IR 05000369/1986029

Insp Repts 50-369/86-29 & 50-370/86-29 on 860929-1003.No Violations or Deviations Noted.Major Areas Inspected:Startup Physics Testing & Closeout of Previous Enforcement Matters

ML20211J824
Person / Time
Site:	Mcguire, McGuire
Issue date:	10/29/1986
From:	Jape F, Long A NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
To:
Shared Package
ML20211J818	List: IR 05000369/1986029 ML20211J811
References
50-369-86-29, 50-370-86-29, TAC-62972, TAC-62981, TAC-62982, TAC-65985, TAC-65986, NUDOCS 8611110324
Download: ML20211J824 (12)

Text

~-

-,

,

an ateo UNITE] STATES

'~

f 'o NUCLEAR REGULATORY COMMISSION

[ 3 REGION li g j 101 MARIETTA STREET, *, ATL ANTA, GEORGI A 30323

' 8

%, * . . . * p .~

-., 4

-

Report Nos.: 50-369/86-29 arid 50-370/86-29 Licensee: DukeiP'ower Company ~

422 South Church Street- 1

- Charldtte, NC'; 2S242 ' ,

^

Docket Nos.: '50-369 and 50-370 -

-

- License Nos.: NPF-9 and NPF-17

. ' ,

Facility Name: McGuire 1 and '

_ Sspection Can'du ted: September 29 - Octotfer 3,1986 Inspector: _

_kM_ - /0 29!80 A. R. Losj Date Signed Accompanying Personnel: F. Jape Approved by: 4 .

7f

/6/,igne/$

F. Jape, Chief, Test Programs Sect Date S d

-

i Engineering Branch Division of Reactor Safety SUM!iARY Scope: This routine, announced inspection involved the review of startup physics testing, and closecut of previous enforcement matter Results: No violations or deviations were identifie PDR ADOCK0500g4 G

s

__m

e

. .

,

.

REPORT DETAILS Persons Contacted Licensee Employees

• N. Atherton,' Compliance

+ E. Bortz, Nuclear Design

1. R. H. Clark, Senior Engineer, Nuclear Design L. F. Firebaugh, Assistant Operations Engineer

+#L. H. Flores, Supervisory Design' Engineer, Nuclear Design

+#S. A. Gewehr, Licensing Engineering Assistant

1. G. D. Gilbert, Operating Engineer R. Gill, Licensing

• B. H. Hamilton, Superintendant of Technical Services

• M, S. Kitlan, Jr. , Reactor Engineer

+L. J. Kunka, Nuclear Production Engineer

• D. S. Marquis, Performance -Engineer T. L. McConnell, Station Manager

~*E. O. McCraw, Compliance Engineer T. W. Mira11a, Performance

+ J. Pierson, Nuclear Production Engineer

• D J. Rains, Superintendant of Maintenance

'#J. H. Randles, Design Engineer I, Nuclear Design

• R. B. Travis, Superintendant of Operations

+#R. Van Namen, Associate Engineer, Nuclear Design Nuclear Regulatory Commission Employees

+ Chatterton, Office of Nuclear Reactor Regulation

+ Dunenfeld, Office of Nuclear Reactor Regulation

+D. Hood, Project Manager, Of fice of Nuc1 car Reactor Regulation

• S. Guenther, Resident Inspector

+W. T. Orders, Senior Resident Inspector

1. Attended Meeting in Corporate Office on September 29,19S6

+ Participated in Conference Call on October 2, 1986

• Attended Exit Meeting on October 3, 1986

. Exit The inspection scope and findings were discussed in a conference c(11 on October 2, 1986 and were summarized in the Exit Interview on Octo')er 3, 1936, with those persons indicated in paragraph 1 above. The inspector described the areas inspected and discussed in detail the inspection finding No dissenting comments were received from the Itcense ,

.

l

-

-

.

By -telephone call on October 9,1986, Duke Power was requested to submit a-letter detailing the sequence of events and justifying use of Duke Power rod swap predictions for McGuire 1, Cycle 4. Duke responded to this request on October 17, 1986. Information from this response has been incorporated into this inspection repor In addition to the inspection findings, the licensee's plans for providing ~

more information in the files of the open item tracking system were briefly discussed at the exit meeting. Information on the completion of corrective action commitments, such as procedure numbers, and dates of personnel briefings, should make the closecut or open items more efficient for both the NRC and the license Pr'oprietary material was reviewed during the inspection, but has not been included in this repor . Licensee Action on Previous Inspection Findings (92702)

The inspector reviewed the corrective actions for the following violations and they will be closed: (Closed) VIO 50-369/82-09-01: Failure to Obtain Prior NRC Approval to Change an Acceptance Criterion in an Essential Test, Dynamic Rod Dro This violation involved an unapproved nonconservative change to the test acceptance criterion on the number of channels required to trip on negative rate in response to a double rod drop. This violated the license condition requiring NRC approval of major modifications of the initial test program as set forth in Chapter 14 of the FSA The test acceptance criterion in the FSAR required demonstration that the reactor tripped on at least three of the four nuclear instrumenta-tion channels. In the original test, the reactor tripped on two of four negativ'e rate channels before a trip response from a third channel occurre In theic April 20, 1982 response to the violation, the licensee made a commitment to rerun the test at the next available outag Their letter of November S, 1982, identified the compliance date as early 198 The dynamic rod drop test was repeated success-fully on May 29, 1983, by disabling the negative rate trip function of one of the power range channel Trip responses were observed on all four power range channels, followed by a reactor tri (Closed) VIO 369/85-20-01 and 370/85-21-02: Failure to Adequately Control and Review Data Used in Safety-Related Procedure This violation involved failure to implement Nuclear Safety Evaluation Checklists (NSECs) for changes to Xenon values in the process compute Xenon worths from the process computer are used in safety related calculations. The licensee committed in their July 29, 1985 response to the violation to implement a procedure requiring (1) the preparation of

.

3-

'

a NSEC and (2) a Nuclear Design Unit review, prior to using new Xenon and/or Samarium worths from the process compute The inspector reviewed -procedure PT/0/A/4600/71 for changing the process computer Xenon and/or Samarium coefficients. The Reactor Group

'at the site prepares -the required- data for the update and performs a Nuclear Safety Evaluation Checklist. The changes are reviewed and -

implemented by the' Process Computer Unit in the General Office. The computer update'is then rechecked by the Reactor Group. Each update is-well documented in the procedure, and implementation of PT/0/A/4600/71 has been performed for the current cycles of both McGuire unit (Closed) VIO 369/85-20-02 and VIO 370/85-21-03: Failure to Maintain -

Adequate Records of Surveillance Technical Specifications require . retaining records of surveillance activities. This violation involved - maintaining calculated shutdown margin values, but failing to maintain the data used to perform the calculation Procedur. , PT/1/A/4600/03B and PT/2/A/4600/038 for daily surveillances, and procedures PT/1/A/4600/03A and PT/2/A/4600/03A for- semi-daily surveillances,-bave been revised to require the attachment of Encicsure 5.3 of OP/0/A/6100/06 when shutdown margin calculations are performe The inspector observed that these requirements are being followe . (Closed) VIO'370/85-21-01: Failure to Follow Procedures for ECP and Unit Fast Recovery The first part of this violation involved exceeding the startup rate limit cf 1.0 decade per minute specified in plant procedure Corrective action commitments by the licensee included moof fying the startup procedure to limit startup rate to a more restrictive decade per min ~ut The inspector reviewed procedure OP/2/A/6100/01, Controlling Procedure for. Unit Startup, and procedure OP/2/A/6100/05, Unit Fast Recovery, and verified that this change is in plac The second part of this violation involved failing to either follow or change a dilution requirement in the startup procedure. When the estimated critical rod position in the procedure was felt to be in error due to problems with the Xenon calculation,. licensee personnel decided not to make a permanent procedural change and proceded with the test without modifying the procedure. In cases like this the licensee is now following a practice. of implementing procedure revisions which are specified as applying to that startup onl . Unresolved items .

Unresolved items were not identified during this inspectio .

-

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

-

.

4 Unit 2 Cycle'3 Rod Swap Evaluation (72700)

The . licensee notified the NRC by letter (Hal B. Tucker to Harold R. Denton,

~

September 15,1986) that one of the model review criteria was not met durin the rod worth measurements on Unit 2 Cycle 3. Model review criteria have no defined safety significance, as opposed to test acceptance criteria, which are based' on verifying safety analysis input assumptions. The NRC Safety Evaluation Report (SER) on the topical report " Rod Bank Worth Measurement Utilizing Bank Exchange" (WCAP-9863-A) requires that the NRC be notified if a review or acceptance criterion has not. been me Rod worths for Unit 2 Cycle 3 were measured at hot zero power using the rod exchange techniqu The rod exchange technique measures the reactivity worth of a control or shutdown bank of control rods (referred to as the test bank) relative to the worth of a reference bank by swapping the banks in the core. Calculated factors are then applied to adjust the measured worth of-the test bank to its worth with no other banks inserted in the core, so that the measurement can be c~ompared to the predicted rod worth on a consistent basis. The reactivity - worth of the reference bank is ' measured by the

-

conventional boron dilution technique. The reference bank is inserted in a series of steps to offset the constant rate of reactivity addition from the boron dilution. The reactivity worths of these steps are summed to obtain the total worth of the reference ban During the Unit 2 Cycle 3 startup, the worth of the reference bank, Shutdown Bank C, was measured by the licensee to be 14.5% below the predicted valu This deviation exceeded the model review criterion of 10%, but did not exceed the acceptance criterion of 15%.

Westinghouse physics data predictions were used for the Unit 2 startup. In accordance with WCAP-9863 A, the test results were forwarded to Westinghouse

, for evaluation. The Westinghouse review (Letter, R. T. Meyer to T. F. Wyke,

"McGuire Nuclear Station Unit 2 Cycle 3 Measurement Discrepancies,"

August 25, 1986) revealed no deficiency in either the calculational models used for the prediction or in the reference bank worth measuremen Westinghouse also confirmed that the discrepancy had no effect on the input assumptions for the safety analysis. The Westinghouse evaluation was edited and included with the September 15 letter to the NR Westinghouse concurred with the editorial changes which were mad The primary purpose of this inspection was to confirm that the reference bank measurement exceeding the model review criterion was not an indication that the plant does not have an inadequate shutdown margi The inspector attended a meeting on rod exchange methodology at the corporate of fice on September 29, 1986 (See paragraph 7) and reviewed the following documents: Letter, Hal B. fucker to Harold R. Denton, Attention:

B. J. Youngblood, September 15, 1986. (Reporting of Rod Worth Discrep-ancy)

.

m e

r

-

.

5 Letter, T. F. Wyke to R. L. Gill, "McGuire 2 Cycle 3 Rod Swap Evalua-tion", - September 9, 1986. (Duke Internal . Transmittal of Westinghouse Evaluation of Rod Worth Discrepancy) Topical Report, Nuclear Physics Methodology for Reload ~ Design, DPC-NF-2010A, June 198 Letter, Hal B. Tucker to Harold R. Denton, Attention: E. G. Adamson,

" Rod Swap-Methodology," August 28, 198 Topical Report, " Rod Bank Worth Measurement Utilizing Bank Exchange",

WCAP-9863-A, May 198 (Westinghouse Proprietary) NRC Safety Evaluation Report on WCAP-9863- McGuire Procedure PT/0/A/4150/11, Control Rod Worth Measurement (Control. Bank C), Performed June 19, 198 McGuire Procedure PT/0/A/4150/11, Control Rod Worth Measurement (Control Bank D), Performed June 26, 198 McGuire Procedure PT/0/A/4150/11A, Control Rod Worth Measurement: Rod Swap, Performed June 19-26, 198 McGuire Unit 2 Cycle 3 Startup Report, Septembei 19, 198 The inspector concluded that the 14.5% dif ference between the measured and predicted worth of the reference bank is not a safety concern because all acceptance criteria for rod worth measurements were me The total predicted worth of all the shutdown and control banks, which is used in determining plant shutdown margin, differed from the measured total worth by-8.8%, which is less than the 10% acceptance criterio There is also substantial additional margin in the input to the safet analysis calculations beyond the values of the acceptance criteria limit This is why a thirty day period is allowed for Westinghouse to evaluate failures to meet a test acceptance criterion, and plant operation is allowed during that time, a

Although the reactivity worths of the other banks are measured relative to the reference bank, it is important to realize that an observed difference between measured and predicted reference bank worths does not necessarily introduce additional error into the measurements of the other test bank Difference.s between measurements and predictions can be due to errors in either the measurement or the prediction. Calculational errors in reference bank predictions can be considered to be independent of the errors between measured and predicted test bank worths. For the reference bank worth to be mispredicted does not necessarily mean that the worth of the other test *

banks is also mispredicte . . . . . - -

,

.>

} .

,

Furthermore, if measurement e rors were to have caused-the apparent worth of the reference bank .to be less than its' true worth, the apparent worth of e

each test bank ~ measured against it.would also be less than its true worth,

' For McGuire 2 Cycle 3, the measured worth of the reference ~ bank was 14.5?s less than the prediction. Asl discussed in more detail later in the report, Westinghouse reviewed the . test results and determined that the most likely sources. of measurement: error 'would have tended to make the reference bank measure less than its ' real wort It then logically follows that if

,

measurement errors caused the Cycle 3 discrepancy, then the actual total worth of all the test banks would have been higher than was measured, and

even closer to the predictio _

.Because the reference - bank measurement was close to the acceptance -

criterion, the inspector independently reanalyzed the reference bank reactivity trace. The inspector's analysis yielded a reference bank worth 4

,

pcm closer to the predicted worth, which is' an insignificant differenc The licensee analyzed these data conservatively (in a way as to decrease rod worth and increase the difference from the prediction). It is not likely that reasonable differences ,n engineering judgement could have led to an- ;

'

analysis of: the measurement which would have violated the 15?; test accept-

ance criterion.

The difference between the measared and predicted reference bank worths appears to be due in part to normal measurement uncertainty, but primarily to using standard rather than more recently approved delayed neutron parameters, and to calculational uncertaint ~

Using refined methods to calculate the delayed neutron parameters for reactivity calculations had been intended for Cycle 3, but due to a misunder-

.

standing Westinghouse provided values determined by standard method Westinghouse determined that use .of the correct set of Betas would have decreased the discrepancy in the reference bank measurement from 14.5?; to about 12?o. (This revised methodology for generation of kinetics parameters

,

was approved by the NRC in DPC-NF-2010A.)

! ~

i The inspector reviewed the agreement between Westinghouse predictions and measured red worths for previous startups of McGuire, and concluded that the agreement has historically been loose. Although the 14.5's error at McGuire 2 was larger than the differences in reference bank worths for the

[

other cycles, differences of this magnitude frequently occurred for other banks of similar siz Calculational uncertainty in the Westinghouse predictions as a major source of the discrepancy is supported by the fact that independent predictions by Duke of the reference bank worth differed from the measured worth by only-9. 44?; . This difference would have been even smaller if the updated delayed i

!

neutron parameters had been used in the reactivity calculation.

i l

i

.

S D

e-

_ ____ ____ -__ _ _ .

.

7'

The Westinghouse codes m'ispredicted which control rod would_have the highest

- worth -and should be used as the reference bank. The Duke calculations correctly predicted Shutdown Bank B as the reference bank. The worth of Shutdown Bank B was predicted by Duke calculations within about 7% of the measurement. Westinghouse predicted the worth of the bank within about 8%

.of the measurement. Either case is within the model verification criterion of 10% and would not have needed to be reported to the NR ,

An additional factor supporting the validity of the measured reference bank t

worth is the comparison of measured and predicted boron worth. Boron worth was measured by the licensee in procedure PT/0/A/4150/11A by dividing the measured reactivity worth of the ' reference bank by the change in bo'ron concentration during the test. This measured boron worth agreed with the prediction to within less than 1 If significant errors had occurred in

,

th'e reactivity worth measurement of the reference bank, it would have been

-

expected to ca'se u a corresponding discrepancy between measured and predicted

~ boron worth, which was not the cas The inspector reviewed the data in procedure PT/0/A/4150/11 -for the reference bank-measuremen The reactivity trace was found to 'obtain no noticeable anomalie The resolution of the data was good, and the noise level in the signal was low. Normal reactivity "undershoots" were observed

-

. in the trace after some but not all of the rod movements. The Westinghouse review of the test data presented the possibility that the absence of the expected undershoots could have contributed to a low measurement of the reference bank worth. The inspector believes spatial effects were adequ-ately'. accounted for in the data and in the analysis. The inspector concur with Westinghouse that inadequate accounting for reactivity undershoot would have resulted in measured rod worth being too low, and would therefore be i the conservative direction for this case.'

,

The possibility that the plant entered the range of sensible heat, which would have induced temperature feedback reactivity effects and invalidated the reference bank measurement, was also reviewed by the inspecto The physics testing flux level range for Cycle 3 was a decade higher than usual

.because .of' excessive noise in the detector signa A review by the inspector of the testing methods used indicates that the sensible heat range would -not have been entered. In addition, no' evidence of temperature feedback effects were observed. by the inspecto'r in the traces of the test

,

data. Moderator ' temperature remained steady throughout the test, and the slope of the reactivity traces remained constant. The _ inspector concurs with the statement in the Westinghouse evaluation of the test results that measurements near nuclear heating flux ~ levels tend to reduce the indicated reactivity, which would be conservative in this_ case.

! Using t'co high of a baron dilution rate is a possible source of measurement error in rod worth measurements. From the reactivity trace for the Unit 2 Cycle 3 reference bank measurement, the boron dil.ution rate was determined to be approximately 700 pcm/hr. This is faster than the maximum of 500

. . , -- . _ _ , - _ ,, _ _ , . -_, - .- _,

_ __

-

,..

,

pcm/hr. recommended by . Wes'ti nghous However, the inspector found no evidence that the . higher dilution rate decreased the accuracy of the measurement. _Even at.this dilution rate the resolution ~of the data was good and allowed adequate accounting for spatial effects in the data redu'ctio The maximum dilution rate was expressed in the test procedure as a' guideline rather than a firm restriction, 'so no failure to follow procedures . is involved. Although the licensee exceeded the procedural recommendation of a dilution rate below 500 pcm/hr, they adhered' to .the procedural limit that the slope of the reactivity trace not exceed- 45 degrees. -The licensee has

~

committed to maintain dilution rates below 500 pcm/hr. in future startups, and did so on the Unit 1 Cycle 4 startu In summary, inspection of the rod worth test showed the test to be valid and the results adequate with respect to safet . Unit 1 Cycle 4 Rod Swap Methodology (72700)

A concern was expressed .by the Office of Nuclear Reactor Regulation (NRR)

that the licensee may not have obtained the proper NRC approval for the test

methods which were used in the Unit 1 Cycle 4 rod worth measurement Rod worths were. measured for Unit 1 Cycle 4 using the rod swap techniqu This technique had previously been used on several other cycles of McGuire 1 and.2. The . change for McGuire 1, Cycle 4 was that Duke Power, rather than Westinghouse,Lgenerated the predictions which were utilized for the rod swap-procedure. The measurement technique and acceptance criteria were essenti-ally unchange Rod swap . test procedures utilize calculations for two different purpose The first type. of' calculated values used in rod swap testing includes the predicted rod bank worths to which the measured values are compared to demonstrate that acceptance criteria have been me The second ' type of

' calculated values used in rod swap includes the. factors used to adjust measured rod worths for thel partial. insertion of. the reference bank. This yields the inserted worth of'each bank with all other banks withdrawn from-the core, so that the measurements can be compared to predictions on a consistent basi . Duke had'been given.NRC approval' to perform the first type of calculations -

the rod worth predictions to which the measurements are compared to verif ' acceptance criteria. Topical report DPC-NF-2010A, Nuclear Physics Methodo-

. logy- for Reload Design, was approved by the NRC and an SER was issued on March 13, 1985. This report approved Duke Power computer codes and methods to perform startup physics test predictions, shutdown margins, and other reload calculations. Because rod swap data was not available for~ the Duke

. units when- the . topical was written, the rod worth predictions were bench-marked in - the~ report against boron dilution measurements rather than measurements made with the rod swap techniqu However, the data demon-strated the ability of the code to predict control rod worth .

I

.

,,~ -

,

,

.

.

,

. Westinghouse. topical report WCAP-9863-A - documents rod swap methodology, including the- second type of calculations previously mentioned - the adjustment factors which are applied to the measured bank worth This topical was approved by the NRC. Although it was not specifically stated in the. Safety' Evaluation . Report, NRR considered rod swap methodology to be approved only if Westinghouse calculations are used and all Westinghouse recommendations in the topical are strictly followe In neither of the topical reports was Duke's calculation of the measurement adjustment factors specifically approved by the NR Duke's plans for performing- the calculations for the startup testing of i4cGuire 1 Cycle 4 included using Duke calculations for the adjustment factors in the rod swap measurements. They decided that a formal revision of the previous topical was not required. Duke prepared a letter to inform the NRC that measurement adjustment factors calculated by Duke would be used in future rod swap measurements, and to provide data' benchmarking the use of Duke factors in the test. This letter, from Hal Baker to. Harold Denton, and dated August 15, 1985, was not received by the NRC as had been intende Since Duke felt that no review was necessary by the NRC, the lack of a response was taken to indicate- agreement instead of non-receipt of the repor During-the U' nit 1 Cycle 4 startup testing, the measured reference bank worth

. and. the total worth of all banks both agreed with the Duke predictions to within 1%. This level of . agreement is considered excellen The' Duke rod worth calculations were documented l and reviewed as safety-related in calculational file MCC-1553.05-00-0006. They were entered into the ; rod swap -procedure in accordance with the administrative procedures normally used to incorporate cycle-specific physics data into startup procedures, and with the same review ~

Although the NRC had not given specific approval for Duke to calculate the measurement adjustment ' factors, the codes which were used had been formally .

approved for rod worth calculations. A preliminary review of the benchmark

. data compi. led by Duke indicates that for the McGuire units, - the Duke calculations give results as good as or better than the Westinghouse calculations. The calculated adjustment factors are related to rod worth

,

ratios, and tend to be easier to accurately predict than the actual bank

worths.

)= . Continuing refinements to the calculational methods described in approved-topical reports is common throughout the industr A preliminary review of the benchmark data compiled by Duke indicates that for the McGuire Units the Duke calculations give results as good as or better than the Westinghouse calculation 'It has not been standard pra.ctice, among either vendors or licensees, to report minor changes to calculational methods to the NRC.

('

L .

_

-

.

' Region II believes. that Duke's use of their own calculations for the adjustment factors in the rod swap test is acceptable with respect to plant safety. -The use of Duke predictions does not appear . to constitute. a significant change to procedures, an unreviewed safety question, or a Technical Specification chang Duke Power Company's resubmittal of the report benchmaking use of their calculations for rod swap has been received by NRR and is currently being reviewe . Meeting on Rod Swap Methodology (72700)

The inspector attended a meeting on Rod Swap Methodology in the Duke Power Company general office on September 29, 1986. The agenda of the meeting included the following: A History of Duke Reload Design I Explanation of Joint Westinghouse / Duke Design Responsibilities III. A Review of Information in Duke Topical DPC-NF-2010 IV. 'Recent Rod Swap Results

<

- Future Plans for Duke Reload Design

,

VI. Action Items The current reload design responsibilities for McGuire Nuclear Station are shared between Westinghouse and Duke Power Compan Duke Power has responsibility for the preliminary and final fuel cycle designs and loading pattern. Westinghouse generates th'e safety-related physics parameters which

.go into the RSAC and RSE and re performs any necessary safety analyses for a reload. Westinghouse has responsibility for the RAOC analyses to determine

'

operating and RPS limits on Power and Axial Flux Difference. Westinghouse also provides theoretical factors, a Nuclear Design Report (NDR), and a Peaking Factor Limit Report. Duke Power supplements the information in the NDR with internally generated physics data documented in the Startup and Operational Data Report. Duke has the responsibility for performing all physics testing and providing the required startup report to the NR The reload documentation provided to Duke by Westinghouse includes a set of-

' generic input used in the safety analysi When Duke performs physics calculations for startup testing, their calculated parameters are compared

<

to the generic safety analysis input. When the Duke calculations are verified during startup testing, the validity of the safety analysis input is confirme Duke Power as licensee for the stations has all final licensing responsi-

bility and may obtain support from the vendo .

'* *

.

_

-- _

g i' . .

11-Duke Power currently - expects to provide internally generated rod swap

'

predictions for'both McGuire Units 1 and 2 next year and continuing into the future.

,

b

$

4 i

,

[ s

!

I i

r

'

i l-

,

.- I