Text

Insp Repts 50-369/85-20 & 50-370/85-21 on 850521-24 & 0603-07.Violation Noted:Failure to Follow Procedure for Reactivity Balance Calculations & Reactor Startup
	ML20132G668
Person / Time
Site:	Mcguire, McGuire
Issue date:	06/20/1985
From:	Burnett P, Jape F, William Orders, Patterson C; NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II)
To:	;
Shared Package
ML20132G627	List: ML20132G624; ML20132G639; ML20132G668;
References
	50-369-85-20, 50-370-85-21, NUDOCS 8508050100
	Download: ML20132G668 (10)
	v • d • e

. .

l

A Mot, UNITED STATES

p o NUCLEAR REGULATORY COMMISSION

[N 7,

'

g

3$

/,j

REGION ll

101 MARIETTA STREET, N.W.

)

't ATLANTA, GEORGI A 30323 '

. $

.....

Report Nos.: 50-369/85-20 and 50-370/85-21

Licensee: Duke Power Company

422 South Church Street

Charlotte, NC 28242

Docket Nos.: 50-369 and 50-370 License Nos: NPF-9 and NPF-17

Facility Name: McGuire 1 and 2

Inspection conducted: May 21-24, 1985 and June 3-7, 1985

Inspection at McGui e Nuclear Station near Huntersville, North Carolina.

Inspectors: 4

g/ C. A' Patterson

b%t2( G h*/PT

yf Date Signed

& d. 9~LaL, 4-2.s- 75

g P. T. Burnett Date Signed

W M %vL6-4.-

pW.T.Ordes

-

b' 10-83

Jate Signed

Approved by: (44{ NM

Frank Ja 6, Section Chief (/ / Date Signed

Engineerine, Branch

Division 01 Reactor Safety

Summary

Scope: This special inspection involved (90) inspection hours on site after

the reactor startup on May 17, 1985, when the reactor was taken critical below

the control rod insertion limits required by Technical Specification 3.1.3.6.

The predicted critical rod position was calculated to be 38 steps on bank D,

but actual critical position was 26 steps on bank C, which was below the inser-

tion limit of 47 steps on bank C. The primary source of error was attributed

by the licensee to be errors in the xenon calculation.

Results: Three violations were identified: (1) Technical Specification 6.8.1.a - Failure to follow procedure for Reactivity Balance Calculations

(Estimated Critical Rod Position OP/0/A/6190/06) and failure to follow

procedure for reactor startup (0P/2/A/6100/01 and OP/2/A/6100/05) (paragraphs

5.c and 5.d), (2) 10 CFR 50 Appendix B, Criterion V - Failure to review changes

to the Xenon Predict Program values used in estimated critical control rod

position calculations and shutdcwn margin calculations, (paragraph 5.e),

(3) Technical Specification 6.10.1.d - Failure to retain records of Surveillance

Requirement 4.1.1.1.1.e for calculation of shutdown margin (paragraph 5.f).

E8"iWR *!mL PDR

. . i

-

.

REPORT DETAILS

1. Persons Contacted

Licensee Employees

T. McConnell, Station Manager

G. Cage, Superintendent of Operations

B. Hamilton, Superintendent of Technical Services

G. Gilbert, Operations

R. Clark, Duke Nuclear Engineering

W. Reeside, Operations

M. Pacetti, McGuire Safety Review Group

W. McDowell, Duke Licensing

W. Jefferies, Duke Nuclear Engineering

R. Phillips, Operations

D. Marquis, Performance Engineer

S. Brown, Reactor Engineering

M. Kitlan, Reactor Engineering

E. McCraw, License and Compliance Engineer

Other licensee employees contacted included engineers, technicians,

operators, security force members, and office personnel.

Attended exit interview on May 24 and June 7, 1985

2. Exit Interview

The initial inspection scope and findings were summarized on May 24, 1985,

with those persons indicated in paragraph 1 above. Two violations and two

unresolved items were identified during the first part of the inspection.

The followup inspection resulted in one of the saresolved items becoming a

violation and clearing of the other. These results were discussed in an

exit interview on June 7, 1985. The licensee acknowledged understanding

of the issues discussed and took no exceptions to the violations.

a. Violation 370/85-21-01: Failure to follow procedures, paragraphs Sc

and 5d

b. Violation 369/85-20-01,370/85-21-02: Failure to review procedures

changes, paragraph Se.

c. Violation 369/85-20-02, 370/85-21-03: Failure to retain records,

paragraph 5f.

Proprietary material was reviewed during this inspection, but is not

incorporated in this report.

o .

'

.

2

3. Licensee Action on Previous Enforcement Matters

This subject was not addressed in this inspection.

4. Unresolved Items

Unresolved items were not identified during thir, inspection.

5. Followup of Critical Rod Position Prediction Error

This special inspection was /cyoted to followup of the reactor startup on

May 17, 1985, when the reactor was taken critical below the rod insertion

limit of Technical Specification 3.1.3.6.

a. Sequence of Events

Time Date Event

Hour

1149 5/5/85 Initial criticality for Unit 2 Cycle two

0430 5/8/85 Zero power physics testing complete

0617 5/8/85 Entered Mode 1 (> 5% power)

0648 5/8/85 Reactor trip from 10% power

1312 5/8/85 Reactor critical (virtually xenon free - no

problems experienced)

0924 5/16/85 Reactor trip from 96% power from failure of

non-safety related power generation

equipment

0609 5/17/85 Reactor critical below insertion limits

Predicted: 38 steps on bank D

Actual: 26 steps on bank C

Insertion limit: 47 steps on bank C

1149 5/17/85 Reactor critical (xenon value in question,

but no changes made to procedure)

Predicted: 228 steps on bank D (plus a

205pcm deficit, i.e. subcritical)

Actual: 90 steps on bank D

1344 5/17/85 Generator on line and power escalation

! continued

!

. .. .. .

. .

t

'

.

3

b. DiscussionofXenon(Xe) Values

Since the beginning of operation of McGuire 1 and 2 the licensee has

felt that the vendor (Westinghouse) underestimated peak and

equilibrium Xe values. This was based on past experience and surveys

of other utilities. A value different from the vendor peak value by

1000 percent millirho (pcm) was used for Unit 2 cycle two. The

values used for Unit 2 cycle two for peak and equilibrium Xe were the

same values used for Unit 1 cycle two, due to similar core designs.

Discussions with the corporate nuclear engineering staff revealed

that the general office usually underpredicts the vendor Xe values

and the plant reactor engineering group supplies a multiplier to the

general office values for values used in the plant computer. The

plant values are usually greater than the vendor supplied values.

The general office felt that the library of cross sections and decay

constants (EPRI library) used in their computer code (PDQ - a two

dimensional diffusion theory code) underpredicted the actual Xe

values in the core. Also, it was felt that the vendor estimates have

improved lately due to changes in the vendor codes and updated

library. This would account for the values used by the plant after

5/17/85 being similar to the vendor values.

Further checking by the reactor engineering group on 5/22/85 found

that the values used in Unit I cycle two were in error. The values

were changed in error in August 1984, based on an estimated critical

on bank D instead of the correct value using bank C. The Xe values

were changed to more closely predict estimated critical rod position.

This error was carried over into Unit 2.

c. Licensee Analysis of 5/17/85 Event Performed on 5/17/85

The previous calculations of critical position were checked and found

correct. Boron samples were also verified as correct.

The plant nuclear engineer assigned to handle the missed estimated

critical position concluded that the error was due to Xe. This was

based on several things. First, the startup on 5/8/85 was essen-

tially Xe free and no problems were noted. Other parameters in the

reactivity balance used to calculate critical position were recently

verified during startup physics testing (control rod and boron

worths). This startup was the only startup in the cycle with signif-

icant amounts of Xe, and the time elapsed between the trip and

startup involved estimating Xe on the peak Xe curve from the decay of

iodine following a reactor trip. Previous estimates of Xe had been a

source of debate and a measurement of Xe in this cycle had not been

performed.

.

- - . - , - . - . -

,. .--- - ,-,,ea

- ,. -- -

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

. .

.

4 .

,

4

. Review of the nuclear instrument chart recorder traces was performed

+

on June 6-7, 1985, at the request of the inspectors. That review

revealed that the startup rate on 5/17/85 at 0600 hours0.00694 days 0.167 hours 9.920635e-4 weeks 2.283e-4 months had been

faster than the 1.0 decade per minute (dpm) specified by the plant

procedures, specifically 0P/2/A/6100/01, Controlling Procedure for

Unit Startup (item 2.3), and OP/2/A/6100/05, Unit Fast Recovery (item

1..

2.1).- The licensee's evaluation of the rate was 1.3 dpm, and the

inspectors calculated 1.8 dpm. Failure to follow procedure was

identified as a violation of Technical Specification 6.8.1.a which

requires that written procedures shall be established, implemented,

and maintained for operating procedures, Violation 370/85-21-01,

Failure to follow procedures.

d. Restart on 5/17/85

For the restart at 1149 hours0.0133 days 0.319 hours 0.0019 weeks 4.371945e-4 months on 5/17/85, a new estimated critical

rod _ position was calculated using plant procedure Reactivity Balance

'

Calculation, Enclosure 5.2 of OP/0/A/6190/06. An estimated position

!- of 228 steps on bank D was calculated. This is all rods out. Step 9

4

of the procedure yielded a negative number for the reactivity worth

i of rods to be inserted and a negative value required an estimated

critical boron calculation per Enclosure 5.1. This would have

required a boron dilution but the results of the enclosure were

ignored and a procedure change not initiated to update the Xe values :

'

believed to be in error. The procedure had been checked by a second

licensed operator.

The dilution was ignored due to the fact that the engineer believed

i the Xe to be in error and a dilution would have resulted in a possi-

ble repeat of the criticality below the insertion limits. The Xe

< worth was thought to be 717 pcm less than as previously used. The

boron concentration also had been increased by 227 ppm.' Using these

l values the critical position was estimated by the engineer to be

around 100 steps on bank D. The actual critical position was 90

steps.on bank D. In effect, the procedure was used to estimate

critical rod position and was used as a test to determine Xe worth.

Failure to follow procedure was identified as a second example of the

violation of Technical Specification 6.8.1.a. which requires that

I, written procedures shall be established, implemented, and maintained

,

for operating procedures (370/85-21-01). This violation was dis-

,

cussed with plant management in the exit meetings on 5/24/85 and

6/7/85.

'

e. ~ Computer Changes after the Startup to Correct Xe Values

'

The value for Xe worth used in step 3 of OP/0/A/6190/06 comes from a

l

computer program. Changes were initiated to the plant off-line

computer which were not in accordance with nomal procedures.

i Nomally, the corporate nuclear engineering staff reviews these

changes and after approval the computer technicians make the changes,

,

_wn,,_.,, ,.n,,_..,., w_,...... . , _ , . . ,..,., , , . _ ,.

~----e--m -,-.-er, , , - , . + , r. ,,,~.,-.,~,,m. ,-, _.

. .

.

5

On 5/17/85 the changes were made to the off-site (corporate) computer

first by computer technicians and then mailed to corporate for.

approval. Discussions with corporate personnel revealed that they

were initially unaware of the event and first learned of the event

days _later from another facility.

The on-line (process computer) was updated with new Xe values on

5/22/85. Changes to the plant computers are made with a Production

Technical Services Service Request Routing and Approval Form (Form

34744 (12-81)). This form requires approval by a technical sponsor.

The reviews of these forms in the past had not identified any errors

in Xe values.

Subsequent analysis of the process through which the Xenon Follow /

Predict program is changed revealed some concerns, which are detailed

below.

Changes to the Xenon Follow / Predict program were made without per-

forming a Nuclear Safety Evaluation of the change. There have been-

-at least five recent changes to this program, none of which had an

associated Nuclear Safety Evaluation until May 24, 1985.

1) June 29, 1984 (Initial cycle values)

2) July 9,1984 (Change to near vendor values)

3) Sept. 11, 1984 (Errorintroduced)

4) April 24, 1985 (Unit 2 cycle two initial values)

5) May 17, 1985 (Correction)

The evaluation for the May 24, 1985, change was performed five days

after the change had been implemented. In an effort to determine why

the evaluations had not been performed, it was detennined that the

licensee did not recognize that changes to the computer calculated

values were considered to be a procedure change. Administrative

Policy Manual 4.2.4.1.e requires all procedures changes to receive a

nuclear safety evaluation prior to use. Since the calculation

derived from the computer program are directly employed in procedure

OP/0/A/6190/06 Reactivity Balance Calculation for determining esti-

mated critical positions _and shutdown margins, which are safety-

related detenninations, the change should receive the same

considerations of technical viability and safety implications as any

other change to a safety-related procedure. Any change, therefore,

should be preceded by a Nuclear Safety Evaluation.

The. failure to implement Nuclear Safety Evaluations for changes to

the referenced computer program is a violation to the requirements of

10 CFR 50 Appendix B_ criterion V. Activities affecting quality are

required to be accomplished in accordance with instructions or

procedures appropriate to the circumstances which embody the neces-

sary acceptance criteria to ensure the satisfactory accomplishment of

the task. Data such as the Xe values used in the procedure are

required to be technically valid and receive the-same review and

. .

.

6

approval as the procedure in which it is employed. This is violation

50-369/85-20-01, 50-370/85-21-02, Failure to review procedure

changes.

f. Inspector Review of Shutdown Margin Calculations

The inspector reviewed the shutdown margin calculation performed on

5/17/85 prior to the 1149 hours0.0133 days 0.319 hours 0.0019 weeks 4.371945e-4 months startup on May 17, 1985, and found

that by using the inflated Xe values in the procedure, errors result-

ed in the nonconservative direction. The procedure, Shutdown Margin

- Unit Shutdown, Enclosure 5.5, allows taking credit for the Xe worth

in the shutdown margin and estimates a time that boration would be

required as the Xe decays. The time was determined to be 2100 hours0.0243 days 0.583 hours 0.00347 weeks 7.9905e-4 months

using the inflated Xe values and 1700 hours0.0197 days 0.472 hours 0.00281 weeks 6.4685e-4 months using the Xe curve with

the updated Xe values.

The inspector requested a review of all past shutdown margin calcula-

tions performed since the introduction of the Xe error in Unit 1

cycle two. These records were not available. The licensee performed

the surveillance calculation but was retaining only the shutdown

margin value and not retaining the data used to perform the calcu-

lation. The inspector stated that this practice was inconsistent

with record retention requirements for surveillances required by

Technical Specifications. The records for the shutdown margin

surveillance, as required by T.S. 4.1.1.1.1.e, while in modes 3, 4

and 5 for the period August 1984 to June 1985 were unavailable. With

the reactor in mode 3, 4, or 5 the shutdown margin is required to be

determined once every 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months considering the following factors:

1) Reactor coolant system boron concentration

2) Control rod position

3) Reactor coolant system average temperature

4) Fuel burnup based on gross thermal energy generation

5) Xenon concentration

6) Samarium concentration

The Technical Specification requirement is implemented in Station

Directive 2.1.1 attachment 5, Document Retention Requirements, which

requires a six year retention of completed safety-related permanent

station procedures that involve documentation of compliance with

acceptance criteria. The failure to retain the records was identi-

fied as violation 50-369/85-20-02 and 50-370/85-21-03.

g. Licensee Reconstructed Data for Shutdown Margin

The licensee was able to reconstruct some of the data used for

determination of shutdown margin. A review of these data was per-

formed to determine if the required shutdown margin had been reduced.

This review was conducted using the updated Xe values and minimum

boron concentration following reactor trips or shutdowns from the

time the peak Xe error was introduced in August 1984 for Unit 1

, .

.

7

cycle two. A check was made to determine if the decay of Xe reduced

the shutdown margin. The licensee also stated that the shutdown

boron concentration curve contained a 100 ppm (1000 pcm) safety

factor.

The results of the review revealed that on 2/5/85 the reactor trip

was 23 ppm short of the required shutdown margin but, by using the

100 ppm safety factor, 77 ppm was available before the shutdown

margin was violated. Also, for the shutdown on 11/23/84 and cooldown

on 11/24/84 it could not be determined if the required boration took

place prior to or after cooldown due to the lack of records of boron

sampling times.

The results of the review were not made available for the inspectors

review until 6/6/85. The inspectors questioned the source and basis

of the 100 ppm safety factor. The safety factor is built into the

vendor supplied table for the shutdown boron concentration. Discus-

sion with the vendor revealed the safety factor is introduced because

of known uncertainties in the vendor calculations such as control rod

worths, and boron worth, and that the safety factor is not an addi-

tional margin available to the licensee. A vendor engineer offered

the opinion that approximately 25 ppmB excess margin was available

because measured boron concentrations at McGuire were consistently

less than the predictions of the Westinghouse calculations.

h. Review of Unit Trip on 6/1/85 and Startup on 6/2/85

The reactor trip of Unit 2 cycle two at 0919 hours0.0106 days 0.255 hours 0.00152 weeks 3.496795e-4 months on 6/1/85 and

startup on 6/2/85 was reviewed for estimated critical rod position

problems. At the time of startup, after the trip, the value of the

Xe worth was greater than 100% power equilibrium. The startup was

conducted using the corrected Xe values implemented after 5/17/85

(nearvendorvalues). The estimated critical rod position was within

500 pcm of the actual rod position and no problems were noted during

the startup.

i. Summary and Conclusions

The corporate nuclear engineering staff conducted an evaluation of

taking the reactor critical below the rod insertion limits and

identified two safety parameters of concern - the available shutdown

margin and the ejected rod worth. The shutdown margin for Unit 2 at

0600 hours0.00694 days 0.167 hours 9.920635e-4 weeks 2.283e-4 months on May 17, 1985 was found to exceed the required margin.

The ejected rod worth was within that analyzed in the Reload Safety

Evaluation.

The startup procedure should have been revised prior to the second

startup when it was found to be incorrect. Further, inadequate

reviews of reactivity parameters used in estimated critical rod

position and shutdown margin calculations led to the problem of

taking the reactor critical below the rod insertion

e

E

. .-

.

8

limits, exceeding the startup rate limit prescribed by procedure and

the possibility of nonconservative errors in the shutdown margin

-calculations. The lack of plant records of previous shutdown margin

calculations made the review process difficult. A complete review of

this event was not conducted by the licensee until prompted by the

-inspectors.

6. Post-Refueling Startup Tests Unit 2 Cycle Two (72700). 1

The following post-refueling startup tests for Unit 2, cycle two were

reviewed:

,

a. PT/0/A/4150/21, Post Refueling Controlling Procedure for Criticality,

Zero Power Physics, and Power Escalation Testing. The test was

approved on April 25, 1985 and performed on May 4-24, 1985. The

impact of core alterations on excore detector signal was evaluated

and found to be acceptable (step 8.3). The predicted shutdown margin

at the rod insertion limit was 3115 pcm, considerably in excess of

the minimum required of 1300 pcm (step 8.4). The shutdown margin for

the rod-swap tests (reference bank in) was acceptable (step 8.5).

More than one decade of overlap between the source-range and inter-

mediate-range nuclear instruments was demonstratea, enclosure 13.2.

Using enclosure 13.3 the point of adding nuclear heat and the upper

power limit for zero-power testing were established. Performance of

the reactivity computer was compared against stop-watch period and i

reactivity measurements for reactor instrumentation using enclosure

13.4, and found to be acceptable,

b. PT/0/A/4150/28, Criticality Following a Change in Core Nuclear

Characteristics. The test was approved April 24,-1985, and performed

on May 4-5, 1985. The chi-squared statistical test was used to

confirm proper operation of the source range detectors used in

providing input for the inverse multiplication calculations that

guided the approach to criticality.

c. PT/0/A/4150/10, All Rods Out Boron Endpoint Measurement, was approved

on April 26, 1984 and performed on May 5-8, 1985. The predicted

all-rods-out boron endpoint concentration was predicted to be 1491

+/- 50 ppmB. The measured result was 1423 ppmB, which was outside

the numerical acceptance limits. The result was evaluated by Westing-

house, at the request of the licensee and found to be acceptable.

d. PT/0/A/4150/12A, Moderator Temperature Coefficient of Reactivity

,

during Startup Mode. The test was approved on April 26, 1984. It

was performed for all rods out on May 6, 1985 and with bank D in on

- May 7, 1985. Although the results were acceptable, the inspector

noted that the differences in results between heatup and cooldown

measurements were greater than observed in the past. The temperature

'

increments were only one degree F. Larger increments would yield

,

more consistent results, and would be compatible with the fuel

1 vendor's recommendations for startup tests,

i

. - _ _ _ .

. . - - - _ - - - .-_-- _ . - . - _ . -. . _ - _ _ - -.

<

s. . .

. .

9

e. PT/0/A/4150/11, Control Rod Worth Measurement, was approved on

April 26, 1984. It was performed on May 6, 1985, for control bank C,

which was to be used as the reference bank in later rod-swap measure-

ments. The reactivity computer traces were independently evaluated

by the inspector, and good agreement with licensee values (within 0.2

pcm per increment) was obtained. The measured worth was 871 pcm,

which was in acceptable agreement with the predicted value of 908 +/-

91 pcm.

f. PT/0/A/4150/11A, Control Rod Worth Measurement, Rod Swap, was ap-

proved on April 30, 1984, and was performed over the period May 6-7,

1985. All measured worths met the acceptance criterion of less than

30% or less than 200 pcm difference from prediction.

g. PT/0/A/4150/02A, Core Power Distribution, was approved on April 19,

1985. The first measured power distribution at.100% power was,

performed on May 15-18, 1985. All thermal and power distribution

limits were satisfied.

h. PT/0/A/4600/02E, Incore and NIS Recalibration: Post Outage, was

approved on May 25, 1984, and performed on May 10-12, 1985. The

results of eleven quarter-core flux maps were analyzed using a

computer program to obtain the correlation between the individual

chamber currents of the power range detectors and axial offset. The

program does not perform an evaluation of the quality of fit of the

data to the correlation line. The licensee stated that the results

are also plotted by hand and a visual evaluation of the reasonable-

ness of the results performed. The inspector had no further ques-

tions regarding this practice.

l

i

4

ML20132G668

Text

Navigation menu

Search