gridap/Gridap.jl

I have a problem with using the Grid function

ljhraid opened this issue · 1 comments

ljhraid commented

Hi, I am very new to Gridap.

I just want to solve a 2D Poisson equation with triangle elements.
But I have a problem with using the Grid function.

Here is my code.

using Gridap
using Gridap.Geometry
using Gridap.ReferenceFEs

xmin = 0.0
xmax = 1.0
ymin = 0.0
ymax = 1.0
nel  = (10, 10) # the number of elements of each sides

domain = (xmin, xmax, ymin, ymax)

model_car = CartesianDiscreteModel(domain, nel)

Grid(LagrangianRefFE(TRI), model_car)

The error message says

MethodError: no method matching Grid(::GenericLagrangianRefFE{GradConformity, 2}, ::CartesianDiscreteModel{2, Float64, typeof(identity)})

But I am quite sure it matches the method "Grid(::Type{ReferenceFE{d}}, ::CartesianDiscreteModel{d})".

supertype(supertype(typeof(LagrangianRefFE(TRI))))
# The output: ReferenceFE{2}

Do I misunderstand something?
Could anyone help me to solve this problem?

Just in case when you are curious about the reason why I use the Grid function, I used it to insert it into DiscreteModel function.

JordiManyer commented

@ljhraid To convert to triangles, use the method simplexify(model_car). Creating a model from scratch is quite involved, and there is no way you can infer the connectivity for TRI elements from a CartesianDiscreteModel like you are trying to do....

Also, we have a gitter channel for this kind of inquiries. Next time, you should try to post there to avoid clutter in the repo issue list, which should be reserved for bugs and/or discussions.

  • 👍1