« PreviousContinue »
HYDRAULIC MODEL TESTING
BY JACOB E. WARNOCK, ENGINEER, BUREAU OF RECLAMATION
with the development of a more economical structure, or (3) an improvement in flow conditions necessitating a more expensive design. Model studies do not always result in a more economical design. In a number of instances, however, the saving effected has many times offset the cost of the experiments. Seldom has the original design of a structure been found so nearly perfect that desirable revisions have not been suggested by the model investigations.
The structures most frequently tested in the laboratory include spillways, stilling pools, outlet tunnels with control valves and Venturi meters, intake structures, needle valves, and hydraulic turbines.
SPILLWAYS AND STILLING POOLS
HYDRAULIC MODEL TESTING was begun by the Bureau of Reclamation in the laboratory of the Colorado Agricultural Experiment Station at Fort Collins, Colo., in 1930. The unprecedented magnitude of the various structures being designed for Boulder Dam was primarily responsible for the introduction of this innovation in Bureau design practice. The subsequent growth of the laboratory plant and the scope of the investigations as well as the increase in personnel have been indicative of the success with which the models of the last 6 years have supplied the information needed by the design department.
In the latter part of 1930 a staff of 12, including engineers, carpenters, and assistants, was assigned to the small laboratory at Fort Collins, with instructions to build a model of one of the two shaft spillways originally proposed for Boulder Dam. Tests on the model showed conclusively that the shaft spillway was not suited to the anticipated operating conditions. As a result, the side-channel spillway was proposed, and models of various modifications of this type were tested intermittently during a period of 2 years. The magnitude of even the auxiliary features of Boulder Dam is difficult to comprehend. Each of the two side channel spillways is designed to receive and to discharge, after a drop of 500 feet, a maximum flow of 200,000 second-feet without permitting destructive erosion.
The experience gained during the foregoing tests demonstrated clearly that the hydraulic model provided practically the only means of ascertaining and eliminating such undesirable conditions as spiral flow in circular tunnels or the choking of shaft spillways. In recognition of the value of models for supplementing the work of the design section, the Bureau of Reclamation now maintains three hydraulic laboratories with a combined staff of about 50 men. These laboratories are situated in the vicinity of Denver and permit close contact with the design section.
Of the types listed in the foregoing paragraph, spillways and stilling pools have been the most frequent. This is evident from the accompanying tabulation in which are listed the major hydraulic structures tested since 1930. Spillways fall naturally into five general classifications (see fig. 1.), the “glory-hole” or shaft-type (photograph 1), the side-channel type (photograph 7), the overflow type (photographs 5 and 8), the open-chute type (photographs 2, 4, and 6), and the enclosed tunnel chute (photograph 3). Of these the first two seldom require stilling pools since they are usually employed in conjunction with large concrete or masonry dams founded on bedrock where no erosion hazard is involved in permitting the water from the spillway to impinge upon the stream bed directly below the dam.
The overflow type of spillway is suited to a wide range of conditions. It may be used either with a high dam founded on bedrock and provided with a natural stilling poof such as is contemplated for the Grand Coulee Dam on Columbia River, or it may be adapted to the smaller concrete or masonry structure such as the Imperial Dam on the Colorado River, which rests on a sand and gravel foundation. In the latter case, a well-considered artificial stilling pool is required.
The open-chute spillway is frequently used in conjunction with earth dams. It follows the general profile of the downstream face of the dam and terminates in an artificial stilling pool. The typical pool comprises a horizontal concrete
PROBLEMS OF THE LABORATORY
Since 1930 approximately 80 models have been constructed and tested in the three hydraulic laboratories. Each of these has resulted in either (1) an improvement in flow conditions through minor revisions of the original design, (2) an improvement in flow conditions combined
apron, on the downstream end of which is constructed a concrete sill or a set of baffles. The apron may rest directly upon sand or gravel, and since its function is to protect that material from erosion, its proper construction is an important requisite to the safety of the dam. A stilling pool is a basin in which the energy of the falling water is violently dissipated without damage to the river bed, it reduces the turbulence and wave action downstream from the spillway and is, therefore, particularly essential to earth dams.
The enclosed tunnel-chute type of spillway is seldom used. A gate section is usually required at the mouth of the tunnel, and the flow conditions produced by the presence of the gate section are seldom satisfactory.
The procedure by which models of hydraulic structures are built and tested in the laboratory before the design is finally adopted and committed to construction is analogous to the manner in which a newly designed machine is thoroughly inspected for defects and imperfections at the factory. The models reveal undesirable features of the design and indicate the proper means for their correction. Due to the frequency with which spillways have been tested in the laboratory, a rather definite and efficient routine of inspection has been developed. This is conveniently separated into the following phases:
1. Inspection of the flow in the approach to the gate section.—If the approach channel is shallow, or if the entrance to the channel is abrupt, either condition will usually effect a reduction in the capacity of the spillway.
2. Inspection of the flow, and measurement of the head loss through the gate section.—A reduction in the loss of head through the gate section will usually result in an increase in the capacity of the spillway. In the majority of cases such a result is to be desired. Slight alterations in the approach, or changes in the piers, will frequently effect a reduction of the crest losses. The character of the flow through the gate section is also largely responsible for the distribution of flow across the face of the spillway, in the chute, or in the stilling basin.
3. Measurement of pressures on the face of the spillway.—For overfall dams it is important that no appreciable subatmospheric pressures shall exist on the downstream face. Pressures on these surfaces are measured in the model by piezometers which are simply short tubes inserted flush and normal to the spillway face. Rubber hoses are used to connect the tubes to manometers. Where subatmospheric pressures are observed, such conditions can usually be eliminated by minor alterations in the shape of the face.
4. Inspection of flow characteristics in open and closed chutes.Most of the difficulties encountered with chutes arise from some form of direction change imposed upon the flow. In some chutes a laterally contracted or “corset-shaped” section is attempted for structural reasons. The spillway in photograph 6 is illustrative of this type. Other forms of open chutes involve horizontal curves which must be superelevated as are the curves on highways and railroads. Such
a spillway is shown in photograph 2. Models of chutes have revealed and made possible the elimination of many unexpected and undesirable conditions such as overtopping of sidewalls, unbalanced combinations of flow, or high standing waves.
5. Investigation of the stilling pool performance. In addition to being the most spectacular part of the spillway, the stilling pool is probably the most important. The failure of the stilling pool to operate properly under all conditions may, in many cases, endanger the entire project, including the dam. The stilling pool must be as efficient in dissipating energy as is the spillway-gate section in conserving energy. Of the many forms of stilling devices in operation, the most successful are those that assist rather than impede the formation of the hydraulic jump. Sills and sloping aprons serve to stabilize the jump within the paved limits of the pool. Because of its importance, the testing of stilling-pool designs is accorded the most exacting consideration and occupies a large proportion of the time devoted to spillway studies.
6. Calibrations.—Each spillway is calibrated to determine its capacity at the maximum designed head and at intermediate heads. Frequently, unforeseen losses reduce the capacity of the spillway below the value specified in the design. Under such circumstances minor revisions can frequently be made which will suffice to increase the spillway capacity to the required amount.
FIGURE 3.-STILLING POOLS. No. 1.-Outlet tunnel and stilling pool, Caballo Dam.
No. 2.-Outlet tunnel and stilling pool, Bull Lake Dam.
When a model has been successfully subjected to each of these six phases of testing, few uncertainties in the hydraulic design remain. OUTLET TUNNELS WITH CONTROL VALVES
AND METERS Probably one of the most difficult tasks for the designer is the computation or prediction of flow conditions downstream from gates or bends in outlet tunnels. If there are more than one gate in the tunnel, the possible combinations of gate openings and the variations in the resulting flow conditions may be numerous. Since the tunnels are usually designed to flow partly full, lateral bends considerably complicate the flow.
Six outlet structures have thus far been tested, including those for the Boulder, Alcova, Caballo, Bull Lake, Owyhee, and Friant Dams. Each project presented entirely different problems, and quite diverse solutions were obtained.
The most outstanding of the six outlet structures tested was that of the Boulder Dam tunnel-plug outlets. Each outlet will comprise a battery of six 72-inch needle valves which will discharge a maximum of 22,000 second-feet into a concrete-lined tunnel 50 feet in diameter. A portion of this model is shown in figure 2. The model needle valves were made adjustable in the vertical and horizon
tal planes so that the setting of each valve, which would be most favorable to smooth flow at the maximum discharge in the large tunnel, could be obtained. The same setting was found satisfactory for flows less than the maximum, provided that the openings of the valves were symmetrical. To provide a check upon the effect of the scale ratio, two additional models were made on different scales. The valve settings finally determined represented a distinct improvement over those originally proposed.
The tests on the Caballo and Bull Lake Dam outlet works were similar in nature. Each involved vertical slide gates operating in tunnels which discharged into stilling pools. The two models are shown in figure 3, photographs 1 and 2. The criteria of a satisfactory solution for this type of model are: (1) smooth flow in the tunnel downstream from the gates under all operating conditions, (2) effective and dependable stilling-pool characteristics, (3) assurance that the hydraulic jump will not travel upstream into the tunnel but will remain within the limits of the pool under all conditions. To obtain satisfactory flow conditions in the Bull Lake tunnel, it was necessary to fix a definite gate-operating schedule that would eliminate unfavorable combinations of openings. The tests on the Caballo outlet showed that substituting a horseshoe tunnel for the originally proposed circular tunnel would
important defects in previous designs and pointed the way to satisfactory revisions. Numerous needle-valve tests have been made, the outstanding of which were those for Boulder Dam previously mentioned. A battery of four 96-inch needle valves for the Friant-Kern canal outlet of Friant Dam has also been tested. These valves discharge directly into a stilling pool at the entrance to the canal. A valve
materially reduce the sinuous flow. On both models a rounded “hydraulic hump” or spreader (photograph 2) was interposed between the tunnel outlet and the stilling pool. This served both to produce a jet of uniform thickness and to prevent the surface roller of the hydraulic jump from moving back into the tunnel at low discharges.
An interesting feature of the Owyhee tests on an irrigation outlet tunnel was the rectangular Venturi meter-flume (fig. 4) included in the tunnel section upstream from four vertical slide gates. This tunnel now serves to control large storage releases from Owyhee Reservoir. Water from the tunnel, which is 3 miles long, is delivered to the open channel of the main project canal. The model was used primarily for the calibration of the Venturi meter-flume. The curves obtained exhibited three distinct phases corresponding to the three ranges of flow, Venturi throat flowing free, throat flowing full but main entrance to tunnel partly full, and tunnel and throat flowing completely full. The result was deemed highly satisfactory, and could have been obtained in no other way except by expensive field measurements.
Unique among the structures which have been tested in the Reclamation laboratories was the model of one of the four Boulder Dam intake towers. This model (fig. 5) afforded a means of observing the flow through the tower under different operating conditions, and made possible the measurement of the losses at several points within the tower. The results provided an excellent check on the design computations.
VALVES AND GATES
With the exception of a series of experiments made in an attempt to improve the mechanical and hydraulic features of Stoney gates, all the tests on gates and valves have been concerned with closed conduit control. An interesting study of ring-follower and cylinder-follower gates revealed
Figu're 6.- Hydraulic turbine, Grand Coulee Dam