ML20128D814

Insp Rept 50-461/93-02 on 930104-15.No Violation Noted.Major Areas Inspected:Licensee Response to GL-89-10, Safety Related MOV Testing & Surveillance

ML20128D814
Person / Time
Site:	Clinton
Issue date:	01/29/1993
From:	Jeffrey Jacobson, Replogle G, James Smith NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III)
To:
Shared Package
ML20128D673	List: ML20128D671 ML20128D814
References
50-461-93-02, 50-461-93-2, GL-89-10, NUDOCS 9302100234
Download: ML20128D814 (14)

See also: IR 05000461/1993002

Text

.

.

U. S. NUCLEAR REGULATORY COMMISSION

REGION III

Report No.: 50-461/93002(DRS)

Docket No.: 50-461 License No.: HPF-62

Licensee: Illinois Power company

_

500 South 27th Street

Decatur, IL 62525

Facility Names clinton Power Station

Inspection Ats- Clinton site, clinton, IL 61727

Ir.spection conducted: January 4-15, 1993

Inspectors: / /Aff / ,_ VS7/93

G. D Replople / Date'

f gg

, ; Wss.Lht -/d V/R

./F.. Smith \

~ -

Date

Approved ys' [ hAMI

J. ; A. Jacobson, chibf

/~ S -9.Q

Date

M te ials and Processes Section

Insoection Summary

Insoection conducted durina Januarv 4-15, 1993 (Recort Ran_

50-461/93002fDRS))

Areas Insoected: Announced safety inspection of-the licensee's-

response to Generic Letter (GL) 89-10',_ " Safety-Related Motor-

Operated Valve (MOV) Testing and" Surveillance" (2515/109).-

Results - No violations were identified.

The licensee demonstrated strengths in the following areas:

• Probabalistic risk assessment (PRA) techniques were used to

-identify the MOVs_most important to safety. These MOVs were

given the highest priority in the program.

• The program reflected significant management attention and-

the coordination between the different participating

departments was excellent.

The trending program for-the evaluation offMOV. failures and.

test results was, excellent.

.

_

9302100234 930201-

gDR ADOCKOSOOp4j1-

___

. -

.

.

s

Inspection Summary 2

The licensos demonstrated weaknesses in the following areast

• A schedule that outlined the testing of MOVs was not in

place at the time of the inspection.

• The corrective actions taken in response to the failuro of

MOV 1SX173B (June 1990) were weak.

. _ _

4

.

TABLE OF CONTENTS

EAGE

1. Persons contacted. . . . . . . . . . . . . . . . . . . 1

2. Inspection of the Clinton Program Developed in >

Rosponse to Generic Lotter 89-10 . . . . . . .. . . . 1

a. Background. . . . . . . . . . . . . . . . . . . . 1

b. Supplomont 3 to Gonoric Lotter 89-10. . . . . . . 1

c. Generic Lotter 89-10 Program Review . . . . . . . 2

(1) Scope of the Generic Lottor Program. . . . . 3

(2) Design Basis Reviews . . . . . . . . . . . . 3

(3) MOV Switch Settings. . . . . . . . . . . . . 4

(4) Design Basis Differential Pressure and Flow

Testing. . . . . . . . . . . . . . . . . 5

(5) Periodic Verification of MOV Capability. . . 6

(6) Schedule . . . . . . . . . . . . . . . . . . 6

d. Associated Programmatic Reviews . . . . . . . . . 6

(1) Design Control for Thormal Overload

Protection . . . . . . . . . . . . . . . . 6

(2) Maintenance. . . . . . . . . . . . . . - . . . 6

(3) Training . . . . . . . . . . . . . . . . . . 7

(4) MOV Failuros, correctivo Actions and

Tronding . . . . . . . . . . . . . . . . . 7

(5) Diagnostics and the Evaluation of Test Data. 10

(6) Walkdown . . . . . . . . . . . . . . . . . . 11

3. Licenseo Self Assessment . . . . . . . . . . . . . . . -11

4. Exit Mooting . . . . . . . . . . . . . . . . . . . . . 12

i-

-'

... - - ,..-. - - .- . - - _

, .. -

. - _ . - - - . _ - . - - _ - . . . __ --. .--. ._

'

,

.

I

i

DETAILS

1. Eersons Contacted

Illinois Power Comoany (IP)

1. J. Perry, Senior Vice President

1. R. Bhat, Supervisor, Mechanical Design

1. J. Cook, Vice President and Manager

1. R. Frantz, Licensing Engineer

1. J. Langley, Director Design and Analysis, Engineering

1. J. Miller, Manager, Nuclear Systems Engineering

Department

1. K. Moore, Director, Plant Technical Staff

1. R. Phares, Director, Licensing

1. J. Puzauskus, Project Manager, Generic Letter 89-10

1. J. Sipek, Supervisor, Regional Regulatory Interface

U. S. Nuclear Reculatory Commission (NRC)

1. P. Brochman, Senior Resident Inspector

1. F. Brush, Resident inspector

1. D. Pickett, Project Manager, NRR

1. Denotes those attending the exit meeting on January 15,

1993.

2. Insnection of the Procram Developed in Resnonse to Generic

Latter-(GL) 89-10 (2515/109)

a. Backaround

A previous inspection of the Clinton GL 89-10 program

was conducted on October 7 through 11, 1991. Due to

the lack of sufficient progress towards program

development at that time, a second inspection was

scheduled. This report details the results from the

second inspection,

b. Suoulement 3 to GL 89-10

During the first GL 89-10 inspection, NRC inspectors

identified four Supplement 3 motor-operated valves

(MOVs) as having marginal capabilities. During the

second inspection, the NRC inspectors verified-that

modifications were planned for the subject MOVs during

the fall of 1993.

The MOV modifications were based on a relatively

conservative valve factor (0.5) and stem friction

coefficient (0.2). However, the differential pressure

1

.

I

.

(dp) used in the calculations included significant flow

losses. The calculated dpa were 836 paid for reactor

core isolation cooling (RCIC) valves 1E51-F063 and

F064, 1014 psid for reactor water cleanup (RWCU) valve

1G33-F001, and 683 psid for RWCU valve 1G33-F004.

The large calculated flow losses were made possible

because of the long distance between the valves and the

location of the postulated breaks. For example, the

break for the RCIC valves was at a point in the piping

which was estimated to be approximately 175 feet from

the valves. The inspectors were concerned because the

approach was not consistent with the spirit of GL 89-10

in that a worst case break might be considered to be

immediately downstream of the valves and not at a more

distant location. Nevertheless, the licensee did not

consider the closer breaks to be within the licensing

basis of the plant.

As a result of the NRC concerns, the licensee performed

an evaluation and determined that the valves could

perform their design basis functions based on a worst

case dp of 1128 paid for the RCIC valves and normal

reactor pressure for the RWCU valves if a less

conservative valve factor of 0.4 and stem friction

coefficient of 0.15 were assumed. For the Clinton

plant, the use of ncrmal reactor pressure, in lieu of

the first relief valve setpoint pressure, appeared to

be justified because tha main steam isolation valves

would not receive an isolation signal during the

accident (reactor pressure would not increase). The

inspectors considered the licensee's analysis to be

adequate, for the interim, since the less conservative

valve and stem friction coefficients have been used at

other nuclear power stations. This issue will be

reviewed during a future inspection when more specific

information is known about the behavior of these

valves.

c. GL 89-10 Prqgram Review

Probabalistic risk assessment-(PRA) techniques were

used to identify the MOVs most important to safety.

Fifty-two MOVs were identified as having a potentially

high impact on safety and were given the highest

priority in the GL 89-10 program. The inspectors

considered the applied use of PRA to be a strength.

Significant resources were dedicated toward completing

the program. Additionally, communication between the

various participating groups appeared to be good and

the staff coordination appeared to be excellent. The

2

.

.

NRC inspectors considered the increased management

attention and staff coordination in the area of MOVs to

be a strength.

(1) Scope of the GL Procram

The scope of the program was reviewed and found to

be acceptable in the previous Part 1 inspection.

At that time, there were 231 MOVs in the GL 89-10

program. Since that time, only one Mov was

removed from the program and justification for-

removal appeared to be acceptable.

(2) Desian Basis Reviews- '

(a) Differential Pressure Reauirements

The. inspectors reviewed Clinton Nuclear *

Engineering Standard ME-03.00 (" General

Guidelines for performing Bounding Valve

Differential Pressure Analysis"), Revision 3,

dated December 18, 1992. - The procedure

described the methodology to be used to

determine the design basis dp for each MOV-

and appeared to be acceptable.

(b) Reduced Voltaae capability

Tne' degraded voltage calculations did not

incorporate the use of locked-rotor current

as specified in.the GL, but used a less'

conservative current from the motor torque-

curves. Although the approach did not appear '

to-be consistent with the licensee's current

commitments, the NRC inspectors did not

pursue issuing a Noticelof Deviation because

- the NRC staff was evaluating the

acceptability of alternative approaches to

degraded voltage calculations. However, when

- this issue is reviewed during-a future

-

inspection the: licensee -will be -expected to

comply.with the docketed commitments.

The inspectors were concerned about theTlack-  ;

of conservatism used in determining the

appropriate motor current for the

calculations. - For example,-additional margin

to account for the inaccuracies in-the motor

curves.was not. included. The-licensee also

failed to account for the difference between

the nominal motor curve voltage.(typically

,

460V) and.the actual voltage applied to~the

3

.

- -

_

s w .e  %

g g -..,e-. a+- ,p gi29y. 9..p- .-.,ww.m.m,..,-,_ - .u.mp u. ,,-.-..$. g -w y y e-- we q w y ,.se+-w-.m y- <-pJ-

.___

.

motor. Since a given motor would generate a

different curve for a different applied

voltage, a motor impedance change would be

expected and would affect the calculations.

Additionally, the calculations assumed that

the locked rotor current power factor would

apply. However, under high loads, at less

than locked rotor conditions, the power

factor is typically higher and would result

in a lower calculated motor terminal voltage.

Cumulatively, the oversights may cause the

calculated motor terminal voltages to be

unrealistically high. The licensee agreed to

address the NRC inspectors' concerns. This

issue will be reviewed during a future

inspection.

The worst case grid voltage used in the

calculations was not always based on the

under-voltage relay setpoint as recommended

in the GL. However, the licensee had-

compensatory measures in place to account for

the differences in the values.- Additionally,

five percent was deducted from the calculated

motor terminal voltages to account for future

electrical load changes. Since the added

conservatisms appeared to bound the above

noted discrepancies, the inspectors

considered the licensee's approach to be

acceptable. This issue will be reviewed

during a future inspection.

The degraded voltage calculation for DC MOVs

only included tne resistance value for one

thermal overload (each circuit contains two

thermal overloads). In response to the

finding, the licensee re-performed the

calculations, incorporated the appropriate

number of thermal overloads and determined

that no significant cafety c">ncerns resulted

from the error. The corrective actions

appeared to be acceptable.

(3) MOV Switch Settinas-

Switch setting calculations and the setting of

switches were performed in accordance with

approved procedures. Switch settings were

recorded in a controlled document and appeared to

be acceptable.

4

j

_ _ _ . _ . . _ _ _ _ _ _ _ _ _ _ _ . __ .

.

.

The stem friction factor and valve factor assumed

in the thrust calculations woro 0.2 and 0.5,

respectively. To support the relatively

conservative factors, the licensee planned to

perform approximately 130 modifications to improve

the capability of the MOV population.

Margin to account for the degradation of the MOVs

was not included in the torque switch settings.

Since torque switches were normally set i

immediately after maintenanco, additional margin

to account for the degradation in performance that

will occur over the specified maintenance period

may be necessary. Typically, the stem friction

factor will increase over the time interval and

will, in effect, decrease the thrust available to

close a valvo at torque switch trip. The licenson

agreed to perform an evaluation of this issue and

stated that appropriate changes to the switch-

setting methodology would be made.

Margin to account for the load sensitive behavior  !

was not included in thrust calculations.

'

Additional margin may nood to be added-to envelop

this effect, when applicable. The licensco

planned to change the switch setting methodology

when more information about the phenomena was

known.

(4) Desian Basis Differential Pressure and Flow

Testing

MOVs found to have the capability to support a

valve factor of 0.5 and a stem friction

coefficient of 0.2 were-considered to have

substantial capability and would not be subjected

to design basis testing. Although the approach

appeared to be consistent with the licensee's

current commitments, the NRC inspectors were

concerned because recent design basis testing at

other facilities has shown that this approach may

not be bounding. In response to the inspectors'

concerns, the licensco-agreed ~to refine the

methodology and consider categorizing the valves

into groups. The methodology would then be shown

to be bounding for each group. If.the approach

could not be shown to be bounding for a given

group then design basis testing would be

considered for these valves.

5

__

p.---7 .y--migu .---u =y-- y 1p y - .s.m -- - - - - - -

.

1

(5) Ecriodie Verification of MOV Canability

The plan for puriodic verification of MOV

capability included static diagnostic testing of

MOVs on an interval consistent with the GL

recommendations. The NRC inspectors informed the

licensco that without adequato justification,

static testing was not an acceptable method of

periodic verification because of uncertainties in

the performance of MOVs under static and design

basis conditions.

The licenson was considering extending the

periodic verification interval for some valves and

exempting ether MOVs from the testing altogether.

This approach was not consistent with

recommendations mado in the GL or the licensco's

current commitments. The licensco agreed that any

chango in the current approach, of a less

conservative nature, would require justification.

This issue will be reviewed during a future

inspection.

(6) Schedula

A schedule that specifies the testing of MOVs was

not in place at the time of the inspection. The

-

licenseo did plan to perform design basis testing

on MOVs within a time framo consistent with the

GL. However, most of the testing would have to be

performed in one refueling outage. Due to the

largo number of valves that may require testing,

completion of the program within the recommended

schedule would be difficult. The lack of a test

schedule was considered to be a weakness.

d. Associated Programmatic Reviews

(1) Deslan Control for Thermal overload Protection

At the time of the inspection, thermal overloads

were intended to be bypassed for all accident

conditions but were active for normal operations

and test conditions. The methods for sizing

thermal overloads appeared to be acceptable.

(2) Maintenance

The NRC inspectors reviewed Clinton maintenance

procedures to evaluate the adequacy of the ,

maintenance program. The program only required

checking the lubrication on valve stems on an 18  ;

1

l

l

l

I

_ .- _. _ - _ ,_

,

.

! month interval and did not require stem cleaning

and re-lubrication. The inspectors were concerned

because visually checking the lubrication may not

identify all forms of degradation. Furthermore,

poor lubrication on unaccessible areas, such as

the stem nut area, may go undetected. For

example, volve 1SX173A failed to stroke during a

routine surveillance. Although the stem appeared

to be well lubricated during the subsequent Valve

Operation Test and Evaluation System (VGTEC) test,

the calculated stem friction coefficient was 1.0,

t'hich was substantially higher than the 0.1 to

0.17 values typically seen on well lubricated

stems. In response to the NRC concerns, the

licensee changed the applicable procedure to

require the cleaning and lubrication of stems on-

an.38 month interval.

A specific schedule for the overhaul of MOVs did

not exist. The need for overhaul was based on

preventive maintenance _and trending results. This

practice may be def!cient because degradation of

MOVs at other nuclear plants was not identified by

preventive maintenance or diagnostic equipment.

Furthermore, degradation which contributed to the

failure of two Clinton MOVs (1SX173A and 1SX173B,

extensive disc guide wear) could have been

detected during an overhaul but would not have

been detected during routine maintenance or static

diagnostic testing. The NRC inspectors encouraged

the licencee to consider the establishment of a

periodic overhaul frequency.

(3) Training

4

The NRC inspectors discussed the Clinton Station

training practiceu with licensee personnel and

reviewed training outlines and records. No

concerns were identified in this area.

, (4) MOV Failures. Corrective Actions and Trendina

The trending program for evaluation of maintenance

problems and MOV test results was excellent. The

program covered an expansive range of data and

appeared to trend all the known trendable

parameters. The inspectors considered the

trending program to be a strength.

Corrective actions taken in response to the

failure of valve 1SX173B (shutdown service water

(SX) pump minimum flow valve)'appcared to be

s 7

. -- - - . .. , - ,

.

'

inadequate in that the licensee failed to take

actions to preclude a similar type failure which

occurred involving the sister valve (ISX173A) in

August 1992. In June 1990, valve 1SX173B failed

to complete its stroke during a surveillance test

of the SX system. After the failure, a root cause

evaluation was not performed. However the

licensee's staff indicated that the valve may have

been used to throttle flow (not a recommended use

for a gate valve) and the throttling was believed

to be the root cause for the damage which was

observed during the December 1990 inspection of

the valve. Steps to address similar degrcdation

which may have existed on-the sister valve were

not taken.

After concerns about the licensee's corrective

actions were raised by the NRC inspectors, the

licensee performed an evaluation and determined

that the failure of the second valve was not

safety cignificant for the following reasons.

(a) If the valve failed to perform its safety

function (establish a minimum flow path for

the SX pump), an annunciator would have

alerted control room operators of-the valve

failure.

(b) An alternate minimum flow path l existed

(through the RHR heat exchanger) and could

have been established in a short time period.

The worst case response time was estimated to

be 10 minutes.

(c) The pump manufacturer was contacted and a

representative confirmed that the pump could

have performed its safety function if it were

run at " dead-headed" conditions for the first

10 minutes.

Based on the evaluation performed by the licensee,

the inspectors agreed that the safety consequences

of the valve failure were minimal. However, the

inspece ors considered the licensee's failure to

'take cbrrective actions to preclude the failure of

the-sister valve 1SX173A to be a weakness.

The root cause evaluation, completed to address

the 1SX173A failure, did not consider all

contributing factors to the event and corrective

actions were not timely (documented in Condition

Report No. 1-92-08-006, " Valve 1SX173A Destroyed

8

- . ..

.

Itself in a Manner similar to 1SX17'20', dated

August 6, 1992). For example, the cause of the

valve failure was determined to be the horizontal

orientation ot the valve which resulted in a high

valve factor (0.6). Additionally, a contributing

factor was an unexplainably high stem friction

f actor (1. 0) that was observed during diagnostic

testing - the stem lubrication, at the time of the

testing, appeared to be in good condition and was

only six months old. The inspectors were

concerned with the evaluation because it failed to

consider, as a contributing factor, the valve's

apparent usage for throttling flow.

Corrective actions, in response to the failure of

1SX173A, were not always timely. For example, to

address potential generic implications that were

associated with other horizontally mounted valves

in the plant, the plan was to wait for the plant's

evaluation of NRC Information Notice (IN) 92-59

(" Horizontally Installed Motor-operated Gate

Valves"), to identify other similar valves prior

to taking additional corrective actions. However,

the due date for the evaluation of IN 92-59 was

June 1993, almost 1 year after the second known

failure. The inspectors did not' con 91 der the

approach to be timely because several years might

expire before an adequate evaluation of the

generic implications reached completion.

Another example of untimely actions was the

response to address the unexplained high stem

friction factor. The licensee planned to wait for

the GL 89-10 trending program to track the

information of interest. The inspectors did not

consider this approach to be adequate because the

trending program was only recently established and

would only analyze test samples taken on an

approximate five year interval (the periodic

verification interval). As such, this approach

would not identify the lubrication degradation

that previously occurred, on the 1SX173A valve,

over a six month period.

In response to the concerns raised by the NRC

inspectors, the licensee modified its corrective

action plans as follows:

(a). An attempt was made to identify other gate

valves used for throttling. None were found.-

9

-. -

_. - _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ _ _ _ _ - _ _ _ -

l

'

.

(b) All horizontally mounted gate valves were

identified prior to the end of the

inspection. The licansee selected a sample

of the valves to test prior to the end of the

next outage. Additional corrective actions

would be considered for this population based

on the results from the testing.

(c) Plans were being developed to monitor the

degradation of the stem friction factor for

valve 1SX173A over the current maintenance

interval. Methods under consideration

included monitoring of motor current as an

indicator of gross degradation and VOTES

testing.

The NRC inspectors considered the licensee's

revised approach to be acceptable.

(5) Diacnostics and the Evaluation of Test Data

The program included the use of the VOTES

diagnostic equipment system during static and dp

testing. The planned use of the diagnostic

equipment appeared to be acceptable.

The NRC inspectors reviewed procedures ME-05 (MOV

Performance Review Following Dynamic Testing"),

Revision 1, dated December 17, 1992 and ME-07

(" Engineering Evaluation of MOV Test Results"),

Revision 1, dated December 17, 1992. Both

procedures provided instructions for the

evaluation of data from design basis testing.

Although the procedures appeared to represent a

good first attempt at evaluating test data,

improvements are needed. The inspectors expressed

the following concerns to the licensee staff.-

(a) Valve factors and motor capability were not

evaluated in the opening direction. Since

opening requirements are often more limiting

than the closing requirements, it would be

appropriate to evaluate MOV capability in

both directions.

(b) Pull-out forces were recorded but the running

loads were subtracted from the term prior to

comparing the values to valve and actuator

structural limits. The pull-out forces

should have been compared to the structural

limits without subtracting the running load.

10

i

I

i

A _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

_ _____-

.

(c) Appropriate accuracy adjustments, in

accordance with recommendations from the

diagnostic equipment manufacturer, were not

made for data outside the specified

calibration range. As a result of the

oversight, the actual pull-out forces may be

higher than originally documented.

(d) The calculated valve factor was

inappropriately based on the thrust observed

at seat contact. This point was not always

the highest thrust recorded during the stroke

of the valve. For example, several MOVs

showed higher thrust readings prior to seat

contact. As such, the calculated valve

factor was lower than the actual worst case

valve factor observed during the dp test.

The licensee agreed with the inspectors'

assessment of the procedures and planned to make

appropriate changes.

The inspectors considered the safety significance

of the procedural oversights to be minimal because

only four MOVs had been subjected to design basis

testing at the time of the inspection. All the

MOVs appeared to have substantial margin.

(b) Walkdown

The NRC inspectors performed a walkdown to observe

design basis testing on 1SXO96A as well ac the

general condition of other MOVs. The inspectors

did not identify any significant concerns.

3. Licensee Self-Assessment

The self-assessment program was acceptable, but narrow in

scope. It thor 'aghly reviewed the areas covered,' identified

shortcomings, and provided a mechanism-for tracking

corrective action and for confirming closure. However, it

was focussed primarily on the administrative portion of the

MOV program and did not encompass the range of programmatic

material normally reviewed by an NRC inspection of the GL 89-10 program. Such expanded coverage is considered to be

essential because it provides an early opportunity for.the

licensee to obviate or mitigate program deficiencies.

11

)

..

._ . .

.

,

4. Exit Meetina

The NRC inspectors met with the licensee representatives

(denoted in Paragraph 1) at the conclusion of the inspection

on January 15, 1993. The inspectors summarized the purpose,

scope and findings of the inspection. The inspectors also

discussed the likely informational content of the inspection

'

report with regard to documents or processes reviewed by the

inspectors during the inspection. The licensee did not

identify any such documents or processes as' proprietary.

12