Solution adaptive mesh generation using feature-aligned embedded surface meshes

In this paper, we present a technique for generating solution adaptive meshes that utilizes feature-aligned surfaces embedded in the computational domain. A feature detection algorithm is first employed to identify regions of the domain where features of interest are located. Surfaces associated with these features are then meshed and a volume mesh that respects these embedded surface meshes is then generated. This approach is promising because it provides a mechanism for controlling local grid quality as well as automatically aligning cell faces with the feature. For flows that include topologically and/or geometrically complex features, we describe a mesh refinement technique that can be used in tandem with the embedded surface meshes to provide a mechanism for refining throughout the computational domain. We include results from three cases to illustrate how the method works.