot=0; [Initial Hot losses]
---------------------------------------------------------------------------------------------------------------------------------
LOOP CALCULATIONS START

---------------------------------------------------------------------------------------------------------------------------------
Number_of_iter=1000;
ntel=0;
Converg=10;
while (Converg > 1e-6) & (ntel < Number_of_iter)

ntel=ntel+1
139
________________________________________________________________________
for i=1:Number_of_nodes
rho_annulus(i)=P3/(R*T_annulus(i)); [Annulus air density [kg/m^3]]
rho_gas(i)=P3/(R*T_GAS(i)); [Gas density [kg/m^3]]
end
Define function for “Flowsplit” subroutine:

---------------------------------------------------------------------------------------------------------------------------------
Pressure_drop,Flowsplits,jetangle,X,U_ref,rho_hole,element_th,Tgx,q_ref,Dref,rho_ref,A_L]=Flowsplit
(Number_of_cooling_slots,Hole_type,Number_of_holes,Hole_diameter,NBlades,BladeTh,K_sw,Theta,Nb,
R,ma,P3,T3,Com,DSw_out,DSw_in,Hole_Position,Nodes,T_annulus,Number_of_nodes,Length,rho_annul
us,rho_gas,dp_hot,Zones,Casing_Diameter,Liner_Diameter,t_L,Discharge_Coeff,slot_height,nCans,Casing
_outer_Diameter,Casing_inner_Diameter,mm,qm);
---------------------------------------------------------------------------------------------------------------------------------
Define function for “Combustion subroutine”:

---------------------------------------------------------------------------------------------------------------------------------
[Tgas,Air_fuel_node,Length,Air_fuel,Equivalence_Ratio,Casing_D,Liner_D,Airflow,Airflow_node,Airflo
w_zone,eta_Rz,eta_Pz,eta_Sz,eta_Dz,theta,EI_CO,EI_HC,eta_comb,Nox,NOx,Comb_eff]=Combustion(T
3,P3,Nb,ma,x,y,Hole_Position,Massflow,Overall_A_F,S_A_F,Total_comb_length,Pz_boundary,Rz_bound
ary,Sz_boundary,nn,Casing_Diameter,Liner_Diameter,Number_of_nodes,Nodes,X,Zones,Length,Hole_typ
e,Rz_flow,Pz_flow,Sz_flow,Dz_flow,Pressure_drop,element_th,nCans,A_ref,Dref);
---------------------------------------------------------------------------------------------------------------------------------
Calculate fuel flow and fuel-air ratio at each computational node:

---------------------------------------------------------------------------------------------------------------------------------
Fuel_flow=(ma/Overall_A_F);
FAR(i)=1/Air_fuel_node(i);
end
---------------------------------------------------------------------------------------------------------------------------------
Calculate annulus air and gas properties as a function of temperature:

---------------------------------------------------------------------------------------------------------------------------------
Ta=T_annulus(i)-273;
Tg=T_GAS(i)-273;
k_a(i)=(2.4186e-2)+((7.8957e-5)*Ta)-((3.7873e-8)*Ta^2)+((1.6685e-11)*Ta^3);
k_g(i)=(2.4186e-2)+((7.8957e-5)*Tg)-((3.7873e-8)*Tg^2)+((1.6685e-11)*Tg^3);
mu_a(i)=(1.733e-5)+((4.5543e-8)*Ta)-((2.2106e-11)*Ta^2)+((7.7219e-15)*Ta^3);
mu_g(i)=(1.733e-5)+((4.5543e-8)*Tg)-((2.2106e-11)*Tg^2)+((7.7219e-15)*Tg^3);
Cp_an(i)=1002.106 + (0.089329*Ta) + ((2.332e-4)*Ta^2) - ((1.2831e-7)*Ta^3);

Cp_gas(i)=1002.106 + (0.089329*Tg) + ((2.332e-4)*Tg^2) - ((1.2831e-7)*Tg^3);
Gamma_air(i)=Cp_an(i)/(Cp_an(i)-R);
Pr_air(i)=(Cp_an(i)*mu_a(i))/k_a(i);
Pr_gas(i)=(Cp_gas(i)*mu_g(i))/k_g(i);
end
---------------------------------------------------------------------------------------------------------------------------------
140
________________________________________________________________________
Define function for “Elements:
---------------------------------------------------------------------------------------------------------------------------------
[Momflux,T_GAS,m_annulus,Hyd_D,dp_hot,Reo,Rei,d_AL,d_ACi,ALi,ALo,Aci,Aco,m_internal,Pattern_
Factor,M_inner,Local_AF,Y_max]=Elements(Zones,Length,Tgas,X,Number_of_nodes,Rz_boundary,Pz_b
oundary,Sz_boundary,Total_comb_length,Massflow,ma,Hole_Position,Nb,Fuel_flow,Airflow_node,Casin
g_D,Liner_D,t_L,t_C,P3,R,T3,U_ref,mu_a,mu_g,element_th,Nodes,Tgx,nCans,q_ref,Pressure_drop,Dref,
Overall_A_F,Liner_Diameter,rho_hole_gas,rho_annulus,V_jet,Cd,Hole_diameter,Com,Number_of_holes);
---------------------------------------------------------------------------------------------------------------------------------
Define function for “Heat_transfer “:

---------------------------------------------------------------------------------------------------------------------------------
[Liner_temp,Casing_temp,New_Tannulus,old_Tannulus,eta_filmcooling,J,distance_downstream_slot,e_g,
Average_haLo]=Heat_transfer(Zones,Nodes,Number_of_nodes,ma,Overall_A_F,T_GAS,R,Liner_D,Casin
g_D,element_th,x,y,Hole_type,Massflow,Number_of_cooling_slots,cooling_slot_height,mu_a,Pz_boundar
y,k_g,T_annulus,sigma,k_L,k_C,t_L,t_C,Airflow_node,Nb,X,mu_g,k_a,Hyd_D,m_annulus,d_AL,d_ACi,T
3,Cp_an,bb,Reo,Rei,Pr_air,Pr_gas,Fuel_flow,FAR,P3,Hole_Position,Tamb,ALi,ALo,Aci,Aco,eLi,eLo,eCi,
eCo,Total_comb_length,Radiation_Model,m_internal,Liner_Diameter,nCans,M_inner,Local_AF,h0,Outsid
e_conv,Outside_rad);
---------------------------------------------------------------------------------------------------------------------------------
Check for convergence:

---------------------------------------------------------------------------------------------------------------------------------
Converg=norm((New_Tannulus-old_Tannulus)/New_Tannulus);
T_annulus=New_Tannulus;
end [End Loop calculations]

---------------------------------------------------------------------------------------------------------------------------------
Define function for “netw “:

---------------------------------------------------------------------------------------------------------------------------------
if qq == 1
[element_length]=netw(Num_elements,Nb,Hole_Position,Total_comb_length,Number_of_nodes,X,Line
r_D,Casing_D,T_GAS,Liner_temp,t_C,t_L,e_g,slot_height);
end
---------------------------------------------------------------------------------------------------------------------------------
Calculates individual cooling mass flow rates:

---------------------------------------------------------------------------------------------------------------------------------
Cooling_massflow=0;
for i=1:Nb
if Holetype == 3
Cooling_massflow=Cooling_massflow+Massflow(i);
elseif Holetype == 4
elseif Holetype == 5
end
end
141
________________________________________________________________________
Per_cooling_massflow=(Cooling_massflow/ma)*100; [Total % cooling flow]
---------------------------------------------------------------------------------------------------------------------------------
WRITE OUTPUT DATA FILES

---------------------------------------------------------------------------------------------------------------------------------
if Com == 2
Comtype = 'Combustor type : Can - Annular';
elseif Com == 1
Comtype = 'Combustor type : Tubular ';
end
if Radiation_Model==1
Radm='Simple';
elseif Radiation_Model ==2
Radm='Accurate';
elseif Radiation_Model == 3
Radm='Off';
end
if bb==1
Cooling='Film Cooled Liner';
elseif bb==2
Cooling='Uncooled Liner';
end
nCann='Number of cans: ';

HoleType='Hole type ';
NHoles='nHoles';
DHole='HDiam [mm]';
Hpos='XPos [mm]';
MF='m/can [kg/s]';
mf='mdot [%]';
Disch='Cd';
Jet='Vjet [m/s]';
Mom='J';
Jangle='Jet ang';
ymax='Y_max [mm]';
Inp=' ********* INPUTS ********* ';

InletP='Inlet Pressure [Pa] ';
InletT='Inlet Temperature [K] ';
Mass='Total Massflow [kg/s] ';
MC='Total massflow per Can [kg/s] ';
Rz='Rz boundary (rel to dome) [mm] ';
Pz='Pz boundary (rel to dome) [mm] ';
Sz='Sz boundary (rel to dome) [mm] ';
TL='Total Combustor Length [mm] ';
Inh='**** HEAT TRANSFER INPUTS **** ';
Lcmodel='Radiation Model ';
Frmodel='Flame radiation modes ';
AmbT='Ambient Temperature [C] ';
Lie='Liner inner emissivity ';
Loe='Liner outer emissivity ';
Cie='Casing inner emissivity ';
Coe='Casing outer emissivity ';
142
________________________________________________________________________
Lthe='Liner thermal conductivuty [W/mK] ';
Cthe='Casing thermal conductivity [W/mK]';
Ld='Liner wall thickness [mm] ';
CD='Casing wall thickness [mm] ';
AF='Overall Air/Fuel Ratio ';
SAF='Stoichoimetric Air/Fuel Ratio ';
Resu='******** RESULTS ******** ';
pres='Pressure Loss [Pa] ';
Persp='Pressure Loss [%] ';
PF='Pattern Factor ';
th='Theta (*10^6) ';
ALiner='Flametube flow area [m^2] ';
Ref_A='Reference Area (Aref) ';
AL_Aref='(Flametube flow area)/(Ref Area)';
Ref_D='Reference Diameter (Dref) ';
Ref_p='Reference density ';
Ref_u='Reference Velocity (Uref) ';
Ref_q='Reference velocity head (qref) ';
CO='EI_CO [g CO/kg fuel] ';
HC='EI_HC [g HC/kg fuel] ';
Nox_L=['EI_NOx(Lefebvre Model) [g/kg fuel]'];
Nox_O=['EI_NOx(Odgers Model) [g/kg fuel] '];
Combeff='Combustion efficiency (figure 2.2)';
eCom='Emission Comb. efficiency [%] ';
Coolingflow='Total amount of cooling air [%] ';
%************** Zone Data *******************
Airflowrz='Mass flow rate in Rz per Can [kg/s]';
Airflowpz='Mass flow rate in Pz per Can [kg/s]';
Airflowsz='Mass flow rate in Sz per Can [kg/s]';
Airflowdz='Mass flow rate in Dz per Can [kg/s]';
airflowrz='Percentage of total flow in Rz [%] ';
airflowpz='Percentage of total flow in Pz [%] ';
airflowsz='Percentage of total flow in Sz [%] ';
airflowdz='Percentage of total flow in Dz [%] ';
Rz_phi='Rz equivalence ratio ';
Pz_phi='Pz equavalence ratio ';
Sz_phi='Sz equavalence ratio ';
Dz_phi='Dz equavalence ratio ';
Rz_eff='Rz combustion efficiency [%] ';
Pz_eff='Pz combustion efficiency [%] ';
Sz_eff='Sz combustion efficiency [%] ';
Dz_eff='Dz combustion efficiency [%] ';
hLocoeff='Average Backside h_coeff ';
Rz_frac='Rz_f';
Pz_frac='Pz_f';
Sz_frac='Sz_f';
Dz_frac='Dz_f';
%%%%%%%%%%%%%%%%%%%% FLOW %%%%%%%%%%%%%%%%%%%%%%
linerd='[Liner_D]';
casingd='[Casing_D]';
x_distance='[X (mm)]';
int_flow_area='[Int_A]';
an_flow_area='[AA] ';
Per='[P]';
Hydrou='[Hyd_D]';
Gasemm='[eg]';
143
________________________________________________________________________
Gas_temp='[Tg(C)]';
Linertemp='[Liner temp [K]]';
AnnulusM='Mannulus';
Liner='LinerD';
if Com==2
lod='LOD [mm]';
cid='CID [mm]';
cod='COD [mm]';
elseif Com ==1
lod='LOD [mm]';
cd='COD [mm]';
end
%*****************************GENDATA.TXT****************************
fid=fopen('Gendata.txt','w');
fprintf(fid,'%s\n',Comtype);
fprintf(fid,'\n');
fprintf(fid,'%s ',nCann);
fprintf(fid,'%6.1f\n ',nCans)
fprintf(fid,'\n');
%***********INPUTS*********
fprintf(fid,'%s\n',Inp);
fprintf(fid,'%s',InletP);
fprintf(fid,'%6.2f\n',P3);
fprintf(fid,'%s',InletT);
fprintf(fid,'%6.2f\n',T3);
fprintf(fid,'%s',Mass);
fprintf(fid,'%6.3f\n',ma*nCans);
if Com ==2
fprintf(fid,'%s',MC);
fprintf(fid,'%6.3f\n',ma);
end
fprintf(fid,'%s',AF);
fprintf(fid,'%6.3f\n',Overall_A_F);
fprintf(fid,'%s',SAF);
fprintf(fid,'%6.3f\n',S_A_F);
fprintf(fid,'%s',Rz);
fprintf(fid,'%6.3f\n',Rz_boundary*1000);
fprintf(fid,'%s',Pz);
fprintf(fid,'%6.3f\n',Pz_boundary*1000);
fprintf(fid,'%s',Sz);
fprintf(fid,'%6.3f\n',Sz_boundary*1000);
fprintf(fid,'%s',TL);
fprintf(fid,'%6.3f\n',Total_comb_length*1000);
fprintf(fid,'\n');
%******* HEAT TRANSFER *******

fprintf(fid,'%s\n',Inh);
fprintf(fid,'%s',Lcmodel);
fprintf(fid,'%6s\n',Radm);
fprintf(fid,'%s',Frmodel);
fprintf(fid,'%6s\n',Cooling);
fprintf(fid,'%s',AmbT);
144
________________________________________________________________________
fprintf(fid,'%6.1f\n',Tamb-273.15);
fprintf(fid,'%s',Lie);
fprintf(fid,'%6.2f\n',eLi);
fprintf(fid,'%s',Loe);
fprintf(fid,'%6.2f\n',eLo);
fprintf(fid,'%s',Cie);
fprintf(fid,'%6.2f\n',eCi);
fprintf(fid,'%s',Coe);
fprintf(fid,'%6.2f\n',eCo);
fprintf(fid,'%s',Lthe);
fprintf(fid,'%6.2f\n',k_L);
fprintf(fid,'%s',Cthe);
fprintf(fid,'%6.2f\n',k_C);
fprintf(fid,'%s',Ld);
fprintf(fid,'%6.2f\n',t_L*1000);
fprintf(fid,'%s',CD);
fprintf(fid,'%6.2f\n',t_C*1000);
fprintf(fid,'\n');
%**********RESULTS***********
fprintf(fid,'%s\n',Resu);
fprintf(fid,'%s',pres);
fprintf(fid,'%6.3f\n',Pressure_drop);
fprintf(fid,'%s',Persp);
fprintf(fid,'%6.3f\n',Pressure_Loss);
fprintf(fid,'%s',PF);
fprintf(fid,'%6.3f\n',Pattern_Factor);
fprintf(fid,'%s',th);
fprintf(fid,'%6.3f\n',theta);
fprintf(fid,'%s',ALiner);
fprintf(fid,'%6.3f\n',A_L);
fprintf(fid,'%s',Ref_A);
fprintf(fid,'%6.2f\n',A_ref);
fprintf(fid,'%s',AL_Aref);
fprintf(fid,'%6.3f\n',A_L/A_ref);
fprintf(fid,'%s',Ref_D);
fprintf(fid,'%6.3f\n',Dref);
fprintf(fid,'%s',Ref_p);
fprintf(fid,'%6.3f\n',rho_ref);
fprintf(fid,'%s',Ref_u);
fprintf(fid,'%6.3f\n',U_ref);
fprintf(fid,'%s',Ref_q);
fprintf(fid,'%6.3f\n',q_ref);
fprintf(fid,'%s',CO);
fprintf(fid,'%6.3f\n',EI_CO);
fprintf(fid,'%s',HC);
fprintf(fid,'%6.3f\n',EI_HC);
fprintf(fid,'%s',Nox_L);
fprintf(fid,'%6.3f\n',Nox);
fprintf(fid,'%s',Nox_O);
fprintf(fid,'%6.3f\n',NOx);
fprintf(fid,'%s',Combeff);
fprintf(fid,'%6.3f\n',Comb_eff);
fprintf(fid,'%s',eCom);
fprintf(fid,'%6.3f\n',eta_comb);
fprintf(fid,'%s',Coolingflow);
145
________________________________________________________________________
fprintf(fid,'%6.3f\n',Per_cooling_massflow);
fprintf(fid,'%s',hLocoeff);
fprintf(fid,'%6.3f\n',Average_haLo);
fprintf(fid,'\n');
fclose(fid);
%************************** HOLEDATA 1/2.TXT************************

fid=fopen('Holedata1_2.txt','w');
fprintf(fid,'%s',HoleType);
fprintf(fid,'%14s',Hpos);
fprintf(fid,'%8s',NHoles);
fprintf(fid,'%13s',DHole);
fprintf(fid,'%13s',MF);
fprintf(fid,'%13s',mf);
fprintf(fid,'%8s',Disch);
fprintf(fid,'%15s',Jet);
fprintf(fid,'\n');
for i=1:Nb
split=Flowsplits(i);
massf=Massflow(i);
hpos=Hole_Position(i);
hdiam=Hole_diameter(i);
nHoles=Number_of_holes(i);
Holetype = Hole_type(i);
vjet=V_jet(i);
Coeff=Cd(i);
if Holetype == 1
Htype='Plain ';
end
if Holetype== 2
Htype ='Plunged ';
end
if Holetype == 3
Htype = 'Splash Cooling ';
end
if Holetype ==4
Htype='Wiggle Strip ';
end
if Holetype ==5
Htype='Machined Ring ';
end
if Holetype==6
Htype = 'Swirler ';
end
fprintf(fid,'%s',Htype);
fprintf(fid,'%10.1f',hpos);
fprintf(fid,'%11.1f',nHoles);
fprintf(fid,'%11.2f',hdiam);
fprintf(fid,'%13.5f',massf);
fprintf(fid,'%14.3f',split);
146
________________________________________________________________________
fprintf(fid,'%11.3f',Coeff);
fprintf(fid,'%11.3f',vjet);
fprintf(fid,'\n');
end
fclose(fid);
%%%%%%%%%%% Holedata 2/2 %%%%%%%%%%%

fid=fopen('Holedata2_2.txt','w');
fprintf(fid,'%s',HoleType);
fprintf(fid,'%7s',Mom);
fprintf(fid,'%12s',Jangle);
fprintf(fid,'%13s',ymax);
if Com ==2
fprintf(fid,'%11s',lod);
fprintf(fid,'%11s',cid);
fprintf(fid,'%12s',cod);
elseif Com==1
fprintf(fid,'%11s',lod);
fprintf(fid,'%12s',cd);
end
fprintf(fid,'%6s',Rz_frac);
fprintf(fid,'%7s',Pz_frac);
fprintf(fid,'%7s',Sz_frac);
fprintf(fid,'%7s',Dz_frac);
fprintf(fid,'\n');
for i=1:Nb
if Com==2
LOD= Liner_Diameter(i);
CID=Casing_inner_Diameter(i);
COD=Casing_outer_Diameter(i);
elseif Com==1
LOD= Liner_Diameter(i);
CD=Casing_Diameter(i);
end
Holetype = Hole_type(i);
momflux=Momflux(i);
jangle=jetangle(i);
yMax=Y_max(i);
rz_flow=Rz_flow(i);
pz_flow=Pz_flow(i);
sz_flow=Sz_flow(i);
dz_flow=Dz_flow(i);
if Holetype == 1
Htype='Plain ';
147
________________________________________________________________________
end
if Holetype== 2
Htype ='Plunged ';
end
if Holetype == 3
Htype = 'Splash Cooling';
end
if Holetype ==4
Htype='Wiggle Strip ';
end
if Holetype ==5
Htype='Machined Ring ';
end
if Holetype==6
Htype = 'Swirler ';
end
fprintf(fid,'%s',Htype);
fprintf(fid,'%7.3f',momflux);
fprintf(fid,'%11.3f',jangle);
fprintf(fid,'%11.3f',yMax);
if Com ==2
fprintf(fid,'%10.2f',LOD);
fprintf(fid,'%12.2f',CID);
fprintf(fid,'%12.2f',COD);
elseif Com ==1
fprintf(fid,'%10.2f',LOD);
fprintf(fid,'%12.2f',CD);
end
fprintf(fid,'%8.2f',rz_flow);
fprintf(fid,'%7.2f',pz_flow);
fprintf(fid,'%7.2f',sz_flow);
fprintf(fid,'%7.2f',dz_flow);
fprintf(fid,'\n');
end
fclose(fid);
%********************* ZONE Data *************************8

fid =fopen('Zone_Data.txt','w');
fprintf(fid,'%s',Airflowrz);
fprintf(fid,'%6.3f',Airflow_zone(1));
fprintf(fid,'\n');
fprintf(fid,'%s',Airflowpz);
fprintf(fid,'\n');
fprintf(fid,'%s',Airflowsz);
fprintf(fid,'\n');
fprintf(fid,'%s',Airflowdz);
148
________________________________________________________________________
fprintf(fid,'\n');
fprintf(fid,'\n');
fprintf(fid,'%s',airflowrz);
fprintf(fid,'%6.3f',(Airflow_zone(1)/ma)*100);
fprintf(fid,'\n');
fprintf(fid,'%s',airflowpz);
fprintf(fid,'\n');
fprintf(fid,'%s',airflowsz);
fprintf(fid,'\n');
fprintf(fid,'%s',airflowdz);
fprintf(fid,'\n');
fprintf(fid,'\n');
fprintf(fid,'%s',Rz_phi);
fprintf(fid,'%6.3f',Equivalence_Ratio(1));
fprintf(fid,'\n');
fprintf(fid,'%s',Pz_phi);
fprintf(fid,'\n');
fprintf(fid,'%s',Sz_phi);
fprintf(fid,'\n');
fprintf(fid,'%s',Dz_phi);
fprintf(fid,'\n');
fprintf(fid,'\n');
fprintf(fid,'%s',Rz_eff);
fprintf(fid,'%6.3f',eta_Rz*100);
fprintf(fid,'\n');
fprintf(fid,'%s',Pz_eff);
fprintf(fid,'%6.3f',eta_Pz*100);
fprintf(fid,'\n');
fprintf(fid,'%s',Sz_eff);
fprintf(fid,'%6.3f',eta_Sz*100);
fprintf(fid,'\n');
fprintf(fid,'%s',Dz_eff);
fprintf(fid,'%6.3f',eta_Dz*100);
fprintf(fid,'\n');
fclose(fid);
%%%%%%%%%%%%% FLOW %%%%%%%%%%%%%%%
fid=fopen('Flow.txt','w');
fprintf(fid,'%s',linerd);
fprintf(fid,'%12s',casingd);
fprintf(fid,'%8s',x_distance);
fprintf(fid,'%10s',int_flow_area);
fprintf(fid,'%10s',an_flow_area);
fprintf(fid,'%8s',Per);
fprintf(fid,'%10s',Hydrou);
fprintf(fid,'%10s',Gas_temp);
fprintf(fid,'%6s',Gasemm);
149
________________________________________________________________________
fprintf(fid,'%20s',Linertemp);
fprintf(fid,'\n');
LD=Liner_D(i);
Cas=Casing_D(i);
xdis=X(i);
I_flow=(pi*Liner_D(i)^2)/4;
an_flow=(pi/4)*(Casing_D(i)^2-Liner_D(i)^2);
perim=pi*(Casing_D(i)+Liner_D(i));
Hydroulic_D=(4*an_flow)/perim;
Gas_emm=e_g(i);
GasTemp=T_GAS(i)-273.15;
Liner_T=Liner_temp(i)-273.15;
fprintf(fid,'%6.3f',LD);
fprintf(fid,'%12.3f',Cas);
fprintf(fid,'%10.4f',xdis*1000);
fprintf(fid,'%10.4f',I_flow);
fprintf(fid,'%10.3f',an_flow);
fprintf(fid,'%10.3f',perim);
fprintf(fid,'%8.3f',Hydroulic_D);
fprintf(fid,'%10.3f',GasTemp);
fprintf(fid,'%10.4f',Gas_emm);
fprintf(fid,'%10.3f',Liner_T);
fprintf(fid,'\n');
end
fclose(fid);
--------------------------------------------------------------------------------------------------------------------------------
**************************************************************************************
---------------------------------------------------------------------------------------------------------------------------------
D.4 FLOWSPLITS
Define function for “netw “:

------------------------------------------------------------------------------------------------------------
function[rho_hole_gas,V_jet,A_ref,Pressure_Loss,RefPressure_Loss,Cd,Massflow,Pressure_drop,Flowspli
ts,jetangle,X,U_ref,rho_hole,element_th,Tgx,q_ref,Dref,rho_ref,A_L]=Flowsplit(Number_of_cooling_slots
,Hole_type,Number_of_holes,Hole_diameter,NBlades,BladeTh,K_sw,Theta,Nb,R,ma,P3,T3,Com,DSw_ou
t,DSw_in,Hole_Position,Nodes,T_annulus,Number_of_nodes,Length,rho_annulus,rho_gas,dp_hot,Zones,C
asing_Diameter,Liner_Diameter,t_L,Discharge_Coeff,slot_height,nCans,Casing_outer_Diameter,Casing_i
nner_Diameter,mm,qm);
--------------------------------------------------------------------------------------------------------------------------------
Convert units from mm to m:

---------------------------------------------------------------------------------------------------------------------------------
Hole_Position=Hole_Position/1000;
Hole_diameter=Hole_diameter/1000;
Liner_Diameter=Liner_Diameter/1000;
Casing_Diameter=Casing_Diameter/1000;
slot_height=slot_height/1000;
Casing_inner_Diameter=Casing_inner_Diameter/1000;
150
________________________________________________________________________
Casing_outer_Diameter=Casing_outer_Diameter/1000;
DSw_out=DSw_out/1000;
DSw_in=DSw_in/1000;
BladeTh=BladeTh/1000;
---------------------------------------------------------------------------------------------------------------------------------
Define end node at each zone:

---------------------------------------------------------------------------------------------------------------------------------
Tgx=zeros(1,Zones+1);
Tgx(1)=Hole_Position(1);
Tgx(2:Zones+1)=Length(1:Zones);
---------------------------------------------------------------------------------------------------------------------------------
Calculate position of each computational node:

---------------------------------------------------------------------------------------------------------------------------------
X=zeros(1,Number_of_nodes);
for i=1:Zones
if i==1
dx=(Tgx(i+1)-Tgx(i))/(Nodes-1);
for j=2:(i*Nodes)
X(1)=Tgx(1);
X(j)=X(j-1)+dx;
end
end
if i == 2
dx=(Tgx(i+1)-Tgx(i))/Nodes;
for j=(i-1)*(Nodes)+1 : (i*Nodes)
X(j)=X(j-1)+dx;
end
end
if i == 3
X(j)=X(j-1)+dx;
end
end
if i == 4
X(j)=X(j-1)+dx;
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
Generate density matrix for each computational node:
---------------------------------------------------------------------------------------------------------------------------------
rho_nodes=zeros(1,Number_of_nodes);
---------------------------------------------------------------------------------------------------------------------------------
151
________________________________________________________________________
Calculate distance between node- defining elements:
---------------------------------------------------------------------------------------------------------------------------------
element_th=zeros(1,Number_of_nodes);
for i=1:Number_of_nodes-1
element_th(i)=X(i+1)-X(i);
element_th(Number_of_nodes) = element_th(Number_of_nodes -1);
end
---------------------------------------------------------------------------------------------------------------------------------
Define air and gas density at each hole set:

---------------------------------------------------------------------------------------------------------------------------------
for j=1:Nb
if (j <= Nb)
if (X(i) <= Hole_Position(j)+(element_th(i))) & (X(i) >= Hole_Position(j)-(element_th(i)))

rho_hole(j)=rho_annulus(i);
rho_hole_gas(j)=rho_gas(i);
end
end
end
end
rho_ref=P3/(R*T3); [Reference density]

-------------------------------------------------------------------------------------------------------------------------------
Qmass=ma; [Total mass flow rate]
Define comnustor type and calculate reference data for selected combustor type:
---------------------------------------------------------------------------------------------------------------------------------
Liner_Diam=sum(Liner_Diameter)/Nb;%Average Liner diameter
Dref=0;
A_ref=0;
for i=1:Nb
if Com ==1 %Tubular

AreaAnnulus(i) = ((pi*0.25)*((Casing_Diameter(i)^2)-(Liner_Diameter(i)^2)));
Dref=Dref+Casing_Diameter(i);
A_ref=A_ref+0.25*pi*Casing_Diameter(i)^2;
elseif Com ==2 % Can-annular
AreaAnnulus(i)= (((0.25*pi)*(Casing_outer_Diameter(i)^2-Casing_inner_Diameter(i)^2)) -
((0.25*pi*Liner_Diameter(i)^2)*nCans))/nCans;
Dref=Dref+(Casing_outer_Diameter(i)-Casing_inner_Diameter(i))/2;
A_ref=A_ref+0.25*pi*(Casing_outer_Diameter(i)^2-Casing_inner_Diameter(i)^2);
end
end
A_L=(pi*(Liner_Diam^2)*0.25);
A_ref=A_ref/Nb;
Dref=Dref/Nb;
rho_ref=P3/(R*T3);
U_ref=Qmass/(rho_ref*A_ref);
q_ref=0.5*rho_ref*(U_ref^2);
152
________________________________________________________________________
---------------------------------------------------------------------------------------------------------------------------------
Calculate total hole area:
---------------------------------------------------------------------------------------------------------------------------------
AreaTotal = 0.0;
ntel = 0;
convergence=1000;
[Initial pressure drop]

for i = 1:Nb;
tmp_dP(i)=0;
dp(i)=0;
[Initial mass flow rate]

if mm==1 [mm=1, calculate specific mass flow rates]
qm(i)=(Qmass/(Nb));
dqm(i)=0;
end
[Calculate total hole area]

if Holetype==6;
Harea(i)=Hole_diameter(i)*1000;
end
if Holetype==4 %Wiggle strip

Wiggle_width=Hole_diameter(i);
Harea(i)=(Wiggle_width*slot_height(i))*Number_of_holes(i);
end
if (Holetype ~= 4) & (Holetype ~= 6)

Harea(i) = pi/4*(Hole_diameter(i)^2)*Number_of_holes(i);
end
AreaTotal = AreaTotal + Harea(i);
end
ntel=0;
---------------------------------------------------------------------------------------------------------------------------------
153
________________________________________________________________________
CALCULATE PRESSURE LOSSES IF MASS FLOW SPLITS ARE DEFINE (mm=2)
--------------------------------------------------------------------------------------------------------------------------------
if mm==2
dP_aver = 0;
ntel = ntel + 1;
Qmpass = Qmass;
for i=1:Nb;
Calculate pressure loss and flow rates for specific holes:

---------------------------------------------------------------------------------------------------------------------------------
if Holetype == 1; [Plain Holes]
Beta(i) = qm(i)/Qmpass;
Alpha(i) = Harea(i)/(AreaAnnulus(i));
Qmpass=(Qmpass-qm(i));
if Discharge_Coeff(i) == 0
Mu(i) = Beta(i)/Alpha(i);
K(i) = 1 + 0.64*((2*(Mu(i)^2))+((4*(Mu(i)^4))+(1.56*(Mu(i)^2)*((4*Beta(i))-(Beta(i)^2))))^0.5);
Cd(i) =(K(i)-1)/(0.8*((4*(K(i)^2))-(K(i)*(2-Beta(i))^2))^0.5);
jetangle(i)=asin((1/(1.6*Cd(i)))*((K(i)-1)/K(i)))*(180/pi);
elseif Discharge_Coeff(i) ~= 0
Cd(i)=Discharge_Coeff(i);
end
tmp_dP(i)=dp(i);
dp(i) = (((qm(i)/(Cd(i)*Harea(i)))^2)/(2*rho_hole(i)))+dp_hot ;
dP_aver = (dP_aver + dp(i)*qm(i));
end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype == 2; [Plunged Holes]
K(i) = 1 + 0.36*((2*(Mu(i)^2))+((4*(Mu(i)^4))+(2.77*((4*Beta(i))-(Beta(i)^2))))^0.5);
Cd(i) =(K(i)-1)/(0.6*((4*(K(i)^2))-(K(i)*(2-Beta(i))^2))^0.5);
elseif Discharge_Coeff(i) ~= 0;
end
tmp_dP(i)=dp(i);
dp(i) = (((qm(i)/(Cd(i)*Harea(i)))^2)/(2*rho_hole(i)))+dp_hot;
154
________________________________________________________________________
end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype == 3; [Splash Cooling device]
Cd(i) = Discharge_Coeff(i);
jetangle(i)=0;
tmp_dP(i)=dp(i);
end
if Holetype == 4 ;%Wiggle strip

jetangle(i)=0;
tmp_dP(i)=dp(i);

end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype ==5; [Machined Rings]
jetangle(i)=0;
tmp_dP(i)=dp(i);
end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype ==6; [Swirler]
Swirl_angle=Theta*(pi/180);
jetangle(i)=0;
tmp_dP(i)=dp(i);
K=K_sw*(((sec(Swirl_angle)/Harea(i))^2)-(1/(A_L^2)));
tmp_dP(i)=dp(i);
dp(i)=((qm(i)^2)*K/(2*rho_hole(i)))+dp_hot; %Lefebvre 1998 swirler formula
dP_aver=dP_aver+(dp(i)*qm(i));
end
end
---------------------------------------------------------------------------------------------------------------------------------
155
________________________________________________________________________
Check for convergence:
---------------------------------------------------------------------------------------------------------------------------------
dP_aver = dP_aver/Qmass;
convergence=0.0;
for i = 1:Nb;
convergence = convergence + abs(tmp_dP(i) - dp(i));
end
end % [end for mm]

---------------------------------------------------------------------------------------------------------------------------------
---------------------------------------------------------------------------------------------------------------------------------
CALCULATE PRESSURE LOSSES AND MASS FLOW SPLITS (mm=1)

---------------------------------------------------------------------------------------------------------------------------------
if mm==1
while (convergence > 0.0001) & (ntel < 1000);
dP_aver = 0;
ntel = ntel + 1;
Qmpass = Qmass;
for i=1:Nb;
Calculate pressure loss and flow rates for specific holes :

---------------------------------------------------------------------------------------------------------------------------------
if Holetype == 1; [Plain Holes]
Cd(i) =(K(i)-1)/(0.8*((4*(K(i)^2))-(K(i)*(2-Beta(i))^2))^0.5);
elseif Discharge_Coeff(i) ~= 0
end
tmp_dP(i)=dp(i);
dp(i) = (((qm(i)/(Cd(i)*Harea(i)))^2)/(2*rho_hole(i)))+dp_hot ;
end
---------------------------------------------------------------------------------------------------------------------------------
156
________________________________________________________________________
if Holetype == 2; [Plunged Holes]
Cd(i) =(K(i)-1)/(0.6*((4*(K(i)^2))-(K(i)*(2-Beta(i))^2))^0.5);
elseif Discharge_Coeff(i) ~= 0;
end
tmp_dP(i)=dp(i);
end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype == 3; [Splash Cooling]

jetangle(i)=0;
tmp_dP(i)=dp(i);
end
if Holetype == 4 ;%Wiggle strip

jetangle(i)=0;
tmp_dP(i)=dp(i);

end
---------------------------------------------------------------------------------------------------------------------------------
157
________________________________________________________________________
if Holetype ==5; [Machined Ring]
jetangle(i)=0;
tmp_dP(i)=dp(i);
end
---------------------------------------------------------------------------------------------------------------------------------
if Holetype ==6; [Swirler]
Swirl_angle=Theta*(pi/180);
jetangle(i)=0;
tmp_dP(i)=dp(i);
K=K_sw*(((sec(Swirl_angle)/Harea(i))^2)-(1/(A_L^2)));
tmp_dP(i)=dp(i);
dp(i)=((qm(i)^2)*K/(2*rho_hole(i)))+dp_hot; %Lefebvre 1998 swirler formula
dP_aver=dP_aver+(dp(i)*qm(i));
end
end
---------------------------------------------------------------------------------------------------------------------------------
Adapt mass flow rate and check for convergence:

---------------------------------------------------------------------------------------------------------------------------------
dP_aver = dP_aver/Qmass;
Qnew = 0;
for i = 1:Nb;
dqm(i) = qm(i)*dP_aver/dp(i);
Qnew = Qnew + dqm(i);
end
for i = 1:Nb;
Qm_new = dqm(i)*(Qmass/Qnew)-qm(i);
qm(i) = qm(i)+(1/Nb)*Qm_new;
%qm(i)=qm(i)+Qm_new;
end
convergence=0.0;
for i = 1:Nb;
convergence = convergence + abs(tmp_dP(i) - dp(i));
end
end %end while loop

end % [end mm]
---------------------------------------------------------------------------------------------------------------------------------
158
________________________________________________________________________
Define jet velocity:
---------------------------------------------------------------------------------------------------------------------------------
Massflow=qm;
for i=1:Nb
if Holetype ~= 4 & Holetype ~= 6

V_jet(i)=Massflow(i)/(((0.25*pi)*(Hole_diameter(i)^2))*rho_hole(i)*Number_of_holes(i));
end
if Holetype == 4
V_jet(i)=(Massflow(i)/(Number_of_holes(i)))/((Hole_diameter(i)*slot_height(i))*rho_hole(i));
end
if Holetype==6
V_jet(i)=Massflow(i)/(Hole_diameter(i)*1000*Number_of_holes(i)*rho_hole(i));
end
end
---------------------------------------------------------------------------------------------------------------------------------
Calculate reference Mach number:

---------------------------------------------------------------------------------------------------------------------------------
Cp_air=1002.106 + (0.089328*(T3-273.15))+ (2.332e-4*((T3-273.15)^2))-(1.2831e-7*((T3-273.15)^3));
Gamma_air=Cp_air/(Cp_air - R);
sonic_vel=sqrt(Gamma_air*R*T3);
Ref_Mach=U_ref/sonic_vel;
---------------------------------------------------------------------------------------------------------------------------------
Outputs to Flowcheck:
---------------------------------------------------------------------------------------------------------------------------------
Pressure_drop=dP_aver;
jetangle=jetangle;
Flowsplits=(Massflow/ma)*100;
Cd=Cd;
Pressure_Loss=(Pressure_drop/P3)*100;
RefPressure_Loss=Pressure_drop/q_ref;
---------------------------------------------------------------------------------------------------------------------------------
159
________________________________________________________________________
D.5 COMBUSTION
Define function foe “combustion”:

---------------------------------------------------------------------------------------------------------------------------------
function[Tgas,Air_fuel_node,Length,Air_fuel,Equivalence_Ratio,Casing_D,Liner_D,Airflow,Airflow_nod
e,Airflow_zone,eta_Rz,eta_Pz,eta_Sz,eta_Dz,theta,EI_CO,EI_HC,eta_comb,Nox,NOx,Comb_eff]=Combu
stion(T3,P3,Nb,ma,x,y,Hole_Position,Massflow,Overall_A_F,S_A_F,Total_comb_length,Pz_boundary,Rz
_boundary,Sz_boundary,nn,Casing_Diameter,Liner_Diameter,Number_of_nodes,Nodes,X,Zones,Length,H
ole_type,Rz_flow,Pz_flow,Sz_flow,Dz_flow,Pressure_drop,element_th,nCans,A_ref,Dref);
--------------------------------------------------------------------------------------------------------------------------------
Convert dimensions from mm to m:

---------------------------------------------------------------------------------------------------------------------------------
Casing_Diameter=Casing_Diameter/1000;
---------------------------------------------------------------------------------------------------------------------------------
Display warning message:

---------------------------------------------------------------------------------------------------------------------------------
if P3 < 100000
disp('Inlet Pressure cannot be less than 10 kPa, is going to set P3=100000')
P3 = 100000;
end
---------------------------------------------------------------------------------------------------------------------------------
Calculate fuel flow:

---------------------------------------------------------------------------------------------------------------------------------
Fuel_flow=(ma/Overall_A_F);
---------------------------------------------------------------------------------------------------------------------------------
Define air mass flow in each segment:

---------------------------------------------------------------------------------------------------------------------------------
Airflow_zone=zeros(1,Zones);
for i=1:Zones
for j=1:Nb
if i==1
Airflow_zone(i)=Airflow_zone(i)+(Rz_flow(j)*Massflow(j));
end
if i==2
Airflow_zone(i)=Airflow_zone(i)+(Pz_flow(j)*Massflow(j));
end
if i==3
Airflow_zone(i)=Airflow_zone(i)+(Sz_flow(j)*Massflow(j));
end
if i==4
Airflow_zone(i)=Airflow_zone(i)+(Dz_flow(j)*Massflow(j));
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
160
________________________________________________________________________
Define air flow in each zone:
---------------------------------------------------------------------------------------------------------------------------------
Airflow=zeros(1,Zones);
for i=1:Zones
if i==1
Airflow(i)=Airflow_zone(i);
end
if i==2
Airflow(i)=Airflow(i-1)+Airflow_zone(i);
end
if i==3
end
if i==4
end
end
---------------------------------------------------------------------------------------------------------------------------------
Define air flow at each computational node:

---------------------------------------------------------------------------------------------------------------------------------
for j=1:Zones
if j==1
if (X(i) <= Rz_boundary)
Airflow_node(i)=Airflow(j);
end
end
if j==2
if (X(i) > Rz_boundary) & (X(i) <= Pz_boundary+0.0000001)
end
end
if j==3
if (X(i) > Pz_boundary) & (X(i) <= Sz_boundary+0.0000001)
end
end
if j==4
if (X(i) > Sz_boundary) & (X(i) <= Total_comb_length+0.0000001)
Airflow_node(i) = Airflow(j);
end
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
161
________________________________________________________________________
Air/Fuel ratio at each computational node:
---------------------------------------------------------------------------------------------------------------------------------
Air_fuel_node(i)=Airflow_node(i)/Fuel_flow;
end
---------------------------------------------------------------------------------------------------------------------------------
Air/Fuel ratio in each zone:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Zones
if i==1
Air_fuel(i)=Airflow(i)/Fuel_flow;
end
if i==2
end
if i==3
end
if i==4
end
end
---------------------------------------------------------------------------------------------------------------------------------
Equivalence ratio in each zone:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Zones
Equivalence_Ratio(i)=(S_A_F)/(Air_fuel(i));
end
---------------------------------------------------------------------------------------------------------------------------------
Determine casing and liner diameter at each computational node:

---------------------------------------------------------------------------------------------------------------------------------
Liner_D=zeros(1,Number_of_nodes);
Casing_D=zeros(1,Number_of_nodes);
Liner_D(Number_of_nodes)=Liner_Diameter(Nb);
Casing_D(Number_of_nodes)=Casing_Diameter(Nb);
for j=1:Nb
Liner_D(1)=Liner_Diameter(1);
Casing_D(1)=Casing_Diameter(1);
if (Hole_Position(j) < X(i))

Liner_D(i)=Liner_Diameter(j);
Casing_D(i)=Casing_Diameter(j);
end
if j < 2
if Liner_Diameter(j) ~= Liner_Diameter(j+1)
if abs(Hole_Position(j+1) - Hole_Position(j)) < 0.000001
Liner_D(i)=(Liner_Diameter(j+1)+Liner_Diameter(j))/2;
162
________________________________________________________________________
elseif (X(i) > Hole_Position(j)) & (X(i) <= Hole_Position(j+1))
if Liner_Diameter(j) > Liner_Diameter(j+1)
Liner_D(i)=Liner_Diameter(j) - X(i)*((Liner_Diameter(j+1)-
Liner_Diameter(j))/(Hole_Position(j+1)-Hole_Position(j)));
elseif Liner_Diameter(j) < Liner_Diameter(j+1)
Liner_D(i)=Liner_Diameter(j) + X(i)*((Liner_Diameter(j+1)-
end
end
end
if Casing_Diameter(j) ~= Casing_Diameter(j+1)
if abs(Hole_Position(j+1) - Hole_Position(j)) < 0.000001
Casing_D(i)=(Casing_Diameter(j+1)+Casing_Diameter(j))/2;
elseif (X(i) > Hole_Position(j)) & (X(i) < Hole_Position(j+1))
if Casing_Diameter(j) > Casing_Diameter(j+1)
Casing_D(i)=Casing_Diameter(j) - X(i)*((Liner_Diameter(j+1)-
elseif Casing_Diameter(j) < Casing_Diameter(j+1)
Casing_D(i)=Casing_Diameter(j) + X(i)*((Liner_Diameter(j+1)-
end
end
end
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
Calculate averge liner diameter in each zone:

---------------------------------------------------------------------------------------------------------------------------------
Sz_number=0;
Dz_number=0;
Pz_number=0;
Pz_Liner_diameter=0;
Sz_Liner_diameter=0;
Dz_Liner_diameter=0;
if (X(i) <= Pz_boundary)
Pz_number=Pz_number+1;
Pz_Liner_diameter=Pz_Liner_diameter + Liner_D(i);
end
if (X(i) <= Sz_boundary) & (X(i) > Pz_boundary)

Sz_number=Sz_number+1;
Sz_Liner_diameter=Sz_Liner_diameter+Liner_D(i);
end
if (X(i) > Sz_boundary) & (X(i) <= Total_comb_length)

Dz_number=Dz_number+1;
Dz_Liner_diameter=Dz_Liner_diameter+Liner_D(i);
163
________________________________________________________________________
end
end
Average_Pz_diam=Pz_Liner_diameter/Pz_number;
Average_Sz_diam=Sz_Liner_diameter/Sz_number;
Average_Dz_diam=Dz_Liner_diameter/Dz_number;
---------------------------------------------------------------------------------------------------------------------------------
LOOK-UP TABLES FOR ADIABATIC TEMPERATURE

---------------------------------------------------------------------------------------------------------------------------------
Pressure=[10000 30000 100000 300000 1000000 3000000];%Pa

Tref=[200 400 600 800 1000];%K
phi=[0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.1 0.15 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5];
for i=1:6
if P3==Pressure(i)
P1=Pressure(i);
P2=0;
elseif (P3 > Pressure(i)) & (P3 < Pressure(i+1))
P1=Pressure(i);
P2=Pressure(i+1);
end
end
for j=1:5
if T3 == Tref(j)
a=j;
T1=Tref(j);
elseif (T3 > Tref(j)) & (T3 < Tref(j+1))
a=j;
T1=Tref(j);
T2=Tref(j+1);
end
end
if P1==Pressure(1)
%Temp rise for Pressure(1)
dT1=[60 88 116 145 174 201 229 288 421 549 794 1017 1226 1450 1608 1766 1894 1968 1973 1922 1853 1779 1713;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1183 1365 1534 1674 1778 1820 1843 1815 1758 1690 1623;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1301 1452 1570 1651 1696 1708 1693 1656 1595 1531;
51 76.5 102 127 152 176 200 250 365 478 695 896 1074 1232 1362 1454 1515 1552 1569 1567 1538 1490 1436;
48 72 96 120 144 167.5 191 235 348 458 664 862 1030 1150 1255 1328 1379 1412 1428 1430 1418 1372 1327];
if T3 == Tref(a)
DT1=dT1(a,:);
elseif (T3 > Tref(a)) & (T3 < Tref(a+1))
DT1=(dT1(a,:)-dT1(a+1,:))*((T3-T2)/(T1-T2))+dT1(a+1,:);
end
end
164
________________________________________________________________________
if P1==Pressure(2)
dT1=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1235 1432 1614 1774 1911 1993 1998 1936 1861 1784 1715;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1184 1370 1540 1687 1805 1871 1882 1840 1773 1702 1629;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1309 1465 1596 1684 1738 1756 1732 1678 1612 1543;
51 76.5 102 127 152 176 200 250 365 478 695 896 1077 1244 1382 1488 1562 1608 1622 1609 1573 1518 1455;
48 72 96 120 144 167.5 191 235 348 458 664 862 1022 1169 1285 1373 1433 1468 1484 1481 1453 1414 1360];
if T3 == Tref(a)
DT1=dT1(a,:);
end
end
if P1==Pressure(3)
dT1=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1234 1432 1614 1784 1930 2025 2012 1944 1868 1789 1713;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1183 1374 1548 1700 1830 1912 1911 1855 1785 1707 1633;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1310 1472 1610 1723 1789 1802 1762 1695 1625 1552;
51 76.5 102 127 152 176 200 250 365 478 695 896 1083 1251 1399 1520 1602 1660.0 1678 1652 1603 1540 1471;
48 72 96 120 144 167.5 191 235 348 458 664 862 1030 1189 1319 1419 1489 1530 1550 1538 1502 1449 1387];
if T3 == Tref(a)
DT1=dT1(a,:);
end
end
if P1==Pressure(4)
dT1=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1239 1438 1609 1789 1940 2043 2025 1942 1868 1790 1720;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1139 1373 1550 1710 1850 1940 1930 1861 1787 1709 1640;
54 81 108 134 160 185.5 211 261 384 501 723 934 1145 1323 1484 1639 1750 1820 1838 1769 1705 1630 1550;
51 76.5 102 127 152 176 200 250 365 478 695 896 1085 1252 1409 1590 1640 1707 1718 1672 1618 1550 1480;
48 72 96 120 144 167.5 191 235 348 458 664 862 1033 1192 1345 1450 1530 1580 1611 1579 1529 1469 1399];
if T3 == Tref(a)
DT1=dT1(a,:);
end
end
if P1==Pressure(5)
dT1=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1239 1433 1615 1789 1949 2059 2030 1949 1869 1789 1710;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1185 1377 1552 1718 1861 1968 1944 1870 1790 1711 1635;
54 81 108 134 160 185.5 211 261 384 501 723 934 1129 1313 1482 1630 1770 1859 1853 1779 1710 1630 1552;
51 76 102 127 152 176 200 250 365 478 695 896 1089 1257 1420 1560 1673 1749 1753 1705 1640 1555 1480;
48 72 96 120 144 167.5 191 235 348 458 664 862 1033 1200 1350 1475 1560 1635 1649 1613 1550 1479 1407];
165
________________________________________________________________________
if T3 == Tref(a)
DT1=dT1(a,:);
end
end
if P1==Pressure(6)
dT1=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1230 1430 1617 1790 1950 2079 2039 1950 1869 1790 1715;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1180 1375 1552 1720 1872 1989 1951 1870 1790 1711 1639;
54 81 108 134 160 185.5 211 261 384 501 723 934 1130 1315 1489 1649 1789 1887 1870 1795 1715 1637 1550;
51 76.5 102 127 152 176 200 250 365 478 695 896 1083 1260 1421 1520 1700 1785 1780 1717 1640 1561 1485;
48 72 96 120 144 167.5 191 235 348 458 664 862 1035 1205 1359 1498 1605 1675 1685 1635 1563 1489 1410];
if T3 == Tref(a)
DT1=dT1(a,:);
end
end
if P2==Pressure(1)
dT2=[60 88 116 145 174 201 229 288 421 549 794 1017 1226 1450 1608 1766 1894 1968 1973 1922 1853 1779 1713;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1183 1365 1534 1674 1778 1820 1843 1815 1758 1690 1623;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1301 1452 1570 1651 1696 1708 1693 1656 1595 1531;
51 76.5 102 127 152 176 200 250 365 478 695 896 1074 1232 1362 1454 1515 1552 1569 1567 1538 1490 1436;
48 72 96 120 144 167.5 191 235 348 458 664 862 1030 1150 1255 1328 1379 1412 1428 1430 1418 1372 1327];
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
if P2==Pressure(2)
dT2=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1235 1432 1614 1774 1911 1993 1998 1936 1861 1784 1715;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1184 1370 1540 1687 1805 1871 1882 1840 1773 1702 1629;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1309 1465 1596 1684 1738 1756 1732 1678 1612 1543;
51 76.5 102 127 152 176 200 250 365 478 695 896 1077 1244 1382 1488 1562 1608 1622 1609 1573 1518 1455;
48 72 96 120 144 167.5 191 235 348 458 664 862 1022 1169 1285 1373 1433 1468 1484 1481 1453 1414 1360];
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
166
________________________________________________________________________
if P2==Pressure(3)
dT2=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1234 1432 1614 1784 1930 2025 2012 1944 1868 1789 1713;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1183 1374 1548 1700 1830 1912 1911 1855 1785 1707 1633;
54 81 108 134 160 185.5 211 261 384 501 723 934 1126 1310 1472 1610 1723 1789 1802 1762 1695 1625 1552;
51 76.5 102 127 152 176 200 250 365 478 695 896 1083 1251 1399 1520 1602 1660.0 1678 1652 1603 1540 1471;
48 72 96 120 144 167.5 191 235 348 458 664 862 1030 1189 1319 1419 1489 1530 1550 1538 1502 1449 1387];
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
if P2==Pressure(4)
dT2=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1239 1438 1609 1789 1940 2043 2025 1942 1868 1790 1720;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1139 1373 1550 1710 1850 1940 1930 1861 1787 1709 1640;
54 81 108 134 160 185.5 211 261 384 501 723 934 1145 1323 1484 1639 1750 1820 1838 1769 1705 1630 1550;
51 76.5 102 127 152 176 200 250 365 478 695 896 1085 1252 1409 1590 1640 1707 1718 1672 1618 1550 1480;
48 72 96 120 144 167.5 191 235 348 458 664 862 1033 1192 1345 1450 1530 1580 1611 1579 1529 1469 1399];
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
if P2==Pressure(5)
dT2=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1239 1433 1615 1789 1949 2059 2030 1949 1869 1789 1710;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1185 1377 1552 1718 1861 1968 1944 1870 1790 1711 1635;
54 81 108 134 160 185.5 211 261 384 501 723 934 1129 1313 1482 1630 1770 1859 1853 1779 1710 1630 1552;
51 76 102 127 152 176 200 250 365 478 695 896 1089 1257 1420 1560 1673 1749 1753 1705 1640 1555 1480;
48 72 96 120 144 167.5 191 235 348 458 664 862 1033 1200 1350 1475 1560 1635 1649 1613 1550 1479 1407];
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
if P2==Pressure(6)
dT2=[60 88 116 145 174 201.5 229 288 421 549 794 1017 1230 1430 1617 1790 1950 2079 2039 1950 1869 1790 1715;
57 85.5 114 141.5 169 196 223 278 407 527 758 978 1180 1375 1552 1720 1872 1989 1951 1870 1790 1711 1639;
54 81 108 134 160 185.5 211 261 384 501 723 934 1130 1315 1489 1649 1789 1887 1870 1795 1715 1637 1550;
51 76.5 102 127 152 176 200 250 365 478 695 896 1083 1260 1421 1520 1700 1785 1780 1717 1640 1561 1485;
48 72 96 120 144 167.5 191 235 348 458 664 862 1035 1205 1359 1498 1605 1675 1685 1635 1563 1489 1410];
167
________________________________________________________________________
if T3 == Tref(a)
DT2=dT2(a,:);
end
end
for i=1:6
if P3==Pressure(i)
dT=DT1;
elseif (P3 > Pressure(i)) & (P3 < Pressure(i+1))

dp_dT1=DT1;
dp_dT2=DT2;
dT=(dp_dT1-dp_dT2)*((P3-P2)/(P1-P2))+dp_dT2;
end
end
for i=1:Zones
if Equivalence_Ratio(i) <= 1.5
Delta_T(i)=interp1(phi,dT,Equivalence_Ratio(i));
else
Delta_T(i)=interp1(phi,dT,Equivalence_Ratio(i),'nearest','extrap');
end
end
Delta_Ts = interp1(phi,dT,1); %temperature rise stiochiometric gas temperature

---------------------------------------------------------------------------------------------------------------------------------
Calculation of Combustion efficiency :

---------------------------------------------------------------------------------------------------------------------------------
if nn==1; [Combustion eff. On]
for i=1:Zones
[Re-circulation zone]
eta_Rz=0.56+0.44*tanh(1.5475e-3*(T3+(108*log(P3))-1863));
[Primary zone]
eta_Pz=0.71+0.29*tanh(1.5475e-3*(T3+(108*log(P3))-1863));
if i==1
Trz=T3+eta_Rz*Delta_T(i);
Tg(i)=((1/3)*T3)+((2/3)*Trz);
end
if i==2
Tg(i)=T3+eta_Pz*Delta_T(i);
end
168
________________________________________________________________________
if i==3
Volume_Pz=((pi*Average_Pz_diam)/4)*(Pz_boundary);
if Equivalence_Ratio(i) <= 1 %Air rich
y=Equivalence_Ratio(i);
n=(2*y^2)/Equivalence_Ratio(i-1);
zeta300=(Fuel_flow/((Volume_Pz)*(P3^n)));
term=y^-1.205;
zeta=zeta300*(10^(-3.054*term))*(T3^(1.2327*term));
%term1=(0.911*log10(zeta)) + (4.56*n) +(Equivalence_Ratio(i-1)/log10(Volume_Pz)) +
(558/T3);%Prediction of wall heat transfer for a gas turbine conbustor
Dstar=0.736-0.0173*(P3/Pressure_drop);
%term1=0.911*log10(zeta)+(8.02*Equivalence_Ratio(i))-1.097 + Dstar;% NREC
term1=0.991*log10(zeta)+(4.56*n)-(1.1*Equivalence_Ratio(i))-1.097+Dstar;%A Simple
Method for the Prediction of Wall Temperatures in Gas Turbines
term2=10^term1;
eta_Sz=1/(10^term2);
eta_Sz=0.9; %Approximate
Tg(i)=T3+eta_Sz*Delta_T(i);
elseif Equivalence_Ratio(i) > 1
eta_Sz = 1/Equivalence_Ratio(i);
Tg(i)=T3+eta_Sz*Delta_T(i);
end
end
if i==4
Volume_Sz=((pi*Average_Sz_diam)/4)*(Sz_boundary - Pz_boundary);
if Equivalence_Ratio(i) <= 1 %Air rich
y=Equivalence_Ratio(i);
zeta300=(Fuel_flow/((Volume_Sz)*(P3^n)));
term=y^-1.205;
% term1=(0.911*log10(zeta)) + (4.56*n) +(Equivalence_Ratio(i-1)/log10(Volume_Sz)) +
(558/T3);
% term2=10^term1;
term1=0.991*log10(zeta)+(4.56*n)-(1.1*Equivalence_Ratio(i))-1.097+Dstar;%A Simple
%term1=0.911*log10(zeta)+(8.02*Equivalence_Ratio(i))-1.097 + Dstar; % NREC
%term1=(0.911*log10(zeta)) + (4.56*n) +(Equivalence_Ratio(i-1)/log10(Volume_Pz)) +
term2=10^term1;
eta_Dz=1/(10^term2);
eta_Dz=0.9; %Approximate
Tg(i)=T3+eta_Dz*Delta_T(i);
elseif Equivalence_Ratio(i) > 1
169
________________________________________________________________________
y=1;
zeta300=(Fuel_flow/(Volume_Sz*(P3^n)));
term=y^-1.205;
%term1=0.911*log10(zeta)+(8.02/Equivalence_Ratio(i))-1.097 + Dstar; %NREC
term1=0.991*log10(zeta)+(4.56*n)-(1.1/Equivalence_Ratio(i))-1.097+Dstar;%A Simple
%%term1=(0.911*log10(zeta)) + (4.56*n) +(Equivalence_Ratio(i-1)/log10(Volume_Pz)) +
term2=10^term1;
eta_Dz=1/(10^term2);
Tg(i)=T3+eta_Dz*Delta_T(i);
end
end
end
Ts=T3+eta_Pz*Delta_Ts;
Tpz_mean=T3+eta_Pz*0.5*(Delta_Ts+Delta_T(2));
end %end nn=1

---------------------------------------------------------------------------------------------------------------------------------
Combustion efficiency off:

---------------------------------------------------------------------------------------------------------------------------------
if nn==2
for i=1:Zones
Tg(i)=T3+Delta_T(i);
end
Ts=T3+Delta_Ts;
Tpz_mean=T3+0.5*(Delta_Ts+Delta_T(2));
eta_Rz=0;
eta_Pz=0;
eta_Sz=0;
eta_Dz=0;
end
---------------------------------------------------------------------------------------------------------------------------------
Calculation of Theta:
---------------------------------------------------------------------------------------------------------------------------------
phiPz=(Equivalence_Ratio(1)+Equivalence_Ratio(2))/2;
if phiPz < 1
b=245*(1.39+log(phiPz));
else
b=170*(2-log(phiPz));
end
theta=(((P3^1.75)*A_ref*(Dref^0.75)*exp(T3/b))/ma)/1000000;
Tgas=zeros(1,Zones+1);
Tgas(1)=T3;
Tgas(2:Zones+1)=Tg(1:Zones);
170
________________________________________________________________________
-------------------------------------------------------------------------------------------------------------------------------
[Emissions]
The following empirical expressions will predict CO within about a factor of 2 for aircraft engines
using pressure liquid fuel atomizers [NREC Vol II]
---------------------------------------------------------------------------------------------------------------------------------
EI_CO=10^(13.477-4.5276*log10(T3)); %[g CO/kg fuel]
---------------------------------------------------------------------------------------------------------------------------------
[Hydrocarbons]
Almost the same comments apply to hydrocarbons as to CO [NREC Vol II]
---------------------------------------------------------------------------------------------------------------------------------
EI_HC=10^(19.730-7.1915*log10(T3));
--------------------------------------------------------------------------------------------------------------------------------
[NOx Emissions]
Combustion Volume (Vc): Assume it to be up to the secondary zone
---------------------------------------------------------------------------------------------------------------------------------
Vc=0;
if X(i) <= Sz_boundary
Vc=Vc+((pi*Liner_D(i)^2)/4)*element_th(i);
end
end
Tmax=max(Tgas);
r_time=0;%(P3*Vc)/(287.5*Tpz_mean*(Airflow_zone(3)/nCans));
%Lefebvre Model
if Nb < 3
Nox = 0;
else
Nox=(9e-8)*(P3/1000)^1.25*Vc*exp(0.01*Ts)/((Airflow_zone(3))*Tpz_mean);
end
%Odgers Model
if Nb < 3
NOx = 0;
else
NOx=29*(10^(-9410/Tmax))*P3^0.66*(1-exp(-250*r_time));
end
--------------------------------------------------------------------------------------------------------------------------------
[Combustion Efficeincy (Calculated from emission levels)]

---------------------------------------------------------------------------------------------------------------------------------
eta_comb=(1-10^(-3)*(0.24*EI_CO+EI_HC))*100;
--------Combustion Efficeincy (figure 2.2 NREC)--------

if theta < 17
Comb_eff=(1.8467 - (0.01634*15)+ (0.000104*15*15) - (18.296/15))*100;
elseif theta > 70
Comb_eff=(1.8467 - (0.01634*70)+(0.000104*70*70)-(18.296/70))*100;
else
Comb_eff=(1.8467-(0.01634*theta)+(0.000104*theta*theta)-18.296/theta)*100;
end
171
________________________________________________________________________
D.6 ELEMENTS
Define function for “elements”:

---------------------------------------------------------------------------------------------------------------------------------
function[Momflux,T_GAS,m_annulus,Hyd_D,dp_hot,Reo,Rei,d_AL,d_ACi,ALi,ALo,Aci,Aco,m_internal,
Pattern_Factor,M_inner,Local_AF,Y_max]=Elements(Zones,Length,Tgas,X,Number_of_nodes,Rz_bound
ary,Pz_boundary,Sz_boundary,Total_comb_length,Massflow,ma,Hole_Position,Nb,Fuel_flow,Airflow_no
de,Casing_D,Liner_D,t_L,t_C,P3,R,T3,U_ref,mu_a,mu_g,element_th,Nodes,Tgx,nCans,q_ref,Pressure_dr
op,Dref,Overall_A_F,Liner_Diameter,rho_hole_gas,rho_annulus,V_jet,Cd,Hole_diameter,Com,Number_o
f_holes);
---------------------------------------------------------------------------------------------------------------------------------

---------------------------------------------------------------------------------------------------------------------------------
Hole_diameter=Hole_diameter/1000;
---------------------------------------------------------------------------------------------------------------------------------
Gas temperature at end of each zone:

---------------------------------------------------------------------------------------------------------------------------------
Fuel_flow=ma/Overall_A_F;
for j=1:Zones+1
if j==1
T_GAS(j)=(((X(j)-0)/(Tgx(2)-0))*(Tgas(2)-T3))+T3;
else
T_GAS((j-1)*Nodes)=Tgas(j);
end
end
---------------------------------------------------------------------------------------------------------------------------------
Interpolate between main gas temperaure values to obtain gas temp at other nodes:
---------------------------------------------------------------------------------------------------------------------------------
for i=1:Zones
if i==1
for j=(i+1):(i*Nodes-1)
T_GAS(j)=T_GAS(i*Nodes)+((X(j)-X(i*Nodes))/(X(i)-X(i*Nodes)))*(T_GAS(i)-T_GAS(i*Nodes));
end
else
for j=((i-1)*Nodes+1):(i*Nodes-1)
T_GAS(j)=T_GAS(i*Nodes)+((X(j)-X(i*Nodes))/(X((i-1)*Nodes)-X(i*Nodes)))*(T_GAS((i-
1)*Nodes)-T_GAS(i*Nodes));
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
Annulus air mass flow rate at each computational node:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Nb
if i==1
172
________________________________________________________________________
M_inner(i)=Massflow(i);
else
M_inner(i)=M_inner(i-1) + Massflow(i);
end
end
for j=1:Nb
m_internal(1)=M_inner(1);
m_annulus(1)=ma-m_internal(1);
if (Hole_Position(j) < X(i))%(X(i) >= Hole_Position(j)) & (X(i) <= Hole_Position(j+1))

m_internal(i)=M_inner(j);
m_annulus(i)=ma-m_internal(i);
end
end
end
if m_internal(i) <= 0
m_internal(i)=ma/1000000;
end
if m_annulus(i) <= 0
m_annulus(i)=ma/1000000;
end
end
Local_AF(i)=m_internal(i)/Fuel_flow;
end
---------------------------------------------------------------------------------------------------------------------------------
Heat transfer area of each element:

---------------------------------------------------------------------------------------------------------------------------------
d_AL(i)=pi*(Liner_D(i)+(t_L))*element_th(i);%Liner mean area
d_ACi(i)=pi*(Casing_D(i))*element_th(i);%Casing inside area
ALi(i)=pi*(Liner_D(i))*element_th(i);
ALo(i)=pi*(Liner_D(i)+(2*t_L))*element_th(i);
Aci(i)=pi*(Casing_D(i)*element_th(i));
Aco(i)=pi*(Casing_D(i)+(2*t_C))*element_th(i);
end
---------------------------------------------------------------------------------------------------------------------------------
Hydroulic diameter of annuls:

---------------------------------------------------------------------------------------------------------------------------------
Average_L_Diam=0;
Annulus_flow_area(i)=(pi/4)*(Casing_D(i)^2-(Liner_D(i)+2*t_L)^2);
Hyd_D(i)=(4*Annulus_flow_area(i))/(pi*((Liner_D(i)+2*t_L)+Casing_D(i)));
Liner_flow_area(i)=(pi/4)*(Liner_D(i)^2);
Average_L_Diam=Average_L_Diam+Liner_D(i);
end
Average_L_Diam=Average_L_Diam/Number_of_nodes;
---------------------------------------------------------------------------------------------------------------------------------
173
________________________________________________________________________
Pressure los due to hot gases:
---------------------------------------------------------------------------------------------------------------------------------
dp_hot=0.5*(P3/(R*T3))*(U_ref^2)*((T_GAS(Number_of_nodes)/T3)-1);
---------------------------------------------------------------------------------------------------------------------------------
Pattern factor:
---------------------------------------------------------------------------------------------------------------------------------
Pattern_Factor=1-exp(-(Dref*q_ref)/(0.07*Total_comb_length*Pressure_drop));
--------------------------------------------------------------------------------------------------------------------------------
Reynolds number in annulus and inside liner:

---------------------------------------------------------------------------------------------------------------------------------
Reo(i)=(m_annulus(i)/Annulus_flow_area(i))*(Hyd_D(i)/mu_a(i));
Rei(i)=(m_internal(i)/Liner_flow_area(i))*(Liner_D(i)/mu_g(i));
end
--------------------------------------------------------------------------------------------------------------------------------
Gas velocity at hole:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Nb
V_g(i)=(M_inner(i))/((Liner_Diameter(i)^2*0.25*pi)*rho_hole_gas(i));
end
--------------------------------------------------------------------------------------------------------------------------------
Momentum flux ratio:

---------------------------------------------------------------------------------------------------------------------------------
for i=1:Nb
Momflux(i)=(rho_annulus(i)*(V_jet(i)^2))/(rho_hole_gas(i)*(V_g(i)^2));
end
--------------------------------------------------------------------------------------------------------------------------------
Max peneteration:
---------------------------------------------------------------------------------------------------------------------------------
for i=1:Nb
m_j(i)=Massflow(i)/(Number_of_holes(i));
d_j(i)=(Cd(i)^0.5)*Hole_diameter(i);
Y_max(i)=(1.25*d_j(i)*(Momflux(i)^0.5)*((M_inner(i))/((M_inner(i))+m_j(i))))*1000;
end
---------------------------------------------------------------------------------------------------------------------------------
174
________________________________________________________________________
D.7 HEAT TRANSFER
Define function of “Heat Transfer”:

---------------------------------------------------------------------------------------------------------------------------------
function[Liner_temp,Casing_temp,New_Tannulus,old_Tannulus,eta_filmcooling,J,distance_downstream_s
lot,e_g,Average_haLo]=Heat_transfer(Zones,Nodes,Number_of_nodes,ma,Overall_A_F,T_GAS,R,Liner_
D,Casing_D,element_th,x,y,Hole_type,Massflow,Number_of_cooling_slots,cooling_slot_height,mu_a,Pz_
boundary,k_g,T_annulus,sigma,k_L,k_C,t_L,t_C,Airflow_node,Nb,X,mu_g,k_a,Hyd_D,m_annulus,d_AL,
d_ACi,T3,Cp_an,bb,Reo,Rei,Pr_air,Pr_gas,Fuel_flow,FAR,P3,Hole_Position,Tamb,ALi,ALo,Aci,Aco,eLi,
eLo,eCi,eCo,Total_comb_length,Radiation_Model,m_internal,Liner_Diameter,nCans,M_inner,Local_AF,h
0,Outside_conv,Outside_rad);
---------------------------------------------------------------------------------------------------------------------------------

---------------------------------------------------------------------------------------------------------------------------------
cooling_slot_height=cooling_slot_height/1000;
---------------------------------------------------------------------------------------------------------------------------------
Assume internal pressure is constant throughout combustion chamber:

---------------------------------------------------------------------------------------------------------------------------------
Internal_Pressure(i)=P3;
end
Beam length l_b:

---------------------------------------------------------------------------------------------------------------------------------
l_b(i)=(3.4*Total_comb_length*0.25*pi*Liner_D(i)^2)/(pi*Liner_D(i)*Total_comb_length);
q(i)=1/Local_AF(i);
end
---------------------------------------------------------------------------------------------------------------------------------
Carbon to hydrogen ratio:

---------------------------------------------------------------------------------------------------------------------------------
C_H=15; %Fuel atomic ratio C/H
L=3*(((C_H)-5.5)^0.75);
---------------------------------------------------------------------------------------------------------------------------------
Calculation of cooling slot distance in terms of computational nodes:

---------------------------------------------------------------------------------------------------------------------------------
if bb==1 [Wall cooling on]
for i=1:Number_of_cooling_slots
slot_lip_thickness(i)=t_L;
end
175
________________________________________________________________________
Massflow through cooling slots:

---------------------------------------------------------------------------------------------------------------------------------
k=0;
for i=1:Nb
if Holetype==3 [Splash Cooling device]
j=k+1;
Cooling_massflow(j)=Massflow(i);
x_cooling_slot(j)=Hole_Position(i);
if j < Number_of_cooling_slots
k=j;
end
end
if Holetype==4 [Wiggle strip device]

j=k+1;
k=j;
end
end
if Holetype==5 [Machined Ring]

j=k+1;
k=j;
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
Downstream distance from slot [x] & Reynolds numbers:

---------------------------------------------------------------------------------------------------------------------------------
for j=1:Number_of_cooling_slots
Inside convective heat transfer coeff.:

---------------------------------------------------------------------------------------------------------------------------------
if (X(i) < x_cooling_slot(1))
eta_filmcooling(i) =0;
NuLi(i)=0.020*(Rei(i)^0.8)*(Pr_gas(i)^0.33333);
haLi(i)=NuLi(i)*(k_g(i)/Liner_D(i));
end
176
________________________________________________________________________
Calculate distance from cooling slot j:
---------------------------------------------------------------------------------------------------------------------------------
if x_cooling_slot(j) <= X(i)
distance_downstream_slot(i) = X(i) - x_cooling_slot(j);
if (X(i) - x_cooling_slot(j) <=0.000001)
distance_downstream_slot(i)=(0.000001);
end
if distance_downstream_slot(i) < cooling_slot_height(j)

distance_downstream_slot(i) = cooling_slot_height(j);
end
--------------------------------------------------------------------------------------------------------------------------------
Momentum flux ratio:

---------------------------------------------------------------------------------------------------------------------------------
J(i)=(Cooling_massflow(j)/(pi*Liner_D(i)*cooling_slot_height(j)))/(m_internal(i)/(pi*(Liner_D(i)^2)
*0.25));
Slot Reynolds number:

---------------------------------------------------------------------------------------------------------------------------------
Res(i)=(Cooling_massflow(j)/(pi*Liner_D(i)*cooling_slot_height(j)))*(cooling_slot_height(j)/ mu_a(i));
Film cooling effieicncy:

---------------------------------------------------------------------------------------------------------------------------------
if J(i) < 1.3
eta_filmcooling(i)=1.1*(J(i)^0.65)*((mu_a(i)/mu_g(i))^0.15)*((distance_downstream_slot(i)/cooling_slot_
height(j))^-0.2)*...
((slot_lip_thickness(j)/cooling_slot_height(j))^-0.2);
if eta_filmcooling(i) > 1
eta_filmcooling(i) = 1;
end
NuLi(i)=0.069*((Res(i)*(distance_downstream_slot(i)/cooling_slot_height(j)))^0.7);
haLi(i)=NuLi(i)*(k_a(i)/distance_downstream_slot(i));
elseif J(i) >= 1.3
eta_filmcooling(i)=1.28*((mu_a(i)/mu_g(i))^0.15)*((distance_downstream_slot(i)/cooling_slot_height(j))^
-0.2)*((slot_lip_thickness(j)/cooling_slot_height(j))^-0.2);
if eta_filmcooling(i) > 1
eta_filmcooling(i) = 1;
end
NuLi(i)=0.1*(Res(i)^0.8)*((distance_downstream_slot(i)/cooling_slot_height(j))^0.44);
haLi(i)=NuLi(i)*(k_a(i)/distance_downstream_slot(i));
end
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
177
________________________________________________________________________
Calculate gas temperature near inside liner wall:
---------------------------------------------------------------------------------------------------------------------------------
Tw_ad(i)=T_GAS(i)-eta_filmcooling(i)*(T_GAS(i)-T_annulus(i));
End
---------------------------------------------------------------------------------------------------------------------------------
Liner outside heat transfer coeff.:

---------------------------------------------------------------------------------------------------------------------------------
if Reo(i) <= 2300

NuLo(i)=3.66;
else
NuLo(i)=0.020*(Reo(i)^0.8)*(Pr_air(i)^0.333);
end
if Reo(i) <= 2300

NuCi(i)=3.66;
else
NuCi(i)=0.020*(Reo(i)^0.8)*(Pr_air(i)^0.3333);
end
end
end [end for film cooled liner]

---------------------------------------------------------------------------------------------------------------------------------
Calculate gas emisivity:

---------------------------------------------------------------------------------------------------------------------------------
e_g(i)=1-exp(-290*(Internal_Pressure(i)/1000)*L*sqrt(q(i)*l_b(i))*(T_GAS(i)^-1.5));
end
---------------------------------------------------------------------------------------------------------------------------------
No film cooling devices used:

---------------------------------------------------------------------------------------------------------------------------------
if bb==2 % Uncooled Liner
eta_filmcooling=zeros(1,Number_of_nodes);
J=zeros(1,Number_of_nodes);
distance_downstream_slot=zeros(1,Number_of_nodes);
Tw_ad(i)=T_GAS(i)-eta_filmcooling(i)*(T_GAS(i)-T_annulus(i));
NuLi(i)=0.020*(Rei(i)^0.8)*(Pr_gas(i)^0.33333);
haLi(i)=NuLi(i)*(k_g(i)/Liner_D(i));
%Backside heat transfer

if Reo(i) <= 2300
NuLo(i)=3.66;
else
NuLo(i)=0.020*(Reo(i)^0.8)*(Pr_air(i)^0.333);
end
if Reo(i) <= 2300

NuCi(i)=3.66;
178
________________________________________________________________________
else
NuCi(i)=0.020*(Reo(i)^0.8)*(Pr_air(i)^0.3333);
end
end
end
---------------------------------------------------------------------------------------------------------------------------------
old_Tannulus=T_annulus;
Heat Transfer process:

---------------------------------------------------------------------------------------------------------------------------------
if i==1
T=T3;
else
T=New_Tannulus(i-1);
end
converg =1;
TTR=T_GAS(i);
TTL=Tamb;
TwCi=T+10;
while converg > 0.000001
TwLi=(TTR+TTL)/2;
[Heat flux to inside liner wall]

QfcLi=haLi(i)*ALi(i)*(Tw_ad(i)-TwLi);
if Radiation_Model ==1 [Simple Method]

QradLi=eLi*sigma*ALi(i)*((T_GAS(i)^4)-(TwLi^4));
elseif Radiation_Model ==2 [Accurate Method]
QradLi=0.5*sigma*(1+eLi)*e_g(i)*(T_GAS(i)^1.5)*ALi(i)*((T_GAS(i)^2.5) - (TwLi^2.5));
elseif Radiation_Model==3
QradLi=0;
end
QwallL=QfcLi+QradLi; [Total wall heat flux]
TwLo=TwLi-(QwallL/(0.5*(ALi(i)+ALo(i))*k_L/t_L));
if TwLo <= 0
TwLo = TwLi - 0.00001;
end
179
________________________________________________________________________
[Backside heat transfer]
haLo(i)=NuLo(i)*(k_a(i)/Hyd_D(i));
QfcLo=haLo(i)*ALo(i)*(TwLo-T);
QradLo=QwallL-QfcLo;
TwCi=T+10;
tel=0;
while tel < 50
tel=tel+1;
term=(eLo*eCi)/(eCi+eLo*(1-eCi)*(ALo(i)/Aci(i)));
hradLo=term*sigma*(TwLo^2+TwCi^2)*(TwLo+TwCi);
TwCi=TwLo-QradLo/(hradLo*ALo(i));
if TwCi < Tamb+0.00001
TwCi=Tamb+0.00001;
end
end
haCi(i)=NuCi(i)*(k_a(i)/Hyd_D(i));
QfcCi=haCi(i)*Aci(i)*(T-TwCi);
[Heating to annulus air]

dQa=QfcLo-QfcCi;
if m_annulus(i) <= 1e-3
Tao= T;
else
Tao=T+(dQa/(Cp_an(i)*m_annulus(i)));
end
QradCi=QradLo;
QwallC=QradCi+QfcCi;
TwCo=TwCi-(QwallC/(0.5*(Aci(i)+Aco(i))*(k_C/t_C)));
if TwCo < Tamb+0.000001
TwCo=Tamb+0.000001;
end
if Outside_conv ==1 [Default convection on]

QncCo=Aco(i)*1.24*((TwCo-Tamb)^1.3333);
elseif Outside_conv ==2 [Define heat transfer coeff]
QncCo=Aco(i)*h0*(TwCo-Tamb);
elseif Outside_conv ==3
QncCo=0; % Convection off
end
if Outside_rad==1 [Radiation on]

QradCo=eCo*sigma*Aco(i)*(TwCo^4-Tamb^4);
elseif Outside_rad==2
QradCo=0;
end
180
________________________________________________________________________
Qlost=QncCo+QradCo;
if Outside_rad == 2 & Outside_conv ==3 % Adiabatic wall

if QwallC < 0
TTR=TwLi;
else
TTL=TwLi;
end
converg=abs(QwallC);
else
[Check for convergence]
if QwallC < Qlost

TTR=TwLi;
else
TTL=TwLi;
end
converg=abs(100*(QwallC-Qlost)/QwallL);
end
end
TLi(i)=TwLi;
TLo(i)=TwLo;
New_Tannulus(i)=Tao;
Tci(i)=TwCi;
Tco(i)=TwCo;
Liner_temp(i)=(TLi(i)+TLo(i))/2;
Casing_temp(i)=(Tci(i)+Tco(i))/2;
end
Total_haLo = 0;
for i = 1:Number_of_nodes
Total_haLo=Total_haLo + haLo(i);
end
Average_haLo=Total_haLo/Number_of_nodes; [Average Backside heat transfer coeff.]
181
APPENDIX E –
Boundary conditions
APPENDIX E Boundary conditions
________________________________________________________________________
TableE.1: Base case boundary conditions
Flow Flow splits Splits to
Fluent Inlet splits [kg/s] individual Directional vector components Hydraulic
Designation [%] 2.37642 holes [kg/s] x y z diameter [mm]
shell swirl 1+2 6.0969862 0.14489
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.96927 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7853494 0.11372 6.828
wigg 2 total 10.607973 0.25209
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027606667 -0.13922 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.0389175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0389175 0.410991 0.899943 0.145567
shell4 hole 3 0.0389175 -0.41099 0.899943 0.145567
shell4 hole 4 0.0389175 -0.64789 -0.7477 0.145567
shell 4 hole
total 6.5506097 0.15567 11.254
shell 4 splash
total 0.8222452 0.01954
wigg 4 total 9.5189403 0.22621
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.063725 0.70639 -0.70639 0.045014
dilute hole 2 0.063725 0.70639 0.70639 0.045014
dilute hole 3 0.063725 -0.70639 0.70639 0.045014
dilute hole 4 0.063725 -0.70639 -0.70639 0.045014
dilute hole 5 4.3321467 0.10295 0.10295 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2435344 0.07708 0.07708 -0.46946 -0.88293 0.007156 15.86
100 2.37642 2.37642
182
________________________________________________________________________
Table E.2: Base Case with blocked cooling devices boundary conditions
shell swirl 1+2 5.376154 0.12776
wigg 1 total 15.62813 0.37139
shell2 hole 1 0.01637 0.63474 -0.73253 0.245984
shell2 hole 2 0.01637 0.969274 0 0.245984
shell2 hole 3 0.01637 0.63474 0.732528 0.245984
shell2 hole 4 0.01637 0.137943 0.959408 0.245984
shell2 hole 5 0.01637 -0.63474 0.732528 0.245984
shell2 hole 6 0.01637 -0.96927 0 0.245984
shell2 hole 7 0.01637 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.8219591 0.11459 6.828
wigg 2 total 10.689188 0.25402
shell 3 hole 1 0.02782 0.93863 -0.27561 0.2074
shell 3 hole 2 0.02782 -0.13922 0.968299 0.2074
shell 3 hole 3 0.02782 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.5120055 0.08346 9.02
wigg 3 total 11.071275 0.2631
shell4 hole 1 0.0392175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0392175 0.410991 0.899943 0.145567
shell4 hole 3 0.0392175 -0.41099 0.899943 0.145567
shell4 hole 4 0.0392175 -0.64789 -0.7477 0.145567
shell 4 hole
total 6.6011059 0.15687 11.254
shell 4 splash
total 0.8285572 0.01969
wigg 4 total 9.5921596 0.22795
wigg 5 total 9.8787251 0.23476
dilute hole 1 0.064215 0.70639 -0.70639 0.045014
dilute hole 2 0.064215 0.70639 0.70639 0.045014
dilute hole 3 0.064215 -0.70639 0.70639 0.045014
dilute hole 4 0.064215 -0.70639 -0.70639 0.045014
dilute hole 5 4.36539 0.10374 0.10374 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2683617 0.07767 0.07767 -0.46946 -0.88293 0.007156 15.86
100 2.37642 2.37642
183
________________________________________________________________________
Table E.3: Case 1 boundary conditions
sw1 0.013171848 -0.52258 -0.15344 0.83867 4.0166
sw2 0.013171848 -0.35666 -0.41161 0.83867 4.0166
sw3 0.013171848 -0.07751 -0.5391 0.83867 4.0166
sw4 0.013171848 0.22625 -0.49542 0.83867 4.0166
sw5 0.013171848 0.45818 -0.29445 0.83867 4.0166
sw6 0.013171848 0.54464 0 0.83867 4.0166
sw7 0.013171848 0.45818 0.29445 0.83867 4.0166
sw8 0.013171848 0.22625 0.49542 0.83867 4.0166
sw9 0.013171848 -0.07751 0.5391 0.83867 4.0166
sw10 0.013171848 -0.35666 0.41161 0.83867 4.0166
sw11 0.013171848 -0.52258 0.15344 0.83867 4.0166
TOTAL
SWIRLER 6.097 0.144890327
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.969274 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7853494 0.11372 6.828
wigg 2 total 10.607973 0.25209
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027606667 -0.139222 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.0389175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0389175 0.410991 0.899943 0.145567
shell4 hole 3 0.0389175 -0.410991 0.899943 0.145567
shell4 hole 4 0.0389175 -0.647886 -0.7477 0.145567
shell 4 hole
total 6.5506097 0.15567 11.254
shell 4 splash
total 0.8222452 0.01954
wigg 4 total 9.5189403 0.22621
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.063725 0.70639 -0.70639 0.045014
dilute hole 2 0.063725 0.70639 0.70639 0.045014
dilute hole 3 0.063725 -0.70639 0.70639 0.045014
dilute hole 4 0.063725 -0.70639 -0.70639 0.045014
184
________________________________________________________________________
dilute hole 5 4.3321467 0.10295 0.10295 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2435344 0.07708 0.07708 -0.46946 -0.88293 0.007156 15.86
100.00001 2.376420327 2.376420327

sw1 0.014489033 -0.61133 -0.19863 0.76604 4.38343
sw2 0.014489033 -0.37782 -0.52003 0.76604 4.38343
sw3 0.014489033 0 -0.64279 0.76604 4.38343
sw4 0.014489033 0.37782 -0.52003 0.76604 4.38343
sw5 0.014489033 0.61133 -0.19863 0.76604 4.38343
sw6 0.014489033 0.61133 0.19863 0.76604 4.38343
sw7 0.014489033 0.37782 0.52003 0.76604 4.38343
sw8 0.014489033 0 0.64279 0.76604 4.38343
sw9 0.014489033 -0.37782 0.52003 0.76604 4.38343
sw10 0.014489033 -0.61133 0.19863 0.76604 4.38343
TOTAL
SWIRLER 6.097 0.144890327
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.96927 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7853494 0.11372 6.828
wigg 2 total 10.607973 0.25209
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027606667 -0.13922 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.0389175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0389175 0.410991 0.899943 0.145567
shell4 hole 3 0.0389175 -0.41099 0.899943 0.145567
shell4 hole 4 0.0389175 -0.64789 -0.7477 0.145567
shell 4 hole
total 6.5506097 0.15567 11.254
shell 4 splash
total 0.8222452 0.01954
wigg 4 total 9.5189403 0.22621
185
________________________________________________________________________
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.063725 0.70639 -0.70639 0.045014
dilute hole 2 0.063725 0.70639 0.70639 0.045014
dilute hole 3 0.063725 -0.70639 0.70639 0.045014
dilute hole 4 0.063725 -0.70639 -0.70639 0.045014
dilute hole 5 4.3321467 0.10295 0.10295 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2435344 0.07708 0.07708 -0.46946 -0.88293 0.007156 15.86
100.00001 2.376420327 2.376420327

sw1 0.018111291 -0.70773 -0.29315 0.64279 5.213
sw2 0.018111291 -0.29315 -0.70773 0.64279 5.213
sw3 0.018111291 0.29315 -0.70773 0.64279 5.213
sw4 0.018111291 0.70773 -0.29315 0.64279 5.213
sw5 0.018111291 0.70773 0.29315 0.64279 5.213
sw6 0.018111291 0.29315 0.70773 0.64279 5.213
sw7 0.018111291 -0.29315 0.70773 0.64279 5.213
sw8 0.018111291 -0.70773 0.29315 0.64279 5.213
TOTAL
SWIRLER 6.097 0.144890327
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.969274 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7853494 0.11372 6.828
wigg 2 total 10.607973 0.25209
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027606667 -0.139222 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.0389175 0.647886 -0.7477 0.145567
shell4 hole 2 0.0389175 0.410991 0.899943 0.145567
shell4 hole 3 0.0389175 -0.410991 0.899943 0.145567
shell4 hole 4 0.0389175 -0.647886 -0.7477 0.145567
shell 4 hole
total 6.5506097 0.15567 11.254
shell 4 splash 0.8222452 0.01954
186
________________________________________________________________________
total
wigg 4 total 9.5189403 0.22621
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.063725 0.70639 -0.70639 0.045014
dilute hole 2 0.063725 0.70639 0.70639 0.045014
dilute hole 3 0.063725 -0.70639 0.70639 0.045014
dilute hole 4 0.063725 -0.70639 -0.70639 0.045014
dilute hole 5 4.3321467 0.10295 0.10295 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2435344 0.07708 0.07708 -0.46946 -0.88293 0.007156 15.86
100.00001 2.376420327 2.376420327

sw1 0.029589091 -0.52258 -0.15344 0.83867 7.4728
sw2 0.029589091 -0.35666 -0.41161 0.83867 7.4728
sw3 0.029589091 -0.07751 -0.5391 0.83867 7.4728
sw4 0.029589091 0.22625 -0.49542 0.83867 7.4728
sw5 0.029589091 0.45818 -0.29445 0.83867 7.4728
sw6 0.029589091 0.54464 0 0.83867 7.4728
sw7 0.029589091 0.45818 0.29445 0.83867 7.4728
sw8 0.029589091 0.22625 0.49542 0.83867 7.4728
sw9 0.029589091 -0.07751 0.5391 0.83867 7.4728
sw10 0.029589091 -0.35666 0.41161 0.83867 7.4728
sw11 0.029589091 -0.52258 0.15344 0.83867 7.4728
TOTAL
SWIRLER 13.696232 0.32548
wigg 1 top 0 normal to inlet boundary 3.27
wigg 1 total 7.7688287 0.18462
shell2 hole 1 0.016274286 0.63474 -0.73253 0.245984
shell2 hole 2 0.016274286 0.969274 0 0.245984
shell2 hole 3 0.016274286 0.63474 0.732528 0.245984
shell2 hole 4 0.016274286 0.137943 0.959408 0.245984
shell2 hole 5 0.016274286 -0.63474 0.732528 0.245984
shell2 hole 6 0.016274286 -0.969274 0 0.245984
shell2 hole 7 0.016274286 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7937654 0.11392 6.828
wigg 2 total 10.626909 0.25254
shell 3 hole 1 0.027656667 0.93863 -0.27561 0.2074
shell 3 hole 2 0.027656667 -0.139222 0.968299 0.2074
shell 3 hole 3 0.027656667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4913862 0.08297 9.02
wigg 3 total 11.006472 0.26156
shell4 hole 1 0.0389875 0.647886 -0.7477 0.145567
187
________________________________________________________________________
shell4 hole 2 0.0389875 0.410991 0.899943 0.145567
shell4 hole 3 0.0389875 -0.410991 0.899943 0.145567
shell4 hole 4 0.0389875 -0.647886 -0.7477 0.145567
shell 4 hole
total 6.5623922 0.15595 11.254
shell 4 splash
total 0.8239284 0.01958
wigg 4 total 9.5361931 0.22662
wigg 5 total 9.8214962 0.2334
dilute hole 1 0.06384 0.70639 -0.70639 0.045014
dilute hole 2 0.06384 0.70639 0.70639 0.045014
dilute hole 3 0.06384 -0.70639 0.70639 0.045014
dilute hole 4 0.06384 -0.70639 -0.70639 0.045014
dilute hole 5 4.3397211 0.10313 0.10313 0.972876 -0.2307 0.017103 20.2
dilute hole 6 3.2494256 0.07722 0.07722 -0.46946 -0.88293 0.007156 15.86
99.999579 2.37641 2.37642

sw1 0.029533636 -0.52258 -0.15344 0.83867 7.4728
sw2 0.029533636 -0.35666 -0.41161 0.83867 7.4728
sw3 0.029533636 -0.07751 -0.5391 0.83867 7.4728
sw4 0.029533636 0.22625 -0.49542 0.83867 7.4728
sw5 0.029533636 0.45818 -0.29445 0.83867 7.4728
sw6 0.029533636 0.54464 0 0.83867 7.4728
sw7 0.029533636 0.45818 0.29445 0.83867 7.4728
sw8 0.029533636 0.22625 0.49542 0.83867 7.4728
sw9 0.029533636 -0.07751 0.5391 0.83867 7.4728
sw10 0.029533636 -0.35666 0.41161 0.83867 7.4728
sw11 0.029533636 -0.52258 0.15344 0.83867 7.4728
TOTAL
SWIRLER 13.670563 0.32487
wigg 1 total 15.509043 0.36856
shell2 hole 1 0.016245714 0.63474 -0.73253 0.245984
shell2 hole 2 0.016245714 0.969274 0 0.245984
shell2 hole 3 0.016245714 0.63474 0.732528 0.245984
shell2 hole 4 0.016245714 0.137943 0.959408 0.245984
shell2 hole 5 0.016245714 -0.63474 0.732528 0.245984
shell2 hole 6 0.016245714 -0.969274 0 0.245984
shell2 hole 7 0.016245714 -0.63474 -0.73253 0.245984
shell 2 hole
total 4.7853494 0.11372 6.828
wigg 2 total 10.608394 0.2521
shell 3 hole 1 0.027606667 0.93863 -0.27561 0.2074
188
________________________________________________________________________
shell 3 hole 2 0.027606667 -0.139222 0.968299 0.2074
shell 3 hole 3 0.027606667 -0.93863 -0.27561 0.2074
shell 3 hole
total 3.4850742 0.08282 9.02
wigg 3 total 10.987115 0.2611
shell4 hole 1 0.03892 0.647886 -0.7477 0.145567
shell4 hole 2 0.03892 0.410991 0.899943 0.145567
shell4 hole 3 0.03892 -0.410991 0.899943 0.145567
shell4 hole 4 0.03892 -0.647886 -0.7477 0.145567
shell 4 hole
total 6.5510305 0.15568 11.254
shell 4 splash
total 0.822666 0.01955
wigg 4 total 9.5193611 0.22622
wigg 5 total 9.8038226 0.23298
dilute hole 1 0.06373 0.70639 -0.70639 0.045014
dilute hole 2 0.06373 0.70639 0.70639 0.045014
dilute hole 3 0.06373 -0.70639 0.70639 0.045014
dilute hole 4 0.06373 -0.70639 -0.70639 0.045014
dilute hole 5 0 0 0 0.972876 -0.2307 0.017103 20.2
dilute hole 6 0 0 0 -0.46946 -0.88293 0.007156 15.86
100.00126 2.37645 2.37645
189

Dissertation

Uploaded by

Copyright:

Available Formats

Dissertation

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Dissertation

Uploaded by

Copyright:

Available Formats

COMBINING A ONE-DIMENSIONAL EMPIRICAL AND

NETWORK SOLVER WITH COMPUTATIONAL FLUID

Johannes Jacobus Gouws

Submitted in partial fulfilment

in the Faculty of Mechanical/Aeronautical Engineering

Supervisor: Mr R.M. Morris

Johannes Jacobus Gouws

Promoter: Mr R.M. Morris

The purpose of this study is two-fold; (1) to develop a one-dimensional incompressible

Philippians 4:13 I can do everything through Him who gives me strength.

CHAPTER 2: COMBUSTOR DESIGN PRINCIPLES:

CHAPTER 3: 1-D NETWORK APPROACH 45

CHAPTER 4: NUMERICAL MODEL 66

CHAPTER 5: NUMERICAL STUDY 80

CHAPTER 6: SUMMARY, CONCLUSION AND RECOMMENDATIONS 105

APPENDIX A: MATERIAL PROPERTIES 119

APPENDIX C: EMPIRICAL SWIRLER DESIGN CODE 124

APPENDIX D: 1-D EMPIRICAL CODE 132

APPENDIX E: BOUNDARY CONDITIONS 182

Gx = Axial momentum flux

I = Total number of nodes

Due to the ever-increasing computing power, numerical methods such as computational

The purpose of this project is to use numerical modelling to investigate possible

The thermal characteristics of the combustor will be modeled by combining a one-

1.2 Overview of the T56 gas turbine engine

Figure 1.2: The T56 gas turbine engine (www.rolls-royce.com)

Figure 1.3: Single T56 combustion chamber

Table 1.1: Operating conditions during take-off

1.3 Combustor design methodology

As the development of computers progressed, it became a viable option for designers to

A commercially available network solver, FLOWNET (www.flownex.com), was utilised

Due to the hostile environment inside a combustion chamber, it is difficult to investigate

The correct evaluation of CFD models is mainly dependent on the specification of

An alternative and more recent approach is to model a combustion system in a fully

1.4 Need for this study

Despite the increased progress made in numerical methods, one-dimensional analysis is

The purpose of this study is two-fold; (1) to develop a one-dimensional incompressible

In addition to the one-dimensional code, a second empirical code was developed to

1.5 Outline of this study

A discussion of one-dimensional models continues in Chapter 3, elaborating more about

2.2 Basic combustor layout and design

Figure 2.1: Basic combustor features (Mattingly et al., 2002)

Figure 2.2: Explanation of exit temperature profile parameters (Lefebvre, 1998)

2.3 One-dimensional flow calculations

where P1 = total pressure upstream of the hole

K=1+0.64 {2ϕ 2 + ⎡⎣ 4ϕ 4 +1.56ϕ 2 (4α-α 2 ) ⎤⎦ } (2.6)

• Calculate the area-weighted pressure drop across the liner

A convenient way of comparing different fuels in terms of their mixing strength is to

Figure 2.4: Temperature rise curves (Lefebvre, 1998)

Tg =T3 +ΔT (2.11)

ηRZ =0.56+0.44tanh ⎡⎣1.548×10-3 ( T3 +108lnP3 -1863) ⎤⎦ (2.12)

ηPZ =0.71+0.29tanh ⎡⎣1.548×10-3 ( T3 +108lnP3 -1863) ⎤⎦ (2.13)

Tout,RZ =T3 +ηRZ ΔT (2.14)

The temperatures in the remaining zones can be calculated by assuming a linear

Tout,zone =T3 +ηzone ΔTzone (2.16)

Figure 2.5: Heat transfer process

The gas emissivity for luminous gases can be predicted as follows:

q = fuel/air ratio by mass

R1 =0.5σ (1+ε w ) ε g Tg1.5 (Tg2.5 -Tw1