Multi-level spectral Galerkin method for the Navier-Stokes problem. I: Spatial discretization (Q2580994)

The authors propose a multi-level spectral Galerkin method in two spatial dimensions with the distinguishing feature that on the fine levels only linearised versions of the Navier-Stokes equation are solved; the fully nonlinear Navier-Stokes equations are only solved on low-dimensional coarse subspaces. The general pattern is that the Navier-Stokes equation at a certain level is linearised around the solution of the next coarser level. The approximating subspaces are chosen from the eigen-expansion of the Stokes operator -- essentially the (vector) Laplacian -- followed by a projection onto the divergence-free vector fields. It is shown that if the dimensions of the subspaces grow in a certain manner, then the optimal convergence rate (which would be achieved by solving the Navier-Stokes equation always on the finest level) may be achieved.