ML20154R385

Revised Insp Results of San Onofre Nuclear Generating Station - 3 MSIV - 8205

ML20154R385
Person / Time
Site:	San Onofre
Issue date:	08/24/1988
From:	Chiu C, Herschthal M SOUTHERN CALIFORNIA EDISON CO.
To:
Shared Package
ML13304A365	List: ML20154R385
References
NUDOCS 8810040241
Download: ML20154R385 (40)

Text

_. _ _ _ _ _ _ _ _ - - _ _ -

1 INSPECTION RESULTS OF l SONGS 3 MSIV 8205 - -

I

!, AUGUST 24,1988 l

l l l  !

! San Onofre Nuclear Generation Station  !

i l

i I l .

1 i

i fS 5 Southern California Edison l

SBA 1882!SS8kke' O

i I

i i

INSPECTION RESULTS OF SON 65-3 MSIV 8205 (Revised) i August 24, 1988

. I 4

f Principal Investigators:

C. Chiu M. A. Herschthal ,

i i

a contributors:

i J. Johnson (WKM) l i  !

] J. Brinkley (WKM)

D. A. Niebruegge (SCE)

N. Quigley (SCE) L

+

3  :

)  !

t l  ;

1 4

e l

f  !

l

- _ _ _ . . . _ . _ . _ . - - . . _ _ _ . . _ _ . = _ _ _ _ _ _ _ __ . - _ . _ . . _ _ _ _ . . _ _ . . _ - . _ _ . _ _ _ _ , _ _ _ _ . , _ - _ _ . . . , _ _ . _ . _ _ _ . . . _ , . . . -

l i

! CONTENTS  ;

M  !

3 INTRODUCTION 1  !

INSPECTION RESULTS 2 l

DISCUSSION OF INSPECTION RESULTS 5 i

4 8

) POSSIBLE FAILURE SCENARIOS  ;

METALLURGICAL EXAMINATION 10 OVERTORQUING TEST PERFORMED BY SCE 11 ,,

) i

) STRESS ANALYSIS 13 l i

l REPAIR OF MSIV 3HV-8205 ,

13 -

1

! JUSTIFICATION FOR CONTINUED OPERATION AND RECOMENDED ]

INSPECTION PLAN 14 .

l BOROSCOPIC INSPECTION IMPROVEMENTS 16  !

l  ;

1 4  :

! STR0KE TIME REDUCTION 17  :

i  !

APPENDIX A STRESS ANALYSIS 19 i

^

4 i  !

l i I

4

11 l

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

i!

LIST OF TABLES AND FIGURES EASA Table 1 Schedule of MSIV Studies and Inspe'ctions 22 Figure 1 A Close-up View of the Chamfer for the North Closing Guide Rail 23 Figure 2 A Close-up View of the Chamfer for the South Closing ,

Guide Rail 24

! Figure 3 Various Views cf Closing Guide Rails and Valve Skirt 25 Figure 4 Various. Views of Closing Guide Rails and Valve Skirt 26 t

Figure 5 A Close-up View of the Chamfer for the North Opening 4 Guide Rail 27 i

Figure 6 A Close-up View of the Chamfer for the South Opening j Guide Rail 28 1

Figure 7 The Location of Three Broken Capscrews 29 l

Figure 8 Fracture Face of the Broken Capscrew Stud 30 Figure 9 Magnetite Build-up on the Fracture Face (1400 X 4

Magnification) 31 1  ;

Figure 10 A Close-up View of the Galling Marks or,the North Aru l Shoe 32 1 -

! Figure 11 A Close-up View of the Gtiling Marks on the South Ars Shoe 33

}

Figure 12 A Distant View of the Galling Marks on the Arm Shoes 34 l

) Figure 13 Borescope Probe Penetration points 35 Figure A-1 Forces Resulting from Interference 36 111 l

-. -- . . -_ a

4 l

. INTR 000CT10N i

j As stated in the companion report, "Root Cause Analysis of the Lev-R-Lock and 3

Guide Rail Interaction Problem for SONGS MSIVs," SCE has concluded that it is  !

unlikely that SONGS MSIVs would experience the failure mechanism that can shear l

i two gate skirt assembly guide rails (the guide rails interacting with the l 0 t lev r-lock arm during closing) that occurred in one of the Waterford-3 MSIVs.

]

This conclusion is based on, among other supporting evidence, a dynamic impact i l

i analysis and fiber optic inspection results on Unit 3 MSIVs 3HV 8204 and s

SHV 8205. The dynamic impact analysis reveals that one of the key parameters  !

determining the magnitude of the shearing energy is the stroke time of the valve I in the close direction. Since SONGS MSIVs stroke approximately two and one half l l i j times slower, it is very unlikely that they are subject to the failure mechanism j l experienced by Waterford-3 MSIVs.  !

a on June 20, 1988, SCE was requested by the NRC to disassemble and inspect Unit 3 4 3HV-8205 to see if it has experienced the Waterford 3 failure mechanism. .The

} inspection results are documented here as an addendum to the root cause l

)

analysis.  !

l l

Since the inspection confirms the main conclusion of the root cause analysis, the root cause analysis together with the safety evaluation included in the root f1 cause analysis (which constitute the bases for JC0, Justification for Continued Operation) remains valid.

1

] The inspections also revealed that three capscrews at the bottom of one of the two segment skirt assembly guide rails were broken. This guide rail is

}

j hereafter referred to as the north segment, or north opening guide rail, and the 1

) other is the south segment guide rail. This failure mechanism is believed to be different from the mechanism that has resulted in shearing of the two gate guide

-1

e rails of the Waterford-3 MSIV. This failure mechanism is analyzed to be much less damaging than tb "Waterford-3 failure mechanism" and it is self-limiting.

The root cause analysis for this failure mechanism is also documented in this 4

addendum, t p

i 18$PECTION RESULTS i  !

) Unit 3 MSIV-8205 was disassembled for visual inspection in July,1988. The 1  :

inspection results are documented below: I i

(1) The two gate guide rails were inspet+.ed. The 45' chaefers, where the lev r-  !

lock arm and the guide rails interact during valve closing, do not show f

. signs of galling., However, wear marks, a result of the lev-r-lock j arWguide rail chaefer contact during the close cycle, are evident. Figures l 1 and 2 show close-up views of these chaefers. l 2 .;

i

, (2) The 18 capscrews ov} the two gate guide rails were inspected. All of them f appeared to be tight and all of the stakes on the capscrew heads wert intact. The guide rails are tightly held against the valve skirt plate. l j Figures 3 and 4 are close-up views from various angles of the guide rail l

l Capstrews.  !

4 i

t (3) The two segment guide rails were inspected. The bottom 45' chamfers show l

]  !

j signs of galling. The galling occurs on the one chaefer corner which I

t j consists of the transition from the bottom end of the rail to the chaefer f face. The length of the area which has displaced metal is only

approximately 1/4" long. Figures 5 and 6 are close-up views of galling j marks for the north and south chaefers.

1 1

.z- .

(4) The capscrews on the opening guide rails were inspected. Three of the 18 l capscrews were broken. All the broken capscrews are located at the bottom of the north opening guide rail. There is approximately a 0.050' gap between the north segment guide rail and the skirt plate at the edge of the l rail where the three capscrews are missing. Figure 7 shows a close-up view of the empty capscrew holes, i

(5) The three empty capscrew holes were inspected for corrosion and  !

l misalignment. No galvanic corrosion pits were found. The misalignment  ;

between the guide rail holes and the valve skirt sockets ranges between 5 to [

10 mils. This is considered insignificant. .

f r

)

(6) All three broken capscrew heads were found at the bottom of the valve cavity. (Approximately 30 etpscrews were broken at Waterford 3. All of  !

l these except one of capscrew heads were found in the MSIV valve cavities or l the turbine stop valve strainer. One capscrew head was never located.) The ,

i broken capscrew heads were visually inspected by SCE and WKM engineers at  ;

the time they were retrieved from the bottom of the valve cavity. No  :

corrosion pits were found. i

- I (7) The studs of the broken capscrews were removed from the capscrew sockets.

They were able to be removed from the skirt plate by hand and visually inspected. No corrosion pits were found.

(8) The fracture faces of the broken capscrews were visually inspected esing a 32X light microscope. General deformation around the fracture face is I evident for all three broken capscrews, indicating the failure mode is ductile overload.

(9) The fracture paths of all three broken capscrews are at the bottom of the capscrews hexagonal hole. Figure 8 shows a close-up view of the fracture face. Significant magnetite build-up was found on the fracture faces of all three screws. This indicates that the failure did not recently occur.

Figure 9 shows this build-up with a 1400 x magnifit.ation.

(10) The total depth of the capscrew hexagonal hole (or keyway for the Allen head wrench) was measured. The depth is approximately 3/8".

(11) The fourth capscrew from the bottom on the guide rail with broken capscrews was found loose; that is, it can be moved by hand. The stake marks on the head of this capscrew appear to be slightly shallower than those for intact capscrews. It was removed to see if it contained any incipient cracks. No flaws were found using a 32X microscope inspection.

i (12) The shoes of both lev-r-lock arms were examined. Galling marks were found on the tops of both shoes (The shoe tops may contact the segment guide rails during the valve opening cycle). The shoe bottoms were smooth (The

shoe bottoes may contact the gate guide rails during the valve closing i l

l cycle). Figures 10 and 11 are close up views of the galling. Figure 12 is 1 1

a distant view of the galling.

(13) The dimensions of the gate, segment, guide rails, lev-r-lock arms, and distance between valve seats were measured. All dimensions were within the

vendor's specifications.

i l

l 4

I (14) The length of the north and south arms were measured. It appears that the north are is about 1/16" shorter at the point of contact with th6 guide rail. However, due to the measurement difficulty, there is some uncertainty associated with this data.

(15) The lev-r-lock arm ear and the segment slot were inspected and showed no signs of excessive wear.

(16) The valve seats were inspected visually. There are several areas on the gate and opening side valve body seats that show signs of wear marks. The maximum depth was measured to be less than 5 mils.

(17) The gate and segment back angles.were inspected visually. Several wear marks were observed.

DISCUSSION OF INSPECTION RESULTS Based on the visual inspection results, it is reasonable to conclude the following:

1) The fracture of the three broken capscrews seems to be a result of interference with the lower edge of the chaefer and the top of the north arm shoe. Figure A-1 illustrates this interference during valve opening. There are galling marks in the area of interference on both the shoe and the chaefer of the guide rail.

2) The galvanic corrosion, if any, is insignificant. This is supported by the absence of any pitting on the broken capscrew threaded areas inspected by fractography.

- - - - A

3) The depth of the capscrew socket is 0.375". The minimum depth is specified by the ANSI 18.3 standard to be 0.220'. Even though there is no maximum depth limit, this 0.375" depth seems to be excessive. To evaluate whether or not these capscrews had experienced overtorquing induced cracks, a torque test was performed. The results are presented later in this report.

Note that the Interim Root Cause Analysis stated the following (refer to page 16 >

in the Interim Root Cause Analysis):

"INSPECTION RESULTS" "The key findings of the boroscope inspection on the Unit 3 MSIV are summarized below.

)

'(1) One broken capscrew was found in the cavity of SONGS-3 MSIV 3HV-8205. The two guide rails (one of the two segment skirt ratis and one of the two gate  !

skirt rails) were in place. The rails were found firmly attached to the skirt plates. They did not exhibit any separation that was observed on the 1 LP&L MS!Vs. Only four capscrews could be inspected inplace. They were ,

, found to exhibit no signs of elongation, deformation, or looseness.

1 j

s

"(2) No broken capscrews were found in the body cavity of $0NGS-3 MSIV 3HV 8204. ,

j The two guide rails that could be inspected (again, one of two segment skirt rails and one of two gate skirt rails) were also in place.

"(3) The chaefers of both upstream and cownstream guide rails for both SONGS-3 MSIVs were inspected for impact marks. Only the chamfer on the downstream,

! or gate guide rails, in MSIV 3HV-8205 has galling marks. The top edge of the chamfers have been rounded by impact and some metal has been rolled l

t 1

l over. This metal roll-over and a relatively high contact stress during l

l

impact could be responsible fcr the observed galling marks. l

. i

' i

. f

*(4) The shoes of the lev-r lock arms for both valves were inspected for impact l j or galling marks. No visible galling marks were observed for the surfaces i that were observed."  !

I l

' i Some of the statements in the Interim Root Cause Analysis Report are refuted as a result of the disassembly and inspection of MSIV 3HV-8205. The are identified I and explained as follows:  !

I -

l (1) Disassembly of the valve revealed that the gate guide rail chaefers were not  !

I i

j galled as concluded after the boroscopic examination. The marks on the  ;

chaefer face gave the appearance of galling as viewed using the boroscope

]

even though the chamfers were found to be smooth with no furrowed metal.

l The corners of the chaefers were no longer sharp but were not galled. The j l wear of the chaefer corners was concluded after the boMscopic examination.  !

l The as found condition of the chaefer was in better condition than had been j determined as result of the remote inspection.

l l I

j Future boroscopic inspections will be improved based upon the. knowledge acquired as a result of these inspections. The improved inspection 5 techniques are detailed 1tter in this report. A more careful remote 1

{ inspection, including profile as well as head on views of the chaefer face, I

j will provide information required to make adequate determination of the i

i material condition of the guide rati in the vicinity of the chaefer.

j (2) The galling on top of the lev r-lock are shoes, the segment guide rail galling, and the slight separation of the north segment guide rail was not i

l -

1 .

concluded / observed after the boroscopic examination. The damaged areas were obscured from view when inspected in place with the MSIV in the closed position. When the valve is lev-r-locked closed, the top of the shoes are swung forward (upstream) and are tucked under the ends of the segment skirt

~

assembly guide rails. The galled areas of the shoes were obscured by the rails. It is possible that galled areas of the rails may not be seen even if the shoes are not in their closed position due to the relative position of the galled surface (downward) and the boroscope probe (also downward).

4 l l

Future inspections will be made with the MSIV in the not fully closed position. This is detailed later in this 'eport, f Item (3) of the Intermin Root Cause Analysis INSPECTION RESULTS above infers that all of the guide rails can be inspected using the boroscope. The item should

have stated that one segment and one gate guide rail of each of the two SONGS-3 l MSIVs were inspected. This would make item (3) consistent with INSPECTION l

RESULTS items (1) and (2) in the Interim Root Cause Analysis. Not all of the

, guide rails can be inspected because the penetration that is used for probe i insertion is located on the surface of the bonnet, directly between the segment i

and gate skirt plates on one side of the valve. This penetration location for SONGS MS!V 3HV 820$ is snown in figure 13. Only the north segment and gate guide j rails of this valve can be examined using the boroscope, i

l POSSIBLE FAILURE SCENARIOS l l

I Several possible scenarios that result in overload fracture of the three

]

l capscrews on an opening guide rail are hypothesized. Based on the inspection

results, only one of the hypothesized scenarios is considered likely. This can l l

i 1

l

not be refuted by any evidenco or observation collected so far. This scenario  :

is stated below.

'At the beginning of valve opening, the frictie on the back angles of the segment and gate prevents the assembly from unwedging, or being "unlev-r-locked." As a result, as the assembly m ves upward, the top of i the lev-r lock arm shoe comes in contact with the bottom edge of the guide rail chamfer. In the first few opening strokes, the resultant tensile force is preferentially imparted to the north guide rail, either because the north l

are is shorter or because the valve skirt is not squarely installed. In .

l addition, the contact area is limited to the lower sharp chaefer edge.  !

This results in sufficiently large stresses such that galling between the I guide rail and shoe occur. The galling results in adhesion and generation of large resultant forces. Consequently, the first few bottom capscrews on i a

l the north guide rail fail by tensile overload. After metal is removed from i

! 1 the bottom edge of the north guide rail chaefer as a result of a few valve l

opening cycles, the interference force begins to be shared by both north and j south guide ratis. The galling process then occurs and removes the sharp l corner on the south guide rail. Once the corners of the chaefers are removed, the contact area between the guide rails and the lev r lock ara l shoes are significantly increased resulting in lower stresses (even though the surfaces have been galled) and discontinuation of the galling and adhesion processes, i

It should be noted that the skirt assemblies will tend to be "self. l l

j aligning." Each skirt assembly is made up of two guide rails bolted to a l plate. The plate is fabricated with a hole with a diameter slightly larger than the valve seat diameter. The skirt assemblies are placed into the l 1

valve body and laid around each seat. The gate and segment assembly is 1

) placed into the valve, between the seats and skirt plates, and between the J

guide rails. The skirt assemblies are free to rotate around the seats within the clearances in the valve body as well as the clearances between the rails and the gate and segment assembly. If one lev-r-lock arm is longer than the other, the shoe connected to the shorter arm will contact the guide rail first. This will tend to place a torque on the segment skirt assembly which will rotate within these clearances in an effort to distribute the load equally between both lev r lock arms. Although the clearances in relationship to the difference in lev r-lock are length could not be correlated, a visual inspection of the alignments and clearances were meticulously performed by SCE and W-K-M engineers. The as left installation was found to be acceptable.

Because the tensile load is now shared by two guide rails and the friction coefficient decreases as the surface gets recontoured after a few instances of interaction, the damaging mechanism stops and no more capscrews fail."

This hypothesis is supported by the fact that a passivated iron coating was found on the fracture surface of all three failed capscrews which indicate that the failures had not recently occurred.

6 The three broken capscrew heads and studs were metallurgically examined. One set consisting of a broken capscrew head and stud was sent to WKM for metallurgical analysis. Two broken studs were sent to Truesdati t.aboratory for metallurgical examination. Two broken capscrew heads were examined by SCE for material composition. The results of this examination are documented here.

1) The saterial of two broken capscrew heads was determined by SCE to be within the specifications of ASTM A193 Gr. 87.

' i

2) The material of two broken studs was determined by Truesdail to be I t

within the specifications of ASTM Alg3 Gr. 87.

3) The material of one broken capscrew head and one stud was P* ermined by WKM to be within the specifications of ASTM Alg3 Gr. B7. ,

i i

l i

?

1 4) The hardness of the material for the capscrew kept by WKM was j i determined to be 302 Brinell. This hardness translates into a tensile '

j 1 strength of 146 ksi. This is greater than the minimum tensile stress i of 125 kst for ASTM Alg3 B7 bolt.

The hardness of the material for a capscrew sent to Truesdail

) 5)

I Laboratory was determined to be 362 Brinell. This hardness translates

}

into a tensile strength of 177 ksi, according to ASTM Specification j A370.

l

6) The fracture faces for all three broken capscrews have significant l magnetite build-up, indi: sting that the fracture did not recently I

j occur.

I l

7) The general deformatten around the fracture face suggests that the

(

failure mode is ductile overload. However, due to the magnetite build-up on the fracture faces, detailed microscopic examination for dimple i marks and fracture face characterization is not possible.

k 1 I OVERT 0ROUING TEST PERFORftED BY SCE i I ,

j The deep capscrew socket, even though its dimension does not violate any capscrew I specification as documented in ANSI Standard 18.3 may have contributed to the l

l 4

1

! 1 j failure of the three capscrews. One scenario is that these deep socket capscrews were broken during initial assembly by overtorquing. To evalud. .:h9her or not l

this scenario is valid, a capscrew was removed from the north segment guide rail  !

and was bench tested. A test bench was constructed with identical dimensions of l

the counter sink in the guide rail. The capscrew was torqued with an

) appropriate lubricant at various torques to see if it developed any incipient j t

?

l cracks or permanent deformation. The results are summarized below. l 1 l i  !

1) Torqued to 150 ft-lbf - no incipient cracks; no plastic deformation

) the Allen wrench starts to defors, no I

i 2) Torquod to 180 ft-lbf - l incipient cracks; no plastic deformation f I

3) Torqued to 250 ft lbf - the Allen wrench completely deformed; no l incipient cracks; no plastic deformation '

4 I r j 4) Torquod greater , l' l than 250 ft lbf - test discontinued because of tool l l

deformation.  !

) Note that the capscrew was originally installed by W M with a torque of 150 ft-lbf. According to the test results stated above, this installation torque I will not cause incipient cracks. As a result, it is reasonable to dismiss the j j failure scenario that these three capscrews (even if the capscrew socket is  !

0.375" deep) were failed due to overtorquing.  !

1 1 i i i

! l 1

STRESS ANALY1Li 4

Based on a stress analysis documented in Appendix A of this report, the maximum tensile force generated by the interaction between the lev-r-lock arm and the guide rail is estimated to be approximately 95,300 lbf. The interaction is assumed to occur with a galling process between two interacting parts. The tensile force needed to fracture a capscrew is estimated to be 32,800 lbf, also considering the co existence of the shear force. 'Since 32,800 lbf is less than 95,568 lbf it is reasonable to conclude that the bottes two to three capscrews  !

J of the only interacting guide rail (north guide rail) will fracture.

) Once the interfering metal was rereoved from the chamfer of the north segment s

guide rail after the first few times of interaction, the friction coefficient of interactioil decreased and the total load started to be shared by both the north  !

l and south guide rails. As a result, the maximum tensile force is reduced to a i 1

level of approximately 16,730 lbf. Note that this maximum tensile load is no l l longer capabic cf fracturing capscrews. j REPAIR OF MSIV 3HV-!(Q1

. i i  !

! The galling marks on the lev r lock are shoes and guide rails were removed by I

j grinding. The chaefer angles were decreased slightly by virtue of the grinding l operations. The chaefer angles and shoes were polished to a better than 63 RiiS surface finish to minimize the coefficient of friction. Thr decreased angles j were identified by the dynamic impact analysis to result 14 decreased interaction

forces between the shoes and guide rail chaefers. The sherp chaefer corners of all four guide rails were rounded off. Removing the chan c er corners will reduce the contact stress concentrations by providing a larger contact area between the shoes and guide rails. The reduced stress betweel the chaefers and the shoes in

i

! l conjunction with the polished surfaces and the reduced chaefer angles should i minimize if not eliminate the g4111ng process initiation.

! Tho back angles of the gate and segment as well as the seating surfaces of the 3

gate and ,egment and the valve body were treated with,a polishing stone. This j treatment ensures that the surfaces are free of burrs and high spots. This also  !

I should reduce the coefficients of friction between the seats and the back angles, i l Minimizing the friction between these surfaces results in a greater likelihood of !

i

gate and seemnt collapsing during the open and closing cycle instead of.the j shoes contacting the guide rails further reducing the likelihood of these 1 {

l surfaces galling.  !

l ,

l MSIV's which are disassembled and inspected in the future will be modified as l l discussed above unless further repairs are required. f 1 '

3 JUST1FICAfl0N FOR CONTINUED OPERATION AND AECWG4EM)ED IMSPECTION PLAN  !

l Based on the inspection results and the discussien above, it seems reasonable to l

! conclude the following: l l

r i

g 1) SCE's M51Vs are unlikely to experience the Waterford 3 M51V failure l 1 l

mechanism because their stroke time is significantly longer than Waterford 3 l j MS!Vs.  !

1  !

l  !

2) The fracture of the bottom three capscrews on one of the two segment guide

]

l rails in Unit 3 M51V 3HV 8205 discovered during inspecticn is likely to be a result of excessive interference between the lev r-lock are shoe and the j guide rail.

}

1 I

3) Based on the failure pattern and the deeper galling mark on the north arm l shoe, it seems reasonable to conclude that the preferential loading of the 1

north guide rail is a reasonable explanation for the fact that all three j broken capscrews are located on the north guide rail. The preferential

)

loading can be caused either by a shorter lev-r-lock arm or by the out of-

squareness of the guide rail skirt plate installation.

l

4) Based on the fact that (1) only three capscrews were fractured after more i

! than one hundred valve openings, (2) the fracture did not recently occur,  ;

and (3) the fourth capscrew from the bottom of the north segment guide rail does not show any signs of cracking or deformation, it is reasonable to i i

. conclude that this failure mechanism is self-limiting. In other words, the l j ,

failure stops once the interfering metal is removed from the chaefer and the galled surface is smoothed out.. i l

j 5) Since the failure mechanism is self limiting, it is unlikely to fracture l l

l ,

more than a few capscrews. As such, it is unlikely to dislodge the whole l l sognent guide rail by this failure mechanism. l l 6) To ensure that this failure mechanism can be corrected before it results in

! significant damage Unit 2 M51V's 2HV 8204 and M51V IHV 8205 will be  !

l Inspected by boroscope in the next Mode 5 outage with sufficient duration I (greaterthan7 days). If broken capscrews are found in the valve cavity and/or galling is observed, an evaluation will be made whether to repair the valve ismediately or if the valve may be repaired in the nett refueling l l

l outage. If no broken capscrews are found in the valve cavity and no l evidence of galling is found, no action will be taken.

i j 15 - -

7) All four MS!Ys for SONGS 2 and 3 will be inspected by borosco!.m during the next two refueling outages or the subsequent two refueling mtages after any repair. If no broken capscrews and no severe galling rarks on both the gate and segment guide rails are found in any of the H51Vs, the inspections i will be discontinued.

BOROSCOPIC INSPECTION IMPROVEMENTS The Interim Root Cause Report stated that only one shwd capscrew head was l

found on the bottom of MS!V 3HV 8205 valve cavity deing tM horoscoot inspection. However, three sheared capscrews wert fwN *:then the valve wat

disassembled. The two additional capscrew heads vere also found on the bottor of the valve cavity and all three were located wit.hin A few inche cf estti othe. ,

The video tape records of the boroscope inspection was reviewM but it c!varty l

identifies only a cingle capscrew head. The most p%ble caus6 of 41 <

l

inspection deficiency is the lack of a systose. tic and mcaping anvcaent # the i i

i boroscope over the entire bottom of ths valve. ,

~

In an effort to more accurately essess the cadition of th9 NP trtere.als, the boroscopic examination technigse will be lonroved as folle m j 1) A thin wire will be Attachsd to the W. uf the video probe. 'This will provide increased video probe m aip:r'ative capb111ty.

2) Th6 inspectier, of the valve castty bottert turt' ace vill be done in a j systematic eteeping nethodology (refer to f',qurs 13 for the prebe i penetration points). For excela, i.he nr,sper. tion will 40 mpde starting at the extremt left side ni the cavity :.nd intpecting front to b><k to front.

The picbe wt11 then he moved about on inch to the right and the front to 16 -

s back to front sweep is repeatad. This sequence is repeated until the entira l lower cavity surfa:a is extensively inapected.

s

3) The inspection of the skirt gu'.d( rails and lev r-lock arm shoes will be enhenced by the use of the wir6 affixed to the and of ths viden probe. The icerasse 1, probe mobility will improve tl.a quality of the insgos.

(

?! The lospartion will be performed with the MS!V la the mid position so that Lbe t'9 Unitca of the lev r rock arm shoe can be inspected and not be obacyred by the segment guide rail.

i Thete improvements in the boroscopic examination procedure will enable us to make more 1ccurate assesunents of th6 MS!V internals.

The proposed inspections discussed above are in addition to the Recommended Actions discussed on page 44 of the Interim Root Cause Report. The evaluations

, i and studits that are proposed art sunnarized in Table 1.

- IULK.T.WLE.IDEILOS One ttes noted in tho Interia Root Cause Analysis is that the MS!Y strokes closed '

f e.:sc when both hy6raulic fluid dump valves are deenergized than with only one ,

dump valve. A flow 9eifice will be installed in the conson discharge piping for eawh KSIV in Unit: 2 and 3 to that the closing stroke time will be virtually  ;

unchanged whether one or two dtsps are deenergized and to ensure that the stroke time is well above three seconds. The impact energies that the lev-r lock are shoe and gate guide rail chaefers undsrSS were determined to be at acceptable levels for stroke times greater than three seconds as identified by W K-M 17 -

engineers and the dynamic impact shalysh performed by SCE engineers. This modification will be implemt.nted no lator than the cycle 5 refueling outages on -

l both units.

The only way an MSIV could stroke with both dump valves deenergized is if it received a Train A and Train 8 simultaneous Engineered Safety F6ature Actuation .

t Signal concurrent with the valve open. The probability of this occurring and the I

valve stroking ta less than three seconds is remote.

i l

4 P

n 1 ,

l i

1

! l i  !

.1 1

I

APPENDIX A itress Analysis 1 1

The purpose of this sih.plified analysis is to determine whether or not the , ,

hypo.hesized scenario is feasible, Three parameters are calculated by a best-estimate method in this Appendix. They are described as follows:

1) The initial tensile fcres experienced by one guide rail.

2) The tensile force needed to tensile fracture one capscrew.

3) The tensile force experienced by two guide rails when galling marks were smoothed out.

Ihtjnitial Tensile,,Eqtsg The maximus opening force. Of, can be datermined by 'oalanci.3 the hydraulic pressure ud the N, pressure:

Of e (400 psi)x((12*)*-2')1.- 176,000 lbf (Ael)

The normal force experienced by the guicts rail, Fn, is calculated from of as follows:

Fn = Of cos # = 171,000 x Cos 40'

- - 135,000 lb/, where 90' is the angle between (A 2) l Of and Fn l

l The tensile force. Ft. is related to Fn by the following fotwilat Ft . Fn e u .y Cos 45'

= 95,300 lbf, where p . 1,0 for a galled surfa n (Reference A) (Ae3)

Reference A "Wear Coefficients - Metals", by E. Rabinowicz, PublisHd in Wear Control Handbook, A?.ME 1980

Fracture Tensile Force During tne interaction, the bottom capscrews will experience both a tensile force and a shear force. The magnitude of the shear force for one capscrew (assuming the loads are equally shared) is:

4 i 2

i Fs Fn Sin 45' / 9 - 10,600 lbf l ,.

Based on the energy distortion theory (Reference B), the capscrew will fail if  ;

the following criterion is met,  ;

l *

((as-ey)'+ey'+as')=2e (A.4) j US j

. where 'US is the ultimate strength.

] ,

l Equation (A.4) becomes the following formula if each ters is multiplied by  !

j . ,

the square of the bolt area.

4 t

((Fs FT )' + FT ' + I S') = 2F US (A.5) k a

, TUS - 125 ksi x 0.5" x e 'a 0.15' - 29,500 lbf l

) Based on Equation (Aa5), FT is deterasined to be 32,800 lbf.  !

1  !

j l

i i

l  !

1 i flt.ffr.t9d l i J. A, Collins, "Failure of Materials in Mechanical Design - Analysis, Prevention, and Prediction,' John Wiley & Sons, Inc. (1981)

- 20 -

l I

l Tensile Force After metal was Removed from the North Guide Rail Chamfer I

After enough metal is removed from the north guide rail chaufer, the tensile load

) will be shared by both guide ratis. Also, the friction coefficient decreases as j thegalledsurfaceisrecontoured(seepage 7,POSSIBLEFAILURESCENARIOS). For this case, ths tensile load for the bottom capscrews for both opening guide l l

i rails is:

l l

2 Fy , Fn u . Ces # l l 2

= 0.35 (Reference C) smoothed galled surface, ,

16,730 lbf (A.6) (

. i i Since 16,730 lbf is less than 32,800 lbf, no capscrews should fail. l l

l i

t h  !

) I i  ;

i

) '

1

) I 1

l

)

Reference C J

Personal Conservation between C. Chiu and Professor Ernest Rabinowicz (N.I.T.),

1 May 28, 1988 1

1 1

1 21 -

4

Table 1 Schedule of M51V Studies and_Insnactio n .

• Install flow orifice in common Nolater}hanCycle5 ,

discharge line Refueling I

• W K M to Reevaluate D-2 Valve 01/01/89 De$ign

-

• Evaluate MS!V Skirt Assembly 02/01/89

- and Lev-r lock Arm Design Enhancements  :

• Perfors MSIV Borescopic Internal .

1 Inspections as follows:

J 2HV-8204 Cycle 5 Refueling 2 Cycle G Refueling ,

2HV C205 Cycle 5 Refueling I )

Cycle 6 Refueling J

4 3HV 8204 Cycle 5 Refueling 3 1

Cycle 6 Refueling 3HV+8205 Cycle 5 Refueling 3 Cycle 6 Refueling -

i ,

l NOTES: 1. Ocifices will be installed no later than Cycle 5 refueling. The i erifices will be installed if an 'utage o of sufficient length l j occurs after engineering and parts are procured. l 3 L

2. Inspection will be made at the next Mode 5 outage of duration, but  ;

no later than the upcoming Cycle 5 refueling.  ;

ll 3. Inspections may be terminated after the Unit 3 cycle 6 refuelics )

outage based upon the results of the previous outage. This is l discussed more in dotati in the body of this report.  !

t j  ;

. )

i l i

I l

l 1

1 1

] 22 -

l l

i l

1 l

1 i

I

• - <7 4 >

• . /

1 ,

..s-

.~",-  ;  ::, ..

. ~ . .

t I/

i

"(/ .

9..

l t[,

.,4 .8 .

~

y

, ~, l ,

,L y g ,,

.  %~.

i i

i 4

i i

1 t

i 1

l l

Figure 1 A Close up View of the Chamfer for the North Cate Guide Rail

l f

i W ~

x;;c. .,

#, H;

i;5-- ,

.:,n,. ;,,.

g; :-

.. ,.

s

.s ,. ,.

(

i

l l

i 1

\

.\

I I

i i

1 l Figure 2 L

l . A Close-up View of the Chamfer for the South Gate Guide Rail i - . _ - . - - . - - - ___ --- . - - . _ __ . --

[

1 J

'* " C i{,

1  :

Ta

\. -

4

• . . ,  : 1, :

. i ,,

p

~. .

. s. , . .

~. *,-

. ;.:' ' ' ~

- +;-.

t

,9 i

l 1

I 1

l

} Figure 3 Various Views of Gate Guide Rails and Valve Skiri i

! 4

--e - - .-,- .- - . . . . . . m_ - _ - . . . . . - - - _ , - - - - ----

P

, Ci"

, . - ~. .

G-

,)

i

~'

?;, _ f

• CS i

! \

4 I

l l

l l Figure 4 Various Views of Gate Guide Rails and Valve Skirt l  !

l 1

I i

i i

a #g %weg 7.rr .

o..'l d a:s,y. y:

e , x

g. v 'L ; t.

f (1 ~'

, c y '. ,-;<

, s r .. .

'g',h > s * ,'

I

. ',5 5'} T f * ',. .it;

',4,'-- '.,( .!

3 g .

7, f .

.g', '. : !. :

' . ' ,!: ^, , - { . ..,. ~- t \ ?, '

, , k.o+ a .

de,'

vi

( s . i:

. ..,.:..,y.)

cd;,

e

'(

^

+ ..

L -1 ;' f

.w

.2 s .a .

\ f m. g\

. 4 ,7f  : 1: 3 ;. : . a ,

?) .

W yG; ' '^r ,

7 % ,s ,. -~g-,. ' '. k-r., s'c'.t' < ; . ',1 s' - n g- v. r

. _ .*g 9 -

2 .,

c~ t..t:1

'I [ .,.

,5- q ,, , !

, y 1

e

% :W.ny .e h -,

,<t'-

s

, , .:., , 4: ',;._ ::;;

.;;g; 3r

. +9 3, w ,,' v, .

. .x . .,

! ' (. v 4 v~ A..seu" . .... . . .

r d

d l

l l

1 1

I Figure 6 2

4 A Close up View of the Chamfer for the South Segment Guide Rail Y

t 28 -

i l,~, -

q.q:c.m,<<. c'y.._,

^

l i ,  ;

4 '

r  :

e,4,<'.,j.ypMcQ.4 3: T , ,* C , ,,,=)4,'..

3 jp i> ,

'} .%j L -

s

/

N * ..

j..,, =- ,- . . . ,

s- r ._, - ..

~

,3f,.7[ ..- u' i ,

' ~ w4 f: <.

.. . , , $4Ffx..v, mL :.,;cm,7 ..)l c.,ym. , - < :p r. ; . , . ,

y: .,: -.,s a,

-.x "L.  % %u ; , , .' ',:,.; L) 9;. . ,;. .

'~~

.--gJ .

. . w..

g.e . . .

..,, r.nga .  ;

.u

,2 r . ,Q i n '-

'f

- r ,_

',:{;]Q s;;,\;

1

. y

..)

- $ ' wL o. -

l l

l l

l 1

l l

Figure 7 I The location of Three Broken Capscrews l

r l

l f

o c. . ..: ,p < ,

gg, g, i- .c, .-

' u st*hn

, y; . . . , g . , ,, ,. ,, p ,

.* &. f, , 'r * : L' 5' *E' '

'Y i; <;n q yi; , '.H + . ,'J; ~:, l  ? .,

'\

^

,l, /,, D

., ..a. .

/

, . ; .... < . ; .' . , e_

!?

ij.. )? '

PS..Tek'm ,

s w f f ri. ";...;, ' v s . ; .- ,

\

p'<. ;o ..t.;

+

.k

~

+q,-

pt

$V pp p

-->- s. w p; n, .. ./.. 3 . z.

.:y - - A c.'V %. . .r;tV- .es, +p. p< > ,. ra

<. -.:3^ . -

.. w:)r ..

.m .

a - -

i . ,- y ,

. w w-

--' ~ ,

. . . '?? , N-g,, .

J s

i

Figure 8 l

Fracture Face of the Broken Capscrew Stud

r. .

a L

. .. se l

'. \. ,.

N

-+ .. fSs s,

.u y

• q, -

-p

~ ..

• ,h'

\ e <, ws ap.

.g z u. . .

D~.

k' k .

.n d

1 I

i i

I Figure 9 i

I Magnetite Build-up on the fracture Face (1400 X Magnification)  ;

\

l l - . .

I

' t y

' <z

.s p .

., f-  %

, e. '.' ,

y

..y l%

^

s

- ~

^ ; '

s ,.

I

" . iE. .s . i ,

- 4l,' 'f)i'

+,Ul' x&

0%: s s

\

> -/

' , . {;:f - .

1 ,  %.g. . \

(#

n *

\

v ,' .

• 5,j, . > -

1. .. ,,4.,

\

} t,.-

s _ x. s s

,~n 3-N

• y .q.

c Qff^ .*

4 ,

. v,

\

.% ;f 7. s s

\ .::

s

\

' g## . , , ' '

\

4 1

l 1

i i

i i

Figure 10 4

A Close-up View of the Galling Marks on the North Arm Shoe 1

l ,

l 4

/

/

r M $4

D. . .

j< , , ,

f' w . ,

i, .

's..

{q t . ',W. , . . ':

.,.f&. . C:::y,. , , l?Y q~f: ff ,, '?

v.q.7 . y: ..

-: n-t..- e { .

.-w: p , , ,

W

/ * ,

[Jj

,?* .

S......

.a I

Figure 11 A Close up View of the Galling Marks on the South Arm Shoe

l l

l l

.~ ..

W

'[4,, . . . ~

~

-:- -m

. + n . ',

- > % ' . 'y.

s

'.3 -

',, :*/_.

.. , . ::&a ~;  ;,

%r~ , -

s.,

s, -

g. ~g 's .

N- '

• N,

\ .

..s_,f, t .,

< .w.

i i

Figure 12 A Distant View of the Galling Marks on the Arm Shoes

a .

!Q r

_L g; FOLLOW PLATE

=? PACKING GLAND q ]

FOLLOW PLATE STUD <

j j; f LIFT RING e c;; o

- 1 .

% . - -9 il 4 1 J STEM SEAL RING #

. .-t ,

8ONNET VENT ACCESS 9 f BONNET ,

YI &

C { l I , SEGMENT STOP p SEGMENT

,[ p[lIQGel5 OY GATE --a -

( ,

, [

LEV-R-LOCK ff a ARM =* y SEGMENT S10E o VALVE SEAT LEV-R-LOCK __ [ }

ARM SHOE J[ h~

[ .

9l

% B

'(

i F 1

.f

/ i GATE SEAT SKIRT GUl0E RAIL lb " ,M

/ -

C SEGMENT SEAT GATE SEAT SKIRT N SKIRT I UPPER ACCESS ORAIN PLUG ' ~ ~ ~ ~ ~t SEGMENT SEAT l LOWER ORAIN PLUG ACCESS ' D' ~] SKIRT Gul0E R AIL i l

l FlouRE 13 l Boroscope Probe Penetration Points j

FRACTlJRE PATH g Fo = OFENING FORCE ,

HYORALLIC PMR9SE FN= NORMAL FORCE N APPLIED TO GUIDE RAIL J

FN = FESULTANT FRICTION w N' FM T

FT = TENSILE FORCE Fs = SFEAR FORCE l

i l

i I '

GALLING l

TOP SUFFACE OF TFE -

LEV-O-LOCK AFN SHOE '

FT Fo FN FN Fs .

PIVOT POINT OF TFE SHOE i

4 i

4 Figure A 1 Forces Resulting from Interference