SMS is not responding

[amy]

Hi,

I want to generate a mesh in a very large domain, based on the bathymetry (SMS 10.0). I am following the ADCIRC tutorial, creating a shallow wavelength function from the bathymetric data. I used the original wavelength function alone. When I try to create the mesh, the SMS doesn’t respond, saying “Paving interior (0%)” for ever. What is the problem? Can it be related with the large extension of the domain?

Thank you!

Amy

[Rusty Jones]

It may be that you are trying to build a mesh with too many elements. I would multiply your size function by 100 using the data calculator and see if that helps. If it builds a mesh, see if it provides adequate resolution and scale the size function appropriately if not.

[amy]

Thank you for your prompt reply.

I continue having the same problem, even multiplying the size function by 1000. Is this method (creating shallow wavelength function from the bathymetric data) not appropriate for large meshes, e.g. 150,000 nodes and 300,000 elements? Which is the most adequate method?

Amy

[Rusty Jones]

By looking at the contours of your size function, you can get some idea of how many elements you will be creating.

The method used in the ADCIRC tutorial tries to create size functions that have better resolution in the area of interest with decreasing resolution as you get to deeper water. I would recommend looking at the LTEA tutorial and using these tools to generate your mesh. The LTEA tools were developed to guide mesh generation using simplified runs of the model itself and uses error estimates from these results.

You can download this tutorial from the link below.

http://www.aquaveo.com/f.php5?s=sms&v=...amp;p=tutorials

If you continue to run into problems building meshes for your domain and it doesn't seem to be a problem with trying to create elements that are too small, please contact technical support (support@aquaveo.com) and get them your files so we can try to figure out the issue.

[amy]

Hi,

I have sent the files to technical support, using gmail. I hope the e-mail is not considered spam.

Thank you.

[John Howlett]

Amy,

I am working on the issue with the files you sent. I found one problem that was causing an infinite loop which is why the meshing was stuck at 0% on your machine. I will keep you up to date on my progress.

- John

[John Howlett]

Amy,

I have found a solution to the problem. The fix will be available in SMS 10.1 Beta which will be released in a couple of weeks.

- John

[amy]

Amy,

I have found a solution to the problem. The fix will be available in SMS 10.1 Beta which will be released in a couple of weeks.

- John

Hi John,

I'm sorry for this late reply. Thank you very for your work with my files. The mesh looks very nice, with a good amount of elements and nodes, however I have a few questions about it:

1- In the original Map Data I sent (CAD), I divided the coast in different feature arcs, because I want different triangles size: 250 m in the Portuguese coast, and 1000 m in the French, Spanish and African coasts. The ocean should have triangles no larger than 50,000 m. I verified that, in the mesh you sent me, there are 2 feature arcs only, and most of the triangles size is around 2000 m, what is much larger than what I wanted, specially for the Portuguese coast. Is it not possible to impose the triangles size at the coast when I build the mesh based on bathymetry (using the Shallow_Wavelength as size function)?

2- I can’t access to the polygon attributes to see which conditions you selected. Was the mesh built based on the data "Shallow Wavelength x 10 - 0.5 limit, 2000 min"? What do these parameters mean?

3- Can I then have access to SMS 10.1 Beta to reproduce your results?

Thank you again.

Regards,

Amy

[John Howlett]

Amy,

1. I attempted to generate a mesh using the parameters in your original file, but the resulting mesh was too large to be generated using the 32 bit version of SMS. SMS used up all the memory allowed for a 32 bit program during the mesh generation and had to exit without completing the mesh. If you have a 64 bit operating system, you will probably want to try the 64 bit version of SMS 10.1 beta to see if you can generate the larger mesh.

2. I scaled the shallow wavelength function up by 10. I then used an area change limit of 0.5. When I generated the mesh I used a minimum truncation limit on the size data set of 2000.

3. We hope to post SMS 10.1 Beta later today - watch for an announcement on the forum.

- John