My PDE is time-dependent and has a coefficient c , which is depending on the gradient of the solution u (due to the usage of Green-Lagrange-strain measure).
Following the documentation „a nonlinear solver is available for the nonlinear elliptic PDE, the parabolic and hyperbolic equation solvers also solve nonlinear and time-dependent problems.“ For elliptic problems it is necessary to explicitly use pdenonlin instead of assempde . But according to the documentation „the parabolic and hyperbolic functions call the nonlinear solver automatically “.
I don't use the PDE-Tool-GUI, but implemented my own method of line with the pdetool-command-line functions. I looked up the codes of the hyperbolic / parabolic -solver and it's subroutines, and I can nowhere find any call of a nonlinear-solver in case of solution-dependent coefficients!
The results of my calculations also strongly imply, that the nonlinearity induced by c is not treated correctly, since the results are the same as in the linear case. So if anyone can tell me, if it is possible to solve nonlinear transient PDEs with the pdetool-command-line functions, I would be very happy, since I am at my wit's end for another error source producing the unexpected results.
Best Answer