Baffling EVAC problem

452 views
Skip to first unread message

markb...@ozemail.com.au

unread,
Apr 15, 2015, 7:30:17 PM4/15/15
to fds...@googlegroups.com

Timo I think is still on holidays as he hasn't responded as yet, so are there any others out there that know how to troubleshoot EVAC input code . In the combined model Fire and Evac, the fire part is working like clockwork, but unfortunately the evacuating agents were not doing as I had expected. So I isolated out the EVAC input code to see what was causing the strange evacuation behaviour. I have attached my latest version of that portion of the input code. I did tweak the evac model further and at one stage, I did have it running on its own but again still not doing what I expected it to do. With further tweaks I have now broken it completely. The trouble is I do not know what this error is referring to, as there are no vents in the model, unless vents = doors but all doors have the correct IOR from what I can see. 

Could you please help me – I think I am so close to resolving this issue, but have spent too many hours trying to troubleshoot this problem. The strange thing is that I have never had much problem with setting up fire models but the fds evac module, although a powerful addition to fds are in my opinion so finicky to get right.  I think I probably too close to the problem to spot why this logjam is occurring.

What I am trying to model is an evacuation scenario that has a common escape corridor that goes to stairs at one end of that corridor then exits out to an escape corridor on ground level, were the occupants walk to escape outside the building.  These agents should exit their rooms and then move to the stair at the end of the corridor and then down the stair and then out through the escape corridor at ground level. At one stage I was also getting agent oscillations in the corridor. This phenomena seems to have settled down so I think it is something to do with the way I set up the stair portion of the model but from what I see all stair doors do have IOR in the correct orientation. So unfortunately the problem may be somewhere else?

Kind regards, Mark

EvacOnlyModelComplexChemFEDV4.fds

TimoK

unread,
Apr 16, 2015, 5:27:55 AM4/16/15
to fds...@googlegroups.com
Hi Mark,

I just downloaded your example. I try to run and see it today. But I am just able to see it
today. I have no time to analyze anything. A commercial project fills my working hours
this week. I have to earn some money to my employer, you know...

Nobody is paying anything to me (nor to my employer) regarding FDS+Evac right now.
This means that no FDS+Evac related project is going on here right now. We are a
project organization.

Wbr,
TimoK
 

Mark Bucktn

unread,
Apr 16, 2015, 6:38:46 AM4/16/15
to fds...@googlegroups.com

Hi Timo;

 

Thank you so much you are my life saver. I suspected something was amiss as I did hear some rumour that EVAC may be coming to an end which if true is very sad for the Fire Safety Engineering community. As soon as I have some cash I don’t mind paying something for your time because you have been so go to me. Just let me know how I can do this and if it is appropriate. I don’t want to jeopardise your employment in any way.

 

Kind regards, Mark

--
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/xFTpm8E8jL8/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.
To view this discussion on the web visit https://groups.google.com/d/msgid/fds-smv/69e6cbbb-a672-4309-a682-f10a8b1e14fc%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

TimoK

unread,
Apr 17, 2015, 6:14:05 AM4/17/15
to fds...@googlegroups.com
See the attached PNG images and the fds input files. I have set the z range of the evacuation
meshes to be different. And I have removed all your doors, exits, evhos, evacs, and evsss.
T_END=0 so you see the geometry. I had to put some HOLEs with EVACUATION=.TRUE:
so that I was able to put an exit to each evacuation mesh. Every evacuation mesh should
have at least one exit or door. If not => error message is shown and FDS is stopped before
the geometry information is written to the Smokeview file. Hence, you need to be able to
run nicely the case for T_END=0 so that you can see the geometry in Smokeview.

The problem is your evacuation mesh geometries. You should change the z coordinates in
the MESH lines so that you get nice "floor plans".

http://www2.vtt.fi/proj/fdsevac/fdsevac_documents.jsp?lang=en

And there "How to debug your input file: Readme_InputErrors.txt":

http://virtual.vtt.fi/virtual/proj6/fdsevac/documents/Readme_InputErrors.txt

Wbr,
Timo

ComplexChemMesh1_s0000.png
TimoEvacOnlyv4T2_s0000.png
TimoEvacOnlyv4.fds
TimoEvacOnlyv4T2.fds

markb...@ozemail.com.au

unread,
Apr 17, 2015, 8:00:17 AM4/17/15
to fds...@googlegroups.com
Hi Timo;

Thank you so much for your reply and help but I am still confused as to your explanation.  What is the specific problem with my file?  Since even your uploaded corrections don't seem to be working as I intended them to work?

Question 1: Is the problem simply that I didn't have EXIT namelists at each EMESH?
Question 2: If so why then did you take all the doors out from the rooms?
Question 3: If a DOOR is say 2 m in height and the EMESH grid is say 3 m in height then according to what I understand that should be OK?  However you seem to be saying that my EMESHs are not correct - what is not correct?
Question 4: The error message complains about IOR but all my IORs are in place why then is there an error or is this a false positive as the real problem is that there are no EXIT name list on each mesh?
Question 5: Why are EMESHES only one cell high i.e. asymmetrical in my case I have a delta_x=0.2 m and delta_y = 0.2 m but according to how I understand EMESHs  can only be 1 cell high - in my case delta_z= 3 m so we would have asymmetrical EMESH grid 0.2 mx 0.2 m x 3 m. why is this schema so different to fire meshes which like symmetrical meshes? Or should all EMESHs be only deltax,y,z=0.2 i.e. symmetric but if so how does agent get through something only 0.2 m in height?    
Question 6: What are nice "floor plans" is not a door within the z-bound of its EMESH nice?  According to what I see I in my file everything looks nice?
Question 7: Are you aware that EVAC will not work with the ISO surface namelist when running a fire simulation that also has Evacuation?

Kind regards, Mark

TimoK

unread,
Apr 20, 2015, 4:30:27 AM4/20/15
to fds...@googlegroups.com
Below are my fast answers to your questions,
TimoK


Question 1: Is the problem simply that I didn't have EXIT namelists at each EMESH?
 
One should have (at least) on EXIT (or DOOR with TO_NODE set) in each of the
evacuation meshes given in the FDS input file. (The main evacuation meshes).



Question 2: If so why then did you take all the doors out from the rooms?
 
These (or some of these) were generating ERRORs. This means that the programme
stopped before it wrote the geometry information to the hard drive (to be seen in Smokeview).
This is the general way of debugging your inputs. Take all things away and see, if the very
simple case is working. This way you see your geometry. Then, add the DOORs (and other stuff)
one by one. I.e. Add one EXIT and see, if you get an ERROR message or not. This way you can
easily see, where you get problems.



Question 3: If a DOOR is say 2 m in height and the EMESH grid is say 3 m in height then according to what I understand that should be OK?  However you seem to be saying that my EMESHs are not correct - what is not correct?
 
Your evacuation mesh XBs may not be correct. Their z-range is quite large. This might be good or bad for you.
If you have some floor or ceiling that touches the z-range of your evacuation mesh => this part of the
floor area is not good for the agents. It is filled with solid obstacles totally. See the figure 34 (page 74)
in the manual.


Question 4: The error message complains about IOR but all my IORs are in place why then is there an error or is this a false positive as the real problem is that there are no EXIT name list on each mesh?

The real problem was not that there was no EXIT in each mesh. The problem is that some of your EXITs or DOORs
are having coordinates (XB) such that the XB is inside some OBST. This is an input error. Or you door/exit is facing
the OBST in the wrong direction (IOR is wrong). But to see this, you need to be able to generate a CHID.smv file,
see the answer 2 above. Activate the grid in Smokeview (and activate the showing of grid locations also), so you
can see, where your OBSTs are in Smokeview.

 
Question 5: Why are EMESHES only one cell high i.e. asymmetrical in my case I have a delta_x=0.2 m and delta_y = 0.2 m but according to how I understand EMESHs  can only be 1 cell high - in my case delta_z= 3 m so we would have asymmetrical EMESH grid 0.2 mx 0.2 m x 3 m. why is this schema so different to fire meshes which like symmetrical meshes? Or should all EMESHs be only deltax,y,z=0.2 i.e. symmetric but if so how does agent get through something only 0.2 m in height?    

The evacuation mesh IJK and XB are totally different than fire meshes. The evacuation mesh is just a
floor plan, that is two dimensional drawing. But the internal form inside the programme code for evacuation
meshes are the same as for fire meshes. So, you should input also the z-direction. But just one grid cell
in the z-direction is enough. The z-direction is not used in the computations where the guiding flow fields
towards the doors is calculated. But for the internal structure of the source code, one grid cell in the z-direction
is needed (the arrays have three spatial indices: x, y, and k (i,j,k).

The z-range of your "one grid cell in the z-direction) should not be the entire height of your floor, see
the figure 34 (page 74) in the manual. The z-range could be 0.00000000000001 m so that you will
get "a real two dimensional floor plan". But the z-range should chosen so, that you will get most
of the obstacles for movement in the floor plan "automatically". See the figure in the manual.


Question 6: What are nice "floor plans" is not a door within the z-bound of its EMESH nice?  According to what I see I in my file everything looks nice?

See the above comments. Pring your geometry floor by floor in Smokeview (make png pictures and print them).
Use dialog "clip geometry" just to show one floor at a time. And move the viewing angle so that you are seeing
the geometry like a bird (directly above). See, if the prints look like "ordinary building floor plans". So that you
see basically just the walls and similar things.

 

Question 7: Are you aware that EVAC will not work with the ISO surface namelist when running a fire simulation that also has Evacuation?


Well, this I do not know. I should test this. I should put ISO namelist and see how it works. Nobody has ever
reportet this bug yet. I do not have all kinds of fire outputs in my simple evacuation verification test cases.
Thanks for pointing this out.

TimoK

upad...@gmail.com

unread,
Apr 20, 2015, 6:02:07 AM4/20/15
to fds...@googlegroups.com
Dear Timo and Mark,

I have the same problem as Mark has stated in 7. I have a working evacuation model but don't seem to be able to write the iso-surface visibility either. 

I am trying to put together a metro-station evacuation video, to show evacuation and the 10 m visibility iso-surface. As this has not been possible , I have run  a separate fire -only model with no evacuation mesh to get the iso-surface. I am now  thinking of a way to combine the sets of files/images together into one video to get round to the problem.

Can you think of a better way for me to do this for the time being? Even better, is this something that can be fixed soon .Unfortunately, time is not on my hands.
I am also exploring on whether to use iso-surface from particle function to get round this. I have tested that the fds+evac model using this and it works although I don't know how to then show the 10m visibility. Lots of question relating to this one problem. 

Please Help. 

Regards
San

Mark Bucktn

unread,
Apr 20, 2015, 6:11:12 AM4/20/15
to fds...@googlegroups.com

Hi Timo;

 

Thank you for your response to my questions but I did clear up these obstructions by adding holes.  I also then simplified the problem as recommended. Now, however, I am only going from the 3 rd floor to the second floor doors and out the new exit on that level which you recently added. As you see FDS still generates the error I described previously. So in fact I have narrowed down the problem. My hunch is that it relates to traversing from one mesh to the other.  In my case 'EMesh_3' to 'EMesh_2'. When I remove stair doors on 'EMesh_2' all works so the room doors on level 3 are all OK in my view. The stair door definitions are these:

 

&DOOR ID='Door_out_2', IOR=+1,

      EXIT_SIGN=.TRUE.,

      TO_NODE= 'EMesh_2'

      XYZ= -0.2, 3.60, 4.6,

      XB=0.0, 0.0,   2.5,4.8,   4.5, 7.5, /

&DOOR ID='Door_in_2', IOR=-1,

      EXIT_SIGN=.FALSE.,

      TO_NODE= 'Door_out_3'

      XYZ=  0.2, 3.60,  4.6,

      XB=0.1, 0.1,   2.5,4.8,   4.5, 7.5, /

 

So as you can see the IOR is Ok and they are not imbedded in obstructions and the TO_NODE definitions appear OK. See my latest version which is attached. So why is it not working?

 

Kind regards, Mark

--

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/xFTpm8E8jL8/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.

EvacOnlyModelComplexChemFEDV6.fds

TimoK

unread,
Apr 21, 2015, 4:36:39 AM4/21/15
to fds...@googlegroups.com
The ISO surface problem you can bypass by making a separate fire calculation without evacuation part.
Set, for example, NO_EVACUATION=.TRUE. on the MISC namelist. You get the files on the hard drive.
And you get the CHID.smv file that you can edit in a text editor.

Run the same case in an another directory (folder) and run the fire+evacuation case.
Copy the fire only isosurface file to this directory and edit the CHID.smv file in this
directory. You should copy from the fire only CHID.smv file the parts that correspond
to ISO files. This you should be able to figure out. And set the file names there
properly for the iso files.

TimoK

TimoK

unread,
Apr 21, 2015, 5:13:45 AM4/21/15
to fds...@googlegroups.com
Mark,

See the attached png images. They are for the mesh at z level 4.5, 7.5,.

I have change you file so that I took all EVSS away => easier to see the geometry.
And I took all of your doors away (the "&" away). And then I did as below:

 Timo: TO_NODE should be some DOOR or ENTR, not a mesh.

 DOOR ID='Door_out_2', IOR=+1,
      EXIT_SIGN=.TRUE.,
      TO_NODE= 'EMesh_2'
      XYZ= -0.2, 3.60, 4.6,
      XB=0.0, 0.0,   2.5,4.8,   4.5, 7.5, /
DOOR ID='Door_out_2', IOR=+1,
      EXIT_SIGN=.TRUE.,
      TO_NODE= 'Door_in_2'

      XYZ= -0.2, 3.60, 4.6,
      XB=0.0, 0.0,   2.5,4.8,   4.5, 7.5, /

 Timo: we change next door to EXIT, so that it is easier to debug, no to_node needed.
 Timo: => ERROR: Specify orientation of VENT   2, MESH NUMBER  3
 Timo: This is the only DOOR/EXIT now that you have (well, the above three EXITs
 Timo: defined by me close to the mesh lines are also present so that on can get
 Timo: a smv-file.
 Timo: So, I ran the case without the EXIT line below, and got a smv-file and there
 Timo: Next step: Open in Smokeview. Take the exterior surface away (show/hide, geometry,
 Timo: surfaces, exterior). Nest DIalogs => Clip geometry and there just the z=4.5, 7.5,
 Timo: exit floor is chosen. Then activate grid locations and activate grid.
 Timo: Next: move the x and y "grid plane pointers" to XB=0.0, 0.0,   2.5,4.8,

 Timo: Next holes are needed. Your have lots of stuff inside the stairs in the
 Timo: evacuation geometry.
 Timo: Note: you need still some other holes for this staircase. See the png
 Timo: image and you notice that there is an OBST at the right landing at
 Timo: the middle (OBST is shown as outline there). You should get rid of
 Timo: this also. And see, if you centere core of the stairs is having
 Timo: correct size also.

&HOLE EVACUATION=.TRUE. XB=-2.5,6.0, 2.5,4.8, 4.4, 7.6,/
&HOLE EVACUATION=.TRUE. XB=-2.5,6.0, 0.0,2.3, 4.4, 7.6,/


DOOR ID='Door_out_2', IOR=+1,
      EXIT_SIGN=.TRUE.,
      TO_NODE= 'Door_in_2'

      XYZ= -0.2, 3.60, 4.6,
      XB=0.0, 0.0,   2.5,4.8,   4.5, 7.5, /

&EXIT ID='Door_out_2', IOR=+1, COLOR='RED',

      XYZ= -0.2, 3.60, 4.6,
      XB=0.0, 0.0,   2.5,4.8,   4.5, 7.5, /

DOOR ID='Door_in_2', IOR=-1,
      EXIT_SIGN=.FALSE.,
      TO_NODE= 'Door_out_2'

      XYZ=  0.2, 3.60,  4.6,
      XB=0.1, 0.1,   2.5,4.8,   4.5, 7.5, /

Do similar checks for other floors also.

TimoK

Timo_FEDV6_s0000.png
Timo_FEDV6_s0000_error.png
Reply all
Reply to author
Forward
0 new messages