Discussion Closed This discussion was created more than 6 months ago and has been closed. To start a new discussion with a link back to this one, click here.
question regarding meshing
Posted Jun 10, 2011, 1:37 p.m. EDT 2 Replies
Please login with a confirmed email address before reporting spam
Hi guys,
I'm new to simulation. And I got a question for a while: does the number of meshing elements significantly affect the simulation result? The reason I ask this is because that when the number of meshing elements is below a certain value, the simulation result doesn't make physical sense. However, once the number of meshing elements is beyond that certain value, the simulation result looks good. I don't know how to explain this. Any input is appreciated.
Cheers
I'm new to simulation. And I got a question for a while: does the number of meshing elements significantly affect the simulation result? The reason I ask this is because that when the number of meshing elements is below a certain value, the simulation result doesn't make physical sense. However, once the number of meshing elements is beyond that certain value, the simulation result looks good. I don't know how to explain this. Any input is appreciated.
Cheers
2 Replies Last Post Jun 10, 2011, 3:29 p.m. EDT
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
There are tons of Finite Element Analysis books out there that explain this in great detail, but in a nutshell it boils down to this:
Say I give you a wildly varying function f(x) defined over a segment [a;b] and I ask you to approximate it with a single straight line. The approximation won't be too good, right? Now, I ask you to approximate it with two straight lines, one over [a; (a+b)/2] and one over [(a+b)/2; b]. You'll be able to get a better approximation, but it still won't be very good, right? Now, continue this process: you're allowed 3 segments, then 4, then 5, etc. Your approximation will get better and better each time. With FEA, it's basically the same thing: the more degrees of freedom you have in a mesh, the closer you can approximate the actual solution of the PDEs that you're solving.
Say I give you a wildly varying function f(x) defined over a segment [a;b] and I ask you to approximate it with a single straight line. The approximation won't be too good, right? Now, I ask you to approximate it with two straight lines, one over [a; (a+b)/2] and one over [(a+b)/2; b]. You'll be able to get a better approximation, but it still won't be very good, right? Now, continue this process: you're allowed 3 segments, then 4, then 5, etc. Your approximation will get better and better each time. With FEA, it's basically the same thing: the more degrees of freedom you have in a mesh, the closer you can approximate the actual solution of the PDEs that you're solving.
Please login with a confirmed email address before reporting spam
Posted:
1 decade ago
There are tons of Finite Element Analysis books out there that explain this in great detail, but in a nutshell it boils down to this:
Say I give you a wildly varying function f(x) defined over a segment [a;b] and I ask you to approximate it with a single straight line. The approximation won't be too good, right? Now, I ask you to approximate it with two straight lines, one over [a; (a+b)/2] and one over [(a+b)/2; b]. You'll be able to get a better approximation, but it still won't be very good, right? Now, continue this process: you're allowed 3 segments, then 4, then 5, etc. Your approximation will get better and better each time. With FEA, it's basically the same thing: the more degrees of freedom you have in a mesh, the closer you can approximate the actual solution of the PDEs that you're solving.
Thanks so much.