See discussions, stats, and author profiles for this publication at: [Link]

net/publication/238183864

Study on Hydrodynamic Torque of a Butterfly Valve

Article in Journal of Fluids Engineering · January 2006

DOI: 10.1115/1.2137348

CITATIONS READS
42 5,271

2 authors:

Ju Yeop Park Myung Kyoon Chung

Korea Institute of Nuclear Safety Korea Advanced Institute of Science and Technology
21 PUBLICATIONS 111 CITATIONS 110 PUBLICATIONS 1,845 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Myung Kyoon Chung on 21 January 2019.

The user has requested enhancement of the downloaded file.

 Journal of
Fluids
Engineering Technical Briefs

Study on Hydrodynamic Torque of a 关1,2兴. Based on a two-dimensional butterfly valve model, the flow
field in the valve was solved by a free-streamline theory 关3兴 with
Butterfly Valve successive conformal transformations and contraction coefficients,
force and torque coefficients were obtained. It is well known that
the free-streamline theory is suited to analyze the plane irrota-
Ju Yeop Park1 tional flow field that includes free or bounded jets. Recently, Has-
e-mail: k385pjy@[Link] senpflug 关4兴 suggested a new two-dimensional butterfly valve
Regulatory Research Division, model that is more precise mathematically and physically than
that of Sarpkaya 关1,2兴. The major difference between two models
Korea Institute of Nuclear Safety, is an introduction of minimum velocity 共see Fig. 1兲 that is located
19 Guseong-dong, Yuseong-gu, in the upper bounding wall to reflect four parameters in the
Daejeon, 305-338, Korea hodograph plane 关4兴. Based on the new model, Hassenpflug 关4兴
solved the problem with the free-streamline theory and found con-
Myung Kyoon Chung traction coefficient but, unfortunately, the hydrodynamic torque of
Department of Mechanical Engineering, the butterfly valve was untouched. Since the hydrodynamic torque
Korea Advanced Institute of Science and Technology, characteristic is an important parameter in designing a butterfly
373-1 Guseong-dong, Yuseong-gu, valve, in the present study, the hydrodynamic torque is investi-
gated based on the improved valve model and the result is com-
Daejeon, 305-701, Korea pared with that of Sarpkaya 关2兴 and recent experiments 关5,6兴.

Since knowledge on hydrodynamic torque of a butterfly valve is

2 Free-Streamline Theory
very important for butterfly valve design, its hydrodynamic torque In order to determine the hydrodynamic torque that works on
is investigated theoretically. For this, a recently developed two- the two-dimensional butterfly valve, whole pressure distribution
dimensional butterfly valve model is solved through the free- on a valve plate is required and this means that a complete flow
streamline theory with a newly devised iterative scheme and the field should be solved. Therefore, an application of the free-
resulting two-and three-dimensional torque coefficients are com- streamline theory is also made in the present study.
pared with previous theoretical results based on the conventional The free-streamline theory is composed of successive confor-
butterfly valve model and experiments. Comparison shows that the mal transformations from a physical plane to a complex potential
improvement due to the new butterfly valve model is marginal. one. The unique feature of the theory is the introduction of a
That is, the three-dimensional torque coefficient is well repre- velocity plane, the hodograph. In the hodograph, an unknown
sented by the new model. Otherwise, the two-dimensional torque free-streamline is transformed to a circular arc exactly and further
coefficient is well predicted by the conventional model. In spite conformal transformations can be done easily. In the present
this fact, the present results can be used in further researches on study, z 共physical兲, w 共hodograph兲, ⌳ 共logarithmic hodograph兲,
butterfly valves because the improved butterfly valve model is and t 共Schwarz-Christoffel兲 planes are used. Equations 共1兲–共3兲 and
mathematically correct and reflects physical reality more correctly Figs. 1–4 show corresponding transform functions and planes,
than the conventional valve model. 关DOI: 10.1115/1.2137348兴 respectively:

Keywords: butterfly valve, valve design, hydrodynamic torque w = u − iv = qe−i␪ 共1兲

w q
⌳ = ln = ln − i␪ 共2兲
VJ VJ
1 Introduction
冕
t
The butterfly valve is widely used in various industrial applica- ⌳=M 共t − tCI兲−1/2共t − tDI兲−1/2共t − tE兲1共t − tDII兲−1/2共t − tCII兲−1/2dt
tions for on-off and throttling services because it is easily closed 0
or opened with relatively low pressure drop. In designing a but-
terfly valve, knowledge on its hydrodynamic torque is required for +N 共3兲
actuator sizing and structural design. where w关⬅dF共z兲 / dz兴 is complex velocity, F共z兲关⬅␾共x , y兲
To the authors’ best knowledge, a only theoretical study on + i␺共x , y兲兴 is complex potential with velocity potential ␾共x , y兲 and
hydrodynamic torque of the butterfly valve is given by Sarpkaya
stream function ␺共x , y兲, z共⬅x + iy兲 is a complex number, i共
⬅冑−1兲 is an imaginary number, u and v are velocity components
in the x and y directions, respectively, q共⬅冑u2 + v2兲 is magnitude
1
Author to whom all correspondence should be addressed.
Contributed by the Fluids Engineering Division of ASME for publication in the
JOURNAL OF FLUIDS ENGINEERING. Manuscript received July 26, 2005; final manuscript of velocity, ␪关⬅tan−1共v / u兲兴 is an angle, and VJ is a jet velocity
received August 10, 2005. Review conducted by Joseph Katz. between wall boundary and free-streamline. M and N are con-

190 / Vol. 128, JANUARY 2006 Copyright © 2006 by ASME Transactions of the ASME

Downloaded From: [Link] on 02/18/2018 Terms of Use: [Link]

 Fig. 1 Two-dimensional butterfly valve model „z plane…

Fig. 3 Logarithmic hodograph „⌳ plane…

stants of the Schwarz-Christoffel transformation and tCI , tDI , . . .
are parameters of the Schwarz-Christoffel transformation.
The complex potential F in the t plane is easily deduced by
noting that a source of strength 2␩Vo and sinks of strength 2nVJ 3 Solution Procedure
and 2mVJ are located at points tA, tDI, and tDII, respectively. It is Although the two-dimensional butterfly valve model is solved
given by conceptually in the previous section, this is not the case. One
2␩Vo 2nVJ 2mVJ should determine proper Cc1, Cc2, and Schwarz-Christoffel pa-
F共t兲 = ln共t − tA兲 − ln共t − tDI兲 − ln共t − tDII兲 rameters such as tCI, tDII, and tCII including complex constants M
2␲ 2␲ 2␲
and N to complete the solution. In the present study, determination
V J␩ of corresponding parameters is done numerically through the fol-
= 共1 − sin ␣兲关共Cc1 + Cc2兲ln共t − tA兲 − Cc1 ln共t − tDI兲 lowing iterative procedure.
2␲
− Cc2 ln共t − tDII兲兴 共4兲 • Step 1. Assume VJ / V MIN to fix a configuration of trans-
formed region in the ⌳ plane.
where Vo is an inlet velocity of channel, ␩ is a plate length and • Step 2. Solve Schwarz-Christoffel parameter problem. Pa-
channel width, n and m are thickness of upper and lower jets, Cc1 rameters such as tCI, tDII, and tCII including M and N are
and Cc2 are upper and lower contraction coefficients, and ␣ is a solved by numerical algorithm of Chung, et al. 关7兴. Values
valve closing angle. For Eq. 共4兲, a continuity of mass and defini- of tDI, tE, and tB are fixed as 0, 1, and ±⬁, respectively. Table
tion of contraction coefficients are used: 1 shows the calculated Schwarz-Christoffel parameters for
␩Vo = nVJ + mVJ 共5兲 various closing angles. Note that closing angles of 0 and
90 deg cases are excluded in the calculation.
2n 2m • Step 3. Assume tA value and find Vo / VJ and Cc1 + Cc2 asso-
Cc1 ⬅ and Cc2 ⬅ 共6兲 ciated with tA. Schwarz-Christoffel transformation deter-
␩ − ␩ sin ␣ ␩ − ␩ sin ␣
mined in Step 2 is used for finding Vo / VJ since ⌳共tA兲
Finally, an implicit solution can be determined as = ln共Vo / VJ兲. Cc1 + Cc2 is determined as

z共t兲 = −
␩
2␲
共1 − sin ␣兲 冕 冉e−⌳
Cc1
+
Cc2
t − tDI t − tDII
−
Cc1 + Cc2
t − tA
dt冊 Cc1 + Cc2 =
2 Vo
共1 − sin ␣兲 VJ
. 共8兲

+ const 共7兲 • Step 4. Determine Cc1 and Cc2, respectively, by the follow-
ing equations 共see Appendix A兲.
where

冕 冉 冊
tD
dz 1 1 I 1 1
= e−⌳ 1− e−⌳R sin ⌳I − dt共Cc1 + Cc2兲
dF VJ ␲ tC
t − tDII t − tA

冕 冉 冊
I
Cc1 = tD
and 1 I 1 1

冉 冊
1+ e−⌳R sin ⌳I − dt
dF VJ␩ Cc1 + Cc2 Cc1 Cc2 ␲ t − tDI t − tDII
= 共1 − sin ␣兲 − − tC
I
dt 2␲ t − tA t − tDI t − tDII
共9兲
have been used.

Fig. 2 Hodograph „w plane… Fig. 4 Schwarz-Christoffel transformation „t plane…

Journal of Fluids Engineering JANUARY 2006, Vol. 128 / 191

Downloaded From: [Link] on 02/18/2018 Terms of Use: [Link]

 Table 1 Parameters of Schwarz-Christoffel transform

␣ t CI tDII tCII M N

0 ¯ ¯ ¯ ¯ ¯
5 −0.00257906512251 5.16772774261455 251.37036829504729 −1.00000000021639 0.00000000000000
10 −0.00268265764988 5.06471845671051 64.12536160056500 −1.00000000000000 0.00000000000000
15 −0.00274481375813 4.92286680162222 29.32203997791210 −1.00000000000000 0.00000000000000
20 −0.00276261168564 4.74812305914738 17.10017335921508 −1.00000000000000 0.00000000000000
25 −0.00272843500847 4.54770620137642 11.42548158028952 −1.00000000000000 0.00000000000000
30 −0.00263713925367 4.32850438304937 8.33504784312660 −1.00000000000000 0.00000000000000
35 −0.00248528120954 4.09709918892896 6.46838537135951 −1.00000000000000 0.00000000000000
40 −0.00227229233408 3.85956841452090 5.25594457556547 −1.00000000000000 0.00000000000000
45 −0.00200257474524 3.62133137548344 4.42489100810807 −1.00000000000000 0.00000000000000
50 −0.00168654216178 3.38716391998115 3.83086788553041 −1.00000000000000 0.00000000000000
55 −0.00134123768751 3.16121003062576 3.39133383630574 −1.00000000000000 0.00000000000000
60 −0.00098982880739 2.94697152569338 3.05599796072864 −1.00000000000000 0.00000000000000
65 −0.00065987892850 2.74723758023603 2.79257688221188 −1.00000000000000 0.00000000000000
70 −0.00037973471009760 2.56396916883108543 2.57950814874032676 −1.00000000000000 0.00000000000000
75 −0.00017305181135684 2.39816032527879861 2.40208354604927172 −1.00000000000000 0.00000000000000
80 −0.00005127533678435 2.24974515396132983 2.25031897642927925 −0.99999999999998 0.00000000000000
85 −0.00000518578360261 2.11763826731841664 2.11766164626554610 −1.00000000000000 0.00000000000000
90 ¯ ¯ ¯ ¯ ¯

冕 冉 冊
tC predetermined error bound, proceed to next step. Otherwise,
1 II 1 1
1− e−⌳R sin ⌳I − dt共Cc1 + Cc2兲 go to step 1 again and try another VJ / V MIN value.
␲ tD
t − t DI t − t A • Step 8. Determine lu and ld with variables obtained in the

冕 冉 冊
II
Cc2 = tC previous steps 共see Appendix B兲 and verify ␩ = lu + ld to
1 II 1 1 check solution quality.
1− e−⌳R sin ⌳I − dt
␲ tD
t − tDI t − tDII

冉 冊
II

共10兲
冕
tC
␩共1 − sin ␣兲 I Cc1 Cc2 Cc1 + Cc2
lu = e−⌳R sin ⌳I + − dt
where ⌳R is a real part and ⌳I is an imaginary part of ⌳. 2␲ sin ␣ t − tDI t − tDII t − tA
−⬁
Cc1 + Cc2 determined in Step 3 is used here.
• Step 5. Compare Cc1 + Cc2 determined in step 3 with Cc1 共12兲
+ Cc2 obtained in step 4. If the difference between these two
values is within a predetermined error bound, proceed to

冕 冉 冊
+⬁
next step. Otherwise, go to step 3 again and try another tA ␩共1 − sin ␣兲 Cc1 Cc2 Cc1 + Cc2
value. ld = e−⌳R sin ⌳I + − dt
2␲ sin ␣ t − tDI t − tDII t − tA
• Step 6. Calculate tA from tDII, Cc1, and Cc2 which were tC
II

determined at steps 2 and 4, respectively. A relation among 共13兲

them is given as follows 共see Appendix B兲:
Cc2tDII where lu is a length from a stagnation point B to CI and ld is from
tA = . 共11兲 B to CII.
Cc1 + Cc2 Table 2 shows determined parameters of the two-dimensional
• Step 7. Compare tA determined in step 6 with tA assumed in butterfly valve model for various closing angles. One can note that
step 3. If the difference between these two values is within a the present calculation agrees well with that of Hassenpflug 关4兴.

Table 2 Parameters of the two-dimensional butterfly valve

␣ VJ / V MIN VJ / Vo Cc1 Cc2 Cc1 关4兴 Cc2 关4兴 lu / ␩ ld / ␩ 共lu + ld兲 / ␩ ld / ␩ 关4兴

0 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 1.0000 0.0000 1.0000 0.0000
5 1.1946 1.1699 0.9630 0.9098 0.9630 0.9098 0.9999 0.00008347 1.0000 0.000083
10 1.4242 1.3721 0.9287 0.8352 0.9286 0.8352 0.9988 0.001115 0.9999 0.00112
15 1.6973 1.6157 0.8968 0.7733 ¯ ¯ 0.9952 0.004743 0.9999 ¯
20 2.0251 1.9126 0.8672 0.7220 0.8672 0.7220 0.9873 0.01260 0.9999 0.0126
25 2.4234 2.2798 0.8396 0.6797 ¯ ¯ 0.9740 0.02588 0.9999 ¯
30 2.9151 2.7410 0.8140 0.6453 0.8140 0.6453 0.9548 0.04514 0.9999 0.0452
35 3.5338 3.3310 0.7901 0.6180 ¯ ¯ 0.9296 0.07036 0.9999 ¯
40 4.3314 4.1025 0.7678 0.5970 0.7678 0.5970 0.8989 0.1010 0.9999 0.1010
45 5.3903 5.1389 0.7469 0.5818 0.7469 0.5818 0.8638 0.1361 0.9999 0.1362
50 6.8485 6.5786 0.7275 0.5720 0.7275 0.5720 0.8252 0.1748 1.0000 0.1748
55 8.9496 8.6653 0.7093 0.5670 ¯ ¯ 0.7842 0.2157 0.9999 ¯
60 12.1577 11.8632 0.6922 0.5661 0.6922 0.5661 0.7421 0.2578 0.9999 0.2578
65 17.4454 17.1445 0.6763 0.5688 ¯ ¯ 0.6997 0.3002 0.9999 ¯
70 27.1434 26.8393 0.6614 0.5742 0.6614 0.5742 0.6578 0.3422 1.0000 0.3422
75 48.0565 47.7515 0.6475 0.5817 ¯ ¯ 0.6168 0.3831 0.9999 ¯
80 107.7647 107.4600 0.6345 0.5905 0.6345 0.5905 0.5770 0.4230 1.0000 0.4230
85 430.1348 429.8307 0.6224 0.6004 0.6224 0.6002 0.5380 0.4618 0.9998 0.4617
90 ⬁ ⬁ 0.6110 0.6110 0.6110 0.6110 0.5000 0.5000 1.0000 0.5000

192 / Vol. 128, JANUARY 2006 Transactions of the ASME

Downloaded From: [Link] on 02/18/2018 Terms of Use: [Link]

 Table 3 Moment arm, torque, and torce coefficients

␣ ␩c Tc Fc共numerical兲 Fc共analytical兲

0 0.00000 0.00000 0.00000 0.00000

5 −0.13887 0.016802 0.12096 0.12098
10 −0.11904 0.025210 0.21179 0.21177
15 −0.10088 0.028299 0.28051 0.28052
20 −0.084411 0.028096 0.33285 0.33284
25 −0.069561 0.025934 0.37284 0.37283
30 −0.056292 0.022711 0.40344 0.40344
35 −0.044575 0.019028 0.42689 0.42689
40 −0.034382 0.015296 0.44487 0.44487
45 −0.025688 0.011783 0.45869 0.45869
50 −0.018452 0.0086606 0.46935 0.46935
55 −0.012619 0.0060274 0.47764 0.47764
60 −0.0080984 0.0039206 0.48412 0.48412
65 −0.0047713 0.0023342 0.48921 0.48921
70 −0.0024865 0.0012263 0.49318 0.49318
75 −0.0010690 0.00053041 0.49618 0.49618
80 −0.00032247 0.00016069 0.49831 0.49831
Fig. 5 Pressure distributions at various closing angles 85 −0.000030083 0.000015029 0.49957 0.49958
90 0.00000 0.00000 0.50000 0.50000

4 Torque Coefficient
To calculate the hydrodynamic torque of the two-dimensional angles. Through comparison of Fc 共numerical兲 with Fc 共analyti-
butterfly valve, a pressure distribution on the plate should be ob- cal兲, one can evaluate the accuracy of the present numerical inte-
tained first. Application of the Bernoulli equation to lower plate gration.
segment BCII gives
p 1 2 1 5 Discussion
+ q = const = V2J . 共14兲 The two-dimensional torque coefficient of the butterfly valve
␳ 2 2
determined by Hassenpflug’s 关4兴 model is compared with that of
Rearranging, Sarpkaya 关2兴 in Fig. 6. Experimental data obtained by Sarpkaya
p
1/2␳V2J
=1− 冉 冊
q
VJ
2
. 共15兲
关2兴 are also included in the figure. It shows that although the
overall tendency of torque coefficient with respect to the closing
angle is similar in two models, the torque coefficient based on the
Since q / VJ = exp关⌳R共t兲兴, determination of ⌳R共t兲 spanning t from Hassenpflug model is deviates more from the experiment than that
t = tCII to +⬁ gives a pressure distribution on the lower plate seg- of Sarpkaya. This is interesting because Hassenpflug’s model is
ment. For the upper segment, a similar procedure should be fol- mathematically more precise than Sarpkaya’s and is believed to
lowed. Figure 5 shows pressure distributions on the plate at vari- reflect physical reality more closely 关4兴. Actually, as mentioned by
ous representative valve closing angles. By using the pressure Hassenpflug 关4兴, Sarpkaya’s model itself cannot give a unique
distribution, a moment arm ␩c with respect to the valve plate stagnation point on the valve plate and, to resolve this anomaly,
center for a specific closing angle is computed as Sarpkaya applied the uniqueness of stagnation point in solving the

冕
problem.
␩
Present results are also compared to the three-dimensional but-
p共s兲共s − 1/2␩兲ds terfly valve experiments 关5,6兴. For this, the two-dimensional
0
␩c = torque coefficient Tc should be transformed to the three-

冕
␩ 共16兲
p共s兲ds
0

where s is a coordinate along the plate length and a numerical

integration is performed by Simpson’s one-third rule 关8兴. Finally,
a hydrodynamic torque coefficient Tc of the two-dimensional but-
terfly valve is obtained by 关2兴

Tc ⬅
T
␳␦␩2V2J
=
1
2 sin ␣
1−
Vo
VJ
冉 冊 2
␩c 共17兲

where T is the hydrodynamic torque, ␳ is the density of fluid, and

␦ is the unit depth of plate. To verify the accuracy of numerical
integration, force coefficient Fc determined by numerical integra-
tion of the pressure and the analytical method are compared. Fc is
given by 关2兴

冕 冉 冊
␩ 2
F p共s兲 ds 1 Vo
Fc ⬅ = = 1− 共18兲
␳V2J ␩␦ 0
␳V2J ␩ 2 sin ␣ VJ
where F is the force on the valve plate. Table 3 shows calculation
results of moment arm, torque, and force coefficients of Hassen-
pflug’s two-dimensional butterfly valve model at various closing Fig. 6 Two-dimensional torque coefficient, Tc

Journal of Fluids Engineering JANUARY 2006, Vol. 128 / 193

Downloaded From: [Link] on 02/18/2018 Terms of Use: [Link]

 Fig. 7 Three-dimensional torque coefficient, Tc3A Fig. 8 Three-dimensional torque coefficient, Tc3B

dimensional counterpart. Sarpkaya 关2兴 suggested a three- Appendix A: Derivation of Cc1 and Cc2
dimensional torque coefficient of butterfly valve Tc3 as The implicit solution of the two-dimensional butterfly valve is
given by Eq. 共7兲 as
T
Tc3 ⬅ 共19兲
␳D3V2J
and showed that it is related to the two-dimensional torque coef-
z共t兲 = −
␩
2␲
共1 − sin ␣兲 冕 冉 e−⌳
Cc1
+
Cc2
t − tDI t − tDII
−
Cc1 + Cc2
t − tA
dt 冊
ficient as
+ const. 共A1兲
8
Tc3 = Tc . 共20兲 Then, it follows that
3␲
Tc3 suggested by Sarpkaya 关2兴 is easily linked to two kinds of
冕 冉
tD
␩ I Cc1 Cc2
widely used three-dimensional torque coefficients as follows: z共tDI兲 − z共tCI兲 = − 共1 − sin ␣兲 e−⌳ +
2␲ t − tDI t − tDII

冉 冊
tC

冊
2 I
VJ
Tc3A = 2Tc3 共21兲 Cc1 + Cc2
Vo − dt. 共A2兲
t − tA
1
Tc3B = 2Tc3 共22兲 Let z共tDI兲 − z共tCI兲 = x + iy and ⌳ = ⌳R + i⌳I. Then y becomes
共1 − Vo/VJ兲2

冕 冉
tD
where ␩ I Cc1 Cc2
y= 共1 − sin ␣兲 e−⌳R sin ⌳I +
T T 2␲ t − tDI t − tDII
Tc3A ⬅ and Tc3B ⬅ tC

冊
I
1/2␳D3V2o D3⌬PLoss
Cc1 + Cc2
are torque coefficients normalized by dynamic pressure and static − dt. 共A3兲
t − tA
pressure drop, respectively 关6兴, and ⌬PLoss is a pressure loss ac-
cross the valve. In deriving Eq. 共22兲, ⌬PLoss ⬅ 1 / 2␳KV2o with K Cc1 is given by
= 共VJ / Vo − 1兲2 is used 关2兴. Here, K is the loss coefficient of the
butterfly valve. Figures 7 and 8 compare the two different calcu- 2n
lations of Tc3A and Tc3B with experimental data. Unlike Tc, Tc3A Cc1 = 共A4兲
␩ − ␩ sin ␣
based on Hassenpflug’s model is better compared with experi-
ments than Sarpkaya’s model. For Tc3B case, both model predic- where n is a normal distance from upper wall to upper streamline
tions are not so good as the Tc3A case. But, here again, Hassenp- that is given by
flug’s model gives a better fit for Solliec and Danbon’s 关6兴
n = 2 ␩共1 − sin ␣兲 − y,
1
experiment. Noting that the pressure loss is included in the defi- 共A5兲
nition of Tc3B and Tc3B is more likely to be influenced by an
accuracy of pressure measurement across the valve, experimental From Eqs. 共A3兲–共A5兲, it follows that

冕 冉 冊
uncertainty may be a source of such discrepancy.
tD
As shown before, the two- and three-dimensional torque coef- 1 I Cc1 Cc2 Cc1 + Cc2
ficients based on the Hassenpflug model did not show a great Cc1 = 1 − e−⌳R sin ⌳I + − dt.
␲ tC
t − tDI t − tDII t − tA
improvement over the two- and three-dimensional torque coeffi- I

cients based on Sarpkaya’s. In spite of this fact, the present result 共A6兲
can be used as a cornerstone in developing any further research on
butterfly valves because the present result is obtained by a more Finally, applying Cc2 = Y − Cc1 to Eq. 共A6兲 and rearranging it with
improved butterfly valve model. respect to Cc1 gives

194 / Vol. 128, JANUARY 2006 Transactions of the ASME

Downloaded From: [Link] on 02/18/2018 Terms of Use: [Link]

 冕 冉 冊 where ⌳共tCII兲 = 共␲ − ␣兲i is used. Since ⌳共t兲 = ⌳R共t兲 + 共␲ − ␣兲i for t
tD
1 I 1 1
1− e−⌳R sin ⌳I −− dtY ⬎ tI Ⰷ tCII, it follows that
␲ tC
t − tDII t − tA

冕 冉 冊
共A7兲
I
Cc1 =
1
tD
I 1 1 ⌳R共t兲 = − ln共t兲 + ␤共tI兲 共B6兲
1+ e −⌳R
sin ⌳I − dt
␲ t − tDI t − tDII
tC
I ⌳I共t兲 = ␲ − ␣ 共B7兲
where Y ⬅ Cc1 + Cc2.
where ␤共tI兲 is given by
For Cc2, a similar procedure reproduces Eq. 共10兲.

冕
tI
Appendix B: Derivation of ld and a Relationship Among 共t − tE兲
␤共tI兲 ⬅ ln共tI兲 −
tA, tDII, Cc1, and Cc2 CII
冑共t − tC 兲共t − tD 兲共t − tD 兲共t − tC 兲 dt.
I I II II
To seek a relationship among tA, tDII, Cc1, and Cc2, one consid-
Application of Eqs. 共B6兲 and 共B7兲 to the integrand of Eq. 共B3兲
ers an exact equation for the plate lower length ld first. It is deter-
yields
mined as follows:

z共+ ⬁兲 − z共tCII兲 = −
␩
2␲
共1 − sin ␣兲 冕 冉
+⬁
e−⌳
Cc1
+
Cc2
t − tDI t − tDII
integrand = eln共t兲e−␤共tI兲 sin共␲ − ␣兲 冉 Cc1
+
Cc2
t − tDI t − tDII
−
共Cc1 + Cc2兲
t − tA 冊
冋
tC

冊
II
Cc1tDI + Cc2tDII − 共Cc1 + Cc2兲tA
−
Cc1 + Cc2
dt. 共B1兲 = e−␤共tI兲 sin共␣兲
t

冉 冊册
t − tA
Let z共+⬁兲 − z共tCII兲 = x + iy and ⌳ = ⌳R + i⌳I. Then, y becomes 1
+O 共B8兲

冕 冉 冊
+⬁ t2
␩ Cc1 Cc2 Cc1 + Cc2
y= 共1 − sin ␣兲 e−⌳R sin ⌳I + − dt for t → + ⬁. For existence of Eq. 共B3兲, the above integrand should
2␲ t − tDI t − tDII t − tA
tC
II more rapidly decay than 1 / t as t → + ⬁. Therefore, the leading
共B2兲 order term of Eq. 共B8兲 should be zero. That is,
and finally ld共=y / sin ␣兲 is given by Cc1tDI + Cc2tDII − 共Cc1 + Cc2兲tA = 0 共B9兲

冕 冉
+⬁
␩ 共1 − sin ␣兲 Cc1 Cc2
ld = e−⌳R sin ⌳I + This gives the relation among tA, tDII, Cc1, and Cc2 as follows:
2␲ sin ␣ tC
t − tDI t − tDII

冊
II
Cc2tDII
Cc1 + Cc2 tA = 共B10兲
− dt. 共B3兲 Cc1 + Cc2
t − tA
Through a similar procedure, one can also determine lu. Now, where tDI = 0.
determine the functional form of integrand of Eq. 共B3兲 as t →
+ ⬁. For this, ⌳共t兲 should be considered first. ⌳共t兲 is given by

冕冑
t References
共t − tE兲
⌳共t兲 = − dt. 共B4兲 关1兴 Sarpkaya, T., 1959, “Oblique Impact of a Bounded Stream on a Plane
0 共t − tCI兲共t − tDI兲共t − tDII兲共t − tCII兲 Lamina,” J. Franklin Inst., 267共3兲, pp. 229–242.
关2兴 Sarpkaya, T., 1961, “Torque and Cavitation Characteristics of Butterfly
For large t共t ⬎ tI Ⰷ tCII ⬎ 0兲, Eq. 共B4兲 is rewritten as Valves,” ASME J. Appl. Mech., 28共4兲, pp. 511–518.
关3兴 Robertson, J. M., 1965, Hydrodynamics in Theory and Application, Prentice-

冕冑
tI Hall, Englewood Cliffs, NJ, Chap. 11.
共t − tE兲 关4兴 Hassenpflug, W. C., 1998, “Free-Streamlines,” Comput. Math. Appl., 36共1兲,
⌳共t兲 = − dt
0 共t − tCI兲共t − tDI兲共t − tDII兲共t − tCII兲 pp. 69–129.
关5兴 Ogawa, K., and Kimura, T., 1995, “Hydrodynamic Characteristics of a Butter-

冕冑
t fly Valve-Prediction of Torque Characteristics,” ISA Trans., 34共4兲, pp. 327–
共t − tE兲 333.
− dt 关6兴 Solliec, C., and Danbon, F., 1999, “Aerodynamic Torque Acting on a Butterfly
tI 共t − tCI兲共t − tDI兲共t − tDII兲共t − tCII兲 Valve. Comparison and Choice of a Torque Coefficient,” ASME J. Fluids Eng.,

冕
tI 121共4兲, pp. 914–917.
共t − tE兲 关7兴 Chuang, J. M., Gui, Q. Y., and Hsiung, C. C., 1993, “Numerical Computation
⬇ 共␲ − ␣兲i −
CII
冑共t − tC 兲共t − tD 兲共t − tD 兲共t − tC 兲
I I II II
of Schwarz-Christoffel Transformation for Simply Connected Unbounded Do-
main,” Comput. Methods Appl. Mech. Eng., 105共1兲, pp. 93–109.
关8兴 James, M. L., Smith, G. M., and Wolford, J. C., 1985, Applied Numerical
⫻dt − ln共t兲 + ln共tI兲 共B5兲 Methods for Digital Computation, Harper & Row, New York, Chap. 5.

Journal of Fluids Engineering JANUARY 2006, Vol. 128 / 195

DownloadedViewFrom:
publication[Link]
stats on 02/18/2018 Terms of Use: [Link]