Thunderhead Engineering Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Forum moved to https://forum.thunderheadeng.com

Author Topic: WARNING TMP_G_N < TMP_D_N  (Read 208 times)

Csaba_90

  • New Member
  • *
  • Posts: 1
    • View Profile
WARNING TMP_G_N < TMP_D_N
« on: July 27, 2020, 12:29:47 am »

I am using Pyrosim v2020.20527. When running a simulation with ESFR sprinklers (or with normal sprinklers) protecting a multifunctional sport hall the simualtion stops (depending on the activation of sprinklers) with the following error message:
WARNING Delta TMP_G. Mesh: 23 Particle: 1
WARNING TMP_G_N < TMP_D_N. Mesh:23 Particle: 1
(I'm looking for the problem source and find that the message means that during the temperature update of a particular grid cell in Mesh 23, the droplet whose index is 1 caused the cell gas temperature (TMP_G) to change too rapidly.)
This error message string is repeated several times for differt particle IDs. The common event the calculation stops for a while (with this error message) and it continues. At random event when the error message occur the calculation stays (at that moment when the message apeear) for days.

I couldn't find the correct solution so far. However i tried the following according to the fds user guide:
- on the MISC. line set PARTICLE_CFL= TRUE (the result was the calculation extremely slowed about 1100 sec the problem occured)
- changed the PARTICLES_PER_SECOND on the prop line from 5000 to 20000 (same result)

This problem didnt occured on the previous versions. Running the case with MPI. Does anyone has the same problem or the correct solution?

Thank you.
« Last Edit: July 27, 2020, 12:55:04 am by Csaba_90 »
Logged

Federico L

  • New Member
  • *
  • Posts: 1
    • View Profile
Re: WARNING TMP_G_N < TMP_D_N
« Reply #1 on: August 15, 2020, 03:51:27 pm »

Hello, I have the same problem when running a simulation with ESFR sprinklers, sprinklers are connected to a dryline with a retard of 60 seconds, when water is discharged begins the problems with particle 1 and finish on particle 5. I´ve changed the PARTICLES_PER_SECOND from 5000 to 20000 and the Jet Stream Offset from 0.05 to 0.1, wich is the mesh length.
After running the simulation again I have a lot of warning messages from particle 1 to particle 1307, just the same like you.
Did you fix this problem?
Regards!
Logged

Bryan Klein

  • Thunderhead
  • *****
  • Posts: 118
    • View Profile
    • Thunderhead Eng. Site
Re: WARNING TMP_G_N < TMP_D_N
« Reply #2 on: August 17, 2020, 01:48:52 pm »

Hello,

"FDS sets a kill radius that is 0.5 % of the mass of the MINIMUM_DIAMETER particle. In your case that is about 3.4 um. This is why you are seeing 6 um in Smokeview.

Within the evaporation routine FDS first tries to evaporate the drop using the current time step. It then does some checks to see if the equilibrium condition is overshot or if the local gas temperature changes too much. If this is the case it tries again at half the timestep. This process is repeated until the droplet passes the checks or the timestep drops too much. If the timestep drops too much a warning is written. In your case the warning is saying that even at a very small sub timestep the gas temperature is still seeing large changes."

From: https://groups.google.com/g/fds-smv/c/H-QjqnueyAs/m/gPuIPTN2AAAJ

Best Regards,
-Bryan Klein
Logged