Real Life Water distribtuion Issue

[waseem19]

Dear All,

I'm kind of new at the water distributing design, so this question might sound silly for some of you.

Everybody models the distribution network in certain software by inputting a demand at nodes, this is really like saying that each node will have a certain head loss which will limit the demand to the required one, but is this true in real life?

If you have an intermittent supply to people tanks say 6 hours every day, so in a certain sector and just before you start feeding these tanks the tanks will be say "30% full” and the valve full open. Is there what prevents water from flowing at its "hydraulically balanced" speed ? this speed could be really high in the pipe work connecting the tanks closest to your source and the source itself.

In the real situation the tanks most closed to your source will be filled a lot earlier than the ones away, which again mean that the velocities in the pipes connecting the source and these nearby tanks are a lot higher than what is in the model.

Will the "small sized" house connections and the ball valves provide this head loss? Even so, did anybody try to actually model the people tanks as tanks not as nodes?

Is it really an issue?

 

[greggy]

Your theory is correct in that water distribution models do not necessarily compensate for the fact that what we call "demand" increases with higher system pressure, since flow from any open spigot increases (slightly) with higher pressure, and vice versa. In real life, unless a customer has a pump, he cannot draw more water from the system than the system can naturally provide to atmospheric pressure.

I don't know if there are any sophisticated models that treat every demand node as an open reservoir, thus depending on the small diameter house service and plumbing to restrict flow to what it actually would be, thus allowing this rate to vary as system pressure changes.

Personally, I believe this phenomenon exists but in the big picture is "small beans". It should be considered in cases where, say, depressurization occurs in a distribution system where the modeler is imposing large demands. For example, if during an emergency the system pressure in a 6-inch water main drops to 10 psi, a customer with a 100-foot, 1-inch service line and an imposed demand of 20 gpm probably cannot receive that full rate of water under actual conditions.

 

[hydrae]

Most models have demand modifiers in the advanced section where you can incorporate a change in demand as a function of the supplied pressure. Use of this function will dramatically increase the solution time, since it adds one more level of iteration...

A good majority of residential uses are a fix volume for a task, flushing a toilet will still use 1.6, 3.5, 5, or 7.48 gallons (depending upon the age of the toilet) whether the pressure is 20 or 80, same goes for the dishwasher, cloths washer, getting a glass of water, or filling the bathtub. Hand washing and showers are some of the few items that will change in flow rate, but some flow restrictors will even modify this rate through active restriction, (the plastic disk with the hole in the center bends towards the aerator plate under high pressure squeezing the resultant hole size)

Leaks on the other hand will be pressure dependent and should be modeled as a function of the square root of the supply pressure

Hydrae