Proceedings of the ASME 2018

International Mechanical Engineering Congress and Exposition

IMECE2018
November 9-15, 2018, Pittsburgh, PA, USA

IMECE2018-88027

AERODYNAMICS ANALYSIS FOR OPTIMIZATION THE DESIGN OF A BAJA SAE

CHASSIS

Jessica Gissella Maradey Lázaro Sergio Andrés Ardila Gómez

Programa de Ingeniería Mecatrónica Programa de Ingeniería Mecatrónica
Universidad Autónoma de Bucaramanga Universidad Autónoma de Bucaramanga
Bucaramanga, Santander, Colombia Bucaramanga, Santander, Colombia
jmaradey@unab.edu.co Sardila180@unab.edu.co

Helio Sneyder Esteban Villegas

Programa de Ingeniería Mecatrónica
Universidad Autónoma de Bucaramanga
Bucaramanga, Santander, Colombia
hesteban@unab.edu.co

ABSTRACT INTRODUCTION
The aerodynamic analysis of vehicles using Computational The Baja SAE competition is a design event for engineering
Fluid Dynamics (CFD) is a powerful tool that is used to explore students organized by SAE International, which promotes
and investigate how air flow behaves around the vehicle, gives a engineering excellence through the design and construction of a
qualitative idea of aerodynamic behavior and provides engineers safe, reliable and economical vehicle that complies with the
with the information necessary to design the pieces that will be established regulations and generates smart ideas for the
tested in the wind tunnel. This technique allows to find the values resolution of problems, through the guidance and support of a
of forces and moments to get an idea of the drag resistance and professor advisor [1]. In general, 100 or more universities
lift forces on the vehicle. It is also possible to determine the participate in each of the events sponsored by SAE, especially in
position of the center of pressure that exerts a vital importance the BAJA SAE category. The regulation emphasizes in the
in the definition of the directional stability of the vehicle in its design of the chassis, the configuration of the vehicle, the
lateral dynamics. On the other hand, it offers the advantage of wheelbase, track width, distance to the ground, wheels, tires,
being able to see many variables of the problem that are very stability of the rollover and in the safety of the vehicle among
difficult to access in reality, for example to see current lines, others such as belts, seats, switches, fast and untimely cutting of
vortex shedding, fields of pressures around the vehicle. Although the current (i.e. battery). Also within the competition are various
the use of CFD has become a common practice in the automotive static tests (i.e. design, cost analysis and presentation) and
industry in recent decades, it is still not considered as a single dynamics such as: acceleration and braking test, strength test,
technique, since one of the limitations is related to the dynamics maneuverability test (including one in water), drag test on rocks,
of a turbulent flow is quite complex so which analytical and and durability test; with which it is rated according to the best
computational study depends on what is called Turbulence performing vehicle [2] [3] [4]. In the case of baja SAE chassis it
Models. is necessary to comply with the principles of aerodynamics to
avoid instability in the vehicle, to protect the pilot, to have
This article aims to show the aerodynamic design process by greater speed and to reduce wind resistance and consumption as
CFD simulation of a Baja SAE vehicle chassis using Ansys much as possible.
software, widely used for simulation of static charges,
thermodynamics, fluid mechanics, multiphysics systems and The simulation is one of the tools of greater application and
vibrations and to provide information that can be validated use in the industrial sector in general, especially in the
experimentally in the future. automotive industry, since it allows to predict the behavior of

1 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 different mechanical components that are subjected to static and with the aerodynamic forces, this in turn with a correct
dynamic loads, as well as to verify the resistance of the materials understanding of the turbulent flow dynamics around them [2].
used, the impact and their modes of vibration. The benefits of Other aspects can include safety, regulations, and aesthetics [8].
the simulation are: considerable reduction of costs in the Fig.1 illustrated all the aspects to take into account when
construction of prototypes, minimization of the time of iterative conducting aerodynamic studies.
processes (test-error), increase in the possibility of optimization,
identification of the segment that tends to fail, characterization
of a situation real emergency, improvements to the
manufacturing design process, achieving a high performance
product.

On the other hand, the simulation is a very important step in

the process of designing a mechanical structure like the chassis
of a vehicle, since it provides an acceptable percentage of error
compared to reality. Also, in many situations, a wind tunnel is
not available to perform experimental validation of the
aerodynamics of the vehicle, and road tests presuppose a very
high cost, a good strategy to limit, approach reality and analyze Fig. 1: Relevant aspects in aerodynamic studies [8].
the phenomena related to it is the Computational Fluid Dynamics
(CFD) methodology which offers a viable and economical The aerodynamic resistance of vehicle is considered to be
alternative to delve into these complex aspects of engineering. a crucial part of automotive design. The aerodynamic drag can
be as high as 65% of the total drag for most passenger vehicles
The dynamics of computational fluids (CFD) is the branch at typical speed limit of 100Km/h. Hence, reducing drag
of fluid mechanics that uses numerical methods and algorithms contributes significantly to the fuel economy of a car [12].
to solve and analyze problems involving external or internal
flows where the medium can be a gas or a fluid [6].Some The aerodynamic drag coefficient is defined as follows[12]:
advantages of CFD are not only the large amount of information 2
provided but also the simulation results can be visualized and can 𝐶𝑑 = 𝐹𝑑 /(0.5𝜌𝑈∞ 𝐴𝑦 ) (1)
be expressed in terms of speed and pressure, additionally It can
include vortices, which are of great importance for the design In this equation, 𝜌 represents the air density, 𝑈∞ is the freestream
and manufacture of a single-seater since it is a tool to enhance velocity, 𝐴𝑦 is the model frontal area, and 𝐹𝑑 is the total drag
and direct the flow around the vehicle, and these analyzes greatly force acting on the car projected on the longitudinal direction.
favor the process of design, manufacturing and performance of
the vehicle . Likewise, you do not need to build a physical model The aerodynamic lift coefficient is defined as follows[13]:
because with the 3D geometry of the object to study, you can
2
start the methodology. 𝐶𝐿 = 𝐹𝐿 /(0.5𝜌𝑈∞ 𝐴𝑦 ) (2)

CFD provides several more approaches to shape In this equation, and 𝐹𝐿 is the total lift force.
optimization, which can reveal areas for aerodynamic
improvement that would not be exposed by neither intuition nor The lateral force coefficient is defined as follows[13]:
experience [8].
2
𝐶𝑌 = 𝑌/(0.5𝜌𝑈∞ 𝐴𝑦 ) (2)
The aerodynamic analysis of vehicles by means of CFD tries
to find the values of forces and moments to have an idea of the In this equation, and Y is the lateral force acting on the car.
forces of resistance to the advance (i.e. drag) and of lift (i.e. lift)
on the vehicle. It is also possible to determine the position of the Pressure coefficient of the flow on vehicle surfaces is
center of pressure that exerts a vital importance in the definition defined as follows [12]:
of the directional stability of the vehicle in its lateral dynamics.
On the other hand, computational analysis offers the advantage 2
𝐶𝑝 = (𝑃𝑥 − 𝑃∞ )/(0.5𝜌𝑈∞ ) (3)
of being able to see many variables of the problem that are very
difficult to access in reality, for example to see current lines, Where 𝑃∞ is the far field static pressure, and 𝑃𝑥 is the local
vortex shedding, fields of pressure around the vehicle[7]. surface pressure.
In general terms, given the importance that currently has an Reynolds Number is defined as follow:
efficient management of energy, an alternative for an 𝑅𝑒 = 𝜌𝜈𝐿𝑐 /𝜇 (4)
improvement in energy consumption in land vehicles has to do

2 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Where 𝜈the velocity of fluid (m/s) is, 𝜇 is the absolut - Animations and export of results to other formats.
viscosity (Kg/ms), and 𝐿𝑐 is the characteristic length (m).
The mesh can be structured or no structured. Also, there are
Reynolds number is dimensionless. If 𝑅𝑒 <1 the viscous the hybrids. The structured mesh only use quadrilateral and
forces are dominant and if 𝑅𝑒 >1 the inertial forces are dominant. hexahedrons elements. The unstructured mesh is characterized
for irregular connectivity and use any element. The hybrid mesh
The critical value of the Reynolds number determines the has structured and structured zones. Tab.1 shows the advantages
threshold between laminar and turbulent behavior in the and disadvantages of structured and unstructured meshes.
movement of a fluid. If 𝑅𝑒 > 𝑅𝑒𝑐 then there is turbulent
behavior, and if 𝑅𝑒 < 𝑅𝑒𝑐 then there is laminar behavior. Structured Unstructured
According to this, there will be laminar flow when Semiautomatic generation Automatic generation
𝑅𝑒 < 400.000 and turbulent flow if 𝑅𝑒 > 500.000. For simple geometries For complex geometries
Alignment in contours Adaptability in contours
The Navier stokes equations are a set of linked, nonlinear More precision Less precision
partial differential equations which describes flow of Newtonian
Less memory ( RAM) More memory ( RAM)
fluids. Assuming that the flow can be considered incompressible
Less calculation time More calculation time
they take the general form [13]:
Tab. 1: Advantages and disadvantages of structured and unstructured
meshes [6].
∇∙u =0 (5)
𝜕𝑢
ρ( 𝜕𝑡 + u ∙ ∇u) = −∇p − μ∇2 𝑢 + 𝜌𝑔 (6) The optimization process of CFD simulation model used in this
project is shown in fig.3.
Where u is the flow velocity field, ∇p is the pressure gradient
and μ is the dynamic viscosity of the flowing fluid. Given Obtaining CAD
Simplification
Generation of the
sufficient initial and boundary conditions these equations fully model of the vehicle computational
of the model
with driver domain
describe the fluid motion, even if the flow is turbulent. However,
it is often not possible to solve them analytically for other than
very simple cases. This is the reason why CFD has to be used to
Obtaining a
obtain numerical approximations of the solutions [13]. set of
Reduction of the
Implementation of the computational
unstructured
computational case domain in half
CFD METHODOLOGY and hybrid
(symmetry)
meshes
The process to simulate with CFD methodology can be seen
in fig. 2.
Calculations Obtaining results
Pre-Processing Solution Post- Processing
Definition of Calculations Results Analysis
mesh
Fig 3. Flowchart of the optimization process of the CFD simulation
model.
Adjustments
The design and modeling was carried out in compliance
Fig. 2: CFD process.
with the requirements of the SAE event, specifically the Baja
SAE category. In order to obtain a better approximation in the
The stage of pre-processing consists to define the geometry calculations, a non-uniform body with the average shape of the
to modeling (i.e. computational domain), mesh generation (i.e
vehicle driver was introduced into the vehicle. Next, we
division of the domain in an enough number of cells that do not proceeded to simplify details of the geometry, in order to avoid
overlap and that cover all the geometry, identification of physical
sharp angles, very narrow areas, complex geometries, abrupt
phenomena, definition of fluid properties and specification of the changes of orientation of surfaces, among others; in order to
inicial and contour conditions of a case study. The solution is a
avoid errors when generating the mesh. As for the computational
key stage of resolution program and to solve through of iterations domain, as evidenced in the literature, it is possible to reduce this
the equations that have been activated in the pre-process (i.e. the
domain by half and thus reduce computational costs and
models). The post-processing has a graphic tools that allow to simulation times.
analyze the results including [9]:
For the optimization of the mesh, it was decided to make a
- Graphic representation of domain and mesh.
set of unstructured meshes formed by tetrahedra in its entirety,
- Contour maps about variables and plotted vector and current
making a refinement in the proximity to curved surfaces (vehicle
lines. walls). Mesh independence is obtained from this mesh set,
- Graphics and distributions.
generating a stabilization of the drag coefficient of the vehicle

3 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 with the lowest possible number of elements. Then, starting from hybrid models that include RANS and LES, achieving a better
this mesh, the necessary adjustments are made for the balance between the computational cost and benefits of each
implementation of a hybrid mesh formed by layers of prisms existing model.
(structured) over the entire surface of the vehicle and tetrahedra
(unstructured) in the rest of the computational domain. The Several works also taking the model of Ahmed [10], which
prisms are created on the surface of the vehicle to be able to more is a simplified model, with a simple geometric configuration,
accurately capture the viscous and pressure phenomena free of attachments or accessories, with smooth surfaces and the
produced in the boundary layer of the same vehicle. aerodynamic analysis can be broken down into two parts, one in
Finally, the computational case is established, as explained in the the determination of the pressure on the skin, to calculate the
following sections; the calculations or iterations of the equations drag force, lift and friction, taking into account the center of
are made and the results are obtained, for their subsequent post- pressure. The other analysis corresponds to the study of the
processing. characteristics and behaviors of the flow trail. Although the first
analysis is more representative from a design point of view, the
MODELS OF SIMULATION latter provides information on where to focus efforts to reduce
coefficients and improve overall efficiency.
The use of CFD has become a common practice in the
automotive industry (vehicle design) in recent decades, which You can see the excellent results of the simulation vs the
has allowed the increase in the understanding of the flow results of the wind tunnel, when comparing the drag and lift
dynamics around land vehicles and thus improving the coefficients, as can be seen in tab.2
prediction of forces and aerodynamic moments. Although the
CFD has certain limitations to be considered as a unique A detailed knowledge of the physical mechanisms involved,
technique for the determination of aerodynamic forces in land such as dissipative structures, diffusion, rubbing, viscosity,
vehicles, these limitations are related to other aspects with which vortex formation, fluctuations and instability, are also essential
the dynamics of a turbulent flow is quite complex, so the for the development of numerical codes [13].
analytical and computational study it depends on what is called
Turbulence Models. Vehicle-A Modified ∆𝑪𝑫 , ∆𝑪𝑳
Vehicle-A
In general, these models are simplified versions of the 𝑪𝑫 Wind tunnel 0.32 0.30 -0.02
reality of the flow and therefore their prediction capacity is Numerical 0.323 0.301 -0.022
limited. However, it can be stated that given the impossibility of Simulation
directly solving (using DNS - Direct Numerical Simulation) the 𝑪𝑳 Wind tunnel 0.20 0.16 -0.04
Navier-Stokes equations in problems of high Reynolds number, Numerical 0.207 0.142 -0.065
then different models have been developed to approximate this Simulation
problem. One of the most used in external flow simulations is the Tab. 2: Comparison of results between the one obtained by wind
Reynolds' Average model of the Navier-Stokes equations (i.e. tunnel vs simulation [14].
RANS). The RANS model comes from the idea that the flow can
be decomposed into two: an average solution and a fluctuating SIMULATION SETUP
solution (Reynolds decomposition). The solution then requires
average and fluctuating values for velocity and pressure. General Simulation Settings (Materials and General Model)
Although the RANS method involves a low computational cost,
it is not able to determine the fluctuating components of speed A Baja SAE type vehicle was modeled using SolidWorks
and pressure with accuracy, since it filters both spatially and CAD modeling software. The geometry was simplified by
temporally most of the turbulent flow dynamics. On the other questions of computational cost reduction and following the
hand, the behavior of the various vortices present in the separated CFD simulation guidelines, where it is always specified to look
flow can not be precisely determined since RANS models (does for a simplified model with the main characteristics of the object,
not solve) all existing vortices in such a way that their behavior in this case the vehicle. It should be clarified that the lower part
is relatively standard [5]. of the vehicle is completely flat and the engine compartment was
not included, as well as a simplified geometry of the driver was
There is another solution method called Large Vortex made. Tab. 3 shows some of the main and important dimensions
Simulation (LES for its acronym in English). This method is of the Baja SAE model used for the present study.
more computationally expensive than RANS but manages to The computational domain consists of a box, as shown in
capture in a more precise way the behavior of the larger vortices fig. 4, from this box the volume occupied by the vehicle is
found in the turbulent separation zones of the flow. LES uses a subtracted. In order to simplify the model and save on
spatial filter to separate the large vortices from the small vortices. computational cost, it takes advantage of the symmetry of the
The large vortices are calculated while the more universal flow around the vehicle [15], therefore only half of the vehicle is
behavior of the small vortices is modeled [5]. We have also used included in the computational domain. Based on the information

4 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 found in the literature [16], the dimensions used in the present mesh. It was possible to obtain an independence of the CD result
study for the computational domain are those shown in tab. 3. with a mesh size (total number of elements) of approximately
2.7x106 .
Dimension Value
Height 1268.21 (mm) After obtaining the best performance mesh, it was decided
Width without wheels 951 (mm) to return to the Cad model of the vehicle, perform a small
Baja SAE optimization of a pointed surface located on the driver's model,
Length 1883.71 (mm)
which was avoiding to generate correctly the layer of prisms on
Front area 0.8422 (m2)
the vehicle. Again, the mesh was generated with the same
Model incident flow 1 x vehicle length parameters of the mesh of 2.7x106 elements, but this time
Model - outflow 3 x vehicle length generating 5 layers of prisms on the entire surface of the vehicle,
Computational
Total width 6 x vehicle width with an initial layer height of 1.5 mm and a height growth radius
domain
Model – roof between 1.1 layers; obtaining in this way a hybrid mesh
3 x vehicle height
computational domain optimized for this specific case, with a smaller number of
Tab. 3: Vehicle and computational domain dimensions elements of 2.0x106 elements. The hybrid mesh obtained is
detailed in fig. 5.

Fig. 5: Hybrid mesh of the computational domain.

Fig. 6 and Fig. 7 shows some details of the 2.0x106 mesh,

which, according to the convergence analysis, better predicts the
drag coefficient with an adequate computational cost.

Fig. 4: Scheme of the computational domain (dimensions in m).

Meshing
An unstructured mesh was generated in the commercial
software ICEM v18.1. This mesh is made up of tetrahedral
throughout its domain, refining the parts closest to the surface of
the Baja SAE vehicle. The maximum average size of the
elements was configured with 0.4 m, the minimum average size
of the elements was set to 1 mm. The mesh method was based
on the Octree Algorithm, activating the refinement parameter in
proximity of surface and curvature. The angles that the different
surfaces of the vehicle possessed did not allow the use of a hybrid
mesh, generating prisms on the surfaces of the vehicle and
allowing a greater refinement of the mesh.
Fig. 6: Detail view of hybrid mesh (symmetry plane).
A set of meshes was generated using different sizes of
Fig. 8 shows the quality histogram for the same mesh. This quality
elements both around the vehicle and in its nearby wake. The is based on a method of evaluating the determinant of the Jacobian
total number of elements in those meshes was in the range of 500 matrix computed for each element of the mesh. It can be observed that
thousand to 4.2x106 . A mesh convergence analysis was when implementing the hybrid mesh, specifically the prism layer on the
performed based on the drag coefficient (CD) obtained for each vehicle, the quality of the mesh is lowered compared to the totally

5 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 tetrahedral mesh created initially, due to the pronounced changes of configured with a constant pressure with the local atmospheric
direction of the surfaces of this vehicle, which generates the pressure value. The lateral, superior and symmetry wall (the
deformation of the prisms and therefore a lower quality of them. latter divides the computational domain into two equal parts),
However, the quality of the generated mesh is acceptable since most of were configured with the condition of symmetry, in order not to
its elements are in the range of 0.7 and 0.95. In other words, a bit of
mesh quality must be sacrificed to obtain a better capture of the
encapsulate or enclose the wake produced by the vehicle and thus
phenomena present boundary layer. avoid inaccuracies in the results of the model. Finally, the floor,
the car and the tires were configured as static walls with non-slip
condition for this case.

The model was implemented in the commercial software

ANSYS-FLUENT V18.1. All simulations were developed under
steady-state conditions. The coupling method used for the
pressure-velocity governing equations was the semi-implicit
method for pressure-linked equations (SIMPLE) with a second-
order spatial discretization. For turbulence, the Spalart-Allmaras
model was used due to its capabilities for the simulation of
external flows with high Reynolds number and with low
computational cost. The convergence criterion was set to 1 ×
10−6 with 1000 iterations.

Numerical Results

One of the parameters to be analyzed is the aerodynamic

Fig. 7: Detailed view of hybrid mesh (i.e. detail of prisms in the drag coefficient (CD) of the Baja SAE vehicle, for this, it is
symmetry).
decided to take the average value of the coefficient of the last
100 iterations of the simulation, obtaining a value of CD = 0.757
for the mesh hybrid selected, which has a higher precision
compared to the unstructured mesh formed by tetrahedra, which
showed a coefficient of drag coefficient of 0.7804; besides the
advantage in the capture of the viscous phenomena and therefore
of vorticity that the hybrid mesh has in comparison with the
unstructured ones.
Fig. 9 shows the evolution of the residuals of the
equations of continuity, velocity and turbulent viscosity,
throughout the 1000 iterations. Also, fig.10 shows the evolution
of the CD. The tendency to stabilize both the curves of residuals
and the CD can be observed, if a greater precision is desired, it
is suggested to take the iterations up to a value of 5000.

Fig. 8: Quality histogram for the mesh (i.e. .number of elements vs.
quality)

Considerations of the model and implementation

In the model implemented, it was assumed that the flow

was incompressible (isothermal), Newtonian, and three-
dimensional. The working fluid is air at standard conditions (T =
25ºC, at 1 pressure atmosphere) giving a Reynolds number of
1.02x106 , based on the wheelbase.

The border conditions implemented in the model are as

follows: the air velocity at the entrance was set constant at Fig. 9: Evolution of the residuals.
30 km/hr. (i.e. 8.33 m/s). The output boundary (back) was

6 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 consequently this increases the drag force of the vehicle due to
the overpressure in the front part and the high pressure drop in
the back of it.

Fig. 10: Evolution of CD

Then, the fields of pressure, velocity, current lines and

turbulent viscosity for the plane of symmetry are shown in fig.11
and fig. 12. These graphs allows to see a good representation of
the behavior of the air flow around the vehicle.

Fig. 12: Speed field on the plane of symmetry and detail view.

So, in fig. 12, the acceleration of the fluid in the lower part of
the vehicle is observed, in the conduit that is formed between the
chassis of the vehicle and the ground, this due to the
strangulation that the air flow suffers in this area. Therefore, to
avoid this phenomenon, the height of the vehicle to the road must
be reduced, taking into account the design regulations according
to the SAE. A similar phenomenon occurs in the area between
the driver and the upper part of the chassis. These phenomena
produce higher velocity gradients in the rear of the vehicle,
generating a separation or recirculation bubble of greater
magnitude. To understand a little better, here is a detailed view
of the fluid stream lines in the plane of symmetry.

As can be seen in fig. 13, two bubbles of recirculation or

detachment (A and B) appear in opposite directions in the rear of
Fig. 11: Field of pressures on the vehicle and plane of symmetry. the vehicle, one with greater intensity than the other, as it is said
in the theory [15]. Additionally, a recirculation is observed in the
Also, in fig 11, the stagnation points of the vehicle are clearly front part of the vehicle (C), due to the straight shape of the
seen, in the front part of the fairing and the driver's breastplate,

7 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 fairing that directly impacts the flow, which generates an
acceleration of the flow in all directions.

Fig. 14: Turbulent viscosity ratio on symmetry plane.

As shown in fig. 15, 16, 17 and 18, there are two pairs of main
vortices that propagate on the surface of the vehicle towards its wake,
they are those produced in the tires and in the upper bars of the roof of
the vehicle according with the theory of vorticity in the aerodynamics
of land vehicles. In addition, a vortex of great magnitude arises in the
rear central part of the vehicle, which propagates 1 meter behind it.
When observing these same iso-surfaces of vorticity but with the
intensity of the TVR, it is concluded that this last mentioned vortex,
presents a great intensity and influence in the general aerodynamics of
the Baja SAE. The vorticity in the normal flow direction is observed in
a vertical plane, located 10 cms behind the back of the vehicle, as shown
in fig.18.

For a correct visualization of the pairs of vortices (positive and

negative) that propagate in the wake of the vehicle, a symmetric
visualization scale should be taken into account, in this case a threshold
of [-30, 30] was raised; the pairs of vortices produced by the upper bars,
the side doors of the vehicle and the higher intensity produced by the
wheels are observed. It is clarified that the vortex of greater diameter
located in the central part of the stela is not reached to visualize because
it arises 1 meter behind the vehicle. As the pairs of vortices move in the
wake, they will be dissipated by the effect of the diffusion of the
vorticity produced by the Laplacian in the Navier - Stokes transport
equation. Fig. 19 illustrates this issue.
Fig. 13: Stream lines and bubbles of recirculation (i.e. detachment in
the vehicle).

On the other hand, the normalized turbulent viscosity

(TVR) contour is obtained (with respect to the unmodified
turbulent viscosity) as shown in fig.14. This property makes it
possible to identify the areas with the greatest tendency to
produce vorticity. It can be verified according to the theory of
aerodynamics of land vehicles that the greatest turbulent
viscosity change arises in the wake of the vehicle.

Finally, with the analysis of obtained results, we proceed to

identify the vorticity fields around the vehicle, for this, it is
applied in the vortex visualization method known as "Q-
criterion", which is defined as the fluid region that it has the
Fig. 15: Isosurfaces around the vehicle (color by intensity of vorticity).
second invariant of the positive velocity tensor [15].

8 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 16: Isosurfaces - side view (color by intensity of vorticity).

Fig. 19: Couples of vórtices in the direction of flow, in the wake near
the vehicle.

CONCLUSIONS
A computational study of the aerodynamics of a modified baja
SAE vehicle was presented, including an analysis of the vortical
structures in the near wake. For this purpose, a structured mesh
was generated using prisms close to the vehicle surface and
tetrahedral elements everywhere else. Several refinement
regions were used in the wake in order to improve the resolution
of the details of the flow in this region. The numerical model was
implemented in the commercial software ANSYS FLUENT
v14.5.
Results of the drag coefficient of the vehicle predicted a value of
Fig. 17: Isosurfaces on the vehicle (color by intensity TVR). CD = 0.757 which will be compared with experimental data in
the future. This result presents a higher precision compared to
the drag coefficient obtained by the unstructured mesh, with a
value of CD = 0.7804.
Although the Baja SAE is not a high-speed vehicle, if better
performance is desired, it would be ideal to decrease the current
CD, for this, it is necessary to reduce the slant angle (of the upper
rear part) of the vehicle, since currently it is close to 90 °, and
the optimal value is 32 ° or less. In this way, the separation
produced in the rear part of the vehicle (nearby wake) is reduced.

By generating a hybrid mesh with layers of prisms on non-

uniform surfaces of the vehicle, the viscous phenomena of the
boundary layer can be more efficiently captured, but the
drawback of the deformation of the elements in these layers is
presented, thus decreasing the quality of the mesh.

ACKNOWLEDGEMENTS
This work is supported by the Dirección de Investigaciones
Fig. 18: Isosurfaces of vorticity - rear view (color by intensity TVR). of the Universidad Autónoma de Bucaramanga, where the
research project has a grant to supports Bachelor´s students in
mechatronics engineering, in design of automotive systems and
fluid engineering.

9 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 REFERENCES
[1][Online]https://es.wikipedia.org/wiki/Sociedad_de_Ingenier
os_Automotrices.
[2] Balaguera, S., Fonseca, J., Jiménez., 2011, “Diseño y
Construcción de un Monoplaza como Soporte para la Propuesta
de una Competencia Universitaria de Vehículos Monoplaza
‘Fórmula U Colombia’”, BSc. Thesis, Universidad Industrial de
Santander, COL.
[3] Serrano, F., Prieto, R., Vázquez, M., 2004, “Diseño y
Construcción de un Mini Baja SAE,” BSc. Thesis, Instituto
Politécnico Nacional, MEX.
[4]2017 Baja SAE Rules [Online]
http://bajasae.net/content/2017_Baja_Rules_1_3_2017.pdf
[5][Online]
http://web.a.ebscohost.com.aure.unab.edu.co/ehost/pdfviewer/p
dfviewer?sid=31894ba9-dd39-41d0-9735
5780d3c00494%40sessionmgr4010&vid=1&hid=4206
[6]López, A., Maya, H, 2015, “Diseño del paquete aerodinámico
de un vehículo de carreras Fórmula Student”, BSc. Thesis,
Instituto Politécnico Nacional, Guanajuato, MEX.
[7] Correa, M., 2010, “ Modelado y Simulación Dinámica de
Vehículos de Competición”, MSc. Thesis, Universidad
Politécnica de Valencia, SPA.
[8]Fryšták, B., 2016, “Fórmula SAE Aerodynamic
Optimization”, MSc. Thesis, Brno University of Technology,
Brno, CZ.
[9]Cajamarca, D., Vásquez, C, 2016, “Diseño y Construcción de
la Carrocería para el Prototipo de Competencia Shell Eco
Marathon”, BSc. Thesis, Universidad del Azuay, Cuenca, EC.
[10]Franck, G.,Nigro, N., Storti, M., D’Elía, J., 2003, “
Modelización del Flujo de Viento sobre el Modelo de Ahmed”,
Mecánica Computacional, 22(1), pp. 124-142.
[11] Franck, G., Carazo, F., Nigro, N., Storti, M., D’Elía, J.,2004,
“Simulación Númerica del Flujo alrededor del Modelo de
Ahmed para un Angulo de Inclinación Crítico”, Mecánica
Computacional, 23(1), pp. 2189-2209.
[12]Altinisik, A., Kutukceken, E., & Umur, H. (2015).
Experimental and numerical aerodynamic analysis of a
passenger car: Influence of the blockage ratio on drag
coefficient. Journal of Fluids Engineering, 137(8), 081104.
[13]Falkovén, A., Nordin, J., Rama, K., Rask, A., Sjostran, P.,
Stadler, M., 2015, “CFD-Analysis of the Aerodynamic Prop-
erties of a Mercedes-Benz 300SLR”, BSc. Thesis, Chalmers
University of Technology, Gotemburg, SE.
[14] Tsubokura, M., Kobayashi, R., Ikawa, Y., Nakashima, T.,
2012, “ Unsteady Aerodynamics of Road Vehicle in Dynamic
Maneuvering”, Proc. 30th AIAA Applied Aerodynamics
Conference, New Orleans, Louisiana, pp. 1-18.
[15] López, O.; Ardila, S.; Blanco, D.; Muñoz, L., 2017,
“Analysis of the Vorticity in the Near Wake of a Station
Wagon” Journal of Fluids Engineering, 139(2), 021105.
[16] Castro, N., López, O., Muñoz, L., 2013, “Computational
Prediction of a Vehicle Aerodynamics Using Detached Eddy
Simulation,” SAE Int. J. Passenger Cars—Mech, Syst.,6(1),
pp.414-423.

10 Copyright © 2018 ASME

Downloaded From: https://proceedings.asmedigitalcollection.asme.org on 03/24/2019 Terms of Use: http://www.asme.org/about-asme/terms-of-use