Page images



Penstocks are required to convey water from reservoirs, forebays, or other PENSTOCKS sources of supply to turbines located in hydroelectric power plants. Where power plants are placed adjacent to concrete dams, the penstocks may be entirely embedded in the dam above the entrance to the power plant. Where the upstream ends of penstocks are some distance from the power plant, the penstocks may be embedded in concrete dams or in tunnels, carried on suitable Supports in open tunnels, supported above ground, or placed underground. Conditions are frequently such that two or more types of installation are involved in the same penstock. Typical penstock installations are shown in Figures 1

and 2.

Pump-discharge lines convey water from pumping plants to tanks, reservoirs, PUMPcanal inlet structures, or similar installations located at elevations higher than DISCHARGE those of the pumps. Where multiple-unit pump installations are made, the pipe LINES lines from two or more pumps frequently converge into a single line through headers, wyes, or branches which may be placed inside the plant, adjacent to the plant, or at Some distance from the plant. The pipe lines may be supported above ground on Suitable supports, placed underground, or installed partially above ground and partially underground. A typical underground installation is shown in Figure 3.

Outlet pipes (also called river outlets) are installed to release water from OUTLET reservoirs for irrigation or other purposes. Where concrete dams are con- PIPES structed, the pipes are usually embedded in the dam between the upstream and downstream faces of the dam. Below the dam, the pipes may be embedded in concrete, carried on suitable supports in open tunnels, supported above ground, or placed underground. Where the dams are of earth construction, the pipes are usually supported in open tunnels terminating in a valve house or similar structure. A typical installation of this kind is shown in Figure 4.

Siphons are usually installed as a part of a canal system to convey water SIPHONS across canyons, gulleys, over or under streams or other obstructions, or through territory where the terrain is not suitable for open-conduit construction. A siphon usually connects two sections of a ditch or canal which are located on Opposite sides of a canyon, gulley, stream, or other depression. Suitable transition structures are required at the upstream and downstream ends of the siphon. A siphon may be supported above ground on suitable supports, placed underground, or installed partially above ground and partially



Hydraulic and economic studies are required in connection with each proposed HYDRAULIC installation for the determination of proper conduit diameter, plate thickness, & and other construction details. Head losses should be estimated as closely as ECONOMIC possible; economic diameters should be ascertained on the basis of the infor- STUDIES mation available; and the pressure rises in penstocks or negative pressures in pump-discharge lines resulting from the interruption of turbine or pump operation, respectively, should be computed. Studies regarding the necessity or desirability of installing surge tanks, surge suppressors, or other pressurereducing equipment in penstocks and pump-discharge lines should also be made.

Head losses in straight pipe may be estimated by using any one of several HEAD LOSSES available formulas, but the use of the Scobey formula 1/is recommended. It IN STRAIGHT was derived from a large number of experiments on steel pipe installations PIPE

and may be written as:

1/ Scobey, Fred C., “The Flow of Water in Riveted Steel and Analogous Pipes,”

Technical Bulletin No. 150, published in 1930 by the U. S. Department of

[merged small][merged small][merged small][merged small][ocr errors][ocr errors]
[merged small][ocr errors][merged small][merged small]

Values of the coefficient Ks depend upon the condition of the interior of the pipe and vary from 0.32 for new smooth pipe to 0.40 for 15-year-old pipe. Friction losses for various flows and pipe sizes computed from Scobey’s formula using a value of 0.34 for Ks are shown graphically in Figure 5.

Head losses in bends may be estimated from the chart in Figure 6, based on
Hinds’ formula, or the chart based on Thoma’s experiments made at the
Munich Hydraulic Institute and shown in the chart of Figure 7. The Hinds’
formula is based on safe average values taken from experiments of earlier
investigators. Although it does not take into consideration the effect of varia-
tions in R/D ratios as does the Thoma chart, it gives more conservative val-
ues than the latter. Investigations have shown that the lowest bend losses occur
with an R/D ratio of about 5.5. However, for practical reasons, ratios between
3 and 4 are usually used by the Bureau of Reclamation. More detailed infor-
mation on this subject may be found in Technical Memoranda Nos. 325, 342,
and 517 and in Bureau of Standards RP-1110.

Head losses in branch connections may be estimated from the charts in Figure
8, which show loss coefficients for cylindrical and tapered outlets, with vari-
ous deflection angles, sharp and round corners, and various diameter ratios as
derived from Thoma’s experiments at Munich. (See also Technical Memorandum
No. 299 for Translation of Experimental Results.)
Head losses in conical increasers may be estimated by use of the formula, 2/

[ocr errors][merged small][merged small]

A1 = area of the upstream end of the increaser A2= area of the downstream end of the increaser v = velocity at the downstream end in feet per second g = acceleration due to gravity. Values of K vary from 0.17 for an included cone angle of 6 degrees, to 0.30

for an included angle of 1.5 degrees. Head losses in conical reducers may be estimated by the use of the formula,

2/ Russell, George E., “Hydraulics,” Fifth edition, 1937, p. 202.

2.9 ERELIMINARY STUDIES (Continued) HF = 0.25 # where the notation is the same as that shown above. The economic diameter of a penstock, pump-discharge line, or other conduit ECONOMIC

to convey water to be used for the generation of electrical energy or to deliver DIAMETERS water to be pumped (excepting discharge lines from turbopumps) is the diameter which will result in the lowest annual cost when charges for amortization of construction costs, operation and maintenance, and replacement reserve are taken into account. An accurate determination of the economic diameter in a given case requires the making of estimates of total annual costs, taking into consideration the value of power lost in friction, for conduits of several sizes. From such estimates a determination of the most economical diameter for each head involved may be made.

Economic diameters may be estimated approximately by use of the chart in Figure 9. After an economic diameter (or diameters) is tentatively selected, further studies are necessary to ascertain whether or not such a diameter will be satisfactory for the plant to be served, taking into consideration operating features and comparative costs of plant equipment. Economics and practicability may both enter into the final determination of the proper conduit size. The diameters of outlet pipes and siphons are usually governed by the total head available to deliver a required quantity of water per unit of time.

Water hammer is a term applied to the phenomenon produced when the rate of WATER flow in a conduit is rapidly changed. It consists of the development of a series HAMMER. of positive and negative pressure waves, the intensity of which is proportional to the spread of propagation and the rate at which the velocity of flow is decelerated or accelerated. An accurate determination of the effects of water hammer on a penstock or pump-discharge line requires a rather complex mathematical analysis. Detailed discussions of water-hammer analysis are contained in the following publications:

“Symposium on Water Hammer” published in 1934 under the joint auspices of the ASME Hydraulic Division and the ASCE Power Division

“Penstock Analysis and Stiffener Design,” Bulletin 5, Part V, Boulder Canyon Project Final Reports, published in 1940 by U. S. Department of the Interior, pp. 113-138

Angus, Robert W., “Waterhammer in Pipes, Including Those Supplied by Centrifugal Pumps: Graphical Treatment,” Bulletin No. 152, published in 1938 by the University of Toronto, Toronto, Canada

Angus, Robert W., “Water-Hammer Pressures in Compound and
Branched Pipes,” Paper No. 2024, published in 1939 by the ASCE,
New York, N. Y.

Approximate estimates of the pressure rises resulting from water hammer in penstocks of uniform diameter may be made by the use of the charts shown in Figure 10. The solutions obtained will give pressure rises at the turbine which may be assumed to reduce uniformly along the penstock to zero at the intake or other point of relief.

In pump-discharge lines, the most unfavorable water hammer conditions usually occur after a power failure at the pumping plant. Immediately after Such a power failure, the pump loses speed at a rate depending upon the pump characteristics and the inertia of the rotating elements of the pump and motor. A time is reached when the head in the discharge line is greater than that produced by the pump, and water flows back through the pump, causing it to

[merged small][merged small][ocr errors]


operate in reverse as a runaway turbine. The maximum increase in head at the pump occurs while the pump is operating as a turbine, except where check valves are installed at the pump; in this event the pressure upsurge resulting from the starting of the pump may exceed the rise in pressure following a power failure. Decreases in head or negative pressures in the pipe line occur prior to the time that the pump starts to operate as a turbine. The negative pressures produced may be of sufficient magnitude to cause the collapse of a discharge pipe if the pipe is not properly designed.

The types of conduit construction used are largely determined by physical
conditions at the site of the installation. In some cases, however, two or more
types of construction would appear to be equally satisfactory or appropriate.
In these cases, costs, safety features, operating requirements, and other per-
tinent items should be considered before selecting the type or types of con-
struction to be used. Where conduit sizes are such that the pipe can be shipped
by rail or truck, it is usually desirable to have the pipe fabricated in an estab-
lished fabricating plant in lengths suitable for shipment and installation. If the
size of a conduit is such that it is impracticable to fabricate it in sections
small enough to ship by rail or truck, the establishment of a field fabricating
plant at or near the site of installation may be necessary. In either event, the
spacing and type of field joints should be such that the fabricated pipe sections
can be transported, unloaded, and installed under the conditions prevailing at
the point of installation.

The different types of steel-conduit construction involved in Bureau of
Reclamation projects usually fall within one of the following general

A. Conduit supported above ground in a manner similar to that shown in
Figure 2, or carried on cradles similar to the one shown in Figure 21.

B. Conduit placed in open tunnels and supported in a manner similar to that shown in Figure 4, or on sliding-type supports similar to the one shown in Figure 20.

C. Conduit embedded in a large mass of concrete similar to that shown in * Figure 1.

D. Conduit placed in tunnels or other passageways and backfilled with
concrete after installation.

E. Conduit placed in open trenches and backfilled with earth as shown in
Figure 3.

Either welded joints or sleeve-type couplings may be used for circumferential
field connections with any of the different types of construction noted above.
The use of sleeve-type couplings is limited to pipe with diameters which are
within the range of coupling sizes furnished. Sleeve-type couplings are a com-
mercial product, a typical section of which is shown in Figure 11. They con-
sist of a middle ring, with pipe stop, follower rings, gaskets, and connecting
bolts. They are furnished without pipe stops if desired, to meet special require-
ments. The couplings may be used to advantage for installations where it is
desirable to eliminate all field welding, and for pipe lines crossing rivers or
canyons on Suspension bridges requiring a flexible construction which will
adjust itself to the changing catenary of the suspension cables, with pipe filled
and empty. Figures 12 and 13 show the 31-1/4-inch diameter Ogden Canyon
Siphon supported on a suspension bridge of 360-foot span. The pipe was made
in 20-foot 3-inch lengths, and connected in the line with sleeve-type couplings.
A stiffener ring was provided at the center of each pipe section which was
bolted to the bridge framing, and alinement saddles were provided near each
end, as shown in Figure 14.

[ocr errors][ocr errors][ocr errors]


Typical welded field girth joints for conduits to be installed under varying TYPES OF circumstances are shown in Figure 15. When there are no clearance restric- CONSTRUCtions or other features precluding their use, double-welded butt joints as shown TION at (a), (b), or (c) should be used. Type (a) joints with exception of the backing (Cont.) weld should, in most cases, be welded from the inside. Where it would be difficult to weld on the outside of the conduit on account of close clearances and where digging of bell holes at girth joints would otherwise be necessary, joints of the types shown at (d) and (g) may be used. Joints of the types shown at (e) and (f) may be used where sufficient clearances can be provided for Outside welding. The type (e) joint is limited to plate thicknesses which are not beyond the range of equipment available to produce bell ends. The pipe taps may be omitted if soap tests are not required. The bell joint provides the desired flexibility for the installation of long pipe lines to comply with ground profiles or to adjust discrepancies in length. Types (f) and (g) joints may be used as closing joints in long tangent lengths after the temperature in the pipe has been normalized by backfilling.


The type of steel selected determines the permissible design stresses. For TYPE OF satisfactory welding, the plates used in the fabrication of conduits should be of STEEL structural, flange, or firebox quality with a carbon content not exceeding 0.35 percent. Plates meeting these requirements are available in either carbon- or alloy-type steels. Limited experience in the welding of alloy-type highstrength steels retarded their use for conduits constructed by the Bureau of


Joint efficiencies assumed for design purposes vary for different kinds of JOINT joints, and different methods of inspection and testing. According to the API- EFFICIENCY ASME Code, joint efficiencies also vary for different types of steels. Where field girth joints differ from shop joints, the kind of longitudinal joint used will usually govern since working stresses in girth joints are usually less than those in longitudinal joints. Under these conditions, and where double-welded butt joints are used, joint efficiencies should be assumed as follows, for all

types of steels:

90 percent for unradiographed welds or for welds spot-radiographed, 100 percent for welds completely radiographed and weld defects repaired. Radiographic tests are usually limited to longitudinal joints.

A design stress of one-half the yield point should be used in the design of DESIGN carbon-steel conduit shells, stiffener and support rings, and other attachments STRESS under assumed normal operating conditions. Where the conduit is embedded in concrete with a minimum cover equal to one-half of the static head, a design stress equal to two-thirds of the yield point may be used. Under emergency conditions expected to exist for short periods of time and at infrequent intervals, a design stress of two-thirds of the yield point may be used. For alloysteels having a small margin between yield point and tensile strength, a design stress of either 50 percent of the yield point value, or 30 percent of the tensile strength, whichever is smaller, should be used.

Shell thicknesses should be such that the maximum equivalent unit stress (see SHELL Figure 16), resulting from the sum of all circumferential stresses combined THICKNESS with the sum of all longitudinal stresses, will not exceed the design stress multiplied by the joint efficiency. Circumferential, or hoop stresses resulting from internal pressure may be computed from the formula,

[ocr errors]
« PreviousContinue »