search

Yinan-Scott-Shi / fds-smv

Automatically exported from code.google.com/p/fds-smv
0 stars 0 forks source link

Different HRR before sprinkler activation #800

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Please complete the following lines...

Application Version: fds 5.3.1
SVN Revision Number:3729
Compile Date: Apr 08
Operating System:Windows Vista

I run two equal simulations with and without sprinkler (attached is a reduced 
version and the comparison file of the reduced version). The HRR before the 
sprinkler activation is strongly different(in the full version stronger than 
in the reduced). I compared the files. They are equal except for the 
sprinklers. Can anybody explain?

Original issue reported on code.google.com by grue...@web.de on 3 Jul 2009 at 11:23

Attachments:

GoogleCodeExporter commented 9 years ago 
The cases are not equal.  Without a sprinkler FDS only has the two mixture 
fraction
species (unburned and burned fuel).  With a sprinkler FDS adds a third species 
for
the water vapor from evaporated droplets.  The initial concentration of the 
species
is set to 40 % humidity.  Thus, there will be subtle differences between the two
since the case with sprinklers has slightly less oxygen (due to the initial 
water
vapor).

Original comment by drjfloyd on 3 Jul 2009 at 4:37

GoogleCodeExporter commented 9 years ago 
Thanks for the explanation.
I tried the case with HUMIDITY=0 and the HRR was the same as in the simulation 
without 
sprinklers.

Original comment by grue...@web.de on 6 Jul 2009 at 7:32

GoogleCodeExporter commented 9 years ago 
But that means also, that simulations without sprinkler always run with 0% 
Humidity and 
yield subtle higher temperatures in a subtle smaller flame.
In my special case a solid object near the burner ignites about 100 seconds 
later with 
sprinkler. That is not a subtle difference.
Why does FDS not account for hunidity in a "normal" simulation
Michael

Original comment by grue...@web.de on 16 Jul 2009 at 1:48

GoogleCodeExporter commented 9 years ago 
If you feel ambient humidity is important to your computation, then include it 
as an
input.  I would be willing to bet that the uncertainty in your knowledge of 
material
properties probably outweighs the impact of ambient humidity.

Original comment by drjfloyd on 16 Jul 2009 at 4:58

GoogleCodeExporter commented 9 years ago 
Jason-

Certainly adding humidity to the MISC line is necessary if it is important in 
the
calculation.  I think the submitter is just pointing out that, even if humidity 
is
not expected to be important, if the simulation has sprinklers, the humidity is 
40%
and if it doesn't, the humidity is 0%.  Assuming you don't want to track water 
vapor
explicitly in the non-sprinkler models, perhaps you just make humidity 0% as the
default for sprinkler spray models and users can set the humidity to 40% or 
whatever
if they choose.  Then the models would run the same on the defaults.  Just a 
suggestion.

Original comment by Stephen....@gmail.com on 16 Jul 2009 at 7:15

GoogleCodeExporter commented 9 years ago 
The trouble with 0% humidity in a sprinklered case is that the droplet 
evaporation 
is affected by the water vapor concentration. I think what we need to do is 
look at 
the gas properties, especially specific heat, with and without humidity.

Original comment by mcgra...@gmail.com on 16 Jul 2009 at 7:23

GoogleCodeExporter commented 9 years ago 
Yeah, thats fine.  I think all the submitter is looking for is consistency.  The
background species is assumed to be the same by default regardless of whether
sprinklers are present or not and that isn't necessarily true.

Original comment by Stephen....@gmail.com on 16 Jul 2009 at 7:41

GoogleCodeExporter commented 9 years ago 
At ambient temperature, 40 % relative humidity is around 1/2 % mole fraction.  
This
really makes no difference in terms of specific heat, viscosity, or density.  
If the
minor decrease in O2 or the minor increase in specific heat is enough to cause 
or
prevent flame spread to another object, what then is the effect of the much 
larger
uncertainties in the thermophysical properties of the object being ignited, the 
fuel
chemistry, the overall calculation uncertainty, etc.

Original comment by drjfloyd on 16 Jul 2009 at 8:29

GoogleCodeExporter commented 9 years ago 
I hear you.  It is likely in the noise.  Maybe just something in the User's 
Guide? 
Right now, it lists that the HUMIDITY on the MISC line default is 40%, and 
while that
is true, really the default is 0% humidity if HUMIDITY is not set and 
sprinklers and
not included.  Perhaps just a sentence indicating that in the description of 
HUMIDITY
in the User's Guide.  Either way, though, I agree with you it likely is not a 
big
deal in 99.99% of all calculations.

Original comment by Stephen....@gmail.com on 17 Jul 2009 at 9:41

GoogleCodeExporter commented 9 years ago 
Yes, that's it Stephen. I do not understand, why the User's Guide default for 
HUMIDITY 
is 40%, but in the real calculation it is 0%. Maybe it is not a big deal, but 
for all 
simulations it would be a better assumption to calculate with the default value 
than 
without. The Humidity of 0% is not a realistic assumption, isn't it?

Original comment by grue...@web.de on 17 Jul 2009 at 10:28

GoogleCodeExporter commented 9 years ago 
This all comes down to how long you want to wait for your results. We can track 
water vapor in every FDS calculation, but that means one more transport 
equation and 
an increase in CPU time of 10 to 15 %. Is that worth it, given all the other 
uncertainties? 

I will clarify the User's Guide -- one idea is to make any specification of 
HUMIDITY 
trigger the inclusion of water vapor as a separate species. I'd have to put in 
words 
of warning about increased CPU.

Original comment by mcgra...@gmail.com on 17 Jul 2009 at 12:13

GoogleCodeExporter commented 9 years ago 
Kevin,

Been pondering this a little bit now.  With the way we now do the mixture 
fraction,
we could account for the presence of ambient water vapor without needing to use 
extra
scalars.  The Y2Z_C vector would just need an initial value for H2O like we do 
for N2
and O2.

Original comment by drjfloyd on 17 Jul 2009 at 12:50

GoogleCodeExporter commented 9 years ago 
OK, but let's put 5.4 out first. We'd have to re-run the V&V. Note the change 
in 
Milestone.

Original comment by mcgra...@gmail.com on 17 Jul 2009 at 12:59

GoogleCodeExporter commented 9 years ago 
yeah, I don't think this is a top priority, and Kevin, I definitely don't think 
that
it is worth tracking H20 with a 10-15% up in calc time.  Like I said, unless 
Jason
has an easy solution, I would just add in some discussion in the users manual 
that
while the HUMIDITY default is 40%, that default is for sprinkler cases.  
Otherwise,
the default is really dry air (HUMIDITY = 0%)

Original comment by Stephen....@gmail.com on 18 Jul 2009 at 3:28

GoogleCodeExporter commented 9 years ago 
The new species framework for FDS 6 now incorporates ambient water vapor into 
the AIR species regardless of the presence of sprinklers.

Original comment by drjfloyd on 10 Aug 2011 at 3:05

GoogleCodeExporter commented 9 years ago 
Closing this issue.  Made obsolete by species changes in FDS6

Original comment by drjfloyd on 18 May 2012 at 7:04