Cfd analysis of turbulence in a gas turbine combustor with reference to the context

INTERNATIONAL JOURNALEngineering and TechnologyRESEARCH IN
International Journal of Advanced Research in
OF ADVANCED (IJARET), ISSN 0976 –
6480(Print), ISSN 0976 – 6499(Online) Volume 4, Issue 2, March – April (2013), © IAEME
ENGINEERING AND TECHNOLOGY (IJARET)
ISSN 0976 - 6480 (Print)
ISSN 0976 - 6499 (Online)
IJARET
Volume 4, Issue 2 March – April 2013, pp. 01-07
© IAEME: www.iaeme.com/ijaret.asp ©IAEME
Journal Impact Factor (2013): 5.8376 (Calculated by GISI)
www.jifactor.com

CFD ANALYSIS OF TURBULENCE IN A GAS TURBINE
COMBUSTOR WITH REFERENCE TO THE CONTEXT OF EXIT
PHENOMENON

a b
P.S. Jeyalaxmi , Dr.G.Kalivarathan
a
Research Scholar, CMJ University, Meghalaya, Shillong.
b
Principal/ PSN Institute of Technology and Science, Tirunelveli, Tamilnadu, Supervisor,
CMJ University, Shillong. Email:sakthi_eswar@yahoo.com

ABSTRACT

The main objective of this investigation is based on the interaction between the
combustor and turbine with respect to the increased temperatures and heat transfer related
aspects. All classic secondary flow models involved a key assumption flux the flow at the
inlet to the turbine was either a well-behaved, two-dimensional turbulent boundary layer with
a uniform mean field at low turbulence, or a constant total pressure field. While these studies
have led to great insight into the nature and driving forces of secondary flows, they neglected
to consider that the flow exiting the combustor may be complicated. Experimental and
computational studies have also been done to attempt for quantifing the effects of a variety of
inlet conditions on the secondary flows, heat transfer, and aerodynamic losses in the turbine.
Some of the conditions considered include Reynolds number, boundary layer thickness,
endwall and vane geometry, and temperature field. These studies have greatly improved the
understanding of secondary flows. Many experimental and computational studies have been
completed to model combustor core flow and exit profiles independently of the turbine. It is
observed that non-uniformities in the flow field exist in the span wise and pitch wise
directions in temperature, pressure and velocity.

Keywords: CFD, Combustor, Turbulence, Endwall, Vane Geometry

1

International Journal of Advanced Research in Engineering and Technology (IJARET), ISSN 0976 –

1.0 INTRODUCTION

It is important to reiterate that a key assumption made in the flow models presented is
that the inlet flow was a two-dimensional turbulent boundary layer. This is a critical
assumption for an engine where the flow exiting the combustor is complex. In addition, the
data and measurements for these studies were taken on a variety of both stator and rotor
vanes with varying turning angle and geometries, which could account for differences in the
evolution of the suction side vortex. By relating the measured surface heat transfer to the
measured flow fields, it was shown that in general high Stanton numbers occurred where
mainstream fluid was brought down to the endwall surface as a result of the downward legs
of the vortices. As the passage vortex turned upwards, the Stanton number is decreased. For
the Reynolds number comparison, a higher peak Stanton number occurred at the leading edge
for the lower Reynolds number. Another area of interest has been the effect of temperature
gradients in the incoming flow on the flow patterns through the passage and heat transfer on
the vane surface and endwall. This effect is critical to understand given the large temperature
variations which would be expected exiting a combustor where 50% or more of the flow may
be cooler fluid from dilution or other forms of cooling. The radial profile caused decreased
heat transfer on the stator and a slight increase on the rotor. The increase was due again to the
segregation of hot and cold fluid with the hot migrating towards the pressure surface. The
pitchwise varying profile was found to cause a large heat load increase, but only if the hot
spot was directly aligned in front of the vane. The geometry used did not have the all the
elements necessary to completely represent a realistic combustor liner; however, it was
clearly identified from this study that changes occurred in the total pressure. Details are not
available for comparing any effects that the two different total pressure profiles had on the
secondary flow field in the vane section. The cascade was located first in a low turbulence
uniform hot gas stream, second in a high-turbulence uniform hot gas stream, and finally in a
high-turbulence, non-uniform hot gas stream downstream of a combustor exit. The effect of
high turbulence was to reduce the cooling effectiveness on the suction surface by 10%, and
the high turbulence and non-uniform flow decreased the effectiveness by 21%.

2.0 NO VANE CASES COMPUTATIONAL DOMAINS

The three combustor model cases with no downstream turbine vane differed only in
the type of cooling schemes employed in the combustor. The baseline case, which directly
modeled the baseline experimental case, utilized the complete cooling scheme including axial
film cooling holes over four liner panels, two rows of dilution jets, and a coolant slot at the
combustor-turbine interface. The second case included the film cooling and exit slot but
neglected to add the dilution flow thus enabling the examination of the effects of the dilution
jets on the combustor flow field. The third case considered film cooling and dilution, but
eliminated the exit slot allowing us to see the effects of the slot flow on the exit profiles. The
three cases with no vane all had approximately the same computational domain with the
exception of the angle of contraction for the slot versus the no slot cases (17.15° versus
15.47°). It was shown computationally that the effect on the flow field of changing the angle
was negligible.

2


3.0 DOWNSTREAM VANE CASES COMPUTATIONAL DOMAINS

The combustor geometry of the first vane case was equivalent to the baseline case; film
cooling, dilution and an exit slot with the addition of the vane. The second case contained film
cooling and dilution, but no exit slot. The last case modeled film cooling, dilution and the exit
slot, but the film cooling holes were oriented with a 45° compound angle in addition to the 30°
stream wise angle. This compound angle model was computed to compare with the axial hole
model in order to determine the differences produced in the combustor exit profile. The three
vane cases had approximately 2.3 million cells each and modeled 640 cooling holes and three
dilution jets. The slot case modeled seven slot feed holes and 52 slot pin fins. The vane stagnation
was located 3.6 cooling slot step heights downstream of the trailing edge of panel four. The
outflow boundary was 1.5 vane chord lengths downstream of the vane trailing edge. An
additional 0.1 chord lengths was added to the boundary in the streamwise direction to avoid
highly skewed cells at the outflow. The compound angle hole case with the turbine vane was
identical to the axial hole case except that the periodic side boundaries, rather than being set
through the centerline of a row of cooling holes, are offset halfway between rows of holes to
account for the 45° angle. The number of holes and width of the domain at one pitch still
remained the same for the compound angle case

4.0 GOVERNING EQUATIONS AND SOLUTION METHODS

The numerical method used in all the solutions was a segregated solution algorithm with a
control-volume based technique. All solutions were computed using Fluent Version 5.0 (Fluent,
Inc., 1998). Fluent also offers a coupled solver which also uses a control-volume based technique.
This general segregated solution technique consists of an integration of the governing equations
of mass, momentum, energy and turbulence on the individual cells within the computational
domain to construct algebraic equations for each unknown dependent variable. The discretized
equations are then linearized and a solution of the resulting system of linear equations gives
updated values of the dependent variables.

5.0 INITIALIZATION

In judging the convergence of the solution, the residuals of several quantities were
monitored. The normalized residuals for continuity, x-momentum, y-momentum, z-momentum,
energy, k and ε were monitored after each iteration. For the RSM turbulence model, the residuals
for the Reynolds stresses were also monitored. As the solution proceeded, the residuals decayed
to some small value and then leveled out. The un-scaled residual was calculated as the imbalance
in the conservation equation of a particular variable over all the computational cells. The residual
was then scaled using a factor representative of the flow rate of the variable through the domain.
For the continuity equation the scaling factor was the largest absolute value of the continuity
residual in the first five iterations. There is no universal law for judging convergence of a
solution. The Fluent default setting requires that the scaled residuals decrease to 10-3 for all
equations except energy, for which the criterion is 10-6. A second approach for judging
convergence is to require that the un-scaled residuals drop by three orders of magnitude. For all
cases computed, the convergence criteria were set at 10-4 for all equations except energy, which
was set at 10-7. Each computation was continued 50-100 iterations beyond convergence to insure
that the residuals continued to decrease steadily and that the solution was actually converged.

3


6.0 TURBULENCE MODELING AND NEAR WALL TREATMENT

Based on these findings, all six models were initially computed using the RNG k-ε
model with non-equilibrium wall functions for the near-wall treatment. For one case, a study
of the turbulence modeling was conducted. The no vane, slot, dilution case was computed
with the original RNG k-ε boundary conditions, with the RNG k-ε model with a different
inlet turbulence boundary condition, and with the Reynolds Stress Model (RSM). Fluent
allows solution-based grid adaption, which is an extremely valuable tool. It provides the
ability to adapt the grid based on specific values or gradients of important flow characteristics
such as velocity, temperature or turbulence. It also allows adaption based on a range of
desired wall unit values, or on specific flow boundaries. A third alternative for adaption is to
target cells with a certain volume or certain volume change between cells since large volume
change between adjacent cells may cause convergence problems within the mesh.

Figure 1. a, b Global and local coordinate systems and reference lengths for combustor and
vane sections.

4


7.0 COMPUTATIONAL RESULTS

The effect of the dilution jets on the flow patterns in the combustor core is clearly
seen by looking at the in-plane velocity vectors without and with dilution in the stagnation
plane cutting through the first row of dilution jets. The disruptions from the dilution injection
to the uniformity of the flow field resulted in a strong secondary flow pattern at the
combustor exit. In addition to the non-uniform velocity field, the dilution jets created non-
uniformities in the temperature and pressure fields at the combustor exit. The normalized
temperatures are calculated based on a mass averaged temperature for each case at the
combustor exit. Negative values of θ are seen where the temperature in the combustor is less
than the mass average temperature of the mainstream flow and the coolant flow. For the no
dilution case, the temperature at the center of the combustor was the same as the mainstream
inlet temperature, and the effect of the film on the liner can be clearly seen. Once dilution
was added with the secondary flows induced by the dilution jet injection caused a reduction
in the film effectiveness close to the wall with hot spots near the vane stagnation regions.

Figure 2. Comparison of Cp contours in plane CS (x/L = 1.02) for case 4 (slot, no vane,
dilution) using the RNG k-and RSM turbulence models with experimental data.

5


Figure 3. Comparison of pitchwise spatially averaged Cp across the span for case 4 (no vane,
slot, dilution) for the RNG k-and RSM models and experimental data in plane CS (x/L = 1.02).

8.0 CONCLUSION

It is seen that the combustor exit profiles affects the development of the secondary
flow field through the turbine vane passage for the downstream vane cases. The secondary
flows were driven by the incoming total pressure profiles created as a result of the various
combinations of dilution, slot injection, and film-cooling hole orientation. The no slot case
shows a relatively uniform total pressure profile approaching the vane passage in the near-
wall region and thus no leading edge vortex was formed. Some vertical motion was present
above 20% of the span due to the increase in total pressure at the midspan. The addition of
the slot produced a significant change in the total pressure profile creating a low pressure spot
in the near-wall region and additional pitchwise non-uniformities. The slot case shows the
development of the leading edge vortex, but still no distinct passage vortex was detected.
Additional vortical motion above 20% of the span was present as in the no slot case due to
the increase in total pressure at the midspan. The addition of the slot also led to increase the
average effectiveness levels on the endwall.

REFERENCES

1. Anand, M. S., Zhu, J., Connor, C., and Razdan, M. K. (1999) “Combustor Flow Analysis
Using and Advanced Finite-Volume Design System,” ASME Paper No. 99-GT-273.
2. Barringer, M.D., Richard, O.T., Stitzel, S.M., Walter, J.P. and Thole, K.A. (2001) “Flow
Field Simulations of a Gas Turbine Combustor,” to be presented at IGTI 2001, New
Orleans, LA.
3. Barringer, M.D (2001), “Flow Field Simulations of a Gas Turbine Combustor,” thesis,
Mechanical Engineering Department, Virginia Polytechnic Institute and State University,
to be completed May 2001.
4. Bicen, A. F., Tse, D. and Whitelaw, J. H. (1988) “Flow and Combustion Characteristics
of an Annular Combustor,” Combustion and Flame 72: 175-192.

6


5. Boyle, R. J. and Giel, P. W. (1997) “Prediction of Non uniform Inlet Temperature Effects
on Vane and Rotor Heat Transfer,” ASME Paper No. 97-GT-133.
6. Burd, S. W. and Simon, T. W. (2000) “Effects of Slot Bleed Injection Over a Contoured
Endwall on Nozzle Guide Vane Cooling Performance: Part I – Flow Field
Measurements,” ASME Paper No. 2000-GT-199.
7. Burd, S. W., Satterness, C. J. and Simon, T. W. (2000) “Effects of Slot Bleed Injection
Over a Contoured Endwall on Nozzle Guide Vane Cooling Performance: Part II –
Thermal Measurements,” ASME Paper No. 2000-GT-200.
8. Holdeman, V. D. (1993) “Mixing of Multiple Jets with a Confined Subsonic Crossflow,”
Prog.Energy Combust. Sci, Vol. 19, pp. 31-70.
9. Kang, M., Kohli, A., and Thole, K. A., 1999, “Heat Transfer and Flow field
Measurements In The Leading Edge Region of A Stator Vane Endwall,” Journal of
Turbomachinery, Vol.121, No. 3, pp. 558-568 (also presented as ASME Paper 98-GT-
173).
10. Kang, M., and Thole, K. A., 2000, “Flowfield Measurements in the Endwall Region of a
Stator Vane,” accepted to the Journal of Turbomachinery, Vol. 122, pp. 458-466.
11. K. V. Chaudhari , D. B. Kulshreshtha and S. A. Channiwala, “Design and Experimental
Investigations of Pressure Swirl Atomizer of Annular Type Combustion Chamber for 20
KW Gas Turbine Engine” International Journal of Advanced Research in Engineering &
Technology (IJARET), Volume 3, Issue 2, 2012, pp. 311 - 321, ISSN Print: 0976-6480,
ISSN Online: 0976-6499.
12. Tarun Singh Tanwar, Dharmendra Hariyani and Manish Dadhich, “Flow Simulation
(CFD) & Static Structural Analysis (FEA) of a Radial Turbine” International Journal of
Mechanical Engineering & Technology (IJMET), Volume 3, Issue 3, 2012, pp. 252 - 269,
ISSN Print: 0976 – 6340, ISSN Online: 0976 – 6359.
13. Ashok Tukaram Pise and Umesh Vandeorao Awasarmol, “Investigation of Enhancement
of Natural Convection Heat Transfer from Engine Cylinder with Permeable Fins”
International Journal of Mechanical Engineering & Technology (IJMET), Volume 1,
Issue 1, 2010, pp. 238 - 247, ISSN Print: 0976 – 6340, ISSN Online: 0976 – 6359.
14. Cherian Paul and Parvathy Venugopal, “Modelling of Interfacial Heat Transfer
Coefficient and Experimental Verification for Gravity Die Casting of Aluminium Alloys”
International Journal of Mechanical Engineering & Technology (IJMET), Volume 1,
Issue 1, 2010, pp. 253 - 274, ISSN Print: 0976 – 6340, ISSN Online: 0976 – 6359.
15. A.Saravanapandi Solairajan and Dr.G.Kalivarathan, “Investigation of Heat Transfer
Through CNT Composites-Focusing on Conduction Mode” International Journal of
Mechanical Engineering & Technology (IJMET), Volume 4, Issue 1, 2013, pp. 66 - 73,
ISSN Print: 0976 – 6340, ISSN Online: 0976 – 6359.

7

Cfd analysis of turbulence in a gas turbine combustor with reference to the context

More Related Content

What's hot (19)

Viewers also liked (9)

Similar to Cfd analysis of turbulence in a gas turbine combustor with reference to the context (20)

More from IAEME Publication (20)

Cfd analysis of turbulence in a gas turbine combustor with reference to the context