ML20011F448

Forwards Revised Operability Evaluation of Salt Svc Water Pumps for Plant.Evaluation Concludes That Salt Svc Water Pumps Operable & Requirements of Tech Spec 3.5.B.1 for Pumps Met

ML20011F448
Person / Time
Site:	Pilgrim
Issue date:	02/20/1990
From:	Bird R BOSTON EDISON CO.
To:	Russell W NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
References
90-024, 90-24, NUDOCS 9003060047
Download: ML20011F448 (21)

Text

.,,

( ..

nosnwanscw .

Ngrim Nuclear Ibwer station Rocky Hill Road Plymouth. Massachusetts 02360  !

february 20, 1990 1 BECo Ltr 90-024 '

Ralph G. Bird '

senior Vice President - Nuclear r

i Mr. William T. Russell Regional Administrator, Region I i 475 Allendale Road King of Prussia, PA 19406 Docket No. 50-293 License No. OPR-35

Subject:

Pilgrim Nuclear Power Station Revised Operability ,

Evaluation of Installed Salt Service Water Pumps  ;

r

Dear Mr. Russell:

This letter transmits the Pilgrim Nuclear Power Station revised (updated) operability evaluation of the Salt Service Hater (SSH) Pumps (Reference BECo Letter No.90-017, dated January lo. 1990). Our evaluation concludes that the SSH Pumps are operable and the requirements of Technical Specification 3.5.B.1 '

for the SSH Pumps are met.

Information concerning the condition of the pump material was provided to .

other licensees-via the Nuclear Network on January 18, 1990 and was updated on '

february 9, 1990 and February 15, 1990.

A copy of the report on the pump material'was provided to the Senior Resident Inspector on February 16, 1990.

Please feel free to contact me or Mr. R. N. Swanson of my staff if you have any questions or need additional information regarding the enclosed operability evaluation. ,

j ,f h. v c.

,crR. G. Bird RLC/bal

Enclosure:

SSH Pump Operability Evaluation, Revision 1 f

P '*kB8e n 888g e Q Q

.e .,

"', Page 2 Boston Edison Company Docket No. 50-293 Pilgrim Nuclear Power Station License No. DPR-35 cc: Mr. H. Fairtile Licensing Project Manager Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission 1 White Flint North 11555 Rockville Pike Rockville, MD 20852 Mr. J. R. Strosnider Chief HPS, Engineering, DRS 475 Allendale Road King of Prussia PA 19406 Sr. NRC Resident Inspector - Pilgrim Station e

F

" /. *

. BOSTON EDISON COMPANY

,t OPERABILITY EVALUATION

'm Revision 1 ,

l.. Initiating. Document 'F&MR 90-8/NCR 90-3

2. Affected Item (System, Subsystem, Train, Component, or Device) 1 Salt Service Water Pumps (P2088, C, D, E)

The salt service water pumps are vertical, single stage, deep well pumps manufactured by Goulds Pumps. .

l

'3. Specified function of the Affected Item l The salt service wat'er pumps supply water at adequate flow and pressure to-the RBCCH heat exchangers as an ultimate heat sink for the RBCCH system for all accident and transient conditions.

4. References ,

FSAR 10.7 Technical Specification 3.5.0, 4.5.B PDCR 77-44 Specification M8 Drawing M8-4 PDC 90-12 ,

S. ' Operability Concern On'l/11/90, while horizontally transporting the assembled pump (without motor) from the maintenance shop to the intake structure, the bowl /

impeller end of SSH pump A was dropped approximately 6 to 12 inches leading to a 360 degree throughwall fracture of the second, 80 inch column section up from the bowl / impeller end. Further investigation indicated that the area of the fracture had reduced material properties (e.g. yield strength, ultimate strength, percent elongation) relative to specification. The operability concern is that remaining pumps in-service may have columns having similarly reduced. material properties that could affect their operability under design loading conditions.

Pump A has been fully inspected, tested and returned to service and no operability concerns exist.

I 6. Operability Evaluation (check one)

[X] Operable L

[ ] Inoperable ,,

[ ] Conditionally Operable i

r" F /. Basis for Lvaludlion

+ '

- - The service water system normally operates at approximately 30-psig and f ,

uses cold ~ sea. water. One train consisting of two pumps is required to

' satisfy the heat removal requirements for all design basis accidents and i transients. Pumps A and B- supply Train A, pumps D and E supply Train B '

F and pump C normally acts as a swing pump.

The' overall length of the assembled pump is in excess of 42 feet (drawing HB-4). The pump was being transported at each end by slings from two  ;

I r forklifts. The fracture occurred approximately 17 feet from the bow 1/ impeller end. Analysis of the stresses resulting from this drop indicate loads of approximately 30 ksi from a drop height of 6 inches and approximately 42 ksi from a drop height of 12 inches. Lab tests of samples from the area of the failure indicated ultimate ~ strengths ranging from 23 to 32 ksi. Therefore, a failure was expected due to the drop.

The pump column ma'terial is an aluminum bronze casting (ASTH B148 C95200) ,

with a required ultimate strength of 65 ksi minimum, yield strength of 25 ,

ksi minimum, and elongation of 20% minimum. A total of four coupons from i the area of the fracture were tested. These tests were performed at two I separate labs and gave consistent results of significantly reduced-  !

O strength. The worst coupon gave an ultimate strength of 23.9 ksi and an i elongation of 0.5%. Lab tests of four samples from a " good" section of a l pump A column indicated tensile strengths ranging from 70 to 85 ksi (ASTH specified at 65 ksi minimum) and yjeld strengths from 27 to 45 ksi (ASTM l Specified at 25 ksi minimum). Thres of the four samples had elongations l of 21 to 34% (ASTM specified at 20% minimum). The fourth sample had an  !

elongation of 16% attributed to a localized crevice / pit. See Table i for  !"

a detailed desi:ription.of lab' test results. i Two factors are believed to have contributed to the fracture. First, the  !

initial. casting process likely produced a susceptible material since i metallurgical analysis has identified a eutectoid condition indicating i that the required single phase of the material was not achieved. In this i initial eutectoid condition, an aluminum-rich phase of the material forms  !

at the grain boundaries. With continued exposure to a corrosive medium ,

(sea: water) over several years, the material dealloys as the aluminum is  !

lost. This results in the formation of a low strength, copper-rich network. This eutectoid condition was in a relatively large section in the vicinity of the flange supporting the conclusion that the weakness was l introduced by casting and not by some other process (e.g. welding), s Metallography has indicated that the fracture area consisted of a substantially interconnected eutectoid region which is significantly more susceptible to dealloying. Metallography also indicated that acceptable material contains the expected microstructure with small amounts of H eutectoid (<5%) which is not interconnected and is much less susceptible 1 to degradation. Slow cooling of the casting is believed to cause the d

eutectoid condition.

i i Second, weld build-up had been done in the upper flange region of this column to restore a spider bearing support surface which had been worn due l to operation. . Subsequently, a crack developed in the weld build-up. This crack was a pre-existing flaw approximately 5 inches maximum in length and l i

2 l

l

l b[ i- -

throughwall. Examination determined that this crack had existed for some

(. .- time prior to application of the impact load caused by dropping the pump.

It was determined that the crack was not the result of fatigue and was not

! propagating circumferential1y during operation.- The primary origin of the fracture was at the weld build-up and it propagated rapidly through the low strength, copper-rich network.

.Because the material in_the area of the break had very low percent elongation, the potential for brittle fracture was considered in this evaluation. For brittle fracture to occur, the material must be .

susceptible. The likelihood of brittle fracture is affected by the size of the load, the type of load, and whether a flaw enists in the material prior to the load being applied. An existing , flaw provides an initiation site for a fracture. The dealloyed condition caused the embrittled-type properties in the failed material making it susceptib'le to brittle fracture. With the pre-existing flaw and high impact load from the drop, the situation was such that a brittle fracture was likely. However, impact loads of sufficient magnitude are not expected during operating or -

accident conditions such that a fracture would occur. .

A fracture analysis of the material with degraded mechanical properties was performed by Teledyne Engineering Services (see attached letters). '

This evaluation provides a measure of degraded materials' ability to withstand shock loading with pre-existing flaws. The testing and analysis-determined the following critical flaw sizes assuming worst identified degraded material (appropriate factors of safety are applied):

Permissible Crack Lenath Throuah Wall Surface (1/4t) ll21 Normal'(2 ksi) 1-1/4" 235 degree 2.2" Faulted (10 ksi) 1/4" 0.8" 1/2" Faulted (6.3 ksi) 0.65" 360 degree 1.1" These results indicate that under a faulted condition stress of 10 ksi, a' 1/4" or greater circumferential throughwall crack or a 0.8" or greater circumferential surface (3/32" deep) crack must exist for the crack to propagate and cause column failure in degraded material. For the maximum calculated stress level of 6.3 ksi, a throughwall crack of 0.65 inches or less or a Xt surface crack of 1.1 inches or less is permissible assuming that the worst identified degraded material conditions existed in the area of the crack. For undegraded material, permissible crack sizes are estimated to be at least four times greater. These flaws would all be readily detectable by visual or liquid penetrant examinations.

Material allowable stress values were adjusted assuming non-ductile and reduced material properties.Section VIII of the ASME code (UCI-23) requires a safety factor of 10 between the minimum tensile and allowable stress for cast iron which is a brittle material similar to the degraded l casting. lhe lowest of four test specimens from the degraded material had a tensile strength of approximately 23.9 ksi. Thus, the allowable would be 2.39 ksi. However, the loading of concern is primarily bending and the l allowable in bending is 1 1/2 times that for tension. Thus, the adjusted allowable value for the degraded material is 3 6 ksi.

3

Table'2 is a summary of calculated versus allowable stress- values for the '

ti normal, operating basis earthquake (OBE) and safe shutdown earthquake (SSE) conditions. Results of Table 2 show that the maximum stress levels would be below the code allowables assuming the worst case identified - '.

degraded' material existed in these pumps. These analyses account for the installed stabilizers and tie-rods (see Figure 1) and the effects of water on mass and damping. Stabilizers have recently been installed on pumps B and C and tie-rods have recently been installed on pump A (all but lowest column) and pumps B, C, D and E (top two columns) to increase margins to allowables. The stabilizers alone improve stress margins by up to 50%.

Therefore, even with an assumed loss of design margin, the salt service water system can still perform its safety function for all accidents and transients described in the FSAR. This analysis is consistent with FSAR structural loading criteria. This evaluation has been reviewed by and has the concurrence of Dr. William Cooper, a consulting engineer with Teledyne Engineering Services, as described in the attached letter.

For all design basis accidents except earthquakes, the pump loads seen during the accident are identical to those seen during normal operations.

Since pump loads during normal operation are low (1.2 ksi), a low probability SSE would have to occur to significantly challenge the pumps.

l The probability of an SSE occurring at PNPS between now and mid-April l

(estimated the range ofnew 1.6column x 10-4 delivery and 1.6 date, x 10-5i..e. approximately

, The likelihood of two months) is in an SSE occurring coincident with an accident is even more remote. Since two pumps in the same train are needed to mitigate this event, multiple pump failures must occur to pose a safety problem.

Also, if only one SSH pump fails, a coincident diesel (or other emergency electrical supply) failure must also occur in the other safety train to challenge plant safety. PNPS diesels have a demonstrated reliability of over 99%. An inservice pump must be in a more severely degraded condition than the worst section of pump A (the fractured pump) in order for a L failure to be considered possible. All of these factors combined l . demonstrate a low risk situation.

Extensive field testing was performed on the pump A columns, the inservice B, C, O and E accessible columns, and new columns-stored in the warehouse.

No other cracking or reduced strength conditions have been identified.

l Testing was performed in the area where the flange joins the column l

because this area is most suscept'ible to slower cooling that leads to eutectoid formation. UT measurements in the areas of post-weld buildup machining ~ indicated some reduction of wall thickness. Assuming safe shutdown earthquake loads, adequate wall thickness was measured using UT.

Liquid penetrant testing of the inside diameter of the other columns of pump A revealed aligned porosity, corrosion porosity, and what is believed to be 360* 1ack of fusion in the upper flange area of one column. This column also had reduced wall thickness at one flange end. These L- conditions also likely resulted from the weld build-up effort. This 5 column was not returned to service in pump A due to time constraints of the rework involved.

4

L 4

-. Should a weld-induced cracking condition exist in conjunction with <

degraded (eutectold) material, the potential for crack propagation exists i , if adequate stress is applied. Pump A operated with the significantly degraded material condition and with a substantial weld-induced crack and did not fail-in service. Without the degraded material condition, colum'n failure by existing crack propagation due to SSE loads is not expected since the base material will arrest the crack growth. Adequate wall thickness is expected in other pumps based on the pump A inspections.

Improvements have been made in the manufacture of replacement column-types 641 and 642 (See figure 1). Thess columns are centrifugally cast

. versus the sand casting used in the failed column. Also, these columns are cast in two smaller pieces which allows for more un1 form cooling and reduces shrinkage and porosity concerns that can result from a single piece casting. The two castings are then joined by a circumferential weld at the centerline-to provide the finished product. Pumps B, C and D have some columns which have been manufactured using the improved processes.

The pump manufacturer knows of no inservice failures of this new design.

This better quality casting is recognizable by the circumferential weld at-its centerline. These columns were visually verified to be installed in pumps B, C and D as indicated by dots (*) on Figure 1. Heat treating and controlled cooling of the castings will reduce the potential for significantly interconnected eutectofd formation in both centrifugal and sand castings. These treatments have been specified for the new replacement columns.

Calculations indicate that the static dead weight loads resulting from the horizontal transport of the pumps are on the order of 5 ksi at the midpoint. The static loads in the columns adjacent to the mid-column of the pump are in the range of 3 to 5 ksi during transport. This is roughly twice the normal operating load. Dynamic effects during transport would have tended to increase these loads somewhat. Therefore, the mid-columns of several inservice pumps have probably experienced loads (during transport for maintenance) that are significantly greater than normal operating loads and a significant percentage of SSE loads. During transport, no failures other than pump A have occurred.

During our investigation, we-learned that Pilgrim had one other column failure in 1977 in pump D. Discussions with a consultant at AMPCo (the company that repaired the column) indicate that this failure was due to fatigue. At that time, the pump tripped on high amperage. On troubleshooting, a 360* circumferential failure of the uppermost column was discovered at the upper flange transition. The pump experienced high vibration and was apparently operated after bearing damage had occurred.

He believe that the cause for the vibration was flow instabilities which were aggravated during certain pump operating combinations. Design changes to the pump suction end and system operational changes have been implemented which have reduced flow induced vibration. Also, a Boston g

Edison records search of work on the salt service water system was performed and no other failures of the older design column were found other than the 1977 and 1990 failures. Goulds has reported no other failures of the older design column. Further research indicates that no other aluminum bronze castings exposed to sea water are used in safety-related applications at PHPS.

Operators would become quickly aware of any pump failures by the high current pump trip and low header pressure indications in the control room. High current pump trip led the operators to recognize the 1977 event.

5 4 . <

Based on inspection and testing of pump A and inservice pump columns, analysis-~0f pump stresses with stabilizers (pump A without crediting

'* stabilizers) and tie-rods, review of pump maintenance records, and the low likelihood of any. seismic loads, the potential for a pump failure due to -

postulated degraded material and/or cracking is negligible and the pumps are considered operable.

c8. -Compensatory Measures / Conditions Required The following actions will be taken as prudent measures to verify-

materials and integrity of the remaining salt service water pumps:

1) Pump A was inspected, tested and returned to service on 2/13/90 at 1734 hours0.0201 days <br />0.482 hours <br />0.00287 weeks <br />6.59787e-4 months <br />.

2) Remaining pumps-(B, C, D and E) will be inspected, tested and returned to service with tie-rods installed on the upper six columns, a) Prior to returning these remaining pumps to service, the #

following inspection plan shall be used and satisfactory results obtained:

Remaining pumps (B, C, D and E) will be inspected and tested in two phases. The first phase consists of external tests and inspections of all columns followed by-tie-rod installation on all.but the lowest column before returning each pump to service. The second phase requires pump disassembly for internal tests.

In the first inspection phase, the following inspections and ,

tests will be performed without disassembling B, C, D and E pumps..

1) Ultrasonic testing wall thickness measurement in the ,

vicinity of the flanges.

2) Visual inspection of the outside diameter.

3) Hardness testing on the outside diameter near the flanges.

on all areas not tested previously.

(Target - 3/18/90 early finish) b) In the second inspection phase, each pump will be removed and disassembled to perform the following inspections and tests:

1) 1.iquid penetrant examination on the outside diameter at the vicinity of the column flanges.

b 2) Visual inspection of the inside diameter near the flanges.

3) Hydro test at 125 psig with a 15 minute holding time.

4) Liquid penetrant examination on the inside diameter in the vicinity of the column flanges.

3) Co' umns found to have degraded properties will be systematically I replaced with satisfactorily tested columns.

I 6

l 1

'- 4 ) -. 'If.an inservice salt service water pump becomes inoperable while one salt service water pump is tagged out of service for inspection and testing and pump column fracture can not be eliminated as the cause, a 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> LCO will be entered.

5) In the longer term, newly ordered columns will replace all coltmns in each: pump.- (Target - 6/30/90 early finish)

9. Performed By N , A #

Reviewed By WM/*

• N //dBate Recommends Approval b A e Date2!I4/90

()(S&SODivision Manager)

Recommends Approval #

Date 2//6/70

(/ (ORC Chairman)

ORC Heeting Number YO ~/P Approved By Date I '/d' 9D

F (P4 ant ManagerW 7 ,

L . .

TABLE 1 (Page 1 of 2)

LAB TEST RESULTS. ,

.- y C ' Tensile Soecimens Sample Ultinate Yield-  %

. Laboratory Samole Location (ksi) (ksi) J(Lang Rockwell'B Specification 65.0~ 25.0 20%

SWEC Tl 6420 mid 85.4 31.1 34% 66~to 75

~SHEC T2 642D mid 70,0 29 16%

< SHEC T3 6420 mid 84.7 31.'l 33%

SWEC- T4 6420 mid 79.5 45 21%-

6420 fracture 31.9 SHEC TS 31.2 (1% 41 to 52 MMR 1 6420 fracture 23.9 21.7 0.5% 58 MMR 2 6420 fracture 28.7 26.3 1.3%

MMR 3 6420 fracture 28.9 .25.9 2.2% ,

MMR 4 6420 mid 82.0 27.2 35%

MMR A 6420 flange 78.0- 36.9 23%

MMR 7 6420 flange 77.0 f HMR 8 6420 flange 76.5 MMR 9 642D flange 77.0 MMR 10 642D flange 79.0  ;

1" from MMR CF1-A 6420 bottom 82.0 28.7 33%

flange 3 1/2" from HMR CF1-A 642D bottom 73.5 25.1 24%

flange Teledyne Degraded 53 to 56 Teledyne Undegraded 61 to 69

l f['

z+ r

- TABLE 1~(Paae 2~of 2)_

" 't

< Chemical Analysis

~

Sample

• Test Lab Location CoDoer (Q Aluminum (%) Iron (1) Other (1)

'MMR' . Specification 88 8.5-9.5 2.4-4 MMR Column Tube 89.17 8.35 .

2.43 .05 MMR Column F1ange 87.22 9.22 3.51 .05 i

~

MMR 1 from 89.4 8.1 2.3 ,

fracture HM'R '- 2from inner 89.4 7.63 2.8 '

pipe ~ surface away from weld MMR 3 from inner 91.5 7.2 1.1 surface of pipe where weld was thought-to be. -

(probably weld, stick material)

.-i a

l

-TABLE 2 (Page 1 of'1) -

EVALUATION OF PUMP COLUMNS HITil AS_ FUMED DEGRADED MATERIAL PROPERTIES *

'(PUMPS B. C. D. E)-

TYPICAL PUMP (Hith Stabilizers.&

' Tie-rods on Top 2 Columns)

Max Calc' Allow Stress Stress Ksi Ksi Normal 1.2 2.4 Operation (NO)

OBE + NO 3.9 5.9 SSE + NO 6.3 8.6

• An allowable of 1/10 of the minimum tensile strength of the four degraded test coupons per ASME VIII UCI-23 was used.

NOTE: Stabilizers have been added to pumps B.and C. Tle-rods have been added to all but the lowest column. of pump A and to the top two columns on pumps B, C, D and E.

s m

, 4 m w~- 4 s -

n. m -k 4 p.g z -

r ,

c 7

V g',

e T__ABLE 3 (Paae 1 of 3) j ,. FIELD TEST RESULTS t t IESI. RESULTS 1HM r

Pumo A All columns except a) LP of OD within 6" of All acceptable fractured column flange .

6420-b) Hydrostatic Test All acceptable c) Flange parallelism to All acceptable ensure proper axial alignment ,

d) LP of ID within 8" of 641 Acceptable flange (644 column sent to 642A Acceptable ,

lab before performing LP) 642B Acceptable  ;

Acceptable 642C 642E Scrapped

• e) VT of 10 within 8" of 641 Acceptable flange -642A Acceptable 642B; Acceptable 642C Acceptable 642E Scrapped

• 644 Acceptable f) UT of OD within 6" of 641 Acceptable-flange (644 column sent to 642A Acceptable lab before performing UT) 642B Acceptable 642C- Acceptable 642E- Scrapped *

-g) Rockwell B Hardness (644 641 Acceptable column sent to lab before 642A - Acceptable l

testing) (Data taken 6428 Acceptable within 6" of flange) 642C Acceptable 642E Acceptable' h) RT within 6" of flange 641 Acceptable 642A Acceptable l 642B Acceptable L

642C~ Acceptable Acceptable g 642E 642E has aligned porosity, corrosion porosity, 360* lack of fusion ,

and a counterbore in the flange area with a reduced wall thickness'.

This column could be repaired except for time constraints and is l' therefore being scrapped.

1, NOTE: Rockwell hardness field test results were checked against lab results i and found to be consistent.

l

- e _ _ _ _ _ - _ _ _ _ _ _ _ _ = _ _ _ _ _ _ _ _ _ _ _ _ -

In ;c s, TABLE 3 (Page 2'of 3)-

[

+  :

. c :.

IIB IESI RESULTS PU'mo A'

' Failed-Column a) Tensile Tests (see Table In fracture area, ultimate (642D) 1). , and yield strength and */.

elongation significantly below ASTH standard. In J middle and other'end of column, same properties -

were within specification.- '

b) Rockwell Hardness (see -In fracture area, hardness '

Table _l). values were below AS1M:

standard. In middle and other end of column, hardness was within specification.

c) Hetallography Results support the .i conclusion of eutectoid l condition being present in_  !

g materials of degraded '

i properties.

d) Chemical Analyses (see _ Analyses in area of Table 1), fracture indicate dealuminization (from corrosion). Middle and  :

other end of column are within specification.

New Columns (In Harehouse) 2-641s a) LP of OD within 6" of All acceptable i 2-642s flange b) LP of 00 within 3"'of both All acceptable sides of circumferential H beltline weld l c) LP of ID within 8" of All acceptable flange d) Hydrostatic Test All acceptable (manufacturers test) e) Rockwell B Hardness within All acceptable .

[ 6" of flanges '

n,

~s l

TABLE 3 (Paae 3 of 3)

, n .

l:

T1SI RESULTS IJJB _ ,

New Columns (In Harehouse) 1-644 'a) Rockwell B Hardness within All acceptable 6" of flanges b) UT wall thickness All acceptable

(.375" minimum)

All acceptable c) VT of ID within 8" of '

flange All acceptable '

d) VT.of OD within 6" of ,

flange

- Note: LP is inconclusive for item 644 which has an unmachined reugh cast 3 i

surface.-

PIECE lE11 RESULTS In-Situ >

~

B,C,D,E a) Rockwell B. Hardness (at Pump B -Acceptable SSH Pumps lower flanges of 641 C Acceptable column and upper flange of D Acceptable 642A column) E Acceptable b) Visual verification of See Figure i for location of new design locations, columns (including those that are submersed)_

s

s d

o

- R N e _

N.NT .'

j _

m- )

I

~

E I y l I r Y J 1 ], . . sd s

_ 7 r

e

_ l ele sl yt as ni n a

N d z _ a e N o _ t _

N R ib bs ia ds nr N e _ a" _ it eg e

sNT j dl t

e3 r

cI

a m ) y I ol ni s _

' e r n D - ,- ..

• aio t u ec t

{ I r Y j1 ] , _

b ir dd 2_

N s r

e 7 __ 5 t

e y a

s u in kr.

i N d _ iz s o l yll =

g_

i R b t lna N e _ t a er s

7 r fo r

do _

8 u 9 c e S L y

a m

s

)' I N. NT Ij L- [

1 w

A ge e

r ak h"

P M

n m

C I--

o .

I sf ley U u lo

- d le 3u ll P C I la f

'R p r k' J i '~

t s

f n 2no oy l E

T A

P m

u N s _ r 7 __ i sn ee miizar

-r 1

e e l

N- d iz _ u b y -

W s n '

R o l i

lo t h ae r u -

E o '

_ b a c St C it. N e t s eAes a -

I V

se _

s e" _ t a pu c mueac i

F R id f N.AT n Pb E D m ) -

- L h1. I S

T n

n u B I

' e e e

L A

'c C

I- l I S p r k' J1 ]_ _

7_ __

m-u -

P r o

' _ e t N s iz l

d K

y c

i.

N R

o _ i b

t a

s ei  :.

N ie a m ) f NkT  ; ,

[

1.

h A  ;-  ;[,  ; .

- fI

]-

r k' ,- '

x N

_ *7 z _

_ il _

'N i b

a t

s e -

"r - -

v__ n o

it a A c '

v__ '

6 1

1 3

3 4 -

" o o p **

t

" o "M 's

's

• _ L m k u a P

• s

• _ p" e

e no r -

lap c

m ) , L" m' I r

le im pu yP T

r k'

~ pl pI e

p m

e g

r a N r^ w_ e hC .b

.D o: M 4 h

i D

c s 'N N

& 2

$ c 4

a _

G 2

4 2

2 (6

4 i3j ggSi-

P ENGINEERING SERVICES 130 StCOND AV(Nut

~8

_1 W41NAu.MA55ACHU$tTf8 Cf364 9195 r

86171460 3350 7WX f71013344tiO4 t January 19, 1990 7257-1 .

Mr. R. V. Fairbank Manager - Nuclear Engineering Department Boston Edison Company 25 Braintree Hill Office Park Braintree, MA 02184 .

Subiect: Operability Evaluation of Salt Water Service Puinps t

Dear Mr. Fairbank:

Teledyne Engineering Services (TES) has reviewed the material property I information and the draf t reports for the subject evaluation, with ,

particular consideration of the structural design criteria applied.

Based on this review, TES concurs with the criteria used in the structural evaluation, specifically ini:luding: the analysis methods used to determine the applied stresse: the use of the as-measured material properties; the -

factors applied to the measured properties in order to determine allowable stress values (use of a. nominal f actor of safety of 10 en the ultimate f

', tensile strength, a multiplying factor of 1.5 because the limiting stresses are primat fly bending, and the multiplying factors of 2.0 for OBE and 2.4 forSSE): and, -he conclusions with respect to structural integrity based on the above listed items.

We particularly want to comment on the procedure and factors used to determine the allowable stresses. The data indicate that the material' properties adjacent to th6 flange are degraded from expected valties. Not only are the yield and tensile strengths lower than normal for this material, but the ductility is significantly reduced. The degraded material approaches the characteristics of the cast irons considered in

. Part UCI of Section VIII of the ASME Boiler and Pressure Vessel Code.

Therefore, we consider it appropriate to use the f actors on the material l properties of the rules cf UCl-(thc values of 10 al.J 1.5) rather than the -

much more liberal f actors applied by the Code to this material in the normal more ductile condition. Having made this correction, we consider it correct to continue to apply the normal Section 111. Class 3 factors for the Service level C and' Service i.evel O conditbns (the values 2.0 and 2.4). Finally, we note that the lowest measured value of the tensile .

strength has been used as the basis of these calculations, and that the next lowest-value is 20; higher. This, plus the alleviating factors which would be demenstrated 11 a more detailed analysis and the f act that the slight overstress condition exists in only one column, contribute to our k.. opinion that reasonable assurance of structural integrity has been demonstrated.

Very truly yours,

.TELEDYNE ENGINEERING SERVICES iff?

W. E. Cooper, PE A h,Brf&pe r*

/ J. C. Tsacoyeanes Consulting Engineer Consulting Engineer t rnws.i n" v.o r uu u.wisr r

5 ..d WTA mYNE T

![y a*

ENGINEERING SERMCES February 2, 1990 7257-2 -

Mr.-R.V. fairbank Manager - Nuclear Engineering Department Boston Edison company 25 Braintree Hill Office Park Braintree, KA 02184

Subject:

Pilgrim Salt Service Water (SSV) Pump Columns -

Fracture Toughness Testing

Dear Mr. Fairbank:

Teledyne Engineering Services has aerformed fracture toughness tests on the SSW Pump A failed column material Specimen material was obtained from column samples in the possession of Massachusetts Material Research, Inc.

The results reported herein are for three-point bend specimens machined from material adjacent to the fracture surface at three circumferential locations. Ncminal dimensions of the . specimens are 3/8" x 3/8" x 1.54".

The procedures. of ASTM E 399, " Standard Test Method for Plane-Strain Fracture Toughness" were used recognizing that the dimensions and nature of the material would preclude full compliance with the E-399 validity-requirements, Specimens were machined with a straight through notch on the side corre-sponding to the outside diameter of the column and precracked by fatigue.

Therefore, crack extension during the bend test would progress through material with the greater degree of de-aluminizing.

Test results are summarized below:

,sogimen Q (XS1 Iin 1 T2 9.0 T3 9.8 T5 9.7 I All bend tests were performed with specimen temperature at 32F. .As expeci.ed, none of the above tests satisfied all of the validity require-ments of E-399. An additional specimen, T1, failed during precracking at an approximate value of 8.6 KS1 Jin . A preliminary value for material considered to be undegraded is estimated at approximately 22 ks1(Tii. This is based on a single specimen which frac.ured during precracking.

e - = .

W IELEDYNE-L/.I E Boston Edison Company-February 2, 1990 ENGINEERING SERVICES. .

[' 1 , .Page 2 L

ilardness measurements obtained on the tested specimens are tabulated below for both degraded and undegraded material.

Dearaded Undearaded Rockwell B 56,56,53,55 69,61 .-

Brinell 88,88,85,86 110,95 y .!

As requested, we have approximated permissible crack sizes based on a i fracture toughness value, Kit of 9,0 ksifin and two crack configurations as- )

follows: ,

j (1) a through wall circumferential crack (2) a surface crack of 3/32 inch depth and of elliptical shape, j (The depth is based on Section 111 postulated crack size of i 1/4 thickness for non-ductile failure evaluation.)

Crack sizes for,the above are calculated for stress levels of 2 ksi .(Normal Conditions) and 10 ksi (f aulted Conditions), in addition, the appropriate saf ty factors of (17 and [2', respectively are applied. The permissible -

crack sizes are tabulated below: j

)

l Permissible Crack length '

Throuah wall Surface Normal 1-1/4" 360 degree l

Faulted 3/4" 0.8" )

i Very truly yours, j TELEDYNE ENGINEERING SERVICES

.f) . 14eumw a; .

/ 1 I

ames C. Tsacoyeanes Engineering Manager JCT /msb j i

l w =m m m nw a33"

. onw co own w s m m

~'

• /eTELEDYNE x "

ENGINEERING SERVICES

,e A'=V ,1 130 SECOND AVENy[

WAL1MAM.MASSACMVSETTS C2254 0196 (617)690 3350 TWX(7101324 76?6 -

February 15, 1990 7257-3 Mr. R,V. Fairbank-Manager - Nuclear Engineering Department Bostor. Edison Conpany ,

25 Braintree Hill Office Park Braintree MA 02184

Subject:

Pilgrim Salt Service Water (SSW) Pump. Columns .-

Dear Mr. Fairbank:

The attachment-(TES t.etter 7257-2) transmitted to you on February 2,1990 reported the results of our testing program on the SSW Pump A failed column material. Also included in the letter were the calculated crack sizes based on- the fracture toughness value of 9.0 ksi /permissibleTii'and two crack configurations as follows:

(1) a through wall circumferent'ial crack (2) a surface crack of 3/32 inch depth (one-quarter thickness) and of elliptical shape

' Subsequent to that work, BECO requested that-. TES perform 'similar calculations for a surface crack of one-half the column thickness, We have-com>1eted: those calculations and have tabulated below the results along wit 1 the previously calculated results, Permissible crack Lencth Through Surface Surface

__ Wall 1/4 T 1/2 T ilarmal (6-2ks') 1-1/4" 235 degrees 2.2" Faulted (SSE):

(6-10ksi) 1/4" 0.8' 0,5" (6 6.3ksi) 0.65" 360 degrees 1.1" I'.Gii:5I:s ANo . ET.u.. : cists

e

. ,nr. n.v; rairbank jg i get;yiyg .

ENGINEERING SERVICES

[ {:

n

.gonEdisonCompany_

' a l Jac * , February 15, 1990 4 Tg ,Page 2 Note that: the value 'of 235 degrees above was previously reported as 350 .

degrees. The new value incorporates the additional condition of limiting the' net-section stress to a value not greater than the allowable stress value for Normal Operation of 2.4 ksi.

Very truly yours, ll TELEDYNE ENGINEERING SERVICES f- J N @ ,.'

mes C. Tsacoyeanes Consulting Engineer i e

F JCT /msb F

v l

I L

1 I

[

l