EPJ Web of Conferences 299, 01014 (2024) https://doi.org/10.

1051/epjconf/202429901014
EFM 2022

Measurement of a ball valve as a control valve

Daniel Himr1* and Pavel Dokoupil2
1
Brno University of Technology, Faculty of Mechanical Engineering, Energy Institute, Viktor Kaplan Department of Fluid Engineer-ing,
Technická 2, 616 69, Brno, Czech Republic
2
Institute of Applied Mechanics Brno, Ltd., Resslova 972/3, 602 00, Brno, Czech Republic

Abstract. There are more possibilities how to control flow rate in a pipeline system. When the flow is
provided by pump, it is very economic to use frequency converter to change the pump speed which leads to
flow rate change. Another possibility is to change number of running pumps, when pump station contains
more than one unit. Control valve is used in systems without pump regulation especially, when the initial
costs has to be low. The paper is focused on measurement of control ability of a ball valve. Generally, ball
valves are not suitable to fulfil flow control function, but in an emergency case, when the control valve is
damaged and the system has to be operated, there is no other possibility, but to use the ball valve.

1 Introduction

Flow through a hydraulic system is determined by pres-

sure drop and resistance to flow. To control the flow, the
option of control the system resistance is logically offered,
which is carried out using control valves, i.e. elements
capable of changing their resistance. The control fittings
must have a dominant pressure loss compared to the rest
of the system to secure an effective operation.
Another requirement is that the pressure loss of the
system should be, if possible, linear depending on the
opening of the control valve, which is rather difficult to ful-
fil in the entire range of opening (0 to 100 %), that is why Fig. 1. Ball valve.
the manufacturers try to have the widest possible control
cannot be controlled, so they cannot be used for flow con-
band, at least.
trol. In pipeline systems, there is often a combination of
Furthermore, the cavitation behaviour of the armature
control and shut-off fittings, where the first serves to con-
is monitored [1], [2], i.e. cavitation resistance and change
trol the flow and the second ensures tightness in the case
in flow resistance due to the cavitating area. Possible
of zero flow.
pressure and mechanical vibrations resulting from flow
This article describes the measurement of the control
through the fitting are also an important factor.
capabilities of a ball valve, which is intended for a spe-
In addition to the control elements, the system also in- cific location, and which definitely does not belong to the
cludes closing/shut-off fittings (e.g. ball valve, see Fig- group of control fittings. However, the system itself has
ure 1), which are not designed to control the flow but have such pressure losses that the maximum flow requirements
good sealing properties. If these fittings are used for regu- would not be met after the installation of the control valve.
lation after all, they may show worse regulation properties The system has been in operation for over ten years and
[3], [4], higher force is needed to control them, and they now it is time to replace the ball valve with a new one. For
have worse cavitation behaviour. There is also a risk of that reason, it was necessary to measure its control abili-
significant mechanical vibrations [5], [6], [7]. ties.
The last significant group of fittings are the one-way Ball valves have minimal losses at full opening and ef-
elements, which should have minimal losses in the direc- fective control is possible only in a narrow opening range,
tion of flow, and they should not allow any reverse flow. which makes the controlled system very sensitive to any
Due to the principle of their construction, these elements change. In addition, with a small opening, there is a risk
* Corresponding author: himr@fme.vutbr.cz

© The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0
(https://creativecommons.org/licenses/by/4.0/).
EPJ Web of Conferences 299, 01014 (2024) https://doi.org/10.1051/epjconf/202429901014
EFM 2022

of cavitation operation and thus higher wear of the ball Table 1. Uncertainty, 95 % confidence level.
valve [8]. Also, higher requirements for turning forces [9]
must be expected. The paper is focused on measurement Quantity Uncertainty
of ball valve characteristics (flow coefficient, torque and
vibration), which are important when the ball valve is used Flow rate ± 15.8 m3 /h
as a control element.
Flow velocity ± 0.06 m/s

2 Description of the measurements Pressure loss ± 0.1 kPa

The behaviour of the ball valve was tested on a closed test Absolute pressure ± 1.9 kPa
circuit, in which the flow was provided by the pump and,
if necessary, controlled by the control valve. See the di- Relative pressure ± 2.3 kPa
agram of the circuit and the connection of the measuring
equipment in Figure 2.
Acceleration ± 1.9 g
The temperature on the supply pipe was measured with
a platinum thermometer (not on the diagram), absolute Position ± 0.5 %
pressure was measured at a distance of 4.8 times the di-
ameter upstream of the ball valve and at a distance of five
Current ± 0.02 A
times downstream the ball valve. Pressures were measured
in the collectors which were connected to the measured
Active power ± 0.02 W
place in the pipe by four hoses uniformly spread on the
pipe perimeter. The flow coefficient Kv was calculated
from these pressure gauges, see equation (1), where Q is
flow, ρ density, pressures pa1 and pa2 correspond to Fig-
ure 2.
ρ 3 Measurement procedure and results
r
Kv = 10 · Q · , (1)
pa1 − pa2 First, the test of functionality was made. The ball valve
The relative pressure was measured 300 mm upstream was opened and closed at the zero flow, so that the re-
and downstream the fitting, the pressure gauges were quired power has been determined. The value of torque
placed directly on the pipe wall at an angle of 30◦ from the T [Nm] was determined by calculation according to the
vertical axis. All pressure gauges were calibrated, hoses equation (2).
Pa
and collectors were vented. T= (2)
At 300 mm upstream and downstream the valve, three- 2·π·n
axis acceleration sensors (at the highest point of the diam- The measurement record is shown in Figure 3. Always at
eter) and strain sensors (deformation – strain gauge) in the the beginning of the movement there is a clear steep in-
circumferential direction (from the side) were also glued. crease of the torque. It is due to overcoming the static fric-
Deformation stress sensor – strain gauge is another way to tion. But it drops down quickly to the operational value.
monitor pressure changes inside the piping. It cannot be Static friction is mainly caused by friction in sliding bear-
used to determine the absolute value of the pressure, but ings, friction of the shaft and ball sealing surfaces. The
it allows monitoring changes in deformation and subse- valve was fully open at the beginning and it was closed
quently evaluating, for example, the frequency of pressure and opened three times. The second and third cycle was
pulsations. One three-axis acceleration sensor was also not fully closed to learn whether there was a significant
placed on the shutter drive gearbox. The degree of open- change in static friction. It was not. The torque in Fig-
ing was monitored using a potentiometer located on the ure 3 shows only the absolute value calculated from elec-
actuator. The movement speed of the valve actuator was tric power, not the direction.
then evaluated from the potentiometer record as n [rps]. During closing, the average value of the torque is
Measurements of electrical quantities (current I [A], 8690 Nm, when opening it is only 8054 Nm. The depen-
voltage U [V],power factor ϕ [-] and active power Pa [W]) dence of the flow coefficient on the opening is shown in
were measured with a wattmeter and current coils. In- Figure 4. It is obvious that the dependence is hyperbolic
dividual measured phases (currents) were derived from in the most of opening range. The measurement was made
a three-phase coupling (380 V). Wattmeter connection: in two directions (opening x closing) without significant
three-wire (three-phase) symmetrically loaded network. hysteresis. Due to the required operating value of Kv , the
The measured quantities were connected to the measuring valve will have to work between 55 % and 66 % of the
system via current outputs of 0-20 mA. opening. Hysteresis appeared when measuring the depen-
dence of power input (torque) on opening. Both quantities
The static characteristics were evaluated from the aver-
were evaluated according to equations (3) and (4)
age values for individual openings under steady-state con-
ditions for a period of 60 s. Uncertainties quantities are Pa
shown in Table 1. kP = , (3)
D3 · ∆p

2
EPJ Web of Conferences 299, 01014 (2024) https://doi.org/10.1051/epjconf/202429901014
EFM 2022

a3
Q G M E

5D 4.8D
DN300 DN300
300
BV 300
pa2 pa1

a2 p a1 p
r2 r1
S2 S1

Fig. 2. Measurement: G – gearbox, M – motor, BV – ball valve, Q – flowmeter, pa – pressure sensor (abs.), pr – pressure sensor (rel.),
S – strain gauge, a – acceleration sensor, E – measurement of electric quantities (voltage U, current I, active power Pa, power factor ϕ).

35 10
5

30
Flow coef. [m3 /h]
Torque [kNm]

25 4
10
Up
20
3 Down
15 10
10
2
5 10
Operation
0
band
10
100 0 20 40 60 80 100
Opening [%]
Opening [%]

80
60 Fig. 4. Flow coefficient.
40
20 1
10
0
0 2 4 6 8 10 12 14 16
Time [min] 0
10
k P [1/s]

Fig. 3. Torque (zero flow rate).

-1 Up
10
Down
-2
T 10
kT = , (4) 0 20 40 60 80 100
D3 · ∆p Opening [%]

where ∆p [Pa] is the pressure loss of the fitting at the given

Fig. 5. Power coefficient, solid line – with flow rate, dashed line
flow rate. The result is shown in Figures 5 and 6, where the
– zero flow rate.
solid line indicates the measurement at flow (see Figure 7)
and the dashed line the measurement at zero flow. The
hysteresis was minimal at zero flow. The pressure coeffi-
cient is almost constant up to 40 % opening and significant and its determination, is very problematic due to the small
increase is apparent from 70 % opening. pressure drops that could be achieved in the test room.
The torque required for the CS drive (conical valve) is
given by the sum of the loss torque without pressure drop 4 Frequency analysis
and loss torque, where we assume linear dependence on
pressure drop. When measuring in the test room, it was A frequency-amplitude characteristic was created from the
mainly the first one that was measured – the constant part accelerometer recording (the sharp peak at 50 Hz, which
of the loss moment, the part related to the pressure drop, corresponds to the frequency of the electric network, not

3
EPJ Web of Conferences 299, 01014 (2024) https://doi.org/10.1051/epjconf/202429901014
EFM 2022

3
10 5 Conclusion
2 The article describes the laboratory measurement of a ball
10
valve, which serves as a regulatory element. From the
k T [-]

1 point of view of CC flow rate, it appears to be a limit

10 Up
value of approx. 50 % to 40 % of the valve opening.
0 Down Below 40 % of the opening, the flow area is so small that
10
-1
the CC is at the limit of usability. Below 45 % of the
10 opening, the ball valve should not be used. If regulation
0 20 40 60 80 100
Opening [%]
with a ball valve is not moved from the worst regulation
band to the more acceptable in terms of vibrations, then,
Fig. 6. Torque coefficient, solid line – with flow rate, dashed line from the point of view of estimating the remaining service
– zero flow rate. life of the valve as a regulatory element, problems with
the overall function of this fitting can be expected in
2500
the future (flow regulation, vibration on the drive, etc.).
The measured values, in particular the dependence of
2000
Flow rate [m 3 /h]

flow rate and Kv on opening, should serve as boundary

1500 conditions in the future for specifying the replacement of
Up the existing ball valve with a new control element. The
Down
1000 current regulation with a ball valve is already outdated
500 and at present such flow regulation is abandoned, and it
would be appropriate to start preparing the replacement of
0 the existing ball valve with a new control element. The
0 20 40 60 80 100
Opening [%] measured values of vibrations and deformation are very
important for possible measurements and analyses, where
Fig. 7. Flow rate. it is necessary to regulate the flow, for example of a power
plant, where the possibility of installing pressure gauges
on pipelines is limited.
plotted). On all sensors, the dominant vibration was in the
vertical direction perpendicular to the pipe axis, therefore Acknowledgements. This paper was supported by Smart
only the recording from this direction is further discussed. Energy Systems ERA-NET project 108786, Digitalization of wa-
ter supply infrastructure to optimize the Water-Energy Nexus
First, the track without the ball valve, see Figure 8,
(DIWIEN).
there is only a record from accelerometers a1 and a2 . The
significant f requency o f 24.3 H z i s o n b oth s ensors. The
a2 sensor picked up similar vibrations at a frequency of References
around 340 Hz and smaller vibrations of 205 Hz, 247.7 Hz
and 251.7 Hz. 1. J. P. Tulis, Cavitation Guide for Control Valves (Divi-
Vibrations were evaluated for a total of 7 flow points, sion of Engineering Office of Nuclear Regulatory Re-
respectively for 7 openings of the cone valve, from a max- search U.S. Nuclear Regulatory Commission Washing-
imum opening of 100 % to a minimum of 5 %. The indi- ton, DC 20555, 1993) pp. 120.
cated frequencies show the behaviour of the conical valve, 2. Y. Wang, C. Zhu, G. Zhang, D. Wang, M. Jiang, Z. Zhu,
but also of the test track, especially the pump and the ty- Z. Lin, Processes, 8, 13 (2020)
pology of the test track. 3. H. Weihua, L. Lianjin, Valve World Magazine, 14, 3
A band of dominant frequencies of 300 Hz to 330 Hz (2013)
on the a1 sensor is evident from the measurements on the 4. B.Cui, Z. Lin, Z. Zhu, H. Wang, G. Ma, Exp. Therm.
track after the installation of the ball valve, see Figure 9. Fluid Sci., 80, 193 – 202 (2017)
The strongest vibration is recorded by the a2 accelerom- 5. H. Wu J.Y. Li, Z.X. Gao, Processes, 7, 13 (2019)
eter around the frequency of 340 Hz, with the frequency 6. I. Javanshir, N. Javanshir, R. Barmaki, M. Mahmoodi,
of 205 Hz close to 250 Hz also being amplified. From the J. Vibroeng., 17, 478 – 486 (2015)
above, it can be concluded that the vibrations in the region 7. K. Sato, Flow Induced Vibration of the Large-Sized
of 300 Hz to 370 Hz can be attributed to the tested ball Sodium Valve for MONJU, Specialists Meeting on
valve. LMFBR flow induced vibrations, 20-23 September,
The a3 sensor captured only oscillations with a fre- 1977.
quency of 24.3 Hz. The listed records for different valve
8. M.J. Chern, C.C. Wang, C.H. Ma, Exp. Therm. Fluid
openings differed only minimally. The strongest vibra-
Sci., 31, 505 – 512 (2007)
tions were at opening 60 %, which is in the middle of the
assumed regulation band. In all cases, the dominant pulsa- 9. G. Gokilakrishnan, S. Divya, R. Rajesh, V. Selvakumar,
tions were only up to 400 Hz, therefore higher frequencies IJRET, 2, 311 – 315 (2014)
are no longer plotted. Measured frequencies are too low to
capture possible cavitation.

4
EPJ Web of Conferences 299, 01014 (2024) https://doi.org/10.1051/epjconf/202429901014
EFM 2022

4
Acceleration [g]
3

2
a1 a2
1

0
0 50 100 150 200 250 300 350 400
Frequency [Hz]

Fig. 8. Amplitude-frequency characteristics without the ball valve.

4
a2
Acceleration [g]

2 a3 a1

0
0 50 100 150 200 250 300 350 400
Frequency [Hz]

Fig. 9. Amplitude-frequency characteristics with the ball valve.