[Gmsh] 2D mesh error message

Yao-Cheng Tien yaocheng2010 at gmail.com
Sat Jan 17 02:37:07 CET 2015


Hi Mikhail,

It's really interesting. I followed your instructions to run my testcases
on Linux machine, and yes, they all work.
I will use Linux instead but just let you know that previous test was run
on Windows. So there might be some issues for the Windows' version.

Thanks for your instant support again!

Best,
YC

On Fri, Jan 16, 2015 at 7:02 AM, Mikhail Artemyev <
artemiev.mikhail at gmail.com> wrote:

>  Hi Yao-Cheng,
>
> I was able to build the meshes for all these 3 files and RandomFactor=1e-6:
> 10K on Linux Mint 64bit, gmsh 2.8.5
> 50K on Linux Mint 64bit, gmsh 2.8.5
> 100K on Mac OS X 64bit, gmsh 2.8.4 (but I think it will work on Linux as
> well)
>
> Best,
> Mikhail
>
>
>
> On 01/16/2015 01:11 AM, Yao-Cheng Tien wrote:
>
> Hi Mikhail,
>
>  I have attached three more testcases. I have tried to embed 10k, 50k,
> and 100k lines(with length 1) to a surface separately. The element size is
> 1.
>
>  The issue is that the 2d mesh(meshAdapt) works for 10k and 50k after
> setting the Mesh.RandomFactor to 1e-6. However, the 100k one fails even
> when I increase the Mesh.RandomFactor to 1e-4. The error message is
> "Identical points in triangulation: increase element size or
> Mesh.RandomFactor"
>
>  My questions are:
> 1. Is there any limit about the number of embedded lines? 100k embedded
> lines are too much?
> 2. Any setting I missed to make it work?
>
>  Thank you!
>
>  Best,
> YC
>
> On Thu, Jan 15, 2015 at 8:59 AM, Yao-Cheng <yaocheng2010 at gmail.com> wrote:
>
>> Hi Mikhaul,
>>
>> My mistake. I'm trying different settings to fix that actually. The
>> original geo file indeed has lc=1 but still fails. I have also tried lc=0.1
>> 0.5.
>>
>> I know it's ok if I have only 1 line or even 10k lines. But in my
>> situation, I probably need to have up to 1 million lines. Any suggestion?
>>
>> Thank you.
>>
>> Best,
>> YC
>> > On Jan 15, 2015, at 6:26 AM, Mikhail Artemyev <
>> artemiev.mikhail at gmail.com> wrote:
>> >
>> > Hi,
>> >
>> > Your lines have length 1, while the characteristic length of the mesh
>> you defined was 10.
>> > What happens if you set lc = 1?
>> > With only one line embedded it works fine for me (and
>> RandomFactor=1e-8), but I didn't try for all 50K of them.
>> >
>> > Best,
>> > Mikhail
>> >
>> >
>> >> On 01/15/2015 02:29 AM, gmsh-request at ace20.montefiore.ulg.ac.be wrote:
>> >> Hi,
>> >>
>> >> I have encountered the issue "*Identical points in triangulation:
>> increase
>> >> element size or Mesh.RandomFactor*" in my 2D geometry file. After
>> looking
>>  >> into previous support mails, I changed the Mesh.RandomFactor to
>> 1e-8, 1e-7
>> >> until 1e-2. However, it still fails. Could you tell me what other
>> option I
>> >> should try since I am pretty new to Gmsh?
>> >>
>> >> My geometry is simple. I have a surface with size 1100 by 210. Also,
>> there
>> >> are 50K lines with length 1 embedded in that surface.
>> >>
>> >> Thanks for your help in advance!
>> >>
>> >> Best,
>> >> YC
>> >
>>
>
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.geuz.org/pipermail/gmsh/attachments/20150116/ff89797e/attachment.html>


More information about the gmsh mailing list