IR 05000213/1986020

Insp Rept 50-213/86-20 on 860708-0814.Violations Noted: Unauthorized Operation of Containment Isolation Valves for Surveillance Testing & Failure to Inspect Portable Fire Extinguishers in Switchgear Room

ML20214L347
Person / Time
Site:	Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png
Issue date:	08/25/1986
From:	Mccabe E NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
To:
Shared Package
ML20214L329	List: IR 05000213/1986020 ML20214L326 ML20214L340
References
50-213-86-20, NUDOCS 8609100101
Download: ML20214L347 (15)

Text

.

.

U.S. NUCLEAR REGULATORY COMMISSION

REGION I

Report No.

50-213/86-20 Docket No.

50-213 License No.

DPR-61 Licensee:

Connecticut Yankee Atomic Power Company P. O. Box 270 Hartford, CT 06101 Facility:

Haddam Neck Plant, Haddam, Connecticut Inspection at: Haddam Neck Plant Inspection conducted:

July 8 through August 14, 1986 Inspectors:

Stephen Pindale, Resident Inspector Paul D. Swetland, Senior Resident Inspector A. G. Krasopoulos, Reactor Engineer Approved by:

&

Or~1e.

Pl25EC E. C. McCabe, Chief, Reactor Projects Section 3B Date Summary:

Areas Inspected:This was a routine resident inspection (141.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />) of the fol-lowing areas:

plant operations, radiation protection, physical security, fire protection, maintenance, surveillance testing and licensee events.

'

Results: Two violations were identified for unauthorized operation of containment isolation valves for surveillance testing (Detail 4.3) and failure to inspect port-able fire extinguishers in the switchgear room (Detail 2.4).

An unresolved item was identified concerning the adequacy of licensee review of the safety implica-tions of installation of a heavy, unrestrained, breathing air bottle cart in the control room (Detail 2.3).

Another unresolved item was identified regarding the adequacy of reactor coolant system 3-loop flow (Detail 4.4).

Continued plant operation with equipment not fully operational was identified when operators did not take timely follow-up on indication of erratic charging pump operation (Detail 4.2).

Also, repairs to the nuclear instrumentation system resulted in indication of excessive axial flux offset at 100% power.

In this case, it appears that the positioning of the flux sensors should have been assessed before 100% power was achieved (Detail 4.8).

8609100101 860829 PDR ADOCK 05000213 G

PDR

.

.

TABLE OF CONTENTS Page 1.

Summary of Facility Activities.......................................

2.

Review of Plant Operations...........................................

3.

Observation of Maintenance and Surveillance Testing..................

4.

Followup on Events Occurring During the Inspection...................

5.

Review of Periodic and Special Reports...............................

6.

Unresolved Items.....................................................

7.

Exit Interview.......................................................

i

,

!

!

!

l

I s

!

!

,

.

.

DETAILS 1.

Summary of Facility Activities At the beginning of the inspection period on July 8, 1986, the plant was operating at 100% power. On July 11, a load reduction was initiated due to oscillation (pressure and amperage) of the "A" charging pump.

The Technical Specification Limiting Condition for Operation for the charging pump could not be met and the unit was placed in a hot shutdown condition (mode 4) for pump. repair. On July 15, the licensee placed the unit in a cold shutdown (mode 5) to accomplish steam generator (SG) tube repairs.

The resulting 3-week unplanned shutdown came about as a result of evaluation of SG eddy cur-rent' test data (performed during the 1986 refueling outage).

On July 23, during the outage, an above limit radiation exposure of a contractor SG worker occurred. The worker received about 10% more quarterly exposure than NRC regulations permit.

(NRC Region I Specialist Inspection Report 50-213/86-22 addresses this item.) On July 26, the outage was completed.

Full power was achieved on August 6.

A subsequent load reduction to 80% power was initiated to perform in-core flux mapping to resolve a discrepancy in axial offset (AO)

'

indications.

The unit returned to 100% power on August 9 and remained at 100%

power through the end of the inspection period.

2.

Review of Plant Operations The inspector observed plant operation during regular tours of the following plant areas:

Control Room Security Building

--

--

Primary Auxiliary Building Fence Line (Protected Area)

--

--

Vital Switchgear Room Yard Areas

--

--

Diesel Generator Rooms Turbine Building

--

--

Control Point Intake Structure and Pump

--

--

Containment Building Building Control room instruments were observed for correlation between channels and for conformance with Technical Specification requirements.

The inspector observed varicus alarm conditions which had been received and acknowledged.

Operator awareness and response to these conditions were reviewed.

Control room and shift manning were compared to regulatory requirements.

Posting and control of radiation and high radiation areas was inspected.

Compliance with Radiation Work Permits and use of appropriate personnel monitoring devices were checked.

Plant housekeeping controls were observed, including control and storage of flammable material and other potential safety hazards.

The inspector also examined the condition of various fire protection systems.

During plant tours, logs and records were reviewed to determine if entries were properly made and communicated equipment status / deficiencies.

These records included operating logs, turnover sheets, tagout and jumper logs, process computer printouts, and Plant Information Reports.

The inspector observed selected aspects of plant security including access control, physical barriers, and personnel monitorin.

.

2.1 The inspector reviewed the status of constantly annunciated control room alarms.

Previously identified alarms (NRC inspection report 50-213/86-16)

which remain annunciated are the pressurizer relief tank temperature and several electrical distribution system high bus voltage alarms.

During this inspection period, there were two additional continuous alarms an-nunciated in the control room: reactor coolant pump seal water high tem-perature at seal and high loop differential temperature.

The reactor coolant pump (RCP) seal water high temperature alarm was constantly an-nunciated due to an instrumentation problem with the No.3 RCP lower bearing water temperature monitoring channel.

The alarm was previously annunciated before the steam generator tube plugging outage, when the resistance temperature detector (RTD) for the No.3 RCP lower bearing water temperature was replaced.

However, this alarm came in again.

In-strument maintenance is continuing to troubleshoot the channel.

Reactor coolant pump seal water return flow and seal differential pressure are being monitored to v3rify that the high reading is due to instrumentation error as opposed to being a bona fide indication. On August 12, the an-

,

nunciator was bypassed to eliminate the possibility of masking actual or new plant conditions.

No unacceptable conditions were identified by inspector review of this alarm.

The inspector questioned the presence of the high loop differential tem-perature (delta-T) alarm. While operators knew the alarm setpoint and the current delta-T readings, the reason for the high delta-T annunciator was not known.

The licensee stated that the delta-T alarm setpoints were raised for cycle 14, due to the new reactor coolant system (RCS) RTOs, to 50(+0,-2) degrees F.

Delta-T readings indicated approximately 48-49 degrees F.

There is no Technical Specification upper limit on loop delta-T.

The variable low pressure scram (VLPS) uses Tave and delta-T inputs to calculate an RCS pressure at which a reactor trip will occur (to prevent departure from nucleate boiling).

The purpose of the high delta-T alarm is to warn operators that, while significant margin exists to the VLPS reactor trip, the cause of the high delta-T should be inves-tigated.

The inspector questioned whether the delta-T alarm was related to the additional steam generator tube plugging performed since the delta-T alarm was not annunciated during operation prior to the shutdown.

The licensee indi ated that this was not the case, and that the delta-T indicator configuration was the cause.

However, the licensee performed a four loop flow test (SUR 5.3-45) on August 12 to determine RCS flow following the SG tube repair and to aid in troubleshooting this problem.

No unacceptable conditions were identified by the licensee or by the inspector's review of this alarm condition.

This matter is further ad-dressed in Detail 4.4 of this report, and the status of unnecessary alarms will continue to be reviewed during routine inspections.

2.2 The inspector attended a Plant Operations Review Committee (PORC) meeting on July 31, 1986.

Technical specification 6.5 requirements for required member attendance were met.

The meeting agenda included procedural changes, proposed changes to the Technical Specifications, and field changes to design change packages.

The meeting was characterized by

,

I

_

.

.

frank discussions and questioning of the proposed changes.

In particular, PORC consideration was given to assure clarity and consistency among procedures.

Items for which adequate review time was not available were postponed to allow committee members time to review and comment.

Dis-senting opinions were encouraged.

The inspector had no further comments.

2.3 During an NRC team inspection of the licensee's fire protection program in June 1986, the inspector observed a cart with breathing air bottles (compressed gas cylinders) within the control room confines. The cart is made with swivel wheels, weighs more than 1500 pounds and was found unrestrained behind the emergency diesel generator control cabinets.

This installation is of concern because the unconstrained mass of the cart and bottles are a potential hazard to control room equipment and personnel.

The inspector asked the licensee for the safety evaluation performed in accordance with 10 CFR 50.59, and the licensee stated that no such analysis was performed.

There was no available written safety evaluation documenting why this installation does not pose an unreviewed safety question.

The licensee subsequently moved the breathing air bottles to a remote location, within the confines of the control room, and the cart was provided with curved wooden blocks at the wheels to restrict motion.

This item is unresolved pending further NRC and licen-see assessment of safety implications of this case (UNR 213/86-20-01).

2.4 On August 8,1986, while performing a routine plant walkthrough, the in-spector observed that several portable fire extinguishers (PFEs) were not inspected in accordance with PMP 9.5-120 (Portable Fire Extinguisher Inspection).

PMP 9.5-120 requires that all station PFEs be inspected monthly. All PFEs located in the switchgear room were last inspected, as logged on the PFE monthly inspection status tag, on June 26, 1986.

Failure to inspect the switchgear room PFEs in July 1986 in accordance with PMP 9.5-120 is a violation (VIO 213/86-20-02).

2.5 The monthly inspection status of the PFEs in the control room indicated that the last three PFE inspections were performed on 5/27/86, 6/5/86, and 7/24/86.

Westinghouse Standard TSs allow a 25% maximum extension of any surveillance interval, and define monthly as at least once per 31 days.

However, Haddam Neck TS 1.21 (Frequency), defines monthly as

"At least once per month".

Nonetheless, performance of inspections at the end of one month and again at the beginning of the next month does not appear to meet the intent of the monthly inspection frequency.

From the three dates on the monthly inpection status tags on the control room PFEs, well over 30 days passed between the June and July inspections.

The inspector brought this discrepancy to licensee management attention.

Licensee action on this matter will be reviewed by NRC during a subse-quent routine inspection.

3.

Observation of Maintenance and Surveillance Testing The inspector observed various maintenance and problem investigation activi-ties for compliance with requirements and applicable codes and standards, QA/QC involvement, safety tags, equipment alignment and use of jumpers, per-

.

.

sonnel qualifications, radiological controls, fire protection, retest, and reportability. Also, the inspector witnessed selected surveillance tests to determine whether properly approved procedures were in use, test instrumenta-tion was properly calibrated and used, technical specifications were satisfied, testing was performed by qualified personnel, procedure details were adequate, and test results satisfied acceptance criteria or were properly dispositioned.

The following activities were reviewed:

--

Maintenance of P-18-1A, P-18-1B Charging Pump (CMP 8.5-141)

Component Cooling Water to Neutron Shield Tank Cooler, P-60 (SUR 5.7-93)

--

Operation of the Flux Map System (ENG 1.7-28)

--

Steam Generator Tube Repair (VP-200)

--

NIS Coaxial Extension Cable Replacement (PMP 9.2-9)

--

Four Loop Reactor Coolant Flow Measurement (SUR 5.3-45)

--

3.1 During the 1986 refueling outage, component cooling water (CCW) system containment isolation check valves failed their respective type C leak rate tests due to silt build-up in the CCW system.

Following installa-tion of a slip-stream filter in the CCW system and a subsequent system realignment which allowed silt accumulation in the CCW system, the valves again failed the test.

The valves involved are CC-CV-885 (CCW supply to neutron shield tank heat exchanger-SUR 5.7-93) and CC-CV-721 (CCW supply to reactor coolant pump seal thermal barrier-SUR 5.7-51).

Fol-lowing cleaning of the valves, both met their respective acceptance cri-teria.

To assure that the CCW valves remained operable throughout the operating cycle, the licensee committed to test these valves during plant shutdowns.

During the outage in July 1986, the licensee performed the committed leakrate tests. On July 19, 1986, the licensee performed leakrate testing on CH-RV-332 (SUR 5.7-60) in addition to CC-CV-721 and CC-CV-885.

Although CC-CV-885 passed as witnessed by the inspector, both CH-RV-332 and CC-CV-721 failed to meet their respective inservice in-spection (ISI) acceptance criteria.

Together, they exceeded the Techni-cal Specification 4.4 (Containment Testing) limit for containment pene-tration isolation valve leakage.

The licensee made the necessary NRC notification via ENS.

Subsequent investigation revealed that the cause of both test failures was backflow through test boundary valves.

Upon ensuring that all test boundary valves were fully closed, the valves were retested satisfactorily without any valve rework.

The licensee has again committed to perform leakage rate testing of CC-CV-721 and CC-CV-885 during the next cold shutdown, thereby assuring valve operability and slip-stream filter effectiveness.

These valves, and others as deemed appropriate by the licensee, will be placed on the plant shutdown work list.

The conduct and results of these tests will be followed by the resident inspectors during a subsequent routine inspection and for re-solution of unresolved item UNR 213/86-06-0.

i

<

.

.

4.

Followup on Events Occurring During the Inspection 4.1 ~ Licensee Event Reports (LERs)

The following LERs were reviewed for clarity, accuracy of the description of cause, and adequacy of corrective action.

The inspector determined whether further information was required and whether there were generic implications. The inspector also verified that the reporting require-ments of 10 CFR 50.73 and Station Administrative and Operating Procedures had been met, that appropriate corrective action had been taken, and that the continued operation of the facility was conducted within Technical Specification Limits.

86-31 -- Failure of the Safety Injection Recirculation Isolation Valve to the RWST (Detailed in NRC Inspection Report 50-213/86-16)

86-32 -- Drift of Dropped Rod Load Runback Setpoint of Nuclear Instru-mentation 86-34 -- Three Loop Flowrate Measurement (Detail 4.4)

86-35 -- Steam Generator Tube Eddy Current Testing Undefined Signals (Detail 4.5)

86-36 -- 1A Charging Pump Shaft Fallare (Detail 4.2)

4.2 On July 7, 1986, the plant was operating at 100% power when the operating charging pump (A) pressure and amperage indications became erratic, but remained within specification.

At 4:00 p.m., the "A" charging pump was secured and the "B" charging pump was started.

Charging flow and related system parameters returned to normal and plant operations were stabilized.

Concurrent with shifting the charging pumps, the positive displacement metering pump relief valve was isolated at 4:15 p.m., as leakage past that valve was evident.

It was initially suspected that leakage past the metering pump relief valve may have caused the erratic "A" charging pump indications; the metering pump relief valve is connected to the chemical and volume control system (CVCS) common discharge piping.

At 6:30 p.m.,

test operation of the "A" charging pump again revealed pres-sure and amperage surges.

The metering pump was subsequently isolated for maintenance on the relief valve.

Inspection of the "A" charging pump was planned for the following morning, with both the "A" and "B" charging pumps considered operable since indications were within specification.

On July 8,1986, at approximately 11:10 a.m., "A" pump inspection re-vealed that the pump shaft had sheared just outside the pump casing.

While the pump would still run and charge water to the system, it was declared inoperabl.-.

,

.

.--- __.

-

-

.-

(

.

,

.

'

4 i

Technical Specification (TS) 3.5, related to CVCS, requires either two charging pumps or one metering pump and one charging pump to be operable while the reactor is critical.

Therefore, with only one charging pump operable, the plant immediately initiated a controlled load reduction as required by 10 CFR 50.36.

The investigation of the metering pump relief valve leakage determined

'

.

that the leakage past the valve was within limits, but the relief valve

'

setting was noted to be incorrect.

The improper relief valve setting was corrected before the metering pump and relief valve were returned

to service. Maintenance on the metering pump relief valve was completed at 12:00 noon.

With the metering pump and associated piping operable,

,

the load reduction was terminated at approximately 90% power.

The unit was then returned to 100% power.

Administrative TS 3.6.(Core Cooling Systems) requires that two charging l

pumps be operable in modes.1-3. The action statement for the associated i

limiting condition for operation allows continued operation for 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> i

with one charging pump operable.

On July 11, 1986, it was evident that the "A" charging pump could not be repaired within the time allotted;

'

therefore, a controlled plant shutdown was initiated at 3:00 a.m. on July 11, and the necessary notifications were made to NRC and the state.

The

"B" charging pump remained operable throughout the event, with the metering pump also operable approximately one hour after the "A" charging pump problem was declared inoperable. After reaching hot shutdown, a

corporate decision was made on July 15, 1986, to place the unit in a cold

!

shutdown condition for steam generator tube repair (Detail 4.5).

The inspector questioned the licensee why the metering pump was tagged out of service on July 7 at 6:53 p.m., while maintenance was not planned until the following day.

This question included concerns about tagging out a second pump (metering) when the operability of the first pump ("A" i

charging) may have been questionable.

The licensee stated that they be-lieved the "A" charging to be operable at the time because cf attribution-of the oscillations to the metering pump relief valve.

However, the "A"

,

charging pump was test run at 6:30 p.m., after the relief valve isolation, and erratic operation still occurred.

Further, no oscillations occurred with the "B" pump operating and the "A" pump secured.

The suspect "A" charging pump was not further addressed for approximately 15 hours1.736111e-4 days <br />0.00417 hours <br />2.480159e-5 weeks <br />5.7075e-6 months <br />.

Prompt investigation could have identified the problem earlier. The in-spector also questioned whether the metering pump should have been tagged out of service when erratic "A" charging pump indications existed.

Previous charging pump failures of a similar nature have occurred.

On May 1, 1985, a charging pump shaft failure occurred at a point on the i

shaft on the threaded portion on which a locking nut is installed to i

compress the 13 stages of the pump.

The July 8, 1986 failure was also

!

in the threaded portion, with the fracture point on a different thread along the shaft.

Causal indications of both failures, upon inspection,

,

,

i

-

- -

-..,

.

.-

.

- -

- -

-

-

--

-,

~~

-

,

....

.

-

%

'

y

.

F

.

- !

were of high cyclic fatigue.

Inspector concerns on this matter include identification of the root cause of.the failures and establishment of measures to prevent-further recurrence.

A representative from the pump manufacturer ( acific) was sent to the site to assist the plant staff in identification of the cause of the

~

event and to establish measures to be taken;during assembly and instal-o lation of the repaired pump.

The licensee had a spare pump shaft and

'

impellers readily available on site.

No significant damage was incurred to the 13 stage pump.

The licensee and company representative formulated s

additional precautionary steps in the pump assembly and installation'

procedure. These included dye penetrant testing of the diffusers and wear rings, blue checks on the threaded portion of-the threaded section of the shaft (to ensure proper-locknut seating),1 hand-fitting of.all mating surfaces which transmit forces through the. shaft (to reduce the-likelihood of inducing high stress regions along the shaft), revised balancing processes for rotating components, and greater attention to detail throughout the reassembly process.

During a torque test of the shaft and locking nut, galling of the rotating element and locking nut occurred and the locking nut was' cut off.

Blue checks were again per-

formed as was a revised torque test.

The pump was satisfactorily re-

'

paired and tested on July 27, 1986.

Additional measures taken by the licensee include submittal of a project assignment (PA) request to investigate the pump design as-applicable to its. intended use and to evaluate the history of the pump's previous operational data.

The licensee hopes to gain information on whether the

"

j'

processes used for the charging pump reassembly and installation were adequate to prevent recurrence.

The PA request also includes a proposal

to investigate whether sufficient margin exists in the current pump de-sign.

With respect to the improper relief valve setpoint, licensee' review of the procedure used to set the relief valve indicated that the procedure

'

did not require a setpoint verification with the-setpoint control log.-

'

The specific procedure (PMP 9.5-114) is to be changed'to address this

concern.

Additionally, to address whether this is a generic concern, all safety-related relief valve setpoints will be verified.

The comple-

,

tion of this review, the PA, and in plant evaluation'results will be reviewed under the routine inspection program.

4.3 On July 8, 1986, the licensee identified that, during surveillance test SUR 5.1-4, four manual containment isolation valves were opened for the i

l test-(SI-V-863 A/B/C/D). These four 1/2" manual loop' recirculation valves are normally locked closed during power operation, and are located i

outside containment.

They serve as the second. isolation boundary in the containment isolation system.

The licensee's~ investigation was to de-

,

termine the reasons for opening the HPSI system common recirculation valve (SI-HCV-1881) whose failure led to the loss of both HPSI trains, I

and was addressed in NRC inspection report 50-213/86-16.

Upon discovery

.

L

T'

.

.

of the opening of.the four isolation valves, it was determined that the valves have been opened monthly for this surveillance test since 1966.

The primary reason for opening the valves is to equalize boron concen-tration and prevent boron stratification in the HPSI discharge piping.

The cpening of SI-863A/B/C/D violates Containment Integrity (Technical-Specification 3.11) as defined in TS 1.8.

Immediate corrective actions-include providing administrative controls to maintain SI-V-863 A/B/C/D valves locked closed until procedural. changes have been made to comply-with TS. The inspector questioned whether administratively locking 3'

a closed the valves had any impact on HPSI system design bases.

On a steam line break accident,_the integrated Emergency Core Cooling System pro-vides water with specified. concentrations of boric acid in solution to

'

.

the reactor coolant system. The licensee stated that, above 32 degrees

'

F, the boric acid would not precipitate out of solution.

The licensee t

also stated that the design of the HPSI system is such that all outside piping is insulated and electrically heat traced, further reducing the likelihood of borated water delivery problems.

Similar TS violations have occurred previously on other containment

isolation valves: operation of manual isolation valve SA-V-413 was iden-

,

tified in NRC inspection report 50-213/86-08; and post-accident sampling system (PASS) valve operations were identified in NRC inspection report

50-213/84-14.

As indicated by the repetitive nature of the violations, l

the individual corrective actions have not resulted in correction of the i

underlying cause(s).

This concern was discussed with licensee management on August 8, 1986.

The licensee included, in the long term corrective

,

l action program, a detailed review of all non-automatic containment iso-lation valves and of applicable procedures in which valves are directed

' Y to be manipulated. The current Haddam Neck TSs do not allow for manipu-lation of these containment isolation valves when the reactor is critical.

.

As indicated by the final corrective action for the PASS valve problem, I

where good cause is shown, TS amendments can be granted to allow manipu-E 1ation of' containment isolation valves under administrative control.

B The inspector will review the licensee's corrective action plan and its implementation'at a later date.

The licensee's failure to adhere to current TS limits regarding SI-V-863A/S/C/D constitutes a violation (VIO 213/86-20-03).

i 4.4 On July 11, 1986, while in Mode 3, the licensee identified problems with

'

'

the 3-loop reactor coolant system (RCS) flow test (ENG 1.7-48) which was performed on June 24, 1986.

The current safety analyses assume a minimum

,.

f 3-loop flow rate of 207,000 gpm.

The minimum flow measured during the

L f

test was 202,520 gpm, about 2.4 percent less than the safety analysis minimum.

Upon evaluation of the test results, the licensee made the

,

g necessary notification to NRC and the State.

Currently, there are no Technical Specifications (TS) for the required 3-loop flow rate.

For the short term, the licensee prohibited 3-loop operation in Mode 1 via temporary procedure changes86-471 and 86-472.

]

$

-

.

i,

-

_ _ _,..

..

.

.. -

.--

.

_

- -

r

.

9,

The licensee also performed a 4-loop flow test (SUR 5.3-45) on May 29, 1986.

These 3-loop and 4-loop RCS flow rate tests were the first actual flow tests performed since initial startup testing.

The acceptance cri-

-terion for RCS. flow rate, based upon mode 1 steady state operation with-four reactor coolant loops operating, is 257,000 gpm. 'TS 3.20 (RCS Flow, Temperature:and Pressure) specifies this minimum 4-loop RCS flow rate for mode 1.

The minimum 4-loop flow rate, as measured in SUR 5.3-45, was 267,067 gpm.

This value was reduced to 259,154 gpm when a 2.963 percent uncertainty factor was incorporated.

That yields a margin of 2154 gpm or 0.8 percent of the TS 3.20 limit.

The present licensee established routine frequency for RCS flow testing is once each refueling.

Since 126 Steam Generator (SG) tubes were

,

plugged during the July 1986 outage, the licensee performed an additional 4-loop flow test on August 12, 1986 after plant startup from the tube-plugging outage.

The test results indicated a 0.5 increase in flow rate during preliminary licensee review.

The licensee committed to re-analyze 3-loop flow to determine if suffi-cient margin exists in the safety analyses to justify 3-loop operation.

Also, the 4-loop flow test final evaluation is to include a determination of the cause of the high differential temperature _(delta-T) alarm which was annunciated following plant startup from the outage (Detail 2.1).

This item remains unresolved pending these actions (UNR 213/86-20-04).

4.5 During the 1986 refueling outage, the licensee performed eddy current testing (ECT) of all Steam Generator (SG) tubes, except for 4 tubes in SG #4.

ECT data identified many tubes (575) with undefined signals (UDSs)

in the tube sheet-area.

A SG U-tube was removed from SG #2 for destruc-tive testing to aid in characterization of the SG tube UDSs. The de-structive test indicated that the UDSs were multiple tube cracks initi-ated inside the tube.

The cause was attributed to primary side stress corrosion.

There were a total of 575 SG tubes with UDS indications, with the crack indications ranging from 20-100% through wall.

The licensee notified the resident inspectors and NRC Licensing of the preliminary findings on June 19, 1986.

The licensee submitted a letter to NRC, dated June 20, 1986, which documented the safety basis for continued operation, concluding that the potential flaws did not represent an undue risk to the public health and safety.

A June 26, 1986 NRC safety evaluation report (SER) concurred with the licensee's position on maintenance of tube structural integrity.

The SER~ established a corrective action plan submittal date of September 30, 1986.

During further licensee review and evaluation of the preliminary ECT findings, 35 of the 575 tubes were assessed as acceptable.

Therefore, 540 SG tubes remained with UDSs ranging from 20-100% through wall.

The licensee performed analyses from which interim SG tube structural ac-ceptance criteria were established.

The analyses assumed that the flaws were in the rolled region of the tubes.

The analyses showed that one inch of sound tube-to-tubesheet roll provides the frictional resistance

f

.

.

required to withstand normal operating, test, and postulated accident loads.

Further analysis demonstrated that, with no defect in the uppermost one inch of the rolled region, any crack below the one-inch point would result in a leakage well below the Technical Specification (TS) limits and structural integrity would be maintained. Therefore, the licensee's proposed interim structural acceptance criteria were 1) no UDSs in the uppermost one inch of hard roll; and 2) any size UDS defect is acceptable below the uppermost one inch of hard roll.

Based upon this information and review of the ECT results, the licensee developed a list of 119 SG tubes that did not satisfy the criteria.

On July 15, 1986, licensee corporate management decided to place the unit in cold shutdown to plug these 119 tubes.

Three of four SGs were affected (SG #2,#3,#4).

The NRC was informed verbally of this action.

Formal notification was made by letter dated July 24, 1986.

The licensee also decided to plug one SG #2 tube with a 55% through wall defect that was inadvertantly left unplugged during the 1986 refueling outage.

The licensee also examined previous ECT results of the four SG #4 tubes which were not inspected during the 1986 refueling outage.

Review and evaluation of previous ECT results indicated that there were no UDSs present.

The licensee has determined that UDSs do not readily propagate and therefore decided not to plug these 4 tubes.

The final plugging list included 120 tubes.

The tubes were plugged in accordance with VP-200.

System hydrostatic tests were then performed to verify the adequacy and leak tightness of the mechanical tube plugs.

The option existed to remove the mechanical plugs and install welded plugs if tube leakage occurred following initial tube repair.

No welded plugs were required.

Six tubes not on the plugging list were also plugged.

Two tubes were inadvertently plugged in SG #4 due to vendor error.

Another two tubes, in SG #2, were plugged because of measurable leakage during the hydrostatic test.

A third SG

1. 2 tube were plugged because of boric acid deposits around and within the tube end.

(Additional ECT on these three tubes indicated tube cracks in the roll transition region of one tube.

The other two tubes did not show any crack indications.) A fourth tube in SG #2 was plugged because special additional ECT during the 1986 refueling outage indicated a flaw in the roll transition zone.

Other 1986 Refueling Outage ECT results were sampled.

No further discrepancies were found.

All SG tube repairs and testing were completed on July 26, 1986.

The NRC reviewed the information and analyses provided in the licensee's

July 24,1986 submittal in support of the interim acceptance criteria.

The NRC subsequently issued a safety evaluation report (SER), dated July 30, 1986, which concluded that there was reasonable assurance that the potential flaws in the rolled region of the 540 SG tubes would not result i

in catastrophic failure in the near term and that any propagation of l

existing flaws will be revealed by an increased primary-to-secondary leakage rate, which is limited by TS.

(As of the end of the inspection l

period, daily measurements indicated no primary to secondary system

!

leakage.) The SER further specified that the licensee complete the UDS et decion to specify final acceptance criteria by the previously speci-

'

,

- - _ _

- _ _ _ _ _ _ _..

-

._

r

.

.

fied September 30, 1986 submittal date, and that a license amendment for the appropriate final acceptance criteria be submitted by October 31, 1986.

4.6 On July 17, 1986, the inspector was notified of a potential problem with the power operated relief valves (PORVs). The PORVs are air-operated and function to limit reactor coolant system (RCS) pressure transients and limit actuation of the pressurizer code safety valves.

The automatic lift setpoint of the two air-operated PORVs is 2270 psig.

A high pres-surizer pressure reactor trip occurs at 2300 psig, and the three pres-surizer code safety valves lift at 2485, 2535, and 2585 psig, respec-tively.

In 1977, the PORVs were replaced with new valves as a result of seat leakage and operability problems.

The new PORVs were initially procured for Millstone Point Unit 3 and have approximately three times more relieving capacity than the original PORVs.

The 2270 psig setpoint did not change with the PORV replacement.

The associated plant design change request (PDCR), No. 248, included a safety evaluation which de-termined that the change could be safely implemented without adversely affecting plant design bases.

On July 17, 1986, the licensee issued a plant information report which identified that, during a loss of load (LOL) transient, the current PORV setpoint may allow liquid to be discharged from the pressurizer code safety valves.

The current LOL transient analysis assumes a high pres-surizer pressure reactor trip to occur within 7 seconds of the transient, with credit taken for the operabiljgy o( thg N ig; g M d g;jg, The inspector expressed concern about using the 2270 psig setpoint with the higher capacity PORVs.

Doing so may delay the high pressurizer pressure reactor trip, reducing the margin to departure from nucleate boiling (DNB).

In response, the licensee and NRC conducte'd'a conference call regarding the setpoint.

NRC questioned why the setpoint should not be changed prior to plant startup, since the higher capacity PORVs, when used with the original setpoint, seem to invalidate the facility de-scription and safety analysis (FDSA) chapter 10 assumptions.

The licensee acknowledged that the PORV setpoints may need to be raised to a value greater than the 2300 psig reactor trip but did not consider it necessary to change them before plant startup.

They further stated that the problem was discovered during the FDSA Chapter 10 reanalysis, during the 1986 refueling outage, and that additional time was required to establish, through analysis, the appropriate setpoint.

They presented arguments to justify postponing the setpoint change, including: 1) the operating moderator temperature coefficient as defined in the cycle 14 reload package is more negative than that assumed in the FDSA (this represents an increase in DNB margin), and 2) the plant normally operates with the PORV block valves closed.

Since the block valves receive the same signal to open as do the PORVs (2270 psig) and require approximately 15 seconds to fully cpen, the reactor would have reached the 2300 psig reactor trip before the PORVs (1.5 second opening time) would have been

g

.

1,

-

'

able to delay the' trip.

The licensee then committed to perform an in-ternal safety evaluation for justification of continued operation.

The safety evaluation was reviewed and accepted by PORC on July 31, 1986.

The licensee implemented administrative controls to assure that the block valves remain closed during operation.

The analysis for the setpoint change _is anticipated to be completed by approximately August 19, 1986.

The inspector will follow the implementation of the setpoint change under the routine inspection program.

4.7 On July 25, 1986, it was discovered that the licensee's emergency noti-fication system (ENS) was out of service.

ENS provides communication channels among the control room, technical support center, NRC resident office and NRC Emergency Operations Centers.

All necessary notifications were made.

On July 29, 1986, following troubleshooting of the ENS by the local telephone company, voices could faintly be heard on the system.

The telephone company is currently investigating the problem. While ENS remains out of service, daily functional checks are performed on the commercial telephone service to the NRC Operations Center in lieu of performing the daily ENS phone checks.

This will continue until ENS service is restored.

The inspector will follow the daily commercial phone checks and ENS repair status during routine inspection.

4.8 On August 6, 1986, after reaching 100% power from the steam generator tube plugging outage, the licensee discovered that the electrical cur-rents associated with three of four nuclear instrumentation system (NIS)

axial offset (AO) channels were inconsistent.

The fourth channel was out of service as permitted by plant technical specifications (TS).

The three A0 channels were investigated because the NIS channel No. 32 A0 channel was in an alarmed condition (negative A0).

Prior to the inves-tigation, the licensee suspected that, during excore detector cable re-placement and sleeving, the detectors were not reinserted in their exact position, creating different nuclear flux readings.

Following the main-tenance, the excore detectors were inserted into their respective hori-l zontal chambers.

Since the detectors monitor a limited area of neutron l

flux (due to horizontal vice vertical alignment), a small offset in their l

placement can cause a change in the flux reading currents.

The unit reduced power to 80% to perform in-core flux mapping to determine whether

,

!

A0 conditions actually did exist in the core and to generate updated i

correlations between incore and excore readings.

The licensee completed

'

the incore flux map and associated evaluations and adjustments subse-quently achieved 100% power on August 9.

The incore flux map indicated that the linear heat generation rate were well within acceptable limits.

i In-core flux maps were performed to verify the new correlations at 100%

l power.

No additional problems were encountered.

!

TS 3.18, Power Distribution monitoring and Control, specifies that ex-

!

core-incore correlation be checked after a major change in excore in-l strumentation using results of incore flux measurements.

As the excore l

detector cable replacement was not viewed as a major change in the system, the checks were not performed.

To prevent recurrence, the licensee com-

r

.;

.

mitted, for all maintenance procedures which affect excore instrumenta-tion, to incorporate a notification to reactor engineering (RE) before any maintenance.is performed.

Therefore, RE can make a determination-as to the effects of the maintenance on the incore-excore correlations.

.The inspector had no further questions.

Implementation of the procedural changes will be reviewed during subsequent-routine inspection.

4.9 On August 8, during a severe electrical storm,.the licensee's security computer system experienced lightning induced problems.

The problems included minor loss of system functions for a short period of time.

Im-mediate corrective actions satisfactorily compensated for this loss.

Normal functioning was restored within nine minutes.

The NRC resident inspector was. informed at the tima of the event.

The event was input

'

to the Security log system, and was not immediately reportable to NRC.

The inspector had no further questions.

5.

Review of Periodic and Special Reports Upon receipt, periodic and special reports submitted pursuant to Technical Specification 6.9 were reviewed.

This review verified that the reported in-formation was valid and included the NRC required data; that test results and supporting information were consistent with design predictions and performance specifications; and that planned corrective actions were adequate for resolu-tion of the problem.

The inspector also ascertained whether any reported in-formation should be classified as an abnormal occurrence.

The following periodic reports were reviewed:

-- Monthly Operating Report 86-06, plant operations from June 1-30, 1986.

'

No unacceptable conditions were identified.

6.

Unresolved Items Unresolved items are matters about which more information is required in or. der

'

to determine whether they are acceptable.

Unresolved items identified during

]

this inspection are discussed in Paragraphs 2.3 and 4.4.

.

7.

Exit Interview

'

During this inspection, meetings were held with plant management to discuss the findings.

No proprietary information related to this inspection was identified.

i.

f l

l r

.

.-.-.

-. - - -

- - - -. -. -

.

..-. -....

..-... -.--.

-. - -.