search

firemodels / fds

Fire Dynamics Simulator
https://pages.nist.gov/fds-smv/
Other
650 stars 618 forks source link

Cases include 'NEAR_WALL_EDDY_VISCOSITY' exhibit inconsistent output after modifying T_END #12373

Closed Bravnewor closed 8 months ago

Bravnewor commented 8 months ago

With consistent Step Size, there are differences between the new test case generated after modifying T_END to 80.920 and the original test case. Specifically, during the third iteration in the out file, there are differences in the Maximum Pressure Error between the two cases:

Step Size: 0.372E-01 s, Total Time: 47.51 s Pressure Iterations: 1 Maximum Velocity Error: 0.10E-02 on Mesh 1 at (11,12,30) Maximum Pressure Error: 0.48E-05 on Mesh 1 at (11,12,30)

Max CFL number: 0.12E-02 at (13,34,22) Max divergence: 0.14E-05 at (27,21,2) Min divergence: -0.13E-05 at (15,27,34) Max VN number: 0.90E+00 at (13,18,31)

This is inconsistent with the modified case:

Step Size: 0.372E-01 s, Total Time: 47.51 s Pressure Iterations: 1 Maximum Velocity Error: 0.10E-02 on Mesh 1 at (11,12,30) Maximum Pressure Error: 0.49E-05 on Mesh 1 at (11,12,30)

Max CFL number: 0.12E-02 at (13,34,22) Max divergence: 0.14E-05 at (27,21,2) Min divergence: -0.13E-05 at (15,27,34) Max VN number: 0.90E+00 at (13,18,31)

This issue has been replicated on other computers, all running FDS 6.8.0 version on Windows 11. It is suspected that this discrepancy may be related to the SURF line parameter 'NEAR_WALL_EDDY_VISCOSITY'. I have attached a scaled-down version of the input cases for reference. InconsistentCase1.txt InconsistentCase2.txt

rmcdermo commented 8 months ago

Thanks for the well-written report and simple input files---this is exemplary.

It looks to me like the time step is not precisely the same even though it looks that way in the .out file. In order to make sure they are identical, we can add DT=0.372E-01, LOCK_TIME_STEP=T to the TIME line. When I do this, the velocity and pressure errors look the same at T=47.51 s with latest source on the "_dv" version. See if this works for you. If not, then we will need to dig into various things like release vs. dv, operating system (I'm on macOS), etc.

Case 1:

       Time Step       3   January 15, 2024  08:40:22
       Step Size:    0.372E-01 s, Total Time:      47.51 s
       Pressure Iterations: 1
       Maximum Velocity Error:  0.83E-03 on Mesh 1 at (10,19,26)
       Maximum Pressure Error:  0.37E-05 on Mesh 1 at (10,19,27)
       ---------------------------------------------------------------
       Max CFL number:  0.12E-02 at (30,27,33)
       Max divergence:  0.14E-05 at (24,38,2)
       Min divergence: -0.14E-05 at (24,9,34)
       Max VN number:   0.90E+00 at (13,13,31)

Case 2:

       Time Step       3   January 15, 2024  08:39:11
       Step Size:    0.372E-01 s, Total Time:      47.51 s
       Pressure Iterations: 1
       Maximum Velocity Error:  0.83E-03 on Mesh 1 at (10,19,26)
       Maximum Pressure Error:  0.37E-05 on Mesh 1 at (10,19,27)
       ---------------------------------------------------------------
       Max CFL number:  0.12E-02 at (30,27,33)
       Max divergence:  0.14E-05 at (24,38,2)
       Min divergence: -0.14E-05 at (24,9,34)
       Max VN number:   0.90E+00 at (13,13,31)
Bravnewor commented 8 months ago

Thanks for the well-written report and simple input files---this is exemplary.

It looks to me like the time step is not precisely the same even though it looks that way in the .out file. In order to make sure they are identical, we can add DT=0.372E-01, LOCK_TIME_STEP=T to the TIME line. When I do this, the velocity and pressure errors look the same at T=47.51 s with latest source on the "_dv" version. See if this works for you. If not, then we will need to dig into various things like release vs. dv, operating system (I'm on macOS), etc.

Case 1:

       Time Step       3   January 15, 2024  08:40:22
       Step Size:    0.372E-01 s, Total Time:      47.51 s
       Pressure Iterations: 1
       Maximum Velocity Error:  0.83E-03 on Mesh 1 at (10,19,26)
       Maximum Pressure Error:  0.37E-05 on Mesh 1 at (10,19,27)
       ---------------------------------------------------------------
       Max CFL number:  0.12E-02 at (30,27,33)
       Max divergence:  0.14E-05 at (24,38,2)
       Min divergence: -0.14E-05 at (24,9,34)
       Max VN number:   0.90E+00 at (13,13,31)

Case 2:

       Time Step       3   January 15, 2024  08:39:11
       Step Size:    0.372E-01 s, Total Time:      47.51 s
       Pressure Iterations: 1
       Maximum Velocity Error:  0.83E-03 on Mesh 1 at (10,19,26)
       Maximum Pressure Error:  0.37E-05 on Mesh 1 at (10,19,27)
       ---------------------------------------------------------------
       Max CFL number:  0.12E-02 at (30,27,33)
       Max divergence:  0.14E-05 at (24,38,2)
       Min divergence: -0.14E-05 at (24,9,34)
       Max VN number:   0.90E+00 at (13,13,31)

Thx for your reply. I will try to specify a DT value in other cases.