See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/222080044

Practical solution to the problem of noise and

vibration in a pressure reducing valve

Article in Experimental Thermal and Fluid Science (EXP THERM FLUID SCI) · January 1995
DOI: 10.1016/0894-1777(94)00074-I

CITATIONS READS

14 673

2 authors:

Amir Amini Ieuan Owen

Spirax Sarco,UK University of Liverpool
9 PUBLICATIONS 47 CITATIONS 104 PUBLICATIONS 866 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Horizontal Axis Tidal Stream Turbines Subject to Wave-Current Interaction View project

The Effect of Ship Aerodynamics on Martime Aircraft Operations View project

All content following this page was uploaded by Ieuan Owen on 31 October 2017.

The user has requested enhancement of the downloaded file.

 ELSEVIER

A Practical Solution to the Problem of Noise

and Vibration in a Pressure-Reducing Valve
A. Amini -.The mechanical vibration that is occasionally found in gas pressure
I. Owen reducing valves can be eliminated by careful design of the valve plug and
University of Liverpool, seat. A pressure-reducing valve was found to be excessively noisy, produc-
Department of Engineering, ing a sound pressure level of 117 dB when throttling air at an inlet-to-out-
Liverpool, United Kingdom let pressure ratio of 15; as a result the valve suffered wear and vibration
damage. By changing the design of the original flat plug and seat, the
problem was significantly reduced. A 60° conical plug and seat produced a
noise reduction of 12 dB (a factor of 4 in actual sound pressure level), the
mechanical vibration was eliminated, and the flow capacity was increased
by about 25%.

Keywords: pressure-reducing valve, vibration, nobe

INTRODUCTION coupled. Figure 1 shows a pilot-operated pressure-reduc-

ing valve in which the area of interest is the main valve
Control valves handling compressible fluids can generate plug and seat. The valve plug is essentially a flat disk that
unpleasant noise, particularly when exposed to high pres- sits on a flat, raised annular face, the two being lapped
sure differentials. The problem arises from the jets of together. Of concern in this valve were the wear on the
fast-moving gas on the downstream side of the pressure- push rod in its guide and the damage to the valve seat and
reducing valve, which mix with slower moving gas, causing plug.
the loss of kinetic energy and hence a pressure drop Analytically, the prediction of aerodynamic noise ema-
across the valve. At certain critical conditions, the conver- nating from the control valve is based on two principles
sion of energy into noise can sometimes reach about [2]: first, the noise created by the control valve during the
150 dB. throttling process, which is a function of the mass flow and
In stating the problem it needs to be recognized that fluid Mach number across the control valve restriction;
there are two categories of noise. The first and most second, the transmission of the sound into the down-
c o m m o n is aerodynamic noise resulting from the Reynolds stream pipe and then radiation of the sound by the pipe
stresses or shear forces created in the flow stream as a exterior into the surrounding environment. Formulations
result of rapid deceleration, expansion, or impingement. predicting the level of control valve noise have been
The second category of noise, and the one that is of proposed [4, 5].
particular concern in this paper, is the mechanical noise The mechanical vibration in a pressure-reducing valve
and vibration that can be stimulated in the moving parts was investigated by Nakano et al. [6], who showed how the
of the valve by the fluid-dynamic pressure fluctuations. flow separating from the valve plug can be unstable and
This vibration, as well as being aurally unpleasant, can oscillatory in nature. Various seat and plug geometries
very quickly lead to damage of the valve stem and seat [1]. and various flow regimes were investigated. An important
The source of these two categories of noise is usually the conclusion from the study was that the level of flow
same, with the principal area of noise generation being instability was dependent upon the geometry of the flow
the recovery region immediately downstream of the vena passage between the plug and the seat. This conclusion is
contracta formed when the flow discharges through the widely accepted and is the basis for the design of "quiet
gap between the valve plug and seat and where the flow valves" such as those described by Seebold [7]. The princi-
field is characterized by intense turbulence and mixing. ple of these designs is not to throttle the flow through a
When there is a large pressure drop across the valve, the sudden expansion as in a flat plug and seat design but to
noise is also affected by the choked-flow shock formation expand it through a number of tortuous passages so that
in the gap between the valve plug and seat [2, 3]. There- pressure loss occurs not only by turbulent dissipation but
fore, although the main concern of the present paper is also by viscous effects. The narrow flow passages normally
the potentially harmful mechanical vibration in a pres- associated with these designs, however, lead to blockages
sure-reducing valve, the two types of noise are directly and to severely reduced flow capacity.

Address correspondence to Dr. A. Amini, Department of Mechanical Engineering, University of Liverpool, P.O. Box 147, Liverpool L69, 3BX,
United Kingdom.
Experimental Thermaland Fluid Science 1995; 10:136 141
© Elsevier Science Inc., 1995 0894-1777/95/$9.50
655 Avenue of the Americas, New York, NY 10010 SSDI 0894-1777(94)00074-I
 Noise and Vibration in Pressure-Reducing Valve 137

valve. A pressure-reducing valve on a compressed air

system, for example, will normally have a limited pressure
reduction because the air is usually required at moder-
ately high pressures for use in actuators, etc. In this case
the associated noise, although it may be unpleasant, is
limited and can be reduced by judicious use of silencers
and insulation [4, 7]. Steam pressure reducing valves, on
Flow d i r e c U o n the other hand, can be required to reduce the steam from
, > mains pressure to atmospheric pressure, or even below,
depending on the temperature required by the process.
(The temperature of saturated steam is governed by its
pressure, and for processes that require heating at 100°C,
for example, the steam might be throttled from a mains
pressure of 10 bar to a process pressure of 1 bar.) Because
of these high pressure ratios, mechanical vibrations are
more prevalent in steam pressure reducing valves. Fur-
thermore, to obtain such a large pressure ratio, the gap
between the valve plug and seat will be small; any vibra-
tion in the valve will therefore cause the plug to "rattle"
on the seat and will lead to damage.
The study reported in this paper was carried out be-
Figure 1. Pressure-reducing valve. cause under certain operating conditions it was found that
the steam pressure reducing valve shown in Fig. 1 suffered
Pressure-reducing valves handling gas flows will suffer from mechanical vibration and wear. The critical operat-
from aerodynamic noise. To a large extent, the level of ing conditions were mainly dictated by the inlet and outlet
noise, and whether or not it will stimulate mechanical pressures. The study focused on modifying the design of
vibration, depends upon the pressure reduction across the the valve plug and seat to provide a valve with reduced

60"
28
I_ -I
18 -
J= 28 -I ~ I

30 A/F HEX_I
I_
50 A/F HEX 50 A/F HEX
I_

" 1
@ _
i? 28x15.6114 =1
16

--I
Stondord Cone Step
Figure 2. Some designs of valve plugs and seats.
138 A. Amini and I. Owen
C
Spectrum Analyser Frequency Analyser

\ \

o
24KhzJ
0 0
1
g 0
0 0 0
z
@ o o
K Accelerometers

o
I o
•
o!m
• I~
I000 m m
_I

o Charge plifi I -

R I Air out
Condenser Microphon

Valve
Figure 3. Schematic diagram of the sound pressure level and vibration m e a s u r e m e n t rig.

S t e a m in
4)

Pd
~I9Pu
Valve ~.~ealel
{5 Weight t a n k
Limit ) / ,
switeh~
0 0
===o 0 0
Control switch
C o u n t e r Timer

Figure 4. Schematic diagram of the steam flow rate measurement rig.

 Noise and Vibration in Pressure-Reducing Valve 139

noise and no mechanical vibration but at the same time 120 I I I I I I o I

having no significant reduction in its flow capacity. For

ease of experimentation, the vibration tests were carried
out using air, with the results being finally confirmed using
steam. Flow capacity tests were also carried out using
115
steam.
m
E X P E R I M E N T A L APPARATUS AND "o

PROCEDURE
g11o
m
The four basic designs of valve seats and plugs are shown
in Fig. 2. The standard design is a fiat disk mating with a =9
ca
fiat thin-annulus seat. The cone design is a matching ca

conical plug and conical seat; the design shown is for a 60°
o. 105
angle; other angles between 15 and 75 ° were also tested.
'~,.- Cone angle
The stepped design is one that was recommended by
Hutchinson [1]; this was included in the tests to provide a o
¢0) \ X'II -- 15 degree
comparison as it employs the "tortuous-path" principle of V/ -4-30 degree
"quiet valves." A number of variations of this stepped
1 O0 "~ -x- 45 degree
design were also tested, but the results are not reported
here because the design is not a good one, as will be -"- 60 degree
shown later, and the inclusion of the additional results -x- 75 degree
would only confuse the issue. -~- standard
95 I I I I I t i I
Figure 3 shows a schematic diagram of the equipment 0 2 4 6 8 10 12 14 16 18 20
used to measure the noise and vibration of the valve
working with high-pressure air. The upstream pressure to P r e s s u r e ratlo, P . / P ,
the pressure-reducing valve was varied from 13.6 barg to
1.7 barg at intervals of 1.7 barg. For each valve seat, the Figure 5. Noise levels for the standard and conical plug and
desired setting to generate sound was achieved by turning seat designs.
the adjustment screw (to obtain critical condition; see Fig.
1). Sound pressure level measurements were obtained by
use of a condenser microphone coupled to a frequency RESULTS AND DISCUSSION
analyzer. The microphone was placed on the centerline of
the outlet at a distance of 1 m downstream of the pres- The results of sound pressure level measurements for the
sure-reducing valve. An accelerometer was placed at the conical valve seats are shown in Fig. 5. This figure shows
top of the pilot diaphragm to record the mechanical how the valve generally produces the most noise (a combi-
vibration in the valve. The output of the vibration spec- nation of aerodynamic noise and mechanical vibration
trum was recorded on a transient recorder. The capacity noise, where it occurred) for pressure ratios in excess of 8.
tests, which were carried out for the standard plug and The reduction in noise between pressure ratios of 4 and 8
seat and for two of the valve plugs and seats that gave the is believed to be due to a change in the flow regime within
lowest sound pressure level measurements, were made the valve. Nakano et al. [6] showed similar effects and
using the steam flow measurement rig shown in Fig. 4. In related them to the flow regime using schlieren photogra-
these tests the upstream pressure, Pu, was kept constant phy. It can be seen that all the conical designs produced
at 7 barg while the downstream pressure, Pd, was in- lower noise than the standard design, and, referring to
creased from 1 barg to 5 barg. Fig. 6, it can be seen that the 60° seat appears to be the
The uncertainty in experimental results was calculated optimum. The reduction in sound pressure level is about
in accordance with [8] and reflects the reliability of the 12 dB at an upstream pressure of 14 bar. It should be
instrumentation and the accuracy with which it could be pointed out that this is a significant reduction, since a
read. The uncertainty was determined using Student's t change of 12 dB, being on a logarithmic scale, represents
distribution at the 95% confidence level. In most cases a reduction in the sound pressure level by a factor of 4.
only one value of the uncertainty in the variable was The reduced noise level of 104 dB is the measurement of
calculated, which represents the total uncertainty account- the jet noise exhausting into the atmosphere. This is
ing for readability, unsteadiness, instrument calibration, obviously still very high, but in practice it should be borne
and any estimated fixed errors. All the error contributing in mind that the valve will be exhausting into a pipeline,
to the uncertainty in the variable was identified and quan- not to the atmosphere. How the sound pressure level of
tified and then combined to obtain the uncertainty in the the valve fitted with the 60° cone plug and seat compares
final results. The range of experimental error in the basic with the stepped and standard designs is shown in Fig. 7.
data in the pressure measurement is +__0.5% of the read- The standard design is clearly the noisiest, with the stepped
ing, the mass flow rate error was calculated to be ___1%, and conical designs being very similar at pressure ratios in
the error in sound pressure level measurement is about excess of 8.
+ 2%, and the error in frequency and amplitude is + 1.5%. The sound pressure level measurements discussed above
The uncertainty in seat angle is due to uncertainty of are very illustrative. However, they do not distinguish
measuring equipment and is about 0.5%. between aerodynamic noise and mechanical vibration, al-
140 A. Amini and I. Owen

120

-• 7 T T T
U p s t r e a m pressure
13.6 barg - I - 1 0 . 2 barg -x- 6.8 bar
v U 70

60
Standard
/

/ 60 deg.cone
/
m 50 170 HZ / iI step
~: 115
,/ I

_o II1
40 l; 712 HZ
.o
"0
W
W /
t 500 HZ 884 HZ
n ~'30
c
~)110
(n
/ ii
20
/ /
' ,L
10

105 ~ - - ± L L J
0 15 30 45 60 75 90 105 120 0 200 400 600 800 1,000
Seat angle, d e g r e e s F r e q u e n c y , Hz

Figure 6. Noise levels from different angles of conical plug Figure 8. Mechanical vibration characteristics (Pu/Pd = 14).
and seat designs.

though an increase in the total noise in the standard valve

design was usually associated with the onset of mechanical
vibration. An accelerometer was placed at the top of the
120 I , ~ ; = , , pilot diaphragm to measure directly the mechanical vibra-
tion. The valve pressure ratio was set to about 14, a
position at which the greatest noise and vibration oc-
curred, and the vibration was recorded on a transient
recorder. The results of this are shown in Fig. 8.
115 The amplitude of vibration for the standard design is
very much higher than for the stepped and cone designs,
m with a number of resonant frequencies being found. What
'ID
this shows, and indeed what was easily discernible during
Q the tests, is that the stepped and conical seats eliminated
g110
the mechanical vibration. The levels of vibration ampli-
.o tude shown for these designs correspond to fluid-dynamic
vibrations and not mechanical ones.
¢~
Q - 1 t The final tests carried out with the valve were to mea-
o. 105 sure its flow capacity when fitted with the different designs
"o
f- of plug and seat. The results of these are shown in Fig. 9.
o As expected, the capacity of the stepped design, due to its
tortuous flow path, is significantly lower than that of the
standard design. What was not expected, however, was the
100 result that the flow capacity with the 60 ° design is signifi-
Seat & plug design cantly better than that of the standard design. This is
Cone(60 deg) believed to be due to the flow passage being more suitable
-~- 6 M e p s for pressure recovery than that of the standard design.
Standard Thus for a given pressure drop there can be a higher flow
95 I I I I I t I I rate, since for a given valve lift the geometry of the plug is
0 2 4 6 8 10 12 14 16 18 20 such that the resultant flow areas are similar.
Pressure r a t i o , P . / P d
It is clear, therefore, that the cone design of plug and
seat offers a solution to the problem of mechanical vibra-
Figure 7. Comparison of noise levels for different seat and tion without incurring any penalties regarding flow rate
plug designs. normally associated with "quiet valves." The laboratory
 Noise and Vibration in Pressure-Reducing Valve 141

300 I I I I I Clearly, one has to be careful in applying general conclu-

Standard ÷ 60 degree cone -x- 6 steps sions.

PRACTICAL USEFULNESS / SIGNIFICANCE

250
Pressure-reducing valves are a c o m m o n feature of all
high-pressure gas systems. These valves all have the capa-
bility of producing undesirable and excessive noise and
200 vibration. In a steam system the pressure of the steam is
selected to provide saturation t e m p e r a t u r e with the pro-
(3rj
cess requirements. Therefore pressure-reducing valves
¢- throttling steam often o p e r a t e over a much greater pres-
150 sure ratio than in the gas systems and consequently can
suffer from greater noise and vibration problems. This
0
U_ p a p e r has shown that by careful design of the valve seat
and plug the acoustic noise can be considerably reduced
100 and the vibration can be eliminated. Standard valves that
suffer from noise and vibration can have their seat and
plug replaced, in service, by the conical design recom-
m e n d e d in this paper, thus saving time and expense for
50
the operator.

CONCLUSION

0 1 2 3 4 5 6 It has been shown how the mechanical vibration in a

pressure-reducing valve can be eliminated by changing the
Downstream pressure, barg design of the valve plug and seat. F o r the pressure-reduc-
ing valve used in the present work, a 60 ° conical plug and
Figure 9. Pressure-flow characteristics of pressure-reducing
seat was found to be optimal. The sound pressure level
valve with various plug and seat designs with steam flow.
was reduced by 12 dB (a factor of 4), mechanical vibration
was eliminated, and flow capacity was increased by about
results were confirmed by modifying a problematic valve 25%.
on an industrial plant. Notwithstanding these results, how-
ever, it is possible that different operating conditions may REFERENCES
require different valve designs to provide an acceptable
solution. The results shown in this p a p e r were for the 1. Hutchinson, J. W., ISA Handbook of Control Valves, 2nd ed.,
valve venting to atmospheric pressure; when the pressure Pittsburgh, 1976.
downstream of the valve was increased, it was found for 2. Reed, C. L., Noise Created by Control Valves in Compressible
the standard seat design that the onset of mechanical Service, Third Control Valve Symposium, ISA, 79-83, 1977.
vibration was slightly different and did not correspond to a 3. Reethoff, G., Turbulence Generated Noise in Pipe Flow, Ann.
Reu. Fluid Mech. 10, 333-367, 1978.
particular pressure ratio or pressure differential. U n d e r
4. Schuder, C. B., Control Valve Noise--Prediction and Abate-
laboratory conditions, the 60 ° conical valve was quiet
ment, in Noise and Vibration Control Engineering, M. J. Ceroker,
c o m p a r e d with the other designs, but it is possible that Ed., 90 94, Purdue Univ., West Lafayette, IN, 1972.
under conditions that could not be p r o d u c e d in the labo- 5. Jenvey, P. L., Gas Pressure Reducing Valve Noise, J. Sound
ratory the vibration p r o b l e m may reappear. It is believed, Vibration 41(4), 506-509, 1975.
however, that the solution will still be found by modifying 6. Nakano, M., Outa, E., and Tajima, K., Noise and Vibration
the design of the valve plug and seat. Related to the Pattern of Supersonic Annular Flow in a Pressure
A n o t h e r consideration that should be borne in mind is Reducing Gas Valve, J. Fluids Eng., Trans. ASME 110, 55-61,
the direction of flow through the valve. In Fig. 1 it can be 1988.
seen that the flow is tending to push the plug onto the 7. Seebold, J. G., Control Valve Noise, Noise Control Eng. J. 24(1),
seat against the action of the push rod. F o r this design it 6-12, 1985.
has been shown that a conical design of plug and seat is 8. Leaver, R. H., and Thomas, T. R., Analysis and Presentation of
better than a fiat one. N a k a n o et al. [6] tested a valve in Experimental Results, Macmillan, London, 1974.
which the flow was in the opposite direction, that is,
tending to push the plug off the seat. In their case they
suggested that the fiat geometry is b e t t e r than the conical. Received November 25, 1993; revised July 29, 1994

View publication stats