Grid-characteristic method on tetrahedral unstructured meshes with large topological inhomogeneities

A key difficulty faced when grid-characteristic methods on tetrahedral meshes are used to compute structures of complex geometry is the high computational cost of the problem. Formally, grid-characteristic methods can be used on any tetrahedral mesh. However, a direct generalization of these methods to tetrahedral meshes leads to a time step constraint similar to the Courant step for uniform rectangular grids. For computational domains of complex geometry, meshes nearly always contain very small or very flat tetrahedra. From a practical point of view, this leads to unreasonably small time steps (1-3 orders of magnitude smaller than actual structures) and, accordingly, to unreasonable growth of the amount of computations. In their classical works, A.S. Kholodov and K.M. Magomedov proposed a technique for designing grid-characteristic methods on unstructured meshes with the use of skewed stencils. Below, this technique is used to construct a numerical method that performs efficiently on tetrahedral meshes.