CORVID TECHNOLOGIES, LLC — Department of Defense SBIR Phase I: AF221-0003

Description

Corvid Technologies will develop a toolset capable of accurately and efficiently generating aerodynamic databases and providing rapid linear and 6DoF analysis. Corvid will demonstrate this toolset using a multitude of validated in-house analysis tools on either the Army-Navy Finner or another preferred airframe. RavenCFD, Corvid’s massively parallel unstructured Navier-Stokes solver, will be used for rapid database generation and toolset validation. Corvid will use virtual geometry insertion (VGI) in Raven to generate sparse datasets quickly and efficiently over a range of Mach, angle of attack, roll angle, and fin deflection angles. The surface mesh will be generated using SAVANT, Corvid’s Surface And Volume Automated Numerical Tessellation tool which completely automates the meshing process. The sparse datasets will be augmented with a neural network (NN) trained to predict computational fluid dynamics (CFD) solutions where RavenCFD is not run, saving time and computational resources while still providing a full database of CFD results. The CFD results will be validated against traditional, high-fidelity conformal grids in Raven to demonstrate that the VGI/NN combination predicts coefficients and stability derivatives with reasonable margins from the high-fidelity grids. The airframe’s stability and phase margins will be analyzed using Corvid’s linear control analysis tools, and the overall airframe performance will be analyzed using VIDOF, Corvid’s in-house six degree of freedom flight dynamics simulation tool. This toolset will be expanded to account for damaged airframes and include a graphical user interface (GUI) as well as validated with the support of Lockheed Martin in either wind tunnel or free-flight testing during Phase II.