Group Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Title: 17.4.3 - Large Eddy Simulation Study of Gas Turbine Lube Oil Nozzles
ALL VOLUMES CITATION THUMBNAILS PAGE IMAGE ZOOMABLE
Full Citation
STANDARD VIEW MARC VIEW
Permanent Link: http://ufdc.ufl.edu/UF00102023/00424
 Material Information
Title: 17.4.3 - Large Eddy Simulation Study of Gas Turbine Lube Oil Nozzles Computational Techniques for Multiphase Flows
Series Title: 7th International Conference on Multiphase Flow - ICMF 2010 Proceedings
Physical Description: Conference Papers
Creator: Sangli, P.
Kane, P.
Fang, N.
Pugh, D.
Moscarino, G.
Corattiyil, B.
Themudo, R.
Publisher: International Conference on Multiphase Flow (ICMF)
Publication Date: June 4, 2010
 Subjects
Subject: oil nozzle
CFD
LES
 Notes
Abstract: Oil nozzles in gas turbines play a critical role in cooling and lubrication of high-speed bearings and other hardware. Overheating, coking or catastrophic failure might be caused by lack of oil supply. Hence, design of a nozzle with good oil stream integrity is crucial to ensure the right quantity of lube oil to be delivered at the right location for safe operation of gas turbine. CFD analytical validation study of two oil nozzle configurations with experimentally known stream characteristics is presented in this paper. The oil jets, with 40958 Weber number (oil jet) and 13364 Reynolds number (oil jet), fall on the boundary of second wave induced and atomization regimes 1. A transient multiphase flow is modeled using FLUENT (a commercially available CFD code), with Large Eddy Simulation (LES) turbulence model to capture the jet break-up due to drag and vortices around the jet. Analysis predictions with Reynolds Averaged Navier-Stokes (RANS) and LES turbulence models are compared. LES is found to be more appropriate in predicting jet breakup. Effects of solver precision (single/double), time step variation, turbulence models and air-oil interface grid size are studied. The analyses are found to be in good agreement with experiments.
General Note: The International Conference on Multiphase Flow (ICMF) first was held in Tsukuba, Japan in 1991 and the second ICMF took place in Kyoto, Japan in 1995. During this conference, it was decided to establish an International Governing Board which oversees the major aspects of the conference and makes decisions about future conference locations. Due to the great importance of the field, it was furthermore decided to hold the conference every three years successively in Asia including Australia, Europe including Africa, Russia and the Near East and America. Hence, ICMF 1998 was held in Lyon, France, ICMF 2001 in New Orleans, USA, ICMF 2004 in Yokohama, Japan, and ICMF 2007 in Leipzig, Germany. ICMF-2010 is devoted to all aspects of Multiphase Flow. Researchers from all over the world gathered in order to introduce their recent advances in the field and thereby promote the exchange of new ideas, results and techniques. The conference is a key event in Multiphase Flow and supports the advancement of science in this very important field. The major research topics relevant for the conference are as follows: Bio-Fluid Dynamics; Boiling; Bubbly Flows; Cavitation; Colloidal and Suspension Dynamics; Collision, Agglomeration and Breakup; Computational Techniques for Multiphase Flows; Droplet Flows; Environmental and Geophysical Flows; Experimental Methods for Multiphase Flows; Fluidized and Circulating Fluidized Beds; Fluid Structure Interactions; Granular Media; Industrial Applications; Instabilities; Interfacial Flows; Micro and Nano-Scale Multiphase Flows; Microgravity in Two-Phase Flow; Multiphase Flows with Heat and Mass Transfer; Non-Newtonian Multiphase Flows; Particle-Laden Flows; Particle, Bubble and Drop Dynamics; Reactive Multiphase Flows
 Record Information
Bibliographic ID: UF00102023
Volume ID: VID00424
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: 1743-Sangli-ICMF2010.pdf

Full Text

Paper No 7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

Large Eddy Simulation (LES) Study of Gas Turbine Lube Oil Nozzles
Pradeep Sangli 1, Prasad Kane 2, Ning Fang 3, Dave Pugh 3, Gary Moscarino 3
Bala Corattiyil 3, Ramon Themudo

'John F Welch Technology Centre, Plot #122 EPIP, Whitefield Road, Bangalore 560066, INDIA
pradeep.sangli@ge.com
former employee of John F Welch Technology Centre, Plot #122 EPIP, Whitefield Road, Bangalore 560066, INDIA
3GE Aviation, One, Neumann Way, Cincinnati, OH 45215 USA


Keywords: Oil Nozzle, CFD, LES


Abstract

Oil nozzles in gas turbines play a critical role in cooling and lubrication of high-speed bearings and other hardware.
Overheating, coking or catastrophic failure might be caused by lack of oil supply. Hence, design of a nozzle with good oil
stream integrity is crucial to ensure the right quantity of lube oil to be delivered at the right location for safe operation of gas
turbine.
CFD analytical validation study of two oil nozzle configurations with experimentally known stream characteristics
is presented in this paper. The oil jets, with 40958 Weber number (oil jet) and 13364 Reynolds number (oil jet), fall on the
boundary of second wave induced and atomization regimes [1].
A transient multiphase flow is modeled using FLUENT (a commercially available CFD code), with Large Eddy
Simulation (LES) turbulence model to capture the jet break-up due to drag and vortices around the jet. Analysis predictions
with Reynolds Averaged Navier-Stokes (RANS) and LES turbulence models are compared. LES is found to be more
appropriate in predicting jet breakup. Effects of solver precision (single/double), time step variation, turbulence models and
air-oil interface grid size are studied. The analyses are found to be in good agreement with experiments.


Introduction

Breakup patterns of the lube oil jet can be sub-divided into
Rayleigh, 1st wave induced, 2nd wave induced and the
atomization regime [1]. The distance at which breakup
start depends on the breakup regime. In the atomization
regime, breakup of the jet starts almost at the point where
the fluid comes out of the nozzle [2]. The phenomenon
causing the jet break-up is governed by instability.
Interaction of surrounding air also contributes jet instability.
Surface of the oil jet becomes wavy while it flows through
the surrounding air before reaching the target.
Perturbations (waves) start growing on the jet with
different wavelengths due to instability. In the
atomization regime jet breakup is governed by
Kelvin-Helmholtz instability theory, due to high relative
velocity between oil jet and surrounding quiescent air (in
the test rig). When the fluids with density differences flow
with large relative velocity, which when exceeds a critical
velocity, waves on the interface start growing and cause jet
breakup [3].

Nomenclature

RANS Reynolds Averaged Navier Stokes
LES Large Eddy Simulation


Geometry Configurations

Figure 1 shows the two geometry configurations that were
studied. Both the configurations have an outlet of 0.110


inches (0.2794 cm). The inclusion of an insert in
configuration 2 increases the length of the outlet arm by
0.035 inches (0.0889cm).





Configuration 1 Configuration 2

Insert



Dia. 0.2794 cm Dia. 0.2794 cm I


Figure 1: Geometry Configurations


Experimental Facility

Figure 2 shows the experimental setup, which shows oil
being injected from a nozzle exit into a quiescent
environment. The oil spray encounters a stationary flat
plate at a distance of 1 inch (2.54cm), which has an oil
collection hole of 0.3 inches (0.762 cm) diameter along the
line from the nozzle exit. Configuration 1 was found to
show a lot of jet spreading and disruption and only 71% oil
got collected at the target hole. Configuration 2 nozzle was
found to be more efficient in maintaining the oil jet
integrity and delivered 91% of oil at the 0.3" target hole.






Paper No


Nozzle assembly


Flat plate

0.3" hole


Figure 2: Experimental oil nozzle test setup

The CFD domain replicates the experimental set up. The
nozzle hardware is modeled as Domain 1 and the jet in
quiescent air is modeled as Domain 2 as shown in figure 3.
The radius 5-inch (12.7cm) of Domain 2 is determined
with a sensitivity study.


T Domain 1


[ umcuui Domain 2 J__



Figure 3: Cross-section of the mesh and domains

CFD Model

Resolution of interface region is very important to be able
to see the jet breakup and correctly capture oil spread
effect [3][4]. Figure 3 shows a section with distribution of
hexahedral mesh in the domain. Total number of elements
in the domain 2 are 0.7 million. Most of the mesh is
concentrated in the region of jet and its interface with
surrounding air. In the radial direction mesh size is
controlled at 25 microns and at 100 microns in the axial
direction. The mesh is controlled near the wall for a y+ of
30.
First, a steady state single-phase (oil only)
analysis of Domain 1 is done. The profiles of three velocity
components (u, v, w), turbulent kinetic energy (TKE) and
turbulent dissipation rate (TDR), are extracted and
imposed on the domain interface as inlet boundary
condition for Domain 2 which is modeled as a transient
multiphase analysis.


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

CFD Multiphase Model and Turbulence Model
Selection

Capturing the air-oil interface was crucial for this analysis.
Hence, the Geo-reconstruct VOF model was selected
which captures the air-oil interface in a piece-wise linear
way in every cell. The analysis time and jet breakup is
found to be very sensitive to the time step. If time step is
very large the instability is not resolved and breakup is not
observed. Time step in the range of 0.0001 sec to 5e-08 sec
is used for the analysis [4]. Analysis is performed with
both single and double precision solvers. Ponstein's theory
of secondary instability says that a vortex ring forms
around the jet, which further breaks into small droplets [5].
To capture this effect, turbulence models used in the
analysis are K-Epsilon (RANS) two equation model and
Large Eddy Simulation (LES) model. In LES large energy
containing eddies are computed and rest are modeled with
Smagorinsky-Lilly sub-grid scale [6][7].


Results and Discussion:

Domain 1: Steady State Single Phase Analysis

Figure 4 shows turbulent kinetic energy (TKE) contours of
Domain 1 steady state analysis results of both
configurations. The flow separation and re-circulation zone
in configuration 1 causes skewness in TKE profile. TKE
plot on the exit plane shows higher turbulence in the oil
flowing through lower part of the nozzle. TKE plot for
configuration 2 on the exit plane of the nozzle looks more
uniform with a concentric pattern.


Sso0..3 Configuration 1
97435+03 05 3
..S... Exit Plane |...




0 0 ......02
I+03 35*03


7 /e+02 25 e02
51:1+02 1 Z::00:02

S7oo 02 I:0
S r00+02 c i a00ti0
r2 e -c ir2u2t02
t 0e+02 2 520L*02
. re-circulation.


Configuration 2


Figure 4: TKE Profiles of both the Configurations


Domain 2: Transient Multiphase Analysis

Figure 5 shows transient analysis oil volume fraction plot
of domain 2-configuration 1 nozzle. We can observe the
breakup phenomena of the jet with droplets separating
from the jet. Another important observation of the oil
volume fraction plot shows that the centre of the jet at 0.3"
outlet has shifted downwards.
Twenty equally spaced post-processing planes
were used as shown in Figure 5 to study the state and
shape of the oil jet as it flows from 0.1" inlet to 0.3" outlet.


r






Paper No


At plane zero (inlet) the oil jet is circular and at further
planes we see disrupted jet with smaller core surrounded
by oil ligaments and droplets. LES turbulence model
proved to be effective in capturing eddies around the jet.


7th International Conference on Multiphase Flow
ICMF 2010, Tampa, FL USA, May 30-June 4, 2010

observed to be more stable with smooth convergence
than Single precision (SP) with 5e-08 sec time step.
SP solver predicted 74% oil at 0.3" outlet and DP
solver predicted results very close to test data with
71.2% oil collected at 0.3" hole. Initial cyclic behavior
of mass flow at outlet stabilizes after about 2.5
physical times.


..." Jet break up ,





SOOe-l01

0 ooe-ol,
15oe-ol






I I:; -l\
0.-0e-0 -- N_


Config. 2


-.- -'~I


Figure 5: Domain 2, Configuration 1: Oil volume fraction
plot: Ligaments and droplets around the jet

The two graphs in Figure 6 show variation of
oil mass flow at section planes from inlet to outlet.
The plane locations are marked in the Figure 5. Both
the figures show three curves for three different mesh
sizes 0.9, 0.7 and 0.5 million. The eddy size resolved
depends on grid resolution in LES model. The
prediction using LES model are closer and mass flow
looks more consistent with physics at most of the
plane locations.


RANS


---0.9M
-m-0.7M
n sM


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Plane location along the jet


LES -0-o
-=-0.7M
0.5M






)6 *-- *--I -*--* t- *- *- *- *- -- S-t- *- *- *-- *-- t
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
Plane location along the jet


Figure 6: Comparison of RANS and LES turbulence
models

Jet physical time, which is 0.001 sec, is the
time required for an oil particle travel from inlet to
outlet with an average velocity. Figure 7 shows the
mass flow rate variation with jet physical time. Double
precision (DP) analysis with le-4 sec time step is


-- FDSP LTS
A- Exp


Ol entenng target Oil lost



-I I
CFD_SPLTS O.26551bm/sec (90 1%) 0.02931bm/s (99A)
Exp 0.26821bm/c '910% 0.0266,b./. (90o)




--CFD_SP_LTS
Con4fig. 1 -Exp
--CFD DP HTS
Oil entering target Oil lost
CFD SP LTS 0.21781bm/sec 74%) 0.07701bm/ (26%)
Exp 0.20901bm/sec (71%) 0.08581bI/ (29%)


0 05 1 15
JetPhyl.i.tim e


2 2.5


Figure 7: Validation of test data with CFD results

Conclusions:

Skewed velocity profiles resulted in poor efficiency of
Configuration 1 nozzle. The Geo-reconstruct VOF
model scheme with LES turbulence model provided a
sharp interface between the two phases and captured
ligaments and large primary breakup. Mesh resolution
at the critical zones was found to be extremely
important to capture the breakup. This work has a
tremendous potential to predict the oil nozzle
performance and reduce the nozzle design cycle time.

ACKNOWLEDGEMENTS
The authors would like to thank GE Aviation for the
support and permission to publish this work.

REFERENCES
[1] S.P. Lin, R.D. Reitz, 1998, "Drop and spray
formation from a liquid jet", Annu. Rev. Fluid
Mech., 30:85-105
[2] T. Funada, D.D.Joseph, 2001 "Viscous potential
flow analysis of Kelvin-Helmholtz instability in a
channel", J, Fluid Mech.
[3] Vineet Dravid, Spring 2000, "Computational
model for instability analysis of a 2-D
non-Newtonian cylindrical ligament", Final
Project, Dept. of Mech. Engg., Purdue Univ.
[4] "Rocket Injector Science", Fluent News, VOL.
XIII, Issue2, Fall 2004
[5] Suk Goo Yoon, Dec. 2002, "A fully non-linear
model for atomization of high-speed jets", Purdue
University, PhD thesis report.
[6] Sung-Eun Kim, Davor Cokljat, "LES is more", ,
Fluent News, Vol. XIV, Issuel, Spring 2005
[7] 1 iung up LES for the sandia flame", Graham
Goldin, Fluent News, Vol. XIV, Issuel, Spring
2005


0.16

S0.14

S0.12

0 o0

0.08
o 08




University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.7 - mvs