Text

TER on IPE Back End Analysis
	ML20140B658
Person / Time
Site:	Clinton
Issue date:	05/30/1996
From:	Meyer J, Hanry Wagage; SCIENTECH, INC.
To:	; NRC
Shared Package
ML20140B649	List: ML20140B642; ML20140B652; ML20140B658; ML20140C337;
References
	CON-NRC-05-91-068, CON-NRC-5-91-68 SCIE-NRC-236-95, NUDOCS 9703200208
	Download: ML20140B658 (39)
	v • d • e

l SCIE-NRC-236-95 Pu Clinton Unit 1 Technical Evaluation Repon on the Individual Plant Examination Back-End Analysis i r..

It t=

t s-

,. H. A. Wagage J. F. Meyer ey -

Prepared for the U.S. Nuclear Regulatory Commission Under Contract NRC-05-91-%8-43 May 1996 SCIENTECH, Inc. ,

11140 Rockville Pike, Suite 500 Rxkville, Maryland 20852 9

e

~'%

Q Dn 31.onLaL > 2 793 -- -

SCIE NRC-236-95 m

Seo Clinton Unit I

_ Technical Evaluation Repon on the Individual Plant Examination Back-End Analysis r.-

It tw I

, . H. . A. Wagage J. F. Meyer -

,?

.a .

Prepared for the U.S. Nuclear Regulatory Commission Under Contract NRC-05-91-068-43 May 1996 SCIENTECH, Inc. .

11140 Rockville Pike, Suite 500 Rockville, Maryland 20852 M*m %w

- Q D') 3.2. 0O.502 2 2 7fs3

.N , .

=i SCIE-NRC-236-95 m

f .e u Clinton Unit 1 Technical Evaluation Report

~

on the Individual Plant Examination .

. Back-End Analysis m

,j t

6.

9 9 H. A. Wagage i J. F. Meyer m .

.J .

M m

,4 w

' r$ '

Prepared for the U.S. Nuclear Regulatory Commission Under Contract NRC-05-91-068-43 '

i May 1996 SCIENTECH, Inc.

p 11140 Rockville Pike, Suite 500 J Rockville, Maryland 20852 sa

!I d

!. .a~ .

4 i '

j TABLE OF CONTENTS

E. Executive Summary ..........................,........ E1 l 1. INTRODUCTION ..............,.................... 1

- 2. TECHNICAL REVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 2.1 Licensee's IPE Process . . . . . . . . . . .. . . . . . . . . . . . . . . . . . 4

2.1.1 Completeness and Methodology .................. 4 i - 2.1.2 Multi Unit Effects and As-Built /As-Operated Status . . . . . . 5 l ,... 2.1.3 Licensee Participation and Peer Review . . . . . . . . '. . . . . . 5 l . 2.2 Containment Analysis . . . . . . . . . . . . . . . . ............ 6 i

2.2.1 Front-end Back-end Dependencies . . . . . . . . . . . . . . . . . 6 i 2.2.2 Containment Event Tree Development .............. 7 2.2.3 Containment Failure Modes and Timing .. .......... 8

" 2.2.4 Containment Isolation Failure . . . . . . . . . . . . . . . . . . . . 14

.; 2.2.5 System / Human Response ...................... 15 2.2.6' Radionuclide' Release Categories.and Characterization . . . ... 16 2.3 Quantitative Assessment of Accident Progression and

. Containment Behavior . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2.3.1 Severe Accident Progression' .................... 17

~ '

2.3.2 Dominant Contributors: Consistency with IPE Insights . . . . 19 2.3.3 Characterization of Containment Performance . . . . . . . . . . . 20 2.3.4 Impact on Equipment Behavior . . . . . . . . . . . . . . . . . . . 20 2.3.5 Uncertainty and Sensitivity Analysis . . . . . . . . . . . . . . . . 21 u 2.4 Reducing Probability of Core Damage or Fission Product Release . . 23 2.4.1 Definition of Vulnerability . . . . . . . . . . . . . . . . . . . . . . 23 2.4.2 Plant Improvements . . . . . . . . . . . . . . . . . . . . . . . : . . 23 2.5 Responses to CPI Program Recommendations .............. 24 2.'6 -IPE Insights, Improvements, and Commitments . . . . . . . . . . . . . 25

3. CONTRACTOR OBSERVATIONS AND CONCLUSIONS ..'....,... 27

. 4. REFERENCES ..................................... 29

,., Appendix: IPE Evaluation and Data Summary Sheet .

.i tu.

>a Clinton Unit 1 Back-End iii May 1996 I

j- . . .. ;

=

l E. - EXECUTIVE

SUMMARY

i . E.1 Plant Characterization ,

' ~

Clinton Nrc, Station, Unit 1-(CPS) is a General Electric boiling water reactor (BWR-6) l rated at 24 MWt with a Mark III containment. De containment includes a large pool j of wmr, i.e., a suppression pool, where steam from the reactor vessel relief valves and 1 pipe breaks can condense. De CPS containment consists of a right circular cylinder with

! a hemispherically domed roof and a flat base slab. De containment wall is constructed j of reinforced concrete, lined intemally with steel plate. De drywell is also a right

! . circular cylinder located within and concentric to the contai,nment. The drywell wall is j i< rigidly attached to the containment basemat and has an annular concrete slab top. A

removable head is bolted over an opening in the top slab for access to the reactor vessel

~

for refueling operations.

}; -

E.2 Licensee's IPE Process ,

l " The Individual Plant Examination (IPE) program f'or the CPS was based on industry-

~

accepted level 1 and level 2 PRA methods and NUREG-1335, " Individual Plant l Examination Submittal Guidance." The IPE team developed containment event trees

" (CETs) to characterize the containment response to severe accidents for the level 2 or j _

"back-end" analysis. De team examined several severe accident phenomena in detail to i

describe their applicability to CPS and possible use in CET headings. The same analysts i performed both level I and level 2 evaluations of the IPE and integrated the two of them j ,, by continuing to use the sequence equations from the level I results through the sequences in the CETs. This process contributed to the continuity, consistency, and accuracy of the overall project.

\ .>

i The Illinois Power Company (IP) perfonned and managed the CPS IPE. The IPE team was located at the plant site and the members were involved in all aspects of IPE l

j ,~ activities. The IP Nuclear Station Engineering Depanment led the IPE.

! A team composed of senior IP personnel, called the IPE Independent Review Team

, . (IIRT), performed an indepenrient review of the IPE products. The IIRT reviewed the interim and final products, which included the Containment Analysis Repon, in order to j ensure accurate representation of CPS design, maintenance and surveillance schedules and

.- recovery actions throughout the IPE.

i ..

A management oversight team consisting of various department managers and a vice i president ofIP reviewed IPE progress and interim product repons. This review team

. provided assurance that results were reasonable and bases for these results were

, documented adequately, facilitating future use by IP personnel, i

The prime consultant for thc' CPS IPE was the Individual Plant Examination Partnership j '(IPEP), made up of Tenera, L. P., Fauske and Associates, Inc., and Westinghouse 1 Clinton Unit 1 Back-End E1 May 1995 i

,.=- - - -- . . -

.Q ' . ,

i

= l l

Electric Corporation. The IPEP project advisor, who reported directly to the IP technical lead, served as the primary interface between IP and IPEP. )

l

- The primary comments of the review teams concerned modeling accuracy and requests for additional explanation and justification of results and conclusions. These comments were incorporated into the interim products at each stage of the project, before approval- ,

of each respective product. l

~

l

, , In preparation for the IPE, the team performed plant walkdowns to ve'ify r the accuracy of

-l system information, identify special or unusual characteristics of iridividual components I a or their location, and to identify potential recovery actions. I 1f i The utility used January 4,1992, as the " freeze date for data used in conducting the l r CPS IPE. l E.3 Back-End Analysis o The CPS IPE team postulated a core damage frequency (CDF) of 2.6E-5 per reactor year based upon the present, as-operated, CPS reactor, plant, and containment capabilities.

The significant core damage contributors were station blackout (SBO) (long and short

- term) and transients contributing to 37.2% and 52.0% of the total CDF. The team calculated a conditional containment failure probability of 5 % from anticipated transients without scram, and some SBO and high-pressure core damage sequences.

The CPS IPE team identified no accident sequence's that would lead to direct containment bypass. Early containment failures at CPS were dominated by containment overpressurization resulted from ATWS sequences. I. ate containment releases were dominated by containment venting and ove: pressurization failures. i 1

'" j The utility's attention to the issue of hydrogen combustion was not complete. As shown

,, on the , quantified CET for SBO sequences, when AC power is recovered after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months there is only a very small probability of containment failure from combustion, this in j spite of a statement in the text that these burns fail the containment by overpressurization. !

,4 ;

When questioned about the validity of the ex-vessel steam explosion results as reported in i the submittal, the utility personnel gave no justification but only noted that the

- methodology used was as proposed by FAI. This is an additional incomplete aspect of the submittal.

E.4 Generic and Contamment Performance Improvement Issues The licensee addressed the recommendations of the Containment Performance Improvement (CPI) Program and the IPE team '

concluded that an alternate power supply l

_ for hydrogen ignitors was not justified.

)

Clinton Unit i Back-End E-2 ,

May 1995

p b l Of the Mark I Improvements that were recommended. CPS is considering making a

, piping modification to the Fire Protection system to create an alternate water supply. -

This hardware change will not be implemented (if at all), until completion of the IPE for l

! external events and development of a Severe Accident Manag'ement Plan.

' ~

l t

, The IPE team included the backup air supply for the automatic depressurization ' system in I

the IPE models to enhance depressurization reliability. This change is to be considered a !

.- part of the Severe Accident Management Plan.. CPS has fully implemented Revision 4 to l the boiling water reactor emergency procedure , guides, which is reflected in the IPE.

l

- The IPE team found that containment heat removal was adequate or not a significant

! ., factor except in anticipated transients without scram (ATWS) scenarios, where, although

! directed by EOPs, venting and suppression pool cooling were not effective in preventing containment failure. 'Ihe IPE team did not take credit for such cooling and still found

! that the CPS containment failure probability although finite was relatively low.

i

" In response to the NRC staff's Request for Additional Information (RAI) the licensee j

! ., notes that improving the containment venting capability with a hardened vent system was

not required.

l .. E.5 Vulnerabilities and Plant. Improvements

! The criterion for identifying vulnerabilities applicable to back-end analysis was as j . follows:

' Are there any new or unusual means by which core damage or containment failure occur as compared to those identified in other probabilistic risk assessments (PRAs)?

! No vulnerabilities were identified.

l

! After evaluating the containment failure cutsets the IPE team identified the following .

situations minimization of which might reduce the risk of radiological releases from 'the j

,. containment: loss of off-site power, delayed recovery of AC power, failure to isolate the

. containment under station blackout (SBO) conditions, and scram hardware failures.

~

i . E.6 Observations i

! The low probabilities of core damige and containment failure were attributed to the fact j that CPS is one of the newest dygned BWR-6 plants and has a Mark III co6tainment.

j Through the conduct of the II'E program Illinois Power has developed an appreciation of

severe accident behavior at CPS. Utility personnel were involved in performing and l .- reviewing the IPE and IP appears to have understood the severe accident sequences most

) , ,

likely to occur at CPS. The IPE ieam performed MAAP analyses of raresentative i accident sequexes and addressed phenomenological uncertainties of accident progression i through sensitivity studies. The submittal demonstrates that the utility has gained more l Clinton Unit 1 Back-End E-3 May 1995 i

l _ .

550 quantitative understanding of the overall probabilities of core damage and fission product

,, release. The back-end analysis results showed that CPS has a major radiological release frequency that is less than the NRC safety goal of IE-6 per reactor year.

- In addition to the " completeness" issues noted in Section E.4,'SCIENTECH notes that the IPE team appears to have significant amount of credit for back-end system / human interactions. However, as noted in the TER of Human Reliability Reviewer recovery

- failure probabilities have been determined using generic results for EPRI RP-3000-34 and ;

the licensee failure to consider CPS-specific factors in repair or restoration of components j in back-end analysis is a limitation in the IPE. SCIENTECH agrees with this

- observation. I

4 By evaluating the back-end analysis results, the IPE team identified the following potential procedure and plant improvements that could reduce the radiological release frequency from CPS: ..

I l

r"'

Procedure changes for the performance of activities involving the switchyard or 1 j the plant connection to the off-site power system during a loss of off-site power. l l

""

Training in the operation of the diesel generator system in recovering AC power.

e- .

Training in isolating the containment under SBO conditions.

-

Training in the maintenance and operation of scram equipment, and equipment i l

modifications to enhance reactor scram in order to lower the frequency of an ATWS.

Overall, IP appears to have achieved the objectives of GL 88-20.

m

,4 i

ts W

a Clinton Unit i Back End E-4 May 1995

'( ,

ai

1. INTRODUCTION ,

1.1 Review Process This technical evaluation repon (TER) documents the results of the SCIENTECH review

, of the back-end portion of the Clinton Power Station (CPS) Unit 1 Individual Plant Examination (IPE) submittal. [1,2] This technical evaluation repon complies with the

- . requirements for reviews of the U.S. Nuclear Regulatory Commission (NRC) contractor o task order, and adopts the NRC review objectives, which include the following:

To help NRC staff determine if the IPE submittal provides the level of detail

.J requested in the " Submittal Guidance Document, NUREG-1335;

~

To help NRC staff assess the strengths and the weaknesses of the IPE submittal;

,. and

,

To complete the IPE Evaluation Data Summary Sheet.

u Based in part on SCIENTECH's preliminary review of the Clinton IPE submittal, the

. NRC staff subihitted a Request for AdditionalInformation (RAI) to the Illin.ois Power w Company on July 21,1995. De Illinois Power Company responded to the RAI in a document dated November 22,1995. [2] nis final TER'is based on the oiiginal submittal and the response to the RAI.

Section 2 of the TER summarizes our review findings and briefly describes th'e CPS IPE l submittal as it pertains to the work requirements outlined in the contractor task order.

Each ponion of Section 2 corresponds to a specific work requirement. Section 3 presents '

y our overall evaluation of the back-end portion of the CPS IPE based on our review.

Section 3 also outlines the conclusions and insights gained, plant improvements identified, and utility commitments made as a result of the IPE. References are given in Section 4.

,, . The appendix.contains an IPE evaluation and data summary sheet.

1.2 Plant Characterization .

p.

Clinton Power Station is a General Electric boiling water reactor.(BWR-6) rated at 2894 MWt with a Mark III containment. The containment includes a large pool of water, i.e.,

.. the suppression pool, where steam from the ~ reactor vessel relief valves and from pipe breaks can be condensed. Figure 1 (a reproduction of figure 4.1-1, page 4-14 of the submittal) shows a cross-section of the CPS containment.

The containment consists of a right circular cylinder with an internal diameter of 124 feet, a hemispherically domed roof, and a flat base slab. The containment wall is

constructed of reinforced concrete, lined with 1/4-inch-thick steel plate. The lower ,

. section of the containment wall acts as the outer boundary of the

=

Clinton Unit i Back-End 1 May 1996 w

e; j . ,

=

i e

' sen CLINTON CONTAINMENT s u' " m f _

~

SIGNFICANT DATA .E _

=

i .

' E - .- .

3 *d . :::

n i *E

=

t 5

/ =

.i . .

' i

'f

\

i

!- - "E i u .5 4

l

~

E=W. N E-a.is w a

=

- 4 i

j ~

E -

Q' -

g .

.m E

,,,5- en a

l. , ,

= _

E t

- . m 1 .

E

=

,a

=

I r V "E b_-=,".~

1

. ii h h.e". ,

y: . I = - nu .a

. ] :-at v.

F OE

- P.

l : ; .: a : L :.

i esasan :-

a*="aa i

ll. ::::!!:i:...:!::!:!::!:i !:::!: ::::::!

!!!: :!:i: !!!.!!.lljij!!!j:.:.:.:.:.......... .;l .

=E-"

a=== =

. . . een4i.=a, .a

.4..... ,. ,

.a 4= . an4 t.e-

) '

9Wer W.B WWENE BM4..t en.D WP46. .r T ggp g ggq

.m .an.,e a = =r em we l

t u

1 6

i

.t .

. .J Figure 1. Clinton Unit I containment (reproduced from figure 4.1-1, page 4-14 of the 4 ,, submittal) ,

Clinton Unit 1 Back-End 2 May 1996 4 9 l tJ .

i t

( , -. . . - .

. . . . _ - . _ . . .~ . - _ _ - _ - _ - . . - . . - . . - _ - . . _ - . - . - _

M

4

! - The drywell is also a right circular cylinder, with an internal diameter of 69 feet, located i' within and concentric to the containment. The drywell wall is steel lined, reinforced l concrete that is 5 feet thick. The drywell wall is rigidly attached to the containment :

- basemat and has an annular concrete slab top that is 6 feet thick. A removable head is i

bolted over an opening in the top slab for access to the reactor vessel for refueling operations.

i -

,, The following is a summary of major features of the CPS Mark III containment .

l (section 2.4.1, page 2-10):

4 -

3 o

Steel-lined, reinforced concrete containment, with a volume of 1,550,000 ft ;

l 1

~~ 3 i .

Drywell structure with a volume of 246,500 ft enclosed by the containment; 3

Suppression pool with a volume of 135,700 ft , which is accessible from the i drywell and containment;

]

ee l

Two trains of containment spray, suppression pool cooling or shutdown heat ;

' removal; and i

j

A reinforce:1 concrete basemat over 10 feet in depth.

s .

j -

The following are the systems located in the containment that are required to mitigate a - 1 l ,

severe accident: ,

j i . 1

^

Inboard containment isolation valves for various systems;

Automatic depressurization system; i

Combustible gas control system (drywell and containment atmosphere mixing,

hydrogen ignitors, hydrogen recombiners);

j' ,,

Suppression pool and supprerion pool makeup; Containment vent system; and .

Containnient/drywell ventilation systems.

s~

u i

~

i .

( =

j

'l .

' Clinton Unit 1 Back fad 3 May 1996

.Y *

, m

2. TECHNICAL REVIEW h12 2.1 Licensee's IPE Process j , - LL1 Comoleteness and Methodolorv. .

d The CPS IPE submittal contains a substantial amount of information with regard to the recommendations of Generic Letter (GL) 88-20, its supplements, and NUREG-1335.

j .g The submittal appears to be complete in accordance with the level of detail requested in

! NUREG-1335. The methodology used to perform the IPE is described i:lcarly in the m submittal. The approach taken, which is consistent with the basic tenets of GL 88-20, 1 i ;j Appendix 1, also is described clearly r.long with the team's basic underlying assumptions.

The important plant information and data are well documented and the key IPE results 3

and fmdings are well presented.

2 j- The IPE program for the CPS was based on level 1 and level 2 PRA methods described i P in the following NUREG documents (section 1.3, page 1-5):

! w l

NUREG/CR-2300, "PRA Procedures Guide"; ,

1 m .

' l 1 '

NUREG/CR-2815, "Probabilistic Safety' Analysis Procedures Ginide"; and s

.* NUREG-1335, " Individual Plant Examination Submittal Guidance."

The IPE team developed containment event trees ('CETs) to characterize the containment

. response to severe accidents for the level 2 or "back-end" analysis. The team examined

several severe accident phenomena in detail, using industry or CPS experience, analytical work, and CPS-specific parameters. The team developed phenomenological evaluation 1 summaries for these phenomena to describe their applicability to CPS and possible use in W CET headings.

l ,

The same analysts performed both level 1 and level 2 evaluations of the IPE and .

l integrated the two by continuing. to use the sequence equations for the level I results through the sequences in the CETs. This process was to ensure continuity, consistency, l q

~-

and accuracy of the overall project.

==

[ The front-end analysis evaluated accident sequences over a 24-hour mission time. The l n

! g IPE team defined core damage as " reactor level less than two thirds the length of active l fuel for more than 4 minutes or Modular Accident Analysis Program results with fuel l r; temperature of 2200eF.or more" (section 3.4, page 3-222 of the submittal).

i

^

j The back-end analysis evaluated accident sequences over a 48-hour mission time. The

,3 IPE team defined "early" containment failure as that occurring within 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months of the l J onset of core damage. This definition is seemingly inconsistent with the NUREG-1150 definition of early containment failures for boiling water. reactors, which included failures i n occurring within 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months of vessel breach. However, because the postulated vessel Clinton Unit 1 Back-End 4 May 1996 J

. t.a .

j

~

4 failures occurred generally within 3 hours3.472222e-5 days 8.333333e-4 hours 4.960317e-6 weeks 1.1415e-6 months of the onset of core damage, the CPS '

q definition 'of early containment failures is consistent with and more conservative than the l

NUREG-1150 definition. ,

j ._ LM Multi-Unit Effects and As-Built /As-Ooerated Status. l l

The IPE team found no multi-unit effects of importance for CPS. ,

= i In preparation for the IPE, the team performed plant walkdowns to verify the accuracy of .

! system information, to identify special or unusual characteristics of individual components l l

! - or their location, and to identify potential recovery actions. 'Ihe team performed l

u containment and drywell walkdowns to evaluate building characteristics and validate !

MAAP parameter file information. A member of the IPE team and a conschant l

! - performed simulator walkdowns to verify operator actions, both when an operating crew l . was in training and when no simulations were in progress. The IPE team, located at the

, plant site, performed additional walkdowns as necessary to answer specific questions as

, e they arose.

64 l

The utility responses to the NRC staff's Request for AdditionalInformation (RAI) notes

! the following about the freeze date on plant data used to condu.ct the CPS IPE.(page 3, reference 2):

{

) As each model was built during 1989 through 1991, the revisions of drawings and procedures were recorded. At the end of 1991, a review was conducted of i drawing and procedure changes and maintenance history since each model was i produced. This produced an effective freeze date for the submitted examination of

- December 31,1991.

l

. The IPE team appears to have verified that the models used were for the CPS as built and j operated.

LM Licensee Particination and Peer Review.

i .

The Illinois Poiver (IP) Company. performed and managed the CPS IPE. The IPE team

! was located at the plant site and the members were involved in all aspects of IPE j activities. The IP Nuclear Station Engineering Department led the IPE. l 1 r., l A team composed of senior IP personnel, called the.IPE Independent Review Team

} [ (IIRT), performed an independent review of the IPE products. The IIRT was composed -

j of a director and supervisors from the various on-site depanments. Most of the review team members held Senior Reactor Operator licenses. Like the IPE team, all members of.

4 the review team were located at the plant site. The purpose of the IIRT was to review the interim and final products that included the Containment Analysis Repon in order to j ,

. ensure accurate representation of CPS design, maintenance and surveillance schedules and

{ recovery actions in the IPE study. The IPE Technical Lead trained the IIRT with '

assistance from IPEP. ,

l

Clinton Unit 1 Back-End 5, May 1996 i '

O g

. ** + '

(

=

l A management oversight team consisting of various department managers anil a vice i - president of IP reviewed IPE progress and interim product reports. All of these team 4

members were also located at the plant site. 'Ihis review team provided assurance that i results were reasonable and bases for these msults were adequately documented, j -

facilitating future use by IP personnel.

l . .

The prime consultant for the CPS IPE was the Individual Plant Examination Partnership

! - (IPEP), made up of Tenera, L. P., Fauske and Associates, Inc., and Westinghouse i , Electric Corporation. The IPEP project advisor, who reported directly to the IP technical j lead, served as the primary interface between IP and IPEP. Major responsibilities of the IPEP included the following (section 5.2.4, page 5-7):

j a

j

Assist in implementing and interpreting PRA guidance as applied to CPS;

Train IPE group and review groups; and l '

l j' "

Perform a quasi-independent review of the CPS IPE with an IPEP Senior 4 Management Support Team consisting of senior IDCOR people.

l l The primary comments of the review teams concerned modeling accuracy, and requests for additional explanation and justification of the results and conclusions. Comments i ,_.

were incorporated into the interim products at each stage of the project, before approval l

of each respective product. After its modification to reflect review team comments, the l repor. became more readable.

i ..,

Although Illinois Power was actively involved in performing ani! reviewing the CPS IPE, l no specific comments' related to the IP reviews are listed in the submittal.

3 Co'ntainment Analysis i 2.2 l ,, 2.11 Front-end Back-end Denendencies.

1 l The IPE team categorized accident sequences leading to' core damage into classes and

... subclasses. Grouping or binning into classes was based on the following criteria:

I

~

Containment integrity;

{ .-

Primary system integrity; i

Relative timing of core damage; ;

i

Primary system pressure; and '

l i

Failure of critical functions leading to core damag'e . l l

Core damage bins used in the conduct of the CPS IPE were selected based on the l i

guidance provided in Nuclear Management & Resource Council 91-04, " Severe Accident l . Issue Closure Guidelines." These core damage bins were called accident classes and i served as input to the back-end containment analysis. Table 3.1-6 of the submittal shows j grouping process.

i i Clinton Unit 1 Back-En'd 6 May 1996 i

)

_ ... _ _ -.m - -

ei ,

s .

l l As shown in Table 3.1-7 of the submittal, the five classes were further divided into

= '

- subclasses based upon the unavailability of key functions. The accident classes applicable to CPS are shown in Table 1 along with their frequencies.

l .

Table 1. Core Damage Frequency by Accident Class Accident Class Core Damage Percent

Frequency Contnbution l Transients - high pressure (IA) 9.8E-6 37 %

Station blackout (IB) 9.8E-6 . 37%

u Transients - low pressure (ID) 5.7E-6 21 %

~

LOCAs - high pressure

~

(IIIB) 1.3E-8 0%

LOCAs - low pressure (IIIC) 1.lE-6 4%

l ATWS events (IV) 1.4E-7 1%

l , Containment bypass (V) < l .0E-9 0%

8 Overall CDF . 2.6E-5/rx y.r l -

l By grouping core dam' age sequences into accident classes and using those as input to the

- . back-end CETs, the IPE team adequately considered the front-end back-end dependencies

, applicable to CPS.

1 i- - 112 Containment Event Tree Develooment.

The IPE team developed a CET for each accident class listed in Table 1, except for

.- " Containment bypass," which has a core damage frequency (CDF) value of less than l j_ IE-9, the truncation limit. CETs were constructed emphasizing the system behavior the operator can see and control, such as containment pressure and temperature and system e operation. Apart from a CET top event to address hydrogen gas control in the -

containment, the phenomenological issues of accident progressi.on were addressed separately. In those cases in which phenomenological issues were judged applicable to n CPS, their effects were considered in quantifying the CETs. Table 2 lists CET top

,,, events along with applicable CETs. Figure 2 (a reproduction of Figure 4.5-1, page 4-62 of the submittal) shows the CET used for accident class IA.

o .

_ Progression of an accident sequence through' the CETs resulted in a plant damage state (PDS) (CET end state). (This definition of PDS is diffccr.t from that used by most other IPEs and NUREG 1150, which defm' ed PDS as the status of systems and equipment important for the source-term analysis at the onset of core damage.) Each PDS was represented by a four-letter code which signified the reactor pressure vessel and containment status as well as sequence timing. Table 3 (a reproduction of table 4.3-3,

- page 4-29 of the submittal) lists the PDS codes.

l l Clinton Unit 1 Back End 7 May 1996 I

i J

l '

o;< ,

a .

Table 2. CET Top Events and Applicable CETs Top Event IA IB ID uns IIIC IV Containment isolation x x x x x Reactor primary system depressurization x x x x l

I

, Injection before vessel failure x x x x x x Late injection x x x x

- Contamment spay in event of pool bypass x x x x x x

Hydrogen gas control x x x x Containment venting x x x x. x l

Containment fails above suppression pool x x -x x x x

'" x x Long term containment heat removal x x x x AC power recovery to prevent vessel failure x I. ate power recovery and injection x Suppression pool cooling x

.. Split fraction values for CET top events reflected finite probabilities of different paths and thus they included uncertainties.

.- CPS containment event tree development appears to be' complete and in accordance 'with the GL 88-20 and NUREG-1335.

r. W Containment Failure Modes and Timine.

Section 4.4, pages 4-33 through 4-55 of the submittal describes the various potential m containment failures mechanisms and summaries of the evaluations which were performed d to determine the applicability of those mechanisms to the CPS containment. The following are the differem mechanisms considered:

m

~ Direct containment hvoass. The IPE team considered the break locations of all piping extemal to the containment that tie in directly to the RPV or recirculation system piping, including main steam and feedwater piping in the steam tunnel. The screening criterion for direct containment bypass sequences (class V) was IE-7 per reactor year. For CPS, '

all of the class V sequences truncated out in the front-end analysis at IE-9; l - -

Reactor oressure vessel blowdown. The team investigated whether jet forces created during the high-pressure vessel blowdown were large enough to cause vessel movement

(

Clinton Unit 1 Back End 8 May 1996

I i i i : : .

1 1 l 1 . t i 1 1

. a class lA_CET . .

,.,'.. 'A . . * . .' . " 'M&".*'

. * " ' " =".0. 9p,gwa t.2.'c!t LT "'T.".,."." 1"n'.','f1" " l'#'

.*1.a7.."" ."l"Of9 .

.ti t g, c .

. . , e c. . . .

tt 9. 9-, .

' t et.w Im et.ee-o t *

t. et.

r

, news safe e.een-e ao sen I tags et.88-9 tatt e t .084 esses tage 4.738-4 es est I tatt e t .eE4 3499 et.08-9 IA89 et.98-9 tage es.st-9 l taas et.se-e I lett 49.0E4

, l tatt et.9E-9

, teSe et.9E-9 I Inge e t .et4 seet et. eta 5400 et 08-9 l CLASS IA CET - This CET begins with the containment building intact at the time core melt i

begins. The reactor vessel is at high pressure. The core. melt at high pressure sequences' .

from all non-IDCA non-ATHS transient initiators in the level 1 PRA were combined to establish the input frequency for this CET.

Figute 2. CET used for accident class IA (repnxluced from figun 4.5-1. page 4-62 of the submittal) i Clinton Unit i Itack-lind g May 1996 l

l

.- i s

! Table 3: Plant Damage State Codes 4

Plant Dammae Etate Codes j

A is the reactor status, either:

4 R - Recover in vessel l L - vessel penetration at low RPV pressure j y B - Vessel penetration at high RPV pressure

t. .

j BS is the containment status:

XX - containment intact

. vs - Vent through suppression pool j ,.. VB - Vent bypassing suppression pool'

! OD - Overpressure failure due to decay heat

! '- OA - Overpressure failure due to ATWS.

! OH - Overpressure failure due to hydrogen combustion -

~

l ov - Overpressure failure due to loss of vapor suppression

.. CI - containment isolation failure .

CB - Isolation failure with suppression pool bypass j

l c is the timing of the event:

~

X - Not applicable

. E - Early (< 6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months )

I - Intermediate (6 to 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ) l L - Iate (> 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months ) .

l 4

! ., coor - a Ba e i

i

$ =

~

l l

i J

1 1

. Clinton Unit i Back.End 10 May 1996 l

i .

o 'd ,

n on its foundation and to tear drywell and containment penetrations. The forces from the ,

e5 upper bound combination of jet thrust from the blowdown and lift forces from differential !

, ,j pressure between the pedestal area and the drywell were calculated to be 577,500 lbf.

Based on MAAP predictions the team assumed a 2 psi differential pressure between the

, q pedestal area and drywell. The vessel holddown force from 120 bolts was calculated to l w be 102.6 million pounds and the vessel / internals were estimated to weigh 1 million pounds. Therefore, the holddown forces greatly exceeded the forces that could cause ,

' P" vessel movement. t V l In- and ex-vessel steam exnlosions. After reviewing the Mark III containment design, the IPE team concluded that in-vessel steam explosion is not a. credible containment failure ,

w mechanism for CPS. The team indicated that the above conclusion was in agreement i

with the findings of the NRC-sponsored Steam Explosion Review Group which concluded l that the likelihood of an in-vessel steam explosion. leading to an alpha mode containment failure was very unlikely. -

" Assuming the following about the. material that would be involved in an ex-vessel steam d explosion. the team calculated the pressure rise in the containme'nt and shock wave pressure exened on the drywell wall and pedestal wall (page 4-37):

" n Molten material = 1/2 the core; 1/2 the lower core plate; 1/2.CRD mechanism;

_ 1/2 the lower vessel head; and '

Water = vessel pedestal full of water up to the l'ower edge of CRD can opening.

The team calculated a drywell pressure rise of 1.036 psi, a shock wave pressure at the drywell wall of 3.6 psi, and a shock' wave pressure at the pedestal wall' of 43.4 psi, e These pressures were not sufficient to cause failure of the containment wall er the l j pedestal wall. .

y In response to questions addressing the validity of the above results, the utility personnel noted that the methodology used was as proposed by FAI (additional RAI of the NRC, dated March 27, 1996). However, details of the calculation would be needed for review.

m J Thermal attack on containment nenetrations. Table 4, reproduced from Table 4.4-1, page 4-54 of the submittal, lists the various nonmetallic materials in the CPS drywell and

.= containment penetrations, the tested temperature for each material, the expected

, temperatures during severe accident conditions, and the anticipated life of the materials at the expected temperatures as calculated using the' Arrhenius equation. The life expectancy of the penetration seal material for the drywell at 700oF exceeded 2 weeks,

. which is much longer than expected in the case of recovery of core cooling.

- The maximum containment temperature predicted using MAAP was less than the tested

- temperature for all but one s&aling niaterial, Bisco LOCASEAL. This material was.

' actually tested at 355 oF for short periods of time, and was used under pressure, and

.: retaining part of the containment electrical penetrations. Its failure would not in itself i -

Clinton Unit 1 Back-End 1I May 1996 l a

, e. ) s v .

m i cause a leak out of the containment. The life expectancy of this material in the j q containment environment was calculated to be 504 days. ,

i

. Using MAAP, the team predicted that, under worst-case conditions, the drywell temperature would exceed 700.oF about 11 hours1.273148e-4 days 0.00306 hours 1.818783e-5 weeks 4.1855e-6 months into an SBO with no operator actions.

j ,_ The team concluded that a release from the drywell to the containment bypassing the i- suppression pool would occur when the drywell temperature reached 7000F. However, i - such a release would be retained within the containment because no containment j u penetration failure was expected to result from the elevated temperatures.

q Containment isolation. See section 2.2.4 of this report.

o l Direct conuinntent heatine. The team concluded that two design characteristics would

~

l significantly limit :he magnitude of the pressure rise associated with direct containment i ... heating (DCH): the reactor depressurization system and suppression pool. The CPS

} plant is equipped with 16 safety relief valves (SRVs) any one of which is capable of

" ensuring low reactor pressure (< 200 psia) at the time of vessel penetration and thus j

4 W would make DCH ineffective. The suppression pool would remove the debris mass and i energy during the blowdown phase. De team calculated the suppression pool l temperature resulting from a reactor vessel blowdown at high pressure using the l ~ -

following assumptions: 1/2 the core,1/2 the lower core plate, and 1/2 the lower vesse' l head were ejected; the suppression pool was at an initial temperature of 122.SoF; and all

'~

l debris energy was transferred directly to the suppression pool, with no energy lost to the ,

surrounding structures. De results showed a suppression pool temperature increase of I 22.2eF. This increase did not raise the temperature'of the suppression pool to saturation, and therefore, was not expected to produce containment pressurization effects.

l

Molten core-concrete interaction. The IPE team considered the following failures l ,,

resulting from core-concrete interactioni (1) the reactor pedestal walls would lose their

^**

load-carrying capacity, (2) the basemat would be penetrated and core debris would exit l ,. the containment; or (3) sufficient noncondensible gases would be generated to fail the ;

containment on overpressure. Because the effect of noncondensible gas buildup in the

containment was included in the MAAP analysis, it was not considered funher.

1 Because of the pedestal design feature of Mark HI containment, the team concluded that the wakening of the pedestal wall resulting from core-concrete interaction would be

.. neg'Jgible: De floor elevation of the pedestal region was far below that of the drywell j ; floer, which was 8.1 feet. De largest debris depth was expected to be 15% of this I difference, or less than 1.25 feet. Therefore, even after a sideward MCCI erosion of the l entire width of the pedestal wall, (i.e., 5.67 feet at CPS) the pedestal wall would rernain attached to the drywell floor across a vertical distance.of at least 6 feet around the entire

circumference. (See figure 1.)

i

) ,_ After vessel failure, the corium ivould split between the pedestal and the sump. De thickness of the basemat is 10.2 feet under the sump and 13.2 feet under the pedestal.

j Assuming noncoolable debris beds, the team calculated that containment failure as the Clinton Unit 1 Back-End -

12 May 1996 i -

e

.' =

a 9 j result of basemat penetration in the pedestal cavity area would occur in 19.8 days, and in

=

l the floor drain sump area in 8.1 days. Therefore, the team concluded that the potential for MCCI-induced fission product releases from the primary containment would exist only long after much more rapidly occurring mechanisms, such as containment

! pressurization, caused containment failure.

e ,.

, Hvdronen combustion. The hydrogen ignitor (HI) system is used to maintain post-L ", accident hydrogen concentration below 495. The HI system contains 115 glow plug type -

o ignitors located throughout the drywell and containment with at least one ignitor for each division located with a maximum separation distance of 30 feet. The hydrogen ignitors j "l were expected to remain operable during all events with the exception of SBO, therefore j iJ limiting hydrogen combustion to temli-I burns instead of global burns and detonation.

' ~ '

Hydrogen combustion early in an SBO (within about the first 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months ) was not expected

l to cause. containment failure because of the limited time available for hydrogen generation. A top event for hydrogen control was included in the CETs.

3

! The utility's attention to the issue of hydrogen combustion appears to be. incomplete. As

! , shown on the quantified CET for SBO sequences, when AC power recovered is after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months , the probability of containment failure by hydrogen combu.stion is less than 1 in

~

a 1,000 (Figure 4.5-2, page 4-63.of the submittal). In response to the additional RAI of l _ the NRC, dated March 27,1996, which asked for more information on this issue, the ,

l utility responded by providing two pages of the submittal. [1] Of these, page 4-69 noted that sequences TL 52 and 53 " consider power recovery at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months with an essentially

. simultaneous hydingen burn that fails the containment by over pressurization." These

" two sequences however were involved withfailures of power recovery at 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months and resulted in containment failure because of pressurization that was caused by loss of decay j r ,

heat removal from the containment.

. Containment overoressurization. The IPE team considered pressurization challenges to l ~ the CPS containinent from the following

1 1

Vessel blowdown; I "

Steam generation from ATWS (SRVs);

w

Ex-vessel steam explosion;

Hydrogen combustion;

"

Noncondensible gas generation by MCCI; and l ,

Steam generation from decay heat.

1 .

4 Of these, only two events, hydrogen combustion and steam generation from ATWS, were

.. found to have the potential to threaten containment integrity during the mission time of i 48 hours5.555556e-4 days 0.0133 hours 7.936508e-5 weeks 1.8264e-5 months . Failure of the containment from hydrogen combustion could occur in 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months

'i after event initiation, and failure from steam generation resulting from ATWS could 4

- occur in 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months after event initiation.

i 1- .

I Clinton Unit 1 Back End 13 May 1996

} I. .

l

p. . ;

=

De IPE team used the CPS USAR overpressurization analysis, section 3.8.1.4.8, (which i j y, is based on containment calculations performed by Sargent and Lundy and the

containment Equipment Hatch and Personnel Airlocks calculations performed by Chicago Bridge and Iron) and the results of the Sandia National I.aboratories 1/6th scale test of

- Reinforced Concrete Containments as the basis for probable failure locations and

pressures.

l'

! - ne team found a median containment failure pressure of 93.8 psig, with the failure j ,,

mode most likely to be in the liner in the vicinity of a containment penetration. The team

estimated that the containment shell (rebar) would begin yielding at 95 psig at the hoop t - reinforcement at mid-height of containment, and be expected to fail at a significantly ,

, g higher pressure (value not reported). De team estimated that the containment equipment ;

j hatch and personnel airlocks would be stronger than the containment liner. Using Monte. !

l Carlo methods, the team calculated the cumulative failure probability curve as shown

,. figure 4.4-1, page 4-55 of the submittal.

" Members of the team assumed that the failure of the containment shell or equipment l .) hatch would be gross (i.e., they would be large failures that would rapidly depressurize .

the containment). ney assumed that failure of the containment liner would be limited in !

'9 size, such that further containment pressurization would be prevented, or a gradual l u containment depressurization might occur. .

i ne IPE team considered containment failure modes consistent with those listed in Table I

! - 2.2 of NUREG-1335.

l L2d Containment IsolationLFailure. .

~

f- All of the isolation valves were assumed to fail under severe accident conditions.

I 1 However, except for SBO sequences, all of these valves either would move to their W required positions early 'in an even' or would already be in the required positions and l therefore would successfully complete their required safety functions. For success, only i

9 one of two (inboard / outboard) isolation valves must move to the required positions.

)

Except for SBO' sequences, the IPE team used a fault tree, accounting for instrumentation j , power dependency, operator actions, and valve failure, to evaluate the probability that at l least one of these valves would move to the required position early during an event, and

thus successfully complete the safety function.

i m i For the SBO sequences, all valves except two (IF007 and IF008) were in the required

~

! position, or were an integral part of a closed-loop system. Valves IF007 and IF008

[ were in the containment penetration line between the upper pool skimmers and Fuel Pool i - Cooling and Cleanup surge tank that would require ope'rator action to isolate. Existing EOPs address operator actions to check and manually close, if necessary, isolation valve

]

- ,q IFC008.

j The IPE team modeled failure of containment isolation as a top event in the CPS CETs.

! .a Only one PDS involving containment isolation failure had a frequency higher than the 1 J i me 1 Clinton Unit i Back End 14 May 1996

, n

i el o .

l

. ~.

1 . t i screening limit of IE-9 per year. This was m accident class IB (involving SBO) and had

,,, a frequency of 7.0E-7 per year. .

l It appears that the IPE process assessed and identified contributors to containment isolation failure. -

i

) 2M System / Human Reqponse.

{

Y As described in Section 2.2.2 of this report, the CPS CETs were constructed emphasizing i the system behavior the operator can see and control, such as containment pressure and l

- temperature and system operation. CPS has fully implemented the recommendations of

,, Revision 4 of the BWROG EPGs in its EOPs. CPS has fully verified and validated the

! procedures, and conducted extensive training of appropriate personnel (including

- simulator training). The following is a summary of recovery actions reported in Section i 4.6.2 of the submatal: )

i l

! r= Power Recovery to Prevent Panetor. Vessel Failure Followine Core Dammee. Based on !

e ., MAAP results showing that vessel failure would occur at 2.6 hours6.944444e-5 days 0.00167 hours 9.920635e-6 weeks 2.283e-6 months after the initiauon of i an SBO, the IPE team assumed 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months were available in which to recover AC power.

! " This would be r6ecessary to recover the injection systeins in time to prevent reactor vessel

! .. failure following core damage in high-pressure sequences. The power recov'ery factors l used for combinations of loss of offsite power and specific' additional events are shown in

Tables 4.6-1 and 4.6-2'of the submittal.

Power Recoveries to Prevent Containment Failure for CETs in which Cantninment

Isolation is Successful or for hee Iniection for Debris Cooline or Scrubbine on the Non-j ~ Isolated Cases. Recovering power beyond a 4-hour time frame could cause hydrogen

burn of sufficient magnitude to initiate a pressure spike that could fail the containment.

The IPE^ team used a conditional failure probability of 0.469 for the recovery of power at t' 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months .

! Recovery of Iniection Svetams Before and After Panetor Vaccal Falkire Followine Core i

i Damage. All of the core damage sequences resulted from failure to depressurize or to recover injection systems in time to prevent core damage. Vessel failure can be

, prevented, even after core damage, if injection systems can be recovered within 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months j l in high-pressure sequences. If the vessel is depressurized, only 15 minutes is available l , for recovery. l

! . Even if injection systems are not recovered before vessel failure, containment failure can 4

! still be prevented in most cases if injection is recovered within 13 hours1.50463e-4 days 0.00361 hours 2.149471e-5 weeks 4.9465e-6 months (4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months for SBO).

j ...

l - Failure to Initiate Containment Sorav. Because containment spray is to be initiated j ,

manually, an human error probability was obtained using human reliability analysis i screening method., A conditional recovery probability of 0.3 was used for containment i sprays. .

~

! May 1996 Clinton Unit i Back End 15 i

1 e

, .- o )

I i

l Failure to Isolate Containment in Case of SBO. He task of manually closing valve l

- IFC008.in case of an SBO is directed in procedure CPS 4200.01, and a HEP of 0.4 was i assigned. . l Failure to Recover Ione-Term Containment Heat Removal in 48 Hours. By this time all

,~ of the resources would be available, (i.e., the CPS resources; state, local, and national j agencies; the Institute of Nuclear Power Operations; and General Electric Corporation). 1

- The IPE team used a conditional probability of IE-3 for failure to recover. , !

Failure to Open ADS Backuo Air Bottles Isolation Va!v2 on Loss of Power. The ADS r- motor-operated backup air supply isolation valves are opened from the Mam Control n Room if normal instrument air supply to the SRVs is lost. During an SBO, power is unavailable to open MOVs, the operators must open the valves manually before the air

~

accumulators are depleted. The team used an HEP value of 0.12 for this action.

Failure to Vent Containment. The team used an HEP value of 0.25 for failure to vent

'T the containment.

The IPE team appears to have significant amount of credit for back-end system / human interactions. 90 wever, as noted in the TER of Human Reliability, Reviewer recovery failure probabilities have been determined using generic results for EPRI RP-3000-34 and ,

the licensee failure to consider CPS-specific factors in repair or restoration of components l in back-end analysis is a limitation in the IPE. SCIENTECH agree with this observation.

216 Radionuclide Release Catenories and Characterization.

The IPE team defined the release mode, describing the type of release for source term

, binning. for each accident sequence in the CET that was not truncated. The release modes are alphanu'meric (e.g.,'Cll) and defined as follows (section 4.3.4, page 4-25):

A - Containment or reactor. vessel is intact at accident termination;

B - Containment failure occurs with release scrubbed through the suppression pool;

C - Containment failure occurs before or at vessel failure, and the suppression pool is bypassed;

D - Containment failure is delayed after vessel failure, and the suppression pool 3- is bypassed; and

E - Radionuclides release to the environment directly through an unisolated LOCA outside containment. .

The team identified subcategories of the release modes tiy the numerical designators of 0 through 12, as shown in Table 4, which is reproduced from Table 4.3-4, page 4-30 of the submittal. ,

w .

m.

! Clinton Unit 1 Back End 16 May 1996 i

e '

i l The team categorized fission products released to the environment into three' groups as

follows: (1) noble gases, which included inert gases; (2) volatiles, which were composed j , of CsI, Rbl, TeO ,2 CsOH, and Te2; and (3) nonvolatiles, which were composed of SRO, l moo 2, BaO, lanthanides, CeO 2, Sb, and uranium /transuranics. As shown in Table 5,

, )

j reproduced from Table 4.3-5, page 4-31 of the submittal, the team defined level-2 release categories based on fractional releases of the above fission product groups.

l

! - ' Of the four release categories, Release Category I did not have any accident sequences

{ with frequencies higher than the truncation. The IPE team noted that the source. terms j

resulting from the containment failure sequences were fairly large (all classes II and III) '

i but the source-term determination involved a number of conservatisms (not taking credit.

l ,

for scrubbing of volatile and nonvolatile fission products at the spent fuel pool; assuming that both the inner and outer containment penetration seals would fail when the drywell temperature reached 7000F).

It appears that the IPE team adequately characterized the source term that would be l generated at CPS as a result of severe accidents.

I j 2.3 Quantitative Assessment of Accident Progression and Containment Behavior 3

.- L1.1 Severe Accident Proeression.

l' Section 4.6.1 of the submittal describes severe accident progression analyzed in the CETs j -

used for different accident classes:

j CET IA. This CET analyzed high-pressure transient, non-LOCA, and non ATWS l

~

accident sequences. This CET resulted in three significant sequences (frequency higher

, than the truncation value of IE-9 per reactor year) but none would involve containment

releases. ' Two of these sequences would recover in-vessel. The third sequence would j involve vessel failure at high pressure (> 200 psi) but no containment failure because the

containment pressure would reach only 25 psia,. which is well below the containment j failure pressure. ,

1 3

l CET IB. Four of the seven significant sequences in this CET would result in a release 4

from the containment (TL51,52,53,54). All of these sequences would involve failure j of AC power recovery to prevent vessel failure. One of these sequences (TL54) would

involve containment isolation failure resulting in a category III release. Sequence TL51

! would involve a delayed containment venting release (manually initiated) which bypassed i the suppression pool in order to prevent failure of the containment by ove pressurization.

l This sequence would result in a category II release. The remaining two sequences

would involve a category III release. One sequence of this CET, TIA1, would recover i -

in vessel.

j , CET ID. This CET analyzed low-pressure transient, non-LOCA, and non-ATWS l j accident sequences. This CET resulted in four significant sequences' but none would

]

involve containment releases. In two of these sequences, ID47 and ID49, venting would Clinton Unit i Back End .

17 May 1996 l l

i

e d i

He l l

4 e

Table 4. Release Modes ,

i -

J

i

. l

, l 1 -

l i

g;gggg gy;;Ag MCDffAMGff PA&iff i .

, maAmAm erAns vesse, arcusvesse. i esAmerm aro u vussa.

PALURE WEE 8EL FALL 8E PALLpg vesELPAt1NG a

acLA m A0 i

l . mAcT A2 vswrun . A1 _

B1 , B2 FMLED wy C1 C2

! a wemru. ,

g D2 g3 C4 4

SPRAY m I

wrmuu. C5 66 l ana '*. D3 - D4

wa-a ouser.* C8'

T emeu. C7 C9 C10

! m wr m eu. g

'"'C " C11 C12 emwsu.

f E2

+

- cour E1 l

4 w Ass 1 ,

4 3 i 1

1 a

f 4 .

1 1

1

)

i =

,l i ~

18 May 1996 Clinton Unit i Back End l

l 1

o ,:

~ ,

I Table 5. I.evel 2 Release Categories Category Noble Gases Volatiles Non Volatiles NR 0 0 0 I s 100 % s 1% 50.1%

D s 100 % 1 - 10 % 0.1 - 1.0 %

III s 100 % > 10% > 1.0% -

f be available if required, but was assumed unused because containment pressure would reach only 32.7 psia, a value significantly lower than the containment structural capacity of 109 psia (94 psig).

~

CET IIIB This CET analyzed accident sequences invoh 6g a small to medium LOCA that did not depressurize the RPV. Neither of the two significant sequences in this CET would involve a containment release.

CET IIIC. This CET analyzed accident sequences involving depressurization of the RPV. either by a. medium to large LOCA sequence , or else by an operator action.

Neither of the two significant sequences in this CET would involve a containment release. Both of these sequences involved either successful injection before vessel failure ;

or late injection. .

CET ATWS. This CET analyzed accident sequences involving containment failure from ,

overpressurization before core damage or vessel breach.

W Dominant Contributors Consistency with IPE Insiehts.

The CPS IPE team postulated a CDF of 2.6E-5 per reactor year based upon the present, as-operated, CPS reactor plant and containment capabilities. The significant core damage contributors' were SBO (long and short term) and transients contributing to 37.2% and i

52.0% of the total CDF. The team cal 2nlated a conditional containment failure '

probability of 5% from ATWS, and some SBO and high-pressure. core damage i

sequences.

Table 6 shavs the SC ENTECH comparison of dominant contributors to the CPS containment failure pn.bability with the r sults of other plants with Mark III containments. W CDF t CPS i; .'s '.he mid range between the CDF at Perry and Grand Gulf but the probability of comainment failure, both early and late, at CPS is significantly lower than that at other two plants. . A detailed assessment of the reasons . .

i why the CPS results are so much differen' is beyond the scope of this TER. However, the impact of hydrogen burns and steam explosions on containment failure-areas found

. incomplete in this TER-may be a contributing factor.

. E t--- -

, e v .

2JJ Characterization of Containment Performance.

! 1

The modular Accident Analysis Program was the primary code used for the containment i analysis. ' The IPE team used CPS-specific data, including containment, drywell, and j, suppression pool parameters as input to the MAAP parameter file. The team used

! " Table 6. Containment Failure as a Percentage of CDF: 1 Comparison with Other PRA Studies

Containment Failure Mode Perry Grand Gulf Clinton CDF per rx yr 1.3E-5 4.lE-6 2.6E-5 I Early failure with pool bypass, no 16.0 16.5 0.5 )

spray i

! Early failure with pool bypass, 0.5 4.9 -

! with spray .

l Early failure with no pool bypass 8.2 21.8 - l

. I. ate failure 7.4 28.4 5 1.8 l . Venting 29.3 3.8 2.7 )

! Intact 39.1 '23.0 95.1 )

l i

Computer Aided Fault Tree Analysis and Set Equation Transformation System for the containment systems and event tree sequence quantification (Section 4.2.1, page 4-18).

The IPE team characterized the CPS containment performance for each of the CET end states by assessing containment loading calculated using the MAAP computer code. The j team performed a MAAP sensitivity analysis to address phenomenological uncertainties of accident progression.

2JJ Imoact on Eauioment Behavior.

Section 4.6.3.1 of the submittal describes severe accident impact on equipment behavior. l At CPS, most of the equipment necessary for accident control is located outside the l containment, and therefore, would not be exposed to extreme environmental conditions m 1 severe accidents. The following is a summary of the availability / survivability of thd )

equipment located inside the containment:

Inboard Isolation Valves. These valves and valve actuators are qualified for accidents under the provisions of 10CFR50.49, " Environmental qualification of electrical

. equipment important to safety for nuclear power plants.* All of these valves are either in the Itquired positions to perform their required safety functions or move to the required positions early in an event (exceot during an SBO), and are expected to coaiplete their

~

safety functions liefore any degradation occurs.

Clinton Unit i Back-End 20 May 1996

y J

, Automatic Deoressurization System Safetv/ Relief Vah.c.5. Safety / relief valves are qualified to operate in system / containment conditions until vessel failure occurs beyond

[ -

which SRVs are no longer required.

]

Combustible Gas Control and AssociateA Comoonents. De IPE team did not model the 4

drywell and containment mixing compressors because of their limited capacity. The vacuum breaker elastomer seals are qualified to 500eF for 65 hours7.523148e-4 days 0.0181 hours 1.074735e-4 weeks 2.47325e-5 months . Under severe

- accident conditions in the drywell, the inboard (drywell side) vacuum breaker seal is expected to fail early and the outboard (containment side) later. Seal failure and suppression pool bypass are assumed after the drywell temperature reaches 700eF. The

hydrogen ignitors were tested in 1009E steam at 330oF for 3 hours3.472222e-5 days 8.333333e-4 hours 4.960317e-6 weeks 1.1415e-6 months , plus 300eF for 3 ;

l. ;

days, plus 250oF for 7 days.

~

Sunoression Pool rnd Sunoression Pool Makeuo. The IPE team did not model the suppression pool makeup system and did not take credit for its operability. Under certain i accident conditions the suppression pool would reach saturation temperature but it would ;

} not affect the ability of'the pumps drawing suction from the pool because those pumps i are designed to pump, saturated water. l l Electrical / Mechanical Penetrations. The drywell penetrations are expected to survive for i

'- more than 2 weeks at 700cF The containment penetrations are not expected to fail due to temperature, humidity, and radiation. ;

i

! Containment Vent System. Motor-operated valves are the only parts of the Containment i l

2 Vent System that would be affected by severe accident conditions. The valves are

. qualified for the same environmental conditions as for inboard isolation valves.

Containment /Drvwell Ventilation Systems. The Containment HVAC System, Drywell f harge System, and Drywell Cooling System are not required or designed to operate under i

severe accident conditions, with the execption of. containment isolation valves.

g-1 Instrument Reauired for Recoveries. The instrumentation required for recoveries

includes RPV level instruments, RPV pressure instruments, containment' pressure instruments, containment hydrogen instruments, suppression pool level and temperature, and containment isolation valve position indication. These inst'ruments are all qualified to l

the requirements of 10CFR50.49

! i Performing a detailed analysis of equipment' vulnerability to the adverse environmental l

- conditions of severe accidents is a strength of the CPS IPE. It :.ppears that the IPE team i- has adequately addressed equipment vulnerability to severe accidents at CPS.

W Uncertainty and Sen.Jtivity Analysis. j i

1 i The IPE team prformed EPRI-recommended MAAP sensitivity analyses to aM'ess issues of severe accident progression. The following were the specific issues addressed:

[

l i

i .

21 May 1996 l

Clinton Unit 1 Back-End 4

f'

.en, ,w.v. . - - - - - - - <-

.. d .i x

'

Hydrogen production sensitivity to channel blockage:

Source term sensitivity to containment failure (vent) size; I 1

l

-

Effect of containment perfonnance if in-vessel recovery fails; ,

l

Effect of using higher (than CRD) capacity recovery systems;

]

Effect of varying LOCA size in class IEC sequences;

=

Effect of leak before break on the source ter74- !

.a . i

Effect of degree of revaporization on the source term; j

Effect of core melt progression on revaporization/ source term; j

"

Effect of debris coolability on containment performance;

,a

Effect of rapid steamir.g period following RPV failure; ;

...

Effect of varying vent timing on release source term;

~

Effect of drywell penetration size on release source term;

Effect of power recovery timing on containment performance;

'

Effect of a stuck-open SRV concurrent with an SBO; 1

Effect of a large bitak LOCA with concurrent SBO;

Effect of alternating hydrogen cor.centrations in SBO sequences; and

Effect of reduced containment overpressure failure threshold.

t' The overall conclusion from the sensitivity study was that, in comparing worst-case l scenarios to the base case, only one change in assumptions (debris bed coolability)

.- significantly altered the parameters that could challenge containment integrity. However, the IPE team concluded that, if this assumption were applied to all CET sequences, it q would not significantly increase the probability of CPS containment failure.

i

_ Performing a comprehensive sensitivity study to address phenomenological uncertainties is a strength of the CPS IPE. It appears that the IPE team's treatment of severe accident ];

progression at CPS is complete. -

Clinton Unit 1 Back End 22- May 1996

- - .- .. - - - - - . . - . - - . - - - - - . - . . ~ . - - . . - . . - . - . - - . - -

+. m l

^

! 2.4 Reducing Probability of Core Damage or Fission Product Release 4

e n

} 211 Definition of Vulnerability. ,

1 l

Section 3.4.2, page 3-228 of the submittal, describes how the IPE results were screened i to determine if new vulnerabilities were discovered during the performance of the CPS IPE. The criterion applicable to back-end analysis was as follows:

l 5.. Are there any new or un' usual means by which core damage or containment failure l occur as compared to those identified in other probabilistic risk assessments >

! (PRAs)?

W No vulnerabilities were identified.

T l ~ 212 Plant Imorovements.

t

The IPE team evaluated the containment failure cutsets to analyze those basic events or

- independent subtrees with the highest importance measures. Table 6-4 of the submittal shows the basic events or indanandant subtrees with the highest Fussel-Vesely values for

" f

! the containment failure cutsets. The following is a summary of the important eatures sdentified: -

i -

! Imss of Off-Site Power. SBO sequences, which originate from the loss of off-site power (LOOP), would cause containment isolation failure because containment isolation vaives would fail to open under loss of power conditions. ' Operators would have to manually ;

isolate containment isolation valves in order to ensure that no radioactivity was released l from the containment. Under certain conditions, a hydrogen burn could cause a rise in i ... pressure sufficient to fail the containment, but under non-SBO conditions the hydrogen j .

ignitors would be able to prevent containment hydrogen concentration from reaching a level at which containment failure could occur from hydrogen burn.

1 l r j In order' to lower the frequency of LOOP, care should be given to the performance of activities involving the switchyard or the plant connection to the off-site power system.

i ,a The IPE team has transmitted this insight to the CPS Nuclear Training Department for i

emphasis to the operators. ,

i

.. Recoverv of AC Power. The back-end analysis. included events representing power i (,, recovery in time to protect radioactive release barriers once fuel damage has occurred

~

(e.g., power recovery in time to prevent reactor vessel breach). Recovery of AC power involves recovery of off-site power, or recovery of emergency diesel generators, or both.

In addition to being able to prevent a LOOP or SBO,. the ability to recover power within

a reasonable time is also important. A strong understanding.of diesel generator system u operation is useful in recovering AC power. The IPE team transmitted this information j to the CPS Nuclear Training Department.' !

l tems Clinton Unit 1 Back End 23 May 1996 I J

p i

Failure to Isolate the Containment 'Under SBO Conditions. During an SBO the operators

- would have to manually isolate valve IFC008 in order to isolate the containment. The IPE team emphasized the ke~y nature of this valve to the CPS Nuc1 ear Training Department, which is to evaluate ways to highlight this requirement in training programs concerning SBO. .

Scram Hardware Failures. Although ATWS sequences contribute relatively little to the

- overall core damage risk, they are important to the risk of radicactive release from the containment. Scram hardware failures (c. g., failures of.the scram discharge volume) are the primary causes of an A*IWS. In order to lower the frequency of an ATWS, care

'. should be used in the maintenance and operation of scram equipment, and any design

.. changes to this equipment should be carefully reviewed for possible reductions in i reliability. The IPE team emphasized the impact of scram system failures to the CPS

" Nuclear Training Department, whkh is evaluate ways to emphasize this information in j the training programs concerning scram hardware.

" The IP effort to understand any required CPS plant improvements is a strength of the l

.' IPE. It appears that the proposed operator training improvements would enhance safety at CPS. .

- 2.5 Responses to CPI Program Recommendations He following is a summary of the Clinton Power Station's responses to recommendations

- made by the CPI Program:

" Hydronen Innitor Backuo Power. The IPE team found that having an alternate source of .

" power with 90% availability to the hydrogen ignitors would reduce the frequency of

, containment release from 1.28E-6 to 8.76E-7. He same source would reduce the

.; frequency of a large release (class III) from 7.52E-7 to "7.4E-7." Because the calculated large release frequency was below the NRC safety goal of less than IE-6, the IPE team concluded that 'an altemate ignition power supply was not justified.

~

Mark I Imorovements:

'g (a) Alternate Water Supply. De IPE team modeled the Fire Protection (FP) system as a long-term core cooling system for low-pressure transient sequences in which some other

~

rn core cooling system runs for a period of time. The FP system has disadvantages because it is a low-pressure system, which depends on reactor depressurization and it takes time to align it to supply water to the reactor. The FP system could be made more useful for core cooling by a piping change, which would involve installing a bypass line with a bypass valve. CPS is considering this hardware change as a possible future improvement in the plant design. His hardware change will not be implemented, however, until the i completion of the IPE for external events and development of a Severe Accident i a Management Plan.

24 May 1996 Clinton Unit 1 Back-End v

d .

[, 4j . & l i -

l l

(b)' Enhanced Depressurizaticn Reliability. He IPE team included the backup air supply m for the ADS in the IPE models. .No backup was available for the depletion of batteries l for the ADS function. The team found that such a backup would reduce the CDF of SBO l l .

l sequences by 25% and overall CDF by 10%. This change is to be considered as a part

) of the Severe Accident Management Plan. :

I r.

j (c) Emergency Procedures and Training. CPS has fully implemented Revision 4 to the l j BWR EPGs, which is reflected in the IPE.

1 ". a

! Mark II Containment Heat Removal. The IPE team found that adequate containment heat ,

i removal was not a significant factor in the CPS IPE except in ATWS scenarios. I

'd Although directed by EPGs, venting and suppression pool cooling were found not !

effective in preventing containment failure in ATWS scenarios. Derefore, the team did )

not take credit for such cooling in the IPE analysis, and still found that the CPS

~ containment failure probability was relatively low.

1

]

The IPE team identified three paths of sufficient size to vent the containment, which do not include a hardened vent (Section 3.2.1.11, page 3-81) [1]: (1) to the spent fuel pool via the RHR systems, (2) to the spent fuel pool via the Fuel Pool Cooling and Cleaning system, and (3) to the duct work through the Continuous Containment Purge system. In response to the NRC staff's RAI, the licensee noted on page 60 that improving the containment venting capability with a hardeneid vent system (see GL.88-20, Supplement no.1, Enclosure 2) was not required. [4]

No transient sequences with loss of containment heat removal (IW) are significant for core damage because all emergency cooling systems am capable of taking suction from a saturated suppression pool. Consequently containment venting has e, no impact on core damage frequency. In only one conaistaent event tree j sequence does failure to vent lead to containment faikre, if venting were to fail in this sequence, the containment over pressure failure frequency.would increase y by about 4.lE-7, and the large release frequency would change from 7.5E-7 to 1.2E-6, to 4.5% of core damage events . . . . Since this value is within the 4

" Severe Accident Issue Closure guidelines" of NEI 91-04, Revision 1, for i- handling with severe accident management guidelines, there is no intent to perform any modification to address improvement of the containment venting capability.

It appears that the IPE team has adequately resppnded to the recommendations of CPI Program. .

M IPE Insiehts. Imorovements. and Comrvitments.

The following is a summary,of major features of the CPS Mark III containment as !

u reported in the submittal:

h

Steel lined, reinforced-concrete containment, with a volume of 1,550,000 ft 25 May 1996 Clinton Unit 1 Back-End

,J.

- - -. .- .- _ .. . . _ _ - - . - . . . . - - . _ _=

{ , * .n 1-1 i 3

-

Drywell structure with a volume of 246.500 ft enclosed by the contim. >. ment;-

3 l .

Suppression pool with a volume of 135,700 ft , which is accessible from the drywell and containment;

] ,

Two trains of containment spray, suppression pool cooling or shutdown heat

- removal; and 1

A reinforced-concrete basemat over 10 feet in depth.

t- Members of the IPE team identified no vulnerabilities. However, they identified several

imponant featums of the CPS, which were tansmitted to the Nuclear Training i Depanment for emphasis to the operators.

4

~

i -

4 4 i i

p .g E

e$

e-

~

hee Clinton Unit i Back End 26 May 1996

- - . . - - - _ - . - - - - . .- -. - . - - _ - = . . - - . . . . . - --

{ 4: &

3. CONTRACTOR OBSERVATIONS AND CONCLUSIONS '

~

The C!in' ton Power Station IPE appears to meet the intent of the Generic Letter 88-20.

SCIENTECH noted the following limitations: -

~

The utility's attention to the issue of hydrogen combustion was not complete. As shown on the quantified CET for SBO sequences, when AC power is nwcered

- after 24 hours2.777778e-4 days 0.00667 hours 3.968254e-5 weeks 9.132e-6 months there is only a very small probability of containment failure tcam combustion, this in spite of a statement in the text that these burns fail the' i containment by overpressurization. :

g When questioned about the validity of the ex-vessel steam explosion results as reponed in the submittal, the utility personnel gave no justification but only noted that the methodology used was as proposed by FAI. 'Ihis is an additional incomplete aspect of the submittal. ;

The IPE team appears to.have significant amount of credit for back-end system / human interactions. However, as noted in the TER of Human Reliability Reviewer recovery failure probabilities have been determined using generic results for EPRI RP-3000-34 and the licensee failure to consider CPS-specific factors in o repair or restoration of components in back-end analysis is i limitation in the IPE.

SCIENTECH agree with this observation. 1

... . In genem1, the submittal shows an understanding of accident progression, phenomenology, conservatisms, containment response, and radiological source terms.

SCIENTECH noted that:

The utility personnel were involved in the development and application of PRA

'~

techniques to CPS ~. Through plant ivalkdowns and peer review the utility -

4 confirmed that the IPE presents the as-built, as-operated plant.

~

The IPE team performed scrisitivity studies to address phenomenological l uncenainties of accident progression. The process used for the back-end analysis is capable of identifying severe accident vulnerabilities.

-

The licensee responded to the CPI Program recommendations.

~

By examining the back-end results the IPE team identified potential hardware or procedure improvements that could be implemented to funher reduce core damage or containment failure.

The CPS IPE team postulated a CDF of 2.6E-5 per reactor year based upon the present, as-operated, CPS reactor, plant, and containment capabilities. The significant core a damage contributors were SBO (lo'ng and shon term) and transients contributing to a7.2%

and 52.0% of the total CDF. The team calculated a conditional containment failure Clinton Unit i Back-End 27 May 1996 M

, 44.,19

.m '

l probability of 5% from ATWS, and some'SBO and high-pressure core' damage

, sequences.

1

~

05%

seus P*

e ..

L4 m

i-Fe

' s.

N I

e-r.

e- ;

i w

e Clinton Unit i Back-End 2R May 1996

. - . . . . . _. ._ . . - . ~ _ - - . . _ . . - . ._. _ _ . . _.. -_.

y.e &.

1

} 4. REFERENCES i ,

, 1. Georgia Power Company "Clinton Power Station IPE, Final Repon,"

September 1992. .

,, 2. Georgia Power Company, " Response to the NRC Request for Additional Information on Clinton Power Station IPE," November 1995.

, su . ,

i l

! 1 m S td i

l l

-Y *

- iJ 9 l u

s. ,

9

- l i )

.w i

e Clinton Unit 1 Back-End 29 '

May 1996

. wJ W i

APPENDIX: IPE EVALUATION AND DATA

SUMMARY

SHEET

1 -

BWR Back-End Facts .

Plant Name i.,

, Clinton Unit 1 Containment Type

=

g Mark III Unique Containment Features CPS has a strong containment design in that it has the largest free air volume and l

+- suppression pool volume,to rated thennal power of any U.S. Mark III ,

,j containments. These factors contribute to slower containment p'ressurization for a i given accident sequen'ce. l m

.x Unique Vessel Features

" None Number of Key Plant Damage States (Core Damage Bins) n

.* 6

! ~

Ultimate Containment Failure Pressure -

94 psig (median or 50th percentile value)

Additional Radionuclide Transpon and Retention Structures

'l None cred.ited

==

, Conditional Probability that the Containment Is Not Isolated ;

" 0.027 9 M

m,d - ,

~

Clinton Unit 1 Back-End A-1 . May 1996 J .

w -

Imponant Insights. Including Unique Safety Features Loss of off-site power is the dominant initiator leading to radioactive release from the containment.

AC power recovery (off-site power or diesel generators) in time to prevent vessel breach could significantly reduce radioactive release from the containment.

During an SBO a failure to manually isolate valve 1FC008 could'cause containment isolation failure, thus a large radioactive release.

,a Although ATWS sequences contribute relatively little to overall core damage risk, they are important contributors to the risk of a radioactive release from the containment. In order to reduce the frequency of occurring ATWS it is important ,

,. to rnaintain and operate a high,y reliable scram system. l l

" Implemented Plant Improvements sa The IPE team has emphasized the radioactive release risk dominance of loss of !

off-site power, failure to recovery of AC power, failure to isolate the containment l w under SBO conditions, and failure scram system hardware to the CPS Nuclear Training Department for emphasis to 'the operators.

- C-Matrix n

PDS* CDF per Early Late Intact.

94 rx yr n

.1 IA 9.8E-6 1 w

IB 9.8E-6 0.07 0.04 0.89 M

j .ID 5.7E-6 1 IIIB 1.3E-8 1 J IIIC 1.1E-6 1

,9 IV 1.4E-7 I w .

Consistent with many PRAs. PDS in this table is defined as the status of systems and equipment important for the source-tenn analysis at the onset of core damage; the IPE team defined PDS as CET end state.

ma i

GuuS Clinton Unit 1 Back-End A-2 May 1996 J

ML20140B658

Text

SUMMARY

SUMMARY

Navigation menu

Search