Agent speed effect from a narrowing of a passage way

[Kristofer Klefbohm]

Hi,

If I use a default agent speed in FDS+EVAC of 1.25 ±0.30 m/s and I have a narrowing of the passageway, do I then need to adjust the agent speed in my code or is it automatically done?

Assuming I would have a 900 mm wide passageway, and then a 10 m long passageway with an obstacle that makes it 700 mm wide. I am aware that a passageway of 700 mm can be considered not to fulfill 848/2017, but it is possible to simulate. If it is only 600 mm wide the agents get stuck.

 

 Thank you for your help

[TimoK]

Hi, sorry for a late answer, I have been too busy, but now couple of DLs have gone. The speed (the unimpeded walking speed v0,i) is not changed according to the passage way width. If the passage way is narrow, it might slow down the agents, when the agents shoulder circles are (or almost) touching the walls. This generates, e.g., friction forces. But if you want to reduce the speed in a narrow passage way, you should do this explicitely in the input file, use &EVSS namelist and give the FAC_V0_HORI (check the manual for the correct name of the keyword) parameter a value less than one. And set the XB to be the area, where the agents should slow down.

And as you said, if the passage way is just 60 cm wide, then the agents (at least the largest ones) get stuck. So, it is nice to have at least, say, 80cm wide movement spaces for the agents. If there are narrower ones that should be used, then modify the geometry. Move the OBSTs and/or create HOLEs so that you get at least 80cm wide corridors etc. Well, it is nice to set the XB and IJK of the evacuation meshes so, that the dx and dy are "nice numbers", like 0.2, 0.3 0.4, 1.0, or so. The evacuation mesh dx,dy does not affect the accuracy of the agent coordinates, the coordinates are continuous (double precision floating point numbers in Fortran). But the OBSTs are in the grid dictated by the XB and IJK. I like to use dx and dy values 0.2, 0.3, and 0.4 typically. Too small dx,dy, like 0.1, is not too good, it might produce some artifacts in the guiding flow field calculation for the agents. With dx=0.3m you can easily get 0.9m and 1.2m wide doors, that are typically found here in Finland (due to the building code, of course).

If your dx,dy does not match nicely your geometry everywhere, then you can add EVSS and reduce the speed, where you have

to use larger width of the passage ways due to the dx,dy values.

TimoK

[Kristofer Klefbohm]

Hi, thank you Timo