EXTERNAL FLOW ANALYSIS OVER AHMED BODY USING ANSYS FLUENT

OBJECTIVE:

The objective of this project is to perform an external flow simulation over the Ahmed body using Ansys Fluent and to study its velocity and pressure profiles in detail. A grid dependency test will also be conducted to study the changes in the velocity profile accordingly and the significance of negative pressure values will also be studied and explained.

AHMED BODY-DESCRIPTION:

Ahmed body is a generic 3D model representing a passenger car used to find the flow over an automobile.It captures the essential flow features around the automobile.It was first described by S.R Ahmed in 1984. It 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 Ahmed body helps in validating the numerical model before designing a new model of a car. It is useful in trusting our experimental data.It also helps in define the turbulent flow around the car like body.

Fig-1: Ahmed body dimensions

Significance of Ahmed body:

The Ahmed body helps in validating the numerical model before designing a new model of a car. It is useful in trusting our experimental data.It also helps in define the turbulent flow around the car like body. It is very much useful in studying the wake region behind the back surface as it is helpful in studying the recirculation areas.

SIMULATION PROCEDURE:

The simulation will be conducted for 3 set of grids. First a baseline mesh is created and simulation is conducted, then the mesh element size is decreased and the quality of the mesh is also increased and then a much refined setup is created and simulated.

Setup - 1 : BASELINE MESH

As a first step, the ahmed body model is imported into Spaceclaim. Then to perform an external flow anlaysis, a virtual wind tunnel has to be build. This is done by creating an enclosure with appropriate dimensions. It is as below.

Fig-2:Ahmed body model with the enclosure

Then this model in incorporated into meshing module. There a basic mesh is created using the parameters already been specified by the software. The parameters are as follows,

1. Mesh element size - 0.413m

2. Inflation layer - Nil

3. No.of nodes - 16654

4. No. of elements - 84443

5. Type of mesh elements - Tetrahedral

Fig-3:Baseline mesh setup

Fig-4:Baseline mesh setup (closer view)

As it can be seen from the above two figures, the baseline mesh setup is very coarse.This indicates less accuracy of the results during simulation. Then an important step has to be carried out, which is naming the faces which is turn is very helpful during specifying the boundary conditions.

Setup - 2 : REFINED MESH:

Before refining the mesh, a change preferred in the modelling the enclosure itself. Our area of focus in majorly on the car wall and the area around it. So, to capture the flow field variables in these areas, another enclosure is being created as shown in the figure below.

Fig-5:Ahmed body with two enclosures

Then this model in imported in mesh modeller to carry out the meshing process. Here, a lot of changes is included for refining the mesh. It is listed below.

1. Mesh element size - 0.1 m

2. Inflation layer - 5

3. Inflation layer maximum thickness - 0.0125 m

4. No.of nodes - 59444

5. No. of elements - 196509

6. Type of mesh elements in outer enclosure - Hexahedral

7. Type of mesh elements in inner enclosure - Tetrahedral

Fig-6:Refined mesh setup

Fig-7:Closer look on inflation layer

So, a mesh model with much smaller element size is created with change in the type of mesh elements. Also, a method called inflation layer is created which is a body fitter mesh. It gives accurate solutions on and around the car body.

Setup - 3 : HIGH-REFINED MESH:

This time, no change is made in the model or in the enclosure,but the mesh setup is pushed to the extend as the element size and the number of elements are specified to the extreme values possible. The values are as follows.

1. Mesh element size - 0.05 m

2. Inflation layer - 7

3. Inflation layer maximum thickness - 0.02 m

4. No.of nodes - 202790

5. No. of elements - 500554

6. Type of mesh elements in outer enclosure - Hexahedral

7. Type of mesh elements in inner enclosure - Tetrahedral

8. Mesh element size in inner enclosure - 0.036 m

Fig-8:High refined mesh setup

Fig-9:Inflation layer in high refined mesh

So, as the three mesh setups are ready, the next step is to simulate the flow in ANSYS Fluent.

SIMULATING THE FLOW IN FLUENT:

In Fluent, the physics setup for all the three mesh models will be same which means, the boundary conditions, cell zones, viscous models, solver type, solving method and number of iterations are all same. Here are the details.

1. Solver - Density based steady state solver

2. Physical model - Viscous model : Realizable k-epsilon model with standard wall functions

3. Working fluid - Air

4. Boundary conditions

a) Inlet velocity - 50 m/s

b) Outlet gauge pressure - 0 pascal

c) Car wall and enclosure lower wall - No Slip

d) Temperature - 300 K

5. Solution method - Implicit

6. Initialization - Hybrid

7. No.of iterations - 500

Eventhough the mach number for above flow is 0.15, a density based solver is used to consider the compressibility effects and also for high accuracy. So the calculation is done for 500 iterations and the residual plots are as follows.

Fig-10:Residuals of baseline mesh setup

Fig-11:Residuals of refined mesh setup

Fig-12:Residuals of high refined mesh setup

RESULTS IN CFD-POST:

Then the results are post-processed using CFD post. A mid plane is created a z axis and the velocity contour is displayed on it to study the effect of grids on the velocity of the simulation. Also, vectors are also included in order to understand the direction and magnitude of flow.

Results of baseline setup:

The velocity contour of the baseline mesh setup is as follows. In this following figure the plane is coloured with velocity and the car wall is coloured with pressure.

Fig-13:Velocity contour of baseline setup(1)

Fig-14:Velocity contour of baseline setup(2)

Fig-15:Velocity plot of baseline setup

Inference:

In fig-13 and 14, it can be noted that the velocity around the car wall and the lower wall of the enclosure are zero. This is due to the no slip boundary condition. The contour looks a bit coarse due larger element size. The maximum velocity obtained is 100.2 m/s. In the contour, it is visible that the velocity is much higher at the front round edges of the wall. The velocity profile plot in fig-15 looks pretty neat for such a baseline setup. But the boundary condition for the lower wall of the enclosure is diffused. Also, the wake region does not seems to be accurate.

Results of Refined setup:

Fig-16:Velocity contour of refined setup(1)

Fig-17:Velocity contour of refined setup(2)

Fig-18:Velocity plot of refined setup

Inference:

The velocity contour is much similar to that of the baseline mesh. But the key difference is the vector concentration. It is high compared to that of the previous setup due to the refinement. But there is change in the velocity profile. The velocity abruptly changes from zero to 45 m/s from the wall. This is due to the no slip boundary condition. It is well captured in particular setup.

Results of High refined setup:

Fig-19:Velocity contour of high refined setup(1)

Fig-20:Velocity contour of high refined setup(2)

Fig-21:Velocity plot of high refined setup

Inference:

In this highly refined case, the boundary conditions are very well captured as seen in the velocity plot in fig-21. The velocity contour looks much finer compared to both cases.

Grid Dependency:

This simulation is grid dependant as it can be seen from above plots and contours. Mainly, the quality in the processing of boundary conditions is increasing as the element size is decreased. In the final setup, the wake region is aslo pretty accurate showing the correct recirculation regions. Moreover, the contour is also much refined and clear in the last setup proving that it the optimal mesh setup to carry out this particular simulation.

Wake region:

The following fig-22 shows the wake region as accurate as possible among the three setups. It can be clearly seen there are two recirculation regions. One from the flow over the body and the other forming due to the flow below the body. They both seperates from the body and is recirculated at the rear end causing the zero velocity there.

Fig-22:Wake region in high refined setup

The fig-23 below shows the formation of vortices at the rear end of the ahmed body. This is due to the slant angle at the rear end. The flow on the sides of the body when interrupted by the slanting surface, it gets deflected inside and get sucked to the bottom due to low pressure region forming the vortices.

Fig-23:Vortex formation due to slant angle at the rear side

SIGNIFICANCE OF NEGATIVE PRESSURE:

Firstly, the negative pressure obtained is not the absolute pressure values. These are relative pressure values based on the pressure in neighbouring areas of flow. From the fig-24, it can be seen that the pressure is at the highest at the front end. This is due to impact of air straight on the body

Fig-24:Pressure on the car wall(1)

And it goes negative at the front rounded edge. This is because of the boundary layer seperation in that area causing a low pressure region. It's the same case at the start point of the slant angle. If noted correctly, all the edges are having this low pressure region with relatively negative values. This in turn causes the adverse pressure gradient which is the reason for the formation of vortices. This huge pressure difference between the front and rear end of the body, it experience a huge amount of drag which is highly undesirable. So the car body should be designed in such a way to reduce this pressure difference.

Fig-25:Pressure on the car wall(2)

CONCLUSION:

The ahmed body is highly helpful in understanding the behaviour of any moving body through fluid. As said earlier, the mesh setup with 500554 elements is the optimal setup to run this simulation. Also, it is helpful in explaining how the negative pressure affects the efficiency of an automobile since it increases the drag thus consuming more fuel.