Redesigning Ford Mustang: A CFD Approach to Aerodynamic Efficiency

This report presents a comprehensive study on the aerodynamic redesign of a Ford Mustang to enhance fuel efficiency and reduce emissions. The study employs Computational Fluid Dynamics (CFD) simulations to analyze airflow characteristics and optimize the vehicle's shape. The initial design is assessed, and modifications are implemented to minimize pressure differences, reduce drag, and manage lift. The methodology includes vehicle modeling in Solidworks, pre-processing in Ansys Fluent, mesh generation, and application of appropriate boundary conditions. Results, including velocity and pressure contours, demonstrate the impact of the redesign on aerodynamic performance, highlighting improvements in airflow and pressure distribution. The analysis provides valuable insights into optimizing vehicle aerodynamics for enhanced performance and sustainability. Desklib offers a wealth of resources, including past papers and solved assignments, for students studying similar topics.

Redesign a vehicle:
Abstract:
The world is increasingly becoming more sensitive in achieving a sustainable environment thus
causing a threat to the development and exploitation of oil and gas which leads to more pollution
on the atmosphere. This limitation has led to the increased fuel prices across the world because
the resource is getting depleted thus there is constant need of manufacturing more fuel efficient
vehicles that cause less hydrocarbon emission. This study will analyse the shape of Ford Mustang,
assess the aerodynamic optimisation techniques and implement a new design shape for the
vehicle. A Computation Fluid Domain (CFD) simulation was carried out to assess the flow of air
around the vehicle by determining the air characteristics and the strategies that need to be
implemented to improve the aerodynamics efficiency of the vehicle. This analysis will help
identify the best method that can be used to minimise the pressure difference between the pressure
stagnation at the front and the base drag at the rear of the vehicle. The fluid dynamic of the air
flow also have an effect on the vehicle body by causing the pressure difference between the lower
surface and the upper surface of the vehicle. The observable pressure difference caused an
induced lift on the vehicle. The optimisation of this ford model was caused by fluid dynamics
flowing on top of the vehicle and the process used resulted into a great improvement of the
aerodynamics abilities of the vehicle.
1.1. Computational Fluid dynamics (CFD):
CFD uses a numerical calculation method to analyse the fluid dynamics flowing around a body. The system

simulation involve very small finite volume elements subdivided from the main physical domain where certain
equation are solved numerically. There are three equations involved in fluid dynamics; that is momentum equation,
continuity and energy equations which are derived from the laws of physics which include the conservation of
energy, momentum, and energy. In CFD, the car simulation is only done on the external part of the car hence the
energy equation can be ignored to some extent because vehicles generally travel at low speed and at a constant
temperature, therefore, we can make an assumption that the flow is isothermal and incompressible (Levin and
Rigdal, 2011).
1.1.1. Navier-Stokes equations:
This is also referred to as the momentum equation that applies the Newton’s second law of fluid dynamics.
The equation is used to represent the conservation of momentum. The Navier- stokes equation is mainly
used to predict the fluid pressure and velocity in a given geometry (Levin and Rigdal, 2011)

Figure 5: Navier-Stokes equations
1.1.2. Continuity equation:
This equation is based on the conservation of the mass principle as shown in figure 6. It is also acceptable to
assume the incompressibility of the fluid given in the continuity equation as shown in figure 7. Combining the
two equations of continuity and Navier – stokes gives rise to four unknowns which can be calculated using the
differential equations. (Levin and Rigdal, 2011)
Figure 6: Continuity equation
Figure 7: Continuity equation for incompressible flow

1.1.3. Reynolds averaged Navier-Stokes (RANS):
The Navier-stokes equation can be broken down into RANS equations and solving these equations ease the
possibility of simulating the fluid flows. The process of breaking down the equation creates new terms such as
Reynolds stresses regarded as the function of velocity variations. The increased number of the unknown
equations relative to the number of equations result into a closure problem. To overcome this problem, a
model of turbulence is generated to produce solvable equations. The RANS equations need less calculation
requirements than the original equation (Levin and Rigdal, 2011).

1.1.4. Turbulence flow and modelling:
Turbulent flow is a stream comprising of different changes in pressure and velocity. Turbulent
flow is unpredictable and shifts through existence, these elements makes it difficult for CFD
reproductions to compute turbulent flow. With current PC limits the Navier-Stokes condition and
progression condition are unsolvable. By the utilization of turbulence models the stream field can
be determined with less registering force. These models will change the equation and considers
just the normal impacts of turbulence. A turbulence display cannot give a definite solutions,
instead it will give a gauge to the solution. The choice of turbulence would rely upon the capacity
of the PC and the dimension of accuracy required.
The k-epsilon (k-ϵ) is the most widely recognized turbulence model used to simulate turbulence
stream attributes in CFD simulation. The turbulence in the k-ϵ show is demonstrated by including
the turbulence viscosity (μt), making the k-ϵ display an Eddy Viscosity show. The k-ϵ is a semi-
experimental technique developed on how the active vitality is transported and the rate at which it
dissipates. The biggest eddies gets their active vitality from the main stream, this vitality is
transmitted into littler vortexes and after that winds up to internal energy. The k-ϵ show is a
RANS-model and it utilizes time normal terms, subsequently the model will pass up a great
opportunity contrasts in slope amid moment time steps. In the k-ϵ display the stream is ventured to
be totally turbulent, making the model powerful for just these conditions. (Levin and Rigdal,
2011)

2. Methodology:
2.1. Vehicle modelling:
The designing of the model of the vehicle geometry was generated using Solidworks 2015 software. The car was
drawn in two parts thus having two symmetrical parts that portrayed the end product. The design below in figure
8 is the depiction of the exterior part of unmodified model of the Ford Mustang but has left out minor details and
parts of the car such as wipers. As much as the left out details may affect the aerodynamic forces, their effect in
terms of drag and lift forces is quite insignificant and could be overlooked. The model was later redesigned after
undergoing a test. The new model is demonstrated in figure 9. After creating the models, hey were saved in the
computer as a STEP file and were later taken to the Anys fluent software were the pre-processing procedures
could be carried out.
Figure 8: Initial Design of Ford Mustang
Figure 9: New model of the Ford Mustang

Figures 10a and 10b show the difference that resulted from alterations that were made from the initial model to
the modified model.
Figure 10a: Frontal surface area of the initial model.
Figure 10b: Frontal surface area of the modified model.

2.2. Ansys 17 Fluent Pre-processing:
2.2.1. Developing the Fluid Domain for simulation.
This was to be taken through a fluid domain to simulate air flow around the car. By the help of the Any’s
Software a fluid domain is created around the car. This was attained by placing the imported vehicle geometry
and the vehicle body in an enclosure. The vehicles body is the removed from the boundary to create a space of the
surface of the vehicle geometry in the walls. By the help of the enclosure an air domain is created is therefore used to measure
airflow in the study of the newly designed model. To achieve best results the enclosure has to be of the set size
since the movement of air and the body of model does not have to be tempered with by the designed
walls . The best fit size was set at the length being that of 6 meters away from the back of the car at from the
length was at 2.5 meters nad the height was set 8.6 Meters. The wideness of the enclosure was at 9 meters.
2.2.2. Mesh generation:
To achieve the best and possibly accurate measurements of the drag and lift force, changes on the set
measurements had to be made. The alteration in the size of the enclosure is done to ensure accuracy. In this
case study a homemade estimation of size was effected to the area so as the area around the model could be
well looked at to achieve a higher accuracy levels . As boundary layer separation has a significant effect on
the drag, to resolve the boundary layer ten layers were introduced to the outer surface of the vehicle. The sum
of elements obtained from the mesh was estimated to be 9.97 million elements.
2.2.3. Boundary conditions and Set-up:

The favourable fettle of the boundary in the simulation are as stated. Main areas of concern are the ground,
velocity inlet and outlet. The ground is simply the vehicle and the walls that make up the enclosure. The inlet
with the magnitude of 70mph(31.3m/s) is defined as the inlet-velocity. Lastly is the outlet which was defined as
pressure-outlet whose pressure was set at zero. The software used which was fluent software has the ability to
use to types of solvers. The software used could be pressure based solver or density based solver. In this study,
the pressure solver was used since it best works for incompressible flows. Since the flow
domain is divided into control volumes, to solve each iteration, governing equations will be
introduced to an algebraic equation system which will be linearized. As the simulation was
steady, the solution system used was simple.
To verify if the results were accurate two solution methods were use; SIMPLEC and COUPLED. When using The
COUPLED solution method more time would be required to run simulations since the computing power needed
will be higher. On the other hand, SIMPLEC solution method may result to instabilities in a steady state simulation.
(Levin and Rigdal, 2011) In the study SIMPLEC solution method was find fit and therefore initiated.
Results:
The residuals plot and other results for the simulation are available in the appendix
Table 3: Comparison of the coefficient of lift and drag between the original model and
Redesigned model:
Model Coefficient of lift (Cf) Coefficient of drag (Cd)
Initial 0.163 0.293

Modified 0.113 0.318
Table 3 displays that the Coefficient of lift is positive but also shows a reduction in the lift coefficient. Table 3
also shows a slight increase in the coefficient of drag, which is undesired.
1.1. Velocity Contours at 70 mph:
Figure 11: Velocity contour of initial modely

Figure 12: Velocity contour of modified model
Figures 10 shows the velocity contours of the initial model while figure 11 potrays the velocity contours of
the modified model . On both models, an area with increased velocity is shown on the curve ends of the
bonnet and the roof. The area with slow air flow velocity is however observed at the front and back of the
vehicle. It is evident when comparing figure 10 and 11 that the flow of velocity at the back of redesigned
model is reduced.
1.2. Velocity Streamline Contours at 70 mph:
Figure 13: Velocity streamline contour of the initial structure

Figure 14: Velocity streamline contour of redesigned mod