NACA 0012 Solver Validation Study
Gradient Dynamics benchmarked the FluxCore compressible RANS solver and standard Spalart-Allmaras turbulence model against the NASA Turbulence Modeling Resource NACA 0012 validation case.
The NACA 0012 is one of the most widely used aerodynamic verification cases. Its simple geometry is paired with extensive experimental measurements and reference CFD results, making it possible to assess force prediction, pressure loading, wall shear, wake development, and the onset of separation within one controlled benchmark.
This study evaluates FluxCore at three increasingly demanding incidence conditions. Predictions are compared with nine established CFD codes, NASA TMR reference solutions, and experimental pressure and lift data to test both integrated engineering quantities and detailed surface behaviour.
Mach number
0.15
Chord Reynolds number
6.0 × 10⁶
Angles of attack
0°, 10°, 15°
Mesh size
229,376 cells
Benchmark and methodology
The computational model uses the modified NACA 0012 profile supplied by NASA TMR, scaled to a one-metre chord with a sharp, closed trailing edge. The Family II, Level-4 structured C-grid contains 897 × 2 × 257 nodes and 229,376 hexahedral cells. Its C-type topology wraps around the airfoil and reconnects in the wake, with the farfield located approximately 500 chord lengths away.
Steady, fully turbulent compressible RANS simulations were run at a freestream Mach number of 0.15 and a chord Reynolds number of six million. The study covers angles of attack of 0°, 10°, and 15°, using second-order MUSCL reconstruction, the van Albada limiter, and the SLAU2 inviscid-flux scheme.
The airfoil is modelled as an adiabatic no-slip wall with low-Reynolds-number near-wall integration and no wall function. Characteristic conditions are applied at the farfield, with symmetry boundaries across the single-cell spanwise direction. Convergence targets were set to 10⁻¹² for the mean-flow residuals and 10⁻¹⁰ for the Spalart-Allmaras residual.
Integrated force results
Across all three incidence conditions, FluxCore remains within the spread of the reference-solver results. The comparison below shows the FluxCore values beside Flow360, one of the reference datasets reported in the study.
| Angle | FluxCore Cᴅ | FluxCore Cʟ | Flow360 Cᴅ | Flow360 Cʟ | Observation |
|---|---|---|---|---|---|
| 0° | 0.00808 | 0.0000 | 0.00808 | ≈ 0 | Exact drag match; symmetric zero-lift solution |
| 10° | 0.01247 | 1.0893 | 0.01244 | 1.0889 | Closely matched forces with attached upper-surface flow |
| 15° | 0.02153 | 1.5451 | 0.02158 | 1.5381 | Both coefficients within 0.5% of Flow360 |
Results at 0° incidence
At zero incidence, symmetry requires zero lift. FluxCore predicts Cʟ = 0 and Cᴅ = 0.00808, exactly matching the reported Flow360 drag value and sitting within the narrow range produced by the wider reference set. Coincident upper- and lower-surface distributions confirm a symmetric, attached solution. The pressure coefficient follows the Gregory and O'Reilly measurements over the chord, while skin friction varies smoothly toward the trailing edge.
Results at 10° incidence
At 10°, FluxCore predicts Cʟ = 1.0893 and Cᴅ = 0.01247, compared with Flow360 values of 1.0889 and 0.01244. The increased incidence produces a strong leading-edge suction peak and a thicker wake. FluxCore captures the measured pressure recovery, while the positive and smoothly decaying upper-surface skin friction indicates that the mean flow remains attached.
Results at 15° incidence
At 15°, FluxCore predicts Cʟ = 1.5451 and Cᴅ = 0.02153. Both coefficients lie within the full reference-solver range and differ from the Flow360 results by less than 0.5%. Suction-side pressure loading increases further, and the upper-surface skin-friction coefficient approaches zero near the trailing edge, consistent with incipient separation reported by the NASA TMR and Flow360 studies.
Surface data and lift trend
The agreement is not limited to integrated lift and drag. Across the three angles, the surface-pressure distributions track the Gregory and O'Reilly measurements, the skin-friction curves follow the CFL3D reference behaviour, and the overall lift curve agrees closely with Ladson's tripped experimental data at a Reynolds number of six million.
Conclusion
FluxCore reproduces the established NACA 0012 aerodynamic trends from symmetric attached flow through high-incidence loading and the onset of trailing-edge separation. At every evaluated angle, its force coefficients remain within the spread of nine established reference solvers, while the detailed pressure and skin-friction distributions show the expected physical behaviour.
Taken together, these results provide evidence that FluxCore's compressible RANS implementation and standard Spalart-Allmaras model are correctly capturing the dominant physics of this benchmark. The study supports use of the solver for attached and moderately separated external-aerodynamic flows, while also establishing a transparent reference point for continued validation on more complex geometries and operating conditions.
