ML20205C129

Submits Findings of Ad Hoc Working Group on Selection of RTNDT Values from 820616-17 Meetings.Method Used by NRC in Deterministic Evaluation of Pressurized Thermal Shock Does Not Contain Overly Conservative Elements

ML20205C129
Person / Time
Issue date:	08/30/1985
From:	Vagins M NRC
To:	Hanauer S Office of Nuclear Reactor Regulation
References
REF-GTECI-A-49, REF-GTECI-RV, TASK-A-49, TASK-OR 820830, PTS02, PTS2, NUDOCS 8509200138
Download: ML20205C129 (41)

Text

.1 i,y, e

n arco UNITED STATES

1. a jo,,

NUCLEAR REGULATORY COMMisslON g

g WASHINGTON. D. C. 20555 AUG 3 n 1982 MEMORANDUM FOR:

S. Hanauer, Director Division of Safety Technology Office of Nuclear Reactor Regulation FROM:

M. Vagins, Chaiman AD H0C Working Group on the Selection of RT Values fiDT

SUBJECT:

FINDINGS OF THE AD H0C WORKING GROUP ON THE SELECTION OF RT VALUES NDT The AD H0C Working Group on the Selection of RT Values convened at 8:30 am, Wednesday, June 16, 1982 inRoom014,YOTthe Nicholson Lane Building of RES. The second day of meetings convened at 8:30 am, Thursday, June 17,1982 in the same place. Those in attendance were:

NRC NRL U of MD P. N. Randall R. Hawthorne G. Irwin J. Strosnider F. Loss A. Taboada M. Vagins ORNL HEDL M. Virgilio J. Merkle G. Guthrie The objective of these two days of meetings was to review present practice and, if necessary, recommend new practices for the detemination of the adjusted RT'.Nt a new method of estimating the adjusted value of for existing plants. After a full day of review, it was detemined f RT was needed. The second day of meetings was used to develop such a prbNdure.

4 In brief, the recommended method of setting the adjusted RT for existing vessels, to be used in the deteministic analysis gpTvessels subject to PTS transients, is as follows.

1.

For each weld procedure (weld wire, flux) identified as existing in a plant, choose the initial RT for that procedure from the gener$0T (RT data N$p) lied by the NSSS as the mean value N

vendors.

f 4

A l

1. [p-z mm=R:

AUG 3 01982 S. Hanauer 2.

For a given fluence, calculate the mean value of the RT shift NDT from the equation:

[-10 + 470 Cu + 350 CuNi] [ fg]0.27 A RT

=

NDT 10 Where Cu nercent by weight of copper

=

percent by weight of Ni Ni

=

fluence at I.D. of belt line region f

=

3.

The adjusted RT then becomes riDT 2ho

+ 06 RT

" 3DTo DT

+

+

NDT Where TT DTo Mean value of initial RT

=

NDT Mean value of RT shift

=

NDT NDT "o

1 standard deviation of mean RT generic data base

=

NDTo 1 standard deviation of the RT shift data base.

oa

=

NDT For plants considered:

2[8 2 + oa2 59'F 0

=

Recognizing the present over conservatism built into the HEDL developed trend lines for higher fluences, additional steps are required, at least until the curves can be corrected. Thus the following:

18 2

4.

For those cases of fluences greater than approximately 3.0 x 10 n/cra,

calculate the ART from the upper limit line defined in Reg. Guide NDT 1.99, Rev. 1. Figure 1.

Compare the value of ART thus developed NDT in step 2.

If the HEDL curve derived value of ART"be value of is lower, choose that value and proceed to step 3 above.

If ART derived from Rg. Guide 1.99, Rey, 1. Figure 1, with fluences NDT greater than 3.0 x 10 n/cm is lower use that value and proceed to step 5 as follows.

5.

The adjusted RT is then:

flDT RT

" N DTo

+ ART

+

NDT NDT(1.99) 20, A

m

AUG 3 01982 S. Hanauer Mean value of RT from generic data WhereKDTo

=

NDTo RT shift fran Reg. Guide 1.99, Rev. 1 ART

=

riDT(1.99)

NDT 1 standard deviation fran generic data base

=

og The value of 2b has been datenniaed to equal 34 F.

Thus, the g

alternate fonn of the ad,iusted RT becomes:

t4Di RT

+ ART 34 F RTNDT NDTo NDT(1.99)

=

6.

The b DT calculated by the above method shall be conpared to available Charpy V notch tests data fran plant pecific (or for B&W plants, integrated) surveillance tests. Any significant deviation fran the predicted values must be rationalized.

It is the finding of the AD H0C Working Group, that the methodology used by the NRC staff in the detenninistic evaluation of PTS scenarios does not contain overly conservative elements.

This groJp further finds that, due to unresolved uncertainties that exist in plant parameters, the use of mean values for RT and RT by then'selves might lead to t

NnT The YD kor 2/ag ac 2 + oa2 was added to reflect unconservative results.

these uncertainties. The attached reports presents a detailed discussion of the AD H0C Committee's findings. The report was reviewed by the AD H0C Committee members and their conments have b appropriately incorporated.

m

,ilton V ins AD H0C v rking Group on the Selection of RT Values t4DT

Enclosure:

As stated 1

[

[

THE FINDINGS OF THE AD H0C WORKING l

GROUP ON SELECTION OF RT VALUES NDT I.

INTRODUCTION The methods of determining the adjusted RT values for materials NDT in the beltline region of operating plants are presently defined by the provisions of NRC Branch Technical Position, MTEB 5-2.

This document was developed in the recognition of the fact that plants 1

i constructed prior to August 15, 1973, would, in most cases, not l

have sufficient material documentation to allow compliance with the I

l regulations established as of that date, to govern the material properties of plants to be constructed. Thus, methods had to be developed by which to use available data to estimate the fracture toughness of the pressure vessels of existing plan +s to ensure that the pressure-temperature operating limits imposed on older plants provide at least the same margins of safety as will be applied to new plants. Because of the sparcity of data available, MTEB 5-2 was, per force, developed as a conservative procedure for evaluating these data. The methods described therein have, since first published, been criticized by the nuclear industry, and others, as being too conservative and thus too limiting for long plant operation. With the advent of the recognition of the Pressurized Thermal Shock Problem (PTS), the importance of the selection of RT vahes has NDT taken on added significance. The luxury of maintaining a possibly overly conservative measure in the light of the extrer.e economic l

l

'2 burden this action could place on.the public, is a luxury we can no longer afford.

Based on this reasoning, it became the mission of the Ad Hoc Working Group on.the Selection of RT Values to critically examine the NDT present methods used to select RT values, to quantify, if possible, NDT the degree of conservatism involved in those methods, and if needed, to reconnend a new procedure to be used to evaluate RT Such a NDT.

new procedure would be less conservative, and yet by its application, still help ensure the safe operation _ of older plants under both normal r.r./ postulated accident conditions.

To aid in this work, a group of knowledgeable and experienced personnel from the national laboratories, the Naval Research Laboratory, the University of Maryland, and the NRC were assembled as an Ad Hoc Group to review this issue and reach a consensus position. This was done during a two-day meeting held in Rockville, Maryland, on Wednesday and Thursday, June 16 and 17,1982. Those in attendance were:

t

A 3

NRC NRL L,0F MD P. N. Randall F. Loss G. Irwin J. Strosnider R. Hawthorne A. Taboada HEDL M. Vagins ORNL M. Virgfilo G. Guthrfe J tbrkle l

e

-_=

4

.1 I

II. INITIAL.RTNDT ( NDTo}

II.1 Measuring RT NDTo Article NB-2331 of Section III of the ASME Boiler and Pressure Vessel Code specifies the requirements and acceptance standards for i

determining initial RTNDT(RTNDTo). The procedures require first determining the nil-ductility transition temperature (NDT) by drop weight tests described in ASTM E 208.

If charpy impact tests, e

conducted at a temperature no greater than 60 F above the NDT, exhibit at least 35 mils lateral expansion and 50 ft-lbs absorbed

{

energy, then the RT equals the NDT.

If charpy tests conducted

~

NDTo at NDT + 60'F do not exhibit the required lateral expansion and I

impact energy, additional tests in groups of three specimens are j

required to be tested at higher temperatures until the lateral expansion and impact energy requirements are met. The RT is NDT then taken as that temperature minus 60*F.

If charpy tests are not a

performed at NDT + 60'F or when charpy tests at NDT + 60*F, do not exhibit a minimum of 50 ft-lbs and 35 mils lateral expansion, the i

1 temperature representing these values may be obtained from the i

lower bound of a full charpy impact curve if it exists. Since these requirements were not in place when some of the presently operating plants were designed and built, the Materials Engineering 4

Branch, NRR, developed a Branch Technical Position, MTEB 5-2, f

l

" Fracture Toughness Requirements," which describes acceptable procedures for. making conservative estimates of the RT for such NDTo l

5-4 plants. The estimation schemes'are based on correlation data for a j-large number of heats of vessel material and are presented for the i

most commonly used material; specifically, SA 533 Grade B, Class 1, 4

and SA 302-B(mod) plant and weld material; and SA 508, Class II forgings. A significant conservatism in MTEB 5-2 is that estimates 1

{

of RT derived from charpy curves for SA 533 materials cannot be NDTo

}

lower than 0*F.

i-f II.2 Existing Data i

As stated above, two types of tests, drop weight and charpy, are required to determine initial RT in accordance with accepted NDT i

procedure. However, the majority of operating reactor pressure j

vessels were fabricated before these procedures were established and for most of these vessels only charpy data was required at the time. Typically, the data available for each plate, forging

]

and weld consists of three charpy test results at 10*F and complete l

charpy curves for the base material and surveillance weld.

It j

should be noted that in many cases the surveillance welds were not 1

always totally representative of the vessel welds.

]'

Based on the limited available data, several methods have been used to estimate the initial RT The NRC has used the guidelines of NDT.

l

6 Branch Position MTEB 5-2, described above, to obtain an estimate of RT Westinghouse (WCAP 10019) has also used MTEB 5-2 to establish NDTo.

RT for licensing applications. Combustion Engineering has f4DTo The first is to set RT preposed two methods for estimating RTf4DTo.

f4DT at the temperature 60*F below the 50 ft-lbs level determined from the mean value full charpy curve. The second approach is to set RTf1DTo at a temperature two standard deviations above the mean value of generic data applicable to the weld of interest. The generic data consists of all initial RTf4DT measurements from welds made with the same flux, but possibly different weld wire heat numbers. The CE data is presented in Table I.(I)* Babcock and Wilcox (B&W) has also proposed using the upper bound of generic data for setting the RT in the welds.(2)

The data included in the CE and B&W data basas are genericully derived from weld metal qualification tests. There are very,few cases where the actual vessel wall prolongations were utilized.

The samples from the qualification welds were not full thickness test pieces and, furthermore, except for a relatively small amount of archive material, most of the qualification welds were made at a later date than the vessels.

• fiumbers in superscripted parenthesis are references.

7 III. SHIFT IN RT NDT III.1 Measuring ARTN DT_

l The shift in RTNDT ( ARTNDT), due to irradiation effects is measured l

as the difference in the temperature at which the unirradiated and irradiated materials exhibit 30 ft-lbs energy in a charpy V notch impact test.

Estimating the A RT from plant specific surveillance NDT results is very difficult for two-reasons.

First, the surveillance weld often does not exactly match the critical vessel weld due to l

the fact that they were fabricated from different weld wire heats.

This is significant because in many cases copper was introduced into the weld from a coating on the weld wire, such coatings varying from heat-to-heat of wire. Second, there is significant scatter in the charpy data in both the unirradiated and irradiated materials.

Further, when one considers that there are only eight charpy specimens tested from each surveillance capsule at time, the ability to achieve a statistically valid value for the charpy energy at 30 ft-l lbs is open to serious question. Thus, there has developed a preference for using trend curves developed from a generic data i

base rather than individual, plant specific data.

l The magnitude of the ART is a function of fluence and material chemistry. Chemical elements to which ARTNDT is known to be sensitive include copper, nickel, and phosphorus. Other trace elements such I

d 8

as tin may also affect ART In general, the relation between riDT.

A RTriDT, fluence and chemistry can be expressed in the form:

A RT

= [A1 + A "1 + * * + An "n-1 + Bj aj aj +...+B ok "1 3E 3

2 n

19 NDT where A constants

=

s B

constants

=

s at,j = weight percentage of chemical constituents f

fluence, neutrons per centimeter squared

=

constant c

=

The coefficients, Aj, Bt and the exponent, C, are determined by multiple regression analysis using data from the commercial reactor surveillance program and from test reactor data. Until recently, the NRC used the relation presented in Regulatory Guide 1.99, to determine ARTNDT. This relation considered copper and phosphorous as the most significant contributors to radiation sensitivity.

However, evidence that low-nickel materials are less sensitive to neutron irradiation has been accumulating for several years. Based on this observation, an updated reactor surveillance data base was reanalyzed by HEDL for the NRC in the Fall of 1981.

This analysis included nickel content in the regression fit. The proposed equation for the mean value of ART is given by:

NDT

9 i

t ART

= [-10 + 470 Cu + 350 CuNi] [

19 ]0.27 (1) f -

NDT 10 l

]

where Cu precent by weight of Copper

=

1 Nj precent by weight of Nickel

=

fluence neutrons /CM2 (E > 1MeV) f

=

i To obtain an upper bound on values obtained using this equation, 48'F, corresponding to two standard deviations, is added at a I9 2

fluence of 10 neutron /cm. The upper bound equation is given by:

ART

= [38 + 470 Cu + 350 CuNi] [

g ]O.27 (2)

I NDT Thus, the upper bound curve data lies 26*F above the mean value i

curve given by equation (1) at a fluence of 1018 neutrons /cm and 89'F at a fluence of 1020 2

neutrons /cm. Some refinement of the upper bound curve is considered necessary because it appears overly l9 2

f conservative at the higher (greater than 10 neutrons /cm ) fluences.

However, for the range of shifts in RT currently of interest NDT in operating plants, the upper bound curves provide an acceptable fit to the data.

Utilization of equations (1) or (2) requires knowledge of the fluence level and the copper and nickel contents for the weld of interest.

I i

i It should be noted, that Reg. Guile 1.99 has been shown to be i

-. ~..--...-- -. -.

10 i

i unconservative as regards fitting the data base, for fluences lower

- than approximately 1019 n/cn8 -for high copper, high nickel welds, l

j and for a somewhat lower fluence for lower copper and lower nickel i

welds. This takes on particular significance when it is noted that it is at these lower fluences that KIa is determined.

All recent PTS submittals have used Reg. Guide 1.99 to establish their crack arrest positions, and thus, all such submittals are suspect.

I

[

h i

t 4

u 1

i i

1 e

1 b

h 1

1 P

i

11 IV. DISCUSSION OF OVERALL PTS SAFETY MARGINS 1

The process of choosing a rational safety limit involves a critical examination of all the explicit and implicit conservatisms and uncertainties involved in the methodology used to arrive at that choice.

J IV.1 _ Crack Driving Energy j

I The crack driving energy or the challenge to the pressure vessel, is, in the terms of the analysis used, composed of two parts: the h

pressure-temperature time history of the assumed transient and the 1

i shape of the flaw. This is true because in the format of fracture mechanics analysis, the load is given in terms of K, (MP/m) (Mega Pascals-square root meters).

Fracture mechanics is predicated upon the existence of a flaw, or crack like defect in the structure I

i l

being analyzed. The first part of this term is developed from the stress field generated in the pressure vessel by the temperature i

{

gradients and pressures imposed on that structure.

I 4

t In order to examine the effects of final water temperature during the transient (assumed to occur in the beltline region of the I

1 t

i

12 downcomer), the rate of cool down, and the effects of pressure in i

the vessel an idealized form of the transient was assumed as:

k i

Tf+ (Tj - T )e

• T-(t)

=

f i

P(t) = Constant where T(t)= Downcomer water temperature as a function of time, Tf= Asymptotic final temperature of water in downcomer Tj= Initial water temperature (taken as 288'C (550*F))

Exponential decay constant 8

=

Tine in transient t

=

P Pressure in vessel during transient.

=

It is recognized that the overcooling transients that have occurred are really not characterized by smooth functions such as that r

i above. Neither is the pressure during the transient a constant.

These functions are used to allow a very large number of calculations to be run (well over 6000 runs at last count and continuing) such

{

that the full gamut of postulated transients can be effectively enveloped and studied. The choice of a reference transient with T and 8 is based upon the judgement, arising out of an examination of actual plant experience, that such a transient is indeed possible.

for all types of PWRs presently operating in the U.S., and that the 4

i t

i l

13 i

probability of such an occurrence is of the order of 10-2 to 10-3 occurrences per reactor year of operation.

TheselectionofP(t)=

2500 psi is predicted upon the observation that for most transients i

observed, the systems could very well have been repressurized to the PORV, or safety valve settings. Thus, the choice of this l

transient involves some small degree of conservatism, but not much.

i

{

This is especially true when it is noted that stratification of incoming cooling water could allow the inner wall of the beltline region to be subject to cold water temperatures of 40' to 70 F.

Thus, weighing all factors, the reference transient assumed is deemed to be appropriate with little or no conservatisms.

l The second part of the crack driving energy, K, involves the type of flaw assumed to exist. The analyses, in all cases, assumed the existance of a long flaw (essentially defined by one dimension, depth) penetrating through the cladding into the base material of the pressure vessel wall. This assumption must be regarded as conservative. The likelihood of such long flaws existing in operating reactors must be small.

It is more probable that small flaws, where the length of the flaw is approximately of the order of six times the depth of the flaw, could exist. This issue involves the physical characteristics of the cladding, particularly its fracture toughness, as a function of neutron irradiation.

If the cladding, which is assumed to be very tough at the start of life for the vessel, remains tough regardless of the neutron irradiation to 4

m..

14 which it is subjected, then it is extremely unlikely that a small flaw will initiate.

If, however, the toughness of the cladding i

does degrade with fluence, than it is likely that a short flaw will initiate, become a long flaw and then propagate into the vessel wall as would be the case if a long flaw existed from the start.

t' From the above, it is judged that some degree of conservatism exists in the way that the reference Ks. or crack driving energies, are developed for the PTS study. Because of the lack of data on irradiated cladding toughness (Type A 308 cladding in particular) no real bounds can be placed on this degree of conservatism.

It could be significant or it could be very small indeed.

Several different research programs are addressing this issue with results forthcoming in approximately one year's time.

It should be noted that no explicit factor of safety has ben applied to the developed K's.

When carrying out deterministic analyses for design purposes it is traditional to apply factors of 7

+

safety to the K's.

For instance, in Appendix G of Section III of the ASME Boiler and Pressure Vessel Code, it is noted that the i

pressure induced primary membrane stress carries with it a factor f

of 2.0 when allowable pressure are being calculated, i

IV.2 Fracture Toughness-Crack Initiation i

1 For deterministic analyses conducted, crack initiation was judged l

1

4 15 4

to occur when K as calculated equaled K as derived from the K g

k Ic curve shown in Appendix A of Section XI of the ASME Boiler and Pressure Vessel Code. This curve is the lower bound of all fracture mechanics specimen data, with loads applied slowly and specimen sizes ranging from crack widths of 1-inch to 12-inches.

If the 4

initial flaw were actually small and limited in length, the use of the K curve would be conservative, the actual initiation value Ic i

l would be probably more like the mean value of all these data rather than the lower bound. However,experiencehasshown(theHSST Thermal Shock Experiments iSE-5, SA, and 6 and one similar experiment I

carried out by Framatome in France) that if the initial crack were i

j.

long, or to become long, early in the transient, then the K curve g

i i

)

or lower bound value does realistically represent the initiation value for long flaws. This situation, again, emphasizes the importance of the fracture toughness of the irradiated cladding, and makes it extremely difficult to quantify the level or degree of conservatism 4

in the analysis. Sensitivity studies have shown, that next to the RT value used during the analyses, the use of mean values or t4DT i

lower bound values for fracture initiation play the most significant role in deterimining at what fluence level a given transient will j

cause crack initiation. Thus, as before, the cladding toughness l

)

effect will determine the degree of conservatism in the choice of 1

fracture toughness values; significant or very little, if any, t

i

]

i i

l

=-

16 r

IV.3 Fracture Toughness-Crack Arrest As for crack initiation, crack arrest values were developed from the crack arrest or K, curve shown in Appendix A of Section XI.

g This curve, too, is a lower bound value of all the dynamic initiation and crack arrest data developed from the data base of laboratory specimens. Though the use of lower bound values for crack arrest I

l also could carry with it a degree of conservatisn, it would be less than that value for crack initiation.

Further, since in the analysis i

long flaws were assumed to exist for crack initiation, the use l

lower bound values for crack arrest is consistent and appropriate, again as shown by the results of TSE 5, 5A and 6. What is controversial, i

l but not necessarily conservative, is the assumption that during the l

analysis, if crack arrest, using the K, curve, could not be calculated g

before the K 's at the unstably propagating crack tip reached a g

l value of 220 MPadii(200 ksi/fn), then the crack did not arrest and the crack would penetrate the full thickness of the vessel. This l

I was taken as vessel failure.

t i

This approach is incorporated in Appendix A of Section XI. There

(

h is much debate about the value of this approach.

It is pointed out j

that as the crack rapidly penetrates the vessel wall, it will enter r

vessel material that is at upper shelf temperatures, as determined by Charpy V Notch impact tests.

It is postulated that at these l

1 P

.. - - - -. =_

i i

17 1

temperatures, Linear-Elastic Fracture Mechanics (LEFM) is no longer valid and that the crack would arrest because of large crack tip a

plasticity effects. This plasticity phenomenon has, as yet, to be shown either analytically or experimentally for a thick section i

condition wherein the crack is being driven with a rapidly increasing K field while penetrating an increasingly (up to a upper shelf g

value) tough material. There does exist some data that indicates that for low value upper shelf material (say of approximately 20 j

ft-lbs), the ability of the material to resist a rapidly propagating i

crack in a rising K field does not exist, and flat fast fracture g

i j

surfaces are generated. On the other hand, for materials that i

exhibit a high charpy upper shelf value, say in excess of 150 ft-lbs, then the K or K, curve could be extrapolated upward to some 4

IR g

value.

(The Japanese have shown this in a least one series of l

tests for values of K, of approximately 400 MPa 8( 370 ksi/1n).

~

j g

Evidence is lacking as to the relationship between the upper value 4

of the K or K, curve and the charpy V notch upper shelf toughnsss IR g

values. This takes on added significance when one considers the reduction in upper shelf values caused by neutron damage. All the vessels under consideration have signficant fluence and some of these vessels had low (below 75 ft-lbs) upper shelf values at start i

i of operation.

j t

From the preceding discussion, the use of 200 ksi An, as the limiting i

I I

T f

18 i

j value for crack arrest, may or may not be conservative. However, due to the limited data available, it should be used as a limiting value until it can be conclusively shown to be incorrect.

IV.4 Warm Prestressing Effect I

1 l

Warm prestress effect (WPS) has been shown (TSE-5A) to be effective j

in either preventing initiation or limiting reinitiation after an j

arrest event in a large class of postulated overcooling scenarios.

This class is characterized by a K field that typically reaches a g

maximum value in time during the transient and then decreases.

It l

is defined by the statement that somewhere in the transient, the 1

K 's at the crack tip will be reducing when the fracture toughnesses g

of the material at that crack tip is reduced to the value of the t

l applied field; i.e.,

t a

j at time t = ti l

KIl Eli 0.0,

> 1.0 (4) hti KIci

\\

If a transient were to be well defined, and the K 's well known g

throughout the time period of interest, then the staff would be j

willing to use the concept of warm prestress to evaluate the s

l 1

i i

m

19 probability of crack initiation. One example of such an application, is the Large Break Loss of Cooling Accident (LBLOCA).

In this scenario, the pressure in the reactor vessel is rapidly reduced to near ambient with no possibility of repressurization. The developed K field will exhibit a maximum, reducing to zero as the tirne in g

transient progress.

This is typical of thermally induced stress fields.

In such an application, if WPS were shown to limit crack initiation, the NRC staff would have no problems in accepting this position. However, in the postulated PTS scenarios, the possibility of repressurization, with pressures up to the safety valve settings, does exist. Such repressurization could cause the K field, as g

developed during the transient, to show no clear cut maxima. This being the case, equation (4) could not be applied. What this really means is that if the transient cannot be well defined, no benefit can be taken, before the fact, for wann prestressing.

This position must be assumed to be conservative for it has been shown that in many of the actual transients that have occurred to date, WPS either did play or could have played a role if the vessels involved had more severely degraded material toughness than was the actual case.

It should be noted that for many of the thousands of PTS scenarios run for the parametric PTS study, WPS would have intervened in each case early on in the transients. This occurred due to an artifact of the way the scenarios were developed.

The key issue 9

20 was the imposition of a constant pressure, thus making each problem, essentially, thermally driven.

It should also be noted that for each of these cases, if the degradation of the vessel material were greater, then the vessel would have failed before the intervention of WPS.

From all the above, then, the NRC position of not considering the beneficial effects of_WPS must be taken as an unquantified, and in fact, an unquantifiable conservatism.

IV.5 Probabilistic Sensitivity Studies Probabilistic sensitivity analyses performed by NRC also provide some insight regarding existing margins and the effect of uncertainties associated with the methodology used to establish the RT NDT These studies were not meant be used in an absolute manner, but rather to help put bounds on the uncertainties involved in the deterministically developed analysis.

The probability distribution associated with crack depth directly affects the estimate of reactor pressure vessel failure probability.

If a probabilistic model with only crack depth as the random variable is considered, the failure probability will decrease from 1.0, in the deterministic case, to whatever the probability of existence of critical size cracks is estimated to be.

However,

21 since flaw distribution is the random variable with greatest uncertainty the degree of conservatism associated with the postulated existence of the critical flaw size is difficult to quantify. This is.particularly true because the critical flaw sizes are relatively small and the current NDE techniques are not considered highly reliable for near surface defects.

It should be noted that Section XI of the ASME Code requires an explicit factor of safety of Ein evaluating the acceptability of known defects. Lack of an explicit safety factor in pressurized thermal shock analyses conducted to date acknowledges, to a degree, the conservatism in assuming that critical flaw sizes exist with certainty.

Regarding fracture toughness, sensitivity studies conducted using the NRC probabilistic model indicate that reactor pressure vessel failure probability will decrease by one to tw orders of magnitude if K and K are assumed to be distributed about the mean value Ic h

of the fracture toughness data used to develop the ASME Section XI toughness curves. As stated previously, the fracture toughness exhibited will be related to the flaw geometry and cladding condition.

Other results of NRC probabilistic analyses indicate that uncertainties expressed in terms of variability, in other' random variables such as fluence, copper content, and initial RT can change the NDT reactor pressure vesselifailure probability by approximately an order of magnitude. Uncertainties in the variability of random 3

1

22 variables, however, should not be confused with uncertainties in the mean value of the random variables which has a much more dramatic effect.

Figure 2 presents the failure probability of a reactor pressure vessel given that it is subject to a specified pressurized thermal shock transient versus the parameter T -RTNDT' f

where T is the final, minimum downcomer water temperature in.the f

transient and RT is the adjusted RT based on the mean values NDT NDT Three different cooldown NDTo, copper content, and fluence.

of RT rates, 8, are considered in the figure and a constant pressure of 1000 psi is assumed. The significance of the results is not in the absolute failure probabilities calculated, but in the sensitivity of the calculated failure probabilities to T -RT or a of f

NDT.

S of 0.15 min-I, a 15 F difference in T -RT causes approdmately f

NDT an order of magnitude change in failure probability. Thus, relatively small errors in T or RT can cause significant changes in reactor f

NDT vessel failure probability.

IV.6 Adju,ted RTNDT Many elements of uncertainty are involved in estimating the adjusted RT NDT (RTNDTo* A NDT) f r a specific vessel weld. These include uncertainty is estimates of RTNDTo; unc n y. n uence esd mates; and uncertainty in weld chemistry.

These uncertainties are the result of inappropriate or insufficient weld certification and surveillance g.

23 material, experimental measurement error, and true material and mechanical properties and chemistry variability. An accurate estimate of the uncertainty in estimating RT requires considering NDT the entire estimation procedure from start to finish and evaluating the error at each step, and carrying the error through the analysis in an appropriate fashion. To provide a fair estimate of the RT NDT' caution must be exercised to not consider the same contributor to uncertainty more than once.

As indicated previously, there is a desire, and in most cases, a necessity, to establish RT on the basis of generic data. Table NDT I presents the data bases for initial RTNDTd grouped genericaHy according to weld flux type. For simplicity, the working group assumed normal distribution for the data supplied although it is recognized that this may not be accurate for all cases. However, the assumption is made to be verified at a later date, and the following procedure is developed.

Figure 1 illustrates the RT estimation process based on generic NDT data. The figure indicates that variability exists in RT as NDTo well as inaRT An important consideration is whether or not NDT.

variability associated with RT NDTo with ART The working group decided that these variabilities NDT.

could be. treated as though associated with independent statistical distributions.

This assumption is considered acceptable because

24 the initial RT and the ARTNDT are measured from different sets NDTo of tests, with different physical origins as described previously.

This assumption of independence allows the overall variability in adjusted RT to be calculated as follows:

NDT 2

R ART NDT NDTo NDT The variability associated with RT can be M e M ned in a NDTo relatively straight forward evaluation of the data. The variability associated with the ARTNDT, however, is more complex.

This is so because estimation of ARTNDT, from equation (1) requires knowledge of the fluence level and copper and nickel contents, all of which have uncertainties associated with them. The regression analysis performed to develop the ART trendlines considered allowance for NDT error in fluence. This was accomplished by assuming that the reported fluences were accurate to within + 35 percent and allowing the fluence to seek a value within the 35% error band that minimized the sum of the squares in the regression analysis. Thus, the standard deviation of 24*F associated with the trendlines developed by HEDL accounts for uncertainties in fluence.

Variability in copper and nickel content are not easily incorporated in the surveillance data based estimates of ART NDT' "9

data base certaintly includes some component due to uncertainties

-~

2-y a-p.

w w

25 associated with the reported values of these elements, it is believed that the value of these elements is-known more accurately in the surveillance data base than in the real RPV welds.

Tables III and IV indicate the variability associated with copper content in real RPV welds. Table III is based on a Babcock and Wilcox analysis and Tables IV is based on chemical composition analysis of i

reactor pressure vessel prolongations conducted by ORNL. The Tables indicate that the copper content can have a standard deviation as large as.07 percent by weight.

Equation (1) indicates that a

.01 percent by weight error in copper content results in a 6 to 8 degrees Farenheit difference in ART' DT. Uncertainties of this N

magnitude are very significant when operating in the transition region of the fracture toughness curve where the toughness drops very rapidly with relatively small increases in RT NDT' i

3 w

P m

e 4

1. .v,

,r gv

~r u

a

26 o

V.

GUIDELINES FOR DETERMINING PLANT SPECIFIC ADJUSTED RTNDT Based on all the preceding information, the working group decided there did not exist an overly conservative margin in the analysis carried out by the NRC staff and considering the importance that the adjusted RT plays in this analysis, the use of mean values NDT for RT and ART only would not supply the necessary margins NDTo NDT of safety required during postulated PTS scenarios.

Thus, it is recommended that the adjusted RT for a specific weld from the NDT generic data base should be accomplished by using the mean RT N DTo '

and the mean ART values from the HEDL trend lines and adding a NDT value of 2do2,,2 to this total. Or in stepwise format as follows:

1.

For each weld procedure (weld wire, flux), identified as existing in a plant choose the initial RT as d e NDT NDTo mean value of the appropriate data from the set of generic data supplied by the NSSS vendors.

2.

For a given plant fluence (at the I.D. of the beltline region),

use the following equation to determine ARTNDT*

A RT

= [-10 +'470 Cu + 350 CuNi] [

b (1) f NDT

]9)g where Cu = percentage by weight of copper Ni = percentage by weight of nickel f = fluence at I.D. of beltline

27 The values for the copper and nickel will be as reported by each plant, taken from the generic set of data, or defaulted to.35 for Cu and.75 for nickel if no other data exists.

3.

The adjusted RT is then:

NDT NDT NDTo + ARTNDT + 2 RT RT 0 + 'A

=

where RT

= Mean value of RT from Generic Data NDT NDTo A RTNDT = Mean value of RT Shift (from equation (1))

o = -1 standard deviation from generic data base o

ca = 1 standard deviation from surveillance data

base, i

The value of 2do 2

+ A has been determined to equal 59 F.

Thus, the adjusted RT ecomes RT NDT N DT "

NDTo (c = 17*F*, o3 = 24 F)

+ ARTNDT + 59 F.

o It is the finding of the Ad Hoc Working Group that this figure of 59 F fairly reflects the remaining uncertainty in the selection of the adjusted RTNDT, and reflects the other uncertainties inherent in the method of analysis used to evaluate the effects of postulated PTS scenario.

• This is true for all cases except for B&W weld WF-70.

For WF-70, ao = 30*F

28 I

Recognizing the present over conservatism built into the HEDL developed trendlines for higher fluences, additional steps are required, at least, until the curves can be corrected. Thus, the following:

4.

For those cases of fluences greater than approximately 18 3.0 x 10 n/cm, calculate the RT fr m the upper NDT limit line defined in Reg. Guide 1.99, Figure 1.

Compare the value of ART thus developed to the value of ART NDT NDT derived from the HEDL curves as described in step 2.

If the HEDL derived value of ART is lower, choose that NDT value and proceed as in step 3.

If the value of ARTNDT derived from Reg. Guide 1.99, Rev.1, with fluences 18 2

greater than 3.0 x 10 n/cm is lower use that value and proceed to step 5.

5.

The adjusted RT is then:

NDT RTNDT = RTNDTo + A RTNDT (1.99) + 2o, where RT

= Mean value of RT from generic data NDTo NDTo ART Shift from Reg. Guide 1.99, Rev.1 NDT (1.99)

NDT ao = 1 standard deviation from generic data base.

The value of bo has been determined to egial 34*F.

Thus, the l

29 f

s alternate form of the adjusted RT eco m NDT NDT NDTo NDT(1.99) +

+^

Finally, for any given plant, the RT developed as above shall NDT be compared to available charpy V notch data on plant specific material and any significant deviation must be rationalized.

l r

4 r

(G f

u p j. % ~ 3

.ll } v.) y r

s

, pD TABLE G;4

~.

~

[, -

t/CA Initial Reference Temperature for

&e

.g, &e,v q~ t.1

q..t

' Representative Submerged Arc Weldments Drop Weight RTNDT 50 55' Material Identification NDTT (*F)

(*F)

'Te - F-T-4 Wire Ht. !51912, Linde 0091 Flux

-50

-50 7 /Os8 U

Wire Ht. !83640, Linde 0091 Flux

-70 /0 /'#

7F Wire Ht. #83642, Linde~ 0091 Flux

-80

-80

-z o 72 J/

Wire Ht. !83653, Linde 0091 Flux

-80 r o fa 6B Wire Ht. !83648, Linde 0091 Flux

-80

-80

-r o rf rV Wire Ht. i4P5174, Linde 0091 Flux

-50

-50 */ o 40 78 Wire Ht. #83637, Linde 0091 Flux

-50

-50 V / o /z5 ff Wire Ht. #83650 Linde 0091 Flux

-40

-40 er,

/r/

T3 Wire Ht. !5P5622, Linde 0091 Flux

-40

-40 o a ff 4o Wire Ht. #83646, Linde 0091 Flux

-40

-40 / so /z/

8#

. J. E Wire Ht. !2P5755, Linde 0091 Flux

-50

-50 + /o /!r. -/S Wire Ht. !4P6052, Linde 0091. Flii-~

-50 7 50

-/ o x

Wire Ht. !87005, Linde 0091 Flux

-10

-10

-5 5

Wire Ht. !87600, Linde 0091 Flux

-70 to

/co 44e Wire Ht. #88118, Linde 0091 Flux

-70 ro BV SS Wire Ht. !4P6524, Linde 0091 Flux

-70 / o /o5 4/

Wire Ht. !87011, Linde 0091 Flux

-50

-50 e/o Br 53

~

~'

'--~ Wife Ht'.' !86998; Eihde 009f'F1'ux

-10 T10 oo

+ze Wire Ht. !88112, Linde 0091 Flux

-70 9-70

-r5

- y.5 cp -70 Wire Ht. !88114, Linde 0091 Flux

-70

.y o

.f c

-40 eio

+4 Wire Ht. !87000, Linde 0091 Flux

-40

-50

-fo

.-So Wire Ht. !V89476, Linde 0091 Flux

-50 p r F Y{. 5

-50_

-50 j.,s

-Wire Ht. !30502, Linde 0091 Flux Wire Rt. !4P6519, Linde 0091 ~ Flux T0 -

-60 o gg Wife Ht7i90077, Linde 0091' Flux

-60

-60

- f :. -

~y Wire Ht. !90128, 'Linde 0091 Flux

-60

-60

_fr

.y Wire Ht. #90069 Linde 0091 Flux

-60

-60

- z. o Wire Ht. !90146, Linde 0091 Flux

-50

-50

-.y

,_ o Wire Ht. #90130, Linde 0091 Flux

-60

-60

z. s

- t y' Wire Ht. i90211, Linde 0091 Flux

-60

-60

-y4

., 3 Wire Ht. !90149, Linde 0091 Flux

-60

-60

z. y

-/4 Wire Ht. !87003, Linde 0091 Flux

-20

-20

+ / r-

1. f4-Wire Ht. i90157, Linde 0091 Flux

-60

-60

- 3. o

-se Wire Ht. 490132, Linde 0091 Flux

-60

-60 zy

-fg Wire Ht. #90209, Linde 0091 Flux

-50

-50

-34

-3o Wire Ht. #89204, Linde 0091 Flux

-60

-60

- e s.

-u (f,f' Wire Ht. !SP7388, Linde 0091 Flux

-30

-30

- f.i.

~'-

.r.,

f.

'[ s[pp,,,4 9

Wire Ht. #3P7150, Linde 0091 Flux

-30.

-30

.- g

-g Wire Ht. #89833, Linde 0091 Flux

-50 3 ;-

,3 g Q

F Wire Ht. i90159, Linde 0091 Flux

-40

-40

-yy

.: y v

>u /

Wire Ht. i90067, Linde 0091 Flux

-70

-70

-6y

-g!

t Wire Ht. #90154..Li.nde 0091 Flux

-40

-40

-yg

_ y y" p

Wire Ht. !89827, Linde'0091 Flux

-80

-80

- f g, nh[A

'Q W

Wire Ht. #89828, Linde 0091 Flux.

-50

-50

.3 o Wire Ht. !4P7656, Linde 0091 Flux

-70

-70Tro

~ pf "g l

\\

y g.J g c, hj)D"_

V '7 (

I 6-13

I 1 I G

jG="(

V Q w

1

(,

f

'yl ),

e- ~.

\\

0gf M ).

TABLE 6;"(Cont'd)

/w g

l Drop Height RTNO

(*F)T [W4 3s Material Identification NOTT (*F)

(

ni; I. t j t 6 !

rre. h

.I Wire Ht. !89022, Linde 0091 Flux

-30

-30 Wire Ht. !90071, Linde 0091 Flux

-80 n

-z c, !

Wire Ht. !87603, Linde 0091 Flux

-60 So

- ro; Wire Ht. !90069, Linde 0091 Flux

-70

-70

-4 o

-7o}

Wire Ht. !89408, Linde 124 Flux

-60 zo

-/5!

Wire Ht. #3P7246, Linde 124 Flux

-60 /p

2. y i

._ -S0 z o i

Wire Ht. #3P7317, Linde 124 Flux

_ z.

Wire Ht. !E56906, Linde 124 Flux

-80

-30 e z z.

,r., o ;

Wire Ht. !651 A708, Linde 124 Flux

-80

_80

-zo zc Wire Ht. !91764, Linde 124 Flux

-60

-60 g sy(/ ypy; Wire Ht. i91762, Linde 124 Flux

-50

-50 yg

_z a j Wire Ht. !4P7927, Linde 124 Flux

-80 ap

- 2 c, ?

Wire Ht. #4P7869, Linde 124 Flux

-70 /z

-ty Wire Ht. #5PS366, Linde 124 Flux

-80

-80

-2=

- z. 4 !

Wire Ht. f69025, Linde 124 Flux

-70

-70

-yo

-yo Wire Ht. !39833, Linde 124 Flux

-60

_-60

-S

,3 Wire Ht. !90144, Linde 124 Flux

-50

-50 /- t o

+g Wire Ht. i3P7802', Linde 124 Flux

-30 z 4 i - so f,

50

-/5

-/s t

Wire Ht. #5P9028, Linde 124 Flux

-80 4 ' 50 Wire Ht. #3P3013, Linde 124 Flux

-30

-/S !-/s' Wire Ht. i4P3632, Linde 124 Flux

-30

-30

- yg

,.go Wire Ht. #3P3013, Linde 124 Flux

-60

-60

- / z..

fg Wire Ht. !5P974?, Linde 124 Flux

-80 /o l f a Wire Ht. !LP5P9744, Linde 124 Flux

-60

. 3o 32 Calvert Cliffs il Surveillance Weld

-30

-80

z. g !

3o Calvert Cliffs #2 Surveillance Weld

-60 z o

_3 3 Fort Calhoun Surveillance Weld

-50

-50 o

i f6

{

l-Millstone #2' Surveillance Weld

-60 s 7 '

Forked River Surveillance Weld

-70 y7 !,2 7g

-50 3 g! _,,pf SONGS !2 Surveillance Weld

-50 SONGS #3 Surveillance Weld

-60

'-34 f z4 i f fo Waterford #3 Surveillance Weld

-80 go ' -gg ANO-2 Surveillance Weld

-10

-10 #ff

1. fg Maine Yankee Surveillance Weld

-30

-30 p-/ g Weld-2, Low Copper Program (Ref. 6-5)

-50

-50

+5 i

yo -

o Weld-3, Low Copper Program (Ref. 6-5)

-70 f4 -35

+5

'o r

Weld-4, Low Copper Program (Ref. 6-5)

-30 s-70

-/o

_fo I

6-14 6

t

E Tchle &.

Measured Vs Estimated RT f N"ld*

ET Diff. between Cy properties RT"

'F measured and Drop vt

Temp, Energy, Lat exp, estimated TNDT, F F

ft-lb mils Meas Est RT ;LT., AF t

Case

. Croup A - Automatic and Manual Submerged-Arc

, yo 'D7 Welds Using Linde 80 Flux f6o i?*

+

JV/*-70

- yo Nf'"g ~~ w Al

-40

+20 56,57,54 53,55,53

-40~ -

+20

+60 kY A2

-40

+80 54,50,51 51,47,51

+20 /

+20 0

A3

-30

+50 53,54,56 48,52,54

-10

+20

+30 4

A4

-40

+70 53,52,55 51,52,58

+105

+20

+10 A5

-20

+00 50,50,53 42,49,4'6

+206

+20 0

A6

-40

+70 52,52,60 38,39,47

+108

+20

,+10

,50,51,57 44,43,55

+10-

+20

+10

'~

A7

-20

+70 A8

-30

+70 61,54,51 56,48,45

+108

+20

+10 A9

-30

+50 53,65,59 44,45,59

-10

+20

+30 A10

-20

+40 65,68,68 35,36,40

-20

+20

+40 Croup B - Automatic and Manual Submerged-Arc

// s/d 8'^ # Y 1"

VfI

. h-Welds Using Linde 0091 Flux B1

-50

+10 98,88,'80 71,66,61

-50

+20

+70 p

B2

-70

-10 110,50,85 83,44,65

-70

+20

+90

~

B3

-60 0

110,121,95 80,86,68

-60

+20

+80 B4

-80

.-20 56,50,85 45,35,66

-80,

+20

+100 B5

-60 0

90,92,91 55,65,65

-60

+20

+80 1-B6

-60 0

90,91,97 52,52,54

-60

+20

+80 4

B7

-70

-10 72,74,82 54,58,65

-70

+20

+90 l

B8

-80

-20 102,103,107 67,74,81

-80

+20

+100 B9

-70

-10 89,92,102 65,70,76

-70

+20

+90 B10

-60 0

54,80,88 43,52,59

-60,

+20

+80 Group C - Manual metal-arc welds C1

-90

-40 52,70,85 46,60,71

-90

+20

+110 C2

-70

-10 114,120,127 63,83,88

-70

+20

+90 C3

-100,

-40 50,53,78 40,46,59'

-100

+20

+120

[r C4

-70

-10 51,75,79 43,61,83

-70

+20,

+90 C5

-80

-20 78,82,112 64,73,91

-80

+20

+100 3-20 Babcock 8.Wilcox tv

&fif a

s

ntration for 177-FA sp Beltline k' elds Cu concentration, ut Z lf Std Category v/m TR

• Mean dev'n Ab g

J, 3

0.20

.M2 W 181 0.07 3'

O.11 0.27-0.31 1 0.07

.*4 31 0.18 0.27-0.31 O.07 1

3I 0.16 0.27-0.31(q0.07 g

g' ;,

g k3 0.22 0.27-0.31C.

0.07 1

0.29 0.35 0.03 1

0.46 0.35 0.03 1

0.27 0.35 0.06 1

0.22 O.31 0.02 f

1 0.20 0.31 0.02 1

0.19 0.31 0.02 2

0.11 0.18 0.07 2

0.19 0.18 0.07 2

0.19 0.18 0.07 2

0.14 0.18 0.07 1

0.21 0.24 0.03 1

0.26 0.24 0.03 O

~

1 0.17 0.25 0.05 1

0.29 0.25 0.05 1

0.20 0.26 0.05 1

0.19 0.26 0.05 2

0.16 0.23 0.07 1

0.27 0.24 0.05 3

0.21 G/2T-4 3b 0.07

'1 0.25 0.21 0.03 2

0.22 0.29 0.07

\\

e

/

S k

Babcock s, Wilc0x f 'b

i&

(AIW5..

STATISTICAL EVAlljATION OF COPER CURENT EXENSIVE GiARACTERIZATION OF TE CEMICAL'COEFOSITION

. IN SEVEN EBS WAS CONDUCIED BY OPliL*

i-

~

.EG

'feu 6Eu av 0.286 0.020 26 61W f.'""'62W 0.181 0.020 281 63W 0.307 0.013 139

. i E#,4 0.359 0.018 239 i

65W 0.215 0.019 248

.: u-66W

~

0.423 0.058 128

=.

.i

j 67W 0.265 0.030 99

~

c.

'N ~ f o

., g.g ETANDARDJFviATIONGEu = 0,01Y

• ERSONNEL C0FTHilCATION: G.D. MilTFAN TO J. STROSNIDER FARCH30,1981 ji l-I I

e 4

-4

.m.

+-

1

---p4--

+a a

ue r.;

b Je4--

J+4 -

A+-

t e

C d

g 1

g 4

S

~l

'I I

1'

%<ia:cty4 M,Cabb i< A A

.. t

+

~

f$b?llk V/ff/f/I!

t

i i

.~~

s a#J;<r l

• I, NT

! } --

1 k

OfGWf 1 1

.f F.

)

t 1

i i

P 1

1 4

0 i

I i

i4-4 I

1 l

i r

1.

GIG 2

-t s

MdMe..

4-

-+c.

_+

s..

1 m.

%.,,.. 6

/m

,s...

.s-..

c.,.4

,\\..,.g i

k N sQ 7

F_ w r.-

4 -= s - - - - - -

.s-4 z.. =.-i. - g u a.=: rT' \\ '

r' 4 ' ' r- ' ' i-* -JerM"M i I - - "

f= :r _- _.. _..'.. " - - -

===ir =1==\\==T = \\ = =.i.: ~== :1: =_

i-

- --a. W W., ), w w g"

.-J3.M M -

g p_:_L_T_ - Y__-.5_31_. _.-. 2 ~

_=.i___L_s_ _6_ r. g _- = =_.E, 7' _ _

. _. __.'_.. - ____.._ _. - _3.]. __ _ - 4....--. MN h-d n.

.__..-,.y.

g,

3 g

... 6. --n o gh v...,i....,.6.,

• a w-=.m ih.... i

--,,, --. 4 :. 4 w_. \\ g -,.,

=.,

. = = g;;; n

= ~..a s,-

' s.

.... -w=3 = : =- --, = = =n = _--- $ _ _- - \\ g :-.:--n., ; r= =. - -

=

e

' if " " "* '#M52.d.-N ~~_~.~-._~ 55

^=.5-d=I-5 bO,._I -k, P'_Z'.--'=": r-N;.'5E= N, =M=~ @m~M. b.'" #-

O7

_~~j

' m-

~.

-. -... - - - = - -

- :.=:-

2

- - - - - - -. _. _ c

_ _A -- _ _T r;= ' E_12_- 3 = - -- =_ _

=_ _-.;i E~- T.:~'7""#---=-t---2

=

=

g

..a

, m

>L-. V w

\\.

=

v 2

\\~

1

,i

+-

t 0

m,

i.

'w

. i 1

a iii i.

pn

.3 i

ie

.F

... i i

p.. v _a.. _,

m. -,

J.T.% -...

• - -. ~" r1 c - - m.-

_-.s.....

i % - ____. y.m _ _x c_-

,, ie - -., _.: A.g. 3 r-s w

,-..p r.k s a se,..

-. m.

n-

' - ;--Q :.G.= g_.;.^ -

p-w,=2.'-'=p

= = -::-' a : --

g

==, _ _ w

_,,, w

.w= & =

_ =.i- __

_= = = 4.,. _ q - -

. =. :

y* i :-. - ------ ^: ^.-- m _n --- _..i--- - -------._.== L ^ :

. = - _. = - _ = _ _ '..L---..-._.'-.,._.

. w : _=._: e _ _ _ = _

I r~

--=.---~~W mas

- q=.w.. :. a mu =_;.- _

-. m.;_ _ ;

_:----..=_-___._=-=.-_.2.___:_:.

- -==.=_._;=m=t-v-

~=-._=:_.- -.

a_

m g

___ 4_,, --4,. _ a __ :; _----- -

=-----m-===_:=-=__5.

--..:._;_=.- _--_: _

1.-

-, ** -L"- # "W-

..W -~K ~= '. ~~ ~ -

3

^*

L?-

M -*

t*--"-.'.".t.~K'=1..=t*"...=.~*.=m'X.""_*"l'm p u.

- n &

7-m,.

-c-t_':

" r. ~rw.

i w,~w ~.

se tw-. =:-._-== :.. - - c

.n -

u a

- T-4 = Y_*. 'S~hW_ '~ P_

bh_==i # _=.=__L_k_-~S_ _ ___._=.- _?_.'_~.= _

_'.'A_-A_ g~_.=-E. _==_ ___~r.- _4_ =_ _.

V. _- =. - ?:EM_ _==h. _6_~' :_.O. _-.=-~--3.-.

a

_ ~ _.

s V'".~_5?'. ~'t^._=. - t.I". '15 ^. '5-~=~S 5 'l-M355&W-~%.MU~- _k~^%W-5==?= ~~-.~.~' % ---?P'"..=#

'~~~5 h-

~= L= - ' ?~- - ~

5+

i

..- -==-:------

k

--.__.==_::=~-,---_--=-=-=aH

=2-_-1$_...-..---.-=.c=----=-=-=----=__=_:==.=-=-.=---

.:-_--1-_

. - - - - - - = =. = = = - - -

_t

=

y t-

- - - + - -._.

_=r_-

y p= n- --._.

• -1 D

i

-,m r

\\

l. c 5

-m - %

i i

c r

~.--s

.ug.. >

g

. g a

i f

1 i

i c.u m-wm

~ = %... _ u-- %, m 3 c_ %.. _..

a...-.f..<

=s.~-~-'-~"_.

h u.~-3,

==_==+t c.:- --; ; ----=..

.t s.u ~-. --L.. _-,->^~. -.:.1-^

~ - - =T~ r. L : - '. D -: s.= V-. er s -- -

a 'a-g n

=-.

=

r--

- =-_ _ = _ -.=_:=.=c-=__=-.

n

..m

. _ = _....

._n_=_=_-=m_--x=_.==_?_-="=-_=-~

u

- - - - - ~

rw n===1-y w._ =_---==_==_=w_.._.__=_;w_1=_._=___=._=i=_+.=-=--.

-~ :

c:_=__-=_=_m______-x_=__=_-===_..=_n..__u_=_-mu_-_==__..=_.__=a m:: m.,._=_.3 m,

g

-- - -====

=

>-=..==------_=-_==::===_-=-==4.-_.==.===:.:_=-n_np:=:::-_-.-==_,.==-_:..==.==.

=

==-

- = - -

w

-=

_ = - - =. _-=.._an-- ~-W.,

,,m\\ -c.:m,w.

=

w-.m.

w.

x x. - - a =.- x,

-.m.m

- -~ - w n- =--

W=w..

W'_".P..+-^~U====~~"='.'.",.E.M_

- - = _

w' --b =._ m:-E -,'-,*-.!""_^=-"._***'^",-.. *.. " " ~

_ _ '..*'-~.%.=+v Wc

+

r S

.- =-

w

=w.x c,.,

=. -

_._ ====- === = = = -

w-.

'===m#hMMW=..N==M=:MM-H NWM= =MW.-=5 Sc=r_W M-zren.-- -=WL X:N17

_---=_~_~_..,-r._-.~~'=-..'...__--._._--2.*~~

=- _=.-

=--

-.. -_~= -__ " ' " = * = _ _ - -.. ~ ~ * ~ = = = = * * - * ' ~ ~ ~.~

._=.y-^;==.,'.'_.--.___-_-_~*

_';*._.'*'"r.__-=

r___-_

t 1,

., u..

..om m

-- -.m__._- %._., %.~

e _w[~-.~_.o

.. - ~....

tr-=---===w=--'-==--

-==..------:-++m=.~=-=,--=?=m.:--

=-=m

=-_ : c -

---=--+um-,-~-

m=~-vm m m

^

e 455~ =..^.~5h=5 :_=- Odi. =E5k h. '== --A '.~5i~?5 5. 5.?-~ = w5lra

=.

... ' ". - ~. '. _ = ' '.

. m:

' ~~

1 h_- tW=3=Wr.G ='----~

.... = - -

~

-=__.-_==._=_.=--

. = m...; x _===

==. =-

-==.-=--===1===---=.-=-.--=-.==--_-

s rg_:===_ -== - = = m=-== = = =- -;- =x===.= ; w=_ _ _5__=m p.; - = = - = = _..====g== =;__-;_-_...._u___;;--2

=;-m

_ = =

w.

__=__

__;3---,

m-__

• L

. *-.,r a n-- - w _ = =, n a. w----

.~w,-.u.

.- w w _

,w

+ --- ~

n.-. -: n..._

" J M AnmM.

--a-.h'r-h_-.-'2-"i.M f -.7=

• h M *Ji C. y _1 s p2 Jull 4 '*.4 } y a'*

^ ~", f t *- a# : _ J t F {.'

f M WS * *- J.*R er**- * : _ rc. -"

t 7_-'T*'

'*'h

['Y ~2l.'~^-A^'5~ 0 WW2h AD- = ' 5_ = ' ?=~h 0 W. ' '_~

l. %^~'~

E,7 9~~ t

• '3***~'*=.

~ ~ ~

T--_"_? - ---- ~~ - i - =_=._=_ht J ~~~~Y - M =.5-E~~ F-"'B Y=3 5 p.C-d i' 9 3 8 f O rt W _'O,'-

w_ _+-E ~- =v

~

6=__

--_._ -=::..-:=..=.---;-=:--

3

< = _. :-=

.=

__. _. = - - _. :===:=..--.---

=.

w.:==-=.=..---.:

- - - - - - = -. =

_ _.. _ _ __- = = - _ _._. _.m=.---

. - _ - - __ - _ - - -. - _ _. =., - -. - - - _

3

- _. _-- _g.=_---

_ = - - - - - -.a_=-----=

+:

__ _ _ _ _ _ - - - - =

h -_h

. = *

+

a-_

i

.L 1

.1 1

1, 1

a ary-#. y usa-s.wa yum.uma_sr= = - -'emi e eg @m-4 a3-or=

e-e rv,.w.4*.+4_*

r 4 n,. e-. a v.

~

a s.

aJ

- ( =m.e t *- - e -a -r* ] s 9 -

- 6.. a s..

^ ' " -

'**'4*'.***~~'.'*'

• ' ~

• %='"&'.~r=~r~="'

E_ '..,.'.4*-. ' ' ' '....._~^'~*^~==~=^W-'"-W-._"-M'~~'"~'. = =

. _ - _ _._:._-_-_-__=---_-,-..-'_"W.

"[^*

_-- " " ---~:=--~---~=-.=_--.::=---.**~M-

~~

~

~T b^

C*'*-~

^*..E' T*

- - - =

I

_. _ - _ - - - -.* * * ~.. -... =_--_.._---.=!

b -b Z -.-

• • ~'

. k

~

..---..=-_==b'"_"_--.=._;__;_,.__:-.____.,____-.-

-.h

_ b

?

g

___._T.:-'.=-

=

__ --_-_=.

-,w

,.m. s t c.. :m v m s -:.e=,,-., - - m ww. -s- "w. n-m n

=r -

. ~--a-"

=3

~

~

- ~ ~ ~ '- - - ~ - -

r-W-

s -

- ^ ^ - - - = - - - = -

ez--=

-[

[';

h

"-^

~- - -

=

.r-m.- = -r: =.=

- - - -==

- - - - - =. - - -. - =.---

... =.- -

?==~.-.

= ;-- = m -

c--~^ m *

= =s' M=__...-===.=-.=xt.--.==_=_=-=w=.-.=----....

.. - - - = = ~ - - ~. -= - - - = = -

~ - - - - -

. - - - = -.

^

'"Z"~~~~.-*--

L^1-.

._=.__--.__;_=____.,

_ - =

._C

.. = * " * * - ~ = = - ' ~ ~ _ - - ____

,.;..._____7 2---

_ ;_.; = ; '._ L.T 0.7_*

• 7J _.

-~**~*T'.""""=;--~~=*="

~L=.'"-

= = ~ ~ ~ - ~ * ' "

- - - - ' - _,. _ _ _ =. =. _ -. _. - -.

- - ~ ~ ~

1..,_^~...'.,--.\\%,

-~

^

-4T- '- = -

mx n

}

~

~

.P.

,1

- r

-G O

-F

%g*

$],Je a

REFERENCES :

1.

Evaluation of Pressurized Thermal Shock Effects Due to Small Break LOCAs with Loss of Feedwater for the Combustion Engineering NSSS by Nuclear Power Systems Division Combustion Engineering, Inc.

December 1981.

2.

Methods of Compliance with Fracture Toughness and Operational Requirements of 10 CFR 50, Appendix G, BAW-10046P, March 1976 (B&W Proprietary) 3.

LWR Pressure Vessel Irradiation Surveillance Dosimetry Quarterly Progress Report for period January - March 1982. HEDL-TME 82-18, to be published about September 1, 1982.

0