Home Search Collections Journals About Contact us My IOPscience

Transient Flows in a Pipe System with Pump Shut-Down and the Simultaneous Closing of a

Spherical Valve

This content has been downloaded from IOPscience. Please scroll down to see the full text.

2016 IOP Conf. Ser.: Earth Environ. Sci. 49 052001

(http://iopscience.iop.org/1755-1315/49/5/052001)

View the table of contents for this issue, or go to the journal homepage for more

Download details:

IP Address: 184.174.98.195
This content was downloaded on 15/01/2017 at 14:05

Please note that terms and conditions apply.

You may also be interested in:

Analysis of Wigner energy release process in graphite stack of shut-down uranium-graphite reactor
E V Bespala, A O Pavliuk and S G Kotlyarevskiy

The Concept of a Sub-Prompt-Critical Nuclear Reactor and Its Applications to the Solution of
Safety-Related Neutronic Problems in Reactor Kinetics
S. Das

Numerical methods and transition investigation of transient flows around a pitching hydrofoil
Q Wu, G Y Wang and B Huang

The transient flow of heat through a two-layer wall

Miriam V Griffith and G K Horton

The application of lasers to time-resolved flow visualization

L H Tanner

The effects of area contraction on the performance of UNITEN's shock tube: Numerical study
A M Mohsen, M Z Yusoff and A Al-Falahi

Acoustic transparency in two-dimensional sonic crystals

José Sánchez-Dehesa, Daniel Torrent and Liang-Wu Cai

Experimental study of the cold front propagation in the plasma shut-down experiment in the J-TEXT
tokamak
Yanhua Huang, Yi Tang, Yihui Luo et al.

Transient Flows Through Micro-annuals

B Wiwatanapataphee, Yong Hong Wu, Maobin Hu et al.
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

Transient Flows in a Pipe System with Pump Shut-Down and

the Simultaneous Closing of a Spherical Valve

Zh. Zhang
ETH Zürich: Laboratory of Hydraulics, Hydrology and Glaciology (VAW),
Hönggerbergring 26, 8093 Zürich

[email protected]

Abstract. Because of the limited value of the wave propagation speed in water the propagation
of a pressure surge in transient flows can be tracked in the time series. This enables both the
pressure head and the flow velocity in pipe flows to be determined as a function of both the
coordinate along the pipe and the time. The propagation of the pressure surge includes both
wave transmission and reflection. The latter occurs where the flow section is changed. The
wave tracking method has been demonstrated as highly accurate and subsequently was applied
to much more complex hydraulic systems, in which the pump is shut off and the spherical
valve is simultaneously progressively closed. A combined four-quadrant characteristic of the
pump and a spherical valve has been worked out, with which the computational procedure for
the transient flow in the complex system could be significantly simplified. It has been
demonstrated that not only the pressure surge in the hydraulic system but also the rotational
speed of the pump could be satisfactorily computed. The computational algorithm has been
demonstrated as quite simple, so that all calculations could be performed simply by means of
the Microsoft Excel module.

1. Introduction
Transient flows, for instance, in hydro power plants are encountered at each start and shut-off of
hydraulic machines as well as by regulating the operating points of the involved machines. Since such
flows represent highly complex flow dynamics and always cause rapid pressure rises in the system, for
safety reasons the related flow phenomena and pressure increases have to be considered already in the
earlier stages of construction. The operation as well as its variation of each hydraulic system are then
all subjected to the conditions, which are specified based on the transient flow analyses.
A rapid pressure rise in a transient flow caused, for instance, by closing a valve is called hydraulic
shock or water hammer. For calculating transient flows, there basically exist two available
computational methods which are based, respectively, on rigid and elastic water column theories. The
former theory relies on the momentum equation i.e. Newton’s second law of motion in mechanics. For
a complex network which, for instance, consists of x pipes of different diameters and connections, x
momentum equations have to be simultaneously and dependently solved. This requires iterative
numerical computations and is always associated with high computational expenses. The latter, i.e. the
elastic water column theory takes the compressibility of the water into account. It is considered a
special field in hydromechanics and, therefore, has widely been investigated since the initial
investigations carried out by both Joukowsky and Allievi. Departing from the basic equation of

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

Joukowsky h   a g  c for the local pressure rise in the flow, the propagation of the pressure wave
is the dominant variable and, therefore, has to be followed when calculating transient flows and the
related pressure surges in the system.
The compressibility of water as an additional phenomenal aspect in transient flows, actually,
functions as a useful feature which contributes to simplifications of the calculations. This is simply so,
because, for the limited propagation speed of the pressure surge, the flows at different locations do not
simultaneously affect each other. At each time step in a time series, the flow at any given location can
then be separately computed, each time by only solving one equation.
The method of computing transient flows has been progressed from the graphical approach
(Schnyder-Bergero) for simple flow systems to the much more efficient and powerful computational
standard based on nowadays computer technology. The most widely applied method is, obviously, the
characteristic method of solving the wave equation. The method is based on dividing the pipe length
into n sub-intervals of equal length, as can be found for instance in [1, 2]. In many cases, however, the
sub-intervals of equal distance do not appear to be necessary when only the local pressure rise in the
flow, for instance, at the regulation valve, is in focus. For this reason the method of tracking the
pressure wave seems to be much more accessible.
Basically, the wave tracking method had been indicated in the early stages of investigations by
Allievi [3]. It has also been explained in [4-8]. This method does not require dividing the pipe length
into equal-distant sub-intervals, except for the graphical method at earlier times. The advantage and
applicability of this method will be demonstrated in this paper regarding transient flows in a complex
hydraulic system.
Typical transient flows in a hydraulic system of hydropower plants with pumps or turbines are
often encountered at the start and stop of the machines. The hydraulic system usually comprises both
low and high pressure pipe systems, each with one or more surge tanks. The related dynamic process
is very complex when, for instance, the shut-off of the pump is associated with a simultaneous closing
of the spherical valve on the pressure side of the pump. The calculation of the related transient flow,
therefore, requires knowledge of the hydraulic characteristics of both the pump and the valve. As can
be expected, it would be very convenient if the hydraulic characteristics of both the pump and the
spherical valve can be combined to one characteristic. The current paper shows the computational
method and applications of such a combined characteristic based on the use of four-quadrant diagrams
of the pump.
As a matter of fact, transient flows and their calculations have been regarded as highly complex, so
that a simple computational method still appears to be inaccessible to everyone. Nearly all available
commercial software is based on long-term developments. The circumstance of using commercial
software highly restricts the engineer’s activities in dealing with engineering transient flows. For this
reason, the current paper provides a well validated method of simply using the tabular computations,
e.g. based on the use of Microsoft’s Excel program. Due to the restriction to the paper length only few
calculation results in the computational example are shown. More results will be presented in a journal
paper.

2. Basic Equations of Transient Flows

From the momentum and continuity equations for viscous flows in a pipe of constant cross-section
(Fig. 1), the wave equation is derived as

(A) x (B)
F
f F c A
f
Fig. 1 Pressure wave propagation LA LB
Injector
in the pipe of constant section L

2
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

d   
h  h0  c  c 0   
a
c c  c0 c0  (1)
dx  g  2 gd

The subscript 0 refers to steady state before a pressure surge is generated. For frictionless flows one
concludes that the quantities h  h 0   a g c  c 0  remain constant while traveling with the wave
propagation speed a along the pipe in two opposite directions. Else, they change by a value (friction
term) which depends on both the flow direction and the flow velocity, in terms on the r.h.s of Eq. (1).
Equation (1) will be integrated along the pipe in both the positive x-direction from x A to x and the
negative direction from xB to x (Fig. 1). Thus, with LA  x  xA and LB  x  xB  , both are positive,
one obtains

c  c0   h  h0  a c  c0   LA c c  c0 c0  ,
a  
h  h0  (2)
g  g  A 2 gd

c  c0   h  h0  a c  c0   LB c c  c0 c0  ,
a  
h  h0  (3)
g  g  B 2 gd
In the integration of the friction terms in both equations, the “First Mean Value Theorem for
Integrals” has been applied. For simplicity, equal mean velocities in both equations have been used.
The resulting error is one order smaller than the friction terms and thus two orders smaller than the
pressure terms in the above equations.
The above two equations indicate that from the known flows at A and B the flow at any other
locations x at a certain later time can be determined. The time delays are tA  LA a and tB  LB a ,
respectively.
For further calculations the following wave parameters are introduced:

2 F  x  at   h  h0  c  c0  , 2 f x  at   h  h0  a c  c0  ,
a
(4), (5)
g g

   
2 FB   h  h0  c  c0  , 2 f A  h  h0  c  c0  .
a a
(6), (7)
 g B  g A
If the positive velocity c is defined in the other direction, then the sign related to c has to be changed.
With the above definition the wave parameters F and f travel, respectively, in the negative and the
positive x-directions (Fig. 1). Much often, F is called the primary wave as it arises from the regulation
organ. f is the reflected wave from the reservoir. Both F and f are independent of the definition of the
x-direction. For the case of Fig. 1 and at the location of the injector valve there is a time delay
t  2 L a between F and f.
Departing from the above definitions and in using the flow rate Q=cA, Eq. (2) and (3), respectively,
are written as

f  x  at   f A  RA Q Q  Q0 Q0  , F x  at   FB  1 RB Q Q  Q0 Q0  .
1
(8), (9)
2 2
The resistance constants are given by
 LA
RA  2
and RB  LB 2 . (10)
2 gdA 2gdA
For computing transient flows one needs only to track the parameters F and f which propagate by the
wave speed a in two opposite directions. At a given position in the flow, both the pressure head and
the flow velocity can then be calculated from Eqs. (4) and (5) as

3
 28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

h  h0  F  f , c  c0 
g
 f  F . (11), (12)
a

3. Generation of the Pressure Surge

The transient flow may be caused by starting and stopping the machine or by regulating the operating
point via regulation organs. For simplicity, the flow rate regulation in a Pelton turbine is considered by
regulating the injector opening (Fig. 1 and 2, the Pelton wheel is not shown). By using the Bernoulli
equation for the jet speed the flow rate through the injector is calculated as

 1 Q2 
Q  c0 Ajet  AD0 2ghtot or Q  AD0 2 g h  .
2 g A2 
(13), (14)

The term with Q2 under the square root sign represents the dynamic pressure head in the flow ahead of
the injector. The regulation of the flow rate occurs by varying the needle position in the injector nozzle
(Fig. 2), leading to the variation of the jet section. The ratio of the jet section to the opening area AD0
of the nozzle is called contraction factor (  ). It represents the injector characteristic and can be
considered as known [11]. From Eq. (14) the flow rate is then resolved as
2 gh
Q  AD0 . (15)
1  AD0 A 
2

This equation is inserted into Eq. (5) with Q  cA , leading to

a  AD0
h  h0*  2 f   0 ,
2g
h (16)
1  AD0 A 
2
g A

in which, the initial stable flow condition has been expressed as

a
h0*  h0  Q0 . (17)
gA

Equation (16) represents a quadratic polynomial for the parameter h to each given value of  . One
point to be mentioned is that one will obtain a quadratic polynomial for the flow rate Q, when Eq. (15)
is combined with Eq. (5) for eliminating the head h.
With the solution of h from Eq. (16), thus, the primary wave F is obtained from Eq. (11) as
F  h  h0  f . (18)
Within the time t  2 L a , according to Fig. 1, there is f = 0 in both Eqs. (16) and (18), so that the
primary wave is simply calculated by F  h  h0 . After a time delay of t  2 L a the wave f, which is
the reflection of the primary wave F at the reservoir, travels back to the injector and is available
because f t    F t  2 L a  .

N
D0

Needle
htot
Fig. 2 Injector nozzle of Pelton s
turbines and its dimensioning

4
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

In most cases of hydraulic systems, the initiation of transient flows can be simulated by a Pelton
injector as a regulation organ, without having to be concerned with the fluid machines and their
characteristics.

4. Verification of the method

The computational algorithm presented above has been applied to a hydraulic system of Pelton
turbines (Fig. 3a) by comparing computational results with experimental measurements [9]. The
pressure head measurement was accomplished while the Pelton injector was getting closed within 45
s. As can be confirmed from Fig. 3b, computational results very well agree with the measurements.
Compared to the simple hydraulic system shown in Fig. 1, the algorithm related to Fig. 3 is more
complex because of the presence of the surge tank. In the realization, a T-type branch connects the
surge tank with the main pipe. The wave propagation through such a T-type branch includes wave
transmissions and reflections. According to the sketch in Fig. 3a, the incoming waves F1, f2 and f3 are
considered as known. The departing waves comprise both the transmission and the reflection of
incoming waves and can be calculated as follows [10]:
A1  A2  A3 2 A2 2 A3
f1  F1  f2  f3 , (19)
A1  A2  A3 A1  A2  A3 A1  A2  A3

2 A1 A  A1  A3 2 A3
F2  F1  2 f2  f3 , (20)
A1  A2  A3 A1  A2  A3 A1  A2  A3

2 A1 2 A2 A  A1  A2
F3  F1  f2  3 f3 . (21)
A1  A2  A3 A1  A2  A3 A1  A2  A3
These three equations represent the fundamental equations in the wave tracking method for computing
the transient flows in each complex hydraulic system. The above equations also apply to a pipe with
abrupt section change. One only needs to set A3 = 0.

Surge tank A3 Lake

Q3

A1, Q1 A2, Q2

to Pelton
turbine f3 F3

F1 F2

(a) f1 f2 (b)

Fig. 3: Comparison between the measured and computed pressure responses in the hydraulic system
of a Pelton turbine after the shut-down of the turbine (Gr. 1)

5. Unification of Characteristics of the Pump and the Spherical Valve

Large pumps in pumped hydro power plants are always installed in connection with a spherical valve
close to the pump on its high pressure side, as sketched in Fig. 4. While the pump, for instance, is
running down, the spherical valve should be progressively closed. For computations of so resulted
transient flows one must know the characteristics of both the pump and the spherical valve. It seems to
be very convenient to consider the pump and the spherical valve as an integrated hydraulic unit and to
use the unified characteristic. As is expected, the unified characteristic should show the flow rate as a

5
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

function of the rotational speed of the pump and the opening degree of the spherical valve. On the
other hand, the unified characteristic has to be represented in a suitable mathematical form, in order to
achieve easy applications.

H12
unified
Pump
Sph. valve

Hpu h
Fig. 4 Unification of a pump and a
spherical valve H12

For the sake of demonstrating the computational method only the case of shutting down the pump by
simultaneously closing the spherical valve is considered. This also includes the reverse rotation of the
pump and the back flow through the pump.
To the pump characteristic only four-quadrant diagrams are applicable. Accordingly, the following
dimensionless parameters for the rotational speed, the flow rate and the shaft torque are defined
d 2 n Q M
n11  , Q11  , M11  . (22)
2gH pu d 2
2 2gH pu d gHpu
3
2

In these definitions, n is in 1/s and Q in m3/s. d 2 is the pump impeller diameter.

Fig. 5 shows the four-quadrant diagrams of a pump, which are measured from laboratory tests.
They will be used in the present paper to compute the related transient flows. To obtain the combined
characteristic, the following 4 steps, separated in Sects. 5.1-5.4, should be performed.

(a) (b)

Fig. 5 Four-quadrant diagrams of a pump (no dimensions are shown because of company policy)

5.1. Pump Characteristic: Flow Rate

For later convenience of calculations the four-quadrant diagram in form of Q11  f n11  in Fig. 5a
should be approximated by the following quadratic equation
Q112  q  m1 n11Q11   m2 n112 . (23)
The reason of the mathematical form in this equation will become clear below. The constants q, m 1
and m2 can be determined by the least square curve fitting method with two independent parameters

6
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

n11 Q11 and n112 , which can be obtained from the available data Q11  f n11  from Fig. 5a. Because the
calculation may cover through a range of rotational speed from negative to positive values, the
characteristic Q112  f n11Q11 , n112  , in the current case, must be approximated at least in two areas as for
n11  0 .88 and n11  0 .88 , respectively.
Equation (23) is then written in the explicit form
1  Q2 Q 
H pu   4  m1n   2 m2 n 2 d 22  . (24)
2 gq  d 2 d2 
This form of the pump characteristic applies to a pipe system with one pump. For two or more pumps
in the pump house the flow rate in the above equation must be related to one pump.

5.2. Pump Characteristic: Torque

The torque characteristic, M, in the combined unit (pump and valve) remains the same as for the pump
alone. For computational purposes the relation that is shown in Fig. 5b can be directly used.

5.3. Spherical Valve Characteristic

The hydraulic characteristic of a spherical valve is represented by the head drop in the flow through
the valve as a function of the opening degree  , with   90  for the fully open valve. With respect
to a possible back flow, the head drop at the spherical valve is given by
cc QQ
h  cp  cp 2
. (25)
2g 2 gAsph

The sectional area of the spherical valve is usually equal to sectional area of the pipe. For nearly all
spherical valves in the hydropower plants, the pressure drop coefficient has been approximated by
Zhang [10], in most appropriate form as given by
10 6
cp  1. (26)
0 .108  2
 1 .35  
2

5.4. Unification of Characteristics of the Pump and the Spherical Valve

According to Fig. 4 the total pressure head over the combined unit is given by
H12  H pu  h . (27)

Inserting Eq. (24) and (25) into this equation yields

 1 c q Q  2 m1n
  p2 Q  Q  m2 2 n 2 d 22  2 gqH 12   0 . (28)
d 4
A Q  d
 2 sph  2

This describes the unified characteristic of the pump and the spherical valve. It represents a quadratic
polynomial of the dependent variable i.e. the flow rate Q , which can, thus, be immediately
calculated. This simply justifies the reason of the mathematical expression that is proposed in Eq. (23).
In Eq. (28), Q Q  1 , depending on the flow direction. With respect to cp  f   from Eq. (26)
the unified characteristic can be represented in the general functional form as
QPu  f H 12 , n,   . (29)
It should be reminded that the constants q, m1 and m2 are applicable each time only to an appropriate
area in the four-quadrant characteristic. This has already been explained at Eq. (23).

7
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

The flow rate, according to Eq. (29), also depends on the rotational speed which by winding down
the pump is not constant. The development of the rotational speed, in turn, depends on the total
moment of inertia (J) of the rotor system. From the momentum law and with respect to the resistance
torque M exerted on the rotor system the angular momentum balance, after the motor is shut down, is
given by
dn
2 J  M , (30)
dt
With the aid of Eq. (22) one obtains
dn   5 n2
 d 2 2 M 11 . (31)
dt 4 J n11

As for M 11  f n11  , each time when n112 and, thus, M 11 have been computed, then dn dt can be
calculated straightforwardly. In the numerical method, the rotational speed in the current time step is
then given by ni  ni-1  dn dt i-1 t .

5.5. Generation of the Primary Waves

The wave generation on both the suction (index 1) and pressure (index 2) sides of the pump should be
considered, at which the flow rate can be assumed to be equal (Q1=Q2). On the suction side of the
pump, the pressure head increase is given, as from Eq. (5), by

h1  h1, 0  2 f1 
a
Q  Q0  (32)
gA1
Within the reflection time of the first pressure surge, the reflected wave is zero (f1=0).
On the pressure side, where the positive flow rate is in the direction of the primary wave, the
pressure-head increase is followed from Eq. (5) to be

h2  h2, 0  2 f 2 
a
Q  Q0  . (33)
gA2
The pressure difference on both sides of the combined unit is then obtained as
a 1 1
h2  h1  h2 ,0  h1, 0   2 f 2  f1     Q  Q0  . (34)
g  A2 A1 

5.6. Pump Flow in Association with the Wave Parameters

Combining Eq. (34) with Eq. (28) one obtains, with H 12  h2  h1 and H12 , 0  h2 , 0  h1, 0 according to
Fig. 4, the final determining equation as
 1 c q Q  2  m1n  1 1 
  p2 Q    2aq     Q
d   d 2  A2 A1  
4
 2 Asph Q  . (35)
a 1 1 
 2 gq   Q0  m2 2 n 2 d 22  2 gq H 12 , 0  2 f 2  f1   0
g  A2 A1 

This equation remains of quadratic form for the flow rate Q. From tracking the wave parameters f1 and
f2 the flow rate at each time can immediately be calculated. In the numerical computations, the actual
value Q Q  1 in the above equation can be set to be equal to that of the last finite time step. The
rotational speed n is obtained directly from d n d t in that time step.

8
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

As soon as the flow rate is calculated, the primary waves both on the suction and the pressure sides,
respectively, can again be computed from Eq. (12) as

F1  f1 
a
Q  Q0  , F2  f 2  a Q  Q0  . (36), (37)
gA1 gA2

6. Application of the Method

Figure 6 shows a pumping system in the Oberhasli Hydroelectric Power Company (KWO). At this
pumping system a test of emergency shut-down of the pump with simultaneous closing of the
spherical valve was executed [10]. A part of the measurement results will be used to confirm the
numerical calculations based on the MS Excel tool.
Lake
Tilted Surge
surge tank tank

Surge
tank
Lake
Pump house

Fig. 6 Pumping system

The basic equation for the transient flow computations is Eq. (35). For the wave parameters f1 and f2,
one then needs to separately track the propagation of the waves F and f in both the pressure and
suction sides of the pump. It is actually the same computational process as that shown in Sect 2.
However, the computations must be extended to the current case with three surge tanks and eight pipes
of different diameters. For calculating all relevant flow parameters and the time-dependent rotational
speed of the pump, more than 40 columns in the scheme of tabular computations are necessary.
Figures 7 to 8 show the parts of computational results which are compared with measurements. Not
only the pressure response in the hydraulic system but also the rotational speed of the pump could
precisely be predicted by the computations. Such a high level of computational accuracy relies on the
fact that highly reliable assumptions have been applied. Because the MS Excel tool is in this case
powerful and, thus, appropriate for such computations, on-line investigations of all flow and operating
parameters can be easily performed.
Because of the restriction of the paper length more detailed computational results will be presented
in a journal paper.

Fig. 7 Pressure response on the suction side Fig. 8 Rotational speed of the pump

9
28th IAHR symposium on Hydraulic Machinery and Systems (IAHR2016) IOP Publishing
IOP Conf. Series: Earth and Environmental Science 49 (2016) 052001 doi:10.1088/1755-1315/49/5/052001

7. Conclusion
Based on the investigations and applications carried out by the author the wave tracking method
among diverse computational algorithms has been demonstrated to be very easily applicable, also to
highly complex system with unsteady elements like the T-branches as well as with regulation organs
and hydraulic machines (pumps, turbines etc.). By following the propagations of the wave parameters
F and f (as solutions of the wave equation) in the system and through unsteady sections, the transient
flow can be very easily and accurately calculated, for instance, by simply using the MS Excel-sheet.
As demonstrated, the method also applies to the transient flow caused by the pump shut-down with
simultaneous closing of the spherical valve. The key elements of the precondition are to work out the
combined characteristics of the pump and the spherical valve on the one hand and to obtain a quadratic
polynomial for easily solving the flow rate on the other hand.

Acknowledgement
The author wishes to thank the Oberhasli Hydroelectric Power Company (KWO) for the projects
concerning transient flows and flow calculations. He gratefully acknowledges the Laboratory of
Hydraulics, Hydrology and Glaciology (VAW) at ETH Zurich for wider supports for completing this
paper.

References
[1] Tullis J 1989 Hydraulics of pipelines – pumps, valves, cavitation, transients (New York: John Wiley &
Sons)
[2] Horlacher H and Lüdecke H 2012 Strömungsberechnung für Rohrsystem (Renningen: Expert)
[3] Allievi L 1902 General theory of the variable flow of water in pressure conducts Theory of water-hammer
translated by Halmos E 1925
[4] Dubs R 1947 Angewandte Hydraulik (Zürich: City-Druck)
[5] Jaeger C 1949 Technische Hydraulik (Basel: Birkhäuser)
[6] Giesecke J and Mosonyi E 2009 Wasserkraftanlagen 5. Auflage (Heidelberg/ Dordrecht/London/New
York: Springer)
[7] Sharp B 1981 Water hammer, problems and solutions (London: Edward Arnold Ltd)
[8] Chaudhry M 2014 Applied hydraulic transients 3rd edition (New York Inc. Springer)
[9] Zhang Z 2009 Druckstossberechnung beim Schliessen der Peltonturbine in Grimsel 1 Technischer Bericht
(Innertkirchen: KWO 16. Juni 2009)
[10] Zhang Z 2015 Private information
[11] Zhang Z 2009 Freistrahlturbinen (Berlin: Springer)

10