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.

partial differential equation

Please login with a confirmed email address before reporting spam

How to specify the Neumann boundary conditions in a coefficient form partial differential equation? (The derivative of u specified by the constraint is affected by the size of the grid)


1 Reply Last Post Nov 16, 2023, 9:08 p.m. EST
Robert Koslover Certified Consultant

Please login with a confirmed email address before reporting spam

Posted: 12 months ago Nov 16, 2023, 9:08 p.m. EST
Updated: 12 months ago Nov 16, 2023, 9:18 p.m. EST

If you use a sufficiently fine mesh and/or sufficiently high order discretization then you can usually successfully/correctly compute solutions that may have strong gradients (spatially or temporally) in their derivatives. Under those circumstances, the specification of u (or its derivative) on a boundary will not introduce a mesh-dependent effect. If you see a strong mesh dependent effect on your solution, then you need to either use a finer mesh, a higher-order discretization, or both. (And in time domain, you might need to take shorter time steps.) Note: if you are able/willing to describe a more specific math or physics problem, then you may possibly receive more useful/specific advice on this forum.

p.s. I'm assuming that by "grid" you are referring to the finite element mesh, or perhaps to size of steps in time. If you meant something else, then my answer may not apply.

-------------------
Scientific Applications & Research Associates (SARA) Inc.
www.comsol.com/partners-consultants/certified-consultants/sara
If you use a *sufficiently* fine mesh and/or *sufficiently* high order discretization then you can usually successfully/correctly compute solutions that may have strong gradients (spatially or temporally) in their derivatives. Under those circumstances, the specification of u (or its derivative) on a boundary will not introduce a mesh-dependent effect. If you see a strong mesh dependent effect on your solution, then you need to either use a finer mesh, a higher-order discretization, or both. (And in time domain, you might need to take shorter time steps.) Note: if you are able/willing to describe a more specific math or physics problem, then you may possibly receive more useful/specific advice on this forum. p.s. I'm assuming that by "grid" you are referring to the finite element mesh, or perhaps to size of steps in time. If you meant something else, then my answer may not apply.

Note that while COMSOL employees may participate in the discussion forum, COMSOL® software users who are on-subscription should submit their questions via the Support Center for a more comprehensive response from the Technical Support team.