The Parametric Modeling of

Local Resistance and Pressure

Dan Wang
State Key Laboratory for Strength
Drop in a Rotary Ball Valve
and Vibration of Mechanical Structures,
School of Aerospace, In this paper, a theoretical study of ball valves is carried out for investigating the local
Xi’an Jiaotong University, resistance and pressure drop of ball valves in operating process. An equivalent model of
Xi’an 710049, China ball valves is proposed based on the inherent mechanism of the resistance loss, which
e-mail: [email protected] can be divided into three equivalent throttling components: a thick orifice, two variable-
opening eccentric orifice plates, and a Z type elbow. Through analysis of the flow resist-
Changqing Bai1 ance of the three components, a general parametric modeling of ball valves is presented
for the flow resistance analysis, and then an analytical formula of pressure drop is dem-
State Key Laboratory for Strength
and Vibration of Mechanical Structures,
onstrated. The results obtained from the presented model are compared with the prior
School of Aerospace,
test data to validate this model, and good agreement is observed. Indicate that the pre-
Xi’an Jiaotong University,
sented model has high accuracy in predicting the resistance and pressure loss in various
Xi’an 710049, China
openings. The results show that the influences of thin orifice plates play an important role
e-mail: [email protected]
in the total flow resistance coefficient and pressure drop, especially in the small opening.
The effects of thick orifice plates and the Z type elbow gradually increased as the valve
opening rises and becomes significant when the opening is more than 70%.
[DOI: 10.1115/1.4037946]

Keywords: ball valve, equivalent model, flow resistance, pressure drop

1 Introduction flows passing through a globe valve by numerical simulations and

experiment. Lorenzini and Biserni [3] setup a hydraulic circuit to
Ball valve, as a necessary component in the pipeline system, is
investigate a possible use of the vapotron effect for the cooling of
commonly used to regulate flow rate and flow direction or control
electronic devices, and the ball valves were used to regulate flow
the stopping and starting of flow in pipes. It has been widely
rate in this novel study. Numerical simulations do provide a
applied in the petroleum, chemical, aerospace, medicine, and
detailed and comprehensive analysis for the valve and they have
other engineering fields owing to its highlighted advantages such
been proved to be an effective method to capture valve parame-
as large flow capacity and operating range [1–3]. One of the key
ters. However, the calculated time of numerical simulations is too
parameters required for the valve design is the pressure drop,
long and it even cannot be completed sometimes, especially in
which is not only the important constituent part of the total pres-
complicated pipelines with nonlinear factors. Furthermore, the
sure loss of a pipe system but also the cornerstone for calculation
results of numerical simulation are significantly affected by
of other fluid properties such as the flow resistance. Furthermore,
the modeling and grids technique. And experiments can prove the
the valve pressure drop has a direct influence on the operating
credibility and reliability of results, but it is also time-consuming
condition and performance of a pipe system. For example, during
and only applies to one condition. Both of them are still in a
the valve operating process, the pressure and velocity of fluid in
semitheoretical and semi-experimental stage, lacking in a unified
valve vary with the changing opening, which will lead to a series
theoretical basis. With the wide application of ball valves, the
of problems such as the severe structural vibration and noise and
methods relied on simulations and experiments can hardly satisfy
seriously affects the working stability of the valve. Therefore, the
the demand for the new pipeline design and fault analysis.
local resistance and the pressure drop of a ball valve have received
In order to obtain the accurate results and reduce the calculated
widespread attention in the last three decades.
time, some data-driven, semiphysical, physical models have been
The flow in valve opening process is a pressure-drive variable-
developed. The data-driven models are based on the input and out-
section unsteady flow. The numerical simulation and experiments
put responses of control valves. Black-box and gray-box models
are the most convenient approaches to analyze, predict, and
as the most generic data-driven methodology have received con-
control the flow characteristics of valve. Many studies about
siderable attention and performed well in the prediction of per-
valves were based on the computational fluid dynamics simula-
formance of hydraulic systems [12–14]. They mainly pointed at
tions and experimental tests. Moujaes and Jagan [4] and Chern
the closed-loop hydraulic systems composed of pump, valves,
and Wang [5] investigated the flow patterns and performance
actuators, and sensors; few efforts have been made to explore the
coefficients in ball valve at different valve openings based on
characteristics of a single valve, especially for the local resistance
three-dimensional numerical simulations and experiment. Lee and
and the pressure drop of valves; furthermore, the results need to
Lee [6] also presented an approximate numerical approach to pre-
be experimentally verified. Semiphysical models [15,16] and
dict the flow resistance coefficient of the ball valve. Van Lookeren
physical models [17,18] are based on the Newton’s second law of
Campagne et al. [7] simulated flows containing bubbles in ball
motion; it mainly studies the mechanical characteristics of the
valves. In addition, Chern et al. [8] and Tabrizi et al. [9] investi-
valve and the fluid characteristics analysis is absent. In addition,
gated the valve performance in the presence of cavitation. Chern
Valdes et al. [19] develop a methodology for the parametric mod-
also designed a cage to control the cavitation inside valve. Davis
eling of the flow rate in hydraulic valve systems; the model still
and Stewart [10,11] employed a closed-loop system to investigate
has some variables that need to be provided by experiments.
Although some contributions of valve models have been done,
1
Corresponding author. models of local resistance and pressure drop for ball valves that
Contributed by the Fluids Engineering Division of ASME for publication in the
JOURNAL OF FLUIDS ENGINEERING. Manuscript received April 12, 2017; final
can be applied directly are absent. Therefore, it is necessary to
manuscript received September 2, 2017; published online October 19, 2017. Assoc. analyze the local resistance and pressure drop of ball valves based
Editor: Daniel Livescu. on its inherent mechanism.

Journal of Fluids Engineering Copyright V

C 2018 by ASME MARCH 2018, Vol. 140 / 031204-1

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 In the fluid-transfer process of pipeline network, the ball valve different views on the flow resistance of elbow bends. Idel’chik
achieves its throttling effect through the variable area of the valve et al. [25] found that resistance coefficient of elbow depends not
spool, and it is commonly partially open to regulate the desired only upon flow Reynolds number but also upon geometric param-
flow rate, pressure, and the flow direction. During the valve rota- eters such as bend angle, curvature radius, and ratio of inlet to out-
tional process, the flow area of a ball valve is less than the cross let. Schlichting et al. [36] hold that the resistance loss of elbows
section of pipeline. So, when the fluid in pipe flows into the valve, should include the drag of elbow itself and the drag caused by the
it will undergo a sudden contraction and expansion, and the direc- varying section in the outlet of the elbow.
tion of the fluid in pipe changes at the same time. In a broad sense, In this paper, in order to study the local resistance and pressure
a ball valve can be regarded as a throttling element, which not drop of ball valves, the equivalent model of ball valve is presented
only has the ability to achieve sudden contraction expansions like first. Then a novel mathematical model about the flow resistance
orifice plates but also has the ability to change the flow direction and pressure drop of the ball valve in the operating process is
like elbows. Therefore, ball valves can be regarded as a certain established based on its inherent mechanism. Finally, the reliabil-
combination of orifice plates and elbows. ity of the presented models is verified by comparing with prior
In general, orifice plates can be classified according to their experimental results.
thickness ratio t/d, where t is the orifice thickness and d its diame-
ter. The thickness ratio directly affects the flow state in orifice 2 Parametric Modeling of Ball Valves
plates, which has a direct influence on the flow resistance
and pressure drop. Chisholm [20] defines the orifices as thin if In order to introduce a parametric modeling process for the
t/d < 0.5 and as thick if t/d > 0.5. In recent years, orifice plates local resistance and pressure drop of ball valves in the operating
have been investigated quite extensively due to their practical process, the projected area of orifice in a ball valve should be ana-
relevance. Roul and Dash [21] and Wang et al. [22] investigated lyzed first, and then the equivalent model of ball valves could be
numerically the pressure changes caused by the abrupt expansion proposed to establish the mathematical model of the flow resist-
and contraction of flow area in mini circular orifice channels for ance and pressure drop of ball valves in the operating process
two-phase flow. Chen et al. [23] investigated the pressure drops in based on its inherent mechanism theoretical analysis.
the abrupt contraction and the sudden expansion, respectively.
Liou and Lin [24] investigated the effect of Reynolds numbers on 2.1 Projected Orifice Area Analysis. The throttling area is a
the flow patterns in rectangular channel using lattice Boltzmann key parameter for a valve, which not only determines the flow
method for single phase. In addition, there have been many theo- rate but also influences the flow resistance and pressure drop. Fur-
retical studies to evaluate the pressure drop of orifice plates, which thermore, the dynamic characteristics of ball valves are directly
is caused by sudden contraction expansions. Idel’chik et al. [25] related with the orifice area. Therefore, the determination of flow
and Grattan and coworkers [26] studied the resistance characteris- area is an essential prerequisite for estimating the other perform-
tics of orifice plate based on the theory and experiment, and put ance parameters of ball valves. Thus, a general mathematical
forward some useful theories on the resistance of orifices. Kojasoy equation on the orifice area of ball valves should be established.
et al. [27] proposed a mathematical model of pressure drop for the As plotted in Fig. 1, a Cartesian coordinate system, featuring x-,
two-phase flow through the thin and thick orifice plates. Fossa y-, and z-axes, is adopted here. The origin is fixed at the center of
et al. [28,29] also studied the pressure drop of single-phase and the ball, O. The x-axis is defined parallel to the pipe flow direction
two-phase flow by the experiments. Their works provide valuable and y-axis, perpendicular to the pipe flow direction. The z-axis
references. coincides with the stem of ball valve, which connects the external
In addition, during the rotational process of ball valves, the handle to the valve spool.
pressure loss due to the abrupt change of flow direction could not Assume that the fluid in the pipe flows from the left to the right.
be ignored. Elbow bends are used to accommodate the pipe in the The rotation process of the ball valve is schematically indicated in
direction of pipeline alignment, so the abrupt change of flow Fig. 1, where the sphere radius of ball valve is R, the radius of
direction can be matched with elbow bends in the meaning of the valve orifice is r. When the valve is fully open, the cross section
resistance and pressure drop. A number of researchers have inves- of circular valve flow channel projected in the x–y plane is a circle
tigated the hydraulic characteristics of pipe bends by means of of radius r, whose center locates in O1 . And when the valve spool
theoretical, experimental, and numerical methods [30–35]. The rotates an angle of h deg, the cross section of circular valve flow
pressure loss in elbow bends is due to a combination of frictional channel projected in the x–y plane is an ellipse with a long axis of
effects, secondary flows, and separation, and researchers have r and short axis of r cos h, and the center locates in O2 . So, when

Fig. 1 Orifice area of ball valves: (a) longitudinal section of a ball valve and its coordinate
system and (b) projection of orifice area on x–y plane

031204-2 / Vol. 140, MARCH 2018 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 the valve spool rotates at h deg, the flow area of valve orifice pro- Generally, the openings of ball valve can be defined with three
jected in the x–y plane is the intersection of the circle and the different approaches. Two of them aim to measure the valve open-
ellipse, as the shadow shown in Fig. 1, in which M and P is the ings in terms of the rotation time or the flow area of valve spool.
points of intersection of the circle and the ellipse, F is the point of While the last one is widely applied in practical engineering due
intersection of the circle and the vertical line passed through O1 , to its simplicity and practicality, and it will be adopted in Sec. 3,
E is the point of intersection of the ellipse and the vertical line which defines the valve opening by rotation angles as
passed through O2 , N is the point of intersection of MP and EF, L
states the distance between the center of the circle and the ellipse, K ¼ ð1  h deg=90 degÞ  100 (3)
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
L ¼ O1 O2 ¼ R2  r 2  sin h. The flow area of valve orifice pro-
jected in the x–y plane at h deg, defined as SFlow , can be combined As shown in Fig. 2, looking from the direction of the flow, the
with S_ and S_ as orifice area increases with the increasing opening of ball valves in
MEP MFP the opening process, where the shaded area indicates the flow area
and the unshaded area is the valve spool. The shape of the
SFlow ¼ S_ þ S_ (1) unshaded area is crescent-like. Its area is gradually decreased as K
MEP MFP
reduces and becomes zero when the valve is fully open.
According to the derivation, as shown in Appendix A in detail,
the mathematical equation used to obtain the orifice area of ball
valve can be expressed as 2.2 Equivalent Model of Ball Valves. When the spool of a
common full bore ball valve rotates a certain angle, the top view
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !
of the valve flow channel is shown in Fig. 3, in which D is the
2 R2  r 2  sin h
SFlow ¼ r arccos diameter of outer pipe, d is the diameter of flow channel in the
r ðcos h þ 1Þ ball valve, L1 and L3 are the length of upstream and downstream,
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi respectively, L2 is the length of ball valve, and L4 is the length of
ðR2  r 2 Þ  sin2 h pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 r2   R2  r 2  sin h spool. The pressure and the flow rate of inlet and outlet are usually
ðcos h þ 1Þ2 tested at the distance of five times and ten times of pipe diameter
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi  before and after ball valves in the engineering that signifies
2 R2  r 2  sin h  cos h L1 ¼ 5D and L3 ¼ 10D.
þ r  cos h  arccos
cos h þ 1 According to the pivotal geometric construction of ball valves
v pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi !2 1 related to the flow resistance and pressure drop, the ball valve can
u
u 2 R2  r 2  sin2 h A be regarded as a combination of several throttling elements. As
tr  (2)
cos h þ 1 plotted in Fig. 4, the equivalent model of the ball valve can be
extracted for investigating the flow resistance and pressure

Fig. 2 Variety of the orifice projected area with increasing opening (the shaded area indicates the flow area,
and the unshaded area is the valve spool)

Fig. 3 Top view of valve flow channel

Journal of Fluids Engineering MARCH 2018, Vol. 140 / 031204-3

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 4 Equivalent model of a ball valve

Fig. 5 Three types of throttling elements

drop characteristics. In general, the diameter of outer pipe is equivalent to two variable-opening eccentric thin orifice plates,
larger than that of valve flow channel (D > d); thus, a sudden whose shape is crescent-like. In addition, during the valve rotation
contraction–expansion occurs when the fluid in pipe flows into process, the pressure loss due to the abrupt change of flow direc-
and out the valve channel. While during the spool rotation pro- tion cannot be ignored. Elbow bends are used to accommodate the
cess, the valve port could be equivalent to two variable-opening pipe in the direction of pipeline alignment. So, the abrupt change
eccentric thin orifice plates, whose shape is crescent-like. Note of flow direction can be equivalent to elbow bends. Thus, the last
that the area of the crescent-shaped orifice plate changes with the component is Z type elbow (see Fig. 5(c)).
valve opening, K, as shown in Fig. 2.
According to the equivalent model of ball valves shown in
Fig. 4, it can be found that the resistance loss and pressure drop of 2.3 The Mathematical Model of Local Resistance. Denote
valves occur in several pipe fittings, including pipe entrance and the resistance coefficient caused by the first part as n1 , the second
exit, sudden expansion and sudden contraction, and turning flow, part as n2 , and the last part as n3 . The total local resistance of ball
etc. Therefore, a ball valve can be disassembled into several inter- valves, l1, can be achieved by analyzing the loss coefficients cor-
connected subsystems. Based on the inherent mechanism of the responding to the three different components. It can be written as
resistance loss, a ball valve can be divided into three throttling
components to study flow resistance and pressure drop character- l1 ¼ n1 þ n2 þ n3 (4)
istics, as shown in Fig. 5.
The first throttling component is a thick orifice plate (see
Fig. 5(a)). Since the flow in pipe undergoes a sudden contraction 2.3.1 Thin and Thick Orifice Plates. As Fig. 5 shows, the sud-
and expansion before and after the ball valve, the energy loss den contraction and expansion before and after the ball valve is
caused from this is matched with that caused by a thick orifice matched with a thick orifice plate in the meaning of the resistance
plate; the second components are thin orifice plates, as shown in and pressure drop. And the inlet and outlet of the valve spool can
Fig. 5(b). During the spool rotation process, ball valve is partially be equivalent to two variable-opening thin orifice plates. Ai et al.
open, so the flow goes through an instantaneous sudden [37] investigated the local resistance of circular orifice plates and
contraction–expansion when the fluid flows into and out the flow crescent-like orifice plates by experiment, and the results indi-
channel of valve spool. Therefore, the valve port could be cated that the resistance coefficients of both are almost the same.

031204-4 / Vol. 140, MARCH 2018 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 6 Orifice plates [29]: (a) thin orifice plate and (b) thick orifice plate

among rc and some geometrical parameters is available. Such

relationships usually obtained by experiment and only require the
flow area ratio, r. The relationship adopted in this paper follows
the well-known Chisholm expression
1
rc ¼ 0:5
(7)
0:639ð1  rÞ þ 1
Therefore, the resistance coefficient of the thick orifice plate
caused by the sudden contraction–expansion can be expressed as:

Fig. 7 Turning flow n1 ¼ nthick þ n01 (8)

And the resistance coefficient of two variable-opening thin orifice
In this work, circular orifice plates being adopted to calculate the plates can be expressed as
flow resistance in place of crescent-like orifice plates.
When the flow occurs through an orifice plate in a pipe, the n2 ¼ 2nthin þ n02 (9)
flow contracts with negligible loss of mechanical energy, to an
abrupt vena contracta of area Ac before expanding again to the where n01
and n02
are the corrected coefficients, which vary with
pipe wall. At this location, the jet flow area reaches its minimum valve openings.
value so that the flow velocity attains its maximum [38]. As such,
the fluid dynamics of orifice flows is controlled by the contraction 2.3.2 Abrupt Change of Flow Direction. As indicated in
and subsequent expansion of the jet [38]. For the flow through a Fig. 7, during the valve spool rotational process, the direction of
thin orifice, as schematically indicated in Fig. 6(a), the jet con- fluid in pipeline changes when the fluid flows into and out of the
tracts throughout the orifice and starts expanding downstream; the valve spool. This leads to a remarkable resistance, n3 , that should
vena contracta is located downstream of the orifice. While for the also be responsible to the total resistance, l1 .
flow through a thick orifice, as schematically indicated in The energy loss caused by the abrupt change of flow direction
Fig. 6(b), the jet expands within the orifice, the flow reattaches can be equivalent to that caused by a Z type elbow, as schemati-
following the vena contracta and can even develop a boundary cally indicated in Fig. 8(a). Since Z type elbow can be regarded as
layer flow before leaving the orifice. Therefore, the flow state in a combination composed of two elbows as Fig. 8(b) shows, two
thin and thick orifice plates is different, which has a direct influ- elbows in place of a Z type elbow are used to estimate the resist-
ence on the resistance loss and pressure drop. ance loss of turning flow in ball valves.
According to the research of Chisholm [20], the resistance loss The well-known formula proposed by Fried and Idelchik [39]
coefficient through a thin orifice plate can be defined as a function can be employed to calculate the resistance coefficient of elbows
of the flow area ratio r ¼ SFlow shown in Fig. 8(b), which can be denoted as nelbow
S and the contraction coefficient rc as
 2 nelbow ¼ a  b  nM (10)
1
nthin ¼ 1 (5)
rrc where nM is the resistance coefficient of bending elbow (see
Fig. 8(c)) that can be defined as
If the flow occurs through a thick orifice plate, the resistance loss
coefficient can be expressed as h h
nM ¼ 0:95 sin2 þ 2:05 sin4 (11)
  2 2
 2  
1 2 1 1 Fried and Idelchik [39] used the bending elbow in place of the
nthick ¼ 1 1 2 1 (6)
rrc r2 rc r elbow to calculate the pressure loss caused by the change of fluid
direction, and then corrected it using corrected coefficients of a
According to Eqs. (5) and (6), the resistance of the thin and the and b. a is the function of turning angle h, that is a ¼ f ðhÞ; b
thick orifice plates can be predicted, provided that a relationship is the function of r1 =d1 , that is b ¼ f ðr1 =d1 Þ, where r1

Fig. 8 Elbows: (a) Z type elbow, (b) elbow, and (c) bending elbow

Journal of Fluids Engineering MARCH 2018, Vol. 140 / 031204-5

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 approximately takes as L4 sin h, d1 approximately takes as the Therefore, the total resistance coefficient of the ball valve sys-
length of EF, as shown in Fig. 1 tem, l, can be expressed as
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
d1 ¼ r þ r cos h  sin h  R2  r 2 (12) l ¼ l1 þ l2 (19)

The relationships of a  h and b  r1 =d1 are shown in

Appendix B (see Tables 6 and 7). Fitting the data of Tables 6 2.4 Pressure Drop. Pressure drop, closely related to the
and 7, the relationships of a and b can be expressed as resistance loss, l, is an important indicator to evaluate the valve
performance. It can be characterized by flow resistance coeffi-
a ¼ f ðhÞ cient; Kirik and Driskell [41] provided formula as
   2
h h
¼ 2:953  1:618  2:089 Dp ¼ lqV 2 =2 ¼ ðl1 þ l2 ÞqV 2 =2 (20)
90 deg 90 deg
 3
h V ¼ Q=ð3600SÞ (21)
þ 1:972 ð0 deg < h < 90 degÞ (13)
90 deg
where Dp is the pressure drop of valve (MPa); q is the density
b ¼ f ðr1 =d1 Þ of fluid (g/cm3); V is the mean velocity (m/s); Q is the flow rate
   2 (m3/h); and S is the cross-sectional area of outer pipe (m2).
r1 r1
¼ 1:139  0:1529  0:01918
d1 d1
 3   3 Validation and Discussion of the Proposed Model
r1 r1
 0:0008672 0:025 < < 8:0 (14) In the recent work of Cui et al. [42] and Shi et al. [43], the
d1 d1 dynamic characteristics of ball valves with the nominal diameter
of 50 mm and 75 mm during the operating process have been
Considering that the resistance coefficient of a Z type elbow in investigated by experiments, and the relationships of resistance
ball valve is not exactly equal to the sum of that of two elbows, n3 loss and pressure drop varying with the valve openings were ana-
can be amended as lyzed by experimental data and numerical simulations. The pro-
posed model of flow resistance and pressure drop of ball valves
n3 ¼ 2nelbow þ n03 (15) can be validated by comparing with these experimental data. First,
the orifice areas in different openings are calculated according
Eq. (2) and the relationship between flow area and valve opening
where n03 is the resistance influence coefficient of two elbows,
is plotted. Then the mathematical model of resistance coefficient
which varies with the valve opening.
and pressure drop described in Secs. 2.3 and 2.4 will be verified
n01 , n02 , and n03 can be summed up as one correct coefficient n0 ,
by comparing with the experimental data.
which is also the function of valve openings, namely, n0 ¼ f ðKÞ.
Therefore, the total local resistance coefficient of ball valve can
be described as 3.1 Calculation of Orifice Area. The detailed size of ball
valves in Fig. 3 is shown in Table 1. According to Eq. (2), the
flow areas in different openings can be obtained. And the orifice
l1 ¼ n1 þ n2 þ n3 ¼ nthick þ 2nthin þ 2nelbow þ n0 (16)
area versus valve opening, K, and rotation angle, h, are plotted in
Figs. 9 and 10. The calculated results indicate that the orifice area
In which nthick , nthin , and nelbow can be determined using Eqs. (8), of ball valves is gradually increased as the valve opening
(9), and (15). increases and the rotation angle decreases. In the same valve
opening, the greater the valve diameter, the greater the orifice
2.3.3 Frictional Loss. In addition to the local resistance loss
area. And the relationships between orifice area versus rotation
of ball valve, in fact, the frictional drag of pipeline accounts for a
angle and valve opening are nonlinear. Due to the nonlinear rela-
small part of the share; denote it as l2 . The frictional drag mainly
tionship between the local resistance and the orifice area, the rela-
includes the friction of the pipe wall to the water flow and the fric-
tionship between the local resistance and valve opening would be
tion among flow layers, and it is determined by the wall roughness
strongly nonlinear even if the orifice area versus valve opening is
and Reynolds number. According to Darcy–Weisbach formula
weakly nonlinear.
[40]

L L1 þ L2 þ L3 3.2 The Validation and Discussion of the Resistance

l2 ¼ k ¼k (17) Model. According to the detailed size of the ball valves shown in
D D
" pffiffiffi # Table 1, the resistance coefficients of thin and thick orifice plates
1 Re k 3:7D and turning flow, nthin , nthick , and nbend , are calculated, respec-
pffiffiffi ¼ 2 log þ (18) tively, according to Eqs. (5), (6), and (10). Thus, the total local
k 2:51 K1 resistance coefficient of ball valve, l1 , is achieved by Eq. (16); it
should be noted that the corrected coefficient, n0 , is not considered
where k is the frictional drag coefficient; L is the total length of here. The frictional loss coefficient, l2 , is also calculated accord-
pipeline; D is the pipe diameter; K1 is the pipe wall roughness; Re ing to Eqs. (17) and (18). The total resistance coefficients l of ball
is the Reynolds number, Re ¼ VD=c; V is the mean velocity; and valves with the nominal diameter of 50 mm and 75 mm can be
c is kinematic viscosity of fluid. obtained based on the mathematical resistance model presented in

Table 1 The detailed size of DN50 and DN75 ball valves

Ball valve D (mm) d (mm) R (mm) r (mm) L1 (mm) L2 (mm) L3 (mm) L4 (mm)

DN 50 [42] 52 50 37.5 25 250 200 500 86.6

DN 75 [43] 80 75 68.6 37.5 375 200 750 92

031204-6 / Vol. 140, MARCH 2018 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Eq. (19). All theoretical results calculated earlier are shown in
Tables 2 and 3. And the experimental resistances in Refs. [42] and
[43] are also shown in the corresponding tables.
As Tables 2 and 3 show, the frictional coefficient, l2 , is a small
amount. And the local resistance coefficient, l1 , accounts for the
vast majority proportion of the total resistance coefficient, l in the
range of the relative opening 0–80%. As the valve opening
increases, the proportion of l1 is gradually decreased. The propor-
tion of l2 gradually increases as the valve opening rises and
becomes significant when the opening is more than 80%. Once
the contracted structure of a ball valve is designed, the resistance
coefficient of the thick orifice plate, nthick , would be determined,
which is a constant that does not change with the valve opening
while the resistance coefficients of the thin orifice plates and
turning flow, nthin and nbend , are changing with different valve
openings. Compared with nthin , nbend is a small amount in small
openings. In the same valve opening, the resistance coefficient of
Fig. 9 The relationship of flow area and valve opening ball valves with the nominal diameter of 50 mm and 75 mm is dif-
ferent. This shows that the local resistance coefficient of ball valve
is not only related to the valve openings but also closely related to
the valve diameter. This relationship is mainly embodied in the
orifice area.
When the valve is opened, the flow resistance coefficient is
comparatively large, and gradually becomes small with the
increasing of opening, especially at the opening 70–100%. This is
the reason that the pressure of fluid upstream has been increasing
under the strong retardative effect before the valve is opened.
Once the valve opened, especially in a small opening, high-speed
jet flow has an interaction with the surrounding fluid for a strong
shear, which results in a violent vortex movement in the valve.
With the valve opening increases, the jet effect decreases gradu-
ally, but there are still vortexes exist. When the valve is fully
open, the resistance loss of the valve is still not zero due to the
friction loss and the loss caused by the sudden contraction and
expansion.
The variation of flow resistance coefficient with increasing
opening obtained from theoretical method proposed in this paper
Fig. 10 The relationship of flow area and valve rotational angel and the experiment and simulation in Refs. [42] and [43] are
shown in Figs. 11 and 12. As can be seen from Tables 2 and 3 and

Table 2 The local resistance of DN 50 ball valves in different valve openings

 
 l exp  l 
   100
K nthin nthick nbend l1 l2 l ¼ l1 þ l2 lexperiment [42]  l 
exp

20 165.9113 0.0428 0.8562 333.5777 0.3495 333.9272 290.972 14.6

30 48.4985 0.0428 0.6754 98.3906 0.3495 98.7401 94.7742 4.14
40 18.2069 0.0428 0.5145 37.4856 0. 3495 37.8351 37.7622 0.09
50 7.6316 0.0428 0.3685 16.0429 0. 3495 16.3924 18.8602 13.3
60 3.3168 0.0428 0.2398 7.1560 0. 3495 7.5055 7.8201 4.5
70 1.4055 0.0428 0.1333 3.1205 0. 3495 3.4700 3.7683 9.0
80 0.5326 0.0428 0.0540 1.2160 0. 3495 1.5655 1.7671 13.7
90 0.1445 0.0428 0.0061 0.3440 0. 3495 0.6935 0.7076 7.6
100 0 0.0428 0 0.0428 0. 3495 0.3923 0.4022 12.4

Table 3 The local resistance of DN 75 ball valves in different valve openings

 
 l exp  l 
   100
K nthin nthick nbend l1 l2 l ¼ l1 þ l2 lexperiment [43]  l 
exp

36 142.2118 0.0832 0.5552 285.6172 0.5158 286.1330 246.88 15.9

46 44.5349 0.0832 0.4276 90.0082 0.5158 90.5240 94.07 3.8
57 16.2739 0.0832 0.3082 33.2475 0.5158 33.7633 36.93 8.6
66 6.6645 0.0832 0.2058 13.8238 0.5158 14.3396 16.57 13.5
79 1.8934 0.0832 0.0888 4.0475 0.5158 4.5633 5.52 17.3
91 0.1981 0.0832 0.0036 0.4866 0.5158 1.0024 1.20 16.5
100 0 0.0832 0 0.0832 0.5158 0.599 0.60 0.0

Journal of Fluids Engineering MARCH 2018, Vol. 140 / 031204-7

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Fig. 11 Comparison of resistance coefficients obtained from Fig. 13 Comparison of n1 , n2 , n3 , and l of DN50 ball valve
the presented model and the results in Ref. [42]

Fig. 14 Comparison of n1 , n2 , n3 , and l of DN75 ball valve

Fig. 12 Comparison of resistance coefficients obtained from
the presented model and the results in Ref. [43]
contribution is 0 while the resistance of the thick orifice plate and
the friction loss do not change with the valve openings. Therefore,
Figs. 11 and 12, the theoretical resistance of ball valves calculated the total resistance loss is not zero at 100% opening.
by mathematical flow resistance model proposed in Eqs. (16) and
(17) is in good agreement with the experimental and simulated 3.3 The Validation of Pressure Drop. The average veloc-
resistance measured by Cui and Shi, even without considering the ities V in different valve openings can be obtained according to
influence of the corrected coefficients. And the resistance errors of the flow area calculated in Sec. 3.1 and the known flow rates Q
DN50 and DN75 ball valves shown in Tables 2 and 3 are less than obtained from experiment in Ref. [42] are shown in Table 4.
15% and 18% in the range of the relative opening 20–100%. This According to Eq. (20), the total pressure drops Dp of the DN50
indicates that the mathematical model proposed in this paper is ball valve in different valve openings are obtained, as shown in
applicable and effective, and it can meet the needs of practical Table 4. And the experimental pressure drop in Ref. [42] is also
engineering even if the influence of the corrected coefficients is shown in Table 4. The comparison of theoretical pressure drop
not considered. and experimental pressure drop is shown in Fig. 15.
To investigate the contribution of the resistances corresponding Because the variation of the inlet and outlet pressures of ball
to the three equivalent throttling components to the total resist- valve is small in the range of the relative opening 0–20%, the
ance of ball valves, according to Eqs. (8), (9), and (15), the resist- pressure drop of the ball valve is analyzed in the relative opening
ance coefficients, n1 , n2 and n3 , of three throttling components from 20% to 100% in the opening process. With the opening of
proposed in Sec. 2.3 are obtained; it should be noted that the cor- the ball valve increases, the flow area of the valve increases and
rected coefficients are not considered here. The comparison of n1 , the jet stream gradually moves backward. This leads to an
n2 , n3 , and l in semilogarithmic coordinates is shown in Figs. 13 increase in kinetic energy loss. Based on conservation of energy,
and 14. The results show that n2 plays an important role in the the pressure drop decreases gradually with the increasing valve
total resistance coefficient l in the range of opening 20–70%, and opening.
its influence decreases with the increasing opening, The effects of As Table 4 and Fig. 15 show, the theoretical pressure drop of
n1 and n3 gradually increase as the valve opening rises and the ball valve obtained from Eq. (19) agrees well with the experi-
becomes significant when the opening is more than 70%. The mental and simulated results in Ref. [42]. The error of pressure
resistance coefficient of the thick orifice plate, nthick , is a constant drop shown in Table 4 is less than 5% in the range of the relative
that does not change with the opening. When the valve opening is opening 20–100%, which indicates again that the parametric
100%, the rotation angle is 0 deg. The equivalent thin orifice model presented in this paper is applicable and effective in practi-
plates and the Z-type elbow would disappear and their resistance cal engineering.

031204-8 / Vol. 140, MARCH 2018 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 4 Q, V, and DP of DN 50 ball valve in different valve openings
 
 Dp exp  Dp 
   100
K (%) SFlow (m2) Qexperiment (m3/h) V (m/s) DP (KPa) Dp experiment [42] (KPa)  Dp 
exp

4
20 2.2647  10 3.1476 0.4116 28.29 27.75 1.9
30 3.8768  104 5.52209 0.7222 25.75 26.498 2.8
40 5.7324  104 8.57933 1.1220 23.82 24.795 3.9
50 7.8068  104 12.2215 1.5983 20.94 21.4668 2.5
60 1.0065  103 16.6082 2.1720 17.7 18.1025 2.2
70 1.2455  103 20.4039 2.6684 12.35 11.794 4.7
80 1.4909  103 22.4234 2.9325 6.731 6.75 0.3
90 1.7337  103 24.8092 3.1138 3.362 3.5045 4.1
100 1.9635  103 24.3715 3.1873 2.043 2.0435 0

Fig. 15 Comparison of the pressure drop obtained from pre- Fig. 16 Comparison of resistance coefficients obtained from
sented model and the results in Ref. [42] the presented model and the experimental data under cavitation
in Ref. [44]

3.4 The Discussion of the Influence of Cavitation on the the resistance coefficients and pressure drop do not vary with the
Resistance Model. The flow pattern inside ball valves is very Reynolds numbers in the range of 0.64  105 to 1.59  105 and
complicated. Three vortices can be observed when the valve is not the inlet velocity in the range of 2–5 m/s. Shi et al. [43] also
fully open. Two vortices are inside the ball valve, and the other studied the influence of Reynolds numbers on the resistance coef-
one is behind the exit of the ball valve. The one outside the valve ficient of ball valves with different nominal diameter by experi-
is larger than the two inside the valve [44]. As the valve opening ments and simulations. The same results obtained when the inlet
decreases, it can be found that the number and the region of vortex velocity in the range of 1–8 m/s and Reynolds number is greater
increase [42,43]. than 0.5  105. Idel’chik et al. [25] in his book pointed out that
Cavitation, the nucleation of vapor bubbles within a liquid, may the local resistance coefficient can be considered independent of
happen inside the valve theoretically at the critical condition that Reynolds numbers in the range of 105 to 2  105. In conclusion,
the local pressure is lower than the saturated vapor pressure. Cavi- the local resistance coefficient and its corresponding pressure drop
tation is a destructive phenomenon in hydraulic components and cannot be affected by the Reynolds numbers in the range of
systems. It not only results in undesirable effects such as intensive 0.5  105 to 2  105 and the inlet velocity in the range of 1–8 m/s.
noise, vibration, and erosion of solid surfaces but also causes In order to investigate the validity of the resistance model in the
more pressure drop [9]. In other words, more energy is lost due to presence of cavitation, the resistance coefficient in cavitation
these growing vortices. Therefore, the influence of cavitation on obtained by experiment in Ref. [44] will be compared with that
the resistance model presented in this paper should be discussed. calculated by the proposed model in this paper. The critical valve
The two main factors affecting cavitation are valve openings opening at which the cavitation happens is 44.4%. The nominal
and the inlet velocity (or Reynolds number). It is easier to observe diameter of ball valve used in Ref. [44] is 50.8 mm (2 in), and the
the cavitation occurring in the high velocity regime [44]. Chern diameter of the pipe is 38 mm. The comparison of the experimen-
et al. [44] investigated the influence of Reynolds number on the tal data under cavitation and the calculated results obtained by the
resistance coefficient and its corresponding pressure drop in resistance model is shown in Table 5. And the variation of flow
the presence of cavitation by experiment; the results indicated that resistance coefficient with increasing opening is plotted in Fig. 16.

Table 5 The comparison of the experimental data in cavitation and the calculated results obtained by the resistance model

K 44.4 55.5 66.6 77.7 88.8 100

l exp under cavitation in Ref. [44] 55.0239 27.1923 12.6888 7.2300 3.5325 1.3000
l obtained by the model 61.0100 33.5809 19.6219 12.0041 7.6827 1.534

Journal of Fluids Engineering MARCH 2018, Vol. 140 / 031204-9

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 As Table 5 and Fig. 16 show, the results of flow resistance Funding Data
calculated by the proposed model are consistent well with the
experimental data in Ref. [44] in the presence of cavitation. Their  Aeronautical Science Foundation of China (Grant No.
maximum relative error is between 0.234 and 6.933 in the present xjj2015105).
range of valve openings, which can satisfy the demand of the
practical engineering. Furthermore, when the cavitation occurs,
the system has to pay more energy to maintain the same volumet- Appendix A
ric flow rate and the same inlet velocity. While as Fig. 16 shows,
the flow resistance calculated by the proposed model is slightly The projected cross section can be expressed as Eq. (1). S_
and S_ are shown in Figs. 17(a) and 17(b), respectively. MEP
larger than the experimental data under cavitation. This is benefi-
MFP
cial for designers and users and can provide reference for them to According to Fig. 17, there are
prevent cavitation. Therefore, the resistance model proposed in
this paper is also applicable under cavitation conditions. 1
S_ ¼ SMO2 P  MP  O2 N (A1)
MEP 2
4 Conclusion
1
In this paper, based on the inherent mechanism of flow resist- S_ ¼ SMO1 P  MP  O1 N (A2)
ance, an equivalent model of a ball valve is proposed to study the MFP 2
resistance loss and pressure drop, which regards the ball valve as
/MO1 P
a combination of three types of throttling elements: a thick orifice, SMO1 p ¼ SO1 (A3)
two variable-opening eccentric orifice plates, and a Z type elbow. 2p
By analyzing the three types of throttling elements, a novel para-
/MO2 P
metric model of the resistance coefficient and pressure drop of SMO2 p ¼ SO2 (A4)
ball valves is presented. 2p
By comparing with the prior experiment data, the good agree-
ment of the flow resistance coefficient and pressure drop is SO1 ¼ pr 2 (A5)
observed. The errors of flow resistance coefficient and pressure
drop are less than 18% and 5%, respectively, in the range of the SO2 ¼ pr 2 cos h (A6)
relative opening 20–100% while the corrected coefficients are not
considered here. Indicate that the presented model has relatively The equation of a circle O1
high accuracy in predicting the resistance loss and pressure drop
of ball valves in various openings during valve operation progress, x2 þ y2 ¼ r 2 (A7)
and it can meet the needs of practical engineering even if the
influence of the corrected coefficients is not considered. The equation of an ellipse O2
The contribution of the resistance coefficients corresponding to
the three types of throttling elements to the total flow resistance x2 ð y  LÞ2
coefficient is discussed. The results show that the resistance coef- þ ¼1 (A8)
r 2
ðr  cos hÞ2
ficient of thin orifice plates play an important role in the total flow
resistance coefficient, especially in the small opening, and its Vertical coordinate of point M obtained from Eqs. (A7) and (A8) is
influence decreases with the increasing valve opening. The effects
of thick orifice plates and the Z type elbow gradually increased as pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
R2  r 2  sin h
the valve opening rises and becomes significant when the opening y¼ (A9)
is more than 70%. 1 þ cos h
The effectiveness of the model of the flow resistance coefficient
and pressure drop in the presence of cavitation is also investi- So
gated. By comparing with the publicly available experimental
/MO1 P ¼ 2MO1 N ¼ 2 arccosðO1 N=rÞ ¼ 2arccosðy=rÞ (A10)
data in cavitation, a good consistency is also observed in the range
of the relative opening 40–100%. Their maximum relative error is
between 0.234 and 6.933 in the present range of valve openings, /MO2 P ¼ 2MO2 N ¼ 2 arccosð O2 N=O2 MÞ
0 1
which can satisfy the demand of the practical engineering. Fur- Ly
thermore, the flow resistance calculated by the proposed model is ¼ 2arccos@qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiA (A11)
slightly larger than the experimental data in cavitation. This is r 2  y2 þ ð L  yÞ2
beneficial for designers and users and can provide reference for
them to prevent cavitation. Indicate that the resistance model pre-
sented in this paper is also applicable under cavitation conditions.
The presented model is adapted to different geometric designs
and operating conditions and can be applied to different types of
ball valves. It provide a convenient and accurate approach for cal-
culation of the local resistance and pressure drop of ball valves,
which would provide references for the design and application of
ball valves. It should be point out two points in this study. First,
the medium used in this study is water. In fact, the local resistance
is related to the viscosity of the medium, so it is worth further
exploration in the future research. Second, the equivalent model
shown in Fig. 5 is proposed for the first time in this paper. The
model provides a methodology for the parametric modeling of the
local resistance and pressure drop in different types of hydraulic
valves. This paper is focused on the O-type rotary ball valves.
Further research needs to be carried out to study the equivalent Fig. 17 S_ and S_ : (a) S_ and (b) S_
models for the different types of hydraulic valves. MFP MEP MFP MEP

031204-10 / Vol. 140, MARCH 2018 Transactions of the ASME

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use

 Table 6 The variation of a with h [14] Casoli, P., and Anthony, A., 2013, “Gray Box Modeling of an Excavator’s Vari-
able Displacement Hydraulic Pump for Fast Simulation of Excavation Cycles,”
h (deg) 0 20 30 45 60 75 90 110 130 150 180 Control Eng. Pract., 21(4), pp. 483–494.
[15] Fang, L., Tang, L., Wang, J., and Shang, Q., 2016, “A Semi-Physical Model for
Pneumatic Control Valves,” Nonlinear Dyn., 85(3), pp. 1735–1748.
a — 2.5 2.22 1.87 1.5 1.28 1.2 1.2 1.2 1.2 1.2 [16] He, Q., and Wang, J., 2014, “Valve Stiction Quantification Method Based on a
Semiphysical Valve Stiction Model,” Ind. Eng. Chem. Res., 53(30), pp.
12010–12022.
[17] Mahrenholz, J., and Lumkes, J., 2010, “Analytical Coupled Modeling and
Table 7 The variation of b with r1 =d1 Model Validation of Hydraulic On/Off Valves,” ASME J. Dyn. Syst. Meas.
Control, 132(1), pp. 91–93.
r1 =d1 0.25 0.5 0.75 1 1.5 2.0 3.0 4.0 5.0 6.0 7.0 8.0 [18] Xu, B., Ding, R., Zhang, J., and Su, Q., 2014, “Modeling and Dynamic Charac-
teristics Analysis on a Three-Stage Fast-Response and Large-Flow Directional
Valve,” Energy Convers. Manage., 79, pp. 187–199.
b 1.1 1.07 1.04 1 0.95 0.9 0.83 0.78 0.75 0.72 0.71 0.7 [19] Valdes, J., Rodriguez, J., Saumell, J., and Putz, T., 2014, “A Methodology for
the Parametric Modelling of the Flow Coefficients and Flow Rate in Hydraulic
Valves,” Energy Convers. Manage., 88, pp. 598–611.
[20] Chisholm, D., 1983, Two-Phase Flow in Pipelines and Heat Exchangers, Long-
Arrange the equations; above all, the mathematical equation of man Higher Education, London.
[21] Roul, M. K., and Dash, S. K., 2011, “Two-Phase Pressure Drop Caused by Sud-
orifice area of the ball valve is obtained den Flow Area Contraction/Expansion in Small Circular Pipes,” Int. J. Numer.
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi ! Methods Fluids, 66(11), pp. 1420–1446.
2  r 2  sinh [22] Wang, C., Tseng, C., and Chen, I. Y., 2010, “A New Correlation and the
R
SFLOW ¼ r 2 arccos Review of Two-Phase Flow Pressure Change Across Sudden Expansion in
r ðcosh þ 1Þ Small Channels,” Int. J. Heat Mass Transfer, 53(19), pp. 4287–4295.
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi [23] Chen, I., Wongwises, S., Yang, B., and Wang, C., 2010, “Two-Phase Flow
ðR2  r 2 Þ  sin2 h pffiffiffiffiffiffiffiffiffiffiffiffiffiffi Across Small Sudden Expansions and Contractions,” Heat Transfer Eng., 31(4),
 r2   R2  r 2  sinh pp. 298–309.
cos h þ 1 [24] Liou, T., and Lin, C., 2014, “Study on Microchannel Flows With a Sudden
pffiffiffiffiffiffiffiffiffiffiffiffiffiffi , Contraction-Expansion at a Wide Range of Knudsen Number Using Lattice
2 R2  r 2  sinh  cosh Boltzmann Method,” Microfluid. Nanofluid., 16(1), pp. 315–327.
þ r cosh  arccos [25] Idel’chik, I. E., Malyavskayafs, G. R., Martynenko, O. G., and Fried, E., 1994,
cosh þ 1 Handbook of Hydraulic Resistance, 3rd ed., CRC Press, Boca Raton, FL.
v pffiffiffiffiffiffiffiffiffiffiffiffiffiffi ! 21 [26] Simpson, H. C., Rooney, D. H., and Grattan, E., 1983, “Two-Phase Flow
u Through Gate Valves and Orifice Plates,” International Conference on Physical
u 2 R 2  r 2  sin2 h
tr  A (A12) Modelling of Multi-Phase Flow, Coventry, UK, Apr., pp. 25–40.
cosh þ 1 [27] Kojasoy, G., Landis, F., Kwame-Mensah, P., and Chang, C. T., 1997, “Two-
Phase Pressure Drop in Multiple Thick- and Thin-Orifice Plates,” Exp. Therm.
Fluid Sci., 15(4), pp. 347–358.
[28] Fossa, M., and Guglielmini, G., 2002, “Pressure Drop and Void Fraction Pro-
files During Horizontal Flow Through Thin and Thick Orifices,” Exp. Therm.
Appendix B Fluid Sci., 26(5), pp. 513–523.
[29] Fossa, M., Guglielmini, G., and Marchitto, A., 2006, “Two-Phase Flow Struc-
The variations of a  h and b  r1 =d1 are shown in Tables 6 ture Close to Orifice Contractions During Horizontal Intermittent Flows,” Int.
and 7. Commun. Heat Mass Transfer, 33(6), pp. 698–708.
[30] Spedding, P. L., Benard, E., and McNally, G. M., 2004, “Fluid Flow Through
90 Degree Bends,” Asia-Pac. J. Chem. Eng., 12(1–2), pp. 107–128.
References [31] Naphon, P., and Wongwises, S., 2006, “A Review of Flow and Heat Transfer
[1] Valdes, J. R., Rodriguez, J. M., Monge, R., Pena, J. C., and Putz, T., 2014, Characteristics in Curved Tubes,” Renewable Sustainable Energy Rev., 10(5),
“Numerical Simulation and Experimental Validation of the Cavitating Flow pp. 463–490.
Through a Ball Check Valve,” Energy Convers. Manage., 78, pp. 776–786. [32] Crawford, N., Spence, S., Simpson, A., and Cunningham, G., 2009, “A Numeri-
[2] Bagherifard, S., Fernandez Pariente, I., and Guagliano, M., 2013, “Failure cal Investigation of the Flow Structures and Losses for Turbulent Flow in 90
Analysis of a Large Ball Valve for Pipe-Lines,” Eng. Failure Anal., 32(3), pp. Elbow Bends,” Proc. Inst. Mech. Eng., Part E, 223(1), pp. 27–44.
167–177. [33] Crawford, N. M., Cunningham, G., and Spence, S. W. T., 2007, “An Experi-
[3] Lorenzini, G., and Biserni, C., 2003, “A Vapotron Effect Application for Elec- mental Investigation Into the Pressure Drop for Turbulent Flow in 90 Elbow
tronic Equipment Cooling,” ASME J. Electron. Packag., 125(4), pp. 475–479. Bends,” Proc. Inst. Mech. Eng., Part E, 221(2), pp. 77–88.
[4] Moujaes, S. F., and Jagan, R., 2008, “3D CFD Predictions and Experimental [34] Sudo, K., Sumida, M., and Hibara, H., 1998, “Experimental Investigation on
Comparisons of Pressure Drop in a Ball Valve at Different Partial Openings in Turbulent Flow in a Circular-Sectioned 90-Degree Bend,” Exp. Fluids, 25(1),
Turbulent Flow,” ASCE J. Energy Eng., 134(1), pp. 24–28. pp. 42–49.
[5] Chern, M., and Wang, C., 2004, “Control of Volumetric Flow-Rate of Ball [35] Dutta, P., Saha, S. K., Nandi, N., and Pal, N., 2016, “Numerical Study on Flow
Valve Using V-Port,” ASME J. Fluids Eng., 126(3), pp. 471–481. Separation in 90 Pipe Bend Under High Reynolds Number by k-e Modelling,”
[6] Lee, J., and Lee, K., 2010, “Prediction of the Resistance Coefficient in a Seg- Eng. Sci. Technol. Int. J., 10(5), pp. 2221–2226.
ment Ball Valve,” J. Mech. Sci. Technol., 24(1), pp. 185–188. [36] Schlichting, H., Gersten, K., and Mayes, K., 2000, Boundary-Layer Theory, 8th
[7] van Lookeren Campagne, C., Nicodemus, R., de Bruin, G. J., and Lohse, D., ed., Springer, Berlin.
2002, “A Method for Pressure Calculation in Ball Valves Containing Bubbles,” [37] Ai, W., Yu, Z., Xu, T., and Hui, S., 2000, “Performance Experimental Study of
ASME J. Fluids Eng., 124(3), pp. 765–771. the Different Resistance Components for Measuring Pulverized Coal Concen-
[8] Chern, M., Tseng, P., Hsu, P., Cheng, Y., and Hu, C., 2013, “Numerical Study tration,” Power Eng., 20(5), pp. 892–895.
on Cavitation Occurrence in Globe Valve,” ASCE J. Energy Eng., 139(1), pp. [38] Cioncolini, A., Scenini, F., and Duff, J., 2015, “Micro-Orifice Single-Phase Liq-
25–34. uid Flow: Pressure Drop Measurements and Prediction,” Exp. Therm. Fluid
[9] Tabrizi, A., Asadi, M., Xie, G., Lorenzini, G., and Biserni, C., 2014, Sci., 65, pp. 33–40.
“Computational Fluid-Dynamics-Based Analysis of a Ball Valve Performance [39] Fried, E., and Idelchik, I. E., 1989, Flow Resistance: A Design Guide for Engi-
in the Presence of Cavitation,” J. Eng. Thermophys., 23(1), pp. 27–38. neers, CRC Press, Boca Raton, FL.
[10] Davis, J. A., and Stewart, M., 2002, “Predicting Globe Control Valve [40] Lyons, J. L., 1982, Lyons’ Valve Designer’s Handbook, Van Nostrand Reinhold
Performance—Part I: CFD Modeling,” ASME J. Fluids Eng., 124(3), pp. Co., New York.
772–777. [41] Kirik, M. J., and Driskell, L. R., 1986, Flow Manual for Quarter-Turn Valves,
[11] Davis, J. A., and Stewart, M., 2002, “Predicting Globe Control Valve Rockwell International, Pittsburgh, PA.
Performance—Part II: Experimental Verification,” ASME J. Fluids Eng., [42] Cui, B., Lin, Z., Zhu, Z., Wang, H., and Ma, G., 2017, “Influence of Opening
124(3), pp. 778–783. and Closing Process of Ball Valve on External Performance and Internal Flow
[12] Tordal, S., Klausen, A., and Bak, M., 2015, “Experimental System Identifica- Characteristics,” Exp. Therm. Fluid Sci., 80, pp. 193–202.
tion and Black Box Modeling of Hydraulic Directional Control Valve,” Model. [43] Shi, X., L€u, H., Zhang, K., Zhu, D., Sun, B., and Cao, B., 2013, “Analysis on
Identif. Control, 36(4), pp. 225–235. Resistance Characteristics and Flow Mechanism of PVC Ball Valve,” Trans.
[13] Witters, M., and Swevers, J., 2010, “Black-Box Model Identification for a Chin. Soc. Agric. Eng., 29(4), pp. 95–101.
Continuously Variable, Electro-Hydraulic Semi-Active Damper,” Mech. Syst. [44] Chern, M., Wang, C., and Ma, C., 2007, “Performance Test and Flow Visual-
Signal Process., 24(1), pp. 4–18. ization of Ball Valve,” Exp. Therm. Fluid Sci., 31(6), pp. 502–512.

Journal of Fluids Engineering MARCH 2018, Vol. 140 / 031204-11

Downloaded From: http://fluidsengineering.asmedigitalcollection.asme.org/ on 05/31/2018 Terms of Use: http://www.asme.org/about-asme/terms-of-use