Non-Intrusive Computational Method and Uncertainty Quantification Tool for Isolator Operability Calculations

Computational fluid dynamics (CFD) technology plays a strong role in the design and development of aerospace and defense vehicles such as high-speed applications where testing under the correct operational conditions is not yet viable. Despite decades of research towards making CFD predictive and reliable, it has not proven so due to the significant uncertainties in physical models, initial/boundary conditions, computational mesh, numerical schemes and methods. In the proposed effort CFDRC in partnership with Virginia Tech and UTSI, aims to directly address these issues by integrating dimensionally adaptive sparse grid uncertainty quantification (UQ) method with an existing reacting CFD solver. The proposers demonstrated this approach to be suitable for achieving this objective during Phase I on a NASA-LaRC nozzle-isolator lab-scale setup. The proposed effort will deliver a practical user-friendly automated software tool combining UQ with CFD (UQCFD), capable of identifying and characterizing regions of high-uncertainty in the CFD code and the associated work-flow, and thereby, provide guidance to the CFD modeler to increase fidelity of those regions. UQCFD software has the potential to make significant impact on a wide variety of application utilizing CFD predictions including design and development of next generation supersonic and hypersonic flight vehicles.