1

APPLICATION OF BUTTERFLY VALVES FOR FREE DISCHARGE,

MINIMUM PRESSURE DROP, AND FOR CHOKING CAVITATION
Butterfly valves are commonly used as control valves in applications where the pressure
drops required of the valves are relatively low. Butterfly valves can be used in applications as
either shutoff valves (on/off service) or as throttling valves (for flow or pressure control). As
shutoff valves, butterfly valves offer excellent performance within the range of their pressure
rating. Typical uses would include isolation of equipment, fill/drain systems, bypass systems,
and other like applications where the only criteria for control of the flow/pressure is that it be on
or off. Although butterfly valves have only a limited ability to control pressure or flow, they
have been widely used as control valves because of the economics involved. The control
capabilities of a butterfly valve can also be significantly improved by coupling it with an operator
and electronic control package.
Butterfly valves allow high flows with relatively low pressure loss from the valves, and
are typically used for flow control for valve openings from 30 to 70 degrees of full open. At
valve openings greater than 70 degrees, the pressure loss of a butterfly valve is too low to
produce any significant effect on flow or the energy loss of a flow system. Two special
applications for a butterfly valve include the use of a valve for free discharge and the use of a
butterfly valve for flashing or choking cavitation service. Free discharge typically produces high
pipe velocities at moderate pressure drops, and choking cavitation typically produces high
velocities with large pressure drops. The following paper also includes a review of flow
coefficients and the application of non-line sized valves, or valves installed in larger diameter
pipes with the aid of pipe reducers.
Butterfly Valves Types
The butterfly valves used in water works applications vary from traditional resilient
seated butterfly valves to the more specialized AWWA valves as detailed in AWWAs C504
standard. Regulations do not generally require the use of AWWA valves in water works service
but most operators will specify AWWA valves for the majority of applications based on their
experience with AWWA valves and their reputation for long service life.
Basic resilient seated valves are designed in accordance with industry standards such as
MSS SP-67 and most frequently come in lugged and wafer configurations. They are considered
by most to be a general service type of valve and as such, are usually restricted to use in the
ancillary systems within water treatment and distribution systems. They are usually limited in
size to around 48 inches in diameter. and pressure ratings for these valves are typically 75, 150,
or 300 psi. The design of the valves with different ratings are very similar. The stem diameter,
disc thickness, and disc to seat interference are the main features which differ as the rating
changes. Flow characteristics and limits are not generally defined in the standards/specifications
nor are they defined by manufacturers in most cases.
2
AWWA valves are considered by most users to be a more robust valve design and in
most cases have greater body wall thicknesses, larger shaft diameters, and thicker discs than
resilient seated valves with equivalent pressure ratings. In addition, AWWA valves would most
frequently be supplied in a double flanged configuration rather than the lugged or wafer styles
although all of these configurations are available in the marketplace. Currently the AWWA
C504 standard defines three shutoff pressure ratings and two flow velocity sub-ratings. Pressure
rating breaks are made at 25, 75, and 150 psi. Flow velocities are defined for the AWWA valves
as specific breaks are made in the line velocities at 8 and 16 feet per second. An increasing
number of applications however, are pushing beyond these traditional pressure/flow velocity
limits with applications at 250 and 300 psi becoming common.
A third category of butterfly valve, the high performance butterfly valve, is occasionally
utilized in water treatment and distribution applications. These valves are characterized by
higher pressure and/or temperature ratings, different disc seating/sealing technology, and
premium materials. These valves frequently follow the design guidelines outlined in MSS SP-
68. The primary distinguishing design feature is a stem which is offset from the centerline of the
pipe. This offset design generates a camming action which allows for different types of seating
technologies when compared to resilient seated valves. These alternate seating technologies
frequently incorporate advanced seal materials, such as PTFE, RTFE, and other advanced
polymers which help the valve achieve higher pressure and temperature ratings. A high
performance butterfly valve is therefore a butterfly valve which has extended the performance
boundaries (in terms of pressure, temperature, and/or corrosion resistance) of traditional butterfly
valves.
Butterfly Valve Characteristics
Within the quarter turn family of valves, ball valves are dominant in the sub-2 inch size
category. Above 2 inches in size, butterfly valves become much more prevalent primarily for
economic reasons. Butterfly valves exhibit a number of characteristics which make them
especially well suited to use in general service type applications. Those characteristics include:
Economical: Because butterfly valves can be designed in a physically compact package
where the closure mechanism (disc) is partially contained by the system piping in the open
position, they are very economic in terms of material usage. Since material costs are a
significant portion of the product cost, the final valve is also very cost efficient.
Compact: The compact package also makes for a valve which is comparatively light and
therefore relatively easy to handle. The compactness lends the valves to installation in places
where other valves will not easily fit while the lightness makes installation and maintenance
relatively easy.
Configuration: The variety of body styles in which butterfly valves are available lends the
valves to a variety of installation configurations and preferences. Typical body style
offerings include wafer, semi-lug, lugged, and flanged valves.
Performance: Butterfly valves are generally good performing valves in addition to being
3
economical. Performance characteristics for most designs include bi-directional sealing, low
pressure loss at high flow rates, simple operation, and reliable sealing.
Actuation: A variety of actuation options exist for butterfly valves including handle
operation (typically only valves 8 and smaller), manual gear operators, and a wide variety of
power operators including pneumatic, electric, and hydraulic.
Butterfly Valve Seats
For butterfly valves where the seat is mounted in the valve body, four basic types of seat exist:
1. Boot seats are the traditional seat style. This type of seat is a separate component which
typically has a U shaped cross section and completely lines the ID of the valve body. They
are retained in the valve body by interlocking dovetails at either face of the valve body.
The primary advantage of this type of seat is its replaceability but they are generally also
available in a wider variety of materials.
2. Cartridge seats also completely line the ID of the valve body but are generally rigid seats by
virtue of a hoop of rigid material encapsulated by a elastomer jacket. The resulting
assembly forms a rigid cylindrical shaped seat. These type of seats are also replaceable and
can offer advantages in vacuum applications.
3. Molded in seats are physically bonded to the valve body during the rubber molding process.
This style of valve offers good performance characteristics but is generally more limited in
material options and seat wear or damage is not repairable. The primary advantage of this
type of seat is its pressure rating and its ability to be used in dead-end or end of line
applications where no end flange is supporting the valve.
4. Mechanically retained seats are most common in high performance valves and in the water
works industry where large diameter valves make the other seat styles described above
impractical to manufacture. Seats are retained by a number of methods including mechanical
metal fasteners and retention keys formed with advanced adhesives. A few valve designs in
the industry mount the elastomeric seat on the disc edge. In this style of valve, mechanical
seat retention is used almost exclusively.
The boot, cartridge, and molded in seats are most common in the smaller valve sizes (through
roughly 24) and define the range of seat configurations for resilient seated butterfly valves
. The mechanically retained seats are typically only used on larger diameter (24 and larger)
AWWA style valves and in high performance valves.
Seat materials for butterfly valves vary but the predominant elastomers utilized would be EPDM
and NBR. Additionally, other elastomers in common use include CR (trade name Neoprene),
SBR, and occasionally more exotic materials such as fluoroelastomers, silicones, or others.
When high performance valves are utilized, thermoplastic seats are also frequently used.
PTFE and RTFE are the most common materials but other material types are used for some
applications.
4
In general terms, any of these materials described here are suitable in water service. The
conditions which lead to a preference of one material over another are just temperature or
other chemical constituents in the water. For example, EPDM gives excellent service in high
temperature (up to approximately 300 degrees F) water and low grade steam. In this
environment its performance is usually significantly superior to NBR which is generally
restricted to somewhere in the 200 degree F neighborhood. EPDM is not, however, a good
performer in environments where hydrocarbons are found while NBR is excellent in
environments where there are hydrocarbons. Generalizations such as this about these
material families can be made but the actual performance of the seat material varies widely
even within the material family due to differences in the compounding of the material.
Because of this, it is extremely important that application data be provided to the
manufacturer in order for the best seat material recommendation to be made. The primary
information that is needed to make the material selection is as follows: temperature (min. /
max. and typical), operating cycle, and chemical makeup of the water including
concentrations. From this information, a recommendation can usually be made or detailed
questions can be generated to help finalize the selection.
Minimum Pressure Drop
A common misconception in sizing large butterfly valves is the need for a minimum pressure
drop or large flow coefficient at full open. Most high performance (low pressure loss)
butterfly valves have flow coefficients large enough that the need or selection for minimum
pressure drop and maximum flow coefficient is irrelevant. The pressure loss of a full open
butterfly valve is so small in comparison to the overall loss of the flow system, that it has
very little effect. The larger the flow coefficient at full open, the smaller the range of
openings for flow control. The concern for minimum pressure loss is often misguided
because of the uncertainty of the pressure losses for rest of the flow system. The minimum
pressure loss of a butterfly valve is usually less than the variation or accuracy of determining
the pipe friction loss for the system.
Free Discharge
Free discharge from a butterfly valve is similar in nature to the condition of choking or
flashing cavitation. Both conditions choke the flow passages and separation zones of the
valve with vapor or air. The result of choking is then a limit of flow for a given upstream
pressure and a reduction in the efficiency or flow coefficient of the valve. The free discharge
flow and dynamic flow torques are less than those calculated for the same pressure drop of a
non-cavitating or non-free discharge condition. Knapp (1950) has classified cavitation as
being either a vaporous or gaseous type. Free discharge is similar in nature to gaseous
cavitation in that free air is drawn into the separation zones from downstream of the valve
instead of from the liquid flow itself.
5
H K
V
g

2
2
C
Q
P Sg
V
GPM

/
Choking Cavitation
Choking or flashing cavitation in a valve is a maximum design limit that results in a level of
extreme cavitation. Free discharge at lower pipe velocities and flows will not produce
undesirable effects. Dynamic flow and actuator torques are lower than expected, and any
cavitation will be significantly reduced. The presence of free air will cushion the cavitation
collapse and in many cases will prevent cavitation vapor from even forming. The major
disadvantage of free discharge is the reduction in flow and flow coefficient. However, it is
important to note that free discharge conditions can produce much larger pressure
differentials and velocities than from similar conditions in which downstream piping will
limit flow and pressure drop. Care must be taken in designing for free discharge from a
butterfly valve to avoid pipe velocities that exceed the maximum design velocities (ie 12 to
16 fps for some classes of AWWA valves).
Choking cavitation occurs when the local pressure inside a valve decreases to the vapor
pressure of the liquid, and the contracted flow through the valve flashes to vapor. At choking
cavitation, the maximum flow for a given upstream pressure (regardless of downstream
pressure) is reached. Typically, the flows to produce large vapor pockets downstream of the
valve are very large, and often exceed the maximum design flow and design structural
stresses for the valve and actuator. Flashing cavitation can also produce precipitous effects to
downstream piping and other flow components. It is not recommended to operate butterfly
valves at or near any limit of choking or flashing cavitation
Flow Coefficients
The most common formula used to relate the pressure differential to the flow is the Darcy-
Wiesbach equation (Equation 1).
(1)
where K is referred to as the resistance coefficient, H is the net pressure differential or head
loss in feet (meters), V is the flow velocity at the inlet in fps (m/s), and g is the gravitational
constant 32.2 fps (9.806 m/s). The parameter V/2g is also referred to as the velocity head.
As simple as the Darcy-Wiesbach equation is, the coefficient K has been used to denote a
number of flow phenomena such as the cavitation parameter and numerous flow coefficients
and constants.
A flow coefficient equation widely used by the valve industry (per ISA S75.01) is Equation 2.
(2)
6
P P Sg f
Q
d
net ISA
0008986
2
4
.
C
d
C
d
Sg f
C
d
V net
V ISA
V ISA
2
2
2
2
2
2
1 0008986

1
]
1

1
]
1

1
]
1

.
C
d K
V net
2
2
8906032

1
]
1

.
Where P is the pressure drop in units of psi, Q
gpm
is in units of gpm, and Sg is the specific
gravity of the fluid.
It is very important to note that the P used in the ISA Cv
ISA
equation is not the net pressure
drop P
net
. The pressure drop used in the ISA Cv
ISA
equation is determined from testing
requirements that include an additional 8 diameters in length of pipe and friction loss in the
measured pressure drop. ISA S75.01 does not allow for the subtraction of the manifold loss
between pipe taps and test valve from P
ISA
. For valves with a flow coefficient Cv/d
2
less
than 20 (where d is in units of inches), the difference between P
net
and P
ISA
is negligible.
The difference between pressure drops can be significant for Cv/d
2
values greater than 20.
Most water works valves have values of Cv/d
2
greater than 20. Equations 3 and 4 were
derived (ref Rahmeyer and Driskell) for converting the ISA test pressure drop (including 8
diameters of manifold loss) to a net pressure drop.
(3)
Where Q is in units of gpm, d is in units of inches, and P is in units of psi.
(4)
Another difficulty with using Cv as a loss coefficient is that Cv is not dimensionless. The
units of Cv are a function of gpm and psi. However, the net resistance coefficient K is
dimensionless, and can be converted directly to Cv
net
by Equation 5.
(5)
A flow coefficient that has limited use by several researchers (Tullis and Ball 1973) is the
coefficient C
d
(Equation 6). C
d
is dimensionless and is dervied directly from K (Equations 6
and 7), and has two important advantages in its use. The range of C
d
will always vary from
0 to 1, where the value 1 represents an ideal valve with no pressure loss. Most types of
numerical analysis or modeling require a relationship of flow coefficient as a function of an
independent variable such as valve opening. Experience has show that the generation or
curve fit of such relationships produces much simpler and accurate relationships if the flow
coefficient is in the form of C
d
.
7
C
V
g H V
K
d

+

+
2
1
1
2

K
C
d

1
1
2


+
P P
P
V 1
2
1

2
2

P P
P
V

(6)
(7)
Cavitation Parameters
The cavitation parameter or index is a dimensionless ratio used to relate the conditions which
inhibit cavitation (P
2
-P
V
) to the conditions which cause cavitation (P). There are numerous
forms of the dimensionless number or parameter which have been used to mathematically
describe cavitation. The most fundamental form is
2
of Equation 8 which uses the
downstream valve pressure (P
2)
, the vapor pressure of the liquid (P
V)
, and the pressure drop of
the valve (P). It is necessary to use
2
for determining specific cavitation effects such as
size and scale effects. However,
2
can be difficult to use for scaling or sizing a control valve
because of the lack or uncertainty of the value of the downstream valve pressure P
2
. Equation
9 is an alternate form of the cavitation parameter that allows the use of the upstream
pressure P
1
instead of the downstream pressure. This form of is much better suited for
scaling and sizing valves. It is also important to notes that
2
can be converted to by just
adding the value of 1. Equation 9 is the cavitation parameter that will be used for this paper.
(8)
(9)
The cavitation parameter can be used to predict the pressure drop or discharge at which a
control valve or orifice will begin to experience a given level of cavitation. If the
calculated for the actual operating pressures of a valve or orifice is less than the value of for
a cavitation limit, the valve or orifice will experience a level of cavitation more severe than
that associated with the limit.
The choking cavitation limit represented by
CHOKED
can be related to the Pressure Recovery
Factor F
L
by Equation 10. F
L
is a form of the choking cavitation parameter that is used by
most valve manufactures and valve design procedures (ISA 1985).
8
F
Q
C P P
L
GPM
V V CHOKED

1
096
1
.

CONSTANT
K
+ 1161
94364
.
.

CHOKED
K
+ 10851
20762
.
.
(10)
Butterfly valves have been tested by many different researchers and valve vendors for
cavitation limits. Equations 11 and 12 show the best fit equations from numerous tests of all
sizes and models of valves for the cavitation limits of constant and choking cavitation.
(11)
(12)
Equations 11 and 12 had minimum linear regression coefficients of R
2
=0.97 and a scatter or
accuracy of 10 to 15% accuracy when applying the equations for different sizes and models
of valves. Equations 11 and 12 have been provided as an aid for the preliminary sizing of
butterfly valves for constant cavitation and for free discharge. The most accurate method for
predicting cavitation limits is to still use actual test data for a specific valve type and size.
Coefficient of Dynamic Flow Torque
Dynamic flow torque is the flow torque produced by the flow forces on the valve disc about
the valve shaft. Most butterfly valves produce a dynamic flow torque that has the rotational
direction that will close the valve. Some types of eccentric butterfly valves that are installed
with the valve seats upstream of the valve shaft will experience flow torque above 70 to 80
degrees of opening that will tend to open the valve. The dimensionless flow torque
coefficient that has been used by many valve vendors and designers is C
TDP
of Equation 13. It
must be cautioned that there are other definitions and applications of C
TDP
that are not
dimensionless. For example, it is common to derive C
TDP
from units of ft.lbs of torque,
inches of disc diameter, and psi of pressure drop. This application is not dimensionless and
produces coefficient values that are 1/12 th that of a dimesionless coefficient.
Another dimensionless coefficient for dynamic torque is C
TV
of Equation 14. C
TV
differs
from C
TDP
in that it is based on velocity instead of pressure drop. The advantage of C
TV
is
that because velocity is used to calculate dynamic torque, C
TV
can be applied directly to
applications of installing small sizes of valves in larger pipes and for use in free discharge.
Equation 15 can be used to calculate C
TV
from values of C
TDP
. It must be cautioned that the
9
C C F
C
V V L
V
CHOKED
*

C C F
C
TDP TDP L
TDP
CHOKED
*
2

C
Dynamic FlowTorque
Pd
T
Pd
TDP


3 3
C K
C
TV
TDP

2
C
T
V d
TV

2 3
P
P P
V
CHOKED
*

1

K K
CHOKED
*
units normally used with C
TV
are foot pounds of torque instead of inch pounds as used with
C
TDP
.
(13)
(14)
(15)
Effect of Free Discharge and Choking on Flow and Torque Coefficients
The flow coefficients and torque coefficients for a butterfly valve are determined by testing
the valves in a test setup with pressurized flow both upstream and downstream of the test
valve. Both types of coefficients are based on the pressure drop associated with the flow
profiles and separation zones produced in pressurized flow. As discussed, choking cavitation
and free discharge cause changes to the flow profiles and separation zones, and almost always
reduce the efficiency or increase the pressure loss of the valve.

To predict flow and dynamic flow torque for the conditions of choking cavitation, equations
16, 17, 18, and 19 were derived from the choked cavitation parameter to correct the flow and
torque coefficients for choked flow.
(16)
(17)
(18)
(19)
10
Q C P Pa
GPM FREE DISCHARGE V
*
1
V
g H
K
FREE DISCHARGE

2
*
( ) T C P Pa d
FREE DISCHARGE TD
*
1
3
Where K*, Cv* and C
TDP
* are the corrected coefficients to be used to calculate flow and
torque for choking cavitation. It can be rationalized that free discharge is a form of choking
cavitation in which the flow is choked by free air instead of vapor. Tests have verified the
use of the corrected coefficients (Equations 16 through 19) to calculate flow and torque for
free discharge conditions. Equations 20, 21, and 22 are then used to calculate the pipe
velocity, flow, and dynamic torque for a free discharge condition.
(20)
(21)
(22)
Where P
a
is the atmospheric pressure. Table 1 compares calculations for free discharge with
actual test data for a 12-inch symmetrical and a 36-inch eccentric butterfly valve.
11
Table 1 Free Discharge Test Data and Calculations
12" SYM. BFV 36" ECC. BFV
Standard Valve Test Coefficients .................................... ....................................
Valve Opening 50 degrees 30 degrees
Cv 1,645 6,950
K 6.829 29.699
C
TDP
0.0783 0.0171

CHOKED
1.985 1.1815
F
L
0.71 0.92
Free Discharge Tests .................................... ....................................
P
1
-P
a
20 psig 42 psig
Q
MEASURED FREE DISCHARGE
5,137 gpm 41,405 gpm
V
MEASURED FREE DISCHARGE
14.57 fps 13.33 fps
T
MEASURED FREE DISCHARGE
1,423 in. lbs 26,813 in. lbs
Calculations .................................... ....................................
Cv* 1,167 6,394
K* 13.5554 35.0893
C
TDP
* 0.03945 0.01446
Q
CALCULATED FREE DISCHARGE
5,220 gpm 41,437 gpm
V
CALCULATED FREE DISCHARGE
14.81 fps 13.34 fps
T
CALCULATED FREE DISCHARGE
1,363 in. lbs 27,485 in. lbs
12
C
K
d
D
d
D
S

_
,

_
,

1
]
1
15
1 067 033
2
2
4
4
.
. .
C
d
D
K
R
+

_
,

1
15 1
2
2
2
.
P valveand reducer C P net valve
R
*( ) ( )
K KC
R
*
K K
d
D
* . +

_
,
15 1
2
2
2
( ) ( ) P P P P C P
V V S 1 1
* +
Non-line Sized Valves
Control valves are not always installed in the same line size as the valve. Non-line size valve
installations create additional connection losses which need to be added to the net valve loss.
Therefore, it is necessary to correct the net pressure drop of the valve (P) to the total loss
(P*) of the valve and pipe reducers. Equation 23 shows the Pressure Correction Factor used
with Equation 24 to convert the net pressure loss to the total loss of a non-line sized valve.
The equations to correct non-line sized valves are based on the assumption that the pipe
reducers will be standard symmetrical pipe fittings (Hutchison 1976 and ISA 1985).
(23)
(24)
The total flow coefficient K* of the valve and reducers can be calculated from either
Equation 25 or 26.
(25)
(26)
The upstream pressure P
1
* of the non-line sized valve and reducers can be calculated from
Equation 2 and the Cavitation Correction Factor, C
S
, of Equation 28.
(27)
(28)
13
C C C
TDP TDP R
*

*( )
( )
valveand reducers
net valve C
C
S
R

+
( ) +
R
PSE SSE 1 1
( )
( ) 1 1
R
PSE
( )
( )
PSE
P P
P P
V
V
R
n

1
]
1
1
1
1
The cavitation parameter for the combined valve and pipe reducers can then be calculated
from the net cavitation limit with Equation 29. If the valve is line sized (d=D), the correction
factor C
S
will equal 0 and the correction factor C
R
will equal 1.
(29)
The coefficient of dynamic torque used with a valve and reducers, C
TDP
*, can be calculated
from Equation 30. It must be cautioned that the application of C
TDP
* is based on the diameter
of the valve and the pressure drop of the combined valve and pipe reducers. To avoid any
confusion in calculating dynamic torque, the use of the coefficient C
TV
is recommended
where d is the diameter of the valve and V is the velocity of the valve inlet.
(30)
Equations 23 through 30 are based on theoretical derivations and on the procedures presented
by the ISA S75.01 Standards (1985) for control valves. The equations in ISA and in this
paper have been verified with limited testing.
Cavitation Scale Effects
(31)
(32)
(33)
Values of n for incipient and constant limits of cavitation for typical butterfly valves are n =0.25
to 0.28. The value of n for incipient damage is n =0.18 for a butterfly valve, and n =0 for
the limit of choking cavitation.
14
K
t
d
t
d
t
d
t
d
t
d
O O O O O
+

_
,

_
,
+

_
,

_
,
+

_
,
214 8024 10073 5996 16797 18031
2 3 4 5
. . . , , ,
SSE
d
d
R
Y
+

_
,

Y K

0159
18
.
/
C C
d
d
V V R
R

_
,

2
2
(34)
(35)
From numerous tests of butterfly valves, the cavitation limits of incipient and constant cavitation
can be scaled with Equation 36. The value of Y for the limits of incipient damage and
choking are Y =0.

(36)
Valve Size Scale Effects
For different sizes of geometrically similar valves, the flow coefficients K, Cv/d
2
, and Cd are
constant. The flow coefficient Cv is not constant with valve size and has to be scaled by pipe
area.
(37)
However, butterfly valves of identical design and model, typically will vary slightly in
geometry for different sizes because of the structural requirements for the disc and disc shaft.
The valve seats also differ in size ratio because many of the valve sizes will have the same
height and thickness of seat. J ames Ball (Ball and Tullis 1973) presented a study of the
variation of flow resistance for similar butterfly valves with differences in valve disc
diameter (d
o
) and thickness (t). The valves that were different models and types of butterfly
valves that included high performance, with seats upstream and downstream, and
symmetrical and boot seated butterfly valves. His data was presented in a plot of the flow
coefficient Cd versus the ratio t/d
o
. Equation 38 is the regression of the data presented in his
plot. The regression analysis had a R
2
of 96.3% and a maximum scatter of data of 8%. It is
suggested to use Equation 38 as a means to extrapolate know values from one valve size to
another.
(38)

R
SSE