On Wed, 26 Apr 2006 13:06:44 -0700, Grant Erwin
wrote:

I just ran into my old boss. He is now retired, and an avid gardener. He put in
a 12-zone automated watering system, and there is a problem in it which is
driving him bonkers. I'm posting this because you guys know everything worth
knowing.

His water system starts at the meter, then it runs to a tee, and off of one leg
goes his entire outside watering network. The other leg of the tee goes to his
house. Where the water enters his house, he has a pressure reducing valve. On
the outside watering leg, there is a backflow prevention valve to ensure that
even in the event of negative city water pressure his gardening water can never
wind up in his neighbor's water glass.

At 4AM his watering system turns on. When the first zone valve cuts on, the
water pipes *in his house* begin hammering loudly. Eventually it dies down and
then the first zone valve cuts off and the second one cuts on, no more problems.
But the water hammering wakes him and his family up every morning, and they
worry about it damaging their house piping because it sounds so violently loud.

They have had the water department guys out to check his pressure reducing
valve, his backflow prevention valve, and the water pressure everywhere, and
they say everything checks out OK. He has had two different irrigation
consultants come out and other than suggesting things like reprogramming to try
a different zone valve to come on first (didn't help) they were similarly unable
to come at the root cause of the problem, nor could they suggest a workaround.

I don't really understand the phenomenon of water hammering, but I do understand
that it is a pressure oscillation which is characteristic of an underdamped
mechanical system. I suggested that he try adding resistance (some kind of flow
reducer, maybe a gate valve) or capacitance (one of those bulb thingys) but he
is on fire to find the actual root cause and solve it, not just find a
workaround. I suggested he try shutting the gate valve, the main water shutoff
valve to his house, about 95% tonight to see if the added resistance might do
the trick.

Anyone got any bright ideas? This guy is no dummy; he was an engineering manager
and has a EE degree from Rensellaer Polytechnic ..

Grant Erwin
Kirkland, Washington

I set up and played with a SPICE simulation (EE circuit simulation
program) for a while, using transmission lines with various
impedances and mismatched load and driving-point Z's, switches to
simulate timed valves, and a diode for the checkvalve. I didn't try
to simulate his pressure regulator. I do see some pressure
variations due to reflections in mismatched lines from the transient
event of first valve opening, in some cases over 50% of supply
pressure peak-to-peak that rattles for a second or two. Adding some
capacitance (damping) definitely helps.

The event where one valve closes and another opens doesn't generate
much of a transient. This is because I simulated those lines as
having fairly well-matched load impedances, my logic being that they
are sprinkler lines hence lines with distributed loss. The primary
oscillations in this simulation are in the line from the meter to
the T and in the house plumbing with the largest pressure excursions
occuring in the house plumbing.

This simulation is far from accurate, just behavioral in nature as a
matter of curiosity. Zo of a transmission line is sqrt(L/C) where L
and C are series inductance and shunt capacitance per unit length.
These would correspond to contained water mass and elasticity of the
pipe per unit length. Voltage is analogous to pressure, current to
flow rate. I included no loss term corresponding to pressure drop
per unit length as fn of flow rate.

I would think that the pressure regulator would keep pressure
variations from entering the house, but it may not respond quickly
enough to deal with transients and oscillations. It being a
mechanical closed-loop feedback control system, it could actually
exacerbate them at certain frequencies. I strongly doubt that the
water department tests transient response or frequency response of
pressure regulators. Root locus plot? "Uh, yeah, they work fine
if there ain't any roots locusated so they clog the pipes."

Without any regulator I do see the largest excursions in the house
plumbing, probably because it is like an open line having infinite
termination impedance. The largest pressure excursions are, not
surprisingly, at the far end. If that's near the bedrooms and
there are some bends in that region with pipes that aren't rigidly
secured, I could imagine that it could get rather noisy.

Possible remedies:

1. whut you said -- restriction i.e. damping. That may not be
desirable because of loss of irrigaton volume flowrate.

2. whut others have said: capacitance, as a standpipe or small
bladder tank.

3. A first valve that opens slowly -- perhaps a motor-driven valve
rather than a solenoid valve. You're a machinist, he's an EE, might
be an interesting little project using surplus parts. I'd put this
in series with a solenoid valve that could be shut abruptly when
second valve abruptly opens. That event doesn't seem to create much
of a transient either in my sim or his experience.

FYI. Your EE neighbor may find this interesting, if not
particularly informative or helpful.