The degrees for the Nedelec and Raviart-Thomas elements should now start at 1 (instead of 0). I'm surprised that you are not getting an error -- are you sure that you have updated FIAT?
-- Marie hheum...@math.ethz.ch wrote: > Hello, > > I upgraded my fenics-installation on ubuntu jaunty to the latest version > and observed some problems with lowest order Nedelec elements in 3D. Since > I was not sure, is I'm using the new Expression-Class properly I wrote > some simple 2D test cases and it seems that already the L^2-projection for > lowest order elements is wrong. In the attached sample code the value of > the projection does not converge to the value of the function. If I use in > contrast in Projection.ufl higher order Nedelec-spaces it converges. > > Regards, > Holger > > Projection.ufl: > W1F = FiniteElement("Nedelec 1st kind H(curl)", "triangle",0) > vL = VectorElement("Lagrange", "triangle", 3) > > v = TestFunction(W1F) > u = TrialFunction(W1F) > f = Function(vL) > > a = inner(v, u)*dx > L = inner(v,f)*dx > > and > > main.cpp: > #include <dolfin.h> > #include "Projection.h" > > using namespace dolfin; > int main() > { > class Solution : public Expression > { > public: > Solution() : Expression(2,2) {} > void eval(double* values, const double* x) const > { > values[0] = 1; > values[1] = 0; //sin(DOLFIN_PI*x[0]); > } > }; > UnitSquare mesh(2,2); > Solution solution; > for (unsigned int i=0; i<6; i++) > { > mesh.refine(); > Projection::FunctionSpace VP(mesh); > Projection::BilinearForm m(VP,VP); > Projection::LinearForm LP(VP); > LP.f = solution; > VariationalProblem problemP(m,LP); > Function T(VP); > problemP.solve(T); > double x[2] = {0.123, 0.23441}; > // Evaluate user-defined function f > double value[2] = {0.1,0.1}; > solution.eval(&(value[0]), x); > info("exact value f(x) = (%g,%g)", value[0],value[1]); > // Evaluate discrete function g (projection of f) > double value2[2] = {0.1,0.1}; > T.eval(&(value2[0]), x); > info("value of projection g(x) = (%g,%g)", value2[0],value2[1]); > } > return 0;} > > _______________________________________________ > DOLFIN-dev mailing list > DOLFIN-dev@fenics.org > http://www.fenics.org/mailman/listinfo/dolfin-dev > _______________________________________________ DOLFIN-dev mailing list DOLFIN-dev@fenics.org http://www.fenics.org/mailman/listinfo/dolfin-dev