Application of Bubnov-Galerkin formulation to orthogonal grid generation (Q1184646)

A Bubnov-Galerkin formulation is used to solve an elliptic grid generation system \({\partial\over\partial\varepsilon}[f{\partial x\over\partial \varepsilon}]+{\partial\over\partial \rho}[{1\over f}{\partial x\over\partial\rho}]=0\), \({\partial\over\partial\varepsilon}[f{\partial y\over\partial\varepsilon}]+{\partial \over\partial\rho}[{1\over f}{\partial y\over\partial \rho}]=0\), \(f^ 2=g_{22}/g_{11}\), \(g_{ij}\) being elements of the covariant metric tensor, by using linear and quadratic isoparametric elements. The system was developed by \textit{G. Ryskin} and \textit{L. G. Leal} [J. Comput. Phys. 50, 71-100 (1983; Zbl 0579.65123)]. Orthogonal grids of good global accuracy in symmetric and asymmetric regions are obtained. Results are discussed and compared with known ones.