Modelling of wind and turbulence profiles

[Robert Taylor]

Dear Kevin,

 I am modelling ground flares and low vents for sour gas (natural gas with H2S) and have used FDS5 with a good deal of success. Thank you very much to the FDS team. It has been possible to validate the calculations by comparison with photographs and with hand held gas detectors. I adress the questions I have to you Kevin, because of your earlier publications in modelling large pool fires, but please correct as necessary.

 In order to make the calculations reflect actual atmospheric conditions it is necessary to establish a wind speed profile. To do this I set the initial wind speed in the &MISC line, and a VEL and PROFILE in a wind source &SURF line. This works well. However it takes some time for the simulation to establish a wind speed profile accross the complete set of meshes (I use a field of about 2 km with varying and well matched meshes to allow for finer modelling of the flare). Question 1: Is there any way of establishing the wind speed profile accross all meshes, in the same way that average wind speeds can? This is a time saver in the simulations only, so is not critical, but would be a great convenience for all simulations of outdoor fires.

 In order to make the calculations reasonable, I have assumed that the atmospheric turbulence field should also be well establiehed and "stable" to the extent that the background turbulence does not change in statistics accross the area of the flare and smoke plume. I have tried this in two ways. The first is to use a long upwind reach, in which turbulence due to ground friction is generated. It is also possible to account for unstable atmospheric conditions by providing convective heat release at a ground &SURF, matched to the measured heat release due to ground temperature from solar radiation. Trying to establish stable atmospheric conditions by using negative heat release does not work - unsurprising, and it would probably take a very long upwind reach to establish a good atmospheric temperature profile anyway - in nature establishment of a stable atmospheric layer often reqires many kilometeres to establish a stable layer.

 The other way which I have tried is to establish a turbulence and temperature profile at the "wind source" &SURF, with a ramp to change wind speeds with time, using variations derived from a preliminary simulation using FDS. This is like simulating a long upwind reach, then storing it and reusing it. It has been possible to match the ramp to publications of wind speed statistics and to meteorological mast measurements. However, This is quite clumsy, and establishing a vertical profile is quite difficult (requires many small vents). So question 2: Is there any simpler way of specifying vertical turbulence and temperature profiles for wind?

 

With kind regards,

  

Robert Taylor

 

 

[F-Sim.de]

Please mind LAPSE_RATE.
FDS 6 lets you use a synthetic turbulence inflow condition (N_EDDY, L_EDDY, VEL_RMS).

[Robert Taylor]

Dear F-sim, Many thanks. Using FDS5 I was not aware of turbulence inflow setting facilities. I will obviously need to upgrade.

Robert Taylor

On Mon, Jan 16, 2012 at 1:50 PM, F-Sim.de <In...@f-sim.de> wrote:

Please mind LAPSE_RATE.
FDS 6 lets you use a synthetic turbulence inflow condition (N_EDDY, L_EDDY, VEL_RMS).

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/Fo5Xu6Ju18oJ.
To post to this group, send email to fds...@googlegroups.com.
To unsubscribe from this group, send email to fds-smv+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/fds-smv?hl=en.

[F-Sim.de]

But be aware that FDS 6 is 'work in progress' and not officially released yet......

[Robert Taylor]

Dear F-Sim,

I am fully aware of that FDS& is a work in progress, and will take care to use simple options. The facilities provided though are precisely those needed , so the effort in adapting to FDS6 should definitely be worthwhile. I generally validate results against some test cases I have from observations in the field, so that any anomalies which there might be should show clearly.

With kind regards, Robert Taylor

On Tue, Jan 17, 2012 at 9:03 AM, F-Sim.de <In...@f-sim.de> wrote:

But be aware that FDS 6 is 'work in progress' and not officially released yet......

--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.

To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/WM7S2J0knfwJ.

[Randy McDermott]

Robert,

If you download and compile the latest source in order to use Jarrin's synthetic turbulence inflow conditions, also compile the latest User Guide in the repository and it has instructions on how to set the turbulence parameters.  You need to specify, at a minimum, a turbulence length scale and a turbulence intensity.  But you can also specify anisotropic length scales and the Reynolds stresses if you know them.  I'm interested to hear how well it works.

Another option I have had some luck with simulating atmospheric flows is to specify a mean forcing with periodic boundary conditions.  This is also described in the latest User Guide.  You can say, for example, that you have a mean wind speed of 15 m/s in whatever direction (U0 = ..., V0 = ..., MEAN_FORCING(1:2) = .TRUE.,.TRUE., on MISC).  Then let the ground level turbulence develop in time.

Fair warning that introduction of species concentrations are handled differently in v6 (input files will probably not directly translate), and so you may need to do some reading on that front as well, and possibly look at some of the latest verification examples.

Best,

Randy

[DavidShep]

In many of our simulations we have to run the model for a very long
time in order establish the initial flow field prior to introducing
our fires. For example, in a large room with HVAC where the goal of
the model is to predict smoke detector activation from a very small
fire.

It would be very useful if we could run the model once to define what
the initial flow field is then use the final flow field from the first
run as the initial conditions for subsequent runs. I understand that
the saved initial flow field would only work for models with identical
geometries.

[Randy McDermott]

It's on the "to-do" list.  Still working on a couple of fundamental issues with v6.  Then we'll knock out some of these enhancements.  Sorry it is taking so long.

http://code.google.com/p/fds-smv/issues/detail?id=730

[Robert Taylor]

Dear Randy,

Thank you very much for your reply, which is just what I needed. My work is primarily in safety engineering, so I will need to catch up with Fortran programming again - last serious work with Fortran  was in 1969. It does not look too difficult to get the mechanics going again, from inspection of the Wiki's.

 

The use of periodic bounding functions was described in outline in the paper by Yvon Mouilleau and  Anousone Champassith, CFD simulations of atmospheric gas dispersion using the Fire Dynamics Simulator (FDS), Journal of Loss Prevention in the Process Industries 22(2009) 169–176.  This gives a quite resonable result if you give enough distance for secondary turbulence to develop, but gives vortex shedding frequencies from sharp edged hills, equipment ettc which are uncomfortably contant. The turbulence statistics from this seem to be good enough, and the average toxic exposure over several minutes gives good agreement with the simulations with turbulence developed by a large fetch distance. Using a forcing function based on earlier simulations with varying times between RMS peaks looks better, and shows some additional flow effects. Note that the "look" of a simulation is still important in convincing engineers of the validity of calculations, even when you have good quantitative validation. They have the real thing to compare with, and are quite sensitive to visual differences between simulation and reality.  

 

I have been running the development of flare simulations in difficult terrain for a few of months now and am getting results which validate quite well, as far as visible flow patterns and some concentration measurements are concerned. I am writing this up as a guide so that colleagues can make use of the results, and will gladly forward a copy when I have had time to incorporate your suggestions. One thing which will be important is the performance of flares under inversion conditions - in the desert environment concerned, inversion is frequent (daily) for some months of the year. Using the lapse rate parameter allows stable conditions to be specified, especially if accompanied by cold ground surface, but the very rapid rise in temperature at typically 60 to 100 m. which we observe is difficult to account for in this way. I have tried using two "wind source" surfaces, one above the other, with a little heat flux on the upper to create a warm upper air layer, and providing an upwind fetch distance to allow the layer interactions to settle down. Some experimenting is needed to get the profiles right. The results are visually convincing, but I do not yet have observations to allow validation. Do you have any comments?

 

Thank you for the warning about species on FDS6. Considering the amount of work your team has been devoting, I find the degree of backward compatibility very encouraging.I had anticipated needing to redevelop all the models completely. I will in any case carry out revalidation against observations.

 

With kind regards, Robert

--

You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.

To view this discussion on the web visit https://groups.google.com/d/msg/fds-smv/-/qdEm5G9gA2kJ.