Dear Prof. Bangerth and Prof.Wick Thank you so much for your reply and suggestion as always, I really appreciate your time. As Prof. Wick has mentioned we have checked the code and the problem was related to the nonlinear part.
Best Regards Shahin On Sat, Dec 18, 2021 at 2:02 PM Thomas WICK <[email protected]> wrote: > Dear Shahin, dear Wolfgang, > > I had a closer look into the technical details. The implementation with > regard to deal.II functionalities is fine. Only for > the nonlinear AFC method, the algorithm needs to be worked out. > > Best regards, > > Thomas > > -- > --- > ++++--------------------------------------------++++ > Prof. Dr. Thomas Wick > Leibniz Universität Hannover (LUH) > Institut für Angewandte Mathematik (IfAM) > Arbeitsgruppe Wissenschaftliches Rechnen (GWR) > > Welfengarten 1 > 30167 Hannover, Germany > > Tel.: +49 511 762 3360 > Email: [email protected] > www: https://ifam.uni-hannover.de/wick > www: https://thomaswick.org > ++++--------------------------------------------++++ > --- > -- > > > > Am 16.12.21 um 21:04 schrieb Wolfgang Bangerth: > > On 12/10/21 03:23, Shahin Heydari wrote: > >> > >> raw_flux_matrix(i,j) = > >> mass_matrix(i,j) * (solution_ul(i) - solution_ul(j)) > >> - > >> mass_matrix(i,j) * (old_solution_u(i) - old_solution_u(j)) > >> - > >> time_step * theta * matrix_D(i,j) * (solution_ul(i) - > >> solution_ul(j)) > >> - > >> time_step * (1-theta) * matrix_D(i,j) * (old_solution_u(i) - > >> old_solution_u(j)); > >> but the obtained solution is very smeared and the solution does not > >> do one complete rotation as it supposed to do. I think the problem > >> might be that I have made a mistake in computing mij, dij and ul in > >> the global indices. > > > > I suspect that you won't get anyone here who has the interest, > > knowledge, and time to go through your code to check out what is the > > problem. So it comes down to how you can debug the problem and my > > suggestion always is to pick simple cases: For example, choose an > > initial condition that is a step function and for which you can > > compute everything on a piece of paper, run one time step, and compare > > what you get with what you expected. > > > > In any case, the code snippet you show that loops through the matrix > > entries is at least conceptually correct, though I will point out that > > this... > > > > for(auto jt = sparsity_pattern.begin(i); jt != > > sparsity_pattern.end(i); jt++); > > const auto j = jt->column(); > > > > ...looks funny to me: The semicolon at the end of the for(...); line > > means that the loop has an empty loop body. But then the variable 'jt' > > would not actually live outside the loop body, and I would have > > expected that you get an error message in the second line. In any > > case, the following line that computes raw_flux_matrix is not within > > the loop body because there are no curly braces, and so does not seem > > to be executed for every 'j'. That doesn't seem right, but I infer > > that the code does not compile that way, and so it must not be exactly > > what you have in your code. > > > > > >> As I know, when we are inside the cell and want to access the value > >> of global matrix or vector we can compute the value at the quadrature > >> point for example as follow: > >> std::vector<double> old_solution_u_values(n_q_points); > >> fe_values.get_function_values(old_solution_u , old_solution_u_values); > > > > I don't get that comment. I thought you were just working with nodal > > values of the solution for the AFC. Why do you need the values at > > quadrature points? > > > > > >> Now I am just wondering how I can access the entry of the global > >> matrix and vector in this case. > > > > And I don't understand this question because you have already figured > > this out. > > > > Best > > W. > > > > -- > The deal.II project is located at http://www.dealii.org/ > For mailing list/forum options, see > https://groups.google.com/d/forum/dealii?hl=en > --- > You received this message because you are subscribed to the Google Groups > "deal.II User Group" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > To view this discussion on the web visit > https://groups.google.com/d/msgid/dealii/7bbd3d1a-829e-486d-a9c2-02134ba29b18%40gmail.com > . > -- The deal.II project is located at http://www.dealii.org/ For mailing list/forum options, see https://groups.google.com/d/forum/dealii?hl=en --- You received this message because you are subscribed to the Google Groups "deal.II User Group" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To view this discussion on the web visit https://groups.google.com/d/msgid/dealii/CALUbCFwdWmzKjs8N_FUS8U41mZB-COx2eNHvarvyn_aGoBM88g%40mail.gmail.com.
