ML17312A497

~

One unresolved

item (529/9521-01)

was

open concerning

problems

experienced

with the essential

chillers in Unit

1 and Unit 3

(Section 5.5).

One unresolved

item (528/9521-02)

was

opened

concerning

a condition

potentially outside the design basis

which could lead to the turbine

driven

AFW pump tripping on overspeeed

(Section 7.2)':

One non-cited violation was identified concerning

inadequate

inservice

testing procedures

for the Unit 3 containment

spray

pump (Section

6. I).

One non-cited violation was identified concerning the failure of

operators

to follow procedures

when removing the Unit

1 Train

N

Auxiliary Feedwater

pump from service

(Section

8. 1).

This closed

Unresolved

Item 528/9514-02.

~

Unresolved

item 528/9431-01

was reviewed

and remains

open

(Section

9. 1).

~

Licensee

Event Reports

528/93-011,

Revision

1 and 528/95-001,

Revision

0

were closed

(Section

10. 1).

Attachments:

1.

Persons

Contacted

and Exit Meeting

2.

List of Acronyms

1

PLANT STATUS

1.1

Unit

1

Unit

1 began the inspection period at

100 percent

power.

On November 26,

1995,

a reactor trip occurred following a loss of condenser

vacuum (Section 2).

The

unit was returned to 100 percent

power operation

on December

1.

On December

7, reactor

power was reduced to 40 percent to identify and isolate

a potential circulating water leak into Condenser

Hotwell

1C.

On December

9,

a reactor trip occurred after

a startup transformer tripped (Section 3).

On

December

13, the unit was returned to 100 percent

power operation

and operated

at this power for the remainder of the inspection period.

1.2

Unit 2

Unit 2 began the inspection period at

100 percent

power arid operated

at this

power for the remainder of the period.

1.3

Unit 3

Unit 3 began

the inspection period with the core offloaded to the spent fuel

pool.

On November 6, the unit began the core reload

and entered

Mode 6.

On

November

28 the reactor

was taken critical.

On November 30, the unit ended

a

site record

47 day refueling outage

and synchronized

to the grid.

The unit

ended

the inspection

period at

100 percent

power.

2

REACTOR TRIP FOLLOWING A LOSS

OF CONDENSER

VACUUM UNIT 1

(93702,

71707)

On November 26, Unit

1 experienced

a main turbine trip from 100 percent

power

when

a condenser

vacuum breaker failed open causing

a low vacuum turbine trip.

Following a seven minute loss of load transient,

a high steam generator level

condition initiated

a reactor trip and

a main steam isolation signal.

During

the initial loss of load transient,

steam

bypass control valves to the

condenser

were only available for the first 13 seconds,

due to their closure

by condenser

vacuum protection logic.

After closure of the steam

bypass

control valves to the condenser,

secondary

pressure

increased

and three main

steam safety valves lifted.

One of the main steam safety valves remained

open

following the reactor trip and subsequently

reopened

at'pproximately five

percent

below its set pressure.

The inspector's

review of the licensed

operator

response

to the transient is discussed

in Section

2. 1,

During the event

and during subsequent

post-trip recovery,

the unit

experienced

several

equipment

problems.

These

included the following:

~

The condenser

vacuum breaker to the

C hotwell spuriously

opened

(Section 2.2).

Although Hain Steam Safety Valve 561

opened within one percent of its

set pressure, it remained

open for approximately ll minutes

and reset at

a pressure

below its expected

blowdown range.

It subsequently

reopened

at

a pressure

approximately five percent

below its setpoint

(Section 5.3).

~

Wh'en operators

attempted

to start the Train

N auxiliary feedwater

(AFW)

pump, it tripped

on low suction pressure.

~

Unrelated to the trip, the Train

8 essential

chiller tripped

on low

refrigerant temperature

(Section 5.5).

The Train

B AFW pump was

declared

since this essential

chiller provided cooling for

the

AFW pump room.

As

a result,

Technical Specifications

required the

licensee

to enter

Mode 4.

~

After the licensee

removed the Train

N AFW pump low suction pressure

trip from service,

they could not reopen the downcom'er isolation valves,

which had automatically closed

on the main steam isolation signal

(Section 2.3).

2.1

0 erator

Res

onse to the Turbine Tri

and Reactor Tri

Transients

The licensee initiated

an incident investigation

team,

lead

by Nuclear

Assurance,

to review the cause of the reactor trip and the operator

response.

The inspector

performed

an independent

review of the event

and interviewed the

control

room supervisor

and shift supervisor.

The inspector discussed

the

preliminary results of the licensee's

investigation with their- team

on

December 12.'he licensee's

preliminary conclusion

was that operators

had

performed properly and

had met management

expectations.

The inspector

discussed

concerns

regarding the following issues:

Operators

opened

the atmospheric

dump valves to control the secondary

pressure

transient

approximately four minutes into the event.

The slow response

of the feedwater control

system

(FWCS)

appeared

to

contribute to the high steam generator level condition.

A high level over-ride

(HLO) signal,

which terminates

feedwater flow on

high steam generator level,

appeared

effective in one steam generator

and not in the other.

Instruments

indicated that reactor trip setpoints

were approached

in

three instances

and operators

did not manually trip the reactor.

Following the reactor trip, it appeared

that secondary

pressure

and

primary temperature

exceeded

emergency

operating

procedure

guidance.

2. 1.1

Atmospheric

Dump Valve Control

The steam generator

atmospheric

dump valves

(ADVs) used at Palo Verde operate

solely on the

demand signal manually positioned

by the operators.

The ADVs do

not have

an automatic capability to relieve steam generator

pressure

and take

approximately

30 seconds

to open after operators

position their controller.

The inspector noted that the main steam safety valves,

which first lifted

approximately

one minute into the event,

were

open for approximately

two

minutes before the operators

opened the

AOVs.

The inspector discussed

the

delay in opening the

ADVs with the licensee.

The licensee

stated that the

operators

had

been trained with an operating philosophy that manual

action

should only be taken

once the transient

cause is understood.

They noted that

during the time the safety valves were open,

the operators

had focused

on

reactor coolant

system

(RCS) pressure,

which was also high.

In addition, the

secondary

side operator

was evaluating the loss of condenser

and the loss of

six of the steam

bypass control valves.

The licensee

indicated that although

the operator

response

to open the

ADVs was not quick, the licensee

determined

the operat'or

response

had met their expectations.

The inspector

agreed with

this assessment.

e

2. 1.2

Level Control

The inspector

reviewed several

parameter

trends to evaluate

the cause of the

high steam generator

(SG) level reactor trip.

The inspector

compared

feedwater flow to reactor

power and noted that the feedwater flow to each

was approximately

500 gallons per minute

(gpm) greater

than the amount

required for the given reactor

power.

In addition, the inspector

noted

a slow

SG level increase.

The inspector

noted that the operators

had not expected

a feedwater transient

and

had not recognized that the

SGs were being overfed until late in the

transient.

Approximately 30 seconds

before the reactor trip, the operators

closed the

SG economizer isolation valves to limit any leakage

past the

economizer control valves.

However, the operators

did not take manual

control

of the feedwater

system to control the feedwater flow before the high level

trip occurred.

The inspector

reviewed the system description

manual for the

FWCS.

The

inspector

noted that the

FWCS has

a lag network to effectively delay the

complete

implementation of a level deviation signal in single element control

(power less

than

15 percent).

The inspector questioned

engineering

about the

response

of the single element

FWCS during the plant transient.

The engineer

indicated that the system

was not designed

to handle the transient that was

experienced

on November 26.

The system

was designed

to be slow,

and not

overreact,

to allow stable operation during reactor startups.

The engineer

indicated the

FWCS responded

as designed.

0

The inspector discussed

the overfeeding observation

and

FWCS response

in

single element control with the licensee's

incident investigation

team.

The

team reviewed the inspector's

findings and evaluated

the trends.

The licensee

subsequently

determined that the

FWCS was not designed

to mitigate this event

before the

SG water level reached

a reactor trip setpoint.

The licensee

determined that,

in single element control, the

FWCS was designed

for power changes

of up to one percent

per minute.

Actual power had

been

changing at approximately

two percent

per minute.

The licensee

determined

that, in three element control, the

FWCS was designed for power changes of

five percent

per minute.

Actual power had

been

changing at approximately

eight percent

per minute when in three element control.

The inspector questioned

several

operators

in all three units about their

knowledge of the

FWCS in single element control.

The inspector

noted

weaknesses

in the operators'nowledge

of the limitations of the

FWCS.

The

inspector

found that operators

expected

the

FWCS to be responsive

in single

element control

and that they were not aware of the limits of the

FWCS in

single element control.

The inspector

noted that operator training seldom

includes plant operational

transients

at low power operations.

In addition,

during plant shutdowns,

plant operators

manually trip the reactor from

approximately

20 percent

power and, therefore,

single element control is not

utilized.

The inspector discussed

the operators'nowledge

of the

FWCS with the

licensee.

The licensee's

investigation

team

had determined that operators

had

responded

appropriately during the event.

They had noted that the event

was

complex with several

parameters

'in transient conditions throughout the event.

They concluded that these conditions would have

made it difficult for

operators

to conclude

what manual

actions, if any,

were appropriate.

However, they indicated that they would evaluate

the licensed operator

knowledge of the

FWCS and determine if additional training of the

FWCS

limitations at low power may enhance

operator

performance.

The licensee

also

planned to evaluate

the current

FWCS design

and determine if it can

be

improved or optimized.

2. 1.3

High Level Override

The inspector

noted that the function of the

HLO was to close the feedwater

economizer

and downcomer valves

when the level in the affected

SG reached

88 percent to prevent excessive

moisture carryover to the main turbine.

The

reactor trip and main steam isolation signals

are set at 91.5 percent level.

The inspector

reviewed the post trip sequence

of events

and noted that

approximately

two minutes prior to the reactor trip on high level in

SG 11,

the operators

had

a high level in

SG 12.

The inspector

found that the

HLO

isolated

feedwater to

12

and prevented

the reactor trip.

The inspector

noted that the

HLO for SG ll occurred three

seconds

before the reactor trip on

high level in

SG 11.

The

HLO did not prevent

an automatic safety function

actuation

in the three

second

time span.

The inspector discussed

the

HLO response

with the licensee's

incident

investigation

team.

The incident investigation

team indicated that the

HLO

responded

as designed

and that the

as found setpoint

was within the setpoint

tolerance.

The inspector questioned

whether the

HLO was nonfunctional

since

it failed to isolate feedwater flow to the

SG ll before the reactor trip and

main steam isolation occurred.

The incident investigation

team indicated that

they would review the setpoint

and setpoint tolerance

to determine if the

setpoints

or function could be enhanced

to prevent reactor trips.

The

inspector

concluded that the licensee

action to further evaluate

the

HLO

design

was appropriate.

2. 1.4

Hanual Operator Actions

The inspector noted that the shift supervisor

and control

room supervisor

discussed

manually tripping the reactor three times during the event.

The

first time occurred

when the pressurizer

pressure

increased

to .the pretrip

setpoint;

however,

pressure

turned

and remained

under control.

The second

time occurred

when

12 high level increased

to the high level pretrip;

however,

the

HLO turned the

SG level.

The third time occurred

when

11 high

level increased

to the high level pretrip setpoint;

however,

the reactor

automatically tripped 'before the operators

could manually trip the reactor.

Licensee

management

stated that operators

had appropriately

implemented the

expectations

that manual

action not be taken if it appeared

that control

systems

were responding

appropriately.

The inspector noted that licensee

management

planned to evaluate

management's

expectations

for taking manual

actions including manual control of automatic control

systems

during plant

transients.

The inspector

concluded that the licensee's

corrective actions

were appropriate.

In addition, the licensee

planned to evaluate their

guidance to operators

on when to initiate

a manual trip.

2. 1.5

ADV Control Following the Reactor Trip and Hain Steam Isolation

The inspector noted that following the reactor trip, the reactor coolant

system temperature

and

pressure

gradually increased for

approximately four minutes" before the operator

reduced

the steam generator

pressure

to allow the open main steam safety valve to close.

In addition,

RCS

temperature

and

pressure

increased

to above the

band specified

in emergency

operating

procedures.

The inspector discussed

the secondary

operator's

performance

in controlling plant parameters

with the licensee.

The

licensee

reviewed the post trip trends

and discussed

the requirements for

plant control that training uses to evaluate

operator

performance

on the

simulator.

The licensee

determined that the operator

performance

was

satisfactory,

but not optimal.

The inspector

agreed with the licensee's

assessment

of the performance.

2.2

S urious

0 enin

of the Condenser

Vacuum Breaker

The condenser

has three

vacuum breakers.

They are air to open,

spring to

close butterfly valves that can

be opened

from the control

room.

Prior to the

event,

an auxiliary operator

(AO) was touring the turbine building and

had

just ensured that the

vacuum breaker for the

C hotwell

had

an adequate

water

seal

on its upstream

side

when

he heard

a "loud sucking noise"

from the vacuum

breaker.

He observed that the vacuum breaker

was in the mid-position instead

of the required closed position.

He took control of instrument valves

on the

actuator

and eventually closed the breaker,

although his action

caused

the

vacuum breaker to open slightly farther than it already

was for a short period

of time.

The inspector

reviewed the licensee's

evaluation of the AO's performance

and

the evaluation of the cause of the failure.

2.2. 1

Auxiliary Operator

Performance

The licensee

concluded that the

AO should

have contacted

the control

room

prior to attempting to close the

vacuum breaker.

The licensee

noted that

AOs

were not trained in the operation of the

vacuum breaker actuator.

Operations

management

stated that it was their expectations

that the

AO contact the

control

room prior to taking actions

in the field on equipment that the

AO did

not have either written instructions or had not had specific training.

The licensee's

investigation

team found that the

AO had previously observed

that there

appeared

to be leakage

past the solenoid,

and that this condition

was not observed

on the other vacuum breakers.

However, the

AO had not

initiated

a work request

to have the condition investigated,

as expected

by

management.

'he licensee

planned to review this event with the

AOs and reinforce the

appropriate

requirements

expected of the

AOs before

and after taking action in

the field.

In addition,

the licensee

planned to evaluate

the current task

analysis

and training for AOs and determine whether additional training was

warranted to enhance

knowledge of air operated

valves

and actuators

similar to

the

vacuum breaker operator.

2.2.2

Vacuum Breaker Failure

The licensee

determined that the apparent

cause of the vacuum breaker solenoid

failure was attributed to component

age.

The licensee

inspected

the solenoid

valve and noted that the middle insert gasket (0-ring) within the solenoid

appeared

degraded

and flat.

The seating

surfaces

on the piston/guide

subassembly

also

showed indications of wear.

The licensee

checked

the other vacuum breaker solenoid valves for leaks

and

noted

two additional

solenoid valves in Unit 2.

As an interim corrective

action,

these

valves were replaced.

Valve Services

Engineering

planned to

perform

a root cause

analysis of the solenoid valves.

Based

on the results of

the analysis,

the licensee

planned to make

a determination of any preventative

maintenance

requirements

or recommend

an alternative corrective action.

-10-

2.3

Downcomer Isolation Valves Failed Closed

On the day after the reactor trip, with Unit

1 in Mode 4, the licensee

completed actions to disable the low suction pressure trip for the Train

N AFW

pump

and successfully

started

and ran the

pump.

Operators

proceeded

with

actions to place the Train

N AFW pump inservice.

The Train

N AFW pump provides

condensate

storage

tank water to the upstream

side of the downcomer'feedwater

control

and isolation valves.

The Train A and

B AFW pumps provide condensate

storage

tank water downstream of these

valves.

For each

steam generator,

there are two (spring to close, air to open)

isolation valves which close

on

a main steam isolation signal.

When operators

attempted to open the downcomer isolation valves,

they could only get one out

of three

open

and could not establish

flow from the Train

N AFW pump to either

steam generator.

Operators,

with the assistance

of the shift technical

advisor,

attempted

various combinations of venting the upstream

and downstream

pressures.

After

troubleshooting for four hours,

they were able to open the remaining three

downcomer isolation valves

when they established

discharge

pressure

of the

Train

N AFW pump to the upstream

side of the valves.

At the end of the inspection period,

the licensee

had not established

a root

cause of the failure to open.

The inspector

noted that Technical Specification 3.7. 1.2 for the

AFW system required that the Train

N AFW pump

and its associated

flow path were required to be operable.

On December

12,

the inspector requested

the licensee

to explain why they considered

the

'rain

N AFW pump flow path operable

in light of the problems experienced

following the reactor trip.

On December

15, in a conference call with the Region

IV management,

the Site

Shift Hanager stated that the flow path

was considered

operable

since the

downcomer isolation valves could

be opened with the assistance

of the

discharge

head of the Train

N AFW pump.

The Region

IV staff questioned

whether this sequence

was explicitly covered in operating instructions.

The

licensee

noted that operating

procedures

allowed this sequence,

but did not

specifically require it.

They stated that

a night order would be issued to

all units discussing

the downcomer event

and the success

path

used

by the Unit

1 operators

to open the valves.

The inspector

subsequently

reviewed the night order

and found that it

discussed

the failure of the downcomer valves to open,

speculated

that it was

caused

due to pressure

binding,

and noted that applying pressure

from the

Train

N AFW pump to the downcomer isolation valves

had allowed operators

to

open the valves.

The licensee

concluded that they had

a high degree of confidence that with or

without the night order, plant operators

could have

opened

the downcomer

valves in similar circumstances.

To assess

the validity of this conclusion,

the inspector discussed

the night order with control

room supervisors

and

-11-

shift supervisors

in all three units.

The inspector noted that prior to the

night order,

the supervisors

in Units

2 and

3 were

aware that there

had

been

problems with the downcomer valves in Unit 1, but were not cognizant of the

success

path for opening the valves.

The inspector

found that although the

night order did not provide specific instructions,

each supervisor

appeared

to

understand its message.

Additionally, the supervisors

interviewed

appeared

confident that they would have taken this course of action if the night order

had not been provided.

The inspector

found that while there

appeared

to be reasonable

basis for the

licensee's

assertion

that other crews could have

reopened

the downcomer

isolation valves in a timely manner,

the night order to communicate this issue

was prudent.

The inspector noted that the licensee

was investigating the cause of the valve

failures.

At the

end of the inspection period, the licensee

considered that

pressure

binding of the flex-wedge gate valves to be

a probable

cause.

Additionally, the licensee

was evaluating the design basis for these

valves to

establish their function to support the Train

N AFW flow path.

The inspector

will review the licensee's

root cause

evaluation in a future inspection.

2.4

Conclusions

The inspector concluded that the licensee's

actions,

both completed

and

planned,

to assess

this event

and its implications were appropriate.

3

REACTOR TRIP

FOLLOWING STARTUP TRANSFORMER TRIP UNIT 1

(93702,

71707)

On December

9, the Unit

1 reactor tripped from 40 percent

power on low steam

generator level.

Two minutes prior to the trip,

a ringtail cat

(a desert

mammal similar to a raccoon)

caused

a momentary

phase to ground path

on the

startup transformer

NAN-X03, which resulted in a loss of power to the Unit

1

bus

PBA-S03 (vital Train A 4160 volt) and the Unit 2 bus

PBB-S04 (vital

Train

B 4160 volt).

Both the Unit

1 Train A diesel

generator

and the Unit 2

Train

B diesel

generator

started,and

operated

as expected.

In Unit 1, the

120 volt ac control

power bus

NNN-D11, which supplies

the

FWCS

and the steam

bypass control

system, failed to complete

an automatic transfer

from its "alternate"

power supply bus

PBA-S03 to

a "normal" non-class

power

supply,

causing

a loss of power to bus NNN-Dll.

As

a result of the loss of

power to the

FWCS, the main feedwater.

pumps went to minimum speed

and

feedwater control valves closed,

causing

levels to decrease.

Steam generator

12 level decreased

below the reactor trip setpoint,

resulting

in a reactor trip.

The licensee classified the event

as

an "uncomplicated trip" and all systems

responded

as expected.

Bus NNN-Dll was re-energized

when it automatically

transferred

back to bus

PBA-S03 after its auxiliary transformer

powered

normal

power supply de-energized

as

a result of the turbine trip.

In addition,

two

reactor coolant

pumps

and two circulating water

pumps tripped following the

-12-

turbine trip since they could not fast transfer from the auxiliary transformer

to startup transformer

NAN-X03.

This resulted

in a loss of condenser

vacuum

and the unavailability of six of eight steam

bypass control

system

(SBCS)

valves.

Secondary

side pressure

rose to 1246 psia

and

one main steam safety

valve lifted and relieved for approximately two minutes.

The licensee

subsequently

determined that the safety valve lifted and reset within its

design margins;

The licensee

determined that

a mechanical

linkage problem prevented

bus

NNN-Dll from completing its transfer

sequence

from its alternate

power supply

to its normal

power supply,

as explained

below.

The licensee

has

had

bus

NNN-

Dll aligned to its alternate

power supply since initial startup.'uring

the

post-trip review, the licensee

determined that it would be preferable to re-

align NNN-Dll to its normal

power supply

and confirmed through testing that it

would reliably automatically transfer

from its normal to alternate

power

supply.

On December ll, at 0605, Unit

1 went critical

and returned to full power on

December

13.

Unit 2 remained

at

100 percent

power throughout the event.

The

resident

inspectors

responded

to the site following the reactor trip and

followed the licensee's initial post-trip review.

On July 17,

1995, Unit 2 experienced

a similar event

when

a breaker operation

error resulted in the loss of 13.8

Kv bus

NAN-S05.

During this event,

NNN-Dll

transferred correctly from its alternate

supply to its normal supply.

However, the transfer

was designed

as break before

make and,

as

a result,

during the brief period the bus

was de-energized,

the

FWCS switched to manual

'ith zero demand.

This event

was discussed

in Inspection

Report 95-14

and

Licensee

Event Report

(LER) 50-529/95-05.

During the initial post-trip review following the Unit 2 trip, the licensee

suspected

that bus NNN-Dll had not properly transferred.

They subsequently

discovered that it had transferred

as designed.

They concluded that

an

initial design of the bus transfer did not provide for uninterrupted

power to

NNN-D11, resulting in an inconsistent transfer of the

FWCS and the

SBCS.

The

licensee

had established

a December

29,

1995,

due date for modification

recommendations

to address

this design

weakness.

At the time of the Unit

1

trip, the licensee

had completed

most of their review, but had not finalized

'heir

recommendations.

In their review of the bus NNN-Dll transfer logic, the licensee

determined

that the original design required the

bus

be supplied

by the auxiliary

transformer,

through

bus

NAN-S01.

However, during initial plant startup,

the

licensee

had established

a practice of lining up the bus to its alternate

source,

from bus

PBA-S03.

This change

was apparently

made after reliability

problems with the automatic fast bus transfer,

between

bus

NAN-S01 (supplied

by the auxiliary transformer)

and

bus

NAN-S03 (supplied

by the startup

transformer),

caused

frequent challenges

to the bus

NNN-D11 transfer.

At the

time, the alternate

supply was considered

a more reliable power source.

-13-

On December

9, the licensee

determined that aligning bus NNN-Dll with its

normal

supply was preferable

to aligning it with its alternate

supply.

They

noted that in recent years

power had

been interrupted

on the startup

transformers

more frequently than

on the auxiliary transformers.

Additionally, the fast transfer of the 13.8

Kv buses

had worked reliably.

The

licensee

also noted that further efforts to address

the reliability of power

to sensitive non-vital

AC instrument

and control loads were still in progress.

4

OPERATIONAL SAFETY VERIFICATION

(71707)

4. 1

Midloo

0 erations

Unit 3

On November

15, the inspector

observed

several

aspects

of the midloop

performance.

The inspector

observed that the control

room staff maintained

positive control of the evolution.

The inspector

noted that the licensee

had

again

used

two control

room supervisors

from another unit,

who had

been

involved in recent midloop operations,

to supervise

midloop operations

in

Unit 3.

The continuity established

appeared

to be effective.

The inspector

reviewed the refueling water level indicating system

(RWLIS)

instrumentation calibration

and noted

no discrepancies.

The inspector

performed

a walkdown of the

RWLIS inside containment

and noted that the system

was properly aligned.

The inspector

noted

one of the

RWLIS instrument tubing supports

was loose.

The inspector

informed the licensee of the problem.

The licensee

promptly

corrected

the deficiency.

In addition, the inspector

noted that the condition

of work areas

and the general

housekeeping

in containment

were not as

good

as

the inspector

observed

in previous outages.

The inspector

informed the

licensee

about the observations.

Maintenance

management

determined

the major

contributor to the problem was contract workers

and took actions to correct

the work practices.

The inspector

concluded that the licensee's

response

to

the deficiencies

was appropriate.

4.2

Containment

Closeout

Ins ection

Unit 3

On November 22, the inspector

performed

a containment

closeout

inspection.'he

inspector

found several

small debris

items

and noted that the containment

cleanliness

was not as

good

as it has

been

observed

in the past.

The

inspector

noted that the reactor coolant

pump bays contained

the majority of

the debris.

The items discovered

in containment

included:

tape, plastic, tie

wraps,

and leather gloves.

The inspector

concluded that the size

and volume

of the items would not impact the containment

sumps

and, therefore,

the debris

items did not constitute

a safety concern.

The inspector discussed

the observation with the licensee.

The licensee

performed additional

walkdowns prior to startup.

The licensee

planned to

evaluate

the inspector's

walkdown findings prior,to the next refueling outage.

4.3

Use of Com uter to Monitor Chan in

Plant Conditions

Unit 3

On November

17, the inspector

observed

a start of a reactor coolant

pump in a

solid plant condition.

The inspector noted that the operators

reviewed the

reactor coolant

pump operating

procedure

and discussed

when the

pump should

be

manually tripped before

RCS pressure

decreased

below the minimum net positive

suction

head

(NPSH) pressure

requirement.

The inspector

noted that the

operator

used the emergency

response facility data acquisition display system

(ERFDADS) to monitor

a single suction pressure

value.

The inspector monitored

the pressure

transient

on the four safety channels

of pressurizer

pressure.

During the

pump start,

the inspector

noted that the pressure

dropped

below the

minimum NPSH pressure

for the

pump before the

pump tripped

on

an unrelated

speed

sensor failure.

The inspector noted that the

ERFDADS pressure

monitored

by the operator still indicated

adequate

NPSH pressure

when the

pump tripped.

The inspector questioned

the

ERFDADS response

with the system engineer.

The

system engineer

indicated that the

ERFDADS has

a minimum of a 1.5 second

delay

compared to the analog safety channel

indication.

The inspector

informed the

operations

department

leader of the observations

and expressed

a concern that

the operator

used

a time delayed indication in a transient condition.

The

operations

department

leader

agreed with the inspector

and issued

a night

order to all operators

indicating that

ERFDADS should not be solely used for

indication during

a transient condition.

The inspector

concluded that the use

of the

ERFDADS in this instance

had

no safety impact

on the'lant

and that the

licensee's

corrective actions

were appropriate.

4.4

A

roach to Criticalit

Process

Review

The inspector

reviewed the licensee's

corrective actions in response

to

'inspector identified weaknesses

in the licensee's

performance of 1/m plots

for monitoring the approach

to criticality during

a Unit 2 reactor startup

(Inspection

Report 95-14).

The inspector

noted that the licensee

evaluated

crew performance,

including the shift technical

advisor

and reactor

engineering,

during several

plant startups

on the simulator.

The licensee

identified performance

and knowledge weaknesses

of both the crew and the

training staff.

The licensee initiated

a condition report/disposition

request

(CRDR) to evaluate

and correct the weaknesses.

The inspector

concluded that

the licensee

performed

an extremely thorough

and indepth

assessment

of the

weaknesses

in both the performance of personnel

and the startup

process.

At the exit meeting,

the inspector

noted that the licensee's

review of the

startup

process

identified

a need for continued training

on complex evolutions

that are performed infrequently.

The licensee

agreed with the inspector 's

statement

and planned to continue evaluating operator training requirements.

5

MAINTENANCE OBSERVATIONS

(62703)

5. 1

Emer enc

Diesel

Generator Tri

s

Unit 2

On November

16, the Unit 2 Train

B diesel

generator

experienced

a

non-emergency trip during

an operability surveillance test.

This trip was

similar to the previous three trips that

had recently occurred

on this diesel

-15-

generator.

During the last inspection period,

a multi-discipline diesel

generator

task force team

was established

by management

to determine

the root

cause of failure of the recent trips.

The inspector

documented this effort in

Inspection

Report 95-18.

The inspector

observed

troubleshooting efforts by the electrical technicians.

The task force team leader developed

an extensive four stage action plan to

determine

the cause of the trips, verify proper restoration of the diesel

generator

following maintenance

activities, monitor diesel

generator

performance for spurious trips,

and obtain additional data for assessment.

The diesel

generator

was returned to service

on November

18, following

maintenance activities

and

a satisfactory operability surveillance test.

The inspector

concluded that the task force team appeared

thorough in their

review and the troubleshooting efforts observed

were adequate.

5.2

Char in

Pum

Power

End

Re lacement

Unit 2

On November 21, the inspector

observed

some portions of the work that

mechanical

maintenance

technicians

were performing

on the Train A charging

pump.

The technicians

were replacing the power end of the

pump.

The

inspector

concluded that the work performed

was adequate

and well controlled.

In September,

an auxiliary operator identified

a low lube oil pressure

condition

on the Train A charging

pump.

The

pump was declared

and

mechanical

maintenance

engineering

was notified.

The power end internals

were

inspected

and found to have significant damage.

The licensee initiated

a

CRDR

and

a formal root cause investigation.

The inspector attended

meetings with the maintenance

engineering staff to

discuss

the status of the investigation.

The inspector

noted that maintenance

engineering

had developed

an extensive

action plan to troubleshoot

and

investigate

the cause of the power end internal

damage.

The main contributor to the problem was

an alignment problem that existed

between

the

pump block and power end of the

pump,

as

a result of maintenance

performed several

years

ago.

The shifted alignment

was caused

by defective

power end studs

and installation errors

on this

pump which do not appear to be

applicable to the other pumps.

Maintenance

engineering

management

decided to

replace the power end of the

pump.

The inspector

concluded that appropriate

actions

were taken

by mechanical

engineering.

The investigation effort was extensive

and thorough.

The

pump

was placed in service

and

has since maintained oil pressure

in the acceptable

ranges.

5.3

Main Steam Safet

Valve Res

onse Followin

Reactor Tri

Unit

1

During the secondary

plant pressure

transient following the November

26

reactor trip (Section 2), three

main steam safety valves

(MSSVs) lifted.

All

-16-

three safety valves

were set to lift at

1250 psig

+ 3 percent.

Plant

recorders

indicated that the three valves lifted within tolerances

and that

pressures

had not exceeded

1250 psig + 3 percent,

indicating that the fourth

valve, set at

1250 psig, could have remained

seated

and still been within

setpoint tolerances.

Although two of the safety valves reseated

within three minutes,

HSSV 561

appeared

to continue to modulate

open

and closed until pressure

dropped

below

1130 psig.

Plant recorder data indicated that the reseat

pressure

would have

been approximately

10 percent

below the lift pressure.

The licensee

concluded

that. the blowdown could be expected

to vary between three

and nine percent.

Additionally, they noted that the valve had

been

open for approximately

ll minutes

when it reseated

and considered

that the heating of the valve could

have lowered the reset

pressure.

Following'he reactor trip, operators

allowed secondary

pressure

to increase

(Section

2. 1.5).

HSSV 561 reopened

at

a pressure

between

1155 psig

and

1181 psig (5.5 to 7.6 percent

below its setpoint).

The valve was subsequently

declared

and gagged

closed.

While the unit was in Hode

3 and prior

to restart,

a licensee contractor,

Furmanite,

performed on-line testing using

their Trevitest method.

HSSV 561

was tested

11 times, with a total of five

separate

adjustments.

The licensee ultimately was able to meet their criteria

of three successive

tests within one percent of 1250 psig.

The inspector reviewed the test results with the inservice testing engineering

staff.

They stated that

HSSV 561

had initially lifted at

1197 psig.

The

inspector

noted that this was considerably

lower that its initial set

pressure,

as well as the pressure it appeared

to lift at during the plant

transient,

considering that no adjustments

had

been

made to the valve.

The

licensee

could not explain the change

in the setpoint.

After the first test lift, the valve nut was turned three flats of the nut.

Although historical trends indicate that this adjustment

should

have raised

the lift point approximately

30 psi, the next two tests

were at

1194

and

1192 psig.

The licensee

could not explain why this had occurred,

except that

the safety valves

do not respond consistently to nut adjustments.

Subsequently,

the valve appeared

to lift at

a consistent

pressure

after each

adjustment.

The licensee

noted that this safety valve had

been

placed in service in

November,

1993 after having been refurbished.

The inspector

observed that in

November,

1993

a similar valve in Unit

1 had lifted approximately

40 psi lower

than its setpoint.

As documented

in LER 528/93-0ll, Revision

1,

HSSV 572 was

gagged

closed for the remainder of the operating cycle when

a review of test

data

had indicated that it had not responded

consistently to adjustments

made

during testing.

The inspector questioned

how the performance of HSSV 561

was

different than the November

1993 performance of HSSV 572.

The licensee

noted

that

HSSV 572 had behaved

more erratically during testing than

HSSV 561.

The

inspector

reviewed documentation of the evaluation for HSSV 572

and concurred

with this assessment.

'

-17-

5.4

Auxiliar

Feedwater

Pum

Governor Valve Deficiencies

Unit 3

On November 23, the licensee

entered

Mode

3 following the refueling outage

and

began to perform testing

on the turbine driven

AFW pump.

Technical

Specifications

require that the

pump

be tested within the first 12 hours1.388889e-4 days 0.00333 hours 1.984127e-5 weeks 4.566e-6 months of

establishing

normal operating

temperature

and pressure

conditions.

During the

testing,

the licensee

observed that, after operating for extended

periods,

the

governor

had

a tendency to lock up at speeds

between

2000

and 3000 rpm.

The licensee

developed

an action plan

and

began investigating the problem.

The troubleshooting

and repair took over 70 hours8.101852e-4 days 0.0194 hours 1.157407e-4 weeks 2.6635e-5 months of the

72 hours8.333333e-4 days 0.02 hours 1.190476e-4 weeks 2.7396e-5 months allowed

outage time.

The licensee ultimately discovered that the governor valve

bonnet

had

been rotated approximately I/4 inch,

as indicated

by match marks,

when it was reinstalled during the refueling outage.

This resulted

in a

misalignment of the linkage between

the hydraulic servo

and the governor stem,

which caused

the linkage to bind.

The binding appeared

to be enhanced

as the

turbine heated

up and thermally shifted.

The licensee

subsequently

performed minor adjustments

to the mounting of the

hydraulic servo

and the linkage to align them with the rotation of the

governor valve bonnet.

They performed

subsequent

testing which demonstrated

that the linkage operated

smoothly.

In addition, in the following week, the

licensee

tested

the Units

1 and

2 turbine driven

AFW pumps to ensure that

similar binding was not occurring.

The inspector

observed

portions of the troubleshooting

and repairs

and

determined that they were well controlled.

The inspector

discussed

the

troubleshooting effort with maintenance

management.

They had determined

during the troubleshooting that there

was additional data which could

be

'athered

during the setup of the governor valve which would aid in future

maintenance

and troubleshooting efforts.

This included establishing

a

baseline of the electronic signal into the governor module

and the hydraulic

pressure

output

as

speed

demands

are changed.

In addition, maintenance

management

stated that they planned to address

the misalignment of the

governor valve bonnet

as

a maintenance

performance error and initiate

corresponding

corrective actions.

5.5

Essential

Chiller Problems

Units

1 and

3

During the inspection period,

the inspector

observed portions of

troubleshooting efforts by the heating, ventilation

and air conditioning

(HVAC) team to resolve

problems experienced

by the Unit

1 Train

8 essential

chiller and the Unit 3 Train A essential

chiller.

The

HVAC team was comprised

of representatives

from design engineering,

system engineering,

maintenance

engineering,

and the maintenance staff.

The essential

chilled water system supplies chilled water to the essential

air

cooling units

and air handling units in the Control Building and the Auxiliary

Building during essential

equipment operation.

The essential

chiller system

consists

of two separate,

redundant,

safety-related

flow trains.

The

-18-

essential

chiller system starts automatically

when essential

equipment

operation is required.

5.5.1

Essential

Chiller Trip Unit

1

On November 27, the Unit

1 operators

were recovering

from a reactor trip

(Section

2)

and

had started

the Train

B essential

chiller to support

room

cooling to the Train

B auxiliary feedwater

pump.

Approximately 30 minutes

into the run, the chiller tripped due to low refrigerant temperature.

Operators

complied with Technical Specification action statements,

initiated

a

CRDR,

and contacted

HVAC personnel.

The

HVAC team initiated

an evaluation of the chiller trip.

The system

engineer

stated that the "low refrigerant temperature" trip appeared

to have

been

caused

by

a low refrigerant level in the cooler.

The cooler is

a shell

and tube heat exchanger.

Low pressure,

liquid refrigerant

on the shell side

absorbs

heat from chilled water flowing through the tubes

'as the refrigerant

flashes to vapor.

The refrigerant level prior to the trip was approximately

3.5 inches.

The

HVAC technicians

identified

some minor freon leaks

and restored

the

refrigerant level to approximately

6 inches.

The chiller was tested

and

returned to service.

The leaks

were subsequently

repaired during

an on-line

outage of the chiller.

The inspector discussed

the chiller trip with the system engineer.

The system

engineer

stated that during cold weather conditions the refrigerant

experiences

a "stacking"

phenomena.

During normal

oper ation,

vapor

refrigerant should condense

to

a liquid in the condenser

and flow to the

cooler.

However,

when essential

cooling water,

which flows through the

condenser,

is colder,

a low pressure

condition exists which causes

the liquid

refrigerant to stay in the condenser.

It appeared

that

a high level of

refrigerant in the cooler was necessary

to ensure that the stacking

phenomenon

would not result in a chiller trip.

The inspector discussed

the issue of refrigerant level with the system

engineer.

It appeared

that there

was competing guidance

on the appropriate

levels of refrigerant in the chillers.

~

Engineering evaluation request

(EER) 88-EC-003,

determined

the

recommended

refrigerant levels required during shutdown conditions.

The

Unit

1 Train

B essential

chiller refrigerant levels were 4.88 inches

minimum and 6.0 inches

maximum.

',

EER 90-EC-031,

which apparently'uperseded

EER 88-EC-003,

specified

a

shutdown refrigerant level

band for each chiller of 3.0 inches

minimum

and 7.0 inches

maximum.

This level

band

was incorporated

in the weekly

preventive maintenance

task.

-19-

During the Unit I fifth refueling outage,

HVAC technicians

identified

pitting on the compressor first stage impeller.

Design engineering

speculated

that the pitting was caused

by liquid refrigerant carryover.

The compressor

is

a two-stage,

centrifugal type.

It takes

suction

on

the cooler shell,

increases

the refrigerant

gas pressure,

and discharges

the refrigerant

gas to the chiller condenser.

The design

engineer

recommended

that refrigerant levels

be maintained

in the lower region of

the level

band to prevent impeller damage

by liquid refrigerant

carryover.

~

During discussions

with the system engineer

on December

21, the engineer

could not conclude that

a chiller could

be considered

operable with a

level of 3.5 inches.

~

On December

21, the inspector identified that

one of the chillers in

Unit 3 had

a level of about 3/8 of an inch above the

7 inch maximum.

This level did not appear to be consistent

with any of the guidance

provided.

The inspector

noted that the weekly preventative

maintenance

task still specified

a level from 3 inches to 7 inches.

The licensee

immediately lowered the level to 6 inches

and initiated

a

CRDR to

document

and resolve the problem.

The inspector

concluded that the licensee

did not have

a good understanding

of

the appropriate

levels for the chillers and

had not implemented sufficient

measures

to assure that chiller refrigerant levels were properly maintained.

5.5.2

Train A Essential

Chiller Low Lube Oil Unit 3

On October 2, the Unit 3 Train A essential

chiller was

removed

from service

for a maintenance

outage.

During this outage,

the rear motor bearing

seal

to

the compressor

was inspected

and the 0-ring seals

were replaced.

The

technicians

had problems with. the replacement

of the rear seal

and the 0-ring

had to be replaced

again.

The compressor

and motor are refrigerant cooled

and oil lubricated.

A motor

driven,

compressor lubricating oil pump and reservoir are located in the

compressor

base.

There are two sight glasses,

an upper

and lower bulls-ey'e,

that represent

the minimum and

maximum oil levels.

The licensee

has

established

that the chiller should not be operated

with the oil level below

the lower sightglass,

which corresponds

to approximately 7.5 gallons of oil,

or with an oil level in excess of 25 gallons.

On October

29, the chiller was returned to service following the maintenance

outage with an initial volume of 15 gallons of oil.

On November

27, the

Train A essential

chiller was declared

inoperable after

an

AO noted that the

oil pump appeared

to be cavitating

due to low reservoir level.

HVAC

technicians

added

more oil and retested

the chiller.

The chiller was then

returned to service.

-20-

The licensee

subsequently

determined that from November

10, through

November 27,

17 gallons of oil had

been

added to the system to maintain the

minimum oil level requirements.

Operations initiated

a

CRDR and the

HVAC

personnel

developed

an action plan to resolve the apparent

loss of oil.

The

HVAC system engineer

explained to the inspector that during cold weather

conditions the oil tends to migrate with the refrigerant.

Once the chiller is

operated

under loaded conditions,

the oil will heat

up and return back to the

oil sump.

The guidance

given to operations

by the

HVAC engineer

was to have

an operator present

during any start of the essential

chiller.

Since there

was

a total of 32 gallons of oil in the system,

once the chiller is fully

loaded the oil would migrate back through the oil return system

and the

operator

would need to drain the oil. If the reservoir

becomes

too full of

oil,

a "high bearing temperature trip" could result.

On December

5, the inspector

observed

HVAC technicians

perform maintenance

activities

on the Train A essential

chiller.

The oil was removed from the

system.

The technicians

removed

and replaced

the rear motor bearing seal.

The technician

added

12 gallons of oil and the chiller was tested.

After a

successful

four hour run, the chiller was returned to service.

The repairs

appeared

to have addressed

the oil migration problem.

The inspector

was concerned

that the repair performed during the refueling

outage

may have

caused

the oil migration problem.

Additionally, the inspector

was concerned

that the licensee

had not established

an appropriate

basis for

considering

the chiller operable after it was identified on November

27 to

have excessive oil.

5.5.3

Summary

The licensee initiated

CRDRs to address

both the Units

1 and

3 problems.

Additionally, the Executive Vice President initiated

a Level

1 action item to

resolve these

issues.

The licensee initiated measures

to determine

the

refrigerant level required to assure chiller operability and resolve the

inconsistent

guidance.

The inspector considered this to be

an Unresolved

Item

(528/9521-01).

6

SURVEILLANCE OBSERVATION

(61726)

6.1

Containment

S ra

Pum

Testin

Unit 3

On November 5, while Unit 3 was in an outage,

operations

personnel

performed

surveillance test

73ST-9SI15,

"Containment

Spray

Pump Full Flow Inservice

Test,"

on the Unit 3 Train A containment

spray

(CS)

pump.

The

pump failed

both the

ASME Code Section

XI and design basis full flow tests.

The licensee

checked

the test instrumentation

and,

although they were able to demonstrate

some

improvement,

the

pump again failed its design basis flow test.

The

licensee

disassembled

the

pump

and

was unable to identify significant

degradation

or wear.

Mechanical

maintenance

technicians

replaced

the impeller

with another impeller to improve performance.

The licensee

performed

a

-21-

Section

XI test to establish

the

new pump curve

and found that the

pump had

improved performance.

The inspector

observed

portions of the corrective maintenance

and surveillance

testing,

reviewed inservice testing procedures

and performance history,

and

discussed

the cause of the apparent

degradation

with inservice testing

engineers,

maintenance

engineers,

and system engineers.

6. 1. 1

Testing Hethodology

The licensee

performed several

inservice tests of the

CS pumps.

The two

Section

XI tests,

a monthly test performed

on mini-flow recirculation

and

a

refueling outage test at 3525

gpm, were to verify that the

pump performed to

its baseline

pump curve.

A design

basis flow test to establish

that the

pump

could develop

a minimum differential pressure

at 4000

gpm was

added

in 1994

and

was being performed

on the Train A CS

pump for the first time.

The

acceptance

criteria for the full flow Section

XI test,

at 3525

gpm,

was

differential pressure

between

219 psid

and 233.6 psid.

The acceptance

criteria for the design basis-test,

at 4000

gpm,

was

a differential pressure

of greater

than

211 psid.

The Section

XI November

5 test result

was

217.3 psid

and the design

basis test result

was

198.3 psid.

During their investigation,

the licensee identified that the correction factor

used in both tests to compensate

for instrument

gauge location

was incorrect.

The correction factor used in the surveillance test procedure

was 1.3 psid.

This correction factor failed to properly compensate

for gauge elevation

differences

and the pressure

drop across

a flow orifice between

the

pump

discharge

and the discharge

pressure

gauge.

The licensee

determined that the

correction factor should

have

been

11.7 psid for the Section

XI test

and

12. 1 psid for the design

basis test.

With the

new correction factor the Train A CS

pump would have passed its

Section

XI test with a differential pressure

of 227;7 psid,

but would not have

passed

the design

basis flow test,

which was being performed for the first

time.

6. 1.2

Correction Factors for Section

XI Tests

The licensee

was unable to determine

the basis for the 1.3 psid correction

factor used in the Train A CS

pump full flow tests.

They did discover that

the correction factor used in the mini-flow recirculation test

was correct.

They found that the correction factor for the full flow tests

had

been

established

in 1989

when full flow testing

was initiated.

The error in the

correction factor had not been identified when the procedure

was revised in

1994 to add the design basis flow test.

The licensee

subsequently

reviewed all other

pump tests

and found similar

errors in the full flow test correction factors.

However, the licensee

determined that these differences

had not impacted the conclusions of these

-22-

tests.

Additionally, the licensee

found minor discrepancies

with the

correction factors of other Section

XI tests.

At the end of the inspection period,

the licensee

was establishing

the basis

for the correction factors

used in all Section

XI tests with the intent of

maintaining the basis.

The inspector

found these

actions to be appropriate.

6. 1.3

Hotor Replacement

In 1987, the licensee

replaced

the Train A CS

pump motor with a motor having

a

lower nameplate

speed

(1757 versus

1779 rpm).

The licensee

did not rebaseline

the

pump curve with the

new motor.

This change

would have

had

a greater

effect on the

pump at higher loads.

The licensee

stated that the design basis

full flow test

loaded the motor to two-thirds of its capacity

and would not

have

had

a significant impact

on the

pump curve.

The inspector discussed

some uncertainties

that

may exist with the system

engineer

concerning the design

basis surveillance test.

He stated that in

addition to using the appropriate correction factor, if additional

uncertainties

were removed, for example,

taking credit for the miniflow valve

that does

not

come full open

and using

a more accurate

flow instrument, it

would be expected that the

pump would have marginally passed its design

basis full flow test.

6. 1.4

Conclus'ions

10 CFR 50, Appendix B, Criterion

V requires that procedures

shall

include

appropriate

acceptance

criteria.

The

CS full flow test procedure did not have

appropriate

acceptance

criteria in that the instrument correction factors

used

to determine differential pressures

were in error and this constituted

a

violation of more than minor safety significance.

The inspector

reviewed the

licensee

evaluation,

discussed

the corrective actions with management,

and

concluded that the corrective actions

were appropri ate.

Additionally, the

inspector

noted that the problem was identified by. the licensee

and that there

were

no similar violations identified by either the inspectors

or by the

licensee

which could have reasonably

prevented this occurrence.

This licensee

identified and corrected violation is being treated

as

a non-cited violation,

consistent

with Section VII of the

NRC Enforcement Policy.

7

ONSITE ENGINEERING

(37551)

7. 1

Steam Generator Modifications Unit 3

The licensee

performed modifications to the Unit 3 steam generators

during the

refueling outage to reduce the dryout region in the upper portion of the tube

bundle.

The modifications included extending

and lowering the elevation of

the downcomer feedring,

removing orifices in the moisture separators

on the

hotleg side,

and cutting 45 three

inch diameter holes in the shroud

between

the downcomer

and the tube bundle.

-23-

Due to unanticipated

problems

associated

with the modification:

~

The licensee

cut only 27 holes in one

shroud

and

no

holes in the other generator

due to problems experienced

with the

cutting tool.

~

The licensee

discovered that

a cyclic low energy water

hammer developed

in the

new downcomer feedring

when low steam generator levels

and

AFW

flow below 200

gpm were established.

~

Operators

noted that the feedwater level control

system developed

oscillations of approximately

10 percent level

peak to peak with reactor

power between

25 and

45 percent.

7. 1. 1

Technical Specification 3.0.3,

Shroud

Holes

The licensee

attempted to cut 45 holes in the shroud

betwe'en the downcomer

and

the tube bundle using

a three inch diameter electrical

discharge

machining

(EDM) tool.

The job was performed remotely by lowering the

EDM tool

on the

downcomer side of the shroud.

Each cut was expected

to take

up to ll hours to

perform.

The licensee

found that each cut was taking substantially longer than expected

to perform., In addition,

the

EDH tools,

which were expected

to last for up to

three holes were eroding after one hole

and in one instance

an

EDH tool

head

fell off the

EDH tool into the area

between

the shroud

and the tube bundle.

The majority of the

head

was retrieved

and

a safety evaluation

was performed

for the remainder which concluded that it could not damage

the steam

generator.

The licensee

had previously mocked

up the modification process

at

a vendor's

facility.

They determined that the conductivity of the water in the steam

generator

was greater

than the conductivity at the mock up facility and that

this difference

had contributed to the, problems in the rate of cutting and the

life of the

EDH tool.

The licensee

was evaluating the cause of the

differences

in conductivity.

The licensee

completed only 27 holes in steam generator

32 and elected

not to

cut holes in steam generator

31.

They revised the

10 CFR 50.59 evaluation to

disposition this interim condition.

The inspector

reviewed the evaluation

and

noted

no discrepancies.

The licensee

concluded that the condition was

bounded

by the evaluations

performed for the original design

and for the completion of

the entire modification.

Additional calculations

were performed

by the vendor

to address

flow velocities at the tube bundle

and found that the

new flow

velocities would be acceptable.

7. 1.2

Feedring

Water

Hammer

-24-

After the licensee

had entered

Mode 3, they lowered level in one steam

generator

as part of a chemistry control evolution.

Engineers

working on the

turbine driven

AFW pump noted

a cyclic banging in the

AFW discharge lines.

Engineers

subsequently

determined that the water

hammer condition had

been

established

in the modified feed ring under certain flow conditions.

The downcomer feed line rises

up the outside of the steam generator to the

nozzle.

On the inside of the generator

there is

a slip fit between the nozzle

and the feedring.

The line has

a 90 degree

bend

as it turns

downward

and tees

into the lines that form the feed ring.

The licensee

determined that with no

flow to the downcomer line,

a 40 inch section at the nozzle dries out.

At low

flow conditions, this portion did not completely fill.

As a result,

the vapor

in this space

would condense

in the environment of the cold

AFW water causing

the water

hammer.

The inspector

observed

the licensee

evaluate

the problem

and establish

a

special test to determine

the conditions where the water

hammer developed.

In

addition, the inspector

observed

portions of the testing.

The inspector

noted

that although the testing

was performed over

a holiday weekend,

appropriate

levels of licensee

management

were involved.

The testing determined that the

water

hammer

was occurring with steam generator levels below 25 percent of the

narrow range

instruments

and at flows less than

200 gpm.

Additionally, they

monitored pipe movement,

installed

an accelerometer,

and inspected

the

insulation

around pipe hangers

and determined that there

was

no discernable

movement in the piping.

Based

on these observations,

they concluded that the

water

hammer

was of low energy

and would not damage

the steam generators.

The licensee initiated

a

CRDR and planned to take action to revise operating

instructions to restrict

AFW flows to above

200

gpm when below 25 percent

SG level.

7. 1.3

Flow Oscillations At Low Reactor

Power

During reactor startup,

with reactor

power between

25 and

45 percent,

feedwater flow oscillations would cause

steam generator level swings of up to

10 percent

peak to peak.

Engineers

obtained data to evaluate

the cause of the

oscillations

and to establish if changes

to the

FWCS could be made to reduce

the oscillations.

The licensee

also provided guidance to operators

on actions

to take in this region if power is reduced.

7.2

Potential

Loss of Two AFW Pum

s Followin

a Steam Line Break

On December I, the licensee

made

a four hour non-emergency

report to the

NRC

in accordance

with the requirements

of 10 CFR 50.72 for a condition outside of

their design basis.

The licensee

had identified that, following a steam line

break of a specific size

and

a loss of offsite power,

a sequence

of events

could occur with design basis

assumptions

that could lead to the turbine

driven

AFW pump tripping on overspeed.

The licensee

postulated that with the

coincident single failure of the motor driven

AFW pump .and with no credit for

-25-

operator actions within the first 30 minutes of the event, that emergency

feedwater

would be lost to the intact steam generator.

The licensee

planned to describe

the details of the event in Licensee

Event

Report 50-528/95-013.

The inspector

had discussed

the potential for this

sequence

of events with the licensee

in Inspection

Report 95-12.

At the. time,

the licensee

had not determined if there

was

a break size that could lead to

the event.

The licensee

has noted that the sequence

of events did not consider operator

action

and

assumed

some conservative initial conditions outside of normal

operating

parameters.

The inspector reviewed the initial conditions

and

operating

procedures

and found that they provided reasonable

assurance

that

emergency auxiliary feedwater

would be available following a steam line break.

This is an Unresolved

Item (528/9521-02).

8

FOLLOHUP OPERATIONS

(92901)

8. 1

Unresolved

Item 528 9514-02

Closed

Two Auxiliar

Feedwater

Pum

s

~Ino erabl e

This unresolved

item involved the failure of Unit

1 operators

to recognize

that one auxiliary feedwater

pump was already inoperable

when they prepared

to

remove

a second auxiliary feedwater

pump from service to perform

a breaker

modification.

Technical Specification 3.7. 1.2 required that, with two

auxiliary feedwater

pumps inoperable,

the unit be in hot standby within six

hours.

Hoth

AFW pumps were inoperable for approximately 5.5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months .

On August

9 at 6:00 a.m.,

the night shift removed the Train A spray

pond from

service prior to shift change.

Operators

entered

the action statement for

Technical Specification 3.7. 1.2 for the Train A AFW pump (the turbine driven

pump) since without the spray pond, the essential

ventilation for the Train A

was inoperable.

Around 9:15 a.m. the day shift crew removed the Train

N AFW

pump to perform

a modification on its breaker.

They did not recognize that

this placed

them in the six hour action statement.

At approximately

3:00 p.m., operators

racked in the breaker for the Train A spray

pond which

subsequently

tested satisfactorily.

Operators

on the subsequent

night shift recognized that the day shift had

removed

both

AFW pumps from service.

They returned the Train

N AFW pump to

service at around 8:00 p.m.

and initiated

a

CRDR.

The licensee

designated

the

CRDR as significant

and established

an investigation

team to review the event.

The licensee

subsequently

determined that they had met the Technical

Specification requirement

in that they had established

operability within the

six hours allowed.

The inspector

reviewed the licensee's

investigation.

The inspector

concurred

that the licensee

had not exceeded

the six hour shutdown requirement.

The

inspector

noted that the investigation

had identified tliat the work on both

the Train A spray

pond

and the Train

N AFW pump were pre-planned for the

same

-26-

day.

While the work on the Train A spray

pond was reviewed

as

a significant

job, the Train

N AFW pump had not been considered significant work.

The

investigation

focused

on the work control process

weaknesses

which had

resulted

in this oversight.

The inspector

reviewed this evaluation

and the

proposed corrective actions

and found them to be thorough.

The licensee

also noted that

a contributor to the scheduling

problem was the

inappropriate application of the recently

implemented probabilistic risk

assessment

matrix.

The matrix provided risk insights to operators for

emergent

work.

The matrix established

that combinations of equipment

which

place the plant in Technical Specification 3.0.3 are not allowed regardless

of

the risk significance.

However, the matrix did allow the more restrictive

Technical Specification requirement for two AFW trains inoperable,

in that it

suggested

that removing the Train

N AFW with the Train A spray

pond

had

provided

no additional risk.

The control

room supervisor

had referred to the

matrix prior to removing the Train

N AFW pump from service.

One of the

corrective actions

was to reinforce with operators that the matrix was

a tool

to be used in conjunction with Technical Specifications

and did not include

all restrictions.

The inspector

found this to be appropriate.

The inspector

noted that ultimately it was the operating

crews responsibility

to recognize that removing the Train

N AFW pump from service

was not

appropriate.

The licensee's

investigation identified that the control

room

supervisor,

who had authorized

the work,

had not discussed

the decision with

either the shift supervisor or the shift technical

advisor.

When

he

subsequently

discussed it with the shift supervisor,

the shift supervisor

had

not recognized that two trains of AFW were inoperable.

Additionally, the shift technical

advisor

was responsible for maintaining the

technical specification

component condition record log to track equipment

out

of service.

The procedure

governing this log required that equipment that is

out of service through

a shift change

be logged.

While the Train

N AFW pump

was not expected

to be out of service through the shift change, it ultimately

was.

The inspector considered

that this was

a failure to follow procedure.

The Operations

Department

Leader discussed

this weakness

and the expectations

on communications with all licensed operators

during the subsequent

training

cycle.

The licensee

also counselled

members of the crew involved.

The inspector considered

that the failure of operators

to follow procedure

40DP-90P23,

"Technical Specification

Component Condition Record

Sims

Procedure,"

and log the Train

N AFW pump to be

a violation of Technical Specification 6.8. 1 of more than minor safety significance.

The inspector

found that the licensee's

actions

were appropriate.

Additionally, the

inspector

noted that the problem was identified by the licensee

and that there

were

no similar violations identified by either the inspectors

or by the

licensee

which could have reasonably

prevented this occurrence.

This licensee

identified and corrected violation is being treated

as

a non-cited violation,

consistent with Section VII of the

NRC Enforcement Policy.

0

e

-27-

9

FOLLOWUP - ENGINEERING/TECHNICAL SUPPORT

(92903)

9. 1

Unresolved

Item 528 9431-01

OPEN

Letdown Isolation Valve Leaka

e

On December

8, the licensee

closed Unit 2 letdown isolation valve CHA-516 and

determined that it did not leak with approximately

2000 psi differential

pressure

across it.

This test supported

the licensee's

analysis

which

concluded that the three letdown isolation valves in Unit 2 could close

against

pressures

established

during

a letdown line break

and prevent

downstream flow from exceeding calculated limits.

The letdown isolation valve

operators

in Units

1

and

3 were replaced

during the past refueling outages

to

resolve the concern that they were undersized.

Earlier in the inspection period, the licensee identified that, in April 1995,

the as-found condition of Unit

1 letdown isolation valve CHB-515 had not been

accurately

described

to engineers

performing

an operability determination.

Upon recognizing the flawed Unit

1 operability determinati'on,

the licensee

reperformed

the operability determination

and developed

a reasonable

assurance

that the Unit 2 valves were operable.

Subsequently,

they determined that

an

online test would establish

conclusively that all three Unit 2 valves were

operable.

The inspector

concluded that the decision to perform the test

demonstrated

conservative

engineering

and found that the licensee

had

performed substantial

reviews

and exercised

proper caution in performing the

test.

The inspector noted that the failure to properly characterize

the as-found

condition of Unit

1 CHB-515 was another

example of weak engineering.

10

ONSITE REVIEW OF LERs

(92700)

10.1

Closed

LER 528

529

530 93-011

Revision

1:

Potential

Safet -Related

E ui ment Problems

Due to

De raded Grid Volta e and

Closed

LER 528 95-001

Revision 0:

Entr

Into TS 3.0.3

Due to De raded Volta e

10.1. 1

Licensee

Reports

LER 93-011,

Revision

1,

was issued

February 6,

1995,

and noted that

a

previously unanalyzed

condition could occur due to low grid voltage.

The

unanalyzed

condition involved double sequencing

of safety-related

pumps during

an accident.

Following an emergency

safety features

actuation,

the licensee

noted that the potential

existed to not only start sequencing

safety-related

equipment

onto preferred offsite power, but also to initiate load shedding

due

to the Class

1E 4. 16 Kv undervoltage

relays dropping out due to low grid

voltage

and not resetting,

and then resequencing

the equipment

onto the

emergency

diesel

generator.

The licensee

stated if grid voltage were kept above

an administrative limit,

their calculations

indicated that the double sequencing

would not occur.

The

licensee

also discussed

long term corrective actions

and stated that pending

completion of these

actions

the grid would be kept above

100 percent.

-28-

However,

on February

15,

1995, the licensee

entered

due to grid voltage falling below 100 percent during

planned

switchyard evolutions.

The licensee

issued

LER 95-001,

Revision 0,

on

Harch

15,

1995, to report the February

15,

1995, occurrence.

10.1.2

Licensee Actions

0

The licensee

issued

a new procedure

specifying operator actions in response

to

a degraded grid and

added

degraded grid voltage to the Unit

1 Plant Monitoring

System displays.

The licensee

stated that the following long term corrective

actions

would make all three units operable

above

98 percent grid voltage,

the

design lower limit.

~

Removal of loads

from Unit 1, the most heavily loaded unit.

~

Addition of an accurate grid voltmeter in Unit 1,

so licensed operators

will have ability to routinely monitor grid voltage.'

Transformer

upgrades

and sequencer

changes.

~

Automatic block of fast

bus transfer.

10. 1.3

Summary of Inspector

Actions During Inspection

Report 95-12

The inspector

reviewed both

LERs, the licensee's

procedure for response

to

degraded grid voltage,

the licensee's

review of the potential for further

occurrences

of degraded grid voltage,

the licensee's

long term corrective

'actions,

and the requirements

the licensee

provided to grid operators.

The inspector

reviewed

Procedure

41AO-IZZ57, "Degraded

Grid Voltage,"

Revision 4.

This procedure

provided operator guidance for all three units for

response

to degraded grid voltage in various plant modes.

The basic action

was to restore

one emergency

bus in each unit by blocking the fast bus

transfer of nonsafety-related

loads to the startup transformers

which supplied

the safety-related

loads.

The inspector

reviewed the affect of blocking the

fast

bus transfer

and determined that licensee calculations

indicated this

block would allow safety-related

voltage to remain

above the degraded

voltage

relay setpoint.

The inspector

reviewed licensee

Letter File 95-005-419.8,

"PVNGS Expectations

Regarding

Evaluations Potentially Affecting Grid Voltage

Range Limits," dated

February

23,

1995,

and noted that grid operational

expectations

were clearly

stated.

The inspector

noted that the licensee

had

used the site main generators

to

raise grid voltage during the February

15 event

and questioned

the licensee

as

to whether grid operators

could have raised this voltage external

to the site

'n

February

15, or during future occurrences

when there

was

no site

generation.

The licensee

provided the inspector

a grid study titled,

"PVNGS

-29-

525

Kv Voltage Regulation

Study Report."

This report concluded that the grid

could reliably be maintained

above

98 percent,

without site generation

support.

10. 1.4

Inspector Actions During This Inspection

and Conclusions

The inspector

reviewed licensee

progress

on the long term corrective actions

and determined that the licensee

was still within their internal

schedule for

completion of these actions.

The inspector questioned

control

room personnel

in Unit

1 and determined that they were familiar with offsite power

requirements

and procedure

41AO-1ZZ57.

The inspector

reviewed

updated

licensee calculations.

Based

on this review and the more detailed review

described

in Inspection

Report 50-528/95-12,

the inspector

concluded that the

licensee

had taken corrective actions to identify when

a low voltage condition

existed

on the grid,

and specified proper actions to take to ensure that Palo

Verde units remained

operable

and complied with Technical

Speci.fication

shutdown criteria.

The inspector also noted that the lice'nsee

had provided

reasonable

assurance

of their-intent to complete the long term corrective

actions discussed

above.

I

'

I

,

f

I

0

ATTACHMENT 1

1

PERSONS

CONTACTED

T

B.

R.

R.

B

R.

W.

A.

J.

D

W.

G

C

J

M.

Arizona Public Service

Com an

Cannon,

Department

Leader,

Nuclear Engineering

and Projects

Chapin,

Department

Leader,

Mechanical

Maintenance

Flood,

Department

Leader,

System Engineering

Fullmer, Department

Leader,

Nuclear Assurance

Grabo,

Section

Leader,

Nuclear Regulatory Affairs

Hazelwood,

Engineer,

Nuclear Regulatory

Affairs'de,

Director, Operations

Krainik, Department

Leader,

Nuclear Regulatory Affairs

Levine, Vice President,

Nuclear Production

Mauldin, Director, Maintenance

Montefour, Senior Representative,

Strategic

Communications

Overbeck,

Vice President,

Nuclear Support

Seaman,

Director, Nuclear Assurance

Vellota, Director, Training

Winsor, Section

Leader,

Mechanical

Maintenance

Enginee'ring

1.2

NRC Personnel

e

D. Kirsch, Chief,

Region

IV Reactor Projects

Branch

F

K. Johnston,

Senior Resident

Inspector

D. Garcia,

Resident

Inspector

J. Kramer, Resident

Inspector

1.3

Others

F. Gowers, 'Site Representative,

El

Paso Electric

R. Henry, Site Representative,

Salt River Project

Denotes those present

at the exit interview meeting held

on December

20,

1995.

The inspector also held discussions

with, and observed

the actions of, other

members of the licensee's

staff during the course of the inspection.

2

EXIT MEETING

An exit meeting

was conducted

on December

20,

1995.

During this meeting,

the

inspectors

summarized

the scope

and findings of the report.

The licensee

acknowledged

the inspection findings documented

in this report.

The licensee

did not identify as proprietary

any information provided to, or reviewed by,

the inspectors.

ATTACHMENT 2

ADV

CRDR

EDH

ERFDADS

gpm

HLO

HVAC

'ER

HSSV

NPSH

RCS

RWLIS

SBCS

LIST OF ACRONYHS

Atmospheric

Dump Valves

Auxiliary Feedwater

Auxiliary Operator

Condition Report/Disposition

Request

Containment

Spray

Electrical Discharge

Machining

Engineering

Evaluation

Request

Emergency

Response

Facility Data Acquisition Display System

Feedwater

Control System

Gallons

Per Hinute

High Level Over-ride

Heating, Ventilation and Air Conditioning

Licensee

Event Report

Main Steam Safety Valve

Net Positive Suction

Head

'

Reactor Coolant System

Refueling Water Level Indicating System

Steam

Bypass Control System