ML20141E216
| ML20141E216 | |
| Person / Time | |
|---|---|
| Issue date: | 04/11/1986 |
| From: | Zech G NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| To: | Bronson B DRESSER INDUSTRIES, INC. |
| Shared Package | |
| ML20141E221 | List: |
| References | |
| REF-QA-99900054 NUDOCS 8604150092 | |
| Download: ML20141E216 (2) | |
Text
- - - - - - - - - - - _ . . . - - - - - - - - - - - -
f ~
l c,
)
i
[ ease o,, UNITED STATES 8 o NUCLEAR REGULATORY COMMISSI n ; wassencrou. o. c. osss
%, *****/
Aoril 11, 1986 I Docket No. 999000054/85-01 Dresser Industries, Inc.
ATTN: B. G. Bronson QA Manager Post Office Box 1430 Alexandria, Louisiana 71301 Gentlemen: ,
Thank you for your letter of February 10, 1986, in response to our letter dated January is needed.
17, 1986. As a result of our review, we find that additional information Specifically, in regards to the Davis-Besse safety valve (serial number BM8635) that was returned to Dresser in 1981 and to the Diablo Canyon Units 1 and 2 safety valve (serial number BN1741) that was manufactured in 1971-1973, answers to the following questions would be appreciated:
Toledo Edison Company / Davis-Besse (safety valve serial number BM8635) 1.
When was this valve manufactured?
2.
What was the carbon and nickel content of the collar (CB-30)? What was the carbon content in the Type 304 stainless steel cotter pin analysis? What was the method of chemical analysis used to generate your alloy compositions?
- 3. Where exactly were your hardness readings taken?
4 Were the failed spindie threads flat, bright and shiny?
5.
What was the chemical analysis of the metal chips that were scattered on the top surface of the lower spring washer and the top surface of the cover plate?
- 6. I Were there any unique cualities about the spindle's post-failure microstructure?
7.
Did the cotter pin show any evidence of necking down?
Did the fracture surfaces show evidence of beach markings?
Pacific Gas & Electric Compa,ny/Diablo Canyon, Units 1 and 2 (serial number BN1741).
We agree with your conclusion that the evaluation of the Diablo Canyon safety valve failure was not completely stated and clear in regards to why "the 1 00h > < y p -
r
(
o.
C Dres'er s Industries, Inc. April 11, 1986 failure is an isolated case." Your explanation is acceptable, however, the following questions remain concerning your evaluation.
- 1. Was there a metallurgical failure analysis performed to determine the cause of the defective disc collar / spindle threads? If so, what were the results?
- 2. What are the major changes that have occurred in Dresser's quality !
inspection program since 1973 to reasonably assure future nonrecur- !
rence of the subject problem?
- 3. !
Which other nuclear power stations, excluding Davis-Besse, have safety valves of the same type and design (as mentioned in your response paragraph 1.7)?
Nonconformance (85-01-07)
It is agreed that the four Section III rod ovens were calibrated according to schedule. However, the nonconformance addressed a failure of Dresser's quality control to certify the-calibration completed on June 6,1984 by either a stamp or a signature. As a result of a discussion with your Mr. Rod Thomas, we understand that the situation has been identified and corrected. No additional response is required.
please provide the additional information within 25 days of the date of this letter. The response requested by this letter is not subject to the clearance procedures of the Office of Management and Budget as required by the Paperwork Reduction Act of 1980, PL 96-511.
Sincerely,
_ , , . l A G s LJ Gary . ech, Chief Vendor Program Branch Division of Quality Assurance, Vendor and Technical Training Center Programs Office of Inspection and Enforcement l
t i
9 VAL S G#####
DRESSER VALVE AND CONTROLS DMSION Desser indusmos. Inc.
industno vane North Amencon Operchons PO. Box 1430 + Aleonso. Louiseano 71301 February 10, 1986 Mr. Gary G. Zech Chief Vendor Quality Assurance Nuclear Regulatory Ccanission Washington, DC 20555
Reference:
Your Docket No. 99900054/85-01 Dear '4r. Zech This letter is in response to your audit report for the NRC Audit of our facility September 30 thru October 4,1985. We have investigated each finding noted by the audit team.and responded accordingly.
If additional information or clarification is required, please cb rot hesitate to contact me.
Respectfully, i
/1 4 Ms"o; R. R. Thomas Quality Assurance Engineer, Systerrs and Audits (318)640-6028 RRT/jf 9602180227 860210 PDR GA999 ENVDRSI 99900054 PDf ,
,.oh
. : =...
.m,s> _ . -so.m m... _ m. _
)yGq
Dodoet No. 99900054/85-01 Reported Violation (85-01-01):
, Dresser's Part 21 file no. 84-01 was not adequately evaluated. File No. 84-01 identified a problem with disc collar failure for 3707RA safety valves. The file stated this problem was not reportable since "the failure is an isolated case." However, a letter dated 07/02/84, in the same file states this failure has occurred twice since 1971 Once in August 1981 at Toledo Edison / Davis-Besse and the second time at Diablo Canyon Unit #2 in May,1984.
Information in file no. 84-01 also indicates that in May,1984, durirg blowdown testing by Nyle Laboratories, Huntsville, Alabana, on the main steam safety valves from Pacific Gas & Electric /Diablo Canyon 2 (Dresser valve type 3707RA),
valve S/N BN1741 lifted with simultaneous shearing of the disc collar / spindle threads and cotter pin. This occurred during the first actuation, at approximately 1065 psig. The disc and spindle deflected sidewise upon closing and the disc became edged between the nozzle seat area and the bottm of the disc holder. The valve became mechanically jamed and could not open. It also leaked severely because the disc was not seated properly.
As a result of this inadequate Part 21 evaluation, Dresser has failed to notify the NBC or it's custmers of this mportable Part 21 item (85-01-01).
Draanar's numerits:
Dresser's evaluation of File No. 84-01 is correct and should not have been reported under Part 21 Dresser contends that: (1) the failure of the disc collar / spindle threads in one safety valve for Diablo Canyon was due to a defective spindle thread, is an isolated event, and is not reportable to other customers under Part 21; and (2) the failure of the disc collar / spindle threads in two safety valves for Davis-Besse was not due to defective parts, but was the result of unusual system vibrations causing physical damage to the threaded joint, which is a system problem at Davis-Besse and is not reportable by Dresser under Part 21.
Although the failure trades for items 1 and 2 were similar, the causes are separate and unrelated. The destruction of the threads is only one of several problems with the safety valves at Davis-Besse due to system vibrations.
Pacific Gas & Electric /Diablo Canyon, Units 1 & 2 The subject safety valve (serial number BN1741) was manufactured in 1971-1973 along with forty-two (42) other valves for PG&E/Diablo Canyon, Units 1 and 2.
The safety valves were shipped to the site and placed in storage because of construction and licensing delays. In May 1984, twenty-two safetys were shipped to Wyle Laboratories, Huntsville, Alabama, for blowdown testing.
-Continued-
l f
Docket No. 99900054/85-01 il Page 2 Reported Violation (85-01-01) .I i
i on 5/15/84, safety valve serial number BN1741, on the first actuation at- '
approximately 1065 psig, lifted with simultaneous shearing of the disc '
collar / spindle threads and cotter pin. On closing, the disc holder stayed in the full open position as the disc moved downward with the spindle out of the i pocket in the dise holder. We disc and spindle deflected sidewise with the disc becoming wedged between the nozzle seat area and the bottcm of the disc holder. Thus, the valve became mechanically janned where it could not open again.
i Subsequent inspection following disasserrbly indicated that the threads on the spindle were defective (split threads) as manufactured by Dresser, so that l
proper thread engagement between the disc collar and spindle did not exist. The i threads in the disc collar were acceptable. The split threads on the spindle were easily detectable by visual inspection. No other defects were noted during the inspection.
he remaining twenty-one safety valves were disassembled and inspected. No '
other spindles were found with defective threads. All valves were then blowdown tested and returned to site. Both Bechtel Energy Corp., and Pacific Gas &
Electric Co., witnessed all inspections and tests. Twenty of these safety valves are now in service for the first time on Unit 2. It is Dresser's understanding that the remaining twenty-one safety valves in storage at the site for Unit 1 were either to be seat to Wyle Laboratories for refurbishment and blowdown testing or were to be disassembled and inspected on site.
All other nuclear power stations excluding Davis-Besse, having safety valves of the same type and design have been operational for several years without reported failure of these safety valves during overpressure conditions.
Considerable changes have occurred in Drescer's quality inspection since 1973 and we conclude that additional action is not warranted.
Therefore, we conclude that the failure of safety valve IN1741 belonging to Pacific Gas & Electric Co. for Diablo Canyon was due to a defective spindle i thread and is an isolated event. ,
However, Dresser's 10CFR21 File No. 84-01 was reviewed. We agree that it was confusing in that we did not clearly state why "the failure is an isolated case". Engineering Report SV221 was revised, adding a sumary and Sketch 5 l
relating to thread dimensions. A copy has been placed in File 84-01. A copy is l attached.
j Toledo Edison Co/ Davis-Besse i
In August 1981, safety valve serial number BM8635 was returned to Dresser. The disc holder was found in full open position with the disc locked between the nozzle seat area and the bottom of the disc holder. Thus, the safety valve was locked in a closed position.
-Continued-I
Docket No. 99900054/85-01 age 3 Raported Violation (85-01-01)
I Subsequent disassembly and inspection indicated that the safety valve was '
severely damaged, as detailed in the attached reports. j l
In September, 1985, during blowt'own testing at Wyle Laboratories, safety valve ;
serial number BM8634 failed after 65 cycles in the same manner as previously i described. This safety valve had been in service prior to blowdown testing.
Dresser's 10CFR21 File No. 85-04 relating to Davis-Besse is still open and is sdieduled for completion on 2/28/86. Preliminary investigation indicates that the system vibrations are considered responsible for these failures. Valve parts are adequate for normal loads and were not defective as manufactured by Dresser. A formal report remains to be written which will include a recomnended i i
preventive maintenance program for Davis-Besse.
I f
f 1
i I
l l I l
i T
' i 1
1 i
_ j; Il
~ l
~
DRESSER v.- <
INDUSTRIES Revision 1 IN DUSTRI AL VALVE O P E R ATIO N S 2 B OX 14 3 0 2 ALEXAN DRI A. LOUISt AN A 713 01 TEL 2:a eac asso 2 Tws: 51097as733 TELEX: se4423 C a m LE. Divio ,
Docket No. 99900054/85-01 Reported Violatio.1 (85-01-01):
PE000Cr ENGINEERING REPORT SV-221 3707RA DISC COLLAR FAIWRE PACIFIC GAS & ELECTRIC COMPANY DIABID CANYCN UNIT 2 Prepared By: W kb 1 R. S. Huf f:nerf Date !
Sr. Product Engineer Reviewed By: I
- 7 #I -
~
F. P. Bolger 'Q Chief Product Engineer Date
~
1 l
Approved By: [}p'l s Y.,A. Lai '
m
'/(
?
-' s / if _5h Date Director, Product Engineerirg 4
Page 1 of 16 9
CON 3OLIDATED I HANCOCK I' DEWRANCE I
i SV -221 Docket No. 99900054/85-01 Reported Violation (85-01-01):
Revision Description 0 File SV-221 was set up containing letter dated 7-01-84 and sketches 1 thru 4.
1 To clarify cause of failure, a formal report was set up adding (1) cover page, summary and sketen 5.
1 1
1 l
l l
l 1
l 1
i
s Engineering Report SV-221 Dodtet No. 99900054/85-01 i
Page 2 of 16 Reported Violation (85-01-01):
- 1. SUIGU M
.' 1.1 %e failure of the disc collar / spindle threaded joint was due to defective (split) threads on the spindle (see Figure 5). Wis -
event is considered an islated quality problen because all other valves shipped in the sane time frame have been in service for many years without reported failure. We subject valve was
\ manufactured in 1973 and placed in storage at site prior to j testing at wyle Laboratories.
I 1.2 This event is not equivalent to the disc collar failure at Toledo Edison, Davis-Besse which was due to system vibrations causing physical damage to the threaded joint.
1.3 The subject safety valve (serial ntaber m 1741) was manufactured in 1971-73 along with forty-two (42) other valves for PGEE/Diablo Canyon, Unito 1 and 2. T:e safety valves were shipped to the site and placed in storage because of construction and licensing delays. In May,1984, twenty-two ..M) safety's were shippe! to Wyle Laboratories, Huntsville, Alaonna, for blowdcwn testing.
1.4 On 5-15-84, safety valve serial unber m 1741, on the first actuation at approximately 1065 psig, lifted with simultareous shearing of the disc collar / spindle threads and cotter pin. On closing, the disc holder stayed in the full open position as the disc noved cbwnward with the spindle out of the rocket in the disc holder. The disc and spindle deflected sidewise with the disc beccning wedged between the nozzle seat area and the bottom of the disc holder. Thus, the valve became mechanically jarened where it could rot open again (see Figures l' thru 4).
1.5 Subsequent inspection following disassembly indicated that the threads on the spindle were defective (split threads) as manufactured by Dresser, so that proper thread engagement between the disc collar and spindle did not exist. The threads in the disc collar were acceptable. We split threads on the spindle were easily detectable by visual inspection. No other defects
, were noted during the inspection (see Figure 5).
1.6 Se ruaining twenty-one (21) safety valves were disassembled and inspected. No other spindles were found with defective threads.
All valves were then blowdown tested and returned to site. Both Bechtel Energy Corporation and Pacific Gas and Electric Company witnessed all inspections and tests.
valves are now in service for the first time on Unit 2.'Nenty (20) of these safety It is Dresser's' understanding that the remaining twenty-one (21) safety's in storage at the site for Unit I were either to be sent to Wyle Laboratories for refurbishnent and blowdown testing or were to be disassembled and inspected on site.
l Engineering Report SV-221 Docket No. 99900054/85-01 Page 3 of 16 Reported Violation (85-01-01):
1.7 All other nuclear gnwer stations, excludirs Davis-Besse, havig safety valves of the same type and design have been operational for several years without reported failure of these safety valves durirg overpressum conditions.
1.8 Considerable changes have occurred in Dresser's quality inspection since 1973 and we conclude that a$ditional action is not warranted.
1.9 Therefore, we concitrie that the failum of the safety valve BN 1741 belongiry to Pacific Gas and Electric Company for Diablo Canyon event.
was due to a defective spirdle thread ard is an islated F
b
v l r --
--- - - J Cover ,, ate Lug , a Dodtet No. 99900054/85-01 Reported Violation (85-01-01):
Disc Holder N
( d Spindle .
N N
l
\ m
\_
Cotter Pin - p R .______
t-- . . - _ _ _ _
.4 Disc Collar - --
I l
-4
/
\
w
% /
I Disc
~
. ~
l lA_
j Figure 1 - Valve in normal Nozzi,deat M '
closed position w
Docket No. 99900054/55-41 -~
Reported Violation (85-01-01):
.k s Y
- fi 4 Cotter Pin ( N - '
6
--____4
. _ _ _J p _ . . . _ .
Threads
/
[ 7 N.
_ . . . . I i
)
+
l+'
Figure 2 - Valve opens with simultaneous failure a of cotter pin and disc cotter / spindle threads.
g i s
r -
e nrw*st No. 9990Aa4/BS-01 Reported Violation (85-01-01 ):
.LN *
.. - s i
N
~
I i
e I
/
i i
%N Figure 3 - Disc moves downward with ,
the spindle out of the disc "
holder pocket. The disc and spindle are now free to deflect sidewise.
I T
N
4-:.2 ..
9 Dodtet No. 99900054/85-01 - - -
l Reported Violation (85-01-01):
N j i
i s
i 1
{
6 l
I i
q
\
i i
)
(
4 g
1 i . . . ;
l s I
l
-. r' f
, . . . . . L~.'.':
, Figure 4 - Disc and spindle , _ (7 \ 1 deflects sidewise. ,
N Disc becomes trapped
- between the nozzle and N
\
the bottom surface of the s \ s.
disc holder. N ,
'5g' I
\ \
\ 'N I
's x.
- - - - - - - - , . . , - . - - - - , , , , , -.-- . ~ . . - . . . - - - - - , - - , , , , - , , , - - . . , . - - . , ,.,,n ,.-- _ _ , - ~
1
~
Dodtet Noa 99900054/85- S V-221 Reported Violation (85-01-01):
)
l 2-16UN-2 Thread Theoretical Thread Profile ,
Disc Collar Visual Evidence of Dwg. VJ800 Split Threads e
d /
s' \
Spindle
's ' Dwg. 442504 s
t
/
s' l
. 02835" approx.
on disc collar' after failure ( cr ests are ' -
slightly torn) . g i
I Note 1: Height for the external thread per ANSI B 1.1 is 0.027 to 0.032. Therefore, disc collar thread was <
correct as manufactured.
I l
l Figure 5 Disc Collar / Spindle Threads After Failure
1
~
Dodnet No. 99900054/85-t,.
Reported Violation (85-01-01):
'IO: R. A. Cedel DATE: July 2,1984 DIERSSER D00STRIES, INC.
FROM: R. S. Huffman
SUBJECT:
DISC COLLAR FAILURE Industrial Valve Division 3707RA SAFETY VALVE COPY 'IO: F. P. Bolger B. G. Brunson J. P. Watz I
Possible 10 CFR Part 21:
In May,1984, the main steam safety valves from Pacific Gas & Electric /Diablo Canp n 2 were sent to Wyle Laboratories, Huntsville, Alabama, for blowdown testing. These valves were Dresser type 3707RA.
On 5/15/84, valve serin number BN1741, on the first actuation at approximately 1065 psig, the valve lifted with simultaneous shearing of the disc collar / spindle threads and cotter pin (see Figures 1 & 2). On closing, the disc holder stayed in the f;. open position as the disc noved downward with ,
the spindle out of the pocket in the disc holder (see Figure 3). The disc and '
soindle deflected sidewise with the disc becm.ing wedged between the nozzle seat area and.the bottom of the disc holder (see Figure 4) . Thus, the valve became mechanically ianmed where it could not open again, although it was leaking severely because the disc was not seated properly.
Following the cynpletion of the test program, the failure was verbally reported to Mr. R. Cedel and Mr. B. Brunson on May 23, 1984. This letter follws up that verbal report.
This is the second reported failure in this node. 'Ihe failure occurred in August,1981, at Toledo Edison / Davis-Besse. Report is attached.
{ t R. S. Huffman Sr. Product Engineer RSH/sc l
l 3
Dodcet No; 99900054/85-01.
Reported Violation (85-01-0*(e Industrial Valve inter-Office Operations Correspondence re, (R.S. Huffma'n $
carte October 22, 1981 F Aov. R.D. Walsh, Jr.
coevrom F.P. Bolger sus.ccr a Maxiflow Spindle Collar - Nuclear Service A lift stop collar, Figure 1, and a cotter pin, Figure 2, were sub=itted for metallurgical evaluation by Rolland Huffman. It was reported that .
when the valve was disassembled, the cotter pin pieces were found en top of the guide and that the collar had to be cut off of the spindle.
The visible evidence strongly suggests that the collar was repeatedly impacted upon with extreme force. i This then caused the collar's threads to be stripped and subsequently, the corter pin was sheared. The basis ;
for this conclusion is described below.
l Figure 3 is a profile view of the lif t stop collar. It was noted that the top portion of the collar has been flared outwards. This is indicative of nu=erous i= pacts of coderate to extreme magnitude. l l
The eventual result of the impacting is docu=ented in Figure 4. Here the threads are shown to have sheared and been flattened from the impacting.
Once again, many cycles are indicated by the smooth nature and regional variation of the damage. l There is also evidence that the cotter pin was shea:ed oft in the process of thread failure. The indications of this behavicr are; a straight pin !
length equivalent to the stop collar I.D. and a marked " ironing" of one pin end. l See Figures 2 & 5.
The pieces of the pin were probably then
" blown" onto the guide shelf by discharging =edia. )
As in any failure analysis, the caterial properties were examined. It was found that the cotter pin was cold formed 304 stainless steel, and the collar itself was cast 430 (C3-30) stainless steel. Both the chemistries and hardnesses of the co:ponents were correct and the substantiating test results are displayed in Figures 6 & 7.
In conclusion, installation irregularities are considered responsible for this f ailure.
These parts are designed to be able to acco=odate i= pacts.
However, the degree cf defor=ational damage is not nor=al for this assembly, and gives witness to the magnitude of the pounding they received. An ex-a=ination of the customer's operating and setting practices, with appropriate corrective action, should eliminate similar occurrences.
, c 9 , J.
R. D. Walsh, Jr.
eds
Docket No. 99900054/85H,.
Reported Violation (85-G1-01): I l
i i
FIGURE 1 Lift Stop Collar Exhibiting Thread Damage (Component Was Cut Off Spindle For Examination)
-=-
'N F
Il llIl illilll.IIlIl1lllll !!!ilI. ((! ,
IflCHES l '1 I l' '2{ ' <'
~
FIGURE 2 Cotter Pin Pieces Pieces Were Found In Upper Recess Of Valve i
f
. , . . - _ _ ,.,w.m.,
~
Dcxiet No. 99900054/85-6.
Reported Violation (85-01-01):
I!Illji!Ml.iIdjjilii!l!-
INCHES I ' 1l l I i l'
l l i
r.,
4
' ,i, p .4 41 f, ' i ,
i I
.- a FIGURE 3
-Lift Stop Collar Profile View Note " Flaring" On Left End Lue To High, Repeated Loading i
1
~
_ '.b
~~ W:5::A.,."y ' W 4. $r'N i.'h mae
=
P 1
FICURE I.
Spindle Collar Threads Stripped Due To High Loading Exhibited in Figure 3 i
--- -- - - - - - - - , - ~ - - - - - . - _
Dodt;t No. 99900054/854 Reported Violatim (85-01-01):
t y..
- , _.yg). ., .
c;; ~ ; - . . r, ._ L . -
-. ~ . . s,. e .. r .
g.;..
' 4r' 2.)_hy Q .
L4-C 4
- e. .,-
7
'4 ~
i E [ \ P. I g *.
. . . _ _ _ . _ 'l,e.,, , . .
..~% ; 3,
~
k I
?
r
. . _ _ _ _ _ _ _ _ _ _ . .i FICL'RE 5 Cotter Pin Segment Shows " Ironing" Effect Of Stripped Lift Stop Collar 8
{
l 9
- 1I
~ 1 1
1 Docket No. 99900C54/85-01 Reported Violation (85-01 -01):
1 I
KEl . 'E.'-l ~
h f ! +..
< ' r i L -i.e.'sH F ,
. 4 i i t -i. s..i 4 1i t a t z :z t t 4 2 211: s z t
- t t tittatsattt:4AttAat
. 't
,t 1
- L j t
- s. < 19.V. ., ,9; , j ,
,iI e I: 1
- - r:4 B.
, ;5 ,.;_ p, l
- i. u t.. u
>4t .. . :. ..
o
,,1 t '. i
- I u,
Etts szts t s22 x:tt ts ss sausag ssgggggggggs i
FIGURE 6 FIGURE 7 Cot ter Pin Analysis Collar Analysis 304 Stainless Steel CB-30 (430) Stainless Steel Hardness = 27Rc Hardness = 87R B
Dodtet No. 99900054/85-t
- Reported Violation (85-01-01):
INSPECTION REPORT valve: 3707RA Serial No.: BM8635 The valve was disassembled in the Clean Room on Wednesday, September 30, 1981. The disassembly was witnessed by the writer.
The disassembly was documented on video tape. *:o customer witness.
Valve Data: Serial No.: BM8635 Nameplate Set Pressure: 1070 psig Base Heat No.: HT8129 Valve Type: 3707RA-RT21-XLP Utility: Toledo Edison Company Site: Davis-Besse Sequence cf Disassembly and Comments:
- 1. Visually examined valve. The following was noted:
- a. ~he seal wire for the ring pins was broken.
- h. The cap assembly and release nut were not widi the valve .
- c. On the yoke, opposite the valve outlet, were hammer marks.
- d. Body.was a casting.
- e. Cotter pin hole for the release nut was near the end of the spindle. (Through the second thread.)
f.
Threads on spindle on one side above the compression screw were flatten.
g.
Adjusting rings were not #rozen but rattled easily.
Rinc cins were engaged.
Dodcet No, 99900054/85-( i Reported violation (85-01-01):
h.
The disc holder was in full open position and against the cover plate.
- i. The disc was on its seat in closed position.
- j. The spring was extended in closed position.
k.
The vertical lengths from the_ top of the yoke (the surface immediately below the head of the compression screw l to the top of the compression screw and to the end of the spindle were 1-29/64" and 4-1/16", respectively.
Measured with scales.
- 2. Loosen the compression screw locknut. Installed the jacking device on the valve to remove spring load from the upper yoke rod nuts.
Removed apper yoke rod nuts and yoke assembly.
The following was noted:
- a. Spindle threads at the spindle / compression screw inter-face were severely flatted (primarily on two sides) .
b.
The icwer inside diameter (spindle /ccmpression screw interface) was severely galled.
- 3. Removed the tcp spring washer and spring. The following was noted:
- a. Metal chips were scattered on the top surface of the lower spring washer (this would be within the spring inside cavity) and the top surface of the cover plate.
Most chips were singular and crescent in shape. How-ever, one chip was a fine wire, spiral shaped 8
piece.
This piece was saved for analysis.
b.
When the spring was removed the disc holder dropped to closed position.
~
Docket No, 99900054/85-0 Reported Violation (85-01-0.,:
- 4. Removed the bottom spring washer. Tne radius of the spindle (the spindle / bottom washer bearing surface) was slightly galled.
- 5. Removed cover plate bolting and cover plate. The following was noted:
a.
The head of the spindle (cover plate / spindle interface) was severely galled (all ar ound) ,
b.
The spindle was a two-piece spindle. There were two drive pin holes in the spindle head. The drive pin was in the lcwer hole. The threads on the spindle stem above the head was excessively long. The assembly was tight.
c.
l Two fragments of a cotter pin (approximately 1-7/8" long) l l
were found in the crevice between the cover plate inside diameter and the upper outside d;2 meter of the guide.
Both fragments were laying in the crevice side by side.
- 6. Removed the spindle / disc / disc holdet assembly. Attempted to unscrew the disc from the spindle, but could not. Noted the following: 1 1
1 a.
No scoring of the disc holder o.D. nor the guide I.D. was noted.
But oblong rust color water spots existed on i
these surfaces. 1 b.
The nozzle seat was scored (large circular marks across l the seat).
These appeared to be impressions left by the disc seqt but were,off center. Estimated four sep-arate impressions.
The seat gave no evidence of valve chatter.
- c. No visual damage to disc seat.
Dodest No. 99900054/85-01 Reported Violation (85-01-01):
- c. Disc holder. The bottom surface of the disc holder had a circular imprint where the disc holder contacted the backside of the disc. The disc holder / disc collar interface in the disc holder _was brinelled.
- d. Disc. The disc pocket bearing surface indicated that the disc was of f center. Threads could not be inspected because they were removed during the cut-off operation.
l
& b h/
Ff R. S. HUFFMAN Sr. Product Engineer RSH:rsl I
l l
l
)
l l
8 1
\
l
Dodcet No, 99900054/B54
.. Reported Violation (85-01-0)):
- d. The disc collar was flopping on the spindle. This al-lowed the disc holder to move upward away from the disc.
- e. The lower adjusting ring was 8 notches belcw the nozzle seat.
- f. The vertical dimension from t_he underside of the guide flange (guide / body interface) to the bottcm of the upper adjusting is 6.860". Measured with dial calipers.
The vertical dimension from the guide / body interface in the body to the nozzle seat was 7-1/32". Measured with scales. The delta value is 7.031-6.860 = .171".
- 7. Cut the spindle / disc / disc holder assembly apart. Cut the i l
disc collar with a oxygen-acetylene torch. Crack in disc holder resulted from this operation. Removed the disc by machining. The following was noted:
- a. Disc collar. The inside diameter was severely upset.
All threads were hammered to a flat surface (no thread profile). The upper portion of the inside diameter showed some evidence of the criginal threads, but the lower portion was a smooth surface. The bcttom surface was upset and flared to the outside.
- b. Spindle. The relief diameter irmediately passed the first thread on the spindle tip was imprinted with the disc threads at one point. The first threads on the spindle were damaged. The second threads on the spindle tip (disc collar / spindle interface) were severely flattened.
1 l
I
.. 3 NONCONEORMANG: (85-01-02)
Dresser obtained calibration services and certifications of furnace thermocouples from unapproved verx3 ors.
CAUSE:
Purchase orders for thernoccuple wire was pla d with approved sources, Honeywell, Baton Rouge, IA who in turn, purchased from other Honeywell <
facilities. !
ACTION TO CORRECT DEFICIENCIES:
Honeywell, Corp. Drive, Houston, TX, was audited, approved and added to i Dresser's Approved Vendor List. Orders since that date and future orders ;
will be plamd directly fran this approved supplier.
i AcrION 10 PREVENT RECURRENCE:
Calibration mrtifications are being reviewed cn receipt to assure the calibration facility adiress agrees with the Approved Vendor List.
l
\
l l
i l
1 .
NONO N ORMANCE: (85-01-03)
Four containers of type 7018 nuclear weld rods, lot 3C504Yoz, heat 76175, were not stanped (or verified) by the Receiving Inspection personnel.
l CAUSE:
1 The four cans of weld rod were purchased for non-Section III use and was I
inadvertently placed in the storage area assigned for Section III weld rod. The area assigned for non-Section III weld rod is located adjacent l to the Section III weld rod storage area and as the rod was noved from Receiving Inspection to storage, it was placed in the Section III area.
ACTION TO CORRECT DEFICIENCIES:
The rod was noved to the cntreet cage.
. ACTION 10 PREVENT RECURRENCE:
The Section III eld rod storage area will be noved to a different location to prevent recurrence.
N ME: The shop traveler for Section III orders requires the inspector to record rod heat and lot numbers an the traveler and obtain Quality Assurance approval prior to use; therefore, the rod would not have been used on Section III orders.
l l
l I
i
_ l NONCJNFOINANCE: (85-01-04)
As of October 4,1985, there was no objective evidence that calibration of the Charpy V-Notch Impact Testing Machine has been performed once in each six-month interval, during the last five years.
CAUSE:
The impact machine has not been in use.
ACTION TO CORRECT DEFICIENCIES:
The inpact machine has been tagged indicating calibration required prior to use. grI-33 has been revised to require calibration prior to use.
ACTION TO PREVENT RECURRENCE:
This machine has been added to Dresser's QAR-5 " Monthly Report of Equipment Calibration" to assist in highlighting calibration requirements.
i I
i 6
eme-NONCONFOINANCE: (85-01-05)
As of October 4, 1985, calibration certifications #1343 and 6141 had no documentation of a calibration standard serial number. Certification i
- 1343 indicated calibration to 50% tolerance of full scale reading instead of the specified 5%.
CAUSE: )
These were errors on tr. e part of the technicians completi'ig the 3 report.
ACTION 'IO CDRRECT DEFICIENCIES:
Wrench S/N 1343 was recalibrated 11/26/85 with acceptable certification.
Wrench S/N 6141 is no longer in use.
1 ACTION 'IO PRFVENT RECURRENCE:
(1) The technician and supervisor were made aware of the errors.
(2) The OA Engineer will review each report in detail and indicate acceptance by stamp and date.
l 8
)
l
. l i
NON00NEORMANCE: (85-01-06)
As of October 4,1985, there was no objective evidence that calibration i for WR-12 Carbon Determinator (models 761-100 and 761-200 (BCD) is trace-able to National Standards on equipment manufacturer's recomended stand- i ards. The WR-12 Carbco Determinator is used to determine carbon and sulfur content.
CAUSE:
Proper standards were acquired with the equipment; however, w fallect to follow up assuring certification to National Standards.
ACTION 'IO CORRECT DEFICIENCIES:
Certifications were obtained from Leco ard are on file.
ACTION 'IO PREVENT RECURRENCE:
Future standards will be purchased with proper certifications.
This equipment has been added to Dresser's QAR-5 "Mcnthly Report of {
Equipment Calibration", to assist in highlighting calibration requirements.,
b l
'lI t
l NONCONEORMANCE: (85-01-07)
As of October 4,1985, there was no objective evidence that calibration of nuclear welding rod men bi-metal thermometers ( W 2, 10-4, TG-6, IG-8, W11, TG-20 & W21) was carried out in accordance with QACP 3-1.
OBSERVATICN:
This appears to be an erroe in documenting audit notes. Only four of these thermometers are used in the Section III rod ovens and according to our records, were within the required calibration schedule.
NTE: This subject was discussed with Mr. Zedt and he advised us to respond and explain our position as above.
l l
l l
l
-_ . -._= . --- . _ . . . .. - .. .
I l
Nonconformance: (85-01-08)
Dresser's Part 21 file no. 85-01 did not identify / list the required pertinent data (85-01-08).
Cause:
Error on the part of Dresser's evaluation team to identify / list pertinent data for Dresser's 10CFR21 investigation file number 85-01.
i Action to Correct Deficiencies:
, Records are being generated to identify / list pertinent data for this investigation of 3050 diaphragm valves.
Action to Prevent Recurrence:
) '
Evaluation team members have received latest copy of Dresser's Part 21 procedure which contains above requirements to identify / list pertinent data j
in paragraph 3.2.8.
1 w
2 i
i 1
1 v
1 l
l 8
t .
)
i i
)
i
.