Smoke detector activation anomaly

[Matt Freeman]

Greetings all,

I have been recently been comparing smoke detector activation times
for a number of different t-squared fires and ceiling heights. My
issue is that in some instances the results indicate that a smoke
detctor on a high ceiling will activate prior to a smoke detector on a
lower ceiling. For example for an ultra-fast t-squared fire with a
detector radius of 14.1m the model with a 5m high ceiling yeilds an
activation time of 50.5 seconds, whilst the model with an 8m high
ceiling yeilds an activation time of 47.2 seconds. The model is simply
a burner, a detector (see below) and a flat ceiling and the only
difference between the models is the height of the ceiling.

&PROP ID='Cleary Ionization I1', QUANTITY='spot obscuration',
ACTIVATION_OBSCURATION=8.00, ALPHA_E=2.50, BETA_E=-0.70, ALPHA_C=0.80,
BETA_C=-0.90/
&DEVC ID='Cleary I1', PROP_ID='Cleary Ionization I1',
XYZ=10.00,10.00,4.95, LATCH=.FALSE./

Cleary these results are nonsense. I have a few ideas as to why this
might occur, but I would welcome any of your ideas. More importantly,
how can i prevent this from happening?

Cheers

Mathew

[BlairS]

Are you moving the detector up with the ceiling height? e.g.
XYZ=10.00,10.00,7.95 for the second model?

Regards,
Blair

[Matt Freeman]

Yes Bair, that is correct, I am moving the detector up with the
ceiling height.

Mathew

[Chris]

You could use the single parameter Heskestad Smoke Detector model to
check the error.

&PROP ID='Acme Smoke Detector', QUANTITY='spot obscuration',
LENGTH=1.8, ACTIVATION_OBSCURATION=3.28/

Cheers
Chris

[Kevin]

I would not jump to conclusions so quickly. Yes, certainly check all
the obvious things in your calc, but then consider what FDS is doing,
and then what the Cleary or Heskestad model is doing. FDS is just
transporting smoke from the fire to the vicinity of the detector. So
FDS predicts the time history of smoke concentration and velocity
OUTSIDE of the detector chamber. So at a minimum, look at these
outputs via the DEVC utility. The velocity determines the time lag for
smoke entering the detector: dt = L/u. So, the faster the velocity u,
the shorter the time lag. Maybe the higher ceiling case has a higher
velocity and shorter time lag.

Another thing -- the detector algorithm activates when the chamber
obscuration gets above the given threshold. Maybe in the high ceiling
case you spike above the threshold for some short time.

In any case, consider the quantities that FDS calculates and feeds
into the detector algorithm before reaching a conclusion.