# Differences

This shows you the differences between two versions of the page.

 — solvers:2dnodalgradient [2017/07/18 14:25] (current)fgillet created 2017/07/18 14:25 fgillet created 2017/07/18 14:25 fgillet created Line 1: Line 1: + =====  ===== + ==== General Information ==== + + * **Solver Fortran File:** ''​Compute2DNodalGradient.F90''​ + * **Solver Name:** ''​ElmerIce_MeshAdapt2D(Compute2DNodalGradient)''​ + * **Required Output Variable(s):​** ​ + * (1) ''​g''​ (dofs=2) + * **Required Input Variable(s):​** ​ + * (1) ''​v''​ + * **Optional Output Variable(s):​** None  ​ + * **Optional Input Variable(s):​** None + + ==== General Description ==== + This is a pseudo solver (i.e. it is not solving an equation). This solver compute the nodal 2D gradient vector <​m>​g_i={{\partial v}/​{\partial x_i}}​ of a variable //v//. + + This is used for example by the mesh adaptation procedure to compute the hessian of //v// ([[mesh:​meshadaptation|Mesh Adaptation]]). ​ + + By default (//FE consistent average = Logical True//), this is done using a //​L²//​-projection on the FE mesh; If //FE consistent average = False//, at a given node the derivative is simply the average of the derivatives evaluated at the node in each elements sharing the node. + + ==== SIF contents ==== + <​code>​ + + Solver 2 + Equation = "Nodal Gradient"​ + Variable = -dofs 2 "​g"​ + Procedure = "​ElmerIce_MeshAdapt2D"​ "​Compute2DNodalGradient"​ + + Optimize Bandwidth = False + + Variable Name = string "​v"​ + FE consistent average = Logical True + End + + ​ + + + ==== Example ==== + Examples for anisotropic mesh adaptation can be found under ''​[ELMER_TRUNK]/​elmerice/​Tests/​MMG2D_Aniso1''​ and ''​[ELMER_TRUNK]/​elmerice/​Tests/​MMG2D_Aniso2'',​ where the mesh size is adapted using 1 or 2 variables (i.e. combining metric informations),​ respectively.