EDP Sciences, ESAIM: Proceedings, (29), p. 89-107
DOI: 10.1051/proc/2009057
Full text: Download
We employ a Fourier-Galerkin method to solve the 2D incompressible Euler equations, and study several ways to regularize the solution by wavelet filtering at each timestep. Real-valued orthogonal wavelets and complex-valued wavelets are considered, combined with either linear or nonlinear filtering. The results are compared with those obtained via classical viscous and hyperviscous regularization methods. Wavelet regularization using complex-valued wavelets performs as well in terms of L 2 convergence rate to the reference solution. The compression rate for homogeneous 2D turbulence is around 3 for this method, suggesting that memory and CPU time could be reduced in an adaptive wavelet computation. Our results also suggest L 2 convergence to the reference solution without any regularization, in contrast to what is obtained for the 1D Burgers equation.