hi maybe i missed something but why not just use the gradient of the interpolation functions
\grad f = \sum a_i \grad \phi_i as a_i are constant? You would then have the gradients at the quadrature points. You would probably need to average the results if you wish to project the gradients to the nodes. Andrew On 20.05.2011, at 09:36, James Avery wrote: > Dear All, > > Given a finite element function f = \sum a_i \phi_i, I need to > calculate the spatial derivatives df/dx_i, or at least a decent > approximation of this function, preferably with an other FE-function > as result. In the documentation, I have only succeeded in finding the > ApproximateDerivative class, which, while sounding promising, seems to > only calculate derivative norms. Can someone point me to an efficient > way of calculating just the regular derivatives, i.e. the gradient? > > Thanks very much in advance! > > Best regards, > > -- > Med venlig hilsen, > James Avery <[email protected]> > _______________________________________________ > dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii _______________________________________________ dealii mailing list http://poisson.dealii.org/mailman/listinfo/dealii
