Re: Emergency Water Turn-Off?

["Robert Green" <ROBERT_GREEN1963@xxxxxxxxx>]

"Prof Reid" <reid@xxxxxxxxxxx> wrote in message
news:d5o0of$2vjm$1@xxxxxxxxxxxxxxxxxxxxx
> Anyone made their own emergency water-turn-off
> system?  Like a WaterCop or WaterBug?
> Something like a ball valve with an attached
> drive wheel and motor to turn the water off.

No, but I've looked into this before and been sent to Grainger's to look for
automated valves.  IIRC, they are available for $25 to $50 but it's been a
while since I've researched this particular HA facet.

I'm going to install one of those valves just to be able to shut off the
watermain without have to go downstairs.  There are a lot of issues to
consider.  The valves I looked at were 24VAC.  For a leak system to be
reliable, it would have to work during a power failure.  That alone makes
the problem considerably more complex than appears at first blush.  You
might be able to design a system with a solenoid valve that always closes
upon power loss, but that means another 24x7 current draw and the
probability that the valve will start humming when things loosen up.

> Also have you tried to get water flow information
> from acoustic transducers on water pipes?  I can
> "hear" the water flowing inside a pipe with a
> stethoscope, but I wonder if I can get quantitative
> enough flow information from a strapped-on "mic".

Again, if I recall correctly there's already a commercial product that did
this.  But when I started to set down and really think about it, the only
time you want an acoustic transducer to trigger a shut-off is when no one is
home.  Even then, a mike would cause a system to shutdown in any empty house
because an icemaker caused a brief water flow.  Evaporation of water in the
toilet tank will trigger water flow if the float drops below a certain
level.

(Speaking of toilets I saw Smarthome flogging a new toilet overflow sensor.
It's a device that I am sure has one of the lowest Spouse Approval Factors
in the world because part of it sits within the bowl itself and the other
part gets hung from the shutoff valve.  As a person who has faced some
significant opposition to various HA gizmos, I think wiring up the toilet
just won't be popular with the ladies.  Back to your problem.)

> I'd like to use the HA computer, multiple-acoustic sensors,
> and an smart algorithm to watch for the acoustic patterns
> that could uncover a running toilet (or other undesirable water flows),
> and turn the water off.  This flow monitoring, as opposed to the usual
> alternative of looking for undesirable water accumulations.

The problem is manifold.  Water pressure varies in a typical home based on
concurrent water usage by neighbors, the amount of water in the local water
tower, etc.  That means that the sounds created by various fixtures would be
different depending on the water pressure and depending on whether someone
was operating two different fixtures at high flow rates.

I agree that by the time a sensor detects that your basement carpet is wet,
it's too late.  But the real rub lies in separating the sounds of legitimate
use from a leak.  Are you going to be able to design a sound processor
that's able to separate all the variants of a toilet flush or a dishwasher
run and act accordingly?  I really doubt it.

I had a washing machine hose rupture recently.  Fortunately I was home.  But
I can't think how I could have differentiated that event from normal water
use without moisture sensors of some sort.  There's just too broad a range
of events to trap for with an acoustic sensor to make it really workable.
At least that's MHO.

With all that said, I know that there are commercial leak detectors that
operate in the ultrasound range.  But I think all of the same  problems
listed above apply.

--
Bobby G.