w**d
发帖数: 2334 | 1 does anybody have some experience with the 3D FEM(unstructued mesh)
code?
I am wondering how complicated it would be if I start from the very
beginning, i.e., generate the mesh(from some software) and reading
the mesh(my own code). And the equation is non-linear elliptic equation.
Is that reasonable to be a project of MS degree? Any suggestion is highly
appreciated. |
l**p
发帖数: 474 | 2 do you have any visulizartion software for display 3D unstructured meshes?
【在 w**d 的大作中提到】
: does anybody have some experience with the 3D FEM(unstructued mesh)
: code?
: I am wondering how complicated it would be if I start from the very
: beginning, i.e., generate the mesh(from some software) and reading
: the mesh(my own code). And the equation is non-linear elliptic equation.
: Is that reasonable to be a project of MS degree? Any suggestion is highly
: appreciated.
|
r*r
发帖数: 8 | 3 I feel it is tough for a MS degree.
non-linear elliptic equation in 3d will lead a big algebraic equation
system in general. Since it uses a unstructed mesh, it is hard to
perform a Newton Iteration directly. Besides, an initial guess will be
crucial.
【在 w**d 的大作中提到】
: does anybody have some experience with the 3D FEM(unstructued mesh)
: code?
: I am wondering how complicated it would be if I start from the very
: beginning, i.e., generate the mesh(from some software) and reading
: the mesh(my own code). And the equation is non-linear elliptic equation.
: Is that reasonable to be a project of MS degree? Any suggestion is highly
: appreciated.
|
t*****r
发帖数: 448 | 4 and how "huge" will your model be?
I finished my MS in china on 3D FEM programming, in the end of the
programming, I found out PC is too weak to run my program. Hard drive was not
big enough, and the CPU took it forever to ...
BTW: I used 20-node 3D cubic elements, plus 9-node shell elements and 3-node
beam elements.
【在 r*r 的大作中提到】
: I feel it is tough for a MS degree.
: non-linear elliptic equation in 3d will lead a big algebraic equation
: system in general. Since it uses a unstructed mesh, it is hard to
: perform a Newton Iteration directly. Besides, an initial guess will be
: crucial.
|
w**d
发帖数: 2334 | 5 how about if I know how to get a good initial guess and I know how to
do the Newton iteration ? How much work will there
be for reading mesh? I only wrote 2D code before. Have no idea how hard
the 3D problem will be.
【在 r*r 的大作中提到】
: I feel it is tough for a MS degree.
: non-linear elliptic equation in 3d will lead a big algebraic equation
: system in general. Since it uses a unstructed mesh, it is hard to
: perform a Newton Iteration directly. Besides, an initial guess will be
: crucial.
|
w**d
发帖数: 2334 | 6 No idea. There are some sofeware in my dept. I don't know which one could
be used to visualize 3d mesh. Is Tech plot ok?
【在 l**p 的大作中提到】
: do you have any visulizartion software for display 3D unstructured meshes?
|
w**d
发帖数: 2334 | 7 I guess I don't need to worry about the memory and cpu. My dept has some good
machines.
【在 t*****r 的大作中提到】
: and how "huge" will your model be?
: I finished my MS in china on 3D FEM programming, in the end of the
: programming, I found out PC is too weak to run my program. Hard drive was not
: big enough, and the CPU took it forever to ...
: BTW: I used 20-node 3D cubic elements, plus 9-node shell elements and 3-node
: beam elements.
|
w**d
发帖数: 2334 | 8 I guess it is relative easy to code with cubic elements.
【在 t*****r 的大作中提到】
: and how "huge" will your model be?
: I finished my MS in china on 3D FEM programming, in the end of the
: programming, I found out PC is too weak to run my program. Hard drive was not
: big enough, and the CPU took it forever to ...
: BTW: I used 20-node 3D cubic elements, plus 9-node shell elements and 3-node
: beam elements.
|
