ML18153A301

Safety Evaluation Denying Licensee Assessment of Reactor Vessel Structural Integrity

ML18153A301
Person / Time
Site:	Surry
Issue date:	04/20/1998
From:	NRC (Affiliation Not Assigned)
To:
Shared Package
ML18153A300	List: ML18153A299 ML18153A301
References
NUDOCS 9804230078
Download: ML18153A301 (14)

Text

e e

UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D.C. 20555-0001 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION ASSESSMENT OF REACTOR VESSEL STRUCTURAL INTEGRITY SURRY POWER STATION, UNITS 1 AND 2 VIRGINIA ELECTRIC AND POWER COMPANY

. DOCKET NOS: 50,.280 AND 50-281.

1.0 Introduction

• By letter dated September 25, 1997, Virginia Electric and Power Company, the licensee for Surry Units 1 and 2, responded to a May 22. 1997, NRC request for additional information (RAI). This RAI, issued as part of the NRC closeout for Generic Letter (GL) 92-01, Revision 1, Supplement 1, requested an assessment of the application of the ratio procedure, as described in Position 2.1 of Regulatory Guide (RG) 1.99, Revision 2, to the pressure temperature (PT) limits curves and the low temperature over pressure (L TOP) limits for Surry Power Station Units 1 and 2. The staffs assessment of the licensee's response is provided below.

2.0

Background

The NRC issued Generic Letter (GL) 92-01, Revision 1, Supplement 1 (GL 92-01, Rev.1, Supp. 1), "Reactor Vessel Structural Integrity" in May 1995. This GL requested licensees to perform a review of their reactor* pressure. vessel (RPV) structural integrity assessments in order to identify, collect, and report any new data pertinent to the analysis of the structural integrity of their RPVs and to assess the impact of those data on their RPV integrity analyses relative to the requirements of Section 50.60 of Title 10 of the Code of Federal Regulations (10 CFR Part 50.60), 10 CFR 50.61, Appendices G and H to 10 CFR Part 50 (which encompass pressurized thermal shock (PTS) and upper shelf energy (USE) evaluations), and any potential impact on low temperature over pressure (L TOP) limits or pressure-temperature (PT) limits.

More specifically, in GL 92-0.1, Rev. 1, Supp. 1, the NRC requested that addressees provide the following information in their responses:

(1) a description of those actions taken, or planned, to locate all data relevant to the determination of RPV integrity, or an explanation of why the existing database is considered complete as previously submitted; (2) an assessment of any change in best-estimate chemistry based on consideration of all relevant data; ENCLOSURE

,,-----=~=-~==-~---

9804230078 980420 --- -----

PDR ADOCK 05000280 P

PDR

2 (3) a determination of the need for the use of the ratio procedure in accordance with the established Position 2.1 of Regulatory Guide (RG) 1.99, Revision 2, "Radiation Embrittlement of Reactor Vessel Materials," for those licensees that use surveillance data to provide a basis for the RPV integrity evaluation; and -

(4) a written report providing any newly acquired data as specified above and (1) the results of any necessary revisions to the evaluations of RPV inte-grity in accordance with the requirements of 10 CFR 50.60, 10 CFR 50.61, Appendices G and H to 10 CFR Part 50, and any potential impact on the L TOP and PT limits in the technical specifications, or (2) a certification that previously submitted

• evaluations remain valid. -- - - * - --

Virginia Electric and Power Company, the licensee for Surry Units 1 and 2, responded to GL 92-01, Rev. 1, Supp. 1, by letters dated August 10, and November 20, 1995. The November 20, 1995, letter indicated that portions of the Surry response to GL 92-01, Rev. 1, Supp.. 1 were contained in Babcock & Wilcox (B&W) Topical Evaluation, BAW-2257, Revision 1, "B&W Owners Group Reactor Vessel Working Group Response to Generic Letter 92-01, Revision 1, Supplement 1," dated October 1995. This topical report indicated that all licensees addressed in the report had determined the best estimate copper and nickel contents of their plant's beltline and surveillance welds* (Part 2 of GL 92-01, Rev. 1, Supp. 1 ). The report also indicated that the ratio procedure described in Position 2.1 (pages 1.99-3 and 1.99-4) of RG 1.99, Rev. 2, need not be applied to the PTS assessments and USE assessments of RPV beltline welds made from Linde 80 fluxes (Part 3 of GL 92-01, Rev. 1, Supp. 1 ). To discuss these conclusions of BAW-2257, Revision 1, a meeting between the NRC and the B&W Owners Group (BWOG) was held on April 30, 1996.

During the April 30, 1996, meeting, the utility representatives indicated their understanding that the ratio procedure requirement was established to accommodate utilities whose* plant-specific surveillance programs did not include weld metal fabricated from the same heat of weld wire as the limiting beltline material (i.e., the beltline material which is predicted to experience the most limiting transition temperature shift). The BWOG representatives disagreed with the practice of using the RG 1.99, Rev. 2 Position 1.1 correlation to modify measured transition temperature shift data (~RTNoT) obtained from samples fabricated from the same heat of weld wire as the limiting reactor vessel beltline material. Such modification, it was argued, was unnecessary since the unmodified transition temperature shift measurements obtained from samples fabricated from a single heat of weld wire already represented unbiased estimators of the mean transition temperature shift for beltline welds fabricated from the same heat of weld wire. It was further argued that the application of the ratio procedure to measured transition temperature shift values obtained from surveillance specimens fabricated from the same heat of weld wire multiplies the effects of uncertainty associated with (a) the measurement of surveillance specimen chemical composition, (b) the calculation of the mean beltline weld chemical

. composition, (c) the determination of the initial (or unirradiated) value of RT NOT (the reference temperature for nil ductility transition), and (d) the correlation of transition temperature shift with bulk chemical composition.

The NRC informed the representatives of the BWOG at the April 30, 1996, meeting that the information presented was insufficient to support the owners group's proposal. The BWOG

3 representatives indicated that the owners group would consider performing additional work, and would consider submitting another topical report for NRC review.

On March 6, 1997, further discussions regarding the statistical and physical validity of the RG 1.99, Rev. 2 ratio procedure as a means for quantifying and accommodating the effects of variation in beltline material chemical composition in RTNoT and RTPTs (defined below) were held. During this meeting, it was indicated that licensees were expected to either apply the ratio procedure as specified in RG 1.99 Rev. 2, or to provide an acceptable alternative

. approach for quantifying and accommodating the effects of differences in reactor vessel beltline

• and surveillance weld chemical composition.

On May 22, 1997, the NRC issued the GL 92-01, Rev. 1, Supp. 1, closeout response to Surry which indicated that the staff considered the RPV integrity data for the Surry Power Station to be complete at that time; however, it requested an assessment of the application of the ratio procedure, as described in Position 2.1 of RG 1.99, Rev. 2, to the PT limit curves and L TOP limits for Surry Units 1 and 2. This assessment was requested to be submitted by September 30, 1997. By letter dated September 25, 1997, the licensee for Surry Units 1 and 2 provided the requested information. The staffs assessment of the licensee's response is provided below.

3.0 Evaluation*

3.1 General Methodology RG 1.99, Rev. 2 and 10 CFR 50.61 describe the methodology used by the staff for calculating the effects of neutron radiation embrittlement on RPVs currently used in light-water cooled reactors. The following equation is used to determine the reference temperature for a reactor vessel material (note that the terminology used is from 10 CFR 50.61 ):

where:

. RT NOT = RT NOT(U) + M + LiRT NOT (1)

RT NOT (adjusted reference temperature, ART, in RG 1.99, Rev. 2) is the reference temperature for a reactor vessel material under any condition. It can be thought of as an estimate of the temperature at which a material transitions from brittle failure at lower material temperatures to ductile failure at higher temperatures.

RTNoT(Ul (Initial RTNoT in RG 1.99 Rev. 2) is the reference temperature for a reactor vessel material in the pre-service or unirradiated condition, evaluated according to the procedures in Paragraph NB-2331 of Section Ill of the. ASME Code (or other methods approved by the NRG).

M (Margin in RG 1.99, Rev. 2) is the margin to be added to account for uncertainties in the values of RT NOT(Ul* copper and nickel contents, fluence, and the calculational procedures. M is evaluated as follows:

M=2*Ja +a TT t:,.

(2)

~.'

where:

e 4

Ou (01 in RG 1.99, Rev. 2) is the standard deviation for RT NOT

cr~ is the standard deviation for LlRT NOT LlRT NOT is the mean value of the transition temperature shift, or change in RT NOT* due to irradiation and is calculated as follows:

LlRT NOT = (CF)*FF (3) where:

CF is the chemistry factor which is a function of copper and nickel content. The method for determining the appropriate CF is discussed below.

FF is the fluence factor and is evaluated with the following equation:

FF= f co.2a-o.10*109 1)

(4) where:

f is the best estimate neutron fluence in units of 1019 n/cm2 (E > 1 MeV) at the clad-base-metal interface on the inside surface of the vessel at the location where the material in question receives the highest fluence for

• the period of service in question.

The chemistry factor is determined by one of two methods in accordance with 10 CFR 50.61 and RG 1.99, Rev. 2. For the first method (Position 1.1 of RG 1.99, Rev. 2), the chemistry factor ("table CF") is evaluated from separate tables depending on the product form of the material (weld or base metal), and the best estimate copper and nickel contents of the steel.

The second method (Position 2.1 of RG 1.99, Rev. 2) is used when two or more credible

• surveillance data points become available from the reactor in question. The latter method yields values of "surveillance CF".

Determination of the credibility of surveillance data is a multi-step process. First, the following criteria must be satisfied, in accordance with 10 CFR 50:61 (c)(2)(i):

(A)

The materials in the surveillance capsule must be those which are the controlling materials with regard to radiation embrittlement.

(8)

Scatter in the plots of Charpy energy versus temperature for the irradiated and unirradiated conditions must be small enough to permit the determination of the 30-foot-pound temperature unambiguously.

(C)

Where there are two or more sets of surveillance.data from one reactor, the scatter of LlRT NOT values must be less than 28 ° F for welds and 17 ° F for base metal. Even if the range in the capsule fluences is large (two or more orders of magnitude), the scatter

5 may not exceed twice those values.

(D)

The irradiation temperature of the Charpy specimens in the capsule must equal the vessel wall temperature at the cladding/base metal interface within +/-25°F.

(E)

The surveillance data for the correlation monitor material in the capsule, if present, must fall within the scatter band of the data base for the material.

If the surveillance data meet these criteria, then the need for adjustments in the measured data due to the chemistry of the surveillance weld must be determined in accordance with the ratio procedure in Position 2.1 of RG 1.-99, Rev 2*, and 1 O CFR 50.61 (c)(2)(ii)(B): (NOTE:

adjustments for the chemistry of base materials is not required.) The PTS regulation and RG 1.99, Rev. 2, indicate that if there is clear evidence that the copper or nickel content of the surveillance weld differs from that of the vessel weld (i.e., differs from the average for the weld wire heat number associated with the vessel weld and the surveillance weld), then the measured values of aRT NoT should be adjusted by multiplying the measured value of aRT NOT by the ratio of the chemistry factor for the vessel weld to that for the surveillance weld (these chemistry factors are determined from the tables used in Position 1.1 of RG 1.99, Rev. 2). This procedure is frequently referred to as the ratio procedure, and gives chemistry-adjusted values of aRT NOT* With these chemistry-adjusted values of aRT NOT (if necessary), or the as-measured values of aRTNoT, determine an interim value of chemistry factor.from the following:

where n

~

(0.28 - 0.10 log I,)

~

[A, X J;

]

i=l CF~~~~~~~~

n

~ (O.S6 - 0.20 log /,)

~

[f,

]

i=l n is the number of surveillance data points Ai is the as-measured or chemistry-adjusted value of aRT NOT fi is the fluence for each surveillance data point (5)

Next, determine the absolute differences between the A values and the companion values of aRT NOT determined from f1 and the interim value of chemistry factor determined above. Each surveillance data point is considered credible (and the interim value of CF becomes CF) if this absolute difference is less than 28°F for welds and 17°F for base metal, or less than twice these values if the fluence range of the data is large (i.e., two or more orders of magnitude).

A positive determination of the credibility of surveillance data can affect the determination of RT NoT in two ways. As noted above, the existence of credible surveillance data results in the use of surveillance CF. In addition, with the existence of credible surveillance data, a reduced

e 6

valu.e of oA can be used in the determination of the margin term, M. In an absence of credible surveillance data (i.e., table CF is used), oA is assigned a value of 28°F for welds and 17°F for base metal, but need not exceed 0.50 times the mean value of LiRTNoT* With the existence of credible surveillance data, oA can be reduced by one-half, to 14°F for welds and 8.5°F for base metal. If credible surveillance data exist and the surveillance CF methodology (including reduced values of oA) yields a higher value of RTNoT than that calculated using table CF, then the surveillance data must be used. If the use of surveillance CF yields a lower value of RT NDT*

then either method may be used.

To determine RTPTS* the end of license fluence (the best-estimate neutron fluence on the expiration date of the operating license) is used as "f' in the above equations and RT p,-5-is equal-.

to the calculated value of RT NDT*

3.2 Licensee's Position Regarding Ratio Procedure The licensee questions the validity of the r~tio procedure for modifying measured values of LiRT NDT for several reasons. First, the licensee believes that the ratio procedure need not be applied if the surveillance data are from the same heat of material as the reactor vessel beltline material which the surveillance data is being used to assess. The licensee further states that surveillance specimens have historically beeri assumed to be the same material as the reactor vessel beltline when they were fabricated with the same weld wire heat number.

The licensee does agree that adjustments to the measured values of LiRT NDT may be warranted to ensure that they adequately (or conservatively) address the actual value of LiRTNoT experienced by the beltline material when:

(a) the chemical composition distribution (mean and standard deviation) of the surveillance specimens is not representative of the chemical composition distribution for a reactor vessel beltline material fabricated from the same heat of weld wire, (b) the difference between the beltline and surveillance weld chemical composition distributions results in a difference in LiRT NDT* and (c) the resulting difference in LiRT NDT reduces the margin of safety inherent in RG 1.99, Rev. 2 estimates of LiRTNoT for a given heat of weld wire.

The licensee described and critiqued four methods for quantifying and accommodating the impact of differences in chemical composition between surveillance and vessel welds fabricated from the same heat of weld wire. These four methods are:

(1) application of the RG 1.99, Rev. 2 (Position 2.1) ratio procedure, (2) application of selection criteria to eliminate surveillance results obtained from surveillance specimens which are considered not representative of the belt!ine material, (3) using all data regardless of differences in chemical composition between surveillance

e 7

specimens and the vessel weld, and demonstrating that the margin of safety inherent in RG 1.99, Rev. 2 is not diminished by the chemical composition differences, and, (4) ranking the surveillance results per RG 1.99, Rev. 2, Position 1.1 and then selectively adding surveillance data and observing changes in the calculated chemistry factor.

With respect to method (1 ), the licensee concluded that the ratio approach introduces additional uncertainty into the measured values of ~RT NDT* principally as a result of erroneous assumptions. The principal erroneous assumption cited by the licensee is that correlation of

~RT NDT with bulk chemical composition means that bulk chemical composition is the cause (licensee's emphasis) of transition temperature shift. The sources of uncertainti cited by the licensee include chemical composition measurement uncertainty, chemical composition sampling uncertainty, simplifying analysis assumptions, and uncertainty in the RG 1.99 Rev. 2

~RTNoT correlation. The licensee did indicate that the RG 1.99, Rev. 2 ~RTNoT correlation's tendency toward higher ~RT NDT values for higher copper and nickel concentrations is undisputed, but argues that use of this correlation to modify measured values of ~RT NDT is inappropriate.

The licensee also stated that RG 1.99, Rev. 2 indicates that the ratio procedure should be applied if there is clear evidence that the copper and nickel content of the surveillance weld differs from that of the vessel weld; however, the licensee states that if there is such clear evidence of different copper and nickel contents, then the surveillance weld is not representative of the vessel weld, and data from the surveillance weld should not be used to assess the condition of the vessel material (presumably under any conditions).

The licensee's second method for quantifying the differences in chemical composition involves the use of a selection criteria based upon the measured copper and nickel contents of each surveillance capsule and their relative ranking with respect to table CF. In this procedure, the first step is to exclude data obtained from surveiillance specimens which are judged to be not representative of the reactor vessel beltline material. To accomplish this, the licensee provided several criteria for determining "clear evidence" that the copper or nickel content of the surveillance weld differs from the average for the weld wire heat number associated with the vessel weld and surveillance weld, and then excludes data where such clear evidence is found.

Two of the licensee's five criteria for when there is clear evidence of chemical composition differences between the surveillance and vessel welds include:

i.

statistical evidence that individual surveillance material copper and nickel concentrations are significantly different from the mean copper and nickel concentrations of a reactor vessel beltline weld, or ii.

anecdotal evidence that differences in the weld wire, coating, welding procedure, specimen fabrication, or any other relevant weld feature may have resulted in differences between the beltline and surveillance material chemical compositions.

The licensee went on to analyze their data for heat 299L44 using criteria I, above. This evaluation suggested that the Surry surveillance specimens may not be representative of the larger population of 299L44 weld material on the basis of the observed chemical compositions.

e e

8 Nonetheless, the licensee concluded that differences in bulk chemical composition, by.

themselves, provide insufficient evidence to reject the hypothesis that measured values of

.6RT Nor from available surveillance materials are unbiased estimators of the.6RT Nor for a beltline weld fabricated from the same-heat.of weld wire.

The third method discussed by the licensee involves assessing the variation (i.e., standard deviation) of the measured.6RTNor values (without chemistry adjustment) from estimated values of.6RT Nor using the surveillance CF (Eq. 5). If this variation is less than oil (28 ° F for welds), then the margin of safety inherent in RG 1.99, Rev. 2 calculations is maintained and it is assumed that no other corrections to the data are needed. In contrast to the ratio procedure,*

this method replaces variability in-chemical composition with variability in measured.6RT 1>1or data as the key indicator of the representativeness of the measured surveillance data in representing the vessel material. The licensee concludes that this method is acceptable.

The fourth method proposed by the licensee involves the following steps:

1.

Determine surveillance CF (Eq. 5) using the as-measured surveillance data,

2.

Determine the standard deviation between the surveillance data and calculated values of.6RTNor using the surveillance CF,

3.

If this standard deviation is greater than 28°F, then a value of oil is assigned in the margin term to account for this excessive deviation.

4.

The surveillance results (i.e., measured.6RTNor) are then ranked in order of descending RG 1.99, Rev. 2, Position 1.1 chemistry factors (i.e., table CF). If two or more surveillance results have the same Position 1.1 chemistry factor, these surveillance results are ranked in order of descending neutron fluence.

5.

The surveillance CF is then determined for the 2 data points with the highest table CF values.

6..

If additional surveillance results are available, the surveillance result with the next highest table CF is included in the data set to determine a second surveillance CF.

Additional surveillance data are added and additional surveillance CF values are calculated until all of the surveillance data have been considered.

7.

The maximum surveillance CF is used to assess the material condition of beltline materials fabricated from the same heat of weld wire as the surveillance materials whose data was used in the chemistry factor calculations.

In summary, the licensee concluded that there are significant technical questions regarding the statistical validity of the ratio procedure and that standard statistical practices would suggest elimination, rather than modification of non-representative data. Furthermore, the licensee concluded that differences in bulk chemical composition between surveillance specimens and vessel welds provide insufficient evidence to reject the hypothesis that measured values of

.6RT Nor obtained from the surveillance specimens are representative of the mean.6RT Nor

e 9

calculations performed for the limiting Surry Units 1 and 2 beltline welds fabricated from the same heats of weld wire. Finally, the licensee concluded that if this hypothesis is not rejected, then the margin of safety inherent in RG 1.99, Rev. 2, Position 2.1 ~RT NoT calculations performed for the limiting Surry Units 1 and 2 beltline materials is maintained without -

adjustments to measured data to compensate for observed chemical composition variability.

3.3 Staff's Evaluation of Licensee's Submittal The observed shift in reference transition temperature (~RT NoT) of a material is attributable, in part, to the chemical composition of the material and the irradiation environment. Research has

-indicated that the copper and nickel concentration and the product form-(weld or-base-metal) - -

are the primary material variables in assessing the radiation embrittlement of RPV steels. With respect to the radiation environment, fluence and irradiation temperature are primary variables controlling radiation damage. RG 1.99, Rev. 2, correlates embrittlement to the chemical composition (copper and nickel concentration) and product form of the material, and the fluence to which the material was exposed.

To account for differences in chemical composition and irradiation environment between su*rveillance specimens and the vessel beltline weld, an adjustment procedure is necessary to ensure that the measured values of ~RT NOT obtained from the surveillance specimens can be used to conservatively predict the ~RT NOT for the vessel. The NRC, in Regulatory Guide 1.99, Rev. 2, provides the ratio procedure as a method to account for differences in chemical composition of weld metals. 10 CFR 50.61 states, in part, that "to verify that RT NOT for each vessel beltline material is a bounding value for the specific reactor vessel, licensees shall consider plant-specific information that could affect the level of embrittlement: This information includes but is not limited to the reactor vessel operating temperature and any related surveillance program results." To account for differences in irradiation environment (more specifically, irradiation temperature), the NRC has made adjustments (on a case-by-case basis) to surveillance data to account for differences between the irradiation environment of the surveillance specimens and the vessel in question.

The licensee's proposed alternatives to the ratio procedure in RG 1.99, Rev. 2 fail to account for the differences between the chemical composition of the surveillance specimens and the chemical composition of the vessel weld. As indicated by the licensee, there can be chemical composition differences between a surveillance weld specimen and a vessel weld fabricated from the same heat (and/or same spool) of weld wire as a result of variations in the amount of copper coating between, and along the length, of any given spool of weld wire of that heat. As a result, licensees need to account for these potential differences in chemical composition to ensure that the surveillance data can be used to conservatively predict the ~RT NOT which will be

• observed in the vessel. Other observations with respect to the licensee's alternatives are provided below.

The licensee's alternative to the ratio procedure which involves applying selection criteria eliminates "extreme" values rather than accounting for differences in the data and adjusting the

-data accordingly. Excluding data because it deviates from the mean value by a specified amount may be non-conservative since integrity (or failure) analyses typically are driven by these "extreme" values. In summary, data should only be rejected as a result of an invalid test.

10 The licensee's alternative which ranks surveillance results appears to simply show the variation in the chemistry factor as a result of additional data rather than assessing the effects of differences in chemical composition.

In addition to not correcting for chemical composition differences, the licensee's alternatives fail to account for the effects of irradiation environment (specifically temperature) when assessing the surveillance data. For irradiation temperatures around 550°F (Surry has an irradiation temperature near 536°F), a decrease in irradiation temperature has been associated with an increase in the observed 6RTNoT at equivalent fluence levels. EPRI report NP-6114 reported that for a large set of surveillance data there is a factor of an additional one degree of RT NOT shift (Le;,*f1RTNot) per one degree of irradiation temperature decrease. In addition, *ASTM

• STP-1046, Volume 2 indicated that there is a trend of about 0.7 degrees per degree change in irradiation temperature for Linde 80 welds similar to those found at Surry. As a result, surveillance data should be adjusted to account for the differences in irradiation temperature to ensure that the resultant chemistry factor will be appropriate. In the case of heat 299L44 (one of the more limiting heats in Surry 1 ), surveillance specimens of this heat at Surry 1 were irradiated at temperatures of 17°F to 21 °F lower than the balance of the surveillance specimens (which were irradiated at Three Mile Island Unit 1 or Crystal River 3). Specifically, the Surry data was irradiated at a cold leg temperature of approximately 536°F while the remaining data was irradiated at approximately 556°F. As a result, the measured aRTNoT for the non-Surry data should be adjusted upward approximately 1 ° F per ° F difference in irradiation temperature. This would result in increasing the measured 6RTNoT by 20°F for the

.non-Surry surveillance data.

In summary, the staff concludes that differences in chemical composition and irradiation

. environment between surveillance specimens and vessel beltline welds must be accounted for to ensure that analyses of the surveillance data will result in a conservative estimate of the 6RTNoT observed by the vessel. The ratio procedure (or another NRC approved alternative) is necessary to account for chemical composition differences. Specifically the ratio procedure accounts for differences in the chemical composition of surveillance and vessel welds to ensure that the measured values of 6RT NOT from a particular surveillance specimen are indicative, or conservatively represent, the actual value of 6RT NOT experienced by the vessel beltline material.

As a result, the licensee's alternatives to the ratio procedure are not acceptable since they fail to account for chemical composition and irraqiation environment (e.g., temperature) differences.

4.0 Staff Assessment of Radiation Embrittlement at Surry To ensure that the measured values of 6RTNoT from a surveillance capsule are indicative of or conservatively represent the actual embrittlement experienced by a vessel beltline material, adjustments to the data should be made to account for differences in chemical composition and irradiation environment between the surveillance capsule exposure and the vessel operating experiences. After these adjustments have been made, a surveillance value of CF can be determined according to Eq. 5. If the surveillance satisfy the credibility of RG 1.99, Rev. 2 and 10 CFR 50.61, then this chemistry factor can be used to project the 6RT NOT expected for the vessel at a given fluence, with ot1 reduced by one-half (14°F for welds and 8.5°F for base metals). If the evaluation of the surveillance data indicate that the surveillance data set is not credible and the measured values of 6RT NOT are less than the projected mean of the table CF

e e

11 plus the generic 2o,I!., the chemistry factor may be calculated using either Position 1.1 or Position 2.1 of RG 1.99, Rev. 2; however, the full margin term must be applied. The method

. chosen must bound all of the surveillance data to be in compliance with 10 CFR 50.61{c){2).

At Surry Unit 1, heat 299L44 is analyzed as part of the reactor vessel surveillance program and is.also one of the more limiting heats of material in the reactor vessel. This same heat has been studied as a part of the reactor vessel surveillance programs at several other plants.

Overall, results are available from surveillance capsules irradiated at Surry Unit 1, Three Mile Island Unit 1, and Crystal River 3, as summarized in Table 1. Surry Unit 1 {and Unit 2) has Westinghouse as its nuclear steam system supplier {NSSS), whereas the other plants have Babcock and Wilcox {B&W) as their NSSS.-- One particular difference* in the irradiation -

~nvironment of these plant.sis the irradiation temperature between the Westinghouse {i.e.,*

Surry) and the B&W plants. Specifically, the *cold leg temperature at Surry is approximately 536°F, compared to a cold leg temperature at the B&W plants of approximately 556°F. As discussed above, a decrease in irradiation temperature has been associated with an increase in

.6RT NDT at equivalent fluence levels {approximately 1 ° F change in.6RT NDT per 1 ° F change in irradiation temperature). If the remaining differences in irradiation environment variables {e.g.,

. flux, spectrum, etc.) are considered negligible, an adjustment of approximately 20°F should be made to the B&W data to make all of the surveillance data consistent with the actual vessel operating conditions at Surry. These irradiation temperature-adjusted values for the surveillance data are listed in the second-to-last column of Table 1.

TABLE 1:

AS-MEASURED AND ADJUSTED SURVEILLANCE DATA FOR HEAT 299L44 Capsule Fluence Measured Irradiation Temp. Adjusted Ratio/Temp.

Desig.

%Cu %Ni (1019 n/cm2)

L'lRTNDT Temperature (536°F) L'lRTNDT Adjusted L'lRT Nor (OF)

(OF)

(OF)

(OF)

CR3-C1 0.37 0.70 0.779 214 556 234 220.6 TMl1-E 0.33 0.67 0.107 124 556 144 147.6 TMl1-C 0.33 0.67 0.866 203

.556 223 228.6 TM12-1 0.33 0.67 0.830 182 556 202 207.1 TMl2-2 0.33 0.67 0.968 222 556 242 248.1 Surry 1-T 0.24 0.66 0.281 165 536 165 199.0 Surry 1-V 0.24 0.66 1.940 240 536 240 289.5 The chemical compositions listed in Table 1 are based on a letter from Framatome Technologies, Inc. {FTI) to the NRC dated July 10, 1997, regarding weld chemistries for Linde 80 copper-coated weld wires. Chemistries for surveillance specimens are based on the average of all available data from the source. The best estimate chemical composition of the Surry vessels is considered to be 0.34% copper and 0.68% nickel per the above referenced FTI document. Accounting for these differences in chemical composition in accordance with the ratio procedure in RG 1.99, Rev. 2, and 10 CFR 50.61, along with irradiation temperature,

12 yields the last column in Table 1.

The staff analyzed the as-measured data and the adjusted data several different ways. These analyses involved evaluating all of the surveillance data together regardless of NSSS vendor and evaluating the data in subgroups based on NSSS vendor. In addition, when the data were obtained from the s_ame source (i.e., from one plant), the data were analyzed without applying the ratio procedure until after the credibility analyses had been performed.

After reviewing the results from the various analyses, the staff believes that surveillance data from a given plant is the most appropriate data for assessing the material condition of that particular heat-in that plant's vessel. This is-due to a consideration that the irradiation* * *

• environment conditions to which the surveillance material was exposed are more applicable than the conditions at another plant. As a result of using just plant-specific data, no or limited corrections for irradiation environment are necessary in assessing the surveillance capsule results since the irradiation environment of the surveillance capsules should be similar to that of the vessel. In the event that data for a specific heat is not incorporated in the plant's surveillance program, data from other sources should be used as the primary data in assessing the plant's vessel. The staff believes, however, that surveillance data that requires fewer adjustments for normalizing the data to the plant's vessel will introduce less error into the assessment.

The above notwithstanding, standard engineering practice and 1 O CFR 50.61 require that "all data" for a particular heat of steel should be considered for verifying that RT NDT is a bounding value for the reactor vessel material under review.

For the case of heat 299L44 for Surry 1, the staff evaluated only the Surry 1 surveillance data for assessing the integrity of the Surry vessel. In this analysis, no adjustments for irradiation environment were necessary, and adjustments for chemical comp*osition were made as discussed below. The staff found the data to be credible in accordance with the procedures from RG 1.99, Rev. 2, and 10 CFR 50.61, as outlined above in Section 3.1. No adjustments for irradiation environment were necessary, but to account for chemical composition differences between the surveillance weld and the vessel weld, the staff applied the ratio procedure to the surveillance chemistry factor determined for the as-measured data. The resultant chemistry factor (259.0) was used in an assessment of RTNoT at 1/4 of the waff thickness (i.e., 1/4T) and RT PTs for heat 299L44 for Surry 1. The RT NDT at 1 /4 of the wall thickness (i.e., 1 /4T) is used in the determination of a plant's PT limit curves.

In addition to heat 299L44, heat 72445 is one of the more limiting heats of material for Surry 1.

The Surry 1 surveillance program does not have this heat of material; however, surveillance results are available from Point Beach 1 (Westinghouse NSSS) and Crystal River 3 (B&W NSSS), as described in Table 2. The irradiation temperature at Point Beach 1 is approximately 542°F and the irradiation temperature at Crystal River 3 is approximately 556°F.

13 TABLE 2:

AS-MEASURED AND ADJUSTED SURVEILLANCE DATA FOR HEAT 72445 Capsule Fluence Measured Irradiation Temp. Adjusted Ratio/Temp.

Desig.

%Cu

%Ni (1019 n/cm2)

.llRTNDT Temperature (536°F).llRTNDT Adjusted.llRT Nor (OF)

(OF)

(OF)

(OF)

Pt B 1-V 0.23 0.62 0.502 110 542 116 110.3 Pt B 1-S 0.23 0.62 0.829 165 542 171 162.7 Pt B 1-R 0.23 0.62 2.380 165 542 171 162.7 Pt B 1-T-0.23 -0.62 2.420 180 - -

- -542

-186

.. -.. 176.-9 CR3-1 0.22 0.58 0.510 148 556 168 168.0 CR3-2 0.22 0.58 1.670 168 556 188 188.0 As with heat 299L44, the staff assessed the heat 72445 data several different ways including separating the data into two subgroups based on NSSS vendor. In this case, the Point Beach 1 data required less adjustment to account for the temperature differences between the surveillance specirnens and the Surry 1 vessel. As a result, the staff evaluated only the Point Beach 1 surveillance data for assessing the effects of irradiation of heat 72445 at Surry 1. The Point Beach 1 surveillance data were adjusted by 6°F to account for irradiation temperature differences; all of the temperature-adjusted data are listed in the second-to-last column of Table 2.

After the temperature adjustment to the surveillance data for heat 72445, chemistry differences between the surveillance weld and the vessel weld were accommodated using the ratio procedure. In this case, the best estimate chemical composition of heat 72445 was considered to be 0.22% copper and 0.58% nickel per the FTI document mentioned previously. The final adjusted values of ~RT Nor applicable to heat 72445 at Surry, accounting for irradiation temperature and material chemistry differences, are provided in the last column of Table 2.

Analysis of only the Point Beach 1 surveillance data indicated that the data are credible in accordance with the procedures from RG 1.99, Rev. 2, and 10 CFR 50.61, as outlined above in Section 3.1. The resultant chemistry factor (144.0) was used in an assessment of RTNor at 1/4 of the wall thickness (i.e., 1/4T) and RTPrs for weld 72445 for Surry 1. The RT Nor at 1/4 of the wall thickness (i.e., 1/4T) is used in the determination of a plant's PT limit curves.

Table 3 provides a summary of the staff's assessment of RT Nor at 1 /4T and RT Prs for heats 299L44 and 72445, using the appropriate surveillance data described, with appropriate adjustments due to differences in irradiation temperature and chemical composition.

5.0 Conclusions Based on the above evaluation, the staff concludes that the licensee has not properly accounted for chemical composition and irradiation environment differences in their assessment of the integrity of their RPV. As a result, the staff concludes that the licensee should resubmit their evaluation.

I 1*

I 14 TABLE 3:

RTPTS AND RTNDT (1/4T) COMPARISON FOR SURRY WELD HEATS 299L44 AND 72445 Heat299L44 Heat72445 RTPTS RTNDT (1/4T)

RTPTS RT NOT (1/4T)

1. of 2

4 Capsules Ratio Yes Yes

. Procedure _

T adjustment No Yes

1. Credible 2

4 RTNDTU

-7.0

-5.0 Cu 0.34 0.22 Ni

• 0.68 0.58 CF 259.0 144.0 CF Method Surveillance Data Surveillance Data Fluence 0.8745 0.7421 1.354 1.240 Factor

.l\\RTNDT 226.4 192.2.

194.9 178.6 01 20.6 19.7 01,.

14.0 14.0 Margin 49.8 48.3 RT PTsfRJ NOT

. 269.2.

--* 235.o: *-.

238;3...*..

_j