ML20055J411
| ML20055J411 | |
| Person / Time | |
|---|---|
| Site: | Yankee Rowe |
| Issue date: | 07/30/1990 |
| From: | Odette G CALIFORNIA, UNIV. OF, SANTA BARBARA, CA |
| To: | NRC |
| Shared Package | |
| ML20055J406 | List: |
| References | |
| NUDOCS 9008020219 | |
| Download: ML20055J411 (13) | |
Text
__
- I 1
I 7/30/90 LN' 1990 Shift Estimates For the Yankee Rowe Vessel i
G. R. Odette Professor of Nuclear Engineering and lingineering Materials University of California, Santa Harhara 1
1
- 1. Introductory Comments This report is in response to a request of the U.S. Nuclear Regulatory Commission (US.
NRC) [A. Taboada, N. Randall and D. Elliott)) to: 1) comment on documents pertaining to estimates of referenec temperature shifts experienced by the Yankee Rowe Pressure Vessel (YRV) as of 1990; and 2) provide my own evaluation of these shifn VUt document briefly describes the results of my evaluations. In part, due to the re7'
. thort time available, the analysis cannot be viewed as comprehensive or defin!* : Nwever, it-1 should be clearly stated at the outset that, even give more time,it would nw H possible to gencrute rigorous or reliable shift predictions based on either what we know about the YRV
{
itself or the existing methodology to account for the lower irradiation temperature (= $00 F or less). For example, there is only one valid YRV surveillance data point (Qt = 0.22 L
known with any certuinty. The other (upper p) late) data point, again uncertain fluence,is for a high flux / lead factor level which disquallfles it as being a valide su veillance data poim (1 also suspect that the actualirradiation temperature may also be
' higher in this case due to in-core gamma heating). Hence, one is forced to seek other sources ofinformation.
Recent progress in understanding embrittlemect at higher temperatures (e.g. 550 F) has
' clearly demonstrated that the phenomenon is caused by a number ofinteracting mechanisms which are mediated by the combination of a number of metallugical and environmental 1
variables Rigorous interpolation and extrapolation of the large surveillance data base will
~
ultimately require validated physical models. Simple analytical prediction procedures should be viewed as an interim engineering expedient which usefully represent data trends over limited variable ranges. For exampic, if a stradiation temperuture sensitivity of 1 F/F is aroposed,it should be understood that this is not a description of some fundamental schavior, ht <ather a representitive observation pertinent only to a well defined set of conditions.The paucity of retevant data near 500 F forces one to rely even mse on such engineering rules of thumb.
h Hence this assessment, while guided by what is currently known about embrittlement mechanisms, was based on the engineering approach described above.1 have reviewed the infonnation pertinent to the YRV provided in the report to the NRC by Yankee Atomic,
i the draft analysis prepared by Eser2 and other relevant information in the literature (primarily, data for arts Jia.bns at or near 500 F) ched below. Most of the irradiation data is from acccicrated m ;.e. '.s test reactor (MTR) experiments; hence, I have, of necessity, ignored possible. acts of neutron flux. The 500 liirradiation data, encom,ussing a modest range c's;eci (bs e and weld) compositions,is both limited and "unscrub sed".
9008020219 900730 PDR ADOCK 05000029 P
PDC e
e
t In the case of the Yankee Rowe (YR) plates, I used various approaches to try to identify "self. consistent" shift estimates, in the case of the YR welds, I recommend'use of a preliminary trend curve which may account for uncenainties both in the metallurgical composition of the stecls and due to the lower ofirradiation tem;>craturc *lhe recommeruled shifts are based on the 1990 vessel Ouence estimates of Hiser2, l
My shift estimates may be on the conservative side of nominalin the sense that one cun identify data which fall below the predictions. On the other hand it would be relatively easy to propose much higher numbers, and I emphatically believe that there are many good reasons to be prudently cautious.
- 2. Comments on Assertions in the Yankee Atomic Report on Grain Size Effects There ate limited data lo the literature (e.g. Ref 3) which support the contention that irradiation embrittlement sensitivit The Yankee Atomic Report (YAR)y of ferritic steels increases with increasing grain siz asserts that this is related to a higher concentration of irradiation induced defects due to a presumed lower sink strength in coarse grained stects, t
However, this is a very unlikely mechanism since, for among other reasons, the internal defect sink strength, due to features such as dislocations, is typically many orders of-magnitude higher than the grain boundary sink strength itself'8 Any simple interpretation of such data trends is confounded by the complex interaction of a number of underlying embrittlement processes and the fact that, m general, several microstructural parameters vary along with grain size. However, a much 3. are likely explanation of a grain size effect is related to the micromechanics of defortnation and fracture: viz. - the temperature shift due to a specific irradiation induced change in microstructure - which occurs on a very fine (nanometer) scale is larger in coarse versus fine grained steels. Indeed, the most plausible mechanism can be deduced from micromechanics models of brittle fracture: a given change in yield stress can cause larger temperature shifts in coarse versus fine grained steels 5,0. Yield stress changes are largely 1
controlled by the fine scale microstructure, composition and the environmental variables (fluence, temperature, flux and spectrum), and are not expected to be directly influenced by t
gram size.
1 The latter conclusion is directly supported by data shown in Figure 1 from a controlled, (single variable) experiment on a single heat of model A533B pressure vessel steeld. Within data scatter, the austentizing temperature variation from 900 to 1000 C, corresponding to grain sizes of about 15 and 50 pm respectively, has essentially no effect on the irradiation induced yield stress increases. Note that this and other data in reference 4 has been 4
misinterpreted in the YAR.
In particular, contrary to the assermon in the YAR, both nickel and irradiation temperature influence yield stress changes in coarse as well as fine grained steels4 Figure 2 shows the optical micrographs for two steels which both have relatively coarse prior sustenitic grain-sizes (-50pm): alloy PK contains 0.41% Cu and 0.86% Ni and was austentized at 1000 C; alley LA contains 0.40% Cu and 0.001% Ni and was austentized at 900 C Both steels-received the same semper and stress relief treatments. Figure 3a sh3ws the temperature dependence of peld stress increases for alloy LA at an imer}.t ated fluence of l1' l
1.1x1019n/cm2. lhe observed sensitivity of 2.5 MPa/C is actually sligN1y larger than nonnal, but well within typical scatter. Yield stress changes for coarse grrmed LA and PK at a fluence of 0.9x1019n/cm2 and an irradiation temperature of 305 C are plotted as a l
' function of nickel content in Figure 3b; an essentially " typical" nickel sensitivity fo thes r
e conditions of 50 MPa/Niis observed. Due to the likelihood of.n increased ratio of-l C" 3' 90 10: 19 AM 704
.. - ~,.
transition shifts to yield raress changes, the effects of nickel and temperature would be expected to increase, rather than decrease,in coarse grained steels.
- 3. Embrilliement Data for Irradiations at $00 F in order te more directly address irradiation embrittlement at 500110 F 1 carried out a brief literature survey which yicided shift data for a modest ran,;c of steels (base and weld)in
" relevant" com high copper (>. position range: (e.g. low to-intennediate nic cel (<1%) and intennediate to-15%))? 10. ne data, plor-4 in l'igure 4, a evaluate the effects of the metallurgical vru..a. 8 that are known to be scificant at higher temperatures (e.g. copper, nickel and, pot.sibly, phosphorous). However, de 500 li data can be effectively bounded by a Reg. Guide 1.99 Rev 2 fluence function (f,0 combined with a chemistry factor (CF) of 300 F. Hence, AT = 300f,t (1) i provides a seasonable peliminary 500 F trend curve. I do not believe that it is neccessarily (or should be considered) a bounding shift curve tince the small data base is '
not sufficient to establish either physically or statistically validated uneenainty limits.
s
- 4. Upper Plate Shift Estimates There may. be a valid dosimetry basis for the fluence revisions proposed in the YAR.
There is insufficient detail presented to suppen this jud;;cment. However, the suggestion that a comparison of YR and M3 shift data supports acjusting the fluences and ignoring temperature effects is not clearly demonstrated by the data. Ln.In plots of the combined.
YR/BR3 data set for various combinations of temperature and fluence corrcetion factors are shown in Figure 5. The open circles are the BR3 data corrected to common temperature of 500 F based on a 1 F/F additive ternt The closed circles are the BR3 data adjusted to 500 F using a multiplicative correction factor, Ct C =1 + 0.006(T 500)
(2) t which yields a nominal value of 1.3 for T = $50 F, The squares and triangles are the YR upper plate data assuming the nominal and twice nominal fluences respectively, if the low Quence (< 1x1019n/cm ) BR3 data are ignored, the combination of the nominal fluences 2
and'the C adjustment factor provides the best fit over the largest fluence range. 'A least t
square fit of the in in data (sohd line) yicids an upper plate shift expression AT =.183$to.315 (p)
(3)
This gives a shift estimate at $t = 2.3x1019 of 238 F versus an estimate based a In In
-interplation of the nominal YR data of 245 F. Assuming the YR fluences are twice nominal is more consistent with'the low fluence BR3 data and a 1 F/F temperature correction ternt A least square fit to this data (dashed line) for $t > 1x1019 ields y
AT = 1740to.283 (p)
(4) and a 1990 shift estimate of 220 F, C "t. 31. 9 0 30:15 AM pct
~
Recommendation -
i Considering the various uncertainties it is not possible to discriminate between the various 4
approaches to data conection (or manipulation). Hence, I recommend use of the (slightly) more conservative proecdure of directly interpolating the nominal upper plate data to the 1990 YRV Guence. As noted above, this gives a nominal shift estimate for the.
upper plate of about 245 F.
5.1,ower Plate Shift Estimates The lower plate contains slightly more copper (0.2%) and significantly more nickel (0.63%) than the upper plate (nominally 0.18%Cu and 0.18% NI). Hence, shift estimates must account for the combination oflower irradiation temperature (nominally 500 F), the higher nickel content and the (apparently) greater embrittlement sensitivity of the YR stecls.
Several approaches were considered.
Approach 1 One approach would be to oerine a composite correction factor, C nis, based on the ratin of the estimated u t
2.3x1019 /cmEper plate shift of 245 F to the 110 F shift predicted by RG 1.99 Rev 211 at for 0.18% Cu and 0.18% Ni plate; thus C nis = 2.23. Multiplying the n
t Reg.Gulde Rev.2 predicted shift of 181 F (for the lower plate parameters. 0.2%Cu and 0.63% Ni at 2.05x1019n/cm2) by C nie 2.23 yields a total shift estimate of 404 F. This 1
t considerably exceeds the estimate from Eq.1 of about 360 F, hence, seems to be overly ConserVallVe.
Approach 2 An additive correction term based on the difference between the estimated and R Rev. 2 predicted upper plate shift (245-110 = 135 F) combined with the RG 1.99 Rev. 2 -
prediction for the lower plate yields a total shift estimate of 316 F.
Approach 3 Since the primary difference between the upper and lower plate is the nickel concentration it wotild seem useful to look for other sources of information to estimate the ext l-embrittlement associated with a difference of 0.45% (0.63%-0.18%)in nickel co f
a) Reference 12 shows the yield stress increase due to irradiation varies with nickel content -
as t
Acys = A +)Ni (5)
I where the nickel coefficient B is a function of temperature, Cuence (and, possibly, co neutron flux and heat treatment). A slight extrapolation of the data given in Figure 13 of reference 12 to 500 Fyields a value of B - 150 MPa/%Ni for a fluence of about 1.5x1019 n/cm2. The mab'nitude of B would be exaected to increase at higher fluence of the lower plate (2.05x10 n/cm2) to about 175 M)a/%Ni. A nominal average shift to yield stress change ratio, AT/4cys, is about 1 F/MPa (note, this may be higher for coarse grained stccls). Hence, a difference in nickel content of 0.45% corresponds to a shift difference of
' 79 F.
c'
- 01. 31, 90 10:19 AM ICE 7
+
. - =.
-n.
b) The shifts predicted by. Reg. Guide 1.99 Rev 211 for for the u plate (181 F) compositions and respective 1990 fluences differ bhper (110 F) versus lower 71 F.
1 c) Generally accepted theories suggest there are at Icust two contributions to embrittlement J
arising from: i) copper rich precipitates; and 2) so called rnatrix defcets At215. The i
4 i
copper contribution is Hieved to be relatively temperature insensitive while the matrix contribution increases with increasing mckel t215 and decreasing irradiation temperature Albl5. li,quation 22 of reference 13 can be used to estimate the effect of an d
increase of 0.45% nickel on the matrix contribution at 500 F and 2.05x1019 n/cm evaluation yleids a increment of 99 F; note this may be somewhat high since the i
temperature sensitivity assumed in E observed in the range of 500 to 550 F (quation 22 of reference 13 is larger than normally typically, about 1 F/F).
liased vr. these evaluations the shift in the lower 11 ate can be estimated by adding an 80 F i nicke1 increment to the upper plate estimate (245 ?) yiciding a total of 325 F.
Appmach 4-
-i A third approach is to apply a both temperature and stcel sensitivity corrections to Reg.
Guide Rev 2 predictions of the lower plate shift estimated from the ratio of the shift of the YR upper (plate to the corresponding s compositionally similar ASTM A302B Reference Correlation Monitor (RCM) steel irradiated in the same capsulel. This approach assumes that the behavior of the RCM steel is consistent with predictions of the Reg. Guide 1.99 Rev 2, which is At the higher fluence the RCM/YR upper plate ratio is 320/225 = generally the caset6, 1.42. The nominal temperature correction is assumed to be 1 F/F (sce below). This procedure yields a lower plate shift estimate of AT = 1.42x(181+50) = 328 F.
Approach 5 Data on low copper steels (<0.1%) from the French reactor' surveillance program can be used to estimate the temperature dependence of embrittlement and the matrix (copper independent) contribution at 500 F for low flux irradiation conditionsl?. The results of are shown in Figure 6. The data at various temp /cm2 The temperature square fit) to a common fluence of 2.7x1019 eratures were interpolated (by a In In least n
y over a wide range is a nominal 1.07 F/F. The shift (assumed to be primarily due to the matrix contribution) at 500 F is about 157 F. The nominal shift contribution for 0.2% Cu is ab 160 F13. lience this approach yields a total estimated shift is 317 F.
Recommendation Dased on the consistency of Approaches 2 to 5, I recommend a 1990 shift estimate for the YR lower plate of 325 F. Note that this is somewhat less than the preliminary trend curvc estimate of about 360 F but somewhat higher than what would be found from using a nominal "cyeball" mean estimate for the mixed data set shown in Figure 4.
i
- 6. Welds Shift Estimates Given the uncertaintics about the composition of the welds, temperature effects, initial properties, copper nickel synergisms and the like,I would recommend using the Eq. I to estimate for the axial weld of 230 F and for the circumferential weld of 360 F.
Assuming a matrix shift contribution of 160 F the circumferential weld shift is consistent with a precipitate contribution from a plausable effective copper content of 0.34E Note tt ?l, 90 le;19 M'
FM
=
some stects shown in Figure 4 with relatively high copper und nickel contents have lower shifts. Indeed, the " bounding" curve was set by a plate containing only 0.21% Cu and 0.17% Ni, presumably due to some undefined source of extra "sensuivity", llence, stects with both extra sensitivity and higher nickel and copper contents may ?xperience evenL higher shifts than the bounding curve in Figure 4
- 7. Uncertainties.
As noted la the introduction the shift estimates given in this repon are subject to very large uncertainties. I believe the values rnay be somewhat conservative, but in my judgement not excessively so. Indeed, much higher plausable estimates could be generated in some cases (e.g. see Hiser's analysis ). Large uncertaintics are associated with the effects of 2
flux and other variables not considered, temperature coastdown the very limited data base at 500 F, lack of information on(irradiations below 50019, compositions and initial properties of the actual vessel materials and the potential for very le w up ier shelf energies j
(which 1 believe could well be less than 30 ft lbs in some cases. ln view of these considerations,1 would recommend that a one standard devttion uncer)tainty estima least) 40 F be inchided in any probabilistic or bounding b',havior analysis.
- 8. Summary To summarize, my recommendations on 1990 YRV shift estimates are:.
I Upper Plate 245 F LowerPlate 325F Axial Weld 230F' Circumferential Weld 360 F I further recommend use of a minimum error estimate of i,
One Standard Deviation 40 F
- 9. References
- 1. Reactor Presure Vessel Evaluation Report for Yankee Power Station, YAEC No.1735 (July 1990)
- 2. A.L. Hiser, personal communication
- 3. G. M. Gordon and li.H. Klepfer, in Effects of Irradiation on Structural Mnterials, ASTM STP 426, p48 (1967)
- 4. G. R. Odette and G.E. Lucas, Irradiation Embrittlement of LWil Pressure Vessel Steels, EPRI NP-6114, (1989)
- 5. R. A. Wullaert, D. R. Ireland and A.S. Tetelman,in Irradintion Effects on Structural Allovs for Nuckadeactor Aonlientions, ASTM STP 484, p 20 (1970) d C'L.31. 9 0
'1C21G AM PCE-
9 5
- 6. G. R. Odette, P.M. Lombrozo and P. A. Wullaert, in Effects of Radiation on j
Materials: Twelfth Intemational Svoosiun!, ASTM STP 870, pS40 (1985)
- 7. T. J. Williams, P. R. Burch, C.A. English and P.H.Nide la cour Ray, in Ploc of the Third 1mernationni Symnosium in Environmental Derradntion of Noelear Power Systems, p 121 (1988) i
- 8. A. L. lawe, in Effects of irradiation on Materinit ASTM STP 1046, p201 (1990) 14th International Svmnosium Vol. 2 9.1 Pachut, Evaluation of HR3 Materinis Metallurgical Trend Curves. Draft Report, S 1985 l
- 10. J. R. llawethorne, Rndiation Fffects Information Generated on the ASTM Reference -
Correlation Monitor Steels. ASTM Data Series Publication DS 54 (1974)
I1. Radiation Embrittlement of Reactor Vessel Materials, Regulatory Guide 1.99 Revision 2, U.S. NRC (1988)
- 12. G. R. Odette and G. E. Lucas, in Effects of irradiation on Mnterials: 14th Internationn1 Svmnosium.V2 ASTM STP 1046, p323 (1990) -
i
- 13. G. R. Odette and G. E. Lucas,in Radiation Embrittlement of Nuclear Reactor Pressttr.c h
Vessel Steelt: An international Revits Y2, ASTM STP 909, p206 (1986) a
- 14. O. R. Odette, in Proe of the Second International Symnosium in Environmental D gmdntion of Nuclent Power Svstems, p 295 (1986)
- 15. G.E. Lucas and G. R. Odette, Ibid p. 345
- 16. F. W. Stallmunn, Analysis of the A302B and AS33b Stindard Reference Materials in SEYigilance Caosules of C' ommercini Power Reactors, NUREG/CR 4947:ORN!/IM-10459 (1988)
- 17. C. Brillaud and F. Hedin, In service Evaluation of French Pressurized Water Reactor Vessel Steel, preprint of paper submitted for publication in Effects of Radiation ou Matcrbils: 15th Svrnnosium ASTM f,TP(to be published)
I
- c 0 ?. 11. 90 10:19 AM 705
200 g
i j
t l
l f
U 4
e o
o gar,0 t$
g l
~
J*
1
$t = 0.14 x 10"n/cm*
$t = 0.9 x 10nicm*
i. T - 299C
$t = 0.36 x 10"n/cm2 T = 305'C T = 285*C
.I I
I I
i
-I t
.30 900 1000 1100 800 900 1000 1100 800.
900 1000 1100 i
Austenitizing Temperature (*C)
Austenitizing Temperature (*C)
Austenitizing Temperature (*C) i i
Figure 1 Yield stress changes as a function orausterizing temperature of a UCSB plate containing - !
about 0.40% copperand 0.86% nickel.
7 h?D
1 i
"ptical Microstructures of Coarse Grained M do e A533B Steels.
)
Temper 664 C,4 hrs, AC/ Stress Relieve 600,40 hrs, FC J
Cu Ni P
Cu Ni P. T e
T = 900 C 0.40 0.001 0.005 A
0.41 0.86 0.005 ^ = 1000 C 1
[~'
[i pC < y I
I' S'
5ET' r-1~..
- Y* *,th~f '
k N',
^^
n
.l
~
- qy sht
~_%a W
~
5<'.4_
. ~' o.
e
\\
.~t},
?
,, n% < -1 C
Cf Be-'
~.'.
Ta, _
ky
~
'p;.
A y,*
\\
3
- f ' l
$. {. Q,hphy
..q l
.. }b E,fS e.l h:- u R.Uf::fl g 50 m gs
-q...
~
J.;
- i
- s0 m
- ..r w
-c
,[. k j}h Figure 2 Optical micrographs (x400) showing the similarity in the coarse grain size of alloys PK (intermediate nicke]) and LA (Iow nickel).
CAD A &7.7F r.
a 150 i
e i
i i
(a)
LA 0.4Cu/0.001Ni/0.005P Qt = 1.1 x 1019 n/cm2
[
g2.5 MPafC 2 100 I
k i
s 50 280 290 300 310 T (*c) 150
- i i
i i
i 6
i i
(b)
PK-0.41Cu/0.005 P R
50 MPa/%Nig 100 M
Qt = 0.9 x 10'8 n/cm2_
(%
T = 305 C.
i i
LA-0.40Cul0.005 P 50 O
0.2 0.4 0.6 0.8 1.0 Ni (%)
Figure 3 a) The temperature dependence of the irradiation induced yleid stress increase of a course grained steel (LA); b) the nickel dependence of irradiation induced yield stress changes in coarse grained steels. (LA and PX) so
.>,n C *t. fd. 9 0 lo:is a 712
500 I
I I
I I I g-c 300f4t a
g
+
e
(
h
++
e+
g
_[
O> M O
O
.4 g
g+
+-
e v
k 8
m
+
100 1
+
- e of
-e s
e 50 i
t i
iiiiii 0.1 1
10 9
2 ot("0 n/cm )
Cu/NI/P Prod. Form Ref.
Cu/Ni/P Prod. Form Ref, o 0.19/0.55/0.011 B
9
+ 0.17/0,12/0.016 W
7 A 0.23/0.56/0.013 W
9 e 0.20/0.18/0.011 B
10 e 0.30/0,56/0.021 W
9 e 0.26/0.28/0.012 B
10 0 0.26/0.56/0.020 W
9 e 0.21/0.54/0.016 W
8-0 0.32/0.67/0.017 W
9 e 0.35/0.66/0.014.
W 8
o 0.19/0.55/0.011 W
9 e 0.22/0.60/0.015 W
8-A 0.15/0.09/0.025 B
7
- 0.42/0.60/0.018 W
8
+ 0.19/0.07/0.017 B
7
+ 0.40/0.59/0.011 W.
8
+ 0.24/0.25/0.024 B
7 o 0.18/0.18/0.011 B
1 m 0.21/0.17/0.033 B
7 (YR-Plate)
Figure 4 Shift data for 500110*F irradiations versus fluence and preliminary recommended trend curve.
C40 A 4 7/3?
k h *f
'41 C *t.
t r.
10 AV Tti
l l'.
<900 n
l s
g u
400 4
-?
_ 200 t
- .L A8 o BR3 + 1F/F
<l 100 eBR3xCi o
O YRP-1x$tnom.
50 e
A YRP-2x$tnom.
25 i
i l'
i 0.1 1
10 100.
1 ot(10'9n/cm, E>1MeV) 2 s
Figure 5 Yankee Rowo Upper Mate (YRP) and adjusted BR3 shift data versus Ouence.
I l
l l-l 1
l
. orcas 7tt i'
Il ' M J I M ^* '"'
9
~
i i
t H
l l
90 i
Cu/Ni/P'
$t - 2.7 x 10"n/cm e 0.09/0.65/0.012-8 O 0.08/0.65/0.009 80 O 0.09/0.64/0.017 Chooz w f.07*C/*C Plate C 1
1
_P-70
<F1 60 ChoozhD
~
Plate B Bugey 2 Plate C2 L 50 250 260 270 280 290 T (*C) i Figure 6 Interpolated shifts at 2.7 x IC*n/cm as a function.of temperature from 2
the French surveillance data base.
4
' CC80 A b7.17
.-m.,
...y 4
.