Data structure of mesh for DG methods

1 month 1 week ago #3658 by Younghigh
Hi, all

When DG methods are employed in the Stokes equation, we tend to use some codes like:
mesh = Mesh(geo.GenerateMesh(maxh=0.1, quad_dominated=False))
V = VectorL2(mesh, order=3, dgjumps=True)
n = specialcf.normal(mesh.dim)
u, v = V.TnT()
mean_grad_u = 0.5 * (grad(u)+grad(u.Other()))
jump_v = v-v.Other()
a = BilinearForm(V)
a += InnerProduct(mean_grad_u.trans*n, jump_v) * dx(skeleton=True)

My question is: for a fixed edge or facet, how do we know its inner element or outer element because we use
jump_v = v-v.Other()
, and further the direction of its unit normal? Are the both consistent?


Best,
Di Yang

Please Log in or Create an account to join the conversation.

1 month 6 days ago #3659 by christopher
Hi,
n and .Other are consistent.

Best
Christopher
The following user(s) said Thank You: Younghigh

Please Log in or Create an account to join the conversation.

1 month 6 days ago #3661 by Younghigh
How to make uniform triangle mesh on the square [-0.5,1.5]*[0,2], not regular domain [0,1]*[0,1]?

Please Log in or Create an account to join the conversation.

1 month 6 days ago - 1 month 6 days ago #3669 by hvwahl
Hi Di Yang,

you can make you own python function based on MakeStructured2DMesh in python/meshes.py. You should be able to get the desired effect by just scaling the mesh points appropriately (line 111 of the file).

Best,
Henry

Please Log in or Create an account to join the conversation.

1 month 6 days ago #3670 by schruste
You don't need to implement the method yourself, see ngsolve.org/forum/ngspy-forum/1360-how-t...ain-0-5-1-5-0-2#3663 .

Best,
Christoph

Please Log in or Create an account to join the conversation.

© 2019 Netgen/NGSolve