Smoke Evacuation Issue

1,160 views
Skip to first unread message

Dan L

unread,
Aug 8, 2014, 1:47:41 PM8/8/14
to fds...@googlegroups.com
Hello all, i'm looking for some help.  I'm using FDS to simulate a mechanically ventilated smoke exhaust system to verify a prescriptive design calc I did as per NFPA 92.


The problem I am having is I cannot get my smoke layer to stabilize at the "Smoke Layer Height Above Highest Level of Egress" point.  I have set up the design fire size and HRR at 10x10 & 20 btu/s/ft2, and set my exhaust vents to match the exhaust volumetric flow rate of ~60,000 CFM required by code.  I have 5 vents uniformly distributed in location & exhaust rate.  My fans are triggered by a beam detector located approximately at my "highest level of egress" cutoff...the fans should activate at full blast at 25' and plugholing should be nullified due to the thick smoke layer beneath.  Yet I cannot seem to stabilize my smoke layer - it continues to descend during the exhaust process.


The reaction chosen is a modified polyurethane reaction, with a defined Soot yield of 0.07 g/g.


Can anyone provide insight on why this might be happening?  Did I misuse the exhaust vents, or pick an improper reaction?   Would appreciate all help, i'm becoming quite frusturated.


FDS file attached.


Also, on a side note, I cannot get Plot3d to function, both using the input file and through using Pyrosim.  Does anyone know what I need to put into the file to get Plot3D to show up in Smokeview?
Geometry_Export.fds

Kevin

unread,
Aug 8, 2014, 2:00:21 PM8/8/14
to fds...@googlegroups.com
I bet. There are many things that you could be doing wrong, but it would be very difficult for anyone to figure out by reading hundreds of lines of input. I suggest you create a 10 to 20 line input file that consists of the shell of the building, one vent, one inlet, and some diagnostic that will demonstrate whether or not your fan is working. If you still can't figure it out, and you think there is a bug in the code, submit the simple case to the Issue Tracker.

Dan L

unread,
Aug 8, 2014, 2:03:45 PM8/8/14
to fds...@googlegroups.com
Ok, I will try that.  Do you recommend any particular diagnostic?  I ask because I'm already recording mass flow from the fire & exhaust, and I have more flow from the fans then is created by the fire.

Kevin

unread,
Aug 8, 2014, 2:10:12 PM8/8/14
to fds...@googlegroups.com
Remove the fire from your test case. Just leave the empty building, a single fan, and your mass and volume flow diagnostics. Confirm that the volume flow is what you want. As for the fire, When you put the fire back in, I would not be surprised if your resulting visibility does not agree with simplified analyses, where you just sum volume flows into and out of the hot gas layer.

Dan L

unread,
Aug 8, 2014, 2:14:22 PM8/8/14
to fds...@googlegroups.com
Ok, thank you Kevin.  I will try and verify.


For anyone who also performs smoke evacuation models, what do you usually set your reaction values at?  I may be using a fire that creates an inappropriate amount of soot.

dr_jfloyd

unread,
Aug 8, 2014, 4:44:50 PM8/8/14
to fds...@googlegroups.com
The NFPA calc merely tracks the layer interface using some very simple assumptions.  Soot yield really has no meaning if your only goal is comparing FDS to NFPA 92..  You are just comparing the interface between air with no combustion products and air with combustion products. If you were using FDS for a performance based design, then soot yield becomes important.  Depending on the size of the atrium compared to the fire, how makeup air is being supplied, etc you should expect that there could be  significant differences between a hand calc using an idealized plume correlation and FDS which accounts for all the air movement in the atrium.

Yatin Munnsad

unread,
Aug 11, 2014, 3:26:19 AM8/11/14
to fds...@googlegroups.com
I do not think your exhaust fans are extracting at all. If you are using FDS 6 you may want to change the following line to:
 
      VOLUME_FLOW to VOLUME_FLUX
 
I had an issue previously on one of my previous models when I noticed that FDS 6 does not recognise VOLUME_FLOW as an exhaust surface and disregards it as an inert surface. Anyway it worked for me and hope it works for you.

Dan L

unread,
Aug 11, 2014, 9:14:13 AM8/11/14
to fds...@googlegroups.com
Hi Yatin,

I get the message "ERROR: Problem with Surf: Exhaust Fan. VOLUME_FLUX is depreciated; use VOLUME FLOW".  I have not been able to use volume flux, and have exchanged it for volume flow in FDS 6.  Do you not receive this message?

Randy McDermott

unread,
Aug 11, 2014, 9:15:31 AM8/11/14
to FDS and Smokeview Discussions
If VOLUME_FLOW does not produce a flow, please submit an issue with a simple test case.  Thanks.


--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/f7662a2b-7647-445b-b6c7-0f751f750fbc%40googlegroups.com.

For more options, visit https://groups.google.com/d/optout.

Brad Casterline

unread,
Aug 11, 2014, 3:26:47 PM8/11/14
to fds...@googlegroups.com
Dan L,

It looks like your exhaust fan is a vent with the exhaust surf properties, but the vent is located on a solid obstruction where a hole is created upon beam detector activation.
Could this be causing the problem? You do not have to create a hole for the smoke to go out, it just 'disappears' at the vent.

Brad Casterline

Brad Casterline

unread,
Aug 12, 2014, 5:19:40 PM8/12/14
to fds...@googlegroups.com
Hi Dan L,

I confirmed that removing the solid obstruction that the exhaust vents are on (by creating a hole) affects the exhaust rate-- instead of the expected ~58,000 CFM I get about 25,500 CFM and after about 4 minutes the Layer Height is ~12 feet lower than it "should be". I noticed the Z plane of your measuring DEVC did not match the Z plane of your exhaust vents so I moved the DEVC's to the face of the vents. The 3 fans farthest from the fire where exhausting ~4 m3/s (instead of 5.47) and the 2 nearest the fire were .00?? m3/s.
I guess a vent not only has to start on a solid obstruction, it has to stay on one!
I also saw what Dr. Floyd was saying about make-up air affecting the "Layer"-- with 58,000 CFM make-up air coming through just a few doors, the velocity is 4 or 5 times faster than 92B would allow just doing the hand calcs, anyway, that really keeps things stirred up, and the plume leans quite a bit :)

thanks, Brad

Brad Casterline

unread,
Aug 13, 2014, 5:56:33 AM8/13/14
to fds...@googlegroups.com
Oops, your side note!
Starting with v6 Plot3D is no longer written out by default. On the &DUMP line, give the time interval, i.e., DT_PL3D=150, and if you want to replace some default quantities with CO, CO2, and O2, it is a little more involved than it used to be-- see 16.7 in the User's Guide.

Dan L

unread,
Aug 13, 2014, 7:41:10 AM8/13/14
to fds...@googlegroups.com
Hi Brad,

This is excellent advice, I sincerely thank you.  I didn't even think about the holes affecting the flow - my other modeling experience (other programs) have surfaces that act just as fans, not as fans and exit boundary for the model. Something I should have understood after reading the manual.  I've started running a new model with the fans on the ceiling with no holes created, and will see how that shapes up.  Also boosted my mesh resolution to help define the flow in that drop ceiling.

Regarding Dr. Floyd's apt comment, I had modified my input file to increase the amount of supply air to the model - and i'm actually supplying air, vs. just having openings.  The model was reporting 2.5 m/s through the doors, which I knew was above the 1 m/s maximum required by code.  I just didn't think originally that i'd have such an impact.  However when I ran the model with a tighter resolution, I did the the plume lean near 45 degrees towards the 'east' door!  Definitely not appropriate, and (in accordance with Dr. Floyd's comment) made me appreciate how incorrect my makeup air location & vent size was.

On the DEVC location - is it ok for them to overlap?  I was under the impression that they would not function properly if it was on the same plane as the Exhaust fan.

dr_jfloyd

unread,
Aug 13, 2014, 7:51:53 AM8/13/14
to fds...@googlegroups.com
There are two volume flow device types.  One for measuring in the gas (VOLUME FLOW) and one for measuring on a solid surface (VOLUME FLOW WALL).

Brad Casterline

unread,
Aug 13, 2014, 8:00:03 AM8/13/14
to fds...@googlegroups.com
My pleasure Dan, thanks for saying. It may not be totally correct, we'll see.
Yes about the DEVCs on the face of the vents because without the holes all 5 vents show 5.47 or 5.48 m3/s.
I also added a velocity slice with VECTOR=.TRUE. at PBY=3.6 to verify the fans were turning the right direction :).
Keep up the good work!

Brad

Dan Lampke

unread,
Aug 13, 2014, 8:20:48 AM8/13/14
to fds...@googlegroups.com
Dr. Floyd -

I wasn't aware of two different devices.  Do I specify by just replacing my VOLUME_FLOW with VOLUME_FLOW_WALL?  I didn't see this device in the V6 users manual, and therefore wasn't aware this changed.


--
You received this message because you are subscribed to a topic in the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/fds-smv/L6DPOKnbY7w/unsubscribe.
To unsubscribe from this group and all its topics, send an email to fds-smv+u...@googlegroups.com.

To post to this group, send email to fds...@googlegroups.com.

dr_jfloyd

unread,
Aug 13, 2014, 8:58:01 AM8/13/14
to fds...@googlegroups.com
The _WALL outputs (mass, volume, heat) were added in 6.1 to make it easier to monitor extraction vents in simulations like you are running.  

Randy McDermott

unread,
Aug 13, 2014, 9:13:55 AM8/13/14
to FDS and Smokeview Discussions
Brad,

I apologize that the WALL variants were buried in the user guide.  I changed this to make them have their own subsubsection, but it still does not show up in the table of contents.

R


Brad Casterline

unread,
Aug 13, 2014, 9:30:00 AM8/13/14
to fds...@googlegroups.com
My Gosh Randy, no need to apologize! But thanks, I will check it out today at lunchtime. The Code Division is caught up on fire modeling projects so I am back full time designing fire sprinkler systems...

wbr,
Brad

Brad Casterline

unread,
Aug 13, 2014, 2:11:54 PM8/13/14
to fds...@googlegroups.com
VOLUME FLOW WALL works like a charm when I run this simple file:

&HEAD CHID='vol-flow-wall'/
&TIME T_END=30/
&MESH IJK=10,10,10, XB=0,1,0,1,0,1/
&VENT MB='YMIN', SURF_ID='OPEN'/
&SURF ID='exhaust', VOLUME_FLOW=1.0, COLOR='GREEN'/
&VENT XB=.4,.6,.4,.6,1,1, SURF_ID='exhaust'/
&DEVC XB=.4,.6,.4,.6,1,1, QUANTITY='VOLUME FLOW WALL', IOR=-3/

but I get 'problem with device' on your input file Dan L, and I think that is due to your vent and devc Z plane being 11.5824,11.5824, FDS is having trouble snapping to the same solid surface. I would make your fans a shallow solid of their own with the Z plane a know Z grid coordinate!

good luck, Brad

Dan L

unread,
Aug 13, 2014, 2:16:05 PM8/13/14
to fds...@googlegroups.com
Yes, I get the same error too when I was running.  So you are suggesting that I give my exhaust fans a (minor) thickness/make them a solid, and place the DEVC on the face of that surface?

Brad Casterline

unread,
Aug 13, 2014, 2:36:09 PM8/13/14
to fds...@googlegroups.com

Yes.

I would make the XB of your new fan OBST say, 4 cells by 4 cells by 1 cell, with the Z a know grid location just below the bottom of the roof deck and put the vent and devc on the bottom. In fact you do not even have to have a vent, you could give the fan OBSTs min Z the exhaust surf properties, and create them by the beam detector activation, like you were creating the holes, because the vent in this case is really just used to apply a SURF property, not to simulate a traditional air vent!!


dr_jfloyd

unread,
Aug 13, 2014, 2:50:10 PM8/13/14
to fds...@googlegroups.com
7.2.2 Thin Obstructions

Obstructions can have zero thickness.....This feature is fragile, especially in terms of burning and blowing gas. A thin sheet obstruction can only have one
velocity vector on its face, thus a gas cannot be injected reliably from a thin obstruction because whatever is pushed from one side is necessarily pulled from the other. For full functionality, the obstruction should
be specified to be at least one mesh cell thick. Thin sheet obstructions work fine as flow barriers, but other features are fragile and should be used with caution. To prevent FDS from allowing thin sheet obstructions,
set THICKEN_OBSTRUCTIONS=.TRUE. on the MISC line, or THICKEN=.TRUE. on each OBST line for which the thin sheet assumption is not allowed.

Brad Casterline

unread,
Aug 13, 2014, 3:06:36 PM8/13/14
to fds...@googlegroups.com

Jason, I thought the OBSTs on the Mesh boundaries created by default were different than other zero thickness OBSTs because we can put vents on them, but we cannot put vents on zero thickness OBSTs in the interior of the MESH. Is it a good idea to thicken the boundary OBSTs also?

 


From: dr_jfloyd [mailto:drjf...@gmail.com]
Sent: Wednesday, August 13, 2014 1:50 PM
To: fds...@googlegroups.com
Subject: Re: [fds-smv] Re: Smoke Evacuation Issue

 

7.2.2 Thin Obstructions

dr_jfloyd

unread,
Aug 13, 2014, 3:31:07 PM8/13/14
to fds...@googlegroups.com
In general I would recommend not putting a thin OBST on the boundary. Just use VENT and follow the VENT precedence rules in the guide.   If you have already defined the boundary with a VENT and then try and place a thin OBST with a different SURF, the initialization logic is messy.  Better to just use VENT.

Brad Casterline

unread,
Aug 13, 2014, 3:40:45 PM8/13/14
to fds...@googlegroups.com

Ok, thanks. And I do not use &VENT SURF_ID='OPEN' anywhere anymore :)

Brad Casterline

unread,
Aug 13, 2014, 4:06:51 PM8/13/14
to fds...@googlegroups.com

CORRECTION! unless it is MB. In v5 it was common to leave the thin boundary OBSTs and then make doors and windows with SURF_ID='OPEN'.

 


To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/8727E745E6714510BEC3BD55368C0208%40office.local.

Hindborg

unread,
Aug 14, 2014, 10:50:04 AM8/14/14
to fds...@googlegroups.com
Hi guys,

I really like the addition of the flow device type '... FLOW WALL' for measuring the flow through the mechanical exhaust system, which was new to me too. I am still, however missing the functionality for heat and mass flow like the original device 'VOLUME FLOW', which works both on a solid surface and in the gas. This would come in handy in situations where I'd like to know the flow through a door or a window, which initially is closed (blocked by an obstacle) and then opens upon detection of smoke. Is there a chance that this is something you would consider adding at some point? Maybe along with '... FLOW WALL +' and '... FLOW WALL -'?

All best,
Henrik

Randy McDermott

unread,
Aug 14, 2014, 1:59:30 PM8/14/14
to FDS and Smokeview Discussions
I am a bit confused by the question of the heat and mass flow at the wall.  We have heat, mass, and volume for both gas and wall devices.

We can add + and - versions for the wall.  I added an Issue https://code.google.com/p/fds-smv/issues/detail?id=2225

Can't you do what you want with 2 devices that start and stop when the obst disappears?  




--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.

Hindborg

unread,
Aug 15, 2014, 7:33:32 AM8/15/14
to fds...@googlegroups.com
Hi Randy,

My question is based on the fact that a 'VOLUME FLOW' device works both in the gas as well as on a solid surface, whilst no device that measures heat or mass flow has this dual functionality.

I'm not quite sure how you would suggest to turn devices on and off?
I can put a 'MASS FLOW WALL' device on the surface of my OBST, which then just measures 0 after the OBST disappears, but FDS refuses to run, if I place a 'MASS FLOW' device in the same spot, even though I have a CTRL_ID on the device line, which is off until the OBST disappears.

I have made a simple test of the problem. If the the last device line is commented, it will not run.

&HEAD CHID='Test' TITLE='Testing control of FLOW-devices' /

&TIME T_END=40 /
&MESH IJK= 20 10 10 XB= 0 4 0 2 0 2 /

&SURF ID='FAN' VOLUME_FLOW=-2. COLOR='BLUE' /
&VENT MB='XMIN' SURF_ID='FAN'  /
&VENT MB='XMAX' SURF_ID='OPEN' /
&VENT MB='YMIN' SURF_ID='OPEN' /
&VENT MB='YMAX' SURF_ID='OPEN' /
&VENT MB='ZMIN' SURF_ID='OPEN' /
&VENT MB='ZMAX' SURF_ID='OPEN' /


&DEVC ID='TIMER_OBST' QUANTITY='TIME' SETPOINT=20 INITIAL_STATE=.TRUE. XYZ=1,1,1 /
&OBST XB= 1.8 2 0 2 0 DEVC_ID='TIMER_OBST' /

&DEVC XB= 2 2 0 2 0 2 QUANTITY='VOLUME FLOW' ID='V_FLOW' /
&DEVC XB= 2 2 0 2 0 2 QUANTITY='MASS FLOW WALL' ID='M_FLOW_WALL' /

&DUMP DT_DEVC = 1.0
DT_CTRL = 1.0
DT_SLCF = 1.0 /
&SLCF PBY=1 QUANTITY='VELOCITY' VECTOR=.TRUE. /

&DEVC ID='TIMER_DEVC' QUANTITY='TIME' SETPOINT=20 XYZ=1,1,1 /
&CTRL ID='CTRL_DEVC' FUNCTION_TYPE='ANY' INPUT_ID='TIMER_DEVC' INITIAL_STATE=.FALSE. /
DEVC XB= 2 2 0 2 0 2 QUANTITY='MASS FLOW' ID='M_FLOW' DEVC_ID='CTRL_DEVC' /

&TAIL /

I hope you can help.

/H

Randy McDermott

unread,
Aug 15, 2014, 11:17:59 AM8/15/14
to FDS and Smokeview Discussions
Henrik,

Looks like we'd have to do a bit more with the control logic for measurement devices to do what I suggested.  Sorry about that.

For the time being I have changed the code to allow the gas devices to be placed near the wall, but a value of 0.00 is returned if the wall is present.  This allows for measurement if the OBST disappears.

It is important to remember that the flow variables in FDS are staggered: velocity components live on faces and scalars (density, mass fraction, temperature) live at cell centers.  This is why VOLUME FLOW does not care if it is solid or gas.  But for scalars we have to do an interpolation.  We do not have a reliable value on the inside of the solid.  This is why the WALL variant is a fundamentally different measurement---we output precisely the wall values used as the boundary conditions.

R


--
You received this message because you are subscribed to the Google Groups "FDS and Smokeview Discussions" group.
To unsubscribe from this group and stop receiving emails from it, send an email to fds-smv+u...@googlegroups.com.
To post to this group, send email to fds...@googlegroups.com.

dr_jfloyd

unread,
Aug 15, 2014, 11:26:34 AM8/15/14
to fds...@googlegroups.com
As a work around until a new version is released, put your removable OBST one cell away from the HOLE.  Have the HOLE be there for the entire simulation.  This way there are gas cells for the device.  

Like below where G is a gas cell, W is the wall that is always there, D is a gas cell where you are placing your measurement device, and O is the wall you remove.  

G G W G G
G O D G G
G O D G G
G G W G G

Hindborg

unread,
Aug 18, 2014, 3:36:05 AM8/18/14
to fds...@googlegroups.com
Randy,

The changes you're making allowing the gas devices to be placed near a wall sound great. That would allow me to do exactly what I wanted.

Jason,

Thanks. That's what we came up with too, when having to work around it.

/Henrik
Reply all
Reply to author
Forward
0 new messages