Feature-Driven Cartesian Adaptive Mesh Re nement in the Helios Code

Feature-detection methods are applied to drive Cartesian-based adaptive mesh re nement for improved vortex resolution in aerodynamics ow elds. Several approaches such as the Q-criterion, the eigenvalues of velocity gradient tensor and the identi cation of local pressure minima are considered. Speci c attention is given to the automation of the renement process so that it is applicable to a wide range of ow elds without need for user intervention. This is achieved by careful normalization of the methods and by applying appropriate threshold values to identify speci c regions for mesh re nement. The methods are implemented within the Helios code, which features a dual-mesh paradigm based on using unstructured meshes in the near-body region and adaptive Cartesian meshes in the o -body region. The resultant adaptive code is evaluated for both theoretical and practical test problems, and the results are compared with those obtained by uniform mesh resolution. It is demonstrated that the adaptive solutions provide comparable accuracy for a fraction of the cost of using uniform meshes and thus show promise for e cient and accurate simulations of vortex-dominated aerodynamics ows.