Ellis,J. Pressure transients in water engineering, A guide to analysis and interpretation of behaviour

Подождите немного. Документ загружается.

24.3.5 Remedial measures

Remedial measures were considered and one simple measure was to

ensure that the ﬂoat valves were able to close fully. This was achieved

by increasing the height of the overspill weir within the BPC.

Eliminating spillage and the accompanying variations of water level in

the chamber allowed the valves to remain shut after initial closure.

Another measure considered was to alter phasing of ﬂoat valve

response by introducing extra throttling between the ﬂoat chambers

and the main tank of the BPC.

512

Head u/s of inlet valves at BPC

0 1 2 3 4 5 6 7 8 9 10

Time

(

min

)

Head (mAOD)

380

360

340

320

300

280

260

240

Fig. 24.11. Head variations upstream of BPC

Maximum and minimum head

along gravity main

d/s reservoir

Helvetica Flush left

u/s reservoir

350

300

250

200

150

100

Chaina

e (km)

Pipeline profile

BPC

Head (mAOD)

0 2 4 6 8 10 12 14 16 18 20 22 24 26

Fig. 24.12. Envelope curves along gravity main

Pressure transients in water engineering

24.4 Small hydro station

The ﬁnal example concerns behaviour observed at a small hydropower

station in New South Wales and which is operated by Snowy Hydro

Limited. Unstable behaviour has been observed since commissioning

of a submerged discharge valve (SDV) which is sited in parallel with

a hydro turbine. The hydraulic system consists of an intake tower

leading to a tunnel of nominal diameter 3750 mm and length 356 m.

At the downstream end of the tunnel is a trifurcation. Two of the

branches downstream of this trifurcation supply free discharge cone

valves of diameter DN 1800 and DN 1200. These valves were shut

during the events being considered. The ﬁnal branch downstream of

the trifurcation is a penstock having nominal diameter 2000 mm and

length around 120 m. The penstock terminates at the small hydro-

station. The station houses the small hydro unit rated at 1.105 MW

and an SDV of the type shown in Fig. 9.13. Inlet to the SDV is

ND 800 with an ND 800 bend before a taper reducing diameter to

ND 600. Plan arrangement of the tunnel and penstock system is

shown in Fig. 24.13. When fully open the valve is capable of discharging

3.5 m

/s. Flows from the SDV and from the small hydro unit enter

individual sumps with ﬂow leaving each sump via an overﬂow weir

and entering a common outﬂow channel. The small hydrostation and

outﬂow channel can be seen in Fig. 24.14.

24.4.1 Observed behaviour

Observations of hydraulic transient behaviour within the penstock and

small hydrostation were reported by O’Connor (2007). Pressure

513

Free-discharge valves,

DN 1200 and DN 1800

Tunnel 3750f Length 356 m

Trifurcation

Intake tower

Small hydro turbin

1.105 MW

Submerged discharge

valve 800 ¥ 600

Penstock, 2000f Length 120 m

Fig. 24.13. Small hydro pipeline system

Flow instabilities

measurements were made within the ND 2000 penstock both with and

without the small hydro turbine in operation. When ﬂow through the

SDV was 2.7 m

/s or higher, the following effects were observed:

(a) a pulsating sound within the small hydrostation

(b) the noise of cavitation from within the SDV in the area of the

DN 800 bend (Fig. 24.15)

(d) more pronounced turbulence in the SDV sump evidenced by

surface wave action (Fig. 24.16)

(e) large pressure ﬂuctuations recorded on pressure gauges at a hydrant

close to the penstock at the small hydrostation.

Recordings of pressure ﬂuctuations in the penstock showed a

maximum head range of about 9 mWG. While a range of frequencies

of oscillation was recorded, it was noted that some oscillations had a

period of around 1.6 s which is close to the fundamental period 4L=a

of the conduit system as a whole. Other much higher frequencies

were also present.

514

Fig. 24.14. Turbine/valve house and outﬂow channel

Pressure transients in water engineering

24.4.2 Comments on observed behaviour

In considering unsteadiness in pressure conduit systems Jaeger (1977)

concluded that no pipeline network is entirely safe against develop-

ment of harmonic vibrations and that there is no reason to conclude

515

Fig. 24.15. Submerged discharge valve bend

Fig. 24.16. Turbulence in SDV sump

Flow instabilities

that any system is also safe from possible resonance developing.

Efforts of designers have been directed towards suppressing any possible

source of incipient vibration from valves, gates, guide vanes, etc.

Generally, the longer a conduit system the greater will be the amount

of damping present and the greater will be the amount of ‘exciting’

energy required to initiate development of an oscillation. In his

studies of resonance, Jaeger (1977) noted that oscillations were more

prone to occur in shorter systems with incidents reported for conduit

lengths of 698, 720, 1430 and 4214 m. The present system has a

relatively short length and also tunnel diameter was sized for operation

of the free discharge valves at greater ﬂows than the SDV. Damping

in the system is therefore quite low. Conﬁguration of branch serving

the SDV is also noteworthy. A step change in diameter from

2000 mm down to 800 mm is unusually large. A higher-frequency

pressure wave oscillation in the SDV branch will largely be reﬂected

from the large section change while longer period oscillations, say

with the overall system period of around 1.6 s, will not be affected in

this way.

Attention was focused on the SDV sump as a source of pressure

ﬂuctuations. The radial jet issuing from the SDV will produce intense

shear and turbulence as it mixes with the surrounding water in the

sump. A state of continuous instability is created with a wide range of

frequencies of oscillation being present in the sump. Tritton (1980)

notes that large eddies have formed from what would seem to be

entirely insigniﬁcant perturbations or local instabilities in the mean

velocity proﬁle. The life of these large eddies can be comparatively

short before a ‘normal’ ﬂow is resumed Al Naib (1992).

24.4.3 Modelling behaviour

It was hypothesised that pressure ﬂuctuations caused by vortex action

within the SDV sump caused by the radial jet motion might be

producing velocity and pressure oscillations at the base of the SDV.

These ﬂuctuations would then be propagated at acoustic velocity,

through the valve and into the system as a whole. To examine the

possible effects of such transients a computer model of the system

was formed with the exit from the SDV subject to a range of frequencies

of oscillation. The frequencies used generally ranged from around 1/

40 s to 1.6 s. The higher frequencies represented ‘typical’ ﬂuctuations

within the radial jet and the longer frequencies were from larger-scale

oscillations.

516

Pressure transients in water engineering

Using this approach it was possible to produce variations of head

of around 9 m in the penstock (Fig. 24.17). This amplitude was

maintained throughout much of the penstock and it is thought likely

that it is the longer-period oscillations that are responsible for the

pulsating sound observed. The longer-period oscillations were not

considered capable of creating cavitation in the SDV. The higher-

frequency components were largely reﬂected from the change of

cross-section (800 mm to 2000 mm). A large amplitude of oscillation

was predicted to occur in the vicinity of the SDV bend. Based on

averaged ﬂow conditions across the pipe cross-section, cavitation was

not thought likely to occur. A more detailed consideration of ﬂow

variation across the bend section was thought necessary. Ideally a

CFD analysis of the bend coupled with information from the transient

model of the system as a whole would be used to yield data on the 3-D

variation of ﬂow and pressure at the bend. Due to limited time, a much

simpler approach was adopted to give some indication of the bend

effect. Rankine’s free vortex approximation was used to produce a

velocity variation across the bend from the outside radius to the

inside of the bend. Figure 24.18 shows a typical result with much

greater velocities and correspondingly lower pressures predicted close

to the inside wall of the bend. It was considered that a transient

cavitation might be occurring towards the inside wall of the bend

produced by a combination of low pressure during oscillations and

reduced pressure during curvilinear ﬂow.

517

Trifurcation

SDV tee

SDV bend

Time (s)

mAD

Fig. 24.17. Head behaviour in SDV system

Flow instabilities

24.4.4 Remedial measures

A number of remedial possibilities present themselves, ranging from

air injection to introducing increased compressibility to sump modiﬁca-

tions to alter the nature of pressure ﬂuctuations. Alteration of the

sump would require considerable time and expense, while air injection

would need to be more or less continuous and involve use of a

compressor.

The simplest remedy is to install a ‘pulsation damper’. This is in

essence a small pressure vessel ﬁtted with a diaphragm and containing

a gas charge. Given that head room may be a problem, a relatively

ﬂat pattern would seem appropriate, such as is illustrated in

Fig. 13.2c. The installation of this damper is shown in Fig. 24.18.

Sizing of the vessel requires computer simulation and a number of

studies were conducted to determine a lower limit of capacity. Figure

24.19 shows results of a simulation. Head variations in the penstock

(SDV tee) and at the SDV bend are shown to be quite modest for

the chosen gas charge of 5 litres. Overall capacity of the damper was

10 litres.

518

Velocity distrbution across bend

Gas charge connection

Diaphragm

Pulsation damper (accumulator)

Connection to SDV branch

Low pressure towards

inside of bend

DN 600

DN 800

Flow

Fig. 24.18. ‘Damper’ installation

Pressure transients in water engineering

24.4.5 Final comments

This installation represents a fairly unusual SDV installation given the

large change of cross-section between the penstock and the SDV

branch. Coupled with the very low-capacity damping system as a

whole, this factor would seem to have allowed turbulent pressure

ﬂuctuations to produce unsatisfactory behaviour within the system,

which would not have been manifest in other circumstances.

The potential for a small gas pocket to amplify transient effects has

been demonstrated in section 18.5 of Chapter 18. In conditions

pertaining at this SDV installation, a small pressure vessel has been

included to alleviate a local problem within a much larger system. It

is desirable to conduct analyses to ensure that no signiﬁcant ampliﬁca-

tion problems can arise within the system generally as a consequence of

including the small gas volume. In this connection studies of SDV

closure were carried out to verify that there was no risk of ampliﬁcation

in this instance.

519

Time (s)

SDV tee

SDV bend

60.0

59.5

59.0

58.5

58.0

57.5

57.0

56.5

56.0

55.5

55.0

54.5

Fig. 24.19. Head variation in SDV system with damper

Flow instabilities

References

Allman, W. B. (1975) Operating Sections Proceedings. American Gas Associa-

tion Conference, Catalogue No. x-50875, May 5—8.

Bishop, R. R. and Lang, D. C. (1984) Design and Performance of Buried Fiber-

glass Pipes — A New Perspective. In Pipeline Materials and Design (Schrock,

B. J. (ed.)), pp. 1—12. New York: ASCE.

Camichel, C., Eydoux, D. and Gariel, M. (1919) Etude The

orique et Expe

rimen-

tale des Coups de Belier. Paris: Dunod.

Carlstrom, B. I. (1981) Structural Design of Underground GRP Pipe. Inter-

national Conference on Underground Plastic Pipe, 30 March to 1 April

(Schrock, B. J. (ed.)), pp. 56—78. New York: ASCE.

Colebrooke, C. F. (1939) Turbulent ﬂow in pipes with particular reference to

the transition region between the smooth and rough turbulent pipe laws.

Journal of the Institution of Civil Engineers, 11, 133.

Collier, S. L. and Hoerner, C. C. (1983) A Facility and Approach to Performance

Test of Check Valves. New York: ASME paper No. 82-PET-8, Vol. 105, pp.

62—67.

Dawson,P.(c.1980)The Design of Descending Pumped Pipelines with Particular

Reference to Efﬂuent Disposal Sea Outfalls. Leeds: Hydraulic Analysis Limited.

Donsky, B. (1961) Complete pump characteristics and the effects of speciﬁc

speed on hydraulic transients. Transactions of the American Society of

Mechanical Engineers, Journal of Basic Engineering, December, 685—699.

Ellis, J. (1980) Study of Pipe—liquid Interaction Following Pump Trip and

Check Valve Closure in a Pumping Station. Proceedings of the 3rd Inter-

national Conference on Pressure Surges, Canterbury, 25—27 March. Paper

E3, pp. 203—220. Cranﬁeld: BHRA.

Ellis, J. (1994) Hydraulic Instability in Water Supply Networks. Proceedings of

the 2nd International Conference on Water Pipeline Systems, Edinburgh, Scot-

land, 24—26 May, pp. 335—357. Cranﬁeld: BHR.

Ellis, J. and Khairulla, L. M. (1974) Oscillations in a surge tank with upper and

lower expansion galleries. Water Power, November, 359—364.

520

Ellis, J. and Mualla, W. (1986) Selection of Check Valves. Proceedings of the 5th

International Conference on Pressure Surges, BHRA Fluid Engineering

Centre, Hannover, Germany, 22—24 September, pp. 213—222.

Ellis, J. and Tint, K. M. (1976) Simulation of pressure transients at valves and bifur-

cations. International Water Power and Dam Construction,8:8,24—27August.

Esleek, S. H. and Rosser. (1959) Check Valve Water Hammer Characteristics.

Proceedings of an American Water Works Association Society meeting,

November.

Fedosoff, F. A. and Szpak, E. (1979) Cyclic Overpressure Effects on High Density

PE Pipe. Proceedings of Western Canada Water and Sewage Conference,

Regina, Saskatchewan. Engineering Digest, September, 35—38.

Fok, A. T. K. (1980) Design Charts for Surge Tanks on Pump Discharge Lines.

Proceedings of the 3rd International Conference on Pressure Surges, Eng-

land, March, BHRA, Paper J3, pp. 445—472.

Gaube, E., Mu

ller, W. and Falcke, E. (1974) The Statics of Rigid Polyethylene

Drain Pipes. Proceedings of Kunststoffe, 64, April, 193—196.

Glass, W. L. (1980) Cavitation and Corrosion in a Pumping Main. Proceedings of

the 3rd International Conference on Pressure Surges, Canterbury, 25—28

March, Paper H3, pp. 401—414. Cranﬁeld: BHRA.

Gordon, J. L. (2004) The untested computer program. Hydro Review World-

wide, 12:13, July, 32 and 34.

Gordon, J. L. (2006a) Nintendo engineers. Hydro Review Worldwide, 14:4,

September, 46 and 51.

Gordon, J. L. (2006b) The hunting impulse turbine. Hydro Review Worldwide,

14:1, March, 46 and 54.

Gotham, K. V. and Hitch, M. K. (1975) Design considerations for fatigue in a

PVC pressure pipeline. Pipes and Pipelines International, 20:2, February, 10.

Graze, H. R. (1968) A Rational Thermodynamic Equation for Air Chamber Design.

Proceedings of the 3rd Australian Conference on Hydraulics and Fluid

Mechanics, Institution of Engineers Australia, 25—29 November, p. 57.

Graze, H. R. and Forrest, J. A. (1974) New Design Charts for Air Chambers.

Proceedings of the 5th Australian Conference on Hydraulics and Fluid

Mechanics, Christchurch, New Zealand, December 9—13, Vol. II, p. 34.

Graze, H. R., Schubert, J. and Forrest, J. A. (1976) Analysis of Field Measure-

ments of Air Chamber Installations. Proceedings of the 2nd International

Conference on Pressure Surges, London, Paper K2. Cranﬁeld: BHRA.

Greatorex, C. B. (1981) The Relationship between the Stiffness of a GRP Pipe and

its Performance when Installed. Proceedings of an International Conference

on Underground Plastic Pipe, New Orleans, 30 March—1 April (Schrock,

B. J. (ed.)), pp. 117—129. New York: ASCE.

Heinsbroek, A. G. T. J. and Tijsseling, A. S. (1994) The Inﬂuence of Support

Rigidity on Waterhammer Pressures and Pipe Stresses. Proceedings of the 2nd

International Conference on Water Pipeline Systems, Edinburgh, UK,

24—26 May, pp. 17—30. Cranﬁeld: BHR Group.

521

References