search

firemodels / fds

Fire Dynamics Simulator
https://pages.nist.gov/fds-smv/
Other
663 stars 622 forks source link

FDS6: Screen (drag) - parameter sensitivity #1781

Closed gforney closed 9 years ago

gforney commented 9 years ago 
FDS Version: 6
SVN Revision Number:14299
Compile Date:20 Dec 2012
Smokeview Version/Revision:
Operating System:Windows 7

I have been looking at the implementation of metal grates/screens in FDS 6 using Lagrangian
particles through a few test cases with simple geometries. As a parameter sensitivity
study I have been examining the effect of modifying the 'screen' parameters while keeping
all other variables constant, and examining the effect on the predicted flow through
the particles.

From the information in the technical reference guide the two parameters of interest
relative to drag are the free area fraction and the screen thickness. The volumetric
flow rate through the area of the screen was examined as an output to see the effect.

Modification of free area fraction yielded behaviour that was expected. However, modification
of the screen 'thickness' has resulted in behaviour that is not expected.

Examining thicknesses in the range of 1 to 2 cm, the flow rate generally decreased
with increasing thickness. However, at certain thickness values the flow rate significantly
increased over very small incremental changes in thickness, and then continued to gradually
decrease with increasing screen thickness (as shown in the attached figure). Examining
the transition in more detail, increasing the specified screen thickness as little
as 5E-07 m (from 0.0131965 m to 0.013197 m) resulted in ~21% increase in flow rate
for the configuration examined.

The file used for examining the screen thickness (screen_601) is attached. The behaviour
was observed using grid sizes of 0.25x0.25x0.25 m, 0.125x0.125x0.125 m, and 0.05x0.05x0.05
m.

In the above tests the specific heat of the material for the screen (aluminum) was
left at an incorrect value of 900, as the thermal properties were not of interest for
an isothermal examination of drag. However, modification of just the specific heat
for the screen material, keeping all other parameters the same, resulted in significant
changes to the behaviour of the flow through the screen (plot of flow rate vs time
is attached for screen thickness = 0.01778).

Original issue reported on code.google.com by adrianmilford on 2013-01-16 19:22:11

gforney commented 9 years ago 
(No text was entered with this change)

Original issue reported on code.google.com by mcgratta on 2013-01-16 19:52:35

gforney commented 9 years ago 
I'll take a look when I get back in the office next week.

Original issue reported on code.google.com by drjfloyd on 2013-01-16 20:00:04

gforney commented 9 years ago 
Something is amiss with how the particle diameter is getting set and that is the root
cause of the behavior.

Original issue reported on code.google.com by drjfloyd on 2013-01-18 23:20:47

gforney commented 9 years ago 
There was an error in getting the diameter when calculating screen drag when material
properties were given.  The routine was picking the thickness of the first layer of
the screen wire rather than the total thickness.  The strange behavior was a result
of how the size of that first layer varied with the RADIUS or the SPECIFIC HEAT.

Original issue reported on code.google.com by drjfloyd on 2013-01-21 15:51:43

gforney commented 9 years ago 
Reply to email shown below (note in the future please keep communication in the tracker):

The issue is related to how your MATL and SURF input is being processed.  The test
case you submitted is isothermal so the MATL and SURF properties (other than the wire
radius) don't have any effect once the bug is fixed.

Until we release either 6 or RC4 delete your PART, SURF, and MATL lines and add:

&RADI RADIATION=.FALSE./
&SPEC ID='LIQUID', DENSITY_LIQUID=1000., VAPORIZATION_TEMPERATURE=3000., HEAT_OF_VAPORIZATION=1.E10,
SPECIFIC_HEAT_LIQUID=10., MELTING_TEMPERATURE=100./
&PART ID='SCREEN',SPEC_ID='LIQUID', DRAG_LAW='SCREEN', FREE_AREA_FRACTION=0.5, STATIC=.TRUE.,
DIAMETER=35560., MONODISPERSE=.TRUE.,ORIENTATION=1,0,0/

The use of a liquid drop avoids the bug and the properties for the liquid will keep
any evaporation from occurring.

Email from OP:
Good Morning Dr. Floyd,

Thank you for your prompt response regarding my issue tracker post (Issue 1816). Would
it be possible to get an updated executable once the fix is implemented? I'm looking
to continue some testing of screens/meshes for a number of models that we have run
previously and would greatly appreciate the chance to keep going on it as soon as I
can. 

Thanks for your time.

Best regards,

Original issue reported on code.google.com by drjfloyd on 2013-01-22 16:33:53

gforney commented 9 years ago 
verified that changing cp by factor of 1E3 doesn't change results

Original issue reported on code.google.com by drjfloyd on 2013-02-26 19:14:31