dolfin team mailing list archive

[Pietro Maximoff <question95601@xxxxxxxxxxxxxxxxxxxxx>]

New question #95601 on DOLFIN:
https://answers.launchpad.net/dolfin/+question/95601

Hi all

I am trying to compute the error given a known solution. (C++). This is what I did.

I created two .ufl files; the first to genrate a functional:

Error.ufl
=======
element = FiniteElement("Lagrange", "triangle", 3)
v = Coefficient(element)
w = Coefficient(element)
M= ((w - v) * (w - v)) * dx

and the second to represent a higher order function space onto which the exact solution will be interpolated:

HigherOrder.ufl
=============
element = FiniteElement("Lagrange", "triangle", 3) 

Then in main.cpp, I had

Function u_exact_high(V_high_order);
u_exact_high.interpolate(u_exact);
Error::functional M(mesh, u_exact_high, u);
double error = std::sqrt(assemble(M));

1. is this the correct way to do it?

2. Is there a function or class that I can use? [I've searched but couldn't find one].

3. Do I follow the same procedure to calculate the H1 seminorm and other norms?

Best wishes

Pietro


-- 
You received this question notification because you are a member of
DOLFIN Team, which is an answer contact for DOLFIN.