IR 05000335/1988026
| ML17222A588 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 11/04/1988 |
| From: | Burnett P, Jape F NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML17222A587 | List: |
| References | |
| 50-335-88-26, 50-389-88-26, NUDOCS 8811220494 | |
| Download: ML17222A588 (18) | |
Text
"tCy~
Wp
+
Cy
~
p%
I
g
Cy Cy
~O UNITED STATES NUCLEAR REGULATORY COMMISSION
REGION II
101 MARIETTAST., N.W.
ATLANTA,GEORGIA 30323 Report Nos.:
50-335/88-26 and 50-389/88-26 Licensee:
Florida Power and Light Company 9250 West Flagler Street Miami, FL 33102 Docket Nos.:
50-335 and 50-389 Facility Name:
St.
Lucie 1 and
License Nos.:
DPR-67 and NPF-16 Inspection Conducted:
September 6 - 30, 1988 Inspector:
P.
T. Burn t Approved by F. Jape, Section Chief Engineering Branch Division of Reactor'Safety ri M/g$
Date Signed
/I Date Signed SUMMARY Scope:
This routine, unannounced inspection addressed the areas of post-refueling startup and power escalation tests for both units.
Results:
The procedure for initial criticality following refueling was found to be sound in principle, but requiring improvement in implementation in confirming operability of neutron monitors, plotting inverse multiplication, and establishing a criterion for stopping dilution-paragraph 2.
The zero power physics test program was found to be basically sound, but needing improvement in calibration of the reactivity computer, labelling of chart records, and evaluation of the results of rod wor~th. measurements paragraph 3,
The;-".r'eactor engineering power ascension program appeared to he ade'quate and well documented paragraph 4.
No violations or deviations were identified.
PDR A
4 881115 8811'~
DOCK 05000335 Pvu
REPORT DETAILS Persons Contacted Licensee Employees'J.
H. Barrow, Operations Superintendent
"G. J. Boissy, Plant Manager
- S.
G. Brain, Independent Safety Engineering Group
"C.
L Burton, Operations Supervisor
"R. Dawson, Assistant Plant Superintendent, Electrical
"T. A. Dillard, Maintenance Superintendent J.
A. Dyer, Quality Control
"J.
B. Harper, Quality Assurance Supervisor K. N. Harris, Vice President
"C.
F. Leppla, I 8 C Supervisor
"N. G.
Roos, Quality Control Supervisor
"D. H. West, Technical Staff Supervisor
"C, L. Wilson, Assistant Plant Superintendent, Mechanical
"E. J. Wunderlich, Reactor Engineering Other licensee employees contacted 1ncluded engineers, technicians, and office personnel.
NRC Resident Inspector
"G. L. Paulk, Senior Resident Inspector
- Attended ex1t interview Acronyms and initialisms,used throughout this report are listed in the last paragraph.
Initial Criticality after Refueling (72700)
The completed procedures for the most recent (Unit I, Cycle 9 and Unit 2 Cycle 4')',. post-refueling startups were reviewed.
These were 1(or 2)-
0030221"-.-"(Revision 15 or 7), Unit I (or 2) Initial Criticality Following Refuel1ng.
The procedures are essentially identical and define an accept-abl e phi 1 osophy of control l ing and monitoring react1vi ty changes in bringing a
new core 1oad1ng to 1ts first criticality.
However, the implementation of that philosophy could be strengthened in the procedures in the following areas:
a.
The chang1ng core reactivity is mon1tored by the traditional plot of 1nverse multiplication or ICRR versus a reactivity variable such as
rod position, RCS C,'r dilution time.
For the majority of the approach to criticality, the count-rate monitoring instruments are the SRNIs.
The pre-critical calibration of the SRNIs amounts to nothing more than the setting of bi stables, exempts the neutron detectors from the calibration process, and does nothing to confirm that the total SRNI system or channel is responding stably and proportionately to the existing neutron flux.
Hence, the calibration gives no configence that the channels will respond proportionately to increases in flux.
b.
C.
Licensee concern for establishing true operability of the SRNIs is reflected in Step 8. 13, which requires additional procedure steps if a SRNI does not appear to be responding at a time multiplication is expected.
That concern could be better resolved using the methods discussed below. True operability of pulse counting systems, such as the SRNIs, can be established using statistical tests, which have been in use in the nuclear industry since its inception.
One of these tests, the Chi"squared test is an example, should be employed whenever the SRNIs are the sole or primary reactivity monitoring instruments.
Reference material on statistical tests was provided to the licensee following the inspection.
In the recent Unit 1 test, it was observed that the plot of ICRR versus time was too steep for good time resolution.
Criticality was obtained in about three hours of dilution, but the time axis was scaled for 0 to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />.
Since the dilution time can be estimated within an hour or two, the time axis scaling can be chosen more appropriately.
Step 8. 14 secures dilution when inverse multiplication is reduced to 0.'05.
At other facilities, a termination criterion this small, at dilution rates greater than
gpm, has lead to significant over dilution requiring prompt action by the operators to insert rods to control flux level.
3.
Zero Power Physics Tests (72700, 61705, 61708, 61710)
Zero power physics test were controlled by operating procedure no.
0110052 (Revision 7), Zero Power Physics Tests after Reload.
The actual measure-ments were conducted as prescribed in the following attachments to the procedure:
A.
Reactivity Computer Checkout:
The minimum flux level for testing was established by increasing flux until the peak-to-peak amplitude of the reactivity recorder trace was less than 2 pcm.
The maximum flux level for testing was determined by finding the level at which sensible heat was detected by observing an increase in T-AVG or a
decrease in reactivity measured by the reactivity computer as a
result of doppler feedback.
The maximum flux was then the bottom of the power decade in which heating was observed.
Positive and
negative period checks were'erformed by comparing reactivity computer indicated reactivity with that obtained by measuring reactor period manually with a
stopwatch and solving the inhour equation.
Only one positive or negative reactivity swing was performed at
pcm.
The acceptance criterion was agreement within 10 X.
During the Unit 1, Cycle 9 tests, the agreement was within 2.4% for both cases.
Prior to use, the reactivity computer was required to be checked out electronically using I&C Procedure 1400165.
For the Unit 1, Cycle 9, tests, a
new digital reactivity computer was used.
Procedure 1400165, which had been reviewed by the FRG and approved by the plant manager, was deleted from the test requirements by TC¹l-88-205 and the vendor manual substituted.
The vendor manual did not receive an equivalent review prior to use.
The accepted Topical guality Assur-ance Report, Appendix C, in its exception to Paragraph 5.3.5(4) of ANSI N18.7 appears to permit the substitution, which then over rides the following requirements and considerations:
TS 6.8. l.c requires written procedures be established, imple-mented, and maintained for surveillance and test activities of safety related equipment.
TS 6.8.2 requires that such procedures be reviewed by the FRG and approved by the plant manager prior to implementation.
TS 6.8.3 allows temporary changes to such procedures providing the intent of the original procedure is not altered, the change is documented and reviewed by the FRG and approved by the plant manager within 14 days of implementation, This issue will be discussed further with Region II management.
With respect to the quality of the bench test of the reactivity computer, neither the vendor manual nor the procedure provide very convincing calibrations.
Both use a step change in input test signal to stimulate a time-varying, simulated-period output.
Other manufac-turers of reactivity computers provide true exponential signals, with both positive and negative exponents for calibration of the computer.
An; alternative to better bench tests is to perform more extensive comparisons of reactivity computer solutions with those obtained from stopwatch period measurements.
Facilities that have chosen that approach typically use nominal test reactivities of 20, 40, and
pcm for the comparisons.
The acceptance criterion is typically 4X agreement, and use of the reactivity computer is procedurally limited to the range in which the acceptance criterion is satisfied.
These features are absent form current procedures at St.
Lucie.'ther activities performed using this attachment included symmetry checks of the CEA ARO Critical Boron Concentration had an acceptance criterion of 100 ppm B between measured and predicted boron concentration.
It was performed in conjunction with data sheet 3,
which accounted, for differences in temperature and rod position between 'actual and predicted conditions.
The procedure required that only positive differences between measured minus design CB be entered in the Reactivity Deviation Log.
For Unit 1, Cycle 9, the ARO critical boron concentration was 1452 ppm B, which was in acceptable agreement with the design value of 1407 ppm B.
For Unit 2, Cycle 4, the ARO C
was 1499 ppm B, which was in accept-able agreement with the predicted value of 1566 ppm B.
Isothermal Temperature Coefficient was measured at conditions approx-imating ARO for both heatup and cooldown of the RCS and averaged without weighting for the magnitude of the temperature change.
The measured MTC was obtained by subtracting the design value of FTC from the ITC.
The design MTC was adjusted for differences in design and actual CB before comparing with the measured MTC.
The acceptance criteria were that the measured MTC was less than the values speci-fied in TS 3. 1. 1. 1.4.a and within 2 pcm/F of the corrected design value.
For the Unit 1, Cycle 9 startup tests, the ITC, the average of four measurements, was +2.55 pcm/F, and the MTC was +4.07 pcm/F, which was less than the TS 3. 1. 1.4 limit of +5 pcm/F and within tolerance of the design value.
The Unit 2, Cycle 4, test yielded an MTC of 3.45 pcm/F, and was equally acceptable.
CEA Group Worths (Rod Swap Method) required that first a reference group be designated and its worth measured during boron dilution.
Then with the reference bank at the LEL, criticality was adjusted by manipulation of the lead group, which was about 120 inches withdrawn and fixed in position for the remainder of the test.
Then the other rod groups were inserted alone, one group at a time, and criticality maintained by adjustment of the reference group position.
The measured worth of any test group was taken as the withdrawn worth of the reference group above the LEL.
For comparison with the predicted worth of the test group, it was necessary to find the predicted worth as function of the position of the reference group.
For Unit 1, Cycle 9,
the necessary design information is contained in ANF-88-049(P),
St'.
Lucie Unit 1, Cycle 9,
Startup and Operations Report, July, 1988.
However, neither this procedure nor this reference contain instructions on how the design data are to be manipulated to obtain the prediction, and this procedure does not
capture the calculations or make them a part of the record.
Based.
upon experience at other facilities, the inspector was able to make independent calculations of predicted worth, without reference to a
procedure, that were in agreement with those final values recorded by the licensee.
For Unit 1, Cycle 9, the agreement between measured and predicted values of rod group worths ranged from 2.7 5 to 14.3
%, with the measured worths 'larger in all cases.
The sum of the measured worths exceeded the design value by 7.5 X.
The concomitant values of boron worth were 9.67 and 9.63 pcm/ppm 8 for measured and predicted values, respectively.
For Unit 2, Cycle 4, the agreement between measured and predicted values of rod group worths ranged up to 7.7 I, with the measured worth smaller in all cases.
The sum of the measured worths was less than the sum of the design values by 3.7 X.
Since the worth determination of the reference bank during boron dilution is the basis for the entire rod worth measurement process, the inspector independently reevaluated the reactivity computer chart records for both the Unit 1, Cycle 9, and the Unit 2, Cycle 4.
In that review, several poor practices were noted:
(1)
Generally, the analog traces were not labelled, nor were the scales used noted on the chart.
(2)
The reactivity computer was being used outside the calibrated range defined by the span of stopwatch period measurements compared with reactivity computer solutions.
(See the discus-sion of Attachment A above.)
The calibrated spans were nominal-ly 20 pcm, and typical reactivity increments ranged from +30 to-30 pcm, with a -few extremes reaching
+ or -50 pcm.
(3)
No plot of rod group differential reactivity worth was made for comparison with prediction.
Only an integral plot was made for use in the rod swap measurements.
The microcomputer program SUPERCALC3 was used by the, inspector to manipulate the raw data and to plot the results, with the licensee's results for comparison, as differential worth curves.
The agreement between results was excellent, but the curves, Attachments
and 2, show far more internal structure than observed in similar curves for other PWRs.
As shown in Attachments 3 and 4, this information is not obtained if only integral worth curves are plotted.
The cause of the structure has not been determined.
It does not correlate with the internal grid elevations in the fuel bundles.
The licensee will investigate this finding furthe E.
CEA=-Group Worths by Dilution (No Overlap)
was not performed for Unit 1, Cycle 9.
This test is optional when worth measurement by rod swap is successful.
In addition to the procedures and attachments discussed above, the follow-ing documents were reviewed in the inspection of the zero power physics test programs:
St.
Lucie Unit 2, Cycle 4 Startup Physics Report, issued February 29, 1988; ANF-88-049(P),
St.
Lucie Unit 1, Cycle 9,
Startup and Operations Report, issued July 1988; XN-75-27(P),
EXXON Nuclear Neutronics Methods for Pressurized Water Reactors, Supplement 5, issued September 8,
1986, which describes and justifies the rod-swap methodology; C-E Notebook, Plant Startup Test Predictions and Physics Data Book, which applies to Unit 2, Cycle 4.
No violations or deviations were identified.
4.
Power Ascension Testing (72700)
The be innin of c cle ower OP 3200020, Primary System Manual Calorimetric, OP 1200051 (Revision 2), Nuclear and Delta T Power Calibration, OP.'200022 (Revision 11), Periodic Surveil-lance of Incore Detection Sy'tem, OP 3200053 (Revision 1), Surveillance Requirement for Azimuthal Power Tilt (T q )
OP 3200058 (Revision 2), Surveillance Requirements for Total Planar (F
) and Total Integrated (FR ) Radial Peaking Factors, T
T xy IKC 1220052 and OP 1200023, Linear Power Range Channel Calibration, g
g y
p escalation was controlled by procedure 0010133 (Revision 4),
Reactor Engineering Power Ascension Program, for both units.
The program is divided into test plateaus at 25, 50, 80, and tl98X RTP.
The program appeared to be well conceived and well documented and to assure required surveillances were performed.
For the recent Unit 1 startup, the inspector confirmed that the following calibration and surveillance procedures were completed successfully as and when required by the program:
OP 3200050 (Revision 2), Calculation and Adjustment of Fixed Incore Detector Alarm Points, and OP 3200057 (Revision 10),
Power Distribution Comparison with Design.
No violations or deviations were identified.
5.
Exit Interview The inspection scope and findings were summarized on September 30, 1988, with those persons indicated in paragraph
above.
The inspector de-scribed the areas inspected and discussed in detail the inspection find-ings.
Dissenting comments were received from the licensee in response to the inspectors comments on use of unr eviewed vendor manuals, which i s discussed in paragraph 3.A.
Proprietary materials were provided to and reviewed by the inspector during this inspection, but are not incorporated into this report.
6.
Acronyms and Initialisms Used in this Report ANF ANSI-ARO C-E CEA FRG
'
FTC gpm ICRR "
ITC LEL NTC OP pcm ppm B-PWR RCS RTP SRNI "
T"AVG-TC TS Advanced Nuclear Fuels Company American National Standards Institute all rods out Combustion Engineering Company control element assembly Facility Review Group fuel temperature coefficient gallon per minute inverse count rate ratio isothermal temperature coefficient lower electrical limit of CEA insertion moderature temperature coefficient operating procedure percent millirho parts per million boron pressurized water reactor reactor coolant csystem
,rated thermal power source'ange nuclear instrument average temperature of RCS temporary change Technical Specification Attachments:
~
2.
3:
4.
St.
Lucie St.
Lucie St.
Lucie St, Lucie 1, Cycle 9, Reference Group A, Differential Worth Curve 2, Cycle 4, Reference Group B, Differ'ential Worth Curve 1, Cycle 9, Reference Group A, Integral Worth Curve 2, Cycle 4, Reference Group B, Integral Worth Curve
~ Inspector
~ LIcensee H I I HLJI1I'ICIHI ST. LUCIE 1, CYCLE 9, REFERENCE GROUP A D~fferenhal Fo'rN, Curve
80
~i5
~
N 10
15 SQ
- r5
i05 186 f35 Gramp Average Withdrawal {inches)
%1
~ Inspector Licensee ATTACHMENT 2 ST. LUCIE 2, CYCLE 4, REFERENCE GROUP B
84fferen,N,al N'orth, Canoe
Q
~A N 80 i5 SQ
6D V5 SO f05 180 f.35 Group Average 'Withdrawal (inches)
ATTACHMENT 3 ST. LUCIE 1, CYCLK 9, REFERENCE GROUP A fntegral Forth Cu~e f085 V
690
~r 575 4c 460 845 lX j
~
45 BO VG
Group Withdrawal (inches)
105 180 f35
ATTACHMENT 4 ST. LUCIZ g, CYCLE 4, REFERENCE GROUP B
Integral Forth Cu~e 2li5 18SO 1645 V X4<0
~~ l176 940 705
45
VG
Group Withdrawal (inches)
I