Zhe Lin

National-Provincial Joint Engineering Laboratory

for Fluid Transmission System Technology,
Zhejiang Sci-Tech University,
5 Second Avenue,
Xiasha Higher Education Zone,
Hangzhou 310018, China
e-mail: [email protected]

Dapeng Yin
National-Provincial Joint Engineering Laboratory
for Fluid Transmission System Technology,
Zhejiang Sci-Tech University,
5 Second Avenue,
Xiasha Higher Education Zone,
Hangzhou 310018, China
Effect of Shaft Diameter on the
e-mail: [email protected]
Hydrodynamic Torque of
Junyu Tao
National-Provincial Joint Engineering Laboratory
for Fluid Transmission System Technology,
Butterfly Valve Disk
Zhejiang Sci-Tech University,
5 Second Avenue, In this study, computational fluid dynamics (CFD) software and detached eddy simulation
Xiasha Higher Education Zone, turbulence model were used to simulate butterfly valves with different designs. The effects
Hangzhou 310018, China of shaft diameters on the value and the fluctuation of valve disk torque were studied, and
e-mail: [email protected] the physical reason was discussed. The simulation results were verified by comparing
with the experimental data. The findings revealed that with the closing of the valve, the
Yi Li hydraulic torque of the valve disk first increases and then decreases. Meanwhile, the tor-
National-Provincial Joint Engineering Laboratory
que decreases gradually with the increase of the shaft diameter. The variation of torque
for Fluid Transmission System Technology,
is caused by the change of pressure on both sides of the valve disk. The result also indi-
Zhejiang Sci-Tech University,
cates that the fluctuation of torque is induced by the flow separation phenomenon occurs
5 Second Avenue,
on the valve disk. The fluctuation is significant for the valve opening from 20% to 60%.
Xiasha Higher Education Zone,
The strength of the torque fluctuation is greater for the smaller shaft diameter. This study
Hangzhou, 310018, China
provides a theoretical basis for the design and optimization of butterfly valves.
e-mail: [email protected]
[DOI: 10.1115/1.4047795]

Keywords: butterfly valve, shaft diameter, hydrodynamic torque, fluctuation, numerical

Jin Sun simulation
Bray (China) Control System Co., Ltd.,
No. 6 Gaoxin Road,
Xiaoshan Economic and Technological
Development Zone,
Hangzhou 310018, China
e-mail: [email protected]

Zuchao Zhu1
National-Provincial Joint Engineering Laboratory
for Fluid Transmission System Technology,
Zhejiang Sci-Tech University,
5 Second Avenue,
Xiasha Higher Education Zone,
Hangzhou 310018, China
e-mail: [email protected]

1 Introduction the control structure. When the valve is opening, the difference in
fluid pressure in different areas of the valve disk forms a torque
Availability of energy is one of the most important issues in
on the rotating shaft. The value and change law of the torque have
today’s society. Energy exploitation, transportation, and utiliza-
a great influence on the design and use of the valve body, the
tion have always been popular research topics [1–6]. Valves are
transmission mechanism, and the control mechanism. Therefore, a
indispensable components in pipeline systems for energy transport
lot of research has been conducted on the valve disk torque and
and are used for the regulation of the transmission media. The but-
the flow characteristics in the butterfly valve [7–20].
terfly valve is a common regulating valve with a rotatable disk as
Caraballo [7] used three-dimensional topology and genetic
algorithm to optimize the butterfly valve disk. Sun et al. [8] stud-
1
Corresponding author. ied the effect of friction coefficient on the flow coefficient of a tri-
Contributed by the Fluids Engineering Division of ASME for publication in the eccentric butterfly valve by numerical simulation. Toro et al. [9]
JOURNAL OF FLUIDS ENGINEERING. Manuscript received October 21, 2019; final
manuscript received July 2, 2020; published online August 7, 2020. Assoc. Editor: investigated the accuracy of using computational fluid dynamics
Praveen Ramaprabhu. (CFD) technology to predict the performance of butterfly valves.

Journal of Fluids Engineering Copyright V

C 2020 by ASME NOVEMBER 2020, Vol. 142 / 111202-1
The author concluded that the numerical simulation results are
accurate at medium openings. Song et al. [10] integrated a met
model with finite element method and CFD analysis and used
Taguchi orthogonal array technique and Kriging model to obtain
the minimum weight of butterfly valves. Yi et al. [11] proposed an
optimal design method for butterfly valves by considering the
structural safety and flow characteristics. Wang et al. [12] placed
a double butterfly valve in a curved channel and investigated the
flow in the pipe and the aerodynamic torque of the valve disk with
different closing times. Lattouf and Huynh [13] analyzed the flow
field of a large caliber butterfly valve by using the CFD method
and discussed the maximum torque acting on the valve shaft.
Farid and Mahdi [14] also used the CFD method to calculate the
torque of a large butterfly valve with different working conditions
and openings. The effects of disk shape and deformation, surface
roughness, upstream/downstream pressure, and disk offset were
also acquired. Morris and Dutton [15] experimentally studied the
compressible flow fields inside circular and parallel butterfly
valves. The variations of torque on the surfaces of valve disks
were also investigated and analyzed. Leutwyler and Dalton [16]
used the CFD method to investigate the subsonic and supersonic
internal flow fields of a butterfly valve. The aerodynamic torque
of valve disk was analyzed with an opening of 45 deg and 70 deg.
Morris and Dutton [17] measured the aerodynamic torques of but-
terfly valves by employing experimental methods and obtained
the effects of valve disk shape, flow separation, and reattachment
position on the aerodynamic torques of butterfly valves. Nasera-
dinmousavi and Natraj [18] analyzed the opening and closing
processes of a butterfly valve. The results indicated that the hydro- Fig. 1 Structure of the butterfly valve
dynamic torque plays an important role in both processes. Liu et al.
[19] examined the effects of inlet speed, particle mass fraction, and is mainly composed of four parts: valve body, rotating shaft, seal
solid particle diameter on the performance of a butterfly valve. ring, and disk. The valve disk rotates from 0 deg (fully closed) to
Baran et al. [20] conducted cavitation experiments on a DN100 90 deg (fully open). The working medium is water, and the density
butterfly valve with different valve openings and discussed the is 998.2 kg/m3. The tests and simulations were initially performed
cavitation evolution. for the original valve disk to verify the simulation.
In summary, most research on butterfly valve focuses on inter- A nondimensional parameter Z was used to represent the rela-
nal flow fields and the valve disk torque by using CFD and experi- tionship between the rotary shaft and the valve disk
mental methods. But the research on torque is mainly in steady-
state. According to the flow around the blunt body, it was found Dshaft
Z¼ (1)
that the flow separation around the bluff body caused pressure pul- Ddisc
sation, which in turn caused transient effects [21,22]. The flow in
the butterfly valve is also similar to the flow around the blunt where Dshaft denotes the diameter of the shaft, whereas Ddisc
body, and pressure fluctuations will also occur on the valve disk. denotes the diameter of the disk. In this study, The Z value of the
Long-term instantaneous pressure pulsation will cause torque fluc- original model is 0.36. Three valve disks were then created on the
tuations that will affect the strength of the rotating shaft, valve basis of the original model to investigate the effect of the shaft
disk and actuator, and cause fatigue fracture. The rotating shaft is diameter. The valves examined had Z values of 0.1, 0.2, and 0.4.
an important part for the transmission of valve torque. Its diameter In the simulation, 14 pressure monitoring points were set on the
affects the structure of the valve disk (a large shaft diameter midsection of the valve disk, as shown in Fig. 2. A1, A2, G1, and
results in a thick valve disk), thus changing the flow performance G2 denote the end points of the valve disk. B1, B2, F1, and F2
of the butterfly valve. However, in previous studies, the variation denote the midpoints of the disk. C1, C2, E1, and E2 denote the
of valve disk torque with the diameter of the shaft has not been junction points of the shaft and the flat disk. D1 and D2 denote the
obtained. Therefore, it is necessary to analyze the fluctuation of vertexes of the shaft.
the valve disk torque with different shaft diameter.
The CFD method has been effective in investigating the inter-
nal flow and the corresponding problems in hydraulic compo- 3 Experimental Method
nents, such as pumps, valves, tubes, etc. [23–41]. Thus, in this Figure 3 illustrates the schematic and shows a photograph of
study, butterfly valves with different shaft diameters (different the experimental system. The valve test system is a closed cycle
valve disks) were explored by transient CFD simulations. The hydraulic loop and includes a water tank, a variable-frequency
time-averaged disk torque was compared and analyzed. The effect pump, many circulation pipes, certain pressure sensors, a flow-
of shaft diameter on the valve disk torque was obtained. To meter, a torque meter, and a test valve. The main purpose of the
explain the reason for the variation of the disk torque, pressures experiment is to measure the pressure at the inlet and outlet and
on the surfaces of disks were also extracted for in-depth analysis. the torque of the valve disk, so as to provide the basis for the accu-
The fluctuation of the instantaneous torque under different open- racy of the subsequent numerical simulation. The pipe diameter
ing degrees was also studied for the first time. This study provides (D) of the tube is 50 mm. The test range of the flowmeter is
a reference for the optimization of butterfly valves. 0–60 m3/h. The test range of pressure sensors is 0–150 kPa, with
the measuring accuracy of 0.2%. The range of the torque meter is
0–20 Nm, with the measuring accuracy of 0.001 Nm. Four pres-
2 Problem Statement sure holes (top, bottom, left, and right) were drilled around the
This research used a DN50 butterfly valve as the research cross sections (5D upstream the valve and 10D downstream), and
object, and the structure is illustrated in Fig. 1. The butterfly valve the average pressure for each section was calculated. The rotating

111202-2 / Vol. 142, NOVEMBER 2020 Transactions of the ASME

 Fig. 2 Position of monitoring points

Fig. 3 The structure of the computational domain

shaft of the butterfly valve was connected to the torque meter. A

bearing was added to the shaft to reduce the friction torques and
to make them smaller than the corresponding hydrodynamic tor-
ques. Pressures and torques were recorded when they were basi-
cally stable.
In the experiment, the idling torque (friction torque) for each
valve opening was measured first. Then, the torque of the working
state as recorded. Finally, the actual hydrodynamic torque was
calculated by adding idling torque with working torque. The
uncertainty of torque measurement is caused by turbulence fluctu-
ation and the measurement accuracy. To reduce the experimental
error, each experiment was repeated ten times, and the average
values were calculated. Uncertainty evaluations were performed
for ten tests, and the results are shown in Fig. 4. The maximum
uncertainty of the torque measurement is 61.6% (Relative to the
mean torque). The experimental uncertainties of the pressure, the Fig. 4 Mesh of the computational region

Journal of Fluids Engineering NOVEMBER 2020, Vol. 142 / 111202-3

 Fig. 5 Schematic of the experimental system

Fig. 6 Uncertainty of experimental results

mass flowrate, and the valve flap angle are 61.1%, 60.4%, and
60.7%, respectively. The Reynolds number of all working condi- Table 1 Boundary conditions of different valve openings
tions in this paper ranges from 6831 to 88826.
Open angle (%) Inlet pressure (kPa) Outlet pressure (kPa)
Uncertainty U was calculated as follows:
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 10 97.67 15.36
U ¼ Ua2 þ Ub2 (2) 20 97.10 16.41
30 95.55 16.95
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 40 90.85 21.76
ui¼1n
uP ðyi  yÞ2 50 79.52 16.65
t 60 68.92 18.42
Ua ¼ (3)
nðn  1Þ 70 49.30 21.34
80 41.18 23.93
pffiffiffi 90 38.03 25.25
Ub ¼ a= 3 (4) 100 36.81 25.52

where y is the measured value of each experiment, y is the average

of ten tests, n is the number of tests, and a is the measurement
accuracy of the instrument. adopted. This method combines the advantages of the large eddy
simulation (LES) method and the Reynolds-averaged Navier–-
Stokes (RANS) equation method. It handles the motion of large-
4 Numerical Method and Validation scale vortices in the flow field by using the LES method, while
To obtain the flow characteristics inside the butterfly valve, calculates the boundary layer of the appendage by using the
the CFD method was used with the commercial software ANSYS- RANS method. This method has also been applied and verified in
FLUENT 15.0. In the simulation procedure, the computational the simulations of the transient internal flow properties and the
region was divided into three parts: upstream pipework, butterfly flow-induced problems of pumps and valves [46–48]. The solver
valve, and downstream pipework, as displayed in Fig. 5. The type was pressure-based, and the pressure–velocity coupling
length of the upstream pipe is five times of the pipe diameter, and scheme was simple. Spatial discretization selected the second-
the downstream pipe is ten times of the pipe diameter. order upwind, whereas the momentum chose bounded central
To ensure the quality of the grid and improve the calculation differencing.
efficiency, the mesh in the valve part was refined, as shown in The grid independence test was performed for the original
Fig. 6. The performance characteristics of the butterfly valve with butterfly valve. Table 2 shows the deviation of average torque
different opening were achieved using transient simulation. between two adjacent grids. After grid sensitivity checks, it is
Steady numerical results obtained by the realizable k–e model
were set as the initial condition. To achieve adequate resolution of
Table 2 Change of torque with grid number
unsteady simulation results, time-step is set as Dt ¼ 1  104 s,
and convergence criteria of the momentum and mass are also Grid number M Discrepancy Experimental
set as 5  105. The total pressure inlet and static pressure outlet (million) (Nm) (%) value (Nm)
were adopted for the inlet and outlet boundaries (the pressure val-
ues at the inlet and outlet of different opening degrees are shown 1.03 0.266535 0.74
in Table 1). Nonslip boundary conditions were used for wall 1.46 0.268512 12.57
surfaces. The pressure values were determined in accordance with 2.11 0.307119 2.60 0.3116
the experimental values under different openings. 3 0.315327 0.29
By the comparison of the transient simulation methods, in order 3.83 0.316256 0.11
to ensure the calculation accuracy and reduce the calculation 4.55 0.315896
resources, the detached eddy simulation method [42–45] was

111202-4 / Vol. 142, NOVEMBER 2020 Transactions of the ASME

found that the discrepancy of the torque is smaller than 0.29% The mean square deviation of disk torque between 0.1 s and
when the overall mesh element exceeds 3  106. Thus, the grid 0.45 s is calculated to express the intensity of torque fluctuation,
number of 3.8  106 was applied for the whole computational as shown in Fig. 10. The larger the mean square deviation, the
region. The mesh must guarantee the y plus <30 in the critical more violent the fluctuation. The figure illustrates that the torque
computational domain. fluctuation is relatively large when the valve opening varies
Numerical torques of the original valve disk for different valve between 20% and 60%. The maximum value occurs at the valve
openings were obtained and compared with the experimental opening of 30%. When the opening is larger than 60%, the fluctu-
results, as displayed in Fig. 7. The simulation results agree well ation intensity decreases as valve opening increases. Meanwhile,
with the experimental results, and the largest relative difference the small shaft diameter leads to a greater torque fluctuation.
between them is 4.5%. Therefore, the present simulations are
accurate and reliable.
5.3 Valve Disk Pressure Distribution. The change of valve
disk torque is caused by the pressure distribution of the valve
5 Results and Discussion disk. Figure 11 shows the distribution of the static pressure on the
In this research, the numerical simulation method is used to valve disk when the opening is 30%, 60%, and 90%. For the valve
study butterfly valves. The changes of the valve disk torque and opening of 30%, the pressure in the middle of the front side is rel-
the surface pressure distributions of the different shaft diameters atively high, and the value decreases toward the edge and shows a
of the butterfly valves are discussed. The details are as follows. convex trend. The pressure distribution on the left side is symmet-
rical with that on the right side, but the high-pressure area on the
5.1 Instantaneous Torque Fluctuations. Figure 8 shows the left side is slightly larger than that on the right side. For the valve
time-history fluctuation of torque for the valve opening of 30%, opening of 60%, the pressure on the front side is unevenly distrib-
60%, and 90%. A large shaft diameter makes a relatively small uted, and the pressure at the inlet end is greater than the pressure
fluctuation amplitude, that is, the fluctuation is the strongest for at the outlet end. For the valve opening of 90%, the pressure at
Z ¼ 0.1. Meanwhile, the torque fluctuation decreases as the valve both ends of the front side is high, the pressure of the shaft part is
opening increases. When the valve opening is 30%, the fluctuation low, and the whole body has a concave distribution. However, the
amplitude for Z ¼ 0.1 is 4.45% (relative mean torque). The fluctu- pressure at the inlet end is greater than the pressure at the outlet
ation amplitude changes to 1.42% when the valve opening end. The low-pressure zone easily appears at the top of the front
becomes 60% and changes to 1.41% when the valve opening turns side rotating shaft, and the high-pressure zone easily appears at
to 90%. The figure also demonstrates that the fluctuations are the junction of the rotating shaft and the valve plate. This phe-
irregular for most situations. For the cases of the valve opening of nomenon is evident as the diameter and opening of the rotating
90% and Z ¼ 0.2, torque periodically fluctuates. In this case, the shaft increases.
deviation of the leading edge from the central axis is exactly equal The back side of the valve disk is in the separation zone, and
to the radius of the rotating shaft. When Z increases to 0.4 (valve the pressure of the back side is smaller than that of the front side.
opening of 90%), the rule of periodicity changes, and the uniform The pressure at the edge of the valve disk and the shaft is low.
periodic fluctuation turns to the repeat of a period of stochastic This phenomenon is also evident as the opening and diameter of
fluctuation. the shaft increase. The uneven distribution of the valve disk pres-
sure leads to the change of the valve plate torque. The more
uneven the pressure distribution on the upstream and downstream
5.2 Time-Averaged Torque. The disk torque fluctuates sides of the valve disk, the greater the valve plate torque.
around a value after 0.1 s, so the torque of different valve open-
ings between 0.1 s–0.45 s was averaged, as shown in Fig. 9. It also
covers the time response of all key flow structures. As valve open- 5.4 Valve Disk Centerline Pressure Change. To obtain the
ing increases, the torque coefficient increases first and then specific pressure value on the valve disk, the pressure distribution
decreases. The maximum torque occurs at the valve opening on the valve disk centerline is extracted. Figure 12 illustrates the
between 50% and 60% due to the variation of pressure distribution static pressure distribution of the monitoring points along the
on the valve disk. This variation is caused by the change of flow valve disk centerline, in which the longitudinal coordinate repre-
fields. These factors are discussed in Sec. 5.6. By comparing the sents the time-averaged static pressure value of the monitoring
results of different shafts, the maximum torque decreases as Z point. In addition, the abscissa represents the monitoring point.
increases. For the valve opening of 30%, the difference in the pressure distri-
bution of the shaft portion is small. As the valve opening
increases, the low pressure occurs in the shaft portion, and the
low-pressure value on the shaft decreases as the diameter of the
shaft increases. The maximum pressure can appear at the junction
of the shaft and the valve disk (point C). The pressure distribution
in the middle of the valve disk (from points B to F) is symmetri-
cal, and the pressure values at points A and G are quite different.
The pressure at both ends of the valve disk has a greater influence
on the torque of the valve plate. On the front side, the pressure at
point A1 is greater than the pressure at point G1. On the back
side, the pressure at point A2 is less than that at point G2, and the
pressure difference between points A1 and G1 is greater than that
between points A2 and G2. Therefore, the pressure distribution on
the front side has a greater influence on the torque than that on the
back side.
The pressure difference between the front and back sides is cal-
culated. The pressure difference in the middle of the valve disk is
found relatively symmetrical and has little effect on the torque.
The pressure difference at the inlet end of the valve disk (A1–A2)
is the largest, and the pressure difference at the outlet end of the
Fig. 7 Comparison of simulation results with experimental valve disk (G1–G2) is the smallest. Therefore, the pressure differ-
data ence between the two ends of the valve disk has a greater impact

Journal of Fluids Engineering NOVEMBER 2020, Vol. 142 / 111202-5

 Fig. 8 Time history variation of butterfly valve torque under different opening degrees

on the valve plate torque. For the opening degree of 30%, the Pressure fluctuation in different areas of the valve disk causes
pressure difference at the inlet end (A1–A2) of the valve disk with the fluctuation of the valve plate torque. The pressure fluctuation
different Z values is basically the same, and that at the outlet end on the valve disk centerline was extracted. Figure 13 shows the
(G1–G2) increases as the Z value increases. Hence, the torque of static pressure fluctuation of the monitoring point along the valve
the valve disk decreases as the diameter of the shaft increases. For disk centerline when the opening is 30%, 60%, and 90%. The lon-
the opening of 60%, the pressure difference at the inlet end gitudinal coordinate represents the variance of static pressure fluc-
(A1–A2) increases as the Z value decreases and that at the outlet tuation at the monitoring point, and the abscissa represents the
end (G1–G2) increases as the Z value increases. Thus, the disk tor- monitoring point. When the opening degree is small, the pressure
que decreases as the diameter of the shaft increases. For the open- fluctuations at both ends of the valve disk are severe. As the open-
ing of 90%, when Z ¼ 0.4, the pressure difference at the inlet end ing increases, the maximum pressure fluctuation intensity in the
(A1–A2) is the largest, and the pressure difference at the inlet end valve disk decreases. However, the influence of pressure fluctua-
of other valve disks is small. The pressure difference at the outlet tion on the shaft area can be evident, especially the pressure fluc-
end (G1–G2) of the different disk is basically the same. Therefore, tuation at the top of the shaft as the opening and Z value increase.
the valve plate torque decreases as the diameter of the shaft Therefore, the pressure fluctuation at the end of the valve disk and
increases. at the top of the shaft is large. Moreover, the pressure fluctuation

111202-6 / Vol. 142, NOVEMBER 2020 Transactions of the ASME

 set at the front and rear edges of the valve disk, and the pressure
difference between the front and rear edges of the valve disk is
obtained. The time variation of the pressure difference between
the front and back of the valve disk over the period 0.1–0.45 s is
shown in Fig. 14. The DP represents the differential pressure at
corresponding monitoring points A and G. For the opening degree
of 30%, the pressure difference at the inlet end (A1–A2) fluctuates
greatly, and the amplitude of the pressure difference fluctuation of
the Z ¼ 0.1 disk is up to 21% (relative mean pressure difference).
The amplitude decreases as shaft diameter increases. The pressure
difference at the outlet end (G1–G2) is small. The amplitude
decreases as the shaft diameter increases. Therefore, the amplitude
of the torque fluctuation decreases as the shaft diameter increases.
For the opening of 60%, the fluctuation intensity of pressure dif-
ference between inlet and outlet is less than 30%; hence, the fluc-
tuation of torque is small. The pressure difference of Z ¼ 0.1 is the
most severe, so the torque fluctuation of the small diameter disk is
also severe. Moreover, the pressure difference at the inlet end of
Z ¼ 0.4 valve disk fluctuates periodically, but the irregular fluctua-
Fig. 9 Torque coefficients of valve disk for valve openings tion at the outlet destroys the periodicity. Thus, the torque does
not fluctuate periodically. For the opening of 90%, the pressure
fluctuation at the inlet end and outlet end is small. At this time,
the pressure difference fluctuation of the valve disk with Z ¼ 0.4 is
intense. Under the influence of the shaft, the flow of the valve disk
with large diameter can be unstable, which can cause additional
fluctuation of the pressure difference. The pressure difference
between the inlet and outlet of the valve disk with Z ¼ 0.2 has per-
iodic fluctuations because the deviation of the leading edge from
the central axis is exactly equal to the radius of the rotating shaft.
At this time, the medium flow state is suitable for the valve disk,
and the flow can be stable. As a result, periodic pressure fluctua-
tions occur, as shown in Fig. 10, at which time the disk torque
also produces periodic fluctuations.

Fig. 10 Torque fluctuation of valve disk under different open- 5.6 Flow Field Distribution. Figure 15 displays the pressure
ing degree distribution on the cross section near the valve disk. Figure 16
Velocity streamline distribution shows the streamline distribution
at point G1 is greater than that at point A1 on the front side, and near the valve disk. For the opening of 30%, the restrictive action
the pressure fluctuation at point A2 is greater than that at point G2 of the valve disk reduces the flow area, the flow velocity near the
on the back side. front side is low, and a large part of the high-pressure area appears
in front of the valve. Flow separation occurs at the edge of the
valve disk. Hence, the edge velocity of the valve disk is high, and
5.5 Valve Disk Front and Rear Edge Pressure Difference the pressure is low. Behind the disk, the pressure is small because
Changes. From the above analysis, the pressure changes of the of the flow separation. This condition also leads to the pressure
front and rear edges of the valve disk have a great influence on the distribution illustrated in Fig. 11. For the opening degree of 60%,
torque of the valve disk. Therefore, pressure monitoring points are the change of flow area leads to the change of velocity near the

Fig. 11 Valve disk pressure distribution

Journal of Fluids Engineering NOVEMBER 2020, Vol. 142 / 111202-7

 Fig. 12 Pressure distribution on the valve disk centerline

Fig. 13 Pressure fluctuation on the valve disk centerline

front face, which makes the pressure field in front of the valve the flow area increases, the flow velocity increases, the pressure
nonuniform. At this time, the influence of the shaft increases, the value before the valve decreases, the pressure difference between
flow separation at the shaft becomes more serious, and the low the front and back sides decreases, the flow separation phenom-
pressure can appear in the shaft. At the junction of the shaft and enon at the shaft becomes more serious, the pressure at the shaft is
the valve disk, stable vortices can be found, and the fluid can stag- the smallest, the diameter of the shaft increases, and the flow sepa-
nate, thereby indicating high pressure. For the opening of 90%, ration phenomenon at the shaft becomes more evident. Comparing

111202-8 / Vol. 142, NOVEMBER 2020 Transactions of the ASME

 Fig. 14 Time difference of pressure difference between the front and rear edges of the valve disk

Fig. 15 Pressure distribution of butterfly valve

the pressure distribution of butterfly valves under different work- at the outlet end of the front side is low. In the vicinity of the sepa-
ing conditions, the separation point at the inlet end of the valve ration point, the pressure fluctuates greatly. Hence, the pressure
disk is found close to the back side, whereas the separation point fluctuation at the outlet end of the front side in Fig. 13 is greater
at the outlet end is close to the front side. Therefore, in Fig. 12, than the inlet end, and the pressure fluctuation at the inlet end of
the pressure at the inlet end of the back side is low, whereas that the back side is greater than the outlet end.

Journal of Fluids Engineering NOVEMBER 2020, Vol. 142 / 111202-9

 Fig. 16 Velocity streamline distribution

6 Conclusions [5] Zheng, X., Lin, Z., and Xu, B., 2019, “Thermal Conductivity and Sorption Per-
formance of Nano-Silver Powder/FAPO-34 Composite Fin,” Appl. Therm.
In this study, the effects of shaft diameter on the value and the Eng., 160, pp. 114055–114094.
fluctuation of valve disk torque were analyzed by numerical simu- [6] Li, X., Chen, B., Luo, X., and Zhu, Z., 2020, “Effects of Flow Pattern on
Hydraulic Performance and Energy Conversion Characterisation in a Centrifu-
lation. Three kinds of shaft diameters and ten kinds of openings gal Pump,” Renewable Energy, 151(5), pp. 475–487.
are discussed. The validity of the simulation results was verified 
[7] Corbera Caraballo, S., Olazagoitia Rodrıguez, J. L., Lozano Ruiz, J. A., and Alvarez
by comparing the time-averaged torque of butterfly valve disk Fernandez, R., 2017, “Optimization of a Butterfly Valve Disc Using 3D Topology
with the experimental results. and Genetic Algorithms,” Struct. Multidiscip. Optim., 56(4), pp. 941–957.
[8] Sun, X., Kim, H. S., Yang, S. D., Kim, C. K., and Yoon, J. Y., 2017, “Numerical
The results show that the hydrodynamic valve disk torque Investigation of the Effect of Surface Roughness on the Flow Coefficient of an
increases first and then decreases with the closing of the butterfly Eccentric Butterfly Valve,” J. Mech. Sci. Technol., 31(6), pp. 2839–2848.
valve. Meanwhile, the hydrodynamic valve disk torque decreases [9] Toro, A. D., Johnson, M. C., and Spall, R. E., 2015, “Computational Fluid
gradually with increases of the shaft diameter. This is caused by Dynamics Investigation of Butterfly Valve Performance Factors,” J. Am. Water
Works Assoc., 107(5), pp. E243–E254.
the pressure distribution on the front and back sides of the valve [10] Song, X. G., Wang, L., and Park, Y. C., 2009, “Analysis and Optimization of a
disk. The more uneven the pressure distribution on both sides of Butterfly Valve Disc,” Proc. Inst. Mech. Eng., Part E, 223(2), pp. 81–89.
the valve disk, the greater the torque. It is also indicated that the [11] Yi, S. I., Shin, M. K., Shin, M. S., Yoon, J. Y., and Park, G. J., 2008,
torque fluctuation of the valve disk is intense and disordered when “Optimization of the Eccentric Check Butterfly Valve Considering the Flow
Characteristics and Structural Safety,” Proc. Inst. Mech. Eng., Part E, 222(1),
the opening is small, and it becomes regular with the increase of pp. 63–73.
valve opening. The strength of torque fluctuation decreases with [12] Wang, P., Liu, Y., and Hu, Z., 2016, “Rapid Close of a Butterfly Valve Placed
the increases of the shaft diameter. The results reveal that the fluc- in a Curved Channel: A Computational Study of Unsteady Steam Flow and
tuation of the torque is induced by the pressure fluctuation at both Aerodynamic Torque,” ASME Paper No. UNSP V008T26A003.
[13] Lattouf, D., and Huynh, B. P., 2017, “Flow Across a Butterfly Valve in a Dam
ends of the valve disk, which is caused by the flow separation Penstock,” ASME Paper No. IMECE2017-71322.
occurs on the valve disk. [14] Farid, V., and Mahdi, M., 2012, “Analysis of the Hydrodynamic Torque Effects
on Large Size Butterfly Valves and Comparing Results With AWWA C504
Standard Recommendations,” J. Mech. Sci. Technol., 26(9), pp. 2799–2806.
Funding Data [15] Morris, M. J., and Dutton, J., 1989, “Compressible Flowfield Characteristics of
Butterfly Valves,” ASME J. Fluids Eng., 111(4), pp. 400–407.
 National Natural Science Foundation of China (Grant No. [16] Leutwyler, Z., and Dalton, C., 2008, “A CFD Study of the Flow Field,
51876193 and Grant No. 51976197; Funder ID: 10.13039/ Resultant Force, and Aerodynamic Torque on a Symmetric Disk Butterfly
Valve in a Compressible Fluid,” ASME J. Pressure Vessel Technol., 130(2), pp.
501100001809). 1030–1036.
 Key Research and Development Program of Zhejiang Prov- [17] Morris, M. J., and Dutton, J. C., 1989, “Aerodynamic Torque Characteristics of
ince (Grant No. 2020C03099; Funder ID: 10.13039/50110 Butterfly Valves in Compressible Flow,” ASME J. Fluids Eng., 111(4), pp.
0004731). 392–399.
[18] Naseradinmousavi, P., and Nataraj, C., 2011, “Nonlinear Mathematical Model-
ing of Butterfly Valves Driven by Solenoid Actuators,” Appl. Math. Modell.,
35(5), pp. 2324–2335.
Nomenclature [19] Liu, B., Zhao, J., and Qian, J., 2017, “Numerical Study of Solid Particle Erosion
in Butterfly Valve,” Second International Conference on Design, Materials, and
D¼ pipe diameter (mm) Manufacturing (ICDMM), Beijing, China.
M¼ torque (Nm) [20] Baran, G., Catana, I., Magheti, I., Safta, C. A., and Savu, M., 2010,
Pavg ¼ average pressure (kPa) “Controlling the Cavitation Phenomenon of Evolution on a Butterfly Valve,”
SD ¼ mean square error IOP Conf. Ser.: Earth Environ. Sci., 12, p. 012100.
[21] Deng, Y. F., He, C. X., Wang, P., and Liu, Y. Z., 2020, “Unsteady Behaviors of
Z¼ ratio of rotary shaft diameter to valve disk diameter Separated Flow Over a Finite Blunt Plate at Different Inclination Angles,”
U¼ uncertainty Phys. Fluids, 3(32), p. 035111.
DP ¼ pressure difference (kPa) [22] Zhang, Q. S., and Liu, Y. Z., 2017, “Separated Flow Over Blunt Plates With
Different Chord-to-Thickness Ratios: Unsteady Behaviors and Wall-Pressure
Fluctuations,” Exp. Therm. Fluid Sci., 84, pp. 199–216.
References [23] Li, X., Shen, T., Li, P., Guo, X., and Zhu, Z., 2020, “Extended Compressible
[1] Song, T., Hartge, E.-U., Heinrich, S., Shen, L., and Werther, J., 2018, Thermal Cavitation Model for the Numerical Simulation of Cryogenic Cavitat-
“Chemical Looping Combustion of High Sodium Lignite in the Fluidized Bed: ing Flow,” Int. J. Hydrogen Energy, 45(16), pp. 10104–10118.
Combustion Performance and Sodium Transfer,” J. Greenhouse Gas Control, [24] Rafiee, S. E., and Sadeghiazad, M. M., 2017, “Experimental and 3D CFD Inves-
70, pp. 22–31. tigation on Heat Transfer and Energy Separation Inside a Counter Flow Vortex
[2] Li, X., Li, B., Yu, B., Ren, Y., and Chen, B., 2019, “Calculation of Cavitation Tube Using Different Shapes of Hot Control Valves,” Appl. Therm. Eng., 110,
Evolution and Associated Turbulent Kinetic Energy Transport Around a pp. 648–664.
NACA66 Hydrofoil,” J. Mech. Sci. Technol., 33(3), pp. 1231–1241. [25] Tao, J., Lin, Z., Ma, C., Ye, J., Zhu, Z., Li, Y., and Mao, W., 2020, “An Experi-
[3] Lin, Z., Liu, Z., Liu, Q., and Li, Y., 2020, “Fluidization Characteristics of mental and Numerical Study of Regulating Performance and Flow Loss in a V-
Particles in a Groove Induced by Horizontal Air Flow,” Powder Technol., 363, Port Ball Valve,” ASME J. Fluids Eng, 142(2), p. 021207.
pp. 442–447. [26] Lin, Z., Ma, G., Cui, B., Li, Y., Zhu, Z., and Tong, N., 2016, “Influence of Flash-
[4] Mitomi, M., and Nagano, H., 2014, “Long-Distance Loop Heat Pipe for Effec- board Location on Flow Resistance Properties and Internal Features of Gate
tive Utilization of Energy,” Int. J. Heat Mass Transfer, 77, pp. 777–784. Valve Under the Variable Condition,” J. Nat. Gas Sci. Eng., 33, pp. 108–117.

111202-10 / Vol. 142, NOVEMBER 2020 Transactions of the ASME

[27] Lin, Z., Sun, X., Yu, T., Zhang, Y., Li, Y., and Zhu, Z., 2020, “Gas–Solid Two- [38] Wei, Y., Wang, Z., Dou, H., and Qian, Y., 2017, “A Novel Two-Dimensional
Phase Flow and Erosion Calculation of Gate Valve Based on the CFD-DEM Coupled Lattice Boltzmann Model for Incompressible Flow in Application of
Model,” Powder Technol., 366, pp. 395–407. Turbulence Rayleigh-Taylor Instability,” Comput. Fluids, 156, pp. 97–102.
[28] Lun, Y., Lin, L., He, H., Ye, X., Zhu, Z., and Wei, Y., 2019, “Effects of Vortex [39] Zhang, W., Li, X., and Zhu, Z., 2019, “Quantification of Wake Unsteadiness for
Structure on Performance Characteristics of a Multiblade Fan With Inclined low-Re Flow Across Two Staggered Cylinders,” J. Mech. Eng. Sci.,
Tongue,” Proc. Inst. Mech. Eng. Part A, 233(8), pp. 1007–1021. 233(19–20), pp. 6892–6909.
[29] Chen, X., Li, X., and Zhu, Z., 2019, “Effects of Dimensional Wall Temperature [40] Yang, H., Yu, P., Xu, J., Ying, C., Cao, W., Wang, Y., Zhu, Z., and Wei, Y.,
on Velocity-Temperature Correlations in Supersonic Turbulent Channel Flow 2019, “Experimental Investigations on the Performance and Noise Characteris-
of Thermally Perfect Gas,” Sci. China Phys., Mech. Astron., 62(6), p. 064711. tics of a Forward-Curved Fan With the Stepped Tongue,” Meas. Control,
[30] Jin, Z. J., Qiu, C., Jiang, C. H., Wu, J. Y., and Qian, J. Y., 2020, “Effect of 52(9–10), pp. 1480–1488.
Valve Core Shapes on Cavitation Flow Through a Sleeve Regulating Valve,” J. [41] Liu, Y., and Tan, L., 2020, “Theoretical Prediction Model of Tip Leakage Vor-
Zhejiang Univ. Sci. A, 21(1), pp. 1–14. tex in a Mixed Flow Pump With Tip Clearance,” ASME J. Fluids Eng., 142(2),
[31] Liu, Q., Ye, J., Zhang, G., Lin, Z., Xu, H., Jin, H., and Zhu, Z., 2019, “Study on p. 021203.
the Metrological Performance of a Swirlmeter Affected by Flow Regulation [42] Spalart, P. R., 2000, “Trends in Turbulence Treatments,” AIAA Paper No.
With a Sleeve Valve,” Flow Meas. Instrum., 67, pp. 83–94. 2000-2306..
[32] Liu, Q., Ye, J.-h., Zhang, G., Lin, Z., Xu, H.-G., and Zhu, Z.-C., 2019, [43] Strelets, M., 2001, “Detached Eddy Simulation of Massively Separated Flows,”
“Metrological Performance Investigation of Swirl Flowmeter Affected by Vor- AIAA Paper No. 2001-0879.
tex Inflow,” J. Mech. Sci. Technol., 33(6), pp. 2671–2680. [44] Spalart, P. R., Deck, S., Shur, M. L., Squires, K. D., Strelets, M. K., and Travin,
[33] Qian, J., Hou, C., Wu, J., Gao, Z., and Jin, Z., 2019, “Aerodynamics Analysis A., 2006, “A New Version of Detached-Eddy Simulation, Resistant to Ambigu-
of Superheated Steam Flow Through Multi-Stage Perforated Plates,” Int. J. ous Grid Densities,” Theor. Comput. Fluid Dyn., 20(3), pp. 181–195.
Heat Mass Transfer, 141, pp. 48–57. [45] Li, X., and Ren, Y., 2004, “Detached-Eddy Simulation k–e Turbulence
[34] Qian, J., Chen, M., Gao, Z., and Jin, Z., 2019, “Mach Number and Energy Loss Model and Its Application in Fire Simulations,” J. Tsinghua Univ., 44(8), pp.
Analysis Inside Multi-Stage Tesla Valves for Hydrogen Decompression,” 1126–1129.
Energy, 179, pp. 647–654. [46] Zhou, P., Wang, F., and Yang, M., 2012, “Internal Flow Numerical Simulation
[35] Qian, J. Y., Chen, M. R., Liu, X. L., and Jin, Z. J., 2019, “A Numerical Investi- of Double-Suction Centrifugal Pump Using Des Model,” IOP Conf. Ser.: Earth
gation of the Flow of Nanofluids Through a Micro Tesla Valve,” J. Zhejiang Environ. Sci., 15, p. 032051.
Univ. Sci. A, 20(1), pp. 50–60. [47] Yuan, S., Si, Q., Xue, F., Yuan, J., and Zhang, J., 2011, “Numerical Calculation
[36] Lin, Z., Ma, C., Xu, H., Li, X., Cui, B., and Zhu, Z., 2017, “Numerical and of Internal Flow Induced Noise in Centrifugal Pump Volute,” Paiguan Jixie
Experimental Studies on Hydrodynamic Characteristics of Sleeve Regulating Gongcheng Xuebao/J. Drain. Irrig. Mach. Eng., 29(2), pp. 93–98.
Valves,” Flow Meas. Instrum., 53, pp. 279–285. [48] Zheng, L., Liu, G., Mao, J., Yuan, Q., Wang, S., Wei, L., and Wang, Z., 2015,
[37] Wang, Z., Wei, Y., and Qian, Y., 2020, “A Bounce Back-Immersed Boundary- “A Novel Numerical Simulation Method to Verify Turbulence Models for Pre-
Lattice Boltzmann Model for Curved Boundary,” Appl. Math. Modell., 81, dicting Flow Patterns in Control Valves,” J. Fluid Sci. Technol., 10(1),
pp. 428–440. p. JFST0007.

Journal of Fluids Engineering NOVEMBER 2020, Vol. 142 / 111202-11