ML20085H480
| ML20085H480 | |
| Person / Time | |
|---|---|
| Site: | Saxton File:GPU Nuclear icon.png |
| Issue date: | 07/14/1969 |
| From: | Cindy Montgomery SAXTON NUCLEAR EXPERIMENTAL CORP. |
| To: | Kirkman R US ATOMIC ENERGY COMMISSION (AEC) |
| Shared Package | |
| ML20083L048 | List:
|
| References | |
| FOIA-91-17 NUDOCS 9110280287 | |
| Download: ML20085H480 (20) | |
Text
'
SAnTON NUCLEAR EXPERIMuNTAL CORPORATION H) m GENERAL P U B LIC UTILITIES SYSTEM i $NfC Janstv Ca= tent Poesa a Lio ev Comeae.i Ns* Jsents Poesa a Liowt Com*4=v ee==,vtv4
.a steeva.c C o-.a=v Id t T 9eDob tT a = (D+ 60 Com p4 = 9 P. O. Box 542, Reading, Pc. 19603 Telephone Area 215 929 3601 P. O. Box 99, $axton, Po.
16678 Aroo 814 635 2937 Saxton, Pa.
July 14, 1969 Mr. Robert W. Kirkman, Director Region 1 Division of Compliance U.S. Atomic Energy Commission 970 Broad Street Newark, New Jersey 07102
Dear Mr. Kirkman:
Please refer to your letter of June 24, 1969, with attached Form 592, relating to the observed surf ace discoloration on the flange support of the lower cc e support barrel.
We have completed our engineering investigation and have concluded that the observed condition occurred during manufacture and therefore does not constitute an unsafe condition. A report of this investigation it. enclosed.
The enclosed report will be included in tb primary coolant system inspection report as Appendix D.
Very truly yours,
.7-7 h
/flhh pn C. R. Montgomery General Manager cc: w/ene.
G. F. Trowb ridge, Esquire D. J. Skovholt R. F. Swift 1
I 7110:eo s7 910^24 PDR FOIA pg~ a DEKOO 1-1, 1
APPENDIX D
LOkTR CORE SUPPORT BARREL INSPECTION REPORT ON DISCOLORATION AND DISCONTINUI17 OBSERVED ON Tile OUTLET N0ZZLE 1.
OESERVATION (Refar to Figure 1)
During visual inspection of the lower core support barrel some surface discoloration was observed on the flange support (Item 7) of the lower core support barrel in the downstream direction from the veld that joins the flange support (Item 7) to the lower barrel (Item 4).
The discoloration was parallel to the water flow tapering off at the downstream end.
It initiated at the veld which joins the flange support to the lower barrel.
Subsequent inspection by T.V. camera and binoculars confirmed there was a surface discontinuity at the origin of the discoloration.
The area was T.V. tape recorde.1 (Figure 4).
The size of the discontinuity was estimated to be 1/4 irch wide, 4-1/4 inches long, and on the order of 1/16 inch deep.
It was oriented in the direction of the weld, i.e.,
perpendicular to the direction of flow.
The surface within the discontinuity was not abrupt or indicative of any stress-related struc-tural defect.
2.
OBJECTIVES OF EVALUATION Since there was no immediate explanation of these observations, an investigation was carried out with the following objectives:
1.
An engineering evaluation td analysis of the condition as to its nature and'possible"cau.cs.
2 '. An evaluation'of'the possibility of further degradation.
3.
Assui.ng '
nozzle, wall could,become, penetrated, would the a ddi,t.o*
by7 pass,, flow create a, potential, hazard?
.......o I / I, b..
sl.
s..
s.
..s i....,p
.6.
.i l
. w i
w r.
- ~ +,
.g.
APPENDIX D
3.
SCOPE OF EVALUATION It is clear that the observed discontinuity was either present when the reactor was first commissioned, or that some removal of metal occurred subsequently.
The investigation was therefore directed at the following areas.
1.
A check on original records to determine if there was evidence of the existet.cc of the discontinuity prior to original commissioning.
2.
of possible mechanisms which could account for the appear-i,
ance of the discontinusty during service.
3.
An evaluation of each mechanism to see if it would account for the observations, and if so, what prediction could be made as to the future growth of the discontinuity.
4 An evaluation of the stresses existing in the vicinity of the dis-continuity 5.
An evaluation of the additional by-pass flow resulting from a possible penetration of the nozzle wall at the discontinuity.
4 CHECK OF ORIGINAL RECORDS This check disclosed the existence of a photograph (Figure 2) showing the Lower Core Support Barrel during fit up of the internals in the manufacturers plant. An enlargement of the nozzle area is shown in Figure 3 and may be compared with Figure 4.
The comparison shows that i
both the discontinuity and some features of the discolored area (lines parallel to flow) can be observed in both' pictures.
The discontinuity extends round the circumference of-the flange support (Item 7), to the interpenetration with the conical section of the barrel (Item 3), at which l
em
,.a
.-m,sm..u.e
--__=
w*e.4
- mat.me==
- em w same e=e=w...
e r.
=ew--
3 APPENDIX D
point a misalignment can be observed in both Figure 3 and Figure 4 The discontinuity appears to have been caused by a misalignment of the parts (Items 3, 4 and 7) during manufacture.
It appears that a fillet veld was deposited in this area to smooth out the profile.
No immediate explanation can be given of the lines parallel to flow in the discolored area, but they may be grinding marks which did not clean up during machining of the nozzle.
5.
EVALUATION OF MECHANISMS WHICH COULD ACCOUNT FOR THE DISCONTINUITY DURING SERVICE The mechanisms considered weret a.
Mechanical interaction.
b.
Corrosion or erosion.
Mechanical interaction was ruled out because of the absence of any ccaceivable mechanism whereby any other part could have come in evn-tact with the discontinuity during service.
The possible corrosion-erosion mechanisms are reviewed below:
l 5.1 Uniform Corrosion (Assuming that the aaterials complied with the assembly drawing, i.e., ASTM A240 Type 304 stainless steel was welded with Type 308 stainless steel weld metal). For Type 304 stainless steel in reactor purity water, Wanklyn and Jones report an average corrosion penetration of 0.06 mils per year and a maximum of 0.12 mils per year. Variations of composition among the austenitic l
grades of stainless steels had little effect on the penetration l
l depth, therefore the same value is applicable to the Type 308 weld, under the assumption of uniform corrosion (galvanic effects are considered below). Variation in water flow rate from nearly l
e.
a
,n-
-4 APPENLIX D
stagnant. to 30 feet per second (fps), variations in surf ace co'.dition, variations of the water pH between 7 and 11, ad-ditions of dissolved hydrogen or oxygen, water temperatures of up to 572*F, and additions of up to 1500 ppm of boric acid did not significantly modify the corrosion rate. For compari-son, the Saxton conditions at the area of interest are:
tempera-ture is about 510*F, flow rate is estimated as 20 fps, pH (cold) is 6.0, boron is 0-1600 ppm -- boric acid equivalent is 0-9120 ppm.[2,3) On the basis of the experimen'.a1 results for 1500 ppm of boric acidI 3, it is estimated on a conservative basis that the actual boric acid concentration in the Saxton cooland would result in a rate of corrosion not exceeding 0.24 mils per year.
Neutron irradiation (in-pile) has been found to enhance the cor-rosion of Type 304 stainless steel clad by a factor of about two, in experiments using nitric acid as a corroding medium. '
Since the outlet nozzle is not subjected to the intense irradiation as experienced by fuel rods, the effect of neutron irradiation is negligible in this case.
In summary, the expected normal uniform corrosion of the speci-fled materials is conservatively estimated as 0.24 mils per year under Saxton conditions and is negligible for time at temperature experienced thus far by the Saxton Reactor.
A time of 18,000 EFP*
hours plus a very conservative estimate of at least an equivalent time period at temperature for training and physics purposes would yield about 1.1 nils corrosion. This mechanism does not therefore account for the observed discontinuity.
- Effective Full Power
~
. -. -~. - ~ _. _ ~ ~._-
1 APPENDlX D
5.2 Veld Metal Contaminants The only veld metal contaminants that could possibly occur would be slag formed during the welding operation and, less likely, sulfur in the electrode forming metal sulfides.[5]
I3 In the case of slag, experience has shown that for a situation where sufficient water flow exists (as in Saxton), corrosion occurs in the area of original' slag deposition and the corroded material is eroded away.
Corrosion continues until the slag is removed.
Such a situation could occur as the result of poor welding practice, but the corrosion rate would be expected to remain relatively con-stant. Any future corrosion penetration, assuming slag contamina-tion throughout the weld, can be calculated from the present rate of assumed corrosion, as was done in Section 5.1 above.
In the case of sulfides occurring as a result of sulfur impuri-ties in the weld material, the expected course of corrosion is simi-lar to that described f or slag, since the sulfide particles would cause preferential corrosion ultimately-leading to removal of the impurity.
In either case, it is normally expected-that such contaminants would occur over the entire veld length and corrosion would not be localized as shown in the photograph.
(Figure 4)
In summary, weld contaminants could cause the supposed corrosion if it is very conservetively assumed that the corrosion conditions, particularly water flow rates, are greater in the area shown, resulting in preferential removal of corrosion products and local-ized,Lfaster corrosion rates than in the rest of the weld.
In this case, it is expected that the rate of c ro rosion would be con-stant.
l
-~ ~~~~ ---- -. - ~ ~ - -~
4
-p-m
- - ';=*--
v M
---g
,ge yge-wg-y..,
e-ig4
,.m-w yr-g
-my e eye + q
-y y my s--O g pvtW try -M-&gy 1r'-w p'F APPENDIX D
5.3 Galvanic Corrosion Two aspects of galvanic corrosion are possible in this case.
in the first, it is assumcd that the specified weld metal (Type 308) is different in composition from the joined material (Type 304) and therefore is theoretically subject to galvanic corrosion due to a difference in eletrochemical potential. However, as noted above, this actual case has been studied and the galvanic cor-rosion effectr. are negligible upon an already negligible overall corrosion rate.
It is expected that the same results would be obtained for any other grade of stainless steel which might have been mistakenly used fer the weld metal.
The second aspect considers weld metal other than stainless steel.
The worst possible case, is considered to be that where a carbon steel weld metal was mistakenly used.
An examination of the gal-l vanic series of metals and alloys shows that carbon steel is decidedly anodic and hence corrodable when coupled to stainless steel. Therefore, one can theoretically expect preferential cor-rosion of the caroon steel, especially since the area of stain-less steel is large by comparison. __In practice, however, Wanklyn53 reports that galvanic couples of mild steel to stainless steel have little effect en accelerating corrosion of the mild steel.
Assuming, first of all, that the above experimental finding is
-correct, one can therefore consider the hypothesized mild steel as corroding-alone. According to-Vanklyn
, corrosion rates for mild steel are roughly ten times greater than for stainless steels under the conditions described in Section 5.1, Uniform Corrosion.
This amounts to about 2.4 mils corrosion per year.
Conservatively assuming the_ time at operating temperatures as =
twice the EFP hours, the expected corrosion is about 11 mils com-pared with the stated observation of 1/16" or 62 mils. No infor-mation is available to justify a future corrosion rate in excess
7_
APPENDIX D
of that aAready hypothetically established.
Secondly, if the experimental report of galvanic effects of stainless to mild steel couples is assumed to be incorrect, a corrosion depth greater than that for uniform corroaion of mild steel can be assumed.
There are no data to expect a rate greater than is already assumed'from the depth of the discon-tinuity. For this situation, it should be kept in mind that two
" ifs" are necessary:
"if" a carbon steel electrode was mis-takenly used and "1f" Wanklyn's reported experimental findings are not correct.
In summary, galvanic corrosion is a possible cause of the observed discontinuity only if one assumes two separate errors, and the expectation of continued corrosion in excess of the established rate is not justifiable from the literature.
In addition, it I
would be expected that this type of corrosion would occur over I
l the entire length of weld exposed to the water, rather than in the localized manner shown in the photographs.
This is contrary I
to the observed condition.
5.4 Stress Corrosion Stress corrosion must be considered in any evaluation of corrosion of stainless steels.
In reac. tor waters contaminated with chlorides, oxygen must be present for stress-corrosion cracking to occur.I 3 3
A check of the Gaxton water chemistry shows that the hot water l
chloride and oxygen contents are less than 0.005 ppm, which re-presents the limits of detectability for each species.
Conservative-
,1y assuming that at least 0.005 ppm of each species actually exists in the Saxton coolant water, and comparing with a curve of reac-tor water oxygen and chloride content as related to stress cor-rosion cracking under intermittent wetting 3 (a condition whereby
,s 4
~_
APPEND 1X D
chlorides can be concentrated), it is seen that the chloride content is one to two orders of magnitude and the oxygen con-tent is at least two orders of magnitude too low for chloride stress corrosion cracking to occur.
Another possible source of intergranular stress corrosion is from caustic solution such as the LiOH used for pH control in Berry'7] reviews work which shows that several thousands Saxton.
of ppm of caustic: solution are required for corrosion of_ stain-less steel. Pement found a rhreshold of 0.1 M lithium hydroxide or less at 450-615'F for failures of stainless steel. This amounts to cbout 1470 ppm of LiOH. The Li content in'Saxton coolant is 0.1-0.2 ppmI3l, and therefore LiOH is highly unlikely-to be a source of corrosion.
In summary, stress corzobion is highly unlikely on the basis of the excellent Saxton water chemistry. Recourse to the assumed existing corrosion rate is necessary if one speculates that the chemistry or literature data are incorrect, but in this case there is no reason to believe that the rate of corrosion would increase with time.
5.5 Corrosion - Summary No corrosion-erosion mechanism examined will account for the ob-served discontinuity unless some other unfavorable assumption is made. Ecwever, if it is assumed that the discontinuity was in-fact caused by some corrosion-erosion mechanism, there is no reason to suppose that the rate will increase over that implied by_the known size of the discontinuity.- 11e thickness of the nozzle is 1/2 inch.
If the assumption is made that corrosion could occur from both sides, then a further 3/16" of corrosion (measured from one side) would have to occur before the nozzle l
wall is penetrated. At the assumed rate of corrosion, this would-l l-f i
o, l,.
_,.., -.., ~.. _.
APPEND 1X D
not occur in the expected lifetime of the plant.
6.
STRESS ANALYSIS The structure has been treated as a continuous, constant thickness shell.
The shell is a cylinder with a conic reduction to a smaller cylin-der and a reinforcing ring at each end of the shell.
The computer program " Seal Shell 2" has been used to determine the basic shell stresses.
In the region of penetration the basic stress is essentially a uniaxial hoop stress of 300 psi. A resolution of all the component stresses on a Von Mises criteria or on an octahedral stress basis gives a resolved stress of 350 psi.
Because of the complex nature of the nozzle construction, no accurate stress concentration factor is available. A factor of 5 has, therefore, been used.
The basic stress is of the order of 2% of the material yield stress and is compressive. Allowing a stress concentration of 5, the stress in the nozzle area is still less than 10% of the material yield strength at temperature.
On the basis of such low steady state stresses, it has been concluded that there is adequate margin for safe operation of the reactor.
6.1 Input for Seal Shell 2 Calculation Loading Conditions Temperature 510*F l
Vessel Differential Pressure 11.3 psi
=
Core Differential Pressure 4.1 psi
=
~.
r r
1 m
i.
APPEND 1X D
Self Weight _s_
(
Lower Core Barrel 2750 Baffle Assembly u
1230 Core Support Assembly 1750 1
=
21 Fuel Assembly 2310
=
21 Dummy Assembly 1210
=
TOTAL 9,250 lbs.
l Pressure Drop over Core s.1 pai
=
41" 1/D core area 1320 in2
=
Lift force 5,420 lbs.
=
Load on Core Barrel 3,830 lbs. axiai e
Materials Barrel ASTM 240 T 304 Grade 5 Electrode ASTM 298 E 308 Specified minimum yield 30,000 psi Specified ultimate strength 75,000 psi Yield strength at temperature 18,200 psi 6.2 Basis for Estimation of Stress Concentration Factor From " Seal Shall 2" Printout by inspection, the stress field across the penetration at mid-plane is effectively uniaxial and constant at 350 psi.
Points of interest are at 90' and 135* as noted (discontinuity lies in this region) 1.e., a nozzle type reinforcement in uniaxial stress.
<7 r
3[-
^
% *gj d/D
.259 t
.67
=
=
l T
I
.045 T=
.017 t =
h// / /
/ //
E O
D Reinforcement Factor
.203
=
D Penetration Angle = 42.4* Lateral e-
1 i
i APPENDIX D
I17l 1.
From.the reference by R. T. Rose the stress concen-tration factor for a completely unreinforced nozzle (t = T )
d D
for T =.017 and d/D =.259 is given as 5.0.
D The nozzle actually has 20% reinforcement therefore the factor 5.0 is high, 2.
From the ASME Code 68 the recommended factor is 2.6.
For the lateral connection of a cylinder at an angle the stress concentration index = K {1 + (tan $)4/3)whereK=
inside stress index.
In this case K = 1.0 tan $%1.0andthereforestressindex=
2.0.
The stress concentration factor is d,1fficult to estimate for this case since there are no directly applicable references.
A minimum value of 2.5 and a maximum value of 5.0 for stress concentration appears to be a valid range.
This range is based on judgement from the list of references given.
The resultant maximum stresses in this region of the nozzle are'in the r'ange'825 psi to 1750 ps'i.'
7.
EVA1.11ATION OF BY-Pass 'FiOW The only poten,tially serious consequence of a penetration of the nozzle wall is the increase in flow of primary coolant bypassing the core.
Calculations have,been made which s,how that a bypass hole in the,pozzle would leak flow at the rate of 1% of the total e a
. to se le s
- i h*
se e e.
y y
= >,
i
. APPENDIX D
system flow per square inch of leakage area.
The Saxton Core III design values of 7880 GPM total system flow and 15% core bypaJs flow contain at least 5% conservatism, i.e., the design value of heat trans-fer flow in the core at.85 x 7880 GPM has at least 5% design margin.
Thcrefore, the core design is not jeopardized by a 5 in2 hole in the outlet nozzle.
8.
CONCLUSIONS This evaluation indicates that the discontinuity occurred in manu-facture. No mechanism has been identified to account for the appearance of the discontinuity in service.
However, even if such a mechanism did exist, the evaluation demonstrates that no hazardous condition could develop during the plant lifetime.
I t
l
.,.--,,_._.,-m,
,,_-c
. APPENDIX D
9.
REFERENCES 1
Wanklyn, J.
N.,
and Jones, P.
J., "The Aqueous Corrosion of Reactor Metals", J. of Nucl. Materials 6, No. 3 (1962) 291-329, 2.
R. Stansfield, personal communication, 6/10/69.
3.
R. Swift, SNEC, personal communicetion, 6/16/69.
4.
Duncan, R.
N., " Stainless Steel Failure Investigation Program",
Final Summary Report, GEAP-5530, February 1968.
5.
Leo Marti-Balaguer, personal communications, 6/11/69 and 6/16/69.
6.
Metala Handbook, 1948 edition, p. 559, 7.
Berry, W. E. "Some Facts About Stress Corrosion of Austenitic Stainless Steels in Reactor Systems", Reactor Materials 7, (1964-1965), p. 4.
8 Pement, D. C., Reactor Chemistry and Plant Materials, WAPD-BT-16, December 1959, 1
9.
Taylor and Lind, 'IAM 270, University Illinois'.
10.
Hardenbergh and Zamrik, " Experimental Investigation of Stresses in Nozzles in Cylindrical Pressure Vessels".
l 11.
Stepanek, " Stress Concentrations in Nozzle Ring of a Pressure Vessel", Nuclear Structural Engineer 2 (1965).
12.
Rose, " Stress Analysis on Nozzles in Thin Walled Cylindrical Pressure j
Vessels".
l
/\\
4
_ _..~.
n.--
_-e-~.*
- ~ = ~
v t*"'
+
r' N
- tve w
i.
APPEND 1X D 13.
Waters, " Stress Near a cylindrical Outlet in a Spherical Vessel".
14 Lind, " Estimation of Elastic Stress Concentration of a Nozzle in Spherical Pressure Vessel".
15.
Clare and Gill, "Effect of Diameter / Thickness Ratio of Flush Nozzles in Cylindrical Pressure Vessels".
16.
Leckie and Penny, " Stress Concentration Factors for Stresses at Nozzle Intersections in Pressure Vessel".
17.
Rose, Design Method for Pressure Vessel Nozzles, Eng., June 20, 1962.
18.
Kitching and Peckins, " Stress Analysis of Rim Reinforced Openings in Pressure Vessels".
~
~ ~
19.
Friedrich, C. M. " Seal Shell 2, A Computer Program for the Stress Analysis of a Thick Shell of Revolution with Axiaymmetric Pressures, Temperatures and Distributed Loads", WAPD-TM-398, Westinghouse Bettis Atomic Power Laboratory, Pittsburgh, Penna. (August 1963).
20.
- Kraus, H., "A Review and Evaluation of Computer Programs for the Analysis of Stresses in Pretsure' Vessels", Welding Research Council Bulletin No. 108.
l'
\\
1 1
1
. I APPENDIX D
10.
FIGURES 1.
Drawing 646J809 2
Photograph 458 3.
Enlargement 4.
TV Picture 5.
Stress Distribution 6
Penetration Details O
e a
e t
fG
i i
l l
g,M v t, i
l 4
w i
l e'
l g
j t,,
'J -
CulDE Pih h.
. fr y
f' t
J
, 4.i
' ~~ $$lbhk
?
%.n p3:*cm.,,
kA y'
AREA th
' _ QUESTION 3 e'M !
+
L'I Fill:0 LUG l
f p
g. fc.3p
.... =
p[
s$y +
p-c g
f, LOWER CORE PLATE A
._6, f &,M._,f 5
a a
t...
r- --
'i.. e -' e._., 3 t
3 s
i e,
60,
\\
,i --
9
_9
- .~ e.
LOWER CORE SUPPORT BARREL I
w r
s.
v
., r, SN "58
. -.c -
FIGURE II.
SAXTON INTERNALS ASSEMBLY I
l i
e i
- i, - _
e 4
- j
.,,k'th'
/.
r
- .s
~ * '
s et li 4',[
.*Q
.w =
a.
G;.
n,
- S G
P O
- z. x *:1,.
o wic-6 Pee,.- -
.I. ff%. I '
0 m
Li!l%.
J hr-Z
~*
w
_.,.. 7
- l;4$Yla
- f., Y 4f O
4.y,,*
gg e
u O
3.
- #.h
[*
"w" xt r
w 3p b
~,'
?I_
,,T$
l v
w4'i.
9 m>
~
u.
o H
4 e
Z ww!,L.,,;ji...
g
. 's.
m a
fs 9'
u 4 h e. i.
%f.;
f s~
.,N.+
~
- f;,
% cq
,4 CC 2
?e O
-~
L.
w 1
~y E
e
,y.
.A L
b l
I I
l e
ARE A IN QJESiloti 4 4 e
% y,
- i fp s. g r
q df Ni r ;
. #,[3 %,1, T~
'6 e
g5
, p k
e4 n mJ'?
..y*
W p.
s 9'
I I
'5
,r
+
N-
-iq
.s e
9
- 1
,3 :3 7, ga
+-
- w -
e i
h'
- N l1%
g; 4,
g 4
+
FIGURE IV.
PHOTOGRAPH OF AREA FROM TV TAPE s
.x
'd I
1377-1 PENE1 RATION DETAILS 0
REPLACEMEN T FACTOR
.94 REPLACEME NT FACTOR
.667 h
N 45 i
/\\
I
/
\\
/
\\
/
\\
/
REPLACEMENT
\\
/
FACTOR.203
\\
/
90 i
/'
s REPLACEME NT FACTOR.203 i
135 REPLACEMENT 180 FACTOR 1.14 s
Figure 6.
Saxton Lower Core Barrel
<\\
-