Skip to content

Conversation

@finsberg
Copy link

@finsberg finsberg commented Oct 20, 2024

Adding comments to book

  • The finite element method
  • An introduction to the Unified Form Language (UFL)
  • The FEniCSx Form Compiler (FFCx)
  • DOLFINx Deep dive
  • Meshes in DOLFINx
  • Applications

# 1) The triangulation covers $\Omega$: $\cup_{j=1}^{M}K_j=\bar{\Omega}$
# 2) No overlapping polyons: $\mathrm{int} K_i \cap \mathrm{int} K_j=\emptyset$ for $i\neq j$.
# 3) No vertex lines in the interior of a facet or edge of another element
# "No vertex lies"?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed

# "No vertex lies"?
#
# We will call our polygonal domain $\mathcal{K}={K_j}_{j=1}^{M}$.
# Maybe add some brackets around K_j?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed

# using the mapping $F_j:K_{ref}\mapsto K_j$.
#
# We define the Jacobian of this mapping as $\mathbf{J_j}$.
# I would add a definition of the Jacobian here
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

You mean writing out the complete definition with basis functions and derivatives?

Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I do have it as an exercise:)

basix.CellType.triangle, 8, basix.LatticeType.gll, exterior=False, method=basix.LatticeSimplexMethod.warp
)
# -
# Is this code snippet suppose to produce some output?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No, not sure why I made it a dropdown. will revert

# as the linear combination of the three basis functions on the reference element $X$.
#
# $$x = F_j(X)= \sum_{i=0}^3 \mathbf{p}_i \phi_i(X).$$
# I guess i = 0 to 2?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Fixed first it here.
Can we discuss point 2 when Im back?

# ```{admonition} Can we use a similar kind of mapping on a quadrilateral/tetrahedral/hexahedral element?
# :class: dropdown tip
# Yes, for any polytope that we can describe with a Lagrange finite element, we can define a mapping from a reference to a physical element.
# I had to google the word polytope, maybe add a definition here or add it to the glossary
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I've simplified the definition.

# As we have already seen, we can describe any cell in our subdivided domain with a mapping from the reference element.
# However, as we want to integrate over each element individually, we need to map the basis functions to and from the reference element.
# We call this map: $(\mathcal{F}_j\phi)(x)$.
# Can we add domain and codomain to the map?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Would need to think about this one, as it should hold for an arbitrary basis function (in any space),

# $$
# (\mathcal{F}_j\phi)(x)= \phi(F_j^{-1}(x))=\phi(X).
# $$
# I didn't really ge this. What does `j` represent?
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

There is a unique mapping for each cell in your physical mesh. It ties back to the definition of $F_j$ and $J_j$ earlier in the section.

# :class: dropdown, tip
# Their tangential component is only non-zero on the edge they are associated with, but the normal component can be non-zero there too.
# This implies that they are normal to the other edges.
# I would maybe add some equations here to make it more clear
Copy link
Owner

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's not easy to see from equations (at least not for a simple-minded person like myself).

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants