作者: I. Lepot , P. Geuzaine , F. Meers , J. A. Essers , J. M. Vaassen
DOI: 10.1007/978-3-642-58312-4_21
关键词: Euler equations 、 Multigrid method 、 Mathematical optimization 、 Jacobian matrix and determinant 、 Applied mathematics 、 Linear system 、 Discretization 、 System of linear equations 、 Solver 、 Mathematics 、 Generalized minimal residual method
摘要: The aim of this paper is to investigate the interest multigrid techniques used in conjunction with a Newton-Krylov solver. Newton’s method linearize system equations resulting from an implicit discretization Euler on unstructured meshes. These linear systems are solved by preconditioned jacobian-free GMRES solver [1]; storage lower-order jacobian however required for preconditioning purposes. To increase radius convergence method, pseudo-transient continuation and mesh sequencing procedure implemented.