ML20126E019

Response to Gessar Source-Term Issue 9:Suppression Pool Bypass in Bwrs

ML20126E019
Person / Time
Site:	05000447
Issue date:	04/04/1985
From:	GENERAL ELECTRIC CO.
To:
Shared Package
ML20126E006	List: ML20126E003 ML20126E019
References
FOIA-84-175, FOIA-84-A-66 NUDOCS 8506150271
Download: ML20126E019 (57)

Text

. ": ' . . ' '

~.

GENERAL ELECTRIC

. PRO.P RIETARY INFORMAT10N RESPONSE TO GESSAR SOURCE TERM ISSUE 9

$UPPRES$10N P00L EYPA55 IN SWR'S W

(

l l

l U5j61pgy1050404 g

CUNNAN64-A-66 PDH /

(f i

.  :. GENERAL ELECTRIC PROPRIETARY INFORMATION

• CONTENTS

( <

EMLt

1. INTRODUCTION 1-1 .

2. IDENTIFICATION OF BYPASS PATNS 2-1 2.1 Liquid Release Pathways (Figure 1. Path A) 2-1 2.2 Airborne Release Paths 2-2 2.2.1 Potential Release Pathways outside Secondary 22 Containment (Figure 1 Path B) 2.2.2 Potential Release Pathways Inside Secondary 2-4 Containment (Figure 1, Path C) 2.2.3 Potential Suppression Pool Bypass Path: 25 Inside Containment (Figure 1 Paths 0 and E)

3. SIGNIFICANCE OF SUPPRES$10N P0OL BYPASS 3-1 3.1 Probability of Suppression Pool Bypass 33 3.2 Bypass Flow Splits 3-3 3.3 Evaluation Results 35

4. F15510N PRODUCT RETENTION ON BYPASS PATHWAYS 41 4.1 Primary System Plateout 4-1 4.2 Building Plateout (Rain Forest) 42 4.3 Conclusion 4-3 i

5.

SUMMARY

/ CONCLUSIONS 5-1 8

111

GENERAL ELECTRIC PROPRIETARY INFORMATION

( , TABLES Tgb,1,3 Title ggge 1-1 Summary of Risks From BWR/6 PRA

• l-3 2-1 Pathways which Terminate Outside Secondary Containment 2-7 2-2 Pathways which Terminate Inside Secondary containment 2-9 2-3 Pathways which Terminate Inside Secondary Containment 2-11 3-1 Sypass Probability Evaluative Probabilities 3-7 3-2 Summary of typass Probabilities and Flow & Splits 3-8 ILLUSTRATIONS

& Ficure Title f,aje a

2-1 BWR Potential Bypass Pathways 2-13 31 Evaluation Methods 3-12 32 Release Pathway Event Tree 3-13 l

O I tv l

e. '

. 1

~

GENERAL ELECTRIC PROPRIETARY INFORMATION~

1. INTRODUCTION ,

Recent discussions between the General Electric Company and the US Nuclear Regulatory Commission and the National Laboratories havt stressed the fission product retention capability of SWR suppression pools. For events which transport the fission products to the suppression pool, this

• retention capability has been shown to reduce the inventory of fission products available for release to the environment to levels far below levels prescribed by the US Regulatory Guides. As a consequence, there are potentially significant impacts on BWR Emergency Planning, location of equipment, and Probabilistic Risk Assessments (PRA) - all supporting the extreme safety of the BWR and low consequence of potential release to the general public. Table 1-1 sumarizes the results of a recently completed PRA showing the extremely low levels of public risk (0.265 manrem/ reactor year). -

p A key concern following this assessment was the degree to which there may be pathways for fission products to be released to the environment which do not pass through (bypass) the suppression pool, although some bypass was treated by the PkA as shown on Table 1-1. The concern was that there may be pathways not addresses '/ the risk assessment which could have a large impact on exposure of the public.

This study was thus conducted to address these concerns and to show that the BWR design effectively removes potential bypass pathways from the standpoint of public risk to fission products. Although previous work was based on a BWR/6 with a Mark III containment, the study was extended to evaluate the impact of bypass on Mark I and II containments.

This study shows that small bypass lines are severely restricted so that an insignificant release of fission products occur. The design of the containment isolation system on larger lines is sufficiently reliable to make the risk of other bypass paths far less than the pathways which include the suppression pool. The study also shows that for all pathways 1-1

l l .

' ~ GENERAL ELECTRIC PROPRIETARY INFORMATION there are natural ~ fission product removal mechanisms which effectively eliminate these pathways from concern.

The conclusion of the study is that BWR suppressio~n pool bypass pathways are not a concern from the standpoint of public risk from fission product j exposure. -

- 1 i

f.

i l

l l

l l

I l

?

L 1-2 l

~

i  : .

GENERAL ELECTRIC 1

PROPRIETARY INFORMATION i

2. IDENTIFICATION OF BYPASS PATHS l

Fission products released from the fuel in a severely degraded accident may be released to the environment by several pathways as indica.ted in Figure 2-1. Both liquid and gaseous release pathways are considered in the figure. ~

l i

As shown, the dominant release pathways are to the suppression pool either through the safety relief valves (SRV) or through the drywell to wetwell vents. The dominance of these paths is assured by the plant design which isolates the significant release paths to direct the flow of RPV effluent to the suppression pool. ~

l l

Other release paths are possible, however, through small leakage paths or paths where the isolation system has failed to function. The following subsections discuss these bypass paths.

i 2.1 LIQUID RELEASE PATHWAYS (FIGURE 1, PATH A)

Liquid leaks or condensate from steam leaks are normally collected by sumps in all areas which could potentially contain radioactive material.

These sumps are discharged to the radwaste building where they are processed and recycled back to the plant. The radwaste building contains {

a basemat designed to withstand design basis seismic loadings so that even if tank failures occur in the Radwaste Building a release to the environment does not occur.

Once in a subcooled liquid form, liquid effluents are limited in reaching the general public by inadvertent releases, or vaporization. These methods are either unlikely or involve small release fractions.

For the above reasons, liquid releases to the environment are not considered a particular hazard.

Furthermore if an inadvertent release to a river or lake were to occur the contamination could be quickly diluted and alternate l

2-1 k

V

' GENERAL ELECTRIC j

PROPRIETARY INFORMATION drinking water supplies could be temporarily used to avoid an impact on

• the general public.

Potential releases from fluid systens have thus not been considered further in this study.

2.2 AIRBORNE RELEASE PATHS The airborne releases during routine plant operation or following transients or accidents consist of noble gases, halogens and particulates. In the most severe accidents nearly all of the fuel inventory of fission products may be released from the fuel. However the halogen and particulate fission products are substantially retained within the vessel, associated piping, containment air or the suppression pool which limits the amount of activity which is available for release to the environment. As for noble gases, although there may be some holdup in plant buildings prior to release, most are expected to ultimately be released to the environment following a severe accident.

This study has concentrated on the halogen and particulate releases l because these are the types of radioactive material which pose the greatest hazard to the general public if they are not retained. The release of even 100%

of the core inventory noble gases has been shown to cause negligible health effects.

2.2.1 Potential Release Pathways Outside Secondary Containment (Figure 1, Path B)

The secondary containment boundary in BWRs contains all potentially radioactive systems except for systems supplied by the main steam system and its condensate. The lines which pass outside of this boundary and which communicate with the RPV or drywell are identified for a BWR/6 Mark III design in Table 2-1. Other product lines are similar in that the lines contain main steam and feedwater, RPV drains, HVAC exhaust, sump discharge, and cooling water supply and return lines.

1 2-2 L

I

l

' ~

. GENERAL ELECTRIC PROPRIETARY INFORMATION i, Table 2-1 shows that all lines contain primary containment isolation valves which can be closed by remote manual control or which receive automatic closure signals in response to conditions representing a potential break outside containment. The lines in Table 2-1 can be l viewed in three groups.

a. Lines which connect directly to the reactor pressure vessel are designed to General Design Criteria 55. These lines provide a j potential release path for all severe accident events if the contain-i ment isolation system fails to function.

b. Lines which connect to the drywell atmosphere or a system (such as RWCU) not directly part of the coolant pressure boundary are designed to General Design Criteria 56. These lines represent a potential bypass pathway only for events which cause a break of the RPV pressure boundary inside the drywell. The RWCU lines represent a bypass path only for events where the RWCU system fails to isolate.

D

c. Lines which are closed inside the drywell are designed to General Design Criteria 57. These lines represent a bypass pathway only if there is a break in the closed system to cooling water interface combined with system failures such that a pressure difference favors release to the cooling water lines.

These potential bypass pathways occur only if the isolation system fails to function and, for groups b and c, if another system break has occured.

For lines in group 1 a third remote manual isolation valve is also provided to provide an additional level of reliability beyond the primary containment isolation design.

For lines which pass outside secondary containment in the 238 Nuclear Island design, a positive leakage control system is also provided to reduce potential release due to isolation valve leakage. Other product lines contain MSIV leakage control systems which direct leakage to the standby gas treatment system.

(s 2-3

l ':

~. '

GENERAL ELECTRIC .

. PROPRIETARY INFORMATION The lines which are potential bypass pathways to areas outside the l secondary containment are protected by a system design which provides an j extremely high level of isolation reliability. These high reliabilities I make the risk to the general public extremely small. Section 3 discusses the probability of release in more detail. Section 4 discusses the retention of fission products in these lines. .

2.2.2 Potential Release Pathways Inside Secondary Containment (Figure 1, Path C)

Lines which connect either to the RPV or to the drywell atmosphere and which terminate inside the Idecondary containment on all product lines are identified in Table 2-2. Table 2-3 identifies other lines which terminate in secondary containment in Mark I and Mark II containment designs. On Mark III designs these lines terminate inside the primary containment, l but outside of the drywell.

The release of radioactive material through small leaks in any of these '

M '

lines to the environment is treated by the Standby Gas Treatment System (SGTS) which may be initiated in response to a process radiation system signal. Thus although these pathways may bypass the suppression pool the SGTS provides a degree of fission product removal for any particulates or halogens which might be released. Since the decontamination factor of q l the SGTS is about 1000, it provides a degree of removal nearly equivalent l to that provided by suppression pool scrubbing.

I Large breaks of the lines in Table 2-2 may cause overpressurization of their local compartments unless a protective feature such as blowout panels or equivalent are provided in this design. In any case, such l large breaks present a potential release pathway to the environment which l would not be treated by the Standby Gas Treatment System. l l

As shown in Tables 2-2 and 2-3 all lines which communicate with the reactor pressure boundary or drywell atmosphere also contain primary containment isolation valves which may be remote manually closed to limit the duration of any detected leak (instrument lines on Mark I and II i

2-4 w

1

C..

GENERAL ELECTRIC .

l PROPRIETARY INFORMATION  :

i plants contain excess flow check valves). In addition the non essential lines automatically close on response to a detected system leak. These i features limit the risk to the general public from releases through these pathways.

The suppression pool suction lines in general do not contain the same degree of isolation as do the other lines on Table 2-3. This is acceptable, however, because they are a liquid source (refer to Section 2.1) which l does not pose as great a hazard to the general public as do the other sources.

The suppression pool suction lines, however, are also potential pathways which could lower the suppression pool water level and reduce the effective-

~

l ness of pool scrubbing. While breaks of these lines outside of the containment are a possibility, the break possibility is extremely remote

-6 yr) because they do not contain high pressure fluid.

(<10 Furthermore should a leak occur, (except at the containment boundary), it would be detected by the plant Leak Detection System and the appropriate line could be quickly isolated by the operator. On the Mark III design, an automatic suppression pool makeup systeu, actuated on low suppression pool water level, gives additional assurance that the suppression pool level is maintained for a period of time. Suppression pool, suction lines are not considered further in this report.

Finally, there are release paths from the drywell through structural features such as hatches or penetrations. Periodic containment testing l ensures that leakage through these paths is kept to a minimum. But

! regardless of the leak rate, except for catastrophic structural failure, they would be expected to provide a highly restricted flow path. Retention of fission products in such pathways is discussed in Sections 3 and 4.

l l 2.2.3 Potential Suppression Pool Bypass Paths Inside Containment (Figure 1. Paths D and E)

As mentioned earlier, the BWR design directs the dominant flow in transients or in breaks inside the drywell to the suppression pool. The safety l

l 2-5

l .. .**.

GENERAL ELECTRio

'. PROPRIETARY INFORMATION I '

relief valves direct. the decay heat steam to the suppression pool while any excessive pressure in the drywell (as a res, ult of a break for instance) is relieved to the suppression pool through the vents. Bypass paths to the containment may be a concern if they provide a release path directly to the environment from breaks inside the drywell or wetwell or from drywell to wetwell leakage. .-

For Mark I and II containment designs, there are no components connected directly to the RPV in the wetwell above the suppression pool water level. Thus wetwell breaks are not possible. In the Mark III containment design there are several such systems which can provide a bypass path inside containment. These were identified in Table 2-3. In addition there are several high energy lines which pass through the wetwell in l Mark III containments. All these lines contain guard pipes, designed to l direct any line break flow back to the drywell.

i a

Drywell to wetwell airspace leakage is possible whenever the differential pressure between the two zones favors it. It is limited, however, by the 4

head created by the submergence depth of the vents. For all BWR designs L

this is only a few pounds of pressure (psi). Thus there is never a large driving force for drywell to wetwell air leakage.

For Mark III containments the drywell is completely surrounded by the wetwell and suppression pool. Only Mark I and II containment designs have the potential for direct drywell bypass leakage to the secondary containment buildings. For these designs only gross structural failure, such as from an external event

• or hydrogen explosion or isolation valve failures can cause a significant pathway. Since Mark I and II l containment designs are inerted during normal operation, only external j events or isolation system failures are reasonable contributors to this type of pathway. Further holdup and retention of fission products in the secondary containment mitigates these releases as discussed in Section 4.

• Seismic, tornado, tsunami, etc.

2-6 l

- -_.m----

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

us.te4.0enDamame-U PROPRIETARB INFORMA1105 e esiW '

TABLE 2-1 -

• PATHWAYS WHICH TERMINATE OUTSIDE SECONDARY CONTAlt04ENT LEAKAGE- ISOLATION BARRIER DESIGN VALVE LINE FLUID TYPE (1) CRITERIA (2) TYPES (3) i From RPV

! 26" Main Steam Steam PC, 3IV, LCS, GP 55 (a) A0 Globe (I,0,0)

(4 lines) 3'" MSL Drain Water / Steam PC, SC, LCS, GP 55 (a) MO Gate (I,0,0) l 20" Feedwater Water PC, 3IV, LCS, GP 55 (a) A0 Check (1,0) -

! (2 lines) 4" RWCU To Main Cond. Water PC LCS, 2LCS 56 (b) MO Gate (I,0) 3" Drywell Susps Disch.* Water PC, LCS 56 (c) MO Gate (I,0) l 2" RWCU Backwash Drain

• Water PC, 31V, LCS 56 (c) A0 Globe (I,0,0) l l Post Accident Liquid Sample Water PC, RO 55, 56 MO Gate (I,0) j .

l From Drywell 1

I 6" Chill. Water from Water PC, V, LCS 57 (c) MO Gate (I,0)

Drywell 4

l Post Accident Gas Sample Air PC, R0 55, 56 MO Gate (I,0) 1

" Liquid drain path - see Section 2.1 /

m T - m_-

' GENERAL ELECTRIC PROPRIETARY INFORMATION MOTES TO TABLE 2-1 (1) Leakaae Barrier Types PC = Primary Containment Isolation Valves Provided SC ; = Secondary Containment Isolation Valves Provided WL = Water Leg Seal V = Vented to Secondary Containment LCS = Leakage Control System Provided R0 = Flow Restricting Orifice Restricts Bypass Flow 31V = Third Isolation Valve Provided (Remote to Manual) 'e - .

2LCS = Secondary Containment Leakage Control System Provided GP = Guard Pipe Between Drywell and Secondary Containment (Mark III only)

(2) Isolation Sianals (Remote Manual Plus)

(a) Low RPV Water Level (L1) High Turbine Building Temperature

High Radiation High Turbine Building Steamline Temp.

High Steam Flow Low Condenser Vacuum High Steam Tunnel Temp. Low Main Steam Line Pressure (Run Only)

(b) Low RPV Water Level (L2) Interlocks with Pump or Valves High Drywell Pressure High Differential Flow l High Steam Tunnel Temp.

(c) Low RPV Water Level (L2)

High Drywell Pressure (3) Valve Actuator Types M0 = Motor Operated I = Inboard Primary Containment

A0 = Air Operated 0 = Outboard Primary Containment 1 .

?

l 1

l l

2-8

+

UENEnML E EtaIfHb ~

~

PROPRIETARY WFORMATION J

TABLE 2-2 ,

PATHWAYS WHICH TERMINATE INSIDE SECONDARY CONTAIPMENT -

LEAKAGE BARRIER ISOLATION DESIGN VALVE LINE FLUID TYPE (1) CRITERIA (2) TYPES (3)

From RPV 14" RHR LPCI Mode Water PC, 2LCS, CL 55 (a) A0 Stop Check (I)

M0 Gate (a)

, (3 or 4 lines)

I 20" RHR 50 Cooling Line Water PC, GP 55 (b) M0 Gate (I,0) 10" RCIC Steam Line Steam PC, LCS, GP 55 (c) MO Gate (I,0) 6" RCIC Pump Discharge Water PC, 31V, CL, GP 55 A0 Stop Check (I,0) 12" LPCS Pump Discharge Water PC, CL 55 (a) MO Gate (I,0) t 12" HPCS/HPCI Pump Discharge Water PC, CL 55 A0 Stop Check (I)

M0 Gate (I,0) 1 l 6" RWCU Pump Suction Water PC, GP 55 (d) M0 Gate (1,0) 6" RWCU Return to FW Water PC, GP 56 (d) MD Gate (I,0) l From Drywell

2" Drywell Bleedoff Vent Air PC 56 (e) M0 Gate (I,0)(2 lines)

SUPPRESSION POOL SUCTION LINES 24" RHR Pump Suction Water CL, 2LCS 56 MD Gate 8" RCIC Pump Suction Water CL 56 (c) MD Gate 12" LPCS Pump Suction Water --

56 MO Gate 24" HPCS/HPCI Pump Suction Water CL, 2LCS 56 MDfiate 12" SPCU Pump Suction Water PC, CL 56 (e) MO Gate 8" SPCU Return Water 2 LCS 56 (e) MD Gate

GENERAL ELECTRIC .

PROPRIETARY INFORMAil0N

( .

NOTES TO TABLE 2-2 (1) Leakage Barrier Types PC = Primary Containment Isolation Valves ~~

CL = Closed Loop Inside Secondary Containment SC = Secondary Containment Isolation Valves 3IV = Third Isolation Valve Provided l 2LCS = Secondary Containment Leakage Control System (2) Isolation Signals (Remove Manual Plus) i (a) Injection Valve Pressure (>450 psi)

(b) Low RPV Water Level (L1) High RPV Pressure (>150 psi)

(c) High RCIC Room Temperature High Turbine Exhaust Press Low RCIC Steam Pressure p (d) Low RPV Water Level (L2) High RWCU Room Temp. -

High Drywell Pressure High RWCU Differential Flow High Steam Tunnel Temp.* Interlocks *

[ (e) Low RPV Water Level (L2) High Drywell Pressure l

I

(*Except RWCU Pump to Demin.)

(3) Valve Types M0 = Motor Operated I = Inboard Primary Containment A0 = Air Operated 0 = Outboard Primary Containment 2-10

ttREMML *g.EGIMiG -

PROPRIETARY DIFORMATION '

TABLE 2-3 PATHWAYS WHICH TERMINATE INSIDE SECONDARY CONTAINNENT (MARK I, II) AND INSIDE PRIMARY CONTAINMENT (Nark III) ,

LEAKAGE . ISOLATION BARRIER DESIGN VALVE LINE FLUID TYPE (1) CRITERIA (2) TYPES (3)

From RPV 1" CR0 Insert / Withdraw Water PC 56 A0 Ball (Incl. SDV) (177 lines) 1-1/2" SLC Supply Water PC 55 Check (I),

M0 Stopcheck (0)

Expl. (0)

TIP Guide Tube Air PC, AP 56 (a) 50 Ball; XO Shear '

{ (6 lines) 3/4" Instrument Lines Water, steam XF, R0 55, 56 XF Check (Mark I, II only)

(62 Lines)

Instrument Lines Air / Steam 56 (19 Lines)

(SRV Tailpipe Pressure) 3/4" Instrument Lines Air PC 56 Manual Globe (DW Press & dp) (4 Lines)

From Drywell 18" Vacuum Relief or Air PC 56 (b) A0 8' fly (vacuun relief)

DW Purge MO 8' fly (H2 NI*f"9) 1/2" Instrument Lines Water XF, RO 55, 56 XF Check (Mark I, II only)

(4 Lines)

Drywell Air Lock Air T N/A N/A' Drywell Equipment Hatch Air T N/A N/A Through Wall Leakage Air N/A N/A N/A

GENERAL ELECTRIC i

• PROPRIETARY INFORMATION NOTES TO TA8LE 2-3 I

(1) -Leakane Barrier Type CL = Closed System. These lines terminate inside primary -

containment (Mark III only).

PC = Primary Containment Isolation Valve (Mark I and II only).

AP = Air purge of lines limits the amount of drywell to i containment bypass.

XF = Excess Flow Check Valve (Mark I and II only). -

GP = Guard Pipe between Drywell and Secondary Containment (Mark III only).

1 R0 = 1/4" reducing orifice limits amount of bypass leakage.

T = Periodic Leak Test. ,

(2) Isolation Sianals (Mark I and II only: remote Manual Plus)

\-

(a) RPV Water Level 2 Drywell Pressure High (b) RPV Water Level 2 Drywell Pressure High Exhaust Radiation High (3) Valve Types A0 = Air Operated I = Inboard Primary Containment MO = Motor Operated 0 = Outboard Primary Containment j XO = Explosive Operated XF = Excess Flow l

l 2-12

_ .. - - __2

l . . .

I .

~

1

. GENERALELECTR!C COMPANY PROPRIETARY INFORMATION l 6

ENVIR ONMENT CONTAINME NT i

~I6 WI

., DRYWELL WETWELL CON A NMENT e=- 0 / \

r n N  : 5GT,5

\ "N' ntacToR sRv

[h E

$h =

nry

/

VENT

'"" "" sump sump surenEssioN l '**'

Wt>M 7722 77M M DWAITE ] f p [

g N ANT RELEASE

" E0"lE ,'Ir",!!

l Figure 2-1. BWR Potential Bypass Pathways l

l l

2-13 l

statunt tuvims PROPRIETARY INFORMATiGi

, 3. SIGNIFICANCE OF SUPPRESSION POOL BYPASS i

• One way to assess the significance of the potential bypass paths identified in Section 2 is to evaluate their contribution to the annualized general '

public risk. The 8WR/6 PRA evaluated the general public risk and, as shown in Table 1-1, found that it is dominated by transient events where no bypass paths occur. For these events, the offsite public exposure is largely dominated by the noble gas dose because the suppression pool effectively removes the particulate and the halogen fission products.

The annualized general public risk is dominated by these events because of the relatively high event frequency for transients.

An assessment of .the bypass paths identified in Section 2 has been made based on determining their contribution to annualized general public risk. Expressed mathematically, the annualized general public risk can j be described as:

.. R= E*F (1)

All events f t

Where:

R = Annualized general public risk (manrem/ year) ,

E = Exposure per event (manrem/ event)

• F = Frequency of event (events / year)

The exposure to the general public is composed of two parts: exposure to the noble gas cloud (EN ) and exp sure to particulates and halogens (Eg ).

l In simplified ta'rms, the risk can thus be expressed as:

R= (EyF + EyF) (2)

All

(, events 3-1

GENERAL ELECTRIC PROPRIETARY INFORMAT10fi

, Where the N and I subscripts refer to noble gas and iodine / particulate i releases, respectively. Since pool bypass paths only affect the second of these expressions, the remaining discussion will~ focus on i the EgF term.

l The significance of each of these release pathways is evaluated by j consideration of both the frequency of the event (F) taking any specific

! pathway and exposure terms (Eg ). As shown on Figure 3-1, the pathways

i. which bypass the suppression pool and those which pass through the

[ suppression pool combine give a certain amount of activity available for release from the plant. As shown, there is a " resistance" to release of l fission products through any pathway which is analogous to parallel I

electrical resistances. If the total " resistance" of the bypass pathways is much greater than the resistance of the suppression pool pathway, then i the bypass paths are not significant contributors to overall risk.

1 I For bypass pathways the " resistance" can be thought of in two parts:

1) the bypass probability (P B) f r a given pathway occurence concurrent with a core damage event (see Section 3.1) and 2) a factor (F ) which B

represents the fraction of core damage release from the reactor pressure vessel which takes the bypass pathway. This factor (FB ) reflects the fact that small pathways are not capable of passing the full vessel release flow. Since the remaining flow is transferred to the suppression

pool, this term is also referenced to on the flow split (see Section 2.2).

When the bypass probability and flow split are combined they represent a release fraction which may be compared with the decontamination factor of the suppression pool to determine the significance of the pathway.

As discussed in Section 4, there are other retention mechanisms which are in effect for both the pathways which pass through the suppression pool and those which do not. These retention mechanisms function reduce the

. significance of core damage events on the general public. However, since

, they are separate phenomena and independent of the bypass or suppression 3-2 A _ _ ___ _ _ _ _

GENERAL ELECTRIC i

PROPRIETARY.INFORMATION pool pathway, they do not contribute to the significance of bypass pathways relative to the pool pathways.

3.1 PROBABILITY OF ' SUPPRESSION FOOL BYPASS (P g) t j The probability of bypass for each line identified in Section 2* bas been estimated. The probability values were obtained from the failure data '

l used in the 8WR/6 PRA (GESSAR 15D.3) and are summarized on Table 3-1. In j general, these probabilities considered failures. of the primary containment l isolation valves and a line break probabilities as independent failures.

The overall bypass probability also takes into account the types and j number of valves, line size, and number of lines among similar or redundant

[ groups of lines.

1 i

For certain lines, such as low pressure ECCS injection lines, which have a high pressure to low pressure interface, the failure of the control  !

system logic (Iow pressure interlocks) was used in lieu of piping system l

g failure. Such a failure could result in the possibility of over pressur-ization (and failure) of low pressure piping while the plant is still at high pressure.

The bypass probability represents the conditional probability,'given a core damage event, that a certain bypass pathway may exist. The values .

obtained in this evaluation are summarized in Table 3-2.

3.2 BYPASS FLOW SPLITS (FB )

To determine the flow split values, potential bypass flow rates were estimated from two sources and the more Ifatting was used on the evaluation:

l 1) formulae for flow of compressible fluids in pipes and crifices and l 2) plugging of aerosols in small holes or cracks.

l 3-3

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

m -,

r

^~

l . *

~ ~ ' GENERAL ELEDTRIC PROPRIETARY INFORMATION i

4 The results of the bypass flow split and plugging fraction evaluations are included on Table 3-2. It should be noted that the flow split evaluations are conservatively based on the full pipe size diameter.

Restrictions due to valves and pipe crack exit effects would be expected to further restrict the potential effluent flow through bypass pathways.

3.3 EVALUATION RESULTS The results of the bypass probability (P g ), flow split (F g) evaluations and plugging fractions are shown on Table 3-2. In order to evaluate the importance of these bypass paths the product of the flow split (Fg) and

~

bypass probabilities (P g) should be compared against the fission product 3-5 I

GENERAL ELECTRIC PROPRIETARY INFORMATION

(, retention which occurs in the pathw y s which pass through the suppression pool as shown in Figure 3-1. -

Figure 3-2 is a simplified event tree showing the overall resistance (1/PggF ) as compared with the decontamination factor expected in the pool due to pool scrubbing. The bypass lines listed on Table 3-2 have been further grouped to show the relative significance of different release path types. The drywell pathway pool decontamination factor is based on the vent discharge to a saturated pool while the others are based on quencher discharge. It can be seen from Figure 3-2 that in all cases there is substantially greater resistance to release through potential bypass pathways than through the suppression pool. Consequently, source terms used to evaluate the consequences' of severe accidents need only be concerned with the dominant release paths which are through the pool.

Reviewing the flow split and plugging fraction data on Table 3-2 shows that for the containment release pathways from the RPV, all lines contain f significant restriction. By considering the conservatism in the methodology used, the conclusion can be reached that these pathways are not likely to realistically be a concern.

l l <

This evaluation also shows that the release pathways from the drywell are

~

dominated by the TIP guide tubes and the guard pipe failure. The drywell I '

vacuum breaker pathway, although it is potentially a large bypass pathway, does not significantly contribute to overall general public risk due to the low bypass probability, i

I I

l '

l i

3-6

' GENERAL ELECTRIC

, PROPRIETARY INFORMATION 4

Table 3-2 BYPASS LINE PR08 ABILITIES AND FLOW SPLITS 8YPASS

1. OF ISOLATION PROBA- FLOW PLUGGING LINE LINES BARRIERS 81LITIES SPLIT FRACTION NOTES LINES TO OUTSIDE SECONDARY CONTAINMENT From RPV 26" Main Steam 4 4 4x10

-10

1. 0 A 20" Feedwater 2 -12 4 <10 1.0 3" Main Steam Drain 1 3 1x10

-12 -1 2x10 A 4" RWCU To Main Condenser 1 -12 ~1 6 <10 5.8x10 A Post Accident Liquid 2 4 1.6x10

~4 5x10

-6 1x10

~3 8

Sample From Drywell

-12 6" Drywell Cooling Water 1 3 <10 4.4x10 -2 C,K

-4 Post Accident Gas Sample 2 4 1.6x10 4.5x10 -3 2x10

-4 8,K LINES TO SECONDARY CONTAINMENT From RPV 20" RHR Shutdown Cooling 2 4

-12

<10 1. 0 D,E 10" RCIC Steam Lines -12 1 3 1x10 1.0 6" RCIC Pump Discharge -12 1 3 <10 1.0 12" HPCS Pump Discharge -12 1 3 1x10 1.0 14" LPCI/LPCS Discharge ~9 4 4 4x10 1.0 0 6" RWCU Lines 1 5 <10

-12

1. 0 F From Drywell 4.4x10-2 -2 2" Drywell 81eedoff ~7 g 2 2 2x10 1.8x10 3-8

T

. GENERAL ELECTRIC PROPRIETARY INFORMATION Table 3-2 (Continued)

BYPASS

1. OF ISOLATION PROBA- FLOW PLUGGING LINE LINES BARRIERS BILITIES SPLIT FRACTION NOTES LINES TO CONTAINMENT From RPV 1" Instrument Lines 80 1 1.8x10-2 2.2x10

-3 1x10~4 I,J

-2 1-1/2" SLC Line 1 3 2.3x10 3 3x10 2.6x10 -2 3 4.1x10 -2 -5 '3 1" CRD Lines 177 3 2x10 1x10 G,J 3/8" Sample Line 1 3 2.3x10 -10 5.0x10

-3 4x10

~4 J

t From Drywell 10" Vacuum Relief 4

-8 1.0 l 2 4x10 J,K

+H 2 Mixing ~

TIP Guide Tubes 5 0 1.0 1x10

~4 1x10

-6 H,J,K Airlock / Equipment Hatch 2 1 2x10

~4 2x10

~3 2x10

-2 J,K,L,M Guard Pipe Failure -3 -2 -2

., 1 1 1x10 4.4x10 118x10 J,K,N ,

L Unidentified Drywell -

0 1.0' 2x10

~3 3x10

-5 g,g,g Leakage i

i l

1 3-9 i

GENERAL ELECTRIC PROPRIETARY INFORMATION.

NOTES TO TABLE 3-2 A. Release path through main condenser failure; condenser vacuum loss assumed. Drains presume discharge above condenser water level.

B. Flow split assumes flow restricting orifice; liquid sample line assumes scrubbing in line (10-3).

C. Flow split assumes small (<2") opening in heat exchanger tubes.

D. Assumes check valve and low pressure interlock failure at high pressure causes break of low pressure RHR or LPCS piping.

E. Operator error to inadvertently open valve while at high pressure is l -

assumed.

F. Probability is based on RWCU pump suction line. Other lines are less likely to be bypass paths.

i G. Flow split assumes in-vessel scrubbing (10-3) prior to vessel l failure likely due to bottom entry; Not a likely pathway after RPV failure.

j H. No isolation provided in Mark III design. Flow restriction severe l due to probe left in guide tube. Mark I and II designs contain dual ~

barrier protection.

I. Mark I and II designs have excess flow check valves; bypass probability is about 10+3 lower.

J. These pathways discharge to primary containment in Mark III designs; containment failure also required to provide bypass path.

K. Pathway from drywell air is most likely after vessel failure.

Bypass probability assumes probability of release in drywell at 1.0.

3-10

. . .- s s

. L. Pathway consists of numerous small paths. Equivalent path less than 125 in.. assumed (conservative).

M. Bypass probability presumes a core damage event which generates sufficient drywell pressure to cause excessive leakage. l N. 8ypass probability assumes failure of the guard pipe due to hydrogen j burning inside the Mark III containment.

GENNRAL ELECTRIC PROPRIETARY INFORMATION l

(

i i

l 5

i l

r 3-11 <

i 1

~

EVALUATION XTH005

~

P OF P Pack g P ACTIVITY DAMAGE .

W AVAILABLE FOR

RELEASE 1

). M l 1 0F 8

BIZAsi i

P, P, = PROBABILITY OF POOL PATIMAY >

'Ps = PROBABILITY OF BYPASS E*'

ulrg C=3 DF, = SUPPRESSION POOL SCRUBBING OF O p*""

I

0Fe

= 1/ FRACTION OF FLCW BYPASSING POG. (Fa ) E4 Q sec: a.

@ On

.. yc OVERALL ' RESISTANCE" "6

• a2C k=

~

+

ks =h + P,F, f.s J >

P J Ql%

La. **C

J

.eD_. .

1 BYPASS IS EGLIGIBLE IF: Q: CC

yg O

I

r. r. - g e[ g j .

4 i (

! Jun 5 I T

3-12 i

1. m - _ - ---- -- ~,,._ ,...___.m.. _ _ _ . . . . - _ __ . _ . . _ _ . _ . _ _ _ . - - - _ _ _ _ _ _ - - - _ _ . - , . _ _ . _

tit JetNALtpl.tGIH10 ' '

PROPRIETARY 1NFORMATI0lt FIGURE 3-2

• RELEASE PATWAY EVENT TREE No Pipe No RPY No Drywell Overall . Seggwession Break Isolation Isolation Pesistance Pool Valve Valve Bypass Path Failure Failure 104 No

-0 Turbine 5x10 Building  :

104 No Core Damage Probabi11ty 8.3x10e y,,

w L Secondary -

Containment  !

104 No ,

hiah oressure 9.9x1081 Yes low nressure 2.5x10a ye, Containment 104 No

4x10s Yes i, ,

Drywell  !

r 102 No I

1.9x104 Yes i

? " ' :-

GENERAL ELECTRIC "

PROPRIETARY INFORMAfl0N

, 4. NATURAL FISSION PRODUCT RETENTION ON BYPASS PATHWAYS

.There are other natural removal mechanisms which prevent release of particulates and halogens inside the reactor pressure vessel and along

, the release pathway. These mechanisms are in effect in all sevdre accident related paths and further reduce the source term of radioactive material to which the general public may be potentially exposed.

Several principle mechanisms have been identified which are in effect.

l First, plateout and/or deposition of the material inside the reactor l vessel, its components such as separators, dryers, and channels, and also

] plateout along the vessel piping pathways provide a mechanism for retention t

prior to release to the environment. Secondly, once released to the

building outside the primary containment (or within the primary containment for breaks) a significant retention in the highly humid environment and i relatively cold surfaces would also be expected due to condensation or deposition. These naturally occurring mechanisms are currently being i

studied by various National Laboratories. The following sections summarize j the current estimates of the potential retention factors which may

) ultimately be demonstrated by these programs.

l i

4.1 PRIMARY SYSTEM PLATEOUT I.

A computer code (TRAP-MELT) developed by Battelle-Columbus for the NRC has been used to estimate the amount of fission produce retention in the primary system. Overall invessel retention ranging from DF of 1 to 10 are expected for paths which include vessel separators and dryers and steam lines. TRAP-MELT verification testing is being conducted by the Oakridge National Laboratory. Early results support the models included in the TRAP-MELT Code.

4-1

l

. GENERAL ELEC1RIC PROPRIETARY INFORMATION I

An internationally sponsored testing program is being conducted at the Marvikin facility in Sweden which is attempting to provide test data to support the level of in-vessel and ex-vessel piping plateout assumed.

This work is a multi year program which was initiated in 1983.

An NRC sponsored program with INEL is conducting "in pile" tests'with release of hot aerosols.through piping to show the degree of removal

, which occurs. Significant amounts of piping deposition are being observed.

Finally, the retention of fission products is highly restricted pathways such as cracks, leaking hatches, or leaking valves is expected due to agglomeration of the solid fission products. Several NRC sponsored research projects are developing models and testing aerosol behavior to verify these models.

l The result of the above studies are providing technical justification for assuming a high level of primary system plateout. Mechanisms of deposition

@ in piping bends, gravity settling in low flow regions and condensation on cold pipes are expected to be in effect and provide a removal fraction of

-1 to 10 -2 j 10 in addition to the suppression pool scrubbing values on the majority of paths.

4. 2 BUILDING PLATEOUT (RAIN FOREST)

Stone and Webster is including a treatment of rain forests in BWRs in its source term paper for the ANS. They have published a paper

• based on PWR studies which shows decontamination factors greater than 30 for a break outside of containment.

l

" Assessment of the radiological consequence of particulate reactor accidents, CSA Warman, November 1982, Presented at Second Internals and Conference in Nuclear Technology Transfer, Buenos Aires, Argentina.

4-2

' "~

SENERAL ELECTRIC .

. PROPRIETARY INFORMATION 9

A similar assessment for SWR line breaks outside containment would be expected to show substantial fission product retention inside building / rooms outside containment due to de' position on wet building surfaces and gravity settling of larger particles. Another removal fraction of 10'1

• to 10-2 for buildings is expected.

4.3 CONCLUSION

The ongoing studies identified above are expected to show that substantial removal mechanh s exist in BWRs independent of the suppression pool.

The judgements used in PRAs, the Stone and Webster paper, and the ongoing NRC research programs, lend confidence that a removal fraction on the order of 10'2 can be justified for large bypass pathways due to inherent retention mechanisms other than suppression pool scrubbing.

7 l

4 4-3 i

f'-

GENERAL ELECTRIC PROPRIETARY INFORMATION

, 5.

SUMMARY

/ CONCLUSIONS The potential pathways for fission products to be released to the environment which bypass the suppression pool have been identified. These pathways are of concern because they do not benefit from the fission product retention capability of the suppression pool.

In examining each potential line it was found that the general public

~

risk is not greatly affected by bypass lines either due to the relatively low amount of flow which would pass through the small pathways or due to the high reliability of the containment isolation design for the larger lines.

As a consequence source terms based solely on soppression pool pathways with credit for suppression pool scrubbing are justified.

s

, Several studies are in ' progress to show that the risk is also not significant t

$En because of natural renoval mechanisms which are in effect independent of the containment isolation system design.

)

Based on these conclusions and the likelihood that the continuing studies  ;

on natural removal mechanisms will confirm the presence of significant retention in bypass pathways, the suppression pool bypass pathways are  !

~

not considered a source of concern for BWR's.

b l

l l

l l l 4

5-1

'*'.'. ** GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY R v. 2

. PROPRIETARY INFORMATION Class III ()

O( k?

H.4 TYPICAL BWR CONFIGURATIONS 1

b Boiling water reactors (Figure H.4-1) have a core support con-2 figuration in which the control rod drive (CRD) guide tubes support the core from below. There is essentially one CRD tube for each group of four fuel assemblies such that the support

, is not only from below but it is also localized.

Given this core support configuration,-rhich i- illu:tr:ted-4n-

.Eigure M.d-2, it is virtually impossible to conceive of a sequence l whereby a degraded core would catastrophically collapse into water. In addition, with the extensive CRD guide tube structure,

, it is equally difficult to envision any process whereby rapid and intimate mixing could occur. The specific details of this i

reasoning process are given below.

l. Under normal operating conditions, the guide tube structure is designed to support the entire core. The major change in the material properties occurs when substantial overheating takes place, but this can only occur in the absence of water. If water is absent steam explosions are not possible.

2. In addition, each group of assemblies is, in effect, individually supported and if a degraded core condition is assumed, the most likely way in which molten core material would migrate to the lower plenum is through the assembly orifice located within the support tube. This would undoubtedly be an incoherent process and the molten core material would flow into the interstitial spaces between the CRD guide tubes and perhaps contact the steel wall and freeze. However, thermal attack of the tube i

itself would not begin until the water had been boiled away inside of the tube. Consequently, not only would 2

15.D.3-696

GESSAR II 22A7007 238 NUCLEAR ISLAND R3v. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III i

H.4 TYPICAL BWR CONFIGURATION ( Continued) the melt progression be incoherent, but the core material could not participate in a global interaction until the water was vaporized. This eliminates the potential for -

any steam explosion.

3. If all the above physical restraints are completely dis-regarded and one assumes that coherent core collapse occurs in any event, then one must consider the forest of support tubes, control rod thimbles, and instrument tubes which exist below the core. This massive, cold structure,thich could freeze the core debris on contact,

! would prevent any large scale, intimate mixing of the

. molten debris and coolant.

\ - These three points, all dealing with the below-core structure, show that catastrophic collapse in the presence of water cannot occur, the downward progression of any postulated scenario would be incoherent and occur within the support tubes (and only in the absence of water), and large scale, intimate. mixing could not be achieved. Therefore, large scale steam explosions involving substantial masses of core material can be ruled out on geometric considerations alone. In addition, these can be con-i i

sidered remote in light of the massive, coherent interaction required in WASH-1400 before vessel failure was calculated. The L below-core structure was ignored for the WASE-1400 BWR analyses.

1 6

One can be equally critical of the slug formation, displacement, and impact model from NASH-1400 as it relates to the actual l

design.

l. With the below-core structure segmenting the water with the core support tubes, the formation of a continuous, q overlying liquid slug can also be discarded.

l f

! 15.D.3-697 0

OL_ _ .- .___ _ _ _.. __ __ -

_ . ~. - .

• GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMAITON Class III H.4 TYPICAL BWR CONFIGURATIONS (Continued)

2. If such a slug is postulated the core grid at the top of the fuel assemblies and the upper plenum dome would des-troy the coherence as the material travels upward i through the vessel.

! 3. Steam separators, located above the core as. '--- in _

.EiffE "f W are large structural components which do i

not provide straight-through flow paths. This would also prevent the upward transmission of a coherent liquid 4 slug. ,

I

4. Steam dryers are positioned above the steam separators.

'These components, like the steam separators, also have

~

a tortuous flow path, and thus, provide another barrier 4

to the postulated coherent behavior.

! 5. In addition to destroying the coherency of a liquid slug, L

the mentioned structures will also attenuate the energy

of dispersed material.

These arguments have been formulated on the basis of specific components available in the reactor vessel but ignored in the Reactor Safety Study. As discussed, these differences are indeed extensive and the discussion of each shows that their neglect in

.WASE-1400 grossly overestimated 1) the likelihood of an event, l 2) the amounts of material involved, and 3)-the damage potential represented by an event. Considerations of the structural

' components allows one to individually rule out 1) catastrophic collapse, 2) rapid and intimate mixing, 3) coherent slug forma-tion, 4) coherent slug transmission, and 5) coherent slug impact.

As summarized in Table H.2-1, all of these are required for the WASH-1400 analysis to predict steam explosions. However, there

)

e i

I 15.D.3-698 s

(

- CESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION Class III H.4 TYPICAL BWR CONFIGURATIONS (Continued) is even a more fundamental misrepresentation in WASH-1400 and that is the characterization of steam explosion themselves. This is addressed in the next section.

W.

4B f

i l

4 f

i L.

1 15.D.3-699 L

1 d - _ _ _ _

- GESSAR II 238 NUCLEAR ISLAND . 22A7007 CENERAL ELECTRIC COMPANY R3v 2 PROPRIETARY INFORMATION Class III H.5.2 Low System Pressures (Large Break Sequence) (Continued)

Since the major concern of this evaluation is the damage potential represented'by in-vessel steam explosions, one must evaluate the amounts of material which can come into contact and mix on an intimate scale prior to the onset of an, explosive interaction.

To make this assessment, necessary criteria for achieving a

~s ignificant interaction in the RPV must be defined and each evaluated with respect to governing physical principles.

If a steam explosion is conceived to be a physical process whereby the reactor pressure vessel integrity can be violated and as a result also violate the containment integrity, several specific criteria must be satisfied for the physical processes to achieve such a magnitude. These are listed below in essentially their chronological sequence and each is discussed individually. As k will be shown in this discussion, each physical process represents a highly improbable if not impossible condition, and as such this li'sts provides a description of why steam explosions are of no practical importance in the containment assessment.

H.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel Explosion To develop a steam explosion of sufficient energy to violate a reactor pressure vessel requires a) sufficient molten corium poured into the lower plenum, b) a sufficient molten condition at the time of initiation of the explosion, c) insufficient pressuriza-tion to permit inter penetration of the melt and water, d) coarse intermixing to a sufficiently small scale prior to the explosion, e) a sufficient trigger to mix these materials on an explosive time scale (s10 muec), f) either sufficiently high shock pressures to rupture the lower head or g) a slug formation and h) transmis-sion upward through the RPV with a coherent impact on the vessel head. Each of these is discussed in more detail below.

15.D.3-709 s

GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION Class'III

)

H.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel Explosion (Continued) a) The principal consideration is not whether steam explosions can occur inside the RPV but whether they can be of such magnitude as to fail the RPV and thereafter violate the containment integrity. There-fore the primary question is whether sufficient molten material is available to provide the necessary work for violation of the reactor pressure vessel. Typically i this would require a minimum of 1.2 tons of molten core material for a theoretically perfect thermal interaction I

and realistically considering the ability to vaporize the liquid in close proximity to the fuel, one should

l. consider several times this amount of material, i.e.

perhaps 12 tons. This material must be available to the water on the time interval sufficiently short such that )

pressurization resulting from steam generation during film boiling as the material enters the water does not provide sufficient forces to prevent penetration of the material. This will be addressed later, but it is closely interwoven with this issue of sufficient corium availability.

b) To initiate an explosive interaction, the corium must be in a molten condition when it contacts the water and it must maintain this molten state during the pre-explosion stage. As a result, it requires that this coarse frag-mentation and mixing take place in a sufficiently short time that the surface does not solidify. In fact, they should not even appronch freezing since corium mixtures become viscous as they approach the liquidus point.

This limitation on surface temperature is even more restrictive since the material must be molten and have ')

l 15.D.3-710

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

l

• GESSAR II I A 007 238 NUCLEAR ISLAND ***

j GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION )

Class III j N.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel Explosion (Continued) a sufficiently low viscosity so that the pre-explosion, coarse fragmentation process can proceed with the limiting hydrodynamic forces associated with liquid-liquid film boiling. The rate at which the surface cools'is dependent upon the particle size, but in all cases is typically on the order of magnitude 1 sec. As a result, l this is coupled with the sufficient corium requirement i discussed in item (a) and these combined establish the rate at which the molten material must be added to the water in the lower plenum to provide a material state i.

for a sufficient explosive interaction.

c) As the material enters the water in a film boiling state,

( the energy transfer in film boiling tends to pressurize the water which in turn attempts to separate the water and core debris. In assessing the needs for a sufficient explosive interaction, the conditions must be such that the pouring or dropping process is of sufficient char-acter that pressurization of the water and corium does not occur at the interface and thus preclude the inter-penetration of the overheated material into the water.

This is related with issues (a) and (b), but is also separate since it represents the ability for continued interpenetration of the corium debris into the water regardless of the failure mechanism which provided the i material pour. ,

d) After sufficient material has been generated and released to the water under conditions providing for its global penetration into the water on a short time scale, the

' material must have sufficient time to undergo fragmen-

' tation (in the film boiling state) to the level dictated 15.D.3-711 l

L

GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 l PROPRIETARY INFORMATION Class III )

f H.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel

Explosion (Continued)

! by the hydrodynamic stability of the water and core j l debris. In this regard the film boiling fragmentation i model provides for a descriptive formulation of the level .

of such pre-fragmentation that can occur. Another means of assessing the sufficient fragmentation size in the liquid-liquid film boiling state is to identify an

. available trigger (and its energy level) in the reactor j system to initiate such an explosive interaction and equate this to the mixing energy, which will be discussed i

below. With this energy level, the size required for

establishing a triggerable system can then be determined which sets the scale of the pre-explosion fragmentation.

]

Comparisons can then be made between the scale of this necessary fragmentation and that achievable in liquid-

)

i

liquid film boiling.

e) Given that such large quantities of molten materials can be available and added to the water over a suf-l ficiently short time interval and premixed to sufficient

level while still in a molten state, a sufficient trig-l ger must then be available at the appropriate time to l

provide the necessary mixing energy to carry out the explosive interaction. This can be addressed in terms j of the available pre-explosion fragmentation size, the j amount of water that must be interacted, and the effec-  ;

tive drag coefficient for rapidly intermixing materials on the size and time scales necessary for such large thermal energy transfers. This is a particularly crucial question since it can be principally based on the avail-ability of a sufficiently large trigger as opposed to the statistical question of whether a trigger is )'

delivered at the appropriate time.

- 15.D.3-712

~

, ,, GESSAR II 238 NUCLEAR ISLAND .

GENERAL ELECTRIC COMPANY 22A7007 PROPRIETARY INFORMATION Rev. 2 Class III i 1

H.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel Explosion (Continued) f) Given the completion of all the above necessary steps to achieve a sufficient explosive interaction for reactor pressure vessel failure, the first structural question is the integrity of the RPV lower head. Fail-ure of this part of the, reactor vessel is independent of slug acceleration and impact which will be discussed later and would only be the result of a very strong pressure wave. A typical operating pressure for a BWR system is 7 MPa (1015 psia) as compared to an upper bound pressure of a steam explosion during the expansion phase of approximately 10 MPa (1450 psia). As a result, the failure of the lower head would require explosive pressures on a sustained level essentially equal to the maximum values observed to date. If such a failure

( is feasible it would still have to be considered highly unlikely.

g) If an explosion is to be considered and the lower head of the reactor vessel remains intact, the other vessel failure mechanism considered as potentially leading to the loss of containment integrity is the acceleration of

[

i a continuous overlying liquid slug upward through the re-actor pressure vessel and impingement of the slug on the

} RPV upper head. This requires both the formation of such f a continuous slug and the transmission of this slug in a i sufficient coherent fashion to impact and fail the upper head. The question which would be asked at this stage is whether such a continuous overlying slug could indeed be formed. Imbedded in such an evaluation is the effect of rapid steam formation during the pre-explosion frag-mentation interval. The steam formed in this time h

15.D.3-713 4

_ . ""*WM'a*www- , _ _ _ _

GESSAR II 238 NUCLEAR ISLAND GENERAL ELECTRIC COMPANY 22A7007 i PROPRIETARY INFORMATION Rev. 2 l Class III N.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel )

Explosion (Continued) interval must either be transmitted through this slug region, transmitted up the by-pass region or stored

, within the water. If the steam is transmitted up the downcomer, the water removal is far greater than'has been suggested by the above questions, and the avail-ability of water for sustained interaction of the fuel is considerably less than was discussed above. If this steam remains within the water, the pressurization as a result of corium entering into the coolant is far greater than was alluded to above. If the steam is transmitted upward through the overlying slug, the slug cannot remain continuous and an evaluation of the steaming rates and the transmission of this steam through the

slug require that this material would have a considerable void fraction, i.e. one not identifiable with a continu-

)

ous overlying slug. As a result, the formation of such a slug would be highly questionable.

h) If all the above restraints were assumed to be violated 4

so that a sufficient explosion was conceived with a continuous overlying slug formed, this slug must then be transmitted upward through the remainder of the

. original core configuration, through the upper core j support plate and through the upper internals before it could coherently impact upon the upper head of the f vessel. This transmission must be such that sufficient

} energy is retained to cause failure of the head. This transmission must be sustained through the remnants of the core, the upper core plate, the steam separators and '

the steam dryers, all of which represent siza$le energy absorption capacity. As a result, the transmission of l

i

. 15.D.3-714

- GESSAR II 238 NUCLEAR ISLAND 22A7007 .

I*V* 2 l GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION l Class III l l

H.5.2.1 Prerequisites for Loss of RPV Integrity by In-Vessel Explosion (Continued) the slug would be such that it could be broken up, dis-persed, and deliver incoherent impact forces with a total energy less than that provided by the explosion itself.

Each of these above points must be satisfied before an explosive interaction can be sufficiently energetic to result in failure of

' the reactor pressure vessel and eventually the loss of contain-ment integrity. In the following subsections, these individual 5 behaviors are quantified to provide the basis for an engineering evaluation of the likelihood for such an event.

l H.5.2.2 Quantitative Evaluation of the Prerequisites for the l

BWR/6 RPV Loss of Integrity by In-Vessel Steam Explosion H.5.2.2.1 Sufficient Molten Material l

The slug impact energy required to fail a BWR reactor vessel head has been estimated to be 500 MJ (References H.5-13, H.5-14). To accomplish the work by a steam formation process requires the vaporization of a water mass which is dependent upon the actual path involved.

l e.

'\

i 1 15.D.3-715 g

g

~

GESSAR II 238 NUCLEAR ISLMD 22A7007 GENERAL ELECTRIC COMPMY Rev. 2 PROPRIETARY INFORMATION Class III The calculated mass of 1200 kg (2650 lbs) is a conservative estimate (by an order of magnitude) of the material mass required to initiate an explosive interaction which could threaten the RPV.

Considering the theoretical density to be 7000 kg/m3 (437 lbm/ft3 3,

- 15.D.3-716 C

.. .. GESSAR II 238 NUCLEAR ISLAND .

22A7007 GENERAL ELECTRIC COMPANY R3v. 2 PROPRIETARY INFORMATION Class III i

H.5.2.2.1 sufficient Molten Material (Continued) this would be a spherical accumulation of a molten debris 0.7 m (27.5 in.) in diameter. This dimension is much greater than either the CRD tubes or their pitch, and as a result, the downward movement would occupy several CRD channels and their interstitial spaces. Coherent downward migration would be extremely unlikely in such a loosely coupled system.

H.5.2.2.2 sufficiently Molten and coarsely Fragmented These are considered together since the two behaviors are intimately coupled. As the material fragments, the cooling rate increases and as the corium approaches the liquidus point the fragmentation process in a liquid-liquid film boiling state becomes more difficult. However, as will be shown, the corium t- decreasing temperature enables the coarse fragmentation to continue to a smaller scale.

i

\ *

! 15.D.3-717 O

i

GESSAR II 22A7007 238 NUCLEAR ISLAND Rev. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III H.5.2.2.2 Sufficiently Molten and Coarsely Fragmented (Continued)

If this analysis is applied to a reactor accident scenario in which 1200 kg (2650 lbm) of core material is assumed to fall into the lower plenumr

• k- 'r ;;;ntetica limit-for e aWR vevmetry with y c=Nide%s- is given-in--Table -M.5-4. Fer-illustration purposes--ealculationa_.ase presentedl in_Talde J.5-A for the 4ame Rometry inMeNhsenOf dNtD g0idedubm Q shown the particle sizes are very large for such a large amount of very hot material in a small area. Obviously such large particles would not exist, but the calculation demonstrates (by orders of magnitude) that in a reactor system water cannot remain in the presence of fine particulation. As a re,sult, finely dispersed configurations in intimate contact with water are physically unattainable.

i Another feature of the BWR system noted early in this report which f also is relevant in assessing the potential for intermixing of

! molten corium and water is the extensive below-core structure. In l the analyzed BWR plant the-core is supported from below by_177 con-trol rod guide tubes.

I The CRD flow ~inside these tubes is

~

separated from the inlet plenum water outside the forest of tubes, and except for minor leakage at the inlet to the fuel assembly, these two sources of water do not mix below the top of the core.

j As a result, the only cross-sectional area available for the i intermixing process is that restricted area between the CRD tubes making fragmentation even more difficult. h l "fhbdb In the hypothetical case, the water would be dis-placed by the downward moving corium and any initiation of frag-t mentation would only drive the water away faster. It should be

, noted that only gravity retains the water and that it can readily be displaced into colder (outer) regions of the core, backwards through the jet pumps, etc.

15.D.3-722

e

GESSAR II 22A7007 238 NUCLEAR ISLAND Rov. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III 6

H.5.2.2.3 Slug Dispersal and Pressurization While somewhat out-of-phase, these two phenomena are considered together since they both involve the effect of steam generated in the film boiling, coarse dispersal and intermixing process.

In one instance, sufficient time is available to allow the steam to escape upwards through the overlying pool and the necessary conditions to allow this escape are evaluated. For the second case the steam is assumed to be retained within the pool, thus pressurizing the system.

As the mixing and inner dispersion progresses, the hot and cold liquids, are in liquid-liquid film boiling. Since the molten core debris is at a temperature of 2500'K or greater, the principle mode of energy. transfer would be via radiation from the hot particles to the water. This energy transfer can be expressed as -

15.D.3-724 r

GESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION Class III H.5.2.2.3 Slug Dispersal and Pressurization (Continued)

Therefore, in a slowly developing dispersion (time scale of 1 sec l or longer) the vapor throughput would be substantial and preclude the formation of a continuous overlying liquid slug. If the vapor is assumed to be retained in the pool, the pressurization would disperse the pool, hence no slug formation. Without the con-tinuous slug formation, the only pressure imposed on the vessel is that due to the explosion itself, which experiments have shown to be a few MPa typically, and could conceivably be as high as '

10 MPa. However, such pressure levels do not even threaten the integrity of the vessel.

4 4

J 15.D.3-720 C

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

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

CESSAR II 238 NUCLEAR ISLAND 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION Class III H.5.2.2.4 Rapid Liquid-Liquid Mixing (Continued)

To summarize, the amount of material required to rupture a BWR vessel would be a minimum of 1200 kg (2650 lba) of molten debris.

This high temperature material must then coarsely fragment in the water contained within the lower plenum. Because of both the high material temperature and liquidus point, the fragmentation must occur in film boiling and can only proceed as long as the water can remain in place, i.e. the hydrodynamic stability limit of the water cannot be exceeded. At typical BWR accident conditions, the spherical fragment sizes are essentially the same as the CRD tube pitch, and therefore, greater than the characteristic f

dimension of the interstitial space between the tubes where the water is located. Rapid liquid-liquid mixing from this coarsely '

fragmented state down to a size capable of rapid thermal response can require substantial mechanical work depending primarily on the initial material size and the length over which the mixing )

occurs. Though the proposed analytical models for the' pre-explosive fragmentation and mixing energy are based on mechanistic considerations, their application in the case of a reactor

requires verification against available large scale experiments.

?

This is done in the next section.

H.5.3 References t

H.5-1 R. E. Henry and H. R. Fauske, " Nucleation Processes in large Scale vapor Explosions," Trans, of ASME, Journal of Heat Transfer, Vol. 101, pp. 280-287, May 1979.

H.5-2 D. J, Buchanan, Journal of Physics DJ Applied Physics, j Vol. 7, pp. 1441-1457, 1974.

H.5-3 R. E. Henry and L. M. McUmber, " Vapor Explosion Experi-ments with an External Trigger," Second CSNI Experts Meeting on the Science of Vapor Explosions, Grenoble, France, September 1978.

.)

15.D.3-734 J.

.* GESSAR II 238 NUCLEAR ISLAND .

22A7007 GENERAL ELECTRIC COMPANY ev. 2 PROPRIETARY INFORMATION g Class III H.8 STEAM EXPLOSION - EX-VESSEL H.8.1 Explosion Scenario If, in a defined accident sequence, water cannot be supplied to the reactor vessel to establish in-vessel removal of the decay power from the damaged core, then eventually the core will melt along with the fuel channels, the core support plate, and the core support tubes. This mass of molten material will accumulate in the lower head of the vessel and will thermally attack the vessel wall and vessel penetrations and result in the corium penetration of the vessel head. This would lead to the discharge of the molten material collected in the lower plenum of the PRV into the pedestal cavity below.

For accident sequences such as a large break LOCA, the pedestal cavity may be covered with up to 5 feet of water from the blow-down of the vessel. In this case, as the molten material is released from the reactor vessel it will encounter water and the potential for a steam explosion would exist. Such a steam explo-sion could be triggered when the molten material contacts the wetted floor of the pedestal cavity. In this section, the maxi-mum work potential of such an explosion and its effect on the drywell boundary are estimated. The physical processes and the specific criteria associated with ex-vessel explosions are the same as in the case of the in-vessel explosion discussed earlier.

H.8.2 Molten Corium/ Water Interaction The structure below the vessel of a Boiling Water Reactor lassembly (Figure H.8-1) consists of the bottom head insulation support beams, the control rod drive housing support beams (Figure H.8-2) in-core flux monitors, and the forest of control rod drives and 15.D.3-767

GESSAR II 22A7007 l

J 238 NUCLEAR ISLAND *#*

GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION ,

Class III , j H.S.2 Molten Corium/ Water Interaction (Continued) r~.

the associated hydraulic drive lines QFigure H.8-3p Thus

,, j even if one were to assume the RPV bottom head failed and

consequently spilled the molten material into the space below the downward progression of the molten material would not be in one coherent mass. Molten metal discharged from the vessel will

first encounter thermal insulation which will not provide any significant resistance to continued penetration of the high i temperature discharge. The structures mentioned earlier are below the insulation and would temporarily break up and dis-

! perse the debris causing some material to be discharged through the pedestal windows, and in addition it would distribute the i debris in a fairly uniform manner.

l The size of the globules reaching the water in the pedestal cavity }

af ter a freefall below the control rod drives, will exceed 9 inches in an extreme case of bottom head collapse.

It was shown in Appendix H.2, in connection with the interaction of molten corium and water inside the vessel, that the surfaces of such large particles would freeze rapidly as they attempt to

mix coarsely with water due to intense radiation heat transfer.

Furthermore the downward penetration would be limited by the l

pressurisation of the pedestal water as the corium enters and the steam formation from the film boiling in the coarsely mixed state l would disperse the incoming corium and the generated steam through the CRD door and the hydraulic line tunnels. Once again, as in the case of the in-vessel interaction, the energy required to 1

rapidly six the coarsely fragmented debris far exceeds that of a l

realistic trigger and also exceeds the mechanical work delivered l by the emplosion itself.

l l 1

l I 15.D.3-768

GESSAR II 2)8 NUCLEAR ISLAND , 22A7007 GENERAL ELECTRIC COMPANY Rev. 2 PROPRIETARY INFORMATION g

Class III H.8.2 Molten Corium/ Water Interaction (Continued) on the other hand, molten corium could penetrate through the RPV bottom in a less dramatic manner. Boiling Water Reactors have a forest.of penetrations in the lower head because the control rods are driven from the bottom and the incore instrumentation also ,

enters the vessel from the bottom. For the BWR/6-238 plant there are about 177 control rods, each with its own penetration, 55 penetrations for in-core neutron flux monitors and a reactor vessel drain. The weld area around these penetrations would be subject to a three-dimensional thermal attack in the presence of a significant accumulation of degraded core material. Because of the large number of penetrations and the three-dimensional type i of melting attack that these would experience, as opposec to the i

essentially one-dimensional melting at the vessel wall, en: ~_uid expect these penetrations to be the first element of the primary k system pressure boundary to fail and admit molten corium into the pedestal cavity.

In the event that a control rod drive support is melted through and the mechanism is ejected the resulting vessel breach would be approximately 7.5 cm (3.0 inches) in diameter. Thus for an assumed failure of one C'D R penetration the total breach area would 2

be around 44 cm (7 square inches). Consequently, the amount of degraded core material in contact with water in the pedestal, at the time its front contacts the pedestal floor, would essen-tially be the breach cross-sectional area times the water depth 3

(5 ft), i.e., about .0067 m and 47 kg (0.24 ft 3 and 104 lbs).

If this is at a temperature of 2200*C and the water is at 100'C, the thermal energy contained within the melt is 60 MJ. Using the experimental data reported in Reference H.5-24 the upper bound .

on the efficiency of such interactions, which were conducted with an iron-aluminum oxide thermite and melt quantities of this t t 15.D.3-769

r GESSAR II 22A7007 238 NUCLEAR ISLAND Rev. 2 GENERAL ELECTRIC COMPANY PROPRIETARY INFORMATION Class III H.8.2 Molten Corium/Wated' Interaction (Continued) magnitude, was 14 of the thermal energy of the melt. This would yield 0.6 MJ of mechanical work which is a negligible level com-pared to that required for failure of the containment boundary and its major effect would be to displace the water from the pedestal cavity.

The immediate reaction of an ex-vessel steam explosion would be to disperse the water and degraded core material through the dry-well. This would enhance the contact between the two media and result in rapid steam production. The remainder of the material released from the vessel at this point in time, while not partic-ipating in the explosion, could be rapidly quenching as a result of this dispersion process.

In summary, ex-vessel steam explosions could occur for those )

defined sequences where water is available in the pedestal cavity, but the amount of material would be very limited. In fact, the major effect would be a rapid quenching of that material which had been released from the vessel at the time of the event. Con-sequently, it is concluded that a loss of containment integrity will not occur as a result of ex-vessel steam explosions.

)

15.D.3-770

N SENERAL ELECTRIC CDMPANY ,

I WII (

g.

PROPRIETARYINFORMATION :

^

Class Ill

• i

.- N l

-. .s p' T

= - * . ~ . . . . . .

/ - -

i

/ i, f'sc'J c- = .

\ ,

pl.

0

- t s

9 M

e i

= l r , .

A i 2(2

^

=

I

\

5 g; ,. ,,,

wAsw 54ao sum conspostra s:TE (As REPORTEDI

\ -

l n t

8 m.

i e

WASH 14s0 suuR AT l I' -

s:Tss seenAccaos f

[ id*l wass sacoswR conspos:Ts s:Ts .

.. sa.cnAc caos... .____. ,, , {

.g -

. W.y -

m.

bain g

m.

ii.

l s

, , i ..' s s

. i.4TowT eATA,uvias een vsAR ex ,

as ,

~ ..

.. ... .6lpwr F.42. Star Ommsmerinen of I nrrt FeenUry Owwee surWAS$.1400DWM I

1h 'W

. /

.15.0.3440't5.D.3400

' ~~ --

g. . .

glov,1g 22 NUCLEAR ISLANO .

l GENERAL ELECTRIC COMPANY

. Pfl0PRIETARY INFORMATION

$ll'88 Ill .

O sia w' ,

A '

d to l

. _ , ca.o.nu .

SITE (AB MEPORTED) e wAsw seco own cowositt slT - os enAc coos

'f to', $ WASH seco WWR AT ent.- on enac coo N

1b'*1~

I -

iod se*,

se' io'

'77 si$

EARLi FATAlfTits (XI

' ses4Hs

, .fW f.4=t 3099 g Compednen of Eaty fadity Omen for WASH 1400 OWM t is.o.iees

-33m m i

, N G8NERAL ELECTRIC COMPANY *11

'* **

• PROPRIETARYINFORMATION -

.._ Class ill .

,**** *~ - ,,

7 -

/  :.. . . . _ . _. . .

=

/ -

/

7 -

l .

I B s-l

=,

" 1 M

A 1d =- =

l [ WASH 1400 WWR COMPOSITE

\ g

,,,, SITE (AS REPORTED) 3l ,. -

WASH 1400 BWR AT -

l' b E

~

SITE S _ GE CRAC CODE f WASH 1400 SWR COMPOSITE SITE - A

_ pE cRAc coot.... __,

{

~

i

-18 r _

f f C

~

i l

10 i 7 _

~

l i.

1 -

~

~.el I l l 11111! I ltillll! l i 111lll 11lll! I I l lllll th: 10 1

10 2

_ _ 10 3

10 4

. LATENT PATA,UTIES PER YEAR (X) , =

aasssey

.. ... .f.bwe F.41 Seer Connenrinen of Laennt FoamUry Omen & WA5N.1400 BWR

.is.o.34si/15.D.3400 f

i ih