Ahmed Body Challenge
Topic
Simulate the flow of Aerodynamics study of Ahmed body and grid dependency test over Ahmed Body using Ansys Fluent
A project
Submitted in partial fulfillment of the requirement for the award of the certificate
Of
Master Program
(CFD Masters)
By
Nitesh Negi
(Enrolment ID:3081)
Under the supervision of
CFD Team Leader Director of product
Pritam Goel Sarangarajan Iyengar
5th floor, Bristol IT Park
Main Rd, South Phase, Sathya Nagar, Gandhi Nagar, Chennai, Tamil Nadu 600032, +918939850851
CANDIDATE'S UNDERTAKING
I, Nitesh Negi (Enrolment Id:3081) state that the contents and findings of the project entitled “Simulate the flow of Aerodynamics study of Ahmed body and grid dependency test over Ahmed s Body using Ansys Fluent ”, are original. To the best of my knowledge, the current project does not infringe upon anyone’s copyright nor violate any proprietary rights and that any ideas, techniques, quotations or any other material from the work of other people included in my project, published or otherwise, are fully acknowledged in accordance with the standard referencing practices. Even though later on in case of any complaint pertaining of plagiarism, I am solely responsible for the same and understand that as per norms.
ACKNOWLEDGEMENT
Primarily, I acknowledge this moment of time which has gone just forever but made up my mind to write the word for those who have been with me with their over-whelming gestures throughout this assignment. With deep gratitude, I express my earnest thanks to my respected guides Mr. Pritam Goel, CFD Team Leader, and Mr. Sarangarajan Iyengar, Director of product skilllync for his constant involvement, energetic efforts, and proficient guidance. Without his wise counsel and encouragement, it would have been impossible to complete the project work in this manner.
I would also thank to all the members of Ansys Fluent Skill-Lync what are app group members for support and express my gratitude.
ABSTRACT
Simulate the flow of the Aerodynamic Ahmed body and grid dependency test over Ahmeds Body using Ansys Fluent. The drag value is very much dependent on the element size and the Y+ value of the inflation layer.
What we have observed that at 200 y+ value the coefficient of drag is very high and further reducing it to the 30 will get us the drag value close to 0.33. Further, we have found that the combination of the meshing that is using multizone meshing, using different face size, using sharing topology will effect to get the accurate results. Since we have done the simulation with very refined mesh also which does not provide accurate values.
The computational fluid dynamics (CFD) package FLUENT as a fluid flow analysis tool to solve the governing equations, prominently equations of continuity, motion, and energy. Numerical results can be compared with experimental results and which was done by research scholar since due to model setup and lab constraints we will only compare the simulation results of different cases as per our objective. The mesh size is selected to do that the numerical solution is taken considering the computation time and simulation processing time.
Keywords: Numerical simulation, Fluent, CFD, k-epsilon,
CONTENTS
Candidate’s undertaking i
Acknowledgment ii
Abstract v
Contents vi
List of figures vii
List of table’s viii
Nomenclature viii
Contents
CHAPTER 1: INTRODUCTION
-Introduction
-History
-Application
- What is a Ahmed Body?
CHAPTER 2: LITERATURE REVIEW
- Introduction
- Review of Literatures
- Objectives
CHAPTER 3: METHODOLOGY
-Flow chart for CFD simulation
-Procedure/Steps for the CFD simulation
CHAPTER 4: RESULT AND DISCUSSIONS
CHAPTER 5: CONCLUSIONS AND FUTURE SCOPE
REFERENCES
LIST OF FIGURES
Figure: Shows the details of the velocity, drag, wake region, pressure, eddy viscosity
LIST OF TABLES
Table: Shows the grid independency and variation of the simulation model at different elements size.
NOMENCLATURE
Symbols Significance
CHAPTER 1
INTRODUCTION
The three-dimensional flow around a vehicle has become a subject of significant importance in the automotive industry. One apparent technique of improving the fuel efficiency of a vehicle is to reduce the aerodynamic drag force by optimizing the body shape. Execution of fine aerodynamic design under stylistic constraints requires an immense understanding of the flow pattern phenomena and, especially, how the aerodynamics are impacted by changes in body shape. The flow area which presents the major contribution to the drag of a car in the wake flow backing the vehicle. The point at which the flow separates directs the size of the separation zone, and as a result the drag force. Clearly, a more accurate simulation of the wake flow
Why Study the Drag Coefficient in Cars?
The Drag Coefficient quantifies the resistance of an object in a fluid environment. It is not an absolute constant for a body’s shape because it varies with the speed and direction of flow, object shape and size, and the density and viscosity of the fluid. The lower the drag coefficient of an object, the less aerodynamic or hydrodynamic drag occurs. In terms of a car, the lower the drag coefficient, the more efficient the car is. As well as affecting the top speed of a vehicle, the drag coefficient also affects the handling. Cars with a low drag coefficient are sought after, but decreasing the drag drastically can reduce the downforce and lead to a loss in road traction and a higher chance of car accidents.
Most cars have the average value of drag between .30 and .35. Boxy cars have a higher number, like the HUMMER® H2 vehicle at .57, while more streamlined and agile cars have a lower number, like the Mercedes-Benz® C-Class® vehicle at .24. Again, this is just an average measurement. The exact drag coefficient of a car varies with the Reynolds number and various other factors.
There are certain ways to modify a car to optimize its aerodynamics and decrease the drag coefficient. To streamline the exterior, you can remove certain aftermarket items such as the roof rack, mud flaps, spoilers, and radio antenna. Professional car racers also take off their windshield wipers and side mirrors, but this isn’t recommended for the average driver! You can also add wheel covers, a partial grille block, an under tray, fender skirts, and a modified front bumper to improve the drag coefficient and make your car stand out in the process.
History
A lot of studies regarding airflow and turbulence behavior on a car involve the use of the ‘Amhed body’, which was first devised by S, Ahmed in 1984. The shape of the Ahmed body is designed in such a way to mimic the typical flow field features of a car bluff body, namely “a strong three-dimensional displacement flow in front, relatively uniform flow in the middle and a large structured wake at the rear” (S.R Ahmed, 1984) (rear flow features illustrated in figure 1, next page) whist being simple to model. It has had a significant impact on research in the automotive industry, used to study and validate flow over an automotive in both wind tunnel and CFD environments. In the case of CFD, its simple shape reduces the required computational power and time taken to model the body and run simulations. The geometry of the body, as originally described by S, Ahmed, can be seen in figure 2. All dimensions are fixed apart from the rear slant angle ‘π’, which can be set at different angles, typically from 0 to 30 degrees to study the difference this makes to the rear wake. Ahmed
The accurate modeling of turbulence has always been a considerable challenge for engineers throughout history. Nelkin, M (1992) states that turbulence is “One of the great unsolved problems of classical physics […] in the sense that a clear physical understanding of the observed phenomena does not exist.”. In simplest terms, turbulence can be defined as the erratic, fluctuating flow of a fluid whose Reynolds number has exceeded 4000. Turbulent flow is considered the natural state of most flows in both an engineering setting and in nature, with laminar flow being an exception to this, rather than the rule. This fluctuating nature of turbulence makes a deterministic approach very difficult, so most current methods of solving turbulence related problems and modeling tend to be more approximating in nature. (Ting, D, 2016)
Application
The use of computational fluid dynamics (referred to as ‘CFD’ for short) has grown considerably over the years, becoming an integral part of the design process components in the automotive industry. CFD is primarily used in the automotive industry on components exposed to outside flow such as a car body, with the aim to detect drag and turbulence forces which could affect a vehicle's fuel efficiency and drivability. On the matter, Ashton, N et al (2015) states that “Computational Fluid Dynamics has increasingly provided an important design tool for the automotive industry, often used as a supplement to experimental studies. [e.g.: wind tunnels]” The main advantage of using experimental studies such as a wind tunnel alongside CFD methods is a wind tunnel results are very likely true to life; as the experiment is undertaken in a real-life setting. The main issue with these experiments however are often very expensive to undertake, with bulky testing equipment and a lot of time is taken producing a scale model of the test piece and undertaking the experiment. In some cases, wind tunnel results inaccuracy can be affected by flow against the tunnel walls, particularly with smaller, cheaper tunnels. CFD in comparison is a much more cost and time-effective solution with less to no wall-flow interference, set up time, and better visualization of results, and because of this, it is becoming a much more popular method of testing. As computational power continues to improve, the accuracy of CFD studies is likely to become increasingly accurate and quicker to perform than wind tunnel tests. It is becoming less of a “supplement to experimental studies.”, like Ashton, N et al (2015) states and more relied upon as the primary testing method, reducing the necessary testing time and cost. As a result of this, the CFD market has seen considerable growth over recent years and the market is estimated to grow a further 13.6% in the 2015-2019 period according to analysts Cision US Inc., (2015). Despite this, one main disadvantage that has consistently remained for CFD however is that the chaotic nature of the turbulent flow. This chaotic nature has always made prediction mathematically complex, computer results may differ from reality, so cannot be completely relied on alone. Accurate modeling of turbulence is particularly important in the case of turbulent flow over an automotive, which can affect the drivability of the vehicle and fuel efficiency. To tackle the task of modeling turbulence, various methods have been devised, including the K-epsilon model, the K-omega model, and the Reynolds Stress Model of which this report aims to validate compared to wind tunnel values using a simplified car body known as the ‘Amhed body’, first devised by S, Ahmed in 1984, which has been used in various other CFD and wind tunnel studies, providing a great deal of data of which to compare with
What is an Ahmed Body?
The Ahmed Body was first created by S.R. Ahmed in his research “Some Salient Features of the Time-Averaged Ground Vehicle Wake” in 1984. Since then, it has become a benchmark for aerodynamic simulation tools. The simple geometrical shape has a length of 1.044 meters, height of 0.288 meters, and a width of 0.389 meters. It also has 0.5-meter cylindrical legs attached to the bottom of the body and the rear surface has a slant that falls off at 40 degrees.
The simple geometry of an Ahmed body.
Modeling Airflow Over an Ahmed Body
In the Airflow Over an Ahmed Body verification model, our Ahmed body has a 25-degree slant and is placed in the following domain, measuring 8.352-by-2.088-by-2.088 meters, to compute the flow field.
Computational domain and boundary conditions for the fluid flow simulation.
The front of the body is placed at a distance of 2 car lengths (2L) from the flow inlet. To reduce the computational cost, asymmetry plane is introduced to model half of the model.
The flow for this model is turbulent, which is based on the Reynolds number determined by the body length and inlet velocity. The simulation solves for the turbulent kinetic energy in addition to the velocity field. For this simulation, we need a larger mesh size than what is usually common to resolve the turbulent flow. More specifically, we use a finer mesh downstream of the model to capture the wake zone.
The flow for this model is turbulent, which is based on the Reynolds number determined by the body length and inlet velocity. The simulation solves for the turbulent kinetic energy and dissipation in addition to the velocity and pressure fields.
The Ahmed body was described originally by S. R. Ahmed in 1984 [1]. Three main features were seen in the wake:
- The A recirculation region that is formed as the flow separates at the top of the vertical back surface of the model
- The B recirculation region is formed due to the separation at the base of the model.
- The c-pillar vortices that form as the vorticity in the side boundary layers roll up over the slant edges.
The wake was shown to be highly dependent on the slant angle. For slant angles less than 12°, the flow remains attached over the slant. The flow is essentially two-dimensional and has low drag. Between 12° and 30°, the flow becomes much more three-dimensional as the c-pillar vortices form. These reach maximum strength at 30°. The drag increases significantly as the low-pressure cores act on the rear surfaces. Past 30° the flow separates fully off the slant. This results in a sudden decrease in drag and weaker c-pillar vortices.
Effect of Aspect Ratio: The aspect ratio of the rear slant had a significant effect on the wake. The wider bodies ceased to reattach at slant angles of 25°, suggesting that the critical angle lowers as aspect ratio increases. They also provide a vortex core location from experimental data that can be used for validation.
Description of the test case: Ahmed's body was first defined and its characteristics described in the experimental work of Ahmed. Two configurations with slant angles of 25° or 35°are considered as a test case. The distance between the body and the plate representing the ground is 50 mm. In the experiment performed by Ahmed, flow velocity was taken 25 m/s, the Reynolds number was based on the model length and use the Y+ values to get more accurate results using the K-epsilon model.
CHAPTER 2:
LITERATURE REVIEW
- INTRODUCTION
To begin the proposed work it is required to make a review of research papers in order to discover what work has been done already about the related topic and which procedure or methodology has been adopted to perform the simulation/experiments and conclude the results.
A review of the various published literature on any specific topic is the first important step to be taken to get,
- Knowledge about the state of art.
- Familiarity with the established methods and procedures for conducting work.
- To stop duplication of work.
- LITERATURE REVIEW
There are several numbers of studies conducted by researchers regarding the Steady-state Simulate the flow of Aerodynamics study of Ahmed body and grid dependency test over Ahmed Body using Ansys Fluent , some of them are discussed below:-
1-Flow and Turbulence Structures in the Wake of a Simplified Car
Model (Ahmed Model) H. Lienhart, C. Stoots, S. Becker Lehrstuhl für Strömungsmechanik (LSTM) Universität Erlangen-Nürnberg Cauerstr. 4, 91058 Erlangen, Germany
2- Experiments and numerical simulations on the
aerodynamics of the Ahmed body
- Meile1, G. Brenn1, A. Reppenhagen2, B. Lechner2, A. Fuchs2C
1Institute of Fluid Mechanics and Heat Transfer (ISW), Graz University of Technology,
Inffeldgasse 25/F, 8010 Graz, AUSTRIA
2The Virtual Vehicle Competence Center (ViF), Inffeldgasse 21a, 8010 Graz, AUSTRIA-OBJECTIVE
Simulate the flow over a cylinder and explain the phenomenon of Karman vortex street.
PART-I
Simulate the flow of Aerodynamics study of Ahmed body and grid dependency test over Ahmed s Body using Ansys Fluent
PART-II
Q1. Describe Ahmed's body and its importance.
Q2. Explain the reason for the negative pressure in the wake region.
Q3. Explain the significance of the point of separation.
Expected Results:
1. Velocity and pressure contours.
2. The drag coefficient plot for a refined case. ( For velocity of 25m/sec, the drag coefficient should be around 0.33).
3. The vector plot clearly showing the wake region.
4. Perform the grid independency test and provide the values of drag and lift with each case.
CHAPTER 3
-METHODOLOGY
The complete methodology that has been adopted to conduct the present project work is shown in the form of the flow chart in figure 3.1
Figure 3.1: Flow Chart for Methodology
-Procedure for CFD simulation of Mixing Tee
Selection of material
The material property is not necessary for this simulation since we are not considering the heat transfer through the wall i.e. is an adiabatic process, hence the energy term not utilized in the simulations.
Further, we will know that in this type of simulation we consider the wet volume and hat is extracted using the Space claim tool.
Cad Model Preparation
Step1: Open the workbench & double click open Flow Fluent
Step2: Double click on geometry to open space claim> open the file from saved location.
Step4: Go to prepare tab and select the volume generation tool and select the edges for complete volume generation from inlet to outlet.
MESHING:
The meshing has been done by applying the multizone meshing method to the outer enclosure
Type of meshing element: Outer enclosure - Hexa dominant mesh
Different element size for the body sizing has been provided to vary the number of elements
and to perform the grid independence test.
Mesh Element size: 100 mm(Enclosure)
Type of element: Tetrahedral
Number of Elements: 8000 to 1400000+
For the Reynolds number calculated with the inlet velocity and the body length of the Ahmed Body,
the target Y+ value was 30-300. Hence to obtain this target Y+ value, the height of the first cell from the wall was calculated to be .44 to 3 mm.
Inflation: Number of inflation layer - 5
First layer height -.44 to 3 mm
Growth rate - 1.2
Inner enclosure - Tetrahedral
Step5: Close space claim workbench and save files if required.
Preprocessing
Step6: Double click on Mesh from workbench>Generate Mesh by right-clicking the mesh icon and click generate mesh option and mesh will be generated; here you can change mesh size from element size option in the mesh tree.
Step7: Give a name to required parts by face section/edge selection>click update by right-clicking on the mesh>check the mesh metric matrix by select element quality should be not less than 5%.
Step8: Can cut the section using the section tool from the above toolbar. Then close mesh workbench which will automatically save the data generated.
Setup/Processing
Step9: Double click on setup to open the workbench
Step10: select for Time -steady-state solution, Solver Type -pressure-based solvers, Velocity Formulation - Absolute and Physics-energy from the model.>select the Viscous Model -the k-epsilon model for turbulent flow based upon the condition.
Step11: setup the boundary condition as per the given data.
STEP 3 : SET UP & SOLUTION
Fluent Setup
- Solver type: Pressure based
- Check the Energy equation
- Turbulent Model: K-Epsilon – standard (Range of Y value is 30 - 300)
- Fluid Material: Air
Boundary conditions
- Inlet velocity : 25 m/s
- Outlet pressure : 0 pa
- Symmetry: For sidewalls
- Wall: Bottom
NOTE:
In the density-based approach, the velocity field is obtained from the momentum equations Continuity equation is used to obtain the density field Pressure field is determined from the equation of state. IT is originally designed for high-speed compressible flow. As we know that velocity is 25 m/s. which is not so high, but we can expect that it can increase in some regions. So we have used a density-based solver.
Setup & Reference Values
- Type - Density-based.
- Velocity Formulation- Absolute.
- Time - steady
- Energy : Viscous model (turbulence model):k-epsilon(2 eqn)
- Inlet velocity - 25 m/s.
- Outlet : Gauge pressure - 0 pascal(1 atm)
- Hybrid Initialization
- Material: Fluid-air
- Density - 1.225 kg/ (constant)
- Viscosity - 1.7894e-05 kg/m-s(constant)
- Lower wall - Wall Boundary condition
- Side wall - Symmetry condition
- Inlet, Outlet as other B.Cs
Step12: The velocity of air> 25m/s>Pressure at the outlet of the geometry 0 gauge.
Step13: In solution select the report >initialize the solution>run for the calculation by entering the no.of iteration
Post-processing/Results
Step14: Open results workbench by double-clicking on the result option>Open the geometry by unchecking the named parameters>create XY plane to plot the contour of velocity and temperature>plot for temperature & velocity by drawn the line along & across the pipe
Step15: Open the workbench & double click open Flow Fluent
Step16: Double clicks on geometry to open space claim> open the file from saved location.
Step17: Go to prepare tab and select the volume generation tool and select the edges for complete volume generation from inlet to outlet.
Step18: Close space claim workbench and save files if required.
The above image shows the velocity profile in the vector form and also about the formation of the wake region.
CHAPTER 4:
RESULT AND DISCUSSIONS
The results of various Simulation tests conducted for Mixing TEE efficiency as well as are calculated and explained with the help of different graphs and charts which are plotted by using Ansys Fluent & MS-Excel software. Comparisons of results are also done in Case I and Case II. The compared results are summarized as the following:-
Case 1
β.Element size:415 mm
β.Body sizing:--mm
β.Face sizing:--mm
β.No of Element:84437
β.Number of Nodes:16633
Residual plot:
Cd plot for car wall:
Cl plot for car wall:
Velocity contour:
Pressure contour:
vector plot clearly showing the wake region
Case 2
β.Element size:100 mm
β.Body sizing :25mm
β.Face sizing :10mm
β.No of Element:1389751
β.Number of Nodes:353711
Residual plot:
Cd plot for car wall:
Cl plot for car wall:
Velocity contour:
Pressure contour:
Effect of the Eddy Viscosity
Turbulent Kinetic Energy plan
vector plot clearly showing the wake region
- Case II
To perform the grid dependency test
- T plot Velocity and pressure contours.
- The plot velocity vector plot clearly showing the wake region.
- To obtain coefficient Drag & Lift
Table
|
Sr no. |
Elements |
Y+ |
Y |
Drag Coeff. |
Lift Coeff. |
|
1 |
84437 |
- |
- |
0.0919 |
0.0623 |
|
2 |
866393 |
200 |
2.9 |
0.3762 |
0.2895 |
|
3 |
873248 |
80 |
1.17 |
0.3739 |
0.3111 |
|
4 |
1045444 |
30 |
.44 |
0.3545 |
0.3015 |
|
5 |
1389751 |
30 |
.44 |
0.3260 |
0.2902 |
|
6 |
1410189 |
30 |
.44 |
0.3232 |
0.2770 |
Case 1
β.Element size:100 mm
β.Body sizing:30mm
β.Face sizing:8mm
β.No of Element:1410189
β.Number of Nodes:405578
Residual plot:
Cd plot for car wall:
Cl plot for car wall:
Velocity contour:
Pressure contour:
Turbulent Kinetic Energy
vector plot clearly showing the wake region
CHAPTER 5
CONCLUSIONS AND FUTURE SCOPE
In this chapter, we will discuss the outcome from a different Simulation performed i.e. Case I, Case II, and the graphs, plots. There are some results that support the previous outcome of research work and finds that the drag value is a very important factor for the Vehicle movement. The drag and lift should be maintained so that they get a smooth ride. Based on observation, and analysis made in the study of Mixing Tee following conclusions are drawn:-
Factors on which accuracy of the predictions depends.
- Accurate modeling of the geometry.
- Geometry should be in good agreement with actual, that flow physics should not be disturbed.
- Mesh accuracy should be according to standards and refined enough to give a converged
- Residuals should be given according to applications and falling those below the given threshold gives an accurate solution if the overall modeling is correct.
- The selection of solution models and methodology also has a great impact on the final solution.
- And the Mesh sizing with Y+ value affects greatly in the accuracy of the model.
Further increasing in mesh elements can only increase in the calculation. We can say that solution is dependent on the value of y+.Y+is very important for near-wall treatment when we are studying or measuring flow at the boundary layer. Also finer the mesh, the wake region will be captured more fine and clear. So meshing is a major factor in the study of the flow over an Ahmed Body.
Negative Pressure
Negative pressure at the nose indicates that lift is produced.
Also, negative pressure can be due to high velocity. At high velocity, kinetic energy increases and so
the pressure will be decreased. So pressure at high velocity can be negative.
Also when the boundary layer separates, a space called wake generates at that area, and due do very
few mass at there, negative pressure can be generated.
From the above vector diagram of velocity magnitude, we can see near the portion of the Ahmed body,
flow area decreases, so velocity increases, and also kinetic energy increases.
So pressure will reduce along the direction of flow.
Now if x is the flow direction, then pressure gradient along the direction of flow, .
So due to the kinetic energy, it will overcome the surface friction and so the boundary layer will not
be destroyed.
For a lower number of grid points also, we will be able to obtain the required results, but the regions of
the wake is clearly visible when the meshing profile with a higher number of grid points is higher.
There is a variation in the pressure along the Ahmed body from the front to the rear side, one of the
reasons for the negative pressure at the rear is due to the separation of flow at the rear and the
other reasons are as follows due to the recirculation of the air.
From the plot for the coefficient of drag, it is clearly visible that the value of Cd becomes almost the
same at a particular number of grid. Thus, instead of increasing the number of the grid, which in turn
increases the computational cost, we can use a lesser number of the grid which will give the same result. And vary the Y+ value.
Negative pressure in the wake region
Within the wake of the Ahmed body, due to high turbulence intensities, the mean velocities
approach zero. A pair of vortices are developed in this region very strongly. This counter-rotating
trailing vortices are responsible for the attached flow.
When the flow separates at the lower edge of the slant, strong underpressures are created. This
results in an increase in drag.
Significance of point of separation
A flow separates as soon as it detaches from a surface into awake. It occurs when the flow slows
and there is an increase in the pressure after passing a bluff or streamlined body.
The flow becomes detached when the speed of the boundary layer relative to the surface is zero
and reverses its direction. This results in the formation of eddies or vortices. This causes an
increase in pressure drag and a reduced lift.
FUTURE SCOPE
Can perform the simulation for the new upcoming models of cars even the moving machinery.
REFERENCES
Ahmed, S., Ramm, G. and Faltin, G. (1984). Some Salient Features of the Time Averaged Ground Vehicle Wake. SAE Technical Paper Series.
Andersson, B. (2012). Computational fluid dynamics for engineers. Cambridge: Cambridge University Press.
Bates, P., Lane, S. and Ferguson, R. (2005). Computational fluid dynamics: Applications in Environmental Hydraulics. Wiley, pp.54-56.
Belo-Millan, F., Makela, T., Parras, L., del Pino, C. and Ferrera, C. (2016). Experimental study on Ahmed's body drag coefficient for different yaw angles. Journal of Wind Engineering and Industrial Aerodynamics, Volume 157, October 2016, Pages 140-144
Caridi, D., Cokljat, D., Schuetze, J. and Lechner, R. (2012). Embedded Large Eddy Simulation of Flow around the Ahmed Body. SAE International Journal of Commercial Vehicles, 5(1), pp.208-214.
Comsol Inc. (2018). Finite Element Mesh Refinement Definition and Techniques. [online] Comsol.com. Available at: https://www.comsol.com/multiphysics/mesh-refinement
Corallo, M , J. Sheridan, M.C. Thompson, Effect of aspect ratio on the near-wake flow structure of an Ahmed body, Journal of Wind Engineering and Industrial Aerodynamics, Volume 147, 2015, Pages 95-103, ISSN 0167-6105, https://doi.org/10.1016/j.jweia.2015.09.006.
Frei, W. (2017). Which Turbulence Model Should I Choose for My CFD Application?
[online] COMSOL Multiphysics©. Available at:
https://www.comsol.com/blogs/which-turbulence-model-should-choose-cfdapplication/ [Accessed 8 Nov. 2017].
Gilliéron, P. and Chometon, F. (1999). Modelling of stationary three-dimensional separated air