ML20056C924

GGNS GL 89-10 MOV Program

ML20056C924
Person / Time
Site:	Grand Gulf
Issue date:	06/30/1993
From:	ENTERGY OPERATIONS, INC.
To:
Shared Package
ML19310D630	List: ML20056C923 ML20056C924
References
GL-89-10, PROC-930630, NUDOCS 9307270155
Download: ML20056C924 (46)

Text

. - - - - . -. .. . . _ . .-= . ..

1 l  !

1  ;

)

t i

.l 1

l l

l l l 1

l l

Attachment to GNRO-93/00078 Grand Gulf Nuclear Station (GGNS)

Generic Letter (GL) 89-10 Motor-Operated Valve (MOV) Program June 1993 j

l l

i 1

l 1 I

l 9307270155 930723 l PDR ADOCK 05000416 {

P PDR ,_ l

~ , , . - _ . . .

___,,y,,- .,-m -m,,. _..,,-,y, y,..9,.yrn,y. ,.my. y-yw, .y- % ,

p.

f Table of Contents  ;

i 1.0 Introduction i

2.0 MOV Program Description t

4 3.0 Completion of Program Development {

4.0 Conclusion I i

5.0 References ,

i Appendix A Flow Test Program Details  ;

-j Appendix B Retest Program Details l Appendix C The Role of In-Situ Flow Tests in Periodic i Verification of MOV Capability ,

i l

t f

i e

i l

1 i

i l

l

)

I

.. I i

1 ;

1 1.0

Introduction:

1 Grand Gulf Nuclear Station began the implementation of a  !

ccmprehensive motor operated valve program in October 1989 i in response to Generic Letter 89-10. Project SMART  !

(Systematic Motor Actuator Reliability Testing) was  !

initiated to ensure both short and long term MOV concerns were properly addressed in the most effective manner. The program developed from experience gained while addressing 153 85-03 and has continued to evolve. ,

A significant number of problems have been identified and -

corrected including incomplete / inaccurate design basis,  ;

marginal switch setting control, overthrust /underthrust conditions, damaged, worn, or missing components, undersized actuators, improper valve orientation, and occasional grease i degradation. It is clear that the additional attention given MOVs under this program has been justified. l I

Careful consideration of the many technical issues has enabled GGNS to clearly define the boundaries for an i effective MOV program. When fully implemented, this '

program, together with other plant programs, will meet the intent of GL 89-10 by eliminating the great majority of MOV problems found during the development phase.

Having addressed the major technical problems, it is  ;

expected that final resolution of developing technical issues will result in few, if any, additional MOV hardware ,

changes.  ;

I This document gives an overview of the GGNS plan and l schedule for completing MOV program development. Additional  !

detail is given where GGNS has elected to pursue alternative  !

actions to those suggested in the generic letter. Within 30 days after program development is complete, GGNS will ,

submit closure notification as requested in GL 89-10,

recommended action (m).

i I

i h

i i

i

e 2

2.0 MOV Program Description The basic building blocks of the GGNS MOV program are illustrated in Figure 1. Each program component is briefly i described below.

2.1 Design Specification The design specification is the central control mechanism of the MOV program. It establishes the operating conditions and the switch setting and sizing criteria needed to perform flow tests, static tests, and design changes. The specification reflects the ,

results of design basis reviews performed for all program MOVs.

The specification may require revision if a baseline static test and inspection reveal hardware ,

discrepancies. Revisions may also be required if flow '

test data were to show that sizing calculations do not envelope design basis conditions.  ;

2.2 Static Testing l Control switch settings are established during initial testing with diagnostic equipment while stroking the '

, valve under static conditions. Prior to testing, most actuators are refurbished.

2 Refurbishment includes a complete teardown, cleaning, and inspection of internal actuator components, including replacement of lubricants / components as necessary. Springpacks are checked for relaxation and force vs displacement curves are developed prior to reassembly.

Stem thrust and springpack displacement are monitored I during gate and globe valve testing so that both thrust i and torque measurements are obtained. In cases where considerable margin is available, direct stem thrust measurements may be omitted. .

Switch setting windows are established in accordance with plant procedures using calculated thrust / torque '

from the design specification. The window is .

established with consideration for equipment accuracy, l possible dynamic effects, and degradation.  ;

r r

6

1 6

MOV Program Building Blocks

,A;x;y x A Static x ,

Testing JL l

i 1

s_ s s2 e xx xx ;

east 5j: Design Design  ;

g spec en noes A

y -

/x Flow Testng [

2 Figure 1 l

l l

l l

l s

4 1 i

2.3 Design. Changes Design changes required during program evolution fall into two categories; those required prior to flow testing and those required after flow testing.

2.3.1 Untested Valves j In some cases design calculations have identified  :

MOVs that lack desirable margin. Both traditional  !

and more recent sizing assumptions support this conclusion. Modifications for these valves may be planned and implemented before static or flow test data is available for the valve. Operability determinations are made on a case by case basis.

2.3.2 Tested Valves Calculations for some valves show their theoretical capability to be acceptable based on traditional sizing methods but not acceptable based on recent methods. MOVs that fall in this capability range are considered operable.

However, if flow data is obtained that clearly demonstrates a need for additional capability, the valve is considered inoperable unless an interim justification for operability can be developed.

Static test data may also provide information which suggests a need to perform MOV modifications. This may happen when switch setting windows are very small.

s 2.4 Flow Testing The GGNS flow testing process is shown in Figure 2. It is a five step process which begins with an MOV' performance analysis.

Siemens Nuclear Power Services was retained to perform y

steps 1 and 2 below because of their experience on similar MOV projects in Germany.

l 1

i

i Flow Testing Process

/ s \

'Dimensionsb / rientaton,4 O

l MOV Performance Matees' " 4 Piping System Analysis Operatng Configuration Condtons/

b \ /

Y Form Valve Families and identify

• for 258 Required Tests Movs I

@ Y

/ Review Ali\ rs3 U*#'

Perform Required Flow m Tests and Select Tests

~ ~

25 Possible Supplemental Flow fjg,8 s Tests - Testing N / l

/ @ Y

/ Adjust N N ,

separate Apply Test Data Effects > e Flow Data to Untested MOVs Data as Needed N j \ / 1

@ y Specification and >

Hardware Adjustment as Required Figure 2

E [

P 2.4.1 MOV Performance Analysis  ;

1 The performance analysis evaluates the travel of the valve disk over a complete stroke, calculating ,

critical stresses, deflections, and disk stability .

along the way. With this information, potentia' I problems during a stroke under design basis  ;

conditions can be identified. The performance -

analysis considers valve internal dimensions and geometry, materials, operating conditions, valve orientation, piping configuration, and adjacent valves, elbows, tees, free pipe length, etc. ,

2.4.2 Selection of Valve Families and Test ,

Candidates ,

1 Once valve performance has been evaluated, families of valves that will exhibit similar behavior can be identified. Test candidates in l each family are selected with the intent that the tests will envelope expected conditions for other valves in the family. Figure 3 shows that a minimum of 31 tests are recommended to cover all program MOVs. Additional testing is planned in -

some cases where useful data is obtainable and the 1 test is practicable.

]

)

2.4.3 Perform Flow Testing Flow testing is performed at design basis conditions where possible. Tests results obtained at less severe conditions are extrapolated to design basis conditions when evaluating the data.

In cases where no test is possible, offsite testing may be performed unless engineering justification for not performing a test can be developed.

I l

l 1

1 I

1 I

_____a

.- GGNS FLOW TESTING

SUMMARY

! i

1. OF i

1. OF GGNS i % OF FAMILY PRESS l VALVES INi # OF TESis j TESTS ' PLANNED TO FAMILY I SIZE MFG. CLASS ' FAMILY REQUIRED l PLANNED BE TESTED GL1 3 12 WP ' 150 300 ' 8 2  : 2 25% >

GL2 14 18 WP > 300 i 4 1 i 4 100%

GL3 i 18 WP 300 2 1 2 . 100%

GL4 ! 4-6 ' WP ! 900 3 1 , 1 1 33%

!  ; 50%

GLS  :

12 WP 900 2 1 1 GL6 ! 10 12 i AD i 900 4 i 1  ! 2 i 50%

I j 13%

GL7 I 3/42 ) Y i ,

I 1500 84 l 2  ! 11 I i GLS 3l4-2I !Y l 1500 l i

12 2 1 1

1 8% -

!  !  ! i i l

NXL1 8-24 HP ! 150 l 28 3 I 11 i 39%

NXL2 ' 30 . HP 150 4 ,

1 0 t 0%

NSR '

10  : HP 150 i 4 1 1 1* I 25%

GA1 4 ' WP 300 1 5 i 2 i 5 { 100%

! 4 31%  ;

GA2 18 24 WP 300 , 13 1  !

GA1 '

2%-8 WP ! 150 t 40 3 I 2 i 5% l GA2 10 24 ' WP ! 150 i 5 l 2 i 0 1 0% 1 GA1 ' 4 28l3-24 I WP ' S00/900 21114 '

4 1 4" l 11%

GA1 4 AD i 900 1 . 1 i 1 i 100%

GA2 12 ' AD ! 655 <

1 1  ! 1 6 100%

GA3 l 18-24 i AD i 150 i 2 1  ! O t 0%

) i I  ! 1

! i  !

257 -

31 i 53 4 21%

, i 1 4 i

• INEL Test NUREGICR 4048!  ! i

" One GGNS Wyle Test. One INEL Blowdewn Test i

i FIGURE 3

i 8

The total number of tests expected to be performed l l is about 53. This number does not constitute a  !

~

GGNS commitment and may vary somewhat as the test -

program progresses. The minimum number of l required tests, the number of tests actually planned, and the populations of each valve family ,

are plotted in Figures 4 and 5.

The performance of each valve tested is evaluated I

, in a timely manner in accordance with plant  ;

procedures.

i J

i 2.4.4 Apply Test Data to Untested MOVs

' After all flow tests in a family are eterieted, determination of valve factors and recci';ed  ;

thrust / torque can be made for the unte ated valves. <

This process is currently under development.

l a

i 2.4.5 Specification & Hardware Adjustment as  ;

Required 1

l .

, The last step is to revise the specification where '

empirical data indicates it is required. Actuator capabilities and field settings will be reviewed ,

and any hardware adjustments necessary will be j made at the earliest available opportunity.  ;

1 Additional details regarding the above GGNS MOV flow testing program are included in Appendix A.

i 4

)

J T

d

. . _ . . _ _ , , . , , _ ._ ,,_.m _ , , __

GGNS Flow Testing Summary Fam.ly i i GL13-12 WP 150-300 luungsinummmE$$$$$2hM&j22]ZZ$pjg]

Smaller Families, GL214-18 WP 300 6 GL318 WP 300 M GL4 4-6 WP 900 E@FM GLS 12 WP 900 M@

GL610-12 AD 900 MM GL8 3/4-2 Y 1500 M g pf M VR3$$$$$$$$146$$'d)$@$f M@l NXL2 30 HP 150 MM NSR 10 HP 150 EMM GA14 WP 300 GA218-24 WP 300 ' ' ~ )JJJJJJJJJfff$$266$$$&K$9&hWiMMMJ?!MS!]

GA210-24 WP 150 GA14 AD 900 GA212 AD 655 GA318-24 AD 150 i

M I . I i l i l i I . I i 0 2 4 6 8 10 12 14 Number of Valves FT Min Req'd Test Population @) No. of Valves in Family E GGNS Tests Planned Figure 4 .

t i GGNS Flow Testing Summary Larger Families Family GL7 3/4-2 Y 1500 ENF*SM M46&X4X R N 6M4M45M6 M 231 NXL18-24 HP 150 _ G ' MXMGi GA121/2 - 8 WP 150 ENXNWM GA14-28/3-24 WP 6/9 ER956$6$&M65Sj i i 1 1 0 20 40 60 80 100 Number of Valves

" ' Min Req'd Test Population x. No Valves in Family GGNS Tests Planned Figure 5 t

?

i 11  !

2.5 Retest l

2.5.1 Retest Periods  ;

Retest periods are summarized in Table 1. The  ;

basis for these periods is a combination of maintenance experience and the valve's relative risk significance.

The GGNS IPE was utilized to determine the ,

relative risk significance of each MOV. Using the Fussell-Vesely importance measure, four priority .

groups were established. (See Table 1) I Retest periods were selected to be consistent with relative importance and are bounded by GGNS maintenance experience.

• These periods may be increased or decreased in cases where service conditions warrant this adjustment. ,

Additional detail on the basis for retest periods  ;

i is given in Appendix B.

2.5.2 Retest Method Static testing is adequate for routine retesting  ;

since: ,

a. static settings are established to cover l cbserved dynamic effects, equipment l repeatability, and degradation,

b. the MOV performance analysis does not  ;

predict severe, progressive degradation  :

of seats and guides, and,.

c. positive impact on core damage frequency l is likely outweighed by additional safety system down time and potentially abnormal system configurations.

Additional detail on the basis for static rctesting is given in Appendix C.

• Actuator refurbishment is normally a prerequisite for initial baseline static testing. After an average of 15 years in operation, worn components and degraded grease have

- been found only in a few cases.

Priority Frequency No. MOVs Description / Basis 1 3 RFO 6 o 99% CDF 2 4 RFO 26 . 99% - 99.99% CDF l 3 6 RFO 92 . Remaining O PRA MOVs

. Auto Ctmt Isolation MOVs 4 GGNS 135 . Remaining Safety Discretion Related MOVs 1

Note: Retest Frequency May Change With Experience

-includes 1 Non-Q MOV Table 1 B

I I

o i

-~

13 2.6 Schedule i The project schedule is shown in Figure 6. Initial  !

testing will be complete for all program MOVs by the end of RFO7 in June 1995, consistent with original GGNS ,

commitments approved in Reference 1. Any modifications identified during the evaluation of flow test data will be implemented at the earliest possible date according to valve importance ranking and the amount of design margin <

available. j Testing of valves in the Main Steam Isolation Valve -

Leakage Control System has been halted pending the results of a study investigating the feasibility of converting this to a passive system. Static baseline l testing will be complete by RF08 for any program MOVs  ;

that retain an active safety function in the final system configuration. The current flow test plan does not require flow testing any valves in this system.

3.0 Completion of MOV Program Development i l

3.1 overview ,

l The purpose of this section is to show that when '

implementation of the GGNS GL 89-10 MOV program described j in Section 2.0 is complete, the basic goal of the generic letter will have been achieved, despite the existence of unresolved technical issues. j 3.2 Unstable Program Acceptance Criteria GL 89-10 has fostered a broad expansion of industry knowledge regarding the design and raintenance of MOVs.

Implementation of our MOV program has led to the identification and correction of various design and maintenance problers. This, together with stronger programmatic controls, has resulted in significant improvement in the operational readiness of MOVs.

Industry knowledge and regulation have traveled parallel paths to their current levels. Ideally, a complete knowledge about the source of MOV problems would have ,

preceded the regulation. However, the problems were l critical enough to warrant immediate action. As a result, the entire utility industry is approaching completion deadlines for GL 89-10 while new information  ;

f rom EPRI, INEL, and other sources continues to surface.  !

i

- - - - . _ - _ _ _ _ _ _ . ._ _ - - - _ ~ _

GGNS MOV Program Summary Schedule Jan 90 91 92 93 94 95 96 97 98 RF'O4I RFOS RF'06 RF'O7 RF'O8 I)esign llases lieriews/ Spec Deveh >pment g __ . _ _ . _ . _ _, . 3 Statec lestuig 5 -

'M indial MOV Mods E -

?

Iinal Groutonq Analysis _ p ;ow 1est f low T est l'roca(hues I kiw Testwig g ._

..g. _ _ _ _ =

"'"" Alipty Ilow Data to Jrdested MOVs 9- .. _. . . . _ . g l'evioikc flote ts Figure 6

15 ,

Considerable action has been taken in good faith by the

' industry to solve MOV problems. However, the complete  ;

solution to several relevant questions regarding MOV i sizing equations has yet to be established. This i situation places the industry in the position of having i to achieve the specific GL 89-10 MOV program goals .

a without stable regulatory acceptance criteria. ,

l The lack of stable regulatory acceptance criteria has ,

j resulted in multiple program restarts in design and testing. This leads to excessive resource expenditures c in both dollars and dose. Without such criteria, the l 1

establishment of MOV programs can easily become a ,

l protracted effort. ,

t

GGNS has been a leader in the analysis of MOVs and the j development of justification to support MOV families.  ;

We have closely followed and supported EPRI research  !

efforts directed at understanding MOV problems. We believe that enough is currently known about MOVs to  ;

permit the establishment of reasonable and stable ,

acceptance criteria that meets the intent of GL 89-10.

.i For GGNS, this document constitutes acceptance and

_tosure criteria for GL 89-10.

,  ?

i 3.3 Putting Emerging Technical Issues in Perspective .

i Some emerging technical issues may have the potential to adversely affect MOV operations if completely j ignored. However, reasonable steps can be taken now ,

that will mitigate the impact of these issues, once they are finally resolved. This approach permits the MOV program to achieve significant early progress in l

these areas while minimizing resource losses due to overly conservative assumptions.

9 1

Figure 7 shows that the issues in question are all related to the thrust specification. Major problems

i with design basis information, control of switch (

settings, and actuator capability have been resolved.  !

Questions about rate-of-loading, aging, stem factor, j disk factor, and temperature effects are still ,

emerging. However, if reasonable steps are taken while  ;

3 developing the program these factors will have little 1 or no impact on program success.

1 I

l l

l Putting the Unsettled Technical Issues in Perspective

,[

[ GL 89-10 Problems 13%

7jfh E Improper Bypass Time 3% ,',f C High Motor Current

! @ Unbalanced Torque Sw i

\

\ %e

/

,./ 13%

3 Valve Backseatng C incorrect Thrust 7'kg[ *] Torque Sw Problems N U Misc Problems

'N d Incorrect Thrust g Thrust 39 %

- Spec y%/

yj e

~

ch Design

\ Settng f Bas:s \

//

({f "****{

j,/

Thrust Spec Q f 13%

i

{ Structural f t m ts

[!D*9 rod ** veta

• Va!ve Factor -y Bottom Line

. FStem Factor GGNS MOV Program Will

\

\ Temp Eliminate The Vast

\ hs Majority of MOV

• \/ ' I Problems

,

• Conservative Trends Figure 7 l l

i 17 [

1 3.4 Conservative Treatment of Emerging Technical Issues ,

3.4.1 MOV Sizing Equations  !

Preliminary results from GGNS flow testing, EPRI Performance Prediction Program testing, and INEL testing all suggest that use of typical globe and butterfly valve sizing equations have normally j resulted in an adequately sized MOV. The contrary t is sometimes true with respect to gate valves. l Industry gate valve sizing equations traditionally used a disk factor of 0.3 to relate stem force to [

horizontal force on the disk. Early in the GGNS  :

l MOV program, the project team felt this assumption ,

I may be unconservative in some cases. As a result,  ;

l program MOVs have been reassessed with a 0.5 disk  ;

factor.

l As expected, many MOVs are already sized and powered to meet these requirements. For the minority of valves not so sized or powered, we have concluded that an interim minimum disk factor i of 0.3 is acceptable to support operability. In all cases, flow testing results are expected to i provide the basis needed to select an appropriate l valve factor. '

i Therefore, using the MOV performance analysis, the valve family concept, and type test data, GGNS will determine what disk factors should be applied in each gate valve family. Once determined, any deficient as-left settings will be identified and corrected or modifications performed where required.

3.4.2 Stem Factor Degradation GGNS does not utilize stem factor in sizing equations because both torque and thrust are determined during testing. Its only use is in calculating VOTES correction factors and determining Limitorque torque switch repeatability.

j l

I 1

18 Mcbilux EP1 is used for actuator gear box and stem lubrication. This grease has an excellent performance record, maintaining consistent lubrication qualities over time and multiple strokes. GGNS stem lubrication inspections are scheduled every 18 months. If the inspection shows evidence of degradation or contaminants the stem is relubricated.

During separate effects testing conducted under the EPRI Performance Prediction Program, Mobilux EP1 was subjected to 500 consecutive strokes under 10 kips stem load. The stem friction coefficient maintained a flat profile, remaining between 0.14 and 0.13. See Figure 8. Figure 9 shows a summary of EPRI separate effects test results for a number of common lubricants. The friction and wear properties for Mobilux surpassed those of nearly all other lubricants tested. In addition, it experienced one of the lowest changes in coefficient of friction.

Recently GGNS performed flow testing on a 14" William Powell gate valve on Wyle Lab's simulated pumped flow blowdown facility in Huntsville, Alabama. (Reference 2) The valve was refurbished prior to the test. Stem relubrication was performed only once about halfway through static preconditioning strokes.

The valve was subject to 50 static strokes, 6 pressure lock strokes, and 79 strokes under 100 to 600 psi DP with flows exceeding 9000 gpm. Stem friction factor was calculated for each flow test.

As can be seen from Figures 10a - 10d, stem friction factor maintains a flat profile, remaining between 0.11 and 0.15 in the close direction and between 0.07 and 0.10 in the open direction except for two anomalous excursions.

f I

i 19 ,

4 I

1 1

Mobilux EP1 has demonstrated very stable

'ubrication properties, even af ter numerous strokes l l and some aging. These properties., together with .l lubrication frequency make the the GGNS stem current stem factor assumptions conservative.

3.4.3 Rate-of-Loading Factor Rate-of-Loading is defined as the difference i between thrust at torque switch trip for static vs j flow test conditions. This effect has been I observed in varying degrees during many flow tests conducted throughout the industry. Although  ;

several theories exist, the mechanisms that cause a i reduction in thrust at torque switch trip during I flow tasting are not completely understood. l

}

l The inconsistency of this phenomenon can be illustrated by observing the trend and magnitudes 2

of ROL effects observed during the GGNS Wyle Labs test series conducted on a 14" William Powell gate i valve. ROL is bounded by 5% but shows no real trend with load (DP). (See Figure 11) In at least 11 of the 28 closing strokes, the effect was smaller than the 3% measurement error.

GGNS believes industry research will eventually

show ROL to be a function of multiple parameters.

Since it is not clear how industry ROL effects should be applied to GGNS MOVs, the best current approach is to base ROL margin on GGNS flow testing.

1 i

, 3.4.4 Emergent Issues Current industry research will undoubtedly generate new issues with regard to MOV operation. For

example,.Limitorque recently announced the results of motor testing that shows AC motor torque output

to be adversely af fected by rising temperatures.

l This issue may require yet another special review i of program MOVs to determine its impact.

i '

_ }

I

_ 0 a

0 5 __

4

_ 5

_ 4 L

0 s 0 t 4 n

e S K g T A A N F g E I G T, n 6 t i

l A 5 3

U

_ l e G M G N

_ p 1,

I  ;

o L S E - -

0 G 0 3

m P u R H i A E T ht O R J I

_ i L S M A U R r

o M 0 5 N R l 8 f U I 2 E I E 4

)

I K l u

s I O G T

+

b R I l I, F I T K S R

0 _

0 0 O -

0 F F 2 X

O S U

C L H, I

I N 4 O 9

K 0 5

M 0

e r

l 1 u @ =

_ g F i

F O C

.  : 0 0

, 8 4

5

_g _

8

_ 3 5 2 5 I.

5 9 9 2 9 1 9 9 S

9 0 _

v ;

FRICTION AND WEAR PERFORMANCE

SUMMARY

COF VS STROKE NUMBER LUBRICANTS WEAR AVERAGE COF e 10K I BS.

COF (10K) BEGIN END CilANGE COAfAfrRCIAL MILS COF COF COF  %

1. MOLY 101, TEST 1 3 0127 0 12 0 132 10

2. MOLY 101, TEST 2 2 0 115 0 117 0104 -11

3. SRI 4 0133 0.122 0137 13

4. MOBILUX <1 0 135 0137 0132 -3

5. NEVER.SEEZ 165, TEST I 4 0 141 0 167 0 121 -28

6. NEVER-SEEZ 165, TEST 2 4 0.137 0.172 0 104 -40

7. FEL-PRO N-5000 3 0 103 0.143 0 077 -46

8. NEVER-SEEZ 160, TEST I 3 0172 0 172 0 1351 -22

9. NEVER-SEEZ i60, TEST 2 4 0 119 0 19 Otm -58

10. DURA-LITil 2 0.122 0 129 0 12 -7

11. MULTIFAK EP-1 1 0 123 0 124 0 126 1

12. LUBRIPLATE 930A A 1 0 128 0144 0 125 -13

13. RF GRAPIIITE 5 0.104 0.iB6 0(29 -52

14. DOW CORNING 44 3 0.142 0 185 O(N7 -48

15. MULTI-MOTIVE 1 0 089 0.109 0(b9 -37

16. N EBULA 2 0 061 0.1 0 073 -27 NEW TORAfULA TIONS

17. DARIN A , TEST I 1 0 067 0.103 0 074 -28

18. DA RIN A, TEST 2 2 0.114 0.121 0 108 -11

19. POLYUREA 1 0 095 0 13 0(E4 -35 NUMBER COF e TST SOLID LUBRICANTS STROKES BEGIN END

20. SPIRE SPUTTERED MoS2 ODSERVE 72 0 149 0257 72

21. EM E-620 t MoS2) OBSERVE 4 0 154 0 193 25

22. WillTE SPtTITER SILVER OD5ERVE 4 0173 0 338 95

13. IlOllMAN M2036 (MoS2) ODSERVE 76 0(b9 0 133 93

24. GR ArilITE INSERT NA NA NA NA NA

25. NFoi UBE 2 OHSFRVF 10(9) 0135(S) 0169(S) 25 NOTES: 1. (S)is stolic ordy tests

2. Severe wear was observed in all cose with btsis. obrasion and coatings worn off.

porticles of the coatings were observed below the strem tuJf.

FIGURE 9

g. .

OPENING VERTICAL STROKE 8 0.7000 l I l

-+--___.

___i _.

0.6000 ,/  ! l l N ,/ ,y-i~~

' ~, ,s' \ ,. - - - - - ..

}.

' t, N , M 0.5000 4

7 1 \ ____j / f l

' f >__..

0.4000

{'

SIE DISK F ACTOR I - ---- HMAC DISK FACTOH 0.3000 STEM WJ 1

0.2000

$ I 0.1000 .-

! l

-- i - 6- f --4-~ l-0.0000 l 1 21 23 25 27 13 15 17 19 3 5 7 9 11 1

3TROKE #

FIGURE IOa

- - - - - _ _ _ - _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - - - - ' ~ " " '"' - ,,

h i-  ;

CLOBING VERTICAL STROKES 0.7000 -

SIE DISK FACTOR f

-- --- - NMAC DISK FACTOR '

0.6000 ST EM MU i t

~

l l 7s /

0.5000

! s' N l l

' N ' x /i l/

OA000 't -- -

/^ q\

/s - - - - -

~

-R- . ~

.j,j t

3 0.2000 0.1000 l

! i 0.0000 } h I t +- -I 6 8 10 12 14 16 18 20 22 24 26 28 2 4 STROME #

FIGURE 10b j

N R

OPENING HORISONTAL STROKE 8 0.7000 j  !  !

s l

i  !' 's i < ( s

____,_ s I f ,_

---

~ ==__.. s c' ,-f- --y, / ,/

0 NKY) y '

s[;

y N i ____fiN  :

t N I 0.5000

/

s --

/ /,/ w.J 4

1

j. j -

0.4000 i ,

i SIE DISK FACTOH '  !

0.3000

---

NMAC DISK FACTOR f i SIEM W '

0.2000 i {, ,

I  !

l

• j t 0.1000 -

t 1 1-- -i i 0.0000 15 17 19 21 23 25 27 1 3 5 7 9 11 13 STROME #

FIGURE 10c

CLOSING HORI2ONTAL STROKES 0.7000 ,

-l l -l l SIE DISK FACTOR j ,e

/

0.6000

---

NMAC DISK FACTOR l//

4 0.5000 ,/N; STEM MU <

' s /i

}'<

N.

0.4000- /

s ,r'  ;

___ ,_ -; y----* _.1 - _. - -- p s  ;

' ~

y \ j -

, s N . N lt '

~

\-

0.3000 ,

\ \ l

\ f 02000 \- /

, / i i I I l i

0.1000 I

i I  ;

+

0.0000 . f I-- 1 2 4 6 8 10 12 14 16 18 20 22 24 26 28 STROME8 FIGURE 10d

l l

Rate of Loading Trend ,

for  !

i William Powell 14" Gate Valve i ROL (Ibs) 4,000 ROL < 5% (11 Strokes Showed No ROL) i

- O  :

O l 1

O l l

3,000 -

i i

O  !

2,000 -

O ,

D .:

0 0  !

O -l O Da j 1,000 -

O O y O i i

O O  :

0 100 200 300 400 500 600 700-O ]

DP (psi) _.

l Figure 11 1

.. . - . .. . ..- -. \

F 27 ,

It is essential that these type issues be handled -

in such a way as to ensure that operability is not jeopardized while minimizing unnecessary dose and i

cost additions to the baseline GL 89-10 program.

i Therefore, GGNS intends to address emergent MOV l issues independently from the MOV program.  !

Emergent issues will be treated in the same manner  :

as any other potential safety issue i.e.,

safety / risk significance will be assessed, the  :

effect on valve operability determined and, if operable, the issue will be addressed on a ,

priority, and at a cost, commensurate with its i safety significance at GGNS. I With an effective program in place, GGNS is well  ;

equipped to handle current and future technical l issues regarding MOVs. Figure 12 shows graphically that new MOV issues will be evaluated

• by existing GGNS programs for processing industry j

notices. A sound MOV program, coupled with existing programs to resolve technical issues,

will ensure that all emerging technical concerns d

are addressed in a timeframe consistent with their ,

importance.

3.5 GGNS MOV Program Effectiveness Meets GL 89-10  ;

Objectives i J

l To determine how effective the MOV program must be to meet the goal of GL 89-10, GGNS conducted a sensitivity study utilizing the GGNS IPE. Core damage frequency i

i. (CDF) was calculated for various MOV failure rates from l l 87/1000 down to 0/1000 to find the point of  !

diminishing returns in safety improvement. The results, plotted in Figure 13, clearly show negligible CDF improvement after reaching a 7/1000 failure rate.

Similar results are obtained when considering dose

] consequences. )

We have applied various quantitative measures as well as reasonable engineering judgment to conclude that y

implementation of the GGNS MOV program will result in MOV failure rates on the order of 3-7 per 1000 demands.

For GGNS, CDF improvement for failure rates below

7/1000 are on the order of 2 E-06. These values are

well within GL 89-10 improvement goals and also represent thresholds beyond which additional MOV work is not likely to be cost effective.

i

.a

Temperature Vane EHects Factor Trends

, r Numeroes Procedures for Resoluton p ROL g vajuatione

- of Safety ,

Equipment l S:=c ,

Issues l Accuracy j'ests Al lf

-<n sk, m i****

I lo w scac

~ pow .

,c l .

l A '

[1f 7

l r o,, 2  !

Design  %, j Vendor

! s indus i

1. Documen' \ .

' Reviews l Conformance Future issues Aging ProgrES k

b Pressure LocknQ b

Lubrication The MOV Program, Coupled With Existing Programs, Ensures Unsettled Technical Issues Will Be Addressed ,

4 l

Fioure 12

7 8 y-o-- ,

0 07 i 0 0 1 '

3 0 0

1 y 6 /

/

e t t i

v 4

a

', R i '

t i

s e r

n e , u e ' ^ l S ' ,

7 ia f F F

D 7/ , V y C 5

• O d 2

- ,N

. Y; 3M 0

/ ;> nI rrE%

t u

S 0 /

1

)

j,np$(2, $ nap;gP*r% oNi?isipbv.nwN'@

r y -

i t

8 f .y:

. n;pM ~l,. r o

i v - - ~ -

t 0 i

0 a o

• 2 s 1 L @ ut b O* v c) e rn E j 3 n L . 1 e e r

S u n ig i

o o F

t ,

0 c

u d 6 /

e 2 0 3 0 R y

/ 0 1

/

k t f e i

s i i

v 1

4 ,

4 a t

R t i

s l R n )( e r

e 7 u S (

7 i l a

e F s 3 , ,

o 4 V

D 4, $~z,a9;'C 0 3 O 1

M 4 ,<  :

}

8 v 8, .,q g h?f ,7 '

,!m,+O%e k 7 1, .ywvj@qsa?qdl - :3ldisi I -;.

8

- ~

0 0 0 o 0 0

0 0 0 0 0 5 0" 5 5, 2 '

2 i

)

Q5g*'.

COpOaDa)c

- c $oa

)

1 I

.. l l

l l 30 'I l i i

3.6 GGNS MOV Failure Trend A recent review of NPRDS failures for all types of safety related MOVs at GGNS shows a steady decrease in the number of cases since 1985, the first year of commercial operation (See Figure 14).

While failure rates under normal operation cannot be compared directly with failure rates under design basis l conditions, it is reasonable to assume that improved i performance under normal operations translates to ,

improved performance under design basis conditions as I well.

As suggested by this data, we believe that improvements in GGNS MOV practices are responsible for the positive trend in MOV reliability.

i 4.0 Conclusions The GGNS MOV Program, " Project SMART" follows the guidance of GL 89-10, except as modified by positions taken in this document. When implemented, this program fully achieves the safety improvement sought in the 1 generic letter by returning MOV failure rates to  !

l acceptable levels.

Completion of the MOV program will not be contingent on i resolution of emerging technical issues since l negligible reduction in risk is expected by doing so. l Instead, emergent issues will be evaluated under normal l plant programs on a priority and cost basis consistent j vith the issue's significance to GGNS. i s

GGNS will submit closure notification as requested in GL 89-10, recommended action (m), when actions discussed herein are complete.

l

GGNS NPRDS MOV Failure Trends MOV Failures 18

O 16 -

Trend Indicates 85-03 and 89-10 Programs Making Positive impact

~

E 14 -

O O 12 -

O 10 -

8 -

O O rioieded i

6 -

O 4 -

O I I '

i 2 94 84 86 88 90 92 Year Figure 14 -

_ _ _ _ .________ _ ________________ _____ . . _ . . _ _ _ _ . _ _ _ _ _ _ _ _ _ __ _ . _ _ . _ . . _ . - - . _ _ . . _ _ _ . . . . . ~ . _ - _ - .

~~

i [

32 ,

t 5.0 References MAEC-90/0301, Memo from L.L. Kintner, GGNS NRC Proj ect l 1.

Manager to W.T. Cottle, V.P. GGNS Operations, dated 12/4/90

2. Wyle Laboratories Test Report 43008-01, 2/18/93, " Flow .

l Loop Differential Pressure and Pressure Lock Tests on a [

l 14" William Powell Gate Valve for Entergy Operations" j l r

3. Value Impact Analysis for Implementation of GL 89-10 at Grand Gulf Nuclear Station, Rev. O, February, 1993. ,

4. NUREG/CR-5140, Value Impact Analysis for Extension of  :

NRC Bulletin 85-03 to Cover All Safety Related MOVs,  !

July 1988, Brookhaven National Laboratory l j

5. Siemens Report 6300/93/001, " Similarity Analysis for Globe Valves: Powell, Anchor Darling and Yarway,  ;

3/30/93

6. Siemens Report 6300/92/118, " Similarity Analysis for Butterfly Valves: Pratt 150#", 3/30/93  ;

l l

7. Siemens Report 6300/92/114, " Similarity Analysis for '

Gate Valves: Anchor Darling Flex Wedge", 11/10/92 l

8. Siemens Report 6300/92/103, " Similarity Analysis for William Powell 900# Gate Valves", 6/1/92 ,

i

9. Siemens Report 6300/92/104, " Similarity Analysis for l William Powell 600# Flex Wedge Gate Valves", 7/10/92 l

10. Siemens Report 6300/92/105, " Similarity Analysis for William Powell 300# Flex Wedge Gate Valves", 9/14/92

11. Siemens Report 6300/92/106, " Similarity Analysis for l William Powell 150# Flex Wedge Gate Valves", 10/8/92

12. Siemens Report 6300/92/100E, " Analysis of Gate Valve Internals", 9/1/92

13. NUREG-0933, "A Prioritization of Generic Safety Issues", July, 1991

14. SECY-91-270, " Interim Guidance on Staff Implementation of the Commissions Safety Goal Policy", 8/27/91

15. Westinghouse Engineering Mechanics Division Memorandum 5683, R/1, 3/31/82, "EPRI Summary Report: Westinghouse Gate Valve Closure Testing Program

i t

t Flow Test Program Details Appendix A  !

i D

i e

k i

i k

)

L l

i p

e i

t I

l 4

l l

l i

i l

l 1

J 1

i l

k i

i 1

1 1.0. Background

, The generic letter requests utilities to perform design j basis flow testing on all valves practicable in order to confirm the adequacy of control switch settings. This '

recommendation raises three basic concerns.

a. In-situ flow testing is not possible for all MOVs in l the program. Therefore a method must be developed to justify the extension of flow test data from one valve to another, i.e., type testing. i N

b. Where in-situ testing is possible, maximum attainable i differential pressures (DP) are often less than the ]

calculated maximum design basis DP. This creates a ,

need to extrapolate test results to higher DPs than l tested. l I

c. GL 89-10 and supplements do not endorse the use of valve families to limit the number of flow tests to be y

i performed. In other words, if all valves in a family

- subject to similar conditions can be tested, they should be. This position creates the possibility that unnecessary cost and dose.will be expended for DP testing that may not add useful information to the data pool.

Because of these concerns it was clear to the GGNS MOV project team that a better understanding of valve operation would be required to extend flow test results to other j valves, to extrapolate data, and to avoid unnecessary duplicate testing of MOVs.

Therefore on September 5, 1990, GGNS met with the NRC to l

discuss the feasibility of developing an analytical approach j to the problem. An MOV performance analysis that determines how well a valve will perform under design basis conditions was proposed.

l I

I i

a a

d I

i

~

i 2

After this initial meeting, Siemens Nuclear Power Services was retained for both the conceptual work and the actual MOV performance analysis. This work is now complete. A minimum l population of valves to be tested has been selected based on the analysis which identifies MOV tests that envelope other MOVs in the family.

A summary of related memos, meetings, and milestones is provided for information in Attachment A1.

2.0 MOV Performance Analysis An MOV performance analysis opens a window to valve internal operation that provides the broad understanding of performance factors required to justify type testing. The major advantages to this approach are as follows:  !

a. Identification of performance problems related to disk tilting and seat / guide damage  ;

b. Establishment of an engineering basis to apply test l results to untested MOVs

c. Establishment of an engineering basis for extrapolation of data to higher DPs

d. Provides insight for aging evaluation [

e. Eliminates unneeded duplication testing and associated l dose / costs.

Unless sufficient analytical work is performed, little can be said about why a valve behaved as it did during flow testing. Without this understanding of valve behavior, it -

becomes difficult to justify extrapolation of test data to higher DPs or other valves.  ;

I

3 3.0 Performance Analysis Methodology The performance analysis for GGNS MOVs was conducted as follows.

3.1 Valve Data and Service Conditions With assistance from the valve vendor, a comprehensive data base containing important valve internal dimensions, clearances, tolerances, materials, service conditions, orientation, piping- configuration (butterfly valves only), failure history, safety function, and operator data was asserbled for each MOV.

The assembled data was used as input to the evaluation of program MOVs. A complete list of data fields required for gate, globe, and butterfly valves is found in Attachment A2.

3.2 Field Verification of Critical Gate Valve Dimensions Gate valve guide dimensions, clearances, and tolerances are especially critical in predicting MOV operation. It is reasonable to expect these dimensions to be within tolerances on production units, especially since the valves were supplied under a nuclear QA program.

Two William Powell gate valves have been inspected and  ;

critical dimensions were found to be within tolerance ,

except for some minor variations. The valves inspected '

were a 10" 900# valve tested by INEL for blowdown  !

isolation, a 14", 600# valve tested by GGNS at Wyle Labs. l These results give no reason to expect gross, widespread,.

manufacturing problems with William Powell gate valves.

Similar inspection results for Anchor Darling gate valves will be reviewed. Dimensional checks on globe and butterfly valves are not considered necessary since small variations in as-built geometry have less potential for impact on valve performance.

I 4

3.3 Valve Evaluation Once necessary valve data was obtained, a DP Build-up vs

% Stroke curve was developed. This curve is based on ,

valve Cv and flow driver characteristics which envelope expected conditions f or the system. Blowdown conditions were assumed where appropriate.

i 3.3.1 Gate Valves Gate valve families were developed based on a comparison l of valve design, system requirements, and performance over the stroke. t The main design criteria used to define the gate valve families are:

a. Vendor / pressure class / size / type

b. Materials / dimensions / clearances / tolerances

c. Guide type (male / female)

d. Stem to disk connection  ;

r

e. Hard facing of sliding surfaces i Important system parameters are:

a. Line pressure and maximum expected l

differential pressure

b. Fluid media and temperature l

c. Flow rate, velocity, and driver (pump / blowdown)

d. Orientation

e. Stem speed l l

l 1

I i

r 5

Similarity of valves was determined based on a full stroke analysis performed using Siemens' proprietary computer code, "GATEVLV". The following were ,

investigated or calculated over the entire stroke at 1%  !

intervals. ,

a. Initial flow contact and flow cut-off points

b. Disk axial and lateral tilt angles

c. When disk stability occurs on body guides and seat

d. When disk contacts seat

e. Time trace of DP build-up

f. Time trace of disk tilting moment

g. Disk and guide deflection

h. Disk, seat, and guide bearing stresses

i. Disk and guide bending stresses

j. Weld stress for welded body guides

k. Interaction between disk and body guides Other important parameters evaluated are:

a. Effect of stem and body orientation

b. Evaluation of manufacturer's disk and seat hard-facing processes and material i

I 3.3.2 Butterfly Valves Butterfly valve families were developed based on a comparison of the basic design features. System parameters, including components up and downstream, and internal stresses of key components were used to identify valves for testing.  ;

i l

6 ,

i

)

The main design criteria used to define the butterfly l valve families are:

a. Vendor / pressure class / size / type

b. Materials / dimensions _

c. Disk shape / disk offsets /hard stop design /  ;

sealing design / bearing design / packing I Important parameters used in the similarity analysis are:

a. Max differential pressure /line i pressure / temperature

b. Fluid media, flow driver, flow rate , and j flow velocity

c. Safety requirement / maintenance history .,

d. Up and downstream components and distances / component orientation / valve orientation / disk facing / rotation direction of disk .

i The similarity of valves is determined when considering l the above criteria in a full stroke analysis._ Similar valves must show similar static and dynamic behavior, I sealing properties, and stress ranges in keys, pins, and bearings.

3.3.3 Globe Valves Globe valve families were developed based on a comparison of functional parts. System parameters such as flow velocity, max differential pressure, and flow l l

over or under the disk were also used to furthe"  !

identify valves for testing.

The main design criteria used to establish globe valve-families are: ,

a. Valve vendor / pressure class / size / type l l

b. Design features such as anti-rotation device / disk guiding mechanism / disk and seat angle / disk shape i

l l

~,

~

7 l

c. Macerial and dimensions j 1

Important system parameters used in the similarity-analysis are: )

a. Max differential pressure /line pressure / fluid temperature  ;

b. Flow velocity / flow rate / fluid medium / flow ,

i driver / flow direction (over or under seat)

c. Safety requirement / maintenance history

d. orientation / stem speed i

The similarity of valves was determined by examining f various combinations of design criteria and system ,

parameters. Some of the more important analytical considerations are:

a. High stem speed, a steep disk / seat angle, and flow over with a safety requirement to open.

b. No disk guiding with a high flow velocity and l DP. ,

l' C. Friction forces and bearing stresses between the stem and the anti-rotation key where high ,

DP is expected,

d. Compression / tension forces in the stem due to -

DP. ,

i

, 3.4 Identification of Test Candidates i i

Based on the above information, each valve was evaluated to determine which had the most limiting [

stresses, orientation, piping configuration, etc.  !

Justification for applying test data from one valve to another is provided in the Siemens reports for each  ;

family. (References 5 through 11) I s

In general, MOVs with the most critical conditions were  ;

chosen since these test results will envelope the conditions for other valves.

i i

6 1

8 4.0 Flow Tests Required Figure 3 summarizes the flow testing requirements established f rom the above analysis. It gives a brief description of each group, the number of valves in the group, and the number of tests required. Also shown are the number of tests planned by GGNS for each group. Minor changes in scope may be necessary  :

before the flow testing portion of the program is complete.

i Figure 3 is graphically summarized in Figures 4 and 5. In addition, the DP levels achievable for the proposed testing are summarized in Figure A1.

1 5.0 Performing Flow Tests Flow testing is performed in accordance with special test instructions. A typical flow test involves the following.

a. The system configuration that produces the maximum system pressure and flow is determined.

b. The valve is stroked at the most challenging normal service conditions achievable and at several other lower flow / pressure conditions.

c. Static tests are performed before and af ter flow testing.

The pre-test is used to evaluate ROL and to ensure thrust / torque are within the setting window. The post-test is an operability check and is also used to evaluate ROL.

d. Temporary instrumentation is installed where possible to record DP if system instrumentation is not adequate.

Temperature and flow are recorded to the extent practicable by existing instruments.

In a timely manner after the flow test is complete, the test engineer performs an initial evaluation of MOV capability in accordance with plant procedures. Post test data evaluation includes assessment of valve factor, stem factor, rate-of-loading, extrapolated thrust requirements at MEDP, degraded voltage capability, and valve flow cutoff capability. j If acceptance criteria are not satisfied, a nonconformance document is generated for engineering disposition. If the initial assessment by the test engineer is satisfactory, the data is transmitted to design engineering for final evaluation.

e J

/

. s

\ O

j. o P 1, 5~ - - N

'o, 0 #

e x o'.

Y @

• 0 o

• O

\

o 9 e o  %

0+O99O x  ;;

\

o m\ayNf

- y;

3 _

sc c y

(h r q@ sl%e{alb s

+

Eh g$ .

t Et  %

Z o +g O Q s

\ g f

O .

,/

f s

g

\

\

. i

, ~.

10 6.0 Evaluation of Design Specification Against Test Results Flow test data for individual valves is transmitted to design engineering for review in a timely manner and when required, is incorporated into the design specification. l After all flow testing is performed for a group of valves the data is evaluated to determine whether the design  !

specification is acceptable for the untested valves in the group. If thrust / torque requirements in the design ,

specification do not envelope test results, the specification will be revised accordingly. Revisions are made as soon as 1 possible after the evaluation of untested MOVs in the group.

This process of revising the specification to reflect empirical data will conclude as soon as practicable after the '

last flow test scheduled for RFO7 in June, 1995.

l 7.O Determine New Thrust Window When design engineering revises the design specification to j incorporate new, empirically based, thrust / torque l requirements, Performance & System Engineering will calculate l new thrust windows if required. Specified minimum and maximum limits will be adjusted for operator and diagnostic equipment error, Rate-of-Loading (ROL), and degradation. ,

i The amount of margin added will be based on the actual l magnitudes of ROL observed during flow tests conducted for th I

specific valve family.

i 8.0 Identify Deficient As-Left Settings i

As-left switch settings will be reviewed to ensure they fall ,

within any new thrust windows calculated. If the as-left setting falls outside the window then a non-conformance document will be generated with a supporting justification f or operability. If required, the MOV will be reset or modified at the earliest practicable time to support the new empirical requirements.

1

^"" * * "' # '

GGNS Milestones l Development of Valve Families September 5,1990 GGNS/NRC Meeting to Discuss MOV Program ,

September 20,1990 NRC Letter Expresses Reservation Concerning Use of Valve Families October 1,1990 GGNS Offers Additional Explanation on Developing Valve Families December 4,1990 NRC Letter States Approach Seems f Feasible But Justification Required i January 29,1991 RFPIssued May 24,1991 Contract with Siemens/KWU to Develop Pilot Valve Family ,

October 31,1991 GGNS/NRC Meeting to Discuss Results of Siemens Pilot Valve Family NRC Considers Method Promising l

February 14,1992 NRC MOV Insp. Cites Valve Family l Approach as a Strength - Cautions  :

Against Use in Lieu of "All Practicable" l l

l March 1,1992 GGNS Contract with Siemens to -

I Develop Valve Families for All GL MOVs March 1993 Siemens Completes Valve Families GGNS Flow Test Scope Finalized

.~,

1