wind direction

[sapiterbang]

I want to create a simulation of the propagation of fire in the wind
changing direction, what should I expect? in the FDS simulation can
make a change of wind direction??

thx before

[dr_jfloyd]

It depends on the change in direction. You cannot take a domain
boundary and change it from OPEN to a fixed velocity or vice versa
during a computation. So if the wind switches from N to S, you would
not be able to do that at the domain boundary. If you had a wind that
shifted from N-NE to E-NE (where at T_BEGIN both the N and the E
boundaries had a fixed velocity) that wind shift you could do by
assigning RAMP_Vs to the two boundaries.

Now all that being said, you could try having all the exterior
boundaries OPEN and use POROUS OBSTs near the boundaries (look at the
jet fan example). You could turn on and off the OBSTs as needed with
a TIME DEVC and assign a RAMP to the SURFs for the OBSTs to give the
time dependence. So if your wind shift from N to S via E, you would
start with an OBST for N boundary, then add an OBST for E boundary,
then remove the N OBST, the add the S, and finally remove the E (each
OBST would need a SURF with an appropriate RAMP_V). I don't know if
this has ever been tried, so I'd start off with a simple test case.

[rmcdermo]

One of the biggest challenges with external, atmospheric flows is specification of the boundary conditions.  I have played with many different bcs, both sides and top bc, in FDS and the results are extremely sensitive to this "parameter" of the model.  A much better approach to handling a mean wind speed is to use the MEAN_FORCING approach:

&MISC MEAN_FORCING(1)=.TRUE., U0=10., MEAN_FORCING(2)=.TRUE., V0=10.

This specifies a northeast wind.  Use OPEN bcs everywhere else.

So, step #1 is getting a wind you can trust.  Then we would have to work on making the wind time dependent.  U0, V0, W0 would have to linked to a DEVC.

[sapiterbang]

if i using &MISC MEAN_FORCING(1)=.TRUE., U0=10.,
MEAN_FORCING(2)=.TRUE., V0=10

it means in computational..wind always blowing in every second to
northeast until the computational done?

[Chris]

What is the difference between MEAN_FORCING and defining a U0 without
this parameter?

Thanks
Chris

> > making the wind time dependent.  U0, V0, W0 would have to linked to a DEVC.- Zitierten Text ausblenden -
>
> - Zitierten Text anzeigen -

[rmcdermo]

sapiterbang: Yes.

Chris: Without MEAN_FORCING(1)=T, U0 just sets the initial flow field to U(I,J,K) = U0.  Wish MEAN_FORCING(1)=T, you are also continually nudging MEAN(U(I,J,K)) toward U0.

In the future, as we enhance our "data assimilation" capabilities, the user will have more control over this feature using DEVC like a weather monitoring station.  And you could have several stations within one domain, each specifying data which varies in time.  What we have with MEAN_FORCING, is the most rudimentary form of data assimilation imaginable, but it's a start.

[dr_jfloyd]

Does MEAN_FORCING try and impose a constant U0 with height or will it
give a z-dependent U0?

[Chris]

So MEAN_FORCING is more or less a linking to a DEVC and an adjustment
to that DEVC-value? To keep the value more constant over time?

> > start.- Zitierten Text ausblenden -

[sapiterbang]

i using &MISC TMPA=25,MEAN_FORCING(1)=.TRUE.,U0=12.85/

and the result is ERROR: Problem with MISC line

is that right using mean_forcing?? because i didn't found it in user
guide

[rmcdermo]

You are probably using an old version of the code.  MEAN_FORCING is in latest SVN.  The latest guides have been updated, but not released.  You can compile them from the repository.

Jason, there is no z dependence on U0 at the moment.