IR 05000335/1993003
| ML17227A777 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 03/12/1993 |
| From: | Burnett P, Crlenjak R, Curtis Rapp NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION II) |
| To: | |
| Shared Package | |
| ML17227A776 | List: |
| References | |
| RTR-NUREG-CR-5560 50-335-93-03, 50-335-93-3, 50-389-93-03, 50-389-93-3, NUDOCS 9303300019 | |
| Download: ML17227A777 (8) | |
Text
UNITED STATES NUCLEAR REGULATORY COMMISSION
REGION II
101 MARIETTASTREET, N.IN.
ATLANTA,GEORGIA 30323 Report Nos.:
50-335/93-03 and 50-389/93-03 Licensee:
Florida Power and Light Company 9250 West Flagler Street Miami, FL 33102 Docket Nos.:'0-335 and 50-389 License Nos.:
DPR-67 and NPF-16 Facility Name:
St.
Lucie Units 1 and
Inspection Conducted:
February 8-12, 1993 Inspector:
e Date Signed Inspector
.
W.
Rapp Accompanying Personnel:
R.
V. Crlenjak (February 12, 1993)
ate Signed Approved by:
R.
V.
rlenjak, Chief Operational Programs Section Operations Branch Division of Reactor Safety iDat Signed SUMMARY Scope:
This routine, unannounced inspection was conducted to review the event of non-conservative errors in Unit 2 thermal power measure-ment, as a consequence of coherent drift of six resistance tempera-ture detectors in the feedwater system.
Results:
The inspectors concluded that feedwater temperature monitoring was consistent with standard industry practice for a non-controlling parameter and that no method of earlier detection of coherent drift of the six resistance temperature detectors was available from that practice.
The licensee's immediate corrective actions included reducing the span of tolerable difference between detectors from 10'F to 5'F and installing a computer-generated'larm whenever any detector exceeded the nominal 435 'F feedwater temperature by 2'F.
The licensee is 9303300019 930312 PDR ADOCK 05000335,
evaluating the possible benefits of additional corrective actions, such as performing a beginning-of-cycle precision heat balance, or serial replacement of the detectors to install one newly calibrated detector at the start of each fuel.cycle.
Extensive root cause analysis of the detector drift is underway in the laboratories of the licensee and the vendor of the detectors.
However, there is no certainty that the root cause will be identified.
Unit 2 operated at 101% of rated thermal power for an extended period of time.
This extended over-power operation appears to satisfy internal NRC management guidance for issuing a violation.
However, the precipitating cause of the over-power operation is quite different from those causes considered by NRC management in
'issuing that guidance.
Therefore, with the concurrence of Region II management, no violation will be issue REPORT DETAILS 1.
Persons Contacted Licensee Employees
- G. Boissy, Plant General Manager
- C. Burton, Operations Manager
- R. Dawson, Maintenance Manager
- R. Englemeier, Site equality Manager
- T. Glenn, Instrumentation and,Control, Lead Supervisor
- J. Hoge, Instrumentation and Control, Engineering Supervisor
- J. Holt, Licensing Engineer
- L. HcLaughlin, Plant Licensing Manager
~A. Henocal, Mechani'cal Maintenance Department Head
"G. Novarro, Instr'umentation and Control Engineer
- E. Ordway, Instrumentation and Control System Supervisor
- C. Pell, Services Manager
- L. Rogers, Instrumentation and Control Supervisor
- J. Scarola, Engineering Manager
- D. West, Technical Manager J.
West, Operations Supervisor C.
Wood, Operations Support Other licensee employees contacted during this inspection included engineers, operators, technicians, and administrative personnel.
NRC Resident Inspectors
- S. Elrod, Senior Resident Inspector H. Scott, Resident Inspector
- Attended exit interview on February 12, 1993.
2.
Thermal Power Monitoring and Feedwater Temperature Measurement Errors (61706)
a.
Introduction On November 5 and 6, 1992, St.
Lucie Unit 2 underwent loop-by-loop replacement of RTDs used in feedwater temperature measurement.
As each set of new RTDs was placed into service, the calorimetric calculation of the DDPS indicated power in excess of 100% RTP.
In each case, power was reduced conservatively until the cause of the change could be determined.
The immediate cause of the calculated increase power increase was the lower temperature indicated by the new RTDs and input into the calorimetric calculation.
One of the original RTDs was destroyed in removing it from its well.
Subsequent measure-ments of the performance of the remaining five of the six original RTDs revealed that, in a water bath at the nominal feedwater temper-ature of 435'F, these RTDs indicated temperatures ranging from 436. 1'F
~
Report Details to 447.5'F.
The average error in feedwater temperature was 5.57'F, which the licensee equated to an over-power error of 1.04%
RTP.
This inspection was conducted to review the licensee's evaluation of the magnitude and extent of the overpower error and the root causes of the non-conservative errors in the temperatures indicated by the feedwater RTDs.
b.
Documentation Reviewed The following documents were reviewed in the course of the inspection:
o Engineering Evaluation for St.
Lucie Unit 2 Inaccurate Feedwater RTDs (JPN-PSL-SEIS-92-018),
o LER 50-389/92-08, Digital Data Process System Calorimetric Error due to Instrument Calibration Error.
o NUREG/CR-5560, Aging of Nuclear Plant Resistance Temperature Detectors.
o Licensee-prepared note books for Units 1 and 2, which contained the maintenance and calibration history for all instruments and detectors used in either the manual calorimetric calculation or that performed by the DDPS.
c.
Potential for Earlier Detection As confirmed in NUREG/CR-5560 and by the licensee's experience in removing RTDs from the thermal wells, it is not industry practice to remove the RTDs for periodic calibration because of the significant risk that the detectors will be damaged during removal.
However, the remainder of the temperature measuring instrument circuits were calibrated with a sesquiannual periodicity.
Standard industry practice is to compare the output of each RTD with the others in a similar environment.
Such intercomparison worked once, in early 1985, to identify a problem with RTD A03, which was replaced.
But, intercomparison is not useful for coherent changes in the performance of the detectors being compared.
The figure on the following page shows the indicated full-power temperatures for each of the six RTDs for the period from February 1984 to September 1992.
The early failure of A03 is obvious in the plot.
After the replacement of A03, the general trend of RTDs is toward increasing indicated temperature.
With a 10'F intercomparison criterion, the trend was not identified.
Industry experience with both nuclear grade and commercial grade RTDs was summarized in NUREG/CR-5560.
No experience with coherent drift was identified in that report, and most of the failure mechanisms
Report Details identified therein led to temperature indications lower than actual rather than higher.
450.0 448.0 446.0 444.0 Feedwater Temperature (')
Indicated at Full Power 442.0 440.0 438.0 438.0 434.0 A01
-- A02
~-* A03
-- 801
-- B02 432.0 430.0 Increasing Time The inspectors reviewed the licensee's trending of heat rate to see if that revealed an apparent continuing increase in plant efficiency.
The heat rate was found to be too grossly affected by variations in sea water temperature to resolve the far more subtle effect of a small increase in power level.
The licensee did routinely perform a manual calorimetric as a check on the DDPS calculation.
Feedwater temperature for the manual calculation was obtained from installed thermocouples rather than the RTDs.
The manual calculation lacked the precision to resolve small errors in the DDPS calorimetric calculation.
The acceptance criterion for agreement between the two calculations was 2%.
Report Details C.
The maintenance history of both units gave no indications of generic problems with the RTDs.
The one that failed in 1985 was replaced with a plant spare, which was ordered separately but came from the same manufacturing lot.
One instance of the two calorimetric calculations failing inter-comparison was determined to be Speedomax recorder problem (PW08 6187, 6199). 'hat recorder is used to obtain the feedwater thermocouple reading.
The licensee does not trend the feedwater temperature.
Although an important input to the calorimetric calculation, there is no control function from feedwater temperature RTDs.
An after-the-fact search of NPRDS history showed no similar events of coherent drift of RTDs by the manufacturer of these RTDs; although, other failures were reported.
The inspectors concluded that earlier detection of the drift in RTD outputs was not possible following normal industry and plant surveillance practices.
Hagnitude of the Error Following review of Engineering Evaluation for St.
Lucie Unit
Inaccurate Feedwater RTDs (JPN-PSL-SEIS-92-018)
and discussions with licensee personnel involved in the evaluation, the inspectors concluded that the error was correctly quantified at 1.04%
RTP for operation at a nominal 100% RTP.
The evaluation did not consider normal power variations about the control point.
The inspectors attempted to determine the sum of the typical variations and the RTD calibration error, -but were not quantitatively successful.
Digital data logs of power are maintained for 1-hour averages; so, no record of power variation over a shorter period was available.
The hourly averages rarely exceeded 100. 1% RTP, and usually, the hourly variations were less than 0. 1% RTP.
Unit 2 was in a ma'intenance outage during this inspection.
Observations in the Unit 1 control room indicated that the power calculated by the DDPS and shown on a continuously updated digital display did not vary by more than +0. 1% RTP from the 100%
RTP setpoint, Discussions with operators at both units confirmed that the observed power variation was typical of all operation.
Accident analyses assume operation at 102% of RTP.
The RTD-induced error plus the system uncertainty of 0.95% calculated by the reactor vendor was barely less than 2%.
Power variations may have led to operation above 102% RTP.
However, conservative errors in feedwater flow from venturi fouling probably kept the unit within the bounds of the accident analysis at all times.
Experience at other facilities
Report Details indicates that the conservative error in feedwater flow may be as great as 2% RTP.
d.
First Stage Corrective Action The licensee had implemented the corrective actions described below on both units by the time of this inspection:
(1)
The believable value comparison of the RTD outputs was reduced to 5'F vice the 10'F used earlier.
(2)
The DDPS will generate an alarm whenever any feedwater RTD output exceeds 437'F.
e.
Root Cause Analysis Inspection of the environment of the feedwater RTDs showed it to be benign, with little potential for damage from unrelated work activi-ties.
The similarly installed RTDs in Unit 1 have shown no problems.
Post mortem examinations of the failed RTDs are being conducted, in parallel, by both the vendor and the licensee.
The inspectors observed some of the activities in the licensee's laboratory, includ-ing the use of an electron microscope to view sensor wire.
Investi-gation techniques were developed using an uninstalled plant spare.
Some of the early, but unevaluated, observations are that the Al,0~
insulation in the installed RTD was grayish in color, but that in the spare RTD was white (chemical analysis will be done later),
and that resistance measurements on the sensor end only found the installed RTD to be
ohms greater than the spare RTD.
Although extensive and intensive effort is being given this analysis, the root cause may not be identified.
Additional Corrective Action 9,
Additional corrective actions may be prudent, and the licensee is considering the following actions:
(1)
Perform a precision heat balance at. the beginning of the cycle using specially installed instrumentation.
(2)
Replace one RTD per cycle for calibration purposes.
Enforcement Action Operation in excess of 101%
RTP for an extended period of time meets internal NRC management guidance for issuing a violation.
However, the circumstances leading to this overpower operation are different
Report Details from those considered when that guidance was issued.
Thus, with the concurrence of Region II management, the inspectors did not propose a
violation.
The inspect'ors did express one subject of concern to licensee management, and that was related to the lack of narrative detail in the licensee's official records.
The operations log books for November 5 and 6, 1992, contain no mention of the need to reduce power or the amount of power reduction in response to the indicated power increases when the new RTDs were placed in service.
Some of the PWOs examined did not contain enough detail to determine what had failed or the cause of the failure.
3.
Action on Previous Inspection Findings (92702)
(Closed)
URI 50-389/92-21-06:
Potential operation above the licensed power level.
Operation above the licensed power level, but below the power level assumed in safety analyses, was confirmed.
No enforcement action was undertaken.
4.
Exit Interview The inspection sc 5.
Acronyms and Initialisms ope and results were summarized on February 12, 1993, with those persons indicated in paragraph 1.
The inspectors described the areas inspected and discussed in detail the inspection results and conclusions.
Proprietary information is not contained in this report.
Dissenting comments were not received from the licensee.
,
RTP URI digital data processing system nuclear plant reliability data system plant work order resistance temperature detector rated thermal power unresolved item