Week 11- Broadband Noise modelling over Ahmed body

In this project, an Ahmed Body will be simulated in ANSYS Fluent, and the relavant results will be studied. An acoustic modelling simulation of the same body will also be performed.

Ahmed Body is a simplified vehicle body used for capturing basic yet charecteristic features seen in objects used in the automobile industry. The flow around the body consists of all types of flow, from flow accumulation to flow stagnation. There are numerous experimental and simulation data for the body, due to its significance in the CFD community.

The Ahmed Body can be used for validating experimental turbulence models/parameters, to check the validity of one's CFD code, etc.

In this project, the Ahmed Body will be simulated for an inlet velocity of 31 `m//s` or 110 `km//h` along a moving road.

Geometry

The geometry is extremely simple and can be created in any CAD program, this is the drawing of the dimensions of the Ahmed Body:

The slant angle (`varphi`) is 25 degrees for this project.

For this project, the geometry was imported into SpaceClaim. Since the Ahmed Body is symmetrical along the XZ plane, the body can be cut along the XZ plane and a "half-body" can be simulated to save computational resources and time.

A wind tunnel must be created to enclose the body. My geometry already contained the wind tunnel. The geometry is checked for errors, overlapping surfaces and if topology is shared.

This is the final geometry seen in SpaceClaim:

Note: The inner enclousure is a mesh refinement region where a smaller element size will be used to capture data near the body in greater detail.

Meshing

The default mesh with a sizing of 1m was used. Multiple controls were utilised to modify the mesh, they are:

1. Controls -> Sizing -> Body Sizing

This control is to make the refinement region (density region) have a lower element size than the base mesh. The size is 0.03m.

2. Controls -> Inflation -> ahmedbody-walls

This control is to ensure the boundary layer is accurately captured. The height of first cell is calculated from https://www.cadence.com/en_US/home/tools/system-analysis/computational-fluid-dynamics/y-plus.html. For a y+ of 1, length of 1m, and an inlet velocity of 31 m/s the required first cell height is 0.000011m.

3. Controls -> Inflation -> road

This is to ensure the boundary layer near the road is captured accurately. The sizing does not need to be specific since no data is extracted from the road.

This is the final mesh:

Zooming in,

Inflation layers,

The mesh contains 410200 elements and 100511 nodes.

Mesh Metrics

Most elements have a quality greater than 0.7, this indicates that the current mesh is of sufficient quality.

Some elements have a lower quality, but these elements are part of the inflation layers and hence can be ignored.

Simulation Setup

General

The simulated used a steady-state, pressure based solver.

Turbulence Model

k-epsilon Realizable was used with a production limiter to reduce unsteady-ness in the solution. Energy was checked to see the temperature changes.

Acoustics Model

The number of realisations was set to 100, this is the amount of data used to calculate the acoustics properties.

Boundary Conditions

ahmedbody-walls - wall

inlet - velocity-inlet

outlet - pressure-outlet with a gauge pressure of 0 Pa.

road (bottom of wind tunnel) - moving wall with a velocity of 31 m/s.

top and bottom - symmetry

symmetry-plane - symmetry

Reports

Two reports were generated.

1. Lift Coefficient of ahmedbody-wall.

2. Drag Coefficient of ahmedbody-wall.

Solution Controls

The courant number was limited to 10.

The solution was initialised from the inlet using standard initialisation.

Reference Values

The reference values were calculated from the inlet and the area was adjusted to 0.112032 `m^2`(= `0.288xx0.389`, frontal area of Ahmed Body)

Results

The simulation ran for 190 iterations, and these results were obtained from Fluent.

Residuals

The residuals have dropped below 1e-3 and are continois, hence the solution can be said as converged.

Pressure Contours

Another view,

There is a high pressure in the front of the body due to accumulation of oncoming air. This is immediately followed by a low pressure region due to separation of flow. The same can be said for the low pressure region in the region region.

Coefficient of Drag

Value: 0.16662797

Since the value of the coefficient has not significantly changed in over the past 50 iterations, the simulation can also be said to be converged.

Coefficient of Lift

Value: 0.14097649

Since the value of the coefficient has not significantly changed in over the past 50 iterations, the simulation can also be said to be converged.

Velocity Contours

There is a noticable increase in velocity on the top and bottom of the body, this is due to collision with oncoming air and constriction in flow of air, respectively. The low velocity region right behind the body is the wake regiom, it is a region with low velocities and two vortices swirling in opposite directions. This can be seen clearly in a velocity vector plot in the wake region as seen below,

Note: Velocity on the ahmed body was not included in the contour since the velocity on the surface of the body is 0 due to  the boundary layer effect.

Acoustics - Acoustic Power Level

This parameter tells us the amount of sound emitted from the domain in decibels.

Most of the sound emitted from the wake region, due to collision of various streams of air of different velocities. This creates disturbances which causes sound waves.

Acoustics - Surface Acoustic Power Level

On the body, less sound is emitted from the front of the body, this can be said due to accumulation of air in that region.

Conclusion

A simple simulation of an Ahmed Body was performed in ANSYS Fluent. Acoustic data from the body was also generated and explained. The data obtained was qualitatively analysed and reasons were outlined, a quantitative analysis could not be performed due to mesh limitations imposed by the student license.