Re: [deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Dear Pasha, You’re going to have to provide more information in order to help work out where the problem is. Please run your program in debug mode and post the stack trace thats printed to the command line. Regards, Jean-Paul P.S. I sent an email to your listed email address yesterday. Did you

Re: [deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Dear all Regarding using MappingQEulerian class in parallel, I did what you said as bellow: dealii::TrilinosWrappers::MPI::BlockVector rel_solution2(partition_relevant); rel_solution2 = solution; rel_solution2 *= -1.0; MappingQEulerian q_mapping(1, dof_handler, rel_soluti

Re: [deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Pretty much exactly as I said. You first need to store the total displacement vector, but this would be the displacement from the spatial to the material configuration, accumulated by subtracting your incremental update from this global vector. You then create an Eularian map

[deal.II] Re: Computation of shape function gradients at the reference cell

Ok. Thank you. I´ll get the gradient directly from the FiniteElement. In case someone else has the same doubt, just to review: - From *FEValues*, one can get the gradient of the shape functions* with respect to the real cell coordinates*. ( https://www.dealii.org/8.4.0/doxygen/deal.II

Re: [deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Yes, exactly. How I can define FEFaceValue in reference frame when the mesh has been moved to the current configuration? On Tuesday, January 17, 2017 at 1:40:03 PM UTC+3:30, Jean-Paul Pelteret wrote: > > I must admit that its now a bit unclear to me as to what information you > do and don’t hav

[deal.II] Re: Computation of shape function gradients at the reference cell

Dear Tulio, This is the type of information that you can get directly from the chosen FiniteElement, e.g. this function provides the shape function gradient on the unit cell. You could also

Re: [deal.II] Computation of shape function gradients at the reference cell

Hi You can get them from your fe space. E.g., if you have FE_Q fe then you can do fe.shape_value(...) fe.shape_grad(..) etc. See https://www.dealii.org/developer/doxygen/deal.II/classFiniteElement.html Best praveen On Tue, Jan 17, 2017 at 4:25 PM, Tulio Ligneul < tulio.lign...@petrosoftdesi

[deal.II] Computation of shape function gradients at the reference cell

Hi, I've been trying to implement some stabilization parameters in a vector valued problem. In one of them, i need the gradients of the shape functions in relation to the reference cell coordinate system and i am not sure how to get these values from deal.II. Can anyone help me? Regards, Túlio

Re: [deal.II] Re: How to solve a linear system with two RHS and a same matrix?

I must admit that its now a bit unclear to me as to what information you do and don’t have during assemble. But to summarise what I understand to be your problem, I interpret that you have a nonlinear updated Lagrangian formulation but want to apply a fixed external load as defined on the refere

[deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Sorry, the local_rhs is available for the second global RHS during assemble process of the glbal matrix and the first global RHS. On Tuesday, January 17, 2017 at 1:26:29 PM UTC+3:30, dealii...@gmail.com wrote: > > Thank you for the answer. > > I am working on a solid mechanical model with large

[deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Thank you for the answer. I am working on a solid mechanical model with large deformations. The first RHS is the residual load vector (external load – internal load). The second RHS is the global external load vector. Since I use update Lagrangian formulations, the mesh is changed during eac

[deal.II] Re: How to solve a linear system with two RHS and a same matrix?

Dear Pasha, At first glance, it would appear that there is a further ConstraintMatrix::distribute_local_to_global function designed with this in mind. It takes in the local_matrix but d

[deal.II] How to solve a linear system with two RHS and a same matrix?

Dear all I want to solve a linear system with two RHS, a same matrix and a same constraints matrix. Since the code works in parallel, *ConstraintMatrix::distribute_local_to_global*