Ahmed Body: External Aerodynamics

Background: The Ahmed benchmark

The Ahmed body was introduced by S. R. Ahmed in Some Salient Features of the Time-Averaged Ground Vehicle Wake (1984). It is a standard benchmark for automotive-style external flow: simple enough for controlled CFD studies while retaining a slanted rear deck, cylindrical supports, and a realistic bluff-body wake.

Representative dimensions (from the original study notes):

Length ≈ 1.044 m, height ≈ 0.288 m, width ≈ 0.389 m
0.5 m cylindrical legs; rear slant 40° from horizontal

The Ahmed body is used here to predict the turbulent flow field around a simplified car geometry using a k-ε turbulence model. Typical post-processed quantities include velocity magnitude, pressure, wake structure, and force metrics (drag and lift).

Geometry and computational domain

Ahmed body reference geometry, slanted rear and cylindrical supports — Fig. 1, Reference Ahmed geometry used as the solid surface for the external flow solve.

Inner and outer fluid enclosures around the Ahmed body — Fig. 2, Inner and outer enclosures: mesh is generated in the fluid region outside the body to capture approach flow, boundary layers, and wake.

The model is enclosed so that mesh can be built outside the solid: an inner zone resolves flow over the Ahmed surface; an outer zone provides far-field extent. A grid-dependency study was planned with element counts driven toward the academic license limit (~512k cells).

Meshing and solver setup

Mesh strategy

Multizone method to separate outer and inner domains, concentrating resolution where air flows over the Ahmed body.
Hexa-mapped mesh topology in the mesher.
Sizing on the inner domain with three element sizes from coarser to finer for grid refinement.
Inflation layers on the Ahmed surface to resolve the near-wall gradient.

Solver

Steady, density-based formulation with implicit time marching (as documented in the original project notes).

Boundary conditions (documented values)

Surface / region	Setting
Inlet	Air velocity 50 m/s
Inlet temperature	300 K
Ahmed body walls	No-slip wall (typical external aero setup)
Far-field / outer boundaries	Pressure outlet / freestream extents per enclosure sizing

Case 1 (coarser inner mesh), documented mesh metrics

Item	Value
Inner domain element size	50 mm
Cell count (inner study)	188,471

Case 1, Mesh, convergence, and force / scalar fields

Hexa-mapped volume mesh for Case 1 — Case 1, Volume mesh (multizone / hexa-mapped topology).

Inflation layers on the Ahmed body surface — Inflation layers clustered at the Ahmed surface for boundary-layer resolution.

Residual history Case 1 — Residuals, Case 1 (density-based implicit steady solve).

Drag coefficient convergence history — Drag coefficient history. **Note (from source figure):** reported Cd values on the plot should be multiplied by **10⁻³** to obtain the physical coefficient scale.

Velocity magnitude distribution Case 1 — Velocity magnitude distribution, Case 1.

Pressure distribution on Ahmed body Case 1 — Pressure distribution on the Ahmed surface, Case 1 (suction on legs and separated regions as discussed in the original write-up).

Velocity vectors in the wake behind the Ahmed body — Velocity **vector** field highlighting wake structure and turning of the flow downstream of the body.

Refined mesh, additional residuals, Cd, and contours

The same post-processing sequence was repeated on a finer inner mesh (second mesh level in the grid study). Representative views:

Refined hexa mesh for Ahmed body study — Refined volume mesh, higher cell count toward the licence limit.

Residuals refined mesh case — Residuals, refined mesh case.

Drag coefficient refined case — Drag coefficient, refined case (**same 10⁻³ scale note** applies as on the earlier Cd figure).

Velocity contours refined mesh — Velocity contours with the refined mesh, smoother gradients and clearer wake definition.

Pressure on Ahmed body refined case — Pressure on the Ahmed surface, refined mesh. Regions of **negative gauge pressure** appear on the cylindrical supports and other separated zones, consistent with acceleration over convex surfaces, separation bubbles, shear work, and local recirculation (as discussed in the original narrative).

Velocity vectors refined mesh wake — Velocity vectors, refined mesh, emphasising wake resolution with finer cells.

Comments on grid dependency and results

Takeaways (from the original study text)

Grid refinement smooths the solution. As inner cells are increased toward the licence limit, velocity profiles and vector plots become progressively smoother, a practical sign of reduced discretisation error.
Wake structure becomes clearer. Finer meshes resolve the wake and recirculation regions with less numerical diffusion.
Negative wall pressures are physical, not a solver defect. Low-pressure pockets on the legs and separated zones align with Bernoulli-type acceleration, separation bubbles, shear-related work, and recirculation, all called out explicitly in the source write-up.

Disclosure: This page migrates a legacy Google Sites export into the same long-form structure as the other ANSYS Software Studies case studies. Wording is tightened for clarity; numerical claims follow the figures and notes preserved from the original document.