Zone and leaks
290 views
Skip to first unread message
Martin
Jun 10, 2008, 7:23:03 AM6/10/08
to FDS and Smokeview Discussions
Hi
I have some trouble understanding the leakage function. From what I
understand from the manual and the discussions on this forum the
leakage is just a simple explicit calculation based on Bernoulli’s
equation and the background pressure The flow is then distributed
evenly on the surface which is defined to leak. I also have some
trouble understanding the pressure concept in FDS. I assume that the
quantity “PRESSURE” is a sum of both pressure components (background-
and so called flow inducedpressure )? Is the background pressure a
global value applied to the whole Zone volume, computed on the basis
of an “apparent” temperature and the net massflux into the pressure
zone? Since the background pressure is global will there be problems
related to the leakage function where relatively small hydrostatic
pressure differences are important?
Please correct me if I got this wrong.
The reason for these questions is that I am trying to use the leakage
function in a large atrium with an enclosed and leaky elevator shaft
and stairwell in the atrium. I am trying to compute the leakage from
the atrium through the elevator doors and into the shaft. I am using
both the leakage- and “POROUS” function to simulate the movement of
smoke though the door cracks. (similar set up as the example files,
except for the fans). However the “LEAK AREA” does not appear to
restrict the movement of smoke through the porous surface. It appears
that the flow through the porous surface is restricted by the area of
the porous surface.
Also the “PRESSURE” seems to overshoot as a function of time and still
it appears to be a flow from low to high pressure?
Any repose to these matters will be highly appreciated.
Regards
Martin Hagen
I have some trouble understanding the leakage function. From what I
understand from the manual and the discussions on this forum the
leakage is just a simple explicit calculation based on Bernoulli’s
equation and the background pressure The flow is then distributed
evenly on the surface which is defined to leak. I also have some
trouble understanding the pressure concept in FDS. I assume that the
quantity “PRESSURE” is a sum of both pressure components (background-
and so called flow inducedpressure )? Is the background pressure a
global value applied to the whole Zone volume, computed on the basis
of an “apparent” temperature and the net massflux into the pressure
zone? Since the background pressure is global will there be problems
related to the leakage function where relatively small hydrostatic
pressure differences are important?
Please correct me if I got this wrong.
The reason for these questions is that I am trying to use the leakage
function in a large atrium with an enclosed and leaky elevator shaft
and stairwell in the atrium. I am trying to compute the leakage from
the atrium through the elevator doors and into the shaft. I am using
both the leakage- and “POROUS” function to simulate the movement of
smoke though the door cracks. (similar set up as the example files,
except for the fans). However the “LEAK AREA” does not appear to
restrict the movement of smoke through the porous surface. It appears
that the flow through the porous surface is restricted by the area of
the porous surface.
Also the “PRESSURE” seems to overshoot as a function of time and still
it appears to be a flow from low to high pressure?
Any repose to these matters will be highly appreciated.
Regards
Martin Hagen
Kevin
Jun 10, 2008, 8:16:27 AM6/10/08
to FDS and Smokeview Discussions
Your understanding of the pressure is quite good. I'll elaborate a bit
--
> From what I
> understand from the manual and the discussions on this forum the
> leakage is just a simple explicit calculation based on Bernoulli’s
> equation and the background pressure The flow is then distributed
> evenly on the surface which is defined to leak.
Yes
> I also have some
> trouble understanding the pressure concept in FDS. I assume that the
> quantity “PRESSURE” is a sum of both pressure components (background-
> and so called flow inducedpressure )? Is the background pressure a
> global value applied to the whole Zone volume, computed on the basis
> of an “apparent” temperature and the net massflux into the pressure
> zone?
Yes -- the low Mach number assumption in FDS essentially means that we
decompose the pressure field into a background, time-dependent only
field, and a small perturbation that drives the flow. Because the
background pressure is not spatially dependent, it allows us to
propagate information infinitely fast, not at the speed of sound, and
thus means that we do not have to use a tiny time step to track these
pressure waves. The background pressure applied to an entire ZONE, and
yes, it is just PV=nRT. The Tech Guide discussion of the divergence
and background pressure can fill in the details.
> Since the background pressure is global will there be problems
> related to the leakage function where relatively small hydrostatic
> pressure differences are important?
> Please correct me if I got this wrong.
Leakage only depends on the global background pressure. Local effects
are "unresolvable". What this means is that if you have two rooms
connected by a closed door, and neither room is "sealed", and each
room opens up somewhere else to the outside, then the smoke from a
cigarette leaking through the door crack cannot be modeled in FDS, at
least not with the LEAK and ZONE functions.
>
> The reason for these questions is that I am trying to use the leakage
> function in a large atrium with an enclosed and leaky elevator shaft
> and stairwell in the atrium. I am trying to compute the leakage from
> the atrium through the elevator doors and into the shaft. I am using
> both the leakage- and “POROUS” function to simulate the movement of
> smoke though the door cracks. (similar set up as the example files,
> except for the fans). However the “LEAK AREA” does not appear to
> restrict the movement of smoke through the porous surface. It appears
> that the flow through the porous surface is restricted by the area of
> the porous surface.
> Also the “PRESSURE” seems to overshoot as a function of time and still
> it appears to be a flow from low to high pressure?
>
I'd have to look at a very simple example to comment.
--
> From what I
> understand from the manual and the discussions on this forum the
> leakage is just a simple explicit calculation based on Bernoulli’s
> equation and the background pressure The flow is then distributed
> evenly on the surface which is defined to leak.
> I also have some
> trouble understanding the pressure concept in FDS. I assume that the
> quantity “PRESSURE” is a sum of both pressure components (background-
> and so called flow inducedpressure )? Is the background pressure a
> global value applied to the whole Zone volume, computed on the basis
> of an “apparent” temperature and the net massflux into the pressure
> zone?
decompose the pressure field into a background, time-dependent only
field, and a small perturbation that drives the flow. Because the
background pressure is not spatially dependent, it allows us to
propagate information infinitely fast, not at the speed of sound, and
thus means that we do not have to use a tiny time step to track these
pressure waves. The background pressure applied to an entire ZONE, and
yes, it is just PV=nRT. The Tech Guide discussion of the divergence
and background pressure can fill in the details.
> Since the background pressure is global will there be problems
> related to the leakage function where relatively small hydrostatic
> pressure differences are important?
> Please correct me if I got this wrong.
are "unresolvable". What this means is that if you have two rooms
connected by a closed door, and neither room is "sealed", and each
room opens up somewhere else to the outside, then the smoke from a
cigarette leaking through the door crack cannot be modeled in FDS, at
least not with the LEAK and ZONE functions.
>
> The reason for these questions is that I am trying to use the leakage
> function in a large atrium with an enclosed and leaky elevator shaft
> and stairwell in the atrium. I am trying to compute the leakage from
> the atrium through the elevator doors and into the shaft. I am using
> both the leakage- and “POROUS” function to simulate the movement of
> smoke though the door cracks. (similar set up as the example files,
> except for the fans). However the “LEAK AREA” does not appear to
> restrict the movement of smoke through the porous surface. It appears
> that the flow through the porous surface is restricted by the area of
> the porous surface.
> Also the “PRESSURE” seems to overshoot as a function of time and still
> it appears to be a flow from low to high pressure?
>
Martin
Jun 10, 2008, 10:32:42 AM6/10/08
to FDS and Smokeview Discussions
Hi
Thanks for the quick reply.
I took the liberty of uploading three files to illustrate parts of the
problem. ( not sure if I am allowed to upload, if not I will delete
them) I made two simple and identical cases ( 7a and 7b)except for the
"LEAK_AREA". I also took the liberty of making a simple Excell file
with a plot of the same point measurements in both cases. These show
that the inflow of combustion products do not change as a function of
"LEEAk_AREA".
Regarding the pressure overshoot, I see from this forum that it might
be related to numerical errors due to the explicit method and too
large "LEAK_AREA"?
Martin
Kevin
Jun 10, 2008, 11:12:52 AM6/10/08
to FDS and Smokeview Discussions
Yes, you are allowed to upload files -- that's the whole point of that
utility.
I'm running the cases and I looked at your plots. The pressure
difference between the compartments is small, and thus there is little
convective flow due to the leak, which is why your two cases look
similar. The POROUS attribute means that the blue sheet does not stop
the smoke from diffusing through. Thus, you are seeing smoke diffuse
through the area of the blue sheet, which is the same in both cases.
Neither the exact area of the blue sheet nor the "turbulent" diffusion
coefficient of the LES model are controlled by you, the user. So I
would not claim that the rate of smoke entering that compartment is
related to your specified leak area. I think that this is just a
numerical artifact.
The leak feature only has meaning when it is the principal driver of
the flow through the "crack". I think you'd want to see tens or
hundreds of Pa difference in pressure before using (or interpreting)
this feature.
> > I'd have to look at a very simple example to comment.- Hide quoted text -
>
> - Show quoted text -
utility.
I'm running the cases and I looked at your plots. The pressure
difference between the compartments is small, and thus there is little
convective flow due to the leak, which is why your two cases look
similar. The POROUS attribute means that the blue sheet does not stop
the smoke from diffusing through. Thus, you are seeing smoke diffuse
through the area of the blue sheet, which is the same in both cases.
Neither the exact area of the blue sheet nor the "turbulent" diffusion
coefficient of the LES model are controlled by you, the user. So I
would not claim that the rate of smoke entering that compartment is
related to your specified leak area. I think that this is just a
numerical artifact.
The leak feature only has meaning when it is the principal driver of
the flow through the "crack". I think you'd want to see tens or
hundreds of Pa difference in pressure before using (or interpreting)
this feature.
>
> - Show quoted text -
Martin
Jun 10, 2008, 2:24:40 PM6/10/08
to FDS and Smokeview Discussions
From what I understand, by embedding one zone (zone 1) in a larger
physical domain, the physical domain outside zone 1, will be treated
as Zone 0 with constant exterior pressure? However will FDS consider
the hydrostatic pressure difference in zone 0 when using the leakage
function bewteen zone 0 and zone 1? This refers to the setup used in
the enclosed files.
Is there any way of retrieving the background component from the total
pressure?
Once again I would like to thank you for the assistance on these
matters. This short discussion clarified several details.
Kevin
Jun 10, 2008, 5:42:22 PM6/10/08
to FDS and Smokeview Discussions
Only the background pressure is used for leakage and fan flows between
ZONEs. There currently is no way to separate background pressure.
ZONEs. There currently is no way to separate background pressure.
Liangzhu (Leon) Wang
Jun 10, 2008, 11:39:29 PM6/10/08
to fds...@googlegroups.com
I have questions here. If both background and perturbation pressures are available and the Bernoulli's equation still applies, the "total" pressure (background+perturbation pressure) should be used as driving force for Bernoulli's, right?
The "Compartment Zone" model in FDS assume the perturbation pressure is negligible (which is true for most of leakage airflows, where the leakage doesn't produce much perturbation in a room) so Bernoulli's equation uses the background pressure. Am I right?
The background pressure (both inside building or ambient) should be height and temperature(which is represented by density difference) dependent otherwise buoyancy-dominant flows can't be correctly modeled even for leakage flows. I guess the "compartment zone" model can do that.
The "Compartment Zone" model in FDS assume the perturbation pressure is negligible (which is true for most of leakage airflows, where the leakage doesn't produce much perturbation in a room) so Bernoulli's equation uses the background pressure. Am I right?
The background pressure (both inside building or ambient) should be height and temperature(which is represented by density difference) dependent otherwise buoyancy-dominant flows can't be correctly modeled even for leakage flows. I guess the "compartment zone" model can do that.
Kevin
Jun 11, 2008, 8:08:27 AM6/11/08
to FDS and Smokeview Discussions
Background pressure does include stratification of the atmosphere.
We don't use the total pressure to drive leak flows because most often
leaks are spread throughout the boundary and it would not be easy to
choose where to obtain the perturbation pressure. Even for a crack
under a door, using the local perturbation of pressure to drive a flow
through the crack is a numerical problem that can create local
instabilities and spurious flow patterns. I'm not even sure that the
physics are appropriate either. Using only the background pressure is
far more stable and more justifiable from a physical perspective.
Finally, there is only so much a model can do to simulate air flow
through tiny cracks when the grid resolution is several orders of
magnitude greater.
We don't use the total pressure to drive leak flows because most often
leaks are spread throughout the boundary and it would not be easy to
choose where to obtain the perturbation pressure. Even for a crack
under a door, using the local perturbation of pressure to drive a flow
through the crack is a numerical problem that can create local
instabilities and spurious flow patterns. I'm not even sure that the
physics are appropriate either. Using only the background pressure is
far more stable and more justifiable from a physical perspective.
Finally, there is only so much a model can do to simulate air flow
through tiny cracks when the grid resolution is several orders of
magnitude greater.
Reply all
Reply to author
Forward
0 new messages