Nuclear Engineering and Design 261 (2013) 165173

Contents lists available at SciVerse ScienceDirect

Nuclear Engineering and Design

journal homepage: www.elsevier.com/locate/nucengdes

Water hammer characteristics of integral pressurized water reactor primary loop

Qiaolin Zuo a , Suizheng Qiu a, , Wei Lu a , Wenxi Tian a , Guanghui Su a , Zejun Xiao b
a
b

School of Nuclear Science and Technology, Xian Jiaotong University, Xian, Shanxi 710049, PR China
Nuclear Power Institute of China, Chengdu, Sichuan 610041, PR China

h i g h l i g h t s

Water hammer models developed for IPWR primary loop using MOC.
Good agreement between the developed code and the experiment.
The good agreement between WAHAP and Flowmaster can validate the equations in WAHAP.
The primary loop of IPWR suffers from slight water hammer impact.

a r t i c l e

i n f o

Article history:
Received 21 July 2012
Received in revised form 2 March 2013
Accepted 10 March 2013
Keywords:
Water hammer
Integral pressurized water reactor
Method of characteristic
Experimental verication
Flowmaster

a b s t r a c t
The present work discussed the single-phase water hammer phenomenon, which was caused by the
four-pump-alternate startup in an integral pressurized water reactor (IPWR). A new code named water
hammer program (WAHAP) was developed independently based on the method of characteristic to
simulate hydraulic transients in the primary system of IPWR and its components such as reactor core,
once-through steam generators (OTSG), the main coolant pumps and so on. Experimental validation for
the correctness of the equations and models in WAHAP was carried out and the models t the experimental data well. Some important variables were monitored including transient volume ow rates, opening
angle of valve disc and pressure drop in valves. The water hammer commercial software Flowmaster
V7 was also employed to compare with WAHAP and the good agreement can validate the equations in
WAHAP. The transient results indicated that the primary loop of IPWR suffers from slight water hammer
impact under pump switching conditions.
2013 Elsevier B.V. All rights reserved.

1. Introduction
Water hammer is a kind of shock wave, which is caused by a sudden pressure change within the compressible liquid and pipes. It is
harmful to all kinds of pressure pipelines and happens frequently,
causing valve failure, pipe leak, ow passage components damage and so on. Water hammer is one of the important factors that
threaten the safety of nuclear power plant. For example, in pressurized water reactor (PWR), violent thermo-hydraulic transients
in the primer loop may cause pipe rupture which may induce loss
of coolant accident (LOCA) and lead to reactor shutdown. In 1973,
a severe water hammer accident happened in American Indian
Point 2 nuclear power plant, which caused the containment thermal deformation and main feed water pipe breakage as large as
45.7 cm at the penetrating place through the containment (Liu,
1987). In China, the pipe line has vibrated voilently and led to

Corresponding author at: School of Nuclear Science and Technology, Xian

Jiaotong University, Xianning West Road, No. 28, Xian, Shanxi 710049, PR China.
Tel.: +86 29 82665607; fax: +86 29 82665607.
E-mail address: szqiu@mail.xjtu.edu.cn (S. Qiu).
0029-5493/$ see front matter 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.nucengdes.2013.03.038

alarm signals for many times since Daya Bay and Ling-ao nuclear
power plants were put into operation, which can even cause the
condensate extraction pump to shut down (Zhang and Zhu, 2008).
Therefore, water hammer phenomenon in nuclear power plants has
drawn much attention for nuclear power plant safety and economic
issues in recent years.
In nuclear power plants, water hammer transients can be caused
by both two phase ow and single phase ow transients. Two-phase
water hammer is often associated with condensation-induced phenomenon, which may cause greater damage than single-phase
water hammer (Beuthe, 1997). Two-phase water hammer often
appears in nuclear power plant and has been investigated by
many researchers (Barna et al., 2010; Barten et al., 2008; Beuthe,
1997; Prica et al., 2008). However, the detail mechanisms of the
condensation-induced water hammer have not been fully understood. Much attention has been drawn to single-phase water
hammer phenomenon. Some sophisticated algorithms are developed and applied to specic single phase ow systems.
Water hammer prediction is mainly carried out by the analysis
method, the graphic method (Allievi, 1925), and the numerical
method in the early stage. The classical water-hammer equations
are a set of hyperbolic equations, which are acquired by the

166

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

Nomenclature
a
A
D
f
g
H
I
K
m
n
Nr
Q
t
Ve
V
w
x


sound velocity (m/s)

ow area (m2 )
pipe diameter (m)
friction factor
acceleration of gravity (m/s2 )
pressure head (m)
rotational inertia of the valve disc. (kg m)
effective elastic modulus (Pa)
the dimensionless form of the axial torque
the dimensionless form of the rotational speed
the rated rotational speed
volume ow rate (m3 /s)
time (s)
the container volume (m3 )
uid velocity (m/s)
the angular velocity (rad/s)
length (m)
pipe inclination angle ( )
valve opening angle( )
the dimensionless opening

dimensionless method (Ghidaoui et al., 2005). The solving procedure of the analysis method is very complicated. The analysis
method is used to solve the basic simplied equations and only
applied to simple pipe networks when the loss of water head is
neglected. The graphic method is very complex in drawing and
not accurate enough. The development of computer technologies
makes the numerical method of water hammer simulation in
complex pipe networks become possible.
In recent years, the numerical method is widely used in water
hammer phenomenon study and almost replaced the analysis
method and the graphic method. The main numerical methods
to simulate water hammer events include method of characteristic (Wylie and Streeter, 1993), the nite volume Method (Zhao
and Ghidaoui, 2004), the nite element method (Kochupillail et al.,
2005), Wavelet-Galerkin (Sattar et al., 2009), the uid structure
interaction, and so on. Among those methods the method of characteristic (MOC) is the most popular one, and Afshar and Rohani
(2008) even developed a different MOC procedure IMOC. Some
research indicates that MOC ts experimental data well (Liu et al.,
2005). Ghidaoui et al. (2005) investigated eleven available water
hammer commercial software packages, and found that in eight of
them the method of characteristic was applied.
Flowmaster is one of the most well-known virtual thermouid modeling software to simulate water hammer. It can analyze
uid ow and pressure surge through complex piping networks.
Flowmaster has been widely used in water hammer calculation in
nuclear system in recent years (Marcinkiewicz et al., 2008; Zhang
et al., 2012). Lee analyzed hydrodynamic characteristics of auxiliary feed water system in PWR using Flowmaster and showed a
good agreement between the simulation and the measurements
(Lee et al., 2011).
There are many factors which lead to water hammer in PWR.
Such as power failure, reactor normal shutdown and startup, pump
shutoff and startup, pump blockage, valve closure, open and instability, pipe rupture, rapid condensation, transient void. Generally
speaking, normal startup and shut-off operation could not lead to
excessive water hammer pressure, but sudden power failure of the
pump and accidental pump shutoff often lead to severe water hammer. In order to evaluate the safety situation of IPWR, the method
of characteristic is employed in the present work to predict water

hammer in primary loop during pump startup, pump power failure

and pump switch processes.
2. Structure of the primary loop
The primary system arrangement of the present IPWR is shown
in Fig. 1, the primary coolant water at 15.5 MPa ows from the
pumps to the downcomer, then travels downwards to the lower
plenum and enters the reactor core. The coolant picks up heat and
exits into upper plenum. Heated primary uid exchanges energy
with the secondary uid in steam generator and nally returns
to the pump. In both WAHAP and Flowmaster, simplied models
of the primary loop and its components are developed. And the
schematic diagram of the models in WAHAP is displayed in Fig. 2.
The arrows demonstrate the ow direction. Container A refers to
the chamber at the outlet of the steam generator primary side,
respectively. The symbols (1)(2) refer to the simplied pipelines of
the OTSG. Pipe 1 with restrictive device refers to the core passage.
Pipe 2 and pipe 3 present the core bottom plate and the downcomer.
Pipe 4 with restrictive device is heat exchange tubes in steam generator and pipes 7, 8, 12, and 13 with restrictive devices refer to the
pipelines adjacent to the pumps outlet. The length of the pipes in
demonstrated in Table 1
The structure of the four pumps in parallel is designed in case
of pump failure. Four eccentric buttery valves are installed in the
pump outlet to prevent reverse ow. When pump starts, shut down
or valve closes, water hammer may happen. The valve structure and
the denition of the opening angle are illustrated in Fig. 3. The valve
disc has an natural opening angle of about 70 in the gravitational
eld. The opening angle is more then 90 when the valve closes.

Fig. 1. The primary system arrangement of the IPWR.

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

167

Fig. 2. The schematic diagram of the models.

Table 1
The length of the pipes in IPWR.
Pipe numbers

5/6/10/11

7/8/12/13

Pipe length (mm)

1100

341

1441

1800

1627

1627

616

3.2. The solution method (MOC)

The above governing equations are rst-order hyperbolic partial differential equations. The method based on the characteristic
of the equations is very effective for the hyperbolic equations. MOC
has been developed for a long time and its main idea is the transformation of partial differential equations into ordinary differential
equations. The explicit methods can be applied in MOC to obtain
more rapid solution. (Todreas, 2001). The nite-difference equations of MOC have been thoroughly discussed by Wang (1995)
and Tian et al. (2008), and these equations are shown concisely
below.
The hyperbolic Eqs. (1) and (2) have two real eigenvalues,
 = V a. As demonstrated in Fig. 4, along these two characteristic
lines, the hyperbolic equations can be expressed in the following
ordinary differential form:

Fig. 3. Structure of eccentric disc type check valve.

3. Governing equations and the solution method

3.1. Governing equations

C+ :

Ghidaoui et al. (2005) summarized both the historic developments and the present research in the eld of water hammer. They
discussed both various types of mass and momentum equations
for one-dimensional ow and the assumptions and restrictions
involved in these governing equations. The control equations of
the water hammer can be conventionally expressed as (Tian et al.,
2008; Wylie and Streeter, 1983)
H
H
a2 V
+V
+ V sin  +
=0
g x
t
x

dx

dt = V + a

 
fV V 

1 dH + 1 dV + 1 V sin  +
= 0
a dt

g dt

2g D

(1)

 

fV V 
V
H
V
=0
+V
+g
+
2D
t
x
x

(2)
Fig. 4. The schematic diagram of container model.

(3)

168

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

C :

dx

dt = V a

 
fV V 

1 dH + 1 dV 1 V sin  +
= 0
a dt

g dt

(4)

2g D

Given that V  a, and V sin /a is small enough to be ignored, we

obtain the compatibility equations:

C+ :

dx

dt = a

C :

g dt

1 dH + 1 dV +
a dt

g dt

 
fV V 

(6)

2t + t+t

d2
= tt
dt 2
t 2

2g D

= 0

a
gA

and let R be scaling for equivalent frictional resistance,

Hp H = HP2,1 HP1,NS

fx
R=
2A2 g D
Substituting this term of B and R in Eqs. (5) and (6) and integrating along the characteristic lines, we can rewrite the following
nite difference equations as follows.

C + : Hi,t+t = Hi1,t B(Qi,t+t Qi1,t ) RQi1,t Qi1,t 

C : Hi,t+t = Hi+1,t + B(Qi,t+t Qi+1,t ) + RQi+1,t Qi+1,t 
Assuming that

(7)
(8)

HP2,1 = head of the rst node of the pipe connect to the outlet of
the valve.
HP1,NS = head of the end node of the pipe connect to the entrance
of the pump.
H = head loss of the valve. Hp = head of the pump
The positive characteristic equation at the node (p1,NS) and the
negative characteristic equation at the node (p2,1) are as follows:
HP1,NS = CP1 B1 QP1,NS

continuity equation is shown as follows:

QP = QP1,NS = QP2,1

 

(9)
H = Hr

Then the local pressure head and volume ow rate are expressed
as
CP + CM
=
2

(10)

CP CM
2B

(11)

Qi,t+t =

(16)

where QP = the ow rate of the pump.

The relation between head loss of the valve H and volume ow
rate is given by (Yang, 2000):

we obtain

Hi,t+t

(15)

HP2,1 = CM2 + B2 QP2,1

CM = Hi+1,t BQi+1,t + RQi+1,t Qi+1,t 

Hi,t+t = CM + BQi,t+t

(14)

CP = Hi1,t + BQi1,t RQi1,t Qi1,t 

Hi,t+t = CP BQi,t+t

(13)

4.1.2. Pump model at the operation condition

The head and ow rate of the pump and the valve in each branch
are solved together.
Head balance equation is shown as follows:

Let B be scaling for equivalent characteristic impedance

B=

(12)

(5)

2g D

dx

dt = a

d2
= MRF + MP + MH + MB MV MF + MR
dt 2

Where MRF is the momentum of the reacting force when valve

disc strikes the momentum of the reacting force when valve disc
strikes the valve seat. MP , MH , MB , MV , MR refer to the momentum of
static pressure drop, impact of water, buoyancy of valve disc, gravity, friction and resistance, respectively. I is the rotational inertia of
the valve disc.
Eq. (12) is solved by central difference and the valve opening
angle is calculated as follows:

 
fV V 

1 dV
1 dH

= 0
+
+
a dt

4.1.1. Valve disc movement model

The valve opening angle can be determined by the following
equation (Tian et al., 2009):

4. Component models
Pumps, valves and containers are important components of the
system, and their models are described as follows. Other detailed
boundary conditions such as local resistance device, pipe intersection can be found in Wang (1995).

QP QP 
Qr2  2

= Hr

q |q|
QP
,q =
Qr
2

(17)

where  is the dimensionless opening of the valve, and 0  1. 

is in proportion to the valve opening angle which was described
in Section 4.1.1. Hr and Qr are head loss and ow rate respectively when valve opens widely ( = 1). Considering reverse ow
condition in the pump, the absolute value of QP is used.
Solving the above equations, the head balance equation for the
pump and the valve can be obtained.
H + CP1 CM2 (B1 + B2 )Qr q

Hr |q|q
=0
2

(18)

4.1.3. Pump and valve model at failure and startup conditions

4.1.3.1. Pump power failure model. The rotational equation has the
form of
dw
= Mg M
dt

4.1. Pump and valve model

The valve is installed at the pump discharge in each branch in

IPWR. Therefore the pump and valve in each branch are simultaneously simulated.

where J refers to the sum of the rotational inertia of the running

part of the unit and the rotational inertia of the uid in the running
part of the unit, w is the angular velocity. Mg is the motor torque,

(19)

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

and M is the axial torque. When the pump power fails, Mg decreases
to zero.Given that
w=

nNr
= nwr
30

(20)

M = Mr m = Pr m/wr
where Pr is the rated power, Nr is rated rotational speed. n and m are
the dimensionless form of the rotational speed and the axial torque
which are the basic parameters of the four quadrant characteristic
curve of pumps.
Substituting angular velocity w and the torque M in Eq. (19), we
obtain:
Jr dn
= m
Mr dt

(21)

Inertia time constant of the unit Ta =

(21) is integrated in the form of

n = n0

1
Ta

Jr
Mr

is introduced, and Eq.

169

4.1.3.2. Pump startup model. The main coolant pumps include two
startup modes, high speed startup mode and low speed startup
mode. Their designed head is 34.45 m and 4.01 m respectively.
Suppose the rotating speed of the pump increases linearly to
the rated rotating speed within two-thirds of the total starting
time T,

n=

t
2

2 , t < 3T
T

(31)

1 t T
3

Multi-step iterations should be performed when solving q and

the interpolations of four quadrant characteristic curve of the
pumps are time-consuming processes. Therefore, in this paper, the
relationship among dimensionless rotating speed n, volume ow
rate Q and head H is given by
H = A2 Q 2 + A1 nQ + nHe

t
m dt

(22)

t0

Expend m at t = t0 , we have
0 t +
m = m0 + m

0
m
t 2
2!

(23)

(32)

where He refers to the head when pump ow rate is 0. Both A1 and

A2 are constant in a certain start mode.
Eq. (32) can only be suitable for impeller rotation in the forward direction at the initial stage of pump startup and shutting
down.
Solving Eq. (18) with Eq. (32), the volume ow rate of the pump
can be obtained.

= d2 m/dt 2
= dm/dt, m
m

where t = t t0 is the time step. Substituting m into Eq. (21), then

(A1 n B1 B2 ) +

(A1 n B1 B2 ) 4(A2

QP =

0
t
m
n = n0
(m0 +
t)
Ta
2

2(A2

(24)

(A1 n B1 B2 ) 4(A2

Hr
 2 Qr2

Dene m at t0 t as m and expend m00 at t = t0 , then we have

.

0 t +
m00 = m0 m

t
(1.5m0 0.5m00 )
Ta

(25)

(26)

With the four quadrant characteristic curve of pumps, the head

balance Eq. (18) for pumps and valves are as follows
Hr |q|q
F = Hr (q + n )(WH(x)) + CP1 CM2 (B1 + B2 )Qr q
2
2

=2

(27)

The following method is used to solve q.

F + Fq q = 0

Fq =

(28)

F
q
= CP1 CM2 + Hr {2q[A0 + A1 ( + tan1 )] + A1 n}
n
q
2Hr

|q|
2

)(n He + CP1 CM2 ) 0

(33)

   
q = q + 0.1q/ q q > 0.1
 
q = q + q q 0.1

4.2. Container model

The upper plenum and lower plenum of the core are treated as
containers in WAHAP as shown in Fig. 4. Neglecting the friction and
the inertia, we give effective elastic modulus K as follows.
K=

p
,
Ve/Ve

Ve =

QP + Q0
t
2

(34)

where Ve denotes the container volume, Ve refers to the volume

of uid entering the container within the time t. QP and Q0 are
volume ow rates at the beginning and the end of a time step.
Substituting the above equations into the following equations

K  t

HP = H0 +
(QP + Q0 )

2
gVe

HP1 = CP1 B1 QP1

(35)

HP2 = CM2 + B2 QP2

HP1 = HP2 = HP

QP = Q P1 QP2

(29)

where A0 and A1 are interpolation factors of WH(x). q can be solved

with the following iteration scheme.

)(n2 He + CP1 CM2 )

m0
t 2 ...
2!

Substituting this equation into Eq. (24) and neglecting high

order innitesimal, we obtain
n = n0

Hr
 2 Qr2

Hr
 2 Qr2

(30)

QP can be solved as follows.

QP =

B2 CP1 +B1 CM2

B1 +B2
B1 B2
B1 +B2

H0
+

K  t
2
gVe

K  tQ0
2
gVe

(36)

where HP and H0 are the head at the beginning and the end of a
time step respectively.

170

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

Fig. 5. The schematic diagram of experimental setup.

Table 2
The length of the pipes in the experiment.

Fig. 7. The experimental and simulated outlet pressure of valve 1.

Pipe numbers

Pipe length (mm)

3500

4623

3500

3995

8307

12,540

5. Results and discussions

5.1. Model verication
A simple experiment is performed to verify the methodology
and the models used in WAHAP. The schematic diagram of experimental apparatus is shown in Fig. 5. 16 are the number of pipe.
The length of these pipes is shown in Table 2. The friction factor
is 0.01094. The pump parameters are the same with that mentioned in pump models The apparatus fundamentally consist of
a container, two pumps, two valves, some pipes and associated
measuring instruments. The pressurizer with the high-pressure
nitrogen is used to keep the system pressure at 3.0 MPa.
In the transient, pump 1 starts rst then shuts down at 0 s when
it works stably and pump 2 starts accordingly. The inlet and outlet pressure of valve 1 are measured and calculated in the switch
process. Fig. 6 shows the measured and simulated inlet pressure of
valve 1. Fig. 7 shows the measured and simulated outlet pressure
of valve 1. It can be seen from these gures that three obvious pressure oscillations happen and their frequency is approximately the
same. The simulated pressure oscillation amplitude is slightly different from the measured amplitude. The pump switch inuence

Fig. 6. The experimental and simulated inlet pressure of valve 1.

on the pressure appears about 0.5 s later in the experiment than

in the simulation. That time delay may be caused by the friction
models used in the work. On the other hand, time error is easy to
happen in pumps switch process in the experiment. There is a high
level of agreement between the simulated results and the experimental data, which can prove the correction of models applied in
the WAHAP.
5.2. WAHAP results and discussions
The schematic diagram of the simulated system is shown in
Fig. 1. At the beginning of the transient, four valve discs stay balanced and the ow rates of all branches are zero. Pump 1 started at
low speed at 0 s and then failed at 10 s. Pumps 24 keep closed for
10 s and then start with high speed.
The volume ow rates and valve opening in the transient process
are presented in Figs. 8 and 9, respectively. The ow rate of branch
1 rises with the startup of pump 1. The reverse ow occurs immediately in branches 24, and the maximum ow rate is 35 m3 /h. The
closures of valves 24, cause the pressure drop oscillation through
these valves, which lasts until 10 s (Fig. 11). When pump 1 loses
its power and pump 2, pump 3 and pump 4 start, the pressure
drops in valves 24, oscillate immediately. The oscillations disappear quickly as shown in Fig. 12. The volume ow rate of branch 1

Fig. 8. Volume ow rates vs. time in each branch.

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

Fig. 9. Valve opening vs. time in each branch.

Fig. 10. Pressure drop in valve 1 vs. time.

decreases rapidly from 644 m3 /h. The reverse ow occurs at 10.7 s,

which makes the pressure drop in the valve 1 shake and rise from
22.5 kPa to 67.5 kPa (Fig. 10).
About ten combinations of the four-pump-alternate-startup in
the two star-up modes are investigated by WAHAP, most of them
are not shown in the present paper. From all the results we can
nd that the pressure drop oscillates for three times in the whole

171

Fig. 12. Variation of ow rate.

transient. The rst one happened at about 0.6 s during the startup
of pump 1, It is named as pump startup induced water hammer.
The second one was attributed to that pump 1 lose its power at 10 s.
Although pumps 24 start at the same time, the process is named
as pump shutdown induced water hammer rather than pump
startup induced water hammer. Because the pump startup process
lasts about 0.7 s in high speed startup mode, while the pressure
drops in valves 24 vibrate instantaneously after pump 1 shut off
as displayed in Fig. 12. The third one happened at about 11 s when
valve 1 was fully closed. The pressure drop in valve 1 lasts until the
simulation ends. It is named as valve close induced water hammer.
It can also be found that the valve close induced water hammer and pump startup induced water hammer last for a long time,
while the pump shutdown induced water hammer disappeared
quickly. On the other hand, both the pump startup and shutdown
induced water hammer happen in the other three branches rather
than the branch that the acting pump located on, while the valve
close induced water hammer happens in the same branch that the
shut-off valve located on.
It can be found in Figs. 10 and 11 that pressure drops are small
compared to system working pressure 15.5 MPa, and there is little
distinction in other combinations of start-up modes. The primary
loop temperature of this reactor varies from 260 C to 300 C, so
it is unlikely that the IPWR primary loop would suffer from twophase water hammer and it would be safe under pump switching
condition.
5.3. Comparisons with owmaster

Fig. 11. Pressure drop in valves 24.

Symmetrical startup simulation is performed with WAHAP, It

is characterized by the startup of pump 1 and pump 2 in the low
speed mode at 0 s and then the startup of pump 3 and pump 4 in
the high speed mode at 10 s.
Fig. 12 shows the variation of volume ow rates in each branch.
When pump 1 and pump 2 start, the uid ows backwards immediately in branch 3 and branch 4 and valve 3 and valve 4 close
simultaneously (Fig. 13), which induces pressure drop uctuations
in valve 3 and valve 4. The opening of valves 1 and 2 changes from
natural state to fully opened state and the ow rates of branch 1
and branch 2 increase rapidly to the designed ow rate. When the
pumps switch at 10 s, the volume ow rate in branch 1 and branch
2 decline and then reverse ow occur which lead to pressure drop
oscillation in valve 1 and valve 2. Valve 3 and valve 4 open gradually
within one second and valve 1 and valve 2 close subsequently.
The results obtained by WAHAP are consistent with those calculated by Flowmaster except for some differences of pressure drop

172

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

in WAHAP. In Flowmaster the pressure source component was

located before pump entrance and the pressure is kept constant
which may ameliorate the water hammer impact on the system.
The differences would not jeopardize the comparison results
with Flowmaster because it is relative pressure that is compared.
The consistency of vibration amplitude and the attenuation trend
can improve that the models and algorithms in WAHAP are correct.
6. Conclusion

Fig. 13. Variation of valve opening.

Water hammer behavior in an IPWR primary coolant system has

been predicted in detail with both newly developed code and commercial software Flowmaster. The new code WAHAP is developed
and its equations and models have been tested. The good consistency of the experimental data and simulation results validates the
models and numerical method used in WAHAP. The four-pumpalternate startup processes are thoroughly investigated by WAHAP.
Three kinds of pressure oscillations were found in the transient.
They are pump startup induced water hammer, pump shutdown
induced water hammer and valve close induced water hammer,
respectively. The results suggest that the primary loop of the IPWR
is safe within pump switching processes. The famous water hammer software Flowmaster is also introduced to test the ability of
WAHAP. The consistency of the results can improve that WAHAP
can be used to simulate water hammer.
Future research will focus on heat transfer module in water
hammer simulation. This work offers a valuable reference for the
prediction and mitigation of water hammer in the integer reactors.
Acknowledgements
This research was carried out under the nancial support of
Science and Technology on Reactor System Design Technology
Laboratory of Nuclear Power Institute of China (KZZJJ-A-201101)
and National Magnetic Connement Fusion Science Program
(2010GB111007).

Fig. 14. Variation of pressure drop in valve 1 and valve 2.

References
oscillations. The reverse ow in branch 1 calculated by Flowmaster
happens later than that calculated by WAHAP, which induce different pressure drop oscillation in valve 1 and valve 2 as shown in
Figs. 14 and 15. The pressure models may cause such considerable
differences. The pressure models in WAHAP and Flowmaster are
different. In Flowmaster, pressure source component is adopted to
achieve pressure convergence and no such component is involved

Fig. 15. Variation of pressure in valve 3 and valve 4.

Afshar, M.H., Rohani, M., 2008. Water hammer simulation by implicit method of
characteristic. Int. J. Pres. Ves. Pip. 85, 851859.
Allievi, L., 1925. Theory of Water Hammer. Typography R. Garroni, Rome.
Barna, I.F., Imre, A.R., Baranyai, G., zsl, G., 2010. Experimental and theoretical
study of steam condensation induced water hammer phenomena. Nucl. Eng.
Des. 240, 146150.
Barten, W., Manera, A., Macianjuan, R., 2008. One- and two-dimensional
standing pressure waves and one-dimensional travelling pulses using
the US-NRC nuclear systems analysis code TRACE. Nucl. Eng. Des. 238,
25682582.
Beuthe, T.G., 1997. Review of two-phase water hammer, Canadian Nuclear Society, Technical Program Proceedings, Powering Canadas Future, 1997 CNA/CNS
Annual Conference.
Ghidaoui, M.S., Ming, Z., McInnis, D.A., Axworthy, D.H., 2005. A review of water
hammer theory and practice. Appl. Mech. Rev. 58.
Kochupillail, J., Ganesan, N., Padmanabhan, C., 2005. A new nite element formulation based on the velocity of ow for water hammer problems. Int. J. Pres. Ves.
Pip. 82, 114.
Lee, S.-K., Kim, N.-S., Shin, B.-S., Keum, O.-H., 2011. Steady-state analyses of uid ow
characteristics for AFWS in PWR using simplied CFD methods. World Acad. Sci.
Eng. Technol..
Liu, S., 1987. Water Hammer in USA Power Plants and Its Evaluation. Nuclear Power
Engneering 8.
Liu, Z.Y., G.L., L., Su, F.J., 2005. Numerical Simulation of Pump-Stopping Water Hammer in Tertiary Circulating Water System of PWR. Nuclear Power Engneering
26.
Marcinkiewicz, J., Adamowski, A., Lewandowski, M., 2008. Experimental evaluation
of ability of Relap5, Drako (R), Flowmaster2 (TM) and program using unsteady
wall friction model to calculate water hammer loadings on pipelines. Nucl. Eng.
Des. 238, 20842093.
Prica, S.S., Stevanovic, V.D., Maslovaric, B.M., 2008. Numerical simulation of condensation induced water hammer. FME Trans. 36, 2126.
Sattar, A.M.A., Dickerson, J.R., Chaudhry, M.H., 2009.
Wavelet-Galerkin
solution to the water hammer equations. J. Hydraul. Eng.-ASCE 135,
283295.

Q. Zuo et al. / Nuclear Engineering and Design 261 (2013) 165173

Tian, W.X., Su, G.H., Wang, G.P., Qiu, S.Z., Xiao, Z.J., 2008. Numerical simulation and
optimization on valve-induced water hammer characteristics for parallel pump
feedwater system. Ann. Nucl. Energ. 35, 22802287.
Tian, W.X., Su, G.H., Wang, G.P., Qiu, S.Z., Xiao, Z.J., 2009. Mitigating check valve
slamming and subsequentwater hammer events for PPFS using MOC. Nucl. Sci.
Tech. 20, 118123.
Todreas, N.E., 2001. Nuclear Systems II Elements of Thermal Hydraulic Design. Taylor
and Francis, New York.
Wang, X.F., 1995. Water Hammer in Industry Pipes. Science Book Concern, Beijing.
Wylie, E.B., Streeter, V.L., 1983. Fluid Mechanics, 6th ed. McGraw-Hill Book Co, New
York.

173

Wylie, E.B., Streeter, V.L., 1993. Fluid Transients in Systems. Prentice Hall, New York.
Yang, K.L., 2000. Hydraulic Transient and Regulation in Power Station and Pumping
Station. China Water Conservancy and Electricity Press, Beijing.
Zhang, S.B., Luo, B.X., Qiao, X.B., Chen, J., Wan, X.D., 2012. Feedwater system transient analysis of nuclear power station. Adv. Mater. Res. Renew. Sust. Energ.,
15571560.
Zhang, X.H., Zhu, J., 2008. Water hammer damage analysis caused by the condensate
extraction pump in PWR. Water Pump Technol. 4, 2732.
Zhao, M., Ghidaoui, M.S., 2004. Godunov-type solutions for water hammer ows. J.
Hydraul. Eng.-ASCE 130, 341348.