NMP2L0896, Interim Deficiency Rept 50.55(e)-86-20 Re Excessive Leakage of Msiv.Initially Reported on 860909.Meeting Scheduled for 861015 in Bethesda,Md to Discuss Rept
| ML20211A933 | |
| Person / Time | |
|---|---|
| Site: | Nine Mile Point  |
| Issue date: | 10/08/1986 |
| From: | Mangan C NIAGARA MOHAWK POWER CORP. |
| To: | Kane W NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| References | |
| 50.55(E)-86-20, NMP2L0896, NMP2L896, NUDOCS 8610170083 | |
| Download: ML20211A933 (92) | |
Text
M V NIAGARA HUMOHAWK NIAGARA MOHAWK POWER CORPORATION /300 ERIE BOULEVARD WEST, SYRACUSE N.Y.13202/ TELEPHONE (315) 474-1511 October 8,1986 NMP2LO896 Mr. W. Kane, Director U. S. Nuclear Regulatory Comission Region I Division of Reactor Projects 631 Park Avenue King of Prussia, PA 19406 Re: Nine Mile Point - Unit 2 Docket No. 50-410 Interin Report for 50.55(e)-86-20
Dear Mr. Kane:
Attached is an interim report regarding the problem with excessive leakage of the Unit 2 Main Steam Isolation Valves which was reported via tel-con to G.
Meyer on September 9,1986, in accordance with the requirements of 10CFR50.55(e).
A neeting has been scheduled for 10:00 a.m. on October 15,1986, in the NRC's Bethesda facility in order to discuss further the bases for this report. Between now and this meetin detailed verification of the report'g, sNiagara bases. Mohawk During thewillmeeting complete we its will identify for the NRC staff any corrections, in addition to any information developed subsequent to this report.
After this meeting, Niagara Mohawk will incorporate in the report all new information, and address the NRC staff's comments and questions. Niagara Mohawk plans to submit a final report on this problem as soon as possible and will provide a submittal schedule at the conclusion of the meeting.
Very truly yours, Ml C. V. MangarF Senior Vice President CVM/CDT/dl (0023C) xc: Director, NRC Office of Inspection and Enforcement M. Haughey, NRC Project Manager (20)
W. Cook, NRC Senior Resident Inspector Public Service Commission (2)
Project File (2) 8610170083 861000 PDR ADOCK 05000410 S. _.
PDR ..-
{( d i'6
INTERIM REPORT 10CFR50.55 (e)
MSIV LEAKAGE NINE MILE POINT UNIT #2 NIAGARA MOHAWK POWER CORPORATION OCTOBER 1986 1 ,
TABLE OF CONTENTS
1.0 INTRODUCTION
1.1 Objectives 1.2 Executive Summary 1.1 Safety Evaluation 7.G BACKGROUND 2.1 valve Description ,
7.2 Valve Selection
~
2.3 Initial Qualification 7.4 Other Operating Experience and Corrective Action 3.G PROBLEM DESCRIPTION 3.1 Problem Discovery 3.1.1 Test Method 3.1.2 Test Data 3.2 Description of Failure 4.G EVALUATION AND ANALYSIS 4.1 Root Cause Analysis 4.1.1 14echanical Analysis of Seat and Ball i 2
1
4.1.2 Stress Analysis 4.1.3 Metallurgical Evaluation 4.1.4 Root Cause 4.2 Description of Tests and Results 4.2.1 Purpose 4.2.2 Test Parameters 4.7.3 Test Combinations 4.7.4 Test Results 5.G CORRECTIVE ACTION AND TECHNICAL JUSTIFICATION 5.1 Corrective Actfon 5.1.1 Recoated Balls 9.1.2 Modified Springs ,
5.2 Mechanical Analysis of New Ball and Modified Springs 5.3 Stress Analysis ,
5.4 Effect of Operating Conditions 5.4.1 Describe Operating Conditions 5.4.2 Steam Effects 5.4.3 Temperature Effects 9.9 Evaluation of Test with Blended Balls 6.0 ADDITIONAL CONFIRMATORY TESTING 6.1 Mid Cycle Leakage Testing 6.1.1 Initial Type "C" Testing 6.1.2 Mid Cycle Test 3
1 1
6.2 Developmental Testing 6.2.1 Test Objectives 6.2.2 Test Organization 6.2.3 Testing Format and Schedule i
i 7.9 CONTINGENCY PLAN 7.1 Leakage Control System 7.2 Y-Pattern Globe Valves R.0 CONCLUSION
/
4.A APPENDICIES i
4
1.P INTRODUCTION During March 1956 the eight NMp2 Main Steam Isolation Valves (MSIV) passed their formal " Type C" Local Leak Rate Tests (LLRT) conducted in accordance with the requirements of Appendix J of 10CFR50.
On September 2, 1986 NMPC conducted additional LLRT's on the MSIV's.
During August 1986 the valves were operated in excess of 100 times.
This was an attempt to ctrrect actustor problems.
The September leak testing was initiated to provide data supporting the acceptability of the LLRT method used at NMP2 for this type of ball valve. During this testing it was discovered that all eight MSIV's had exceeded allowable leakage. Immediately following this discovery an extensive series of inspections, analyses and tests were performed.
1M Objectives This report presents the original bases for selecting these valves for this application, discusses the details of the leakage problem, presents a root cause analysis, and describes the corrective actions that have been taken including those tests performed to justify our resolution of the. leakage problem. The justification for the corrective action is based on analytical and test results, together with additional planned confirmatory testing and contingency plans. This document is the interim report required by 10CFR50.55 5
(e) addressing the MSIV leakage problem. This report supplements information contained within the Application for Schedular Exemption Related to Further Analysis of and Possible Modification to the Main Steam Tsolation Valves s ubmi t ted to the NRC on October 2, 1986 9
8 e
6 ,
1.2 Executive Summary Niagara Mohawk believes that the problem resolution plan presented in this report includes sufficient testing and analyses to demonstrate that the MSIV's will remain leak-tight through the first operating cycle. Furthermore, Niagara Mohawk is confident that the mechanism
- which caused the excessive valve leakage is understood. Specifically,
, it has been determined to be wearing stress (excessive contact stress combined with friction) which causes localized delamination of the tungsten carbide ball coating. The removed material scratches I the stellite seats, causing excessive leakage. Actions are being
. taken to restore the valves to an acceptable condition and to reduce the wearing stress to an acceptable level. On site cycle testing and analysis of operating conditions provide confidence that the coating delamination will not occur during the first operating cycle.
Analyses have been completed which d emo n s tra te that normal and emergency operating conditions, including effects of temperature and valve closure under steam flow, do not add significantly to the condition causing delamination on the ball. Further, comparison of calculated maximum bearing stress and as-tested tungsten carbide bearing strength shows that in situ forces should remain below those necessary to cause for coating.
Niagara Mohawk is committed to continued testing and contingency programs. The long-range testing programs provide further assurance 7
. _ _ _ _ , _ _ . - . . - _ = _ ~ . _ _ _ . _ _ - _ - _ _ - - - - - -
that the root cause of previous failures are thoroughly confirmed and documented. If modifications are required to assure reliable service of the valves throughout plant operating life they will be completely developed and thoroughly tested for implementation during the first refueling outage. Also, design and procurement activities for the addition of a MSIV leakage control system will continue to proceed on an expedited basis should future testing show the need for such a measure. Finally, design and procurement planning for Y-pattern globe valves and their associated leakage control system remain in progress.
In conclusion, the current MSIV ball valves are safe for a minimum of one cycle of unit operation. Ongoing analyses and testing are being expedited to assure early identification of concerns so that any long term modifications can be implemented. Contingency programs are also being expedited to ensure they can be implemented if necessary.
1.3 Rafety Evaluation The quick-closing MSIV's function to isolate the reactor and containment systems in the event of a break in a steam line outside the primary containment, a design basis loss of coolant accident (LOCA), or other events requiring main steam line or containment isolation. In the case of a main steam line break, the isolation valves would terminate the blowdown of reactor coolant in 3 to s seconds thereby preventing an uncontrolled release of radioactivity 8
from the reactor vessel to the environment.
Results of NRC staff standard plant analyses, which use conservative assumptions for considering the offsite consequences of a postulated design-basis LOCA coupled with uncontrolled leakage of the main steam isolation valves above technical specification limits, have indicated that the calculated doses would be in excess of 10 CFR 100 guidelines.
Even though the valves did not meet leakage specifications, they would close and would have terminated blowdown of the reactor vessel.
Had the leak rate through the main steam isolation valves not been corrected, a design-basis LOCA could result in offsite doses in excess of 10 CFR 100 guidelines; or a main steam line break could result in excessive doses in plant occupied areas.
l 9
2.0 BACKGROUND
2.1 Valve Description NMP2 is provided with two MSIV's on each of four main steam lines (see Figure 2-1). The valves are designed to provide isolation within 3 to 5 seconds during such an emergency.
The MSIV's also provide redundant isolation between the reactor vessel and the main turbine generator, thereby permitting normal operation and maintenance of the steam plant systems while the reactor is at operating temperature and pressure. Valve operation under normal conditions takes approximately five (5) minutes to open and twenty (20) seconds to close. A more detailed discussion of normal closure is contained in Section 5.4.1.
4 The MSIV's are ball type valves, welded into a horizontal pipe run of each of the four main steam lines; one valve is close to the inside of the primary containment and the other is located just outside the containment.
A complete description of the valves is detailed in Gulf & Western Topical Report No. G&W FSD 2538 submitted to the NRC on January 24, 1979. Figure 5.4-7 (attached to this section) taken from the NMP2 FSAR shows a cutaway view of an MSIV. Each 24-inch reduced-port (21-inch) ball type valve has a full-ported ball with an integrally cast top and bottom trunnion. At rated steam 10 .
flow through each valve the pressure drop through a valve is calculated to be 1.2 psi.
The valve internals are top loaded into the valve which allows disassembly without removing the valve from the piping system.
The ball, when rotated within the body, is aligned in two roller-bearing assemblies (upper and lower). Valve seal assemblies are fit into the valve body and held in contact with the ball by the force of multiple springs. The seat-to-body interface is sealed by a multiple ring packing seal that is compressed by the seal springs. The seal is produced by the spring-loaded force of the seat against the balls surface. In a closed position the upstream seat's sealing force is aided by system pressure acting against the projected annular area of the seat. A multiple ring packed stuffing box seals each ball trunnion against the body.
The bonnet closure is sealed by a metallic, pressure-type seal ring. The ball is designed with a vent hole between the flow hole of the ball and the bottom of the ball. During operation with the valve open, this equalizes pressure between the body cavity and the steam flow area, thereby reducing seat differential pressure loading. Details of the materials of ball construction are contained in Appendix 9.1.
11 .
2.2 Valve Selection This ball valve was selected as the NMP2 M5IV based on the significant benefits compared with the standard Y-pattern globe valve, and based on the extensive testing and analysis of smaller valves of the same design chat demonstrated their suitability for the proposed application. At the time of purchase these benefits were determined to be long range enhancements to the unit's overall availability since Y-pattern valves were considered a maintenance problem. An MSIV chronology is contained in Appendix 9.2 The main advantages'of the ball valve are summarized below:
o Each valve contains two sealing surfaces, o operating characteristics which result in less wear and stress on valve internals and thus less maintenance.
o Ease of disassembly and maintenance will result in less plant down time and radiation exposure.
o Low pressure drop providing optimum steam conditions for power generation.
o Simplified piping, support and restraint configurations.
o Low operating (rotational) velocity even during emergency 12
closure (5 rpm or ~ l f t/sec maximum rotational velocity) .
o Sealing components are not used in the deceleration and stopping of the valve in the fully closed position, o Seating surface maintenance can be accomplished in remote areas thereby minimizing personnel exposure.
2.3 Initial Oualification Energy Products Group (EPG) Fluid Systems Division of Gulf and Western Manufacturing Company was the original supplier of the MSIV's and publisher of the Topical Report. Most of the initial qualification testing was performed by this organization, and is described in their Topical Report. A summary of this testing is provided in the following paragraphs.
Dynamic qualification of the valve assembly was demonstrated through a combination of testing and analysis which evaluated all parameters i affecting its function. These parameters included flow, temperature l 1
and combined seismic and hydrodynamic loads, in both normal and i accident conditions.
The ball valve and actuator design was subjected to extensive at the EPG facility in Warwick, Rhode Island. In August 1976 a prototype 8 inch valve and actuator were tested with 1000 psig 13
saturated steam. The valve was in the open position with a downstream block valve initially closed. The block valve was then opened to allow steam flow through the 8 inch ball valve while venting to the atmosphere. The ball valve was then closed against the saturated steam flow of approximately 1.8 x ig6 LB/HR. Testing was repeated several times against varying flow rates. During the test, the valve body was subjected to torsional, axial, and bending loads that approximated plastic deformation limits of a la inch pipe. Data were taken regarding the stresses incurred in the body and actuator as well as determining the closure torque requirements.
Subsequent tc the flow test at Hunstville, the valve was shipped back to the manufacturer for leak rate testing. The leakage was found to be excessive. Upon d i s a s s embly it was noted that the ball had sustained galling on its upstream face. This was determined to be caused by debris being carried in the steam and impacting the ball during the closing cycle. Substantiation for this conclusion was the considerable mill scale and oxidation (rust) found in l
the body cavity. For this test, both the 8 inch ball and seats l
were overlayed with Stellite which further contributed .t o the )
galling effect once the coating was damaged on the ball face.
The ball and seats were reworked and the reassembled valve exhibited a very low leakage rate.
l l
14
This testing' resulted in several key conclusions:
o Flow has little effect on required closure torque, o The ball required a more durable coating and properly matched seat material.
o Mathematical models could be developed for valve body and actuator stress analysis.
In addition to the Wyle Laboratorie.s tests, a quench test was conducted on a 3 inch ball at EPG. This demonstrated that the tungsten carbide coating on the ball is suitable for the thermal transients a n ticipa ted during all modes of plant operation. The test was performed using a ball overlayed with stellite and coated with tungsten carbide. The ball was heated slowly to 4750 F and then quenched in water. The ball was visually inspected with no relevant indications found.
7.4 Other Ooerating Experience and Corrective Actions Beaver valley has three ball valves originally manuf actured by EPG and which are identical to those at NMP2. This unit has not operated their valves enough at this time to have useful information about leakage related problems.
The Swiss utility Kern Kraftwerk Leibstadt (KKL) has four of these 15 ,
valves. These valves are the same size, mate ial and p ovided by the same vendor as those at NMP2. These valves are normally open and used only for low pressure containment isolation. Each valve is the third isolation valve in the main steam line; the first two valves are Y-pattern globe type valves.
KKL identified corrosion (rusting) of the carbon valve body under the spool seat packing. A noncorrosive material was overlaid in the packing area. KKL also uses a low salt (chloride) Titan packing to avoid spool bore corrosion. Based on the Swiss experience NMPC developed and implemented a repair procedure to correct this problem. Incoloy.A25 was applied to the valve body adjecent to the spool packing seal. This enhancement was completed in 1985.
KKL has also identified wearing of the tungsten carbide ball coating and measured leakage rates greater than their allowable amount.
The KKL valves have experienced hot steam conditions in the open position; they have not been closed against design steam flow. KKL has sixty-ll5 pound springs equally spaced on their seal ring.
The NMP2 valves will have a modified spring arrangement.
Upon inspection of their valves, KKL observed radial cracks on one of the spool seats. Following discussions with KKL personnel ,
1 it was determined that this was an isolated incident relating to a manufacturing defect. These radial cracks were repaired and returned to service.
16 .
~ - - - - - - -- - - , .
Following rework to remove seat scratches the NMP2 seal rings will be choroughly inspected by appropriate NDE methods also, this seal ring has been successfully analyzed for thermal condition.
1 i
i 17
LINE B I; -
W - -
PEN. -
W -
]
i) A z_3g A 0 2 MSS *HYV78 2 MSS *HYV6B PRIMARY CONTAINMENT LINE A 1: -
W -
PEN. -
W -
)
l i) A z-1A A 0 2 MSS *HYV7A '
2 MSS *HYV6A PRIMARY CONT AINMENT WALL
! l LINE D I; W -
PEN. -
W -
)
A l i) A Z-10 0 2 MSS *HYV7D 2 MSS *HYV6D l
1 LINE C 1 W PEN -
W - )
A
- ~ ~
i) A z-1c 0
) 2 MSS?HYV7C 2 MSS *HYV6C MSIV
- h ARRANGEMENT FIGURE 2-1 euusivuo42
l l
TRUNNION G e BEARING PLATE "" @
LANTERN RING- f .
8 e i BONNET ~_, h / */
g BODY a PRE 55URE SEAL I _1 ,
i ROLLER BEARING~ ! - '
ASSEMBLY BEARING RETAINER
[
N
\
B A LL
}'- ,
I REDUCER SPRING C VER SPRING RETAINER l SPOOL / SEAT FIGURE 5.4-7 MAIN STEAM ISOLATION VALVE CUTAWAY VIEW NIAGARA MOHAWK POWER CORPORATION NINE MILE POINT-UNIT 2 19 FINAL SAFETY ANALYSIS REPORT
3.4 PROBLEM DESCRIPTION 1.1 Problem Discovery 3.1.1 Test Method 1
l l The NMP2 MSIV's are allowed a maximum leakage of 6 standard cubic feet of air per hour (SCFH) per valve using a test differential pressure of 40 psi. This leakage can be measured in either of two ways; in the normal leakage flow direction through the valve (inboard to outboard) 'or between the valve's upstream and downstream I seats by pressurizing the MSIV body cavity. Appendix 4.3 provides additional details on MSIV LLRT.
Recent leakage test results have confirmed that the between seat method is more conservative in determining 1
valve leakage conditions. This method directly tests both seats at the required pressure. Through the valve testing allows a substantial pressure drop through the upstream seat, lowering the driving pressure for downstream seat leakage, thereby resulting in lower values for total measured leakage. Figure 3-1 demonstrates this condition.
20
3.1.2 Test Data Between seat " Type C" tests were performed for informational purposes after valve assembly in April 1985. These tests were conducted to verify proper installation of essential sealing components. Formal, between seat,
" Type C", leak tests were conducted in March 1986.
An estimated 15-20 cycles of the valves occured between April 1985 and March 1986. Most of these cycles were with a anual actuator which takes approximately 5 minutes to stroke the valve fully open or closed. The last 2-3 cycles prior to formal testing were with the actuator.
The valve closure speed during these strokes was within the required 3-5 seconds. All data demonstrated leakage well below the 6 SCFH allowable value. Table 3-1 (attached) provides a summary of these two tests.
On August 28, 1986 the NRC questioned the test ~ method since it did not simulate the in situ flow dirgotion.
The NMPC position as previously stated was that the between seat test is conservative. On September 2, 1986 in an effort to demonstrate this position, NMPC conducted a leakage test between two valves. This test indicated excessive seat leakage. Since our previous Type "C" tests were between the seats, we duplicated our' testing technique and recorded excessive leakage
, between the seats of the valve. All valves failed the 6 SCFH acceptance criteria. Table 3-2 (attached) indicates theactualleakagevgues.
1
From March 1986 to September 1986 all eight valves were stroked during actuator testing. It is estimated that each valve was cycled over 100 times as part of this testing. The tests results dictated that visual inspection of the ball and seats was required. The first ball was removed for inspection on September 8, 1986.
3.7 Description of Failure
~
All MSIV's were disassembled and inspected. Visual inspection revealed that the eight valves exhibited similar conditions, markings and defects. Figures 3-2 and 3-3 show typical damage locations.
o There were scarred areas on the hard surfaced ball where patches of the tungsten carbide were missing. The removed areas were near the ball open position to the right of the ball vertical centerline at the top and bottom of the ficw hole. Approximately 8 to la mils of material had been removed.
There was also some galling in the scarred areas.
o There was some evidence of galling on each ball's machined
~
bottom surface which bears on a 304 Stainless Steel thrust washer. Galling of this surface is addressed in Appendix 9.4 22
o There were some scratches in the direction of ball rotation and there was material on the ball in areas where the seat would seal against the ball. Laboratory analysis of the material on the ball and in the valve body established that this was common rust. .
o The Stellite seal ring showed signs of scratching due to the cutting by tungsten carbide particles. ,
The major portion of damage was evident in areas near the valve open position. The closed sealing surfaces were in relatively good condition. However,.the stellite seal ring scratching was sufficient to degrade the valves sealing capabilities.
23
BODY CAVITY UP STREAM SEAT DOWN STRE AM SEAT P2 hM/M///l W//MAk FLOW
' ~
3 DIRECTION BALL
(/Y [
\
\ / d (
\/
)
g THROUGH VALVE TEST /
P, = 40 PSIG > P, > P3 = 0 PSIG BETWEEN SE AT TEST P2 = 40 PSIG > P, = 0 PSIG & P3 = 0 PSIG i MSIV LE AK AGE TESTING NMMSIVMD39 FIGURE 3-1
N2 O3 I
TE I R DGU V NFI I
1/
SO i i I l l I l l I MC h
L E L o SO A L
i W
B D+) C E i TI i" I
V (Q
i P
O
'i T f
,l l l l l l l l l A A E E R R A A D D E E G G A A M M A A D D I ,
/
x .
_'i i* W E
I I
i 3
I
+ !
3 i
V E
D
\
1 j, .
S I , A E
R A )
D E
)
E D EI D 0 I
GS 4 S A D R MR M V
A A A I S
F = D(F M
( ' M N
C ]
CLOCKWISE v TO CLOSE w J DAMAGED AREAS TO THE RIGHT OF VERTIC AL CENTERLINE cp -
i i
s
- - s MSIV B AJ. L 26 C0NDITION NMMSIVMD 4 6
- FIGURE 3-3
___.__.T _ _._ _ _.__._ . .. ________ _ __. _
l
INFORMATIONAL FORMAL l TYPE "C" TYPE "C" j APRIL,1985 M ARCH, 1986 :
! VALVE LEAKAGE (SCFH) LEAKAGE (SCFH)
! 6A 0.89 1.09 l 68 1.37 0.54 6C 0.321 0.158 l t 6D 0.99 0.215 l
l 7A 0.34 0.084 i
j 78 2.778 1.183 7C 0.798 0.199
! 7D 0.306 0.088 i
l i
INITI AL TYPE "C" '
. TESTING euusivuo4s TABLE 3-1
l 1
TYPE "C" SEPTEMBER 2,1986
- VALVE LEAKAGE (SCFH)
I
! 6A 22 i
l 68 '40.2 j 6C 37.3
'l l : 6D > 42 I .
l 7A 30.3 1
. 78 > 42 i
- 7C 23.6 I
i 7D 16.7 i
SEPTEMBER, 1986 i
TYPE "C" TEST
- NMMSIVMD49 TABLE 3-2 3 ___
4.0 EVALUATION AND ANALYSIS P
4.1 Root Cause Analysis As discussed in Section 3.2, all eight balls showed consistent damage patterns of tungsten carbide delamination at the top and bottom of the flow holes. The location of the damage matched the position of the outer diameter of the sealing face of the seal ring at a small angle of rotation from the full open position.
This damage occured from spring loads on the seats, as the valves were not operated with line pressure.
An analysis of the mechanical interaction of the seat and ball follows, with a stress analysis, and metallurgical evaluation of the damage.
4.1.1 Mechancial Analysis of the Seat and Ball Figures 4-1 and 4-2 show the details of the seal assembly a
and ball with the valve in the full closed position.
The seat is held against the ball with sixty-ll5 pound springs, placed uniformly around the circumference, which exert a total force of 6900 pounds on the ball.
The seal assembly is free to move in the spool bore with a clearance of approximately 40 mils.
Because of the relationship between the seat outer diameter 29 e
and the flow hole inner diameter, when the valve is near the full open position, the seat approaches a point of instability. As shown in Figure 4-3, at this point, almost half of the spring loaded seat is not in contact with 'the ball and the seat may pivot near the edge of the flow hole and " rock" the seat area off the ball.
If this " rocking" were to occur, the entire 6900 pound spring load would be concentrated at the pivot point near the edge of the ball flow hole, which could result in a high contact stress on the ball.
An analysis follows of the mechanical forces on the seat and ball while opening and closing the valve.
valve closing When the valve operates from the open to closed position, rotation of the ball is clockwise as viewed from the top of the valve. Referring to Figure 4-4, there is a friction load on the slightly larger than one half of the seat in contact with the ball. This friction load causes the seat ring to tend to pivot in the spool bore around the seat packing in a counter clockwise a
direction as shown in Figure 4-4. The counter clockwise rotation of the seat ring due to friction assists in keeping the seat in contact with the ball while closing the valve. Therefore, forces do not exist to cause l c
30
instability or " rocking" of the seat off the ball while closing the valve.
Valve Opening When the valve operates from the closed to open position, rotation of the ball is counter clockwise as viewed from the top, and the friction forces are reversed. As shown in Figure 4-5, the friction load on the seat causes the seat ring to tend to rotate in a clockwise direction.
This rotation of the seat "into the hole" could result in instability or " rocking" of the seat off the ball while opening the valve.
Mathematical Model Figure 4-6 shows a simplified mathematical model of the seat ring loaded uniformly by the seat springs.
A calculation was performed to estimate the coefficient of friction required to initiate incipent rocking of the seat while the valve is opening. The calculation (still under review), determines the cof ficient of friction required to cause incipient " rocking" of the seat:
Required Coefficient of Friction = 0.4 31 .
Ongoing evaluation on available data for the coefficient of friction between the tungsten cerbide ball and the stellite seat is in progress. It is concluded, that the friction force caused the seat to " rock" off the ball during valve opening.
4.1.2 Stress Analysis From the preceding section, it was shown that the seat can " rock" off full contact with the ball during valve opening. This section provides an estimate of the resulting contact stress on the ball.
When " rocking" occurs, the full spring load of 6900 pounds is concentrated at two relatively small areas near the edge of the ball hole. Estimating contact area between the seat and ball of approximately 0.055 square inches, the contact stress is estimated as shown below:
Stress = 6900 lb. = 63,000 psi 2x (0.055) sq. in.
This concentrated stress is illustrated in Figure 4-7.
32
4.1.3 Metallurgical Evaluation Preliminary evaluation of tests performed both at Union Carbide and at the NMP2 Site further substantiates the tungsten carbide failure mode. From th'is testing it has been concluded that the tungsten carbide coating failed dc.e to high localized stresses and not due to improper coating application.
4.1.4 Root Cause The root cause of the tungsten carbide coating failure and resulting valve leakage has been determined to occur in the following sequence. This sequence is predicated on the original spring arrangement.
o Friction between the ball and the seat while the ball rotates from the closed to open position causes an i overturning moment on the portion of the seat remaining in contact with the ball. This moment tends to lift the seat.
o The point at which the rocking starts to occur causes the area of the seat in contact with the ball to.become i
very small, yielding high localized stresses on the tungsten carbide ball coating.
o The high localized stress combined with friction caused i
delamination and removal of the tungsten carbide coating.
33
.- _ . _ , . _ . -. . , , _ , . . , _ - . , _ . - ~ _ . - , _yy., . . , . _ _ , - . . . . . , . . , , , ,
o The removed coating becomes trapped between the seat and ball causing the removal of additional coating and scratching of the seats.
o The scratched seats do not seal properly when the valve is in the closed position thus failing the leakage testing.
A discussion of coating failures that occurred during the testing program is included in the following section (Section 5).
9 34
4.2 Description of Tests.and Results 4.2.1 Purpose Based upon preliminary evaluations of the valve failure a test plan was devised. The testing plan included a reasonable number of valve cycles to envelope one plant operating cycle, and included sufficient leakage testing to correlate between seat and through valve Type "C" leakage tests.
Three tests using different ball and seal spring combinations were run. Var.Jables in each of the tests were carefully controlled. The number of cycles, the type of leakage testing and the hydraulic actuator configuration were held as constants.
4.2.2 Test Parameters o our preliminary root cause analysis had determined that the seat rocked on the ball. A mathematical model indicated that our spring forces had to be revised such that counter balancing would not allow the seal ring to rock on the ball. Figure 4-8 shows the modified spring arrange-ment.
o The number of valve cycles was established to be 75 (75 openings and 75 closings strokes). This is the number of cycles conservatively estimated to the first v
35 -
- refueling of the plant, approximately 30 months from initial fuel loading. Table 4-1 provides a basis to this number of cycles.
o The valves were cycled a specified number of times followed by a Type "C" test. Table 4-2 (attached) shows the stroke and leakage test requirements for the test plan.
These requirements ensure sufficient through valve testing would be performed to establish a correlation between the two test methods, o For the purposes of this testing the MSIV actuator was temporarily configured to trip the valve closed in 3-5 seconds by venting the hydraulic cylinder to the hydraulic
. system tank. This venting was done through two solenoid operated valves. This is the configuration that will be used on the MSIV's when the final hydraulic actuator modifications are complete, thereby simulating operating conditions. The actuator modifications will be addressed
'in a separate report.
4.2.3 Test Combinations 4.2.3.1 The first combination consisted of:
c o New Ball (this was a spare ball that had never been installed) 36 .
o Modified Springs Section 5.1.2 discusses the modified spring arrangement, also see Figure 4-8 4.2.3.2 The second combination of ball and seat spring
~
configuration consisted of:
o Blended Ball - (see Figure 4-9)
This is a ball on which the damaged area has been uniformly removed and edges blended into the surrounding areas. The blended areas were approximately 8-10 mils lower than the remaining tungsten carbide coating. Blending was selected as an option to quickly make available test balls with good closed seal surfaces. As discussed in Section 3.2, the damaged area of the ball was located in the 1
$ open seating surface. In both blended balls l
)
tested the mating seal rings were lapped to j the ball in the closed position to assure l initial leak integrity.
I o Original Spring Configuration ;
I l
4 l
37
.- - ._. . . - . _ . .'_~ ----. - - - - - - - - - --
4.2.3.2 The third combination consisted of:
o Blended Ball o Modified Spring Configuration 4.2.4 Test Results o Leakage Table 4-3 (attached) presents the leakage test summary for the three test configurations.
1 In summary, test results show the blended balls behave in a similiar fashion with either an original or modified spring configuration. The modified spring configuration when used in conjunction with a new ball results in no damage to the ball or seats. .
o Visual Examination After testing was completed the three test valves were disassembled and inspected. Technical Consultants from the valve manufacturer (Crosby) and tungsten carbide coating applicator (Union Carbide) were part of the inspection team.
t 38
New Ball, Modified Spring Con'iguration -
There was no evidence of ball coating degradation due to seat rocking. The only indications were slight signs of polishing from the interaction of ball and seat.
i' These indications appeared as lines on the ball. The consultants concluded that this ball would have been capable of acceptable performance with more than 75 cycles. The leakage data also confirms the sealing integrity and continued capability.
Blended Ball, Original Spring Configuration -
4 The tungsten carbide coating adjoining the bler:19d area had been removed. Except for location and extent, the
~
~
damage appeared consistent with that observed on the original eight balls. The seats had trapped some of the removed material and the damaged area extended near the closed seating area of the ball.
, Blended Ball, Modified Spring Configuration -
1 l
The damage was slightly less than on the other blended ball. The revise spring configuration apparently helped
( to stabilize ghe seat, otherwise the tungsten carbide removed area was to the right (facing ball) of the original removal area.
39
WZ .
IO 8
- J&-
4 & y WZ-3 m-m 9 O'
e >- a J G1 O J2W 4 w (n G3 M O (n a
<u N h hIh E
- s
.5
?!
j 1 5 f a
\ ,
- 'N.
/ '\
/ N
/
u4....+
\
( J L
\
\s \ ,
i i!
40 ,
~
DET AIL " A" VALVE BODY SEAL ASSEMBLY SE AL SE AT -
CLEARENCE hwunr-- ENCAPSULATED SPRING SPRING RET AINER MULTIPLE SPOOL SE ALS
//
SPRING COVER ALL BALL & SEAT ASSEMBLY IN THE CLOSED POSITION NMusivMD24 MGURE 4-2
l s
CONTACT AREA !
I P l 1 l l* .
i e
)\) ,
l- : l
/ _
r L E l
' l r a ii r ll1 L li l L lll k lll SEAT APPROACHING '
POINT OF u INSTABILITY suusivuos2 FIGURE 4-3 P AGE 1 0F 3
'~~' -
SPOOL SEAT CONT ACT ARE A ON BALL AT GOVERNING POSITION - 12 FROM OPEN e J SPOOL SEAT CONTACT AREA BALL l
l l
l C
I t i
l i
BALL & SEAT CONTACT AREA 43 J
NMMSIVMD 61 FIGURE 4-3 PAGE 2 0F 3
s llI ll1 1 CONTACT l gl AREA l
1 - s l
/ \
STEAM FLOW l UPSTREAM
[l l l[DOWNST RE AM { _
l I
( '}s
\
s-_-
l lIll i
ll I
lIl N
BALL & SEAT CONTACT AREA NMMSIVMD58 FIGURE 4-3 PAGE 3 OF 3
BALL
/
/
/
I
/ I j
/ I f
/ B ALL ROT ATION [
[ l l 1r /
f
/ [
l 1
- i. l
[ /
/
RESULTANT
/
f
/
/ / FRICTION FORCE CONTACT AREA SE AT ROT ATION RE ACTION TO FRICTION FORCf ,
P ACKING VALVE CLOSING FRICTION FORCE
.3
& ROT ATION suusivuoss FIGURE 4-4 P AGE 1 OF 2
D G 2 AI N F O
OS 2 LO E G
L A NC P O 4 T G I
4 CI N E R
I R U R U G I
F D F m
N M
A E
R T
S N
W O
D l l I
/
l l l l l l IlIlIli N
O I
g T D g CA I
RO s /
A -
FL
- _ N l l N
- s
- N
\
G ,
^
Illlll ,l l l l 1 l I W
O .J_
L F u A
M E A R E T T S S P U
s 9
S D
M V
t S
y .
M M
N
BALL I
I I
ll l l
l l I l B ALL ROTATION j I
I lf T I
I I
l I
l I
l l I
- l. I
/ RESULTANT ;
FRICTION I !
f FORCE I f
I :
l l CONTACT AREA SE AT ROT ATION RE ACTION TO FRICTION FORCE P ACKING VALVE OPENING FRICTION FORCE 47
& ROTATION suusivuos7 FIGURE 4-5
X ct= 8.09 SEAL je -
SPRINGS
-- ~s (OTY 60)
BALL ' b
/ SPRING g DIAMETER g EQUALS 22.11 INCHES g 15 2
- f ' '
C.G. RIGHT
[ 1 1 SIDE ap i
0 @- - -
() ,
O A 4 1 C.G. LEF T <>
l SIDE
\ /
/
f
\
\ /
's ,-
A " '
, 2.63 Xc2 = 5.91 ,
FRICTION MOMENT SPRING LO AD = 115 LBS/ SPRING vv vv vv vv v v v v vv v vv v v v v vv vv v v c A e A
SEAT RING M ATHEM ATIC AL MODEL suussvuoso FIGURE 4-6
t
/
l l I ,
I I /'
~
I j l , I I /
r=, -
m --
qi, i i i /. I e
i
,, i l i i ,, i, l BALL VALVE OPEN B ALL V ALVE AT 12' l
l l A I
I l 1 s 5
a I
i !
l l l $
, I e
, E k I A
k Y l
STRESS DIAGRAM STRESS DIAGRAM 1 i
CONCENTRATED STRESS ON BALL ,
suusivuoss FIGURE 4-7
2 4 -115#
SPRINGS
,- ~s I
/ c REMOVE A f I
XA ADD 8-578 4 - 115 # - - _ _
SPRINGS SPRINGS w I V
N 1
J 2 4 -115#
SPRINGS MODIFIED SPRING FIGURE 4-8 NU MSIVM0 4 7
C D CLOCKWISE C OSE w J I
W//
>- BLENDED AREAS 8-10 MIL S LOWER THAN R E M AINING CO ATING V '
I
- - s MSIV BLENDED l 31 BALL NWMSIVMD 4 5 ' FIGURE 4-9 ,
~ - - - -- -
1 -
1 I
i
! VALVE ASSEMBLY STROKES 5 i
l PRE OPERATIONAL TESTING 5 i
POWER ASCENSION TESTING 5 I
i SURVEILLANCE TESTING 30 l MAINTENANCE CLOSURES 5 i
l CONTINGENCY 25 l TOTAL 75 l
l l
)
- NUMBER OF MSIV
! CYCLES suusivuoso T ABLE 4-L
_- 3 - _ _ _ _ __ __ _ __ _ _ _ _ _.- _ ____
- ACCUMULATED BETWEEN THROUGH i
CYCLES SEATS SEATS 5 X X
- 15 X X l.
) 25 X i
35 X l 45 X 1
55 X l
! 65 X
! 75 X X l X- INDICATED LEAKAGE AT STEADY STATE RECORDED.
i l TYPE "C" METHOD nuusmuosi T ABLE 4-2 4 .
I.
- ACCUMULATED NEW BALL BLENDED BALL BLENDED BALL l CYCLES MODIFIED SPRINGS ORIGINAL SPRINGS MODIFIED SPRINGS 1
BS TS BS
- TS** BS TS
- 5 4.8 2.3 11.2 5.1 18.4 ***
l \
j 15 4.9 2.9 51 5.5 59 -
4 J
i 25 4.7 -
132 10.9 100+ -
l j ; 35 4.4 - - - -
I I
i 45 4.2 - - - -
! 55 4.0 - - - -
i l
! 65 3.9 - - - -
! 75 4.4 3.2 - - -
1 NOTES:
- BETWEEN SEATS - SCFH
- THROUGH SEATS - SCFH TEST j
LEAKAGE l * *
- THROUGH SE AT TESTING NOT COMPLETED DUE TO POOR
< INITIAL TYPE "C" BETWEEN SEAT TESTING.
SUMMARY
nuusivuos2 T ABLE 4-3
5.0 Corrective Action and Technical Justification 5.1 Corrective Action 5.1.1 Recoated Balls Analytical and test results confirmed that blending of damaged areas on the balls would not solve the problem. Recoating is required to provide balls in the same condition as the ball used in the test configuration. The recoating process involves the following:
o- Removal by hand chipping of the old tungs ten carbide coating.
o Machine grinding of the ball to provide a uniform new surface for recoating.
o Local weld repair of Haynes 25 and/or base material and regrinding as required.
o Application of the tungsten carbide.
o Machine grinding to proper tolerance and finish requirements.
o Lap mating seats to new coating.
55
The above steps assure that each ball is returned to originally specified requirements identical to the ball used in testing, e
5.1.2 Modified Springs The modification for reducing the concentrated load to the ball surface consists of removing four 115 pound springs on the left side of the seal ring (facing ball) and adding eight 57 pound springs on the right side. The net sealing forces remain virtually constant.
5.2 Mechanical Analysis of Re-Coated Ball and Modified Springs .
The modified spring pack, as described in Section 5.1, will prevent the seat from " rocking" off the ball and will, therefore, eliminate the high contact stress which damaged the tungsten carbide coating.
As shown in Figure 5-1, the modified spring pack provides a counter clockwise moment on the seat ring which counteracts the clock-wise moment produced by friction when the valve is opening. A mechanical analysis similar to the analysis dicussed in Section 4.1.1 was performed for the modified spring pack. This analysis using a conservative friction of 0.5 verifies that the seal ring assembly will not rock on the ball.
Figure 5-2 shows a simplified mathematical model of the seat ring loaded by the modified spring pack.
56 ,
5.3 Stress Analysis Since the modified springs will prevent the seat from rocking, the contact area of the seat on the ball is considerably large.
A more precise analysis is being performed to determine the maximum contact stress. Preliminary calculations show a reduction of at least an order of magnitude compared to the original spring pack.
o conclusion o The modif.ied spring pack will prevent the seat from
" rocking" off the seat area of the ball.
o The seat spring loading will be distributed over a relatively large contact area on the ball.
o The estimated contact stress is below the allowable value.
o Damage to the tungsten carbide coating will be precluded, o Site testing has demonstrated that the tungsten carbide coating and seal ring are not damaged.
57
a.4 EVALUATION AND ANALYSIS 5.4.1 OPERATING CONDITIONS 5.4.1.1 Normal Opening and Closure During normal Plant Start-up, the MSIV's are opened with no differential pressure across the valves. valve opening time is approximately 5 minutes. It is permissable to open the MSIV's at rated temperature and pressure. However, pressure will equalize across the valves prior to reaching the critical valve open position.
During normal plant shutdown, the MSIV's remain open until reactor pressure reaches approximately 150 psi at which point they are closed and the plant put in shutdown cooling mode, valve closure time is approximately 20 seconds in this mode.
During monthly surveillance testing the valves are partially stroked closed. This test is performed with no dif ferential pressure.
5.4.1.2 Abnormal Conditions o Upset and Emergency closure of the MSIV's is assumed to occur during normal operation with the design flowrate through the valve 58
at 100% (3.79 x ig6 lb/hr) and Reactor Pressure Vessel (RPV) pressure and temperature equal to 1965 psia and 5520 F. Valve closure occurs within 3-5 seconds.
o Faulted The faulted operating condition for the MSIV's is defined by a rupture of the main steam line downstream of the MSIV's. Under this condition, the design flowrate through the valves is limited to 200% of maximum normal flow (7.58 x la6 lb/hr) by flow restrictors in the main steam line. The' valves will close against this maximum flowrate with the RPV pressure and temperature equal to 1965 psia and 5520 F. valve closure occurs within 3-5 seconds.
59
5.4.2 Steam Offects Figures 5-3, 5-4, and 5-5 show three positions in the sequence of valve closure or opening against steam:
- 1. Valve fully open
- 2. Valve partially closed (or open)
- 3. Valve fully closed Provided below is a summary of an evaluation of the ef fects of drag and pressure forces on the seats due to steam and resultant contact, stresses on the ball.
As seen in Figures 5-3, 5-4, and 5-5, there are two seats, an upstream and a downstream seat. These seats are affected differently by steam flow and pressure.
The following paragraphs discuss the effect of steam flow and pressure on these seats during valve opening and closing.
Steam Flow The flow on the upstream seat produces only a drag due to skin f riction since the seat is not protruding into the flow stream. The flow on the downstream seat, on the other hand, produces a direct impingement load on the downstream seat when the valve is partially closed. It should be noted that the impingement force on the downstream seat tends to rotate 60 .
the seal ring in a manner that would keep the seat on the ball and would assist in preventing the seat " rocking" that was previously discussed.
Steam Pressure When the valve is normally open, the differential pressure across both the upstream and downstream seats is zero because of a vent in the ball which equalizes the cavity pressure, as shown in Figure 5-6. When the valve is closed against steam flow and pressure, the differential pressure across both seats wil'1 increase. The pressure on the upstream seat will cause the seat contact stress against the ball to increase.
However, the differential pressure across the downstream seat will reduce the contact stress against the ball and the seat will actually move away f rom the ball at a dif ferential pressure of just over 100 psi. When the valve is opening or closing there is a non-uniform differential pressure across the upstream seat which will tend to rotate the seat ring in a manner that would assist in preventing the seat from " rocking".
valve closing It was concluded in Section 4.1.1, that seat " rocking" will not occur during valve closure. This conclusion is also
, applicable for valve closure against steam flow and pressure.
61 ,
Seat " rocking" is caused by a moment on the seal ring due to friction, which tends to rotate the seat "into the hole",
that is, in a counter clockwise direction when viewed from the top. Drag and pressure forces on the seat due to steam will change the magnitude of seat forces on the ball and, consequently, the mangitude of the friction on the seat ring.
However, the direction of the frictional force and resulting moment on the seal ring is governed by the relative motion between the seat and ball. While closing the valve, with or without steam, the ball is rotated in a counter clockwise direction, as viewed from the top, producing a clockwise rotation of the seat. Therefore, the seat will not " rock".
valve opening As discussed in Section 5.4.1, the valve maybe opened against steam pressure. For this case, the pressure will equalize quickly af ter the ball begins to rotate open, and the dif ferential pressure across the seat will be zero before the seat reaches the postion required to initiate " rocking". In addition, as discussed previously, there will be a non-uniform dif ferential pressure across the seats which assists in preventing " rocking".
' Therefore, the analysis and testing of the modified spring I
pack envelope conditions which will exist while opening the valve during operation.
62
Cnnelusion: Steam Effects o During valve opening and closure, the differential seat pressure assists in preventing seat " rocking" o Drag loads due to steam flow are small on upstream seats compared to seat spring loads o Drag loads on the downstream seat assist in preventing seat
" rocking" o The analysis and 75 cycle test with the modified spring pack are representative and applicable to full steam conditions, including drag and differential pressure across the seats 63
5.4.3 Temperature Effects The following discussion will address the suitability of the internals for the thermal environment and will focus on the effects of the difference in coefficients of thermal expansion between the ball and its coating and the effect of differential expansion between valve componets, o Ball o Kern Kraftwerk Leibstadt The exposure of these valves at reactor operating temperatures has not revealed evidence of cracking of the coating due to any temperature related phenom-ena. These valves are coated by the same process as the NMP2 balls, o 3-inch Ball Test 1980 A quench test was performed on a 3 inch diameter ball in 1980 by Gulf and Western Fluid Systems Division. This test was performed to show the acceptability of the application of the tungsten carbide coating application method. The method tested was the detonation gun process. This method employs high particle velocity to provide a coating 64
with low porosity compared to other application methods. Coating temperatures remain cooler, generally under 3000F, than by other methods, o Thermal Testing of Coating on MSIV Ball To provide additional assurance that the coating of the ballr will not crack at operating temperature a test was conducted on the ball from Valve No. 7D which had been removed for recoating due to damage described in Section 3.2.
Test
Description:
The ball was encased in thermal blankets and heated on the inside of the 21 inch diameter hole of the valve. It was heated at a rate which did not exceed 1000F per hour, up to a temperature of 5000F, while performing a series of tests to check'the effect of temperature on ball coating. The ball was then cooled at a controlled rate which did not exceeding 1000F per hour. Temperature was monitored at four locations both on the bore and by a calibrated surface pyrometer on the surface.
Inspection of the ball surface revealed no indication of coating failure due to the heating process.
65 .
4 o Other Valve Components o Seat Springs The seat springs are fabricated from Hastelloy C276 a trademarked alloy of the Cabot Corp. This alloy was chosen for its good high temperature creep properties and no degredation of the spring force is expected.
66 .
5.5 Evaluation of Tests with Blended Ball As discussed in Section 4.2.4, two tests were performed on balls with a 6 inch' portion of the tungsten carbide coating " blended" that is, ground away, in the area of the observed damage. One test was run with the original seat springs and the other with the modified spring pack.
Although both tests failed, due to further damage to the balls and subsequent excessive leakage rates, the tests further strengthen the root cause evaluation and corrective action. Also, both tests showed subtle dif ferences which are apparent considering the f ailure mechanism.
A blended ball actually causes the seat ring to be unstable as the ball is rotated either open or closed. The instability is strictly due to the fact that more than one half of the spring loaded seat ring is not supported on the seat area of the ball, regardless of the friction effect discussed in Section 4.1. The instability exists for both the test with the original spring pack and the test run with the modified spring pack.
The very rapid deterioration of the coating for both tests was due to the instability occuring during both the opening and closing of the valve. From observations of the two balls, and as discussed in Section 4.2.4, the ball with the original spring pack sustained more damage than the ball with the modified spring pack. In effect, 67
the modified seat springs attempted to counteract the instability, and although the damage was less severe than the ball with the uniform springs, the countering moment was not large enough to overcome the instability.
68
BALL I
I I
I I
I I
I I
I I
I B ALL ROT ATION j l
l V I l
I I I I I I I I l I
l l
l l
l l l RESULTANT FRICTION FORCE,/
hoITE O FRICTION FORCE l
l
' CO c ION ,UNTEesoe 1 0 0 E ET SPRING A A A A AA AA $$iFyEO l
Ap l l l DESIGN NOTE: SEE FIGURE 4-8 FOR SPECIFIC MODIFIED SPRING DESIGN.
VALVE OPENING-ROTATIONAL FORCES ON SEATS NMuslVMD68 69 FIGURE 5-1
X ct = 6.86
, w
^
j O
, - - ~
BALL ' b 7
/ SPRING g DIAMETER N EQUALS 22.11 INCHES l
' 15 2*
j '
i C.G. RIGHT , ,
[ SIDE i >
0 @- - -
A C.G. LEF T j t i l SIDE l ' l
\ / [
/
r
\
N /
A 2.63 X c2= 6.45 FRICTION MOMENT l1 v v v vvv v v v v v v v vv vv vv vv v v v v vv C A B A
MODIFIED SPRING MOMENT SE AT RING M ATHEM ATIC AL MODEL suusivues, 70 FIGURE 5-2
3 Y 5 WL E OLU R U
G L N FI F F E P
MVEO A L
~ >y y r
- E
- T A
% SV
\f-
,lgtl
!ehid hlq <
\ /
\
3
/ _
N .
- s e
,4i i,N lgIll td! p
\ "
(%'-
> b '
c1 1
7 D
M V
t 5
M W
N
c 4 5
W Y E O LDU RG L E L EI F FV LI A S YY#w+. - MA TO AV RL AC E
_ - _ P s4/ T
~ S w - ~- - _
" . V -
y
', - e x
/
i i L111 I.J;)I
'P .l l
\M2 N'mNNNE ,
i- A l_ _
\
E
^_ _
l n'
1 Al'
/ '
~
~ '
- - .P -
'i w zE -
- 1-Q 0 7
D M
V I
S M
M N
E E D 5 5
M LV S E R
A E A OUL G
I F
TVC S -
WY O
N O LL L
N .
_ F U F
b l
N h'
\
\ x x
x
~ -
k b
~
, N x
\
N 2 _
7 _
D _
M V
I S
M M
N
I V4 e L.L3 aa =
<O 8 co I C W 1-
>Z
_J W I >
Al G
I I a
I
/
- 's
\
I i t l
Z1 wo N
>I o
/ =- ======
Al s
5 Ib 9
5 2
h
6.0 ADDITIONAL CONFIRMATORY TESTING 6.1 Leakage Testing 6.1.1 Initial Type "C" Testing Post assembly verification of the valve requires a limited number of cycles (1 to 5 cycles) in order to properly align th'e ball and seats. Local Leak Rate Tests (LLRT)
(between-the-seats method) will be performed subsequent to each cycle as a proof test for proper alignment.
The final,LLRT will serve as the valve's formal LLRT (Type "C") in accordance with 10CFR50 Appendix J.
In preparation for fuel loading, the above procedure will be performed on one valve in each of the four steam lines. However, the valve actuators used to stroke these valves will not yet be modified with the revised hydraulic system. Therefore when these valves' actuators are modified and the valves are then stroked for timing and limit switch adjustment, another LLRT will be performed to ensure the required leak tightness.
The remaining four valves will then be assembled, but may not undergo initial stroking and LLRT until after their actuators have been modified.
75
In this case, these valves wi.'l be cycled for timing and limit adjustment in conjunction with the alignment stroking and LLRT. Again, the final LLRT with acceptable leak rate results will serve as the LLRT for plant operation.
6.1.2 Mid Cycle Test The next leak rate test will be performed during the mid-cycle outage tentatively scheduled for the Fall
'87, approximately 12 months af ter initial power operation.
All MSIV cycles will be logged in order to establish a total number of cycles on each valve prior to this mid-cycle testing. This outage has been scheduled in order to perform surveillance testing as required by the Technical Specification (e.g. , snubber inspections) .
It will also serve as the time period to perform all 18 month required surveillances (of which type "C" testing is a part) prior to continued power operation until the first refueling outage tentatively scheduled for the Spring of 1989.
It should be noted that leak rate testing of the MSIV's will be accomplished with the between-the-seats testing as the preferred method. This method has been demonstrated to be a conservative test compared with normal through-the-valve test method. However, should this test method produce unacceptable test results (due to 76 -
its highly conservative nature), NMPC may emplov the through-the-valve test method in order to satisfy Technical Specification requirements. '
6.2 DEVELOPMENTAL TESTING 6.2.1 Test objectives Niagara Mohawk is initiating a developmental program for both the valve and actuator of the MSIV's. This program includes plans to completely review the existing valve and actuator design, perform additional analyses of specific design features, investigate alternate materials and test a full scale prototype of the valve and actuator.
The prototype testing will include operation of the valve under steam flow conditions. The principal objective of the developmental program is to demonstrate the ability of the MSIV's to perform their intended function beyond the first operating cycle. The program will also identify any changes in design or materials which will improve the long term reliability of the valves.
6.2.2 Test Organization A task force, under the direction of Niagara Mohawk, is being organized to manage the developmental program.
The task force will include representatives from Crosby 77
Valve & Gage Company, General Electric, MPR AsLociates, Stone & Webster Engineering Corporation, and Westinghouse Electric Corporation. These representatives will insure access to a wide range of experience and technical expertise on valves, materials, and mechanical design. This experience and expertise will be used to the extent necessary in defining and implementing the developmental program.
6.2.3 Testing Format and Schedule The developmental program is being structured into 3 general phases. Phase'1 will include a complete review of the existing valve design, materials and operational experience; a thorough re-analysis of problems which have occured; and identification of any additional areas of concern. Phase 2 will involve identification of alternate design features and materials, evaluation and testing of proposed changes, and the selection of specific chan9{es for implementation. Phase 3 will be the detailed des $gn, procurement, f abrication, qualification and installaticp of any identified valve and/or operator modifications. Organization and definition of the developmental program is in progress at the present time. Initial steps have been taken to locate and assemble component [1and facilities required for full scale prototype testing. I Preliminary reviews of the valve and operator relative to long term operation have started. Completion 78
of the Phase 1 reviews and analyses is scheduled for January 1987. Depending on the availability of certain critical materials, a valve prototype should be available for initial testing by January 1987. Testing of a complete valve and operator assembly under simulated operating conditions is scheduled to commence in April 1987.
Completion of the entire Phase 2 effort is scheduled by December 1987. The Phase 3 design work should begin October 1987 with materials for modifications delivered by January 1989. Actual installation of modifications is scheduled for the first refueling outage which is currently scheduled for Spring 1989.
7.0 CONTINGENCY PLANNING In the unlikely event, that the additional testing or developmental testing programs described earlier indicate that leakage or reliability characteristics of the ball valves are not satisfactory, NMPC is actively developing two contingency plans.
7.1 Leakage Control System one plan would provide for the installation of a leakage control system to be used in conjunction with the MSIV ball valves to ensure that any leakage past the seats is collected and discharged in a monitored, controlled manner to the 79
environment through the Standby Gas Treatment System and main stack. This Leakage Control System is being designed using the guidance contained in Regulatory Guide l.96. Preliminary schedule information indicates that the leakage control system could be installed and operable by late January 1987 at the earliest. However, this schedule assumes all activities
- are expedited and has no contingency for unknowns.
7.2 Y-Pattern Globe Valves The second contingency plan involves the procurement of fully qualified Y-pattern globe valves for installation in place of the ball valves. The globe valves would also be provided with a leakage control' collection system meeting the same requirements as the system described above for use with the ball valves. Preliminary schedule information indicates that the globe valves and their requisite leakage control system could be installed and operable to support fuel load by late January 1987 at the earliest. Again, the schedule assumes expedited activities and has no contingency.
8.0 CONCLUSION
S NMPC believes this report demonstrates that the current MSIV design coupled with the modified seal spring arrangement is reliable for a minimum of one plant operating cycle.
80 .
=. _- __
NMPC believes this report demonstrates that the problem resolution presented herein assures the MSIV's will remain leak-tight through the first plant operating cycle. Niagara Mohawk is committed to continued testing and contingency programs to provide further assurance of reliable service throughout plant operating life.
/
1 0
81 -
_ m e * - 6 J -
O n L J L-- 2
-, ---- ,a: -- -h4.+
9.9 APPENDECIES
9.1 DESCRIPTION
OF MATERIALS OF BALL CONSTRUCTION 9.2 MSIV CHRONOLOGY 9.3 MSIV LLRT PROCEDURE
SUMMARY
9.4 BALL POSITIONING (THRUST WASHER) t e
82
APPENDIX 9.1 Description of Materials of Ball Construction o The bore core consists of a type 316L stainless steel casting made to SA351, GR. CF8M.
o A nominal 0.090 inch layer of Ilayn e s 25 is deposited by welding. Haynes Alloy No. 25 is a cobalt-based superalloy with a nominal chemical composition of approximately 50%
cobalt, 10% nickel, 20% chronium and 15% tungsten. It has excellent corrosion resistance. It was deposited on the 316L MSIV balls using-the submerged arc welding process.
The hardness of this as-deposited alloy is approximately 80 to 90 on the Rockwell B scale.
o The tungsten carbide coating (0.008" to 0.010") is deposited on top of the ground Haynes 25 layer using the detonation gun process. Detonation gun process is used because it imparts minimal heat to the workpiece (the ball in this case) and therefore presents the least risk of distortion.
It is a process specifically developed for the deposition of hard, wear-resistant materials such as tungsten carbide.
Typical applications include jet-engine seals and aircraft compressor and turbine blades. The tungsten carbide coating has a Union Carbide coating designation of LW-5 and a nominal composition of 68 W, 22 Cr, 5 Ni, 5 C.
83
--- v- ~ -w- , - -- , ,,g
o As used on the MSIV seats, Stellite 6 is a cobalt-based weld deposited hardfacing alloy with a nominal chemical composition of 1.1% carbon, 28% chromium, 4 % tungsten and balance cobalt. Its microstructure consists of a network of chromium and tungsten carbides in a cobalt-chronium matrix. It was deposited on the 316 stainless steel spool seat using the gas tungsten arc welding process. The average hardness of this as-deposited alloy is approximately 35 on the Rockwell C scale. However, the average hardness may not be completely indicative of the alloy's behavior in wear situations since the carbide microconstituents are extremely hard.
84
Appendix 9.2 MAIN STEAM ISOLATION VALVES CHRONOLOGY August 1976 8" Prototype Ball Valve Testing October 1977 Purchase order for MSIV's was placed with the Fluid Systems Division (FSD) of Gulf and Western (G&W). The Energy Product Group (EPG) of FSD has merged with EFCO Ball Valve Company
,EBV)
( who actually manufactured the valve.
Jcnuary 1979 Topical Report submitted to NRC.
February 1981 Valve bodies arrived at the jobsite.
March 1983 Hydraulic actuators arrived at the jobsite.
November 1984 Crosby Valve and Guage Company (Division of Geosource Incorported) has announced the acquisition of the FSD of G&W.
December 1984 Applied corrosion resistant cladding to spool bore area.
April 1985 All MSIV's passed " informational" Type "C" leak tests.
85 -
August 1985 Main Steam Line & RPV hydrostatic test and system flushing.
Dscember 1985 Installed actuators on valve bodies.
February 1986 MSIV Pre Operational Testing began on site.
March 1986 All MSIV's passed " formal" Type "C" leak tests.
March-August 1986 All MSIV's stroked for actuator testing.
Scptember 1986 All MSIV's failed Type "C" leak tests.
I e
86 ,
APPENDIX 9.3 MSIV Local Leak Rate Test Procedure Summary o Pressurize line or valve body to 40 psig with air, o Vent opposing test boundaries, i o Allow sufficient time for air in-leakage readings to stabilize.
o Record in-leakage value in SCFH.
The leakage test schematic and valve line-ups are shown on the following pages.
i I
e
! 87
PR IM AR Y L O l -40 PSIG CONTAINMENT f
u y ITHROUGH VALVE TEST)
M AIN STE AM L INE RPV u - VENT A VENT B VENT C l
INBOARD OUTBOARD
- M SIV MSIV TURBINE
$ BODY C AVITY > < BODY C AVITY DR AIN L INE DR AIN L INE 40 PSIG > < 40 PSIG IBETWEEN (BETWEE N SE AT TEST)
SE AT TEST)
J
! MSIVTYPE "C "
l TEST SCHE M ATIC APPENDIX 9.3 FL CH.M F.WC1. 0 6
APPENDIX 9.3 TYPE "C" VALVE LINE-UPS o Through Valve Test - Inboard
- 1. Main Steam Line Plug in Place
- 2. Vent "A" Closed
- 3. Inboard MSIV Closed
- 4. Inboard Body Cavity Drain Closed
- 5. Vent "B" Open o Through Valve Test - Outboard-
- 1. Main Steam Line Plug in Place
- 2. Vent "A" Closed
- 3. Inboard MSIV open
- 4. Inboard Body Cavity Drain Closed
- 5. Vent "B" Closed
- 6. Outboard Body Cavity Drain Closed
- 7. Vent "C" Open o Between Valve Seat Test - Inboard or Outboard
- 1. MSIV Closed
- 2. Vent "A" & "B" or "B" & "C" Open as Applicable 89
APPENDIX 9.4 Ball Positioning (Thrust Washer)
Galling of tih e thrust washer during valve movement is due to rubbing of the 304 SS thrust washer with the 316L SS ball. On some valves, the galling was observed 3600 around the thrust washer while on others it was evident only on a portion of the circumference.
The galling was not considered to be excessive considering the normal characteristics of a loaded stainless steel to stainless steel interface. Since the thrust washer is at the bottom of the valve assembly it is also possible that tungsten carbide chips worked their way into the thrust washer area and initiated galling.
In order to eliminatee this to galling a new thrust washer material has been selected.
The material selected is a bronze bearing material. This material is normally used in this type of load bearing application and is not susceptible to galling. The bronze material also will reduce frictional forces on ball movement.
Another critical attribute of the thrust washer is its thickness the washer must be sized to allow thermal growth of the ball without interfering with proper seat and ball interaction. In the cold condition the thrust washer is sized to be twenty to thirty mils 90 ,
smaller than would be required to align the ball bore to the valve body bore (see Figure 9.8-1). This undersize allows for thermal growth and permits the seat to self align with out interference with the body bore. If sized thinner the seat could bottom out on the body bore, if sized thicker the seat could top out on the body bore in the hot condition. In either case seat or ball damage could occur.
91 ,
T =
(1/2 A+B) -
(1/2 C + D)
(0.020" TO 0.030")
l I I I
I I , ,
/ l BALL _,______
f 4k l Ak I
4 0.020" I TO 0.030" l A -
. _ _ C BODY JL BALL
V l , i l 9 l Ak
, l T I
IhLVE THRUST BODY WASHER
! THRUST WASHER SEAT THRUST WASHER SIZING i
NMMSIVMD41 APPENDIX 9.4