FDS 6.3.2 - DT_MEAN_FORCING

[Ed]

Hi All

With reference to FDS 6.3.2 and the mean forcing feature (FDS User Guide Section 6.4.2), I note the recent update has introduced the DT_MEAN_FORCING line representing the time scale for relaxation.  I also note in Table 17.14 the DT_MEAN_FORCING value is 1E10 (time, seconds). 

I had a look through the Technical Reference guide but couldn't find any specific reference to the data assimilation / nudging approach to model wind other than the reference to Kalnay's work. 

Would anyone be kind enough to point me in the right direction of DT_MEAN_FORCING is doing in greater detail than the User Guide?

For information, I am currently looking at modelling wind effect on a structure in an open air area. 

Many thanks!

[Randy McDermott]

Hi Ed,

Here is the brief story... The "nudging" approach in Kalnay has been proposed as a first, rudimentary step in "data assimilation" in FDS.  The main user has been one of our colleagues here at NIST who is working on pollutant dispersion in urban areas.  In short, the method is very much in the research stage.  What I can tell you is that DT_MEAN_FORCING is the "tau" in Kalnay's nudging approach.  You can easily see in the code that the forcing function is a simple relaxation toward the target velocity value:

DU_FORCING = (UBAR-UMEAN)/DT_MEAN_FORCING

UBAR is the target

UMEAN is the current mean for the U component

In the very early stages we set DT_MEAN_FORCING by default.  But, as the method was being used we quickly noticed that the time scale can greatly affect the results for time-dependent mean flows, that is, if the mean wind speed is a function of time.

Here is an example from our verification suite:

https://github.com/firemodels/fds-smv/blob/development/Verification/WRF/wrf_time_ramp.fds

The reason to set the default to a large value is to force the user to think about what this value should be for their problem.  The FDS developers can't know a priori.  There is not much more to it than that, at this point.  I do not plan to write this up in more detail until it is better understood, because anything written in our guides may get misconstrued as "validated".

Randy

--
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 post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/3ac21ee7-6152-4a8d-93e8-68ee6b01b670%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Ed]

Hi Randy

Many thanks for the prompt reply and for the explanation.

This has given me more to think about and I certainly agree with including an obviously wrong value as the default as it's not unusual for users to revert to default values taking that as 'acceptable' without really thinking about what that means.

This is a good case in point as the 1E10 value forced a user (me!) to think about this.

Many thanks.