SU2 Aerodynamic Shape OptimizationNew

Background

Gradient-based shape optimization needs accurate and cheap gradients of drag/lift with respect to hundreds of shape variables. The discrete/continuous adjoint gives this at a cost nearly independent of the number of design variables - the core capability of SU2.

Typical workflow

1. Set up RANS case with SA or SST turbulence model
2. Define objective (drag, lift-to-drag, moment) and design variables
3. Run continuous adjoint to obtain surface sensitivities
4. Shape optimization loop with mesh deformation (FFD or free-form)

Physics-ML extensions

• Learned turbulence corrections and blending models
• Surrogates inside the optimization loop for early-stage trade studies
• Reduced-order adjoint models for real-time design exploration

Use cases

• Airfoil and wing redesign for drag reduction
• Turbo-machinery blade optimization
• Academic benchmarks for adjoint verification

Resources

• SU2 documentation with tutorial suite and adjoint examples
• Reference papers on the adjoint formulation and verification cases