Hi all, I have a math weak form as shown in the following figure and I implemented it in fenics. It gave a result which has big different from that given in the literature, so I’m not sure if the implementation is correct. The main part of my dolfinx implementation is: g = Function(W) g.in…

I put the ferrite post in the error position carelessly and now the result looks similar to that in the literature. Now I have an another question that I have a rectangle domain with a circle domain inside. The tags of the subdomains inside and outside the circle are 6 and 5 respectively. Then I wan…

dokken, thanks for your help. I have seen this post before and test boundary integration with “assemble_scalar(inner(dot(grad(u)(’-’), n(’-’)), v(’-’)) * dS(7) +Constant(mesh, 0) * dx(subdomain_data=mf_triangle))” and it gives an error “TypeError: object of type ‘int’ has no len()”

You Need to supply a minimal working example of this behavior for anyone to help you.

dokken, it’s a bit trouble to create a small example and this post does not allow upload the xmdf files. So I paste the error output instead. --------------------------------------------------------------------------- TypeError Traceback (most recent call last) <ipyt…

My first question is then: is a supposed to be a scalar? As you use both u and v, to me it Seems likely that this should be assemble_matrix. Anyhow, you can still post the full code (using for instance a built in mesh) to illustrate what you would like to do.

Here’s my code and there is no problem to solve it. I tested assemble_scalar(inner(dot(grad(u)(’+’), n(’+’)), v(’+’)) * dS(7)) and it can give a value. While I calculated assemble_scalar(inner(dot(grad(u)(’+’), n(’+’)), v(’+’)) * dS(7) + Constant(mesh, 0)*dx(domain=mesh, subdomain_data=mf_triangle)…

That is because a is not a scalar value. a contains a test and a trial function, and you should use assemble_matrix(a). However, then the additional term Also has to be bilinear, like: a= inner(dot(grad(u)(’+’), n(’+’)), v(’+’)) * dS(7) + inner(u,v)*Constant(mesh, 0)*dx(domain=mesh,subdomain_data=m…

Then why I can calculate assemble_scalar(inner(dot(grad(u)(’+’), n(’+’)), v(’+’)) * dS(7))?

Thats a Good question. But as I said. It does not make sense as u is a trial function and v a test function.

Weak form implementation

dokken March 2, 2020, 11:38pm 3

Hi, see the following post Integrating over an interior surface, which summarizes how to choose the direction of the plus minus restriction.

Topic		Replies	Views
Integrating over an interior surface variational formulation	23	8099	October 18, 2024
Strange behaviour of assemble variational formulation	11	706	October 28, 2020
Consistent side of interior boundaries with dolfinx dolfinx	5	181	October 4, 2024
Integration over subdomain with ufl.grad General	2	65	June 23, 2025
Integration Measure over small dimension General	9	114	October 2, 2025

Weak form implementation

Related topics