ML19347F441
| ML19347F441 | |
| Person / Time | |
|---|---|
| Site: | Crane  |
| Issue date: | 05/14/1981 |
| From: | Hovey G METROPOLITAN EDISON CO. |
| To: | Barrett L Office of Nuclear Reactor Regulation |
| References | |
| LL2-81-0053, LL2-81-53, NUDOCS 8105190278 | |
| Download: ML19347F441 (14) | |
Text
..
s 6
19 Metropolitan Edison Company
^
t-t
[
Post Office Box 480 h ((g
[j Q 9
\\
t Middletown, Pennsylvania 17057 g
s V
0 717 944-4041
[-
7 ~-
,s
@4 May 14, 1981 W
. rg i}
LL2-81-0053
\\-l
'.R'lh Gl/l q
w C3 H
N,r l
Q TMI Program Office
~#
Attn:
Mr. Lake Barrett, Deputy Director U.S. Nuclear Regulatory Commission bE r,
h@
E c/o Three Mile Island Nuclear Station ly Middletown, Pennsylvania 17057 u
^
Dear Sir:
Three Mile Island Nuclear Station, Unit 2 (TMI-2)
Operation License No. DPR-73 Docket No. 50-320 Waste Gas System Summary Report In accordance with an NRC request to evaluate prot,lems with the Waste Gas System prior to and shortly after the March 28, 1979 incident, we have compiled the following report. This report presents an overview of the history of the system and is a synopsis of the previously documented information associated the Waste Gas System. This report also provides an assessment of various problems associated with the system and our solstions to these problems.
The report t reats each of the three major sections of the Waste Gas System, i.e.
the Vent Header, the Gas Compressors, and the Decay Tank and Release Piping as a
~
separate section and further divides each section into their Pre-Incident and, Post-Incident histories.
1.
Vent Header A. Pre-Incident The Vent header portion of the Waste Gas System was hydrostatically tested during the Functional Test Program at TMI-II.
There were no discrepancies noted during the testing.
Af ter being placed in service for plant operation no major problems were encountered. However, during January and February of 1979, all four liquid drain traps in the Vent Header in the Auxiliary Building were found to be leaking.
These traps were repaired, and assumed to be operating properly prior to the March 28, 1979 incident.
Aco/
s
//
~
Metropohtan E hson Company 's a Mem0er of me General Fucic UtSt es Sys ee 810s100 9 7((
l
0 Mr. Lake Barrett f B. Post-Incident During and immediatel:t following the March 28, 1979 inciuent gaseous activity was released from various sources in the Auxiliary Building.
As a result we formed a Task Group to find and correct leaks, including leaks in the Waste Gas System.
Extensive testing was performed which varied system parameters in specifically isolated sections of the system. Results of testing revealed one minor problem with a check valve. The check valve (Wr0-V10), down-stream of the Make Up Tank Vent Valve (MU-V13), indicated sticking.
It appeared that the valva remained in the closed position until sufficient differential pressure existed to force it open. Although an operational inconsistancy, no releases or adverse ef fects were attributed to this problem.
Several attempts were made to determine if the liquid drain 4
l traps were a leakage source. However, due to high radiation levels around these traps, visual inspection and detailed monitoring could not be conducted. Leakage tests performed at a slight positive pressure did confirm the overall integrity of the Vent Header.
II Waste Gas Compressors A. Pre-Incident After completion of Pre-Operational and Functional Testing, the Waste Gas Compressors were turned over to Operations with no "Open" items.
Machinery History reporte indicate two problems of significance af ter the system was in operation. In January of 1979 the "B" compressor developed a metallic scraping sound. In February of 1979 the "A" compressor began to cycle on and of f until it tripped the breaker on thernal over-load.
Maintenance was performed to correct both problems. However, system operation indicated a continuing concern with the "B" compressor due
~
l to an inability to pump down the vent header. This remained an open item prior to the incident.
B. Fost-Incident The rask Group determined that the initial ges leak in the system originated in the Waste Gas Compressors. However, due to the system configuration and radiation levels within the compressor cubicle, it was impossible to determine the precise location of this leak within the compressor portion of the system.
Later testing revealed definite leaks on the "A" Compressor. These leaks were finage leaks which were subsequently repaired. Leakage points are identified in Attachment I to this letter. Major maintenance was performed on the "B" compressor in July 1979 and extensive damage was discovered in the compressor internals. The internals were sent to GPU labs in Reading, Pa.
The results of their analysis are presented in s
t l
\\
Mr. Lake Barrett Attachment 2.
One important conclusion was that the damage was due primarily to cavitation with mechanical damage being a secondary problem.
New parts were installed and maintenance was completed. When the com-pressor was run, it would not pump down vent header pressure and metallic scraping noise was detected. Investigation again revealed internal damage that3h not to the extent found and reported by GPU labs.
The damage was indic&sive of compressor operation with no seal water.
Repairs were made and the compressor placed back in operation.
Seal water was closely monitored and was determined to be at a proper level.
Once again the compressor failed to pump dcwn ent header pressure.
Forther testing indicated a failure of the compressor suction volve, WDG-V833. This valve is an air operated diaphragm valve. The failure
'occured when the diaphragm separated free the valve stem.
As the com-pressor began to pump down the vent header, the diaphragm was pulled down onto the valve seat by the pumping action, which effectively blocked the suction line.
Based on the above in*ormation the apparent sequence for the "B" Com-pressor is as follows; pate Cvent Remarks 1/2/79 Compressor shut down due to Apparent cause was loss of metalie scraping noise.
seal water 2/79 Maintenance complete.
Returned to Operations 2/79 Operations removed from Failure to pump down vent service. Apparently header indicates possible
~
probicms not corrected.
blocked suction 3/28/79 Incident "A" Compressor in Auto-Low.
"B" Compressor OFF.
4/1/79 "B" Compressor placed in All efforts directed to reduce Manual-Run despite pre-header pressure minimizing vious concerns.
gaseous releases from Plant.
5/1/79 "B" Compressor removed from No pumping action probably service due to no pumping due to f ailure of suction action valve WDG-V83B.
(Cavitation confirmed by laboratory analysis.)
5
~-
w y
e-.-
-ne r----
-r y
s Mr. Lake Barrett Date Event Remarks 7/29/79 "B" Compressor tagged out and dismantled 9/14/79 Laboratory analysi:t complete Results confirm two failure modes.
1.
Mechanical due to lack of seal water.
2.
Erosion due to cavitation.
That is, no open suction.
10/5/79 Compressor maintenance New internals installed.
complete.
10/27/79 Compressor returned to Immediate investigation ope ra tion.
Immediate discovered improper seal water metallic scraping noise supply.
and failure to pump down vent header.
11/29/79 Inspection of ecmpressor Confirmed mechanical damage perfo rmed due to lack of seal water.
1/5/80 Parts ordered for repair
' 3/80 Compressor repaired Implemented new seal water requirements Lato procedure.
' 4/80 Compressor returned to service. Indicata4 blocked suction line.
Secured after failure to pump down vent header.
5/21/80 Test conducted to determine Results showed WDG-V83B had problem.
diaphragm broken. Result was a blocked suction line.~~
5/25/80 Parts Ordered For WDG-783B WDG-V83B awaiting testing.
The above sequence clearly indicates two problems with comyressor operation. Inadequate seal water caused compressor overheating resulting in a loss of running tolerance. This malfunction, in turn, caused severe mechanical damage. This was due to faulty level indication on the moisture separator tank. The other problem was a separated diaphragm on suction valve WDG-V83B. Similar problems have occured on other valves of this type. This failure (i.e. blocked suction) was responsible for the cavitation induced erosion in the compressor. Damage by this mode was caused by extended ('4/1/79 to 5/1/79) operation with the suction line
. ef fectively blocked.
a o
w l
l
4 Mr. Leks Barrset III Waste Gas Decay Tanks and Release Piping.
A. Pre-incident The Functional and Pre Operational test on this portion of the system found no significant problems.
Af ter system turn over, several minor problems occured. These had no significant impact on system aperation. Details of these problems are covered in component maintenance history records.
B. Post-Incident The Task Group investigating the system for leaks discovered one source i
I of leakage to be Radiation Monitors WDG-ET-1485 and 1486. Maintenance and/or modifications had been performed on these instruments to facilitate Waste Gas Decay Tank release to the Reactor Building and a cap had been removed on a sensing line. This release point was capped and successfully hydrostatically tested soon af ter discovery. In addition, it was determined that if at any time, the pumps associated with these radiation monitors were not operating and the tanks were being vented to the Reactor Building, a gas leak path exists at the pump seal. Close monitor-ing of the operation of these pumps during a release has been incorporated into the operating procedure for the Waste Gas System.
IV Summary and Conclusions The current status of the Waste Gas System is as follows;
- 1) The system is in a normal line up.
- 2) "A" Compressor is in service.
- 3) "B" Compressor is operable but out of service due to problems with the suction valve (WDG-V83B).
- 4) Waste Gas Decay Iank (A) is in service and pressurized to 26 psig.
Based on our evaluatton of the problems associated with the Waste Gas System we have determined that the root cause was not a defect in the compressor but was a coibination of valve failures and a failure to maintain proper com-pressor seal water. In order to protect the Waste Gas System f rom future failures of this type changes have been implemented to the start-up eaction of the Waste Gas System operating procedure. These changes ensure that we verify that the compressor shaf t can be rotated by hand, the moisture separator tank Hi/ Low level alarm is clear, and that the mositure separator tank preesnre reaches proper pressure to ensure adequate seal water flow within 15 seconds of starting the compressors, thus increasing system reliability.
Sincerely,
/
G. K. Hovey Vice President and Director, IKI-2 I
a Dr. B. J. Snyder Program Director - TMI Program Office cc:
a
~
a -,
-wm w
.y
=
,.9 J
,I a
22at 22:21 at:
~~21 12:12 ;* -
h.s lag a
m ic u m 1
,11:....=
11.
see... e. a.
1., x, p,*ff$ =*
'I 7
3 j
,1 sr f *77 3 )*
- J QS ec888
'80 C'0 3 t.) W A
r NYs
- ' D.'s )
'v *-
J
?.3
/'
w, 2
J s
ese U
O :
g*
tI o w :U 't ud
$ w#,. 3 4 d V
x.8
d
'*2 3. h 3 3'
}3 r
a C0
~. %j $
b o
h*;
4.7 9
AO nn ese nn mn d,6 4 y
7 7.
44 44441 44 4
Itti 32 e
4 ce v
v
,e y,
.2 ww
< c$ { j a
JJ
(
2 e
2 oc
/
.,t w
., 73::"o7.% 2rg P Aj j g ws*WW a
e" e-w w w 4%e ww:
44
'l.W gg-1
__ e7,h arre:
F neh; 3',,.esa i*f**.
f,l E E 55
- een:
- - a '$ :a OO 31 5 "$55 0$,,,28 I
- e,
,,. g p e d
. fgg 2rzzz a m>**
8g : %c22: ::55*: :cygs 97" O
eed a
g
- c- -
0E i
5,8888:
n' o
's--s 2
e,
- 1 e f g d
~
$ 2
,,..d*l ry 1 ~~
f ax
" 2. d d:1 141
~~
g.:112 j:..,..i.. on222 3.-
- r 2 e e c e.o *e o.
2 2. :. e
- g oeoe oeo o.
o o
om
,c,, n ra na
..n-
- Y fQ 2
Q 23 d
u wy 4
'T - r*!$g h,,
i~I v
E I
a o
=
o=
s a
1 z2 5
2 T,
3*f 2 p g$
=
x a
Ot 5
2,-
Ed 1:1 I
~
o TI h*
== 1 v$#'0:
3==
33
- dg 8
82 8
1:
e
- 2 5%:
8~9 4
owB
]* 5 'TY*v ?**2tw&.y&.. 33 3s a ust::
1 g2 3o g
",0 s st lT
- re.
2,
.:.s ::::=
ss :
Ex-h: g:= :
gg
.8 z
- =. nurir ras.a ;; m((s:.3 %
~1e.,:=s n
s_a eg vvvy vvv1-a 5 ve-
- sv
.I 22" f2 5
!::--2 TTTTT TTOTT TTT
- v..e s. :. s. ::...j :..:. = w :
TT-I5H.5::a s.s.:.] a.a:vr.:..:aa*. svan.s e
a d
.2222 ~22%:,.,,-. ::= :f 20
- $E ::::
5:2:02 2 2*
e.e a e w" 'z
'l (,L
- 3 eto P,
Aj'? ad Y
_ {a g
y t-0*
- 3 C
6 i
!. 1 a 2.
w e
u 1 CB C
j Q
y 7
a g2
[ gr e e
ze n
- n l
E,
.,,,Qi[t/i ~,
c (M
3 do 82
$9 2
t e
.g
'o
) -y
-5 g
tc B
/
3 k,_o u
7.
7 og I
e b
i l~-
> $ Ns i
Qc
,k_
G c
g., g Q
3; % -@,
m
,8
.i N
J r
Qi s\\
~
a C#i '
\\ !!
Ow d
g s
s u 92 b
q
(}
4 ),+
r v e
s lal h
E 5 % *g l J J 68 2
8 s!!g 21 18
_J 4
m-e-
a nm t
x 8'a 5 g3
~g*9 2
d2 4' 4 v Q. "t SE i
d 3
h # a'** e x s a ;;1 ;.t. '
56 19i -ba '
9
.6 d
c 2,,
>>,au zo m
=
63 o.so.oi'a a2 4
2 U
o)J tJ
"'T S-8 d
W W Q tk k 'l
.=v-m
Arrrh m r 2 Int:r-Offica M morcndum Cate Septe=ber 14, 1979 Ludservice Sucect TMI #2 Waste Gas System Compressor To D. B. JENKINS Loat:en Reading On August 24, 1979. three failed parts frem a waste gas system at TMI #2 were received for analysis and assigned lab nu=ber 56735. The parts consisted of the lobe (head), rotor (impeller), and cone (Figures 1, 2 and 3).
Two possible reasons for their failure which was advanced at that time were (1) the existence of possible casting defects in the lobe which acted as sites fo erosion and created unbalanced hydraulic forces, and (2) pressure perturbations that may have resulted when the inlet pressure control valve was closed to ensure a vacuum en the vaste gas header. Material specifications were Mil Spec B16576 for the lobe and rotor and Mil Spec B16444A for the cone. The first is an 88Cu-8Sn-4Zn casting alloy while the second is an 85Cu-5Sa-5Pb-5Zn casting alloy.
Visual Examination The lobe had metal loss in areas on its inside surface near the unloader slots. This metal loss appeared to be the result of erosion and had thinned the lobe wall to a point where continued erosion a ' fluid flow had deformed two areas outward (Figure 4). No evidence of eavitation existed on the lobe.
The rotor had suffered heavy mechanical da= age as well as cavitation.
Several blades were completely missing (Figure 5) while almost all blades were cracked at their juncture to the upper and lower plates, with the cracks emanating from the inner radius (Figure 16). Heavy =echanical defe.=ation was present on the leading edges of all blades and also on a few trailing edges. Cavitation damage was present along the leading edge.
The cone had what appeared to be cavitation d2 mage at the two outlet ports near the apex of the cone (Figure 7). No damage was observed on the larger inlet ports.
Lab Investigation The rotor was examined first. A wedge shaped section was cut from the rotor and from this section, a blade that was cracked for three-quarters of its length was cut so that the crack surface was available for scanning l
electron microscopy. The fracture surface was found to be rubbed lightly l
but what appeared to be fatigue striations were observed on the fracture surface (Figure 8).
No evidence of intergranular cracking was found. The cavitation region at the inner radius showed the porous topography typical of cavitation (Figure 9).
Energy dispersive X-ray analysis of the fracture surface revealed no foreign contaminants but it should be noted that all l
parts had been cleaned during decontamination. A metallographic section through one blade was then prepared for optical microscopy. Heavy mechanical da= age was present on the leading edge as evidenced by defor=ation of grains and slip lines observed (Figure 10).
5 GPU SM Corporanon is a subscary of General Pubhc Utst:es Corporat:en l
D. B. JENKINS S:ptcmb:r 14, 1979 Lab Investigation (continued)
The cone was examined next. A metallographic section containing a cavitated region showed a surface that was typical of that resulting from cavitation.
A cross section of the lobe through the small perforation and an adjacent eroded area was prepared for optical microscopy (Figure 11). Heavy deformation typical of solid particle erosion was present on the I.D. surfaces of the regions examined (Figure 12). No evidence of pre-existing casting defects were found.
Energy dispersive x-ray analysis of the bulk metal of the lobe, rotor, and cone revealed energy spectrums consistent with the specifications. The lobe and rotcr were found to c'ensist of copper, ein and zinc. The cone consisted of copper, tin, lead and zinc.
Microbardness readings were found to be inaccurate on these samples; therefore, Brinell hardness values using a 500 Kg load were used. Tne average hardness for the cone, lobe, and rotor were 57.0, 59.5 and 64.0 respeer.svely. Typical hardness values for 85Cu-SSn-SPb-5Zn castings are 54-67 while those for 88Cu-8Sn-4Zn are 68-72.
The hardness values recorded for the rotor and lobe are below typical but it should be noted that hardness measurements are not required in the specifications.
Discussion Failure of the compressor resulted from cavitation, most likely occurring from pressure perturbations due to the inlet pressure contro'1 valve being
-._.c ose. The rcsulting pressure variations created regions of low pressure in l
d
$hich gas bubbles formed and regions of high pressure in which they collapsed, casuing cavitation.
It was not possible to determ., whether cavitation occurred first in the rotor or cone. The damage to the lobe resulted from erosion by cavitated particles and fragments of the rotor blades. The fatigue cracking of the blades was probably caused by impingement of cavitated particles and flow imbalances.
Conclusions 1.
Failure of the compressor tesulted from cavitation. The cavitation occurred because pressure perturbations resulted when the inlet pressure control valve was closed to maintain a vacuum on the waste gas header.
2.
Impingement of cavitated particles and flow imbalances produced I
fatigue cracking of the rotor blades.
3.
Erosion of the lobe occurred by cavitated particles entrained in the, fluid flow.
.@.G M J. W. WOOD, JR.
JWW:bk Atchs.
cc:
J. L. C. Bachofer, Jr.
J. B. Logan t
l R. D. Hopkins/F. S. Giacobbe (Scodi)
R. J. McGoey 6
N. C. Kazanas B. C. Rusche 6
G. A. Kunder R. L. Wayne R. L. Long
- rw u s = =~,*+ g syy_K.~ ~ ~
~w v
--w e r*
~. -
-~~.3<
- o. yy e. w -
, my
-- m
_.*smy r
.L
\\
i l
l
\\
J
.- :u I
j - 1 I
J.
_.y
'.9 l
~-
. s., '
.v j
~~ N
- an y
s
%. s,.
yx s
.c..
w c
g
[,.~,,-
p :
- '. M *.
.... : e -
Qq o.
i-i
_%~..
~
^
r.
%. L "
J'*.
p
.e
..;.w., %, x. h:h'N
{,
p i
- $*T$k,.'.-f
" r;' ^[1_,~ _,. ',
~ %.-
_ W. ' 4 ' '. ',
.? War e w ~
)h
' !h ly'?
6.,. s. g g y
l
. 1.f.. A '
- iG'
?
~
~
-\\g u p e p.. e...
~p
.m~.,
,n n
4^
- p' % + f?*,wt Q& .
r
~
w_---
.z-s.
'l
,, i y ~,
~,, '
.,ar.W -
i..
'T-
%p I
~.
/
JLQQ.-;-7;;;q g
Figure 1.
The lobe as received.
~
- - n,'
l
(
1 t
l
~.
~
AN
.N i
.s h,-[ _
\\
/
. *,.h,.
~.
v.-
t y
f i
i:
i
~,
~
/
...J l.,
M. n/
1-
=
.s s,
j.
/-
8 l
~
./
.%_ p...,--
.x s
l
,, a,.. _
g k
s
. = - - :y a
' N ~. o ;,-..
w~,
5
~
g Figure 2.
The rotor as received.
a P00R ORIGINAL
-- ~
-~~~-,r---
i a,.
s e
)
I
.s
\\
l
- .w.a.:
+
.. c.
(
..s a
3 3..
.. =.
{
..-.y.-.
l
.h ' ~,' s.
.. f.:,
f'-
's.
.. ~.
~
.... :.. ;~
--s
- a.r
, r 1
j 3.y _ :.
' ' g,_ f... %.:Q...l:
-L _ Q
}.
j 4,
4.
u, j.,
.e..'.'.,'...
c-
- er
,g
)
f
- / y m,
.y
..w,x m.),;.
7 v e.
~~ ^
.g -
y,
' '- l'}
-L.
~
.s
- 4 % gf m,..
..., M a; '.,1,a.bn....
N.
W'
, "'.. g& m._.,, i -%%
. -+
.~
... a..-.+d. g3 s,a.w J :.%
.m.
l
~
4^
5 $Ehh
. ?. - [-
h u.
.s~
.~"-:
W,s.q a y %.
g#,/.ve%. r..
a-
.a 3
v.
~yn%.,;t ;. C,..
- <,.m... w% t r
t 4 g :..
..w w6+mc.:t a
=r s.* m --.s!,-w ;..y yp:.y.ggy> k-1.aiitm.$l 2*8 & '
a.
+
t m.m g.
S$$ym,p$bND.,2 :
- 4 w a
- &b.-b,w-mme.u MM R G Y,u.~;w egyg. '._"I M w %. d*VE~~ M :F',W A F R *d_-@~7a,g h n fQ pp 2
w 3
i m e.
.'..?,,]
i Figure 3.
The cone as received.
.... :....? - ;=.O.i
~_...: :%...,,,, w
.... cf.,.-....j -,
.. g,..
t.
.. '..'2. ~a.:. :. ' %.
s.
_ i,;,.. p... -...
n..
s..
a.:,s
....~:
n.-
m
....- ~.
y
- 2. i 4
l l
e-g..%
f j
g.
~
.# s F
h 9.
y a 2,=;,
~~
Figure 4.
Large hole and small perforation existing on the I.D. of the lobe near one unloader slot.
P00R BRIGINM
~
i I
i f ' * "___ -
1 p
I.
n' I
, 1, i
4 e
1 i
k.
I I
Figure 5.
' Side view of the rotor showing a missing blade.
i
~
P90R ORIGINAL
- g., w - 9
_.f -
, g,
-s
,g,,
,.l h
,1 7
. f. s L~
y, y
l (4 -
s.
i< ~
9 l.
- .:4 t-T
, 'f.:
e
'e
?
....)l
,S
,7
,y.
4_
,- c.-
.A 99
- I.
s l
w.- ~ ~
1
_ -min. --,-.
_ s;7
.n
~
~.
y =-.
p.wmmyc A 1 Figure 6.
Cavitation at the inner radius and cracks
~
at the juncture between the blades and the endplates.
1
I
~
t.
(
^
l Figure 7.
Cavitation damage on the cone.
?
- l
- I.L.
f k
h ti,:
'. sy
?
~
Figure 8.
Fracture surface of a rotor blade showing possible fatigue striations (1200X).
6 i
P00R ORIGlNAl.
_ 5;
. =.. _
t l
l l
l i
l
.4-y Figure 9.
Cavitated surface (200X).
l l
y ' *:,..
b~.,,
E
'?' ', g.3 I+V-k 1-
.y a :.xs.
l
, ' ci f.t.
0 2 '"
- c. * *.
' ~. 'j, A y ?... h:,+, ;'~p. ;
f' a
rh Y
?:. ; c.
0 :., L.,y4.,
.q
'uI,*.,',!v1. 5 1 ?. d i !;Ms[N!?[ik N3'
' $O A. h. g *'-
\\
.l
'M^.*
l 4, J? h; k. ;
-..t' i
. ; '(
Q >d'(,'h
- - 7. i
}.,y *.<' _l.g.
s '
4_
r.-
\\
'w i
~,.
7 s
.s_ s.....
.N m..
i l
Figure 10.
Deformation and slip lines existing on the leading edgeof a rotor blade (50X).
s i.
?-
r
...m.
j 4
- g g
-f-l 3 u t,, -
. w.
t, m. i?.'
g; '
3.,.<.i
=w
. -r p.M
- .cf
..n>
y 5.<*,~.l@.
M, a, : X
(&.
_=g m
- m "
~
1 w,.
m
.. 5 +.-%,,,( ?',y.g.,.<,- m.s 5i.D % $d %.h ~ n < -
f,. '-Q. :. 4.
.x -r
- ,w c = c.'...? -
n
'.', :. & *~- ',,;_',-
. 5.m. :.-
Figure 11.
Area through which lobe metallographic section was made is indicated by dashed line.
t i
[EJ~
.h@^..'
]i
~
~
., 4.9 Wf e -9 4' g
.4 -, ' l<.,.w..
a:
~
=
-.f=.t:};pj,,'S. *
., -'4ff r'
6-
^
- ?
./s$
s Ar...
.f...
- p..,
1
~
r
!* Y
'?
E~
.=
k
. i,' k
. h.. '
y.h>...... f_;.-
j s %<
s-
- l' W 5%^.o W. '.' %
$5.i f f.Y 'r-t
.W
',f ',,
- y ' *
{,{
/. P.
<* "- ]
f,y *_'.,s' n
. c'. A
.s i
p.
"." j
/Yw 6
Figure 12.
Deformation at the I.D. surface of the eroded areas on the lobe (400X).
a P00R ORIGINAL
~
.