ML17347A311

Forwards Addl Info Re Reactor Vessel Matl Surveillance Program,Analysis of Capsule V,Per 861121 Request

ML17347A311
Person / Time
Site:	Turkey Point
Issue date:	02/26/1987
From:	Woody C FLORIDA POWER & LIGHT CO.
To:	NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
L-87-94, TAC-62760, TAC-62761, NUDOCS 8703040113
Download: ML17347A311 (33)

Text

REGULATORY T ORNATION DISTRIBUTION 'SYS 8 <BIDS)

I ACCESSION NBR: 87030401 13 DOC. DATE: 87/02/26 NOTARIZED: NO DOCKET 4 FAC IL: 5g-250 Turkey Point Planti Unit 3I Florida Pouter and Light C 05000250 50-251 Turkey Point Plant. Unit 4I F)orida Pouter and Light C 05000251 AUTH. NANE AUTHOR AFFILIATION MOODY~ C. O. Florida Pouter 5 Light Co.

REClr . VaNE RECIPIENT AFFILIATION Document Control Branch (Document Control Desk) 0

SUBJECT:

Forutards addi info re reactor vessel matl surveillance programs analysis of Capsule Vi per 861121 request.

DISTRIBUTION CODE: AOOID COPIES RECEIVED: LTR i

TITLE: OR. Submi ttal: General Di str but ion I ENCL Q SIZE: /6 NOTES:

RECIPIENT COPIES RECIPIENT COP IES ID CODE/MANE LTTR ENCL ID CODE/MANE LTTR ENCL PMR-A EB 1 1 PWR-A EICSB 2 2 PMR-A FOB 1 1 PWR-A PD2 LA 0 PMR-A PD2 PD 04 5 5 NcDONALDi D 1 1 PMR-A PSB 1 1 PWR-A RSB 1 INTERNAL: ADN/LFNB 1 0 NRR/DHFT/TSCB 1 1 NRR/ORAS 1 0 OGC/HDS2 0 01 1 1 EXTERNAL: EQhQ BRUSKEi S 1 1 LPDR 03 1 1

'NRC PDR 02 1 NSIC 05 1 1 TOTAL NUNBER OF COPIES REQUIRED: LTTR 22 ENCL 18

yQ'S'I Sl. i 'l << ),,I Suufv tf j',V <<

I. ~

ff II tt u

S Ef!'t v '"

j S <<<<$ <<u

~ l' jfu I<<'V S'<<

ll V <>I., -. u V,t pffft J !p,"j:-3 '<<Vj~ Sf I ><<jS) p "'/ NSV u f) jS' )t<<>jtj,'<I ~<<" <<

<<I'"lf' 'Ti " v,,u 7 t uf') u,. s tf ) I P<<ftl<<uv.<<;l) " u!f S

'3 ( ~ I 'ff v "Sl<<

urt rt I'll' u

'*'j jj <<u '<< ft

<<uu j

lu '<<"

fv. Wu'4 v Slt "s 'I p g IS tt <<<<S( II r<<

Nu <<$ <<

6-.i)<<'l r

IV 'I

.r

P. O. BOX 0, JUNO BEACH, FL 33408-0420 yah)bj~

FE8RUARY 26 1987 L-87-94 U.S. Nuclear Regulatory Commission Attn: Document Control Desk Washington, D.C. 20555 Gentlemen:

Re: Turkey Point Units 3 and 4 Docket Nos. 50-250 and 50-25l Reactor Vessel Material Survei1 lance Program Request for Additional Information NRC TAC Nos. 62760 and 6276 I Attached is our response to your November 2l, l986 request for additional information regarding the Reactor Vessel Material Surveillance Program for Turkey Point Unit 3, analysis of Capsule V.

If you have any further questions, please call us.

Very truly yours, C. O. Woody Vice President Y'roup Nuclear Energy Attachment COW/TCG/cvb cc: Dr. J. Nelson Grace, Regional Administrator, NRC Region ll Mr. D. R. Brewer, Sr. Resident Inspector, Turkey Point Plant 8703040113 .870226 PDR ,ADOCN, 05000250 P ~

- PDR DD I /NRC/ I an FPL Group company

1

~ I e

V l'

~

tl

= ~ l I ~

I I I 4 V V I vtII" V

~ ~ ' E Il

~ t q,",.e V

'I V

V~

F V

• I

• II" tl I ~

t t ,

FJV ~ . ~

1Vvf ~ V f 'FV ~ F

ANSWERS TO NRC l}VESTIONS ON REACTOR VESSEL HATERIAL SURVEILLANCE PROGRAH ANALYSIS OF CAPSULE V TURKEY'OINT UNIT 3 guestion I: An experimental error analysis should be performed to support vessel fluence uncertainty values.

The estimated experimental error provided relates directly to the activity of the dosimeter at the time of removal (At~r) quoted in the revised Table X (see Response 2) of the submitted report "Reactor Vessel Haterial Surveillance Program for Turkey Point Unit 3; Analysis of Capsule V", SRI Project 06-8575.

The gamma ray emission rates in disintegrations per second (DPS) were determined with a germanium detector with an IT-5400 multichannel analyzer according to Southwest Research Institute (SRI) procedures which reference ASTH standards appropriate to processing dosimetry. In general, these standards are designed to yield gamma ray emission rates with an uncertainty of + 3X at the 68K confidence level. For the specific case of the determination of Ator for Capsule V, the following three sources of error were identified along with an estimate of their magnitude:

a) Random counting error = + 3X b) Systematic counting hardware error = + 5X c) Calibration source error = + 2X These errors are independent and statistically combine to yield a total error on the order of 6% ( 1s).

The weight of the dosimeter used in the determination of the activity per milligram (DPS/mg) was +established with a Hettler balance with a quoted accuracy of .I microgram. Given the range of dosimeter weights in Table X (8.4 to 2107 milligram) the error in Ator due to uncertainnty in the dosimeter weight alone is less than lX and is not considered a significant contributor to the experimental error in Ator.

For the purpose of interpretation of the measurement results relative to calculational results, a total measurement error of 105 is considered to be an upper bound in view of the use of the information quoted by accepted industry standards appropriate to the determination of dosimeter activity and the specific estimates for Capsule V.

E I

'l

Question 2: There is no'justification for the rejection of the dosimeter measurements and the exclusive use of dosimeters from the Charpy bars.

Res onse:

The neutron dosimetry in the surveillance Capsule V contained three neutron dosimeter capsules. They were positioned vertically at top, middle and bottom of the compartments of the surveillance Capsule V. Each dosimeter capsule contained copper, nickel, al-cobalt wire (cadmium-covered and uncovered), Np-237 and U-238 threshold dosimeters. Also, the Charpy test specimens served as iron threshold dosimeters. Only copper, al'-cobalt, and iron dosimeters provide reliable neutron fluence measurement data. All others are regarded with suspicion. The questionable measurement data given in the revised Table X of the SRI report, were not used for comparison between measured and calculated fluence data. Justifications for the rejection of the suspected dosimeters are given below:

Both uranium and neptunium dosimeters were powder specimens and were contained in metal capsules of either brass or steel. During the opening of the metal capsules, both dosimeters were contaminated by metal filings from the sawing of these metal capsules. The contamination increased the weights of uranium and neptunium dosimeters. Thus, the dosimeter's specific activity in (DPS/mg) became questionable due 'to uncertainties in the actual weights.

2; The basis for rejection of the nickel dosimeter which generates the Co-58 gamma emitter is that no expected Co-58 photo-peak was observed from the gamma ray spectroscopy counting. Instead, two 1. 17 and 1.33 HeV photo -peaks characteristic of the Co-60 isotope were observed from the counting. It is believed that a Co-59 wire rather than theNi wire was loaded into the dosimeter capsule. Therefore, intended nickel dosimeter as an integrated fluence indicator was not available.

The revised Table X reflects two corrections from the old Table X.

a."' correction for usi.ng old NBS standard source Co-60 to analyze the gamma counting data.

b. A minor correction for iron weight percentage in the Charpy test specimens.

It should be pointed out that measurement data from iron and copper dosimeters was,used for comparisons.

A question 3: A computational error analysi s should be per forme'd to evaluate the relative value of the computed to the measured fluence.

Res onse:

The error analysis provided is an estimate of the overall computational error resulting from the use of nominal input data which is subject to variability or uncertainty about its nominal value. In this analysis two reactor vessel flux calculation reference cases were established: The first reference case is a one-dimensional (1-0) transport code - ANISN case, and the second reference case is a two-dimensional (2-0) transport code -DOT4.3 case. The two reference cases utilized the nominal values of the analytical model input data.

To obtain a deviation of the nom,ina 1 value of input data, a reasonable data uncertainty range was assigned to the nominal value. An additional reactor vessel flux calculation with new input data (nominal value + data uncertainty) was performed to determine new vessel flux data.

A comparison of the new calculated and reference calculated vessel fluxes was made to obtain the. computational vessel flux data uncertainty.

F i gur e I s h ows a standar'd vessel flux ca cul at iona 1 f ow-chart 1 1 with alphabetic labels for identification of the input data uncertainty studied. Table I lists each alphabeticThe label last associated with the actual input data uncertainty.

column o f Table I indicates the computed vessel flux variation over the reference vessel flux data. An overall computational vessel flux error was found to be + 27.5/. Table 2 presents conditions for the two reference cases. It should be pointed out that the above vessel flux uncertainty analysis was analytical based on a single cycle burnup average core power data. Since methodologies for vessel flux calculations of different fuel cycles are identical, the calculated + 27.5X vessel flux error is applicable to cycles I through 9.

The relative value of the computed to measured fluence for Capsule V was determined to be .88. This was established by fluence analysis performed by FPL. Table 3 presents an'ndependent a compari son o f measured and calculated neutron (E > 1.0 HeV) fluence results for the Turkey Point Units 3/4 since 1975 up in- to present., The measurement data gi ven in Table 3 are from the vessel and ex-vessel neutron dosimetry program at FPL and the calculational data are from the FPL diffusion (PDg-7) + transport (DOT4.3) computer codes package. Table 3 indicates a consistent underprediction of the measurements. The uncertainties and the

0 g I

consistent calculational bias identified above provide a sound basis for the conclusions drawn in Response 4 regarding the adequacy of the Capsule V measurement data.

l}uestion 4: Justify the use of the PI approximation.

An independent Capsule V neutron fluence analysis using the industry accepted P3 approximation was performed by FPL for the purpose of evaluating the measurement results reported by SRI, which relied on a lower order PI cross-section approximation.

The cumulative fluence results reported by SRI were derived from several factors of which only the radial flux gradient correction factor inside Capsule V and the effective spectrum-averaged dosimeter reaction cross-section depend upon the PI approximation. The other factors are either dependent upon direct measurement or industry accepted constants.

Figure 2 shows the neutron flux distributions resulting from the PI and P3 cross-section approximation analyses. The. effect of the higher order cross-section approximation be'comes evident in the capsule region. A 1,2X flux increase is observed with respect to the PI cross-section approximation case. However, the radial flux gradients inside the capsule region are almost the that same for both the PI and P3 cases. The reason for this is the surveillance capsule is sufficiently far from the source of neutrons that the bias between the PI and P3 approximation for all practical purposes is constant in the Capsule V region.

As for the determination of the effective cycle-specific spectrum-weighted average dosimeter reaction cross section, FPL established that this value did not significantly vary over each of the nine operating cycles and was very similar to the value established by SRI.

Table 4 lists the FPL calculated 9-cycle average spectrum-weighted reaction cross-sections for iron and copper dosimeters.

Also provided in Table 4 are the SRI reported cross-sections for comparison. At most, a 3.0$ cross-section difference is noticed between the PI and P3 cross-section approximation cases.

0 0 P

The reason that the calculation of this factor is not strongly influenced by the order of the scattering approximation used it is a spectrum weighted average defined as: is'hat 0

f o.(~)4 i~)dc Pre) dp E~

in which biases due to the order of the scattering approximation in the calculated energy dependent flux /(E) tend to cancel.

The reasons provided above, as substantiated by the FPL performed P3 calculation, justify the use of the PI approximation in the SRI reported measurement results.

The calculated fluence at the Capsule V location for each of the nine operating cycles is presented in Table 5.

As 'indicated in Table 3 of the preceding response, Capsule V actually received the cumulative neutron {E > 1.0 HeV) fluence of 1.25E+19 -(neutrons/cm2) for plant operation of cycle I through cycle 9. The corresponding calculated Capsule V neutron fluence is 1.104E+19 (neutrons/cm2) for plant operation of cycle I through cycle 9. The ratio of calculated to measured neutron fluences is 0.88 which shows a 12K underprediction of measurement.

The uncertainty in the calculated neutron fluences due to uncertainties in the analytical input data and models is estimated to be + 27.5X as gi ven in Table I of the preceding response. The uncertainty in measured neutron fluence at Capsule V is estimated to be + lOX and is shown in the Response to guestion l. If a 12/ of underprediction of measurement is used in the calculated neutron fluence data, the variation of measured neutron fluence at Capsule V is between +2K and 22K which is well enveloped by the variation of the calculated neutron fluences from the range of + 27.5%%d.

This provides additional evidence that the reported measurement.

results {with the exceptions noted in Response 2) are accurate.

tg N

t

g i E 1 i h chemically pure.

Res onse:

In Capsule V, only the iron dosimeter s made from the irradiated Charpy test specimens were not chemically pure. Chemistry of the Charpy test specimens are typically 97% iron and 0.68K nickel, with the remainder being other alloy materials. Oue to neutron activation of the Charpy test specimens during its residence inside the reactor vessel, two prominent gamma emitters were produced. They are the Hn-54, 'product of the Fe54(N,P)Hn54 reaction, and the Co-58, product of the Ni58(N,P)Co58 reaction.

The Hn-54 gamma emitter has a half-life of 321.5 days and emits a 835 KeV photon. The Co-58 gamma emitter has a half-life of 71 days and emits a 811 KeV photon. However, only the iron activation product Hn-54 gamma counting is of interest for dosimetry purposes. Since no chemical separation of nickel from iron was done, the nickel activation product Co-58 gamma can potentially interfere with the Hn-54 gamma counting due to the proximity of the Co-58 photo peak.

In the gamma counting procedure, an IT-'400 multichannel analyzer and a conventional Ge(Li) coaxial detector were used to measure the gamma activity of the Hn-54. Before measurements, the counting system was calibrated by using standard test gamma sources obtained from the National Bureau of Stan,dards.'

typical 0.5 KeV per channel of IT-5400 multichannel analyzer and 0.25 percent of energy resolution (at 835 KeV) of the counting system were used for measurements. There were 48 channels of separation between the Hn-54 and the Co-58 photo peaks (48 channels = (835-811 KeV)/0.5 KeV/channel). 0.25 percent energy resolution provided the estimated 2-KeV full width inferred at half maximum (FWHH) spreadings for both gammas. This in turn that there were 44 channels (44x0.5 KeV) of separation time between the Hn-54 and Co-58 photon distributions. At the of measurement, at least 215 days had expired since Capsule V was removed from the reactor. No significant interference of the Hn-54 photo peak due to the proximity of the Co 58 photo peak is expected for the following reasons:

a. the lower production rate of the Co-58 gamma from nickel impurity at the end of neutron irradiation

b. at the time of counting, the Co-58 emitter had significantly decayed as a result of the shorter half-life (71 days) of Co-58 as compared with the half-life (312.5 days) of Hn-54

c. high energy resolution of the counting system

I j An analytical evaluation was used to estimate the magnitudes of the total gamma counting of the Hn-54 and the Co-58 in order to substantiate, the above statement. Estimation was made based on the production rate of the Hn-54 and Co-58. Results demonstrate that at the time the measurement was performed, the magnitude of the Hn-54 gamma counting is higher than the Co-58 gamma cou'nting by a factor of 36.

Based on the above results from measurement and analytical estimation, there was no significant distortion of the Hn-54 photon distribution due to the proximity of the Co-58 gamma source.

C 1

l,1 I~

j

TABLE 1 Estimated Computational Vessel Flux Uncertaint Estimated Vessel Alphabetic Flux Uncertainty Label In ut Data Uncertaint (res ect to reference)

Downcomer RC ND's variation due + 0.5%

to RC temperature changes + 2'F RC Tave ND variation due to + 0.4% -0.5%

Tave changes + 4'F B Capsule V positional variation -4.0% to +5.0%

by + 1.0 cm Core baffle dimension variation + 1.0%

+ 1/16" Core barrel dimension variation + 1.0%

+ 1/16" Thermal shield dimension variation + 1.0%

+ 1/16" Downcomer region (near vessel) +21% to -24.8%

dimension variation by 0.5" D Peripheral fuel assembly power +3.0% to -3.0%

variation by 1.05 and 0.95 Homogenized core data variation -6.0% to 7.0%.

by 1.05 and 0.95 1-D ANISN mesh size variation

a. In downcomer region 1.07 cm/mesh to 0.539 cm/mesh +0.4%

1.07 cm/mesh to 2.14 cm/mesh -0.3%

b. In thermal shield

'0.969 cm/mesh to 0.678 cm/mesh +0.1%

0.969 cm/mesh to 2.26 cm/mesh -1.2%

2-D SORREL (DOT) Mesh Size Variation 0.94'/mesh to 0.82 /mesh + 0.2%

2-D cycle burnup power variation

a. BOC Pin Power +7.1%

b. EOC Pin Power ~ 0.0%

Azimuthal power Tech Spec 1.02 tilt variation factor applied by +1.8%

to core flat fuels 27.5%

0

'1

TABLE 2 Reference Case Conditions Reference Case Code Condition ANISN Standard practice P3 S12 Neutron Source- homogeneous core RC Tave = 575.4'F Boron Concentration 715 PPM Downcomer RC Temperature 546.2'F Reactor Internals Dimensions-Nominal Values Reactor Vessel Dimension-Nominal Value Turkey Point Unit 3 Cycle 1 data 27-group cross-sections (> O.l MeV)

Mixed U-235 and Pu239 fission spectra Peripheral assembly radial power gradient E

DOT4.3 Standard practice P3S8 R-Theta Model (One eighth core model)

All material surveillance capsules, T,S, and V in calculational mode'ls.

Azimuthal and radial power gradient's-from FPL PDQ-7 pin power files The remaining conditions are the same as reference case l.

TABLE . 3 Comparison of Neutron Dosimetry Data (Prom 1975 throu h l987)

FPL Dosimeter Measured Pluence Ratio of Unit Cycle(s) In-Vessel Ex-Vessel Value B Calculated Measured 1

3 1 Capsule T 5.68(18) Westinghouse 0.95 Capsule T 6.05(18) SRI 0.92 1-9 Capsule V 1.25(19) SRI 0.88 4

1'2 Measured Westinghouse 0.86 Data Data Sources:

l. WCAP-8631 Report

2. SRI Project 502-4221 Report

3. SRI Project 006-8575 Report

4. WCAP-11138 Report

I TABLE 4 Spectrum-Weighted Average Reaction Cross-Sections for Iron and Co r Dosimeters at Ca sule V (barns)

Threshold Reaction SRI (P ) ~PP~

Fe54(N, P)Mn54 0.0786 0.0806 0.98 Cu63(N,H)Co60 0.000885 0.0009088 0.97

TABLE 5 Cycle-Specific Capsule V Neutron (> 1.0 MeV)

Fluence for Turke Point Unit 3 Charpy Test Specimen Cumulative Cycle Length Region Average Flux Cycle Fluence Fluence Cycle (sec) (Neutrons/cm4-sec) (Neutrons/cm ) (Neutrons/cm )<

I 3.609(7)* 4.596(10) 1.659(18) 1.659(18) 4.574(10) 1.121(18) 2.780(18) 2.451(7)'.418(7) 4.908(10) 1.187(18) 3.967(18) 2.462(7) 4.356(10) 1.072(18) 5.039(18) 2.453(7) 4.498(10) 1.103(18) 6.142(18) 1.587'(7) 4.325(10) 6.863(17) 6.828(18) 2.902(7) 5.397(10) 1.566(18) 8.394(18) 4.302(7) 3.612(10). 1.554(18) 9.948(18) 3.263(7)** 3.358'(10) 1.096(18) 1.104(19)

Read 3.609(7) as 3.609xl0

• • Actual cycle length. rather than planned cycle length has been used in the fluence calculation.

FIGURE 1 CALCULATIONAL FLOW-CHART l

BUGLE-80 47-Neutron-Group + 20-Gamma-Group Library Cross-Section Library PPL FPL Revised 27-Neutron-Group Cross-BUGLE-27 Section Library (E > 0.1 MeV, No Gamma)

Library Plant Plant B Dimensions Materials GIP Data Data Plant Dimensions ANISN DOT4.3 SORREL Data H

See Table 1 for the circled alphabetic labels notation PDQ-7 OUTPUT OUTPUT (1-D Flux) (2-D Flux)

0 l

>1

I D FLUX(E>1.0 MEV) AT'CAPSULE T CQMPARI5QN QF Pl VS PO fLVX

.P3 Case Pl Case Charpy Specimen Region 189 eeIAL DISTANCE (CM)

I:I P0 FLVX + P1 FLUX

,FIGURE 2 Comparison of Pl and P3 Pluxes

REVXSHDTMLK X February 24, 1987 SATUBAXED hCTXVITIKS hND DERIVED FLUBNCE RhXES l'Ok ChPSUM V React5on and Radial - Oosiaeter hlOR ASAr "Shl{s) fluce~ Flupnce{b Axial locat1on locat5on Meight {dPS~) (dPSiatoa) hd~usted a te {~ Rate O'S atm n-ca -s 2 1 ll -R. s CR lfc{n,p) Nn S-58 TOP 191.47 lK8.1 1.96E3 4.2lE-15 3.89E-15 5.36E10 4.95E10 S-52 MIDDLE 191.47 2107.3 1.92E3 4.13E-15 3.&ZE-15 5.25E10 4. &6E10 M-2 80TTGN 191A7 19&8 6 1.81E3 3.NE-15 3.59K-15 -

4.95E10 4.57E10 R-48 TOP 192.47 1955.4 1.63E3 3.50E-15 4.20E-15 4.45E10 5.34E10 R-4Z HIDDLE 192.47 1359.% 1.67F3 3.59E-15 4.3]E-15 4.57E10 5.4&EIO II-2 SOTTO@ 192.47 1&27.1 1.52E3 3.27E-15 3.92E-15 4.16E10 4.99E10

@cu(l,a) ~co Cu TOP 192.47 156.496 1.22EZ 3. 53E-17 4.23E-17 3. 99E]0 4.7&f10 Co BOTTOM 192.47 104.709 1.23E2 3.56E-12 4.26E-17 4.02EIO 4.&1E10 5&F5(n.p)5&Co Ni HIODLE 192.47 8.937 Ho cobalt-58 peak-slees bo CO-60 peaks 237wp(g f) l37 ADDLE 191 69 19. 9 2.51E3 9.79E-14 9.79E-lh 3.92EIO 3.92f10 U(n,f)+~cs 59C il MIDDLE (n,.)6OC

" 191.69 19.3. 1.20K-14 1.20E-14 2.22E-I2 3.22EIO 3.22E10 '.94E2 Co TOP 191.47 8.474 1.00E7 1.85E-12 3.03E10 3.03E1Q Co-Cd TOP 191.47 9. 161 4.93E6 9.13K-13 Co MIDDLE 191.47 9.218 9.06K6 1.68E-12 2.02E-12 Co-Cd NIIILK . 191.47 10.689 No. of counts indicates that count tice is 2000 instead of 70.000 Co BORON 191.47 8.616 9.70E6 1.79E<<12 2.15E-12 2.89E10

- C Cd&OTTe 191.47 9.434 4.87E6 9.01f-l3 1.0&K-12 .

~

hSAT GATOR (I) At dasiaetnr l~otion. th> A8 inst ial ceoterliue 'of s eciaeoa.

~ '.k