International Journal of Trend in Scientific Research and Development (IJTSRD)

Volume 4 Issue 3, April 2020 Available Online: www.ijtsrd.com e-ISSN: 2456 – 6470

Comparative Study on Effective Turbulence Model for

NACA0012 Airfoil using Spalart – Allmaras as a Benchmark
R. Allocious Britto Rajkumar1, N. Mohammed Raffic2, Dr. K. Ganesh Babu3, V. Vignesh4
1,2Assistant
Professor, 3Principal, 4UG Scholar,
1Department of Aeronautical Engineering, Nehru Institute of Technology, Coimbatore, Tamil Nadu, India

2Department of Mechanical Engineering, Nehru Institute of Technology, Coimbatore, Tamil Nadu, India

3Chendhuran College of Engineering & Technology, Pudukottai, Tamil Nadu, India

4Department of Aeronautical Engineering, Nehru Institute of Technology, Coimbatore, Tamil Nadu, India

ABSTRACT How to cite this paper: R. Allocious Britto

The present work is dedicated to conduct a comparative study on identifying Rajkumar | N. Mohammed Raffic | Dr. K.
the effective turbulence model in terms of flow outlet velocity, error Ganesh Babu | V. Vignesh "Comparative
percentage, number of iterations and time coefficient using NACA0012 airfoil Study on Effective Turbulence Model for
by considering Spalart – Allmaras as a reference model. The current study NACA0012 Airfoil using Spalart –
considers 12 different turbulence models including Spalart – Allmaras for Allmaras as a Benchmark" Published in
obtaining the output characteristics individually. The turbulence models in International Journal
existence such as Standard K-epsilon, RNG K-epsilon, Realizable variant of K- of Trend in Scientific
epsilon, Standard K-Omega, SST K-Omega, BSL K-Omega, Transition K-KL Research and
Omega, Transition SST, Reynolds Stress (Linear Pressure Strain), Reynolds Development
Stress (Quadratic Pressure Strain), Reynolds stress (Stress Omega) have been (ijtsrd), ISSN: 2456-
utilized for the evaluation and comparison. The NACA0012 airfoil is modelled 6470, Volume-4 |
using CATIA and the meshed model of the airfoil is analyzed using ANSYS Issue-3, April 2020, IJTSRD30824
FLUENT under standard boundary conditions. The results obtained have pp.1049-1056, URL:
shown that the Standard K – epsilon model is found to have less error www.ijtsrd.com/papers/ijtsrd30824.pdf
percentage in comparison to other turbulence models. The count over the
number of iterations taken reveals that the models such as Standard K-omega, Copyright © 2020 by author(s) and
SST K-omega and BSL K-omega has shown the least number of iterations International Journal of Trend in Scientific
compared to rest of the turbulence models for completing the analysis. The Research and Development Journal. This
time coefficient calculation shows that Standard K-omega and SST K-omega is an Open Access article distributed
ranks top by showing less time for conducting the analysis with 77.92 seconds under the terms of
and the maximum time was shown by the Reynolds’s stress models the Creative
considered in the study. Commons Attribution
License (CC BY 4.0)
KEYWORDS: CFD, NACA0012 Airfoil, Spalart – Allmaras, Time Coefficient, (http://creativecommons.org/licenses/by
Turbulence Models /4.0)
1. INTRODUCTION
Computational Fluid Dynamics (CFD) has become a field Power, Turbo machinery, Construction, Hydraulics,
with higher utilization in many engineering disciplines due Biomedical, Sports, Medicine and Space studies makes it an
to its flexibility and capability to handle and solve complex inevitable method of understanding the difficulties and
engineering problems. The domain comes under the branch avoiding failure of the product or process at later stages
of physics which involves the flow analysis of fluids and which may result in huge loss in terms of time, money and
gasses with different boundary conditions. Computational energy. Software’s such as ANSYS FLUENT, OPENFOAM has
Fluid Dynamics (CFD) is the combination of various fields the capability to conduct the flow analysis with varying
such as physics, flow technology, computing power, boundary conditions. CFD has specific applications in
mathematics and fluid mechanics. It is a group of techniques aerospace industry such as aerodynamic design of vehicles,
combined together in solving the NavierStokes equations or combustion modelling, performance of aircraft components
strictly, RANS equations (Reynolds-Averaged Navier-Stokes) such as turbochargers, propellers and cooling fans etc.
by satisfying the conservation of mass, momentum and
energy to predict the behavior of fluidic systems. The Nicolas Pellerin et.al [2] investigated the turbulent flow over
numerical methods such as Finite Volume method, Finite NACA0012 airfoil with Reynold’s number 5×105 by using
Difference method and Finite Element method are involved LBM method with multi - domain grid refinement, cascaded
in solving complex engineering problems and Finite Volume collision operator and considered Spalart – Allmaras model
methods is the most preferred solver in case of CFD analysis to compare the results in terms of force coefficients,
[1]. pressure profiles, and velocity profiles. Olubunmi Popoola
et.al [3] has investigated the accuracy of the turbulence
The most promising nature of the CFD domain in providing models such as Standard, RNG, Realizable k-ɛ Models,
space for the analysis of complex situations prevailing the Standard and SST k-ω Models, Transition k -kL-ω Model and
product design and development of fields like Electronics, the Transition SST Model for the simulation of higher heat

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1049
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
transfer rate for Reciprocating Mechanism Drive Heat Loop k-ω model is the best turbulence model for simulating flow
device by using CFD solver and they have found that the around both clean and iced wind turbine airfoils.
Standard k-ɛ Models provides the least prediction and the
RNG k-ɛ Models has higher accuracy than all other models. 2. Turbulence Models
Manuel Garcia Pérez, Esa Vakkilainen [4] has made a The current sections explains about the various turbulence
comparison on the turbulence models such as URANS k-ɛ models present in the study for conducting the flow
and DES and the two and three dimensional mesh simulations In total the study considers 12 different
approaches in terms of their capability of providing highly turbulence models which are in existence and from that the
accurate results and computational cost by considering the model Spalart-Allmaras has been considered as a standard
unsteady CFD ash deposition tools. A. Riccia et.al [5] has for comparison. Turbulence flow modelling through
investigated the impact of the turbulence model and computational software has an objective to predict the
roughness height selection 3D steady RANS simulations of quantities of interest such as fluid velocity, pressure and
wind flow in an urban environment and they have found that other related characteristics. Turbulence models can be
the turbulence models have more impact comparing to the classified based upon the computational cost, finer the
surface roughness parameters considered. Abdolrahim et.al resolution of the simulation, length of the turbulent scales,
[6] has conducted a study to understand the accuracy of the accuracy.
turbulence models for CFD simulations of vertical axis wind
turbines using the one equation and two equation models 2.1. Spalart – Allmaras
with an additional intermittency transition model (SSTI) From the day of its introduction on 1992, this model is found
models. The authors have reported that models such as SA, to have more application in the field of aerodynamics in
RNG, Realizable k-ε and k-kl-u models failed to reproduce solving problems with turbulent flow. This method is
the aerodynamic performance of VAWTs. The variables of considered to be the most efficient and effective model for
SST model is found to have good agreement with the conducting the aerodynamic flow analysis in structures such
experimental data sets than other models. Tao Zhi et.al [7] as airfoils, wings, fuselages, missiles and ship hulls by the
has studied the effect of turbulence models in predicting the simulation community as it takes lesser time, iterations and
convective heat transfer to hydrocarbon fuel at supercritical cost to provide results of higher accuracy [16] . This one –
pressure and found that SST model and Launder and Sharma equation turbulence models is used in aerospace
model performed well compared to other models. Mohamed applications with wall bounded flows and it has shown
M.Helal et.al [8] studied the numerical simulation of appreciable results for problems with boundary layers
cavitation flow over marine propeller blades using transition subjected to adverse pressure gradients. This method has
–sensitive turbulence model in comparison with the obvious applications in the field of turbo machinery due to
standard K-epsilon model and they have found that the its potential to provide results with higher accuracy.
prediction based upon k-kl-ω transition model has good
agreement with the experimental data set. V.K.Kratsev et.al 2.2. Standard K-ε Model
[9] has reviewed the role of URANS and LES hybrid A two equation model which gives a general description of
simulations in the field of internal combustion engine turbulent flows by means of two transport equation in the
application for the development and optimization. Feng Gao form of PDE’s. The model has more utilization in CFD
et.al [10] attempted to compare the different turbulence softwares for the simulation of turbulent flows at varying
models in simulating unsteady flow and they have found that boundary conditions. The transport equations are not in
robustness of the standard turbulence models are superior integration to the walls but the factors such as production
to low Reynolds’s model. M. Ghafari and M.B. Ghofrani [11] and dissipation of kinetic energy are well specified in the
has studied about the effects of overestimation or near wall using the wall’s logarithmic law.
underestimation of turbulence characteristics at the
interface by using a new turbulence model. Jia-Wei Han et.al 2.3. RNG K-ε Model
[12] conducted a study on various CFD turbulence models on Renormalization Group K-ε model is a mathematical
for refrigerated transport of fresh fruit and found that SST k- technique which results in the modified form of standard K-ε
ω models have shown agreement with measured model to make an attempt for the different scales of motion
experimental data set. Aoshuang Ding et.al [13] has through changes to production term. The standard model
conducted numerical investigation of turbulence models for considers only the single turbulence length scale to
a superlaminar journal bearing by considering 14 different determine the eddy viscosity but in reality all scales of
turbulence models and concluded that the SST model with motion will contribute to the turbulent diffusion
low –re number yields the best results in comparison. Douvi characteristics of the flow.
C. Eleni et.al [14] has compared the various turbulence
models for the NACA0012 airfoil for a two dimensional 2.4. Realizable variant of K-ε Model
subsonic flow at various angles of attack and Reynolds’s The term realizable represents the nature of the model in
number 3 x 106 and highlighted that the areas transition satisfying certain mathematical constraints over the
model prediction and turbulence modelling requires further parameter Reynolds’s Stress which are in consistent with the
investigation as no model is found to be providing more physics of the turbulent model. It is an improvement over
accurate results at higher angle of attack. the standard K-ε Model and differs from the same in two
FernandoVillalpando et.al [15] assessed the flow simulation different ways. The developed. K-ε Model consists a new
of different turbulence models for the wind turbine airfoil of formulation for the turbulent viscosity and a new transport
NACA 63-415 model at various angles of attack by using the equation for the dissipation rate ɛ. For flows with an
one equation and two equation models available in involvement of rotation, boundary layers under strong
commercial packages. The authors have concluded that SA adverse pressure gradients, separation, and recirculation can
model is good in predictions of near maximal lift and the SST be solved with this model with an achievement of superior

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1050
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
performance. The model also demonstrates a superior 2.12. Reynolds stress (Stress Omega)
capability in capturing the mean flow around complex The stress transport model based upon omega equations and
structures involved in the analysis. LRR model which finds applications in modelling flows with
curved surfaces and swirling flows. The model resembles the
2.5. Standard K-Omega K-Omega model in terms of prediction capacity for a wide
The model with two equations which generally attempts to range of turbulent flows. The model can be selected from the
predict turbulence by two PDE for the two variables namely viscous model dialog box in ANSYS FLUENT.
turbulence kinetic energy (K) and Specific rate of dissipation
(ω). The model is a two equation turbulence model which 3. Modeling and Analysis
has a very high closeness with RANS (Reynold’s Averaged The present work considers the NACA0012 (National
Navier – Stokes) equations. Advisory Committee for Aeronautics) airfoil for conducting
the analysis and the 3D model of the airfoil is created using
2.6. SST K-omega CATIAV5R20 software with a total of 164 coordinate points.
The Shear Stress Transport (SST) in combination with K- The airfoil has a chord length of 100mm with a thickness of
omega is a two equation eddy viscosity model which has 10mm.The airfoil is modelled using simple 2D commands
gained more popularity due to its flexibility in solving the such as spline, mirror, connect curve and join in the sketcher
turbulent flow problems involving low Reynolds’s number. module and pad command is used for converting the 2D
The model is recommended by many users for its sketch in to 3D model of the airfoil section. The Figure No 3.1
meritorious good behaviour in the separating flow and shows the 3D model of the NACA0012 airfoil.
adverse pressure gradients. The model has a tendency to
produce a too large turbulence levels in regions with strong
acceleration which is actually less pronounced with normal
K-ε Model.

2.7. BSL K-omega

The Baseline (BSL) model with Bradshaw’s assumption with
two equations designed to provide results similar to that of Fig No 3.1 NACA0012 Airfoil 3D Model
real K-Omega model of Wilcox and it has an identity of
around 50% of the boundary layer with gradual changes. The 3D model of the airfoil created using CATIAV5R20
The results obtained by this model is similar in nature with software is saved in the .igs format and imported to ANSYS
respect to K-omega model with som exceptions. analysis package for further analysis of turbulent models.
The airfoil is considered to behave like a cantilever
2.8. Transition K-KL omega beam.The geometry section and airfoil is meshed in the
A newly developed three equation model which actually ANSYS workbench with a maximum of 3, 00,000 elements
applicable for incompressible flows with a low Reynold’s and 54,548 nodes connecting the elements by adopting fine
number. It actually applied towards the transition of a flow quality mesh which is generally considered to achieve
from laminar to turbulence. The model has equations for results with higher accuracy. The meshed geometry is
laminar kinetic energy, turbulent kinetic energy and specific further exported to FLUENT module for conducting the
dissipation rate. The model can be simulated using analysis. The inlet flow is created and made to flow over the
OPENFOAM software. airfoil section and the output velocity of all the models are
tabulated to comparison.
2.9. Transition SST
The model is primarily applicable to only wall bounded flows
which is not a Galilean invariant and it cannot be applied for
transition in free shear flows. The model is not suitable for
wall jet flows. The model is designed for flows with a defined
nonzero freestream velocity.

2.10. Reynolds stress (Linear Pressure Strain)

The Reynold’s Stress model which considers the pressure
strain term in three different categories namely slow
pressure –strain term, rapid pressure – strain term and wall
reflection term with equations by including the pressure
strain term as default one. The same model when applied to
near-wall flows described with two layer model requires the Fig No 3.1 Meshed View of the Enclosure
modification of pressure strain term.

2.11. Reynolds stress (Quadratic Pressure Strain)

The model is considered as an optional pressure – strain
model proposed by Speziale, Sarkar, and Gatski. The model
provides high performance in a range of basic shear flows,
plane strain, rotating plane shear and axisymmetric
expansion / contraction. The model provides enhanced
accuracy for flows with more complexity in particular with
streamline curvature. Fig No 3.2 Meshed View of NACA0012 Airfoil

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1051
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
The Table No 3.1 shows the NACA0012 Airfoil Dimensions and Inlet Flow Properties considered in the present study.

Table No 3.1 NACA0012 Airfoil Dimensions and Inlet Flow Properties

S. No Parameters / Characteristics Symbol S.I Unit Value
1 Length l mm 100
2 Thickness t mm 10
3 Reynold’s Number Re NA 507596.685
4 Pressure P Pascal 101325
5 Velocity v m/s 75
6 Turbulent Intensity I % 3.06
7 Turbulent Kinetic Energy k m2/s2 8.09233848
8 Turbulent Length Scale L m 0.007
9 Turbulent Dissipation Rate ߝ m2/s2 5.40374578×102
10 Specific rate of dissipation ߱ 1/s 6004.16188
11 Turbulent viscosity ratio ߤ‫ ݐ‬/ ߤ NA 738.164555

4. Results and Discussion

The flow simulations obtained for the various turbulence models have been noted for its value towards the output
characteristic velocity of the flow after flowing over the NACA0012 airfoil and the values are tabulated in terms of both velocity
contour and velocity vector. The one equation Spalart Allmaaras model is considered to be the standard model for comparison
with other 11 different turbulence models considered. The error coefficient value obtained with respect to the standard model
for other turbulence models and the number of iterations occurred for the individual turbulence model and time coefficient is
also compared. The Table No 4.2 shows the results obtained for various turbulence models such as error coefficient, number of
iterations and time coefficient through flow analysis using ANSYS FLUENT module.

Table No 4.1 Results of Various Turbulence Models for Velocity in terms of Contour and Vector
Velocity (m/s)
S. No Turbulence Models
Contour Vector
1 Spalart-Allmaras 8.881×10 8.971×101
1

2 Standard K-epsilon 8.896×101 8.986×101

3 RNG K-epsilon 8.923×101 9.013×101
4 Realizable variant of K-epsilon 8.911×101 9.002×101
5 Standard K-Omega 8.923×101 9.014×101
6 SST K-Omega 8.923×101 9.013×101
7 BSL K-Omega 8.897×101 8.987×101
8 Transition K-KL mega 8.980×101 9.071×101
9 Transition SST 8.910×101 9.000×101
10 Reynolds Stress (Linear Pressure Strain) 8.937×101 9.027×101
11 Reynolds Stress (Quadratic Pressure Strain) 8.932×101 9.023×101
12 Reynolds stress (Stress Omega) 8.900×101 8.990×101

The tabulated output velocity in terms of contour and vector are further used for calculating the error percentage that exists
between the velocity value of the standard turbulence model and other models considered for study. A total of 24 simulations
has been generated through FLUENT solver for both the velocity contour and vector of all the turbulence models considered.
The outlet flow velocity has been calculated for all the turbulence models considered and the values of velocity has been
recorded in terms of both contour and vector form and tabulated for further analysis and comparison. From the simulation
results the outlet velocity obtained for Spalart Allmaras model is found to be 8.881×101 m/s in contour form and 8.971×101
m/s in vector form. The outlet values velocity obtained of other turbulence model is found to the higher in case of both the
contour and vector forms for all the turbulence models. The lowest outlet velocity values in terms of both contour and vector
form is obtained for Standard K –epsilon model (8.896×101 m/s and 8.986×101 m/s) and the highest outlet velocity values in
terms of both contour and vector form is obtained for the Transition K-KL omega model with 8.980×101 m/s and 9.071×101
m/s respectively. The outlet velocity values obtained for RNG K-epsilon, Standard K – Omega and SST K-omega in contour form
is found to be similar (8.923×101 m/s). The simulation of velocity in contour and vector form for the standard Spalart –
Allmaras model has been shown in Fig No 4.1 and 4.2. The sample of few CFD simulations of turbulence models with less error
percentage, less number of iterations for completing the analysis and less time coefficient are shown in Fig No 4.3, 4.4, 4.5, 4.6,
4.7, 4.8 for reference.

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1052
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470

Fig No 4.3 Contour View of Spalart – Allmaras Model

Fig No 4.4 Vector View of Spalart – Allmaras Model

Fig No 4.5 Contour View of Standard K-epsilon Model

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1053
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470

Fig No 4.6 Vector View of Standard K-epsilon Model

Fig No 4.7 Contour View of Standard K-omega model

Fig No 4.8 Vector View of Standard K-omega model

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1054
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
Table No 4.2 Results of Various Turbulence Models for Error Coefficient, Number of Iterations and Time
Coefficient
Error Percentage Number of
S. No Turbulence Models Time Coefficient (s)
Contour Vector Iterations
1 Spalart-Allmaras 0 0 30 59.4
2 Standard K-epsilon 0.1686 0.1669 43 128.87
3 RNG K-epsilon 0.4729 0.4659 43 85.91
4 Realizable variant of K-epsilon 0.3366 0.3455 43 85.91
5 Standard K-Omega 0.4729 0.4793 39 77.92
6 SST K-Omega 0.4729 0.4659 39 77.92
7 BSL K-Omega 0.1801 0.1783 39 119.1
8 Transition K-KL Omega 1.1024 1.114 94 281.71
9 Transition SST 0.3265 0.3232 48 143.85
10 Reynolds Stress (Linear Pressure Strain) 0.6305 0.6240 121 487.51
11 Reynolds Stress (Quadratic Pressure Strain) 0.5742 0.5796 236 707.29
12 Reynolds stress (Stress Omega) 0.2139 0.2117 221 883.11

Conclusion 9. Next to the top ranking models in low time coefficient

The Following points may be used for concluding the present the models such as RNG K-epsilon, Realizable variant of
work conducted through ANSYS simulation software. k-epsilon have shown 85.91 seconds and it is found to
1. The most adopted Spalart-Allmaras turbulence model in be 45.61 % more time than the standard model.
case of aerodynamics flow analysis has been taken as 10. The other considered models such as BSL K-omega,
the standard reference turbulence model for Standard K-epsilon and Transition SST have shown time
comparison with the other 11 different turbulence coefficient values of 119.106 seconds, 128.871 seconds
models considered for studying the flow analysis of and 143.856 seconds respectively which are found to be
NACA0012 airfoil more time consuming than standard models in the order
2. In the estimation of error percentage, Standard K – of 102 %, 118.42 % and 144 % respectively.
epsilon model is found to have less error percentage in 11. The maximum time coefficient was obtained for
comparison to other turbulence models. The turbulence Reynolds stress (Stress omega) model with 883.11
models such as Transition K-KL omega, Reynolds stress seconds followed by Reynolds stress (Quadratic
(Linear pressure strain) and Reynolds stress (Quadratic pressure strain) with 707.29 seconds and Reynolds
pressure strain) have shown higher error percentage. stress (Linear pressure strain) with 487.51 seconds
3. The number of iterations occurred for the turbulence when compared to the standard model with 59.4
model has been considered as one of the output seconds.
characteristic and compared with each other. The 12. In comparison with standard turbulence model the
models such as Standard K-omega, SST K-omega and maximum time coefficient obtained for models such as
BSL K-omega has used less number of iterations in Reynolds stress (Stress omega), Reynolds stress
comparison with other turbulence models. (Quadratic pressure strain) and Reynolds stress (Linear
4. The maximum iterations were shown by Reynolds stress pressure strain) are found to be more time consuming in
(Quadratic pressure strain) model with 236 iterations the order of 1396 %, 1098% and 726.28% respectively.
followed by Reynolds stress (Stress omega) with 221 13. The more time consuming models such as Reynolds
iterations and Reynolds stress (Linear pressure strain) stress (Quadratic pressure strain) and Reynolds stress
with 121 iterations when compared to the standard (Linear pressure strain) are taking more number of
turbulence model with 30 iterations. iterations with higher time and error percentage may be
5. The number of iterations taken by the Reynolds stress avoided in conducting turbulence model analysis. The
(Quadratic pressure strain) model, Reynolds stress Reynolds stress (Stress omega) model is also showing
(Stress omega) and Reynolds stress (Linear pressure higher time coefficient and iterations but in case of error
strain) are found to be taking 8 times, 7.5 times and 4 percentage it is found to be less than other Reynolds
times more than the standard model. stress models.
6. The models such as Standard K-omega, SST K-omega 14. From the current study of flow analysis over an airfoil at
and BSL K-omega has shown only 39 number of subsonic conditions Spalart Allmaras is found to be the
iterations which is slightly higher than the Spalart better one in providing accurate results than other
Allmaras model with 30 iterations. turbulence models.
7. The other models such as Standard K-epsilon, RNG K-
epsilon and Realizable variant of k-epsilon has shown 43 REFERENCES
iterations and found similar with each other. [1] Ram Kumar Raman, Yogesh Dewang Jitendra Rag, “A
8. In case of time coefficient calculation, the models such as review on applications of computational fluid dynamics
Standard K-omega and SST K-omega have shown less “, International Journal of LNCT, 2018, Vol 2(6).
time for conducting the analysis with 77.92 seconds and
[2] Nicolas Pellerin, Sébastien Leclaire, Marcelo Reggio,
ranks top when compared to the Spalart-Allmaras
“An implementation of the Spalart–Allmaras
turbulence model with 59 seconds. The time taken by
turbulence model in a multi-domain lattice Boltzmann
both the models are found to be 32 % higher than the
method for solving turbulent airfoil flows”, 2015,
standard model. The number of iterations for both the
Computers and Mathematics with Applications, Vol 70,
models are found to be same.
3001–3018.

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1055
 International Journal of Trend in Scientific Research and Development (IJTSRD) @ www.ijtsrd.com eISSN: 2456-6470
[3] Olubunmi Popoola, YidingCao, “ The influence of Internal Combustion Engines simulation, Energy
turbulence models on the accuracy of CFD analysis of a Procedia, 2018, Vol 248, 1098 – 1104.
reciprocating mechanism driven heat loop “. Case
[10] Feng Gao, Haidong Wanga, Hui Wang, “Comparison of
Studies in Thermal Engineering (2016), Vol 8, pp 277–
different turbulence models in simulating unsteady
290.
flow”, Procedia Engineering 205 (2017) 3970–3977.
[4] Manuel García Pérez , Esa Vakkilainen, “A comparison
[11] M. Ghafari, M. B. Ghofrani, “New Turbulence modeling
of turbulence models and two and three dimensional
for air/water Stratified flow“, Journal of Ocean
meshes for unsteady CFD ash deposition tools” ,2019,
Engineering and Science, 2020, Vol 5, pp 55- 67.
Fuel 237, pp 806-811.
[12] Jia-Wei Han, Wen-Ying Zhu, Zeng-Tao Ji, “Comparison
[5] A. Riccia,b,c,, I. Kalkmanb, B. Blockena,, M. Burlandoc,
of veracity and application of different CFD turbulence
M.P. Repettoc, “Impact of turbulence models and
models for refrigerated transport“, Artificial
roughness height in 3D steady RANS simulations of
Intelligence in agriculture, 2019, Vol 3 pp 11 -17.
wind flow in an urban environment”, 2020, Building
and Environment, Vol 17. [13] AoshuangDing, XiaodongRen, XuesongLi ,ChunweiGu,
“Numerical Investigation of Turbulence Models for a
[6] Abdolrahim Rezaeiha, Hamid Montazeri, Bert Blocken,
Superlaminar Journal Bearing “, Advances in tribology,
“On the accuracy of turbulence models for CFD
2018 .
simulations of vertical axis wind turbines “, 2019,
Energy, Vol 180, pp 838 – 857. [14] Douvi C. Eleni, Tsavalos I. Athanasios Margaris P.
Dionissios, “Evaluation of the turbulence models for
[7] Tao Zhi, Cheng Zeyuan, Zhu Jianqin, Li Haiwang, “ Effect
the simulation of the flow over a National Advisory
of Turbulence models on predicting convective heat
Committee for Aeronautics (NACA) 0012 airfoil “,
transfer to hydrocarbon fuel at supercritical pressure “,
Journal of Mechanical Engineering Research, 2012, Vol
2016, Chinese Journal of Aeronautics, pp 1247 – 1261.
43, pp 100 -111.
[8] Mohamed M. Helal, Tamer M. Ahmed, Adel A. Banawan,
[15] FernandoVillalpando, Marcelo Reggio, AdrianIlinca,
Mohamed A. Kotb, “Numerical prediction of sheet
“Assessment of Turbulence Models for Flow Simulation
cavitation on marine propellers using CFD simulation
around a Wind Turbine Airfoil”, Modelling and
with transition sensitive turbulence model”, 2018,
Simulation in Engineering, 2011.
Alexandria Engineering Journal, Vol 57, 3805 – 3815.
[16] ANSYS Customer Training Manual, “Turbulence
[9] V. K. Krastev, G. Di Ilio, G. Falcucci G. Bella, “ Notes on
Modeling”, Introductiion to ANSYS FLUENT.
the hybrid URANS/LES turbulence modelling for

@ IJTSRD | Unique Paper ID – IJTSRD30824 | Volume – 4 | Issue – 3 | March-April 2020 Page 1056