ML20081F609

From kanterella
Jump to navigation Jump to search
Part 21 Rept Re Edward Controlled Closure Check Valves & Testable Piston Check Valves Dashplate Bolt Failures Due to Rapid Valve Opening or Impulsive Differential Pressure Surge Across Dashplate.Design Changes Recommended
ML20081F609
Person / Time
Site: Perry, Seabrook  FirstEnergy icon.png
Issue date: 04/23/1991
From: Sook R
ROCKWELL INTERNATIONAL CORP.
To: Rossi C
Office of Nuclear Reactor Regulation
References
REF-PT21-91 NUDOCS 9106110198
Download: ML20081F609 (25)
v • d • e


Text

{{#Wiki_filter:-. DP F EdwardValvesinc e a.,.,8 April 23, 1991 Director, Of fice of 11uclear Reactor Regulation U. S.11uclear Regulatory Commission (OWFfJ) Washington, DC 20555 Attention: !4r. Charles Rossi

SUBJECT:

EDWARD TESTABLE CHECK VALVES EDWARD cot 3 TOLLED CLOSURE CHECK VALVES CODE OF FEDERAL REGULATIO!JS 10 CFR PART 21 Size 12 Edward Testable Check Valves at Cleveland

Reference:

Electric Illuminating Co., Perry Station Size 18x16x1B Edward Controlled Closure Check Valves at Public Service of flew Hampshire, Seabrook Station The purpose of this letter is to inform you that modification of certain Edward (formerly Rockwell) testable check valves and controlled closure check valves (CCCVs) in U. S. nuclear power stations is recommended. Internal bolting failures were experienced in the referenced valves and stations. An evaluation by Edward indicates that this problem may represent a safety hazard as defined in the Code of Federal Regulations 10 CFR Part 21; however, a systems evaluation by owners is required for a final determination. It has been determined that the only other U. S. utilities that were delivered valves of this basic type are: Cincinnatti Gas & Electric Co., Zimmer 11uclear Station; size 18 controlled closure check valves (plant cancelled) South Carolina Electric & Gas Co., V. C. Summer 11uclear Station; size 18x16x18 controlled closure check valves l I 110073 wn. now w: OW'O 'I p); f, DOC r P.O. Box 1961. Roleigh. N C. 27602 1900 South Somden Street, Rotelgh. N C. 27603 919 B32-0525 )

e \\ Di r.act er, Office of Nuclear Regulation l Apr1T 23, 1991

)sgc 2 Please note that the valves furnished for the Summer and Zimmer station had a different internal design with bolting that was not subject to primary stress loading.

It is considered that the Zinser valves are not an issue due to plant cancellation, but Edward recommends review of the Sunner internal design against user-supplied pressure surges when Perry and Seabrook valve designs are reviewed. White a number of CCCVs, most with bolting very similar to that at Seabrook, were furnished for nuclear plants in Belgium, Spain, Taiwan, Korea, and the Phillipines, the Zimmer and Summer CCCVs are the only onec in the U. S. (beside valves furnished for Perry and Seabrook). Edward has never been informed of any internal bolt failure problems in any of the other CCCVs, here or abroad. The attached technical report, EDWARD CONTROLLED CLOSURE CHECK VALVES AND TESTABLE CHECK VALVES --- DASHPLATE BOLT FAILURES, summarizes tne Edward evaluations of the incidents at Perry and Seabrook. Note that the incident at Perry (December 1990) was originally concidered a result of a spurious tasting condition and was not reported. However, the incident at Seabrook (April 1991) resulted in a jamming of the valve that could impair its safety related function. A temporary modification (doubling the number of bolts) was made at Seabrook pending more detailed analysis. The Edward report concludes that bolt failures at Perry and Seabrook were results of " slam-open" impact or unanticipated transient flow or pressure surges. It is not considered that these valves had a design deficiency, but they were exposed to system conditions that were not specified or considered in design. As described in the report, Edward proposes to modify the disk-piston stops to insure that opening impact does not disturb the internal bolting. Edward also proposes increasing the internal bolting area uhere necessary, but user-specified pressure surge design criteria are required for establishment of the bolt sizes and quantity. Size 6 testable check valves and size 20 CCCVs at Perry also have bolting designs similar to that used in the valves described in the attached report. Edward has determined that the bolting in these valves is stronger, proportionately, than that in the valves that experienced failures, but they should be evaluated against user-specified criteria. Responsible parties at the Summer, Perry, and Seabrook stations have been made aware of our recommendation and will be delivered copies of this letter and the attached report. 1 i

s N Director, Office of !!uclear Regulation April 23, 1991 . Page 3 If you have any questions concerning this notice, please direct them to me at (919) 831-3300 or to Roger D. 11orden, Vice President, Engineering, at (91 9) 831-3207. Very truly youru, w l R. Sook President cc R. A. Bandukwala R. D. 11orden J. C. Westendorf l i i l l I n - _. _ _ _,.. _ _ _ _.. _ _.... ~.... _ _ _. _. _ _ _ _ _ _. - _ _. _ _ -. _ - -

L. Doch9 ste Bolt Failures Pace 1 1 EDWARD CONTROLLED CLOSURE CllECK VALVES AND TESTABLE PISTON CHECK VALVES DASHPLATE BOLT FAILURES ABSTRACT: Failures were experienced in dashplate hold-down bolts in one size 12 Edward testable piston check valve at the Perry Nuclear Power Station (Cleveland Electric Illuminating Co.) in December, 1990. Similar failures were found in 2 of 4 size 18x16x18 Edward controlled closure check valves (CCCVs) at the Seabrook Nuclear Power Station (Public Service Co. of New Hampshire) in April 1991. These valves have different functions, but they have similar internal design features. They incorporate "dashplates" designed to withstand a high differential pressure in the downward i direction but have only relatively small " hold-down" bolts to withstand loads in the upward direction. Analyses of these incidents has shown that the observed failures i 1 were caused by either (1) impact as a result of very rapid valve opening or (2) an impulsive differential pressure surge across the dashplate due to unidentified system flow or pressure influences. No fundamental design deficiency exists, because normal operation as a check valve does not pro:. luce significant loading on the bolts in question. Design changes are proposed to prevent occurance of hold-down bolt i failures in other Edward testable piston check valves and I contro12ed closure check valves which may be exposed to opening i impact or impulsive surges. Information is required from users to i _ establish specific design criteria for the changes. EDWARD VALVE DESIGNS INVOLVED: Exhibit 1 is eheet 1 of Edward (formerly Rockwell International) drawing D82-24401-18, which illustrates a size 12 Figure 4094 (WCC)JNOTY testable piston check valve furnished for the Perry Nuclear Power Station (BWR). This is a normally-closed valve with a provision for-opening (for test purposes) by injection of water beneath the piston from an external source. A position indicator with proximity switches permits verification of satisfactory operation. The "dashplate", piece 26, acts as a pressure barrier to permit pressurization of the volume below the piston. Downward pressure loads are reacted by a segmented retaining ring, piece 27, which engages a groove in the valve body. The dashplate is

R.n.nhplate_hlt. rallurca l'aae 2 held in place by 6 cap screws (hold-down bolt s), piece 24, which eBgage a lower lock ring, piece 28. Since no significant upward loads on the dashplate weres expected, these screws were sized for attachment purposes only. Exhibit 2 is sheet. 1 of Edward drawing D84-30371-07 which depict.s a size 18x16x18 rigure 2092 (WCC)DJOTY Edward cont olled closure check valve (CCCV) furnished for the Seabrook lluclear Power Station (PWR). This is a normally open valvo located outside containment in a secondary syst em main feedwater line. Its function is to prevent severse flow in the event of a loss of feedwater supply for any reason, including a full line rupture in the upstream piping. In this valve, the dashplate, piece 3, provides the contrclied closure function; a contolled clearance (small, but with a free-running fit) between the dashplate and the disk-piston, piece 2, provides a " dashpot" function to limit the valve closure speed (and associated water hammer pressure surgo) in the event. of upstream line rupture [ Exhibit 3 is a copy of.\\CtE paper 00-C2/PVP-27 which do. scribes ananlytical and experimental work supporting development of this valve type). !Jote that this valve also incorporates a substantial segmented ring, piece 5, engaging the body t o support high potential downward differential pressure on t.he dashplate, but it also has relat.ively light hold-down bolts (piece 6) designea for attachment only. DASilPLATE BOLT FMLURE I!JCIDE!JTS : Perry liuclear Power Station: During a Unit 1 outage in December 1990, Perry pertonnel contacted Edward by telephone concerning a size 12 testable piston check valve as described above (Exhibit 1). The position indicator indicated that the valve was not fully closed. The valve cover and piston were removed, and Edward was informed that broken hold-down bolts were found between the dashplate and piston; the bolts were interfering with full closure of the disk-piston assembly. Edward was informed that a spurious pump startup during tests in a partially dry system had caused the valve to " slam open". IJote that, in the design shown in Exhibit 1, the opening motion is stopped by the disk striking the dashplate, which can produce tensile impact stresses in the hold-down bolts. Since this failure was believed to be the result of a spurious test incident, and since we understood that the safety function of this valve is to open, it was not considered that this incident was evidence of a substantial safety hazard. Perry personnel reported that they replaced internal parts from this valve with parts from an identical valve procured for their inactive Unit 2.

Dithplate Bolt Pailures P age _, _3, . Seabrook Nuclear Power Plant In April 1991, Edward was informed that one of the four size 18x16x18 CCCVs in Seabrook Unit 1 (Exhibit 2) was stuck partially closed, preventing the achieveme'it of full feedwater flow rate. An Edward engineer and a field service representative were dispatched to Seabrook 6.o assi9t in the diagnosis and correction of the problem. When the cover and piston were removed from the valve in question, the dashplate was found to be cocked in the body bore, jamming the disk piston and preventing opening. Six of the eight hold-down bolts were broken as found; the remaining two were broken during jacking of the dashplate from the valve. Edward personnel initially suspected that the broken bolts had resulted from an impact problem similar to that at Perry, but Seabrook personnel reported no spurious pumping inc. dents. Note also, from Exhibit 2, that the disk stop location ir the Seabrook valves is nominally on both the darhplate and the it.wer locking ring. By design, nominal dimensions of parts are such that the i I bottom surfaces of both parts are line-to-line. Inspection of as-found parts in this specific valve indicated that the dashplate protruded approximately 0.005" below the lower locking ring, and there was evidence of impact on t.ie dashplate. Nevertheless, a 0.005" elongation of the bolts would allow impact load to be transferred to the lower locking plate, relieving loading from the bolts. Since the failed bolts exhibited substantial stretching and necking (Exhibit 4), the evidence did not appear to support impact as the root cause of the problem. The other three feedwater CCCVs at Seabrook Unit I were disassembled during the same inspection period, and Edward representatives reported the following observations: o In one valve, five of eight hold-down bolts were broken. This valve showed evidence that the disk had stopped on the lower locking ring, not the dashplate, in the open position. This added further support to a conclusion that this was not an impact problem, o In one valve, one of eight hold-down bolts was damaged but not broken. Four of the others were slightly loose, indicating overstress. o In one valve, all eight hold-down bolts were both intact and tight. Normal torque was required for bolt removal for disassembly. DISK IMPACT WITH THE DASHPLATE WAS RULED OUT AS A PRIMARY CAUSE FOR TENSILE FAILURE OF HOLD-DOWN BOLTS IN THE SEABROOK VALVES, ALTHOUGH IT MIGHT HAVE HAD A CONTRIBUTORY INFLUENCE. - - - - - - - - - - - - - - - ^ - - - ^ -

Elahplate Pelt Failures Page 4 The CCCV design was next reviewed in detail to determine what other conditions could cause unanticipated tensile loading on the dashplate hold-down bolts. Other ncLential root causes identified were: o Dirt, rust, or other foreign material might jam the close fit between the disk-piston and the dashplate, permitt ing full opening (upward) differential pressure forces on the disk piston to be transferred to the dashplate. THIS WAS RULED OUT BY THE I!1SPECTION OF DISASSEMBLED VALVES. THERE WAS NO UNUSUAL DIRT OR CORROSION IN THE VALVES. o A high upward differential pressure might be induced during valve opening due to the necessity for water to flow through the small clearance between the disk and the dashplate. THIS IS CONSIDERED THE MOST PROBAB Y CAUSE OF THE TENSILE FAILURE 9, BCCAUSE OTHER POSSIBLE CAUSES WERE RULED OUT. As previously noted, the valves were not designed to withstand a reversed (upward) differential pressure across the dashplate. Note: The fact that pene of the Seabrook valves had experienced total fracture of all eight hold-down bolts showed clearly that the loading was brief or "irpalsive"; if a high load had been sustained for any significant period of time, the first fractures would have caused progressive increases in stresses in all remaining bolts, and all eight would have failed. The temporary remedy for the four Seabrook CCCVs was to add eight more identical bolts to the original 8, doubling the tensile strength of the connection of the dashplate to the lower locking ring. While the cause and magnitude of the loading that produced the original failures was unknown, the fact that two of the four valves did not have fractured bolts indicated that the two-to-one margin was adequate for the short term. However, Seabrook personnel requested Edward Valves to evaluate the problem and propose a permanent modification for implementation at the next outage. EVALUATION OF CONDITIONS CAPABLE OF INDUCING UPWARD DIFFERENTIAL PRESSURE ON DASHPLATES: A check valve without an actuator is basically a p_assive device that responus only to forces produced by flow and differential pressure. The force balance on a check valve disk is basically a balance between cravitational forces and flow forces. Exhibit 5 I

l L 5 Dashplate noit railures Pace 5 . with the seat or the full-open stop). It is recognized that local flow disturbances produce non-uniform pressures in some areas (particularly under the disk), but the diagram applies if P, 1 P, and P4 are treated as average values. P,2 3 The force balance on the CCCV disk-pisten assembly illustrated in Exhibit 5 is Forces actino to open Forces resistino oneninc = 2 i (pi/4) (DC) +P3 (pi/4) (D -Dg) P C =P2 (pi/4) (D -DR) C 4 (pi/4) (D + P C I + W sin 450 ... (1) wheres P, etc. are pressures located as shown, Exhibit 5, Ib/in2 1 pi = 3.14159 l DC = Cylinder Diameter, inches ( 14.1 " for Seabrook CCCV) i DR = Disk " Rod" Diameter, inches (4.0" for Seabrook CCCV) i W = Buoyant weight of disk immersed in water, Ib (apptox, 280 lb for Seabrook CCCV) l I Equation (1) may be rewritten ast { (P2 -P ) 3 4 C (4/pi) W sin 450 [ (P1-P) D J/ (D ~U C R ... (2) Equation (2) is a solution for the upward differential pressure across the dashplate. Substitution of the diameters and weight shown for the Seabrook CCCVs results int (P2-P) = 1.09 (Pg-P) 3 4 1.4 ... (3) Sionificantiv, this shows that some system operating condition must produce a high differential pressure across the entire disk-piston assembly to cause a high upward differential pressure l across the dashplate. 1 l m., -m.-_..,_.._.._ -..._.._-.,_,__.._.m., ,...-..m..

J.. l s Dasholate Bolt Palluren Pace 6 , Calculations using some of these dimensions also help to show the 7 order of magnitude of the differential pressure required to cause failure of hold-down bolts. It is obvious that: f 2 2 i -Dg3 jy (pgf4) p SD" (P2-P) (P /4) (DC 3 ...(4) i where: Sg = Stress at minor diameter of bolt threads, Jb/in2 J N = Number of bolts (8 for Seabrook CCCV) l DD = Minor diameter of bolt threads, inches (0.2983") With the Seabrook CCCV dimensions, this reduces to: SB = 256.8 (P2~E) ... (5) 3 The hold-down bolt materigi, A-193 Gr B7, has a minimum yield i strength of 105,000 lb/in. It is easily shown that the conditior. to produce failure (bolt yield) ist (P2-P3).> 105,000/256.8 = 409 psi ... ( 6) Note: Similar calculations using dimensions from other Edward estable check valves and CCCVs with similar dashplate bolting in U. S. nuclear plants reveals the following upward differential pressures as minimum values required to produce bolt yield: Plant Valve Size and Tyne Eg - Py k Perry Size 6 Testable Check Valve 1680 psi Perry Size 12 Testable Check Valve 420 psi Perry Size 20 CCCV 1210 psi 7 Rearrangement of equation (3) shows that this requires a differential pressure across the entire disk / piston assembly of: (P1-P) > (409 +1. 4) /1.09 = 377 psi ... (7 ) 4 .In steady state or near-steady state flow conditions (including normal flow startup conditions) this differential pressure is in the order of the valve pressure drop due to flow; typically, this t _ -. _ _.. _ - -... -. _. -. _ _.. -.. _ _,, _.. -..,.. -. - -. ~. _ _ _... _..,. _. _ _... ~. -. - .__-,._.-e.,.

Qathplat e Bolt Failurer Pace 7 is in the range of 0.5 to 2.0 psi for partially open check valves l (ref Edward Catalog EV-100, pp G 29 - G 45). Thus, to produce a differential pressure across the dashplate that could cause hold-down bolt failure, a system-induced differential pressure surge AcLess the valve in the order of 380 psi or more would have been required. While the analysis above is on, static basis, it can be shown that inertial forces reriet motion and would tend to reduce, not increase, the driving forces which would increase the dashplate differential. In particular, if the valve is filled with water, opening acceleration produces a high value of P4 due to the necessity of accelerating the water column in the equalizer pipe (which connects the chamber above the piston to the outlet of the valve). Since the pipe is much smaller than the piston, the water velocity is much higher than that of the disk piston, and inertial forces require a large transient pressure drop in the equalizei pipe, hence a high value of P4 From equation (3), it is is beneficial. Note howcVer obvious that a high transient P4 that an air or vapor pocket abere the piston would reduce this effect. Edward cannot identify what system condition or combination of conditions produced the hold-down bolt fractures in the Seabrook CCCVs. However, the following are possible sources: o Water hammer surges transmitted through the upstream piping with the valve closed or partially open. o Flow startup with an air or vapor pocket above the piston (an implosive collapse of a vapor pocket could produce a significant differential pressure), o Flow or pressure surge due to stick-slip in a feedwater control component. RECOmiENDED DESIGN CHANGES While the bolt failures at Seabrook are not fully uncerstood, the combination of the experiences at Perry and Seabrook prompted a restudy of the designs by Edward, If the conditions of slam-open impact or large differential pressure surges exist, it is suggested that design changes be made to Edward testable check valves and CCCVs with similar dashplates and hold-down bolting. These are not intended as corrections of design defects, but they will increase margins to assure reliability in the event of system disturbances that were not anticipated. I

Dachplate Erlt Failurer Page 8 Edward recommends the following changes: o FULL-OPEN STOPS ON VALVE COVER Since impact of the disk against the dashplate was the most probable cause of the bolt failure in the size 12 testable check valve at Perry (Exhibit 1), Edward recommends that valve covers be equipped with full-open disk-piston stops that extend from the valve cover as shown in Exhibit 6 (specific design details may vary slightly with individual valve designs). Impact of a piston lift check valve in the fully open position is not a normal design criterion, but it has been observed in a few isolated cases in fossil-fueled power plant applications. Flow startup in unvented systems may cause unexpected " slam-open" action. Other Edward piston lift check valvres use steps that engage the bonnet or cover. While some of the valves now in question have designs like that in Seabrork CCCVs that limit impact stresses in dashplate nold-down bolts, and some designs would not induce any tensile stresses in bolts at all, impact of the disk against either the lower locking ring could loosen the bolt preload. While all of these valves have tab washers to prevent disengagement of bolts, removal of impact loading from this area is desirable. The relocated disk-piston stop will reduce the disk lift very slightly. There will be a very small reduction in the valve flow coefficient (C ), y o INCREASE TOTAL HOLD-DOWN BOLT AREA Increases in hold-down bolt area would not provide a proportionate improvement in solution of an impact problem, because the increased stiffness of the joint would cause some increase in impact stress. However, increases in bolt area will increase the strength of the joint to resist upward differential pressure, and increased bolt area is recommended where necessary to satisfy an upward differential specified by a user. Since the evidence at Seabrook was clear that the loads that produced bolt fracture were impulsive (not sustained), it is considered that minimum bolt yield strength is an adequate design criterion. As noted in the preceding analysis of the Seabrook CCCV bolting, the original design provided a bolt strength at yield equivalent to an upward differential pressure of slightly over 400 psi across the dashplate. The temporary remedy, doubling the number of bolts, increases the strer.gth to an equivalent of over 800 psi. This may be an adequate permanent solution, but this requires evaluation by the user. i

_bishplate Bolt Failures Page___ 2 The Perry and Seabrook incidents are the first cases of hold-down bolt failures reported to Edward on any testable check valve or CCCV; some of these valves were shipped as early as 1980 for plants now in operation and presumably have been in aervice for more than five years. There was nothing in any specification that would have given a design criterion for upward differential pressure on dashplates. Consequently, Edward Valves cannot identify a universal value. Users should evaluate system applications of these valves and advise Edward Valves what opening surge differential conditions can exist in their specific valves. Edward will analy.e specific valves end advise what changes are recommended. Some valves may require no bolting changes. In some cases, this may involve additional hold-down bolts, as employed currently at Seabrook. !!owever, major increases in bolt cross-sectional area may require larger bolts, a wider lower locking ring, and/or dashplate modifications. Submitted by: Approved by: N [- ( E'. A. Bake R. D.14orden ,kesearch Manager Vice President, Engineering j/ Edward Valves, Inc. Edward Valves, Inc. Date: April 23, 1991

i l DAshr>1_at e Bolt Failures P.gon 10. i i LIST OF EX112 BITS j f Drawing, Size 12 Testable Check Valve, Perry WPS l t 2. Drawing, Size 18x16x18 Controlled Closure Check Valvo, f Seabrook 11PS 3. ASME Paper 80-C2/PvP-27 4. Report on QA Examination of Seabrook Dolts ( 5. Free Body Diagram of Dick-Piston Assembly 6. Illustration of Proposed Full Open Stops on Valve Cover i r l l l l I l l

p .-i ._.._t &y'p\\, iih,(d!!!Ei ! M M !$.al..,7.y{d00 ,... n. _,..i,.:r.% )r.l L G h, t 1 <c:.:*jn t;, J:i [, ? ' t., ;,~I . / 'l 5 j)' 9.00 g \\p V k 6 j 'll't ... i t h< 'I. 8 'J lj l*st! h[l t i f s i i, s t A r iq y f('s,,g;i ili:d f!! v'jsh y' b, 'i t iiisM4 i <d'.ipts,s.@f<$r;$'3 r i u E r' ! ; dt- .%D 1 d D" 6

(!Ii !'!

/' 'A dsh y 4!nr / E C;*sicr:rtip l.l;.ill ll' ? !!!T.:3.!. d..y ?i!prN.T:

c. -

iE; s q:I aII;i,.".. p d..ig li5 i %p- !?% .. :':::l! gItrrr.dt !:;'t @[Ch-f., rp a: em 8 see m :. t: "G E6 2 w s. i, } t: ll 1_1 @jI ,,!! E UfM : 5)55sE EE b i 55 :;E nn!!ft 5 $r. 0 4j;I, I E ,i 5.t D'$' l $$HEkta,fj! b !.f M d. t 99

]j< r N

$f 'i 5 d(pt;!::: 8 - -1 gi ) .) s I .- : _ : s t i .ggsp* r:s:se~d = " j '.., !* ;E} : : :*

  • Og g~, ='i-oc r
  • *n. i. yyyyL......(.. t.

U) 9 g U g

j.......

In { t. I r

l;

.I 5 ijr.: yo e t r x y ff g !!) H t.d!i 11 -- %r e j ]ls'15!! i k.45 yi . b(i g -t g 71 i ? I'/ i 2_1.5.91.5 _i.il.lN3 !is.4Ld.iEU agL .< M. ' ~ ~( 4 ~~' " Edf ii qfi' 8 1 r1 : - 4] @ ***** ***ii H I$f3 1l'% EO il '\\. m _ i ~1utJ1~.i. _s

  • ' Ev, T a

l* l$..':n/ a$mg; ~r h,,-n v&r=mp= es cc T' N' 0' s - +.. ' - i a,a l > w-l1 N< A y, u 1 gw:yE p, S y \\ MX 'j e .n Q 5

  • '
  • til gh

< O, -[..!-w ,g. g / [ M (' i d) B' / o / / 9' 3 et 3 !i ia / el o%.ji /-f 7 1l Qp f ;;. k, ( r .E 0 4y,- y t,wMd!:!5'3. ~. '. a! .f y s:}m: p g T=j / , g\\s N \\, s s (f) t-e@ @( N [j dd ND [\\ N w e 2 e

$J H, f 3i

'i'-h J ( G L ye I e g 'm ed '! s r s 5' is la i r q ss : . p .L M.i. o; 1 l

-{a{

-o - I s i'

I;

, $y-j st v w-i gj; 7-r w. : 4G*L g a

  • ,5 Q, s

i 3 2 p; "' '. :.4.,.,pt

'u t

O gam *** 8% M

i> j,. i m ,.' f R #4M[6 w Ga 'I -l u A___.____ 'll1 k+ r N 2 $'\\ ~ - a tl L \\

j d ~C. 2_,_.s^ a A, w O

a h. y ~~' s ({#. s 6 /J Ars t* /' C t n s a 4 (,Ma0~ 7'r g., M*7 - s ' 4 Z r a 4,, y.nj* c h~hgrN1 I r tvL 4 R **a f Jc ~ / v8e7 a d e *~ g,'f.*a/^~[,w.0 a ~ w, - . N .n!4g m ~ n* .*F o T j/ l 6 e t w a r. n I 0 A A m 7: dr n .ew O C a t 4 a m. - tt7 A e a s =f l mO *T, c, T tg A w~ C>X W i. m-e I % e.o,3.T a C ! / '>.. SD 4 s r m O' N 4 t I W& s fv3 PI 3 ."r E */ r G o a ~ n" T w 5 f l lt T r $ W r c >5 I Y ?uG-7m.{l1*1' a *4 a.,a L E e c A s mcuwm. M T .( 2 =:.,E~r> t a BV 5 I x \\li .1 "/, '*' *

  • 2 g?

A L A 4 .d ~ a w 4 m,, h' T M( L C;*2'1Mh. A kaHO

    • a i.

~ O SV C .u. f

  • l),n E

t A e {AI h K eC s p R \\' 8 C e wY E \\}'.,...... '. H \\.. N C 5 N

  1. 5 ww.,

g / /, / /,/ ' " x - mS y d. 0, * .'N, a ,e., j A W [v sys ,u, s \\ s,, I \\ f t /, N ~, y' / ' + s P h /j

h -

v, ~ /,,$,) e,~ L ~e= / /,,, a y t=- e r ' / )f w,, r r w a N ,/ '\\x e e I P i N I n5 ~ y&'J e-w I / o \\ sm. e s 1, s' c e f s e iv a oy.gg, t v c Ta l \\s' N\\ e, y \\N e IV .\\N

o. -

B - y\\ N\\ g I s m-Ja uuMs. k \\ r s' s x \\QN a I g \\ \\ m.<o I c Xe E N N h \\ s\\ \\ C E r , s hwg '", ',p.e e l \\ \\ b e N a m t s s '7 e .e T [' D s, N E ,' e ,U s T E L ' '

  • r s

W y S H N I d' C # t 2 5 s 8 N o C C e vI L tO nbQ o O6 L q* ^ c A lt G 3, I E D r f g t N ls 3G E I l sL-l

  • Ei 2

1 lf e**> e $ y e O c og$ $H g, - j L' Ir

  • c 1\\\\* oMm < ~, s

,{

v s. ,,__ _ _.., y. 3,.,.. d..,3,. 7.y _,h_, h1 i ..no i t. :c..;r.1 ' $ o { <. e

h,
, j.
),{ iN<

g,..i);;.!. J:s < c;.': ? 5 ,o l'. . ':.p::::.. :: ;. y. m s, = : 2.gh 'a. r, y. y: i q; - y yi; y* ,/ , / ,J. i1 w t i i a p r o lS 3' l i s -} l g. w ' I

r t t dt a rst 3,

i.. L' ! g .-~ .,,i, g q-e / )- e II S N,t ". n,. d j z' t i:r r. d l l * ! ? ? ! M.:. u._. p:MI ?!! ?N W h 8 $ '3.11!.Ill 9 3

gIir:r a_u s_: :

.. w:#:.nn... p.. J.g...,_ _ m u w n Qo ,L t5 ~ e e. $l2 l' ': : l t r,!:, :,ir n m , a + .<, v ,+ . /,e : i WW

d;i3HI{-,!:s::-::':g!jEf',s:Ili!if-

=. c. 2-1,*e: i W,,1 99 (ig 5 ;@ -

r,er..::s:.ta.$!I;.'!

2 gi il b h: L F : l I. r g:5 !'(MI::: g ! 'i' iE t 's. -~ l. . <4,. i, : :: 2.. rads:::k ee s 1.* 16 v iv v y

i.,,,

o e .s $

.. <..... L'... %j. 4.

g; + - + ~

v r, l.

a !n L E. P . li ) J ra

l; j

1 5 1 i { 3' I. } i. [5.. [,,, 1,1 r t. i f r.a c, -

3. l s

lY, Y3.w h ( ( ,? ..d. f ,l

  • 1 N

a i) \\,s 5 y m, , _,,a??-M m*W Myni~k W X c0.' ~

n. -

i; it 'n L ? ps ev. Vy, au w, 1 3 I : x y:q.n ! p s\\lQ 1.y. g.J .4 % f e

  • Cdin 3

'/ / /' \\', \\,): ' /.' 7,,l { im., [U. lsi: s at y"' ' / ~ =\\ l) E ~/ / g (g( I dr$}ytI MJ ii [N. f WQd gw ~ i@t "' (E) Q 3 u s il "rE 'N .N 9 2 2 .,jb,; g k' s O' - !Lp,- a g: - s,. g q.l:S e r m ff 5 j ,. dy i, e; i (i}!m i . n,Ll.{.l. r y ,i m-N f, ~

d. j

-w I r wltlr .e L-! 7., g - r g v ..s,. ly'E SE ( . g r=# g.

1 b> 'E l18 F' Id i 3

t o,

n m. y 'r b 6 p-e- ga diL k.d, E!I[; hD 4,! ~ %))3 M ]8 c[a. 1 ,y l ?, o e APEllTURE i ? fi 7.jp Hd n04 e, / e I. n ;!:b tpT yf..

8., y j. A o

calm .,1 a n e oue g 'i J I .t .h Y Me4 gn 4N r l -- q-{t; $ g;p ]l

? F

.-i u, " I E99 gb Ala> Anilable On X < s,w ( ad:: r. ~ d.8 n it ap< ran e nra / ,- ;s.3, %., pd g g i' =

a TE p

e r l{' ': 5 13 _; w m e~,, cn > a h.i i } Mc / V s = e g. l i < <; 4 p

71h ps e, > e L, g e.: - el
gid, s 7.u a

.L .i [D ev7-- -- - l ( #,., ,x o'. '.n'-';.5,I:is.a ~2 e'4 - r ct d e O hJ i ng..t.y, .e .m_x.,. i m , (D f,f

Spo iQ~

^ c a / l

y c

y n. .i e .J.(c ; s' Q. l 1 m n r~ g ,/ ' ' }, l {s f** j>'/,' < >I is. ~,,, ;d,.1 C

/

3 'y f / 4

  • y(f.{

p(.,g ( Q,7 %.-- J n -*--d Il 'yT" CN dI' Wlf. -E g 3 c-y 3 s /Qf y +-- .-.7-[+ lg. ,,,.l-; V {'-. m#' ~. h f '. II 5 . e s 3 s s s x) s ( g.e >gg3 ( i 4 er t N)s N\\NN pi, e- ':gWT 1 4{/; s lW/j Wl},I y> N a / tr p i at s: cd 5; Qi Lh !.;. t II ja.8 / d

a. %ma.,r.y m:o(n wy o

m /l ///,'a s i .I. L, ',)^ l/ d / ij: i .)1 ,/,/ / / .thf }y(9[' 6 'j,e i I ;

  • p m

r 'l / k n,:- u e m n a, ) i '/,/ {,,{"; f o / / y'/ ll

  • ~

i hm a 1 X 1 v> --g ) o s a. e E~ 1_.. - _ w y l 3 4 G n = e t_ _E _ "a o n D G li o \\ W e' in EXilIBIT # 1 Testable Check Valve, Perry !;PS

rr g r

    • a.,

4

  • =*-.ee 9

,6 ./N. s/ - d> {d s N '\\x/ / '\\4} g / l h / r g; [, g' h,i j!' f xp\\ J.mun ali // T' 'N e@ m a cp hp'N N / \\xywi,n /, \\ . e~;; a v.i hm 'o,- i y i 'N 'f t'f ,f ,f3 ,xs$'k /[ Y p. s Q :k /M, % / 5 g kkSh\\[M' -y y w I'

' %)/

g,qq y y ~ f2 y' 'N o, c [sG m .4, ,/'/ xN f .x e s gpN _1 // -3 s. .,p. y - + - - ..n 4 Y/f// l 4 S. t ,8 s 5: "-3k 9 ! hbh"* i l ,1 f i,i t. L_gli_7 y-e 1 9-

_... ~ 'l l n l t- ,,, j ' ' ' ' " = ,. -.. i.. x, n t 3I g= z s u,; i

il l
c ;

AParrtalE c.um r -vc 1, ,o 2-m,M;. .e w i 3

e...

6L

m 1,
y. so :,,awc on ae_.h.,

m o,i I,: e i wea r 5 'y ' 1.']j.Kl { [ t '; D i t; O d.. <a j. .;9 .=.j s

.i

- ['g 4, ;l ; w 8 4 2 4,9 6 n ut-; m.E ,-j, 'II. 1 9 3 N 'x i' ,1 ;ii"E -- e ! :s 1R q s! n

i p n

dp n -i .H N m en w s. x . o i O ,N .v u, w e 1 /, s (, 2, p l 6 'N i 1 N : C B 5-1 / N ,'c fi m ./ eda) } 'i M ,;B _. _ l f o .t a,e .s..x$ r . ga u, i, i y w(-.: e dh .:l, e.!. / i f a"i t i l - Q' o,.,.

; a..
. :...

e.i-p <i e' 1 ast : n n.; + 8 x x d.9 ; 8 . : ;I, 1 l <' m'3 A. a'. s / c 2: a _ 2 i-u c.

== . _./'k M Ijj ^ N f 'M / x i i y2e. f. o (s's m sN / c x x -_i x_v m 1 s .. x ix P s's 'x d, pL. v x r_l. s s \\ \\ .\\ me ** .-en g p - x x v .u. f s

  • 4.*

4 v s. c .) y / ?m o /p/

9'.1

/ 4 /

,L<

0>. \\ . : e. M D>.. v*., i.1 *. e - t3 8!,s 03 ua , --- I..1 N Jy s i .m Q\\o%0\\%~bl:.Fi . =.:.. - ~ - EXilI DI'I ' 2 ValVU contro31cd Closure CheCV-Scabrook I'PS

ppgyW,.. ep...e m..wa 9 y 80.C2/PVP 27 -.~w - e [ h' M<',. d.N A$M.thd.h. h j/ 4, .y .. s '#". d%Q-o pa% 2 .r 6, ,r g T HE At/ERIC AN SOCIETY Or' MECH ANIC AL ENGINEE RS W,j (,/{ 4, yb ',M i / 1$ k %@, ^. 34 E 47 Sta Np W.N1@U j g p~ ?,E n

u. s.x.,,, >,.. w t...w w w. -,~ nw s. s,. v. s., m. v..

v - R 4 r s / i .. mas-om.wwu.w.,aunw mei,- oeer-Sg j m em... t..e e 4,- c r, w, <.m - re, m e a sw y 3.% m a .r 4 _ n, e.m. , n s..,c,,. +. w,w,- r.,r vo.% g .., s 8 \\ 1 W g Q g$ e.e se t e r, u m ~ ' m= t~ - m ~ --s' t &gR

  • hnn 4~ ~ @"3 ~

h op ',; tim /MN': -EXHIBIT # 3 US Mhh c %s g ,f ~. GTs .5e%. A Model for Check Valve /Feedwater $y um m - ~ ne-System Waterhammer Analys.is g ft. g'* R. J. Gradle gud shedm a n.rtuceJlad ater ime h a unanvenalchea sahe can generan da"upnf su'Fe r'nsure Homeser, slon closurc h a controluclewe cked sahe can

g.

Asy: ver m often reduce surge prosure to a solevale lese: Snce a ched sahr a scll actin.c ru ken nrf y e Awe: Vemams ga,isica! r*\\o,lel vla s hed sahr a*1d a nmfMrJ)ccJsater sytem under imebecai , ) 3; mJrutuJe of the surge pressure J<penJs on check vahe'feeJsaier ss stem micractw: vase t nwneenn; sna hnerth m39,,, a y,ng,;jd Ihn model s< uwJ so dmemme the sahe dadret capaaty s l E he+ Corn D,v s on n to reduce the surge freswe to an accerlaNe wlue Data l'vm the r'wdrI a'e nec kxuec We to rumJ tw rp MI@ hsu W N ,,i compartson are docuned h/.:

g a

. i) ,;*9 .;.y IS U M !10N the decirn of such a dashpet is nok a s ug l e t a s. k. M Al t he ag h t he r ec han ic al (.cn f igur at ion e f a da stiet MI Tcr rary years, c enva t ional c%c k val e s have rav bt easy t o de f ine, the string cf the fl.tottling ,(.3 ? y[. een used in tee: water lino tc previde prctectico crificer er flow t es is t ar.c e de vic es tr c btain pir per I ', a[elnSt effects cf reVe!Se flow durinL ncr7al {lant v a ls <: closure t i!tt' inVD}veS the sOlut100 tI COyleX operations inese salveE include tiltint-did, equations which are dependent en the characterittics l 3 swing and piston tyre check salves, and tre cesap M c! the valve installatten. Int e rne t len t'e tweer the to close rar idli be:cre a slyn!!1 cant level cf theck valve and the rest cf the feed.ater systec P '. reverse flew can be established, during valve closure requires the simultaneou [, '" In r uent yeart, attention has focused on the solut i t,n o f t he. equa t ions f or wa ter hete r, valv e f lew ( ) postulated deutle-end.d l i r,e rupture areblen in rate, and c he:k eler.cnt thot ion t e de t e tu te if the n ' icedwater lints in nacicar pcwer plants. Analysis. dashpot is adequately sized. ( indicates that the rapid cichure *v;1 cal of This pat ++ r SWs that an ana l s t i c al trad e l c f a - i conventional ch9ck valves in a line rupture check valve with dast pot can t.e developed in co-situattor can result in large, damagini, surge junction with the waterhaver equations fer the 'Y pressures (waterha v er). Dre rupture tests by feedwater system to predict surgt pressure dt.:a witt .f AS F.A-ATE [ l 3 in IM. en a sirulated feehater reasonatie accuraev, and thus indicate that the y ch ck valve, pressurized dashpet is correctly sired. because of t u ccq lexity i$ system using a tilting-d4' s cold sster, and rapture disks to initiate 'clowdown, of the t ode l equa t t er.s, a cceputer progra: eg ic. iry [' l confirn that sui n prestures ce r e at h date g irig a num rical sclution procedure was written to p rc vide j levels. In these tests, peak pressures of abeat sur ge p r e s su r e ar.d valve behavior da t a. J 4100 psia (250 bar) were reasured fcr an it.ittal Since the intent cf the cetr; ut c r rnodel is to 2 line preswre of 1100 rsia ( M be ). s ue dashpet s for che ck valves t o be installed in j i One way to insure that surge pressures stay f edwat er systems, t he twd e l tn u s t have a b sis fcr j i below damatirF< level 5 1'. to sulstantielly increase qualificatien, }.referably experimertal data ettained g the clo %re time of the check valse [2], in [ 3] for that purpose, te provide t he c.a x ican match-vi {t it is s hWn tP 3t just a slight Ir. crease in valve between mMel and test systeL To this end, a test 7 y't closure time car. actually incrense sur s;e p res sure ; systet was built to develop the experisertal data e an increase in cl s.ure time mere like an order of t,t b e required. C rputer calculated surre pressure h !tagnitude i s r er,uird (e.g. closure tirae of wout dat a ba ned or, the iritial cenditiens defined by tne one seccnd; actual tim depend s en t b c-inttallation), individual test r uns c ould then be compar ed te The add;tler ei a darshpot to a cerventional actual test data te show that the analytical cedel 5 j J. check valve is a feasille meant ef increasing predicted test results with acceptable accuracy. clocure time ty t edac t r4 c lesure speed.

hevrver, The valve used for the model and t he est wa !.

a W ratterr., riston tvre check valve as shavn in I' 5 f mWea h.he b we in.co A W p Do, < r, %,a n e, 93g, ; g$,. valve a s rwrn Mme it W ie h natural snate for the additlen c'i n internal dashpct An the salve neck. The dasnpot is ferred s im; 1 y tv recewed ai mtE Hodme, Wu l' > the insertian of a stationar annular baffle plate. ( me w r,eca e c w % i m is t bM s, i g ^ v h __

h. Fast sw g , M t 0.M t k t 14 l s _N g g, x x s x

,-_ fx N x 1 'NN _'

s om e cdtcr \\ AiV! /" i/ \\ N. N /_ co.,m uituttw. ( - w " 7' / j tussat w . -\\ F15f m 15 M T' N v x, x / / _/ j L'U*..a / L__ / / ~/ h v o. is m.gn s to.; __htun. " new u;. ,[ N*-

  • t w H stat"

.\\\\ X/ J s uu M rm ?* *,V U l{ % Y 4 n'...u' A y [f,M rig. 2 hodel of Teedwater Systre io i / // s, \\ A' // disk. The dashpot acts to reduce the c1c strg speed ns.y / '/'f ,f cf the check cler*ent by nethting umst of the 9.,4' j, g urbalanced closir g f orce throuph pressure b 11d%p } / / ' /,- in the dashpot. Closure t reed t hu s d er+nd s on t he c ), I i '/ / / leakaFe rate c f the liquid fic tr t he dash; e t t hr og h l/ I _,) / // M"N stred clearances at the piston and t he dashpc t plate. , /[Sp z /((/ Description cf instantaneous c he ck element y // / /s *-~ ' [' ' f n e j m a t t er.. a,u r g e pressure and c hoked flow rate c on s / f rom the sitruitaneous soluticr. cf equations gesetning i j [ eh pm. j trie se phenamena, Also, valve cever pressute and '- [' h,', ' ' KP f ') dashpot presrure are calculated from equatter.s I / representing an I ac e t unt ing c.f the mass an thenc ' N <j / taetties, p:ston action, and mass f l o'. ttrough the / ND/ equalizer tube and dashpot clearances. s e A ute ide m ei'* % y Check Elecent M-tion icuations Tig. 1 Model cf Y-Fattern Fisten Check Valve with Internal tmshpct Surn in g ferces on the check element, shown in Tig. 1, and sut s t *. t ut ing in Sewt on's Sec cnd Law ANALYTICAL MODR E I* W dV Descriptien of the analytical model Legins ^d ~ 1^'d d^p' I ^ u cp

  • i k with the feedwater syster, then centinues with the equattuns p~rninE check element moticn. This is where T

= ec e rme u er ne did, followed by the equations fer the flow rate through u the valve and, finally, by the watetha m er equations. { The systet basis fer the anal. is is a check P = pressure in the valve at the ccnnected valve, censtant-p reservoir, and inter-line side, akselute cent.ec ting pipin;,{ essure as shown in Fig. 2. The feed-I dashpet cavity pressure, abs? lute = water line rupture is assuned to occur near the d underseat end of the valve, which is eppcsite to F = valve cover cavity p re ssure, absolute the reservoir end. The valve is considered as a c ~ cavitation-choked nczzle of ever-centracting area Ad

  • underside disk area as it closes. Also ccnsidered are the ins *tia v(di,k OD)'/

effects (wa t e r hastr.e r) in the wa ter c olunn, which = can reduc e the p ressure entering the valve belcw vapor pressure as the fica reserses, and can cause A = cennecti n tube area t surges above the r eservoir pressure as the valve f(tube OD) /4 = completes closure. Frictional effects in the pipe are assumed to exert only riner influences during t'd = urderside di d area valve closure and are negle ct ed. Ad-At = The check valve with the dashoct, depicted in Fig. 1, functions, for an underscat line upperside piston area A = rupture, by generating a closing force on the P check element due to a dif f erence between t..e c(piston OD)*/4 a valve cover pressure and the pressure under the A, p = under s h'e r is t on are a 1 It i t, recognized that t r.is is a slaple and ideal-Ap.At = istic f eedwater syster; the intent cf this paper is F net friction force = to show that a valve model can be developed which gives satisfactory results, not to develop models G gravity force = of complex feedwater systens. Once an adequate 1. = tbeck cler.ent weight valve ::ocel is obtained, it can te rurped with a mc. del of any feedwuter system te estitate surge V = ther.k element velocity pressure data, acceleration of gravity g = 4 time t = 2 1

Chec k t hfer,t lift abovt the body seat (x) 15 piven it is ift per t ant to p e ir,t rs t that th le ss 'l, l4 bvt c oe f f i c i t h t r,. Ec and K,, for the clearances cf the d a s *,p c t connc' te es t in at ed accurat ely u s tr:g i. u a d e t i.;y ( c.n t r ac t ion c r enlar g eu nt data for concentrit ;in s,

  • d x t '

j It~ ire effect cf ary edge thanfers and eccentricity cf i tte clearances is dif ficult te assess, ict this

  • a o

n e t he less caf fichnts e nasutM ushg a b ressures anJ Trrres-test fixture sic-ulating *ne dashpot clearances at 1 var ying degrees c f ec centricit y, and found tc be The tressure-eder-the-disk term. P e r s er~ y, ewhat larger than thest f er sudden are a changes <t tih uy rettvsents tie tite:tise value t t t.e a: tuel in rise. pr e ssure d istr it ut ion ove r the underside area of F itilar ly, the friction facter, f, fer the the dibe f or toeked liquid flow through the valse' dashpot clearances canret be tased cn ripe data for This presbute is deterrined from the empirical t w na m s. Tirst, tM clearaue f h surfa m h relattent nct have randern roughness as in tir es. The surfaces 1, ar e cathined and have r ibt ed sur f ac e roughness (t it s j py )) p u 1 T erpendiculkr to the flow) t ypic al o f turning ( t bcring operations. Second, the clearances are where F1 1s an expirircentally determined function of eccentric t o varying de Frees and this has an effect f'>' disk lift, d isk r esetry, body nat dianter, and on the flew res1st ance ( S) even though the fles area valve internal flow re m try. This function is retains cnnstant. Since trse tycraulie d i a:r e t e r dc e s ' h based on critical gas f lew data and is a ssmed te not c hange with e:rentr ic it), this chu n in flcw h actquately represent condat!cnt undct the disi, resistance must be represented in t r.c friction (' factor. As with the less coefficients, the frictier Felations ei this kind can re ceveleted using a slight 1v rnodified fett of equatio (1) and data f ac t e r was twasured cy e rimentalls using a test y f rom ste ady, c t'e L e d, reserse flo. tests with t hs fixture with c le ar anc e surf aces tut tined to matcr, k' check eletent held stationary, at various did Nn f actbr ing finisne s. % asured friction factors .j lifts, by a lead cell cennected to the ter c f the were much Irwe* than that f or cer:vic ial steel pipe, ,. i I piston. Except (c1 the "p A ' terr, data fer all and decreased at the dettee ci eccentricity incre. sed. ud terms in equatien (1) are teasured including an Anv attent at modeling r.achined clearances, of this -e extra tert ter the 1eW cell ferce (the acceleraticn type should rely on tests to de^elet flew resistance term is retc). The effective disk pressure Tu is factcr. then calculated from equ tion (1). Experimental fne oblique tip end of the e:;ualt rer tube (see .7 l results are given in I",0, the experimental instt cf fig. Il provides an aspiration effect fer procedure is described in [ 4!. nortal flow through the valve at the tube cpening, j which causes the pressure there to be less than the , y% frassure in t he dashpet chamber, Id. and 7 pressure in the valve cc'ver cavity, T are stream static pressure. Een flow reverses due te a c, 'etermined f rom equations relating pressure dro; line break, the reverse effect eccurt causing, flow thr ough the f asbpot clearances and tre effectivelv, a stagnation pressure at the tube 3 equalizer tute, and continuity of mass eq;a t ions crcning. "be pressure at the tube cpening is given fer these ch ebers. Considering the two clearances apptcximatels byt and the equaliser tube, the pressure drcp and flow 3 L ' = lj + '~ E V' (7) velocitita are related bv (see Fig. 11t i g VilVil Id-Ei E IKc+K fp (4) where Vf is the fluid velocity in the valve inlet + e region. vn tys I - I +Ke+ fp, (5) piston velocity and flew velocities in the d

  • s ~p'M t

c clearances nd the equalizer tube are related by the censernrion cf mass law applied to the valve y3;yy (Ec+Ke+ fp (6) cover cavity and the dashpot char,be r. This gives - \\ t F*-F i c ,'E ^ 3 two additional equations: 7 '1 quid density VAp = V,A3 - V,A., (E) where c = p(y /.f entrance flow loss coef f ic ient K-* = YA'p = Y)A1 - V., A., (9) exit flew less c oe f ficient K e flow rath irittien facter p f = where V = pisten velocity, subscripts 1,2,3 re f er e path er clearance lenFth to flow areas and flev velocities asso- ',A L = ciat ed with the dashpc+ caffle plate t tath or clearance hydraulic diameter D = clearance, Tiston clearance, and equa'izer 9 l$%j P * = pressure at eblique tip end of tube, respectively. i equalizer tube, absolute Trictien force T is the product cf the net [ flow path velocity normal force on the check elehnt in reactica to Qq Vj =

1. 2. 3 der.ote s da s hpc t baffic plate bearing en the valve bore and the coefficient of
  • j

j = cleararct., riston clearance, and frictien. This force c or e s free two independent d equalizer tute. res;ectively compenents, one due to flow ferces and the ether Lyh due te gravity cr: %d 6..., obm 3 m .P R b If

l I.3 )' a

  • W

,I t hM e d tm a fles 1 ate threu)h the valvt 5 t r. (lfn whetv w = p A," = c ou 1 valet.t ttatle are e t f valve for t ot f f ic ier.t c! f r ic tior ier the dish / whetV L, a cN bed 11ould fIm, stan i s a funetim bete t.aterial ( r t ina t ier' of disk lttt (ut [ 7] f c t ts11c.! valus). h = e n te r in+r,t all s d e t e rTti r e d function of . t r.r t ant ane ou s 11 ric tre n ure at salve t I p dish lift and valvc internal fios geomtry; it is lastJ on data for li vapor pressure based on initial flute critical gas flew und i s. assuced to be (Cit h ns adequate fer choked liquid flo. See rwrastabilits f ac'er r = [ 4 ] f or v alta s

t. f thtt function and b

descrirtion of the expe r ien t. inside area 0f cennected pipe A c ut e r-: loc k win angle frem positive i x-directitn te trasity 'trter, see

Tig, l.

Tre denetinater tr+1de the ladatal terresents tne fet h ter line ve loc i t v-o f-art ri ac h e f f e t t. The gravity frrce c ore c ne n t in tbv >-Jirection %b 11 N.r t t e thr0ugb t he salvr i s. related tc t he line flN selor.ts (V(.) h the c ent ina t t s i s t he G t e r rL in e pat icn (1) and is ex; t ened ly: equatten, ctt ('

  • V ces

(}}} ~. An (f 1 - riv)/g-- t.,. ,., I Valve Tlow Rate anf Waterherer f 1-( l' Twc of the trimarv concertu in the d e ve l c re er.t f undament al w a t e r hamme r r e l at t or s, t akd ilattic c f t he analytical nxiel' wc r e cetertining tFe chNed, sater eclum. theory ( 21 are: r ever se f im rate of tb-subceoled liquid penerally having a Lign varer t resmr e, tnrough the creck 3,_ i\\f i W valve dur ing c ic su r e, and determining the state cf r the line fluid irrediat ely af t( r line rupture. b;perimental data [ t j 2ndicates that after the 4 c ' {f_ line ru;ture occurs, a tre uure reduction vasc traveh g S throuFh t r1e valv e and up the connected lire at the accultic si.eed of the liquid and, g e rally. caases 21 M m e i m n ! n Aq tte c am W W re s

  • pressure tr.all telew vapcr tressure.

.nitially,

g the fluid takes en a met ar t able- (put(theated) state because of insufficient tire 1er signif1(ant vapor-T instantaneous pire presnure bu%Ie growth.

Thereafter (several rallir.eten:r g in t after decompression wave rassaye), ccmplications set in due to va;ct-butble trout

  • cf nagnituces which Vf= instantaneous fluid velocity vary along the pipe.

This situation, fortunately, a m stic wave speed ef fluid t = lasts only fer a stall fra;' ion et t he valv e clest.r e t ime as valve closure compression waves and re' Tr.c o c ;ua t ions have t he t ene r al selut io n: corTression waves (from the rerervetri cause the pressure te rapidly rise atwve vapor pressure T-T

2) + pI(t +b (13i F (t

= collapsing any tulbles f or r.e d. It in assumed in C t. c. this analysis t hat the fluid remains a superheated liquid durir.g the nr.all time spar, frem rupture until vapor tressure is reached and that, because cf this ,y , L [p h

1) -T 1(t + 1)j y,

(14) ic lI c c snall tire s pa::, the effect of this assutption en the accuracy of the solutlen is likewire snall, wher* I -l e surge s t essure related to initial = Calc ulat ing the c hM ed, rever n t' l cw rate rf liv prusun t (tefore rupture) subcocled water, generally naving a high vapcr e pressure, t hr out h the closing valve is an a;picxt-P rressure disturb anc e (frem initial 3

n. ate procedure chitfly due to retastatilitv effects, line pr e ssure) em nating f rem t t e Tluid tressure at th valve throat can drc; well valve (a function of t - v/c) below vapor trensure before si ptficant varer-t uM le growth occurb.

Tnis allows fcr a corresponding

I tressure distuttance (f rom initial

= line pressure) returning to the valvt inc r e a se i n u.a s s flow rate over that for vapor (a function ef t 4 y/c) pressure occurring in the valve throat. As the difference between vapor pressure and the (lower) Vf initial flow velocity (t+ fore rupture). g throat pressure increases, there is an increased potential for vapor fermtion. Eventually, as this It can be s hown [ 2 ) t ha t when the tbeck vabe pressure dif f e rential grows, vaper will ferr and is cennected to a constant-pressure reservcir, tb+ limit the flow rate. A discussien t e ni t ing f r on. a f} n et ure disturbances are just the negative study of this rnatter is giv-n in [ 7] which also reflections (free the reservoir) of the 11 press.ure recommendh a flow rate M uation that contains a disturbances c uated earlier at the valve. If the r epresen'ative rnetastability f actor to account for pipe length is denoted bv L, then the round trir the incr eue flow r a t e. dut to the metastatility travel time of a pressure disturbancc movinp at effect. Tbe equatlen ig wave Meed c t e. 2L/t, thus: "Ij(I - (15 j (p1 riv) c/r I. c ,Ars

  • l-

,b i 'A1' 4

A]t Ab $ gi I

  • y f, s

Althourn the waterho m t tw ations (13,14) disks instal!cd on the understat end of the valvi, f } afp)y to the entire ; 1 pe l i ne. interest here is The testing has pettorrned with a site 6 valve for twu conf iti J to t he pire cennection at the valve reas.n6, these bein61 1 where y = 0, fer which the e q ua t t eris becomet 1 (M t) IM d vel ed wher egen .p OpA-pe 1 p1 it) ()b) P (1) + valva wa s c h:4en f er t he test as beir.y large .j( = Y $ fo.l.h[l(t) - p 1( t ))J (17) y yfi mises while at the same t in e mit ir,1 r ing the problem of tranagirq t he reaction lead. ' I ((s sc h instantaneous ripe pleuure at valve

' ) lhe particular site t valve used in the test

- / der e F = 3 inlet was the same salve used to develop the ). Vg g instantanecus fluid velocity at valve f unc tional relationships, between valve inlet I ^ } inlet p r e ssur e and (a) the average pressure under t, L)', the disk and, (b) the friction E1de beafiry Sufficient equa r iens have riN been developed to load used in t he food til. Since these functions set up the solutten procedure, but first the init ial ate exactly c orr ec t fer t his valyc, use of 1 c c.nd i t i on s mu s t t e def ined. this valve prevides the best ressible test of f

  • 8 0: * = Ao (initial lift), V= 0 (din t he applic at-ilit y c f t he c hc k ed-p a r f low-At t =

is stationary), p; = 0 (ne reflected pressute waves delived functions to choked subcooled liquid b

21. / c ).
flow, i.e. El c.f equation (3) anJ E2 of h

arrive at t he valve unt il t lo t he c omput er t rnt ar, a solution is ottained equation (10). M ty solving the ateve equat ier s. numeric ally using the } f our t h-ord er kunn -t ut t a roe t hod. Progran results P u t-te the complexity and cost of troviding high gj include the lift, velocity and acc ele rat ieri of the temperature water, which, admittedly, would ,f check element, the.alve cover cavity, dashpot sirmlate actv11 feedwater conditions rare closely. ^' d [@$ chamber, and feedwater line presFures, the Itu s flow these tests were perfrimod with cold hater. t at e thrugh the valvr. and the total mass of fluid Since the test system corresponds closelv to di d arged fro.the valve. the unde riving basis of the analytical model, ( Pteliminary vettiicatten of the cotputet t rogre suc cessf ul co'rparison of the result c obtained f rotn .f wa s ob t ainec in twe sters. Tirst. In the prograr, the test to the predictions of the model is c onsidered equation (l; for check eierent acceleration was to represent qualif icat ion of t he analvt ic al rode l, f forced to have a i n vt. acceleration. The trogram The test provides meat.urement of the reservoir cc :tput ed t he c orr e sp ondig checi elen ent velocity riessure, to check f or a constant value, and aisc [j and displacement, wrich agttai sith the natt provides neasurement cf all other ptessures calculated i soluticn. hcend, input data t' t he t roittam wa s by the model, with the exception of the averate 'g selected such ttst th* valve closur. was, essentially, pressure under the disk, f er comparison t o the Ogood agreement nstantaneous. The resultirg t.u r g e pre *sure was in calculaticns. Check element position during the (3 with the n ac+ result, tran=1ent is also menitored. Due to the unavailarilit) of a suitable f ic.w cicasur ing devic e, the flow velocity QUAllTICATIO" TESTS was not measured. However, total n.ans cf fluid ", g discharged was teasured for comparisen to model An analytical model is considered qualified II"'"' 4,* ] when it can be shown, senerally tnr eug h c eir4 sr 16cn t a e ass pa m g to t est data, that the rnodel p rovides re asonable ineter s measur ed during the t esting of t he internally +] eat ima t es c f the pe r f orausN e of the system it was danhpotted misten check valve provide sufficient datn D-to assure tha' undetlying assumptions made in the ). developed to tepr(sent, ,ubsequent applicatien of t the model to represent actual rhvsical systems development o f t r e tred el a r e reasonably correct, requires a judgerent deciA1cn naJe en a c ase-bv-Additionally, the su cessf ul cotepar isen of the test case basis. ,he qualif ication o f t he model, ano4 c data to the model preutettons demonstrates that the 3-analytical medel prevides reasonable estimates of the iU net its application, is of conce rn here, performance of the system it oss developed to represent. I As discussd earlier, the raidel des:titing the dynamic perfermante c,t the check valve is based on Therefore, the model would be censidered qualif ied for i .jk a constant-pressure r e s ervc i r, the valve, and the gg g ge g, ,g g) b dashpetted check valves in simple ti tng sys t etns with ' J' interrennecting piping. The line rupture in assumed t to be instantaneous enu, te cccur en the underseat f c on s t an t-pre s su re reservoirs. end of the valve. Tre riodel war written t o a c c o t-trwdate subcooled licuid flow, over the range of 7,,, gy,,,, pres sures normaliv er c euntt i ed in feedwater service. In order to provide a test data base for ut.c in 1.ine rupture testsveteperforgedusin t h. test qualifying the model, a test system was constructed facility shovn in lig. 3. The 40 ft (1.1 tr ) watec which corr espond s c lesely to the basis of the u se nc4 r se rved ss t he,,c ons t ant-pres sut e -j analytical model. u sen m,, tank and was russc M at ts to M th j,. The test systern consisted of a constant-pressure 3 ta W an air stcrage field of 600 f t3 (17 m ) A r ese rvoir, a size t, T-t at t ern, internally dasbrotted '"I"CIEY' eneck valve, and the tequired interconoecting pit.ing. Firing was 6 in. ( l'iO enm) schedule 60, with a Pressure was sunlied tc the system frem a tempressed total length f y r'm t he reservoir tank to the valve of air storage f ield with suf f ic ient c ar.a c i t y that the "C "" I " ## i.) reservolt pressure rem ined essentially constant u se n ed r ta alp allowed test runr. with a pipe curing the transient. tariation cf t h'e field pressure "W "t O. imul taneou s rovided variatier, of the initial test tressute. As 6 E Gass 1500 i-pattern pisten type line rupture was provided by rupture 5 74 ' fi l f 4

As the test systen was beir.g f ressur ire d f or a

  • W "M p-/{

) test, t ht? cavity t-etween t he two ruptute disks was [d# , (_ w_m j independently brourbt to an int ermedi at e i t t ssu e by 'i___- en air bottle. This pressure develets tt e recWited m* s a t i t

  • by pressure dif f errt.tlais ac t est t he it.rie t d10 and the cuter disk to prevent pr en.a t ur e t ur t ur t r.t. Tvt ic alb t he disk rupture p r e n t:u r e diffetential is abeut ? ';

l [. percent of the test preu ure. l'isk tressure dif f et ent ials we r e inaintained at at out W percent of ,,ut e ni u n m e y c #1' o i.i h the test pressure. %-g Calibration of the pressure tiansducerb and J' associat ed inst rument a tion was init ially es t ablished ty the tr.anuf ac t ur e r, and chec k ed be f rir e ever y t e at 1v f ollowing the v.anuf ac tur e r's recc tnended proc edut e of applyti.y a knwn DC volt age to sach tr ansducer n, scitst amplifier and enecking the light-team r er orde r deflection. =st s rue p(\\()h Vben the test was ready, tbc strip chart SW-LM snaett r ecerder n,c tor was t urned on and a svitc h t hruvt te m vt gi f remotely open the solenoid valve connected te the ( U t r, __1p,) sk p )" rupture disk cavity. This allowed tie rupture dtO ~ -9 n I~ Y cavity tressure to bleed *f f causit.g the inner disk vtu t tue mtw J te burst, followed by t ursting of the cuter disk. e Dur ins eac h tes t, prest,ures at the ve pcints flg. 3 S t ben.a t i c plagram cf Test Fyster indicated tr, fig. 3 we re it eanut ed, i.e., tressure r e servoir, valv e c over c av i t s, da shpet ch nber, pipe cbs.h valve was used as the test valve. The salve near tbc valve icint, and valve discharge. A.ldi-ses v.odified fer a dannpct as shown in Fig. 1, and tienally, the valve check element resition was f o r.* s t ctu ( n e t s.h e d connected tc the l iston. This recerded. stem prottuded through an 0-ring seal in the valve following the test run, the test system was cover and served as a check element mction indicator isolated from the air storage field and de-pressurited. vnen cannected t o a linear potentiemeter n.aunted to At this tir+, the water volume dischatted f l ott the the top of the valve. test valve during the test wa s the.,sur e d by notinE t h< Line rup ture was simulated using a co ttme rc ially difference in t he wat er tu te r reading af t e r t he wa t et available rupture disk assettiv vhich consisted of level in the reservoir tank was raised back to the two thin n.etal rupture disks held separately by 1cvel switch, flanges. Blowaut are a of the rupts.re disks was larttr than the pipe internal ar ea. KE5plif Transient prensurt s t a sar emerit s a t the points indicated in Fig. 3 were by tie:oelectric pressure pomparisens of measured to calculated suryc transduce rs having high frequency capability to assuta fressure, valve c losing t in es, and quant it le r. o f accurate transient restense, water distr arped f er all of the tests are shovn All pressure transducer signals and t he potent-gt ar tically in Figs. L, 5 and t, and are tabu-iotteter signal were reter ed simultaneously by a Figh lated tu lable 1. Data f or both long and shor t frequency response, light-beat, strip char t recorder. pipe lenttbs are included. kater in the system was treated with a corresten inhibitct te maintain clean 11r. css after initial flushing. g C Test Procedure ?. Friar to each test, the following sequence

u g

,/ G s occurredt The rupture disk assembly was cennec t ed t o the [ underseat end of the teet valve, and the valve epened. I Bleed points in the valve and piring were used to F ** remove any air in the systet.. P.ak e-ur wa t er was pumped through a residential t ype water r.eter and g o into tha bottem o' the reservoir tank until the water level I acbed a level switch mounted just above the 0 I reserv r tank. The valve isclating the water tueter wa s cic aed and the wa t er tneter reading noted. o tow eiet tenta A full-ported pluF valve (see Fig. 3) shown in s mg ,,,,g%,,, the line at the top of the reservoir tank inclates the air storaFe field from the test system u:stil j time for the test. The air **.crage field had been ,/,, 6" w previously pressurized to the desired test p r e s e.u r t. t A bypass (not abovn) around the plug valve allowed the test system tc be brought up to test 1-ressure slow 1v. Once the test r:vster was at test pressure, 71" "I M" und an d Calculat ed the plug valve was cren[J. inis allowed air flow I I I-A U" E*I '5I '" I ' *"" I I

  • f rom the storage field to ruintain the water r ese rvoir pressure nearly constant during the test run.

O

I

  • e N

A I

g - qy= M Wem;, fy,yu pg.a.or

/ l 1 !=

7..

,g { in / r.3 - l. g u, .c g w(., OC 4 '" u 4 i j, { y e y E i. ... i u. du .,j . 1 .. i 'j 0 a .2,

  • o..

r Table 1 Dashpet Check Valve Test Results 5 ) u 0 6ews mi (t=ch reasons for the overtredictien of surge pressure and o sn0tt PIPl 4thGfm fluid loss when long pipe lengths are invQlved, a 4 1 study of the comr ar t s.cn of measured te predic ted g q valve disk position with tirne reveals another reason. . i i o- ~ A typical conparison is shovra i t. Fig. 7 (kun 7 in e c.: e.. ri. e s.e L M k 1). Total time to close the volve is cal- ~ f .igasvegg saat cgosins f =g, sic culated with ve ry good accuracy but the shape cf the [7 curve is verth examining. ,., j Fig. 5 Ceeparison of Measured and Calculated Apparent good agreement is shown betweer. the Valve Closing Time measured and coeputed dish positiens in Tig. 7 However, the vertical separation of these curves is 't CNerall agreement between model and test is significant, especially during the latter portion /.* U' considered good. Dashpot chamber pressure and valve cf closure. During the entire tira of closure, the cover cavity pressure comparisons alse indicate good calculated diss. lift is Freater than the measured agreement but are not shown. value. This means that the calculated effective y valve nestle ares (A ), which varies with disk lif t, j n 1.on e Pipe leng t h is also Isrger and permits larger calculated reverse (. flow velocities. These larger calculated flov [. g SurFe pressure data in Fig. 4 and fluid loss velocities result in prediction of higher surge ( f Aata in Fig. 6 show tnat the model tends to over-tressure and increased fluid discharge from the valve. 'g I )tituate and therefore is conservative. Valve Therefore, it is suspected that improved agreement k osure times shovn in Tig.

  • appear to be cal-tetween the disk position curves vill also result in j

culated with very good accuracy and with no obvious better predictability of the model, j g bias. A sample comparison of measured to calculated j g Eglanation f er the censervatism cf the model line pressure variation at the valve inlet (Pi in 1 7 shown in Figs. 4 and 6 is effered. Although neg-Tig. 3) is presented in Tig. B. This graph, like 2t lecting the effects of pipe friction is among the Tig. 7, is based en data f rom Test kun ? whieb was i initiated at a pressure of 10i0 psia (75.2 bar). W Trie analytical model predicts that the line pressure l increases in steps at the beginning of the transient, g as the pressure waves are reflet'.ed 1.ack and ferth between the valve and the reserveir. The steps in g-the curve are prominent in the beginning because { o o there is little change in flow resistance of the f .no valve initially in the transient, and the system E 1 behaves similar to a constant-pressure reservcir H j ( discharging through a pipe with a fixed flov [ g JD co resistance at the end of the pipe. As the transient

p i Mo o

progresses, the curve becomes smoother because the a continuously increasing flow resistance of the valve

l 3

la becoming dor.inant and causes continuous chanFen in i { ~ / pressure at the valve. Measured line pressure shown in Fig. 6 is traced from a strip chart recording of rav test 5 s data, similar to that shown in Fig. 9, and shows M.3 m otoas PI't ts=m just the essential charatteristics of the rav test onest Pies Lt=ota curve. The lina pressure curve in Fig. 9 shows that ? '!' a clean, initially-steplike curve is net measured. g Reasons for the ragged appearance can be pipe ( { 8 friction, non-instantaneous line rupture simulatten, and mechanical vibrations. These reasons are t atamte 'tu:t nst meaanste La* discussed further below. k O ig. 6 Comparisen of Me6sured and Calculated " E' ""I Yd E** Fluid Mass DischarFed free Valve during Transient (kg = lt,/2.2) s j i 7 ? I

7 w( e e PE St RVOIR Pitt SSUkl ---4<.W w +-. - - - ~ ~ - - - f . 's g. '( Dt5u. Position 'q % Atyt D:SCHARGE w\\ l8 ~ PHt$$URL y q. l ;i.

  1. \\

y ,g' m a am "a Ei \\ DAt e01fPt$5Upt ,M e MtVL COVlh PRIf.5tlRI j# \\ y i 25 (iq ppgggung At a' e rt e iw N. 7 VALVE I \\ - ~ \\,, 6 s e x. I \\ ' Fig. 4 Sart le Strit Chatt hcordinF cf \\ N Inst ant aneous ler.! Data y i7 ~

l s

iL i 5. i <m .t. betseca valve ncrzle area ard disk lift [f -. Im F i g.. 7 L.emr ar tsen of N asure d and Calc ulated the tent salve, nerrie s t e a d e c i t a c r-alh st 11:a ar l* Valve Di n, tw s i t i m. fer lest Fun w i t h disk litt until the dis,e is h i. t a tu vt the 1 :. i s. (see.T a tl e litrr *.>.i. F in.) body teat. At this point nutzle area is trac t tr all; area disk lift relation s +.n r p o l rero; here the no:zle presrure magnitude. a t r u;' t ly. ber a.ae of valve g emetts. and ther T he ahnlv t ic al thodel assuct s a single, ccrtinues with s,tmill slepe 10 foro 44 the d1Ek st u t s t he nu. Kewru flw n b city is 11 While it is true instantanct us f ull line ruptur e. that tbs ruptur e disk s use d in t hese te sts but ut t o rn ina t e d al =.' a t this point of disk 11ft. and aire gemte tra t o din he in tb pemanc e cl a lN e t int.t ant a ne ou s l y (1 or the purpenes of these wwes t a a s t n y' dec eri toulon whic h were reflertec tests), the t ime delav t.etween t he bursttrg of the inntr tnd outer diO 6 wa6 relatively long, typically 'Y I I" no mnpr t-e t n r 30 rat. This delav altwst coincides witn tne pies urt betaast htteng cemrteoston wavem are Fen rated at Se Mn tM o$ t ht M M t ent u wave travel time from the valve to reservett to valve and will undoabtedl" to the source c f,.orat d i t.c r e-flov viincits. I.illattens I tancv. It i s ear.pec t ed t hat thanv et the smaller' high rt exhibit the t c'm nly Onctved pteshure frequency spikes ap pear ing in the n1c atur ed line whict occur after the valse is clemed. 'nac tr t' pressure curve ate due to pressure fluctuatinnn nasund awluude of W m ulatim ese tc indui e d in t he line fluid by the enechanical f riction e f f ects in the rit e is obviouh The vibrations of the pressure boundaries in the test system observed during tha test 6 Tbc piero-e t lects att not included. t r ansduc e r s at e cerg ensated f or electric pressure enechanical vD t at acus ittponed oc the t r anr.duc e r s p g, p and sbould nat produce e xt r aneou6 pressure signals. Imt h t he c alculat ed and ebea surt o line p r essur e Calculatior.s for the tests run s ith t hs 5ft data in!1cate that the peak pressure is reached prier ( 1. 5 TH r i pa length underl'redict the ruasured sur p pressure (!1g. 6), and everpredict t i.- tw a s u r ed t o rotir l et e clonure of the valve as shown in Fig. B. ,r.in is explained by the nature of the relation valve closinr t ir.e (Fif. ,) and the twaeured flutd n.. ins discharge from the valve (Tit. t). Tne overall agreement be t we e n to.od e l ni.d t e s t im net as yo d as with the long pipe length, and is c onside re d to be f due to effects not considered in the analytical rnod e l. l' is net clear what these e f f e c t *6 art, r x c e;. t

  • l

-... n m u: l distuttances a$nociated with the delav fer pressure between t urture disk bursts (fer the shc r t pipe lefip th this was about 10 ens ar.d allcw* fer about i 1 ,/, j five pressure wave r ef lec tier s bet wen disk but h t s). 1 hf e-m m. ne net -'A ^ \\ because the wury presnutch a ksnc i at ed wit t. .l,

ie ;

short pipe hegths are renerslly s.rall, t he non- ^ I \\. Is is _ _. _. ;.,!; Q _. _ connervatism of the surn pr essur e pred t( t ions 7l \\ sea a n -, / 1 rg t c>f p eat c on c e rT1; beweve r, t t st data and model 6g ._._t__..._,.. ([),, / t enp r ov erse n t s are still being i.tudied to identit'v the {p / 3 E Jj Iu ij' s.ource of the discrepantv. if )P y,,. /e m. i 1 lg i l 4, COSCLU.10t:0 d i konult s of the tests verif y the atility c! the I

  • e e.

eo o, c,. , y annivtical rodel presented here to predict con-the servatively, and in rea9enM lv F ood a g r ee rnen t, surge pressure nnd ett er assortated quantitiek Tig. E Cotnparinon o f Measured and Calculated Line tneasured during the line rupture tram ients watt !? { Pressure at Valve Inlet for lest Pun 7 longer pir e length m"re tvpiral of feehnter linet. (see inble 1)(bar = pstall.. 5) 6

MJ 3 a m./el underrt(dicts surge Treasure fer A D.N O.". l DC f.P" ' very sA er t rire lengths; b wver, tr est s u r y t-y' Frgsdres' tend t< te t*.all and nay tiet t+ (f auc! '! ! e auth:rs with tc t h a n e. Mr. L. A. Citrerv an' / c onc e r ti.. The cause fct this be havirr i t, telnr Mr, T. A. De fet their assistante in the c ? r.L ! r t:11 ') i tive 6 t ig a t ted. cf the test faLillts. Mr J. E. hugh ict assistan:t in p he accc unt was later 1 r6 tre ar.alytical Podel runairn the tests, arid Mr. E. A. lave fcr ;tttir,ent (,Jr t he ad,su s t r en t of valve r ar ame t er s sutn at tutresticns concerning the de sign c f tN test ta:111ts, e nonle area to a test-n tatlisned value, er the use c.f etasured f l u i c' r r e; e r t ie s su:h as tressure van

  • j i T' ri T TN ;U' rpeed frcm the line ru;rure terts, to 1F;teve data

~" terrariscns. Indicatio

  • att I na t sa:t use weald artrove the a gr eer ent.

havevct, this a t t e ru t l. En erimenta! !*vestigativo cf Iressurt r e ns ie nt, y Created by a Clesing. Check \\ a l ve,,, kepert he. is net pencrally avallarle when ccnsidering a valvt g for an insta,laticn. (.se c f c a t alc g da t a an., rii 73-473, 1973, ASEA-ATOM. Weden. Iht reference rock data should Trevide saf f icient r ad e l l a ro n iat., J., aterh m t Analyris, D,. v e r accuracy. Tatli:ations, 1*3 .ests descrited in this ;.ger verifv the 4 i abilit y c f slow-clesing check valves tc' redxe 'A. for,, 1. F. , hur l( at C a r.t a l m.c n t cf Icstulated surge prest.ures in feecceter syste 5 u ~, e r g e. I n g e-reedvater Lar.(tre n., Ferert L. \\.-ve;. . E - J. 3 line ruptu,re ll ev a :.T. s 4he testa cited earller bv ASEA-A4tw. 'I were W r4 or a test rig e,. sin 11ar T1ov Line, 3rc. Quar.er, i.,y, Frenell Inter- ,4 s n a t t e r.a,6, l' i t t td a t E r. Ia cenStrLCtion At.d t rNu.. / a s u r p tc Trebsure cf 3., i. tsi (2M tar) fee a:. Initial prissure cf IIK isla L 8' %/{ e,, I. E. M.c le ar F a in d t t r isc,at4cn ( 7.; bari a n,. a size e t i l t in g-d i s e crece valvt. i \\a.4ves Tercrt \\.-Fet ,,c-), l_ i c s ,t ir.c, for the s a r.t. 1 r.i t 14 4 pressure, tests en tnt chece , w. 7 t. Ec ; kvt a,l s g i.y Jrc Ouarter, , r r e r n a t i : ".a,., valvc with dar.hr frcda e_, a surge pressare cf a i I1!tbburfh, l' a. g alm u t 700 psi (-F tari. Tut t be r re:: stier in turge , '5. 4 pressure can t.e cttainea inrreasing the c a s r.t : t ., a c, N. a m,. .' evan, s. F., .,.t t o u t t - T, c k in 4 t. a flev resistanet. e tc centri: and E: centri: An n u,.1 cf f i n t: 3 Fecause o, tne generally y ; c,, arreement Mtween s_4 e a r a n : e s W i t h am,., i t h ou t Eelatise ' tice e t edel and t e st sh vn tere, the a n a l v t i c a a, rode, is . the k un d a r i e s., Trans. ASPE, N:ved er, IMS, i,. 4 (? censiderea qualifle, fer predicting the perferrance 7 3 g g 3 y' * ('N. .M of the t yp e c f ched valvt witn dastrot used in the .M tes ts and f er a

a. i tr i c rl;1ng Ostet with a constant-f.

Levards. A. F. a n, Ma t t e r,,.. J.,,, S c r e i n. S t u :. l e s 4 pressure reurvalt. helated tc the Less-c:.-Co:lant A:cident,,' Tc;1c a' A.., lse cf the analytical tr:de1 fcr the type of Meeting en Kater,keacter Safety, stnf. <hh,, - y valve use in these tests i r..cnjunt'ac. wit' water- ',..'i s AE u,,, Marc * .t .:, 19, a. Salt, ak e C. i t y, t a *. o d ime r eq.a tic ni fce ! cre en;;(x feed. cater syste s [ Muld pic' vide reasrnatie a::uracy frevided the T.;cv sr '), ' N / alalati - D n.,. r r c *, F oc he,3 , n t e r!.a t i c na,, M s c ption cf rc vapr f cr Itacn anr ttre in thc c .,y i i t. T w.w Contrcl,ivisic, Fittsburtt, fa. 19,.,. d oma i n c f the t o t a., n :'d e., is valid. As with th( use ;f any analytical eccel, prelirir.ary study and y[ t judgement Are require-d tc evoluate tre validits cf tne model a;;11:a t ic. i %h 1.$ r,- ' O,, r g $k< a l Mf, m l I 4 i as ) NN. l

  • $*3.

1 P' 1 "h i '1 l l 4'& l I w' w,. '5 s t us., 1(b 0% ' ;v; --%o. i \\ %g$m l .v.G I Y- !M

EXilIBIT #4 ' EDWARD VALVES,1NC INTERNAL LislTER April 22,1991 FROM: II.E. Carothers To: Earl Dake

Subject:

Failed Capscrews Returned from Seabrook 'Ihe failed portions of the two capscrews returned from Scabrook were marked on the head with the grade identifier "I37" and "Til", indicating that they were mawfactured by Texas Bolt. 'Ihey were further identified by the manufacturer's Code Number *F85". These were the material specified and subsequently supplied for the subject components. Both returned capscrews failed in the third thread about 11/8" from the head. Ilased on the failure being in the threaded area root and the observed locallred necking, the mode of failure was obviously overloading in tension. The hardness was checked and found to be llRC 28.5, which is typical of B7 bolts. I have retained the parts for further evaluation, as deemed necessary. n u [ d h l D. E. Carot iers, Manager Metallurgical Process Control cc: R. A. Bandukwala l-l l

De n/\\e\\ '\\ \\ \\ \\ l \\ \\ \\ / \\ Da t \\ \\ \\ r w \\ \\ P1 45 \\ N N NOR'M AL TL O W Exhibit //5 Free Body Diagram of Disk-Piston Assembly f

I"L l l L I \\ l i i i i I l ffkh Nf ~' ~ l \\ ( x l %3/ (( Tis,f t-?* ?% r N N i N s N N N N N I N I PISTON IN l OPEN POSITION 1 l l l l Exhibit

  1. 6 Proposed F ull-O p e n Stops on Valve Cover ES11}}
Retrieved from "https://kanterella.com/w/index.php?title=ML20081F609&oldid=4376072"