Butterfly Valve Sizing

8/12/2019 Butterfly Valve Sizing

1/10

IsBilimi ve Teknii Dergisi, 30, 2, 103-112, 2010J. of Thermal Science and Technology

2010 TIBTD Printed in TurkeyISSN 1300-3615

EFFECT OF THE FLOW CONDITIONS AND VALVE SIZE ON BUTTERFLY VALVEPERFORMANCE

Mehmet SANDALCI1, Ebru MANUHAN2, Emre ALPMAN2, Kurtul KKADA3

1Gedik Casting and Valve Company, Pendik, 34 916 Istanbul, Turkey, msandalci@gedikdokum@com2Marmara University, Faculty of Engineering, Mechanical Engineering Department,

Gztepe, 34 722, Istanbul, [email protected] University, Faculty of Engineering, Chemical Engineering Department,

Gztepe,34 722, Istanbul, Turkey [email protected]

(GeliTarihi: 21. 10. 2009, Kabul Tarihi: 03. 02. 2010)

Abstract: In the present work, two different sizes butterfly valves, DN65 and DN80, were tested according to standardtesting method ANSI/ISA-75.02-1996. The tests were performed at different flow rates such as 2, 3 and 4 m/s and atdifferent valve opening angles such as 0, 10, 20, 30 and 40. The opening angle 0 was considered as the fully openvalve. The flow area percentages, , were calculated for different valve openings as 100, 82.64, 65.80, 50 and 35.72%and pressure drops, P, were recorded for different valve openings at different velocities. Using the experimental datathe loss and flow coefficients,Kand Cv, were calculated and correlations were proposed to giveKand Cvas a function ofthe flow area percentage, . Uncertainty analysis was performed to show the effect of the measurement uncertainties onthe performance coefficients. The proposed correlations provide an effective way to determine the performancecoefficients of two different sizes of butterfly valves. This approach can be applied to other valves for determining their

performance.Keywords:Butterfly valve; Loss coefficient; Flow coefficient; Uncertainty analysis.

AKIKOULLARI VE VANA APININ KELEBEK VANA PERFORMANS

KATSAYILARINA ETKSzet: Bu almada vana sektrnde tesisatlarda abuk ama ve kapama durumlar iin ideal bir tasarm olan kelebekvanalarda ( DN80 ve DN65) farklakhzlarnda 2, 3, 4 m/s (trblanslakkoulunda) farkldisk alarna (=0, 10,20, 30ve 40) karlk gelen ak alan yzdelerinde ( =%100, %82,64, %65,80, %50 ve %35,72) basn kaybllmtr. Kelebek vana diski tam ak pozisyonda (=0) iken borudan geen akkan debisi maksimumdur. Vanamerkezine yataklanmdisk, askadar dndrldnde akkan ak snrlanmaya balar ve as40oolduundaakkan geii ksmen azalr. lmler test metodu ANSI/ISA-75.02-1996da belirtilen artlara uyularak kurulan testdzeneinde yaplmtr. Bu almada, iki farkl ap kelebek vanada basn kayb lm deerleri kullanlarakhesaplanan kayp (K) ve debi katsaylarnn (Cv) disk alarna gre deiimi

gsterildi. lmlerdeki belirsizliklerin

performans katsaylarna (K ve Cv) etkisini grebilmek iin belirsizlik analizi yapld. lm sonular kullanlarakperformans katsaylarn, kelebek vanadaki disk asna bal olarak kolayca hesaplayabilecek korelasyonlar nerildi.retici firma veya kullanclar bu korelasyonlarkullanarak DN80 ve DN65 kelebek vanalarda farkldisk alariinKve

Cvdeerlerini kolayca hesaplayabileceklerdir.Anahtar Kelimeler: Kelebek vana, Kayp Katsays, Debi katsays, Belirsizlik analizi

NOMENCLATURE

Av valve cross section area [m2]

Ao valve flow area [m2]

Cv flow coefficient [m3/s or gal/min]

K loss coefficient [-]

103


2/10

104

Q volumetric flow rate [m3/s or gal/min]r valve radius [m]Re Reynolds numberv velocity [m/s]UK Uncertainty for loss coefficient [-]UCv Uncertainty for flow coefficient [m

3/s orgal/min]UP Uncertainty for the pressure drop [N/m2 or

psia]

UQ Uncertainty for the volume flow rate [m3/s or

gal/min]

Greek LettersP pressure drop [N/m2or psia]P0 pressure drop [N/m

2or psia] density [kg/m3or lb/ft3]0 density of reference water [kg/m3or lb/ft3] disk rotation angle [Degree or radians] the opening angle [Degree]

INTRODUCTION

Selecting the proper valve for piping systems plays animportant role in reducing the energy requirement andthus the operating cost. Various valves are used for on-off control, modulation of the flow rate through thesystem, prevention of back flow and pressure relief assafety devices. One of the most widely used valves is

butterfly valve. The primary aim of butterfly valve is to

regulate the flow rate. The flow characteristics inside anddownstream of the butterfly valves behave differently indifferent angle of valve disk and different inlet velocities.Coefficients (loss and flow coefficients) defining the

performance of a valve are affected by the valve opening.In addition the flow coefficient is a function of the valvediameter. The flow rate and pressure measurementtogether with the particle tracking flow visualizationmethod were used to estimate the performance and flow

patterns of a ball valve by (Chern et al., 2007). Thecorrelation between the flow pattern and valve

performance is discussed. They stated that the inletvelocity and ball valve opening play very important roles

in the flow characteristics of ball valves and theyproposed correlations to determine the loss and flowcoefficients as functions of the valve opening for a ballvalve of 38 mm in diameter. The research on butterflyvalves concerned the investigation of valve performancefor various valve and pipe configurations. The effect ofthe downstream of an elbow on the valve performancewas investigated by (Morris and Dutton, 1991a). Theeffect of the valve/elbow interactions on the pressure dropand flow coefficient was investigated using air as theworking fluid for a butterfly valve of 76.2 mm indiameter. They also investigated the effect of two

butterfly valves mounted in series on the valve

performance (Morris and Dutton, 1991b). Pressure losseswere measured by (Fester et al., 2007) for 5 different sizesof diaphragm valves of diameters ranging from 40 mm to100 mm using both Newtonian and non-Newtonian fluidsfor laminar, transitional and turbulent flow. Empiricalcorrelations were derived to calculate the loss coefficientsfor each diaphragm valve in the fully open position.(Perry, 1997) gives loss coefficient data for butterflyvalves in the nearly fully open (5) and 10, 20, 40,60open position. But, there is no mention of a diameter

effect. The effect of the two different disk configurationssuch as perforated and solid disk plates on the losscoefficient was investigated by (Eom, 1998)for a butterflyvalve of 100 mm in diameter for position of every 10from0to 90.

The aim of this study is to measure and observevariations of performance coefficients (loss coefficientand flow coefficient) of DN65 and DN80 butterfly valvesfor different flow conditions. These two different sizes

being DN65 and DN80 are the most commonly usedsizes in industrial applications. The flow conditions referto 3 different Reynolds numbers for two different sizes ofthe valves at 5 different valve flow area percentage(100%, 82.64%, 65.80%, 50%, 35.72%) corresponding to0, 10, 20, 30, 40, regarding the fully open position(100% open) as 0. The Reynolds number changes from1.555105to 3.12105for DN80 and from 1.255105to2.52105 for DN65 sizes of valves. These values of theReynolds number correspond to 3 different velocitiessuch as 2, 3 and 4 m/s.

EXPERIMENTAL INVESTIGATION

Experimental Set-up

Experimental set-up was installed on the test site ofStandard Pump Company according to the (ANSI/ISA-75.02-1996, 1996) and the experimental set-up wasshown schematically in Fig. 1. Two different sizes ofwafer style butterfly valves, DN65 and DN80, were used.The experimental set-up is a closed system with a waterreservoir of a capacity of 1000 litres and the circulationof water is achieved by using a centrifugal pump(Standart Pump) with a power of 18.5 kW. The

volumetric flow rate, Q, is controlled using two flowbalance valves (Todr and Anderson) and a digitalflowmeter (Khrone) is used to record the flow rate. Theaccuracy of the flowmeter is within 0.1%. The inlet

pressure is measured using a barometer (Pakkens) and theinlet pressure was kept constant at 400kPa. Valve

performance or in other words, the flow coefficient andloss coefficient were computed using the experimentaldata. Pressure drop was measured using a U-tubemanometer with probes installed 2 diameters upstream


3/10


4/10

106

positions (0, 10, 20, 30, 40) of the disk rotationangle, , from fully open, =0, to partially open, =40for DN80 and DN65 butterfly valves, are given in Table1. Fig. 3 shows the variation of butterfly valve flow

areas and flow area percentage versus the disk rotationangle starting from fully open to fully closed positions byusing Eqs. (1)-(3).

Table 1.Valve Flow Area and Flow Area Percentages Corresponding to Different Disk Rotation Angle of DN80 and DN65 Valves

Rotation angle, (Degrees) 0 10 20 30 40

Rotation angle, (Radians) 0 /18 /9 /6 2/9

Valve flow area for DN80 (m2) 0.00478 0.00395 0.00314 0.00239 0.00171

Valve flow area for DN65 (m2) 0.00312 0.002576 0.00205 0.00156 0.00112

Flow area percentage (%) 100 82.64 65.80 50 35.72

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50 60 70 80 90

Angle (Degree)

Flowarearatio(%)

0

0,001

0,002

0,003

0,004

0,005

0,006

Flowarea(m2)

Flow area ratio

Flow area

Figure 3.Variation of Butterfly Valve Flow Area and Flow Area Percentages Versus the Disk Rotation Angles (DN80).

Uncertainty Analysis

In the calculation of the performance coefficients such asthe flow and loss coefficients there are uncertainties dueto the uncertainties in the pressure drop and flow ratemeasurements. The reading accuracy of the manometerand the flowmeter is around 3% and 0.1%,respectively. Therefore, the uncertainty in the pressuredrop is 3% while the uncertainty in the flow rate 0.1%.Using these values the uncertainties in the flow and losscoefficients were calculated and given in Table 2. Theuncertainty analysis was performed using

the method described by (Taylor et al, 1999). Theprocedure for calculating the uncertainties in the loss andflow coefficients was explained in detail in the AppendixA. The percent relative uncertainties forKand Cvvaluesare the ratio of the uncertainties at different angles to thevalues of K and C

v

, such as UK

/K and UCv

/Cv

,respectively. Then the percent relative uncertainty in theloss coefficient is 3.0 % for two different sizes ofvalves. However, the percent relative uncertainty in theflow coefficient is 1.50 % for DN80 while it was foundto be 3.35 % for DN65 size of valve.


5/10

107

Table 2.Uncertainties in the Loss and Flow Coefficients for Two Different Valve Sizes.

Opening angles(deg.) 0 10 20 30 40Flow area percentage(%)

100.0 82.64 65.80 50.00 35.72

DN80 KUK 1.150.035 1.650.050 2.360.071 5.380.16 15.20.46

VCV UC 291.23.94 243.13.29 203.52.75 134.71.82 80.081.08

DN65KUK 1.510.045 1.650.050 2.330.070 4.950.15 16.70.50

VCV UC 149.55.01 142.84.79 120.24.03 82.472.77 44.961.51

PERFORMANCE COEFFICIENTS

In general, a valve is evaluated using the loss coefficientK, and, the flow coefficient Cv. In this work, K and Cvcoefficients were calculated by using the pressure dropand volume flow rate measurements. The pressure dropmeasurements, P, were recorded for five different valve

the flow area percentages (100%, 82.64%, 65.80%, 50%,35.72%) corresponding to 0, 10, 20, 30, 40 atdifferent inlet velocities. The values of inlet velocity forexperiments are 2, 3 and 4 m/s. The Reynolds numberscorresponding to three different velocities are 1.56x105,2.33x105 and 3.12x105 for DN80 valve. Similarly, theReynolds numbers are calculated to be 1.26x105,1.89x105and 2.52x105for DN65 valve at three differentvelocities.

The Loss Coefficient

The loss coefficient is unique to each type of valve and it

is a dimensionless parameter giving the ratio of thepressure drop to the kinetic energy of the fluid. The losscan be given as follows for any valve geometry:

P2

v

1

v2

1

PK

22

== (4)

In Eq. (4), v is the inlet velocity, is the density of thefluid and P is the pressure drop measured between 2diameters in front of the valve and 6 diameters behind thevalve.

The Flow Coefficient

The flow coefficient is defined as the flow capacity of avalve at a standard temperature between 5 and 40Ccorresponding to a unit pressure drop, P0, at an opening

position. The value of P0 is 1 psia, or 1 N/m2

depending on the units of Qbeing U.S. gal/min or m3/s.In the below equation, 0represents the density of the

reference fluid which is water and it is taken as 62.4 lb/ft3or 1000 kg/m3depending on the units of Q.

2/1

0

0

=

P

PQCV (5)

The flow coefficient is generally given in U.S. gal/min inthe literature and it is given for a reference temperature of60F and for a reference pressure loss, P0, of one pound

per square inch at a specific opening position. When P0and/0are taken as unity the above equation reduces to:

2/11

=

PQCV (6)

In the above equation when the volume flow rate, Q, andthe pressure drop are used in units of U. S. gal/min and

psia, the flow coefficient, Cv, is obtained in U.S. gal/min,(Perry, 1997) and (Zappe, 1999).

RESULTS AND DISCUSSION

The pressure drop measurements for five different valveopenings and for three different velocities were used tocalculate the flow and loss coefficients. The calculationswere done for two different sizes of butterfly valves

being DN65 and DN80. In the following sections, theloss and flow coefficients obtained using experimentalmeasurements will be presented.

The Loss Coefficient

The variation of the loss coefficient with the openingangles at three different velocities was shown in Fig. 4.As it can be seen the dependence of the loss coefficienton the velocity is negligible and it is a strong function ofthe opening angles. Fig. 4 shows that the loss coefficientis independent of the Reynolds number in turbulent flow.


6/10

108

0

2

4

6

8

10

12

14

16

18

30 40 50 60 70 80 90 100

The opening angles, (%)

Losscoefficient,K

2 m/s

3 m/s

4 m/s

v (m/s) Re

2 1.55105

3 2.3310

5

4 3.12105

Figure 4.Loss Coefficient for DN65 Versus the Opening Angles at Three Different Velocities.

Variation of the loss coefficient which is calculated usingmeasurements is compared by obtained data from

products of (Hybvalve, 2008) manufacturer in figures 5and 6. It was observed that the loss coefficient is notaffected by the flow rate of the working fluid, but theopening angles of butterfly valve. The value of Kdecreases as the disk angle moves form partially closed

position, =40o, to fully open position, =0o.

The relations can be given to relate the loss coefficient tothe opening angles, . The loss coefficients for a velocityof 4 m/s, and for DN80 and DN65 butterfly valves are

%310074.1 514.25 = K and

%310244.0 269.25 = K , respectively and theyare shown in figures 5 and 6.

0

2

4

6

8

10

12

14

16

18

20

30 40 50 60 70 80 90 100

Flow Area Percentage, (%)

LossCoefficient,K

Measured DN80

HYBVALVE [10]

s (Measured DN80)

K=1.074105-2.514 3.0 %

Figure 5. Comparison of CalculatedKValues at 3.12x105Reynolds number with the Other Literature Data (DN80).


7/10

109

0

1

2

3

4

5

6

7

8

9

30 40 50 60 70 80 90 100


Losscoefficient,K

Measured DN65

HYBVALVE 10

Power (Measured DN65)

K=0.244105-2.269 3.0 %

Figure 6.Comparison of CalculatedKValues at 2.52x105Reynolds number with the Other Literature Data (DN65).

The Flow Coefficient

Variation of flow coefficient, Cv, is the indicator of theflow rate at a certain pressure drop as a function of the

valve opening. The computed values of the flowcoefficient with respect to the opening angles at differentvelocities for DN65 valve size were given in Fig. 7.

0

25

50

75

100

125

150

175

200

30 40 50 60 70 80 90 100


Flowc

oefficient,Cv

2 m/s

3 m/s

4 m/s

v (m/s) Re

2 1.55105

3 2.33105

4 3.12105

Figure 7.Flow Coefficients for DN65 Versus the Opening Angles at Three Different Velocities.

It was observed that the flow coefficient increases as theflow area of the valve increases. However, the effect ofthe velocity is negligible. Variation of the flowcoefficient with the opening angle is not linear but anexponential function as shown in figures 8 and 9. Fortwo different sizes of butterfly valves such as DN80 and

DN65 the flow coefficients are( ) 50.10195.0exp431.42 = VC

and ( ) 35.3018.0exp334.34 = VC , respectively.The results were compared to obtained data from(Valmate, 2008) and (Crtec, 2008) manufacturers.


8/10

110

Differences between the computed results fromexperimental data and the results given in the literature

were observed at higher the opening angles.

0

50

100

150

200

250

300

350

30 40 50 60 70 80 90 100


FlowCoefficient,CV

Measured DN80

CR-TEC 11

Valmate 12

Expon. (Measured DN80)

Cv=42.431 exp( 0.0195 ) 1.5%

Figure 8.Comparison of Calculated Cv Values with the Other Literature Data for 3.12x10

5Reynolds number (DN80).

0

50

100

150

200

250

300

30 40 50 60 70 80 90 100


Flowcoefficient,Cv

Measured

Valmate 12

CR-TEC 11

Expon. (Measured)

Cv=34.334 exp( 0.018 )3.35%

Figure 9.Comparison of Calculated CvValues with the Other Literature Data for 2.52x10

5Reynolds number (DN65).

Comparison of the Loss and Flow Coefficients forTwo Different Sizes of the Valves

The variation of the loss coefficient, K, and the flowcoefficient, Cv, with respect to the opening angles, , for

both valves, DN80 and DN65, were shown in figures 10and 11. It was observed that the flow coefficient is afunction of the valve size but the variation of the losscoefficient is different. Although the loss coefficient isindependent of the valve size at the higher opening


9/10

111

angles it is dependent of the valve size at the lower opening angles.

0

2

4

6

8

10

12

14

16

18

20

30 40 50 60 70 80 90 100


Losscoefficient,K

DN65

DN80

Figure 10.The Change ofKValues Versus the Opening Angles for DN65 and DN80 at the 4 m/s Velocity.

0

50

100

150

200

250

300

30 40 50 60 70 80 90 100


Flowcoeffic

ient,Cv

DN65

DN80

Figure 11.The Change of CvValues Versus the Opening Angles for DN65 and DN80 at the 4 m/s Velocity.

CONCLUSIONS

The valve performance can be determined by using theloss coefficient K, and the flow coefficient Cv, using the

pressure loss and volume flow rate information from theexperimental data.

The loss coefficient is independent of the inletvelocity and it is dependent on the valve size at thelower opening angles. But, each one of the proposedcorrelations, %310074.1 514.25 = K or

%310244.0 269.25 = K , can be used for two

different sizes of butterfly valve at the higheropening angles.


10/10

112

Flow coefficient is independent of the inlet velocitybut it is dependent on the valve size. The proposedcorrelations giving the flow coefficients as functionsof the opening angle are

( ) 5.10195.0exp053.49 = VC and

( ) 35.3018.0exp334.34 = VC for DN80 and

DN65 butterfly valves, respectively. Proposed correlations for Kand Cvwhich are given

above are appropriate for practical use.Manufacturer or designer of butterfly valves can findeasily the correspondingKand Cvvalues for a givenvalve opening angles.

APPENDIX A

The Procedure for Calculating the Uncertainties inthe Loss and Flow Coefficients

The set of input parameters are directly related to the

measured variables in uncertainty analysis. Thecomputed performance coefficients are obtained usingthe experimental measurements. Uncertainties in the

performance coefficients such as K and Cv, aredetermined using the uncertainties in the flow rate Q, and

pressure drop P.

The loss coefficient K, and flow coefficient Cv can becalculated for any valve by using the below givenequations:

P

Q8

DP2

V

1K

2

42

2 == (A.1)

P

Q

PQCV

=

=

5.01 (A.2)

The uncertainty in the pressure drop PU and the

uncertainty in volume flow rate, QU are used to be 3%

and 0.1%, respectively.

The partial derivatives of Kwith respect to measured Qand measured P are called the sensitivity coefficientsand they are given as:

P

Q4

D

Q

K3

42

=

2

42

Q8

D

P

K=

(A.3)

Similarly, the sensitivity coefficients for CV can bewritten as:

PQ

CV

=

1 ( ) 2/32 PQ

P

CV

=

(A.4)

The uncertainty in the loss coefficient in terms of theuncertainties and sensitivity coefficients is given as:

2/12

P

2

QK UP

KU

Q

KU

+

=

(A.5)

Similarly the uncertainty in the flow coefficient due touncertainties in the pressure drop and volume flow rate:

2/12

Pv

2

Qv

VC U

P

CU

Q

CU

+

=

(A.6)

REFERENCES

Chern, M. J., Wang, C. C., and Ma, C. H., PerformanceTest and flow Visualization of Ball Valve, Exp Therm.and Fluid Sci. 31, 505-512, 2007.

Morris, M. J. and Dutton, J., C., An ExperimentalInvestigation of Butterfly Valve Performance

Downstream of an Elbow, J. of Fluids Eng. 113, 81-85,1991a.

Morris, M. J. and Dutton, J. C., The Performance of TwoButterfly Valves Mounted in Series, J. of Fluids Eng.113, 419-423, 1991b.

Fester, V. G., Kazadi, D. M., Mbiya, B. M. and Slatter, P.T., Loss Coefficients for Flow of Newtonian and Non-

Newtonian Fluids Through Diaphragm Valves, Chem.Eng. Res. and Des. 85, 1314-1324, 2007.

Perry, R., H., Perrys Chemical Engineering Handbook(7thEd.), McGraw-Hill Companies, U.S.A., 1997.

Eom, K., Performance of Butterfly Valves as a FlowController, J. of Fluids Eng. 110, 16-19, 1998.

ANSI/ISA-75.02-1996, Control Valve Capacity TestProcedures, American National Standards Institute, 1996.

Taylor, R., P., Hodge, B., K., and James, C., A.,Estimating Uncertainty in Thermal Systems Analysis andDesign, Appl.Therm. Eng. 19, 51-73, 1999.

Zappe, R., W., Valve Selection Handbook (4thEd.), GulfProfessional Publishes, Burlington, MA, U.S.A., 1999.

Hybvalve Loss Calculations, Available from:, 2008.

Valmate Industrial Corporation. Available from:, 2008.

Crtec Flow Calculations. Available from: , 2008.

Butterfly Valve Sizing

Documents

Transcript of Butterfly Valve Sizing