Quasi-ENO schemes for unstructured meshes based on umlimited data-dependent least-squares reconstruction (Q1360386)

The paper concerns with high-order reconstruction of functions from cell-averaged values as a part of ENO-type methods. The least-square approach is used for fixed stencils with data-dependent weights. As a result, the author obtains a set of derivatives in the Taylor expansion series needed for the reconstruction. One of the main feature of the technique is its direct applicability to multidimensional unstructured meshes. Numerical examples for third- and fourth-order methods applied to smooth function reconstruction and the Euler equations are presented. In the latter case, unstructured meshes were used to calculate transonic airfoil flow.