At Texas Tech Aero Lab they were using FORDC-2 (Flow Over RaDiator and Condenser 2-Dimensional),based on the Semi-Implicit Methods for Pressre Linked Equations (SIMPLE) an algorithm developed by Spalding and Patankar,adapted by Ford for exterior flow.It embodies the standard K-E (Kappa-Epsilon) turbulence model.
--------------------------------------------------------------------------
Texas Tech also used an academic version of FLUENT,distributed by CREARE,Inc.,a 3-D,finite-difference code also based upon the SIMPLE algorithm,which also embodies the Kappa-Epsilon turbulence model,limited to 10,000 control volumes and orthogonal grids.
--------------------------------------------------------------------------
As of 1999,GM was using a 'proprietary artificial intelligence-based neural network and CFD software.
--------------------------------------------------------------------------
I don't know what any of that means,but they must have the ability to predict 3-D wake structure,which may not be present in an aerospace CFD package,as according to Hucho,would 'presume' attached flow over the aircraft's entirety,not something which could ever be presumed for road vehicles.
It will be great if 'free' software can be adapted to our needs.