Page 1 of 1

Error using Calving3D: nodes in a calving column out of order

Posted: 28 Jun 2021, 20:30
by jm.munoz
Hello everyone,

I''m currently working on a 3D Flow model and I want to include the Calving3D Solver. I run the test provided on the git branch and everything goes ok, but when I try to apply it on my model, I'm having some problems. I think this is probably due to my mesh, but I'm not sure.

The thing is that I got it to work with just one timestep. In that case, a calving event takes places and after that the remeshing algorithm runs successfully. But when I try to run more than one timestep, the second time that the Calving3D solver runs produce the next error:

Code: Select all

WARNING:: Find_Calving3D: Calving front nodes in a column are out of order. This may not be a problem.
 There are            8  out of           10  nodes out of order.
 
ERROR:: Find_Calving3D: Majority of nodes in a calving column are out of order (z). This is probably due to not using MeshExtrude, or a change in the way in which MeshExtrude operates.
Like I said before, I think that this could be a mesh problem, but I am not sure because I don't really understand what is going on. Any help would be more than welcome. Thanks in advance.

I attach to this message the following files:

PlanMesh.geo: The geo file of my main mesh.
mitest_transient5.sif: The sif file that I am using
output_tr.txt: The output file of the run.

Regards,
José

Re: Error using Calving3D: nodes in a calving column out of order

Posted: 29 Jun 2021, 11:11
by samuel.cook
Hi José

Yes, the problem will be with your mesh. What that error means is, essentially, that, on the second timestep, Calving3D has run into some sort of complicated mesh geometry that it can't deal with. This may be a result of the calving events from the first timestep or, if it's trying to do more than one calving event on the second timestep, it might be to do with one of the earlier calving events in that timestep.

One thing to bear in mind is that Calving3D is (pseudo-)random, so you never get quite the same calving event twice. You could try just running the test a few times and seeing if that randomness generates a way out, but it's more likely that you might have to modify your mesh a bit. Have a look at the output from the first timestep and see if there are any obviously wonky bits of calving front that are causing the problem - the various calving solvers should output several different vtu files that you can examine to see what's going on (you may also have some output from these from the second timestep before the solver crashes, so those would definitely be worth looking at).

Looking at your mesh, if I've got the direction of flow right (bottom to top?), you seem to have a very straight calving front too (which is unlikely to be what the calving front actually looks like). Try making it a bit wigglier and a better fit to the real one and that might help - you'll have a more realistic initial geometry that way that might calve in a more sensible manner.

Hope that helps.

Samuel

Re: Error using Calving3D: nodes in a calving column out of order

Posted: 08 Jul 2021, 11:02
by jm.munoz
Hello, Samuel

First of all, thanks for your answer and my apologies for taking so long to post a reply. The truth is that I've doing some tests with my mesh trying to identify my problem. The good news is that I think I know where is it; the bad is that I am not sure why is that happening.

One of the things that I tried was putting attention at the output files of the first timestep. I realized that before the calving event, the front looks like this

https://imgur.com/XWPwYRr

but after the calving event looks like this

https://imgur.com/R7GcyNa

It seems like something is going wrong with the extrusion after the calving event. When I check the 2D foot print that is generated after the event I don't find any problem. But when I look at the extruded image, I can see that there are some issues. And, of course, with that extruded front, it is obvious now why the simulation fails.

My guess is that I could be doing something wrong with the free surface evolution (I'm trying to identify why that downpeak occurs) or maybe when I calculate the depth or the elevation. But I'm not sure yet.

Thank you for everything. I'll keep working on this and, like I said before, any help would be more than welcome.

Regards,
José