Fire suppression systems

I've been recently hearing from a number of people that "dry pipe" fire
suppression systems were not being recommended any more and that they
were bad because they have been known to fail and not function.  Well we
recently had a meeting (about other things) with Bill Lull and we got
side tracked onto fire suppression systems.  I have to admit that I
don't know much about them (can't be an expert in everything), but I
learned some interesting things.

There are apparently three types of systems that people have been
calling "dry pipe".  The first one is called a "Deluge" system.  This
one uses open sprinkler heads and empty piping in the storage area.  At
the first sign of fire, a valve is opened and the whole room gets soaked
(hence the term "Deluge".)  Bill pointed out that with this system, a
cigarette in the wrong place could be disastrous. The second type is
the real dry pipe system in which all sprinkler heads are sealed and the
pipes are empty. Apparently there is also compressed air used to hold a
flapper valve shut. Where there is fire, the sprinkler heads are opened
(seals melted) and the flapper valve opens thus only the sprinklers in
the area of the fire are open.  Failure may result from loss of
compressed air, but it only means that the sprinkler pipes fill up.
Since the heads are sealed, nothing else happens.  The third type is the
type most often called dry pipe and it is the "Preaction System".  In
the Preaction system, there are fused or pneumatically sealed heads. The
pipes are dry and the preaction valve must be opened to fill them.
Apparently the system uses electronic detectors and valve control
system.  Smoke detectors are put in the storage area in overlapping
zones with 2 detectors per zone.  Apparently *BOTH* detectors must go
off for the system to activate thus is one smoke detector is not
functioning, neither is the sprinkler.  This system also uses compressed
air.  Failure may occur if (a) the pipes are already filled with water,
(b) there is a low pressure condition present (open fused heads,
compressor failure, etc.)  In addition, since the whole system is
computer controlled, it must be verified that what the computer calls
Alarm A really is Alarm A. (Hardware and software must agree.)  Finally
there must be a manual release that is easily accessible.  Bill had
stories of manual releases in the storage area (you have to walk through
the fire to get to the manual release) and in locked maintenance rooms.
One other cause of failure in these systems is from rust.  They are
apparently prone to building up rust inside.  The fire marshals often
require a demonstration to prove that the system is working (ie flooding
the pipes.)  The flooded pipes now contain a mixture of water and rust
sludge that wanders back to the preaction valve and gums it up.

>From what Bill said, it sounded like the preaction systems are the ones
that are failing and not the true dry pipe systems.  Although I'm a
little worried about wet pipes in collections, Bill said that he has
only seen leaks in wet pipe systems from frozen pipes.  There was one
case of a very slowly dripping elbow, but he said that it was easily
fixed and posed not problem.  There was also one case that he had seen
in which the sprinkler heads were knocked off (causing a flood).  This
storage vault had a ramp to move materials in and out of a basement.
The drop (false) ceiling was also sloped, but not quite as much as the
ramp was (it was apparently not easy to see without actually measuring
the floor to ceiling height at the top and bottom of the ramp.)  Of
course, one day a large object was being wheeled up the ramp (pushing in
every ceiling panel along the way and) knocking off all the sprinkler
heads above the ramp.  Bill did not feel that sprinkler pipes were prone
to leaking like (potable) water and sewer pipes were although I didn't
find out why.

I hope that I got everything right.  Bill was talking and I was
furiously scribbling down notes.

-Doug
Image Permanence Institute

                                  ***
                  Conservation DistList Instance 7:35
                Distributed: Saturday, October 23, 1993
                        Message Id: cdl-7-35-002
                                  ***