ML17331B344

Results:

Based

on the results of this inspection,

there

were three violations

identified that pertained

to procedures

(paragraph

5.a); correct'ive actions

to

a previous event (5.c);

and equipment status

(S.d).

There were three

unresolved

items identified that pertained

to pre job briefings (5.a),

training of operators for infrequently performed plant evolutions (S.b),

and

work scheduling

and planning 5.e).

One inspection followup item was

identified that pertained

to lack of level indication between

the bottom of

, the pressurizer

and the reactor vessel

flange (5.d).

The following is

a

summary of the licensee's

performance

during the draining of the

on

February

16 through

18,

1994:

0 erations

The performance

of the operations

department

during the February

16 through

18,

1994, draining of the

RCS was considered

poor.

The

NRC staff is concerned

with the failure of the control

room operators

to properly control

a

significant plant evolution such

as the draining of the

The staff is

also concerned

with the failure of the contr'ol. room. operators

to understand

9400140029

94040i

PDR

ADOCK 050003i5

6

PDR

the importance of maintaining identical pressure

throughout the

RCS when

draining,

and also the effects

on

RCS level during rapid depressurization.

Maintenance

Overall, the work planning

and scheduling

performed for the draining of the

RCS was poor.

An outage

schedule

allowed

an air eductor to be installed

on

the reactor vessel

head vent with inappropriate plant conditions.

This

contributed to the perturbations

in the

RCS level during the draining of the

on February

17,

1994.

There were other concerns

noted with work planning

and scheduling that did not

contribute to the event,

but were indicative of poor work planning

and

scheduling.

Both trains of reactor vessel

level indication system

(RVLIS)

were taken out of service with the

RCS level at 620 feet

and

5 psig pressure,

in preparation

to remove the reactor vessel

head.

Other examples

include:

There

was several

hours of delay in venting sight glass

NGG-100

and

'evel

instrument NLI-112'due to ALARA concerns,

because

the containment

purge

system

was not in service

and workers did not obtain

enough

portable tubing.

There

was

a delay in test'ing the eductor after installation because

the

solenoid valve needed for the operation of the eductor

was installed in

a new'ocation.

This was not considered

during the preparation of the

work package.

,

Persons

Contacted

Details

Consumer

Com an

E.

E.

A. A.

K. R.

L. S.

-~ J.

E.

B. A.

T.

P.

F.

D

L

L. J.

T. K.

S. A.

P.

G.

J.

S.

L. H.

G. A.

D.

C.

H. L.

Fitzpatrick, Senior Vice President-Nuclear

Generation

Blind, Plant Manager

Baker, Assistant Plant Manager-Production

Gibson, Assistant Plant Manager-Projects

Rutkowski, Assistant Plant Manager-Technical

Support

Svensson,

Executive Staff Assistant

Beilman, Haintenance

Superintendent

Carteaux,

Training Superintendent

Noble, Radiation Protection Superintendent

Matthias, Administrative Superintendent

Postlewait,

Design

Changes

Superintendent

Richardson,

Operations

Superintendent

Schoepf,

Project Engineering Superintendent

Wiebe, Safety

8. Assessment

Superintendent

Vanginhoven, Site Design Superintendent

Weber,

Plant Engineering Superintendent

Loope, Chemistry Superintendent

Horvath, guality Assurance

Supervisor

2.

Denotes

those attending the exit interview conducted

on Harch 21,

1994.

The. inspectors

also

had discussions

with other licensee

employees,

including members of the technical

and engineering staffs, reactor

and

auxiliary operators,

shift engineers

and foremen,

and electrical,

mechanical

and instrument maintenance

personnel,

and contract security

personnel.

Descri tion of Event

On February

12,

1994, Unit

1 was shut

down to begin

a scheduled

refueling outage.

On February

17,

1994, preparations

were

made to vent

and drain the reactor coolant

system to approximately two feet below the

reactor vessel

flange to remove the reactor vessel

head in preparation

for core offload.

The licensee

also planned to install

an air eductor

on the reactor vessel

vent line to facilitate the 'removal of conoseals.

Due to inadequacies

in the drain down procedure

and ineffective planning

and control

by operations

personnel,

the reactor coolant system level

was off sacle for approximately

a five hour period.

The inspector concluded that the draindown of the

RCS was not adequately

performed since there

was

a period of five hours

when

RCS level

instrumentation

was off scale.

Fortuitously, there

was

no safety

significance to this event since,

at all times,

adequate

decay heat

removal for the reactor core remained available.

However, considering

corrective action needs to be in place to prevent recurrence

of these

events.

3.

Initial Conditions

Prior to the event,

the residual

heat

removal

(RHR) system

was in

service,

maintaining desired

RCS temperature.

Two pressurizer

power

operated relief valves

(PORV) were open to connect the pressurizer

with

the pressurizer relief tank and the

RCS to maintain the entire

system at

5 psig.

The initial conditions prior to commencing the draining of the

RCS was

as follows:

4.

~

RVLIS train

A was out of service

(OOS).

~

RVLIS train

B wide range indication was in service.

Pressurizer'cold

calibration level

was at 85 percent.

~

RCS was pressurized

to

5 psig.

Se

uence of Events

The sequence

of events

was determined

by interviews with licensee

personnel

on shift at the time of the event,

the review of logs,

and the

review of instrumentation strip chart recorders.

The times noted in the

following sequence

of events

labeled with an "(A)" were approximate

times

based

mostly on personnel

interviews.

Wednesda

Februar

16

1993

Shift supervisor for the

1830 to 0630 shift calculates

volume of water

to be drained,

including steam generator

(SG) U-tubes.

However, the

calculation

was not logged or communicated

to subsequent

shifts.

Thursda

Februar

17

1993

0413

Shift supervisor

authorized

the start of the draindown of

the

RCS to the reactor water storage

tank

(RWST) from

82 percent pressurizer

level with only the wide range of

RVLIS in service.

0615

The

RCS drained to 8 percent

and stabilized.

Shift change

in progress.

0900

(A)

Opened bullseye flow indicator on reactor

head vent line.

0930

(A)

Sight glass

NGG-100 vented

and in service.

0936

1100

(A)

1120

(A)

1402

1830

(A)

Started

draindown from 8 percent at

5 to

10 gallons per

minute

(gpm).

Pressurizer

level at

0 percent.

For approximately

20 minutes with no

RCS level instrument available,

draining

was cautiously performed at

5 to 10 gpm..

4

RVLIS upper plenum on scale at

100 to 99 percent

and

tracking.

Stabilized

RCS level at 620 feet (two feet below reactor

vessel

flange)

and

5 psig.

Bullseye

was isolated which resulted in the reactor vessel

head

no longer in communication with the pressurizer

and

pressurizer relief tank (PRT).

1830

(A)

Started to isolate

RVLIS for head removal.

1830

(A)

Shift change

(1830 to 0630) in progress.

Started installation of air eductor.

The decision to

isolate

RVLIS and install eductor

was the step in the

sequence

that led to the

RCS level perturbations.

1900

(A)

Air eductor installed.

1900

(A)

1900

2030

2045

(A)

RVLIS taken out of service.

End of shift change with RCS level at the 620 feet elevation

and

5 psig.

When the air eductor

was tested,

a two foot decrease

in

level

was indicated

on NLI-112.

By testing the eductor,

the

reactor vessel

pressure

was inadvertently vented to 0 psig

with the

PRT, pressurizer,

and level instrumentation

being

maintained

at

5 psig.

Also,

some draining of the

SG U-tubes

occurred.

Air eductor tested

again;

a one foot decrease

in level

was

indicated

on NLI-112.

2100

2230

(A)

Rapidly opened

PRT vent valve RC-148 to vent

RCS from 5 psig

to 0 psig.

This caused

the Steam Generator

(SG) U-tubes to

drain,

and level

on sight glass

NGG-100

and NLI-112 rapidly

increased

to off-scale high.

Draining of RCS continues.

2000 gallons drained

from RCS based

on

a 0.5 percent

increase

in RWST.

Draining was stopped

since

NGG-100 was

still not back on-scale.

5

2230

(A)

Shift personnel

caucused

and calculated that approximately

13,000 gallons of water

had drained from the

SG U-tubes.

This ca'iculation

was

based

on

an 'assumption

that. 11.5 feet

of 'water from the

SG U-tubes

was drained

due to the venting

of the

RCS from 5 to 0 psig at 2030.

The decision

was

made

to slowly drain

(40 to 60 gpm) until level

was back on-scale

on NGG-100.

Februar

18

1994

0030

(A)

0200

(A)

Recommenced

draining of the

RCS to obtain

an

on scale

reading for sight glass

NGG-100.

Approximately

5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after going off-scale,

level

indication was back on scale

on NLI-112 and sight glass

NGG-100;

0330

(A)

RCS level

was stable

at 620 feet

and

0 psig with the

SG U-tubes-vented.

0455,

Placed air eductor in service to remove conoseals.

Ins ection Results

The inspector's

review of the activities for the draining of the

RCS on

February

16 through

18,

1994, identified the following concerns

in the

areas of procedures,

training, corrective actions to

a previous event,

equipment status,

and work scheduling

and planning.

a.

Procedures

The inspector identified the following concerns with procedures

01-OHP 4021.002.005,

"RCS Drain to One to Two Feet

Below Reactor

Vessel

Flange with Fuel in Core";

PHI-4090, "Criteria for

Conducting Infrequently Performed Tests or 'Evolutions";

and OHI-

221

"Maintenance of Operations

Department Logs,":

1)

Procedure

01-OHP 4021..002.005

"RCS Drain to One to Two Feet

Below Reactor

Vessel

Flan

e with Fuel in Core"

The inspector

reviewed procedure

01-OHP 4021.002.005

to

determine if the procedure

was adequate

to control the

draining of the

This review identified the following

concerns:

The draindown procedure

allowed the shift supervisors

to drain the

RCS without having the

RVLIS in service.

RVLIS, if available,

provided

a mechanism for trending

-

RCS level

between the top the reactor vessel

and the

vessel

Shift supervisors

invoked this option

twice while draining the

The first time,

when

commencing the draindown, train

(A) of narrow level

RVLIS was out of service for the modification to

convert the reactor protection

system from analog to

digital,

and train (B) of RVLIS was out of service for

breaker maintenance.

The wide range indication for

train (A) of RVLIS was available.

Also, all of RVLIS

was taken out of service with the

RCS level at

620 feet

and

5 psig pressure,

in preparation

to remove

the reactor

vessel

head.

This premature

removal of

both trains of RVLIS resulted

in operators

having

no

RCS level indication for a 5-hour period when the

was depressurized

shortly after installing the

eductor.

e

The procedure

did not provide the volume of RCS to be

drained or the volume of water expected

from the steam

generator

(SG) U-tubes

when the

RCS was vented from

5 psig to 0 psig.

Therefore,

when the

RCS was vented,

resulting in an off scale indication

on sight glass

NGG-100, the operators

did not know the volume of

.

water to be drained

from the

RCS to restore

RCS level

indication without performing calculations.

The procedure

did not contain

adequate

instructions to

slowly vent the reactor coolant

system

from 5 psig to

0 psig in a controlled manner.

This resulted

in the

rapid depressurization

of the

and subsequent

lost

of level indication.

The procedure

did not contain

any references

between

plant elevations

and the various level instruments.

Therefore,= operating

crews

used personal

not'es

from

RCS level

based

on

wide range

RVLIS.

The procedure

did not clearly define the plant

conditions required

(RCS at

0 psig) to isolate the

bullseye

and install the air eductor.

2)

Lo

Kee in

and Shift Turnovers

Procedure

OHI-221 "Maintenance of Operations

Department

Logs," Revision

15, step 3.2. I.b for the control

room log

and step 3.3. I.b for the shift supervisor log book, requires

that all significant shift activities

and events,

and all

pertinent information that would assist

in the

reconstruction of a plant event'e

recorded.

The logs were

inadequate

in that the following key plant evolutions were

not properly logged:

~

The calculated

amount of water that was needed

to be

drained

from the

PRT to achieve

a

5 percent level, the

volume of water actually drained,

and the flow path

used to the Reactor Coolant Drain Tank.

The time when

RCS level

was

no longer available

from

pressurizer

level indication.

The time when

RCS level returned

on scale

on wide

range

RVLIS.

The time that

RCS level went off scale

high on sight

glass

NGG-100

and the time

RCS level returned

on

scale.

Additionally, calculations for volume changes

were performed

on scratch

paper

and not passed

from one crew to another

during shift turnover.

These calculations

were not kept

which resulted

in some operators

not knowing the quantity of

coolant to be drained.

For example,

the evening shift on

February

16,

1994, did calculations for the quantity of

water needed

to be drained

from the

RCS to reach the reactor

vessel

flange,

including the water that would drain when

from the

SG U-tubes.

During the event,

when

RCS level

was

off scale

high on sight glass

NG-100, the shift calculated

that 13,000 gallons of water was

needed to be drained to

obtain

an

on scale reading

on NGG-100.

This calculation

was

written on scratch

paper,

and then selected

information was

entered

into the log.

Also, during this draindown,

there

were several

late entries

in the log relating to important

information.

Based

on the'bove,

the failure to have

an appropriate

procedure

to control the draining of the

RCS,

and the failure to record in

the control

room log or shift supervisor

log book significant

shift activities

and events with other pertinent information that

would assist

in the reconstruction of the reactor coolant

system

draindown event is considered

a violation of Title 10 of the

Code

of Federal

Regulations,

Part 50, Appendix B, Criterion V.

(50-

315/94004-01(DRP) )

During the'eview of the event,

the inspectors

determined that

a

pre-job briefing for draining the

RCS had not been

performed

by

the licensee's

management.

Discussion with licensee's

management

determined that procedure,

PNI-4090 "Criteria for Conducting

Infrequently Performed Tests or Evolutions," did not require

a

pre-job briefing with plant management

in attendance

for each

shift involved in the draining of the

The procedure

does

allow management

discretion to conduct

a pre-job briefing for

plant evolutions,

but this discretion

was not invoked for the

,draining of the

on February

16 through

18,

1994.

The

inspectors

were concern that

a pre job briefing was not conducted,

and this matter is considered

an unresolved

item pending further

NRC review (50-315/94004-02).

8

T~rainin

The inspector reviewed the circumstances

that led to the

installation of the air eductor with RCS pressure

at

5 psig.

Through discussions

with licensee

personnel,

the inspectors

determined that the installation of the, air eductor

was placed

on

the schedule without adequate

consideration of the plant status

by

the control

room.

The control room's decision to allow the

installation of the air eductor

was

based

on the refueling outage

schedule,

which identified the necessary

plant condition to

install the eductor

as

RCS level at 620 feet (approximately I foot

below 'the flange).

There was

no consideration

in the schedule for

RCS pressure

during the installation of the eductor.

When the

eductor

was tested,

the reactor vessel

head

was vented,

which

resulted

in the reactor

vessel

being at

a different pressure

(0 psig) than the pressurizer,

PRT,

and the level instrumentation

{5 psig).

This resulted

in an

RCS level indication lower than

actual level.

Also, shortly after testing the eductor,

the control

room directed

operators

in the containment to vent the

RCS through

a connection

to the

PRT.

With the reactor vessel

head at

0 psig; the rapid

venting of the

PRT and pressurizer

resulted

in an

RCS level

change

due to the large volume of coolant drained

from the

SG U-tubes

into the reactor vessel.

The draining of the

SG U-tubes to the

reactor vessel

resulted

in

RCS level going off scale

on sight

glass

NGG-100 for approximately

a 5-hour period with no other

level indication available.

The operators

were not aware of the

effects of rapidly depressurizing

the

This was evident

because

of the operators'ubsequent

action to drain

2000 gallons

from the

RCS in an attempt to restore

RCS level indication.

The

operators

proceeded

in the face of uncertainty

by not assessing

the loss of level indication prior to draining the

2000 gallons.

Based

on the above,

the inspectors

had

a concern with the training

of operators

for infrequently performed plant evolutions.

This

matter is unresolved

item pending further review by the licensee

and the

NRC {50-315/94004-03{DRP)).

The licensee

has

agreed

to

respond to this item by August I, 1994.

of the

During the draining of the

RCS from the bottom of the pressurizer

to the top of the reactor

vessel

head

on February

17, I994,

no

flow through the bullseye

was observed

by an operator

and verified

by

a second operator.

The bullseye

was installed

on the reactor

vessel

head vent to monitor the status of the

RCS level

between

the bottom of the pressurizer

and the top of the reactor vessel

head during draining.

If flow existed through the bullseye,

the

operators

knew that the

RCS level

was between

the bottom of the

pressurizer

and the top of the reactor

v'essel

head.

head during draining.

If flow existed through the bullseye,

the

operators

knew that the

RCS level

was between the bottom of the

pressurizer

and the top of the reactor vessel

head.

The failure to see flow through the bullseye also occurred during

the draindown of the

RCS on August 5,

1993,

when the reactor

coolant system

was being drained to repair

a reactor

head

conoseal.

As a result,

the reactor

was drained to a lower level

than anticipated prior to valving in sight glass

NGG-100.

The

shift's explanation

in August,

1993,

was that when the bullseye

was monitored for flow, the level

had.- already

been, drained

below

the bullseye connection

on the reactor vessel

vent line.

A condition report

(CR 93-1267)

was issued for the August 5,

1993,

draindown event.

The corrective action consisted of a revision to

procedure,

01-OHP 4021.002.005,

"RCS Drain to One to Two Feet

Below Reactor Vessel

Flange'with

Fuel in Core."

The revision

required that sight glass

NGG-100 be placed in service at

a level

of 5 percent in the pressurizer.

The corrective action was narrow

in scope

and did not adequately

resolve the problem,

since

two

operators

did not observe flow thr ough the bullseye during the

February

16 through

18,

1994, draining of- the

The failure to

provide adequate

corrective action to preclude the inability to

use the bullseye during the draining of the

RCS is considered

a

violation of 10 CFR Part 50, Appendix B, Criterion XVI (50-

315/94004-04(DRP)).

E ui ment Status

The inspectors identified that the currently installed

RCS level

instrumentation did not cover the full span

from the top of the

pressurizer

down to the reactor vessel

There

was

approximately

a three foot area

below the pressurizer

and

above

the top of the reactor vessel

where there

was

no installed level

instrumentation.

In addition, the licensee

used

RVLIS

instrumentation, if available,

as

a trending device to monitor

RCS level

between the top of the reactor vessel

and the vessel

At the vessel

flange, the level fell within the scale of

the

NGG-100 gauge glass.

The inspector noted that

01-OHP

4021.002.005

did not require that

RVLIS be available during the

draindown evolution.

The matter regarding full span level

coverage is considered

an Inspection

Followup Item pending further

NRC review (50-315/94004-04(DRP)).

Also, the inspectors

were concerned with the operators'se

of

pressurizer relief tank

(PRT) level instrument indicator,

1-NLA-

351, that

had

a "defective" tag which had

been placed

on the

instrument

on November

18,

1992.

This

PRT level indication was

required to be used during the drain down because

the drain down

procedure,

Ol-OHP 4021.002.005,

required

PRT level to be less

than

5 percent prior to commencing

the draining of the

The

operating

crew drained the

PRT to the reactor coolant drain tank

10

appeared

to track, the use of an instrument that was

known to have

accuracy outside of necessary

limits to monitor level in the

PRT,

is considered

a violation of 10 CFR Part 50, Appendix B,

Criterion XII(50-315-94004-05(ORP)).

e)

Schedule

and Plannin

Poor work scheduling

and planning contributed to the perturbations

.

in the

RCS level during the draining of the

on February

17 and

18,

1994.

The following were examples of poor work scheduling

and

planning:

I'oth

trains of RVLIS were taken out of service with the

level at 620 feet

and

5 psig pressure

in preparation

to

remove the reactor vessel

head.

The air eductor

was installed

when plant conditions were not

appropriate for the activ'ity.

The schedule identified that

the eductor

was to be installed

when the

RCS level

was at

620 feet elevation,

without requiring the

RCS to be

depressurized

to 0 psig.

As

a result,

when the

RCS level

was at 620 feet,

maintenance

workers were staged

in

'ontainment

to install the air eductor.

Expediters

repeatedly

telephoned

the control

room requesting

authorization to install the air eductor.

This external

pressure

was

a contributing factor in the

operators'ecision

to install the air eductor with the wrong plant

conditions.

There

was several

hours of delay in venting sight glass

NGG-

100

and level instrument

NLI-112 due to ALARA concerns,

because

the containment

purge

system

was not in service

and

workers did not obtain

enough portable tubing.

The total

delay in putting these

instruments

into service

was

approximately five hours.

However, this delay did not

impact the draining of the

RCS,

since level

was still in the

pressurizer

during this time.

There

was

a delay in testing the eductor after installation,

because

the solenoid valve needed for the operation of the

eductor

was installed in a new location.

This new location

resulted

in maintenance

workers not having enough air line

to test the eductor.

This change

in location of the

solenoid

was not considered

during the preparation of the

work package.

The licensee

issued

Condition Report 94-290

to evaluate this inadequate

work planning

.

Based

on the above,

concerns with work scheduling

and planning are

considered

an Unresolved

)tern pending further review by the

NRC

(50-315/94004-06(DRP)).

11

l'

Root Cause

Anal sis

The inspectors

determined that the major casual

factors for the

level perturbation

event that resulted

in the loss of level indication

for five hours

on February, 17,

1994 were:

~

A draindown procedure that did not provide adequate

instructions

.

to vent the reactor coolant

system from 5 psig to 0 psig in a

controlled manner.

This resulted

in the rapid depressurization

of

the