The effects of chip-shaped solids on valve head loss characteristics by David Allan Johnson A thesis submitted to the Graduate Faculty in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE in Civil Engineering Montana State University © Copyright by David Allan Johnson (1968) Abstract: The main purpose of this study was to determine the head loss characteristics of different valves in a pipe line carrying a mixture of water and chip-shaped solids. Since the head loss caused by a valve is given by HL = KL (v^2)/(2g), this study involved determining values of the loss coefficient, KL, for each valve at different flow conditions. Observations were also made on the efficiency of each valve with respect to use in a solid-liquid pipe line. The following four valves were tested in this study: ball valve, plug valve, v-ball valve, and pinch valve. Tests were conducted on each valve for various closures at velocities of 4, 6, 8, and 10 fps with chip concentrations of 0, 10, and 20 percent. The loss coefficient was observed to (1) be approximately constant with respect to velocity at velocities greater than 6 fps for a given closure and percent concentration, (2) increase with increasing chip concentration, and (3) increase with increasing valve closure. For design purposes, empirical relationships of the following forms were derived for each valv6: ((KL)c)/((KL)0) = e^bC and ((KL)c)/((KL)0) = 1+mC These equations give the loss coefficient (KL)C for a given concentration as a function of the concentration C and the clear water loss coefficient (KL)0 for a given valve closure. THE E F F E C T S OF C H I P - S H A P E D SOLIDS ON VALVE HEAD LOSS C H A R A C T E R I S T I C S > by DAVID A L L A N J OH NSON A thesis sub mitt ed to the Gra duat e Faculty in partial ful fill ment of the req uire ments for the degree of MASTER OF SCIENCE in Civil E n g i n e e r i n g Approved: H e a d , Major D e p a r W e n t C h a i r m a n , E x a m i n i n g Com mittee Gr a d u a te^Dean MON T A N A STATE U N I V E R S I T Y Bozeman, Montana December, 1968 iii ACKNOWLEDGEMENT This study was one phase of a project inv es t i g a t i n g the hyd raul ics of t r a n s p o r t i n g wood chips by pipe line spo ns o r e d by the Int er m o u n t a i n Forest and Range Exp erim ent Station, U. S 0 Forest Service, and the Department of Civil E n g i n e e r i n g and E n g i n e e r i n g Mechanics, Montana State Un i v e r ­ sity. The valves used in this study were provided by: Rockwell M a n u f a c t u r i n g Company Grove Valve and R e g u l a t o r Company Fisher G o v erno r Company Farris Flexible Valve Corporation. The author wishes to express his gratitude to Dr. Wil l i a m A. Hunt, who provided the guidance for this s t u d y . A special thanks is ext ende d to Mr. Ronald Schmidt, who p r o ­ vided con s i d e r a b l e effort and advice in the app aratus con­ str uction and data col lect ing phases of this project. The a u t h o r ’s a p p r e c i a t i o n is also e x t e n d e d to Mr,. Gary Hen drix for his ass ista nce in this study. Notable contrib ution s to this study were made by members of the Mechanical E n g i n e e r i n g Dep artm ent mac hine shop at Montana State Uni vers ity and mem bers of the stgff of the Montana State U n i vers ity Computing Center, R e c o g n i t i o n is also due to und ergr aduat e assistants, Mr. Clyde Wolf, M r 0 Larry Neal, and Mr. Tom Hedges, who were r e s p o n s i b l e for r e d u c i n g data from film strips to F O R T R A N coding s h e e t s . Special r e c o g ni tion is extended to the author's wife, P a t , for her pat ience and effort spent in the typing and p r o o f r e a d i n g of the manuscript. TABLE OF CONTENTS Page List of Tables ............ . . . . . . . . . . . . . . List of Figures . . . . . . . , ............ . . vii . . . . . viii List of S y m b o l s . . . . . . . . . . . . . . . . . . . . . . . . . Abs trac t I. Int r o d u c t i o n A. B. C. II. III. VII. .. I Ba c k g r o u n d mat e r i a l . . . . . . . . . . . . . . Obj ecti ves of s t u d y ......... L i t e r a t u r e review . . . . . . . . . . . . . . . . . . . . I 2 4 7 A. B. . . . . . . 7 Basic head loss m e c h a n i s m . . . . . . . . Effects of valve closure, chip concent ratio n, and vel ocit y E x p e r i m e n t a l Methods . . . . . . . . . . . . . . . . 8 13 D e t e r m i n a t i o n of loss coe ffic ient . . . . . D e t e r m i n a t i o n of valve areas . . . . . . . . . D e t e r m i n a t i o n of velocity and con cent ratio n . . App arat us Des c r i p t i o n . . . . . . . . . 13 16 18 . 21 Ope ra t i o n of the pipe line system . . . . . . . System components . . . . . . . . . . . . . . . Test section and components . . . . . . . . . . . . Valves tested . , . .. . . . . . . . . . . . . . . . . . . . . . . . 21 23 26 31 Test Pro cedu re A. B. C. VI. .. . xiii . . . . . . A. B, C. D, V. . .................. . . Theory of Head Losses Caused by Valves A. B. C. IV. . . . .. . . . . . . . . . . . . x .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34 P r e p a r a t i o n of equ ipment ........ . . . . . . 34 Data col lect ion procedure . . . . . . . . . . . 36 Data r e d u c t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Data Ana lysis ... . . . . . . . . . . 40 A. B. Computer operati ons . . . . . . . . . . . . . . 40 P r e s e n t a t i o n of data ...................... 44 C. D. Flow pictures . . ...................... . . . . Summary of valve c h a r a ct erist ics . . . . . . . Co n c l u s i o n s and R e c o m m e n d a t i o n s ............... . 52 55 57 vi A p p endi ces A. B. C. D. E. . . . . . . . .......... , ................ Error Ana lysi s ........................ . . M a n o met er Board . . . . . . . . . . . . . . . . S t a t i st ical Equ ations .............. Com pute r P r o g r a m . . . . . . . . . . . . . . . . . . . Summary of C o m pute d Results ................ Li t e r a t u r e Cited . . . . . .................. . . . . . . 60 61 65 70 78. 88 92 vi i LIST OF TABLES Table I. II. III. Page Summary of Com pute d R e s u l t s , Emp ir i c a l Con stants for Valve C h a r a c t e r i s t i c s V-Ball Valve . . . , . 45 Equ atio ns . . . . . . . 50 .............. . . . . . . . 56 IV. Summary of C o m pute d Results, Plug Valve . . . . . . 88 V. Summary of C o m p u t e d Results, Ball Valve . . . . . . 89 VI. Sum mary of C o m p u t e d Results, Pinch Valve 90 VII. . . . . . Sum mary of C o m p u t e d Results, Gate Valve . . . . . . 91 viii LIST OF FIGURES Page Figure 1. Grade lines sho w i n g effect of constri ction .. . . . 2. Pipe and typical hyd raul ic grade line . . . . . . . 14 3. Plot of dim en s i o n l e s s grade line . . . . . . . . . . 15 4. Plot of ind icator rea ding versus percent closed 19 5. Sch emat ic of pipe line system 6. Control console 7. Plan view of test s e c t i o n ......... .... . . . . . 27 8. Loc atio n of pre ssure taps intest .. . . . 28 9. Typical pre ssure tap c o n s tru ction . . 9 . . . . . . . . . . . 22 .. . . . . . . . . . . . . . ................... 24 section .............. . 30 10. Dif f e r e n t i a l m a n o m e t e r b o a r d .............. 11. S c h e mat ic dia gram of valves tes ted . . . . . . . . . 32 12. F O R T R A N coding sheet and sample data . . . . . . . . 41 13. Sample computer output . . . . . . . . ............ 43 14. Plot of versus vel ocity .............. . . . . . . . . . . 46 15. Plot of versus c o n c e n t r a t i o n . . . . . . . . . . 47 16. Plot of (Kr ) / ( K r ) versus c o n c e nt ratio n . . . . . . 49 17. Plot of (K l ) q versus percent closure . . . . . . . . 51 18. Pic ture of sliding bed . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 19. T u r b u l e n c e d o w n s t r e a m from plug v a l v e , 4.4 percent closed . . . .................... . . 53 20. 21. .. . . 29 . . . . . . T u r b u l e n c e d o w n s t r e a m from plug valve, 50. 0 percent closed . . ............ . . ........... 54 Pic ture of b e g i n n i n g of " p l u g - u p ” 54 . . . . . . . . . ix 22. S c h e m a t i c of d i f f ere ntial m a n o m e t e r board . . . . . 23. Sample linear t w o - d i m e n s i o n a l data . . . . . . . . . . . . . . 65 „ 70 X LIST OF SYMBOLS A - Any c r o s s - s e c t i o n a l a r e a . a - S t r a i g h t liffe y-intercept. b - Slopd of straight line. C - V o l u m e t r i c chip concentration. CVAL - Valve dlosure in percent of constriction. COEF - C o r r e l a t i o n coefficient. CCl^ - Carbon tet rachloride. D - Pipe diameter. DPHD - Dif f e r e n t i a l pre ssure head between tap I and any . other tap. E - Sum of s qu ared e r r o r s . e - Base of Nap erian logarithms. E^ - Energy loss. EVAR - E s t imat e of var ianc e about regression. f - F r i ctio n factor. fps - Feet per second. g - Gravitational acceleration. gpm - Gallons per minute. Hl - Valve head loss. Hg - Mercury. HGL - H y d r a u l i c grade line. ID - Inside dia m e t e r of any pipe. Kl - Valve head loss c o e f f i c i e n t . L - Length of any reach of pipe. xi mm - Millimeter. n - Number of observations. OD - Outside dia mete r of pipe. P - Pressure. pcf - Pounds per cubic foot. Q - Flow rate in g p m . S - Spe cific gravity of any fluid. S - St a t i s t i c a l variance. V - Nominal v e l ocit y in fps. XVAL - X - d i s t a n c e c o r r e s p o n d i n g to center of valve in test s e c t i o n . X - Distance from first pre ssure tap in test section to any other t a p . y - Level of fluid in m a n o m e t e r tubes. Z - Height of water above fluid in m a n o m e t e r tubes. A - Dif fere nce in any rea ding or measurement. - Spe cific weight in pcf of any fluid. - T h e o r e t i c a l mean of infinite population. > - Gre ater t h a n . C - Any + error. b p/d* -- Change in pre ssur e with res pect to distance. Subscripts i - Pressure tap n u m b e r . u - Upstream. d Downstream. xi i v - Valve area. p - Pipe a r e a . m - Mixture flow rate. w - Clear water supply flow rate. f - Fluid flowing. c - Carbon t e t r a c h l o r i d e . m - Mercury. n - Some pre ssure tap d o w n str eam of tap I. o - Loss coe ffic ient for clear water, c - Loss coe ffic ient for a given concentration. xiii ABSTRACT The main purpose of this study was to det ermine the head loss c h a r a c t e r i s t i c s of dif ferent valves in a pipe line car ryin g a mi x t u r e of water and ch i p - s h a p e d solids. Since 2 the head loss caused by a valve is given by , this study inv olved d e t e r m i n i n g values of the loss c o e f f i c i e n t „ K^, for each valve at different flow c o n d i t i o n s . O bs erva tions were also made on the e f f icie ncy of each valve with respect to use in a s o l i d - l i q u i d pipe line. The fol lo w i n g four valves were tested in this study: ball valve, plug valve, v-ball valve, and pinch valve. Tests were con duct ed on each valve for various closures at v e l o c ­ ities of 4 , 6, 8 , and 10 fps with chip c o n c ent ratio ns of 0 , 10, and 20 p e r c e n t . The loss coe ffic ient was o b s erve d to (I) be a p p r o xi matel y constant with respect to vel ocity at velocities gre ater than 6 fps for a given closure and percent concentration, (2 ) increase with in c r e a s i n g chip concentration, and (3) increase with i n c r eas ing valve closure. For design p u r p o s e s , emp irical r e l a t i o n s h i p s of the fol lowing forms were der ived for each v a l v 6 : c k L-1C bC c k L-1C r — r— = e and 77;— %— = I + mC. CKL 0 '. CKL 0 These equ ations give the loss coeffic ient (K^ )c for a given c o n c e n t r a t i o n as a function of the con cent ratio n C and the clear water loss co e f f i c i e n t (Kr ) for a given valve closure. CHAPTER I INTRODUCTION A. B a c k g r o u n d mat eria l The t r a n s p o r t a t i o n of solids by pipe line is not n e w . For man y years chemical, mining, and dredging operations have inc luded sho rt-h aul slurry pipe lines. years, however, In the past few inc reas ed att ention has been given to the hyd raul ics of t r a n s p o r t i n g solids by pipe lines over longer distances. Cur rently, con side rable research is being done on solids pipe lines in the United States, Canada, and other fdreign countries. Also, several long distance pipe line t r a n s p o r t a t i o n systems have been built and ope rated s u c c e s s ­ fully . It is g e n e r a l l y con clud ed that because of continuous ope rati on and low ope ra t i n g costs, pipe lines can compete e c o n o m i c a l l y with other exi s t i n g methods. Govatos Zandi and (12) discuss solids pipe lines and give 30 ref erences to articles d i s c u s s i n g solids t r a n s p o r t a t i o n and industrial pipe line installations. A f e a s i bi lity study in 1965 by Hunt (4) showed that a pipe line system could also be used to convey woo d c h i p s , eco no m i c a l l y , from forest to pulp mill. As a result, a res e a r c h pro ject was ini tiated at Mon tana State Univers ity to inv esti gate the hyd raul ics of such a t r a n s p o r t a t i o n system The project was spo ns o r e d through a cooperative agreement by the I n t e r m o u n t a i n Forest and Range Exp erim ent Station, —2" United States Forest Service, and the Department of Civil E n g i n e e r i n g and E n g i n e e r i n g Mec hanics at Montana State University. Up to this date studies have been completed under the above c o o p e ra tive agreement on the fol lo w i n g s ub­ jects ; specific gra vity of sat urat ed wood c h i p s , pump p e r ­ formance c h a r a ct erist ics+ head losses in a x i - s y m m e t r ic pipe expansions, head losses of a m o d i f i e d gate valve, pipe line frictio nal l o s s e s , and a water h am mer analysis of a high pressure chip injection system. A study of the head loss cha rac t e r i s t i c s of four dif fere nt types of valves in a pipe line carrying a mixture of water and chips is pre sent ed in this paper. Results of this study will pro vide design informa tion on valve head losses and hence power req u i r e m e n t s for con stru ction of a ful l-sc ale ope ra t i n g system. Tests were conducted on a v-ball valve, a ball valve, a plug valve, and a pinch valve. Results of the m o d i f i e d gate valve study m e n t i o n e d above are also pre se n t e d in this paper. The tests were per fo r m e d in the civil e n g i n e e r i n g section of Ryon L a b orat ory on the campus of Mon tana State University. B. Obj ecti ves of study Flow thr ough a valve is a r e l a t i v e l y com plex phenomenon and head losses gen er a l l y cannot be det ermi ned analytically. Because of the variety of flow pas sage ways in different valves, head loss cha rac t e r i s t i c s must be d e t e rmi ned -3- experimentalIy. The energy equ atio n for a liquid written between points ups trea m (subscript I) and d o w nstr eam (subscript 2) of a valve is T + 2i + Z. Pg v2 z0 +' *''**"*a losses i_2 if +' --2g +' "2 I where p is pre ssure in psf, (I) ft is the specific weight of the fluid flowing in pcf, v is the nominal velocity, z is the e l e v a t i o n above a datum, g is the g r a v ita tiona l acceleration, and the loss term includes the loss due to pipe line friction and the loss caused by the valve. For a horizontal, constant dia mete r pipe this sim plif ies to l o s s e s 1_2 Pl ~_J2 (2 ) a where p^ and p^ must be d e t ermi ned experim ental ly. losses for a given flow condition are determined, Once the it is des irab le to equate them to a fun ction of some commonly known par am e t e r such that head losses may be e s t i m a t e d for similar flow conditions. The D a r c y - W e i s b a c h equation, giving the head loss as a fun ction of velocity, was d e v e l o p e d for pipe line friction losses. Similarly, head losses caused by valves and other fittings are e x p r e s s e d as lL 2g where K. is an emp iric al head loss coefficient (3) —4— The main obj ective of this study, t h e n , was to determine exp e r i m e n t a l Iy the values of the loss coefficient, each valve at dif ferent flow conditions. , for Velocity, valve closure, and chip con ce n t r a t i o n were varied and the effects on determined. Flat r e c t a n g u l a r plastic chips with a spe cifi c gravity sli ghtly greater than that of water were used to simulate wood chips. In add ition to head loss data, o b s e r va tions were also made on the m e c hani cal e f f icie ncy of each valve with res pect to use in a s o l i d -l iquid pipe line. C. L i t e r a t u r e review A literature survey was con duct ed in an attempt to find inf orma tion on head losses caused by valves in s o l i d-l iquid lines for Com pari son with values obt aine d in this study. V i r t ual ly no i n f o r ma tion was found on valve head losses in s o l i d - l i q u i d pipe lines. In fact, very little recent inf or­ ma t i o n was found on valve head losses in general. Pub lished ana lyti cal and e x p e r i m e n t a l inf orma tion on valve head losses tends to be the result of work done several years ago and on the older types of valves. More recent work has apparently been done by valve m a n u f a c t u r e r s and pri vate firms and not rel e a s e d for public consumption. The fol lowi ng inf orma tion was found on the effect of closure, concent ratio n, and vel o c i t y on the loss coefficient. Rouse (8 ) gives the fol lowi ng loss coefficients, , for various valves at dif ferent valve closures for clear water flow. —5— Gate Valve Fully Open 3/4 Open 1/2 Open 1/4 Open Plua Globe or Stop Valve 0. 19 I .15 5.6 24.0 D i a p hra gm Valve Fully Open 3/4 Open 1/2 Open I /4 Open Fully Open 3/4 Open 1/2 Open 1/4 Open 4.0 4.6 6.4 78.0 Plua Valve with Scr ewed Ends 2.3 2 *6 4.3 21.0 Fully Open 90% Open 80% Open 70% Open 0.77 2.86 9.6 28.0 No i n f o r ma tion about the effect of con cent ratio n on the head loss c o e f f ic ient was found. However, Charley (I) did some work on the effect of ch i p - s h a p e d solids on head losses in pipe expansions. He found t h a t , for flow rates greater than 226 g p m , the loss' coeffic ient for a pipe exp ansi on app e a r e d to dec rease with an increase in chip concentration. The Crane Com pany (2) implies that the loss coefficient, K l ,. does not vary with Reynolds number and hence velocity for fully tur bulent clear water flow. The valve head loss 2 equ a t i o n H l L Tj^. v2 H l = f-jj L 2g is similar to Darcy's equation TThus h u s , a log-log plot of loss coefficient. coefficient, K 1 l, versus R e y n o l d s number would be e x p e c t e d to yield a graph similar to the w e l l - k n o w n Moody dia g r a m for pipe friction losses. The Crane Com pany claims that in the f u l l y - d e v e l o p e d tur bulent region of a log-log plot of K l versus R g or velocity, the loss coefficient, K l , for clear water would be constant - 6- as is the friction factor for M o o d y ’s diagram. i n f o r ma tion was given on how flow of water and s o l i d s . However, no varies with vel o c i t y for CHAPTER II T H E O R Y OF HEA D LOSSES CA U S E D BY VALVES Some knowledge of the cha ract erist ics of flow through a valve is nec essa ry before d e v e l o p i n g the exp erim ental methods used to det ermine the loss c h a r a ct erist ics of the various valves. Flow of a fluid through a valve is a rel ativ ely c om plex phenomenon. The addition of solid particles to the fluid further complic ates the c h a r a ct erist ics of the flow. As ind icat ed in the p r e c e d i n g c h a p t e r , no pub lish ed t h e o r e t ­ ical work on flow of s o l i d - l i q u i d mix tures through valves was found. Thus* the fol lowi ng d i s cuss ion gives a general des­ cription of (I) the head loss mec hani sm, as d e v e lop ed by app l i c a t i o n of the basic laws of fluid mec hanics, and (2) the effects of valve closure, chip concentration, and velocity on head losses as r e p o r t e d in rel ated sol id-l iquid tra nsp o r t a t i o n studies, A. Basic head loss m e c h a n i s m A valve is e s s e n t i a l l y an irr egul ar c o n s t ri ction in a pipe line. For any given flow condition, the e q u a t i o n of con tinu ity requires an acc e l e r a t i o n of the fluid through the con s t r i c t i o n and a dec e l e r a t i o n d o w n s t r e a m from the valve. At the same time, the energy equ atio n requires a reduction in pressure, head to coincide with the inc reasing velocity through the valve. In other words*, the con stri ction caused by the valve results in a con vers ion of pressure head or -8- pot ential energy to vel ocit y head or kinetic energy. In the region d o w n s t r e a m from the v a l v e , the energy con vers ion is r e v e r s e d and pre ssure head is r e c o v e r e d as the fluid slows and again assumes uni form flow conditions. Howeverbecause of energy losses * this t r a n s f o r m a t i o n is less than IOO percent eff icie nt and the initial pressure head is not com pletely recovered. A certain amount of energy is dis sipa ted by turbulence and friction in the valve and in the region d o w n s t r e a m from the valve. Since the continuity equ atio n requires the v e l o c ­ ity head to return to its initial value during the recovery process, this d i s s i p a t e d energy is lost at the exp ense of pre ssure head o n l y . loss. The r e s u l t i n g Ap/fr is known as the head The m a g n i t u d e of this head loss is aff ected by valve closure, chip c o n c e n t r a t i o n , and vel ocity as dis cu s s e d below. A section of pipe line with a valve is shown in Fig. I with the c o r r e s p o n d i n g vel ocit y ^ hydraulic, and energy grade lines. B. Effects of valve closure, chip con centration, and vel ocity As seen in the above discussion, valve head losses are a result of energy d i s s i p a t e d by tur bule nce and friction. Spe cifi c causes of this energy d i s s i pa tion are given below. The con s t r i c t i o n or closure of the valve causes turbule nce in three ways. The r e d u c t i o n in area through the valve causes an a c c e l e r a t i o n of the fluid. As a result of the increased -9- velocity, zones of high shear form which in turn disturb the vel ocity d i s t r i b u t i o n or cause turbulence. con s t r i c t i o n is increased, hence t u r bule nce increase. As the degree of the velocity, shear forces, and The geo metr ic shape of the p a s sage way through the valve also aids in the formation of tur bule nce by d i s r u p t i n g the streaml ines of flow in the pipe line. Tu r b u l e n c e also forms in the region just dow nstr eam from the valve as a result of the constriction. As seen in Fig. I, the c o n s tri ction in the pipe line results in a positive pressure gradient valve. ^ ^ >0 in the region dow nstr eam from the A c c o r d i n g to V. L . Str eeter (IO) this positive pressure gradient induces boundary layer separation along the Total Energy Grade Line £_ + V_ Hyd raulic Grade Line Velocity Grade Line Fig. I. Grade lines showing effect of constri ction -10- pipe walls in this region which results in back flow and the for mati on of add itio nal eddies. The turbulence caused by the valve is e v e ntua lly damped out and the flow is e s s e n ­ tially u ni form again at the point where the h y d r aul ic grade line becomes a slo ping straight line. The dis tance between the valve and this point is r e f erre d to as the settling length. The set tlin g length increases with the amount of t u r bule nce or energy dissipated. A c c o r d i n g to the above discussion, then, tur bule nce and the c o r r e s p o n d i n g di s s i p a t i o n of energy are direct Iy related to the degree of con s t r i c t i o n caused by the valve. Therefore, head loss and hence the loss coeffic ient can be pre dict ed to increase as the valve closure is increased. This is in agr eeme nt with the clear water loss coefficients given by Rouse (8). The presence of c h i p - sh aped solids in the fluid also affects the flow cha ract eristics. H o w e v e r , the effect of solids on valve head losses is not nearly as obvious as that of valve closure. Solid particles have both ben efic ial and det rime ntal effects. A c c o r d i n g to the results of experiments on the flow of s o l i d - l i q u i d mix ture s in pipes by J. W. Daily and T „ K . Chu as r e p o r t e d by M. Hino (3), app ro x i m a t e l y neu tral ly buo yant particles cause (I) an increase in the t u r bule nce int ensity over that for clear water, and (2) a r e d u cti on in the set tlin g length. In other words, solid — 11 — par ticles break up large eddies into smaller, high velocity eddies which, turbulence. in effect, increase the intensity of the A l t h o u g h the int ensity of turbulence is increased, the red u c e d set tlin g length caused by the presence of solid par ticles indicates that the t u r bule nce is damped out quicker than for clear water. This in turn indicates that less energy is d i s s i p a t e d and hence, the head losses are smaller. A c c o r d i n g to B . W. C ha rley (I), these facts account for the o b s erve d decrease in the head losses caused by a x i - s y m m e t r ic pipe e x p ansi ons for flow of a s o l i d - l i q u i d mixture. However, S . L , Soo (9) reports that because of physical contact bet ween the particles and the walls, the addition of solid par ticles to the flow causes an increase in friction and flow r e s i s t a n c e along the bou ndar ies of a conduit. Energy di s s i p a t i o n and hence head losses wou ld app aren tly increase with an i n c reas ing volume of par ticles or chips in the fluid flowing. In a p a r t i a l l y closed valve, chips tend to collect in pockets and crevices, r e s u l t i n g in a sort of "semi-p lugge d" con dition and inc reas ed flow resistance. Fri ction between the chips and t u r b ule nce caused by the relative mo t i o n of the chips also dis sipa te small amounts of e n e r g y . In summary, t h e n , the pre sence of chips appears to cause a red uc t i o n in head losses by i n c reas ing the tur bule nce intensity which in turn reduces the set t l i n g length. On the other h a n d , chips ap p a r e n t l y also increase the head losses because of energy — 12 — d i s s i p a t e d by inc reas ed friction and flow resistance. Thus , without further evidence, a theoretical pre dict ion of the net effect of chips on the head loss cannot be made. P r e l i m i n a r y analysis of the data from the previously m e n t i o n e d gate valve tests indicates that for a given v e l o c ­ ity and valve closure, chi p-sh aped solids increase the head loss over that for clear w a t e r , contrary to the conclusions of Charley (I). The refore, on the basis of trends shown by the gate valve tests, it can be pre di c t e d that the net effect of ch i p - s h a p e d solids should be that of inc reas ing the head loss and the loss c o e f f i c i e n t , „ As r e p o r t e d by Rouse (8 ) and others, the clear water valve head loss coefficient,, K^, is generally ass umed to be constant for any velocity. In other w o r d s , the head loss increases with velocity but the ratio remains approxiv2/2g m a t e Iy constant. In conclusion, t h e n , it can be pre dicted that the valve head loss coe ffic ient should increase with i n c reas ing valve closure and chip con ce n t r a t i o n but remain constant with i n c reas ing velocity. CHAPTER III E X P E R I M E N T A L METHODS In order to ach ieve the objectives of this study, the fol lo w i n g e x p e r i m e n t a l pro cedu res were followed: m i n a t i o n of the head loss coefficient, (I) d e t e r ­ , (2) d e t e r mi natio n of valve areas, and (3) d e t e r m i n a t i o n of the m i x t u r e velocity and percent concentration. A. D e t e r m i n a t i o n of loss coefficient, The head loss, in feet of fluid f l o w i n g , o c c u r r i n g at a valve in a pipe line is m e a s u r e d as the differe nce between the ups trea m and d o w n s t r e a m portions of the h y d r a u l i c grade line when both lines are pro je c t e d to a point over the valve. Fig, 2 shows a typical h y d r aul ic grade line, H G L , and the head loss, H^, caused by a valve. To det ermine the head loss and hence the effect of velocity, closure, and c o n c e nt ratio n on the loss coe fficient, K^,. the hyd ra u l i c grade lines for the various flow conditions must be est ablished. A series of pre ssur e taps a l o n g the test section, as d i s c u s s e d in Cha pter IV, and a d i f f e re ntial m a n o m e t e r board (Ap pend ix B) pro v i d e d information needed to est ab l i s h the h y d r a u l i c grade line. The m a n o m e t e r board gave the pressure head differences, DPHD., between the first pre ssur e tap in the test section and the other pre ssur e taps at distances x^ d o w n s t r e a m of the first tap as shown in Fig. 2. , A set of t w o - d i m e n s i o n a l data (x^, D P H D .) was thus obtained where the -14- subscript i refers to the pressure tap number. Plo ttin g this data would give a curve with the exact shape of the true hyd ra u l i c grade line but upside down and with a y - i ntercept of z e r o . Div iding each of the d i f f e re ntial pressure heads, D P H D ^ , by the vel ocity head gives a set of dim ensi onles s differential pressure heads, DIMPH., which are dependent on distance along the pipe line, . The data (x_, D I M P H ^ ) thus establi shes an upside d o w n , d i m e n si onles s hyd raulic grade line. However, the d i f f e re ntial pressure heads dis cuss ed above are the head losses occ urri ng between the points indicated. P r o ject ing the straight line portions of the dimensi onles s hydraulic grade line as shown in Fig. 3 to a point coi ncid ing with the center of the valve (x = X V A L ) gives the loss coefficient, DPHD Fig. 2. Pipe and typical hyd raul ic grade line , -15- f or the valve because K A n /x I. , as pre viou sly shown. Once the data (x_, DIMPH.) was calculated, the loss coe ffic ient could be d e t ermi ned by using linear reg ress ion met hods to fit straight lines to the data for the linear portions of the dim e n s i o n l e s s grade line. These fitted lines are given by the equ ation y = a + bx where the variables (x, y) c o r resp ond to the data (x., DIMPH^), a is the y-intercept, and b is the slope of the line. By using linear regression, a and b were d e t e r m i n e d such that the establi shed lines were the lines of "best fit." The equations used to cal culate a and b for each line are dev eloped in A p p e n d i x C . The loss coe fficient, , is then given by the difference bet ween the ups tream and do w n s t r e a m lines or, in other words, the dif fere nce between the y values given by the equations AP/tf v2/2g x - d ista nce dow nstr eam Fig. 3. Plot of d i m e nsi onles s grade line — 16— y = a + bx at some common x. The re f o r e f K L = y d “ y u = ' ( a + b x ) d “ (a+ bx)u (4) where x = XVAL and the subscripts d and u refer to the d o w n s t r e a m and ups t r e a m portions of the d i m e nsi onles s grade line as shown in Fig. 3. A computer program (Ap pendix D) was d e v e l o p e d to per form the above operations. T h e o r e t i c a l l y the ups trea m y-i ntercept, a^, should be zero and the slopes of both lines should be equal. However, the r e g r e s s i o n analysis gave values of a^ that varied slightly from zero and slopes that were not quite the same. A d i s cuss ion of the r e s u l t i n g errors is given in A p p e n d i x A. B. D e t e r m i n a t i o n of valve areas Because the head loss varies with the amount of con­ str iction in the pipe line caused by the v a l v e , some means of gau g i n g the closure of each valve was needed before testing could begin. Valve closure is usually exp ressed as a p er­ centage of stem travel or angle of turn of the ope rating m e c h a n i s m on the valve. In this study valve closure was c a l cula ted as a fun ction of the area of the ope ning through the valve and the c r o s s -s ectio nal area of the pipe line. In other w o r d s , the term valve closure indicates the percent age of con s t r i c t i o n in the pipe line caused by the valve. valve closure in p e r c e n t , C V A L , is exp ressed as The -17- CVAL = 100 - -Tjl (100) P where (5) is the area of the opening of the valve at a given position, and A^ is the area of the 4-in. diameter pipe. An indicator, as dis cu s s e d in Chapter IV, was ins talled on the o p e r ati ng m e c h a n i s m of each valve for gauging the closure. The scale on the ind icator ranged from 16 at the wide open pos ition of the operator to zero for the fully closed position. In order to det ermi ne the valve closure cor r e s p o n d i n g to a given indicator reading, a plot of indicator r e a d i n g versus closure was made for each valve. This plot was made from data obt aine d by ca l c u l a t i n g the areas and p e r cent age closed of a few openings c o r r e s p o n d i n g to equally spaced positions on the ind icator scale between 16 and the 100 per cent closed reading. The areas of the openings thr ough the various valves were d e t e r m i n e d in dif fere nt ways. Pictures were taken of the v-ball and ball valves from whi ch drawings of the areas of ope ning were made. These drawings were then pla n i m e t e r e d to det ermi ne the areas in square feet. The openings in the plug and pinch valves were not e n t i r e l y . v i s i b l e through the inlet of the valve and pictures could not be taken. Instead, calipers were used to obtain the dimensions of the openings and a p p r o x i m a t e sketches were made. Areas of the openings — 18— were then c a l cula ted by p l a n i m e t e r i n g the sketches. The s i m plic ity of the shape of the o pe ning in the gate valve made it possible to calculate the areas mat hema tical ly. The plot of ind icat or r ea ding versus per cent age closed for each valve is shown in Fig. 4. From this plot, the indicator setting for any closure chosen for tes ting could be determined. The curves for the v - b a 11, plug, and pinch valves in Fig. 4 do not reach 0 percent closed at an indicator rea d i n g of 16 because the areas through the valves when in the wide open pos ition are less than the area of the 4-in. pipe. Thus, these valves cause a constri ction of the pipe line when in the wide open position. The curves for the v - b a 11, plug, and ball valve reach 100 percent closed at an ind icat or rea d i n g other than zero. This is bec ause the ind i­ cator scale was con nect ed to the o p e r ati ng m e c h a n i s m and ranged from 16 to 0 between the wide open stop and the .closed stop of the operator. For these three valves, the area of opening was red uced to zero before the operator rea ched the closed stop which cor r e s p o n d e d to a zero indicator reading. A plot of ind icator rea d i n g versus percent age closed is known as the "control c h a r a c t e r i s t i c " of the valve. C. D e t e r m i n a t i o n of vel ocity and c o n c ent ratio n The actual nominal vel ocity of the flow dur ing each test was d e t e r m i n e d from the average v o l u m e t r i c flow rate of the mixture, , which was m e a s u r e d in gallons per minute, g p m . -19- 1. 2. 3. 4. 5. V-Ball Valve Plug Valve Ball Valve Pinch Valve Gate Valve Ind icator reading Percent closed Fig. 4. Plot of indicator r ea ding versus percent closed The vel ocity in feet per second was det ermi ned by converting Qm from gallons per min ute to cubic feet per second and div i d i n g by the area in square feet, or gpm 448.8 x 4 x 144 ".ELHl1x „cf s. 'in I .TtzI x D2 W where D is the dia mete r of the pipe in inches. In this study chip con ce n t r a t i o n was exp ress ed as a ratio of the volume of solids to the total mix ture volume. (6) -20- C o n c e n t r a t i o n of the mi x t u r e was de t e r m i n e d from the clear water flow.rate, Q , and the flow rate of the mixture, Qfflt as shown below. The volume of water in the mixture is equal to the total volume of the mix ture minus the volume of chips in the mix ture, or Volume HgO = Volume Mixture - C o n c e n t r a t i o n x Volume Mixture. Since the flow rates are in terms of volume per unit of time, this can be e x p r e s s e d in terms of the flow rates, Qffl and Q , or = Qm - CQm where C is the con centration. R e a r r a n g i n g gives QL M u l t i p l y i n g by 100 gives C as the vol umet ric c o n c e nt ratio n as a p e r c ent age of the total volume, or Q r., 100. (7) CHAPTER IV APPARATUS DESCRIPTION The apparatus de s c r i p t i o n will be given in four parts: (I) ope rati on of the pipe line system, maj or com pone nts of the system, (2) d e s c r ip tion of the (3) des crip tion of the test section and its components, and (4) des crip tion of the valves tested. A. O p e r a t i o n of the pipe line system A sch ematic di a g r a m of the pipe line system is shown in F i g . 5. Because of the need for var y i n g the con cent ratio n of wood chips in the mixture, two sep arate systems are required to supply chips and water. Water is initially pumped from a sump under the l a b orat ory floor to a constant head tank on the roof by an aux il i a r y 6-in. centrif ugal pump. A 3-in. dia meter line from the tank feeds water to the m i x tank as shown in Fig. 5. The clear water flow rate is m e a s u r e d by a 3-in. m a g n e t i c flow met er in the supply line. is used for co n t r o l l i n g the clear water supply. A gate valve O ve rflo wing of the tank is pre ve n t e d by an a u t o m a t i c shutoff valve o p e r ­ ated by a float sensing the level of water in the tank. The plastic chips are stored in the chip bin as s h o w n . They are fed into the m i x tank via a short conveyer belt. The supply of chips is reg ulat ed by a sliding gate on the storage bin. C o n c e n t r a t i o n of the m ix ture can thus be - 22- 11 I. 2. 3. 4. 5. 6. 7. 8. Long Branch Short Branch 3-Way Valve R o t atin g Screen Water Col lect ion Chip Storage Mix Flow Meter Supply Gate F i g . 5. 9. 10. 11. 12. 13. 14. 15. 16. Control Valve Pump and Motor Console Conveyer Mix Tank Ove rflo w Valve Control Valve Water Flow Meter Sch emat ic of pipe line system — 23 ” r e g u l a t e d by c o n t r o l l i n g the amounts of water and chips fed into the m i x t a n k . A hor izon tal 4-in. alu minu m pipe carries the mixture from the m i x tank to the main p u m p . A vertical section connects the pump dis charge to the pipe line which hangs near the roof of the laboratory. The dis charge of the mix ture of wat er and chips from the main pump was con trol led by varying the speed of the pump motor, A d i a p h r a g m valve in the ver tical section was also used for control purposes. Flow rate of the mi x t u r e is m e a s u r e d by a 4-in. m a g n e t i c flow met er above the control valve. The pipe line is made up of two parallel branches con nect ed to the pumping apparatus by t h r e e - w a y valves as shown in Fig. 5, The valves tested in this study were ins tall ed in the short branch. The pipe line dischar ges into the upper end of a tilted^ rotating, c yl indr ical screen which separates the chips from the water. Water is col lect ed in a bin beneath the screen and r e t u r n e d to the lab orat ory sump for recircu latio n. Chips are tumbled out of the lower end of the screen and into the chip storage bin from which they are fed into the system again. B. Sy s t e m components All operations of the system are con trolled from the control console which is shown in Fig. 6. main pump motor, Switches for the c o n v e y e r , rot a t i n g separation screen, flow -24- Fig. 6. Control console charts , and a rheostat for varying the speed of the main pump motor are mou n t e d at the left on the console top. Fuse and starter boxes for the motor are m o u n t e d below the console top. The two large radial charts are Foxboro Dyn alog recorders which give a direct reading in gallons per minute (gpm) of the flow rate of the clear water in the supply line and the flow rate of the mixture of chips and water being di s c h a r g e d from the p u m p . -25- To the right of the flow charts is a m a n o m e t e r tube showing the level of water in the m i x tank. Also to the right in Fig 6, within easy reach of the o p e r a t o r , are the gate on the chip storage bin and the clear water control valve. The control valve on the dis charge side of the pump is vis ible in the picture to the left and rear of the console. The i n s t r u m e n t a t i o n below the flow charts is part of a system linked to a Hew lett Pac k a r d 211 6 Com pute r used in other studies. The main pump used in the system is an All is - C h a l m e r s 4 x 4 x 9.5-in. cen trif ugal p u m p . The pump was equ ipped with a f l a t - b I a d e d i o p e n -fa ced imp eller (NSX t y p e ) which is e s p e c i a l l y suited to h a n d l i n g s o l i d - l i q u i d mixtures. Power to the pump was sup plie d by a 15 hp I 150/2600 rpm General E l e c t r i c 240 volt, d-c shunt wound motor. Flow rates of the clear water supply and the mixture were m e a s u r e d by 3- and 4-in. Fox boro D y n a l o g , m a g n e t i c flow t r a n s mi tters con nect ed to the Fox boro model 9 6 5 OC recorders on the console. O p e r a t i o n of the flow meters is based on Far aday 's law of e l e c t r o m a g n e t i c induction. Flu id moving through a m a g n e t i c field in the m e t e r induces a voltage pro p o r t i o n a l to the velocity. This voltage is t r a n s mi tted to the rec o r d e r which is c a l i bra ted to give a r e a d i n g from O to 400 gallons per minute. cent of full scale accuracy. The m a n u f a c t u r e r claims +1 p e r ­ -26- The plastic chips used in the study were made by Com merc ial Plastics of Chicago. The nominal dimensions of the chips are 1/2 by 3/8 by I/IO in. chips were used. Two types of plastic One type, red in color, had a specific gravity of 1.04 and was made of Cycolac. The other, blue in color, had a spe cific gravity of 1.05 and was made of Ethocel. The pipe used in the system had a nominal inside diameter of 4 inches. Clear acrylic plastic pipe was used extensively, but a l u m i n u m pipe was used where higher strength was needed. V i c t a u l i c couplings were used to join sections of pipe. C. Test section and components The valve test section was located on the p r e v i ou slym e n t i o n e d "short b r a n c h " of the pipe line system. of the test section is shown in Fig. 7. ope ra t i n g pla t f o r m A plan view The test section, m a n o m e t e r b o a r d , and camera were all hung from the roof trusses of the laboratory. The entire test section was made of 4-in. ID clear, plastic pipe. The re­ mai nder of the short branch was made up of a 4-ih. ID flexible rubber hose on each end of the test section con nect ed by 4-in. S m i t h - B l a i r flexible couplings. The section ups t r e a m of the valve con tain ed four pressure taps spaced as shown in Fig. 8. The d o w n s t r e a m section contained 10 pressure taps. Data from taps I through 4 were used to e s t a b l i s h the ups trea m linear por tion of the dim en s i o n l e s s h y d r aul ic grade line and from -27- Valve a O= 6 O Ope rating platform Camera Man ometer board Fig. 7. Plan view of test section taps 9 through 14 for the d o w nstr eam section. The pipe was jo i n e d to the test valves with 4-in. Smith-B lair universal flanged c o u p l i n g s . A typical pre ssure tap is shown in Fig. 9. pressure vents, pipe, The three 1/8 inch in diameter, are located around the 120 degrees apart. Two -inch square plastic blocks were shaped to fit the cur vature of the pipe and glued over the pressure vents. These blocks were drilled and threaded to accept I m p e r i a l - E a s t m a n 1/4-in. brass tee sections. The three tee sections were then con nected by l/4-in. OD tygon tubing. A fourth tee con nected the pressure tap to the line from the m a n o m e t e r board. Fig. 10 shows the differe ntial man ometer board used in this study for d e t e r m i n i n g the h y d r aul ic grade line along the test section. A dev elop ment of the man omet er board equations is given in A p p e n d i x B. Tubing from the pressure taps in the test section was con nect ed at the top of each tube of the Upstream Section i to CD I 5 6 7 8 Q Q Q Q s k 3 0.5 ' Q Q - 3 S1 .0 ' Q G - I @ 4.5' Q 2 @ 5.0’ D o w nstr eam Section Fig. 8. Loc ation of pressure taps in test section -29- Brass Tygon r" tubing Fig. 9. Typical pressure tap construction m a n o m e t e r board through an Imp e r i a l - E a s t m a n brass needle valve ( H ) . The tubes of the m a n o m e t e r board were 48 inches long and made of l/4-in. OD glass tubing. At the top of each tube was a tee section, one branch connected to the tube, one branch con nect ed to the pressure inlet valve, and the third con nect ed to a 1/4-in. air from the tubes. 5/8-in. Nupro purge valve used for bleeding The m a n i f o l d was made up of a section of ID copper tubing with silver soldered Swage Iok fittings c o n n ect ing to the glass tubing. Sta ndar d globe valves (C) and (D) are used to divide the m a n i f o l d when either or both of the -30- Fig. 10. Dif fere ntial man om e t e r board me r c u r y loops (F) and (G) were in u s e . The two mer c u r y loops were con nect ed in a separate stainless steel m a n i f o l d (E). Tank (A) is a res er v o i r for the carbon t e t r ach lorid e used in the m a n o m e t e r tubes. The carbon tet rach lorid e was colored with red Sudan III dye for easy visibility. m a n i f o l d through a globe valve ( B ) . Fluid enters the -SiSim u l t a n e o u s readings of all tubes on the m a n o m e t e r board were made by taking photogr aphs of the manometer. The camera used in this study for r e c o r d i n g data was a Leica MS with 35 mm lens and light met er attachment. an ASA number of 400 was u s e d , Kodak T r i - X film with The high speed of the film a ll owed a shutter speed of I/60 sec at lens openings of f5.6 to f 16. No art ific ial light was u s e d ; the natural light in the laboratory, though somewhat variable, was sufficient. D. Valves tested Four commercial valves were tested in this study: ball, t ap ered p l u g , and pinch t y p e „ v-ball, The previous Iy-me ntio ned gate valve was con s t r u c t e d of ple xigl ass in the laboratory. Fig. 11 shows the pri ncip le of ope rati on of each valve and the general shape of the ope ning through the valve. The ball valve m e c h a n i s m consists of a ball with a 4-in. dia mete r cylindr ical hole through it. Rotation of the ball is about an axis p e r p e n d i c u l a r to the flow and in a horizontal direction. The plug valve operates in the same way as the ball valve, the d i f fere nce being that a plug in the shape of a t r u n c a t e d cone with a tapered r e c t a n g u l a r hole through it replaces the ball. The pinch valve m e c h a n i s m consists of a heavy rubber sleeve, the bottom half of which is forced to the shape of a smooth weir. The upper half of the sleeve moves v e r t i c a l l y in operation, p r o v idi ng a pinching a c t i o n . The v-ball valve m e c h a n i s m consists of a portion of a spherical -32- P r i n c i p l e of Ope ration Flow Area Ball Valve Fig. 11. Sch em a t i c diagrams of valves tested -33- shell with a V - s h a p e d notch in one side. Rot ation of the V - n otched shell is about an axis p e r p e n d i c u l a r to the flow and in a vertical direction. The gate valve consists of a cir cular gate o p e r a t i n g in a vertical direction p e r p end icula r to the flow. All of the com merc ial valves except the plug valve were p r o vide d with p n e u mat ic operators. For simplic ity of o p e r ­ ation, the p n e u mat ic operators were removed and some form of man ual operator installed. A radial ope rating lever was in­ sta lled on the v-ball and ball valves similar to the one r e c e i v e d with the plug valve. All three were closed from the wide open pos itio n by r o t atin g the ope rati ng lever through an arc of about 90 degrees. The pinch valve r e q uire d a screw- type ope rato r with a stem travel of 2.5 in. An indicator was ins talled on each valve for gauging the area of opening^ Each of the indicators read from O to 16, in 16 equal divisions from closed stop to open stop. The pinch valve ind icator ope rated v e r t i c a l l y with res pect to the stem travel of the screw operator. The indicators on the other three valves were con nected to the ope rati ng lever and had radial scales. The gate valve tested in 1965 had a screw- type operator and indicator similar to the ones used on the pinch valve. CHAPTER V TEST P R O C EDU RE The test p r o c edu re was made up of three parts: (I) p r e p ­ aration of e q u i p m e n t , (2) data collection, and (3) data r e ­ duction. A. P r e p a r a t i o n of equipment Data from the m a n o m e t e r board des crib ed in the preceding chapter was used in d e t e r m i n i n g „ T h u s , it was one of the most important pieces of equ ipment used in c o l l ect ing data and as such it r e c e i v e d con side rable attention. Before each valve t e s t , the m a n o m e t e r tubes were drained, partial Iy d i s a s s e m ­ bled, and scr ubbe d with a small nylon brush. At every other cleaning, the tubes were flushed with acetone to remove har d e n e d deposits before being rin sed with water. The mercury tubes were d i s a s s e m b l e d and cleaned in a similar m a n n e r . Before r e f i l l i n g the m a n o m e t e r board, the carbon tet rach lorid e and m e r c u r y were fil tered to remove impurities. After cleaning, the tubes were rea s s e m b l e d on the m a n o m ­ eter board and all tubes filled with water to eli mi n a t e air from the, system. The r e s e rvo ir tank was then filled with colored carfron tetrach lorid e. By ope ning the valve (B , Fig. 10) below the reservoir, fluid was all owed to enter the m a n i f o l d and rise in the glass tubes, d i s p l a c i n g the water. The tubes were filled to a level of app ro x i m a t e l y -35- 24 i n c h e s . Mer cury was added to the two mer cury loops through a sta ndpipe att ache d to the mid dle of the mer cury manifold. Air bubbles in the m a n o m e t e r system affect the level of fluid in the m a n o m e t e r tubes which results in incorrect values of d i f f e re ntial pressure. T h u s , after all tubes in the m a n o m e t e r were filled to the des ired level with fluid, flow was e s t a b l i s h e d in the test section which caused air bubbles to be driven from the manometer. By opening the purge valves at the top of each m a n o m e t e r tube, any air bubbles in the system were forced out by the pressure. Tes ting began after all air bubbles were e l i m i n a t e d from the system. Air bubbles which p e r i o d i c a l l y ent ered the m a n o m e t e r system during the testing were removed, when discovered, by the a f o r e m e n t i o n e d method. Acc urat e flow rate m e a s u r e m e n t was essential to the test p r o c e d u r e . . The a c c u r a c y of the flow rate r e a d i n g s , from whi ch values of c o n c e n t r a t i o n and actual vel ocity were d e t e r ­ mined, was dep ende nt on the cal ibra tion of the flow m e t e r s . T h u s , before each valve test, the pipe line was div erte d to dis char ge into a c a l ibra ted tank to check the flow recorders. Clear water was used for cal ibra tion p u r p o s e s . The procedure inv olved c o n v e r t i n g the dis charge from pounds of water for a m e a s u r e d time interval to a gallons per minute flow rate. The m e a s u r e d flow rate and the flow rec orde r reading were then com pare d and the rec orde rs adj uste d accordi ngly and re c h e c k e d . -36- C a l i b r a t i o n checks were made at flow rates of app ro x i m a t e l y 350, 250, and 100 g p m . With little effort the flow charts could be made to agree to within about one gallon per minute of the cal cula ted discharge. The laboratory water supply system was also cleaned per iodically. Twice during the testing, the laboratory sump was drained, hosed clean, and r e f i l l e d with fresh water. C o l l e c t i o n screens were cleaned of chips and other debris w as hed into the sump from the laborat ory floor. Screens above the water col lect ion bin ^ under the discharge end of the pipe line, were also cleaned of spilled chips and foreign m a t e r i a l . B. Data col lect ion procedure The testing phase of this study involved col lect ing valve head loss data for dif ferent flow conditions set up by varying the velocity, valve closure, and chip concentration. Data was col lect ed at nominal vel ocit ies of 10, 8, 6, and 4 fps for each valve tested. These velocities correspond to flow rates of 382, 305, 229, and 153 g p m . For each vel ocity the valves were tested at three or four dif fere nt closures. All valves were tested in the wide open and 50 percent closed position for purposes of c o m p ari ng the eff ec t i v e n e s s of the different valves in passing a mi x t u r e of wood chips and water. Each valve was also tested at one or two other closures, depending on the ch a r a c t e r i s t i c s of the valve. Tests were run with O i 10, and 20 percent chip con c e n t r a t i o n for each given -37- v e l ocit y and valve c l o s u r e . The 1965 gate valve tests were c o n d uct ed at con ce n t r a t i o n s of O 6 5, 10, 15, and 20 percent. The data col lect ion p r o c edu re involved two m-en. One man ope rate d the system and r e c orde d flow rates from the control console. The other man ope rate d the camera, m a n o m e t e r board, and valve. velocities. Tests were run in order of high vel ocit ies to low The valve was ini tially set in the wide open pos itio n as the ope rato r e s t a b l i s h e d a flow of 10 fps at zero concent ratio n. While the flow con dition was stabilizing, the cam eraman p r e p a r e d i d e n t i f i c a t i o n tags and placed them at the bot tom of the m a n o m e t e r board. These tags ind icated valve type, date of test, m a n o m e t e r board arrangement, velocity, valve closure, and con centration. Each pho togr aph of the m a n o m e t e r board was thus identified. When a steady flow was est ablished, the ope rato r signaled the cam eraman and began r e c o r d i n g ten observa tions of mix and clear water flow rates on p r e v i o u s l y - p r e p a r e d data sheets. As each ob s e r v a t i o n was made, the cameraman was not ifie d and a picture of the m a n o m e t e r board was taken. When ten o b s e r ­ vations were com plet ed for 0 p er cent concentration, the ope rato r added chips to the system and reduced the clear water flow to get 10 percent con ce n t r a t i o n at the same velocity while the cam eraman changed the c o n c e n t r a t i o n ide nt i f i c a t i o n tag on the man omet er. When steady flow at 10 percent con cent ratio n was establi shed* another series of ten observa tions were -38- re c o r d e d . The chip con ce n t r a t i o n was next inc reased to 20 percent by the ope r a t o r and ten more observa tions were made. The system was then cleared of chips as the cameraman set the valve to the next closure and changed i d e n tif icati on tags. A series of ten o b s e r va tions were then made at that valve set ting for O 1 10, and 20 percent concentrations. This process was r e p e a t e d for each desired valve closure. After tests were made for each valve closure at 10 fps, the valve was opened to its wide open position. The flow rate was then red u c e d to 8 fps and the test pro cedure repeated. Tests were run at 6 and 4 fps in the same manner. At the end of each valve test, three reruns at some given velocity, valve closure, and c o n c e n t r a t i o n were made as a check on r e p e a t ­ ability. Values of obt ained from the rerun data were gen er a l l y found to agree with those obtained from the regular test data. During tire testing of each v a l v e , observations wer e also made on the general ope ration of the valve and the ch a r a c t e r i s t i c s of the flow imm edia tely upstream and dow n­ stream of the valve. C. Data red uc t i o n All of the tests on each valve were com pleted in about two days. At the end of each day's testing, the exposed film was taken to a local p h o t o g r a p h i c shop for ove rnight p r o c ­ essing. In this way any data lost due to pho t o g r a p h i c errors could be rerun the next d a y . The films were d e v e lop ed into — 39 — 35 mm neg ativ e strips which were p r o j ect ed onto a screen with a film strip projector. Two und er g r a d u a t e assistants read the m a n o m e t e r levels in each frame of the film strips and rec orde d the l e v e l s , along with c o r r e s p o n d i n g m i x and clear water flow rates from the console operator's data sheets on FOR T R A N coding f o r m s . A sample coding sheet with data is shown in Chapter 6. The data was then pun ched on standard data pro cess ing cards. These cards and a F O R T R A N source program (Appendix D) d e v e l o p e d for data r e d u c t i o n were sub mitted for pro cess ing on the SDS Sigma 7 com puter in the Montana State Uni vers ity C o m p u t i n g Center. The computer output for one test run at a given velocity, valve closure, and chip con cent ratio n is shown in Chapter 6. Generally, two or three runs were re­ quired before all m i s t a k e s and bad data were e l i m i n a t e d and final head loss results were obtained. . Bad data was dis­ covered by the m et hods dis cu s s e d in A p p e n d i x C . CHAPTER VI DATA ANALYSISThe analysis of data col lect ed in this study includes the following: data, (I) com puter operations, (2) p r e s e nt ation of (3) flo w pictures, and (4) summary of valve c h a r a c t e r ­ istics. A. C o m pute r operations Data from the sheets pre pare d by the console operator and from the pictures of the m a n o m e t e r board was rec o r d e d on F O R T R A N coding sheets as shown in Fig. 12. The sample data shown is for one ball valve test at 10 fps, 25 percent valve closure, and 0 percent concentration. The first six lines contain m i s c e l l a n e o u s inf orma tion as exp lained by the comment cards in the computer pro gram (Ap pend ix D ) . The next 18 lines contain obs erve d data, each pair of lines being one o b s e r ­ vation. Columns I through 12 of the first line of each pair contain the flow rates Q the flow recorders. and Q in gpm which were read from Columns 13 thr ough 72 are filled with the m a n o m e t e r tube fluid levels for pressure taps I through 10. The first 24 columns of the second line contain man ometer levels for taps 11 through 14 while columns 25 through 36 contain the levels of fluid in two mer cury tubes. A computer p r o g r a m , written in F O R T R A N II l a n g u a g e , was d e v e l o p e d for ana ly z i n g the above d a t a . The com puter was Party Date /g Col Sc. 5 Soi - B all Ihnze CSfimfiL e )_____ 10 15 20 35 30 25 40 45 50 55 60 7 [5 /,.,8 ,3 ,3 1 4", ,4 / lo , , , , I , , , , I , , . , I . , , . j . . -Oi • <0,0< . i^l# t0<0\ I / <0\ • tO ,0, . ■ /.0 .3 .8 3 --M .L .L V ,ft,L V ,- . 7 . .z z s , . <a , i , a X . 1. . . 3.&. / .37. ,8, , 1 , 7 , .,Co, . . / . f i . l V I S , . , 3, J .7 ,.,is , , I S , . ,?, . /.7, „ 7 ! , ,/,? ..lv , , . . -?.^,-i7, I ! J1J, . ,Y, , ,/,Y l i JT i , /t Y' • ' i' I 1/161 . 1Q 1 /s ...i. ./.7 , . , 7 /,37 »i Yi ,/,z ..6 l» : ,l.1 ,.'.3 S , . . IS ,.,I. , 1,7,.,i . , , / , f i . l v . , . . /,S ..9 S , m,.,is /|3T «i Yi ,1,7,.. i . ,I 3, 8, 3 , , ,9, I I I/ eI .3 ^ ,V $«i i I I /i / 1« I I I /i^ - # I ! . 3 . 0 . . . I ' S . . 3 . 7 . . '. 3 . . . 3 . 1 '. . . L . . . V . a . . , & 3 .6 ...7 , I , 3 1 , . , 0 , , . / . v . . l j r , , 1.1 , . , 3 , , ,/Iv ,.,? , , r—------ Ir . i I »«4« . ,^ ,7 ,. ,3, , ,j.Y i. I^, I ,Y,^, . 7 . ' •3«8Y I• I I ,W1S 1Y , i I ,/,Z 1* , 7» • I / i YI • , • I , i#*, 9 »,?, , , «?,7i • «3, , ,*3,Y, .,Y, i , Y '^ i»' ^ I •V•9i ~ t • I , ,S1Y, • , , , , / , I , • ,Cr, , , I !Y! #, S1 , • ,or, I •• , £, , I , w»5, <»• I , ,3 $ , S?,Yi • 14C I • Y*e?f • . .. I _* ------- 4 T ----- 1 > — I_ 3,6*, »,7 , , ,J 1S1eI »I I t / Y' »'61 I I /iV ' #1^ 137 , /,3 , . I . a s . . 33 . . . <N Q • «?' &*3,» I I I 3|6«aji «I I I I / i / I » .w,8 .2 ,,, , 6 ,& , A a a .,js /,£>,• ,S 'S ’ • ZiT1. 13,8t , / , 7 , < / & , . , 3 , 8 , 0,0 , B ,0 , I . /.f i.lv . . . . ^7» 13|3, 13,6, *, 7, 1 1 1 , 1 1 1 , 1 , , 1 , 1 1 1 , 1 1 S i •«7i 1 ,J 13, »,Y, , ,/,Y ,. ,4, , ,ZiY>»'3? t , / 4 7 . A , ./.fi.lv . , . , ! #?, 4 ,. Y , 1 ,3 ,6 *,» ,7 , 1 1 • 1 1 1 1 1 1 1 1 I 1 1 1 1 1 1 1 1• , , , , / , / ! • !^1 ! ! I !Y ,0 ,3 , I 3 6 ,,,7 , 1 , 3 , 3 , *,9, , , / Y ,51 , , /.^ .,Y , , . / ' . 1 . . . 9 S • ,w«-^i •,?, , «*c«7,«,3, i ,3 ,YI• , ^ 1 I , Y,«P, • 16 1 ■ 3,57 ei 3, 1 13 61 », f l , 1 1 1 1 1 , , 1 1 1 1 1 1 1 , 1 1 , iw,5,w, •, , ,3 i9 i6, «• I , , / , / , • ,6, I I / 1Yi •• 3 I 3,6, •,7, , , 3 , 7 , »-0. , 1/,Y, » 6, 1 , 7^4. *Ji , ■ toiiw,• ,Y A 6 .., 6 ,< , . * • , { , • ,0 , , '8, 57» I ,is,.,0, , , ,< 7 ,. .3 , , ,J.Y ,.,3 ! , ,y .3 ,..7. I • ,"•.,7,• ,3 , ,3 ,& i I I , , J,Y,• ,3,5" IYie?!* ,^1 , i I /1Z, • 18 , , I / , Y « » IY i ,3 , , , /,V,.I a' , , /,V ..4- . , / X a , 3 6 • ,6 , , ,3,2, * ,8 , , , Z,Yi • 17, , 1 /,Y i»161 1 ,/ « ■ . , a , • * , • ,0, , Iflz, 7,. m5 i , , J.YI * IY,5*, I Y3, • ,3*I , Fig. 12. , /,9, .lv. . . . 3^1» 12 I 1Si 6, e, 7, A, I , I I I I I I • I I I I I , I I I F O R T R A N coding sheet and sample data , ,/,f i.lv . . . . — 42 — thus p r o g r a m m e d to per f o r m the f o l l o w i n g sequence of operations in.the c o m p u ta tion of K^: (a) read the m i s c e l l a n e o u s information and observed data for a given v e l o c i t y , closure, and concentration, (b) compute the average m a n o m e t e r r ea ding for each tap and the average c o n c ent ratio n from the observed flow rates, (c) compute the statist ical probable error of each average m a n o m e t e r rea ding and the average con­ centration, (d ) compute d i m e nsi onles s differe ntial heads from average m a n o m e t e r readings, (e) compute slopes and y - i ntercepts of upstream and d o w n s t r e a m portions of the dimensi onles s grade line using regression, (f) compute the c or rela tion coefficient and est imat e of var iance about reg ress ion for each line, (g) compute the loss coefficient, , as the d i f fere nce b et ween the two lines at x = X V A L , and (h) print out the input data and .results of all c a l c u la tions and recycle to (a) for the next set of test data. The computer output for the data shown in Fig. shown in Fig. 13. 12 is Two pages of output were req u i r e d for each test at a given velocity, valve closure, and concentration. — 43 — -»..' .... . -.«d_ $i^***^^* x«*<r8 %:,%%%:, "» -K *" 5*T *•v t-«x " *- ' - " ' "» '* - ■ ■ WItItf ■ ' . ■ . . . . . ■ ■ - , ,%.» C-I *».* i».» I)'* w«» '*«? . ««,. %».. ,.,e, U f I* «M T M.» *.» **«# $#«#W«* *»«» **«»W,' *»» ** » I* *#«»M T T 'M, T m . * . M r.' M.T . M.T c*. -s««..*.fcj» S i.* t» .? 8 * . r » *.» »»«* i i . s s»*« i*« » »*<’ t»«» «»<* *»•» « » * ♦»*’ * * • * w * .5.5 * i*«9 „. WM. f . M T n.$ tt.t «.» 0,1%».$ « «** )M » *»!? %'* *" '*"* " * **"' U « tT .» M t W .« I * . * l» * M l M M IT M I * . * # * M * ' » * * , » *TMMt T M 4 « T M I .T r . T M m/I# M T M .T , M M M M M M M ,T M T M M «« #MT ; T ,., .,M „ . , M .T ' TW .2 : % ' MTMfTt — TTTMTi* .. -!-,rrnr-T .VO MW# T T k T C liT I I T M M TT T M TCC "Mkw* "VUfiy Fig. ^ I* «:*m% 13. '*** Sample computer output -44- Input data and averages are pri nted on the first sheet while the loss c o e f f ic ient results are p ri nted on the second sheet as s h o w n „ B„ P r e s e n t a t i o n of data E x p e r i m e n t a l values of and cor re s p o n d i n g actual v e l o c ­ ities and c o n c e n t r a t i o n s for tests run on the v-ball valve are s u m m a r i z e d in Table I. The same data for the plug, ball, pinch and gate valves is given in Tables IV through VII in Appendix E . Raw data, computer output lists, and other i n f o r ­ ma t i o n p e r t a i n i n g to this study may be found on file in the D e p artm ent of Civil E n g i n e e r i n g and En g i n e e r i n g Mechanics, Montana State University, B o z e m a n , Montana. The v a r i a t i o n of the loss coefficient with velocity at 0, 10, and 20 percent con ce n t r a t i o n s for the various valves is shown by the log-log plots in Fig. 14. The clear water loss coe ffic ient for each valve tends to be constant with res pect to vel o c i t y except for the gate v a l v e . also indicate that The graphs appears to vary somewhat with velocity for higher chip con centrations. H o w e v e r , for each valve, tends to app roach some constant value for vel ocit ies greater than about 6 fps. For design purposes, t h e n , can be ass umed constant with respect to vel ocit y for a given closure and per cent chip concentration. The loss coefficient, , was also observed to increase with i n c r e a s i n g v o l u m e t r i c chip c o n c e nt ratio n for a given TABLE Valve V 4 fps C I Summary of Com pute d Results - V-Ball Valve 8 fps 6 fps V C V C KL KL KL V 10 fps C KL 3.932 0.00 1.033 5.903 0.00 9.68 43. 3% 3.9 56 10.17 1.320 5.911 5.885 19.59 .972 7.877 0.00 1.089 7.861 9.88 1.303 7.869 19.63 .901 9.880 0.00 .981 1.023 1.166 9.867 20.50 1.198 0.00 1 .661 5.903 0.00 9.36 9.93 1.949 5.895 5.877 19.48 1.693 7.840 0.00 1.851 7.877 10. 19 2.2 62 7.911 20.36 1 .663 1.864 2.121 62.5% 5.908 0.00 5.92 7 10. 10 5.909 19.91 4.764 7.859 0.00 4.700 7.846 9.36 5.838 7.914 20.38 4.592 9.843 0.00 4.853 5.011 5.498 9.872 20.55 5.769 75.0% 5.916 0.00 20.104 7.888 0.00 19.951 5.924 10.33 19.801 7.835 9.24 19.197 5.897 19.85 22.604 7.856 20.07 21.305 3.940 50. 0% 3.948 Repeat Repeat 5.867 9. 18 1.867 7.856 20.07 5.885 62.5% Closed 5 0 . 0 % Closed 10.0 10.0 50. O K Ball 50. 0% Gate Pinci I .Ol-7 8 9 10 Vel ocity (fps) 8 9 10 Velocity (fps) Fig. 14. Plot of versus velocity at different concentrations 10.0 a. Ball A Gate (a8.9 2 . 0< r Ba 1 1» i nch + Percent Concent ratio n Fig. 15. Plot of versus concentration at 50 percent closure - 48- valve closure as shown by the s e m i-lo g plots in Fig. 15. Similar trends were obs erved for each valve at any given percent closed. For any set of t w o - d i m e n s i o n a l data (x , y) that plots as a str aight line on a sem i-log graph, an equation of the form y = ae I)X can be derived where a is the y-inter cept and b is the slope of the line. A set of data (C , ) was obt aine d at each valve closure for the valves tested in this study. P l o t t i n g this data gave a set of straight lines for each valve similar to and inc luding those in Fig. 15. By a v e r a g i n g the slopes of these lines, an empirical equation of the f o l l owi ng e x p o n en tial form was dev eloped for each valve: (KL ) c _ " V o bC (8 ) = " In this equ atio n b is the average slope, C is the variable c o n c e n t r a t i o n in decimal form, ( K ^ ) q is the clear water head loss c oe ffic ient or y-i n t e r c e p t for any valve c l o s u r e , and ( K ^ ) c is the loss c oe ffic ient for a given concentration. In other words, this equ atio n gives the approximate value of the loss coe ffic ient as a function of con cent ratio n and the clear water loss c o e f f ic ient at some valve closure. Values of b are given in Table II. Ano t h e r ap p r o x i m a t e r e l a t i o n s h i p for p r e d i c t i n g also developed. Values of (K^) / ( K l ) was computed from the e x p e r i m e n t a l data for each valve are plotted against con­ cen trat ion in Fig. 16. Straight lines fitted to this data 2.0 2.0 tV-Ball I vO Ball Fig. 16. Plot of I V c c k L 1O versus c o n c en t ra tion — 5 0— can be e x p r e s s e d by linear equations of the form I V c ckL 1.0 + mC (9) jO where m is the slope of the line. Values of m for this equ atio n are also tab ul a t e d in Table II. T h u s „ once the clear water loss coeffic ient for a given closure is determined either an e x p o n en tial or linear r e l a t i o n s h i p can be used to app r o x i m a t e the c o r r e s p o n d i n g loss coefficient for any con­ centration. Table II. Emp iric al values of b and m for K l e q u a t i ons Valve Type E x p o n en tial Equation Linear Equation V-Ball Ball Plug Pinch Gate 1.00 0.83 1.02 1 .78 1.57 1.30 1.40 . 92 1.80 2.30 S e m i - l o g plots of (K l ) q versus percent valve closure for the various valves are given in Fig. 17. These curves , when used with the r e l a t io nship s dev el o p e d above, provide a p p r o x i ­ mate values of the loss coefficient, K L , at a given valve closure and chip con ce n t r a t i o n for design purposes. These methods give values of K l that gen er a l l y agree to within 10 percent of the empirical d a t a . Gite 10.0 Ball 10.0 — 51 — Percent Closure Fig. 17. Plot Percent Closure of (K l )q versus percent closure -52- C. Flo w pictures Fig. 18 is a picture of the test section car rying a flow of 153 gpm (4 fps) with a chip c o n c ent ratio n of about 5 percent just ups t r e a m of the plug valve. the picture. Flow is from left to right in Under these conditions, the chips settle and move along the bot tom of the pipe in what is r e f erre d to as a "sl iding bed." The picture in Fig. 19 shows app ro x i m a t e l y the same flow conditions just d o w n s t r e a m of the plug valve, which was in the wide open position (4.4 percent constriction). Since the sli ding bed is only slightly disturbed, minor t u r bule nce is evident. For comparison, Fig. 20 shows the effect on the flow caused by the plug valve when 50 percent closed. The sli ding bed is com plet ely disrupted, a high degree of turbulence. indicating This turbulence was observed to increase with v e l ocit y as well as valve closure. At 75 percent closed, the ope ning through the plug valve was in the shape of a na r r o w slit. The width of this slit was slightly greater than the largest dimension of the chips used. As a result, plu gging of the pipe line occ urre d for flows with r e l a t i v e l y low chip concentrations. Fig. 21 shows the b e g i n n i n g of a plu g-up just u p s trea m of the plug valve at 4 fps, 75 percent closed, and about 8 percent concentration. A mass of chips can be seen in the picture piled up in the pipe just to the left of the valve coupling. -53- Fig. Fig. 19. 18. Picture of sliding bed of chips ups trea m from valve Pic ture of tur bule nce dow nstr eam from plug valve, 4.4 percent closed -54- Fig. 20. Fig. 21. Picture of turbule nce dow nstr eam from plug valve, 50 percent closed Picture of b e g i n n i n g of "plug-up" -55- D„ Summary of valve cha ract erist ics A chart of valve cha rac t e r i s t i c s is given in Table III for purposes of c o m p ari ng the valves tested in this study. The "control curve" refers to the plot of indicator reading versus p e r cent age closed given in Fig. 4. A perfect control curve would be linear from the wide open position to the 100 percent closed position. "Wide open per cent age closed" is the amount of c o n s t ri ction in the pipe line caused by the valve when in the wide open position. self e x p l a n a t o r y . The r e m a ini ng items are TABLE III Valve Charact erist ics Ball Valve Plug Valve V-Ball Valve Pinch Valve Gate Valve Control curve Good Good Good Poor Fair Wide open % closed 0.0% 4.4% 43.3% 28. 5% 0.0% K l @ wide o p e n , 0% chips 0.020 0. 100 0.970 0. 120 0.014 K l @ 5 0 % c l o s e d , 10% chips 6.60 6.10 1.70 0.90 O b s t r u c t e d str eaml ine flow Yes Yes Yes No Yes Pockets or shoulders Yes Yes Yes No No Flow in either direction Yes Te s No Ye s Yes Tight shutoff Yes Yes Yes Ye s Yes Chip she arin g ability Yes Yes Yes No No 4.30 5 6- CHAPTER VII C O N C L U S I O N S AND R E C O M M E N D A T I O N S As predicted, an increase in the loss coefficient, , was obs erve d for each valve as the degree of con stri ction in the pipe line or valve closure was i n c r e a s e d . values of The observed for clear water flow through the gate valve are a p p r o x i m a t e l y the same as those rep o r t e d by H . Rouse (8). R o u s e ’s plug valve values are c o n s i de rably higher than the plug valve data obt a i n e d in this study. This probably is due to dif fere nces in the geo metric c o n f igu ratio n of the p a s s a g e ­ ways through the valves. At a given closure and flow rate, was found to in­ crease with chip c o n c e n t r a t i o n as predicted. For the various valves tested, the loss coe ffic ient at 20 percent chip c on­ cen trat ion was obs e r v e d to be gre ater than that for clear water by factors ran g i n g from 1.18 to 1.46 times. This is contrary to the effects R . W . Ch a r l e y (I) observed for flow through pipe expansions. The valve head loss coeffic ients for clear water were obs erve d to be constant with vel ocit y for a given valve closure as predicted. At higher concentrations, increase sli ghtly with d e c r e a s i n g velocity. greater than 6 f p s , h o w e v e r , tended to For velocities a p p r o a c h e d some constant value. Thus, K l can be ass u m e d constant for design purposes „ -58- The fol lowi ng r e c o m m e n d a t i o n s are made as a result of e x p e rie nce gained in this study. Improvements i^i this p a r ­ ticular study could be made by (a) e l i m i n a t i n g v i c t aul ic couplings in the test section which cause extra und esir able t u r b u l e n c e , (b ) using filters on the pressure inlets of the m a n o m e t e r board to keep foreign mat erial from e n t erin g the s y s t e m , (c ) using fluid in the m a n o m e t e r board that has a p p r o x i m a t e l y the same spe cific weight as carbon t e t r a c h l o r i d e but which is n o n c o rr osive and does not dissolve the pla stic chips used for testing purposes, (d ) c o n d uct ing tests with chips of different specific gravities, (e) using different sizes of chips and/or pipe and valves to check if the ratio of particle size to pipe diameter is a factor, and (f) tes ting the various valves ins talled in n o n c o n v e n t i o n a l positions. To further enhance the technical knowledge of t r a n s ­ por ting c h i p - sh aped solids by pipe lines, the fol lowing res earc h projects should be initiated: (a) a study in which the results of all prior tests using pla stic chips are checked with tests using actual wood c h i p s ; -59- (b) a study aimed at d e t e r m i n i n g critical factors such as flow rate and con ce n t r a t i o n for p r e d i c t i n g "pl ug-u ps" caused by v a l v e s , e x p a n s i o n s , c o n t r a c t i o n s , e l b o w s , and other f i t t i n g s ; (c ) a study to det er m i n e the corrosive effects of wood chip brine on the pipe line, p u m p s , v a l v e s , and other c o m p o n e n t s ; (d) a study to develop a me t h o d for sep arat ing the wood chi p-wa ter m ix ture at the pipe line t e r m i n a l ; and (e) a study to det ermine methods for disposal of the p o l lute d wood chip brine at the pipe line terminal,. 60- APPENDICES APPENDIX A ERROR ANA L Y S I S The accuracy of e x p e r im ental work is limited by errors inherent in the pro cedu res and app aratus used. A dis cussion of the errors a f f e c t i n g the results of this study will be given in three parts: (I) the effect of errors in m e a s u r e ­ ments , (2) errors due to regression, and (3) sum mary of error effects. A. The effect of errors in mea s u r e m e n t s M e a s u re ments of flow rates were made to est ab l i s h con­ cen trat ion and vel o c i t y and of m a n o m e t e r fluid levels to e s t a bli sh the loss coefficient. I. Flow rate mea s u r e m e n t s Flow rate obs erva tions ave rage d less than +5 gpm v a r i a t i o n from the des ired mean values. The m a x i m u m flu ctu­ ations were obs erve d for low flow rates (+15 gpm) because of pump instability. Flu c t u a t i o n s also had a ten dency to in­ crease with i n c r eas ing chip concentration. A c c o r d i n g to a sta tist ical analysis (Ap pend ix C ) , the probable error in the average c o n c e n t r a t i o n d e t ermi ned from ten flow rate o b s e r ­ vations was usually less than +0.5 percent (percent con­ centration) . L a b o r a t o r y c al ibra tion of the flow recorders provided acc urac y of +1 g p m . An error of I gpm in the o b s erve d clear water and mix t u r e flow rates would cause the computed 20 p e r ­ cent c o n c e n t r a t i o n to vary by +1 percent (percent concentration) — 62 — for 4 fps (153 g p m ) and +0.5 percent for 10 fps (362 gpra). A +1 gpm error in the mi x t u r e flow rate would cause an error of +.03 fps in a vel ocit y of 4 fps and +.02 fps at 10 fps. 2. Man om e t e r level m e a s u r e m e n t s F l u c t u a t i o n of the levels of fluid in the man omet er tubes was obs erve d to increase with valve closure and chip con centration. A c c o r d i n g to a s ta tist ical a n a l y s i s , the probable errors in the average m a n o m e t e r readings were about +.05 in. The pro babl e errors in the observed fluid levels of m a n o m e t e r tub.es con nect ed to pressure taps in the highly tur bule nt region just d o w nstr eam from the valve were +.2 in. Errors in the obs erved m a n o m e t e r fluid levels r e s ulte d from couplings in the test section cau sing extra t u r b u l e n c e , air bubbles in the m a n o m e t e r , small leaks in the m a n o m e t e r , dirt in the m a n o m e t e r , and human errors. An error of +.1 in. in the difference between the o b s erve d fluid levels of two m e r c u r y tubes would cause the dim e n s i o n l e s s d i f f e re ntial pressure head, D I M P H , to vary by + . 4 1 6 at 4 fps and +.0 68 at 10 fps. Mercury man omet er tubes were used for high dif f e r e n t i a l pressures just dow n­ stream of the valve. Thus, as shown by the equation developed in A p p e n d i x B , the d o w n s t r e a m portion of the dimensi onles s h y d r aul ic grade line is rel ated to the upstream line by the DIMPH givten by the mer c u r y loop. Any error in the DIMPH given by the m e r c u r y loop raises or lowers the d o w n s t r e a m grade line - 63- with respect to the ups tream line causing to be in error by the same a m o u n t . An error of +.1 in. in the difference between the o b s erve d fluid levels for two carbon tet ra c h l o r i d e tubes would cause the DIMPH to vary by + . 0 1 9 6 at 4 fps and +.0 032 at 10 f p s . The effect on of such an error in one DIMPH would be small since the grade lines are es t a b l i s h e d by regression with several points B. ( , DIMPH^). Errors due to regress ion A c o r r e la tion coefficient, C O E F , and an estimate of var ianc e about regress ion, E V A R , were cal cula ted as a measure of the acc urac y of the straight lines fitted by reg ress ion to the linear portions of the dim e n s i o n l e s s grade line upstream and d o w n s t r e a m of the valve. cussed in A p p e n d i x C . The COEF and the EVA R are d i s ­ In this study values of the COEF for the e m p i r i c a l l y e s t a b l i s h e d grade lines varied b e t w e e n 0.98 and 1.00. A COEF of 1.00 indicates a perfect linear r e l a t i o n ­ ship bet ween the data, x. and DIMPH^. Values of the EVAR com puted for the grade lines ave rage d about .15 which indi­ cates the grade lines fitted the data quite well. An EVAR of 0.00 indicates a perfect fit. T h e o r e t i c a l l y the slopes of the two straight line p o r ­ tions of any given dim en s i o n l e s s grade line should be equal. However, values of the two slopes varied at most by .002 because of e x p e r i m e n t a l errors in the data. The upstream — 64 — y - i n t e r c e p t s should have always been 0.0 , but the actual values varied by +.005. The loss coefficient is given as = (a + b x ) ^ - (a+bx) From this equation it can be seen . that any error in either intercept, a, will directly affect . Likewise, any error in either slope, b, will cause an error of bx in C. where x = X V A L . Error summary From the for eg o i n g discussion, the fol lowing max imum error est imates can be tabulated: Loss C o e f f ic ient Percent C o n c e n t r a t i o n 0.5 % Pro babl e error I .0% R e c orde r error .416 Hg m a n o met er error .020 CCl^ m a n o met er error 1.5% Total .030 Reg ress ion slope error .005 R e g ress ion intercept error +.471 Total The error in c o n c e n t r a t i o n was found to increase with de­ cre asin g vel ocity while the error in crease with d e c r e a s i n g velocity, in c r e a s i n g chip concentration. was obs erve d to in­ inc reas ing valve closure, and APPENDIX B MAN OM E T E R BOARD Ap/ $ i Fluid Flo wing Fig. 22. Carbon Tetrachloride -Ma ni fold Mercury S c h e m a t i c of d i f f ere ntial m a n o m e t e r board — 6 6- The m a n o m e t e r board used in this study was designed for m e a s u r i n g dif f e r e n t i a l pressure heads between the first pre s s u r e tap in the test section and other pressure taps d o w n ­ stream of the first. Fou rteen pre ssur e taps were located along the test section as shown in Chapter 4. Fig. 22 shows a sch em a t i c dif f e r e n t i a l m a n o m e t e r board and its rel a t i o n s h i p to a typical section of pipe line and the hyd ra u l i c grade line Bas ic a l l y two m a n o m e t e r board arr ange ments were used. One con figu ratio n, using carbon t e t r a c h l o r i d e (CCl^) in all m a n o m e t e r tubes, was used when the were all small. A p / $ . as shown in Fig. 22 Because of the r e l ativ ely light weight of CCl^ (specific gra vity = 1.59), dif fere ntial pre ssure heads gre ater than 3 ft of water (1.3 psi) caused deflect ions ( A y ) too great for the length of m a n o m e t e r tubes used. When high dif f e r e n t i a l pre ssure heads occurred between two pressure taps, such as across the valve or between points 2 and 3 in Fig. 22, a' second a r r a n g e m e n t was used in which a U-tube filled with m er cury was con nected to the two pre ssure taps. D e f l e ct ions thus obt ained were smaller because of the greater weight of mer c u r y (specific gravity = 13.55). The man omet er board used in this study was d e s igne d so that one or two mer c u r y loops could be v a Ived into the carbon t e t r a ch lorid e system when needed. - A.' 67 - Dif f e r e n t i a l m a n o m e t e r board for low A p/g- With flow from left to right in Fig. 2 2 ? the pressure is gre ates t at the left end or point I and decreases with d i s ­ tance to the right. T h u s , P ^ > P2 > P3 > P4• pre ssure is constant, Since the m a n ifol d the dif ferent pressures at taps I through 4 force the m a n o m e t e r fluid to the levels s h o w n . With valves A and B closed and valve C o p e n , the d i f f e r ­ ential pressures bet ween the pre ssure taps in the pipe line are given directly by the levels of the carbon t e t r a ch lorid e tubes and the m e r c u r y loop is isolated from the system. Since the m a n i f o l d pressure is c o n s t a n t , the following is true: (IO) R e a r r a n g i n g Eq. 10 gives: Pl-P2 = *f(%2-=l) + ( 11 ) Bc(^-Fl) where p^-pg is the d i f f e re ntial pressure between taps' I and 2. Also note that z 2_ z i = ~ ^ 2 " ' Srl ^ ° E g . 11 can now be rew ri t t e n a s : APl-2 = y2-?l ( #c- ( 12 ) Bf)' The head loss for steady flow in a hor izon tal pipe of constant diameter is e x p r e s s e d as a change in pre ssure head where Ap is the change in pressure in psf and ^ is the — 68 - specific weight in pcf of the fluid flowing in the pipe. T h u s , div idin g A p 1_2 by to give the head loss , , and c o n v ert ing the m a n o m e t e r readings from, inches to feet gives; A P 1-S ' O 2- V ^c" (13) 1-2 The spe cific weight of any fluid may be exp ress ed as ■ I W (S4,) where t ^ W is the specific weight of water and is the spe cific gravity of the fluid. r e a r r a n g e d and T h u s „ Eg. 13 can be can celed giving HL (14) 1-2 In this study the fluid flowing was w a t e r . T h u s , with = 1.0 , the head loss between tap I and any d o w n s t r e a m tap n is given by H Y n- Y 1 tlL 1 = (S -I). 1-n 12 c B. (15) D i f f e r e n t i a l m a n o m e t e r with mer cury loop for high Ap/ft With valves A and B open and valve C closed in Fig. 22, the m e r c u r y loop is inserted and gives the pressure dif ference bet ween tap 2 and tap 3. The m a n i f o l d pressure for tubes I and 2 is now dif fere nt from that for tubes 3 and 4. Eg. 15 holds for the head loss between taps I and 2: HL Ya-Y1 1-2 12 (S -I). c (16) — 69Likewise,, the head loss between 2 and 3 is given by Hr 2 -3 - T T jl lV (17) n which is similar to E q „ 16 except for the use of mercury rather than carbon t e t r a c h l o r i d e . The head loss between taps I and 3 is thus Hr _ Hr , Hr L l-3 L l-2 + L 2-3 1-3 Y2- Y 1 - T T j - (sc - n Yi-Ys + -TTj (s„ - n (18) E q 0 15 can then be used again to find the head loss between tap 3 and any other d o w n s t r e a m t a p , n, or (19) J3-n Using Eq. 18 and E q 0 19 the head loss between tap I and any tap d o w n s t r e a m from tap 3 can be found by Vn “V 2+V 3+Vn 1-n Y 2- Y 1 Yi-Yc - T T j (s c - 1 > + - T T j cV Y n- Y 3 li + - T 2 cs C-15 (2 0 ) R e a r r a n g i n g E q 0 20 gives the final m a n o m e t e r equ ation v.- (21) APPENDIX C STATISTICAL EQUATIONS S t a t i st ical analysis was used to find: (I) the slope and intercept of the d i m e nsi onles s hyd raulic grade lines, (2) the c o r r e la tion coe ffic ient and estimate of variance about r e g r e s ­ sion of the fitted dim ensi onles s hyd raul ic grade lines, and (3) the probable error in the average man om e t e r readings and concentration. A. Slope and intercept of grade lines The dim en s i o n l e s s hyd raulic grade line for fluid flowing in a pipe is r e p r e s e n t e d by the line that best fits a plot of Ap/ft —5--- versus distance along the pipe. The linear regression v /2g methods for d e t e r m i n i n g this line are developed below. A sample set of typical t w o - d i m e n s i o n a l data points (x ., y .) for n obs e r v a t i o n s are shown plotted in Fig. 23. The equation of the straight line fitted to this data is y = a+bx where a is the y - i n t e r c e p t and b is the slope. y Fig. 23. Sample linear t w o - d im ensio nal data -71- To make this line the "line of best fit," a and b must be d e t e r m i n e d such that the error, € ^ is a minimum. The error, €., is the d i f f e r e n c e between the e xp erim ental y. value c o r r e s p o n d i n g to a given equ ation y = a + b x . . and the y value pre di c t e d by the Therefore* ' £. = a+bx.-y^. Any single C. cannot be m i n i m i z e d without a f f e c t i n g the others. Thus, n the sum € . must be made as close to zero as possible. This i= l 1 sum can be made equal to zero by many choices of straight lines for which the pos itiv e and negative errors cancel. the sum of the squares of the However, if £. is minimized, the effect of sign on the error is eliminated. The refore, a and b must be d e t e r m i n e d such that the sum of the squares given by E in the equation n Y E= 9 (a+bx.-y.) (22) A is a m i n i m u m . This pro cedu re is known as the m e t h o d of least squares. E x p a n d i n g E q . 22 gives: E = X j (a^+2abxj-2ayj+b^x.^-2bx.y.+y.^). i= I (23) The m i n i m i z a t i o n cri terion requires setting the partial d e r ivat ive with respect to a and b equal to zero giving: - f 1—1 and (2a+2bXj-2y.) = 0, (24) -72- £ C2a x . + 2bx. - 2x^y^) i= l db (25) R e a r r a n g i n g and div i d i n g both equations by 2 and noting that n ^ a = an gives the fol lo w i n g e q u a t i o n s : i= I n n (26) E V i = an+b E x i 1=1 ^ i=l n n 0 a E x i + b E (xi ) • i=l 1 i=l (27) These two equations are known as the normal equations. Sim u l t a n e o u s solution of the normal equations gives the values of a and b for the line that best fits the data acc ording to the cri teria of least squares or n o n n n (28) n (x ) i= l (2 9 ) n I (x.)2 i i= l Li=*! These equ ations for a and b are rel ativ ely c u m b e r s o m e . E q . 28, h o w e v e r , can be r e a r r a n g e d and sim plif ied to give: a = y+bx (30) where y and x are the average values of y and x, and b is the slope com pute d by E q . 29. -73- E q 1 29 and E q „ 30 were used in the computer pro gram dis cussed in A p p e n d i x D to det ermine the equ atio ns of the dimensi onles s h y d r aul ic grade lines ups t r e a m and dow nstr eam of the valve being tested. Values of y c o r resp ond to dim ensi onles s d i f f e r ­ ential pre ssure heads (DIMPEL) and values of x to distance along the test section (x EL . C o r r e l a t i o n coe ffic ient and estimate of variance about r e g r e s s i o n It is des irab le to have some m ea sure of how well the exp e r i m e n t a l data est abli shes the d i m e nsi onles s hyd raulic grade line. One such indicator is known as the correla tion c o e f f i ­ cient ( C O E F ) . Ano t h e r is the est imat e of var iati on about regression (EVAR). The c o r r e la tion c oe ffic ient is a statistical mea sure of the degree to which two variables relatio nship . (x^, y .) form a linear The COEF varies between -I and +1, depending on the increase or decrease of one variable as the other increases. A c o r r e la tion of -I or +1 indicates a perfect linear rel atio nship . In other w o r d s , a straight line fitted to the data would pass through every point (x ., y .) . There­ fore, a COEF gre ater than -I or less than +1 indicates that not all data points ( x .„ y ^ ) form a linear rel ationship. Since the hyd raul ic grade line should be linear in the regions upstream and d o w n s t r e a m of the valve on a plot of pressure versus d i s ­ tance, a COEF much gre ater than -I or less than +1 would ' -74- indicate bad data. Mis read or incorrect data were determined by inc ludi ng a c o r r e la tion coe ffic ient calcula tion in the com puter pro gram used in r e d ucin g the data. to det er m i n e the COEF is dev elop ed below. The equation used Steele and Torrie (10) give the cor rela tion coefficient as COEF = E xy -/E(X)2 E(Y)2 (31) where Z Z xy (x-x.)(y-y.), i= l Z<x)2 Z Ii-Xi)2. i= I Z(Y)2 - Z Ci-Yi)2. i=l S u b s t i t u t i n g for ^ X Y , E x , and EY in E q „ 31 and exp anding gives the final equation: COEF (32) The est imat e of variance about regression is a st a t i s ­ tical m e a s u r e of the degree to which the data points deviate from the line fitted to the data by reg ress ion methods. In other w o r d s , the EVAR is a mea s u r e of the amount of e r r o r , ( . as shown in Fig. 23. Therefore, the smaller the value of -75E V A R , the closer the data points are to the fitted line. An EVAR of zero wou ld mean the line passes through every point, (Xj, y.). A mea s u r e of how acc urat ely the exp e r i m e n t a l data e s t a b l i s h e d the hyd ra u l i c grade line was provided by including a c a l c u la tion of the EVAR in the data reduction program. equ atio n used to det er m i n e the EVAR is developed below. The Steele and Torrie (10) give the est imate of variance about reg ress ion as E y5 EVAR ( E x y )2 E x 2 (33) n— 2 Making the same s u b s t it ution s as those leading to Eq. 32 and r e a r r a n g i n g gives the final equation: n EVAR n Ln1?iXiyi"1?1Xi i?/* I n (n-2) C. n (34) n 0 r n n y (x .) - y. x i=I Li=I Variance of m a n o m e t e r readings and con cent ratio ns In this s t u d y , readings were made of m a n o m e t e r levels and flow rates and the averages used to est ablish the d i m e n s i o n ­ less hyd ra u l i c grade line and c o n c ent ratio n of the mixture. An infinite number of such readings wou ld be req u i r e d to obtain the true means. Thus, it is des irable to have some mea sure of how closely the o b s erve d averages predict the true means that wou ld be obt ained from an infinite number of observations. Use — 7 6— of the S t u d ent -t d i s t r i b u t i o n as dis cu s s e d by Miller and Fre und (6) provides such an indicator. The process of ass igni ng a + error to an average, x, used to pre dict a true m e a n , yd , such that yU = x + £ , is known as interval estimation. Miller and Freund (6) show that interval e s t i m a t i o n met hods give the error in p r e dict ing the true mean as: (35) in nearly all m a t h e m a t i c a l handbooks and statistics texts as a function of the des ired confidence interval and the number of o bs erva tions n . S u b s t i t u t i n g the equ atio n for the variance of n o b s e r v a t i o n s , x . , into E q . 35 gives the probable error as (36) Previous work by Cha rley (I) d e t e r m i n e d that ten observations for a 90 percent confident interval were needed to reduce this error to less than .1 in. In add ition to p r e dict ing the accuracy of the average m a n o m e t e r r e a d i n g , the probable error was also used to find m i s r e a d or m i s p u n c h e d data. The probable error of each average m a n o m e t e r rea ding was cal cula ted in the computer pro gram used -77- to reduce the d a t a „ One or more bad readings would noticeably increase the error value. Bad data could thus be found by stu d y i n g the error output rather than all of the individual readings. APPENDIX D COMPUTER PROGRAM C C C C C C C C C C C C C C C C C C C C C C C C C WOODCHIP PIPELINE VALVE STUDY HEADLOSS CALCULATION PROGRAM (FORTRAN II) FROM READ-IN EMPIRICAL,DATA THIS PROGRAM CALCULATES THE AVERAGE MANOMETER READING,THE PROBABLE ERROR IN THE AVERAGE USING A 90 PERCENT CONF IDENC E LIMIT,THE DIFFERENTIAL PRESSURE HEAD OF EACH TAP RELATIVE TO THE FIRST UPSTREAM TAP, THE DIMENSIONLESS PRESSURE HEAD AT EACH TAP DEFINED AS (D IFF HD)/(V**2/2G ),THF REGRESSION LINES AND CORRELATION COEFFICIENTS OF THE DIMENSIONLESS HYDRAULIC GRADE LINE UPSTREAM AND DOWNSTREAM OF THE VALVE,THE PRESSURE DROP THROUGH THE VALVE AND THE HEAD LOSS COEFFICIENT AND PROVIDES PUNCHED OUTPUT OF X-DI STANCES, DIMENSIONLESS PRESSURE HEADS, CONCENTRATIONS, AND HEAD LOSS COEFFICIEN TS FOR USE IN OTHER PROGRAMS INPUT DATA ARRANGEMENT CARD I SGC,SGM,D,NS,T90,NTA,NTB,NFV CARD 2,3 X-D ISTANCES CARD 4 NVEL,RIDENT,NVO CARD 5 KVO,CVAL,MB,KU I,KUF,KD I,KDF, NTU,NTD,XVAL CARD 6 NOBS,NC,NML CARD 7+ QM,QW,14Y,0-4YM (FORMAT I ) (FORMAT 2 ) (FORMAT 3 ) (FORMAT 4 ) (FORMAT 5 ) (FORMAT 2) NOTE....PROGRAM WORKS WITH ONE SET OF VALVE DATA AT A TIME. EACH SET OF VALVE DATA MUST BE RUN SEPARATELY. DIMENSION SUY(IS) ,YB II8) ,YA(5) ,SSQY(IS) ,RIDENT(S) ,AYBRO(IS) ,SYSARf 118) ,YERR(18),SUYMI 18 I,SSQYMf18) ,AMBRQt18) ,SYMBRi 18) *YAERR!18) ,DPHD 2(18),DIMPHI I8), X ( 18) ,Y(18,10), Q M ( 10),QW(10),YM(5,10),CONf18 I DIMENSION B (2), A(2I,EVAR(2),COFT(2),HL(2) 1 FORMAT(2F6.2,F6.3,14,F8.3,312) 2 FORMAT (12F6.2) 3 FORMAT (I4,5A4,12) 4 FORMAK 13,F5.1,714,F6.2) 5 FORMAT (314) 4F FORMAT (IHl) 46 FORMAT(3 X ,I B H RUN IDENT IF ICAT ION,5 X ,5A4,5 X ,I2 , 2 IH FPS NOMINAL VELOC I IT Y,4X,6HVALVE , F 5 . I,15H PFRCENT CLOSED«4X,4HC0NC^F7.2,BH PERCENT) 47 FORMAT (2OX,49HDISTANCE FROM FIRST UPSTREAM PRESSURE TAP IN FEET) 48 FORMAT (27X,68H Xl X2 X3 X4 X5 X6 X7 X8 X9 XlO X 111 Xl2 X13 X14) 49 FORMAT (26X,2?F6.I) 50 FORMAT (2 O X,2 5HMANOMET ER READING OF Y(M) Y2 Y3 Y4 YS Y6 QW CONC Yl 51 FORMAT (IX,11IHRUN QM HGl HG 2 HG3 H G 4 ) YG Y9 YlO Yll Y12 Y13 Y14 I Y7 (I3,2F6.0,F6.3,1X,22F5.I) 40 FORMAT 41 FORMAT (IX) 52 FORMAT(21X,19HAVERAGE VALUES FOR ,13,5H RUNS) 53 F0RMAT(3X,2r6»0,F6.3,lX,22F5.1) 54 FORMAT (20X,47HPROBABLE ERROR WITH 90 PERCENT CONFIDENCE LIMIT) 55 FORMAT (IX, 5HERR0R, 9X,F6.2,2X,22F5.2) 56 FORMAT (20X,26HDIFFERENTIAL PRESSURE HEAD) Yl Y2 Y3 Y4 Y5 Y6 Y7 57 FORMAT (3X , 9TH Y9 YlO Yll Yl 2 Y13 Yl 4) I YB 58 FORMAT (3X,18F7.3) 59 FORMAT (20X,27HDIMENS IONLESS PRESSURE HEAD) (I3,2F6.0,F6.3,1X,7F5.1,5X,14F5.1) 60 FORMAT 61 FORMAT(3X,2F6.0,F6.3,1X,7F5.I»5X»14FB. I) 62 FORMAT (IX ,5HERR0R, 9X,F6.2,2X,7F5.2,5X,14F5.2) 63 FORMAT (3X ,7F7.3,7X,10F7.3) 64 FORMAT t3X,I2,21H FPS NOMINAL VELOCITY,10X,SHVALVE ,I3,15H PERCENT I CLOSED) 19 READ I,SGC,SGM,D ,NS,T90,NT A ,NTB,NFV READ 2,(X(I), 1=1,NS) 20 READ 3, NV E L , RIDENT , NVO PUNCH 175,RIDENT,NVEL 175 FORMAT{5 X ,2OHRUN IDENTIFICATION I 3HFPS) PUNCH 176,(X(I),1=1,NS) ,5A 4 ,2X ,I5HN0MINAL VELOC , I2,lX, I NO I c C C C C C C C C C C C C C C 176 FORMAT(5X .I3HX-DISTANC ES ,9F6.2/(5X»9F6.2) ) 21 RFAD 4,KVO,CVAL,MB,KUI,KUF,KDI,KDF,NTU,NTD,XVAL PUNCH I77,CVAL 177 FORMAT(5X,15HVALVE CLOSURE ,F5.1,8H PERCENT) 22 READ 5, NOBS, NC, NML x (I)=d i s t a n c e f ro m u p s t r tap (f t ) n v e l =n o m i n a l v e l o c i t y RIDENT =RUN IDENT IF ICAT KVO=VALVE OPENING IDENTIFICATION CODE CVAL=VALVE CLOSE IN PERCENT MS=MERC LOOP CODE KUI=INIT UPSTR TAP FOR SLOPE KUF=FINAL UPSTR TAP KDI=INIT DNSTR TAP FOR SLOPE KDF=FINAL DNSTR TAP NTU=NO OF TAPS UPSTR SLOPE NTD NO OF TAPS DNSTR SLOPE NOBS =NO OF OBSERVATIONS PER RUN NC=CONCENTRAT ION CODE NML =NO OF MERCURY TUBES USED SGC=SP GR CARB TET SGM=SP GR MERC D=DIAM NS=NUMBER OF TAPS T90=PR08 ERROR TEST VALUE NVO=MAX NO. OF VALVE SETTINGS PER NOMINAL VELOCITY NFv =FINAL VELOCITY OF THE SERIES DO 23 J=I ,NOBS 23 READ 2, QM(J), QW(J), (Y(I,J ),I= I ,NS), (YM(L ,J ),L= I,NML) QM =M IX FLOW QW =CLEAR WATER FLOW Y(I)=CARB TFT MAMOM READINGS YM=MERC MANOM READINGS C AND CC CONVERT MANOM READINGS TO PRESSURE HEADS-CARB TET AND MERC RESPECTIVELY C=(SGC-I.)/12. CC=(SGM-I.1/12. OBS=NORS SSQC=O.OO SUCON=O.00 SUQM=O.OO SUQW=O.OO INITIALIZATION FOR SUMATION DO 24 II= I ,NS SSQY(II)=0.00 24 SUY(II )=0.00 DO 26 JJ = I ,NOBS SUQm =SUQM + QM(JJ) SUQw =SUQW + QW(JJ) SUMMING QMIX AND QWATER DO 26 N= I, NS 25 SUY(N)= SUY(N) + Y(N,JJ) a f C C C C C -81 C SSOY(N)=SSQYtN)+Y(N,JJ)**2. YB(N)=SUYtN)/OBS AYBRQtN)=ABSF((SSQY(N )-SUY(N )#*2./OBS>/(OBS*(OBS-I.))) SYBAR(N)=SQRTF!AYBRQ(N)) SUY(N)=SUM OF MANOM READ SSQY=SUM OF READ SQUARED YB=AVG READING AYBRQ=VARIANCE OF Y SYBAR =STD DEVIAT OF Y YERR =PROB ERROR OF Y 26 YERR(N)=T90*SYBAR(N) DO 28 IK = I»NML SUYM(IK)=0.0 SSQYM(IK)=0.0 DO 27 JK = I ,NOBS SSQYM (IK) =SSQYM (IKH-YM (IK,JK)**2. 27 SUYM(IK)=SUYMtIKHYM(IK,JKl YA(IK)=SUYM(IK)/OBS AMBRQt IK)=ABSFf (SSQYM(IK)-SUYM(IK )**2./OBS)/(OBS*(OBS-I.) )) SYMRRl IK)=SQRTFtAMBRQl TK)) SUYM,SSQYM,YA,AMBRQ,SYMBR,YAERR SAME AS ABOVE BUT FOR MERCURY TUBES 28 YAERRt IK )=T90*5YMBR(IK) DO 30 N=I ,NOBS CON(N)=I.-QW(N )/QM(N ) SUCON=SUCON+CON(N ) CON=CONCENTRAT ION SUCON =SUM OF CONC. ACON =ABSOLUTE VALUE CONC. ACON=ABSFtCON(N) > 30 SSQC =SSQC +AC ON**2 CONC=SUCON/OBS PC ON=C ONC*100. IF (PCON-I.0) 90,90,91 90 PCON=O.OO 91 c o n a r =a b s f (sucoN) TESTA=(SSQC-CONAB**2/OBS)/(OBS*!OBS-I.)) VARNC =ABSFt TESTA) SCBAR=SQRTF(VARNC) SSQC=SUM OF ACON CONC=AVG CONCENTRATION SCBAR=STD DEVIATION OF CONG. CERR=PROB ERROR OF CONC. QA=AVG OMlX QB=AVG QWATER CERR=T90*SCBAR*100. QA=SUQM/OBS OB=SUQWZOSS n r> n n C C C C GO TO (100, HO, 120, 130), MB MB=MERCURY LOOP CODE MB=I NO MERCURY LOOPS USED MB =2 I MERCURY LOOP AND ALL TAPS MB=3 2 MERCURY LOOPS AND ALL TAPS MB=A 2 MERCURY LOOPS NO TAP NO.8 (GATE VALVE ONLY) 100 DO 101 1=1,NS 101 DPHDt I)=C*(YB(I)-YG(D) DPHD=DIFFERENTIAL PRES HEAD REFERENCED TO TAP NO ONE GO TO 129 H O DO 119 I= I ,NS IF (I-NT A )111,112,113 FIRST MERCURY LOOP IS BETWEEN TAP NO. NTA AND TAP NO. NTA-I 111 DPHD(I)=C*(YD(I)-YG(I)) GO TO 119 112 DPHDt I)=C*(YBtA)-YSt I))+CC*(YA(2)-YA(I)) 113 DPHD(I)=C*(YB(A)+YB(I)-YB(S)-YB(I>)+CC*(YA(2)-YA(I)) GO TO 119 119 CONTINUE GO TO 129 120 DO 229 1=1,NS IF (I-NTA)121,122,123 121 DPHD(I)=C*!YB(I)-YB(I)) GO TO 229 122 DPHD(I )=C*(YB (A)-YB(I) )+CC*(YA(2)-YA(I)) FIRST MERCURY LOOP IS BETWEEN TAP NO. NTA AND TAP NO. NTA-I AND THE SECOND MERCURY LOOP IS BETWEEN TAP NO. NTB AND TAP NO. NTB-I GO TO 229 123 IF (I-NTD)163,125,126 12 5 DPHD(I)=C*(YB (A)-YBt I) )+CC*(YA(A )+YA(2!-YA(3)-YA(I)) GO TC 229 126 D P H D (I)=C* (Y B (I)+YB(A )-Y B (6 )-Y B (I ))+CC* (YA (A )+Y A (2)- Y A ( S ) - Y A d )) 229 CONTINUE GO TO 129 130 CONTINUE DO 139 1= 1,NS IF (I-NTA) 131,132,133 131 DPHDf I)=C*(YBtI)-YB(I)) GO TO 139 132 DPHDt I)=C*(YB(S)-YBtl) )+CC*(YA(2)-YA(I)) GO TO 139 133 IF (I-NTB) 134, 135,136 134 DPHDt I)=C* (Y B (5) + YB(7)-Y B (6)-Y B (I ) )+CC*(Y A (2)-Y A (I ) ) GO TO 139 135 DPHDt I)=O GO TO 139 136 DPHDt I )=C* (YB (I )+YBt 7) +YBt 5 )- YB (9 )- YB (6 )- YB (I )>+CC* (YA (4.) + YA (2 )-YA I (3) -Y A ( I ) ) 139 CONTINUE 129 VU=(QA*576.)/(448.8*3.1416*D**2) DENOM=VU**2/64•4 C VU=VEL OC IT Y IN FPS DENOM=VELOC ITY HEAD D IMPH=D IMENS IONLESS PRES HEAD DO 141 1=1,NS 141 DIMPH(I)=DPHDt I)/DFNOM PRINT 45 PRINT 41 PRINT 41 PRINT 46, RIDENT , NVEL ,CVAL , PCON PRINT 41 PRINT 41 PRINT 47 PRINT 48 PRINT 49,(X(I),1=1,NS) PRINT 41 PRINT 50 PRINT 51 PRINT 41 J=NTA+I K=NTB+! GQA=QA/448.8 GO TO (70,75,75,80), MB C PRINTING SEQUENCE— MB SAME AS ABOVE 70 DO 72 L=I,NOBS PRINT 40,L,QM(L) ,QW(L) ,CON(L) ,(Y (I,L )* I= 1,NS 5 72 PRINT 41 , co f PRINT PR INT PRINT PRINT PRINT PRINT PRINT PRINT PRINT PRINT 73 PRINT PRINT PRINT PRINT PRINT PRINT PRINT PRINT PRINT PRINT PR INT PRINT GO TO 75 DO 77 PRINT 77 PRINT PRINT PRINT PRINT PRINT PRINT PRINT PRINT PR INT PRINT PRINT GO TO 41 52 , NOBS 41 53, QA,QB,CONC »(YB(I),I = I,NS) 45 46» RIDENT, NVEL ,CVAL , PCON 41 41 54 55 ,CERR ,(YERR(I),1= 1,NS) 41 41 57 56 58» (DPHD( I),1= 1,NS) 41 59 58, (DIMPHt I),1= 1,NS) 41 41 41 41 95 L= I,NOBS 40 ,L »QM (L )»QVI(L ),CON (L) »(Y( I,L )»1 = 1, NS )»(YM (I 41 41 52 , NOBS 41 53, QA ,QB,CONC,(YB( I),I= I,NS),(YA(I) ,I= I♦NML) 45 46» RIDENT, NVEL ,CVAL , PCON 41 41 54 55, CERR,(YERRI I),I= I ,NS),(YAERRI I),1 = 1,NML) 73 I CD I 80 DO 82 L=IiNOBS PRINT 60»L >QM( L) ♦QW(L) ,CONIL )»(Y( I»1) »I= I >J) »( Y( I »L )»I=K»NS) ,(Y M U I,L),I=I,NML) 82 PRINT 41 PRINT 41 PRINT 52, NOBS PRINT 41 PRINT 61 , QA ,08 ,CONC ,IYB (I),I= I ,J ),IYB (I) ,I=!<,NS ),IYA II ),I= I ,NML ) PRINT 45 PRINT 46, RIDENT,NVEL,CVAL,PCON PRINT 41 PRINT 41 PRINT 54 PRINT 62,CERR,(YERR( I) ,I= I ,J ),IYERRt I),I=K,NS) ,(YAERRt I),I= I,NML) PRINT 41 PRINT 41 PRINT 57 PRINT 56 PRINT 63, (DPHDt I),I= I ,J),(DPHDt I),I=K,NS) PRINT 41 PRINT 59 PRINT 63, (DIMPHt I), I= I,J},!DIMPHt I) ,I=K ,NS) PRINT 41 PRINT 41 PRINT 41 PRINT 41 95 TS=NTU NA=KUI NO=KUF NB=I 145 SXY=OeO SSX=OeO ScIY= O eO SX=0.0 SY=OeO DO '146 I=NA ,NO SX = SXtXt I) I Ol cc I C C C C SY =SY+DIMPH(I! SSY=SSYX DIMPHt I )5**2 SSX=SSX-H X( I55**2 146 SXY=SXY-K X (I5*D IMPHt I55 THIS SECTION COMPUTES SUMS FOR USE IN THE REGRESSION CALCS SYS=SY**2 SXS=SX**2 • AX=SXZTS AY=SYZTS BINB)=(SXY-(SX*SY)ZTS>Z(SSX-(SXSZTS55 A(NB)=AYHAX*BtNB 55 EVAR(NB) = !SSY-(SYSZTS)-I ((TS*SXY)-(SX*SY)5Z( (TS*SSX)-SXS 555Z (TS-2. I5 COET(NB)=SQRTFt ((TS*SXY)-(SX*SY) )**2Z( ((TS*SSX)-SXS5*( (TS*SSY)-SYS I 55 5 IF ICOFTtNB5-1 .05 600,500,500 500 COFT(NB 5=1 .000000 600 CONTINUE NB = I FOR UPSTR SECTION NB =2 FOR DNSTR SECTION B = REGRESSION LINE SLOPE A=Y-TNTERCEPT EVAR=EST OF VARIANCE ABOUT REGRESSION COFT=CORRELATION COEFFICIENT IF(NB-I) 147,147,148 147 TS=NTD NB =2 NA=KDI NO=KDF GO TO 145 148 DO 149 NI = I ,2 149 HL (N I5=A(N I5+ (XVAL*B(NI?) HLCOFT =HLf 2)-HL(I) H L O S S = ( H L C O F T )*DENOM C HLOSS=HEAD LOSS OCURRING AT THE VALVE C HLCOFT=THE HEAD LOSS OOEFFIOIENT=DIFFERENCE BETWEEN UPSTR AND DNSTR C REGRESSION LINES AT THE VALVE WHERE X=XVAL PRINT 150 150 FORMAT(5X,55HREGRESSION INFO UPSTREAM DOWNSTRE IAM ) co f C C C C C C C PRINT 41 PRINT 151,(0(1),I=I,2) 151 FORMAT(5X,5HSLOPE,23X,F9.6,8X,F9.6) PRINT 152,(Al I),1= 1,2) 152 FORMAT(5X,9HINTERCEPT,19X,F9.6,8X,F9.6) PRINT 153,(EVAR( I),1= 1,2) 153 FORMAT(5 X,20HESTIMATE OF VARIANCE,7X,F10.6,7X,F10.6) PRINT 154,(COFTt II,1=1,2) 154 FORMAT(5X ,23HCORRELAT ION COEFFIC IENT,5X,F9.6,8X,F9.6) PRINT 41 PRINT 180, VU 180 FORMAT(5X,17HAVG TRUE VELOCITY,F8.3,IX,IOHFT PER SEC) PRINT 181, QA 181 FORMAT(5X,17HAVG MIX FLOW RATE,F8.3,IX,11 HGAL PER MIN) PRINT 182, GOA 182 FORMAT(5X,I7HAVG MIX FLOW RATE,F8.4,IX ,I3HCU FT PER SEC ) PRINT 41 PRINT 155,HLCOFT, HLOSS 155 FORMAT(5X,3 5HHFADLOSS COEFFICIENT OF THIS RUN IS ,IX ,F9.6,5X ,I2HHEA ID LOSS IS,1X,F9.6,5H FEET) PUNCH 178,PCON,QA,HLCOFT I78 FORMAT(5X,19HCHIP CONCENTRATION ,F6.2,18H PERCENT QMIX=,F5.0,1 I6H HEADLOSS COEF ,F8.4) PUNCH 179, (DIMPHIN),N=I,NS) 179 FORMAT(5X ,2 IMD IMENLESS PRES HEADS ,9F6.2/(5X,9F6.2) ) IF(NC-I) 22,163,163 NC=O READS MORE DATA AT DIFF CONCENTRATION FOR SAME VALVE CLOSURE 163 IF(KVO-NVO) 21, 171, 171 KVO LESS THAN NVO READS MORE DATA FOR DIFF VALVE CLOSURE 171 IF(NVEL-NFV) 20,37,37 NVEL LESS THAN NFV READS MO°E DATA AT DIFF NOMINAL VELOCITY EACH NOMINAL VELOC DATA SET INCLUDES DATA SETS FOR DIFFERENT VALVE CLOSURES EACH OF WHICH INCLUDE DATA SETS FOR DIFFERENT CONCENTRATIONS DATA INPUT IS IN ORDER OF LOWEST NOMINAL VELOC TO HIGHEST, ZERO PER CENT VALVE CLOSURE TO MAX CLOSURE, ZERO CONC TO MAX CONC. 37 CALL EXIT END , co Tj APPENDIX E SUMMARY OF COMPUTED RESULTS TABLE IV Plug Valve Test, July 1968 Valve V 4 fps C KL V 6 fps C KL V 8 fps C KL V 10 fps C KL 3.950 0.00 4.4% 3.927 8.12 4.005 20.21 .009 6.009 0.00 .024 5.987 10.04 .194 6.022 20.40 .192 7.989 0.00 .175 7.989 9.69 .2 68 7.974 19.51 .052 .067 .094 9.968 0.00 9.927 9.41 9.932 19.71 .222 .229 .224 3.939 0.00 25.0% 4.02 6 9.74 3.945 19.14 .226 5.969 0.00 1.002 5.932 9.34 3.380 5.908 18.71 .655 7.924 0.00 .673 7.958 9.66 .763 7.932 19.74 .736 10.008 0.00 .730 10.034 9.20 .837 9.929 19.62 .562 .559 .590 4.008 0.00 5.982 6.013 0.00 50.0% 3.937 8.34 7.501 6. 047 10.98 3.927 18.75 13.462 5.890 18.33 6.700 7.948 0.00 7.442 8.034 10.44 8.856 7.937 19.40 6.003 75.0% 6 fps 70.0% 8 fps Repeat 3.906 8.30 25.0% Closed 6.564 9.932 0.00 6.464 7. 158 10.005 10.06 6.970 8.293 9.990 20.18 7.932 0.00 38.458 7.953 0.00 29.261 7.916 9.74 27.028 Repeat Repeat 10.72 7.124 6. 130 8.023 .036 6.023 10.55 50.0% Closed 50.0% Closed TABLE Summary Valve V 4 fps C KL of Computed Results V 6 fps C V - Ball KL V Valve Test, 8 fps C KL July 1968 V 10 fps C KL 3.965 0.00 -.007 6.013 0.00 0.0% 3.987 10. 16 -.009 6.035 11.05 3.926 19.30 .021 5.987 19.94 -.125 7.991 0.00 .052 7.969 9.90 .027 8.044 20.26 .015 10.048 0.00 .015 10.000 10.06 .003 9.939 19.77 .02 6 .108 .052 3.919 0.00 .705 6.052 0.00 2 5.0% 4.005 11.22 .963 6.065 10.98 3.906 18.34 1.196 6.000 19.65 .586 7.929 0.00 .790 7.979 9.86 .897 7.93 7 19.43 .838 10.039 0.00 .815 10.055 9.52 .901 9.966 19.70 .820 .847 .952 3.940 0.00 6.397 6.02 6 0.00 50.0% 3.940 9.09 7.287 6. 000 10.46 3.929 18.83 9.176 5.963 20.07 6.458 7.927 0.00 7.140 8.005 10.42 7.930 7.914 19.19 6.374 9.942 0.00 6.508 6. 757 9.955 9.56 6.936 7.366 10.029 20.57 7.863 7.961 0.00 22.029 7.984 10.08 23.167 8.008 20.31 24.353 67.5% 5.984 5.969 75.0% Repeat 4.017 20. 61 0.0% Closed 0.00 50.826 9.94 49.767 Repeat .050 6.034 10.61 25.0% Closed Repeat .814 7.953 0.00 6.111 50.0% Closed TABLE VI Summary Valve V 4 fps C KL of Computed Results V 6 fps C KL - Pinch V Valve 8 fps C T e s t , July KL V 1968 10 fps C KL 3.908 0.00 28.5% 3.930 8.85 3.961 15.62 .104 6.039 0.00 .323 5.974 10.07 .5 62 5.908 19.76 .140 7.882 0.00 .144 7.948 9.92 .263 7.935 19.33 .109 10.061 0.00 .115 9.996 9.59 .154 9.913 19.56 .064 .079 .105 3.963 0.00 40.0% 3.895 8.24 3.950 19.75 .169 5.969 0.00 .558 5.987 9.99 .874 6.013 20.26 .2 62 7.950 .354 7.966 .5 78 7.919 .409 10.018 0.00 .429 10.016 10.07 .506 9.937 19.97 .305 .3 68 .447 0.00 9.85 9.94 3.992 0.00 .521 5.979 0.00 .755 7.940 0.00 .859 9.940 0.00 .845 50.0% 3.971 9.54 .754 5.987 10.39 .899 7.914 9.74 .937 10.084 10.60 .917 3.971 19.67 ].5 62 6.031 20.72 L.205 7.961 19.51 1.117 9.911 19.62 I.028 60.0% 70.0% 5.987 0.00 I.484 7.979 0.00 1.894 5.958 9.48 1.737 7.974 10.03 2.055 5.953 20.02 2.323 7.966 19.92 2.296 7.997 0.00 3.631 8.000 10.52 4.264 Repeat Repeat Repeat 3.940 19.03 .514 5.953 9.66 .554 7.901 0.00 .820 28.5% Closed 40.0% Closed 50.0% Closed TABLE VII Summary of Computed Results Valve % Clos ure % of Stem Travel 0.0% 14.2% 38.9% 68.4% V 6 fps C - Gate KL V Valve Test, 8 fps C September KL V 1965 LO fp s C KL 0.0% 5.908 0.00 5.908 4.87 5.935 11.01 5.945 15.56 5.935 20. 70 0.013 0.015 0.017 0.018 0.032 7.869 0.00 7.848 4.73 7.895 10.59 7.843 15.00 7.869 20.27 .016 .017 .017 .023 .028 9.856 0.00 9.804 4.80 9.856 10.08 9.878 15.31 9.856 19.89 0.014 0.027 0.023 0.023 0.025 25.0% 5.882 0.00 5.942 5.40 5.935 10.57 5.961 15.79 5.963 20.21 0.2 62 0.278 0.298 0.316 0.363 7.869 0.00 7.869 5.32 7.895 10.60 7.817 14.38 7.885 19.76 .267 .269 .284 .301 .316 9.882 0.00 9.935 5.79 9.827 9.81 9.856 15.12 9.882 20. 11 0.193 0.277 0.254 0.281 0.336 50.0% 5.908 0.00 5.908 4.87 5.911 10.21 5.935 15.42 5.935 20. 70 1.531 1.674 I .722 2.273 2.720 7.838 0.00 7.874 5.38 7.843 9.67 7.922 15.84 7.840 19.64 1.841 1.920 2. 101 2.359 2.407 9.830 0.00 9.840 4.89 9.804 9.60 9.908 15.57 9.864 19.96 1.980 I .939 2.210 2.356 2.500 75.0% 5.908 0.00 18.008 7.843 0.00 18.117 9.830 5.908 4.87 20.898 7.874 5.04 19.207 9.830 5.924 10.28 23.380 7.848 9.73 20.986 7.854 14.78 18.030 7.872 20.29 20.816 0.00 14.993 5.59 16.336 L I T E R A T U R E CITED 1. Charley, Robert W., The Effect of Chip-Shaped Solids on Energy Losses in Axi-Symmetric Pi pe ~Expans ions , UnpufP"" lished Master’s Thesis, Department of Civil Engineering and Engineering Mechanics, Montana State University, Bozeman, Montana, June, 1965. 2. Flow of Fluids through Valve Fittings and Pipe, Technical Paper No. 410, Crane Company, Chicago, Illinois , 1957. 3. Hi no, Mikio, "Turbulent Flow with Suspended Particles," Journal of the Hydraulics Division, ASCE Proceedings, V0I. 89, No. 4, July, 1963, pp. 161-185. 4. Hunt, William A., An Economic Analysis of Transporting Low Value Forest Products Continuously in Hydraulic Pipe Lines , Unpublished Final Report, Department of Civil Engineering and Engineering Mechanics, Montana State University, Bozeman, Montana, March, 1965. 5. Liptak, Bela G., "Control Valves for Slurry and Viscous Service," Chemical Engineering, April 13, 1964, pp. 185-192. 6. Miller, Irwin and Freund, John E., Probability and Statistics for Engineers, Prentice Hall, Inc., Englewood Cliffs, New Jersey, 1965. 7. Pao, Richard F., Fluid Mechanics, John Wiley and Sons, Inc., New York and London, 1965. 8. Rouse, Hunter, Engineering Hydraulics, Proceedings of the Fourth Hydraulics Conference, Iowa Institute of Hydraulic Research, John Wiley and Sons, Inc., New York, 1950. 9. Soo, Shao L., Fluid Dynamics of Multiphase Systems, Blaisdell Publishing Co., Waltham, Massachusetts, 1967. 10. Steele, Robert G. D. and Torrie, James H., Principles and Procedures of Statistics, McGraw-Hill Book Co., Inc., New York, i960. 11. Streeter, Victor L., Fluid Mechanics, McGraw-Hill Book Co., Inc., New York, 1966. 12. Zandi, Iraj and Govatos, George, "Heterogeneous Flow of Solids in Pipe Lines," Journal of the Hydraulics Division, ASCE Proceedings, Vol. 93, No. 3, May, 1967, pp. 145-159. MONTANA STATE UNIVERSITY LIBRARIES 762 100 4578 ' J631 • cop.2 » I Johnson, David Allan The effects of chip­ shaped solids on valve head loss characteristii :s N A M K a n 6 A 6 6 i* AA» J.:. 6:). KcU^y •'t K * - cq /d/ N r v lK J b u Z r W * I £ dfttf/Hi’tWfUlt %— s V y cat