Understanding a Data Hall Fire Detection Strategy

[Messy]

I am dealing with a large data centre and am having difficulty understanding the fire detection strategy:

The environment?s temperature is controlled by a CRAC cooling system which draws in cool air in at a low level and applies a draft from under the false floor up over the server cabinets and out of a high level grill to be cooled by radiators outside and recycled back again. (aka ?free cooling?).

There is VESDA in the server rooms with inlet pipes fitted over the air outlet grills to catch any smoke caught up in the airflow. There is also point smoke detectors & beacons in the server room. So far so good as it all makes sense.

What I don?t understand is that there are single point detectors under the false floor as well. This area is subject to quite strong airflow (compared to the large server room space above) and I would have anticipated that single point detectors would be far less efficient than VESDA in that under-floor confined space with those strong air currents.

Does anyone know if this is usual practice and if so, why isn?t the VESDA extended under the floor instead of standard point smoke detectors?

Thanks

[kurnal]

I don't know messy but I would expect there is a positive pressure under the floor relative to the space above which would require a dedicated sampling unit sampling and exhausting to the under floor area. Furthermore the pressure may vary depending on cooling load making it difficult to manage the air flows through the unit. Just a guess!

[Mr. P]

Can you ask for a copy or sight of, the fire strategy, that should have been documented prior to agreement and install. That should tell you what considerations/reasons were contemplated for the design of what has been installed.

[Messy]

Thanks for the replies

I am more than happy for the VESDA sampling the exhaust air - this seems to be standard in data centres.

It just seems that if detector heads in a room are supposed to be 1m+ away from AC vents due to excess drafts in order to comply with BS5839-1, the underfloor area of a data hall with its strong airflow seems to be an even more challenging environment!!!

I will ask some more questions re the strategy or standard they are working to.

[John Webb]

Messy - I recall latish in my fire research days that detectors were being placed under false floor areas in such rooms because there was concern about the fire load of the cables. Also because the void height was over the 0.8m maximum (I think) that voids could be without needing detection. (Sorry for the hazy details - can't locate my old copy of BS 5839)