EXPERIENCE IN TURBULENCE IN HYDRAULIC STRUCTURES

HAROLD M. MARTIN, WILLIAM E. WAGNER HYDRAULICS BRANCH

OFFICIAL FILE COPY BIRO OF REiCLAMATION UNITED STATES DEPARTMENT OF THE INTERIOR DENVER, COLORADO — USA

This paper indentifies the ,types of turbulent flow ,experienced in high-head hydraauldc structures and .appurtenances. Turbulent flow in a complex outlet works ;structure creating severe vlibratio•n was studied in the laboratory. Measurements taken .and method of analysis of data leading to an improved structure are presented. Experience with turbu-lence-iduced v'i'bration tin high-head valves is related. Model-prototype verification date are included. Model and prototype studies of medium-head stilling basin are described in moderate detail. Transient pressure measurements made in laboratory studies are reconciled with proitotype observations inolu•ding .cavitation noise and vibration. Finally, laboratory studies of pressure variations in a high-head .penstock Y-branch are described together with prototype observations.

The paper illustrates .the importance .of obtaining all possurble data in studies of turbulence in high-head structures and demonstrates limi-tations of pressure :and vibration measurements. Velocity and turblence surveys will undoubtedly improve the understanding of turbulent flow as improved -techniques become avadlNble.

RE COPY WHEN BORROWED RETURN PROMPTLY

Ce rapport identifie les types de6coulement turbulent trauv6s dans les structures hydrauliques de haute chute et lours ouvrages annexes. L'6coulement turbulent dans une structure com,plexe d'touvrage de sortie, qu'i produit une vibration s&vere, fut etud6 dan.s le laboratoire. Les auteurs donnent 1es m6sures poises -et la m6tho•de utilis6e ,pour analyser les donn6es en vue d'am61d•omLr la ;structure. L'exp6rience avec la vibration induite par la turbulence dans les Vannes de r6glage de haute chute est expoo6e. Les donn6es verdfi6es sur modele et sur place sont comprises. Des Etudes sur modble et sur place d'un :bassin d~amartissemen,t de mo-yenne chute scant d6crites en •quelques d6tails. Les mesures de la pres- siontransitodre effectu&es daps les recherches en laboratodre scnt ac•coc- d6es avec le bruit et la vibration de cavitatien. Enfin, les investigations en laboratoire des variations ide la pressdon dans une conduite force bifurqu6 de haute •chute sont d6crites en jointe avec les observations du prototype.

Ce rapport e ,plique Pdmportance A obtenir toutes les donn6es possiibles par les Etudes de la turbulence dans les structures de haute chute et d6montre les limitations des m6sures de la pression et de la vibration. Suivant que les meilleures techniques commenoent A se d6ve-lopper, les examens de la vdtess.e et de 'la turbulence augmenteront sans do•ute la connaissance de i'6coulement de haute chute.

LrN tri 9-4

.4

INTRODUCTION

Turbulent flow is the result of the disinte-gration of eddies into a random pattern of mixing throughout the fluid. It is characteri-zed by the presence of wandering vortices or eddies whose velocities are irregular and, vary both in direction and magnitude. This inter-mixing of the fluid results in the dissipation of energy accompanied by pressure fluctua-tions which are intensified by abruptly chang-ing a flow boundary, by submerging the jet, or

IAHR - NINTH CONVENTION - BELGRADE 1961

by placing an obstacle, such as a baffle, in the flow path.

Fluid turbulence plays an important part in designing hydraulic structures. The inten-sity of turbulence may be influenced by the boundary roughness which must be conside-red in computing the head loss in a conduit or the capacity of conveyances. Turbulent eddies cause a submerged jet to expand laterally and serve to decrease the jet velocity. Turbulent motion in a hydraulic jump results in a relati-

K

vely high loss of energy and becomes violent in reducing supercritical flow to nominal ve-locities which can be readily handled in na-tural or artificial channels without excessive erosion. Flow boundaries, if improperly de-signed, may establish zones of excessive turbulent eddies which cause large pressure fluctuations with accompanying vibration or cavitation in the system.

Since its inception in 1902, the Bureau of Reclamation has built 158 storage dams with associated spillways, outlet works, and po-werplant structures. In designing these con-trol structures which may operate under head up to 500 feet and develop flow veloci-ties in excess of 150 feet per second, many unusual flow problems are encountered in shaping the flow passages, controlling the flow, and dissipating the energy.

This paper presents several prototype ex-periences by the Bureau of Reclamation in which turbulence has significantly affected the operation of hydraulic structures, and describes the methods of analysis and the measures taken to correct or improve the flow and operational characteristics of the stru-ctures.

Turbulence, as discussed in this paper, refers to the largescale turbulent mixing ex-perienced in a hydraulic jump or a zone of separation rather than the fine-grained tur-bulence found in flow between parallel boun-daries.

VIBRATION IN CONTROL STRUCTURES

Keechelus Dam Outlet Works')

Keechelus Dam was constructed over 40 years ago as a flood control and irrigation storage project on the Yakima River about 75 miles southeast of Seattle, Washington. Ir-rigation releases from the reservoir are made through a I1-foot 10-inch high horseshoe-shaped conduit, approximately 450 feet in length. Water enters the conduit through a vertical intake tower containing an inner cy-lindrical shaft and an outer annular passage-way, separated by concrete walls, Figure 1.

1) Bureau of Reclamation Report No. HYD-342, Hydraulic .model study of the vilbrations of Kee-chelluns Dam Outlet Works-Yakima Project, Wa-shington.

Nine openings 4 feet high and about 3.2 feet wide at two levels, elevations 2415 and 2470, form water passages through the inner cylinder wall. Flow through these openings is controlled by two 12-foot-diameter cylinder gates operated from the gatehouse on top of the inner cylinder. Three guide walls, 120 degrees apart and extending vertically between the le-vels of the cylinder gates, separate the annu-lar ring into three passageways and provide stability to the tower. One of the three walls is located on the outlet center line directly above the entrance of the outlet conduit. The other two walls are continuous between ele-vations 2430 and 2474, except for 4-foot high openings with bottom elevations of 2440 and 2448.

The inlet to the gate tower is located op-posite the outlet conduit and contains six 7-foot high by 3-foot-wide emergency slide gates separated by piers. Water is passed through the slide gates into the annular pas-sageways, through the cylinder gates into the inner cylinder, and out through the outlet conduit whose invert intersects the inner cylinder at elevation 2424.3. A ver-tical pier 2 feet thick and 11 feet long on the outlet center line, divides the conduit en-trance.

The outlet is designed to release a maxi-mum flow of 3,950 cfs under a head of 95 feet. The upper cylinder gate is used for regula-tion when the reservoir level is sufficiently high to release the required amount of water. Normal releases, however, are made through the lower cylinder gate.

Operation of the outlet works over the years has shown that excessive vibration in the gate structure occurs when the discharge is about 1,300 to 1,500 cfs. The vibration was particularly severe when releases were made simultaneously through both lower and up-per cylinder gates.

In 1950, a 1 to 15 scale model of the intake tower and outlet conduit was constructed and tested to determine the cause of the vibration and to develop methods of reducing or elimi-nating it, Figure 2A. The vibration characte-ristics of the prototype structure were eva-luated qualitatively, because it was infeasible to obtain similarity for the physical pfoper-ties of the model and prototype structures.

Initial observations of the model indicated that large random eddies and surges formed in the inner cylinder when water passed

I/1.7-2

6-3 'x 7' Slide gates - Sill E1.2424.0, R.=54'-6"

.~------ 29'-6"----- .

✓: o:. 7 D }

13' 101/g"R.

2:1 Slope

SECTION B-B SECTION C-G

c~ ~: ~

a

oho :.~~.° .~.a:. ";o;•

.' o ( E

4

E..

R ~.. SECTION a-a 4' : °: : =. Oper~~nys 1 .. Y-r i

~:'. 4' .° G p r ° -Upper cylinder gate k- - - - - 1 -12' 4" ~ X - --

::?1-Bottom 0 }o• (NOT SHOWN

I —•'~b'... o~ 10' ': of wallI

o:•;. °.:: ~:. °~~°;'.p. : •,,:r.°~'... ~ ~ .'.

El 2430 (':: .o.

❑ ❑ ol'• E1.2470 _°• -° .. 5'

D. .:~.I'.o 2--6.. .19`R :lco•'

a_ o: A d ;;.L.:~:d A °'_= '-- -----~19,R - o. 5'-10"

12

SEC. E-E °,I:.: SECTION D-D

B ( a.: I: :I: ° (ENLARGED)

El. 2425. El, 2425--,

° -- E1.2424.33 B

'1

C f 4 a ❑ ❑ :C?_?J_~C_y-EI 2415

Lower cylinder gate - :,o::':o• : :`~: ( NOT SHOWN)

:o o.'.o:-a `° 'o.•• '.' LONGITUDINAL :: a....o'. .a, :. O'.'..,.Jt e.'_:.p:'.o:•°.; SECTION

VIBRATION STUDIES KEECHELUS OUTLET WORKS

Fig. 1

I/17-3

Fig. 2a — 1-15 scale model intaike bower

Fig. 2b — Flow disturbance in model gate tower- -discharge 2720 -cfs. Reservoir elevation 2520- -rawer ring gate 501/o open

.through the lower cylinder gate, Figure 2B. The instability of the eddies and surges cau-sed a fluctuation of flow into the outlet conduit. During the period that the outflow from the inner cylinder was retarded, the inflow from the annular passageway under nearly con-stant head was stored and increased the in-ner cylinder depth. With the increased depth of water in the cylinder, more of the inflow energy was absorbed, the eddies subsided, and the outlet discharge again increased.

From the test observations, the surges and quantity of water were sufficiently large to induce unequal and fluctuating pressure forces causing vibration of the gate tower.

Two schemes were tested in the model without a major revision of the tower. The first scheme was to regulate the flow with the six 7-foot high by 3-foot wide slide ga-tes with the lower cylinder gate fully open. The second scheme was to place a regulating gate in the discharge conduit with the slide gates and lower cylinder gate full open. The upper cylinder gate was kept closed for all tests on these two schemes.

The varying turbulent conditions within the intake tower were studied by comparing the rate at which energy was absorbed within the inner cylinder. Since the surging indica-ted an unsteady condition of head loss thro-ugh the cylinder gate apertures, this head

QWh loss was used ,in the equation, P = ---,

550 as a measure of energy dissipation within the tower. In this equation, P is horsepower, Q is the discharge in cfs, W is the water weight in pounds per cubic foot, h is the average head loss between inner and annular passa-geways, and 550 is the number of foot-pounds per second for one horsepower.

Analysis by this method showed that when flow as controlled by the lower cylin-der gate, the maximum average energy dissl- pation within the tower was approximately 13,800 horsepower, or 230 horsepower per foot of inner cylinder depth, and occurred at 52 percent gate opening for a discharge of 2,700 cfs, Figure 3. Similarly, when flow was controlled by the 6 slide gates with the lower cylinder gate open, a maximum energy dis-sipation of 97 horsepower per foot of depth occured at the maximum slide gate opening and a discharge od 3,600 cfs. This method of analysis showed that the energy dissipation within the tower was reduced about 3.3 times by controlling a discharge of 2,700 cfs with the 6 slide gates.

Vibration and pressure oscillograms were also obtained to correlate the model tower vibration with the energy dissipation. An in-ductivetype vibration meter, which measured frequency and displacement velocity, was fastened to the sheetmetal transition bet-ween the tower and exit conduit, Figure 2A.

I/17-4

m bb

i

300 1 300

vi

J

"_,--~Gote opening 44 per-

cent of ful I travel

UJI C) z o

W Reservoir Elevation 2520 250 250

'T

Uj VIBRATION

Uj CE

Cr

200 200— Z Ln ENERGY DISSIPATED PER Q: 0

1 INNER LLI FOOT OFWATER N I,- < < Cr TOWER -~; 'J m U_ >

2 < 150 1500 a- LLJ C) 7"

ILL) -,)"Charge confrollede 0 (f) I- LL_ Cr < oylowercylinder

It 0 o z gate—, Ld m ®b (L Z:

loo < LJ 100 Er m <

C. V 18 R AT ION a-

LL op— Co C)

z

//,-Dishorge controlled by 6- gates

equally opened >_ Uj 50 50 (D

~Cl. ENERGY DISSIPATED PER FOOTOF 0: rr I k UJI Ld WATER IN INNER TOWER CL LLJ

0 0 1000 2000 3000 4000

DISCHARGE-CFS

VIBRATION STU DIES— KEECHELUS OUTLET WORKS Vibration studies using lower cylinder gate and

six slide gates to control discharge

II

1.0 inch Vertical d(

1.6 feet of water

a i do I

I

I 4 2

H®931-25 i

I

1.0 inch Vertical deflection E 1.6 feet of grater

I QP I

i

2

i

I

H931-2 ~

quols

5

Fig. a

1/17-6

The recorded vibration frequency and velo-city of discplacement of the model transition surface qualitatively represented the vibration characteristics of the prototype transition be-cause of the difference in physical proper-ties of the two structures. Traces of pressure variation on the transition invert were also transmitted to the oscillogram by a reactance-type pressure cell connected to a piezometer. Representative traces of the transition vibra-tion for a discharge of 2,500 cfs controlled by the cylinder gate and the 6 slide gates are shown in Figure 4.

The average trace amplitude (average ve-locity) was used to compare the slide gate control to that of the cylinder gate control, Figure 4. The percent indicated vibration for any gate opening is plotted as the average tra-ce amplitude for that opening divided by the amplitude for the full opening.

Good correlation between indicated vi-bration in the tower and energy dissipated per foot of depth was obtained with control by the slide gates or the lower cylinder gate. The curves with the discharge controlled by the cylinder gate indicate that the tower will absorb maximum energy and have maximum vibration at approximately the same disharge and gate opening. With the discharge control-led by the 6 slide gates, the energy dissipated per foot of depth and the indicated vibra-tion of the tower were a maximum at 100 percent gate opening. The indicated vibration with the cylinder gate control was found to be 4.8 times that for the slide gates at a di-scharge of 2,500 cfs.

On the basis of these test results and as a temporary expedient until a more permanent means of regulation is provided, it was con-cluded that releases could be made through the outlet works using "the emergency slide gates for control when the vibration in the tower became excessive. Since the model studies were completed, outlet releases have been less than 900 cfs and the lower cylinder gate has been used for regulating service wi-thout objectionable vibration.

Hovell-Bunger Valves at Ross Dam

An interesting correlation between model and prototype average pressures was obtai-ned on the hoods over the 72-inch Howell-Bunger valves at Ross Dam, Figure 5. The hoods were developed by model studies to prevent excessive, inherent spreading of the

jet and to reduce objectionable spray, Figure 6A.

The spray hoods, developed specifically for the Ross Dam installation, were 12-foot-diameter semicircular steel plates anchored independently of the valves to the concrete walls on either side of the outlets. The hoods were of 1-inch steel plate with 1-1/8-inch shield plates on the upstream end and 1-inch stiffener bands at the center and downstream end, Figure 6B. Pressure taps were installed along the top of the right hood, Figure 7, to correspond to those in the model.

At the time of the field tests, the valves were operating under a head of 200 feet. Re-sults of the pressure tests are shown in Figure 7 where the solid lines represent prototype pressures and the broken lines indicate cor-responding data from the model. Model pres-sures were measured with water manome-ters while prototype pressures were measu-red by mercury-filled »U<< tubes. The compa-rison showed the average prototype pressu-res to be approximately 20 percent higher than those indicated by the model. Also, the pressure spread along the hood was greater in the prototype installation. The latter dif-ference was no doubt due to the air mixing with the water and swelling the prototype jet, while air insufflation was negligible in the model. Pressures at the two taps in the up-stream vertical shield plate were approxi-mately atmospheric.

The hydraulic performance of the proto-type valves was entirely satisfactory, as pre-dicted from the model studies. No vibration was noted in either valve; both valves were receiving an adequate supply of air, and no evidence of cavitation was found in the valves or hoods.

Vibration was obvious, however, in the hoods which were mounted independently of the valves. Intermittent surges in pressure cau-sed the hood to vibrate longitudinally with a noticeable deformation of the downstream end. In approximately 1 month of operation, the vibration intensity was sufficient to crystal-lize and sever ten 1-1/4-inch bolts near the downstream end of the hood. The bolts, which were placed in double rows on 6-inch cen-ters, failed in the threaded section just inside the nuts near the top of the hood flange. Tests on one of the broken bolts indicated an ave-rage hardness of 97 by Rockwell B test or an approximate strength of 100,000 pounds per square inch, a fair grade of steel alloy.

I/17-7

' 11Y1 f, ~ 4 —41 , r•, 1,"

PLAN OF SPRAY HOOD AND VALVE HOUSE FOUNDATION

t 72 P, V-JA

Dnerslon

q -z'p'pQ

N,

IV' 14~4'

SECTION G-G

RI note I it

2'0 Sd Me, 4'-0'- 14Q. I 1~ '1U1

ftskr 4". 0~4' ~J'

ANCHOR BOLT DETAIL- CENTER BRACKET

Si. it. 13,17 A

Elptrdtlrig F-

122325 ---------

N

,AEI 11160

Ulm

pip

72'H

. .... ... ... Arxns 1203 4:

s;k r,. N2V4

.~,I;info ft d" I ------- 14~O" W-O' - - - Mll,.r rods Dr II i

--Stn ti- to

4: Iz

. . . . . . .....

ed 1.1" Anchor rods

Avmxrnklinqi 'JA 64. 1

,

SECTION A-A SECTION B-13 TRANSVERSE SECTION THRU E-LEVATION OF VALVE HOUSE VALVE HOUSE FOUNDATION FOLIWATION

Fig. 5

1/17-8

As a temporary solution, the broken bolts were fused to the hood flange by means of welding. The welded bolts remained intact until the Howell-Bunger valves were repla-ced later by two hollowjet valves.

Fig. 6a — 6-1111th valve d_s•charging with hood removed

This experience demonstrates the impor-tance of edaquate instrumentation for inve-stigating pressure fluctuations in a high-ve-locity jet. Sensitive pressure-cell measure-ments will indicate the frequency and maxi-mum intensity of the pressure forces so that adequate support and steffening for struc-tural members can be provided.

Fig. 6b — Pressure connections in hoods of 72-inch prototype valves

PROBLEMS IN STILLING BASINS

Glendo Dam Outlet Works')

Glendo Dam is a multipurpose structure on the Bureau of Reclamation's Missouri River Basin Project, and is located on the North Platte River in southeastern Wyoming. Normal releases for irrigation and power pourposes are made through a 23-foot-diameter con-duit which branches into two 12-foot-dia-meter power penstocks and three 12-foot-diameter outlet pipes. The outlet works has a desgn capacity of 10,000 cfs at a maximum head of 150 feet and is con-trolled by three regulating slide gates, each 7 feet 3 inches wide and 7 feet 9 in-ches high. Flow from the regulating gates which are tilted downward 15 degrees enter three 20-foot-wide outlet bays separated by two intermediate dividing walls before enter-ing a hydraulic jump stilling basin, Figure 8. The stilling basin is 66 feet wide and 72 feet long. Flow from the outlet works basin com-bines with the power-plant flow and enters a common outlet channel. When power is being generated, approximately 1-1/2 feet of additional tailwater is available at the outlet works stilling basin.

The outlet works structure is unusual in that the regulating gates are placed low rela-tive to the tailwater and the stilling basin has a length of only 2.8 times the conjugate depth. Model studies conducted in 1954 indi-cated that comparatively large chute blocks and baffle piers were required to obtain adequate stilling basin action in the compa-ratively short basin length.

When the outlet works was placed in ope-ration in the spring of 1958, an audible thumping noise was noted in the stilling ba-sin and vibrations were observed in the ba-sin training walls, particularly the left wall which is cantilevered between the basin and powerplant tailrace. The thumping noise had no definite period; at times the »thumps« were barely audible and occurred at time in-tervals of 10 seconds; at other times, a flurry of thups would occur at one or two second time intervals and the louder thumps could be heard 100 feet from the basin. There ap-peared to be a direct relationship between

') Bureau oaf Reclamation Report No. HYD-461, Hy-draulise model studies of Glendo Dam Outlet Works-Missouri River Basin Pmject.

1/17-9

51

----------------- n 0" -------

i I

P,ezameters Flood

a. Flow

Howell - Bunger," Valve-- `-----

o

of valve _~...~.-._...

ELEVATION

LOCATION OF PI EZOMETERS

--L--.I

SECTION

5

a

1

— i I VALVE OPENING-PE GENT

— 100 .90

80-

- 70

0 90- 7

/ 100

e0

1

I 1

- l ~

G J 5 b ] tl y DISTANCE ALONG HOOD-FEET

■■ . ~~ ~,

Protollpl.,

DISTANCE ALONG HOOD-FEET Note: Pressures obtained' with 200 feet

ROSS D A M of reservoir head.

HOWELL.-6UNGER VALVE HOOD-MODEL-PROTOTYPE COMPARISON OF PRESSURES ON HOOD

FOR DIFFERENT VALVE OPENINGS

Fig, 7

I/17-10

4

Gate deck

Tailrace

---------------- 64'-3".

--------------- ----- -----------

'-Div iding walls Outlet works

stilling basin,

a _,

,Elev. Varies

SECTION 8-8

A

Power plant-"'

IL A

22~'_O"

---29*-9 ------

PL AN

7-'3"a 7=9" H P slide gates--:~;

0" D I ~Clp', p~, invert El 4491 75

El 4509.5

ki-E1.4502.0

-Contrection joint

EI 4486.0-, ,EI 4407,0

it 1 4482.0-,

5* 'E1.4479.0

SEC TION A-A

,10-wo 4,41

VIBRATION STUDIES GLENDO DAM OUTLET WORKS

Fig. 8

the pounding noise and the wall vibration, because the greatest wall deflection appeared to accompany the loudest thumps.

The thumping noise and _vibrations persisted through the 1958 irrigation season and beca-mem pronounced during the latter part of the season when the outlet works discharge was about 5,000 cfs or 50 percent capacity. No po-wer was generated during this period.

At the end of 1958 irrigation season, the basin was unwatered and extensive cavifa-tion erosion was found on the surfaces of the chute blocks. It was concluded that the thumping noises and vibration originated from the cavitation occurring on the flow surfaces. By removing the source of the ca-vitation, the vibrations thus would be reduced or eliminated.

A second model study was made and ex-tensive pressure measurements were obtained on the flow surfaces of the chute blocks, baf-fle piers, basin floor, and training walls. Oscil-lograms of typical pressures on the chute blocks, baffle piers, and basin floor immedia-tely -downstream from the chute blocks are shown in Figures 9 to 11. The upper oscillo-grams in each figure are for the original field design and the lower are for the design deve-loped in the second model study.

The model investigation showed that the cavitation erosion on the original chute blocks resulted from insufficient streamlining of the blocks, and a more streamlined block with reduced height was developed, Figure 9B. Al-though no cavitation erosion was found on the prototype baffle piers, the model study indicated that cavitation pressures were pro-bable with the modified chute blocks; the-refore, additional streamlining was recommen-ded for the baffle piers, Figure 1013. Cavita-tion pressures were also found on the model basin floor immediately downstream from the original and the modified chute blocks. At-tempts to supply air to this low pressure re-gion and placing steps between the chute blocks were unsuccessful in relieving the low pressures. It is possible, therefore, that cavitation pressures may occur on the basin floor for gate openings above 50 percent.

During the winter of 1958-59, the struc-ture was modified by changing the height and shape of the chute blocks and baffle piers as determined from the model study.

The powerplant was placed in operation in May 1959, providing about 1-to 1-1/2 feet

of additional tailwater for the stilling basin. The outlet works discharge during this period ranged between 3,000 and 5,000 cfs, except for 6 days when the discharge averaged about 7,000 cfs. No thumping noises or vibrations were reported during these normal releases.

Thumping and vibration were noted, ho-wever, during test releases in June 1959, which were conducted to determine, among other purposes, the stilling basin perfor-mance for the range of discharges up to the maximum of 10,000 cfs. The thumping was first observed when the outlet works dischar-ge reached about 7,000 cfs -and persisted through the remainder of the tests, inclu-ding 7,500 cfs through the outlet works only and combined outlet and powerplant flows of 7.500 up to 9,100 cfs. Following these releases, which were made by in-creasing the discharge in increments of 500 cfs, the outlet works test releases (with no flow through the powerplant) were repea-ted with discharges decreasing from 7,500 to 5,000 cfs. Thumping noises and all vibrations were noted during each of these repeat tests.

It is significant that vibration was not noted in the earlier tests until 7,000 cfs was released, although thumping noises and vi-bration were clearly noted at lower dishar-ges of 6,000 and 5,000 cfs during the repeat tests. Also, slightly higher tailwater eleva-tions were recorded during the repeat test releases.

A possible explanation for the nosies and vibration is the formation and collapse of ca-vitation envelopes on the flow surfaces in the vicinity of the chute blocks. Cavitation enve-lopes collapse audibly with tremendous force which could be transmitted to the walls, but the frequency is considerably higher than the frequency of the noises and vibration obser-ved at Glendo. The fact that no cavitation damage was reported after the 1959 irriga-tion season fails to support this theory. Ho-wever, the basin operated only a compara-tively short period of time at discharges ac-companied by noises and vibration.

Another explanation for the noises and vibrations is the unsteady pressure forces on the blocks caused by variations in the sepa-ration and vortex flow patterns with and without cavitation. These variable loads on the blocks may be transmitted through the structure to the wall; 6r the pressure varia-tions in eddy patterns near the trailing edge

I/17-12

80 00 40

20 220

60 -00

71H~~I1I A111■lfltl r, II mom IH„

Ill

PJIYI

■ ■WE ME ■ ■ snow i

■■■■■■E ■■■ ■

O

0

10

.,

F

2.3 9L 2

DOWNSTREAM END

ELEVATION

PIEZOMETER LOCATION

l■■■■■m■■E■■ ■■■■■■■■■■■ I

p ■■N■mOouO■small~■■■■■■■■u■■

• 'ICnn-ICJ is

giggyFd"'1mmg NA14MM -•

■/■■■1 1■■

/■■■■■■■■■■■■■ ■

ammo■■■ ■E■■■■

Von gor

MEMO mom

a■■■■■■■.■■■■■■■ /m■■■■m■■u■ Ill ■O■

PIEZOMETRIC PRESSURES

A. ORIGINAL DESIGN

k" -----'-------"'---------- 2772•------------------------- - 1 6.9' •b y. ~

~3 • •.•S:t . ~ 4 ~

Q 1

F-1 E J.47 ON

PIEZOMETER LOCATION

n

2 5= 92'x'

F __-S.00---

DOWNS TPLEA END

/m■m■■■m■■■m■■m■ Imm■■■■■p are■fir*11111■mA/1■~ an :■ww■~votl hI~WI1Yi:r4MlLLtfiX,WlGWLL !G1lIIR'n1tl1! ■.■■■■~■■■■■■■■■ I■■■■■■■■

■■■■■■■■■■■MOMOMI NONE■ Imm■m~■ om ■■■■■■■■■■■■■■■■ I■o■■O■■■ ■■■■■m■O■■■■N■■I I■■m■■NN■ /■■■■■mm■■■■■■■■ 1■■■0■■■■ #■■■■■■■■■■■■■■1 ;mmmmmmmm

■ st■■i■r■wn■■■r.~■ ' ■~uclrl,n'~r+ll~l~ir~rwr,rrelr ~1N'" l' I I'I'I'rIYIYi■;.tiY ■SEE WE~NVENE■/ ■■■■■■Imo ■■■■/ mom■=■1■m■m■■■o/ mo■m mlomm■mm■■/ ■m■■■ol■■m■■mone oo■mo■lmoo■m■O■/ i■■■ool■■o■■■mom

B. RECOMMENDED DESIGN

Fig. 9

I/17-13

of the blocks might be transmitted to the wall and initiate the vibrations. Changes in tail-water elevations affect these separation and eddy patterns, which may explain why vi-brations were observed during the repeat test releases and not during the earlier tests.

Piezometer laps were placed on the left training wall in the model to measure the pressure fluctuations on the wall surface. No effort was made to represent the wall rigidity or to measure the wall vibration in the model.

The greatest pressure variation, equivalent to about 45 feet of water occurred about 3 feet downstream from the chute blocks near the basin floor. The pressure variation decreased to about 20 feet of water near the downstream end of the wall and to a few feet of water near the water surface. The magnitude of these pressure variations was sufficient to cause the wall to vibrate, particularly if the frequency of the pressure variations and the natural wall frequency coincide.

The direct source and cause of the vibra-tion in the training wall has not been deter-mined. The structure operated satisfactorily during the 1960 irrigation with no apparent thumping or vibration.

PROBLEMS IN CLOSED CONDUITS

Falcon Dam Outlet Works Y-Branch3)

There has been a trend in recent years to use outlet valves as devices for measuring wa-ter releases from reservoirs. Good accuracy control valve is preceded by sufficient lengths of straight approach pipe. In such cases, stan-dard pressure-discharge relationships have been determined by model studies with the pressure-measuring section arbitrarily placed one diameter upstream from the valve. When the approach pipe contains a bend or is com-paratively short in length, the pressure and velocity distribution in the approach may be nonuniform and a special model study or field calibration is rquired to obtain accurate pres-sure-discharge relationships.

Such a case exists at the Mexican outlet works and the United States outlet works lo-cated at Falcon Dam which was designed by the Bureau of Reclamation. The outlet works

3) Bureau of Reclamation Repert No. HYD-474, Ca-11bratlon of Hallow-Jet Valves and Vibration Studies of Outlet Works Y-Branch--Falcon Dam:

are located on either side of the Rio Grande River channel. Each structure has an unsym-metrical Y-branch followed by horizontal or vertical bends a short distance upstream from the hollow-jet control valves, Figure 12.

An aerodynamic model, including a part of the power penstock, the branch from the penstock, and the approach pipe to the outlet Y-branch was tested to determine the velo-city distribution entering the Y-branch. Velo-city traverses by pitot tube in this model sho-wed that a very symmetrical valocity pattern existed at the entrance to the outlet Y-branch. Therefore, about 20 diameters of straight pipe were placed in the hydraulic models up-stream from the two Y-branch to assure uni-form velocity distribution at the Y-branch entrances.

Initial tests on the hydraulic models indi-cated that varied or no flow through one leg of the Y-branch would not affect the pres-sure-discharge relationship in the other leg. Because the pressure taps were located near the downstream end pipe bends, the pressure would vary around the periphery of the pipe. Multiple taps connected to a manifold for in-dicating pressure head are usually installed at the pressure measuring section. At the Fal-con Dam installations, taps were placed on the pipe periphery at the four points mid-way between the vertical and horizontal cen-terlines.

Measurements at the model taps for the United States outlet valves showed the pres-sures on the outside of the bend were as much as 4 percent higher than those on the inside of the bend. The pressure difference for the Mexican outlet valve installation va-ried as much as 31 percent. The higher per-centage difference measured in the Idexican branch resulted from the greater bend cur-vature in the Mexican instalation. Only in-significant differences in pressure head bet-ween the two taps on the inside of the bend were measured. Similarly, the two taps on the outside of the bend gave almost identi-cal pressure readings.

With a manifold connecting all four taps flow would occur from the higher to lower pressure openings. The magnitude of this ma-nifold pressure would vary depending on the sizes of the tap opening and water passages. An accumulation of foreign material or corro-sion in any opening would change the mani-fold pressure and, thus, the pressure-dischar-

1/17-14

w K 20

15

10 4 5

3 0

a 0 _5 0

pw -15

0000■ ■ ■ aw■ ■■■1■ ■NM■ ...■ 0000■ 0000mommomm■ ■ ~.51■■■■■■■MEN 000■ g■■■■■■■■■■■■■■ ■:0000■■■ monism!41ii

11 a : l.:~ :p ? : ~ me mill

IT oil ~■■■■■■■■■■■■■■■■ =•5011■ 1 mmon on 1:111110 :. 1:1■MIN. 0000■1 N■■■ among ■■u

Ill' 1~~J1%L'.71Y)l~ia'1+~*~Rl JH~*c+I.J

1L0w ~

w j 40

N K 30 W F 20

a 3 10

T o -10 > -20 0. -30

u -40 K 4

DOWNSTREAM ELEVATION

6 .• T2 _r 46 ,79 7•

I

7 2 3°

I

I N

I I 3

I 4

I 5 6 I

1

Br 13

PLAN

a

~Iw■■a■nand■■■.n■■E..■■ . a+•iuurdu~r•I`11~!.a~Ls1..r- ~,•rilia

MENNEN MEMNON

20

III + 71 r

.►'~ -T I 1

__J FLAW

X1415 19f .k- a i9 'A I

Y 0 X37 i 0 3613 3e 1

I N 1 1

AAL~

L4- 6.50'____ ____ yl

SECTION A-A DOWNSTREAM

PLAN RIGHT 61DE LEVATION ELEVATION

VILL. a

Eno

MESON ■■......■ONES No ■H■■M■■.■■.■.■■g■a

an

■■

i/■■ ■1■■■■ESE ■E1mmmmm

. ~.........■.■.■■ imam 11 min 1on 4a.■■■..■■m■MN■■■■a■■.

Mammal no"

i;iz

■■■

E■NON.■■■■ ■ MmMM■MMm

EMENIMME ONE I ME Mimi! 0 Sol Sol a

ri ■~ a■N■mana■Sam.■Eli■■

:mmmmumma0:00:■::

MEN:::■:■:::■~

24

PIEZOME i RIC PRESSURES

Fig. 10 1/ 17-15

,Sf43O175.72, DOWNSTREAM ENO OF BLOCK FLOOR

EL,44/9Ao.

---------5.37'--------~ ',`~ _ _3,83'_ rLow

N

—u —}i2- 7 . O'V----~1 _

}- —

4.60'________~I S 6

—__ y I',53'--+♦F -163-....

PIEZOMET ER LOCATION

i■■■■■■■■■■■■■■■

!1010■/ 10own ■/ ■ ?'W ',V111l0i,17,rip MET," fig"Li NAM 011111 6' w 1■■■■■~■/ ■m■ ■ immsmo s

• i////■ 1010■■■

■■mmmoo■~_~ 1010 1010 1010■■ 1010

I.-

/1■ ■■■■■~ ■ "MIKE 11010 :■ '~► 11^~I/t1~YW6~~r{11~yU'{ ~'af~■ 1'' miM:a s mom■ mmmmmm■■■■■ ■ 1010■■■■oms■ 1010■■■■

!■ MISS

■ m■■mm■I ■.■.■.:■■■■■■ ■ 1010■■/t■ ■■ r

pl4,ll"" uT:1i :11ltU~ ri'~Pil(~'i'

:a a■a■SAW ■C.. "aaaama:aaaa ■:■a:

■■■■■■■~■■■■■■' ■■■1010■■ 1010■■N

■■■■■■■■~■■■■■■■ ~■■■■■■■ 1010■■■■. ■' Wiln■u■i~InRn

~111P, MA~ ~1i.N17!51ti.Yl ■rm fr■■■p■o■■m. ' i■■■ 1010■■■N■■■. p■■■■■ ■■■■■■■■~

SEEN ■■■■■■■■■■■■ I■■ ■. .1010■.■ ■ 1~.■:■■ 1010■■■. an 4a*■1R1 innn■jnilii■u ~H111 i1V1',C ii4~111U~r. 'd ,11.51'1 IJ!III ~tjjl r~li~

"I 111:0.

1aa III aaa::aa:a::

■1110■■1■■■■■■■11 //■■■■■■■■■aaa■, ■■■■■■11■■■a■■I *■■ri ■n1 m

f■n ■■1

LIII~ ~uAtfltt ~`II~~I fI1~ ~S ~1lu, iRUN ■:":S~~:aaal ~aaa::aa:aaaa::I'I

PIEZOMETRIC PRESSURES

A. DOWNSTREAM FROM ORIGINAL CHUTE BLOCK

i■■n■■■n■■■■■■

u ■

■a~~■ a'a'CO Iwo a 1r1 101,11

AT■~If1Al~i.'l•~rf/ 0,

Wells

1■as ■■m■mm■m s■ -30 :a■■aa■■:a■■■iim:

I■■■■■■■■■■■■■■ .mmmon .. ■ .■

l;mmmmmmm

m i~asm ■11 g~F Iflail■ "'7i::a:::::~aa::is■N■■ 1010■■ ■■■i■■m■■m■■■■ /■■ / ■■■m■■ lim

I■~■■■■■■■■■_■■■ ismoniiii III II !/■m■ 1■■■■■~.■■■ri■■ ■ 1 tai*w /~ Iww1~11iR/!.1/1 ur.~■■~. aa~u~~til ■■ 1010■■ aaal

o■■■ ■■a■■■■■/I

,v■~■n■■■■■■■■

as~E=~mm6 ~IN 0.

Me~ ■̀ ,

■~ 1010 ' ■ 1010■■a■■■ a ■.■N. ~ a

■■'■■■■■■■■■■■1 .... ■.■■.....a1 a■■■a .■ ..

a R p 44,11t.11 WIN

a soon mn

n■■■■n 10■u10■■■■■OiI .~■■■■■u ■ 1010■ ■ ■1010■ Now win

::=~:a~:a■aa■■I WINE

■m■■■

a■a■ma■aaCa■■a0

.. ■ .1010 ■■// :::aI as IdIIhI11i■a. ■■11! ■ ii n

'..NMMO :.a:':

PIEZOMETRIC 'PRESSURES

B. DOWNSTREAM FROM RECOMMENDED CHUTE BLOCK

Fig. 11

I/17-16

ge relationship. The flow circulation within a manifold using two taps measuring nearly equal pressures would be negligible and blocking of a tap would have little effect on the measured pressure. Therefore, the two

Fig. 12a — United States Outlets Right Outlet Com-.plete Orifice replacing left hollow-jet valve Model scale 1 : 12

taps on the outer surface of the bend were used in the prototype installation because heads of greater magnitude were involved, resulting in more accurate recordings of head.

Fig. 12b — Mexican Outlets Right Outlet Com-plete Left outlet - - Laib+aratery control valve in position of deft !butterfly valve Model scale 1 : 15

Reports from the field indicate these pressure measurements are reliable in measuring re-leases through the outlet works.

When the Mexican outlet works was ini-tially placed in operation, excessive vibration and noise were observed in the Y-branch. The noises were described as loud and heavy thumping or slapping sounds. The vibration and sounds had a frequency of from 2 to 3 seconder and appeared to originate from tur-bulent eddy conditions in a zone of separa-tion adjacent to the left wall of the left leg of the Y-branch. The condition was more pro-nounced when both valves were discharg-ing simultaneously at large valve openings.

Similar vibration . tendencies were noted when the flow conditions were represented in the 1 : 15 scale model of the Y-branch and control valves, Figure 12B. Because the vi-bration and noise were apparently associated with highly fluctuating pressures within the branch, numerous piezometers were installed, Figure 13, and measurements of transient pressures were made with strain-gage-type pressure cells and recorded by oscillograph.

Rapid pessure fluctuations up to 60 feet of water (prototype) were recorded at Piezo-meters D and E, as compared to 4.0 feet at Piezometer 1 near the branch entrance, Fi-gure 14. Intermittent subatmospheric pres-sures of about 17 feet were measured at Pie-zometer 5. These model observations led to the conclusion that pressures in the separa-tion zone of the prototype momentarily re-ached vapor pressure, formed vapor cavities that collapsed with the sudden rise in pres-sure, and initiated pressure surges of suffi-cient magnitude and frequency to cause the noise and vibration.

Experience with zones of separation of this type has shown that the pressure fluctua-tions will be minimized by roughly shaping a new flow boundary to the outher limits of the turbulent separation zone. Three filler blocks with dimensions and shapes as shown in Figure 13 were tested.

With the sm4Jlest test block (6 inches in length) installed, the maximum pressure fluctuation was unchanged and a minimum subatmospheric pressure equivalent to 27 feet of water was observed at Piezometer 3. The intermediate filler block (8.9 inches) reduced the pressure fluctuations to a maximum of 20

I/17-17

of

orizonfol clia,'A

Radial distance (varies)

SECT̀ ON A-A

T,&Rt r op rr)npn[M,&T;:-'; AXIAL DISTANCE ON Ck

PIPE RADIU5 ((NCWEs)

RADIAL DISTANCE (INCHES)

6.6-3LOC K B.9° BLOCK 10.4'BLace,

0 3.78 0.0 0.0 0.8 V.

3.70 0.64 0.69 0.69

2" 3.61 0.89 1.03 1.05

3!' 3.53 0.89 1-17 1.21

4' 3.44 0.73 1.15 '-1.23 5' 3.35 0.43 1.04 1.16 6" 3.26 0.0 o.84 1.00 7" 3.17 059 0.81

61, 3108 0-30 059

8.9" 3.00 0.0 - 9" B.00 0.35 [a- 3.00 0 17

o

CA,

e-

qt' Ch

yn C* 111\0 I IQ

e,

V

v. 1piezor".t 5., ONT

a Above 0.5 -+~"~ 17

/

hcriz diu

Plow

ri of pipe '

DEVE-LOPED SECTION OF Y - BRANCV-1 -PLAN WEW 6.0'c-_ --- -- - rl 8.9 ----~+

I A 777777-,

Radial disiance

LAxial disfanco

6.0'81oc

----------- Vj 8~911 131 ck- 10.9" Block

A K`-Start of

at pipe inter•5ection al

pipe--,!f-

ENLARGED SECTION OF BLOCKS

Fig. 13

RM

Fa

91

C

0 D

E

w UJ I-- F LU x 0 N

32

12

13

M

17

WITHOUT BLOCK WITH 6.0" BLOCK WITH 8.9" BLOCK WITH 10.9* BLOCK Horizontal scale:-lines at 1/3 second intervals, Vertical scale:-heavy tines at 0.5' H2 0 model intervals

NOTES See Figure 4 for piezometer locations and block shapes. Both 90-inch hollow jet valves fully opened-maximum head Rate of flow-6.68 c.f.s. (5 820 c1 s. prototype) '

rig. 14

1/117-19

feet and the lowest observed pressure was 7 feet of water at Piezometer C. Installation of the largest test block (10.9 inches) resulted in a negligible decrease in the pressure fluc-tuations from those on the intermediate size.

It was concluded that the intermediate- Sized filler block gave optimum pressure con-ditions and the field structure was modified to include this shape.

Subsequent observations on the field struc-ture indicated the restrictive flow boundary eliminated the heavy thumping sounds and vibration in the Y-branch, suggesting a sub-stantial reduction in pressure fluctuation.

CONCLUSIONS

Adequate investigation of zones of turbu-lent eddies is emphasized for studies of structures involving high-velocity flow. The

capabilities and limitations of test equipment must be determined. Maximum benefit from basic and applied studies is obtained through analysis of accurate and reliable measure-ments of the hydrodynamic forces in critical flow regions of the structure. Such test results from a model can be applied quantitatively and qualitatively to predict the operation of the prototype structure.

ACKNOWLEDGMENTS

The studies of turbulence and its effect on various types of hydraulic structures ci-ted in this paper were carried out over a period of 10 years. Bureau of Reclamation Engineers Donald Colgate, Fred Locher, Jack C. Schuster, William P. Simmons, and Willi-am E. Wagner were the principal investiga-tors for the design studies described in the laboratory reports referenced throughout the paper.

1/17-20