Test Procedures for Determining Cavitation Limits in Control Valves

Author(s): William J. Rahmeyer

Source: Journal (American Water Works Association), Vol. 78, No. 11, Distribution System
Problems (NOVEMBER 1986), pp. 55-58
Published by: American Water Works Association
Stable URL: http://www.jstor.org/stable/41290446
Accessed: 14-03-2016 03:46 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/
info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content
in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship.
For more information about JSTOR, please contact [email protected].

American Water Works Association is collaborating with JSTOR to digitize, preserve and extend access to Journal (American
Water Works Association).

http://www.jstor.org

This content downloaded from 130.63.180.147 on Mon, 14 Mar 2016 03:46:10 UTC
All use subject to JSTOR Terms and Conditions
 Test Procedures
for Determining Cavitation Limits
in Control Valves
William J. Rahmeyer

The design and operation of control valves in a closed conduit system require testing of the and damage will occur at or just prior to
different flow characteristics of the valves under experimental conditions. Different testing choking.2-6
standards and recommended testing procedures have been published for characteristics such
Parameters
as pressure loss and pressure recovery. However, discussion concerning the testing
procedures for cavitation and operation limits is scarce. Detailed information is presented on The flow conditions that result in the
testing procedures and the definitions for the different cavitation limits. The effects of the different limits of cavitation can be
variables of pressure, size, and flow coefficients are discussed, along with design equations and mathematically represented by cavita-
data for scaling cavitation information with changes in system pressure. tion parameters or indexes. The cavita-
tion parameter a is derived from dimen-
The design and operation of control Critical cavitation13 is the level at sional analysis and is the ratio of the
valves in a closed conduit system require which cavitation effects become steady fluid forces suppressing (the system
testing of the different flow character- and increase with discharge at a slower pressure) to the forces causing (the pres-
istics of the valves under experimental rate. The effects of critical cavitation are sure drop) cavitation. The lower the
conditions. Different testing standards light vibration and noise, similar to the numerical value of the cavitation param-
and recommended testing procedures sound of bacon frying. Cavitation is very eter, the more likely or the more severe
have been published for characteristics mild, with no objectionable effects such cavitation will be.
such as pressure loss and pressure as structural damage at this level. Several forms (Eqs 1, 2, and 3) of the
recovery. However, discussion concern- Incipient damage34 is the level of cavitation parameter have been derived
ing the testing procedures for cavitation cavitation that results in detectable and used in the testing of control valves,
and operation limits is scarce. Detailed damage to either a valve or downstream and can be converted (Eq 4) to the
information is presented on testing piping. This level is difficult to detect relationship used here, a ' (Eq 1).
procedures and definitions for the dif- and damage is measured4 5 on samples of
ferent cavitation limits. The effects of soft materials such as aluminum. This a, = (Pu-Pv)/dP (1)
the variables of pressure, size, and flow level of cavitation is a good design limit
coefficients are discussed, along with because it represents the mildest and o2 = (Pd-Pv)/dP (2)
design equations and data for scaling most conservative stage of possible
cavitation information with changes in damage. K = dP/(Pu - Pv) (3)
system pressure. Choking cavitation6 is the level at
Most of the information available for which the local, mean pressure just o' = <72 + 1 = 1 /K (4)
the design and testing of control valves downstream of the valve drops to vapor
concerns pressure loss and pressure pressure and a choking of the flow The upstream pressure Pu is the static
recovery characteristics. Little has been occurs. At the point of choking, down- wall pressure measured upstream of a
published, however, about sizing valves stream control or downstream flow valve at a minimum distance equal to
for cavitation or about determining modifications will not increase the flow one inside pipe diameter. The down-
different cavitation limits. rate through the valve. This level repre- stream pressure Pd is the static wall
sents the most severe cavitation because pressure measured at least 10 pipe
Background the maximum effects of noise, vibration, diameters downstream of a valve after
Cavitation is a liquid phenomenon
similar to the boiling of a liquid, except TABLE 1
that it results from a sudden drop in Pressure scale effects for control valves
pressure. Vapor pockets or cavities are
produced by low-pressure sources such Valve Cavitation Limit N
as eddies or vortexes in separation zones. Full ball Incipient 0.39
In higher pressure regions the cavities Critical 0.39

collapse, or cavitate, and produce intense Incipient Damage 0.44

Choking 0.50
energy that results in noise, vibration, Segmented ball Incipient 0.30
and sometimes structural damage. Critical 0.30
The effects of cavitation can be cate- Incipient Damage *
gorized and defined according to several Choking 0.50
Butterfly Incipient 0.39
levels (Figure 1). The first level is incip- Critical 0.39
ient cavitation.1 2 This level comprises a Incipient Damage 0.46
set of flow conditions by which cavitation Choking 0.50
can first be detected, either with struc- Globe (S type) Incipient 0.46
Critical 0.46
tural or audible noise measurements.
Incipient Damage 0.31
Incipient cavitation is intermittent and Choking 0.50
produces a very mild effect of noise and *Not available
vibration, without structural damage.

NOVEMBER 1986 WILLIAM J. RAHMEYER 55

This content downloaded from 130.63.180.147 on Mon, 14 Mar 2016 03:46:10 UTC
All use subject to JSTOR Terms and Conditions
the flow has returned to a fully developed 450 r -I 8000

pipe flow. The downstream pressure

must be corrected or extrapolated to a Choking

uniform flow pressure just downstream

of a valve by subtracting the line loss
resulting from the pipe wall friction 400 - J 7000
between a valve and the downstream
pressure taps. Some testing standards7-8
do not account for friction loss and can
result in errors of 30-40 percent in
/ //C
determining the pressure drop and flow
coefficients. The pressure drop (dP= Pu 350 - j Jf ~ 6000
- Pd) will then represent the actual
pressure loss or drop in energy that
occurs at the valve. is the vapor / // incipient
pressure of the liquid and is a function of / // damage
300 - / / ly ~ 5000
the atmospheric pressure and the tem- / // Critical
perature of the liquid. The variables in
the cavitation parameters are such that
the parameters are dimensionless.
The cavitation parameter is a function
of valve geometry, pressure drop, system J 250 - 7 / / / - 4000 |°
y / / y Incipient
pressure, and the size of the test valve
and system. It is not completely inde-
pendent or constant for different system
pressures and sizes. The degree of
cavitation can increase for a constant 200 ~ / / / y - 3000
cavitation parameter as the pressure or
the size of the system is increased.
Scaling equations and exponents have
been derived to correct or extrapolate
the conditions of the cavitation limits 150 - I - 2000
from one system pressure to another.
The pressure scale effect (PSE) and the
scaling equations 5-7 have been experi-
mentally determined for scaling param-
eters to different system pressures. 100 - / / X/ - 1000

PSE =[(Pu-Pv)/{Pu-Pv)r'N (5)

Q = Qr x PSE (6)
o

o 20 40 60 80 100
a = orf(Pu-Pv)/(Pu-Pv)r'l-2N (7) Valve Opening- degrees

Figure 1. Cavitation limits for a 6-in. (150-mm) butterfly valve (Pu - Pv = 85 psi
In these equations Q is the discharge and
r refers to a known or reference set of [586 kPa])
Discharge- gpm
cavitation data. The exponent N corre-
1000 1400 1800 2200 2600 3000
sponds to the degree of pressure scale
effect. A value of xh for N indicates that Choking
there is no scale effect. Table 1 contains
exponents for some of the types of valves
that have been tested.
The cavitation parameter of a valve - Q/ ' I
will vary with the geometry or opening - - jX ' - 5 |
of a valve. The flow coefficient is a
^ - Critical *■ q*
parameter that can be used to define the
geometry of the valve. A commonly used
form of the flow coefficient is Cv.s

cv = £gpm x (Gf/dP)Vl (8)

1 - jo '
in which the discharge Q is in gallons
per minute, the pressure drop dP is in b - 0.2
pounds per square inch, and G/ is the
specific gravity of the liquid. One problem
Incipient ~ 0.1
with using Cv as a flow coefficient is that
it is not dimensionless and is not in- L

50 75 100 125 150 175

dependent of valve size. When comparing
Cv for different valve sizes, Cv must be Discharge- L's

divided by the diameter squared, and the Figure 2. Incipient and criticial cavitation for a 6-in. (150-mm) gate valve
units of the diameter are usually in
56 RESEARCH AND TECHNOLOGY JOURNAL AWWA

This content downloaded from 130.63.180.147 on Mon, 14 Mar 2016 03:46:10 UTC
All use subject to JSTOR Terms and Conditions
 inches. (Cv does not have a direct metric
equivalent.)
Discharge- gpm
1490 1500 1510 1520 1530 1540 1550 1560 1570
Another flow coefficient that is used
15 0 and derived from the Euler number is Cd
(Eq 9), which is dimensionless and inde-
10.0 -
pendent of size.

80-
cd = Q{2gdH)~Vi/Ap (9)
6.0 -
in which the flow Q , the pipe area Ap> the
o-
pressure drop dH' and the gravitation
' 4.0 " / constant are in units such that Cd is
E/
dimensionless. The coefficient Cv can be
S. y'
converted to Cd by Eq 10.
H

o> / Cd = (890.72 Cv*/d2yVi (10)

E 2.0 ~ /
<o S
O <o /
in which the pipe diameter d is in inches
(millimetres).

Experimental procedures
1.0

The experimental procedures for de-

08 - y/ termining cavitation limits at a given
valve opening and flow coefficient are
based on tests conducted at constant
0.6'

92 93 94 95 96 97 98 99 100
Discharge- L/s
upstream pressures. Only the pressure
Figure 3. Incipient cavitation damage for a 12-in. (300-mm) ball valve drop and discharge are varied during the
tests. To control both upstream pressure
AP ' - - psi and discharge, control valves must be
well upstream and downstream of the
0 2 4 6 8 10 12
test valve. To allow a wide range of
/
downstream pressures and to test for
Choking / choking, the downstream test piping
300 "
if open should have very few sources of pressure
losses such as flowmeters and' elbows.
The downstream piping should be
straight and, if possible, expand to a
250 - Kc L - 4500 larger size with the downstream control
valve being larger than the test valve.
Incipient and critical cavitation limits.
J
Usually both the incipient and the
j
critical cavitation limits are determined
225 ~ / // - 4000 during the same set of test runs. A series
of tests consists of different discharges
at the same upstream pressure and valve
opening, starting at flow conditions not
200 - / - 3500 causing cavitation, and then progressing
through the different levels of cavitation.
The discharge is increased in small
o7f3 increments, allowing the flow to be
stabilized, until cavitation can be de-
175 - II - 3000
tected. Detection can be by sound or
vibration measurement, with careful
consideration given to the difference
/J^
/T^ between flow noise and vibration noise.
150 - If ~ 2^00 The flow conditions corresponding to
Critical
cavitation j m the first detection of cavitation define
the incipient cavitation limit.
After incipient cavitation, the flow is
increased in small increments, and the
magnitude of cavitation noise and vibra-
125 -

tion will appear unsteady and will in-

crease at a rapid rate. At the limit of
critical cavitation, the noise and vibra-
100
tion will begin to increase at a slower
30 40 50 60 70 80 rate. Sometimes it is possible to distin-
^p.Vj-kPa guish critical cavitation by aural mea-
surement. Cavitation at the critical limit
Figure 4. Kc and choking cavitation for an 8-in. (200-mm) globe valve (Pu - Pv = 85 is steady and very mild with light noise
psi [586 kPaJ) and vibration.
Incipient and critical cavitation can be

NOVEMBER 1986 WILLIAM J. RAHMEYER 57

This content downloaded from 130.63.180.147 on Mon, 14 Mar 2016 03:46:10 UTC
All use subject to JSTOR Terms and Conditions
determined graphically by plotting the ficient. It can be graphically represented coefficients. FLt Eq 11, is less compli-
log output of either noise or vibration as in Figure 4 where the discharge is cated to use during testing because the
versus discharge or the cavitation pa- plotted versus the square root of the pressure drop is not used.
rameter. To measure structural vibra- pressure drop. The slope of the straight o1 at choking, Km, and FL are all
tion from cavitation, an accelerometer is portion of the curve is actually the dimensionless parameters and are inde-
attached to the downstream connection coefficient Cv. Kc occurs where the slope pendent of both system pressure and
or piping of the test valve. The output of of the plot ( Cv ) begins to decrease. Some size. Choking data for valves do not have
the accelerometer is read as an accelera- publications imply that Kc is a design to be taken at constant pressures. Be-
tion in inches per second squared (milli- limit for cavitation. For many types of cause the exponent N of Eqs 6 and 7 is
metres per second squared). Figure 2 valves, however, cavitation damage, equal to xh for choking cavitation, the
shows the changes in slope that corres- heavy vibration, and noise occur at pressure scaling equations do not have
pond to incipient and critical cavitation. pressure drops less than that of Kc . to be considered.
For most valves, the difference in a for Choking cavitation. Choking cavitation
is determined experimentally by main- Summary
incipient and critical cavitation is usually
from 10 to 20 percent. taining a high upstream pressure and The methods for determining cavita-
Incipient damage limit. The procedures increasing the discharge by opening the tion limits represent more than 15 years
used to determine incipient damage are downstream control valve to reduce the of testing different types, styles, and
based on the original work by R.T. downstream pressure. Choking occurs three sizes of control valves. The equa-
Knapp to define incipient damage.5 when the Upstream pressure and dis- tions and procedures have been verified
Knapp defined the limit of incipient charge are not affected by further open- by comparisons of test data with field
damage as the flow conditions corres- ing the downstream control valve. data.
ponding to a measured damage rate of 1 Choking is difficult to test for because
References
cavitation pit per minute per square of the high pressures and discharges
inch (square millimetre) in soft alumi- required to produce vapor pressure 1. Ball, J.W. & Tullis, J.P. Cavitation in
num. A cavitation pit is defined as a downstream of a test valve. The down- Butterfly Waives. Jour. Hydraulics Div. -
small pit or crater, formed plastically in stream piping can complicate the testing ASCE, 99:HY9:1303 (Sept. 1973).
the surface of a metal as a result of the 2. Rahmeyer, WJ. Cavitation Noise From
by causing additional pressure losses
Butterfly Valves. Nuclear Engineering
single collapse of a vapor cavity. Later that limit the maximum discharge.
and Design 72, North-Holland Publishing
studies49 verified that this definition is Choking cavitation can be verified by Co. (July 1982).
sufficient because (1) at lesser dis- several methods during the tests. If the 3. Ball J.W. et al. Predicting Cavitation in
charges, damage will occur at an ex- discharges are sufficient at choking, the Sudden Enlargements. Jour. Hydraulics
tremely low rate and (2) at slightly vapor cavities will flash and extend well Div.-ASCE, 101 :HY7:857 (July 1974).
higher flows, damage will occur at a very downstream of the valve. Usually, when 4. Stripling, T.E. Cavitation Damage Scale
high rate to almost all types of materials. the flashing condition begins, the loca- Effects. Doctoral dissertation, Dept. of
Prototype tests29 verify Knapp's con- tion of the cavitation noise and vibration Civil Engineering, Colorado State Univ.,
Fort Collins (1974).
clusion that the damage rate will increase moves downstream and the magnitude
5. Knapp, R.T. et al. Cavitation. Engineer-
to the seventh power with an increase in of noise and vibration appears to decrease
ing Societies Monographs, McGraw-Hill
discharge. (Figure 2). Another indication of choking Book Co., New York (1970).
The test procedure for determining is the presence of vapor and vapor pres- 6. Tullis, J.P. Choking and Supercavitating
incipient damage involves testing a valve sure at the downstream pressure taps. If Valves. Jour. Hydraulics Div. -ASCE,
at long durations and at high discharges the downstream pressure is not at vapor 97:HY12:1931 (Dec. 1971).
to determine the location of all possible pressure, choking can be verified by 7. Instrument Society of America. Control
damage. Damage may occur at several closing the downstream control valve to Valve Sizing Equations. ANSI/ISA S75.01
different places in the valve as well as in increase the downstream pressure by Standard. Research Triangle Park, N.C.
(1981).
the downstream piping. To show the several percent. If the discharge remains
constant with the increase in down- 8. Instrument Society of America. Control
locations of damage, valves and piping Valve Capacity Test Procedures. ISA-
can be covered with thin layers of mate- stream pressure, the valve is choking. S75.02 Standard. Research Triangle
rial that will be easily damaged by The procedures for choking cavitation Park, N.C. (1981).
cavitation but are resistant to high liquid are the same as for determining the 9. Rahmeyer, W.J. Prediction and Modeling
velocities. pressure recovery factor, FL,58 in Eq 11. of Cavitation Damage. Doctoral disserta-
After determining the possible loca- tion, Dept. of Civil Engineering, Colorado
tions for damage, sections where damage Fl = Qgpm(Pu-PvT^Cv (11) State Univ., Fort Collins (1980).
will occur are replaced with soft alu-
minum. Test runs are made at constant in which the discharge is in gallons per About the author:
upstream pressures and valve openings minute (litres per second), the pressures William J. Rahmeyer
and at different discharges, starting with are in pounds per square inch, gauge is an assistant profes-
conditions that will not cause damage. (kilo Pascals), and the coefficient Cv is sor of civil and environ-
After each test run at a slightly higher averaged from test data that was not mental engineering at
discharge, the aluminum is inspected for influenced by cavitation. Utah State University,
damage. The discharge corresponding to Km is another parameter that is used Utah Water Research
a damage rate of 1 pit/min/sq in. can to express the conditions of choking Laboratory, UMC
graphically be extrapolated by plotting cavitation. It is the inverse of o' as 8200, Logan, UT
the log of the pitting rate versus dis- shown in Eq 12. 84322. He received his PhD in 1980 in
charge (Figure 3). Then alD is calculated the field of hydraulics and fluid dynamics
for the flow conditions at incipient Km = l/a, = Fl* (12) of closed conduit flow and has been active
cavitation damage. in consulting and research in the area of
Kc. Another parameter that has been Since it is not possible to accurately valve testing since 1973. Rahmeyer is a
used is Kc. It is defined as the set of measure the pressure drop that occurs member of standards committees for valve
conditions under which cavitation will at choking cavitation, it is necessary to testing in both A WW A and /Sj4 and an
begin to affect the flow pattern in the calculate the pressure drop for Eqs 1 and author of other publications concerning
valve and begin to lower the flow coef- 12 from known discharges and flow cavitation and valves.

58 RESEARCH AND TECHNOLOGY JOURNAL AWWA

This content downloaded from 130.63.180.147 on Mon, 14 Mar 2016 03:46:10 UTC
All use subject to JSTOR Terms and Conditions