ML20236R648
| ML20236R648 | |
| Person / Time | |
|---|---|
| Site: | Prairie Island  |
| Issue date: | 07/15/1998 |
| From: | Rao G WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML20236R629 | List: |
| References | |
| NUDOCS 9807220178 | |
| Download: ML20236R648 (14) | |
Text
. - - _ _ _ _ _ _ _ _ _
Westinghouse Non-Proprietary Class 3 MetallurgicalInvestigation and Root Cause Assessment of Part Length CRDM Housing Motor Tube Cracking at Prairie Island Nuclear Generating Plant Unit 2 Preliminary Summary Report Gutti V. Rao BACKGROUND AND INTRODUCTION L
On January 24,1998, Prairie Island Nuclear Generating Plant Unit 2 was brought to a cold shutdown condition due to the detection of a small (0.31gpm) unisolable RCS leak at one of the part length control rod drive mechanism (CRDM) housing motor tubes. The leak was I
contained in the G-9 motor tube lower weld (G-9A)* at a location slightly above the lower l
canopy seal weld. RT and UT examination of the tube confirmed the presence of l
circumferential indication (s) along the bimetallic weld corresponding to the leak location. A l'
24 inch long segment of the tube containing the leak region (Figure 1) was sectioned out and l
shipped to Westinghouse hot cell facilities for metallurgical investigation. Subsequently, five l
additional tube segments containing similar welds from three part length CRDM motor tubes l
from the unit were shipped to Westinghouse for examinations. These included upper weld in l
C-9 tube (G-98) and lower and upper welds in G-5 and E-7 tubes (i.e., G-5A, G-5B, E-7A and E-7B). The motor tube is made of Type 403 martensitic stainless st eel material and is joined at either end with Type 304 austenitic stainless tube with full penetration bimetallic welds. The bimetallic weld consisted of a 0.25 inch thick Type 309 stainless steel buttering applied to the Type 403 stainless steel tube followed by Type 308L, full penetration weld. The drawing requirements stipulate a postweld heat treatment following buttering (Ref.1). This was not verified pending access to the original records.
This report summarizes the preliminary findings of the metallurgicalinvestigation of cracking and leakage of the G-9A motor tube lower bimetallic weld. The investigation was centered on j
l the 24 inch motor tube segment containing the leaking weld (G-9A). Examinations also 1
included the remaining five other welds from G-9, E-7 and G-5 part length CRDM motor tubes from the unit. The metallurgicalinvestigation included the following major tasks:
I Surface Examinations
)
Dimensional Measurements e
Non-destructive Examinations Sample Sectioning Metallographic Examinations e
7
- The lower and upper welds in the motor tube are designated respectively by 'A' and 'B' for the sake of current discussion in this report.
Reference 1: Royal Industries " Motor Tube Center Section Drawing No.121E005-1," August 1968.
9807220178 980715 DR ADOCK 0 2
1
f l
Westinghouse Non-Proprietary Class 3 i
Fractographic Examinations Chemistry Evaluations l
Mechanical Property Measurements Review of Available Fabrication History Records Residual Stress Measurements The overall purposes of the investigation were to establish the mechanism and cause of the motor tube leakage and further to develop information that would be helpfulin developing corrective actions. A summary of the evaluations and results follows.
EXAMINATIONS AND TESTS Surface and NDE Examinations The as received surface condition of the G-9A tube sample was examined visually and by light optical microscopy techniques for evidence of cracking, deposits, surface attack, and/or other mechanical distress marks. The examinations were carried out on the inside diameter (ID) and outside diameter (OD) surfaces non-destructively and the results were documented digitally by an electronic recording system.
Boroscopic examinations were carried out on the ID surface prior to any sectioning to identify the presence and location of the cracking. UT examinations were conducted (by ABB personnel) employing procedures similar to the ones followed for the on-site tube inspections.
For the purpose of identification of surface locations and samples, a consistent orientation location procedure, looking down on the motor tube from above the vessel head, was established. A site marking noted on the tube identifying ' North' direction was designated
'zero' degree clock location for this purpose. Based on the results of the surface examinations and the NDE examinations, initial sectioning of the tube was carried out to facilitate a more detailed ID surface examination of the tube in the leak region. The initial sectioning consisted of a transverse cut right above the lower canopy seal weld and axial sectioning to remove a 120' circumferential segment of the remaining tube diametrically across from the leak region.
This corresponded to axial cuts at 60' and 300* clock locations. The leak location in the tube corresponded to 185* to 190' clock location. Closer ID surface examinations were carried out by light optical microscopy after initial sectioning to identify the presence and location of cracking more precisely and to define locations for additional sectioning. The results were documented by digital photography.
Surface examinations were also carried out on the G-9 tube upper weld sample (designated G-98) and G-5 and E-7 upper (G-5B and E-78) and lower (G-5A and E-7A) samples for evidence of cracking if any.
i n\\4139 doc 063098 2
Westinghouse Non-Proprietary Class 3 Dimensional Measurements and Sample Sectioning a)
C 9 Tube 1
Prior to and following the initial sectioning, dimensional measurements were i
performed for comparison with the design specifications and further to identify any l
post-sectioning displacements following sectioning that would provide a measure of locked-in residual stresses. An axial sectioning procedure of the tube was then followed based on the results of the surface (and NDE) examinations. Six axial strips, each approximately 0.5 inch wide, were removed at 30 to 60 intervals for metallographic examination. The tube circumferential segments connecting the axial strips containing the crack were opened in the laboratory and were employed in fractographic examinations with the exception of two samples. The segments extending between 0 and 60 location and the segment between 250 and 300* clock location were saved for the purposes of utilizing in the NDE calibration and in machining of mechanical test samples respectively.
The G-9B and the G-5A, G-5B and E-7A, E-7B motor tube welds were sectioned axially at 0*,
120,240 clock locations for metallographic samples.
Metallographic Examinations Light optical metallographic examinations cf the G-9A weld and other welds were performed on the axial strips. The metallographic examinations were carried out in the 'as-polished' and i
l
' polished and etched' conditions. The purpose of the metallographic examinations is to identify the presence, location and depth of cracking, material microstructure, the cracking morphology, and its relation to the local microstructure. The metallographic examinations also established the buttering and weld deposit morphologies as well as the presence of any weld repair.
Fractographic Examinations l
Fractographic examinations of the freshly opened circumferential cracks from the G-9A tube weld were conducted to establish crack initiation sites, propagation directions and the general j
fracture morphology The fractographic examinations were conducted by light opticaland scanning electron microscopy techniques, l
Chemistry Evaluations l
Chemistry evaluation of the ID surface deposits from G-9A tube in the as-received condition were analyzed by energy dispersive x-ray analysis to examine for evidence of coolant j
contaminants that may have played a role in the cracking process. Chemistry analysis of crack deposits at the crack-tip locations was also conducted by energy dispersive x-ray analysis to identify the role of any contaminants in the cracking process.
Auger (AES) and electron spectroscopy chemical analyses (ESCA) were performed on the oxide deposits of the fleshly opened cracks to assess the composition, stoichiometry, kinetics and the I
l l
l o:\\4139. doc-063098 3
L.--.
Westinghouse Non-Proprietary Class 3 l
temperature of oxide formation. Elemental chemistry profiles across the crack in the weld
[
dilution region were established by electron microprobe analysis to examine if the weld L
dilution and /or weld metal composition played a role in the cracking. Bulk wet chemistry l
analyses of the type 309 weld buttering and type 308L weld deposits were performed. The bulk chemistry analyses examine if the weld metals met the specification requirements.
t L
Mechanical Property and Residual Stress Measurements Vickers hard ness traverse measurements were conducted across the weld interface on polished j'
sections of the C-9A weld to assess the approximate strength levels of the base metal and weld j
l metal regions. A segment of the tube (between 250* and 300* clock locations axiaD containing
).
the crack was sectioned into tensile and bend specimens to establish the strength and toughness properties of the remaining ligament. Weld residual stress measurements were conducted on C-5A tube by hole drilling technique and monitoring with strain gage rosette as per ASTM E837-9 procedure.
j Review of Design and Fabrication Records l
The vendor supplied design and fabrication records were reviewed to examine the fabrication l
and inspection sequence employed for the CRDM housing tubes and in particular the C-9A l
weld.
l Mechanistic and Root Cause Assessments I
Based on the results of the various tests and evaluations discussed above, an assessment of the mechanism and cause of cracking was developed These assessments also included l
consideration of the results of the UT examinations conducted at site on similar motor tube l
welds in some of the units that are currently in outage and the results of this review of fabrication records. The results of the evaluations and assessments are summarized below.
l l
SUMMARY
RESULTS l
The cracking is located in the 309 SST buttering at the 403 SST base metal interface of the C-9 tube lower weld (Figure 2).
l l
The cracking is initiated on the ID surface and progressed radially outward.
The cracking is circumferential and covered 360 around the inside diameter circumference.
The cracking depth varied from 65% to 100% depending on the circumferential locations. A circumferential through-wall crack of ~0.5 inches length was observed at 190* clock location l
on the OD surface where the leak occurred.
The cracking is comprised of discontinuous segments connected with intergranular/ intercellular ligaments suggesting that the cracking may have occurred during solidification (Figure 3).
o:\\4130. doc-063098 4
Westinghouse Non-Proprietary Class 3 Evidence of a 0.03 in. to 0.07 in. wide band consisting of descrete and freshly fractured ligaments connecting the two mating crack faces was seen extending around the circumference positioned prior to the crack-tip. This most likely may have contributed to additional strength for the cracked penetration.
l l
The fracture face was covered with dark high temperature oxide all the way up to the OD surface where the leak occurred suggesting that the cracking most likely occurred during welding or post weld heat treatment (Figures 4 and 5),
Based on the results of the examinations conducted to date, no evidence of fresh fracture (or i
e dimpled ligament) was seen at the leak location.
The analysis of ID surface deposits showed no evidence of contaminants such as halogens suggesting that coolant chemistry did not play any role in the cracking process. Utis was also confirmed with the crack deposits.
Chemistry analysis results confirmed the presence of residuals and low melting species such as copper, sulfur, zinc, and boron in the oxide layer on the fracture face.
UT examination of five other similar welds in the remaining CRDM motor tubes showed no l
indications.
Metallographic examination of axial sections through the five similar welds in the three motor tubes (C-9, C-5, and E-7) at Unit 2 confirmed no evidence of cracks.
l Additional Examinations at Other Units UT examination results from two other plants obtained recently on a total of at least 24 similar welds confirmed no cracking.
Fabrication Sequence l
l Type 403 stainless steel tube motor tube is joined to its 304 stainless steel end forgings via a full l
penetration weld. The weld materialis Type 308L The welds are subjected to both radiographic, penetrant test, and visual examination in accordance with the requirements of the ASME code in effect at the time of fabrication. The specified (Ref.1) sequence of welding, heat treatment and non-destructive examination is listed below for the C-9 motor tube.
Machine butter preparation on 403 tube PT machined surface before buttering Butter with 309 weld wire Machine butter on both ends PT examine butter surface o:\\4130. doc 063098 5
l
Westinghouse Non-Proprietary Class 3 RT butter weld Heat treat after buttering per ASME Code Case 1337, Part 3 l
i The records indicated a rework on the weld approximately 10 months after the date of original fabrication.
A review of records confirmed that the vendor (Royal Industries) changed the weld deposit material to Alloy 82 applied to the 403 (with no buttering) approximately one year after the original fabrication of the Prairie Island tubes.
l l
MECHANISTIC AND ROOT CAUSE CONSIDERATIONS Inspection results to date in the industry show that the cracking is isolated to only one weld which is the C-9 lower weld at Prairie Island 2.
The cracking occurred at the dissimilar metal n'terface where the residual stresses are likely to be higher, The cracking followed intercellular morphology.
The fracture face was covered with dark high temperature oxide. The weld surface on the ID which was at service temperatures appeared bright and shiny in contrast to the fracture i
face.
I l
Evidence of low melting species such as copper, sulfur, zinc and boron was found in the oxide layer on the fracture face.
The intercellular morphology, evidence of interconnecting ligaments and dark high i.
temperature oxide on the fracture face, stresses due to dissimilar metals and the presence of the low melting point materials suggest " hot cracking" as the likely mechanism although additional contribution (growth) due to " reheat cracking", " constitutional liquation", and " hydrogen cracking" during post-weld treatment cannot be ruled out. In either case, the cracking appears to be originated from the weld fabrication.
l Hot cracking is generally an isolated event, because the welding procedures and qualifications used for nuclear equipment are carefully conceived, and inspections are extensive. This weld region is subjected to multiple inspections by both liquid penetrant and radiography during the fabrication process. While it is clear that the weld crack at Prairie Island Unit 2 was not detected, this is considered a rare event. Inspections of the other part length CRDM motor tubes that have been removed show no evidence of cracking.
The C-9A weld contained a repair on the ID adjacent to the crack. Experience with making a repair with 308 adjacent to 403 indicates that it could be very difficult, especially with limited access ID locations such as the case here. Although normally not expected, it is conceivable as an isolated occurrence, that while making an unsuccessful weld repair the crack in the buttering could be driven radially outward further inside the weld. The 17T and RT results o:\\4139. doc-063096 6
Westinghouse Non-Proprietary Class 3 L
could potentially be masked by the dissimilar metal interface considering the fact that hot cracks are generally very tight.
l CONCLUSIONS
. Based on the overall results of the investigation it is concluded that:
The observed cracking in the C-9A part-length CRDM housing motor tube weld at Prairie Island Unit 2 station is originated from weld fabrication.
l No evidence of additional service growth of the service defect could be identified at the leak location.
Examination of the welds in the other motor tubes at Unit 2, fabricated in the same time period confirmed no evidence of any defects.
(
\\
f l
I i
l l.-
1 l
i l.
o:\\4139. doc-063098 7
FIGURES Prairie Island Unit 2 Part Length CRDMS Evaluation e-i t
I
\\'
Westinghouse Non-Proprietary Class 3 i
l
\\
'\\.
j
\\
i
\\
\\.
l
\\.
\\
~
n
\\,
N I
\\
\\
\\
g
___4 1
\\.
\\
l
\\
'\\
'\\
5
\\
3 u
i
\\
\\
j 4
j Figure 1. Section Of Prairie Island Unit 2 G-9 Motor Tube Weld Received At The Hot Cell For Examination o;\\4139. doc-063098 9
Westinghouse Non Proprietary Class 3 l
l l
I 309 WELD BUTTERING 1\\
308L WEELD 304 403
(
l
\\\\
--- Flaw Location A
3.9" 10 Figure 2.'
Schematic Of Weld Configuration For Part Length l.
CRDMS, Showing The Location Of The Flaw in Prairie island Unit 2 -
' o:\\4139.dx-0630%
10 l
Westinghouse Non-Proprietary Class 3 fa j ig A
- smg,,w[gbj@wy)b n
3.
v r^ % g n_+v en...} xnJI >
I w
%M g$:-~ hp{? A jw.m y-l
-r w
t 48mn
.[bir Y
e4
^
if w
Q@buPoy!yiliy%.h.
me-h
&@dNkf!MId$hp$W((~M%E; h$E d
34
~
3$l}
?
7'
!$Qh M RW 4 4
k I
fang $ B
.,g 403 stainless
[%rMpdex.yQLggggg4 pg#
pp hmq_g.
- rnpw, 444g0b.gkh$s^HTN uye.
WWAN m
y.
s.Jf.ywylg M me w A m.v.
u steel M
em.aW NQ,&eckyh h:
ma "b$q
&nyy >
w ma.3
&Yjkk&e%g% ia+,
pe:cpeMm$ a%.mp,Gpqs
+p,, ~k
+ag w
wp
_g m
% w' -e emera 4 sweep @n wp yge&g gegy Y wa h
hhhh s$a#$1."yggg nu+t,pr;,a.M (9.4[$j@pM9&gkR4e.:.)&q%s?Q,m+
n a.w'.
nn -
k VbMQt k,
QkQfQg%.Q[*pg.s}pf(Qygggn i fyppy ;
Y[
G m e e A n!@ $u'pn,ih M 5 i$nhMd [L@&@
MMMN iii ay@;v[i;4j! kg4>93qdd%M.i %.
t, J.q.
4 m J Li sgi h,M M L d sNL hg q. p;s Q.. p lv,d @ 4,9. d w;WQ, J pr
- p%W i
S-
.m Q
3 ig t
ggs s r
Lfsn191M 2fL hi i
Figure 3. Metallographic Examination Results Of The G-9 Tube Lower Weld Illustrating The Cracking Morphology in The 309 Buttering.
j o:\\4139. doc-063098 l i l.
Westinghouse Non-Proprietary Class 3 i l i 4 l le 'p Leak; Location (185 -190 ) e s wa,., f ' f.\\/ f'. [W !') b> k - l .<.. ? .;.i ?.
- a.,
~ ','{ b: I ' .'.c . jr. .tm ! 6
- en:,x V>
- 1; y
- 3 w
- a l
. ? A ' / ~p e -19/ ~ ]'.7
- '<" :o ' sm, s R
s-
- t..
(; s ..,;:){ _ Q;9 f.. <;;;p' M ca Pj-
- s..,
f '"; ":) ~ k. r Gi:dxW:' < - [' <}. , y,; .:.3. x ?l.t ;.,_ @ 7 N;k:[ / d N M,pN @[sk $M}N Vy 9.3 . %; 9:@4$$g+W.,ma s -Qf:p4 ;w*r:^ :.k,% ffy I b f m)@ d;:;:..p",p>' %m
- 2:
- i t,
l ~ e' / 45
- m. 7 4 e.
4;qQ .r n.o.: .,n~ r
- p
. f4,w q y... 73s - A. a-S. N. 3;c.; ,Qh 4: x s s. <y
- w
.t s ^I '2 - [ Yftf: s y 4@%p w%'. x::P:^. '. L:. F %o y s9m@x au:2--7:.ylm w%s..etc-
- th, J:ai % ??@:r.g@4:. m w g ;sW,%-
41 lve ai nwganyuawc.s 3:s:%:.: e 4 :, 6 M E M e w %y p %r %- u %ss:h 4 M M A $eg g si q w 3 Dx& f.::. 4-amagglgeg k p;w a ds te is 4'N%W Qq c,anaww lllll l "l l l [T
- ^
- , jp '
t 2. 3 I 4 i Figure 4. Light Optical Fractograph Of The Freshly Opened Crack in The Tube Segment (Between 150 And 220 Clock Location) Containing The Leak Location. o.\\4139 doc-063098 12
n I 4 l l Westinghouse Non-Proprietary Class 3 l i l 1 I i l i 1 Degrees i' (x10] 4. .h ,#.. Z d ji ), 4 Saved for UT Cal. Std. l e.#ep. 2 afg ,v i e,.. Y 7 ..y uf3 f, Q~9 7 3 t.c s - 33 /- \\ / / N, N' 0 4 T. g D, ~ f f
- 4. m>-
22 s i / x '; l ; g. 1 ng a l
- 6).;
"f :. 31 ' z 3. s s. N l i,; / 1 , (1 Y 3 ^"l " 4..] .,y,3 7 s, ,7 ? l l,19 -.t.:. ,Nx A,\\,\\ \\ .n g/ : 1 ;;/ / < ,,,--"~g , h; 29.~~.~.. 'N,~g.s.s N 'h Ps \\ j,/, / ~x"'% ; s w, a; u,,,, y
- m. p y
l ~;; .".11.aL -. ~ ~ ~ " ' g l "'4.:; 28 c --.j.. _.._"...__ _ _ l _3...__ g Test Samples ~ y,_.- g \\,, -g ;6;n sy.,.:;' ~4' h x,C., N._,, 3, ~ ~3-27 ? ts ".,, pi /c 7,,. .is '/,,sv /,li \\ \\\\',.)- '% c d 't' / i $>\\ .1: /, - / / ' $4 l d)^ 12 i:. 25 /.T,k w\\h' .,N-f l 1[., /- /. u, e 33 .i .\\ g 7, / [/ '[ y ( \\( \\, + ~ e 14 j g n s 33 22 _i~~. 2i l- .(J 16 w tr 18 tr l "T,.:-
- w. t..
F c.. .+ m "y, - m.: 4.. _.t . ~ v.- .:.w. t l l Figure 5. Macrophotos Illustrating Circumferential Layout Fracture of the Faces Of Freshly Opened Cracks in The G-9 Lower Weld. o:\\4139. doc-063098 13
i 4 i Westinghouse Non-Proprietary Class 3 l 1 h l l 1 l ) i r-i i i i 1. ) I GS i I RNM LWt g 1 l ~ " ~ ,,,,,.....,,......,xtw EI- ~ i-1 m m i 10 % 0% 0 80 100 tto 300 260 300 330 Anste (Deg4 j l i l l Figure 6. Summary Of Fractographics Measurement Data i Showing Depth Of Cracking And Mixed Zone Width Distribution Around The Circumference in G-9 Motor Tube Lower Weld. o:\\4139. doc-063098 14 t}}