IR 05000317/1987017

Insp Repts 50-317/87-17 & 50-318/87-19 on 870701-31.No Violations Noted.Major Areas Inspected:Facility Activities, Routine Insps,Operational Events,Nrc Notifications,Maint, Surveillance,Lers,Radiological Controls & Physical Security

ML20237K133
Person / Time
Site:	Calvert Cliffs
Issue date:	08/14/1987
From:	Lester Tripp NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I)
To:
Shared Package
ML20237K099	List: IR 05000317/1987017 ML20237K096
References
50-317-87-17, 50-318-87-19, NUDOCS 8709040210
Download: ML20237K133 (16)

Text

.

.

,

.

.

.

J U. S. NUCLEAR REGULATORY COMMISSION

REGION I

Docket / Report: 50-317/87-17 License: DPR-53 50-318/87-19 DPR-69 Licensee: Baltimore Gas and Electric Company i

Facility: Calvert Cliffs Nuclear Power Plant, Units 1 and 2 Inspection at: Lusby, Maryland '

Dates: July 1-31, 1987 Inspectors: T. Foley, Senior Resident Inspector D. T imble, Reside , Inspector Approved by: / _27Nyff]p)gw 8//4/87 g,t/L3 E. Tripp, CTiipt, ~ Reactor Projects Section 3A date Summary: July 1-31, 19871 nspection I Report 50-317/87-17; 50-318/87-19 Areas Inspected: (1) Facility Activities, (2) Routine Inspections, (3) Operational Events, (4) NRC Notifications - loss of off site power, (5) Maintenance, (6) Sur-veillance, (7) Licensee Event Reports, (8) Radiological Controls, and (9) Physical Securit Inspection Hours totalled 276 hour0.00319 days <br />0.0767 hours <br />4.563492e-4 weeks <br />1.05018e-4 months <br /> No violations were identifie Results: An error was found in the conduct of a facility change resulting in~che improper loading of a low pressure safety injection pump nozzle (Section 2d). An uncontrolled boration event occurred on Unit 1 as a result of operator difficulty in diagnosing the event and weakness in the design of the make up mode selector switch (section 3). A total loss of offsite power event occurred on July 23, 1987 (section 4). Itaproper settings were found for certain reactor protection system coefficients (section 4).

8709040210%,,$$$317 PDR ADOCK PDR G

_ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ - _ _ _ _ - _ _ _ _ _ _ - _ _ _ _ _ _ _ - _ _ - _ _ _ _ _ _ _ - _ - _ _ _ - _ _ - _ ~

_ _

___ _ _ . _

..

....

-

I

.

DETAILS Within .this report period, interviews and discussions were conducted with various licensee personnel including reactor operators, maintenance and surveillance technicians and the licenree's. management staff. Night shift'

inspections were conducted on July 15, 23, and 26. Weekend inspections were performed on July 26, 198 . Summary of Facility Activities Unit I had been operating 36 days since its last shutdown'and continued routine operation until-July 14 at 1:45 p.m. when the unit automatically tripped on high steam generator water level. The event was initiated by' a transient in the feed water heaters and aggravated by the. reactor operator incorrectly diagnosing the resulting reactor transient.as a dilutio Operator action resulted in over-boration and oscillations in steam generator level control which caused the high steam. generator water level trip. The unit returned to power on July 15 and continued routine operation until July 23 when a fault on one of the offsite 500-KV power lines (5052) caused a loss of this line, reltying problems in the switch i yard and subsequently a loss of the other 500-KV offsite power line (5051)

resulting in a total loss of offsite power causing both units.to automatically trip. The licensee declared an " alert" emergency condition which was downgraded to an unusual event after confirmation that power was available on one of the offsite lines. During the return to forced circulation, reactor coolant pump 11A failed to start. It was subsequently determined that another motor would be required. This prolonged the Unit 1 shutdown through the end of the report perio Unit 2 began the period in hot standby while completing testing of reactor coolant pump (RCP) 22B and routine start up surveillance-testing. The unit was brought critical on July 1 and paralleled on July 3, however, the unit tripped on low steam pressure due to the turbine control circuitry accepting 100 MWe instead of the 40 MWe which the circuit normally accepts. The resultant excessive steam demand caused the low steam pressure. Later that day the unit paralleled to the grid and power was held at 70% due to excessive vibration on the main turbine No. 11 bearin A shutdown was subsequently scheduled to add a balance shot after sufficient vibration data was obtained. On July 8, the unit reduced power to Mode 2 and added a balance shot to the tuibine and returned to power operations; however, vibration on the exciter remained hig I Additionally, the reactor coolant gross leakage began trending up towards the Technical Specification limit of 10 gpm. Unit 2 c:ntainment average temperature also became excessively high. The-licensee planned a scheduled shutdown to correct these problems. Outside air temperatures during this time had approached 100 F for several consecutive day On July 17, the total identified and unidentified (gross) leakage exceeded 10 gpm. The licensee shut down the unit to fix the leakage, containment air temperature problem and turbine vibratio By July 19, the unit returned to power operation until the July 23 loss of offsite power event

- - _ _ - _ _ - _ _ - _ _ _ . _ _ _ _ _ _ - n

_ _ _ _ _ - _ - _ .

. .

,

.

.

.. 3 as described above. On July 24, Unit 2 returned to power and continued routine operations throughout the remainder of the perio . Review of Plant Operation'- Routine Inspections Daily Inspection During routine facility tours, the following were checked: manning, l access control, adherence to procedures and limiting conditions for l

'

operation (LCOs), instrumentation, recorder traces, protective systems, centrol rod positions, containment temperature and pressure, control room annunciators, radiation monitors, effluent monitoring, emergency power source operability, control room logs, shift supervisor logs, tagout logs, and operating order No unacceptable conditions were note System Alignment Inspection Operating confirmation was made of selected piping system train Accessible valve positions and status were examined. Power supply and breaker alignment was checked. Visual inspection of major components was performed. Operability of instruments essential to system performance was assessed. Inspections included: -

l --

Unit 2 low pressure safety injection (LPSI) checked on July 9, l 198 Unit 2 component cooling water (CCW) checked on July 13, 198 ..

--

Southern Maryland Electric Cooperative (SMECO) electrical breaker line up on July 27, 198 No unacceptable conditions were noted, Biweekly and Other Inspections During plant tours, the inspector observed shift turnovers; boric acid tank samples and tank levels were compared to the Technical Specifications; and the use of radiation work permits and Health l Physics procedures were reviewed. Area radiation and air monitor use i and operational status was reviewed. plant housekeeping and cleanliness were evaluated Verification of several tagouts indicated the action was properly conducte Plant housekeeping has improved dramatically. Considerable effort is being devoted to improving the appearance of the facility both within and outsid This is being done in preparation for the upcoming International Atomic Energy Agency (IAEA) Operational Safety Review Team (0SART) inspectio _=_=__-___

_ - _ . _ _ _ __ _

. .

..

. ..

. 4

, I

,

Error in Facility Change During the week of ~ July.6,1987, the inspector ' reviewed a typical Facility Change Request (FCR). The objectives of the inspection'were to assess the technical adequacy of the change, and determine if

. program controls were adequate and properly implemente The change, .FCR 87-45, Linvolved the removal of a cold spring condition on the suction piping-of the #22 low pressure safety injection (LPSI) pump and the elimination of- a vibration problem in that pum The inspector performed a walk down of the system piping and hangers in the vicinity of the pump and reviewed the following:

--

_ Associated isometric end hanger drawings

--

FCR-package including safety analysis

--

Field work procedures (maintenance orders)

--

Procedure DESP 7, Revision 2, Design and Design Review

--

Procedure DESP 2, Control of Changes, Tests, and Experiments-

.The inspector interviewed licensee systems and design engineering personnel and an architectural engineering firm (Bechtel)

representative associated with the FC The FCR assessed the cold spring' load on the LPSI pump nozzle and ap-peared to correctly determine that, while the loading exceeded a'

system design basis requirement (1'.e., exceeded forces allowed by American Petroleum Institute. Standard 617), stresses remained low in the elastic range for the material involved. The inspector noted that only the stresses in the pump nozzle due to external forces and moments had been calculated. He pointed out to design engineering personnel that the internal stresses (e.g. , pressure) should be combined with the external stresses to assess total stress. The engineer who performed the calculation acknowledged that this should have been done but pointed out that, throughout the process, he recognized that the internal stress contribution would be small relative that caused by external stresses and would not have affected the overall conclusio The weight of the suction piping and valves in the vicinity of the LPSI pump should be supported by two spring hanger assemblies' located above the pum Prior to the FCR, a portion of that weight was also being carried by the #22 LPSI pump suction nozzle. When the suction flange was unbolted the piping moved downward approximately inches. The FCR removed a 0.5 inch section of the vertical piping, resulting in proper alignment of the piping to the sucticn nozzl The inspector noted, however, that after the modificatict had been completed and the piping reconnected, craft personnel had readjusted (relaxed) one of the spring hanger assemblies in such a manner that possibly as much as 900 pounds of loading may have been placed back on the pump nozzle. He pointed this out to the design and systems ,

engineering personnel involved. They discussed the problem with the

- - _ -______

-_ -

. .

,

.

.

-- 5 architectural engineer, agreed that some amount of loading had been placed on the nozzle, and readjusted the spring hanger assembly back to the load condition that existed when the pumping was reconnecte Further engineering analysis will be performed to quantify the amount of load that had been placed back on the nozzle and verify allowable stresses were not exceeded 2 Hanger cold load setting specifications will be modified as necessar The apparent cause of the above event was a lack of Systems Engineering personnel understanding of the design of the particular piping support system. Design personnel, who had a more complete understanding, were not asked to review the field work procedure. It is the licensee's intention that systems engineers have overal control and responsibility for FCR implementation. This places a significant responsibility on both the Systems and Design Engineering groups to ensure that the systems engineer is made aware of design detail This process is complicated when design work is done by outside contractors. In that case, a licensee design engineer (DE)

is designated to interface with the contractor, and the DE must ensure he understands the design and communicates this to the systems enginee .The above problem was aggravated by the fact that the Systems Engineering group involved was laboring under a very heavy workloa That group had lead responsibility for a number of the significant problems that arose during the recent Unit 2 refueling outage (e.g.,

reactor coolant pump suction deflector ring bolt failure, leakage of relief valve piping for LPSI system, mechanical commercial quality replacement parts issue, etc.). A realignment of responsibilities for this group is planne A final weakness noted by the inspector was that the field work procedure did not require verification that, when the suction flange was initially disassembled and the piping allowed to move to its free position, the spring hangers did not bottom out. Pipe movement was not sufficient to do this; however, had bottoming occurred, pipe movement would have been restrained and an inaccurate calculation of cold spring stress could have been obtained (stress was determined from pipe displacement).

No other unacceptable conditions were note . Operational Events Reactor Trip: On July 3,1987, with the reactor at 8%, Unit 2 paralleled I to the grid by shutting the main breaker The generator picked up approximately 100 megawatts, or 12% of rated load, and caused a sudden increase in steam demand resulting in quickly lowering steam generator pressure.

I

- _ _ _ _ _ _ _ - _

. .

.

.'. 6 The operators took action to reduce steam demand. Turbine bypass valves were shut and turbine load was manually lowered, but the reactor tripped on low steam generator pressure. The licensee immediately initiated an'

investigation of the caus Results of the investigation revealed that during the recent refueU ng outage the turbine generator control circuits were calibrated and aligne The procedure, I-20 " Turbine Generator Electro-Hydraulic Controller Calibration / Calibration Check," requires the use of a vendor-supplied test device called a turbine simulatar. The simulator used during the most recent alignment was newly supplied specifically for Calvert Cliffs whereas the previously used simulator had been shared among other power plants within the BG&E syste The Westinghouse simulator sends signals to the turbine controller that normally come from the running turbine generator. Results of I-20 required setting the Initial Valve Position Percent micrometer at 0.16 This setting is supposed to cause the turbine generator to pick up 5%

(approximately 40 megawatts) of rated load when the main breaker is close During the event, power was at 8%, operators were controlling temperature using turbine bypass valves and turbine control was in " Operator Auto" when the main breaker was shut. The generator picked up approximately 100 megawatts, or 12% of rated load instead of 5% and caused a sudden increase in steam demand resulting in lowering steam generator pressure and primary temperatur The operators took action to reduce steam demand closing bypass valves and turbine load was lowered manually. Steam generator pressure continued to decrease. Exacerbating the event was a positive moderator temperature co-efficient providing negative reactivity feedback causing power to decrease, further reducing temperatures and steam generator pressur Post-trip analysis and discussion occurred involving plant management, engineering, technicians, and the turbine vendor technical representativ The Initial Valve Position Percent micrometer setting from the previous performance of I-20 was 0.050. It was decided that, even though using that setting would probably result in an initial load pick up of less than 5%, using that setting was conservative. At that setting, t%e turbine generator picked up approximately 2 megawatts, essentially no load, and more load had to be picked up manuall The licensee has acknowledged that the root cause is still unknown and is in process of performing calibration checks of the new simulator and experimenting to check whether the allowable tolerances of the procedure might have yielded the high Initial Valve Position Percent micrometer settin . ...

GR ?

J2 -

Until the root cause is determined, a micrometer setting lower than the controller calibration procedure result will be used as a conservative

"

approach for future start up :

Reactor' Trip:'At 1:47 p.m. on July 15, 1987, the Unit 1 reactor tripped-from approximately 20% power. The cause of the trip was a high steam generator level condition initiating a turbine trip. The turbine trip, as designed, tripped the reacto ~

Plant conditions were quickly stabilized and plant systems functioned normally following ,the tri _

The sequence of events leading to the trip is described below. A

~

maintenance effort was in pronress to repair a diaphragm on the controller associated with the normal lev 'l control valve for the #16A feed water heater (FWH). That valve was gagged in a partially opened condition to permit a normal flow from the FWH. The high level dump valve was gagged

. closed. Both valves are operated by instrument air. Air to both valves had to be isolated to repair the diaphragm and the gags were installed to prevent valve movement. Water level began increasing in the FWH and caused a high level condition which caused the bleeder trip valve supplying extraction steam to the heater to trip closed. The resulting loss in heater efficiency caused a reduction in feed water temperature to the steam generators and led to an increase in reactor power. The reactor was operating at 100% power at the time. The operator incorrectly diagnosed the event and believed the power increase was due to a small dilution of the reactor. coolant system (RCS). Previously, during the shift he had observed that small amounts of demineralized water from the make up system were leaking into the volume control tank (VCT) with the make up control. valve in the closed position. The operator, therefore, began adding boric acid _to the VCT to counter the effects of dilutio Within a minute.he saw that RCS cold leg temperature was dropping and realized that boration'was not appropriate. He shifted the make up mode

't '

'

selector switch from the " borate" position to the " manual" position. By design, this action should stop the running boric acid pump. However, if the make up mode selector svitch is operated too rapidly, insufficient

'

time is available.for a relay (relay #42) to deenergize and an associated contact to open. This results in the continued running of the boric acid pum Operators'are generally aware of this idiosyncrasy of-the switch and normally additionally verify that the pump did sto In this case, the pump continued to run, and due to concerns regarding the FWH problems, the operator.did not check to ensure the pump had stopped. Boration continued and reactor power decreased. Operators decreased turbine power in an attempt to match' reactor power. By the time the operators realized the cause of the problem (boration) and stopped the boric acid pump, they were experiencing great difficulty matching turbine and reactor powe They elected to manually trip the unit. Before the manual trip could be initiated, the unit automatically tripped due to high steam generator level (level had swelled high due to rapid opening of turbine throttle valves by operators attempting to match turbine / reactor power).

,

n.

'a . y: T a s

1 3

I-1 m

-

~

,

3 )

2:) o s t <l A ,

s 4

,,

+

y .

j )

'

J .: t ., 8

,

,

.

(D i ' ' '

l.ia.wr u ;t >

l.^ NrbrtherMiew by the licentde inUcated that the operator involved must 3 "also have,^1n addition to edding boric acid to the volume control tank,

.z

'

' opened valve 1-MOV- RT, prcH ding a direct flow path from one of the boric

~

s acid storage ttoh (via th'e' running boric acid pump) to the suction of the  !

'

'

.

charging pump No charging pumps w~ere running at the tim '

jN: proximately 400 gallonsaf boric acid had been added to the reactor

, s pot.lst. system due to tiie dnplanned running of the boric acid pum y, <

cN Fo] lowing the event, valve line ups were performed to ensure no improper

'l,3

- paths existed for borfc acid additio he event was discussed with j' operations perscnnel, end the licensee is evaluating the design of the make up mode select switch for possible upgrad l

' P' ant Shut Down Due'to Excessive Gross Leakage: On July 17 at 5:45 Ehe:Ud t 2 reactor c &lant leak rate calculation identified gross leakage I to be 10.3 gp Pursuant to licensee administrative procedures, leakage greater than.10.0 gps requires a plant shutdown. Thus, Unit 2 commenced a controlled shutdown. The licensee reported this event to the NRC Operations Center ai;a shutdown required by Technical Specification s Previously, the licensee had scheduled an outage for July 17 in order to correct turbine and resctor coolant pump vibration problems, con +ainment elevated temperature problems and excessive coolant gross leakage into the reactor coolant drain tank (RCCI), averaging 8-9 gpm since startu During the previous week, the operations staff performed various valve line ups and recorded temperatures to try to identify the source of leakage into the drain tank. 'This resulted in isolating the cause to the

" leak off piping" principally composed of valve packing leak off or vent and drain leak of During the inspector's follow up of this situation, it was noted that, ac-cording to STP 0-27-2 " Reactor Coolant Leakage Evaluation," RCS gross leakage was 10.3137 gpm; RCDT (identified) in-leakage was 9.7289 gpm; net (unidentified) leakage was .5848; identified leakage was less than 10 gpm and unidentified leakage was less than I gpm; a notification to NRC Operations Center was not required; and a shutdown was not required by Technical Specifications. Nonetheless, if the licensee had desired to continue operation, ar operator would typically have been assigned to tighten a valve packing or two to reduce leakage to below the Technical Specifiestion limit and further troubleshooting and repairs mignt have been performed at powe Technical Specification 3.4.6.2 states:

Reactor Coolant System leakage shall be limited to: No PRESSURE BOUNDARY LEAKAGE, GPM UNIDENTIFIED LEAKAGE,

___ __ _____ - ___ _ _________-______

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

. . .

,

.

.. 9 )

l

l GPM total primary-to-secondary leakage through steam generators, and GPM IDENTIFIED LEAKAGE from the Reactor Coolant Syste The bases for the leakage states that:

Industry experience has shown that while a limited amount of leakage j is expected from the RCS, the unidentified portion of this leakage can be reduced to a threshold value of less than 1 GPM. This threshold value is sufficiently low to ensure early detection of .

additional leakag !

The 10 GPM IDENTIFIED LEAKAGE limitation provides allowance for a limited amount of leakage from known sources whose presence will not interfere with the detection of UNIDENTIFIED LEAKAGE by the leakage detection system The inspectors' review determined that the licensee's action in this case was conservative in reporting and prudent to shut down to correct the problems at hand. The licensee subsequently determined that a majority of the 9.7 gallons per minute leak rate into the RCDT was from seat leakage through the regenerative heat exchanger drain valves. The two valves in series were replaced. This corrected the excessive leakag No unacceptable conditions were note Improper Coefficient Setting in Reactor Protection System: On July 17, 1987, the licensee discovered an improper reactor protection system (RPS)

coefficient setting on each uni The settings were both associated with a dynamic response circuit in the Thermal Margin / Low Pressure (TM/LP)

calculator. The circuit determines and uses the rate of change of reactor coolant system (RCS) (differential temperature, dT) (hot leg minus cold leg temperature) and cold leg temperature to provide an anticipatory signal to a calculation of reactor thermal (dT) power. The circuit compensates for the slow response time of RCS temperature detectors (RTO's). Thermal power, in turn, is auctioneered (high select) with neutron flux power and used in the generation of the following three RPS trips: variable overpower (V0P), TM/LP, and axial flux offset (AFI).

The coefficients involved are labeled "a" and "T" (tau). RPS coefficients are tyoically provided to the licensee by Combustion Engineering as part of plant reload analyses. Prior to 1979, specific values were given for

"s" ar<! "T" (a = 0.81 and T = 1.6). Those values were verified te be in P.PS during start up testing on both units. During subsequent periodic survMllt.1ce testing however, only coefficient "a" was procedurally cnncked for proper setpoint. In the 1979 time frame, CF supplied a graph f- te be used for determining acceptable ranges of values of "a" and "T" instead of specific values. The graph showed that at least coefficient "a"

'

would have to be changed. If T were to remain at 1.6, an acceptable value of "a" would be 0.90. At the time the Electrical and Controls (E&C) group

--______ ___ _ _

. .

..

.

L"

l was both responsible for physically changing the coefficient in RPS and '

updating the value listed in the setpoint procedure. That group noted that "a" required change on Unit 1, appropriately updated the setpoint procedure to the value 0.90, and inserted the new setpoint in RP However, it appears that the change was missed on Unit 2, and the "a" i

setpoint was maintained (both in the procedure and in RPS) at 0.8 Six other Unit 2 coef ficients (YO4, b+4, b-4, x4, B4, 84) wera changed in that time frame by CE, however, they were maintained at their former values by the E&C group. The E&C group understood that the former values of the six coefficients were more conservative and apparently made a conscious decision not to change those coefficients. The basis of that understanding is not clear at this tim !

, In January 1986, the licensee established a dedicated group of systems engineers who were tasked with gaining expertise on system  !

design / operation and with coordinating maintenance and surveillance activities. The systems engineers for RPS recognized that the above coefficient differed from CE recommendations and, not feeling confident in their knowledge of the possible effects of these differences, initiated <

efforts to match RPS coefficients with CE recommendation When "a" on Unit 2 was raised to 0.90, swings that normally occur in the output of the dT calculation circuit were amplified and began causing spurious V0P reset alarms in the control room. The systems engineers then verified that "T" was set at the proper value (by visual observation of setpoints screws on a differentiator module). Since Unit 1 was not experiencing similar swings, the "a" and "T" coefficient settings were checked. The "a" coefficient was found to be 0.90 and the "T" coefficient was zero. With T = 0, the dynamic response circuit appeared to be disabled. The problems was reported to the NRC, and the setpoint for "T" was set to the proper 1.6 valu The NRC asked the licensee to determine the effect of the improper "T" and

"a" settings on circuit response time and, based upon this information, assess the effect on RPS performance in providing protection for the design basis accident of concern; i.e. , excess steam deman The licensee had Combustion Engineering (CE) perform this assessment. CE reported that, even with the improper "a" and "T" settings, core design limits would not be exceeded. At the time of this report, details of that analysis had not been presented to the NR The licensee has performed hand calculations using three sets of initial conditions that indicate that leaving the other six coefficients described above at their former values results in conservative trip setpoints being calculated by the TM/LP calculator. At the time of this report, CE was performing a more extensive assessment of the effects of the former setpoints. Results were not yet availabl However, it should be noted that there are other RPS trips which are redundant to the TM/LP trip which protect the core during the accident of concer FSAR section 14. __ _ - _ - _ _ _ _ _ - _ _

. .

,

..

.', 11 states that these can include, depending on the initial conditions, variable over power, axial flux offset, rate of change of power, and low steam generator pressur The licensee reviewed applicable surveillance test procedures and found, as noted above, that the value of the "T" coefficient is not periodically checked which is a deficiency in the procedur The other RPS coef ficients are checked. Additionally, RPS response time testing procedures currently utilized to meet Technical Specifications surveillance requirement 4.3.1.1.3 did not specifically test the time response of the dynamic circuit or otherwise include possible effects of the circuit in determination of overall trip system response time.

l The failure to periodically properly verify the performance of the dynamic

! response circuitry in the TM/LP calculation (including the checking of the l "T" coefficient) and to periodically review all RPS coefficients to ensure proper agreement with CE recommendations is a licensee identified

,

'

violation. This item is unresolved pending confirmation of the NRC's understanding that the RPS was capable of providing adequate protection for the design basis accident of concern and verification that all criteria of 10CFR2 Appendix C, Section V. A. have been satisfied (87-17-01; 87-19-01). The licensee's significant effort in the past few months to seek out and correct problems on a proactive basis, even where potential enforcement consequences could result, are acknowledge . Events Requiring NRC Notification The circumstances surrounding the following event requiring prompt NRC notification pursuant to 10 CFR 50.72 were reviewed. For this event resulting in a plant trip, the inspectors reviewed plant parameters, chart recorders, logs, computer printouts and discussed the event with cognizant licensee personnel to ascertain that the cause of the event had been thoroughly investigated for root cause identificatio Total Loss of Off Site Power At 3:25 p.m. on July 23, 1987, Units 1 and 2 automatically tripped from 100's power on loss of load. This was due to a total loss of offsite power. All three emergency diesel generators automatically started and supplied vital equipment as required. Reactor cooling was maintained with auxiliary feed water motor driven pumps, and RCS natural circulation was established. At 3:30 p.m. an Alert was declared, although the Emergency Plan only required an Unusual Event. This was done in order to activate the ancillary emergency stations; i.e. , Emergency Operating Facility (EOF), Technical Support Center (TSC) and Operational Suppert Center (OSC). By 3:45 p.m. all plant parameters had stabilized. By 4:45 both AFW turbine driven pumps had been started and the motor driven pumps secured. The first attempt to start No. 11 AFW pump failed when the pump tripped on overspeed, apparently due to an adjustment on the governor control linkage. At 5:00 p.m., efforts were in progress to energize a fourth 4-KV vital bus (Bus 21). This was to be done by powering the 13-KV

_ _ _ _ _ _ _ _ _ _ _ _ - - _ _ - _ _ -

_

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

.

• *

x

... ,

,, 12-

.I l

service Bus 21 with the third off! site power supply, a 13-KV line from i

, Southern. Maryland Electric Cooperative (SMECO). This power supply norm-ally feeds the warehouse and training / administrative facilities and is divorced from the plant. Electricians were requested to "make ready" the SMECO line to be " brought in." Electricians closed the normally

'

,

disconnected / racked out breaker 252-2310 and reset the UV and lockout

'

<

relays and reported to the shift supervisor that the breaker was ready for  !

closing. The shift supervisor interpreted this to mean that the requirements'of 01 27-E " Operation of the SMECO 13 KV Power Supply" had been fulfilled. However, one step in the procedure requires electricians to go outside the protected area to Warehouse No. 1 to substation No. 5 and disconnect' the warehouse and other facilities from the line before connecting the power to the 13-KV service bus 23 (on site). This step was  !

not done. It is normally directed to be done by the shift superviso ,

Consequently breaker 252-2301 was closed and the 13-KV service bus 23 was energized but when breaker 252-2303 was closed to energize the 4KV unit bus 21,'the SMEC0 line tripped on overload. At this point, the licensee recognized that the warehouse loads had not been removed and dispatched personnel to isolate the warehouse and called SMECO to'reclose their tripped breaker. By 5:20 p.m., both units had been borated to the re-quired'3% shutdown to ensure proper shutdown margin. Baltimore Gas and Electric had verified that 500 KV was available on the 5051'offsite power supply line and the fault was confirmed to be on the 5052 line; however, further examination of the ring bus breakers was still-ongoing. Because power was.available in the switch yard, the licensee downgraded the event to an Unusual Event. By 5:23 p.m. the SMECO power was successfully brought on site and aligned to power the 4-KV bus 21. The' Unusual Event was terminated at 8:10 p.m. after the 5051 line was restored on site at 8:05 p.m. Reactor coolant pumps were re-energized to establish forced circulation in each unit by 8:45 ~

The event took place just prior to shift turnover, therefore, an abundance of licensed operators and auxiliary operators were available. Operators and plant personnel followed the Site Emergency Plan as if it were routine. In general, the perception of the resident inspectors observing the control room and technical support center noted that the event was handled very smoothly in a calm and efficient manner. Emergency Operating Procedures were followed verbatim. The SMECO operating procedure was not, however, due to miscommunication. After power had been restored, the licensee held a Plant Operations Review Committee meeting to review the sequence of events, identify the root cause of the plant trips and obtain POSRC concurrence to restart the units. NRC regional personnel were privy to this meeting via telephon Resident inspectors also attended the POSRC meeting along with the NRC resident inspectors from Salem Nuclear Power Plant who were dispatched from Region I as a contingency. Restart was' approved providing the root cause of the 5052 line fault was identified and corrected before use. Unit I was not predicted to restart for an undetermined period due to a failed motor on 11A reactor coolant

.

pump. The cause of this motor failure is still under investigation.

,

Subsequently, Unit 2 resumed power operation on July 24. Also on July 24 l

_ _ - . _ _ _ _

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

_-_____

.

..

..

. 13

'

a critique of event was held with all personnel involved during this event in which the following discrepancies were identified:

--

Security diesel fuel transfer pump loss power due to loss of offsite l powe l

--

Automatic ring down phone lights and ring functions did not wor Meteorological tower data was not available directly in the control roo Meteorological Interactive Dose Assessment System (MIDAS) in control room did not functio Initial word to State and counties was incomplet Sicw to alert recovery organization in Baltimore due to incorrect phone numbe No. 11 AFW pump failed to start on first attemp i

--

Protective relays in the switch yard were overly sensitiv TSC computer tripped off lin Communication problems between control room and TSC were note Plant parameter information between EOF and TSC was slo Power for radiological air samples were not availabl Substantive plant parameter information was not available in the TS Telephone communications were generally weak throughou Emergency centers were set up to handle an event at one unit but not both simultaneousl In spite of the above deficiencies, plant personnel overcame these obstacles by using backup methods and performed very wel Inspectors have witnessed numerou" art failures of the turbine driven pumps during maintenance tests, sui millance tests, during plant trips, and now during an emergency situation. Since the pump normally starts on {

the second attempt, a Maintenance Request is not submitted nor is the system logged out of service. The licensee contends that the weakness in the AFW system has been recogni d and reliability has been increased by l adding motor driven pumps powered from a 4-KV vital bus.

Subsequently, Baltimore Gas and Electric Distribution and Transmission Department determined that a tree made contact with the "C" phase of line .

i

)

____ ___________ ___________ - _-

- _ _ - _ _ - - _ _ - _ -

. .

• s

..

. 14

5052 which caused the ground fault. When this occurred, the breakers 41 and 43 opened on the ring bus both at Calvert Cliffs and in Waugh Chapel,

,

isolating the 5052 lin However, as designed, the breakers at Waugh l Chapel attempted to reclose momentarily to determine if the fault still existed. When this occurred the fault was reflected through the ring bus l at Waugh Chapel onto the 5051 line causing the breaker on Calvert Cliffs ring bus (21 and 23) to open. This occurred due to a failure of a circuit card in the " permissive overreaching relay scheme." Breakers 21 and 23 would have also attempted to reclose except that they sensed a frequency difference between the coasting down turbine generator and the line (

'

frequenc On July 25 at 3:45 p.m. the 5052 500 KV line was restored to the Calvert Cliffs ring bu The permissive overreaching relay was taken out of l service since it is redundant to two other protection relays scheme No violations were identifie (

5. Plant Maintenance The inspector observed and reviewed maintenance and problem investigation activities to verify compliance.with regulations, administrative and maintenance procedures, codes and standards, proper QA/QC involvement, safety tag use, equipment alignment, jumper use, personnel qualifications, radiological controls for worker protection, fire protection, retest requirements, and deportability per Technical Specifications. The following a;tivities were included:

--

Oil change / coupling lubrication on No. 11 HPSI pump observed on July 13, 1987

--

Generator bearing change out for No. 11 diesel generator observed on July 29, 198 i No unacceptable conditions were note . Surveillance i

The inspectors observed no special surveillance tests this period due to I I

the abundant reactive activit . Review of Licensee Event Reports (LERs)

An LER submitted to NRC:RI were reviewed to verify that the details were l clearly reported, including accuracy of the description of cause and i adequacy of corrective action. The inspector determined whether further I information was required from the licensee, whether generic implications l were indicated, and whether the event warranted on site follow up. The I following LER was reviewed: j

--- - - - - - - - - - - _ - ----- _

-

_

.

..;

( )

L 1'

'.. 15

)

!

LER N Event Date Report Date Subject  !

Unit 2 87-05 07/03/87 07/30/87 Following the Calibration of the Turbine Generator Controller, Excess Load Results in a Low Steam Generator Pressure Reactor Trip Detailed examination of this event is documented in detail 3 of this inspection repor No unacceptable conditions were note . Radiological Controls Radiological _ controls were observed on a routine basis during the reporting period. Standard industry radiological work practices, confcmance to radiological control procedures and 10 CFR Part 20 requn cments were observed. Independent surveys of radiological boundaries and random surveys of non-radiological points throughout the facility were taken by the inspecto No unacceptable conditions were identifie . Observation of Physical Security Checks were made to determine whether security conditions met regulatory requirements, the physical security plan, and approved procedures. Those checks included security staffing, protected and vital area barriers, vehicle searches and personnel identification, access control, badging, and compensatory measures when require No unacceptable conditions were identifie . Exit Interview Meetings were periodically held with senior facility management to discuss the inspection scope and findings. A summary of findings was presented to the licensee at the end of the inspectio ,'