THE UNCERTAINTY QUANTIFICATION METHODS DEVELOPED UNDER THIS PROGRAM ARE DESIGNED FOR USE WITH CURRENT STATE…

Public award description

THE UNCERTAINTY QUANTIFICATION METHODS DEVELOPED UNDER THIS PROGRAM ARE DESIGNED FOR USE WITH CURRENT STATE-OF-THE-ART FLOW SOLVERS DEVELOPED BY AND IN USE AT NASA. THE PHASE I PROGRAM DEMONSTRATED THE CRISP CFDREG ERROR QUANTIFICATION AND REDUCTION CODE WITH SIMULATIONS CONDUCTED USING THE NASA UNSTRUCTURED SOLVERS FUN3D AND USM3D. PHASE I PROVIDED EVIDENCE SUPPORTING THE SUSPECTED NEED FOR AN ERROR PREDICTION CODE THAT MATCHES THE FINITE VOLUME SCHEME OF THE NAVIER-STOKES SOLVER ITSELF. PHASE II WILL CONTINUE THIS WORK BY EXPANDING OUR ERROR TRANSPORT EQUATION (ETE) SOLVER TO TREAT BOTH CLASSES OF UNSTRUCTURED GRID FINITE VOLUME SCHEMES. SUPPORT FOR THE CGNS STANDARD WILL BE IMPLEMENTED AND PERMIT USE OF THE PHASE II PRODUCT BY A BROADER SPECTRUM OF POTENTIAL USERS. SPECIFIC ISSUES THAT AFFECT NUMERICAL ACCURACY OF THE ERROR PREDICTIONS AND HOW THEY PROPAGATE INTO INTEGRATED QUANTITIES SUCH AS LIFT AND DRAG COEFFICIENTS WILL BE ADDRESSED. REDUCTION OF ERROR FOR LARGE SCALE MESHES IS A MATTER OF EQUAL IMPORTANCE, AND IMPROVEMENTS ARE PLANNED THAT WILL PROVIDE FOR ANISOTROPIC GRID REFINEMENT WITHIN THE EXISTING CRISP CFDREG MESH ADAPTATION CODE. FINALLY, ERROR QUANTIFICATION APPROACHES FOR TRANSIENT APPLICATIONS WILL BE EXPLORED TO EXPAND THESE DEVELOPMENTS TO PROBLEMS THAT INVOLVE INHERENT UNSTEADINESS AND/OR MOVING BOUNDARIES.