Thunderhead Engineering Forum

Please login or register.

Login with username, password and session length
Advanced search  


Forum moved to

Author Topic: Numerical Instability Errors??  (Read 454 times)


  • New Member
  • *
  • Posts: 1
    • View Profile
Numerical Instability Errors??
« on: May 13, 2021, 03:59:55 pm »

Hello all,
We've been trying to run a file simulating a fire in an atrium. Since the atrium is really big, when we first tried to run the test it gave us an expected run time of over 135 days (not sure why). We tried to limit the computational cost by forcing a time step of 1s (previously FDS chose DT=0.112 s or so), but now we're getting this error at 203.13 s:

ERORR: Numerical Instability - FDS stopped (CHID: Hepterium_DF2.fds)

Does anyone know how to fix it? Also, any tips on how we can avoid an unreasonable run time?
Attached is our code.



  • New Member
  • *
  • Posts: 7
    • View Profile
Re: Numerical Instability Errors??
« Reply #1 on: May 15, 2021, 11:06:05 pm »

Hi Aviv,

Using a Long time step may put the results you get into serious doubt.
as for the numerical instability, if the atrium is not open to the ambient environment, you need to configure a zone inside the atrium.