FlexPDE3 mesh generator has "floating... Log Out | Topics | Search
Moderators | Register | Edit Profile

FlexPDE User's Forum » User Postings » FlexPDE3 mesh generator has "floating point error" « Previous Next »

Author Message
Top of pagePrevious messageNext messageBottom of page Link to this message

Jared Barber (jared_barber)
Member
Username: jared_barber

Post Number: 4
Registered: 01-2007
Posted on Sunday, May 27, 2007 - 11:15 pm:   

Hey,

I have what I feel is a relatively simple domain for which I want to solve a set of equations on. The equations are not the problem as one can see from the attached the pde-file which has no equations associated with it, just the domain. In this particular case, I run FlexPDE3 and receive

Floating-Point Divide by Zero
---Called from resolve_features_0
---Called from dominpde_build_features_0
---Called from mesher2_freshmesh_0
---Called from control_do_gridgen_0
---Called from resolve_features_0

This error occurs for the attached pde-file (which actually has no pde in it). It occurs in FlexPDE Version 3.11. It does not occur in FlexPDE 5.0.12. I was hoping there was some simple thing I could do (for instance, I wonder if the problem lies in the "fillet" that I'm using...maybe I should write out the arc rather than using "fillet") to get the mesh generator to somehow take this domain and generate a mesh on it.

I note that if I take out the fillets (making the domain all boxy) that FlexPDE3 generates a mesh without error. It would be desirable not to have to resort to corners (because those corners do not exist in the system I am considering).

Alright, thanks

Jared

application/octet-stream
Test1.pde (2.0 k)
Top of pagePrevious messageNext messageBottom of page Link to this message

Robert G. Nelson (rgnelson)
Moderator
Username: rgnelson

Post Number: 861
Registered: 06-2003
Posted on Monday, May 28, 2007 - 03:13 pm:   

It appears that the point at xc1[8] lies exactly on the boundary that would be there if the corner were not filleted, and that in version 3 this coincidence is treated before the fillet. This causes the phantom line to be split, and generates a weird boundary that fails to disappear when the fillet is applied.

Apparently, Versions 4 and 5 treat fillets first, and do not have this problem.

The problem will probably go away if you use an ARC for this inside corner, since in that case there will be no apparent continuation of the boundary to the interior square corner.
Top of pagePrevious messageNext messageBottom of page Link to this message

Jared Barber (jared_barber)
Member
Username: jared_barber

Post Number: 5
Registered: 01-2007
Posted on Monday, May 28, 2007 - 05:11 pm:   

Thanks

Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action:

Topics | Last Day | Last Week | Tree View | Search | Help/Instructions | Program Credits Administration