ML20148A485

Shoreham Nuclear Power Station PRA W/Supplemental Containment Sys

ML20148A485
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	02/29/1988
From:	LONG ISLAND LIGHTING CO.
To:
Shared Package
ML20148A483	List: ML20148A485 SNRC-1424, Forwards Shoreham Nuclear Power Station PRA W/Supplemental Containment Sys, in Response to NRC Questions.Requests Concurrence W/Quality Group Categorization by 880408
References
NUDOCS 8803180202
Download: ML20148A485 (149)

Text

)

S..0.0+.........W....em...3. 0We

. . . . . . . Sr. ... 0 .. ..

?::0bcif.s:f.c Is.< Assess:::.e:::: Wl::::.

r1', n . . . ,,,

nn .,.. ,. ., .

.n.,. .

c. .n. , .

> ......U s.J J..n.,,

n.,.. ,.U ...... U

. L....

... v V .... . L ...<......... U .... . O.U......

=%.-

. ~ . . .

4

• \ h Submitted by:

Long Island Lighting Company February 1988

n. ,, - . .

s' Y

. a ; : s. ; 2 : z q , ::s = % % s A W , :, ,, , -1 ' "' ' ' . W"<' Y'Mf"~ -

- 2 =- - , . - _ _g y ,.

S...pp.pp d .. .. i .

..p

. . . . . . i %f' . . .

,.eo% tin , O b': ' l T.

gob P

TABLE OF CONTENTS SECTION TITLE PAGE ABSTRACT 1 I.

LII . BACKGROUND AND DESIGN INFORMATION OF 2 SUPPLEMENTAL CONTAINMENT SYSTEM A. Design Goals of Shoreham Supplemental 2 Containment System (SCS) ,

B. Design Criteria For The Supplemental 5 Containment System C. Description Of The. Supplemental 6 Containment System

1. Introduction 6

2. Experimental Test Basis for the Swedish FILTRA Design 7

3. Shoreham SCS Design 7 III. 1983 SHOREHAM PRA ANALYSIS AND PLANT 17 MODIFICATIONS THEREAFTER A. 1983 Shoreham PRA Summary of Results 17 B. Plant Improvements Not Credited in 25 the 1983 PRA IV. PRA METHODOLOGY DESCRIPTION 31 A. Severe Accident Analysis Methodology 31 For The Current Full Power PRA Update Effort B. Core Damage Frequency Evaluation 34

1. Introduction 34

2. Data Update and Modeling Format Changes Since 1983 PRA 34

3. Evaluation Methodology 36

4. Initiating Events 36

5. Major Common Cause Initiators (MCCI) Results 39

6. Core Damage Frequency Results 41

7. Characterization of Uncertainties 49 i

i

SECTION TITLE PAGE C. Containment Event Tree Analysis 60

1. Introduction 60

2. Containment Performance 60

3. Release Pathways to the Environment 63

4. Risk Ranking Methodology for Release Categories 63

5. Release Category Frequency 64

6. Selection of Representative Sequence for Release Categories 65

7. With SCS /s. Without SCS Coaparison 69 D. Source Perms Analysis 83

1. Introduction 83

2. The IDCOR Source Term Methodology 83

3. Source Terms 85

4. Release Category Representative Sequences 85

5. Impact of SCS on Severe Accident Progressic.) 93 E. Offsite Dose Analysis 100

1. Introduction 100

2. Emergency Planning Considerations 101

3. Methodology For Calculation of Doses 103

4. Treatment of Ground Level and Elevated Releases 107

5. Dose vs. Distance Risk Distributions 107

6. DBA Analysis With and Without SCS 109

7. Risk Characterization Using Radius of Injury-Threatening and Fatality-Threatening Doses 110

8. Extended Times for Release Start and Release Duration 114

9. Land Contamination Benefits 115

10. Conclusions 116 F. Summary 139 11 j

h l

LIST OF TABLES Table Title II.C-1 Assumptions used in the 100% Power PRA SCS Analysis II.C-2 Discharge Control and Overpressure Protection Systems Parameters III.A-1 1983 Shoreham PRA Mature Plant Operation Accident Sequence Frequencies by Initiator and Class IV.B-1 Summary of the Core Damage Accident Sequence Subclass (Plant Damage Bins) for the Updated Shoreham PRA IV.B-2 Summary of Changes Incorporated into the PRA Update IV.B-3 Comparison of the Core Melt Accident Sequence Subclasses 1983 PRA Versus FILTRA Plant Configuration (Internal Events)

IV.B-4 Summary of Accident Sequence Frequencies (with SCS)

IV.B-5 Summary of Accident Sequence Frequencies (with 6" Vents)

IV.B-6 Comparison of Venting Sensitivity Analyses IV.C-1 Summary Table Describing the Coupling between Containment Pressurization Rate, Leakage Size, and Break Location IV.C-2 Summary Table Describing the Coupling between Containment Pressurization Rate, Leakage, Size, and Break Location IV.C-3 Shoreham Coritainment Event Tree End State Categoriza'; ion IV.C-4 Summary of Core Melt Accident Frequencies (With and Without SCS)

IV.C-5 Comparison of Release Categories Frequencies Relative to SCS IV.C-6 Shoreham Release Categories and Source Terms Characteristics (With SCS)

IV.C-7 Shoreham Release Categories and Source Terms Characteristics (Without SCS)

IV.D-1 Summary of Shoreham MAAP 3.0 100% Power Representative Sequences (with SCS) iii

- _ _ _ _____ _______ _ _ l

1 LIST OF TABLES (CONT'D)

IV.D-2 Summary of Shoreham MAAP 3.0 100% Power Representative Sequences (Without SCS)

IV.E-1 Release Category Represeatative Sequences, Frequencies and Characteristics IV.E-2 MAAP Release Data Total Rel(ase to Environment IV.E-3 MAAP Release Data Converted to Consequence Model Input Total Release to Environment

)

IV.E-4 EPZ Estimation Consistent with NUREG-0396 Considerations 2 and 3 IV.E-5 Comparison of Pose Exceedance Probability for the NRC and Shoreham (Overwater Corrected) with and without SCS Cases and their Ratios at Fixed Distances IV.E-6 Two Hour Doses at the Exclusion Area Boundary for Design Basis Acciden'ts at Shoreham and the Extrapolation Distance to Obtain Lower and Upper PAG Whole Body or Thyroid Doses iv

LIST OF FIGURES i

Figure Title II.C-1 Supplemental Containment System Conceptual Flow Diagram ,

for PRA-Executive Summary III.A-1 Summary Comparison of the Frequency Point Estimates of Core Vulnerable, Core Melt, and Significant Release IV.A-1 100% Power PRA Project Methodology IV.B-1 The Shoreham Core Melt Frequency (per Reactor Year) Due

) to the Identified Accident Sequence Contributors IV.C-1 Comparison of Released Frequency IV.C-2 Summary of Accident Sequence Frequencies for Various Levels of Potential Consequences IV.D-1 BWR Primary System Nodalization IV.D-2 Shoreham Mark II Containment IV.E-1 Shoreham 100% Power-No SCS Whole Body-1 REM

. IV.E-2 Shoreham 100% Power-No SCS Whole Body-5 REM IV.E-3 Shoreham 100% Power-No SCS Whole Body-50 REM IV.E-4 Shoreham 100% Power-No SCS Whole Body-200 REM IV.E-5 Shoreham 100% Power-No SCS Thyroid-5 REM IV.E-6 Shoreham 100% Power-No SCS Thyroid-25 REM IV.E-7 Shoreham 100% Power-SCS 60 M Whole Body-1 REM IV.E-8 Shoreham 100% Power-SCS 60 M Whole Body-5 REM IV.E-9 Shoreham 100% Power-SCS 60 M Whole Body-50 REM IV.E-10 Shoreham 100% Power-SCS 60 M Whole Body-200 REM IV.E-11 Shoreham 100% Power-SCS 60 M Thyroid-5 REM IV.E-12 Shoreham 100% Power-SCS 60 M Thyroid-25 REM IV.E-13 Average of I,Cs,Te Release Fractions Conditional Mean Radii for Injury and Fatality IV.E-14 Average of I,Cs,Te Release Fractions Conditional Me,an Radii for Injury and Fatality v

1 I. ABSTRACT This report summarizes the methodology and conclusions of a full-scale Probabilistic Risk Assessment (PRA) analysis t performed for the Shoreham Nuclear Power Station. The probabilistic evaluations included an assessment of plant hardware and procedure improvements which have been implemented or planned since the publication of the Shoreham PRA in 1983. These included the subject Supplemental Containment System (SCS), the addition of 3 Colt emergency diesel generators to the existing plant standby power capability, the upgrade to super-enriched boron for reactivity control, improvements to containment isolation capability during station blackout events, and the upgrade of plant emergency operating procedures to the latest BWR Owners Group guidelines. Additionally, the i Major Common Cause Initiating (MCCI) events probabilistic evaluation of 1985 was updated in the area of large earthquake effects on reactor vessel failures and station blackout events.

The PRA update program involved the following major tasks:

a detailed probabilistic assessment was performed of the frequency associated with plant (core) damage states; a detailed containment event tree was formulated to define the frequency associated with a spectrum of radionuclide release categories; the timing and severity of radioactive releases were modeled utilizing the IDCOR MAAP 4

methodology; and, a consequence analysis utilizing the CRAC2 methodology was performed to establish potential dose consequences to the public as a function of distance from the Shoreham Nuclear Power Station. Additionally, the Design Basis Accidents (DBA) described in the Updated Safety Analysis Report (USAR) were reviewed.

A significant component of the PRA effort involved analyzing the performance of the SCS which is a gravel-bed filtered containment vent system based on a modified version of the Swedish FILTPA concept employed at the Barseback Nuclear Power Station. The Long Island Lighting Company established the following three principal goals for the SCS: (1) The SCS is to provide a significant reduction in risk to the general public by enhancing the safety performance of the Shoreham Containment System and by mitigating radioactive releases due to severe accidents, (2) The SCS is to provide improvements in the Shoreham plant's ability to satisfy the technical bases for Emergency Planning requirements, and (3) The SCS is to be implemented without degradation of the design basis safety criteria established for the Shoreham plant in the Updated Safety Analysis Report (USAR).

II. BACKGROUND AND DESIGN INFOPAATION FOR THE SUPPLEMENTAL CONTAINMENT SYSTEM j A. Design Goals of the Shoreham Supplemental Containment System (SCS)

As a step towards further improving Shoreham, the Long Island Lighting Company hss decided to build the Supplemental Containment System (SCS). Insights from risk assessments performed by LILCO, generic findings of related government and industry studies, and the international initiatives on venting resulted in the following goals being established for the SCS:

o The SCS is to provide a significant reduction in risk to the general public by enhancing the safety

)

performance of the Shoreham containment, and by mitigating radioactive releases due to severe accidents.

o The SCS is to provide improvements in the Shoreham plant's ability to satisfy the technical bases for Emergency Planning requirements.

o The SCS is to be implemented without degradation of the design basis safety criteria established for the Shoreham plant in the USAR.

To meet these goals, the SCS is designed to mitigate the consequences of most core melt radionuclide releases by providing a preferred location for a controlled means of venting and filtration. The current PRA update has shown that a structural failure of the primary and secondary containments is avoided for most severe accident sequences because the SCS increases the heat and volumetric capacitance of the primary containment. Furthermore, over-pressure response of the combined primary containment and supplemental containment systems (to a severe accident) has, in general, become more predictable than the response of the primary containment alone.

This concern for "predictability" of the containment response in a severe accident came directly from the 1983 PRA studies. The PRA showed that risk was dominated by severe radioactive releases in which the l

containment failure had been predicted to take place l

in the drywell region of the primary containment. The flow path of the radionuclides to the environment was through the reactor building and resulted in relatively little attenuation. P

)

The SCS design maximizes the benefits of having the suppression pool in the release flow-path by providing a large capacity vent line (i.e., by having a preferred' primary containment vent location) from the i wetwell airspace. This concept is consistent with IDCOR studies which concluded independently that wetwell airspace venting reduced the severity of core melt releases. An important aspect of suppression pool scrubbing in the shoreham PRA was that its success depended upon maintaining the integrity of the drywell-to-wetwell boundary in the Shoreham MARK II containment design. Following melt-through of the reactor pressure vessel, this boundary could be jeopardized primarily due to downcomer melting (effectively bypassing the suppression pool). The damaging consequences of such an event are reduced in

> the proposed SCS design because of the filtering properties of the SCS which is connected in series with the wetwell airspace.

The filtration capability of the SCS gravel bed has been the subject of extensive experimental testing and analytical modeling as part of the Swedish FILTRA program. Experiments performed at Studsvik indicate that very large decontamination factors are to be expected from the gravel bed for particulate releases such as Cesium-Iodide. This subject is further discussed in Section II.C.

It should also be noted that the SCS reduces the frequency of core melt by reducing the impact of postulated scenarios in the Shoreham updated PRA in which containment failure was postulated to occur (due to loss of heat removal capability or ATWS) prior to damage to the nuclear core.

To evaluate the success of SCS to its second goal cf providing improvements in Shoreham's ability to satisfy the technical bases for Emergency Planning, the results of the radiological consequences calculated in the current PRA update effort were compared against the criteria of NUREG-0396. This NUREG provides the technical basis established by the i NRC for developing the generic 10 mile EPZ. Using the l same criteria as a method of measuring the reduction i of risk, distances are obtained for Shoreham with and l without SCS at which the probability of exceeding

certain doses of interest occurs. The results of this analysis indicate that the SCS results in a significant improvement in the dose-versus-distance relationships.

The design of the SCS has been optimized to increase the average time from the initiation of the event to the time of release and also to increase the average duration of the release given a severe accident. Both the average time to release and the average release duration increase for early and moderately early

)

1 releases by a factor of two. Design optimization is achieved through the inclusion of a discharge control and overpressure protection (DCOP) system on the outlet of the SCS. Specifically, the entire SCS

> structure is allowed to pressurize to approximately 40 psig prior to the release of radionuclides to the environment. This results in a significant delay in release times of radioactive noble gases.

The SCS reduces the severity of land contamination given a severe accident. In general, the gravel bed overall decontamination factor associated with particulate releases is greater than 500. This leads to a reduction in land contamination by an average factor of about ten.

To ensure that the SCS adheres to its des'ign goal of maintaining the Shoreham plant design basis, the system is designed to actuate only for low probability severe accidents which are well beyond the containment design basis accident pressures. For all design basis accidents analyzed in Chapter 15 of the USAR, the SCS will remain isolated from the primary containment.

The isolation is accomplished with redundant rupture discs whose setpoints are approximately 80% above the maximum LOCA pressure in the wetwell. In addition, two normally open containment isolation valves are provided. These valves will automatically close in the unlikely event that apurious rupture disc failures occur. Normally closed, redundant containment isolation valves are provided for the two manual drywell vent paths. Thus, the containment isolation system preserves the containment response to design basis LOCA events analyzed in the Shoreham USAR.

1

-4 - 1

?

B. Design Criteria for the Supplemental Containment System In addition to the overall design goals established i for the SCS, LILCO also developed specific design criteria for the system. These criteria were based on scoping analyses performed in conjunction with the current PRA update.

The criteria determined essential to the design of the SCS include:

o The activation of the SCS shall be passive; o The habitability of the Control Room and the Technical Support Center shall not be compromised by the operation of SCS; o The SCS shall be operable during a station blackout (SBO) for a period of 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after a SBO is initiated; o The SCS shall ensure that the primary containment pressure of 70 psig, which corresponds to the maximum containment pressure that will assure the operability of the Main Steam Isolation Valves and Safety Relief Valves, is not exceeded; o The operation of the SCS shall not create a detonable gas mixture in either the primary or secondary containment; o The design of the SCS shall monitor any radiation release; o The design of SCS shall not have a significant adverse environmental impact on safety related equipment within the reactor building secondary containment.

A scoping study was performed on various primary containment vent design configuration options including: (1) a vent to the reactor building: (2) a vent to the plant stack; and (3) a filter or scrubber vent system which terminated outside the Reactor Building. Options 1 and 2 were found to have deficiencies with respect to both the overall design goals and the above design criteria and consequently Option 3 was pursued.

)

C. Description of.the Supplemental Containment System

1. Introduction

As part of the decision process leading to a filtered vent system for Shoreham, LILCO evaluated a number of the systems that other countries have either committed to or have installed at their nuclear plant sites. Among those countries that have committed to such systems are Germany, i France, and Sweden.

The earliest and most aggressive program on containment filtered venting has been in Sweden.

Their first efforts produced a large gravel bed filtering strucrure (i.e., the FILTRA) which was installed at the Barseback Nuclear Station in southern Sweden near the Danish city of

'.openhagen. Recently, Sweden has developed a lower capacity venturi water-scrubbing system which is being installed at the rest of their plants.

The French have decided to install filtered vents at each of their PWRs. Their design is based on the accident scenario which involves core-concrete interaction potentially leading to long term overpressurization of their large dry reactor containment system. Their main objective is to maintain containment pressures close to design values and to provide a decontamination factor of 10 for nuclear aerosols. Consequently, the French decided on a filtered vent system that employs a dual manual valve interface, with the primary containment connected to a sand filter bed roughly 24 feet in diameter and 30 inches deep.

Germany has committed to install a standby filtered vent system for one operating plant (Brokdorf) and is studying a system for its other plants. The Brokdorf system is designed for hookup and operation in 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. It consists of a small vent connection from the hydrogen recombiner system to a packed stainless steel wool filter assembly. This type of system is designed for slow pressurization core melt scenarios that have been postulated in the German Risk Study.

The scenarios typically challenge containment integrity in about four days followLng the initial j accident.

L On the basis of a conceptual design study of l various containment vent options performed by I LILCO in 1987, the system employed in Sweden at the Barseback Nuclear Station was determined to be the most appropriate for use at Shoreham.

t i

\

2. Experimental Test Basis for the Swedish FILTRA Design The Swedish FILTRA design, adopted as the base i model for the Shoreham SCS design, is based on results from the Swedish Research Project

" FILTRA" (IIC-1) initiated in February 1980. The major results and conclusions of the project experiments and analyses showed that:

o A good filtration effect is obtained by usjng a gravel bed with a volume equal to 10,000 m (353,147ft3) .

o The removal efficiency of the gravel bed for aerosol particles and elemental iodine (i.e.,

radioactive substances that could cause long lived land contamination) is 99.9%.

o Condensation of steam in a gravel bed takes place without plugging of the bed by the condensate that has been formed. The condensation front moves slowly downward in the gravel bed providing good conditions for the retention of aerosols and iodine.

o Very rapid pressure transients caused by steam explosions, hydrogen burns and rapid steam formation (flashing) do not constitute a threat to the type of containment studied.

o Filtered venting does not affect other safety functions to any extent, o Venting prevents overpressure failure of the reactor containment for the most probable event sequences that might lead to such failures in the absence of venting. These events include transients and pipe breaks in combination with loss of decay heat removal.

3. Shoreham SCS Design The Supplemental Containment System (SCS) to be installed at the Shoreham Nuclear Power Station is similar to the one installed at Barseback. Some modifications have been made to reflect LILCO's design goals and criteria. The SCS consists of the same three major systems as at Barseback: 1) a wetwell pressure relief system 2) the FILTRA gravel condenser and 3) auxiliary systems (See Figure II.C-1).

- _ _ _ _ _ _ _ _ _ a

)

I

a. Wetwell Pressure Relief The wetwell pressure relief system for Shoreham consists of a 24" diameter relief

! pipe that connects the wetwell airspace (at the existing personnel access hatch at elevation 40' in the reactor building) with the gravel condenser. The pressure relief pipe, which expands to a 30" diameter beyond the secondary containment penetration, is routed in a concrete tunnel and terminates in a manifold embedded in the roof slab of the gravel condenser. Two rupture discs are located in series with two normally open, motor-operated butterfly valves, in the 24" section of the pipe as close to the primary containment as possible. The rupture discs have a setpoint of 60 psig compared to the maximum t.nalyzed wetwell LOCA pressure of 33.7 psig. LILCO intends to perform a series of rupture disc pressure tests in order to finalize burst pressure reliability data.

The two rupture discs, with their combined high reliability, give very high confidence that a single failure in the vent system will not alter the response of the plant to DBAs.

Furthermore, establishing the rupture point of the discs at 60 psig assures that the primary containment pressure will not exceed 70 psig, which is the pressure at which the MSIVs and SRVs become inoperable because of the back pressure on their air operators during severe accident sequences. ,

The 24 inch pipe, between the primary containment up to ar d including the rupture discs, is an integral part of the primary containment pressure boundary. Consequently, this section will be designed to ASME Section III, Class 2 requirements and classified QA Category I. The balance of the pipirq system will be engineered to meet the requirements of ASME Section III, Class 3; however, it will be classified QA Category II. All process piping and civil structures of the SCS will be designed in accordance with Seismic Category I design criteria. An appropriate QA program will be prepared for the implementation of the design and construction of the QA Category II portion of the SCS. The program will ensure that the quality basis developed during the design phase will be carried over into the construction.

_j

Y The control logic for the primary containment isolation valves is designed to automatically close the two MOVs if the rupture discs burst below containment design pressure, i.e., 48 f psig. The valves will, however, reopen automatically if and when the primary containment pressure increases beyond 55 psig.

The MOVs will be powered by the Class _IE Division I and II,_480 VAC power systems. A non-seismic QA Category II battery and an Uninterruptible Power Supply (UPS) will be utilized to provide automatic back-up power through transfer switches, should the voltage of the preferred source at the transfer switches fall below design voltage. Since the Class IE 480V AC system is assumed _ unavailable during a station blackout, the battery will be sized to provide power for a period of 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br /> after SBO is initiated.

In addition to the wetwell pressure relief pipe, the drywell section of the primary containment has been provided with the capability of manual venting to the gravel condenser. To allow for this drywell venting, two 6 inch lines have been incorported into the design. These lines will combine into a 10 inch line in the secondary containment, which then joins the wetwell vent line outboard of the rupture discs. These lines provide the operator with the ability to control the drywell temperature under some severe accident scenarios, and to vent the drywell if flooding of the containment should be desired. The drywell vent lines are provided with remote manual containment isolation valves which are normally closed,

b. FILTRA (Gravel Condenser)

The Shoreham FILTRA building consists primarily of a gravel condenser contained between two concentric concrete cylinders. The annular l walls are carbon steel lined with a 5/16" thick i liner on the outer wall and a 3/16" liner on l the inner wall. The annular space between the cylinders contains 18,700 tons of gravel ranging in size from 1" to 1-3/8" compacted to a void fraction of 35% to 40%. There are two l

penetrations through the inner wall at elevation 25'-7" through which two 26" lines lead into two discharge pipes: a 36" pipe and a 12" pipe. The 36" pipe terminates at a Discharge Control and Overpressure Protection (DCOP) system on top of the gravel condenser.

i The 12" pipe also terminates on top of the gravel condenser but does not have any relief valves. (See Figure II.C-1. Also,'for additional dimensions and other parameters of

( the gravel condenser see Table II.C-1.)

The DCOP, which serves the dual function of discharge control'and overpressure protection, consists of two pressure relief valves and an air operated ball valve. The valves are set up in three pressure set point stages (see Figure II.C-1 and Table II.C-2). The first stage (Stage 1) of the DCOP system is a 12 inch pneumatically operated ball valve, which is air powered to open and spring loaded to close.

The valve is sized to handle 96% of the SCS actuation events and its capacity is as shown in Table II.C-1.

Stage II is a balanced bellows Safety Relief Valve (SRV) discharging to the atmosphere.

This second valve will serve two functions:

1) it will provide additional discharge control and 2) it will serve as overpressure protection for pressure levels above Stage I. The additional discharge control is for the remaining 4% of the GCS actuation events.

These are events that will have mass flows that exceed Stage I capacity. The valve set pressure of 45 psig, or 6 psi above the Stage I operating pressure, is intended to provide sufficient separation between the Stage I valve and Stage II SRV to control the possibility of valve chatter. The valve capacity is shown in Table II.C-2. The combined capacity of Stages I and II are equal to 8% reactor power.

Stage III is also a balanced bellows Safety Relief Valve (SRV) discharging to the atmosphere. This valve will support the SRV capacity of Stage II. The Stage III SRV set pressure is 6 psi above the Stage II SRV to control the possibility of chatter. The Stage III capacity can be found in Table II.C-2. All three stages combined, can pass 16% of reactor power.

Both the Stage II and the Stage III valves are provided with motor operated isolation valves on the upstream side. The isolation valves are normally locked open but can be closed remotely from the control room in the event that either of the valves in Stages II or III is unable to reseat. Position indication is provided for the Stage I, II and III valves.

h' Rupture discs with burst pressures of approximately 17 psia are provided on the discharge side of each valve to afford protection from the environment.

i The decision to adopt the DCOP system at Shoreham was driven by the desire to maximize the: 1) quantity of radioisotopes "bottled-up" in the SCS; 2) hold-up time for radioisotopes to decay; 3) filtration of halogens and aerosols; 4) dispersion of SCS releases by creating elevated releases.

In addition to meeting the above objectives, the design had to accommodate a wide range of flow rates that could be postulated for severe accidents.

Furthermore, the DCOP system on top of the FILTRA had to enable the SCS to satisfy the dose criterion given in the "Manual of Protective Action Guides and Protective Actions for Nuclear Incidents," EPA-520/1-70-00, which states that the whole body dose to onsite personnel should not exceed 25 rem.

Calculations demonstrate that the DCOP system will satisfy this criterion since under Shoreham meteorological conditions at a

. confidence level of 95%, severe accident releases from the SCS will be 89.7 m above sea level. With such elevated releases, the 30 day whole body dose to Control Room and TSC personnel was calculated to be less than 5 rem.

The PRA analyses further demonstrate that the elevated releases will provide sufficient dispersion to reduce offsite doses.

As stated earlier, the Shoreham SCS discharge system also includes a normally closed 12" line with a rupture disk set at 17 psia. The purpose of this line is two-fold:

1. To provide a means to depressurize the SCS when the pressure is below the Stage I set point.

2. To provide a controlled means to esablish filtered venting to the environment, in the unlikely event that a scenario develops, in which the pressure in the SCS is not sufficient to open the relief valves, but the temperature in the drywell keeps on increasing to the point where it threatens the drywell integrity.

11 -

U

r Y

c. Auxiliary Systems The auxiliary systems of the SCS are: (a) the drain-system, (b) the ventilation system, (c) the radioactivity monitoring system, (d) the sampling system, and (e) the process monitoring system.

The drain system collects the condensate created from the 30 inch wetwell pressure relief pipe, following the SCS actuation. The condensate is collected in a tank which empties to the gravel condenser sump with the assistance of pressurized nitrogen gas. The system is passive during the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following SCS actuation.

The ventilation system consists of two subsystems which control the pipe tunnel and auxiliary building temperatures.

The radioactivity monitoring system monitors and registers total gaseous activity that passes through the SCS discharge line to the environment, in the unlikely event that the SCS becomes operational. The systdm also has the capability of collecting samples on filters to quantify the small amounts of halogens and particulates that may be discharged.

The sampling system will provide liquid and gas samples of the gravel condenser'and drein tank.

During standby operation, the gravel condenser atmosphere is sampled in order to monitor its humidity and oxygen / nitrogen contents.

The process monitoring system provides information such as pressure, flow, level, temperature and radiation in the SCS.

The instrument panels for the auxiliary systems are located in the SCS auxiliary building. The information is transmitted via a data link to

! the station computers. Essential monitoring is L

battery powered. Appropriate system

_nstrumentation and displays will be available in the main control room under all conditions and in the Technical Support Center (TSC) and the Emergency Operations Facility (EOF) when electriccl power is available at either the normal or the reserve station transformer.

l 1

)

REFERENCES IIC-1 FILTRA Final Report, November 1982; AB ASEA-Atom and STUDSVIK ENERGITEKNIK AB i

i I

13 -

l

)

TABLE II.C.1 ASSUMPTIONS USED IN 'IEE 100% POWER PPA SCS ANALYSIS

)

1. The FILTRA particulate retention efficiency is asstmed to be 99.83% or to have a DF of 588.2.

2. The FILTRA radiciodine retention efficiency is assumed to be 99.9% or to have a DF of 1000.

3. The FILTRA Design Parameters (or analytical design input parameters) are assumed to be as follows:

Height of FILTRA (without stack) =123ft J Active height of the filter bed =100ft Grade level for FILTRA = 40ft Outer diameter of the filter bed = 65ft Inner diameter of the filter bed =16.5ft Wall thickness of FILTRA =3.28ft 3 Density of FIIHPA wall =145.0lb/f3 Wall area =20,000.0ft Wall thermal conductivity =0.919 B'IU/ft/hg/ F Wall specific heat =0.2199 B'IU/lg/ F Free volume of gravel bed =141,235.0 ft Void fraction of gravel bed =40%

Diameter of spherical Mcles =0.115ft 8 Total no. of particles =2.53x10 7 2

'Ibtal area of particles (= heat transfer area for condenser) =1.05x10 ft 3 Density of particles =165 lbn/ft Total gravel mass =33,068,783.0 (b Specific heat of particles =0.15B'IU/lkn 'T Initial gravel terrperature =70 Initial gas teperature =70 Initial gas pressure in FILTRA =14.7 psia Initial FILTRA humidity =50%

4. The seismic fragility of the FILTRA (SCS) is assumed to be equivalent to the Shoreham reactor building structure.

5. Each rupture disk between FILTRA and primary containment is assumed to fail at 60 psig. ('Ihe reliability of the rupture disk is bounded by 60 + 5 psig.) _

l

)

TABLE II.C-2 DISCIITRGE CONTROL AND OVERPRESSURE PROTECTION SYSTEMS PARAMETERS

)

A. FILTRA DESIGN BASIS 200lb/sec @ 45 psig i

B. VALVE CONFIGURATION AND FLOWS (lb/sec)

VALVE STAGE I STAGE II STAGE III SIZE 39 PSIG 45 PSIG 51 PSIG 12W16 49 54 60 20BB224 168 185 20BB 24 185 C. TOTAL FLOW PER STAGE 49 222 430 NOTE: 1. The relief valve setpoints and release rates are those used in the PRA analysis. Final design values may vary slightly. They will not, however, affect the results of the PRA evaluation.

I

eamen e

, wo-sos->s l

, (L I47 (' _DA CAT I QA CAED_

[L_ij]f St at( 4 - '

y ,.

) ./

C C 2,,C C 4 4'

7$-a6 e e' R

-- E- -

~

) )

C C2-**- CC 4 I

sc wah - ** -sECO*sCA# v - O CON T A.%AT N T CC4 Ta the*(h f Ostwf LL

.D -

vv s y.,

db i.

/ /

Q l,, CC!gt 4

. ... g -,

wa ~

- VJ 7=a.a-winu acw aim w -o

_w __-

I a.

/%, f E.e

%' "3'*.'* t~* n * " ~O T lj 'l~ maisamt.. ,.

I ' li,,r.ao

-= , , ,

- = 1

.rc,.

nu

• • • *i pm amouro . ri. ^

3 9Ay 5 cgy, ,;,,,_

,tesi ht e t

J i

Et' y=

1 I

nu l% a ca.a u-l 7nov,a cmA i _.

l4 i 33' xLO A

l' r.i . 'lli

, y . :_ : -u. n fU v

S

' q ::m:6

. u.2

" v g -r~' ll g ....<. - ..: a a -

s s,,,, .

.-, ur m s

.1 w5% h f fa j cc.a ssart I  :" , J' g ,.I..$$

,s / x oo t w aws' ce a e.

4to a vn 3 o 7o Sic'" " 7 ta= r-"9

-- tm;+a ,j

,;, h phd rf P y TO sap ^ t 80 Cat L.--s eu

.a . . . _

""5 ""

om ,s

~~ ~ _ . _ tt svu

- - - - - - - - - - - - - , - - , ,-- -- - - - - - - - ~

meg e%- 4 * * * * ~ h gl ....... , . - - - - - -

--  : _ . , . s ..c. mA-~ a .-

2.'-, '._ "

2 's K' T*g'2s y f

~ ~ :: 4 ' , ' ' ' %*l , L ' ., *2 T.'

3 , W -+ * *. , ' ' , ' + ' } ' ' G i f '

. . = .

M

g i * -

i i 1 (

LEGEND P( AC - peedARY (04fAhdlief Aftde$h(OC (0*sTR A bYSitw f - TElf O - CEAre v - WNT 4 [D-e w m.

8tD

• ntatyps pgc -

4r S ,

C) 9 h ,

4vf -

i NOTES: _

SEE NL7?E 6 ECEacrE6 1. THtS FLCD DAG A AM #6 A CitAC,A A4SA'C A(Pp(Steifais0%

=, & OF A C0deC(P fLAL L(5% AND 6 frA INf 0AW AT'Oss s :r,y ~.

, TC T" G. .

&*,qi5 m ^ \ n Nde

,.-'~[3 y

-Of 0- of... , ,ge gg _.

24'3 y.")

6 RWTLitt OiSCS A31E PSOwlOEQ Pgt w(af MER P6 0f ECT104 _

• -'2, or sELIEF batvE INTERhALS SET PREi$uRE r7 esta.

vnao L * *

7. 34 a 24 ACM P191 TEE SMALL H A=E AY LE AST A l *") edin AL mLA L TMICahE 55 OF 43/ ser(M 36 FLA84(5 h,j

. g,

I

- ~ -

-l ,,: p SM ALL BC 150 L95 4

< , ,n u - A - ~ .

( gutgyA,ATEo aves carA we l I

V .nt I I

mAvn, eco GAA 4 ecd y e e ' .es.4 G 2 aos cuCuseFTm,adE e atue AW

= i W -

1 ge <

u 6 -

a_.__ #

- l 7.c- s f-w l

_ ti j

~

~ was ,cpw(  !

' W M CN w- - * '

t: %G.'r dt T

' "N7 --W[.l

,e r~

O3/70 t l

ow= ros -

To ma ts . l M.$em 'g* ,e I

<< l il '

i' sn A. m _-

vtNT uaE >

ll - - -e  ;

l FIGURE II .C -l

- eq so crmP.e r ,. .'ecm sc:

. ._e

.a 9.m

-t c a Aa es x ;qgt,tc. To staw<: cL A's i , _ , , , , , , , , 'o j

850 MNe o A CAT I .H h

'^#

SHOREHAM NUCLE AR PO AER ST ATION j M ew' T i LONG ISL AND LIGHilP6 COMPANY HlCKSVil { E , NEW YORK SUPFtD. ENTAL CCt4 tat *WENT SYSTEM

-_..____ y , -

,.y _ . _ _ _ ,

_ _ _ CONCEPTUAL FLON DORAM

....... _ i j__-,-.- i l

FOR PRA-EXECulivE SIMMARY

__._a rl 6  ; _ . J;-R . _, "-' ?%r- = SK- SCS-004 (t. ut:7 99 a . :p p ~ y:p 4- .- . , .:. . ~ -

t =.: =yj ppj:: ( } = (_q =.- 13 .-,-ogl 4

Page 16

III. 1983 SHOREHAM PRA ANALYSIS AND PLANT MODIFICATIONS THEREAFTER A. 1983 Shoreham PRA and Summary of Results

)

The quantification of the risk associated with the operation of the Shoreham Nuclear Power Station was completed in 1983 and consisted of an extensive set of analyses (probabilistic and deterministic) encompassing the requirements of a Level 3 PRA

) (IIIA-1). As a Level 3 PRA, the frequency of severe accidents, the likelihood of loss of containment integrity, the radionuclide source term associated with severe accidents, and the radiological impact on the public were calculated.

The 1983 Shoreham PRA included a detailed Level 1 PRA evaluation of accident sequences that could lead to core damage states. The evaluation utilized a comprehensive set of system level fault trees coupled with event tree diagrams for use in the quantification of accident sequence frequencies. The data used in the quantification was based on industry operating experience since Shoreham had not yet begun operation.

The PRA quantified the different effects on plant response under a variety of initial conditions, ranging from the more common types of anticipated transiente to very low frequency initiators and LOCA sequences.

Five accident classes were chosen to represent the spectrum of accident sequences from (relatively) high d frequency / low consequence events to low frequency /high consequence events. These classes are defined as follows:

CLASS DESCRIPTION I Inadequate Coolant Inventory Makeup II Inadequate Decay Heat Removal III LOCA with Inadequate Coolant Inventory Makeup IV ATWS with Inadequate Containment Heat Removal V Interfacing LOCA The conclusion from the original 1983 Shoreham PRA was that the core vulnerable frequency was low and that there was no single accident sequence or set of sequences that so dominated the spectrum of plant damage states that their alimination would

dramatically change the calculated risk. Table III.A-1, from the 1983 PRA, supports that conclusion because there are a numoer of different potential contributors to core vulnerability which are

) distributed over sequence and initiator type.

The 1983 Shoreham PRA also investigated the core melt processes and subsequent radionuclide releases for the accident sequences which follow from the core damage states that were previously identified. The analysis y included the evaluation of the containment response and radionuclide behavior within containment. Major features of the in-plant consequence assessment include development of an accident progression model specific to the Shoreham Mark II containment and a comprehensive treatment of the systems and

) phenomenological interdependencies in the development of the detailed Shoreham-specific containment event trees. Containment event tree data were used to establish the dot.inant accident sequence progression leading to a release given a core vulnerable event and the relative frequencies of a spectrum of possible releases. Radionuclide release characterization in the Shoreham PRA was performed using sixteen release categories compared with the five defined in the Reactor Safety Study (IIIA-2).

Radionuclide transport models (MARCH 1.1 and CORRAL computer Codes) considered specific dominant accident sequences for the analysis of core melt progress, ion.

Radionuclide transport within the containment models the mechanisms of radionuclide release from the fuel and removal processes from the containment atmosphere using data from experimental work performed since WASH 1400. The processes for fission product removal were generally more realistic than those used in the Reactor Safety Study (RSS), i.e., (1) primary system retention, (2) suppression pool scrubbing and (3) natural renoval by mechanisms such as plateout or settling. The treatment of the first two mechanisms included later (1982) determinations of aerosol behavior. The third mechanism was treated using more realistic and detailed compartment models of the Shoreham containment system. The 1983 Shoreham PRA models did not include revaporization of radionuclides predicted in subsequent IDCOR work with the MAAP computer code. (The current PRA update utilizes the MAAP 3.0 computer code which is described in Section IV.D of this report.)

The 1983 Shoreham PRA processed the data from the system level event tree end states, as described previously, through a set of containment event trees.

The containment event trees modeled the possible range of plant and operator responses during severely 18 -

degraded conditions involving multiple failures and plant conditions which may be substantially outside design.

j The following brief description taken from the 1983 Shoreham PRA provides a useful perspective on the characteri=ation of the accident sequence frequencies (refer to Figure III. A-1) :

o Core / Containment Vulnerable: There are a set of 3

low frequency states which could occur at Shoreham for which the containment or core integrity could be challenged. Bar graph A is the sum of such states, o Core Melt: Given a core / containment vulnerable condition, there is a fraction of those challenges that are recoverable by proper operator action or use of alternate systems. For the Shoreham PRA the containment event tree and system analysis establish that approximately 20% of the core / containment vulnerable conditions can realistically be recovered to prevent core melt, o Radionuclide Release: A subset of the core melt sequences that provides a direct input into the ex-plant consequence analysis is the set of sequences which result in a radionuclide release.

This group covers the spectrum of all postulated releases subsequent to a degraded core condition from cases involving the release of noble gases (the more likely) through cases involving the release of substantial fractions of the core inventory (remote possibilities).

o Significant Radionuclide Release: The determination of public health effects has been shown in past PRAs and studies by the NRC to be sensitive to specific species of radionuclides.

Bar graph D of Figure III.A-1 represents those accident sequences for which substantial amounts of Iodine and Tellurium could be released. These are the sequences for which reduced containment retention occurs, e.g., the suppression pool is partially bypassed or the containment integrity is challenged early in the sequences.

o Large Radionuclide Release: In addition to the lower consequence events, there is also a group of sequences which have been postulated for which significant amounts of radionuclides could conceivably be released to the environment. These sequences have the possibility of causing early health effects in the public if the release occurred. However, as shown in bar graph E of

)

Figure III.A-1, the frequency of such sequences is extremely low, 2 per 10-million reactor years.

This frequency is far below our experience threshold and encompasses sequences which are deemed possible, but are highly unlikely.. Based

)

on the 1983 Shoreham PRA, an EPZ.of 2-3 miles was projected by LILCO (IIIA-3) based on the criteria of NUREG 0396.

)

?

REFERENCES IIIA - 1 Probabilistic Risk Assessment Shoreham Nuclear Power Station, Long Island Lighting Company, June

)

1983.

IIIA - 2 Reactor Safety Study, An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, WASH 1400 (NUREG 75/014), October 1975.

IIIA - 3 Testimony from the Governor's Commission on Shoreham (Marburger Report).

)- ,

i.

)

Teble III.A-1 1983 SHOREHAM FRA MATURE PLANT OPERATION ACCIDENT SEQUENCE FREQUENCIES j BY I'4ITIATOR AND CLASS (EVEhrS PER REACTOR YEAR)

CLASS CLASS CLASS CLASS CLASS SEQUENCE y EVENT INITIATOR I II III IV V TOTALS

)

Transients:

Turbine Trip 2.5E-6 1.OE-6 -- -- -- 3.5E-6 Manual Shutdown 1.4E-6 1.2E-6 -- -- --

2.5E-6 MSIV Closure 7 4E-7 3.5E-7 -- -- --

1.1E-6 i Loss of Feedwater 2.OE-7 4.2E-8 -- -- --

2.4E-7 Loss of Condenser Vacuum 3.2E-6 2.1E-6 -- -- --

5.2E-6 Loss of Offsite Power 9.9E-6 5.7E-7 -- -- --

1.0E-5 IORV 6.8E-7 8.9E-8 --

, - - -- 7.7E-7 Subtotal Transients 1.7E-5 5.9E-6 2.4E-5

.LOCA:

1-Large LOCA --

6.9E-7 1.8E-7 -- --

8.7E-7 Msdium LOCA --

2.7E-7 5.1E-7 3.0E-8 --

8.0E-7 Small LOCA 2.1E-7 2.8E-8 1.5E-8 -- --

2.6E-7 LOCA Outside Containment --

7.2E-9 -- --

3.6"-8 4.3E-8 Raactor Pressure Vossel LOCA -- -- 3.1E-7 -- --

3.1E-7

. Subtotal LOCA 2.1E-7 9.9E-7 1.0E-6 3.0E-8 3.6E-8 '2.3E-6 ATWS:

Turbine Trip 1.2C-6 --

8.5E-10 2.3E-6 --

3.58-6 MSIV Closure / Loss of Condenser vacuum 8.0E-7 --

7.5E-10 7.4E-6 --

8.2E-6 Loss of Offsite Power 7.1E-8 -- --

6.9E 7 --

7.6E-7 IORV 1.7E-7 -- --

1.6E-7 --

3.3E-7 Loss of FW 1.8E-6 --

2.1E-9 3.0E-6 --

4.8E-6 Subtotal ATWS 4.0E-6 3.7E-9 1.4E-5 1.8E-5 22 -

Table III.A-1 (Continued) 1983 SHOREHAM PRA MATURE PLANT OPERATION ACCIDENT SSQUENCE FREQUENCIES BY INITIATOR AND CLASS (EVENTS PER REACTOR YEAR)

CLASS CLASS CLASS CLASS CLASS SEQUENCE y EVENT INITIATOR I II III IV V TOTALS Other Transients:

Cases involving the Releese of Excessive Water 3.1E-6 7.8E-7 --

3.9E-10 --

3.9E-6 Cases Initiated by the Loss of DC Power Bus 2.7E-6 7.4E-8 --

4.4E-8 --

2.9E-6 Cases Involving an Upset Condition with the Reactor Water Level Measurement System 2.4E-6 1.2E-7 --

1.9E-7 --

2.7E-6 Manual Shutdown due to High Drywell Tempera-ture 1.4E-7 -- -- -- --

1.4E-7 Subtotal (Water Level) 2.4E-6 1.2E-7 1.9E-7 2'."8 E- 6 Measurement Failures) 1 css of Service Water Initiated Events 3.lE-7 6.9E-7 4.6E-8 1.0E-6

,3 CLASS TOTAL 3.lE-5 8.6E-6 1.0E-6 1.4E-5 3.6E-8 5.5E-5 NOTES:

1. Core vulnerable frequency totals may not match due to round off.

2. The Shoreham core vulnerable frequency of 5.5E-5/ year represents those sequences where core or contafrn.ent damage is imminent. The core vulnerable category includes some potentially receverable sequences wherein operator action or alternate systems can be used to avert core damage or containment failure. Nonrecoverable sequences progress to core melt. The core melt frequency associated with the 1983 PRA (internal events only) is 4.9E-5 events per year.

3. The 1983 PRA addressed internal events only. \n e::ternal event core J7 melt frequency of F.8E-6 was developed in the 1995 NUS MCCI report g (see Section IV.B for more details). The total core melt frequency 0; (internal and external events) originally developed for Shoreham was n 4.9E-5 plus 9.8E-6 or 5.92-5 events per year.

23 -

ir

l

}

}

1.

s.st.s 4.It.1 3,g,g

.?. . i s

O '

3.N.s u

5 '

s i.d .

3 .

9

= i . x.t e

y i. ' -

E i.' .~ .I -

com m at w it L5 ".'rJc'" .T.c=tustW,'c.,wai,,

.t6ctst A 't C e tretts or rotterin wiesuctix actruc

• 'Ihis core melt frequency estirrate was updated to 4.9E-5

, during tha cueront PftA update, based on a more stringent interpretation of core damasc. (See Table II .B-3 for trore details.)

Figure III..A-1 Sumary Comparison of the Frequency Point Estimates of Core Vulnerable, Core Melt, and Significant Release 1

i

=:a______-____________________

)

I f B. Plant: Improvements Not Credited In The 1983 PRA

1. Introduction

) A primary benefit of the Sh>reham PRA (IIIB-1) is- j that it provides a framework for the assessment of the risk benefit of proposed modifications. This framework has enabled LILCO to focus on potential improvements to risk significant sequences such as h, ATWS and Station Blackout (SBO). _The relative l significance of these' sequences was also confirmed in the SNPS Individual Plant Evaluation (IIIB-2) and the Request for Authorization to Increase Power to 25%(IIIB-3).

Since 1983, LILCO has implemented (or plans to

! implement) several modifications to further r thefrequencyofpostulatedsevereaccidents.gpuce The changes in the Shoreham hardware and procedures th6t have affected the calculated core melt frequency or the public safety, are as follows:

a. Hardware o Installation of the Supplemental Containment System; o Inclusion of an ADS inhibit switch; o Increased boron enrichment for the SLC system; o Addition of three Colt diesel generacors; o Addition of a 71ack Start" 20 MWe gas turbine on-cite; o Modification of the isolation valves on the drywell drain lines.

1/ Based on insights gained during the development of the 1983 Shoreham PRA, LILCO elected to: (1) install a corium ring to direct molten core debris to the suppression pool during reactor pressure vessel (RPV) failure; and (2) modify the MSIV isolation setpoint on low RPV water level to help ensure the continued availability of the main condenser, in conjunction with the lower wat'r levels associated with ATWS reactivity control measures. These modifications are considered in the 1983 PRA and are therefore not discussed

, he,re.

l i

5' . _ _ _ . _ . _ . _ _ _ _ _ _ _ _ _ _ _ _ _

)

l

b. Procedural o Incorporation of the reactivity control guideline from EPG Revision 4;

)

o Direction to maintain the HPCI and PCIC pump suctions on the Condensate Storage Tank (CST);

o Incorporation of containment venting direction utilizing e::isting drywell and wetwell 6 inch lines; o Inclusion of the use of the ADS inhibit switch in both ATWS and non-ATWS accident sequences.

These modifications are discussed below. The Supplemental Containment System has been previously addressed in Section II of this report.

2. Discussion of Modifications

a. ADS Inhibit Switch The addition of the ADS inhibit switch was implemented as part of the ADS logic changes proposed in response to Item II.K.3.18 of NUPEG 0737. The modification consisted of the addition of a manual switch to reset the ADS timer (if initiated) and the appropriate control roota annunciation. The ADS timer will not activate until the inhibit switch is placed back into the "NORM" mode. ,

The use of the ADS inhibit switch has been incorporated into the Emergency Operating Procedures. The switch enables the operator to avoid an uncontrolled depressurization.

Ur. der ATWS conditions , the use of the inhibit switch enables wrter level control at low RPV levels without the continual resetting of the ADS timer. This provides greater assurance that the operator will avoid a power excursion associated with inadvertent ADS opera' cion and uncontrolled 17w pressure injection, as well as the pntential for pentaborate washout from the RPV.

b. 6LC Pentaborate Enrichment The SLC system modification to utilize highlv

- -iched sodium pentaborate was initiated as a re:mit of ATWS Rule (10CPR 50.62) require-ments. The Rule requir+: an increased SLC l

1 l

. injection capability, equivalent in control capacity to 86 gpm of 13 weight percent natural pentaborate, based on a 251 inch I.D.

vessel. The injection rate for the SNPS 218

) inch vessel, based on a chemical concentration of 12% is approximately 76 gpm. There are two basic compliance alternatives; two pump 1

operation or the use of enriched boron.

LILCO examined the risk attributes, as well as

) the complexity of both alternatives, and elected to use enriched sodium pentaborate in-the SLC system. A study was performed to optimize the pentaborate enrichment using the original PRA as a basis.. LILCO specified the use of 85 atom percent enriched sodium

) pentaborate in the SLC system. This high enrichment provides an injection rate that is double the ATWS Rule requirement and approximately quadruple the control capacity of the original system design.

• The use of highly en::iched sodium pentaborate allows an operator action time of at least 25 minutes to successfully initiate SLC injection. This increased time for operator action results in a significant increase in the estimated operator success probability associated with the manual actions to mitigate a postulated ATWS.

c. Additional AC Power Sources The emergency AC power system will include six onsite emergency diesel generators. Three new Colt EDGs are being added to the original plant design in parallel with the existing TDI diesel generators. There are three emergency buses in the system; each bus will be fed by either a Colt or TDI EDG during operation.

One diesel in each pair will be preselected to act as the lead and will automatically start on an accident signal. The alternate diesel and its main breaker will automatically trip in order to ascure that both diesels do not load onto the bus. If the lead diesel degrades or fails to start, the alternate diesel may be remotely started from the Control Room after the engine cooling water is manually aligned.

)

The addition of one diesel to each bus greatly reduces the risk of a station blackout. Only a single diesel on emergency bus 101 or 102 is required to mitigate a potential event. Bus 103 does not power a sufficient amount of safety related equipment to assure plant shutdown.

In addition to the Colt emergency diesel generator modification, power restoration capability at Shoreham following loss of offsite power is augmented by the installation of a 20 MW gas turbine with black-start capability onsite. The black start gas turbine is connected to the 69 KV switchyard which feeds offsite power to the plant through the Reserve Station Service Transformer.

Electrical faults which occur on the grid are isolated from the circuit between the 20 MW gas turbine and the plant.

The gas turbine can be operated in either an automatic or remote manual mode. Starting air is stored in pressurized receivers of sufficient capacity to allow three starting attempts without recharging. An automatically controlled air compressor within the enclosure maintains the compressed air supply. The distribution system has a battery. A battery charger maintains the battery charged at required levels. Fuel is provided from an onsite 1,000,000 gallon storage tank. Two fuel pumps deliver fuel under pressure to the gas turbine,

d. Containment Isolation Valve Modification An additional SBO related modification involves the containment isolation provisions for the drywell equipment and floor drain lines. The Containment Isolaticn Valves (CIVs) on these lines are presently AC motor operated valves. Upon loss of all AC power the valves will remain in the normally open position. In lieu of taking credit for hand operation, LILCO has elected to either repower .

the valves with DC motor operators or install a fail closed AOV on each line just downstream of the CIVs. Either of these options will assure automatic containment isolation under loss of all AC power conditions.

28 -

e.- Emergency Operating Procedures In addition to the hardware modification cited ,

above, the PRA update also considers the h implementation of the Shoreham specific procedures based on the Emergency Procedure Guidelines (EPGs), Revision 4. Specifically, this study credits the incorporation of

-Contingency 5, "Level / Power Control", of the EPGs. Unlike earlier procedures, the operator I is no longer restricted to the high pressure systems for RPV injection during an ATWS. In addition to the low pressure injection system suc7ess path, the lower RPV pressure coi.cidently reduces ATWS core power levels, leading to a greater likelihood that the

) generated power will be within the SCS capability.

A second procedural modification that has been '

included is the requirement to maintain the HPCI and RCIC pump suctions on an external 5 source of water by defeating high suppression pool / low CST level pump suction tranrfer logic. This removes a potential sysiem failure mode due to high lube oil ten.peratures when high temperature suppression pool water is used for.RPV injection.

~

Containment venting procedures through the existing 6 inch drywell and wetwell vents have been included as prevention and mitigation measures as part of the existir.g Emergency Operating Procedures. These procedures provide direction to mitigate overpressure challenge to the containment,

REFERENCES IIIB-1 Probabilistic Risk Assessment - Shoreham Nuclear Power Station, Long Island Lighting Company, June, 1983.

)

IIIB-2 IDCOR Individual Plant Evaluation Method, Applied to the Shoreham Nucle.1r Power Station, Long Island Lighting Company, April, 1986.

)

IIIB-3 Authorization to Increase Power to 25% - Shoreham Nuclear Power Station, Long Island Lighting Company, April 1987.

L i

J i

i l

l l

i r

i

~. .

l IV. PRA METHODOLOGY DESCRIPTION A. Severe Accident Analysis Methodology for the Current Full Power PRA Update Effort

) Since the Shoreham PRA was first published in 1983, LILCO has incorporated, or intends to incorporate, a number of design and procedural modifications to enhance the safety of the plant. An update of the 1983 PRA was performed to evaluate the modifications in terms of public safety. The general methodology

) established to perform the current PRA update parallels closely the original 1983 effort. Figure IV.A-1 presents an overview of the scope of the current evaluation.

The core damage frequency has been determined by I

IT/Delian Corporation and NUS Corporation. The IT/Delian evaluation is a Level I PRA in which the impact of the SCS has been factored into the spectrum of possible plant damage accident classes. New accident subclasses are defined specifically to handle unique aspects of SCS on the accident progression.

The NUS evaluation of external events (seismic, fire) represents a major input to the Levdl 1 probabilistic core damage frequency assessment. The NUS work was originally completed in 1985 and documented in a report entitled "Major Common Cause Initiating Events

- Shoreham Nuclear Power Station" and is commonly referred to as the MCCI report. Portions of the MCCI report regarding large earthquake induced core melts were updated as part of this study.

The Level 1 portion of the analysis includes the definitions and frequencies of Plant Damage States to be used as input by Science Applications International Corporation (SAIC) to the containment analysis.

The containment event tree analysis p.rformed by SAIC results in the definition and quantification of release categories. The release frequency determination originates with the plant damage states defined by IT/Delian Corporation, but factors in containment related events (failure location, size, etc.) which affect the timing and severity of a release and the potential for mitigation. These events are analyzed formally in Containment Event Trees (CET). Fauske & Associates Incorporated (FAI) provides a detailed deterministic analysis of the accident progression by utilizing the !1AAP 3.0 computer code. The MAAP analysis performed by FAI defines the timing and radioactive release characteristics of specific accident sequences utilized by SAIC to represent the release categories.

In addition, MAAP analyses were utilized to define success criteria by which specific accidents could be mitigated in the IT/Delian and SAIC analyses.

31 -

l As input to the containment event tree character-ization developed by SAIC, a structural integrity analysis was performed by Stone & Webster Engineering Corporation (SWEC). The SWEC work updates the ultimate pressure capability analysis originally

)- performed for the 1983 Shoreham PRA to account for postulated temperature induced failures of the primary containment.

Radiological consequence analyses were performed by Pickard, Lowe, and Garrick (PL&G) utilizing the CRAC2

) computer code for each of the release categories defined by SAIC. The release categories are weighted by their relative frequencies in order to define composite radioactive release consequences that represent the spectrum of releases. Consequences are expressed as the conditional probability of exceeding I

a given dose level in the event of a core melt as a function of distance from the Shoreham plant.

LILCO performed radiological analyses relating to Design Basis Accidents and onsite doses.

The methodologies and results of this project will be reviewed by a Peer Review Committee consisting of Dr.

E. Fuller, Dr. J. Hendrie and Dr. R. Lahey. All ree are well known for their expertise in nuclear sate.y.

These experts also participated in the Peer Review of LILCO's 25% Power PRA Analysis performed in 1987.

s 1 .

l l

FI G U R E II. A - l 100 % POWER PRA PROJECT METHODOLOGY l

)

OVERALL CORE DAMAGE FREQUENCY SEVERE AC CID ENT CHARACTERIZATION c r., ANALYSIS AND (IT/DELIAN CORP. ) SOURCE TERMS

)

(FAI) d A

> EVALUATION OF EXTERNAL EVENTS CORE DAMAGE FREQUENCY (SEISMIC, FIRE, ETC )

(NUS CORP.)

~

f CONTAINMENT CONTAINMENT EVENT TREE STRUCTURAL INTEGRITY r AND RELEASE CATEGORIES (SWEC) (S AlC)

I RADIOLOGICAL CONSEQENCES (PL 8 G) l DBA AN ALY SIS RISK ASSESSMENT AND m .NUREG-0396 APPROACH (PL8G)

ONSITE DOSE - INJURY-THREATENING DOSE (NUS)

(LILC 0) l i_ .-- .....-_ - _ - - -.-.. - -_- . - . - . . . . . . - . - _ . - . - - . - . -. --

)

B. Core Damage Frequency Esaluation l

l 1. Introduction This section summarizes the core damage frequency portion (IVB-1) of the current PRA update and provides a discussion addressing: initiator data update /modeling changes; the evaluation methodology; the spectrum of accident initiators; major common cause initiators; a summary of the results and the evaluation uncertainties.

This core damage frequency evaluation includes the major aspects of a Level I PRA and encompasses both internal and external events.

The objectivas of the accident sequence frequency evaluation include:

o Identification of the types of sequences and their characteristics; o Identification of the potential contributors to core damage and calculation of their frequency for input into the calculation of risk; o Identification of the boundary conditions (including containment conditions and mitigating system status) which will affect the containment event tree evaluation; o Identification of dominant contributors to core damage.

2. Data Update and Modeling Format Changes Since 1983 PRA.

Since the 1983 Shoreham PRA (IVB-2), additional operating experience has occurred and data have been collected which assist in quantifying the probabilistic models associated with accident sequences. The principal areas where accumulated data can have an impact are in the following:

o the frequency of anticipated transient initiating events; o the frequency of loss of offsite power initiating events.

In this update effort, each of these frequencies has been updated with the latest available information, i.e., NUREG/CR-3862 for anticipated transients, and LILCO grid specific data for LOOP events.

_ _ . _ _ _ . . . . . . . . _ _ _ _ J

As a result of an analytical evaluation by NUTECH (IVB-3), the conditional probability of an interfacing LOCA 2/ has decreased because of the low pressure pipe systera pressure retaining

) capability.

In addition to the data updating that was performed, there were additional changes in the modeling format. However, these changes in format by themselves, did not result in any

) numerical changeu to the overall core melt frequencies for each of the accident classes defined in-the original PRA. These modeling changes included:

o Minor changes in the event tree model

).

description which collapsed several nodes to one (e.g. high pressure coolant makeup and containment heat removal);

o Expansion of the ATWS and Class II event trees to incorporate the SCS and containment

> unfiltered vent configurations; o The ATWS and LOOP event trees have been updated to ensure that the BWROG EPGs and the Shoreham specific implementation are accurately reflected in the quantification; o The LOOP analysis also includes the expansion of a LOOP specific support state event tree to ensure that unique support states involving DC power unavailability are accurately included.

2/ An interfacing LOCA is a inrge ECCS pressure boundary failure that is postulated when normally closed isolation valves open, allowing the low pressure portions of the ECCS to be exposed to the high temperature and pressure of the RPV.

n' l

1

3. Evaluation Methodology l The methodology used in the current reevaluation of the accident sequence frequencies is the same as used in the 1983 Shoreham PRA. A set of initiating events were identified which may challenge the capability of the plant to reach a safe stable state. Event trees are used to define the principal accident sequences to be evaluated; i.e. the pathways that could lead to

)

different challenges to plant equipment and operating staff actions. System level fault trees are used to quantify the individual system unavailabilities; functional fault trees provide the integration of similar systems for the quantification of event tree nodes.

4. Initiating Events A broad spectrum of initiating events that may lead to challenges to the capability of the plant to reach a safe stable state were defined. The j initiating events include:

o anticipated transients; o LOCAs; o anticipated transients with a failure to scram; 7

o special transient initiators.

The event tree methodology is chosen as the framework to investigate and display the results of the probabilistic engineerino evaluation of the Shoreham plant systems. The etant trees delineate the system and operator response which may follow the initiating events. The sequences are tracked to either: (1) successful hot shutdown of the reactor (which is the case in most of the sequences) or (2) to the point at which a challenge may exist to the core or containment integrity. The sequences are then binned into classes with common containment challenge characteristics. Subsequently, these sequences are then incorporated into the containment event trees to develop estimates of postulated radionuclide release frequencies.

The Shoreham PRA update develops and quantifies separate event trees for those initiating events which may have a markedly different effect on the systems available for accident mitigation and plant cooldown subsequent to the initiating challenge. Based on this philosophy, the following event trees were developed from the accident initiators:

)

a. Transient Event Trees o Turbine Trip o Manual Shutdown y o MSIV Closure o Loss of Feedwater o Loss of Condenser Vacuum o Loss of Offsite Power o Inadvertent Open Relief Valve

b. Loss of Coolant Event Trees

)

o Large LOCA o Medium LOCA o Small LOCA o Large LOCA Outside Containment o Reactor Pressure Vessel Rupture

c. ATWS Event Trees

d. Other Initiator Event Trees

o Water Release onto Elevation 8 of the Reactor Building o DC Bus Unavailability o Malfunctions Associated with the Reactor

> Water Level Measurement System o Loss of Service Water

e. External Event Evaluation

f. Loss of Containment Heat Removal (SCS Summary)

Potential consequences or major interfunctional dependencies are differentiated by the success or failure of the nodes identified in the event trees. Items which affect the nodal functions and their reliability, but do not significantly change the level of consequences are addressed in the fault tree models. The fault tree models of l each system identify the potential modes of l system unavailability due to hardware failure, i human interactions, testing and maintenance.

l System interdependencies are treated either through the Boolean combination of the fault tree models using the WAM series of computer codes or l directly reflected in the dependent conditional l failure probabilities of the event trees.

The event trees constructed for the Shoreham PRA generate several hundred accident sequences for analysis. While a continuous spectrum of potential accidents and consequences can be postulated to occur, the spectrum has been, for the purposes of this analysis, represented by discrete end states. The end states of the systemic event trees may differ in consequences

) and/or effects on system recovery, including accident mitigation. Because a number of the accident sequences have a similar impact on containment and the potential for release of radioactive material, these similar accident sequences are grouped or binned into classes and

) are treated identically in terms of impact on the containment event trees and potential consequences. Table IV.B-1 summarizes the accident classes used in the Shoreham update.

These classes are the same as in the original PRA except for those additional classes required to j accurately encompass the SCS modification or the present non-SCS configuration utilizing the existing 6 inch containment vents.

The principal modeling change to the event tree structure in the updated PRA analyses is in the

)

treatment of the SCS as a containment heat removal path. This treatment is evident for two types of accidents: (1) the Class II loss of decay heat removal (TW) accidents, and (2) the Class IV ATWS. For each of these accident types, a SCS event tree has been developed to assess the

conditional probabilities of the various scenarios possible with SCS installed.

The ATUS SCS event tree is incorporated directly into the ATWS section and is essentially an extension of the event trees. On the other hand, the Class II events are treated such that the TW-SCS event tree is used to assess the Class II event sequences from all event trees. Therefore, all Class II events from the original PRA (with updated quantification) are transferred to a TW-SCS tree for disposition, given that the SCS is a last resort containment heat removal method.

5. Major Common Cause Initiators (MCCI) Results The "Major Common-Cause Initiating Events (MCCI) l Study" for the Shoreham plant was completed in 1985 (IVB-4). The focus of the study was to obtain insights into the plant's susceptibility to, and inherent defenses against MCCIs. Major l common-cause initiating events are occurrences which have the potential to initiate a plant l

transient or LOCA and, in addition, damage one or more plant systems needed to mitigate the effects of the initiated transient or T.OCA. The scope of I the MCCI study included detailed analyses of seismic events and fires through to core damage; i

I ._

1 and bounding analyses of aircraft crashes, windstorms, turbine missiles'and releases of hazardous. materials near the plant. The scope of the study was based on the results of previous

) full scope PRAs available at the time of the MCCI project initiation (e .g. , Limerick, oconee , Zion, and Indian Point).

Based on mean core damage frequencies, the 1985 MCCI study concluded:

o Fires, at 11% of the total frequency of core damage, are significant but not a dominant )

contributor to the core damage frequency; o Earthquakes, at about 4% of the core damage

) frequency, are a less significant contributor; ,

and o All other MCCIs contribute a negligible amount.

)

The 1985 MCCI analyses of seismic and fire induced accident sequences indic seismic and three fire sequencesgped that two account for most of the MCCI core damage frequency contribution. The dominant seismic sequences were seismically-induced failures of the Service ,

Water System (T s SWP) and seismically-induced LOCA or RFV failures which exceed the capability of the ECCS (Tg As )*

3/ The dominant fire sequences from the 1985 MCCI report consist of relay and control room fire initiators that damage safety related equipment. LILCO did not re-examine these fire sequences.

I 39 -

l L

) '

l In the current PRA update effort, a re-evaluation of the known conservative margins associated with the seismically-induced sequences was performed (IVB-5) to develop a more realistic mensare of their risk importance.

)

The following is a summary of the key results from the updated MCCI examination of important seismic issues. The focus of the update to the '

1985 MCCI study was to:

) o Re-examine the seismic margins associated with RPV support structure seismic failure modes; o Re-examine seismic station blackout issues; and

)

o Evaluate the impact of these seismic fragility re-analysis on the SNPS severe core damage frequency.

In addition, a seismic fragility analysis of the y SCS (and the associated piping and components) was not perrormed in this analysis since final construction drawings are not yet available. A general assumption was made (subject to a confirmatory analysis) that the seismic capacity of the SCS is approximately equivalent to that of 3 the Shoreham Reactor Building.

The conclusions from the updated MCCI evaluation are as follows:

o As a consequence of the use of plant-specific data and the more realistic response analysis using a coupled analytical model, the results of the re-analysis show that the RPV and supports are much more seismically robust than estimated earlier. In fact, re-analysis of the RPV support seismic fragility, using plant-specific data, indicatet that the predicted median ground acceleration capacity of the RPV has increased to 1.40g from the 1985 study result of 1.04g.

y o Re-analysis of the Service Water 480V Motor control Centers (MCC) seismic fragility has determined that the predicted median ground acceleration capacity for an assumed functional failure mode of relay or auxiliary contactor chatter has increased to 1.219 from the 1985 study result of 0.83g.

o It was judged that the Colt Industries diesel generator installations have seismic capacities at least as great as the insta'. led

1 TDI units. Therefore, the Colt units can be considered as an equally reliable emergency power source in the seismic risk model.

) As a result of the reanalysis of the seismic fragility of the RPV supports and those components which contribute to station blackout concerns, a reevaluation of the dominant seismic

~

sequences was performed. The results of the reevaluation. include:

)'

o The annual mean frequency of seismically-initiatedseverecore_gamageorcoremelthas decreased to 1.1 x 10 / year. This represents about a factor of 2.3 reduction from the earlier 1985 MCCI estimates.

) o Seismic events, at 3% of the total frequency of core damage, continue to be a minor contributor to core damage frequency, o The T A seismic core damage sequence

)

frequ5n8y has decreased a factor of 2.5 to 3.2.E-7 per year using the revised RPV fragility estimates, o The T SWP seismic core damage sequence frequ3ncy has decreased by a factor of 5 to

2.2E-7 per year using the revised service water MCC fragility assessment.

6. Core Damage Frequency Results

a. Probabilistic Assessment with SCS The PRA update results show that, after requantifying the accident sequences as they are affected by the hardware and procedural modifications that have been, or will be, implemented at Shoreham, the core melt accident frequencies are reduced.

Table IV.B-2 summarizes the physical plant and procedural changes that resulted in tne l

reduction of the total core melt frequency as they affect each accident subclass.

i Table IV.B-3 presents a comparison of the

criginal PRA accident sequence frequencies to l

the results of this evaluation for internal l events only.  ;

1 Table IV.B-4 is a quantitative summary of the contributions to core melt frequency by accident sequence subclass including a characterization of the source of the dominant r .-,_ ._. - - .,- - . . - , . . . _ . . - - - - - _ , - - - - _ . _ _y. ..

,-m - - - - ._-------y

contributors to each of the subclasses. This table summarizes both the internal and external event evaluations.

The following discussion identifies the reasons for the quantitative di#ferences relative to the 1983 PRA by accident sequence class. This summary focuses on the internal event evaluation because the hardware and procedural changes primarily address these accident sequence types.

Class IA: Loss of inventory makeup with the RPV at high pressure.

The initiating event frequencies have been reevaluated using the latest available data.

These changes in frequency do not produce a substantial change in the calculated Class IA frequency.

The addition of the ADS inhibit switch and the accompanying procedural instruction result in an increase in the potential for ADS inhibit; and therefore, the poteatial nesd for manual depressurization. The conditional probability of failure of successful manual depressurization is found to increase. This increase is a strong function of the procedural requirements and the level of training of the operating staff. The addition of the ADS inhibit switch and associated procedures results in a slight increase to the frequency of Class IA sequences leading to core melt initiation. It should be noted, however, that the small increase in the core melt frequency cited here, is more than compensated for by the reduction in the Class IV core melt frequency attributable to the ADS inhibit switch.

The internal initiated flood sequences from Class ID in the original PRA nave been transferred to Class IA because they are judged to be more appropriately treated for consequence evaluation in Class IA. The >

updated Class lA core melt frequency estimate for internal events is 2.6E-5.

l

)

Class IB: Loss of offsite power initiated events.

Loss of Offsite Power (LOOP) initiated events were identified in the 1983 PRA assessment as a substantial contributor to the core melt frequency at Shoreham (approximately 18%) .

Two major modifications alter that estimate by reducing the frequency of LOOP initiated sequences that contribute to core melt. The modifications are:

o The installation of a blackstart 20 MWe gas turbine generator as an additional AC power source; -

o The expansion of the Station complement of onsite emergency diesel generators with the installation of three Colt diesels.

The reassessment of the LOOP initiated event tree results in a calculated Class IB core melt frequency of 4.lE-7 per reactor year due to the internal event evaluation. This represents an order of magnitude reduction in the Class 1B internal events contributor. e These changes are not considered in the seismic induced station blackout sequences, based on a review of the dominant seismi c contributors.

Class IC: ATWS related sequences for which the containment is intact, but at elevated pressure when coolant inventory makeup becomes unavailable.

The procedural modification associated with the use of low pressure makeup systems, the addition of highly enriched boron to the SLC system, and the SCS modification have resulted in a large reduction in the Class IC accident sequence frequency to 1.lE-8 per reactor year.

Class ID: Loss of inventory makeup with RPV at low pressure.

No plant modifications have been identified that significantly alter the frequency of Class ID sequences as calculated in the 1983 PRA. However, the assessed frequency of Class ID is reduced because flood sequences similar to TQUX were includad in the original PRA Class ID. These have been shifted to Class IA in thia update. Therefore, the calculated Class ID accident sequence frequency is revised to 2.0E-6 per reactor year.

__..j

Class II: Loss of Containment Decay Heat Removal The addition of the SCS modification has a significant effect on the frequency of Class II accident sequences compared with the original Shoreham PRA. Because the SCS is a passive system that can effectively remove decay heat from containment, Class II accident sequences have an additional likelihood of being mitigated successfully with no core damage induced. Specifically, the major considerations in the reassessment of the Class II accident sequence frequencies include the following:

o SCS venting is a new method of containment heat removal not considered in the 1983 PRA; o SCS venting does not affect CRD pump '

performance, where the CRD pump is a prime means of long-term RPV inventory control, o Loss of NPSH remains as a potential contributor to some sequences.

The togp1 frequency of Class II sequences is 3.4E per year; however, the SCS heat removal capability results in the vast majority of these sequences being shifted to Class IIF sequences in which the SCS functions to minimize radionuclide releases.

Class III: LOCA Initiated Events In general, the modifications investigated by LILCO and included in the Shoreham design have addressed those accident sequences perceived as risk significant. LOCA sequences have not been identified as risk contributors, and specific modificatione to reduce sequences resulting from LOCAs have not been identified for implementation at Shoreham. Therefore for this evaluation, the core melt frequency remains essentially the same as that calculated in the PRA. Specifically:

4/ Internal Events 1.6E-6/Rx Yr External Events 1.8E-6/Rx Yr Total 3.4E-6/Rx Yr

)

i l

i Sequence Internal Event Subclass. . Frequency (per reactor year)

[.

IIIA 2.9E-7 II1B 7.4E-7 IIIC 3.6E-7

)~. IIID 8.1E-8 Class IV: ATWS Initiated Events ATWS sequences'have previously been identified as among the potentially most risk significant residual sequences, both generically and specifically in the case of Shoreham, as documented in the original PRA.

As an insight from the Shoreham PRA and in

) ongoing probabilistic analyses, LILCO has implemented hardware and procedural modificatigns to substantially reduce the frequency and consequences of the ATWS related sequences. Tne major considerations in the reassessment of the Class IV accident sequence

) frequencies include the following:

o The increase in the SLC system sodium pentaborate enrichment resulted in a reduction of the core melt frequency due to ATWS related sequences; i

!~

i l

l i

i i

\

y o Incorporation of the SCS modification reduced the frequency of ATWS related sequences and probabilistically shifted sequences from potential high releases to

) filtered releases; o The ADS inhibit switch modification increased the probability of avoiding a rapid, uncontrolled depressurization during an ATWS;

) o Emergency Operating Procedures are being modified to incorporate the guidance of Revision 4 of the EPGs.

Class V: LOCA outside containment.

) -

An' evaluation of the pressure retaining capability of the low pressure inventory makeup systems connected to the primary system was performed (IVB-3). This analysis.has identified that there is a very low j probability of system failure as a result of inadvertent exposure of the low pressure piping or components to full RPV pressure and temperature. This results in a reduction in the assessed frequency of the class V events.

) Figure IV.B-1 provides a graphic summary of the accident sequences that contribute to the calculated core melt frequency for the current Shoreham configuration. The accident sequences labeled TOUV and TOUX are composed of a large number of individual sequences which have different outset contributions.

In conclusion, the improvements in the Shoreham plant since 1983, have resulted in an internal core melt frequency reduction of approximately 33%. Additionally, those sequences with the highest potential for severe radionuclide release have been reduced by an even larger percentage (i.e., ATWS by a factor of 10, Class IIA by a factor 75, Station Blackout by a factor of 17). These latter reductions in the sequences of highest potential consequence are significant because of the potential increase in public safety. -

In addition to the above conclusions, this analysis did not identify any new types of internally initiated accident sequences which could lead to core melt beyond those originally identified in the 1983 PRA.

The operation of Shoreham in the configuration documented in the current PRA update results

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

)

in a reduction in overall point estimate frequency of core melt conditions and a potentially 15 ge reduction in public risk as a result of the reduction in the frequency of the most' risk significant sequences,

b. Probabilistic Assessment Without SCS Installed and Comparison with SCS This section provides a synopsis of the sensitivity analysis, representing the use of the existing 6 inch wetwell and drywell containment vents, and a comparison with the baseline SCS evaluation.

The implementation of containment venting has a significant effect on the frequency of Class II accident sequences original Shoreham PRA57.mpared with theventing Containment can affectively remove decay heat from containment. Consequently, Class II accident sequences have an additional likelihood of being mitigated successfully with no core damage induced. Specifically, the major considerations in the reassessment of the Class II accident sequence frequencies include the following:

o Venting is a recently implemented method of containment heat removal, not considered in the 1983 PRA; o Venting does not affect CRD pump performance, where the CRD pump is a prime means of RPV inventory control; 5/ The core damage mitigation capability attributable to the existing 6 inch wetwell and drywell containment vents is neglected for Class IV sequences.

1 o Venting affords a scrubbing mechanism for mitigating the releases. (This is treated in the contair. ment event tree evaluation);

j- o Loss of NPSH remains as a potential contributor to some sequences.

There are five principal types of event tree transfer states in terms of their likelihood of successful mitigation. These sequence dependent trar.sfer states are evaluated

) separately in the VENT. Class II event tree to determine the plant damage states that could result.

Table IV.B-5 summarizes the accident sequence

)- frequency results of the existing Shoreham configuration. The total core melt frequency of 4.2E-5 events per reactor year is low. The core melt frequency estimate of 3.6E-5 (Table IV.B-4) for the SCS represents an additional 14% decrease in the already low core damage

) estimate associated with the non-SCS configuration.

Two major conclusions that can be drawn from the comparison of the SCS and VENT results of Table IV.B-6 are the following:

o Installation of the SCS system resul's t in a majority of the Class IIA and IIB wetwell vent sequences being shifted to Class IIF for consideration because system actuation significantly prolongs the timing of TW accidents. Additionally, the Class II core melt frequency with the SCS installed is reduced by a factor of approximately 2.5.

o Actuation of the SCS not only provides additional capability for ATWS events (illustrated by the reduction in Class IV core melt frequency by a factor of approximately 1.5), but also prolongs the timing of a majority of sequences which progress towards core degradation.

The ramifications of reduced sequence frequencies and extended sequence timing l are realized in the containment and source term evaluations for characterization of l the overall risk profile for Shoreham.

~ _. . ,

)

7. Characterization of Uncertainties l

l The former NRC Safety goal (IV-6), AIF proposed goal (IV-7), and the BWR IPE screening value (IV-8) were established using mean estimates of core damage frequency. In addition, the former NRC safety goal proposed that "the mean probability of large-scale fuel melt should be less than 1 in 10,000 per reactor-year as a goal and less than 5 in 10,000 as an upper limit. The ACRS defines large-scale fuel melt to occur when over 30 percent of the oxide fuel becomes molten.

It has long been recognized that these point estimates of core damage frequency are subject to uncertainty", (IV-6). Therefore, the use of mean values within the Shoreham PRA for comparison with proposed goals is consistent with the evolution of the PRA processes and the recognition of the use of uncertainty therein. It should be noted that current NRC policy statement no longer references a core melt frequency goal.

This section presents a characterization and discussion of the uncertainties inherent in the analytical approach to a PRA of a nuclear power plant. The uncertainties arise from a number of sources including uncertainties in the input data, accident sequence definition, plant systems, and containment response models.

The 1983 Shoreham PRA has previously summarized the status of uncertainties and the impact that they may have on the conclusions. That evaluation indicated that the Shoreham plant does not have any unique features that would result in uncertainty estimates significantly different from other published PRAs. The current PRA update has reached a similar conclusion with one exception, the SCS modification.

Some of the sources of uncertainty associated with previous BWR PRA calculations involved operator actions to recover containment heat removal systems, initiate SLC in short time frames, or vent the containment. The addition of the SCS at Shoreham tends to reduce the upper bound uncertainty associated with such actions because it limits the requirements for active operator response or greatly increases the time window over which the operator can effectively mitigate the accident. Therefore, the proposed SCS modification is a unique feature, and the quantification of this ui cem within the analysis could result in slight differences in uncertainty estimates among accident subclasses.

)

The quantitative evaluation of accident sequence frequencies indicates that the mean core melt frequency is well below the formerly proposed safety goal. In addition, using the uncertainty bounds quantified in both the 1983 Shoreham Nuclear Power Station PRA and in the IDCOR IPE, the upper bound uncertainty is also within the estimates of these safety goals put forth by the NRC and ACRS. With the SCS modification, the upper bound uncertainty should be reduced even further although the precise reduction has not been calculated.

._____-_J

)

REFERENCES 1

l IVB-1 "Shoreham Nuclear Power Station, Full Power PRA, PRA Update: Supplemental Containment System Implementa-tion", Long Island Lighting Company, February, 1988.

IVB-2 "Probabilistic Risk Assessment of the Shoreham Nuclear Power Station", Long Island Lighting Company, June 1983.

IVB-3 "Evaluation of Typical RHR and CS Injection Class 2 Piping to Withstand a Ma::imum Postulated Intersystem LOCA; Shoreham Nuclear Power Station Unit 1," NUTECH Engineers, Inc., December 1986.

IVB-4 "Major Common Cause Initiating Event Study", Volume 1 and 2 (NUS 4617), Long Island Lighting Company, February, 1985.

IVB-5 "Re-Analysis of Seismically Induced RPV Failures and Seismically Induced LOSP Sequences in the Shoreham Nuclear Power Station PRA Update", (NUS-5080), Long Island Lighting Company, February, 1988.

IVB-6 "Proposed Policy Statement on Safety Goals for Nuclear Power Plants", NRC, November 12, 1981.

IVB-7 "A Proposed Approach to the Establishment and Use of Quantitative Safety Goals in the Nuclear Regulatory Process", AIF, May, 1981.

IVB-8 E. T. Burns et al., "The Impact of Uncertainty on Severe Accident Policy Statement Decision Making", Delian Corporation for IDCOR, November 1986.

)

Table IV.B-1 l

SUMMARY

OF THE CORE DAMAGE ACCIDENT SEQUENCE SUBCLASSES (PLANT DAMAGE BINS)

FOR THE UPDATED SHOREHAM PRA ACCIDENT CLASS DESIGNATOR SUBCLASS DEFINITION EXAMPLE CLASS I A Accident Sequences Involving TQUX Loss of Inventory Makeup in which the Reactor Pressure Remains High B Accident Sequences Involving TO E

a Loss of Off-Site Power and Loss of Coolant Inventory Makeup C Accident Sequences Involving TTg C QU Loss of Coolant Inventory Induced by an ATWS Secnence D Accident Sequences Involving TQUV a Loss of Coolant Inventory Makeup in which Reactor Pressure has been Successfully Reduced to 200 psi.; Accident Sequences Initiated by Common Mode Failures Disabling Multiple Systems (ECCS) Leading to Loss of Coolant Inventory Makeup CLASS II A Accident Sequences Involving TW A Loss of Containment Heat Removal with the RPV Initially Intact; Core Damage Induced Post-Containment Breach B Accident Sequences Involving AW Loss of Containment Heat Removal with the RPV Breached But No Initial Core Damage F Class IIA or IIB Except that TW SCS (or the 6" Containment Vents) Operates as Designed; Loss of Makeup Occurs at Some Time Following Venting Initiation. Suppression Pool Saturated but Intact CLASS III A Accident Sequences Leading R to Core Vulnerable Conditions Initiated by Vessel Rupture Where the Containment Integrity is not Breached In the Initial Time Phase Phase of the Accident

_. - - . _ _ J

)

Table IV.B-1 (Continued)

SUMMARY

OF THE CORE DAMAGE ACCIDENT SEQUENCE SUBCLASSES (PLANT DAMAGE BINS)

FOR THE UPDATED SHOREHAM PRA ACCIDENT CLASS DESIGNATOR SUBCLASS DEFINITION EXAMPLE CLASS III B Accident Sequences Initiated S yQUX (Continued) or Resulting in Small or Medium LOCAs for Which the Reactor Cannot be Depressurized C Accident Sequences Initiated AQUV or Resulting in Medium or Large LOCAs for which the Reactor is at Low Pressure

~~ D Accident Sequences which AD are Initiated by a LOCA or RPV Failure and for which the Vapor Suppression System is Inadequate, Challenging the Containment Integrity CLASS IV A Accident Sequences Involving TCC Tg2 Failure to Insert Negative Reactivity Leading to a Containment Vulnerable Condition Due to High Containment Pressure F Class IVA except SCS TCC, Mg~

is Actuated and Operates as Designed; A Loss of Makeup to the RPV Occurs Following SCS Initiation; Suppression Pool Saturated, but Intact G Class IVF Except the Coolant -

Injection Continues for an Extended Time and the Filter Bed Becomes Saturated.

CLASS V -

Unisolated LOCA Outside -

Containment 53 -

)

l Table IV.B-2 l

l

SUMMARY

OF CHANGES INCORPORATED INTO THE PRA UPDATE CLASS MAJOR CHANGES INCLUDED IN PRA UPDATE IA+ o ADS INHIBIT SWITCH ADDED B+ o BLACK START ONSITE GAS TURBINE INSTALLED o SWITCH RCIC/HPCI TO CST PROCEDURE C+ o INCREASED BORON ENRICHMENT OF SLCS o EPG REV. 4 (CONTINGENCY 5) BASED REACTIVITY CONIROL PROCEDURE o SUPPLEMENTAL CONTAINMENT SYSTEM II+ o SUPPLEMENTAL CONTAINMENT SYSTEM IIIA B

C o NO SUBSTANTIAL CHANGES D

IV+ o INCREASED BORON ENRICHMENT OF SLCS o EPG REV. 4 AF (CONTINGENCY 5) BASED REACTIVITY CONTROL PROCEDURES o SWITCH RCIC/HPCI TO CST PROCEDURE o LPCI/ CONDENSATE THROTTLED INJECTION INCLUDED IN PROCEDURES o SUPPLEMENTAL CONTAINMENT SYSTEM V o NUTECH AND BWROG EVALUATIONS OF THE CONDITIONAL FAILURE PROBABILITY OF LOW PRESSURE SYSTEMS

+ The initiating frequency has been updated using the latest available data.

'm K 2 IV.5-3

) GNWYSE OF 15E GEE MLT ACCIIBff amg2 mmart 1981 IsA vtaStB TILtRA HAff GNIGRtIIOl (DfDMEL E19GS) c.RE m.T Itap:ct i AKIIDT (PerYear) l CASS SL2 GASS IETISTIICH IXAMRZ JLNE 83 FIl3 A IMCE CESIO MCR Q ASS I A Acident Sequerees hlving less of I inv etory Makaup in dtich the P4 actor M2 1.4E-5 2.6E-5 Pressare Panains Mid B Me1A,nr Sequences Irwolving a loss of Off-Site Powr ard Ime of Coolmt Tg(LV 6.9E4 4.1E-7

. Iriventorv *kne C Prihr Sequerras Involving a less of em1mt Inventory Irdred by an ANS Ty 4.lE4 1.11 4 S*Nence i

D peia nr Wences Irwolving a loss of Coolm:t Irnntory Makap in diich Reactor Pressure has been Successfully Raired to 200 psi.; #e4 Anne Sequences Initiated by WN 7.lE4 2.24 cmm hde Failure Disabling Nltiple System (IDCS) laaitre to loss of Coolant inventory Kskap C. ASS 11 A k eident Sequences h lving a loss of Ccntairnent Heat Ramval with the RPV W 2E4(IST) 2.7E4 Initially Intact; Core Docage Irduced Post-Ccntalment Brexh B Accident Sequmces h1ving A ices of Cr.taiment Heat Pa:cval with the RPV Ah' NA 5.7E-9 Breachef Ne No Initial Core themge F Class IIA or IIB except that S G Operates as Designed; Ims of Makap Occurs W NA 1.6E4 at See h Folkwing SG laitiatim.

Surpressim Ptol Saturated tut intact C. ASS Ill A Accident Cag a laading to Core Winerable Ccrditims Initiated by Vessel R 3.0E-7 2.9E-7 Rapture dare the Cmtaircent Integrity is rot Breached in the Initial Tica lhase of the Accident B Asident Sequences initiatal or Pandting in Small or Malita 1&As for Eich the S@g 2.W 7. 4 7 Raactor Carrot be Depressurized C Meident Sequences Initiated or Rasulting in Maiita or large IKAs for *1ch tre AOLV 3.6E-7 3.6E-7 Ramtor is at Irw Pressure D Meidant Sequences Wtich are Initiated by a IDCA or RPV Failure ard for dtich the AD 9.1E4 8.lE4 Vapor Sogpressica Systan is TWea.

Omilenging the Ccntairment Intestrity CASS IV A p eta,nr Sequences h 1ving Failure to Insert Negative Reactivity InUng to a Tg 1.4E-5 4.4E-7 Chtaircent Winerable Ccrditicn Due to High Ccntairrant Presman F Class IVA except sci is k tuated asi Operatas as Desiped; A loss of Makaup g NA 3.tZ-7 to the RPY (k: curs at Scre Time Folicwirs G Initiaticn; Syyressica Pool Saturated, but intact G Class I\T except the Sogpressica Pool k'ater Isval is Behv the III:S metian NA 6.4E-7 elevaticn.

C. ASS V - thisolated LOCA Gatside Conrainmt 3.t>E-6 1.4E-d MAL + ,++ 4.9E-5 3.1-5

+ Care emit frequery; rot core or ccntairrant vulnerable frequency.

++ Iniivital subclass frequencias, as wall as the total core nelt frequency (c'E) quoted in the 1983 Shorehan PRA esy vary slightly frta this presentaticri due to the slightly different interpretaticn of core dange. Specifically, the 1983 Storehan ITA c'J inherently a==r** mbstetial fuel simpirs to the tottaa head of the RPV. Um current ITA spdate uses a slightly tore restrictive intet7retatim Aich incitdes accident sequences dare the core toccoes 2/3 tncovered ard core melting has initiated.

W U Table IV.B-4 SUFt1AltY OF ICCIDF2?P SIXXH7KE FIREQUEICIES (WI'111 SCS)

SEISt11C IU11tlAFTP Iffl13d3AI, FIIE IIIDUCED I!JIXJCED CIASS/SUICIASS SIXUF2KI COltE MELT COPE !!Ellr CCl!E ME1T CIASS DESIGilNIDit CATIXXXtY FitECU12X."I+ ETIEQUEIU+ FitEQUE2K'Y+ 'IUPAL Class I A Internal 2.6E-5 1.1E-7 2.1E-9 2.6E-5 B Seismic 4.1E-7 1.6E-8 7.5E-7 1.2E-6 C Internal 1.1E-8 0.0 0.0 1.lE-C D Internal 2.0E-6 1.4E-7 2.3E-8 2.2E-6

, Class II A Fire 2.7E-8 4.2E-8 -

6.9E-8 B Internal 5.7E-9 - -

5.7E-9 P Fire 1.6E-6 1.8E-6 -

3.4E-6 l

I Class III A Internal 2.9E-7 0.0 0.0 2.9E-7 B Internal, 7.4E-7 0.0 0.0 7.4E-7 C Internal 3.6E-7 0.0 0.0 3.6E-7 D Internal 8.1E-8 0.0 0.0 8.1E-8 Class IV A Internal 4.4E-7 0.0 0.0 4.4E-7 F Internal 3.6E-7 0.0 0.0 3.6E-7 G Internal 6.4E-7 0.0 0.0 6.4E-7 Claus V Internal 1.4E-8 0.0 0.0 1.4E-8 StilV Seismic -

0.0 3.2E-7 3.2E-7

'IUTA1,* 3.3E-5 2.1E-6 1.1E-6 3.6E-5

+ All Frequencies per reactor year.

o Totals may rot initcli due to rourx1-off.

l l

W U Table IV.B-5 SUITMRY OF ICCIDDTP SECUE!JCE FREQUE21CIES (WITil 6" VDTIS)

SEISlilC IXY11!!AlTr IiTrEI21AL FIRE 11JDUCID IIHXJCED SIEUEtJCE CORE f EI.T CORE MEI.T CORE t-ELT CIASS CI.NIS/fXfPCI ASS CATEOCRY ITS.)UE2KN+ FRFTJJE21CY+ FREQUEtJCY+ '1 URAL DESIGTJNIUR Class I A Internal 2.6E-5 1.lE-7 2.lE-9 2.6E-5 l 4.1E-7 1.6E-8 7.5E-7 1.2E-6 I Il Seismic C Internal 4.9E-9 0.0 0.0 4.9E-9

! D Internal 2.0E-6 1.4E-7 2.3E-8 2.2E-6 Class II A Fire 2.3E-6 5.8E-6 -

8.1E-6 l

11 Internal 8.0E-8 - -

8.0E-8 F Internal 4.4E-7 0.0 -

4.4E-7 Class III A Internal 2.9E-7 0.0 0.0 2.9E-7 11 Internal 7.4E-7 0.0 0.0 7.4E-7 C Internal 3.6E-7 0.0 0.0 3. 6E--7 D Internal 8.lE-8 0.0 0.0 8.1E-8 Class IV Internal 2.1E-6 0.0 0.0 2.1E-6 Class V Internal 1.4E-8 0.0 0.0 1.4E-8 SRPV Seimic -

0.0 3.2E-7 3.2E-7 3.5E-5 6.lE-6 1.lE-6 4.2E-5

'IvrAI/

+ Al1 Fnquencies per reactor year.

• 'Ibtals may not mtch due to rountl-off.

_ 57 _

J l Table IV.B-6 l

ColiPARISON OF VENTING SENSITIVITY ANALYSES Plant System Configuration Accident SCS 6" CONTAINMENT VENTS Subclass Subclass  % of  % of Frequency (Per Total Class Frequency (Per Total Class Reactor Year) Frequency Reactor Year) Frequency TW SEQUENCES II A 6.9E-8 2% 8.1E-6 94%

II B 5.7E-9 <1 % 8.0E-8 ,

1%

II F 3.4E-6 98% 4.4E-7 5%

TOTAL 3.4E-6 8.6E-6 ATWS SEQUENCES IV A 4.4E-7 31% 2.1E-6(IV) 100%

IV F 3.6E-7 25%

IV G 6.4E-7 44%

TOTAL 1.4E-6 2.1E-6

)l u

)

% )

) 0 r s

% a r 0 0 et o 6 Y u

) 5 ( rb

% ( or i X t t 2 A U cn 9 C O ao

( O T eC L / R e R V r c E U en pe R W O ( m A O T yq u

E P ce Y nS C e R O ut qn O e e MTC Fi rd AA c HER t c l

eA E/ Md R5 - \ e e rf i

OE oi Ct n H6 ne S3 ) h r

oI d

= ee

% r h ct N 6 h A 9 So t E ( e e M / h u TD W )

3 1 1

( ) -

R 7 B.

E 6 I

I T ( )

A  % ) ) e W L E

3  %  % r u

E 3 0 5 g V ( i C E (4 1

( F I

V L 0 R R 8 S D E E S /

O S T A

W O A L W F x G R D L

B x

R 1 i

D C. Containment Event Tree Analysis

1. Introduction The Containment Event Trees (or CETs) provide a consistent methodology for determining the spectrum of possible radionuclide releases to the environment and their associated frequencies for the dominant core damage states (described in the previous section). Each CET end state represents a particular release event or a recovered degraded-core state which may be characterized according to its 1) potential for fission product release to the atmosphere, 2) timing, and 3) release duration; all of which are important to the off-site consequences. The CET release modes are then collapsed into a set of Release Categories (RC) which are sufficient to represent the full spectrum of source terms.

This section describes the salient aspects of the CET analysis which included the following topics:

assessment of containment performance (failure mode, location, size, and timing);

characterization of release pathways to the environment; risk ranking of accident sequences; quantification of release category frequencv; and the selection of representative sequences.

Comparisons of the release spectrum "with-SCS" and "without-SCS" are provided to illustrate the benefits associated with the SCS.

2. Containment Performanca The primary containment provides an important barrier to the release of fission products to the environment. This subsection, therefore, focuses on primary containment performance under severely degraded conditions and potential failure modes and locations which affect releases should the containment be challenged beyond its ultimate structural capacity.

Primary containment performance is evaluated by considering the following factors in the CETs:

o Pressure loads 3ading to overpressure failures including failure location and size; o Thermal loads contributing to over-temperature failure in the drywell due to material strength reduction and seal degradation;

- - - - - - - - _ - _ - J

I J

o Combined pressure and thermal loads leading to a change in the failure location; o The timing of containment failure relative to core melt; o Isolation failures leading to an impaired containment; Stone and Webster Engineering Corp. performed a detailed structural evaluation of containment structural capacity for combined thermal and pressure loads. A temperature dependent leakage relationship was determined and is utilized in the CET and MAAP analyses.

a. Containment Failure Locations and Failure Size For postulated containment failures, the location and size of the breach are important in determining the radionuclide releases because these impact the potential for suppression pool scrubbing and reactor building retention.

Containment failures due to pressure loads (given temperatures in the drywell are not elevated) are modeled following the same guidelines as determ_aed in the 1983 PRA. One of the attributes of this is that containment break size and break location are probabilist-ically coupled. The break size and location are also dependent on the pressure at which the primary containment fails, and the pressurization-rate. A summary of the assumptions used for overpressure events and models of failure appears in Table IV.C-1.

The principal basis for the assumptions delineated in this table is "leak-before-break" (consister.t with IDCOR's conclusion in Reference IVC-2). Catastrophic failure of the containment is considered unlikely, partic-  !

ularly for those scenarios where containment pressurization is gradual. Recent high-pressure testino at a one-sixth-scale model of a reinforced concrete containment at Sandia Laboratory confirms this assumption.

For late containment failure involving combined temperature and pressure loads in the drywell (for temperatures between 500 F and 800 F), the relationship shown in Table IV.C-1 is modified as shown in Table IV.C-2. For l

J temperatures in this range, and pressures l

' approaching the ultimate capacity, the drywell and wetwell are assumed to fail with equal likelihood. This is judged to be reasonable

) since the temperature rise (given that the containment loading from steam oressurization which occurs as a result of suppression pool boiling by the core debris) is not expected to appgoach the drywell critical temperature of Leakages occur in the drywell 800 F.

(principally due to thermal loads from debris radiative heating of the containment structures) with a break size qquivalent to a small leakage area of 0.001 FT' at tge SCS back pressure of 60 psig. Aboye 800 F this increases to 0.05 FT~ and 1 FT with equal likelihood.

b. Containment Failure Timing The time of containment failure is important since it determines the degree of aerosol depletion and the decay of radioisotopes before they are released to the environment.

Consequently, containment failures are characterized in terms of their timing (early or late) relative to core melting, o Early Containment Failure: These include system related events, such as containment isolation failure or loss of vapor suppression function as well as phenomenological events (hydrogen burns, etc). These events can provide the pressure loads that could challenge containment integrity early with respect to core melting. The likelihood of early containment failure is dependent on successful SCS actuation.

o Late Containment Failure: Generally, late containment failures result from long-term thermal and/or pressure loads in containment. Thermal loads are predicted by the MAAP code due to core debris material dispersed into the outer drywell upon reactor vessel failure. This tends to negate somewhat the benefits derived from the SCS venting system since primary containment failure can occur even with successful SCS actuation.

i

3. Release Pathways to the Environment The radioactive release pathways to the environment are considered in the CETs because they influence the magnitude of fission product releases. The release pathway attributes include:

o SCS Gravel Bed Filtration and IIoldup Time The SCS is an efficient filter for particulate releases (DF greater than 500) and provides a significant holdup of noble gases, o Suppression Pool Scrubbing The 1983 Shoreham PRA and subsequent EPRI test results modeled in MAAP demonstrate that the suppression pool, if not bypassed, is an efficient scrubbing mechanism for particulate releases.

o Reactor Building Decontamination MAAP modeling predicts attenuation of the release through natural removal processes such as settling and plate-out depending on the break location and driving force through the building.

4. Risk Ranking Methodology for Release Categories Each plant damage state sequence progresses through a Containment Event Tree (CET) to a set of release events that are described by unique release modes. The release mode characteristics are qualitatively described in Table IV.C-3. For the purposes of radiological consequence presentation, the release events are combined into a limited number of Release Categories (RC) through a risk ranking process described below.

The Release Categories are given in Table IV.C-6 with SCS and Table IV.C-7 without SCS.

The quantified principal attributes of the release categories which are used to represent the spectrum of release events are: 1) release time -

early (0-6 hours), moderately early (6-24 hours),

or late (greater than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />), 2) release duration - short initial puff (less than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />),

long duration (no significant initial puff with constant release greater than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />), or extended and controlled release (very slow i

)

release); and 3) release magnitude - noble gas and particulate release fractions.

For the purposes of risk ranking the release events are screened based on the potential consequences and risk to the public. Two consequence measures were used in this study:

a) Estimates of the potential whole body and thyroid doses to an individual located at some distance from the plant were obtained as a mean value over metecrological conditions.

b) Curves representing the conditional probability of exceeding a specific dose level (for whole body and thyroid dose) were obtained as a function of distance from the plant.

The mean doses were expressed in terms of a unit release of the principal fission proluct groups.

By this means, a limited set of similar consequence impacts, i.e., release categories, were identified for a given combination of noble gas and particulate release fractions and release timing representing a release event.

The dominant sequence contributors to each release category are found by ranking each of the release event contributors in the release category in order of its risk. Risk is defined as the product of the consequence measure and the frequency associated with the release event.

5. Release Category Frequency This section presents the results of the CET evaluation of the frequency associated with a spectrum of accident cequences. The accident sequences have been grouped into twelve release categorias. The frequency of each release category is determined from the product of the plant damage state frequency and the conditional probability of release, given the plant damage state.

The plant damage state frequencies calculated by IT/Delian for the plant configuration with and without the proposed SCS system are summarized in Table IV.C-4.

Table IV.C-5 summarizes the frequency distribution of the release categories which represent the

)

final grouping of accident sequences considered to be important to risk.

< 6 Selection of Representative Sequences for Release

)' Categories Ideally, each combination of core melt accident sequence and CET release mode could be analyzed to provide a complete and thorough consequence assessment of the potential source terms resulting 1 from it. On the other hand, it is recognized that by combining sequences that have similar progressions and fission product transport behavior into groupings called release categories (RC), the source term spectrum can be adequately analyzed. The analysis of the spectrum of

)L accident sequences conducted in this study identified dominant contributors to each RC so that a source term scenario can be selected which can be used to represent the RC.

The source term characteristics are summarized in

) Table IV.C-6 and IV.C-7 for the SCS and non-SCS cases, respectively. A brief description of the RC attributes are described in the following sections:

a. Early Release Events: RCl-RC4

)

Inspection of the early release events and the various contributors to the binned release end states indicates that the potentially high consequence release category RCl is determined principally by the Class IVA (ATWS) sequences 4 leading to an early drywell containment failure bypassing the suppression pool. The leakage rates are expected to be high and the release point to the reactor building would not support significant decontamination on reactor building surfaces.

Release Category RC2 includes wetwell failures for the same plant damage state (Class IVA),

with impairment of the downcomer vent path from the drywell to the suppression pool resulting in pool bypass. However, modest reactor building decontamination is predicted resulting in reduced potential consequences.

I RC3, in which the off-site consequences are substantially reduced due to suppression pool l

scrubbing, is dominated by wetwell failures of j the Class IVA sequences without pool bypass.

i The releases in this case are principally noble gases.

RC4, which could have a moderate-to-high consequence impact, is represented by Class IIID (LOCA) sequences. In this scenario, the containment is initially impaired due to the failure of the vapor suppression function, induced by a large seismic initiating event (SRPV). The radionuclide releases to the environment could occur early, although the magnitudes are significantly less than for RCl due to the absence of substantial fission product driving force from the impaired containment. In addition, the release durations are expected to be longer than for RCl. The volatile fission products may still affect the consequences because the suppression pool is bypassed. The potential consequences are expected to be less than in the case of RC2 due to the lack of sufficient fi.csion product driving force from the impaired containment and low drywell temperatures (which support slow rates of release, resulting in longer duration and reduced release magnitudes). The time of release for these four release categories (RCl through RC4) are relatively early, on the order of at most six hours. The same release categories were found to dominate early release events for the configuration without the proposed SCS.

b. Moderately Early Release Events: RCS-RC7 One of the key findings in this study (which was not identified in the original PRA) is that potentially high temperatures may be obtained in the drywell following vessel breach. This could result in a moderately early or delayed particulate release due to drywell thermal failure.

Inherent in the postulated drywell thermal failure is the implication that high containment and vessel surface temperatures lead to significant revolatilization of initially deposited fission products which results in higher particulate release fractions (although late in the sequence).

These scenarios are represented by Class IBl (Station Blackout) and characterized as release category RC5.

7 a

RC6 has a similar sequence progression, although reactor building retention (or lower drywell temperature) mitigates the potentially high release of the revolatilized fission product species. The release timing is expected to be similar to that for RC5.

RC7 represents vented sequences where the release of noble gases is not controlled, either due to SCS depressurization or saturated bed conditions which preclude noncondensible gas hold-up. For the plant configuration without SCS, RC7 is characterized by core melt sequences vented through the 6 inch wetwell and drywell vent lines. Class ID dominates this release category.

c. Late Release Events: RC8-RC11 The accident sequence evaluation identified degraded core accident sequences in which the containment integrity is challenged by failure of the heat removal function. Core cooling is lost only after an extended period, induced principally by the loss of containment integrity. Following loss of core cooling, the fission products would be released in an initially open containment. The leakage path can occur through the drywell, with potentially severe release fractions.

However, since core degradation occurs only after an extended period (on the order of a day), the decay heat generation rates would be so low that the energy would not be high enough to result in the significant releases expected from an initially failed containment prior to core melt. These accident conditions are characterized as Class IIA. (The core melt progression essentially parallels that of Class IVA, except for the extended timing involved due to the shutdown core state).

These late overpressure scenarios (late due to the late initiation of core uncovery as defined by Class II), are represented by RC8.

Thermally induced failures of the drywell following long-term pressurization of the containment, where revolatilization is expected, are included in RC9. The containment leakage area is postulated to be ,

small since the thermal loads are associated  !

with the low decay heat levels and furthermore

a k

are mitigated by the steaming of the saturated suppression pool. In this group of core melt scenarios with long times to release, relatively low release magnitudes are obtained because removal in the secondary containment would reduce fission product concentrations prior to release to the environment. This release category is represented by late station blackout sequences. Both RC8 and RC9 involve pool bypass containment failure modes.

RC10 represents late noble gas releases uhere the release to the environment occurs after about a day. These are dominated by SCS-mitigated late core melt sequences as in Class IIF, where the particulate fission products are scrubbed. Included in this release category are the Class I and III late overpressure failures in the wetwell airspace, where the suppression pool is not bypassed.

With the SCS installed, overpressure failures of the containment, both early and late, are mitigated. Early containment pressure challenges cause the vent to open, and leakage through the gravel bed is assured. This has two significant implications for the release characterization of the event. First, the fission product aerosols are removed from the gas stream as it is released, and secondly, the timing of the release is extended in terms of release initiation and release duration.

This sequence of events is presented by RCll, in which primarily noble gases are released over a very extended time period.

d. Recovered Events: RC12 An important observation derived from this assessment is that containment integrity can be maintained for a substantial portion of the severe accident spectrum. The containment would not fail and the core melt sequences would lead to a contained release where the environmental consequences would be insignificant. The radionuclide releases, even of the noble gases (which generally are not attenuated in the containment system),

would then be limited to the design leakage rates from the containment. These recovered core melt events, represented by RCll, are principally determined by the Class IA sequences. Other sequences included in this

?

release category include degraded core accident events where the containment may be vented but the core melt progression is eventually recovered. These may be terminated

) either prior to or subsequent to vessel failure, int significant environmental releases c's not occur.

7. With-SCS Vs Without-SCS Comparison A comparison of the release spectrum frequency redistribution for the "Without-SCS" and "With-SCS" cases is shown in Figure IV.C-1. A significant reduction in the frequency of the early-severe release categories (RCl to RC4) is evident with SCS. A major shifting into RCll (w/ filtered release) with SCS is evident.

This risk reduction may be compared with that of the 1983 Shoreham PRA for both plant configurations. A qualitative release characterization of the source terms at Shoreham may be obtained by estimating off-site doses given discrete release fractions. The radiological consequences shown in Figure IV.C-2 of the range of potential source terms are presented below:

Severe means that all of the noble gases are released early and significant fractions of volatiles and particulates are released.

Moderate implies that principally noble gases and some particulates are released to the environment.

The release is moderately delayed from accident initiation.

Minor means that primarily noble gases and very small fractions of fission product aerosols are released. This includes all scrubbed release categories.

Negligible include very small fractions of fission products (noble gases and aerosols) are released since core melt is arrested or that containment leakage is limited.

A similar characterization was done with the 1983 Shoreham PRA and such groupings are used for this comparative risk study. Figure IV.C-2 provides a graphical overview of the risk implications of the different plant configurations evaluated thus far for the Shoreham power station. Note that most of the degraded core accident sequences are mitigated

3 due primarily to .e inherent mitigating design characteristics t the Shoreham containment. As can be seen in F ure IV.C-2 SCS has a major effect on reduc'.1g the frequency of releases with 3 severe and modr. ate radiological impacts.

I P

i I

D REFERENCES f l

IVC-1 Containment and Phenomenological Event Tree Evaluation at

) Full Power - Shoreham Nuclear Power Station, by SAIC, February 1988.

I IVC-2 IDCOR Technical Report 17.5.

)

)

)

)

)

u TABLE iV.C-1 StiMMARY TAllLE DESCR1111NG TIIE COIIPLING BETWEEN CONTAINMENT PRESSIIRIZATION RATE, l.EAEAGE SIZE, AND BREAK LOCATION (FOR TEMPERATl!RES < 500 F)

CONTAllalENT FAILURE DISTRIBIITION ACCIDFUT SEQllENCE TYPES CONSIDERED LEAFACE SIZE BREAK LOCATION DESCRIPTION Containment failure due to Breaks in suppression chamber (IN)

Pressure up to 135 psia. considered more likely than the Slow pressurization: rellatively small breach.

Smal1 Breaks judged to be m (Leakage beyond design basis) drywell region. Breaks below l

' waterline of suppression pool (W) more likely. do not relieve pren:ure adequately to prevent other containment failure locations.

Containment failure due to Breaks in suppression chamber Pressure i'p to 115 psia considered more likely than Rapid and sustained rupture causing a rapid

' depressurization. (Large drywell. Breaks at wetwell wall pressurization: large and wall and basemat juncture small breaks judged to be leakage rates) judged equally likely given equally likely suppression charber failures.

- t2 -

- v v v v-- - -

TABLE IV.C-2

SUMMARY

TABLE DESCRIB5NG THE COUPLING BETWEEN CONTAINMENT PREFSURIZATION RATE, LEAKAGE SIZE, A!!D BREAK LOCATION (FOR TEMPERATURES >500 F)

ACCIDENT SEQllENCE TYPES CONTAINMENT FAIIURE DISTRIBUTION Cof:SIDERED LEAKAGE SIZE BREAK LOCATION DESCRIPTION FOR TD1PERATURE T 500 $ T <800 l Pressure in tSe range of SCS Thermally induced containment Break location is in the back pressure (60-75 pala) failure resulting in a relatively drywell.

l small breach.

1 Pressure approaching 135 psia Small breaks more likely than Breaks in the drywell and l large breaks. wetwell equally likely.

1 FOR TEMPERATURE 2 800 F l Any pressurization. Thermally induced containment Break location is in the-

! Moderate and large breaches failure resulting in moderate drywell.

Judged to be equally likely. breach.

l FOR TD1PERATURE >l200 F l

l Any pressurization. Thermally induced containment Break location is in the l failure resulting in large drywell.

I breach.

l l

l I

l _ 73 _

l,_._.,._--_-.-, . _ ~ .

D 1

TABLE IV.C-3 SHOREFR4 COtCAIM'. cit EVCC TREE END STATE CATEGCRI"JTION i

Rf'M FDDE QUALITATIVE ATTRIBtHES "A States" Core Cooling Pecovered with Al,A2,A3 -

Containment i;. tact (A1); vented (A2); late CF (A3)

"B States" Core cult with containment failure in tPe short term (i.e., prior to or shortly after vessel breach)

B1,B4* -

Slow release through the pool with RB retention B2,B5* -

Slow release bypassing the pool with RB retention D3,B6* -

Slow release bypassing the pool i B7 -

Moderate release through the pcol B9 -

!bderate release bypassing the pool but with P3 (

retention B9 -

Fbderate release bypassing the pool B10 -

Large puff release through the pool B11 -

Large puff release bypassing the pool with P3 retention B12 -

Large puff release bypassing the pool "C States" Core melt with containment loss of integrity including venting long after vessel breach and release frcm the fuel C1,C2 -

Vented release with gravel bed decontaidnation C3,C4 -

Delayed the=al-induced drywell failure bypassing the SCS C9,C10 -

Late the=al-induced drpell failure bypassing the SCS C5,C7 -

Late over-pressure failures through the suppression pool C6,C8 -

Later over-pressure failure bypassing the pool "D States" Core mit with debris cooling ex-vessel D1.D2,D3 - Vented (Dl,D2) or contained release (D3) where design leakage rates determine release m gnitudes of noble gases D4 -

Small leakage rates determine release m gnitudes of noble gases

• S all and large contai.mt failure si:e, respectively. For srall leakage size, the noble gases are esticated to be released in two phases, an initial puff folicwed by a slev release for several hours because of the lack of sufficient fissicn prcduct driving ferce.

l TABLE IV.C-4

SUMMARY

OF CORE MELT ACCIDENT FREQUENCIES (WITH AND WITHOUT SCS)

CORE MELT SEQUENCE FREQUENCIES FROM ALL INITIATORS WITH SCS CASE WITHOUT SCS CASE PLANT TOTAL FRACTION TOTAL FRACTION DAMAGE FREQUENCY OF TOTAL FREQUENCY OF TOTAL CLASS IA 2.6E-05 0.72 2.6E-05 0.62 CLASS IB 1.2E-06 0.03 1.2E-06 0.03 CLASS IC 1.1E-08 3.0E-04 4.9E-09 1.2E-04 CLASS ID 2.2E-C6 0.06 2.2E-06 0.05

'.?SS I 2.9E-05 0.81 2.9E-05 0.70 C IIA 6.9E-06 1.9E-03 8.1E-06 0.19 CLASS IIB 5.7E-09 1.6E-04 8.0E-08 1.9E-03 CLASS IIT 3.4E-06 0.09 4.4E-07 0.01 Class II 3.5E-06 0.10 8.6E-06 0.21 CLASS IIIA 2.9E-07 8.0E-03 2.9E-07 0.01 CLASS IIIB 7.4E-07 2.0E-02 7.4E-07 0.02 CLASS IIIC 3.6E-07 9.9E-03 3.6E-07 0.01 CLASS IIID 4.0E-07 1.1E-02 4.0E-07 0.01 CLASS III 1.8E-06 5.0E-02 1.8E-06 0.04 CLASS IVA 4.4E-07 1.2E-02 2.1E-06 0.05 CLASS IVF 3.6E-07 9.9E-03 CLASS IVG 6.4E-07 1.8E-02 CLASS IV 1.4E-06 4.0E-02 2.1E-06 0.05 CLASS V 1.4E-08 3.9E-04 1.4E-08 3.3E-04

______________ . ____________M_W________4-____ .________..________________

TOTALS 3.6E-05 4.2E-05

- . - - - - - - - - - - - _ _ _ _ J

D Table IV.C-5 COMPARISON OF RELEASE CATEGORIES FREQUENCIES RELATIVE TO SCS Release With SCS Without SCS Category Frequency  % Frequency  %

RC1 4.lE-08 0.11 1.9E-07 0.45 RC2 5.3E-08 0.15 2.1E-07 0.5 RC3 3.7E-07 1.0 1.7E-06 4.0 RC4 3.7E-07 1.0 4.8E-97 1.2 RC5 2.4E-08 0.07 1.3E-07 0.3 RC6 5.1E-08 0.14 4.8E-07 1.2 RC7 9.4E-07 2.6 2.5E-07 0.6 RC8 1.2E-08 0.03 3.2E-07 0.78 RC9 5.4E-08 0.15 4.5E-06 11.0 RC10 3.4E-06 9.5 3.9E-06 9.3 RC11 1.1E-06 3.0 N/A RC12 3.0E-05 82.3 3.0E-05 71.0 TOTAL 3.6E-05 100.0 4.2E-05 l'1.0

TAPIF, IV.C-6 SilOREIIN4 REIFXiE CATEGORIFS AND SOUFCE TETNS CIIARACITRISTICS (With SCS)

Release Category RC1 RC2 RC3 FC4 RCS TC6 Prequency 4.1E-8 5.3E-8 3.7E-7 3.7E-7 4.2E-8 5.1E-8 (per reactor year)

Permnt of Core Fbit .11 .15 1.0 1.0 .07 .14 Rink Ibninant Sequence Plert Damage State IVA(1) IVA IVA I]ID IB1 IA (A'IWS) (A'lWS) (A' INS) (SEISMIC (SPO) ('IUJX)

Releaso Mode Early Early(2) Early(3) W Mod.Early Mod.Early W WW kW Initially W 'Ihermal W 'lhermal Failure Failure Failure Failed Failure Failure (B12) (B9) (B10) (B6) , (BS) (C3) (C4)

Source Tenn Sequence MAAP Secpience IVA.1A IVA.1B IVA.1B IIID.1 SBO.RCSA ' IDIO.BC6A Time to Pelease (br) <6 <6 <6 <6 6-74 6-24 Fission Product Species (percent of Core inventory) tbble Gases 100 100 100 100 100 50-100 Particulates > 20 > 10 < 0 .1 .1-10 > 10 .1-10 (1)SCS system is actuated but is inadeaunte due to high reactor power ( 20%).

(2) Coincident failure of the downccner vent pipe in the pedestal defeats pool scrulhing effectiveness.

(3) Pool scrubbing is maintained following wetvell failure.

TARIE "IV.C-6 (Continued)

SIOREFIAM RELEASE CATEGORIES AND SOURCE TERMS CIIARACFERISTICS (I!ith SCS) l Release Category RC7 RC8 IC9 RC10 BC11 RC12 Frequency 9.4E-7 1.2E-8 5.4E-8 3.4E-6 1.1E-6 3.0E-5 (par reactor year)

Percent of Core Melt 2.6 0.03 .15 9.5 3.0 82.3 Risk Dcatinant Sequence Plant Damage State IVG IIA IIB IIF ID IA (A' INS) ('lW) (Late) ('lW) ('lVt1V) ('IQUX)

Sao)

U Release thle Mod.Early Late Late Late Controlled Vented Slow Pelease Release DW %ermal Release Release Core Melt

% rough  % rough DW Failure  % rough  % rough & Recovered FILTRA FII.TRA FII.TRA Ex-Vessel q (C2,C1) (B11,B12,B9) (C9) (C1,C2) (Cl) (Dl, Al) i Scntree Term Sequence MAAP3.0 Sequence IVG.DC7A IIA.1 IVDUX.RC8 IIA.1 'ITUV.RC11 'IDUX.RC12 (modified timing)

Tine to Release (hr) 6-24 > 24 > 24 > 24 > 24 NCF 1

Fission Product Species (percent of Core inventory)

Noble Gases 100 100 100 100 < 50 <10 I Particulates 0.1 > 10 .1-10 <q 0.1 < 0.1 < 0.1 (4)2e radiological analysis conservatively crsnbines RC8 and RC9.

- 78

. o TABLE IV.C-7 SilOREllAM RF3FASE CATECORIES AND SOURCE TEIMS CIIARACTERISTICS (Without SCS)

PC3 RC4 PCS RC6 Release Category ICI RC2 2.1E-7 1.7E-6 4.8E-7 1.3E-7 4.8E-7 Frequencv 1.9E-7 (per reactor year)

.5 4.0 1.2 .3 1.2 Percent of Core Melt .45 Risk Dminant Sequence IVA IVA IIID IB1 IA Plant Damage State IVA(1) (Early (1UJX)

(A'lWS) (A*IWS) (A1WS) (SEISMIC PIV) SDO)

Early( } Early( } W Mod.Early Mod.Early Release Mode Early N Thermal DN W W Initially N 1hermal Failure Failed Failure Failure Failure Failure (B10) (B6) , (BS) (C3) (C4)

(B12) (B9)

Source Term Sequence IVA.1B IIID.1 SDO.RCSA TDIO.RC6A MAAP Sequence IVA.1A IVA.1B (modified)

<6 6-24 6-24 Time to Release (hr) <6 <6 <6 Fission Product Species (percent of Gore inventarv) 100 100 100 50-100 Noble Gases 100 100

> 0.1 .1-10 > 10 .1-10 Particulates > 20 > 10 (1) Venting system is actuated but is inadequate due to high reactor power (>18%) .

(2) Coincident failure of the downccmer vent pipe in the pedestal defeats pool scrubbing effectiveness. )

(3) Pool scruthing is maintained following wetwell failure.

79 -

y

_ _ .. m y g .

TABIE IV4-7 (Continued)

SilOREllAM RELEASE CATEGORIES AND SOURCE TEP11S CIIARACIERISTICS (Without SCS)

Release Category RC7 RC8

' RC9(6) RC10 RC11 RC12 Fraluency 2.5E-7 3.2E-7 4.5E-6 3.9E-6 N/A 3.0E-5 (par reactor year)

Percent of Core Melt .6 .78 10.7 9.3 71 Risk Dcminant Sequence Plant Damage State ID IIA IIP IIA IA

("IUN) ('IW) (Vented ('lW) ('IQUX)

'IW)

Release Mode Mod.Early Late W and Iate Vented Release Release DW Release Core Melt 6 in W 'Ihrough DW Release Vent W & Recovered Vent to Rx Bldg. Ex-Vessel (C2,Cl) (Bil,B9) (C3,C4) (BIO) (B10) (D1,A1)

Scarce Term Sequence MAAP3.0 Sap 2ence 'IQUV.RC7 IIA.1 'IWC IIA.1(5) 'IUTX.RCl2 (modified Time to Release (hr) 6-24 g 24 > 24 > 24 NCF Fission Product Species (percent of Core inventory)

Noble Gases 100 100 100 100 < 10 Particulates < 0 .1 > 10 .1-10 < 0.1 < 0.1 (4)Not applicable for Shoreham configuration without FILTRA.

(5) Source terms modified to reflect pool scrubbing consistent with this release category (i.e., particulate release fractions are set to<l%)

(6)'Ihe radiological analysis conservatively cmbines PC8 and RC9.

D SHOREHAM FULL POWER PRA UPDATE CET RELEASE BINS

= d$ .

s/q 9

$ /j

!  % f 4

e i

ic.

%q 3 5{s g p

p i; <' a d th, s s' a E S p , a t t t*a gg l

! 4 4 t 4 d

a gl

$ 1sfu $$ $$*h @ h 44gb -

sg s st

, 2 3 4 3 6 7 8 9 10 it 12 C W/ S C S W/O SC3 SHOREHAM FULL POWER PRA UPDATE CET RELEASE Bif(S 80 - f 5 s 7o -

p eo-

! s

S. sk

- 6s%

g so - f e

~

o$(

f st o .

,+ymms S . 3 . , S 338s$3 G W/ SC S G W/O SCS FIGURE IV.C= 1 COMPARISON CF RELEASED FREQUENCY i

O Y~

ACCIDENT SEQUENCE FREQUENCIES

. POTEr4TUL COtiSEQUENCES ,

CORE WlNERABLE HE GIBLE' Ml? TOR MOO TE- SD DRE 9 PRA WOS S -

W/SCS -

FIGURE H.C-2 Sumury of Acctotat Sequence frequencies .for Various tevels or Potential Consequences

D. Source Term Analysis

1. Introduction l

"Source Term" refers to the quantity, timing, physical and chemical characteristics of the radioactivity that would be released to the environment in the event of a severe accident in a nuclear power plant. The principal interest in the deterministic evaluation of source terms is for the evaluation of off-site consequences resulting from postulated core melt events. In order to estimate off-site risk, potential consequences to the public must be evaluated for a spectrum of accident sequences leading to a release of radionuclides to the environment. Accident sequences which have similar off-site consequences are grouped together as discrete release categories. The identification of risk dominant accident sequences to represent each release category was accomplished in the containment event tree analysis. These p

representative sequences for the release categories are chosen to result in a spectrum of release events which will characterize the pote pirl off-site consequences from a hypc theJical core celt accident at Shoreham during full p;wer operatiot..

The source terms which characterize the release p

categories were calculated by use of the IDCOR source term code MAAP3.0.

2. The IDCOR Source Term Methodology The Modular Accident Analysis Program (MAAP) is a computer code which simulates light water reactor system response to accident initiation events. It was prepared as a part of the IDCOR (Industry Degraded Core Rulemaking) program to investigate the physical phenomena which might occur in the event of a serious light water reactor accident leading to core damage, possible reactor pressure vessel failure, and possible containment failure and depressurization. In the current Shoreham 100%

power PRA update, the MAAP code is used to assess the ultimate fission product releases to the environment for those very low probability events which result in core degradation and fission product releases from the containment.

MAAP can predict the progression of hypothetical accident sequences from a set of initiating events to either a safe, stable, coolable state or to containment failure and depressurization. MAAP treats a wide spectrum of phenomena including:

)

o Primary system heat transport; o Primary system loss of water inventory and accumulation in containment;

)

o Emergency core cooling sources and paths; o Operator interventions; o Core uncovery, heatup, hydrogen and fission product release into the primary containment; o Hydrogen migration into the containment; o Fission product transport and settling in the primary and secondary containments; o In-vessel core migration, fragmentation, steam generation, and additional hydrogen formation; o Reactor vessel failure and ablation; o Ex-vessel steam generation and hydrogen formation; o Core debris entrainment and coolability, concrete attack and carbon monoxide generation; o Hydrogen and carbon monoxide combustion in containment; o Ex-vessel heat transport, water inventories, and containment cooling; o Long-term heating of the primary system due to deposited fission products; and o Containment failure or venting and depressurization.

MAAP allows for nodalization of the four major physical control volumes consisting of the RPV, Reactor Containment (including the pedestal cavity, drywell and wetwell), Reactor Building and Supplemental Containment System. (See Figure IV.D-1 and IV.D-2). Also shown in the figure are the modeled flows of steam, water, corium, hydrogen and other gases (CO, C0 7, N2, 02) between the Reactor containment regions.

I

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ ___ J

l l

A simple MAAP model was developed in order to i represent the thermal-hydraulic conditions within

! the proposed SCS. Since the existing reactor building model has the ability to represent the i

thermal-hydraulic and fission product transport within a control volume, it was used to also model the SCS with modifications to account for the relieving valve outlet configuration. A single l

temperature is assumed to represent the entire I gravel mass in this single node model.

! 3. Source Terms Twelve Release Categories (RC) were chosen to represent the spectrum of releases for the plant

! condition with SCS installed. Without SCS, releases are represented by 11 Release Categories.

With the exception of RJ7, RC9, and RC10, the same representative MAAP sequences are used for both the with and without SCS release categorization. The l

MAAP 3.0 results for all the release categories are l summarized and presented in Table IV.D-1 and IV.D-2. CsI is used as a representative particulate release because of its volatility and potential health effects. In general, early releases (RC1 - RC4) occur when containment integrity and core injection are lost soon after accident initiation. Delayed releases (RCS - RC7) are realized by either the containment maintaining its integrity for a period of time after core melt, or by the fact that venting is delayed by the thermal capacity of the suppression pool and/or SCS. Late releases (RC8 - RC10) are characteristic i of containment overpressure failures caused by a

- loss of decay heat removal accident. Controlled releases (RCll) are typical of decay heat driven releases from the SCS. The primary contributors to RC12 are scenarios where core cooling is recovered with the containment intact or vented and core melt sequences with ex-vessel debris cooling where core releases may be contained or vented. Release duration and release fractions are highly sequence dependent. The physical phenomena which dictate these release characteristics are discussed in the next section.

4. Release Category Representative Sequences i

As part of the containment event tree analysis, dominant contributors to each release category are j identified. The binning process enables the selection of one accident scenario to represent each release category. This section will summarize the progression of the accident sequences, as l calculated by the integrated source term code 1 MAAP3.0, which have been defined to represent each  !

of the twelve release categories.

__b

l 3

With the exception of three RCs, the same source term sequences are used to represent the release categories with and without SCS. Release Categories 7, 9, and 10 will utilize different 3 sequences to meet the plant condition with and without SCS in operation. Release Category 11 is defined with only the SCS in operation.

a. Sequences Analyzed with SCS D RCl - Anticipated Transient Without Scram (IV A.lA) :

For this ATWS sequence, the reactor power is not controlled because neither pressure nor water level control is assumed. A core power I of about 30% rated power is predicted by the Chexal-Layman correlation (developed at EPRI) which determines ATWS power as a function of vessel pressure and water level. The wetwell pressure has reached 75 psia in about a half an hour and the wetwell vent to SCS is opened.

P As the ATWS power exceeds the vent capacity of SCS, the assumed containment ultimate pressure capacity of 135 psia is reached at 1.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.

The containment failure location is modeled in the drywell head ring beam region to provide a direct pathway tc the upper region of the reactor building, namely, the refueling floor level. This is a conservative assumption since scrubbing of fission products between the failure location and refueling floor level would be expected. All injection systems are assumed to be lost at the time of containment failure. Reactor pressure vessel (RPV) failure is calculated to occur at 3.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

The RPV high pressure blowdown not only results in dispersal of molten core debris to the drywell but also creates the driving force to expel fission products to the reactor building through the drywell break, bypassing the suppression pool ang SCS. A drywell leakage area of 0.5 ft. , located in the upper region of the reactor building results in a relatively low reactor building decontamination factor (DF) of 1.5 for CsI.

The MAAP3.0 calculation indicates that following vessel failure, a rapid release to the environment of about 40% CsI and 100%

noble gas occurs.

)

RC2 - Anticipated Transient Without Scram (IV A.lB) :

This source term sequence is similar to the 3 one for RC1, with the SCS system bypassed due to containment overpressure failure in the wetwell airspace subsequent to SCS vent actuation. Suppression pool bypass is assumed to occur and is modeled as a failure of the pedestal downcomers at the time of vessel failure. However, for this representative sequence, the release of particulates to the environment is somewhat attenuated, since the assumed containment failure location is in the lower portion of the reactor building.

Consequently most of the building is available for fission product deposition. MAAP3.0 calculates a reactor building DF of approximately 4 for CsI, resulting in 16% CsI and 100% noble gas being released to the environment.

RC3 - Anticipated Transient Without Scram (IV A.lB, mod . ) :

The ATWS sequence fcr RC2 is modified to be used as a surrogate representative sequence for RC3, The release path for RC3 involves the wetwell, as in RC2, but the pedestal downcomers are assumed to maintain their integrity. Thus, the fission product aerosols are scrubbed in the suppression pool, with a negligible amount being released to the environment. Typical pool decontamination factors calculated by MAAP3.0 are on the order of 1000, resulting in insignificant particulate releases.

RC4 - Seismically-Induced LOCA (III D.1):

The seismically-induced RPV breach and simultaneous failure of the drywell head initiate an early release of fission products to the environment. All vessel injection systems are assumed to be lost at time zero.

Although the noble gas release starts early, it takes about ten hours to achieve a 100%

release. CsI release to the reactor building is also slow because of no substantial carrier gas flow. This results in long residence time inside the building and a calculated DF of approximately 5 for the CsI. About 7% of the CsI inventory escapes the reactor building and enters the environment.

h RCS - Early Station Blackout (SB0. RC5 A) :

The accident sequence used to represent Release Category 5 is an early station 3 blackout for which no mitigating systems are available and AC power is not recovered. Core coolant makeup is lost in one hour and the reactor is unable to be depressurized. About 25% of core debris is entrained to the drywell when the RPV fails at high pressure.

' Containment failure is assumed to be thermally induced in the drywell. Essentially 100% of the noble gas is released to the environment within 40 hours4.62963e-4 days <br />0.0111 hours <br />6.613757e-5 weeks <br />1.522e-5 months <br />. As a result of revolatilization due to high drywell temperatures and the bypassing of both suppression pool and the SCS gravel bed, approximately 90% of CsI inventory is released to the reactor building. MAAP3.0 calculates an eventual release of 15% CsI to the environment resulting in a reactor building DF of six.

RC6 - Loss of Coolant Injection (TDIQ.RC6A):

The source term sequence for this release category represents a delayed release event with both the suppression pool and SCS bypassed. As in RCS, the initially deposited volatile fission product aerosols are revolatilized and released through a thermally impaired containment, but some mitigation of the releases is modeled in the form of reactor building retention. Slow leakage from the containment to the reactor building enhances the retention of fission products in the reactor building. Releases from the reactor building include 70% of the noble gas inventory and 4% CsI. A decontamination factor of about 20 is calculated for CsI released into the reactor building.

RC7 - Anticipated Transient Without Scram (IVG.RC7A):

The representative sequence is a vented ATWS scenai'io. With pressure control and water level maintained at near normal level a core power of about 15% of rated, is calculated by the Chexal-Layman correlation. At 0.95 hours0.0011 days <br />0.0264 hours <br />1.570767e-4 weeks <br />3.61475e-5 months <br />, the containment pressure has exceeded 75 psia and the wetwell vent to SCS is actuated. With the RHR system removing about 5% of initial power, the SCS system is able to relieve the rest of the ATWS power. At 1.1 hours1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br /> the hotwell water inventory (220,000 gallons) is depleted and one of the RHR loops is placed

b into the vessel injection mode. With no external make-up to the suppression pool available, the water level drops below the RHR pump suction intake at 4.2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />. At this B point all vessel injections are lost. At 4.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br /> the core uncovers, followed by the onset of core melt at 5.9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br />. Vessel failure is calculated to occur at 7.2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />. The sustained steam flow to the SCS system saturates the gravel bed in about two hours, P well before core heat-up begins. As a result, when the noble gases are released into the SCS at core melt, the gravel bed is already saturated and pressurized and there is little hold-up of the noble gases prior to release to the environment. On the other hand, substantial deposition of the fission product aerosols in the SCS is expected. A SCS DF of 588.2 is assumed for the particulates. This results in fission ogoduct release fractions of less than 1 x 10 for all the particulates. Since the releases from SCS are through SRVs with setpoints high enough to guarantee a high discharge velocity, the release will be elevated resulting in a greater dispersion of fission products.

RC8 - Loss of Decay Heat Removal (IIA.1):

This release category is represented by a Class IIA accident sequence where the SCS system is unavailable and core melt is induced by containment failure, after a long term pressurization due to the loss of decay heat

) removal capability. Loss of core inventory through the SRVs due to decay heat boil off, is compensated by makeup from the ECCS. At 34.7 hours8.101852e-5 days <br />0.00194 hours <br />1.157407e-5 weeks <br />2.6635e-6 months <br />, the containment pressure has exceeded 135 psia and a drywell failure in the head ring beam region wjth an equivalent failure area of 0.1 ft. is assumed. All injection (including CRD flow) is assumed to be lost at the time of containment failure.

Since the RPV fails when the containment is l still at elevated pressures (90 psia), there is a driving force to expel the fission products out to the reactor building, through the pre-existing breach in the drywell. The result is a rapid release of noble gases to the environment and a minimal retention of particulates in the reactor building. The final releases are, 100% noble gases and 9%

CsI.

l 1

l

D RC9 - Late Station Blackout (IVDUX.RC8):

The representative source term sequence for RC9 is modeled after a Class IB plant damage J state with late containment over-temperature failure. All high pressure injection systems are assumed to be lost at twelve hours as a result of depletion of the station batteries.

A high pressure RPV melt through is calculated to occur at 16.3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />. Approximately 25% of the core material is entrained to the drywell.

At vessel failure, a pressure spike in the containment actuates the wetwell vent to the SCS. Core debris in the drywell slowly heats up the dgywell airspace. A leakage area of 0.05 ft.' is assumed to be created in the upper regions of the drywell, when a temperature of 800 F is exceeded. From this time on, the sequence assumes all releases from the containment to be directed to the reactor building, bypassing both the suppression pool and the SCS system. As a result of post-vessel failure revolatilization, CsI release into the reactor building begins soon after drywell thermal failure has occured. However, the small drywell leakage area leads to a slow flow into 4 the reactor building. A decontamination factor of twenty is calculated for CsI. The long residence time in the reactor building also helps the retention of noble gases.

Eventually, only 70% of the noble gas inventory and 3% of CsI are released to the environment.

RC10 - Loss of Decay Heat Removal-SCS Vent Actuated (II A .1, Mod . ) :

A Class IIF plant damage state is a dominant contributor to this release category. Since only noble gas release is expected, and the time of release can also be estimated, a MAAP sequence was not required for this release category. Instead, the source terms for RC8 are modified to reflect the ef fects of SCS venting. Subsequent to the loss of low pressure injection systems at vent actuation, only the CRD flow at 180 gpm is sufficient to provide core makeup. With a CST inventory of 550,000 gallons, the CRD flow is expected to last for about 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br />. After CRD is lost, the core would uncover, melt and eventually fail the vessel in a manner similar to that for the RC8 sequence. As the SCS vent is actuated prior to loss of core cooling, the SCS gravel bed is expected to be saturated during core melting. Thus, significant

D hold-up of noble gases is not anticipated.

The particulates are either scrubbed in the suppression pool or removed by the SCS. Noble gas release is assumed to start at 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br />, 3 with a release duration comparable to the RC8 representative sequence.

RCll - Loss of Coolant Injection - SCS Vent Actuated (TQUV.RCll):

The MAAP scenario which is used to represent RCll involves a transient with early loss of ECCS for which the reactor is depressurized and venting to the SCS is successful.

Significant hold-up and delay in the controlled re.l ease of the noble gases are two of the more visible benefits of the sCS. With a loss of all core injection at time zero, RPV failure (at low pressure) is calculated to occur at 1.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />. At the time of vessel failure the pedestal downcomers are assumed to fail, causing the fission products to bypass the suppression pool before they are released to the SCS. The SCS is actuated at 3.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />.

With the additional heat capacity provided by the gravel bed, it takes the SCS about twelve hours from the time of vent actuation to reach the SCS Stage I relief valve pressure setpoint of 54 psia. This represents a significant delay in the release of noble gases whose release to the environment is further held back by the controlled release through the SCS relief valves. At the end of 60 hours6.944444e-4 days <br />0.0167 hours <br />9.920635e-5 weeks <br />2.283e-5 months <br />, the gravel bed is still subcooled and only about 80% of the total noble gas inventory is released at high velocity to the environment.

Though the suppression pool is bypassed, the SCS is able to retain the overwhelming majority of the particulates (a DF of 588.6 is assumed 5 f r CsI) and a CsI release fraction of 3 x 10 is calculated. Since the RPV fails at low pressure, there is no debris dispersal to the drywell and ghe drywell gas temperature does not exceed 800 F.

RC12 - Loss of Coolant Injection - Recovered After Vessel Failure (TQUX . RCl 2 ) :

The source term sequence for RCl2 represents a recovered state after reactor vessel failure.

Containment failure is prevented and releases are in the order of design leakage only. The sequence is initiated by MSIV closure followed by reactor scram. The high pressure systems are not available and the reactor is assumed to be unable to be depressurized. The core uncovers in 0.9 hrs. followed by core melt at 2.0 hrs. and RPV failure at 3.2 hrs.

g

)

Once the failed RPV has become depressurized, both loops of RHR are realigned to vessel injection through the heat exchangers. With both loops of LPCI injecting into the failed

) vessel, the core region is kept cool and the water leaving the vessel is able to cool the debris on the pedestal and drywell floor. The drywell gas space slowly heats up as a result of heat transferred from the upper vessel internals. About 21 hours2.430556e-4 days <br />0.00583 hours <br />3.472222e-5 weeks <br />7.9905e-6 months <br /> after vessel

) failure, the drywell gas temperature has 0

exceeded 500 F and thg drywell leakage area is increased to 0.001 ft . Significant hold up of noble gases and CsI is observed in the reactor building. Noble gas and CsI release fractions to the environment are respegtively g calculated to be only 0.01 and 3 x 10

b. Sequences Analyzed Without SCS In order to quantify and compare the off-site consequences with and without the SCS g

installed, a spectrum of source terms has also been determined for the plant without SCS.

Except for the following three release categories (RC7, 9, and 10), the representative source terms are common to both the with and without SCS cases. For the

' without-SCS plant condition, there is no corresponding plant damage state that will result in a controlled release which is characteristic of RCll.

RC7 - Loss of Coolant Injection - Wetwell Venting (TQUV.RC7)

This vented sequence represents moderately early release of noble gases without significant hold-up in the reactor building.

This loss of all injection sequence (except CRD) leads to RPV failure at 1.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.

Although the 6 inch wetwell vent is effective in containment pressure control, a thermally induced drywell leakage area of 0.001 ft.' is developed when ghe temperature in the drywell has reached 500 F at 17.8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. In the first 50 hours5.787037e-4 days <br />0.0139 hours <br />8.267196e-5 weeks <br />1.9025e-5 months <br /> of the transient, this small drywell leakage, which bypassed the suppression pool, gives a CsI release fraction of 0.002 to the environment and corresponds to a reactor building DF of 4. Although the whole reactor building is credited for aerosol deposition, the continuous release of steam from the wetwell, reduces the residence time of the particulates in the building. The release of the noble gases is 100% (which includes a percentage that is released through the RBSVS).

1 b

1 RC9 - Loss of Decay Heat Removal - Drywell and Wetwell Venting (TW .C) :

The source term sequence represent a

' successful containment venting case where catastrophic failure of the containment is prevented and the release of CsI is mitigated by the absence of significant post-vessel failure revoldtilization. With a loss of decay heat removal capability, the containment has reached a pressure of 75 psia at 16.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> and both the wetwell and drywell vents to the reactor building are actuated; they represent equjvalent vent areas of 0.0564 ft.2 and 0.0674ft. , respectively. Two hours after vent actuation, at 18.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />, all injection systems are assumed to be lost. Vessel failure is calculated to occur at 24.4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />.

Simultaneous failure of the pedestal downcomers is also assumed. Since the drywell is vented and the suppression pool is saturated, the superheated gas in the containment is constantly being purged from the drywell gas space. As a result, the drywell temperature is kept below the level at which significant CsI revolatilization could occur. Thus, only about 20% of the CsI inventory is released to the reactor building, with an eventual release of about 3% to the environment. This corresponds to a reactor building DF of about 7 for CsI. The whole reactor building is credited for aerosol deposition. The total release of noble gases to the environment is 100%.

RC10 - Loss of Decay Heat Removal (II A.1, mod.):

This release category represents a late scrubbed release of essentially noble gas only. The source terms can be derived from the representative sequence for RC8 by assuming a containment failure in the wetwell instead of the drywell. Hence, only noble gases are released to the environment. The remainder of the fission product groups are assumed to be scrubbed in the suppression pool.

5. Impact of SCS on Severe Accident Progression The Supplemental Containment System is designed to be a passive, engineered vent with scrubbing capability. It affects severe accident progression in terms of both prevention and mitigation.

D Since the release from the vent is external to the reactor building, much of the downside risk associated with containment venting is avoided.

The accessibility to the reactor building is g maintained, even under a vented condition. This improves the likelihood of achieving recovery actions during a severe accident.

The enhanced venting capacity of the SCS, not only increases the time for recovery actions, but also prevents the overpressurization of the containment for many ATWS and all loss of decay heat removal sequences.

The SCS provides additional free volume and thermal capacity to enhance the performance of the primary containment. The SCS is capable of mitigating a loss of vapor suppression incident such as the case of a stuck open vacuum breaker.

The pressure relieving feature of the SCS limits the overpressurization of the containment and allows continued operation of the ADS valves by retaining a sufficiently large pressure differential between the instrument air pressure to the valves and containment pressure. Thus, the RPV remains depressurized as might be required by emergency pressure control procedures.

In terms of off-site consequences, the benefits of SCS are numerous. It reduces the uncertaincy associated with the failure location of the containment because the SCS rupture disc is set at a pressure well below the ultimate pressure capacity of the containment and the SCS helps mitigate releases from the containment which are not scrubbed. The concern of pedestal downcomer integrity is alleviated by the fact that fission product aerosols that bypassed the suppression pool, are now going to be scrubbed in the gravel bed. The SCS is designed to support a minimum decontamination factor that is of the same order as what a suppression pool can provide. For Shoreham, most of the core melt accidents resulted in reactor vessel failure prior to containment being challenged. Consequently, without the SCS, containment venting leads directly to noble gas release for Class I and III accidents. However, with SCS, noble gas release is delayed. For example, in the case of RCll, the representative MAAP sequence shows that following SCS actuation a holdup of 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> occurs before the noble gases are released. With the added thermal capacity and additional free volume, SCS not only delays release but also can retain noble gases.

j

)

The DCOP System on the SCS is designed to assure that the releases, under Shoreham meteorological conditions, will be elevated. This assures a greater dispersion of the released fission

) products. Since the relief valves cycle opened and closed based on pressure, the release duration is expected to be extended also. After SCS vent actuation, the containment pressure is regulated by the relief valves on the SCS. The ability to maintain a containment back pressure improves the 3 likelihood that the ECCS pumps meet their NPSH requirements.

In conclusion, the SCS increases the time for taking mitigative actions before a release occurs.

It also enhances the performance of the primary I containment. Finally, the installation of the SCS reduces some of the uncertainties associated with containment end states which have significant impact on the source terms.

l

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

V LJ Table IV.D-1 Stzmary of Shoretrun PMAP 3.0100% Power Ibpresentative Sequences (with SCS)

Noble Gas Pelease Cs1 Release to Times are in lxxirs to Envirururent (2) Environment Felease MAAP Core Core Vessel Wetsel1 Containment Tine Duration Fraction Time Duration Fraction Category Soluence ADS Uncovery Melt Failure Vent Failure (1) (llr) (Ilr) (IIr) (Ilr)

With SCS IVA.1A N/A 1.7 2.4 3.4 .63 1.6 (DWF) 2.4 3.0 1.0 (R) 3.4 21.0 .40 (R) 1C1 1.7 2.4 3.4 .63 1.6(hMF) 2.4 3.0 1.0 (R) 3.4 21.0 .20(R)

IC2 IVA.1B N/A IC3 IVA.lB N/A 1.7 2.4 3.4 .63 1.6 (WF) 2.4 3.0 1.0 (R) 3.4 21.0 10 (R) l l

(Fixl. ) l

.63 N/A 0.0 (DWF) 0.0 15.0 1.0 (R) 0.0 22.0 .07(R) l RC4 IIID.1 N/A 5.2x10~ .21 1

SDO.ICSA N/A 2.4 3.3 4.3 N/A 15.7 (DWF) 4.8 36.0 1.0 (R) 6.0 36.0 .15 (R)

ICS 0.5 1.1 1.9 N/A 14.5 (DWF) 3.2 36.8 .70 (R) 6.0 14.0 .04 (R)

IC6 ' IDIO.RC6A N/A 1.0 (F) 7.2 38.0 10-5gg)

IC7 IVG.IC7A .17 4.7 5.9 7.2 .96 N/A 7.0 3.0 3.5 37.0 38.8 40.8 N/A 34.6 (DWF) 38.8 4.0 1.0 (R) 40.8 6.0 .09(R) 108 IIA.I 13.6 14.8 16.4 16.4 31.1(DWF) 17.3 23.0 .70 (R) 31.1 7.0 .03(R)

IC9 IVDUX.RC8 N/A

17. 50. 4.0 1.0 (R) 50. 6.0 10-5gp3 1C10 IIA.1 3.5 50. 50. 50. N/A (tkx1.)

.20 .70 1.6 3.5 N/A 15.3 44.7 .80 (F) 34.5 25.5 3x10~ (F)

Rc11 'IUN.RC11 .17 25.0 22.0 .01 (R) 34.0 13.0 3x10-5 (p, PCl2 'IUJX.IC12 N/A 0.9 2.0 3.2 N/A 24.8 (DWF)

Wotes:

(1) (IWF) stanis for drywell failure and (WF) starvls for wetwell failure.

(2) (R) starxls for release to environment through the reactor building.

(F) irvlicates the release is through SCS where a DF of 588.2 is prestrned for the particulates.

1

o -

u Table IV.D-2 Sisimary of Shoreham MAAP 3.0 1001 lbwer Representative Sequences (Without SG)

Noble Gas Release Csl Ikelease to l

l Tirnes are in hoitrs to Envircment (2) Environrient Idlease MAAP Core Core Vesr.e1 Wetwell Containnent Tinn Duration Fraction Tinn Duration Fraction Categor,f Sa]uence ADS Urxx)very Melt Failure Vent Failure (1) (llr) (Ilr) (llr) (Ilr)

Without scs ICI IVA.lA N/A 1.7 2.4. 3.4 .63 1.6 (DWF) 2.4 3.0 1.0 (R) 3.4 21.0 .40(R)

IC2 IVA.IB N/A 1.7 2.4 3.4 .63 1.6 (WWF) 2.4 3.0 1.0 (R) 3.4 21.0 .20 (R)

IC3 IVA.lB N/A 1.7 2.4 3.4 .63 1.6 (WWF) 2.4 3.0 1.0 (R) 3.4 21.0 10 (R)

(Fkxl. )

RC4 1110.1 N/A 5.2x10- .21 .63 N/A 0.0 ([MF) 0.0 15.0 1.0 (R) 0.0 22.0 .07(R) 105 sin.rCSA N/A 2.4 3.3 4.3 N/A 15.7(DWF) 4.8 36.0 1.0 (R) 6.0 36.0 .15 (R) 106 TDIO.RC6A N/A 0.5 1.1 1.9 N/A 14.5 (IMF) 3.2 36.8 .70 (R) 6.0 14.0 .04 (R)

-3 RC7 'lGN. RC7 .17 0.2 0.70 1.6 8.9 17.8 (DWF) 8.9 16.0 1.0 (R) 20. 30. 2x10 49) l 108 IIA.1 3.5 37.0 38.8 40.8 N/A 34.6 (IMF) 38.8 4.0 1.0 (R) 40.8 6.0 .09(R) l RC9 'IW.C 3.1 20.7 22.6 24.4 16.4( I N/A 22.6 12.0 1.0 (R) 24.5 35.5 .03 (R)

Ic10 IIA.1 3.5 37.0 38.8 40.8 N/A 34.6 (WWF) 38.8 4.0 1.0 (R) 40.8 6.0 10-5 (R)

(tr. Only) 25.0 22.0 .01 (R) 34.0 13.0 3x10-5gg) 1r12 'IUTX. RCl 2 N/A 0.9 2.0 3.2 N/A 24.8 (DWF)

Notes:

(1) (IMF) starxis for dt3wll failure and (WWF) starx3s for wetwll failure.

(2) (R) starvis for release to environment through the reactor building.

(F) irxlicates the release is through SC3 where a DF of 588.2 is prestuned for the particulates.

(3) 6" drywell vent is actuated sinultaneously.

3 &SRV r

-wwwwx" b & LOCA O 4 UPPER HEAD

)

Aw3wi Y 3 a g ,

e = LOCA 3

Ak::::::_=:_~~ =A .K E Y l STANDPIPES ~

SEPARATORS e A* O GAS FLOW P A T ri N O.

$ l

" ^""^'"'"*

@ = ====== '"A SURFACE NO.

%,\ \ ,\ ___

2 SHROUD 5 '

HEAT SINK HEAD UPPER A DOW NCOMER .

@ @y y

A m:::t Q xs us sxxs q q

6 LOWERS l CORE $ I b DOW NCOMER R g& ( :::_=_=::= +A q

Q' LOCA 8

REC!RC

^

d t LOOP exuw k O

Y g 7 g

i  ;

e a LOWER HEAD (Flow Path 8 Not

, ,, 2,, ,i sisi s ,is i / / / . . u , mu, App 11 cable for GEBWR's)

O Y

- rigure IV.D-1 DWR primary system nadalization. ._

) .

I l

l l

q, GACTOA VESSEL 2

CONTROL ,,,,,

A00 4048 p . Al . .

j .....rA6 l 4 .

Asesom

... 0.. 6 pas 4AcawAF

..i Nkr o e

L -.

h llp . . f% A f*********

l WETw8LL - j u

~ $

l esAe0==st .:::' ?-4 -

l0

"^'*""

i UU T1 0 0 YU lf . W ,, . . . . i O .

7006

)

.2 . . . ,

l SweAEMAM PalWAAY Conf AlleMEnf

) '..,,

CONTAueW4MY FARWAS O.,'.,e

.e g, S/M/S t=eseessee.4%.a.ed AgLigy y,Lygg g SJw/W /

)

I v4..n..

PRIMARY ,,,, ,,,," ,,,,,,,,,,,,,,

SYSTEM ,,,,, DRYWELL ,,,,, WETWELL Fif 0wAff A &

, CC: a cAe none .....

i ...,.

I i

WESedL #A4 wad lwoMJC/8 l PEDESTAL """'"*"* " " " " " '

sue **e s uon l CAVITY so nanscie

'oo' sowinse _

l

..... ..... i

...... ...~. i l l DOWNCOMERS S/W/H/C/O Figure IV.D-2 Shoreham Mark !! Containment.

(

D E. Offsite Dose Analysis

1. Introduction D The offsite doses resulting from the current PRA update, which are discussed in this section, are provided by utilizing the computer code, CRAC2.

Results are obtained for the sequences repre-senting each of the release categories, which together represent the spectrum of offsite release I modes. CRAC2 utilizes, as input, release fractions for fission products and the timing of the releases obtained from the source term analysis (using the MAAP code) for these sequences. Such results, in the form of the probability of exceeding a given dose at some i distance from the reactor centerline, have provided basic information utilized in the establishment of the plume exposure Emergency Planning Zone (EPZ) (Reference IVE-1). A discussion of emergency planning considerations is first provided as a focus for the general discussion. The methodology utilized in obtaining the offsite dose consequences is then considerea.

An important feature of the consequence calculations for releases from the SCS is the elevation it provides to the exiting plume (in addition to its capacity for fission product holdup and scrubbing). This is due to the uprard momentum of the jet. An understanding of this feature is provided in a discussion of the treatment of elevated and ground releases.

The discussion of the results which are obtained from the offsite consequence calculations is based on curves for probability (conditional on core melt) of exceeding a particular dose as a function of distance from the site, for both with and without SCS. Comparison is made between these two cases and NUREG-0396 results.

Risk characterization is then obtained using NUREG-0396 criteria regarding severe accidents.

NUREG-0396 also provides considerations regarding EPZ size based on design basis accidents (DBA).

An analysis is therefore provided of the application of these considerations to Shoreham.

An alternative risk characterization method that has been recently developed considers a radius for injury-threat- ening or fatality-threatening doses. This methodology is also used in this analysis.

1 1

100 - l l

l

-_ _ _____________. A

)

2. Emergency Planning Considerations

a. History and Introduction

) A major factor in planning for emergency actions in response to an accident at a nuclear power plant is the size of the plume exposure emergency planning zone. Probabi-listic assessments of the offsite consequences from core melt accidents, based on an 3 assumption of no timely emergency response, have been used to aid in determining a generic zone size. These assessments provided estimates of the probability of exceeding selected doses as a function of distance from the plant, given a core melt accident.

An assessment of this kind was performed in 1978 by an interagency task force (U.S.

Nuclear Regulatory Commission and U.S.

Environmental Protection Aaency) as part of the establishment of the planning basis for state and local radiological emergency response plans for accidents at nuclear power reactors and was documented in NUREG-0396 (Reference IVE-1). For this planning basis analysis, the releases of radioactive material from nuclear power plant accidents, and the associated probabilities of those releases, were based on the 1975 Reactor Safety Study (RSS) (Reference IVE-2). The results of the analysis were used to support a recommendation of ten (10) miles as the radius of the generic plume exposure pathway EPZ. This recommen-dation was documented in the Code of Federal Regulations, Title 10, Part 50.47, and was incorporated in NUREG-0654 (Reference IVE-3).

The present work considers a comparable assessment specific to the Shoreham Nuclear Power Station for operation at full power.

The effect of the SCS (Reference IVE.4) was evaluated by performing separate assessments with and without the SCS.

b. Objectives and Considerations The objective of emergency response plans is to provide dose savings (and in some cases immediate life saving measures) for a spectrum of accidents that, without protective action, could produce offsite doses in excess of Protective Action Guides (PAG - the projected radiation dose to individuals that warrants protective action in the event of an accident.

Reference IVE-5).

101 -

. _ _ _ _ _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ m

)

The size of the EPZ is a key factor in emergency planning. This generic distance from the plant to the boundary of the EPZ was determined partly on the basis of analyses of the probability of exceeding certain doses

' assuming no emergency response for a period of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following plume passage) as a function of distance from the plant and partly upon some considerations which were not

, quantitative in nature.

J Although the reasoning used to select 10 miles as the plume exposure pathway EPZ did not include, as an initial step, the determination of clear quantitative criteria, it did incorporate observations drawn from quentitative probabilistic assessments of reactor accident consequences. The observations made in NUREG-0396 are as follows:

1. The "... higher PAG plume exposures of 25 rem (thyroid) and 5 rem (whole body) would not be exceeded (for design basis accidents) beyond...(the EPZ)...for any site analyzed" (p. I-5).

2. The "... doses from ' melt-through' (less severe core melt) releases... generally would not exceed even the most restrictive PAG beyond. . . (the EPZ) " (p. I-6). "Less than 30% of all core melt accidents would result in high exposure outside the recommended planning distances" (p. I-9).

3. "As can be seen from (NUREG-0396) figure I-ll (probability of exceeding dose versus distance, conditional on core melt) , core melt accidents can be severe, but the probability of large doses drops off substantially at about...(the EPZ boundary)" (p. I-37).

4. "The Task Force had to decide whether to place reliance on general emergency plans for coping with the events of Class 9 accidents for emergency planning purposes, or whether to recommend developing specific plans and organizational capa-bilities to contend with such accidents.

The Task Force believes that it is not appropriate to develop specific plans for the most severe and most improbable Class 9 events. The Task Force, however, does believe that consideration should be given

- 102 -

I

)

o to the characteristics of Class 9 events in judging whether emergency plans based primarily on smaller accidents can be expanded to cope with larger events. This g is a means of providing flexibility of response capability and at the same time giving reasonable assurance that some capability exists to minimize the impacts of even the most severe accidents" (p. III-3).

A more concise statement of these consider- ,

ations which led to the selection of 10 miles

's the plume exposure EPZ was included in NUREG-0654 (Reference IVE-3). This was inte".ded to summnrize the planning basis work

) reported in NUREG-0396 as follows:

1. "Projected doses from the traditional design basis accidents would not exceed Protective Action Guide levels outside the zone" (p. 12).

2. "Projected doses from most core melt sequences would not exceed Protective Action Guide levels outside the zone" (p.12).

3. "For the worst core melt sequences, immediate life threatening doses would generally not occur outside the zone" (p.12).

4. "Detailed planning within...(the EPZ)...would provide a substantial base for expansion of response efforts in the event that this proved necessary" (p. 12).

The assessment described in Sections 3 through 5 is of use for the second and third consider-ations in the lists above. The first consideration for Design Basis Accidents is discussed in Section 6.

3. Methodology For Calculatirn of Doses

a. General Description of Methodology The assumptions made in the NUREG-0396 analysis for purposes of calculating radiation doses were used for this study and are provided here. It was assumed that the general public was exposed to airborne radioactive material for the entire duration of plume passage and to radioactive material 103 -

O deposited on the ground for a period of 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following plume passage. The doses computed from exposure to radioactive material in the plume include direct radiation from I that source and the committed doses from inhalation of radioactive material during plume passage. Doses from direct radiation from radioactive materials in the plume or on the ground were reduced to account for the fact that people would be indoors during part

, of the exposure period and that doses would be reduced somewhat by shielding afforded by residential structures. The shielding factor applied was 0.75 for exposure to the plume and 0.33 for exposure to material on the ground.

These factors are appropriate for structures in the Snoreham locale.

The detailed probabilistic analysis of the potential accident doses that was performed makes use of site-specific data on meteorology, demography, topography, and the protection afforded by shielding of local structures.

An accidint resulting in a release is assumed to occur at a random time. The most important factors affecting variation in accident consequences are related to the weather conditions at the time of release and during dispersion of the released material. Wind speed, wind direction, atmospheric mixing, and precipitation determine the area affected and the concentrations of airborne and deposited radioactive material at a given location. For a particular release, one hundred or more meteorological scenarios are simulated to determine the probability distribution for each effect studied. Meteorological conditions are selected from a Shoreham specific data base for corresponding randomly selected start times. Sequential hourly data are used to calculate trajectory changes and concentration changes during transport.

Accident event sequences which would lead to similar releases of radioactive material are grouped into a single release category. This permits the use of a small number of release categories to represent the entire spectrum of all possible accident sequences. The offsite consequence analysis is performed separately for each release category.

- 104 -

D After the completion of the consequence analysis, a probability distribution for the dose exceeding a specified level, as a function of distance from the plant, is O computed for each release category. This conditional risk distribution is then weighted by the frequency of occurrence for the particular release category to define an absolute risk distribution. The total absolute risk from the plant is the sum of the

, absolute risk distributions over all release categories. In this study, risk is then made conditional on the occurrence of a core melt accident. This is the total absolute risk divided by the frequency of occurrence of a core melt accident.

The computer code used in these studies is the CRAC (Calculation of Reactor Accident Conse-quences) derivative code CRAC2 (see Reference IVE-6, Chapter 9). The 1983 Shoreham PRA utilized the CRACIT Code which is also derived I from the original CRAC Code. The CRAC2 code includes treatment of the effects of plume rise, wet and dry deposition, and changes in meteorological conditions (except wind direction) during plume transport. It also allows simulation of the impact of evacuation and other mitigative measures and it allows modeling of doses and health effects from both early and chronic phases of exposure. The CRACIT program incorporatec additional features such as the effect of variable wind direction trajectories, which enables more realistic simulation of site conditions, and the effects of large water bodies near the site on plume trajectory.

The major modeling difference between CRAC2 and CRACIT results is the treatment of doses calculated for uninhabited locations, such as locations over Long Island Sound. In calculating dose-versus-distance probability distributions CPAC2 includes doses for uninhabited locations in the distribution.

The CRACIT program calculates doses only for the populated locations, and thus excludes those cases. The CRACIT approach is more realistic because protective actions in typical emergency response plans would clearly not be needed in the over-water areas.

In the present assessment, the significance of this modeling difference is isolated by showing, in addition to the CRAC2 results

- 105 -

l

- - - _ _ _ _ _ _ _ _ __ n

l 1

O computed directly, CRAC2 results modified to eliminate those simulations in which the straight-line plume trajectory would pass over Long Island Sound, by multiplication by 0.5.

O An earlier study (Reference IVE-7) has shown that CRAC2 modified results are equal to or slightly conservative relative to the CRACIT rc.sults.

b. Input Data and As~sumptions D

Input data related to release magnitude and frequency were drawn from those described in Sections, IVC and IVD. The frequency for each release category, with and without the SCS system is shown in Table IV.E-1. The event D sequences chosen to be the representative for the release categories with and without SCS are also identified in the table. The release characteristics of the representative event sequences are summarized in Table IV.E-2.

> The data in Table IV.E-2 must be translated into a form suitable for input to the CRAC2 code. For example, release fractions for cesium iodide or cesium hydroxide must be translated to separate release fractions for cesium and iodine. Furthermore, only one

> release time and duration can be input to CRAC2 for a given release category. Generally releases of all isotope groups begin at the same time and continue for the same duration.

Whenever required, release start times and durations were conservatively chosen to

) maximize the calculated doses. The data from Table IV.E-2, translated for input into the CRAC2, are shown in Table IV.E-3.

It should be noted that "Time to Release" and "Release Duration" values in Tables IV.E-2 and j IV.E-3 differ from corresponding values in I Tables IV.D-1 and IV.D-2. These differences i are mostly due to CRAC2 modeling requirements.

CRAC2 does not allow ramp changes in release rates; as a result, "step releases" were input into CRAC2 which conservatively model the actual MAAP release results. This is done by initiating the release for CRAC2 at the time when the MAAP calculated release fractions become significant.

106 -

D

4. Treatment of Ground Level and Elevated Releases All releases from the reactor building have been modeled as ground-level releases in the wake of 3 the power plant complex.

The DCOP configuration of the SCS system is designed to provide momentum-driven plume rise to vented releases. An ef fective release height is determined by adding the calculated plume rise to O the release point elevation (50 meters) and subtraction of the maximum elevation of terrain in the vicinity of the Shoreham site (61 meters).

Plume rise is primarily a function of gas exit velocity multiplied by stack diameter (Wo*d), wind speed and stability class grouping (stable vs.

I unstable and neutral) (Reference IVE-10). Utili-zation of Shoreham Joint Frequency Distributions Tables (Reference IVE-8, Section 2.3) , allows one to obtain effective release height distributions, for nitrogen and steam releases. Imposition of a requirement to meet a 40 m. effective height 2at L to Wo*d of 115.5 m /sec the for 95th nitrogen percentile leady and 142.7 m /sec for steam.

Utilizing these values and choosing, conserva-tively, among cumulative probability distributions for stability class groupings, allows for the choice of a realistic ef fective height of 60 meters representing the 70 to 90 percentile range. Both 95 percentile and 70-90 percentile heights are achieved for all three stability class groupings (unstable / neutral, stable and all classes together) . As a sensitivity study, releases from the SCS system were also modeled as ground releases in the wake of the structure housing the system.

It should be noted that the primary source of meteorological data was that for a one-year period from January, 1974 through December, 1974 collected at the 400 ft. SNPS tower. For assess-ment of the effect of plume rise for releases from the SCS system, wind speeds representative of the 150 ft. elevation were used and adjusted to the point of release.

5. Dose Versus Distance Risk Distributions Results of the study are shown in Figures IV.E-1 through IV.E-12 in two sets of six figures each.

The sets represent Shoreham without-SCS (Figures IV.E-1 through IV.E-6 labeled "NO SCS"), and Shoreham with SCS with releases from the SCS system treated as elevated at an ef fective release height of 60 meters (Figures IV.E-7 through IV.E-12 labeled "SCS 60 M").

- 107 -

O Each of the figures shows the dose-versus-distance risk distribution results from CRAC2 and from CRAC2 results modified (which eliminates receptor

, locations over water). Also shown is the compar-able dose-versus-distance probability distribution from NUREG-0396 (curves labeled "NRC"). Each curve represents a dose-versus-distance risk distribution integrated over all core melt accidents for one of the doses of interest

-, described in NUREG-0396. Because of the low dose probability results at Shoreham, special CRAC2 runs were made to define the distributions at distances within one mile. NUREG-0396 results within a mile are unavailable. The first four figures in each set are the results for whole body doses of 1,5, 50, and 200 rem, respectively. The last two figures in each set are the results for thyroid doses of 5 and 25 rem, respectively.

Two methods are used to summarize the risk reduction indicated by the figures. The first method is consistent with NUREG-0396 consider-ations 2 and 3 for establishment of an EPZ radius.

These results are summarized in Table IV.E-4 for the CRAC2 results. The table provides estimated radii, based on the considerations when CRAC2 results are modified so that plumes over Long Island Sound are not considered to provide radiation doses to individuals. Doses unmodified for the overwater correction are also presented in parentheses.

Risk reduction comparisons utilizing considera-tions 2 and 3 of NUPEG-0396 are conducted relative to the 10 mile radius. The without-SCS results indicate a radius of 0.9 mile. The with-SCS results show an estimate of less than 0.3 mile for the radius. Both cases are significant improvements over the NUREG-0396 value.

The second method compares the advantage obtained in the dose versus-distance risk curves for the without-SCS case as compared uith the NUREG-0396 curves (NRC Case) and the with-SCS case as ccmpared with the without-SCS case at different distances from the site. These are shown in Table IV.E-5.

Table IV.E-5 clearly shows that the Shoreham without-SCS case provides significant benefits relative to the NUREG-0396 case. The with-SCS case, in turn, provides significant benefits relative to the without-SCS case. The table shows that the with-SCS improvement over NUREG-0396 ranges from about one order of magnitude to more 108 -

than two orders of magnitude (for doses at the distances indicated).

6. DBA Analysis With and Without SCS As required by NRC regulation, the design of a nuclear power plant must be such that the consequences resulting from design basis accidents are below the guidelines of 10 C.F.R. Part 100.

Traditionally, the worst case design basis

'. accident analyzed has been the loss of coolant accident, since it results in the largest offsite doses of any design basis accident. Thus, the LOCA was the design basis accident considered specifically by NUREG-0396 in discussions of the rationale for the generic 10-mile EPZ.

g For Shoreham, analyses were performed to develop an understanding of the conservatism of the 10-mile EPZ with respect to the consequences of design basis accidents from full-power operation.

Using the Chapter 15 accident analysis from the

} Shoreham Updated Safety Analysis Report (USAR)

(Reference IVE-8), results showed that four design basis events had the highest potential for offsite radiological consequences. All of the remaining 34 USAR licensing basis transients and accidents which were considered would be bounded by these

) four events. The four bounding events included:

o Control Rod Drop Accident; o Steam Line Break Accident; o DBA-Loss of Coolant Accident; 1

o Fuel Handling Accident.

DBA radiological results are not affected by installation of the SCS since the SCS is designed not to be initiated under design basis conditions.

The results of the analyses of these four events, expressed in terms of dose-versus-distance, are depicted la Table IV.E-6. Because the USAR analysis assumptions were established prior to the development of the NRC Standard Review Plan (SRP),

two types of results have been included in the table for the design basis events. Those results labeled USAR stere developed using USAR methodology and those results labeled SRP were developed by incorporating the latest SRP assumptions.

For the design basis accidents included in Table IV.E-6, results are provided for dose versus 103 -

O distance calculations using the extrapolation methodssuggestedbyNUREG-0{9gbasedona geometric attenuation of 1/R . The table q,

provides the distances at which whole body dose is equal to one or five rem or thyroid dose is equal to five or twenty-five rem.

The table shows that the steamline break accident (SLBA) produces thyroid doses slightly larger than D

those for the DBA-LOCA. The SLBA would not exceed two-hour doses equivalent to a lower PAG beyond one mile, with one exception. The steam line break analysis, using the peak Technical Specification assumptions for reactor water Iodine-131 concentration of 4 uCi/gm, predicts two-hour thyroid doses above 5 rem out to about 1.2 miles from the reactor centerline. However, the SRP analysis based on I-131 concentration for unrestricted operation shows that a dose exceeding 5 rem thyroid does not occur beyond the Exclusion Area Boundary. The result obtained for SLBA at 4

, uCi/gm is not significant in applying the DBA rationale in this assessment because: (1) 30% of plants studied in NUREG-0396 had projected doses above the lower thyroid PAG level at the generic 10-mile EPZ; (2) maximum I-131 concentration is permitted by Technical Specifications for only a short period of time; and (3) doses exceeding the 25 rem upper thyroid PAG do not occur beyond 0.4 mile from the plant.

It is, therefore, concluded that the DBA rationale clearly would be satisfied at less than one mile for Shoreham operating at full power.

7. Risk Characterization Using Radius of Injury-Threatening and Fatality-Threatening Doses Another methodology for assessment of risk focuses on the risk of injury and risk of fatality.

Described in a recent issue of Nuclear Safety (Reference IVE-9), this approach consists of predictions of the mean radius out to which injury-threatening and fatality-threatening doses might be seen as a function of source term magnitude (with and without SCS). It is sufficiently different from the NUREG-0396 risk assessment methodology to constitute an independent verification that beyond a fraction of a mile, the risk associated with Shoreham operation at full power is very low.

This approach is based on Figures IV.E-13 and IV.E-14. These figures show the upper bound mean radius at which injury-threatening or fatality-

- 110 -

O threatening doses might be experienced, given a release amount determined by the average volatile fission product relelse fraction. An injury-threatening dose is defined as 55 Rem (whole body), which is the threshold for prodromal vomiting syndrome as defined in the Reactor Safety Study. A fatality threatening dose is defined as 320 rem (whole body), which is the threshold for early fatalities. The injury-threatening or y

fatality-threatening dose is relevant given that a primary goal of emergency planning is to prevent injury or fatality.

The upper bound mean radius for injury-threatening or fatality-threatening doses, as shown in Figures IV.E-13 and IV.E-14 was determined as a function of average source term magnitude. More particularly, fission product source terms of various magnitudes were used in conjunction with CRAC2 to determine the upper bound radius, averaged over the full spectrum of weather conditions, at which early injury (55 Rem) or early fatality (320 Rem) might occur.

a. Application to Shoreham for the Without-SCS Case Figure IV.E-13 shows pessimistic upper bounds on the fatality-threatening and injury-threatening radii. Figure IV.E-13 also shows the average I, Cs and Te release fraction for the SNPS release categories without the SCS.

The principal conclusions are as follows:

(1) Release categories contributing 84% of the core melt frequency lie to the left of the upper bounds on Figure IV.E-13.

Therefore, these categories can never lead to radiation doses that might cause injuries or fatalities beyond the site boundary.

(2) For both RC8 and RC9, the estimated start of release is very long (39 hours4.513889e-4 days <br />0.0108 hours <br />6.448413e-5 weeks <br />1.48395e-5 months <br />); so long that even people who remain for more than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> in the open before evacuating (after the event has started) will not be exposed to the cloud. Furthermore, the long time before release means that there has been significant decay of important radionuclides before they have been released. Therefore, the upper bound injury-threatening radii appropriate for RC8 and RC9 are very small, certainly not more than 0.1 mile.

- 111 -

O (3) The only release categories that could conceivably lead to life threatening or injury threatening dose beyond the site boundary are RCl, RC2, RC4, RCS and RC6, which contribute a total of 3.5%

to the core melt frequency, or an absolute frequgncy value of-6 (0.035*4.2x10 ) = 1.5 x 10 /yr.

Thus, even with the extreme conserva-tisms that are implicit in the upper D bounds of Figure IV.E-13, there is only one chance in seven hundred thousand per year that a release could occur that has the potential to cause life or injury threatening doses beyond the site boundary.

3 By way of comparison, from Figures IV.E-3 and IV.E-4, the traditional NUREG-0396 method of analysis (for Shoreham conditions) shows that the probability, conditional on core melt, of causing life or injury threatening doses beyond the site boundary is about 19%. This is higher than the percen-tage quoted above (3.5%). However, the NUREG-0396 method of analysis is even i

more conservative than the present risk based analysis because it includes RC8 and RC9. The NUREG-0396 type analysis accumulates dose for these release categories, as if people continue their normal activities until the release starts (39 hours4.513889e-4 days <br />0.0108 hours <br />6.448413e-5 weeks <br />1.48395e-5 months <br />), without evacuating the EPZ. As noted above under 2, it is much more reasonable to eliminate RC8 and RC9 as plausible contributors to life or injury threatening doses within the EPZ. In this case, the NUREG-0396 probability falls to 8.9%, within a factor of 2.5 of that predicted by the mean radius approach.

(4) Focusing attention on a distance of one mile (because the NUREG-0396/DBA approach enables one to say that one i mile or less is a reasonable choice for the radius of the EPZ), the release categories that could cause injury threatening doses to extend beyond one mile are the same as those discussed in 3 above for the site boundary.

Therefore, the chance that prompt emergency actions would be needed to prevent injury threatening doses from

- 112 -

Q being accumulated beyond one mile is very small, one in seven hundred thousand per year.

O The release categories that could cause injury threatening doses to extend beyond about 3 miles are RCl, RC2, RC4 and RCS, contributing-a total of 2.4%

of the core melt frequency or 6

1.0xE /yr. Thus, once in one million D years, a release could occur that can have the potential, in extremely unfavorable conditions, for injury-threatening doses to extend beyond 3 miles. Figure IV.E-3 shows a y

comparable 6.5%, of which about 41 is due to RC8 and RC9. If these are eliminated, the percentages predicted by this and the NUREG 0396 methodology are very close, y

b. Application to Shoreham for the With-SCS Case Figure IV.E-14 is similar to Figure IV.E-13, except that the release categories are now those for the case with-SCS. The principal differences are that RC8 and RC9 have all but disappeared, and the more severe categories (RCl, RC2) comprise only 0.3% of the core-melt frequency. The principal conclusions are as follows:

(1) Release categories contributing about 98% of the core-melt frequency (compared to 85% Without-SCS) lie to the left of the upper bounds on Figure IV.E-14. These categories will not lead to radiation doses that might cause injuries or fatalities beyond the site boundary.

(2) The only release categories that could conceivably lead to life-threatening or injury-threatening radiation doses beyond the site boundary are RCl, RC2, RC4, RCS and RC6, which contribute 1.5%

of the total core melt frequency, or an absolute frequency value of 7

5.42x10 /yr. Thus, with SCS, a conservative upper bound on the frequency with which it will be necessary to act promptly to avoid life or injury threatening radiation doses beyond the site boundary is about one in two million per year. Note that the

- 113 -

O corresponding frequency derived from the traditional NUREG-0396 methodology is 4% of core-melt frequency, (see g

Figures IV.E-9 and IV.E-10) (compared to 19% without-SCS).

(3) Similarly, the chance that prompt emergency actions will be required beyond one mile is very small, indeed, about one in two million per vear. The 3 release categories that could conceivably cause injury threatening doses beyond about 3 miles (RCl, RC2 and RC5) to the core contribute melt frequency 0.3}/yr). Thus, (1.2x10 with SCS, a conservative upper bound on the frequency with which prompt emergency measures would be needed to avoid injury threatening doses beyond 3 miles is about one chance in eight million per year. From Figure IV.E-9, the corresponding number is about 1.5%

of the core melt frequency.

c. Conclusion The approach of Reference IVE-9 has been used to show that, with or without SCS, it is extremely unlikely that prompt emergency actions will be needed outside the site boundary to avoid the accumulation of fatality-threatening or injury-threatening radiation doses, and even less likely that actions will be recuired outside 3 miles. The approach of Reference IVE-9 explicitly takes into account a range of uncertainties up to extraordinarily pessimistic assumptions. Even with these conservatisms, the predicted probabilities are lower than those predicted for Shoreham by the traditional NUREG-0396 methodology. This gives confidence that the conclusions obtained based on NUREG-0396 methodology are safely conservative.

8. Extended Times for Release Start and Release Duration It is also instructive to consider the specific results of these Shoreham analyses for full power operation in a somewhat different context. For example, the time available in which to take protective action is not directly addressed in the NUREG-0396 rationale discussed in Section IV.E.2.

Clearly, however, the time until the start of a

- 114 -

D Clearly, however, the time until the start of a release and the duration of a release are of interest in determining the adequacy of emergency g planning.

The planning basis of NUREG-0396, in considering '

the range of times in which to implement protective actions, considered the start of release to be as early as one-half hour after the initiating event (using source terms from the

) Reactor Safety Study). For the spectrum of dominant core melt accident sequences for Shoreham at full power, the releases are generally well delayed. In fact, no containment failure (i.e. <

radiation releases are due to leakage only) is g predicted to-occur for the Without-SCS case for 71% of the core-melt frequency and for the With-SCS case for 82% of the core melt frequency.

For about 92% of the accident spectrum, releases to the environment would not be expected to start before one full day after the initiating event (for both without-SCS and with-SCS cases). Indeed p

93% of the without-SCS accident spectrum and 95%

of with-SCS accident spectrum would not be expected to provide releases to the environment before about 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> after the initiating event.

The major advantage of with-SCS is seen when

,' considering the mean time to release start (for i both plant configurations) for releases beginning less than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after event initiation.

Without-SCS the mean time is 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. With-SCS the mean time is 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br />. The increase in mean time is 5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, a factor greater than 2, A similar advantage occurs for release duration for f releases started before 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after accident initiation. The with-SCS mean duration du 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> as compared with a 4 hour4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> mean duration without-SCS. The increase in mean duration is 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />. Extended duration allows greater time for i isotope decay and for plume meander effects to decreane offsite doses. It can also be concluded that in only a few percent of the accident spectrum (5% for with-SCS, 7% without-SCS) are ,

release start times less than about 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br />. The l duration of release is short (less than 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />) la only 5.3% of the without-SCS and 3.9% of the with-SCS accident spectra, for releases starting less than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after event initiation.

9. Land Contamination Benefits A major reason for the Swedish development of the  !

SCS was to minimize land contamination in the event of a nuclear accident. LILCO has also adopted this goal as a primary reason for i installing SCS at Shoreham.

l - 115 -

O It is duly noted in the literature that the cesium isotopes would be the major contributors to long term land contamination (particularly at Shoreham since cesium is, by far, the largest of all C) contributors to long half-life particulate radioactivity emitted during accident sequences (See Table IV.E-3). It is to be expected that both whole body dose due to particulate deposition and land decontamination costs would be roughly proportional to the fraction of cesium inventory 3 emitted during an accident (See Ref. IVE-9). An estimate of the benefit of with-SCS relative to without-SCS, with respect to land contamination in an accident can be obtained by taking the ratio of the cesium fraction emitted in the without-SCS case to that in the with-SCS case. (Since the D long half life isotopes are of interest, radio-active decay will not play an important role.)

The frequency-weighted results demonstrate a land contamination benefit with-SCS of a factor of 9.6 relative to without-SCS (with frequencies obtained from Table IV.E-1 and cesium fractions from Table

) IV.E-2). This factor provides a further measure of the benefits of SCS in coping with severe accidents at Shoreham.

10. Conclusions

} The following is a summary of the offsite dose results presented in this section,

a. Analysis of bounding design basis events (control rod drop, steam line break, loss of coolant, and fuel handling accidents)

/

demonstrates that 2-hour doses for Shoreham I would not exceed EPA PAG limits at distances greater than 1.0 mile from the reactor centerline.

l

b. For severe accidents, most core melts at Shoreham without the SCS would not result in doses to populated areas in excess of the EPA plume exposure pathway Protective Action Guide (PAG) doses at distances beyond about 0.7 mile. With SCS, such severe accidents would not exceed plume exposure pathway PAG doses beyond 0.3 mile. For severe accidents without the SCS, the probability given core melt of exceeding a 200 rem whole body dose at 0.9 mile is equal to that in NUREG 0396 at 10 miles. With SCS, the probability given core melt of exceeding a 200 rem whole body dose is further reduced to an equivalent distance of less than 0.1 mile.

116 -

O

c. Using an alternative approach, it is demon-strated that with- or without-SCS, it is extremely unlikely that prompt emergency actions will be needed outside the site gp boundary to avoid the accumulation of injury-or fatality-threatening doses. This gives confidence that the conclusions obtained based on NUREG-0396 methodology are safely conservative.

9 d. A significant benefit from installing SCS at Shoreham is that without-SCS, the early release events (releases beginning prior to 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) have a weighted average time to start of release of about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br />. With-SCS, this average time to start of release increases to y about 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br />. This gives increased time for decay of fission products, as well as increased time to take emergency protective measures if required. Similarly, the duration of release for these early release events is about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> without-SCS, and about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />

, with-SCS. This gives greater time for fission product decay, and also increased probability of greater dispersion due to plume meander and wind direction changes,

e. Finally, a benefit for potential land

> contamination by long-lived cesium isotopes has been determined. The frequency-weighted average quantity of cesium released is about 10 times lower with-SCS than without-SCS.

This would result in proportionately lower deposition doses to the public, as well as lower post-accident cleanup costs.

117 -

O REFERENCES O IVE-1 "Planning Basis for the Development of State and Local Government Radiological Emergency Response Plan in Support of Light Water Nuclear Power Plants," NUREG-0396 and EPA 520/1-78-016, U.S. Nuclear Regulatory Commission and U.S.

Environmental Protection Agency, December, 1978.

3 IVE-2 "Reactor Safety Study: An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants," WASH-1400 (NUREG 75-014), U.S. Nuclear Regulatory Commission, October, 1975.

IVE-3 "Criteria for Preparation and Evaluation of Radiological p Emergency Response Plans and Preparedness in Support of Nuclear Power Plants," NUREG-0654, Rev. 1 and FEMA-REP-1, Rev. 1, U.S. Nuclear Regulatory Commission and U.S.

Federal Emergency Management Agency, November 1980.

IVE-4 Engineering Service Scope of Work (ESSOW) for Engineering, 6

Design and Performed Certification of Shoreham FILTRA Unit", Vol. 1, Stone & Webster Engineering Corp., ESSOM No. SCS-1, Rev. O, July 24, 1987.

IVE-5 "Manual of Protective Action Guides and Protective Actions for Nuclear Incidents," EPA-520/1-75-001, U.S.

t Environmental Protection Agency, September, 1975.

IVE-6 "PRA Procedures Guide, A guide to the Performance of Probabilistic Risk Assessments for Nuclear Power Plants,"

NUREG/CR-2300, U.S. Nuclear Regulatory Commission, January 1983.

IVE-7 Shoreham 25% PRA "Request for Authorization to Increase Power to 25%", LILCO Request to NRC, April 14, 1987.

IVE-8 Shoreham Updated Safety Analysis Reports Rev. 1, December 1987.

IVE-9 G.D. Kaiser, "Implications of Reduced Source Terms for Ex-Plant Consequence Modeling and Emergency Planning,"

Nuclear Safety 27, 369-383 (July-Sept. 1986)

IVE-10 "Atmospheric Science and Power Production," Chapter 8, DOE / TIC 27601, 1984

- 118 -

-. c, 1

TN2E IV.E-l RHEASE CAllIT1FY PJ]Rf5HffATIVE SHJMUS, FRHJt2K'IF3 ND O'ARN!!TRISTKS

'O FilJ1% LT111 FILTt^.

FIITASE RFT 10 Ull&R RDIASE RDEASE FRFIKUKX % RFP fC Ull&R RHEASE RHEASE FRFIt1FTIN  %

STAKT IXR (l/YR) m STI) STAKT IXR (l/YR) m CA1111HY SiD (a) 03) (c) (d) (e) (f) (a) (b) (c) (d) (e) (f) 0.4 2 2 1.9E-07 0.5% IVA.lA 1 0.4 2 2 4.lFrO8 0.1%

101 IVA.lA I 0.2 2 2 2. lE-07 0.5% IVA.lB 1 0.2 2 2 5.3E-08 0.1%

IC2 IVA.lBI 1 l

I (g) 2 2 1.7E-06 4.0% IVA.IB-H I (g) 2 2 3.7Fr07 1.0%

IC3 IVA.1IH1 l

1 0.07 10 4.8Fr07 1.1% IIID.1 1 0.07 1 10 3.7E-07 1.0%

l IC4 IIID.1 1 1 0.1%

2 1.3Fr07 0.3% SIllCSA 1 0.1 16 2 2.4E-08 FES SHICSA I 0.1 16 8 4.8Fr07 1.1% 0.7 0.04 14 8 5. lfrO8 0.1%

IdYi V)1QIC6A 0.7 0.04 14 "II)IQlCSA 0.002 9 8 2.5Fr07 0.6% R U C7A 1 (g) 7 1 9.4E-07 2.6%

IC7 BfMC7 1 2 1.2E-4B 0.03%

RfB IIA.] 1 0.1 39 2 3.2E-07 0.8% IIA.1 1 0.1 39 2 4.5Fr06 10.7% IIA.1 1 0.1 39 2 5.4Fr08 0.1%

10 IIA.I 1 0.1 39 I (g) 39 2 3.9Fr06 9.2% IIA.lM2 1 (g) 50 2 3.4E-06 9.3%

1010 IIA.lMI

- 0.0EMD 0.0% BPMCI1 0.3 (g) 15 20 1. IE-06 3.0%

FCI1 - - -

20 3.00E-05 71.1% 'R11XPCl2 0.01 (g) 25 20 3.0E-05 82.4%

IC12 BJIXICl2 0.01 3.0Fr05 25 4.2Fr05 100.0% 3.6E-05 100.0%

TTTM.

(a) ft H si,le gas release fraction (ii) Ull&R-releare fractim for otlier frantoges (c) RUTA 9E START--time letwen initiatirg event susi start of release (lir)

(d) FJ3FASE IIR-tlie release duratim (hr)

(e) FRH)!!IIX-tle release freqviency (per year)

(1) Im-tlie percent of core nelts represented ly tie release category (g) release fraction 1.0E-5

- 119 -

U O TAIME IV.E-2 MMP RHEACE IWTA 1 URAL RHERE 10 DN!!1tNUE 1 .iCIDDTP SHjt1RIE IVA.lA IVA.IB IVA.llfi IIID.1 SitIC5A 'IDlyRGA 'It)tNRC7 IWIC7A IVltDmC3 IIA.1 IIA.lMI IIA.lM2 T13MCll "Rjt!mCl2 El] E R E l'/E # 5'IT E 16.0 14.0 9.0 7.0 25.0 39.0 39.0 50.0 15.0 25.0 Tl!E 1D RUIRE (IR) 2.0 2.0 2.0 1.0 11.0 12.0 6.0 1.0 5.0 2.0 2.0 2.0 20.0 20.0 ta; R1]ENE IIR (IR) 2.0 2.0 2.0 10.0 2.3 8.0 30.0 1.0 5.0 4.0 2.0 2.0 20.0 20.0 Est RHIEE IIR (IR) 1.0 10.0 2.0 10.0 131FASE f1W.TiltG 7.0E-01 1.0E400 1.09 0) 7.0Fr01 1.0B00 1.0900 1.0900 8.0E-Ol 1.0 fro 2 IIMUE CAS 1.0B00 1.0EiOO 1.0B00 1.0ERO 1.0D00 4.0FAJ2 2.Tr03 1.0FAS 2.7Fr02 9.0E-02 1.Tr05 1.0E-05 3.TM5 3.0Fr05 l 135I111 1(UIIE 4.0FM)1 2.0Frol 1.0E OS 7.0FAI2 1.0E-01 i

1.0Fr05 1.0E-05 1.0 fro 5 1.0Fr05 1.0FA15 1.0Fr05 1.0FrOS 1.0Fr05 1.0Fr05 1DJJRIlli 1.0FM)5 1.0FM)5 1.0FM)5 3.0Fr05 1.0Fr05 l 1.0Fr05 1.0E-05 1.0Fr05 1.0E-05 S114Tirit!! 2.0E-03 1.0Fr03 1. Tr05 1.8FrC3 3.5FM % 8.TMM 1.Tr05 1.TM5 7.0FA% 6.0FA%

l 3.Or'-05 2.0Fr05 1.Tr05 2.0Fr05 2.0E-04 1.0FA)5 1.0E-05 1.0E-05 1.0E-05 13m!INIlti 6.0E4% 3.2FA % l.0Fr05 5.0FMM 1.0FMM 4.Tr02 2.0E-03 1.0FM5 3.0Fr02 1.T-01 1.0FM5 1.0FM15 4.0FM5 4.0FM)5 U S1111 IrilstIKii t 5.fE-01 2.3FM)1 1.0E-05 7.0FA12 2.0Fr01

- 120 -

TAl2E IV.E-3 MAP RF1D5E IATA GRA1RIID 10 (I? Nip 2EE ttitL IfGUT .

1 TRIAL RE2fASE 10 IIN115ttsNr l

l #UIDmr SI192CE IVA.1 A 1VA. I B IVA.1IN IIID.I SIMICSA 1DI(PUA 11M107 IV307A IVIUXRW IIA.1 IIA.IM1 IIA.IM2 1ISTICII KtIXRCl2 IGIIASE lYER571RS 16.0 14.0 9.0 7.0 25.0 39.0 39.0 50.0 15.0 25.0 1 Tlf t: 10 Rf]ENE (100 2.0 2.0 2.0 1.0 7.5 1.0 5.0 2.0 2.0 2.0 20.0 20.0 REIERE ItEATitti (10:) 2.0 2.0 2.0 10.0 2.0 8.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 k'Aldt!U TIM UIO 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 IGL isAT (cal /9E) 0.0 0.0 0.0 0.0 0.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 10.0 RFirASE lEl(3fr (N) 10.0 10.0 10.0 10.0 10.0 l

103frE 1RNTitiS 7.OE-Ol 1.0EMD 1.0E4(O 1.OE400 8.0 Fall 1.0 fro 2 1stEE GASFS 1.OEHX) 1.0EMO 1.0EMD 1.0EHX) 1.0E+00 7.0FM)1 1.0E400 1.0E400 0.0E40) 0.M400 0.OE400 0.0E400 0.M400 0.CERU (44ANIC lttilfE 0.0EiOO 0.0E+00 0.0EMO 0.E40) 0.0E*00 0.T400 0.0E*00 0.M+00 2.7E-02 9.0FA)2 1.0FA)5 1.0FA)5 3.0E-05 3.0FM15 111WANIC ImlfE 4.0 Fall 2.0FA)I 1.0E-05 7.0FA)2 1.0 Fall 4.0FA)2 2.0FM)3 1.0FM)5 1.E-05 3.9Fe05 2.9E-05 3.0E412 9.0E-02 1.M-05 0 31111 4.9FA)I 2.3E4]l 1.OE4)5 7.0E-02 1.9FM)1 4.0FA12 2.M-03 1.0FrOS 1.0FA)5 1.0FA)5 1.0&OS 1.0FA)5 1.0E-05 1.0E-05 11]JJEliti 1.0&OS 1.0FM)5 1.0E4)5 3.0FM)5 1.0FA)5 1.0FA)5 1.0E-05 1.0&OS 1.0FAIS 1.0E-05 1.0FA)5 7.0&4% 6.0E-04 1.0FA)5  ;

SintirIIM 2.OE-03 1.0E4D 1.0FA5 1.8FA)3 3.5&O'e 8.M4% l.0&OS 1.0FA)5 2.0FA)5 2.0FM% 1.0FA)5 1.0&O5 1.0FM15 1.0&OS

}:IT11ENIlt! 6.OE4% 3.2FM% 1.0&OS 5.0EM4 1.0FA% 3.0&OS 2.0&O5 1.0E-05 0.0Ef00 O.0E400 0.GEKX) 0.0E4al 0.Mf(D 0.0E100 IRmWAtl 0.(E400 0.0ElG) 0.0EMD 0.MHO 0.0E40) 0.0E400 0.0E400 0.UEMD G51111 - 0.0f%*Cs14 0.916*Ces11; ImtfE - Cs1

- 121 -

-' O O TABLE IV.F-4 ESTIMATED DISTANCES OBTAINED FROM NUREG-0396 CONSIDERATIONS 2 ANC 3 Based on Whole Body Dose 1 REM (30%)** 5 REM (24%)** SO REM (10%)** 200 REM (3%)**

<Cn.1 mile 0.1 mile 0.1 mile 0.9 mile WIT 110UT SCS (0.4 mile) * (1.7 mile) * (1.9 milel*

(0.4 mile)*

WITli SCS <CO.1 mile <CO.1 mile <CO.1 mile <CO.1 mile 1 (0.2 mile)* (<CO.1 mile)* (<C0.1 mile)* (<0.1 mile)

• l l

l Based on Thyroid Dose 5 REM (23%)** 25 REM (14%)**

WITilOUT SCS 0.7 mile <CO.3 mile (1.8 mile)* (9 mile)*

WIT 11 SCS <CO.3 mile <CO.1 mile (0.8 mile)* (0.3 mile)*

• Distances in parenthesis are without over water correction.

• • Percent values refer to dose probability associated with NUREG-0396 curves at a ten mile distance from the site.

- 122 -

'O O TABLE IV.E-5 COMPARISON OF DOSE EXCEEDANCE PROB ABII.I TY FOR Til E NRC AND SIIOR Ell AM (OVERWATER CORRECTED)

WITII AND WITil0UT SCS CASES AND TilEI R RATIOS AT FIXED DISTANCES Total body Dose (Rem) Thyroid Dose (Rem) 1 5 50 200 5 25 Case 2 Miles Distance 1

NRC Probability 0.580 0.440 0.155 0.0750 0.380 0.190 Without SCS Probability 0.121 0.076 0.041 0.013 0.101 0.087 j 0.024 0.015 With SCS (60 m.el.) Probability

• 0.043 0.019 0.006 0.002 l Ratio of NRC to without SCS** 4.8 5.8 3.8 5.9 3.8 2.2 Ratio of Without 2.8 4.0 7.0 6.7 4.3 5.9 to With SCS Ratio of NRC to 13.4 23.2 26.3 39.5 16.0 13.0 With SCS**

10 Miles Distance NRC Probability 0.300 0.240 0.100 0.03 0.230 0.140 Without SCS Probability

• 0.0711 0.0331 0.00095 0.00005 0.079 0.0651 With SCS (60 m.el.) Probability

• 0.0157 0.0063 0.00025 0.00005 0.0108 0.0069 Ratio of NRC to 4.2 7.3 105.3 600 2.9 2.2 Without SCS**

Ratio of Without 4.5 5.2 3.8 - 7.3 9.5 to With SCS Ratio of NRC to 19.1 38.1 400 600 21.3 20.3 SCS

• For No overwater Correction Multiply by 2.

• • For No overwater Correction Divide by 2.

l

- 123 - l

TABLE IV.E-6

~1VO HOUR DOSES AT THE ENCLUSION AREA BOUNDARY FOR DESIGN BASIS ACCIDENTS AT SHOREHAM AND THE EXTRAPOLATION DISTANCE TO OBTAIN LOWER AND UPPER PAC VHOLE BODY OR THYROID DOSES Two-Hour Dose (Rem) at Distance (miles) From Reactor Center To Obtain Design Basis Exclusion Area Boundary (EAB) Lower or Upper FAC Whole Body or Thyroid Dose Accident Whole Body Thyroid Whole Body Thyroid (For 1 Rem) (For 5 Rem) (For 5 Rem) (For 25 Rem)

Control Rod Drop 1 28 0.19 mile 0.1 mile 0.61 mile 0.21 mile Accident (CRDA)

USAR Analysis Steam Line Break 0.06 3.8 0.1 mile 0.1 mile 0.16 mile 0.1 mile Accident (SLBA)

Continuous I-131*

SRP Analysis St.BA Peak I-131** 1.2 75 0.22 mile 0.1 mile 1.18 mile 0.40 mile SRP Analysis INA 3.9 43 0.43 mile 0.15 mile 0.81 mile 0.28 mile SRP Analysis Fuei Handling 5.5 16 0.60 mile 0.21 mile 0.42 mile 0.14 mile Accident (FHA) USAR Analysis

• Technical Specifications Continuous I-131 Reactor Water Concentration limit is 0.2 uC1/gm at Shoreham.

• • Technical Specifications Peak 1-131 Reactor Water Concentration is 4 uC1/gm at Shoreham.

- 124 -

3 3

SHOREHAM 100% POWER--NO SCS e WHOLE 80DY--1 REM (PROBABILITY CCNDmONAL ON CORE MELT) 1.000 .

O-o.Os B 1 oN Os .

,'.s c% O .

^'h s

-..._,~. gO '

A ~'te, D

g o,joo ON 8'  : g,g%3 \  :

O  :  :

5 . 4 ^ \. .

> y - -

@ . A e

O

$ 0.010 - --

. A ,

'O.001

O.1 1.0 10 100 DISTANCE (MI)

O - O NRC . - . C RAC2 A- A CRAC2 MOD FIGURE F. E-l

-125-l

, i

O 3

SHOREHAM 100% POWER--NO SCS e WHOLE BODY--5 REM (PROBASILITt CCNDmONAL QN CORE MELT) 1.000 . ., .

O -o,O'O ' O '

7 'c%

'A

-.-._.*w*.~,'e %O i

@ 0.100 -

A-a_a'%s s* -

^'^- A **

@  ! As .  !

A u .

O e'

U -

A -

w . o. s\ .

8 ^*

a -

a -

ca O

\

A

$ 0.010 -

. A .

0.001 ' ' ' ' ' ' ' ' ' ' ' ' ' ' o'

O.1 1.0 10 100 DISTANCE (MI)

O - O NRC 4 - 9 CRAC2 A- A CTMC2 MOD FIG U R E '.EZ . E- 2

- 126-

e e.

J SHOREHAM 100% POWER--NO SCS l WHOLE BODY--50 REM (PROSABILITY CONDITIONAL ON CORE MELT) 1.000 .

)  : .

e ---

, N os

) O-O-O-O'O o o,joo  %~e N g% O a e

-a s  :

w ^%

'D g '. N

• d

) d . 'o -

w .

a .

M o e a .

CD O 6

) $ 0.010 _-

. 4

. \ \, .

O a\0 0.001 ' ' ' ' ' ' ' ' ' ' ' "\ ' ' ' '

O.1 1.0 10 100 DISTANCE (MI)

O - O NRC 9 - 9 CRAC2 6- A CRAC2 MOD FIGURE II. E-3

- 127 -

.O 1

e SHOREHAM 100% POWER--NO SCS

, WHOLE BODY--200 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 .

O 8

w 0.100

/s

.N

\*%** ,*O-O-o-O No A~a'As 4'.

o . s .

5 .

cb .

w . A \

Q .

$ Os * '

\

a . b o .

m O

\ \

b \

$ 0.010 .- O -

e  :

b 0.001 ' ' ' ' ' ' ' ' ' ' ' '0' ' ' ' ' ' ' ' '

O.1 1.0 10 100 DISTANCE (MI)

O-O NRC 4 - 4 CRAC2 A- A CRAC2 MOD FI GURE 23E.E- 4

-128 -

i

)

SHOREHAM 100% POWER--NO SCS

) THYROID--5 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 . ., ., .

9: N* s ,'*  :

) ^N 3 0-o

^ ~ a _,g .. '  %,  :

w ,

_,\,

) 0 0.100 -

'a-a --

e  : he a-3 0  : Na  :

x y .

, w . O en o -

a .

CD o o

E 0.010 -

0.001 O.1 1.0 10 100 DISTANCE (MI)

O - 0 NRC 9 - . CRAC2 6- A CRAC2 MOD FIGURE E. E-5

-129-

)

O O r SHOREHAM 100% POWER--NO SCS

THYR 0lD--25 REM (PROBASILITY CONDmONAL ON CORE MELT) 1.000 .

is..s 5  %

• %' ' **88 o ' e --Q-C -

a 0.100 -

. 'A h -A-A-a 3, D o***N,  :

d o

A o sN -

d . g .

w f.A 9 O -

O -

(D O -

a, 0.010 -

E D

E D

8 D

~

?

o,co, . . . . . . . . , . . . . . . . . , . . . ....X 0.1 1.0 10 100 DISTANCE (MI)

O - O NRC 9 - 9 C RAC2 6- A CRAC2 MOD FIGURE N. E- 6

-130-

I D

0

, SHOREHAM 100% POWER--SCS 60 M WHOLE BODY--1 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 .

O-o.O'OsO D -

4 ~.' s% .

cd O

^%g

.N -

I N 4,*% ' e 8 0.100 a --

x o- .,N 8

w o

N o .-

, d . 'a'as , o'e ,  :

w . s .. .

M s .A N o og -

O -

g CD 66

' 0 -

O.010 -

3 ,

! \O  !

e 0.001 O.1 1.0 10 100 DISTANCE (MI)

O--O NRC 4 -e CRAC2 6- A CRAC2 MOD FIGURE E . E 7

- 131 -

O O

m5m 8oM moqmDld0* $ r yo m @o<siw gm5 S@hG o96 Bog Oz oA tb

" .ooo - - -

2 -

~

- a OI9O fo /o I .

o .

O D o' oo ^  :

E .

D

./o/.i E .

E C " .

X I* .

E o f, E ./ b/ /, .

S b1 O lP /, .

D B

O bfp /p '4* /

R p ,

P o'o o .

i

. b ' .

. D o. .

y. .

. V, .

o'oo- . - - - _ _ . - .:- d.__mo n

P.- g mo "o 9MQzOmnEc OIO, el

• oWo" b lb ok~ rOO n5cmm %

- m.*

iMI

O-O SHOREHAM 100% POWER--SCS 60 M

g. WHOLE BODY--50 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 .

O - O-O - O s.,O ~

) @ 0.100 -

a w

O y .. .

O .

d 4!N O -

,N

) o o 0 %g he m as s i O 63 ,N

$ 0.010 -

A a e

- \

. g .

. N -

6 e .

. A -

O

' ' ' ' ' ' ' ' \' ' ' ' ' ' ' ' '

0.001 O.1 1.0 10 100 DISTANCE (MI)

O - O NRC 9 - 9 CPAC2 A- A CRAC2 MOD FI G U R E 33E . E - 9

-133-

O O

SHOREHAM 100% POWER--SCS 60 M e WHOLE 80DY--200 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 .

) . .

@ 0.100 - -

w o

O-O-o-Os os .

N -

w Q .

\

8 9 O -

8 e 0.0,0 .~.,s. \o -

m-e s Os ._0 .

sg .

\

3 .

0.001 ' ' ' ' ' ' '

O.1 1.0 10 100 DISTANCE (MI)

O - O NRC . - . CRAC2 6-6 CRAC2 MOD FIGURE N. E-l0

-134 -

d e

SHOREHAM 100% POWER--SCS 60 M e THYROID--5 REM (PROBABILITY CONOmONAL ON CORE MELT) 1.000 g 9

.N  :

d

1. s,N 0-o

.N a *s.

s O

N 'g No gDO .

3'A O 0.100 - o\g9 -

W

\\

ao s

w -

\ 9 -

w .

\

%e s .

O As o a a s Gege** e cn o s#

ct:

a. 0.010 -

a%o \, --

sg\ .

3 .

~

9 0.001 O.1 1.0 10 100 DISTANCE (MI)

O - O NRC $ - 9 CRAC2 6- A CRAC2 MOD FIGURE *E. E-l l

- 135 -

D D

SHOREHAM 100% POWER--SCS 60 M D THYR 0lD--25 REM (PROBABILITY CONDmONAL ON CORE MELT) 1.000 .

4 -

N 0-0 e

' O _

O.100 t ,

O w -' \e U

x

\A s e

N

~

, E .

w O

O A'A* ~...~O g ,

e'^-a s 3 e'**4. ,,

g N x*

0.010 c.

g%

m ,g .

A .

0.001 O.1 1.0 10 100 DISTANCE (MI)

O - O NRC e -9 CRAC2 A- A CRAC2 MOD FIGURE .E. E- 12

-136 -

a w

=8 KEY:

2 RCl,RC2, - INDICATES SNPS RELEASE CATEGORY C I I I

~] (O.5%) ETC. INDICATES CONTRIBUTION TO CORE MELT FREQUENCY N 10 '

1 o

s/

a E

8 q.

S b a ?a t,

' 8 8 s< l

1. c E

k w .6 3e 2 g g g g g gj g

• g a

s 5 -

w. -

8v1 4 x 2 8 6  : xe S 6 5

g 6

e 5 e

t e

8

/e e em e 2 0 5e a

z O.I

) I

-4 1

-3

-2 Il l Il -

l I O 10 ' 10 10 10 10 ' 10 AVERAGE OF I,Cs,Te RELEASE FRACTIONS i

CONDITIONAL MEAN RADil FOR INJURY AND FATALITY-SNPS RELEASE CATEGORIES -

{

w/o SCS FIG. IE.E-13

't 1I

. C C L

)

S E

L I

M

(

Y T

I L

A T

A F

O D N N

A Y

R 0+o U

J

+

N p 44 I

1. 4 p,0

- p l 9, i

D 09 O

A m i

R

~ )

N  %

A 9 E 2 ) )

M 1

(

0 5

1 8 2 L C ( )

R, 0 A 2  % ) ( ) )

N 7 C

l C

I %9 0 C

• /o % )

O l

1 R, R, O. .

1 I

(

l R, (0 0 0 T 3 l (

N

(

4 8 5 2

. (6 l

C C C DI R C C C C Cl R R R R R R N

O

- C e - %R - - - - - _

I E _d -o 6 iU

-O~

_O l -

-o B

m%>Om ms.O=.mjm$M  !

I h5ZW oZ2jOZ>r

- Em Z 1 .

=>9_ mO2 Zc.C gC n >#E+f M "mFMbm o> MOON mV) I (7 O ni mb M*$.

il i' ,I1l l ll

O F. Summary In a continuing effort to enhance safety and maximize gp defense-in-depth against postulated severe accidents, LILCO has implemented or plans to implement additional modifications and procedural changes to the Shoreham plant.

The significant hardware modifications since the g Shoreham 1983 PRA publication include:

o Installation of a 20 MWe black-start gas turbine onsite for auxiliary power; o Upgrading the Standby Liquid Control (SLC) system with highly enriched sodium pentaborate which provides 200% of the ATWS rule (10 CFR 50.62) SLC control capacity requirements and approximately 400% of the control capacity of the original system design; o Installation of an ADS inhibit switch; o The addition of three Colt diesels in addition to the three original diesels for emergency AC power; o Installation of a Supplemental Containment System which provides the following banefits in the unlikely event that core damage were to occur:

- Increases the effective heat capacity of the primary containment; Increases the best estimate success probability associated with operator actions to recover containment heat removal systems or initiate SLC because it increases the time which the operator has to mitigate the accident; Provides an added contairment volume of 141,235 ft.3 for storage of non-condensable gases;

- Provides scrubbing for all radioactive material except noble gases;

- Increases the time available for decay of radioactive isotopes including noble gases;

- Provides an elevated release and consequently lower doses from any radioactive releases;

-139 -

1 f

.O l

- Assists in preserving the release pathway through the suppression pool by providing a preferential primary containment vent location in the wetwell airspace; e

- Eliminates the need for operator action to open the vent for containment over-pressure protection; D

- Mitigates the potential for containment failure;

- Improves accessibility into the reactor building under venting conditions.

p The major procedural changes since the Shoreham 1983 PRA publication include the following:

o Ticorporation of the reactivity control guideline from EPG Revision 4;

> o Direction to maintain the HPCI and RCIC pump suctions from the Condensate Storage Tank (CST);

o Incorporation of containment venting utilizing existing drywell and wetwell 6 inch lines; o Inclusion of the use of the ADS inhibit switch in both ATWS and non-ATWS accident sequences.

Insights gained from the current PRA update analysis have been integrated into the desiga of the Shoreham plant. One such change is the redeeign of the containment isolation valves for the drywell equipment and floor drain lines, to assure automatic containment isolation under loss of all AC power conditions (station blackout). Another insight is the installation of the Discharge Control and Overpressure Protection system on the SCS to enhance the retention and dispersion of any radioactivity releases.

LILCO has concluded, based on the PRA, that the station modifications, procedural changes, and insights gained have a significant impact on the ability of the plant to mitigate radiological consequences and, as a result, have a dramatic benefit on the ability of Shoreham to respond to severe accidents. The major conclusions that can be drawn from this PRA update are:

o The dose-versus-distance probability distributions for Shoreham are distinctly below those in NUREG-0396 (which are, in part, used to establish emergency planning requirements) . Specifically:

-140 -

3 In the event of a core melt accident at Shoreham with the SCS, the probability of receiving a dose of 200 Rem (the dose g resulting in serious illness in the months following an accident) at a distance of five miles from the site is 200 times less than the probability given in NUREG-0396. Without SCS, the corresponding probability for such doses at Shoreham is pbout 55 times less than those At a distance of 10 miles, i in NUREG-0396.6 the corresponding probability for such doses for the plant witn and without SCS is greater than 600 times lower.

- For severe accidents, most core melts at Shoreham with SCS would not result in doses to populated areas in excess of the EPA plume exposyyedoses pathway Protective Action Guide at distances beyond about (PAG)-

three-tenths of a mile from the plant.

Without SCS, such severe accidents would not exceed EPA plume exposure pathway PAG doses beyond seven tenths of a mile from the plant.

- For severe accidents without the SCS, the probability given core melt of exceeding a 200 rem whole body dose at 0.9 mile is equal to that in NUREG 0396 at 10 miles. With SCS, the probability given core melt of exceeding a 200 rem whole body dose is further reduced to an e aivalent distance of less than 0.1 mile.

o Doses to populated areas under the USAR Design Basis Accidents will meet PAG dose standards at a distance of less than one mile.

6/ The Shoreham dose has been modified to include the effects of the plume trajectory over Long Island Sound.

7/ The Protective Action Guide (PAG) is defined as the projected dose to individuals in the general population that warrants Protective Action in the event of an accident. The projected dose is the dose that would be received from the accident if no protective actions were taken. The PAG level doses are low and range from 1 to 5 rem whole body and from 5 to 25 rem thyroid. "Manual of Protective Action Guides and Protective Actions for Nuclear Incidents", EPA-520-1-75-001 (revised June 1980).

-141 - j

O o The frequency of potential core damage conditions with SCS installed hg 1983 PRA cglculated valuebyp approximatelybeen reduced 39% to 3.6 from the g x 10 per reactor year. Without the SCS installed (and assuming that the existing containment venting capability can be implemented), the frequency of core damage conditions has been reduced from the 1983 PRA -5 calculated value by approximately 29% to 4.2x10 per reactor year.

3 o No new types of internally initiated accident sequences which could lead to core melt were identified beyond those originally discussed in the 1983 Shoreham PRA.

o A large fraction, i.e., 82%, of all core damage conditions with SCS lead to containment integrity being maintained and result in a calculated leakage rate on the order of design basis accidents. Such leakage rates have low radiological consecuences. Without SCS, this percentage of core damage conditions is 71%.

o A significant fraction, i.e., about 92%, of all potential core damage conditions (with and without SCS) are characterized by a slow accident progression which results in a release to the environment that begins after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> from the initiating event. Furthermore, for these same damage conditions, the frequency weighted average release duration is 18 hrs, with the SCS, and 16 hrs. without the SCS. These long times to release coupled with the long average release duration and the small halogen release fractions, help greatly to enhance public health and safety by increasing the effectiveness of the emergency plan for Shoreham. The accident progression of the remaining 8% of the potential core damaae conditions (which can be characterized as "early and moderately early releases") is such that the frequency weighted average release start time (after the initiating event), is about 9 hours1.041667e-4 days <br />0.0025 hours <br />1.488095e-5 weeks <br />3.4245e-6 months <br /> with the SCS, and about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> without the SCS.

The frequency weighted average release duration for the same 8% of potential core damage conditions is about 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> with the SCS and about 4 hours4.62963e-5 days <br />0.00111 hours <br />6.613757e-6 weeks <br />1.522e-6 months <br /> without the SCS.

8/ The 1983 PPA calculated value for the core melt frequgncy per reactor year without external events is 4.9 x 10 per reactor year. The 1985 MgCI report external event contribution of 9.8 x 10 per reactor year is added to this valug to give a total core melt frecuency of 5.9 x 10 per reactor year.

-142 -

P o Over the spectrum of all severe accidento, the I frequency weighted land contamination with SCS will be about 10 times lower than without 3CS.

O o Under severe accidents with the releases through the SCS, the whole body dose to the Control Room-and Technical Support Center personnel will be less than five rem with a confidence level of at least 95%; the thyroid dose will be negligible.

o Using an independent method of risk characterization, LILCO has obtained resulte similar to those found via the NUREG-0396 methodology. In fact, this evaluation has shown that it is extremely unlikely that any accident will occur at Shoreham for which prompt emergency actions will be needed beyond the site boundarv.

-143 -

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