Tanks, Siphons, and Weirs
Water leaving a tank moves as fast as if it had fallen from the surface.
Before this Lesson
Drill a small hole in a rain barrel, 1 ft below the water surface, and the water squirts out at 8 ft/s. Drill one 1 ft below the surface of the reservoir behind a dam: 8 ft/s. Fill the barrel with oil: still 8 ft/s. That's exactly how fast a dropped stone is moving after a 1 ft fall. Only the height matters, and nozzles, siphons, and weirs all run on that one fact.
Height turns into speed
From Continuity and Bernoulli's Equation, between two points in steady, frictionless flow with no pump between them:
Each term is a head, a height of the flowing liquid. Pressures are gauge, so a point open to the air has .
Pick two points: point 1 on the free surface, and point 2 in the jet just outside the opening, whether that's a hole in the wall or the end of a pipe.
Both points are open to the air, so both gauge pressures are zero. A jet in open air has no walls to hold any other pressure. The surface sinks so slowly that its velocity head is negligible: by continuity, a tank with 100 times the hole's area sinks at a hundredth of the jet's speed, with a velocity head 10,000 times smaller. Only height and speed are left:
That's the falling stone's speed. The liquid doesn't matter, because enters only through the pressure heads, and both are zero. The result is Torricelli's theorem, and the Handbook prints it, unnamed, under Jet Propulsion.
A pipe changes nothing. Without friction, its length and route drop out, and is the outlet's depth below the surface. The flow is through the exit, a nozzle's if there is one, and the velocity in the pipe follows by continuity. The Energy Equation adds the friction real pipes have.
A siphon climbs before it falls
A siphon is a drain that climbs first: a tube from a reservoir up over its bank and down to an outlet below the water surface. Fill it with water, and it keeps flowing on its own. The Handbook has no siphon section, and needs none. It's Bernoulli between the right points.
The flow comes from the surface and the outlet, both at atmospheric pressure: . The crest isn't in it, since the water gets back every foot it climbs on the way down.
The crest sets the pressure instead. From the surface to the crest:
Both terms are negative, so the crest is under suction. In a tube of one diameter, all along it, and the crest's pressure head works out to , minus its height above the outlet.
Nothing pulls the water up there; water isn't a rope. The atmosphere pushes on the reservoir's surface, and the water rises toward the lower pressure at the crest, like a barometer's column. That push has a limit. When the crest's absolute pressure falls to the water's vapor pressure, , the water boils, cold as it is. Vapor fills the crest, and the siphon stops.
The Handbook's Properties of Water tables, read as in Fluid Properties, give as an absolute pressure: their 212°F row reads 14.70 psi, a full atmosphere. So compare absolute with absolute, . For 60°F water, the most suction the crest can take is ft of water, and that has to cover both the crest's height above the surface and the velocity head.
| Temperature | ||
|---|---|---|
| 40°F | 62.43 lbf/ft³ | 0.12 psi |
| 60°F | 62.37 lbf/ft³ | 0.26 psi |
| 80°F | 62.22 lbf/ft³ | 0.51 psi |
| 10°C | 9.804 kN/m³ | 1.23 kPa |
| 20°C | 9.789 kN/m³ | 2.34 kPa |
| 30°C | 9.764 kN/m³ | 4.24 kPa |
Source: FE Reference Handbook, Properties of Water (English Units) and (SI Metric Units), selected rows
Real openings: the jet shrinks and slows
A real opening falls short of in two ways.
The jet shrinks. Water heading for a sharp-edged hole arrives from every direction, some of it sliding along the wall straight at the edge. It can't turn that corner at once, so the jet keeps narrowing for about half a diameter outside the hole. Its narrowest section is the vena contracta, where the streamlines run parallel and the pressure is atmospheric, so that's where point 2 really belongs. The Handbook writes its area as , with the opening's area and the coefficient of contraction.
The jet slows. Friction at the edge costs it a little energy: , with the coefficient of velocity. Real area times real speed is the real flow:
The product is the coefficient of discharge. This is the Handbook's Orifice Discharging Freely into Atmosphere, with from the liquid surface to the centroid of the opening. Its table of Orifices and Their Nominal Coefficients, printed with the flow meters, gives typical values:
| Opening | |||
|---|---|---|---|
| Sharp-edged | 0.62 | 0.98 | 0.61 |
| Rounded | 1.00 | 0.98 | 0.98 |
Adapted from the FE Reference Handbook, Fluid Flow Measurement, Orifices and Their Nominal Coefficients
A rounded entrance guides the water through the turn, so the jet fills it and . Keep each coefficient on its own quantity: the velocity takes alone, while the flow takes . Both grow with the square root of the head.
Below the water, or under pressure
Two variations change only the head that drives the jet.
A submerged orifice sits in a wall between two tanks, below both surfaces. Its jet is surrounded by still water, so at the vena contracta its pressure head is , its depth below the second tank's surface. From the first tank's surface, with the datum at the orifice:
That's the Handbook's Submerged Orifice Operating under Steady-Flow Conditions, , with the opening's area. Raise or lower the orifice, and both depths change together: only the difference in water levels drives the flow.
A pressurized tank has air at a gauge pressure above the water, which adds a pressure head at point 1:
The is the water's, because a head is a height of the liquid that flows. So 10 psi of air adds ft. For a target flow, run it backward: the driving head is , and the problem tells you how much of it is depth and how much is air.
Weirs: an orifice with no top
A sharp-crested weir is a wall across a channel, topped with a sharp-edged plate, the crest, that the water spills over. The water backs up until it stands just high enough above the crest to pass the flow, so that height, the head , measures the flow. Read it upstream, before the surface dips toward the crest, and from the crest, not the floor. The water below crest level isn't going anywhere.
Why should the flow follow ? Slice the overflow into thin horizontal strips, an orifice with no top. Each strip discharges under its own head, at about for a strip a depth below the surface. So the average speed grows as , while the area, for a crest of length , grows as . Area times speed gives . The Handbook's Weir Formulas, under Hydrology/Water Resources, fold the constants and a contraction like an orifice's into one coefficient:
Here for feet and cfs, and for meters and m³/s. It carries inside it, which is why it changes with the units: .
A suppressed weir spans the channel wall to wall, so its ends don't contract. A contracted weir's crest is shorter than the channel is wide, and the water turning in around each end pulls the sheet narrower, as at a sharp-edged orifice. The Handbook charges of crest length for the two ends. The usual convention is per end, so a weir with one end against a channel wall gives up .
Working backward is direct for a suppressed weir's head, , and for a contracted weir's crest length, . A contracted weir's head appears twice in its equation, so solve for it with your calculator's equation solver.
The V-notch, for small flows
Cut the crest into a 90° V, and the opening widens as it fills. At a head , it's wide at the surface, a triangle of area , as the right panel of the weir figure above shows. Times the same speed, that gives . The Handbook's V-Notch formula, , takes in USCS units and 1.40 in SI, for a 90° notch only.
That narrow bottom suits small flows. At 0.10 cfs, a 2 ft rectangular crest runs at a head of just 0.73 in., hard to read precisely, in a sheet thin enough to cling to the plate instead of springing clear as the coefficient assumes. A 90° notch stacks the same flow 3.3 in. deep.