Minimum density clip message

[matt.b...@gmail.com]

Hi

I am running simple tunnel simulations - 5 MW fire and up, airspeed 3 m/s or thereabouts. I have noticed recently (using Nov 30 nightly release) a message about minimum density clipped to 0.12 kg/m^3 appearing. I have not encountered this before - has anything changed recently that might be causing the message to appear, or is there maybe an issue to delve into?

THanks

Matt

[Randy McDermott]

These clippings have always happened, but we decided to start showing warnings because of a particular case that took us a long time to diagnose.

However, as you noticed, we quickly saw that this is happening all over the place.  So, we have now implemented a new algorithm to prevent it from happening most of the time.  You should be able to run with last night's test bundle to check.  Hot off the press!

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/c01cb2f9-f154-47e8-a866-ed5769deba62n%40googlegroups.com.

[Glenn Forney]

fyi:  the windows nightly bundle didn't get built last night.  most likely because the PC where it gets built rebooted.  I will check it out.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/CAC6wax3Vk7_4EdvVYqoHCZn3J%2BANMCoLYZnP%3Dg%2BRLBLsZee2yw%40mail.gmail.com.

--

Glenn Forney

[matt.b...@gmail.com]

Thanks - just saw that it appeared now, will give it a try

[Glenn Forney]

the windows bundle is up there now .  

https://drive.google.com/drive/folders/1X-gRYGPGtcewgnNiNBuho3U8zDFVqFsC?usp=sharing

The PC had rebooted probably to install security patches.

On Friday, December 18, 2020 at 9:39:27 AM UTC-5 Glenn Forney wrote:

[matt.b...@gmail.com]

Is there an issue ID so I can read a little more on the history of the change?

[Dan]

Dear All, 

I am running a model with 4m/s wind and was getting the same error "WARNING: Minimum density, 0.120 kg/m3, clipped in mesh.." when I'm running the model through FDS (Linux). I'm using the FDS.6.7.6 May 27th version. But I don't get the error with my windows machine when I run it through Pyrosim (FDS 6.7.5). I have fewer cores with the windows machine and prefer to use Linux as I have double the number of cores there. 

Could I please know whether there is a workaround for this? 

Thanking you in advance. 

Dan

[Dan]

And just for your reference, I tried running the model with the nightly bundle's version and ended up with the same error at the same timestep. 

Thanks again. 
Dan

[Kevin McGrattan]

It is just a warning that we added in 6.7.6. This can sometimes be an indication of trouble, and this trouble would have existed in previous versions too. We decided to print it out to warn that there might be a possible numerical instability. FDS will adjust the time step to try to avoid this trouble, but the print out can be helpful for diagnostic purposes.

[Dan]

Thank you very much for your prompt response, Dr McGrattan. 

I somehow couldn't run the model at all with 6.7.6. It tried to fix the time step but stopped after 2, 3 steps with the instability error. 

But surprisingly, I could run the same model with 6.7.5. Would you suggest that the model would still have some errors within although it ran for the entire period (3600 seconds) ?

Thanks again. 

Cheers

Dan

[Kevin McGrattan]

Try this

&CLIP CLIP_DT_RESTRICTIONS_MAX=0 /

[Dan]

Thanks again for the suggestion, Dr McGrattan.  

The model seems to be running now while providing warnings for each step and is now at 446.99 s/3600s. I'm running the model with 32 cores, and it seems the model is slow compared to the model I ran with 6.7.5 version. 

Out of curiosity, could I please know whether there are any possible reasons which are leading to the warnings? I was also looking at a GitHub thread initiated by Emanuele Gissi and planning to run a separate case with some of the suggestions mentioned there. 

Thank you very much again for all the help given so far.  

Kind Regards, 
Dan

On Monday, October 11, 2021 at 5:32:36 AM UTC+11 Kevin wrote:

Try this

&CLIP CLIP_DT_RESTRICTIONS_MAX=0 /

[Dan]

Dear  Dr McGrattan, 

I tried running the "&CLIP..." as well as  CFL_MAX=0.9 CHECK_HT=T / just to see any of these would work, but unfortunately, it crashed again. 

I also tried extending the domain almost two times which didn't do anything at all. 

And interestingly, this happens only with a certain wind direction. I'm performing a parametric study these days and was looking at the same wind speed in different directions. So this minimum density error occurs when the wind is applied at 180 degrees, whereas all the other models with others run ver well. 

I would be grateful if you could give me some idea of what may be leading to this minimum density error so that I can try to work around the issue by some other means. 

Thanking you in advance. 
Dan

[Marcos Vanella]

Hi Dan, is this case an outdoor flow? If so is it using a GEOM terrain? If so, post an issue in the github tracker and upload the case.

We are seeing instability issues with level set on these, and are working on this feature.

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/37a5e65d-a3b3-4d08-aaf9-7bca95f5dbbfn%40googlegroups.com.

[Dan]

Hi Marcos, 

Thank you for your response. 

Yes, it is an outdoor flow, but not using GEOM. 

I will post the issue to the GitHub tracker now. 

Thanks again. 

Cheers

Dan