RBG-25975, Ro:On 870413,personnel Discovered Core Thermal Power Not Changing Proportionally to Core Flow Changes.Caused by Feedwater Averaging Function of Computer Program Not Updating Points Values.Procedure EDP-CC-0001 Implemented

Ro:On 870413,personnel Discovered Core Thermal Power Not Changing Proportionally to Core Flow Changes.Caused by Feedwater Averaging Function of Computer Program Not Updating Points Values.Procedure EDP-CC-0001 Implemented

ML20214K540
Person / Time
Site:	River Bend
Issue date:	05/19/1987
From:	Booker J GULF STATES UTILITIES CO.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
RBG-25975, NUDOCS 8705280477
Download: ML20214K540 (4)

Text

_ . - ,

GULF STATES UTILITIES COMPANY PoSf OFF3CESOX220 + ST, F ft A N C i S V a L L E . L o u 4 S 1 A N A 7 0 7 7 $

AREA CODE SO4 635 3237 387-4257 May 19, 1987 RBG- 25975 File Nos. G9.5, G9.25.1.4 U. S. Nuclear Regulatory Commission Document Control Desk Washington, D. C. 20555 Gentlemen:

River Bend Station - Unit'l Docket No. 50-458 Per discussions with your staff, please find enclosed a voluntary report describing an event which occurred recently at River Bend Station-Unit i regarding thermal power level. If'you have any questions please contact Mr. E. R. Grant at (504) 381-4145.

Sincerely, gm Rdd hJ.E. Booker Manager-River Bend Oversight River Bend Nuclear Group h[

JEB/TFP/PDG/DAS/ehs Enclosure cc: U. S. Nuclear Regulatory Commission 611 Ryan Plaza Drive; Suite 1000 Arlington, TX 76011 NRC Resident Inspector P. O. Box 1051 St. Francisville, LA 70775

~

8705280477 870519 PDR ADOCK 05000458 S PDR f iY da

INFORMATIONAL REPORT REPORTED CONDITION:

At approximately 1506, on 4/13/87 with the unit in full power operation, program modifications were installed on the Plant Process Computer. Control Room Operatione personnel were verbally made aware of the changes and a Process Computer system restart was initiated to complete the implementation. After the installation, Control Room operators made minor control rod and core flow adjustments to maintain the unit at 100 percent power.

At approximately 1830, Operations personnel discovered that core thermal power, as calculated by Nuclear Steam Supply (NSS) sof tware, was not changing proportionally to core flow changes.

At 1900, at the direction of the Control Operating Foreman (C0F),

Operators reduced reactor power by reducing core flow and the software personnel were called in. Software personnel discovered that the feedwater averaging function of the computer program was not updating the averaging point's values or their current status. At 2015 reactor power was reduced to ensure that thermal power remained below the licensed limit of 2894 MWt, as determined by the Average Power Range Monitor (APRM) readings.

At 2105, the modified program was replaced with the known working version originally installed. To complete the installation, the Process Computer was restarted. At 2109, after system restart, the Feedwater flow average points were updated by the computer program and checked for accuracy against the analog inputs to the averaging function. The Feedwater average values were found to be accurate. At 2200, the problem was considered fully resolved and Operations resumed power ascension to full power.

INVESTIGATION:

A conservative estimation of the peak power and duration of the event in which the reactor thermal power exceeded 100 percent rated, was made based on the main generator output. This worst-case estimation resulted in a peak of 2926 MWt (101.1 percent) based on a generator output of 1000.4 MWe. It was also estimated the time spent above 2894 MWt was approximately 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> and 15 minutes. However, a trace of the APRM readings indicated that this estimation is conservative. Another analysis performed by the GSU Core Analysis Group, based on the APRM readings during the event, determined the peak thermal power to be approximately 2916 Mwt (100.8 percent).

The core thermal parameters at the peak thermal power were evaluated to be acceptable. Ample margin within the Technical Specifications limits was verified to have been maintained during the entire event. The 12-hour " shift average" power for the two ,

shifts, during which the event occurred (from 0600 hours0.00694 days <br />0.167 hours <br />9.920635e-4 weeks <br />2.283e-4 months <br /> to 1800 )

hours for the day shift and from 1800 hours0.0208 days <br />0.5 hours <br />0.00298 weeks <br />6.849e-4 months <br /> to 0600 hours0.00694 days <br />0.167 hours <br />9.920635e-4 weeks <br />2.283e-4 months <br /> for the

night shift), were calculated to be 2894.3 MWt and 2892.8 MWt, respectively.. These values were developed using worst-case methodology which was expected to provide very conservative results. Exceeding the limit by 0.3 MWt during the day shift was negligibly small (0.01 percent) considering the conservatism given in the estimation. In addition, the "24-hour average" power, from the initiation of. the event until the end of the ensuing 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> period, was 2885 MWe, 0.3 percent below the licensed power limit.

Consistent with the guidance given by the NRC and the magnitude and duration of the event, this event has been considered to not be reportable pursuant to the River Bend Station - Unit 1 Operating License. However, this voluntary report is being-submitted for general information.

The Process Computer problem was in the area of calculating the hourly rate-of-change type data points. This rate-of-change function is processed by the Transformed Variables (TRV) program.

This software is included as part of the BOP Information Package.

The rate-of-change function is used primarily for heat-up/ cool-down monitoring and is in no way utilized by any NSS calculation.

A TRV program change was implemented -which was expected to correct a rate-of-change problem in the current version of the software. The change to the TRV program was provided by the software vendor from the source code of another BWR/6 operating plant and was believed to be fully tested and operational. .The change was reviewed prior to implementation. There appeared to be no possibility of system integrity deterioration, especially in the area of NSS calculations. The program to be modified is not included in the NSS package and would not normally be expected to affect the NSS calculations.

Upon initial system check out after the software modification, there did not appear to be any adverse effects on the operation l of the Process Computer. There were no external operating system indications of any program stalls. Data base alsrming was checked for any new " low-confidence" type point alarms which are likely to occur if a change of this type is not operational.

There were no new alarms.

The rate-of-change points were examined. There did not appear to be any problems associated with these specific point types.

Since the plant was at steady state operation, the. rate of change values were not fluctuating, which is functionally correct.-

Subsequent heat balance (0D-3) and core thermal limit (P-1)

computer runs showed no erroneous data, calculation .results, or bad inputs (no new failed sensors). The calculated results of the P-1 .were very much in line.with the current plant operating conditions. .The steady state mode of operation, however, could not have emphasized any problems in this area. Furthermore, it

is not advisable nor prudent to simulate new values for any input into the NSS calculations in order to test for unreasonable results. With the information gathered and the known operating conditions present, the change to the TRV program was believed to be successful.

Upon investigation into the software problem with the TRV program, it was determined that the rate-of-change logic, as modified, was continuously looping through these data point types and not allowing any of the averaging logic to be executed.

Since the looping was not " inhibited", no time-out errors from the operating system indicated a looping problem. The averaging points were not updated with new values since the averaging calculations were not running. Although only two of these averaging points are utilized by NSS calculations, they can affect the calculation results considerably.

It was also found that the other operating BWR/6 plant, which was also using this sof tware modification, had developed an update which eliminated the problem experienced by River Bend Station.

However, GSU was not aware of this modification at that time.

CORRECTIVE ACTION:

The corrective actions taken to prevent recurrence of this type event will include a review of the quality assurance as applied to the Process Computer software. The scope of the review will include software testing, configuration control and documentation standards.

River Bend Procedure EDP-CC-0001, " Software / Database Change Control," has been implemented which requires operations notification prior to implementing program changes. An independent signoff has been included to ensure that on-shift Operations personnel are cognizant of changes made to computer systems that could impact plant operation. Procedure EDP-CC-0001 also includes the Operations Department in the review cycle for proposed software or database changes.

Additionally, CSU is considering having the vendor validate all modifications to the process computer which do not have previously documented validation.

SAFETY ASSESSMENT:

As stated above in the investigation, core thermal parameters at the peak thermal power were evaluated and found to be acceptable.

The Technical Specification thermal limits were maintained.

Therefore, there was no impact on the health and safety of the public as a result of this event.