We propose an adaptive-mesh, real-gas Euler finite-element solver for rapid prediction of high speed environments. Mesh adaptation is output-based, providing reliable control of numerical errors associated with quantities of engineering interest, such as drag or integrated wall temperature. The adaptation provides a feedback loop to automatically produce anisotropic meshes that are aligned with shocks, reducing errors induced by mesh-shock mismatch, and eliminates laborious human-in-the-loop mesh generation.
Robust simulation of the strong bow shocks associated with high-speed flows is achieved through the use of a PDE-based artificial viscosity augmenting the conservation equations. The PDE-based articial viscosity provides superior dissipation for shock capturing by distributing artificial viscosity in a smooth manner between neighboring elements. The capability effectively reduces entropy noise
in the post-shock region compared to shock operators based solely on local quantities.
Metric-conforming mesh adaptation will be incorporated into the industry leading Pointwise mesher using established local cavity operators and mesh curving procedures. The metric-conforming meshing algorithm will first be performed under the assumption of linear elements (Q1). For higher order solves, the mesh will then be curved to better approximate the geometry.
Existing finite-rate and multi-species non-equilibrium chemistry models will be analyzed as to their suitability for the higher order mesh adaptation. A specific challenge to be addressed in this project is adjoint compatibility (i.e. well posedness of adjoint PDEs) of the non-equilibrium relations and adjoint consistency of the resulting discretization. Essential to this process is establishing consistency between the mathematical and physical entropies associated with the models, and may require minor adjustments to modeling correlations.
NASA applications include high-speed air vehicles, space vehicle launch, assent, and entry, parachute deployment, and complex multi-species reacting flows.
Military applications include vehicle aerodynamics and store separation. Civil applications include vehicle aerodynamics, low sonic boom aircraft, propulsion integration, rotorcraft, medical device, power generation, and complex reacting flows. There is essentially no market limitation for output-based mesh adaptation coupled with a robust solver.