GLOBE-VALVE FLOW COEFFICIENTS FOR VALVE METERS A. EDGAR KROLL' AND H. V. FAIRBANKS Rose Polytechnic Institute, Term Haute, Znd.
A
The conversion of a n ordinary commercial valve into a flowmeter is described. The valve flow coefficient, Ke, from which t h e rate of liquid flow through a valve may be calculated, is defined in the following modified form of the fundamental flow equation:
VALVE meter is an ordinary commercial valve converted into a flowmeter by attaching pressure connections upstream and downstream from the valve to a manometer or any other suitable pressure measuring device, as shown in Figure 1. The difference between the pressures upstream and downstream from the valve, caused by the smaller valve opening, can be measured by means of the manometer. The use of a valve as a flowmeter was recently suggested by one of the authors. Data were given on a single 2-inch globe valve which was calibrated and used successfully to determine and control the rate of flow in a pipe line (6). In continuation of this idea it seemed of interest (a) to obtain further information regarding the use of valves for determining and controlling the rate of liquid flow in pipe lines, (b) to determine valve flow coefficients for valves of different sizes, and (c) to investigate the feasibility of using commercial valves as flowmeters without calibration by virtue of predetermined flow coefficients. Investigations in this field appear to be lacking, since a search of the literature revealed no information relative to the use of valves as flowmeters. This paper describes the results obtained with various sizes of brws globe valves manufactured by the Crane Company.
K, = 1.496 ul,Jl/pdpr Valve flow coefficients for various valve settings on different sizes of Crane brass-globe bevel-seat brassdisk valves are presented, together with details of the method by which they were obtained. These coefficients were determined with Reynolds n u m bers, referred t o t h e pipe, ranging from 30,000 to 150,000. I t is shown t h a t the use of a globe valve as a flowmeter is practical. Valve flow coefficients and 1-inch valves for three different '/P, were found to have a maximum deviation of 5 % ; for l'/r-inch valves, flow coefficients had only a 3% maximum deviation. Use of these valve flow coefficients for computing rates of flow through other valves, of the same type and make, without calibration is feasible where the above maximum deviations are permissible. The valve flow coefficients presented agree within 3.0% with values calculated from data obtained by Corp and Ruble and by t h e Crane Company for loss of head through fully open, new Crane globe valves.
- - ,
APPARATUS A$D METAOD OF TESTING
Crane brass-globe valves, Xo. 1, with bevel seat, brass disk, and screwed ends were used; sizes were '/z to 11/, inches, inclusive. Runs were made on three different valves of each size. All the valves were new except one ll/,-inch and one I-inch valve, which were used but in good condition. The results obtained on the used valves were in good agreement with those on the new valves. Water temperature was 60' F. in all cases, with the flow upward through the valve orifice. Figure 2 is a diagram and Figure 3 is a photograph of the apparatus. In all tests the valve was preceded by fifteen pipe diam-
Pointer
I Preaent address, E. I. du Pont de Nemoura & Company, Ino., Childersburg, Ala.
,-Valve under t e s t
Figure 1 (Above).
Diagram of Valve Meter
Figure 2 (Right). Diagram of Apparatus
1
2:
1
L - 4 4 - d
588
Z N D U S T R I A L A N D E N G I N E E R I N G CHEMISTRY
Jun, 1945
589
TABLE I. SUMMAFLY OF CALCULATED RESULTSON GLOBE VALVES
No. of Turns Open
Valve Siie'. Inch1/1
I/( 1/ 1
1
Il/, 2
a
4*
1
Mean Value of Ks
Deviation, % Av. Max.
0.021 0.043 0.074 0.090 0.09s 0.100 0.100
4.4 2.1 1.0 3.3 1.0 0.4 0.7
6.3 3.1 1.4 4.8 1.4 0.7 1.0
0.062 0.108 0.207 0.248 0.283 0.326 0.341
4.9 1.7 0.9 3.0 2,s 3.1 2.3
6.1 2.5 1.3 4.8 6.0 4.9 3.4
0.072 0.150 0.278 0.361 0.423 0.470 0.507
2.2 1.0 1.2 2.0 1.8 1.9 1.2 1.9
3.2 2.5 1.8 3.1 2.8 2.9 1.8 2.9
0.s20
,
Three valves were wed for eaoh settin b Approximate. Different valves v a r i e h g h t l y in the number of turns to fully open.
Figure 3. Photograph of Apparatus Used in D e t e r m i n i n g G l o b e - V a l v e Flow Coefficients
DISCUSSION AND RESULTS
The working equation for .computing the actual rate of liquid flow through an orifice, adopted by the A.S.M.E. Special Research committee on fluid meters (8, 7), is
0
eters of straight galvanized steel pipe and was followed by fifteen pipe diameters of similar pipe and a gooseneck (1) in which the rise WM always greater than six pipe diameters. Some of t8he valves used are pictured in Figure 4. The rate of water discharge was measured a t various pressures and valve settings, taken as the number of turns open. For convenience in setting the number of turns open, only round valve handles were used and were equipped with a narrow strip of metal bent at a right angle, pointed on one end and having a hole in the other so that it could be fastened on the valve stem together with the valve handle as shown in Figures 1 and 3. The zero point was determined by first opening the valve wide, allowing water to flow for a few seconds, and then closing the valve slowly until the flow just stopped. The meral pointer was attached and the number of turns open were set from this point. The purpose of this method of determirring the zero point was twofold: The valve was flushed clean of a ~ foreign ~ y material which might obstruct the flow until the valve was fully opened, and any error due to slack in the valve stem was eliminated. The taps for determining the pressure drop across the valves were made by first welding on a '/r-inch half-coupling, SCI;wing in a short nipple, and then drilling holes with a '/,-inch drill (6) two and a half pipe diameters upstream and eight pipe diameters downstream from the valve. Great care was taken to remove all burrs from the edges of the holes on the inside wall of the pipe. These openings were connected to calibrated Bourdon pressure gages for large differential pressures and to a mercury manometer for small differential pressures. Readings on the pressure gages were taken to the nearest 0.5 pound per square inch on the upstream gage and to the nearest 0.2 pound on the downstream gage. The manometer readings were taken to the nearest 0.1 om. The quantity of water discharged was determined by weight on Fairbanks platform scales of 1000-pound capacity. In determining the rate of discharge, the time was measured by a Kodak timer. The Reynolds number, referred to the pipe, for all runs varied from 80,OOO to 150,000. The source of water supply for the investigation was a blow case which furnished the various pressures, ranging from 10 to 40 pounds per square inch. During any one run, constant p m u r e was maintained.
For this study Equat,ion 1 was modified t.o
whcre K., the valve flow coefficient, as defined here, differs from tho standard orifice flow coefficient (9)only in that it includes the cross-sectional area of the discharge opening and is not dinicnsionlcss. Simplifying Equation 2,
K
1.496 Ill
-- dZ3
Globe-valve flow coefficients calculated from data obtained in this study are shown graphically on logarithmic paper in Figure 5. These valve flow coefficients were calculated from Equation 3 using values obtained by measuring the rate of water discharge and differential pressure a t various valve settings. Table I summarizes the calculated results and shows the average and maximum deviation from the mean value of the flow coefficient. One valve reproduces the results of another, usually within less than 5% on the l/*-, 3/4-, nntl I-inch valves; results on the I'/c-inch valves varied lesq than 3%. The largest devia-
Figure 4. Photograph of Globe Valves
590
INDUSTRIAL AND ENGINEERING CHEMISTRY 60
40
Y
9f 30
c
20
he
P
$
Q
/0
b
8G 6 $
4
2.3 2
/
1 4
VAL1.E Figure 5.
L 2
2
/
3
4
rate of flow, w, is calculated from Equation 3. While it is known that there is little variation in orifice flow coefficicnts with Reynolds numbers above 30,000 ( T ) , it should be noted that the valve flow coefficients presented here were determined with Reynolds numbers, referred to the pipe, between 30,000 and 150,000. Figure 6 shows the flow characteristics of the globe valves used. The values for percentage of total flow and percentage of total turns ppen were calculated from the mean values of the flow coefficients given in Table I, assuming a constant pressure drop. Table I1 compares valve flow coefficients obtained in this study with those calculated from data obtained by Corp and Ruble (3') and the Crane Company (4) for loss of head through fully open Crane globe valves. The present data compare favorably for the 3/4and 1-inch valves tested by Corp and Ruble in 1919, being 0.5% higher and 2.9% lower, respectively. The '/=inch valve is 20.5% higher. This unfavorable agreement may be due to the fact that the '/Finch valves were tested by Corp and Ruble with Reynolds numbers less than 20,000. The Crane Company's value for the l/rinch valve flow coefficient was calculated from k = 10.0 in the equation H = kVz/2g given in the reference for globe valves (4). Agreement within 3.1% is shown here. SUMMARY AND CONCLUSION
S€TT/NG, N& O F TURNS OPrN
Globe-Valve Flow Coefficients a t Various Valve Settings
tiom were obtained at turn open, which were probably due to variations in the valve settings or flow characteristics at close throttling conditions. The valve settings are less critical as the opening increases. There is also danger of wiredrawing the seat when the flow is throttled down to I / , turn open. For these reasons use of the I/, turn open setting is not recommended. Even with the favorable agreement shown above, it should be pointed out that valve flow characteristics are influenced considerably by manufacturing variations. design details such as type of disk, contour of body, etc., and direction of flow through the valve orifice, so that use of the flow coefficients given here should be restricted to the type and make tested, since their use with other types and makes may lead to serious error. For example, valve flow coefficients determined for a composition disk valve varied as much as 30% from the brass-disk bevel-seat type. It is felt, however, that with consistent design and manufacture the average flow coefficients for the commonly used globe valves could be determined; this would make possible their use as flowmetere without calibration, provided the deviation and accuracy are stipulated. The valve flow coefficient curves given in Figure 5 may be used for computing rates of flow through these Crane valves without calibration, where the above maximum deviations are permissible. The pressure taps should be installed according to Figure 1, the direction of flow is upward through the orifice, and the weight
Vol. 37, No. 6
1. Valve flow coefficients have been given for various valve settings on Crane I/*-, a/,-, 1-, and 11/4-inchbrass-globe valves No. 1 with bevel seat and brass disk. The flow coefficients were determined with Reynolds numbers, referred to the pipe, ranging from 30,000 to 150,000. /A9
90
80 70
60
22 ? I
58 h
$
44 b 3u
2
q
20 /Q
0 0
/u
50 68 To 80 20 30 # P E R CENT OF TOTAL f L O W
90
lo4
Figure 6. Flow Characteristics of Globe Valves with Flow Upward through the Orifice and Based upon a Constant Pressure Drop
INDUSTRIAL AND ENGINEERING CHEMISTRY
June, 1945
s91
ACKNOWLEDGMENT
TABLE 11. VALVE FLOWCOEFFICI~NT~ CALCULATED FROM EQUATION 3, USINGDATAOBTAINEDBY CORPAND RUBLZ(8) AND THE CRANE COMPANY (4) FOR Loss OF HEADTHROUGH GLOBEVALVES~8 COMPARED TO COEFFICIENTS OBTAINEDIN THIS STUDY Fully Open Re for Valve Siae of: 1/1
Corp and Ruble (1922) Crane Co. Kroll and Fairbanks (1944) a
-
-
*/a in. 0.208
1 in. 0.861
o:lbb
o:iiio
Used valves at Reynolds numbers leas than 20,000. k 10.0 in II kVV2o. One wed valve.
b Calaulated from e
in.
0.088@ 0.097s 0.100
Thanks are due to B. E. Hunt for the drawings and to F. M. Lundgren who took the pictures of the valves and apparatus used.
A*
- -
NOMENCLATURE
cross-sectional a m of discharge opening, sq. ft. D = diameter of pipe, ft. = acceleration due t o avity, 32.17 (ft./sec.)/sec. = loss of head, feet of &id k: = coefficient, dimensionless K = flow coefficient, dimensionless K. KA2 valve flow coefficient, s ft. PI, p z = pressures a t upstream and 3ownstream pressure taps, respectively, lb./sq. ft. aPd = = Uerential pressure, lb./sq. in. 144. V = velocity of water, ft./sec. w = weight rate of discharge, Ib./sec. p 5 density, lb./cu. ft.
%
’ a
2. The ordinary commercial globe valves of the type and make used in this study may be used as flowmeters without calibration by use of the valve flow coefficient, K,, from which the rate of flow may be calculated, where a maximum error of 5% is permissible. 3. Greater precision with larger size valves is indicated by the fact that the maximum deviation for the l’/r-inch v@lveSwas only
3%. 4. The principal advantages of a valve meter are: (a) It may be used aa an adjustable orifice for widely varying conditions of flow; (b) the valve is already in the pipe line and hence there is no increased resistance to flow; (c) little initial cost is incurred by converting a valve into a flowmeter; and ( d ) the upkeep is small.
LITERATURE CITED
(1) Am. Boo. Meoh. Engrs., “Flow Measurement”, p. 45, par. 121 (1940). (2) Am, Boo. Meoh. Engrs., “Fluid Meters Report”, 4th ed., p. 48, par. 155 (1937). (3) Corp and Ruble, Univ. Wisconsin Eng. Expt.Sta., Bull. 9, No. 1 (1922). (4) Crane Co.,Catdog 41,p. 630 (1941). (5) Oeaa, Louis, Instrumatation, 1, No. 2,26 (1944). (6) Kroll. A. E..C h . & Met. Em.. 51, No. 7. 114 (1944). (7j Perry, J. H., Chemical Enginiers Handbook, ind ed., p. 848. New York,McGraw-HiU Book Go., 1941.
Interchain Order and Orientation in Cellulose Esters-Correction A
B
C
D
An unfortunate error occurred in connection with Figure 6 of this article by W. 0. Baker in the March, 1945, issue of INDUSTRIAL AND ENQINP)ERINQ C ~ M I S T R Y ,page 252. Each x-ray diagram of this figure rn printed in the original article should be considered as rotated by 90’ so that the fiber axis, shown horiaontally, becomes vertical. Poor reproduction, due to the quality of paper now available, destroyed the diagrams; Table I11 (page 253) gives the actual features. Figure 5 is reproduced here, placed and captioned correctly.
Figure 5.
X-Ray Diagrams of Starch Triestere
A . Tapiwm atamh triawtate, unoriented B Comstuoh triloetate drawn u n i U I d y C: T8plocr .taroh tribu&ate, unoriented D . Potato at& tripmapionate, oriented
