evac flow visualization

[Luis Lencina]

Hello everyone,

i am having some trouble with agents move, there is a room that the agents dont move. In the manual to define a slice  &SLCF PBZ=x.x, QUANTITY='VELOCITY',VECTOR=.TRUE./ to see the evacuation flow, but after running 0.1s i can not see nothing else than noise in this slice. Please, can someone help me?

Thanks in advance Luis

[TimoK]

Well, you should add "EVACUATION=.TRUE." to the slice file ouput definition

&SLCF PBZ=x.x, QUANTITY='VELOCITY',VECTOR=.TRUE., EVACUATION=.TRUE./

Then run the case, for example 0.1 s, so the initialization phase of the evacuation calculation will be done and the guiding flow fields towards the exit doors should be outputted. The PBZ of your SLCF should be within the evacuation mesh z range. Load the evacuation mesh SLCF and see, if you have some vectors in your room. If not, then the room is "sealed" in the evacuation geometry. So, there is not way to any exit door. So, you should remove some OBSTs so that there is a way to walk to the exit doors. Note, the "hole in the wall" should be larget than 60cm, say, 70cm or larger is usually fine. The largest (male) agents can be wider than 60cm, so they stuck to holes that are 60cm or less.

Use lines like

&HOLE EVACUATION=.TRUE., XB=x,x,x,x,x,x /

to make the necessary holes in the geometry. This way they just go to the evacuation geometry and the fire geometry is not affected.

Note also, that add a keyword "EVACUATION_DRILL=.TRUE." on the MISC namelist. This way you will see just the evacuation geometry in Smokeview and "debugging" the evacuation geometry is easier this way.

If you see some guiding flow field vectors in your room, the there might be something wrong in your EVAC namelist inputs for these agents. The agent are not moving, if they have no target door. So, specify some known doors (at this evacuation mesh) for these agents. If an agent does not know any door on this evacuation mesh and can not see any door ==> it does not move (actually this evacuation mesh and this "evacuation zone", where "evacuation zone" is something like "pressure zone" in fire calculation). Well, sometimes the agents will go to some door, if no visible and no known doors are found. This depends on the parameters in the PERS namelist. If I remember correctly, the default parameter is such that the agents will not stand still, they will go to some door on the current evacuation zone, if they do not know or see any door.

Check also your detection and reaction times, should the agents be moving already? You can see in the text output file CHID_evac.out the initial positions of the agents, their (random) detection times, reaction times, body size, v0 velocity etc information.

TimoK

[Luis Lencina]

Thanks again Timo,

I added the evacuation and now the the vector visualization is working. I can see that the agents dont move because the vector inside the rooms are too small (i think). I'm using these parameters

&PERS ID='Male',

      FYI='Child diameter and velocity',

      DEFAULT_PROPERTIES='Male',

      PRE_EVAC_DIST=7,PRE_MEAN=5.0,PRE_LOW=1.0,PRE_HIGH=9.0,

      DET_EVAC_DIST=1,DET_LOW=30.0,DET_HIGH=60.0, /

&EVAC ID = 'Male_PA', 

      NUMBER_INITIAL_PERSONS = 50,

      FYI = '100 pers, male 20%',

      XB =  1.5, 14.0,  0.6, 14.0, 5.31,5.51

      AVATAR_COLOR = 'BLUE', AGENT_TYPE=2,

      PERS_ID = 'Male'  /

I dont know how to solve this problem. 

Thanks in advance Luis

[Luis Lencina]

I change the agent type from 2 to 1 an the agents evacuate.

[TimoK]

Well, the size of the vectors does not matter. The vectors are normalized later in the evacuation part. I.e., the evacuation algorithm uses the shown vectors after a normalization (all vectors have same length that is equal to unity). Once I modified the evacuatioin SLCF output so that the vectors were shown as unit vectors, but this was  not a good idea. There was some problems that made it a little bit tricky and not useful always. So, you can change the lengths of the vectors in Smokeview, see the Smokeview user guide how to change the vector lengths. And see also how to change the number of shown vectors (every "grid cell", every second grid cell and so on). Sometimes there "are too many vectors" to see, if you increase the length of the vectors and your grid resolution is quite fine (or your building really large). So "zooming in the vectors" in the room where the agents were not moving => you'll see that there are vectors indeed.

So, changing agent type 2 => agent type 1: Agent type 1 knows all doors, so they can move. Agent type 2 does not see any door in some places of the geometry => they do not move (if no known door given).

So, you should (usually) give the agents some known doors (in each mesh that they are using to go out of the building). Or you can give on the DOOR and EXIT namelist "KNOWN_DOOR=.TRUE." keyword, so these doors are then known to everyone. Default is .FALSE. A good tactics would set the KNOWN_DOOR true fo the main exit doors of the building and at different floors the doors leading to the "main staircase" (whatever this would be, "a staircase that is in the elevator lobby, for example". And give then on some EVAC lines some other doors also to be known (for some agents), so that also these doors will be used. If there is just one known door in a floor => all (type 2) agents will be using this. The preference is visible known door first, then non-visible known door, and then non-known visible doors (well, that holds for fire drill case, i.e., no smoke). Note: agent type 1: all doors are known => preference is to use first (closest) visible door, then non-visible doors.

TimoK

TimoK

[WP]

Excuse me.  

Is it possible to use SLCF to plot the the preferred walking direction vector field as shown in Figure 4 in FDS+EVAC User Guide?  

在 2018年7月26日星期四 UTC+8下午4:17:29，TimoK写道：

[Gregor]

Excuse me.  

Is it possible to use SLCF to plot the the preferred walking direction vector field as shown in Figure 4 in FDS+EVAC User Guide?  

Yes, it's possible.

for example: see chapter 8.1, page 75:

[...] The guiding flow fields for evacuation meshes should always be checked by using Smokeview, when building the geometry of the model. So it is a good practice to add a SLCF output (EVACUATION = .TRUE., QUANTITY = VELOCITY, VECTOR = .TRUE.) at the levels of the evacuation meshes and see these vectors in Smokeview. [...]

or search for SLCF in the user guide.

[WP]

Thanks a lot for your guidance.  

Yes.  I got the velocity vector field when the things are zoomed in sufficiently.  The vectors are very dense and small, but they are not easy to observe directly.  It will be easier to observe U, V, W with colorbar.  Is that right?  

In addition, when I added the SLCF line to an existing example, I got the following error message.  

Your comments or suggestions are much appreciated.      

Time Step:   -149,    Evacuation Initialization Time:    -1.490

At line 7945 of file ../../Source/dump.f90 (unit = 18)

Fortran runtime error: Specified UNIT in FLUSH is not connected

在 2018年8月6日星期一 UTC+8下午2:57:14，Gregor写道：

[TimoK]

Well, which version are you using? The FLUSH issue was some years ago. The present version should not have these problems. Well, at least one "flush issue" was corrected a year or so ago.

Note, that the &SLCF should have EVACUATION=.TRUE., VECTOR=.TRUE., QUANTITY='VELOCITY'. And it should have the z-level (PBZ) somewhere between the evacuation mesh z_min and z_max.

Read the Smokeview user guide (check the keyboard short cuts, so you do not need to read whole guide). You can make just every second vector to show, every third (or is the fourth, i.e., powers of 2?) to show, etc. And there is shortcuts also for the vector lengths.

TimoK

[Glenn Forney]

Appendix c has a list of shortcuts. Help menu has list also 

--
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/452c384d-99f5-4fa8-b028-b9d5be5f1cc2%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[WP]

I mainly use FDS6.5.3.  I got the above error message in two testing examples.  

One is a standard doorflow example that I downloaded.  There is only one door exit in width of 2m in a room 15*15.  I think the example was previously wrote for FDS5, and I ran it well in FDS 5, but when I tried to run it again in FDS 6.5, I got the error message.  Maybe it is a problem caused by different versions.  

The other one is an example I recently modified to test group dynamics.  I ran the example well in FDS6.5.3.  However, when I added SLCF to observe the evacuation flow field.  I got the error message.  Probably SLCF has something to do with the error.  

I supose PBZ should be within the evacuation mesh.  As below please find how I add SLCF in my existing example.  I add SLCF at the end of the input file.  Any mistake there?  If so, just let me know.  Many thanks.  

&HEAD CHID = 'PreEvac2018_Compare_EvacFlow3' /

&TIME T_END=60.0 /

&MISC EVACUATION_MC_MODE=.true.  EVACUATION_DRILL=.true. /

&MESH XB= 0.0 9.0 0.0 13.0 0.0 3.0 IJK=90 130 10 /

&MESH ID='main' XB= 0.0 9.0 0.0 13.0 0.9 1.1 IJK=90 130 1 EVACUATION=.TRUE. EVAC_HUMANS=.TRUE. EVAC_Z_OFFSET=1.2 / 

...

...

...

Next line could be used to plot the evacuation flow fields:

&SLCF PBZ=1.5  QUANTITY='VELOCITY' VECTOR = .TRUE. EVACUATION=.TRUE. / 

&TAIL /

在 2018年8月28日星期二 UTC+8下午4:59:46，TimoK写道：

[TimoK]

Well, there seems to be something that it is not really nice in your input. These things might cause problems or not:

 * EVACUATION_MC_MODE=.true.  (.true. => .TRUE. i.e., all CAPS. I do not know, how this is written in Fortran standard. To be safe, use uppercase letters.)

 * You are not using "," (commas) between different inputs within a NAMELIST line:

    XB=0.0 0.9 ...  => XB= 0.0, 0.9, .... (and same is true everywhere in you input file)

 * Your evacuation mesh is at the z-level: z_min = 0.9, z_max= 1.5. And your SLCF file has PBZ=1.5. So it is not within your evacuation mesh.

Correct these things and see if your input case works. If not, download the latest FDS version (6.7.0) and see, if it works using that version.

TimoK