ML20215K713

Considerations Re Certain Aspects of Severe Accident Mitigation Afforded by Operation of Shoreham at Reduced Power, Final Ltr Rept

ML20215K713
Person / Time
Site:	Shoreham File:Long Island Lighting Company icon.png
Issue date:	06/12/1987
From:	Harrington R, Hodge S OAK RIDGE NATIONAL LABORATORY
To:	Office of Nuclear Reactor Regulation
Shared Package
ML20215K635	List: ML20215K633 ML20215K713
References
CON-FIN-A-9465 NUDOCS 8706250423
Download: ML20215K713 (44)

Text

r CONSIDERATIONS REGARDING CERTAIN ASPECTS OF SEVERE ACCIDENT MITIGATION AFFORDED BY OPERATION 0F SHORERAM AT REDUCED POWER

. S. A. Hodge 4 R. M. Ha r rington Boiling Water Reactor Severe Accident Technology Program Oak Ridge National Laboratory Oak Ridge, Te nness c e l

Le t t e r Re po r t June 12, 1987 Research sponsored by the U. S . Nuclear Regul a to ry Co mmi s-sion Of fice of Nuclear Reactor Regulation under Interagency ,

Ag re eme nt DOE 0554-0554-Al with the U. S . De pa r tme nt of l Energy under contract 00E-AC05-8 40R21400 with the Mar ti n '

Ma rie t ta Ene rgy Sys t ems , Inc.

NOTICE This report was prepared as an account of work sponsored by an agency of the United States Government. Ne i t he r the United States Government nor any agency thereof, or any of their employees, makes any warranty, exp ressed o r implied, or assumes any legal liability or responsibility for any third party's use, or the results of such use, of any in-formation, apparatus, product or process disclosed in this report, or represents that its use by such third party would not inf ringe privately owned right s.

e 8706250423 870612 PDR P

ADOCK 05000322 PDR

1 i

i J

l CONSIDERATIONS REGARDING CERTAIN ASPECTS OF SEVERE ]

ACCIDENT MITIGATION AFFORDED BY OPERATION l OF SHOREHAM AT REDUCED POWER j i

S. A. Hodge a R. M. Harrington Boiling Water Reactor Severe Accident Technology Program l Oak Ridge National Laboratory Oak Ridge, Tennessee l

Le t te r Re po r t June 12, 1987 Research sponsored by the U. S . Nuclear Regulatory Co mmi s-sion Office of Nuclear Reactor Regulation under Inte ragency 4 Agreement DOE 0554-0554-Al with the U. S . De pa rtme nt of Energy under contract DOE-AC05-8 40 R21400 wicb the Martin f Marietta Energy Systems, Inc.

?

s l

~

f

)

NOTICE l, I

i This report was prepared as an account of work sponsored by an agency of the United States Government. Neither the  !

United States Government nor any agency thereof, or any of  !

their employees, makes any warranty, expressed or implied, or assumes any legal liability or responsibility for any I third party's use, or the results of such use, of any in-fornation, apparatus, product or process disclosed in this  ;

report, or represents that its use by such third party l

would not infringe privately owned rights. i 1

{

i I

1 I

i

)

r ,

i I

l CONTENTS

)

l 5

Page l

)

i

1. INTRODUCTION ................................................ 1.1 j

2. MAJOR POTENTIAL SEVERE ACCIDENT MITIGATION ASPECTS OF l OPERATION AT 25% POWER ...................................... 2.1 j 2.1 Delay in the Timing of Severe Accident Events .......... 2.1 I i

2.2 Reactor Response to Sudden Reactivity Insertion g Under ATWS Conditions .................................. 2.1 2.3 Reduced Opportunity for Core Debris-Concrete  !

Interaction ............................................ 2. 2 .

l

3. CALCULATIONS FOR LOSS OF INJECTION (TQUV) SEQUENCES ......... 3.1 ,

+  !

3.1 BVRSAT Program Severe Accident Models .................. 3. 2 i 1

3.2 The Total Loss-of-Injection Sequence ................... 3.2 l

3. 3 Comparison with Total Loss of Injection From j 100% Power ............................................. 3.4 )

- 4

4. ATWS CALCULATIONS ........................................... 4.1 l 4.1 Operator Actions if the Poison Injection Function Fails .................................................. 4. 2 4.2 Consequences of Low-Pressure Criticality ............... 4.3

4. 3 The C9D ATWS Sequence .................................. 4. 5 l 1

5. CALCULATIONS FOR LARGE-BREAK LOCA WITH LOSS OF ALL ECCS (AE) SEQUENCES .......................................... 5.1 l

6. SUMHARY AND FINDINGS ........................................ 6.1 l i

7. REFERENCES .................................................. 7.1 APPENDIX ........................................................ A.1 >

l l

O

l. INTRODUCTION This report describes the results of analyses performed at Oak Ridge National laboratory (ORNL) to assess certain aspects of the ef fec-

- tiveness of operation of the Shoreham Nuclear Power Station at reduced power as a Seve re Ac cident mitigation technique. This wo rk was per-formed at the request of the Risk Applications Branch of the Nuclea r Regulatory Commission (NRC) Of fice of Nuclear Reactor Regulation in re-sponse to a current request of the Long Island Lighting Company (LILCO) to operate the Sho reham facility at 25% of full rated powe r. (The facility would be so operated until certain emergency planning issues related to the granting of a full power license are resolved. )

Tne specific calculations performed in this work are those speci-fled by the NRC technical sponsor. These include representation of both Loss of Inj ec tion and Anticipated Transient Without Scram (ATVS), the two accident classes that in general dominate the risk of core melt at boiling water reactor f acilities. This general status applies also to Shoreham, for which LILCO has indicated, in its request for 25% power operation, that los s-o f-inj ec tion type accident sequences contribute about 70% and ATWS type sequences contribute about 14% of the total estimated core melt frequency.

Other calculations have been performed to estimate the timing of events for the large-break LOCA accident sequence compoanded by loss of all reactor vessel injection. Although this combination of initiating events represents a much lower fraction of the overall estimated core melt frequency, it is of interest because it leads to the shortest esti-mated time to the onset of core degradation. All accident sequence s tudies include, where appropriate, a calculation of the event timing for the case of initial operation at 100% of rated power, for comparison with the results f or 25% power operation.

l The authors of this report have been involved with severe accident calculations for boiling water reactor facilities since late 1980.

These previous efforts were performed as part of the Severe Accident Sequence Analysis (SASA) Program from October 1980 until September 1986 i and for the follow-on Boiling Water P4 actor Severe Accident Technology (BWRSAT) Program f rom October 1986 to the present. Bo th of these pro-grams have been conducted at ORNL under the auspices of the Office of ,

Nuclear Regulatory Research. A list of the SASA and BWRSAT program re-

~

potts dat the authors of this letter report either wrote entirely or  ;

co-aut?ored is provided as Refs. 1-15. 1 I

All of the calculations whose results are represented in this '

letter report were carried out on a bes t-es tima te basis. This includes the models employed for representation of the decay heat power as a function of time.

, 1.1 j I

i i

2. MAJOR POTENTIAL SEVERE ACCIDENT MITICATION ASPECTS OF OPERATION AT 25% POWER The purpose of this chapter is to provide a brief discussion of

. each of three maj or areac in which operation of the Shoreham Nucle a r Power Station at 25% of rated power would be expected to decrease the severity of postulated severe accident sequences.

2. I Delay in the Timing of Severe Accident Events Reactor scram from 25% power operation as opposed to scram from 100% power operation translates directly into a factor-of-four reduction in the subsequent decay heat power levels. 'Ih i s c anno t , however, be ex-pected to simply be reflected in a f actor-of-four increase in the time periods between successive major events of a severe accident sequence.

Because of the effects of chemical energy release caused by oxidation of the me tals within the core region and the complicated action of the l

{

safety / relief valves in releasing steam f rom the vessel, code calcula-tions are necessary to provide estimates of the extended event timing.

' Nevertheless, there can be no question that the reduced level of decay heat power will have a significant effect upon extending the time available for operator action, equipment repair, or personnel evacua-tion. Re sults of calculations of the severe accident sequence event timing for total loss-of-injection sequences are provided in Chap. 3, Anticipated Transient Without Scram (AWS) event timing is discussed in i Chap. 4, and event timing for large-break LOCA with f ailure of ECCS is provided in Chap. 5.

l l

2.2 Reactor Response to Sudden Reactivity l

Insertion Under ATWS Conditions l The extent of control blade insertion into the core is greater for operation at 25% power than it is for operation at 100% power. Acco r d-ingly, under ATWS conditions, less of the total negative reactivity, which serves in the overall reactivity balance to maintain the core critical, is due to voids. It f ollows that perturbations to the crit-ical state that tend to collapse voids in the core region will insert less positive reactivity with the control blades in their 25% power positions than with the control blades in their 100% power positions.

This matter is important because required operator actions in the f unlikely event of failure of the sodium pentaborate neutron-absorbing poison function are complicated and difficult to accomplish properly.

The operators may' lose control of the situation resulting ht the sudden inj ection of large amounts of cold water or in pressure excursions, either of which causes sudden void collapse. The more sluggish response of the core to void collapse with the control blades in their 25% power position is discussed in Chap. 4.

2.1

r 2.3 Reduced Opportunity for Core Debris-Concrete Interaction The Shoreham thelear Power Station is fitted with four downcomers immediately beneath the reactor vessel within the pedestal region of the

. d rywell. All potential flow pat hways f rom the inpedestal region. to the expedestal region of the drywell are blocked to a level about two feet above the floor. Since the upper surf ace of the inpedestal downcomers I is flush with the floor, the majority of any molten material emanating from the reactor vessel would be expected to flow through the downcomers and be quenched in the pool.

If there is no significant buildup of core debris on the drywell floor, there will be no significant degree of core debris-concrete in-teraction. It is important to recognize that in order for fission prod-ucts d)her than noble gases'to reach the outside plant environment, they must be propelled from the plant by some motive force. This motive force , night be provided by the pressure stored within the reactor vessel before vessel bottom head f ailu re, by the pressure stored within the primary containment before drywell or wetwe.11. pressure boundary failure, by flasning of an overheated pressure suppression pool upon primary con-tainment depressurization, or by the gases released from concrete by the ablation process driven by the hot core debris. In general, BWR severe accident sequences proceed in such a manner that most of the volatile l fission products are trapped in the pressure suppression pool or are condensed upon reactor vessel surfaces; furthermore, the reactor vessel

. and containment pressure sources have been dissipated by the time the core debris begins to leave the reactor vessel. In these cases, it is the core debris-concrete interaction that provides the motive gases, which sparge through the core debris and carry a small fraction of the remaining fission products to the outside environment, bypassing the pressure suppression pool via the previously-failed primary containment boundarf.

Thr lower decay heat power of core debris resulting f rom 25% power operatien would, of course, slow the rate of concrete ablation. On the other hsnd, if all of the core debris were to leave the reactor vessel as molten liquid and flow through the inpedestal downcomers to the pool, there would be no core debris-concrete interaction at all. These matters are discussed in connection with the total los s -o f-inj ection calculation results in Chap. 3.

.. 2.2

L CALCULATIONS FOR LOSS OF INJECTION (TQUV) SEQUENCES In the TQUV (Transient-induced scram, f ailure of normal feedwater system to provide core make-up wate r, failure of high pressure (H PCI, RCIC) and low pressure (RHR, Core Spray) systems to provide reactor ves-sel injection} accident sequence, there is a transient followed by reac-tor scram, followed by f ailure of function of all of the systems that would normally be relied upon to deliver cooling water into the reactor vessel as necessary to keep the core covered. If the reactor scram occurs from 100% power, the uncovering of active fuel begins in

<1/2 h.6 The ";QUV recident sequence has been among the dominant sequences in ainost all BWR probabilistic risk assessments (PRAs).6 Indeed, in the LILCO submission for operation of the Shoreham Station at 25% of full power, the loss-of-injection class of accident sequences is indicated to constitute ~70% of the total. risk of core melt.

It is important to recognize that only a small injection source is necessary to keep the core covered when the only energy release within the reactor vessel is that due to decay heat. This point is disctissed in Chap. 1 of Ref. 6 for the case of scram f rom 100% power, and, of course, even less injection is required in the case of scram from 25%

power. For example, with the reactor vessel remaining at pressure, an injection rate of just 7 7. 5 gpm taken f rom the condensate storage tank at a temperature of 90*F would be sufficient to maintain a constant reactor vessel water level above the core for times greater than 10 min after scram. Since the core would not be uncovered during the 10 min period before injection is assumed to be initiated, an injection rate of 77.5 gpm, which is less than the capacity of the control rod drive (CRD) hydraulic system when the reactor is scrammed, would preclude core un-covery.* Thus, we must include loss of the CRD hydraulic system in addition to the loss of function of the standard inj ec tion systems specified by the TQUV definition if we are to define a basis for consid-eration of a severe accident situation. We will term this heightened level of pumping system degradation " total loss of injection."

• Plant tests at Shoreham have demonstrated a CRD hydraulic system injection capacity of 112 gpm with the reactor scrammed and the reactor vessel at pressure. However, if the control room operator should reset the scram, this inj ection rate would automatically decrease to about 5 7 gpm. This demonstrates why plant procedures and operator training should emphasize the relation between the condition of the shutdown reactor (scrammed or scram-reset) and the available CRD hydraulic system injection flow.

3.1

r .

l l

l l

l 3.1 BWRSAT Program Severe Accident Models 1 l

l The Boiling Water Reactor Severe Accident Technology (BWRSAT) Pro-gram at Oak Ridge National Laboratory employs the seiere accident model-

. ing strategy outlined in Table 3.1. The models desed bed in this table were developed at ORNL by L. J. Ott and have been incorporated into the l Boiling Water Reactor Severe Accident Response (BWRSAR) code. It should '

be carefully noted that this methodology results in significantly longer times to reactor vessel bottom head penetration failure than other methods because of the contention that the large amount of water in the BWR bottom head must first be boiled away and that the debris must then reheat to about 1800*F before the reactor vessel bottom head penetra-tions can fail. It should also be noted that, in general, the debris in the bottom head is not molten at the time of penetration failure and therefore the debris begins to pour from the vessel some time after penetration failure. If the reactor vessel is pressurized at the time of penetration f ailure, then the gas blowd own through the bottom head debris has a cooling ef fect upon the debris.

3.2 The Total Loss-of-Injection Sequence l The sequence and timing of events following a postulated total loss of reactor vessel injection at the Shoreham station as calculated by the BWRSAR code are provided in Table 3.2. It is assumed that the reactor had been operating at 25% power at the time of scram and, in spite of the long times involved, it is assumed that no injection source is ever l recovered. For conservatism in the analysis, there is no modeling of l pressure suppression pool cooling or operation of the drywell coolers.

The reactor vessel is assumed to remain at pressure. l l

Plots of certain key parameters representing events within the re-actor vessel are provided in Fig. 3.1. These plots represent events from just after core uncovery, which occurs at time 138 min, until 1500 min, which is some 24 min af ter reactor vessel bottom head penetra-tion f ailure and the beginning of reactor vessel blowdown to the dry-well. The effect of bottom head penetration failure at time 1476 min upon reactor vessel pressure can be seen on the extreme right side of Fig. 3.la, which also shoss the effect of earlier safety / relief valve actuations that occurred before bottom head d ryout puts an end to reactor vessel steam generation at time 1181 min.

The swollen reactor vessel water level, which includes considera-tion of the effect of voids, is shown in Fig. 3. l b. The calculated level in general follows a smoothly descending curve, with perturbations caused by safety / relief valve actuations, as water is boiled off by decay heat. However, after time 395 min, molten core debris begins to

, fall into the remaining water above the core plate, accelerating the rate of decrease in level. Af ter core plate dryout at time 483 min, the water level remains relatively constant, just below the core plate (whose upper surface is at 199.6 in.). Howe ver, the BWRSAR code does 3.2

recognize the displacement of water in the bottom head whenever lirse debris masses are introduced into it; this is manifested most clearly in Fig. 3. lb by the temporary water level increase shown when radial region two of the core collapses into the bottom head at time 644 min. This temporary level increase is followed by a rapid major decrease of almost 100 in. caused by steam generation as the mass of radial column two is quenched. Subsequently, the decay heat associated with the f uel pellets of radial column two causes a boiloff of the remaining water in the bot-tom head and bottom head dryout occurs at 1181 min.

Figure 3.lc represents the gradual heatup of the fuel rods in radial column three and the subsequent large reduction in temperature that follows collapse of radial column (zone) two at time 644 min. The quenching of the mass of radial column two in the bottom head releases a large amount of steam, which causes accelerated zirconium-water reaction and a concomitant temperature increase in the outer regions of the core where zirconium metal still exists in the cladding. However, all of the remaining clad st ructure in radial column three is zirconiaa ouide at this time, so the steam release associated with the collapse of radial )

column two produces only rapid cooling there. This ef fect is important, l because wi thout it, radial column three would soon reach temperatures that would cause its remaining zirconium oxide clad structure to lose strength and it would also fall into the bottom head, instead of remain-ing in place.

1 i

Plots (d), (e), and (f) of Fig. 3.1 pertain to the extent of hydro-gen generation by the zirconium-steam reaction. Some 51% of the clad J and 39% of the channel box wall is predicted to be oxidized during the 1 accident sequence, producing about 1840 lbs of hydrogen. (This demon- I strates the wisdom of maintaining an inerted primary containment atmos- l phere during reactor operation.)

)

Selected primary containment response characteristics are displayed in the individual plots provided in Fig. 3. 2. The primary containment l pressure (Fig. 3.2a) shows significant increase in response to the col-lapse of radial column two into the reactor vessel bottom head at time 644 min and the associated discharge of noncondensible gases. Th e c on-comitant drywell atmosphere temperature response (Fig. 3.2b) is slight because this discharge f rom the reactor vessel is via the safety / relief valves and is cooled during its passage through the pressure suppression pool. Both the pressure and temperature response are marked, however, following the inception of reactor vessel blowdown at time 1476 min when the release is directly into the drywell atmosphere. thvertheless, the peak pressure is not of a magnit ude to threaten primary containment pressure boundary integrity and the temperature excursion is rapidly terminated as the heat sinks absorb the released energy. This does not threaten the integrity of the drywell liner, however, since its peak calculated surface temperature in response to the reactor vessel blow-

, down is only about 270'F, as shown in Fig. 3.2c.

Certain calculated conditions within the wetwell are shown in Fig. 3.2 plots (d) through (f). As indicated in Fig. 3.2e, a large mass 3.3

1 l

I l

\'

of hydrogen is collected in the wetwell airspace when the reactor vessel

~

is flushed by the large release of steam from the bottom head at-the time of collapse of radial column two. However, it should be recognized  ;

that this is not all of the hydrogen in the containment; some 450 lbs i are predicted to have been passed back into the drywell via the werwell-to-drywell vacuum breakers.

Figure 3. 3 indicates the integrated mass of molten material pre-dicted by the BWRSAR code to have been released f rom the Shoreham reac-tor vessel as a function of time following reactor vessel pressure boundary failure at time 1476 min. There is a small initial release of f molten stainless steel immediately upon bottom head penetration failure, l but the passage of blowdown gases cools the debris and significant addi- f tional release does not occur until about time 1910 min. The calcula-tion is carried out to time 2880 min (48 h af ter scram), although the l calculated reactor vessel wall temperature at the end of the calculation j (about 2210'F at the connection with the vessel skirt) is much too high i for the wall structure to have survived. Indeed, it must be expected that the reactor vessel wall would fail shortly af ter time 2100 min, -

when the average wall temperature in the vicinity of the vessel skirt is  !

about 1600*F.

~

For accident sequence time 2106 min, the BWRSAR code predicts that some 101,000 lbs of molten metal eutectics have left the reactor vessel whereas about 556,000 lbs of solid debris remain within the vessel, in-cluding all of the original inventory of UO2 fuel. Should the reactor vessel bottom head wall fail by creep-rupture at about this time, a great deal of solid debris would be deposited onto the drywell floor.

The results of these calculations do not support the contention of the LILCO submission that all debris f alling onto the Shoreham drywell floor would be molten, subject to immediate passage through the downcome rs into the pressure suppression pool.

3. 3 Comparison with Total Loss of Injection From 100% Power In order to clearly demone.trate the ve ry significant delay in severe accident event timing associated with operation at 25% of full power, the total loss of injection sequence was recalculated with all parameters the same except for the initial power, which was set at 100% ,

of rated (8.312 x 109 Btu /h). The dif ferences in timing of the maj or  !

events of the accident sequence are indicated in Table 3.3. l I

In addition to the large delays in severe accident event timing associated with operation at 25% 'of f ull power, there are two other sig-nificant differences in the calculated plant response. Firs t , more hydrogen is generated in the case of total loss of injection f rom 25% >

power, 1840 lbs as opposed to 1620 lbs for the 100% power case. The relative fractions of clad, channel box, and control blade - stainless  :

steel sheath oxidation are provided in Table 3.4. The major source of I

\

I 3.4 l

the additional hyd rogen is seen to be the additional oxidation of the

, channel box walls, while they are slowly heated at oxidizing temperature ,

in the 25% power case. (In the 100% power case, the channel box wall  !

reaches its melting temperature more rapidly and more of the zirconium metal is relocated downward without oxidizing.) Howe ve r, a great deal of hydrogen is predicted to be generated in either case so the addi-tional 14% in the 25% power case, while of interest, does not constitute a special concern.

The second significant dif ference in calculated plant response is also derived from the much slower heatup rate in the 25% power case. j Af ter bottom head dryout, the debris in the reactor vessel bottom head is heated by its internal decay heat generation. For the 100% power case, the debris heats rapidly and its temperature greatly exceeds that  !

of the reactor vessel wall, which is heated only by conduction f rom the debris. Fo r the 25% power case, however, the debris temperature in-creases much more slowly and the average reactor vessel wall temperature does not lag so far behind. The upshot of this is that relatively [

little of the debris has become molten and lef t the vessel at the time that the average wall temperature reaches the creep-rupture f ailure range in the 25% power case whereas most (but not all) of the debrio has left the vessel by this time in the 100% power case. A comparison of l

. the calculated quantities is provided in Table 3. 5. This matter is  ;

impo r tant bccause the debris leaving the reactor vessel in the molten liquid phase is expected to immediately enter the pressure suppression

~ pool via the downcomers whereas most solid material falling on the dry- (

well inpedestal region floor upon creep-rupture f ailure of the bottom [

head would not. >

k It should be noted that the models now employed for the BWRSAR cal- l culation for definition of the formation of eutectic mixtures within the ,

core debris and the melting temperatures of these mixtures are current best-estimate and s ubj ect to revision upon completion of certain per- f l

tinent experiments now planned by Dr. Dana Powe rs at Sa ndia Na tional l Labora torie s .

i i

I l

i 1

1 l

3.5 I I

l

i Table 3.1. BWRSAT program me thodology employed to represent events between onset of core degradation and reactor vessel f ailure for BWRs

1. As canister and control blade material becomes molten, it is relo-cated onto the core plate. This causes:

a. a temporarily increased steaming rate

b. core plate dryout and cessation of steaming ,

c. buildup of mass on the core plate and core plate heatup.

2. Each radial region of the core plate fails due to loss of strength when its calculated temperature has increased to 12 7 5'F (964 K).

Each -core place region and its accumulated debris falls into the lower plenum, producing a burst of steam and lowering the water level there as the fallen material is quenched.

3. Molten Zr metal flows downward over the lower core fuel rod nodes, leaving the UO2 fuel pellets encased in thin Zr02 sheaths. Steam rising from the lower plenum cools the core nodes from which all un-oxidized Zr has been removed. On the other hand, the rising steam j causes energy release in the core peripheral nodes where Zr metal at  !

elevated temperature still remains. l i

4. The standing portions of the core f all into the lower plenum by ]

radial column. Each core column collapses when its average clad j temperature reaches 4250*F (2616 K), at which time very little of '

the UO2 mass in the region has become molten. (The actual failure  !

mechanism is weakening, by overtemperature, of the Zr02 sheaths sur-  !

rounding the UO2 fuel pellets.) The f alling mass is quenched by the water in the lower plenum until the time of bottom head dryout. 1 Af ter bottom head dryout, the debris'begins to reheat.

5. The structure of the control rod guide tubes in the. lower plenum is a heated by the surrounding core debris and is weakened to the point  ;

of failure when its temperature reaches 1400*F (1033 K). Failure of j the control rod guide tubes causes all remaining standing portions I of the core to immediately collapse.  !

6. Bottom head penetrations fail by a simulated creep-rupture mechanism as the debris mass in their vicinity is reheated to about 1800*F 'i (1255 K). The reactor vessel depressurizes and equalizes with dry- l well pressure.  !

7. The individual components of the debris mass leave the vessel only after they have reheated to their melting points and thereby become liquid.

P

7 Table 3.2. Major events of the total-loss-of-injection accident sequence f or scram f rom 25% power ime ter scram Event (mins)

MS1V closure and total loss of 0 injection The swollen water level f alls below the 138 top of active fuel

' Structural relocation of core material 395 begins Core plate dryout 483 Core plate region i fails 522 Core plate region 4 f ails 566 Core plate region 2 fails 641 Radial region two of core collapses 644 Core plate region 3 fails 644 Core plate region 5 fails 856 Bottom head dryout 1,181 Radial region one of core collapses 1,476 Bottom head penetrations fail and blow- 1,476 down into drywell begins High temperature weakens control rod 1,476 guide tube structure and remaining portions of the core collapse. A small amount (13,500 lbs) of molten steel leaves the vessel. Die re-mainder of the core debris is frozen Beginning of continuous pour of molten 1,910 material from the reactor vessel Average reactor vessel wall temperature 2,100 exceeds 1600*F in vicinity of attach-ment to supporting skirt. Creep rupture failure of vessel wall and partial separation of lower portion of bottom head is probable O

i i

Table 3.3. Comparison of the timing of major events j in the total-loss-of-injection accident sequences i for initial powers of 100% and 25% j Time af ter scram  !

Event (mins) 100% power 25% power J MSIV closure and total loss of injection 0 0 Swollen water level falls below top of 25 138  !

active fuel 4 Structural relocation of core material 68 395 begins Core plate dryout 74 483 Radial region two of core collapses 110 644 Bottom head dryout 231 1181 Bottom head penetrations fail, blowdown 267 1476 ,

into drywell begins '

Beginning of continuous pour of molten 267 1910 3

, material from the reactor vessel j Average reactor vessel wall temperature 610 2100 exceeds 1600*F in vicinity of attach-ment to supporting skirt. Creep rup-ture failure of vessel wall and partial separation of lower portion of ,

bottom head is probable "

l i

Table 3.4. Comparison of the calculated hydrogen generation parameters in the total-loss-of-injection sequences for initial ;owers of 100% and 25%

Parameter 100% power 25% power Fraction of clad oxidized 0.48 0,51 Fraction of channel box wall oxidized 0.23 0.39 Fraction of control blade stainless steel 0.06 0.09 1, sheath oxidized Total hyd rogen generated, lbs 1620 1343

)

i

~

e dw l k i e

e S.

i t h

l l

l I

i Table 3.5. Comparison of the calculated quantities l of core debris invessel and exvessel at the j time that the reactor vessel wall temperature exceeds 1600*F j 1

e 100% power case 25% power case l time 610 min time 2100 min Debris constituent Invessel Exvessel Invessel Exvessel l (1b) (1b) (1b) (1b)

Zr 22,600 48,700 44,400 18,000 Fe 27,200 156,200 127,200 55,900 i Zr02 35,600 17,100 65,200 0 UO2 42,600 235,600 278,800 0 1 Other# 14,100 53,600 39,300 27,600 l Total 142,100 511,200 554,900 101,500 Cr, N1, B4C, Fe0, Fe304, Cr203, NiO, and B203 l i

i 1

l 4

4 l

e~ E

• n .

" " " a u

. = = -

=

• =

8$ -

- R W .@ A m .t S .

.:s ) R y

• N E (

)C O f U t

(

"U M .EM I Bi l

T ei h c t a

" - y f

b r a

• - d e el t c

• - c u

" . . - i n

• = = .

.a = m = " "' - = m .

. d s

m 8 . .

,"3 em ra ph

{ zb =P52 5;oja2i< ': g,e:!me fN - e sr ao h

l S e

s e s" sh et

"' , v

" = ,* . t ra o

t r

- c e

. a n

aw eo _

. . r p

"* ~

e% _

S h5 _

• g .I N

W( )

t2 .

• u m

(

E b

) .

E M

(

e nm i o

( .I

.T h r

• Mt T tf .

i wma

• sr t c ns e _

• vg _

eni rw _

= .

= _ _ a. -- oo

• . m

= = . . s . . fl l

g I- so g_= a!a xo5g , ef ub 2dsI9 a#g E 5d: E r un t o ai nt gc 2 $. i e sj v

a n

n

. n t i

u. - n ef d o

. . i

a. cs c s

. A o

.i l i _

.N S .l S .ts 1 a .

.N ) . t O

t u) t d

(

3 o E a .Eu M t _

M I

( gr T

. / i o

. Ff m e d

o a

. c

. . o a o. _ -- - n -.

n "" - "* ** " m i

. = .

e 2@,

g a: 2 ,o6<t

,=gE d@>

b s s = = =

.

. =

n = "' =

t a

s g1 c -

J se -a

- R

, A

. s o = S e i R

W

. B

. . o s

=

a =

e

, S

.N sI

)

M

)

c f -

=a x t

u

)

g h

t y .

( ( t

( by E

.EM a M t

. .I sT -

=T I di el ti s

^

= cc ia df e

s

= rr pa a

e

.e a sl i

"' ac a =

n a

n i = n i

a n = = ua . ,

a u

• l n b E l em wa gz3b 5ags'"$ggx e g EI z ;: . =Egeg th ee wro dh b E- nS a

e

. =

s =

s

• a. =

u . = = =

a

= ,

, y. lh p .

s sa lt e

. 1 a

wt

. . ya

s. . .

n r a

dr e

m ew s

a i ho _

t p _

.N

)

S

) .es

. ) f%

e t

M b .t u (

o5

(

( ( 2 E E

.M

.I

" =ti M

em so vT ..

> nr of _

s u p sm _

ea

u. .

r r c -

e s w.,a

-. = . . * . g h

t g

n .

. n = s

• 3 '

= . _ = s .

. " ri _

ow 6 S3 f o l HE < 2 $" % gg e cE$ g U*  !

o ]5 l t

so e

r f -

u t n 6 g a io nt gc

= = - 3

= . = *** = u. = i u n s je

a. 3 n , . t S .

s a ua t i n -

. . n sa ni ef d o i

cc ss a

1 6

f t

M )

7 .

)

..tw. )

A o l

( a (u d

.l e .T E

M

( E M

(

2. at

- I s .

.T t 3o

.t a -

. . gr .

s .

= i o Ff

= .

n e

d _

s o

.a l.

c a .a .

a ,, .a . . . y-u = = = a* .

e 3$~ e-js. .

% d.$O % g>Bc [0 u:h< 2 5 a ucg2}s--

Ifj 1!i, i 1

aR ed rs cAhaei Y oT y- seSt ewh ajR h ot nW f l 2

0 2 2 2 2 lib e o oerm e t n e 4 2 sf e 0 th o 0

6 soh Tl e l t o t fe 2

vs Ab t b s a -

ro . l ol y e h l t e 0

t cl i t .s s s

. 0 4

2 a eti a lrvgv )s e h es roc t a oeh r u i e f hh o ha t ot 0

t r sc

. 0 2 t oe4f a8 n a N 2 f is e O S e) c l( hir _

I t T A NI l su en ew sierhe TNM gt nua m untut t

S06 llr7 l

0 0 in _

. EC1 0

2) vih neam ame iz r o r moe r WE 1

S h OJT7 N 0 o rpf PNA8I 0 I

M s8h0 f m ed lFl S 1, 0( i4S8 t

0E r2 8 nt e

. AOl5 E A2 8

M b e2 i t _

LSF 1 eoh al a CSlY I

T dt( t eemla r ap _

UDlA AM re t imr w s e e NIE LI os at MA I 0 ca tod ATL . 0 8 h r et t e ev lOE 1 np wad a ETSS l

e el ah t t r o f

0 F i l ep'd ct uc 1 V ot %ei l t 1

S 0 mt td a s

. 0 4

1 fl a5aec2 nr u m o i p sho r em m h e d st r e lat a e mg f t 0

a ,h 0

ne Y li csr e r mo 2

E arn t uewt e tt ao v 0

od u q T

n ew o ee n v h ob ms o n 0

- - 9 i

1 '

.i t i

h e

6 5 5 i 1 s.

n 5

5 s.

2 s,.

s.

a 3 3tt nl a 2 et 2 2 1 m i 1 3f e uO i f

.od cU t o i

7ET >

MS- gnc al l

ioc ca nio Fia sn

- 4$y*p Mo n ej% yeQ7eam$ n gg<$ 0- g8 t i h i .

_ cnegt t e nodi uiort oi rm ft coapt

4. ATWS CALCULATIONS This chapter describes the predicted response of the Shoreham

, Nuclear Power Station to a postulated complete f ailure to scram follow-ing a transient event that has caused closure of all main steam isola-tion valves (MSIVs ) . This accident sequence is the most severe of a class of sequences commonly denoted "ATWS", the acronym for " Anticipated Transient Without Scram." With the MSIVs closed, almost all of the steam exiting the reactor vessel would be passed into the pressure sup-pression pool through the safety / relief valves (SRVs); the remainde r would be used to drive the High Pressure Coolant Injection (H PCI) or Reactor Core Isolation Cooling (RCIC) system turbines during their periods of operation and then, as turbine exhaust, would also enter the pressure suppression pool. Since the rate of energy deposition into the pool can greatly exceed the capacity of the pool cooling equipment, the possibility of excessive pressure suppression pool temperatures leading to primary containment f ailure by overpressurization is of primary con-cern during the analysis of ATWS accident sequences.

As in all designs, the criticality of the Shoreham nuclear reactor depends upon a complicated set of f actors that simultaneously introduce positive or negative reactivity. Whether there is a power increase, constant powe r, or a power decrease at a given point in time depends upon the particular reactivity balance at that instant. In BWR studies , j

. it is necessary to recognize the importance of the void coefficient of  !

reactivity. In the BWR, boiling takes place within the core and " voids" are created by the steam bubbles f ormed within the core volume. The moderation or slowing-down of neutrons is much lesc in steam than in liquid water so increased voiding has the effect of re 'ucing the supply l of thermal neutrons. The ref o re, an increase in voids inocoduces nega- J tive reactivity . and a decrease in voids introde on posit ve reactiv- )

ity. Since the BRR operates with the water moderator at saturation con- '

ditions within the core, negative or positive reactivity insertions caused by the creation or elimination of voids are a natural and  !

important result of reactor vessel pressure changes.

1 Provision is made for rapid reactor shutdown under emergency condi- '

tions by neutron-absorbing control blades that can quickly and auto-matically be inserted (scrammed) into the core upon the demand of the reactor protection system logic. When inserted, the control blades introduce enough negative reactivity to ensure that the reactor is main-tained suberitical even with the moderator at room temperature and with zero voids in the core.

Although all transient-initiated accident sequences can most easily be brought under control and terminated by successful scram of the con-trol blades, they can also be brough t under control and terminated by 1 apprc9riate operator action. In other words, given properly trained operacors and properly functioning equipment, the f ailure of the scram function can be considered to be merely a nuisance requiring a more com-plicated and time-consuming method of achieving shutdown.

4.1

1 One important action that the operators are directed to take in the event of ATUS is to initiate the standby liquid control system (SLCS) if the pressure suppression pool temperature reaches 110'F. This system injects a neutron-absorbing poison of sodium pentaborate solution into the reactor vessel by means of positive displacement p ump s. Previous studies ll at ORNL have demonstrated that with operator action limited to successful initiation of the SLCS, the ATWO accident sequence can be brought under control. If a postulated ATW5 accident sequence is to degenerate into a severe accident, then, we must also assume f ailure of the SLCS.

4.1 Operator Actions if the Poison Injection Function Fails The BWR Owners Group Emergency Procedures Guidelines (EPGs) provide a strategy for operator actions to deal with the MSIV-closure initiated ATWS that can be summarized as follows: At temp t manual scram and, if not successful, begin manual insertion of control blades. Initiate the SLCS anc pressure suppression pool cooling. Reduce core power by taking manual control of the reactor vessel injection systems and lowering the reactor vessel water level to the top of the core; this reduces core inlet flow by interrupting the natural circulation path from the core through the steam separators and back through the jet pumps in the down-comer region. The result is increased voiding in the core and increased t emp e rature of the flow at the core inlet; the latter because the lowered water level uncovers the feedwater spargers through which the HPCI and RCIC systems inject, thereby providing heating of the injected droplets as they all through a steam environment.

5 The EPGs are intende$ to be symptom-oriented instructions to the contuoi room operator thats are comprehensive and cover every eventual-i ty ., To maintain assuranle that the thermal energy released from the pritspry system can be cond:nsed in the pressure suppression pool, there is a+ requirement that reactor v ssel pressure be reduced as the pressure supdfession pool temperature inc eases. For the Shoreham reactor, pres-surd reduction must begin when the suppression pool temperature exceeds 150]? and continue in accordance with a graph of permissible maximum reactor vessel pressure vs suppression pool t emp e rature. (Previous studies at ORNL7 demonstrate for Brovns Ferry that once reactor vessel dertiessurization is begun, it must be continuous because each increment of : energy deposited in the pool during depressurization increases the suppression pool t emp e ratu re to the extent that, f ollowing the graph, further depressurization would be required.) There is no suggestion in thu EPGs that the graphical schedule for reactor vessel depressurization as pressure suppedssion pool temperature increases should not be fol-lowed in the event of ATWS.

Recent preliminary work at Rensselaer Polytechnic Institute (RPI) indicates that the Shoreham nuclear reactor would be suberitical with the reactor vessel water level lowered to the top of the core and the reactor vessel depressurized to 200 psi or less, given the control 4.2

~

blades stuck in their normal 25% power operating positions.16 In other words, the small rate of water injection to the reactor vessel required to maintain the vessel water level at the top of the core would be converted to steam by decay heat alone, and the associated voids within

, the core region would keep the core suberitical. If subsequent analyses confirm this finding, then, should the SLCS system f ail, the Shoreham reactor could be brought suberitical simply by lowe ring the reactor vessel water level and pressure, while keeping the core covered. It is emphasized that this would not be the case if the control blades were stuck in their (more withdrawn) normal 100% power operating positions.

4.2 Consequences of Low-Pressure Criticality The question as to whether or not the Shoreham reactor would main-tain criticality with the water level lowered to the top of the core and the reactor vessel depressurized is of some imp o rtance because of the potential for rapid insertion of positive reactivity under such condi-tions. This might occur either by uncontrolled injection of cold water by the large-capacity low pressure injection sys t ems or by any small action that initiates a reactor vessel pressure increase. As indicated

' by the data provided in the steam tables, the change in steam vapor specific volume for a given increment in pressure is much greater at low pressures.

If it is considered that the Shoreham reactor would remain critical in a depressurized state with the water level lowered to the top of the core, then it can be demonstrated, in general, that the power and pres-sure transients resulting f rom sudden insertions of positive reactivity would be less severe with the control blades stuck in their normal posi-tion for 25% power operation. To this end, two sets of demonstration calculations with the BWR-LTAS code 9 have been performed f or ATVS situa-tions identical except for the control blade positions. The determina-tion of the additional control blade reactivity associated with 25%

power operation is described in the Appendix. In both cases, no opera-tor action is assumed and the calculations are based upon Browns Ferry parameters.

The calculated results for the two cases are shown in Figs. 4.1 and 4.2. Without operator action, HPCI, RCIC, and CRD injection maintain reactor vessel water level high above the top of the core until the HPCI turbine is lost upon high lube oil temperature (the system lube oil is cooled by the water being pumped, and without operator action, the sys-tem suction would automatically be shif ted f rom the condensate storage tank to the_ pressure suppression pool; HPCI system f ailure is assumed at a pool temperature of 210*F).

' With HPCI system f ailure, reactor vessel water level decreases even though the RCIC and CRD hydraulic sys tems continue to inject. The automatic depressurization system (ADS) is enacted when the combination of low reactor vessel water level and high drywell pressure is sensed.

4.3

l l

l l

This is much later for the case of ATWS f rom 25% power (Fig. 4. 2b vs.

, 4.lb, where six open SRVs represent ADS actuation).

As the reactor vessel is depressurized into the regime in which the large-capacity low pressure injection sys t ems are able to pump cold water into the vessel, large cycles of injection - no injection begin l (Figs. 4.lc and 4.2c). Each period of cold water injection produces void collapse within the core and subsequent reactor power spikes (Figs. 4.1d and 4.2d). However, the power spikes produced for the case l with the control blades in their normal 25% power position are much less '

severe and less numerous than for the case with the normal 100% power control blade positions.

Each power spike produces a pressure increase within the reactor vessel that quickly terminates further water injection by the low-  ;

pressure systems. Without injection, the core rapidly becomes uncovered l and power drops to decay heat levels. Howeve r, as long as the ADS valves remain open, the pressure then rapidly decreases and another cycle of cold-water injection begins.

If the increasing drywell pressure approaches the available control air pressure, the safety / relief valves will close, the reactor vessel

, will repressurize in a perntnent f ashion, and no additional low pressure l inj ection cycles will occur. As shown in Figs. 4.lc and 4.2c, this '

happens at about time 39 min in the case of an ATWS f rom 100% power, but

~

not within 45 min for ATWS f rom 25% power. The associated drywell pres-sure traces are shown in Figs. 4. le and 4.2e.

The conclusions obtained by this discussion and the associated cal-culations are that adequate mitigation of an ATWS accident sequence, '

given f ailure of complete shutdown, is much simpler for the control room operators if the control blades are in their configuration for 25% power operation as apposed to their configuration at 100% powe r. Since the relative negative reactivity contribution of the core voids is less with the blades in their 25% power configuration, any subsequent change in voids also has less effect. Ferhaps the best evidence of the overall effect of this is the difference in predicted pressure suppression pool temperatures at time 40 min in Figs. 4.lf and 4.2f.

It should be noted that the results presented in this section for ATWS f rom 25% power operation were calculated on the basis of a negative 0.015 delta-K/K control rod reactivity differential associated with operation at 25% power as opposed to 100% power. This is considered to be a conservative basis since, as discussed in the Appendix, the best-estimate negative reactivity differential is -0.019. A test BWR-LTAS calculation was performed using -0.019 delta-K/K control rod reactivity, but the calculated reactor power was so low following loss of the HPCI system that the low reactor veesel water level required for ADS actua-tion was not reached within the 45 min period represented by the calculation.

4.4

4.3 The C9D ATWS Sequence Together with their application for Shoreham to operate at 25% of rated power, LILCO submitted the results of several accident sequence calculations in which events were assuced to occur in such a manner that the accident procee.ed beyond core melting and reactor vessel pressure bounda ry f ailu re. among these is the C9D ATWS accident sequence, in which the control blades are assumed stuck in their 25% power operating con figu ra tion.

The sequence of events for the C9D accident sequence up to the time of loss of injection as described by the LILCO submission is described in Table 4.1. Several conservative assump tions were employed by the LILCO subcontractors in determining this event sequence including fail-ure of the SLCS function and loss of all reactor vessel injection sys-tems upon the inception of wetwell venting, as well as the basic assump-tion that the reactor would be critical when depressurized with the water level at the top of the core. There would, of course, be no severe accident if any one of these assumptions is not valid.

To determine the sequence of events involving core degradation and reactor vessel f ailure that would occur af ter loss of all injection to the reactor vessel, a BWRSAR code calculation has been performed at ORNL for the period of the accident sequence af ter time 90 min. Co re powe r was controlled by user input and the control blade positions were established so as to app roxima t e the actual 25% powe r configu ration.

The predicted timing of events is provided in Table 4.2. Calculated hydrogen generation is 1866 lbs, resulting f rom oxidation of 57% of the clad and 41% of the channel box walls.

The ORNL be s t-e s tima te methods of calculating degraded core and reactor vessel f ailure events (Table 3.1) predict a much longer time to reactor vessel pressure boundary f ailure ( 3 0. 8 h vs . 10. 4 h) than does the MAAP code method used in calculaticg the results of the LILCO sub-mission. (This is prima rily because of the BWRSAT Program contention that core debris relocated into the reactor vessel bottom head is quenched as long as there is sufficient water remaining to do it.)

Also, the LILCO submission predicts that molten core debris begins to leave the reactor vessel at the time of pressure boundary f ailure where-as the ORNL method delays the initial pour until the debris in the bot- '

tom head has reheated to its melting point. Th u s , in the ORNL results calculated with the BRRSAR code, the initial pour of molten metal f rom  ;

the reactor vessel does not occur until time 37.5 h. l If it is desired to compare the event timing given in Table 4.2 for the ATWS case to the timi ng provided in Ta ble 3.4 for the loss-of-injection cases, one must remember that only about 24% of the total con-trol blade mass is represented as being in the core for the ATWS case-i this has a significant effect upon the calculated results. l l

I e

, 4.5 l l 1 i l l l l

I

l I

As a final comment in regard to the lessons learned by review of the C9D ATWS severe accident sequence, it should be noted that the pri-  ;

mary containment atmosphere would be depleted of noncondensible gases  ;

af ter several hours of wetwell venting. If the situation within the  !

. reactor vessel were subsequently brought under control and the wetwell l' vent was then closed, a vacuum greater than the design basis might be j drawn within the prima ry containment as the steam therein condensed. i This is why plant emergency procedures should ensure that the composi- j tion of the prima ry containment atmosphere is ascertained and that  ;

nitrogen is introduced as necessary before the vent is closed.

l l

l i

J i

I e

4.6

Table 4.1 Major events of the C9D Anticipated Transient

, Without Scram (ATWS) accident sequence up to the time of loss of injection Time after Event MSIV closura (mins) i MSIV closure and f ailure of scram function 0.0 1

HPCI/RCIC/CRD maintain reactor vessel level between 0.0-18.0 Level 8 and Level 2 l Two RHR cooling loops aligned for pressure suppression 10.0 pool cooling (

Pressure suppression pool temperature reaches 200*F, 18.0 ADS is manually actuated, vessel depressurization causes loss of HPCI operational capability. Op e ra to rs j lower reactor vessel water level to the top of the core i and maintain it there by use of RCIC and CRD. Re actor power is 107. of full power. The ADS valves automati-cally open and close to maintain the reactor vessel pressure between 50 and 75 psi higher than that of the j wetwell. '

f

~

Containment precsure reaches 75 psia. Operators vent 90.0 l the wetwell airspace causing harsh environmental condi- i tions in the secondary containment. All reactor vessel j injection capability is assumed lost. Also, since the drywell pressure exceeds the available control air pressure, the ADS function is no longer operable and the (

reactor vessel repressurizes. He reaf ter, the SRVs will i open only automatically, on high reactor vessel pressure.  ;

1 1

i i

l i

l l

I Table 4.2. . Maj or events of the C9D Anticipated Transient

- Without Scram (ATWS) accident sequence af ter loss of injection

--~

Time af ter.

Event scram (mins)

The reactor vessel repressurizes quickly and the core 90 is uncovered rapidly as core power is restored to 10%

of full power by void collapse  ;

100 The swollen water level f alls below the top of active fuel

)

Low water level and increasing voids restore core power 103 to decay heat levels

.8 Structural relocation of core material begins 292 1 i

Core plate dryout 518 j Radial region two of core collapses 565

]

Radial region three of core collapses 920 Core plate region 5 fails 923 i

. Core plate region 1 fails 1062 i Bottom head dryout 1337 Overtemperature weakens control rod guide tube structure 1723 and remaining portions of core collapse Bottom head penetrations fail and blowdown into drywell 1850 i begins I Molten core debris begins to pour f rom the reactor vessel 2250 j Average reactor vessel wall temperature exceeds 1600*F .2700 i in recently of attachment to supporting skirt. Creep rupture failure of vessel wall and separation'of lower portion of bottom head is probable

l l

l g MM 0**

• 4 W SPCI i

, 1

. . . . . . . . . . _ . . . . . . . . . . . . g, l w w . anes eva

[ .

~ N, M

'l=

.. . . . . --.*=*..-.'*m*******

I . i4i w <t  : :gg L -

.B I

MAAwwuuq  ! !!

=

Yrw l

.; . . i. i. - - - -

me man memm (a) (b) 1

~ , 1 l

. u .

. M

• l T

M F .u.

.5 . .

\

E ,, go . i' d

um w een z

u -

i i i ,

j

)

i

.g . . i AWAM ,

l u .

= u .

. . = = -

_ [1

=

4.<44<<<<<w mswm ms oam [

1 (d)

(c) i

" su l- .

- .t.

s' s .

,.p'- g Y

j i-g-

g .

y./_. .

,,, g s-i .

B.

.t j; $

+ .,, .

g .

L.

. ,,*,/*

/

. . s. i. .. a. -..

mamm . . a i. - - u -

mx onn (e) (f)

Fig. 4.1. Accident signatures for the MSIV-closure ATWS from 100%

power operation, first 45 minutes with no operater action.

u. . ua

~

WM'

. s . ua

, g. . . . . . . . . . . . . . . . . . . . . . . . . . .

. m -. n. g

u. -

I i.

g .. . T

( . .. .n.

.h

• b 3

JMl11.1111 .

.' 5 A.

. i. .

newwm

l M E WIN) 7043 WIN)

(a) (b)

. u .

. u .

. u .

. u .

E= .

$u .

3 --

3 . 2 ..

. . )

5 .

fu

) )

m. . u .

u I

-+- J. .

~-

(((

. . .. . . . . . . . i. . .

w Woo Maomo (c) (d)

n. . u.

i-1u .

- m

g. .

n==-...-.....-.- gn - . l

g. .

x -

[a ,, . .

.,,..,,- . i. i, l

. . ',, /,. .

i.

'. . i. .

gg . . a .

a

• MS NIN)

(e) (f)

Fig. 4.2. accident signatures for MSIV-closure ATWS from 25% power operation, first 45 minutes without operator action.

5. CALCULATIONS FOR LARGE-BREAK LOCA WITH LOSS

, OF ALL ECCS (AE) SEQUENCES The large-break LOCA initiated by double-ended rupture of a 28-in. I (outside-diameter) recirculation pump suction line serves as the design basis accident to establish the adequacy of the Shoreham containment ,

with regard to drywell internal pressure, suppression chamber internal l pressure, drywell floor differential pressure, suppression chamber temperature, and secondary containment pressure. The configuration of the break and the location of the reactor vessel within the containment are shown in Fig. 5.1. The response of drywell pressure, suppression chamber pressure, and drywell floor differential pressure and the i associated drywell and suppression chamber atmosphere temperatures with- l in the first 30 s af ter the break are shown in Fig. 5. 2. All of the components of these figures were taken from the Shoreham Updated Safety Analysis Report (USAR).

At the 30-s' point, ECCS injection to the reactor vessel would begin l (via core spray) and the accident sequence would be brought under con-  ;

trol and terminated without degradation of the original core geometry. I Nevertheless, in the very unlikely case of failure of the ECCS function,

, the resulting accident sequence (termed AE) provides the shortest period of time to the onset of core degradation of any. Calculations with the BWRSAR models to es tima t e these times for operation of the Shoreham

~ reactor have been performed for both 100% power operation and 25% power i operation. Although the BWRSAR models were never intended to calculate l the complex processes of la rge-b reak reactor vessel blowdown, their )

results for the first 30 s of blowdown are in good agreement with those shown in Fig. 5.2. The calculated event timing for the AE sequence taken f rom the BWRSAR results for both 100% and 25% power operation is provided in Table 5.1.

Less hydrogen is generated in the AE sequences than in the total-loss-of-injection sequences because less steam is retained in _the reac-tor vessel to fuel the metal-water reactions. BWRSAR program results pertaining to hydrogen generation are provided in Table 5.2.

With recognition of the importance of the calculated reactor' vessel wall temperature in estimating the time at which the remaining solid core debris could no longer be supported by the reactor vessel bottom head, a finer bottom head debris nodalization was developed and employed for the large-break LOCA calculations. This involved implementation of a small node structure near the reactor vessel wall to represent the '

effect of debris in close contact with the w'all; the' eff ect is to pro-l duce higher calculated wall temperatures.

The predicted distribution of core debris invessel and.exvessel at the time of probable bottom head failure is provided in Table 5.3. _ Hore of the debris remains within the reactor vessel bottom head in the 25%

power case, but the difference is not as marked as in the calculation

, for total loss of injection, which used the original debris noding 5.1

structure to produce the results showe in Table 3. 5. It is intended

. that all future calculations will use the new debris noding structure.

1

)

1

)

i 1

m 5.2

l Table 5.1. Comparison of the timing of major events in the large-break LOCA accident sequences ,

for initial powers of 100% and 25% 1 Time af ter scram Event (m ns) l 100% power 25% power MSIV closure and total loss of injection 0.0 0. 0 Recirculation gump suction line break 0.002 0.002 (area 4.22 ft )

Swollen water level f alls below top of 0.05 0.06 active fuel Core plate dryout 0.09 0.10 Structural relocation of core material LO.9 59.5 begins First collapse of a core radial region 44.8 200 Second collapse of a core radial region 45.3 201 Bo ttom head dryout 59.2 220 High temperature weakens control rod 59.3 349 l

guide tube structure and remaining i

portions of the core collapse l

First failure of reactor vessel bottom 70.4 469 l

l head penetrations Beginning of continuous pour of molten 80.0 570 material from the reactor vessel m

\

Table 5.2. Comparison of the cales. lated hyd rogen  !

, generation parameters in the large-break LOCA sequences for initial powers of 100% and 25%

Parameter 100% power 25% power Fraction of clad oxidized 0.38 0.35 Fraction of channel box wall oxidized 0.28 0.17 Fraction of control blade stair less steel 0.08 0.04 l sheath oxidized Total hydrogen generated, lbs 1326 1157

~~ .-

Table 5.3. Comparison of the calculated quantities of core debris invessel and exvessel at the time that the reactor vessel wall temperature exceeds 1600'F 100% power case 25% power case j time 200 min time 810 min l Debris constituent Invessel Exvessel Invessel Exvessel (lb) (lb) (1b) (lb)

Zr 58,600 15,700 75,200 6,400 Fe 77,900 105,300 163,900 19,800 $

Zr02 47,900 500 38,500 0 l UO2 270,400 8,400 278,800 0 Other" 43,300 24,700 58,500 8,800 1

Total 498,100 154,600 614,900 35,000 Cr, Ni, B4C, Fe0, Fe304, Cr203, NiO, and B203 l

1 O

e

l 1 1 l1 r

f

. D

. d e r

. m a o A, p ,

,k ,( t_

by ,A 4

r f

DN C

f a

T

{* ke

\ .

n e

ah et A 3 r f

_ i

~ $. " .a t

ty bf o -

A

,t. I 0

_ \.. 1

_ .h  :

e -

Q .. eh f

i *aTN A8s nn

/ .

~

Tr -

n i

ena sin ee

.mP ,

i o li t

J my na N-

=g

_ oc A --

- __ i o O .J tl

- n .

c N -

.,, J - -

ue sh A 0 -

- l [l My,!4$[

_ ele. i p m;w t .

n

\

1 _

uR o pAh

> S s nU N,g -

1  !

l

. o i ms o -

- a._ ,

• t al aha l e urs

. coi rh -

iS m c e

. eet rht sy d s en

=: - ni t a

t n sd e _

4 0 .e 3 oa r sem asn

. Ti a

< Ar p

. y '

Ctat Oi a eua si _

. u pLmi s e K (e y t

h ct t sn e

Rt uAo a P Ep ss Rna t

fd c i o

. NmaE r o

u. '

,n_

t SO 5

E UNn I ELT T Nct IAI ns as L Laa t nnh oet t e id J

. a0

_ . E C RIO .u E. t i n -

. . t I Tea aci s CCe0 t

eEUwP cch

'- l F RSsU

. nat i

.y . : -

I

.a ..

. sw

=.

n.

.isl 1 ae

. _.ra.

. .b s

__ s .

5

. .. .s 3 .

.gv ne

_. ;. 4 s gi i

..=

e t

.a i sr F eo dt

. c ea h e

. t r

i

, se th

• I CafetLL at ,

te le '

gg )

m I en 33 . I g i.

tM 88 '

= 049%CLL PLO42

• PeFFEstattAL i

o#

L. .... ... .:. ...

flMC &# fge ACCT 0$ Nf. SEC040S Fleuat &2,6-6 PRlW ARY CONTAINWENT SHORT TEAW PRESSURC RESPCN$C TO A RfC1ACULafiON LINE 8Af AW SmoetH AM wuCLE AA Post # Sf 4Tt04 We0Argo sutry anasygga agpoor l

i l

I i

310 I w .

04Y2466 see -

ba i m .

I '" '

r'

$LMeission CHAM 4(A g m.1 i sit ..f i is se*' ed ee* m' e8 f ast 17 t E A ACClotMT= $t Conos AA TANWENT SHOAT f TAW ffWPER ATUAE RESPON$$

10 A AfCRCUL AflCN Lhg SA( AX l@u'an9ls**l? s'UA'h'**

Fig. 5. 2. Calculated prima ry containment- pressures and tempera -

, tures for the first 30 s af ter double-ended rupture of a recirculation pump suction line.

The results shown here are taken f rom the Shoreham USAR.

6. SUltMARY AND FINDINGS This ' le t ter report presents the results of calculations performed by personnel of the Boiling Water Reactor Severe Accident Technology (BWRSAT) Program at Oak Ridge National Laboratory. The purpose of these calculations is to examine certain aspects of the potential for mitiga-tion of severe accident consequences afforded by operation of the Shoreham Nuclear Power Station at 25% of rated power. The areas exam-ined are those requested by the NRC-NRR technical sponsor of this work.

Degraded core response calculations have been perf ormed f or total loss of injection accident sequences using the Boiling Water Reacto r Severe Accident Response (BWRSAR) models developed by the BWRSAT program at Oak Ridge. To establish the delay in sequence event timing afforded by the lowe r decay heat powe r levels derived f rom operation at 25%

power, two accident sequence calculations were performed using identical input parameters except for the initial power levels, which were equiva-lent to 100% of rated power in one case and to 25% of rated power in the other. Both calculations were carried out well beyond the time of pre-dicted failure of the reactor vessel bottom head penetrations so that the pouring of molten constituents of core debris f rom the vessel could be analyzed.

Anticipated Transient Without Scram (ATWS) calculations have been i performed with the BWR-Long Term Accident Simulation (BWR-LTAS) code to examine the more sluggish reaction of a nuclear reactor to positive reactivity insertions with the control blades in the 25% power config-uration as opposed to the 100% power configuration.' <

Other ATWS accident sequence calculations were performed with the BWRSAR program models. These calculations were based upon the C9D acci-dent sequence described in the LILCO submission regarding the operation of the Shoreham Nuclear Power Station at 25% power. Their purpose is to examine the difference in predicted timing of severe accident events introduced by the BWRSAT program methodology -(described in Table 3.1) as opposed to that used for the calculations cited by LILCO.

A final set of calculations with the BWRSAR degraded core response models was made to determine the shortest possible period of time be-tween reactor scram and the onset of core degradation. This occurs f or the very low probability large-break LOCA with total loss of ECCS (AE) accident sequence. These calculations were performed both for scram from 100% power and for scram from 25% powe r, and were carried out through the period of release of molten material from the reactor vessel.

The major findings derived f rom the efforts discussed above are as

, follows:

6.1

l. Given the unlikely occurrence of an accident sequence that pro-

, ceeds through core degradation and reactor vessel bottom head penetra-tion failure, there would be a significantly longer period of time be-tween the events of the severe accident sequence following operation at

, 25% power than following operation at 100% powe r. This conclusion is based upon calculations for the total loss of injection sequence and the comparison of event timing provided in Table 3. 3, as well as calcula-tions for the large-break LOCA with loss of all ECCS sequence with its comparison of event timing provided in Table 5.1. There is no reason to believe that a similar conclusion can not be drawn for any other acci-dent sequence for which comparison of event timing might be made.

2. About 14% additional in-core hydrogen generation is predicted for the total loss of injection accident sequence initiated from 25%

power as opposed to the hydrogen generation predicted for the identical sequence initiated from 100% power. The additional hyd rogen generation is primarily caused by increased oxidation of the channel box walls ,

which occurs during the time that they are slowly heated within the tem-perature range at which oxidation occurs. This matter is discussed in detail in Sect. 3. 3. The generation of this additional hydrogen, while of interest, is not a critical consideration because a great deal of hydrogen (>1600 lbs) is predicted to be generated regardless of the

. assumed initial power.

3. Given the continuance of a severe accident sequence beyond the

, point at which molten core debris would begin to pour f rom the reactor vessel, then consideration mu s t be given to the temp e rature of the reactor vessel wall in the vicinity of the bottom head. With the core debris at the decay heat power level derived f rom previous operation at 100% power, the temperature of the debris is predicted to increase rapidly, with the reactor vessel wall temperature lagging f ar behind.

Under these conditions, much of the debris is predicted to have become molten and left the vessel by the time the wall temperature reaches levels at which creep rupture is expected to occur.

On the other hand, with the core debris heated by the lower decay heat power level derived from previous operation at 25% power, the debris temperature increases more slowly, and the vessel wall tempera-ture more closely follows the debris temperature. Under these condi-tions, the wall temperature is predicted to reach temperatures at which creep rupture failure would be expected while a major portion of the core debris is still solid within the bottom head. Thus, the results of the calculations described in this report do not support the contention of the LILCO submission that all core debris would exit the reactor vessel in a molten state.

4. ATWS accident sequences initiated by transient-induced closure of the main steam isolation valves from operation at 25% power should not be of concern. Successful operation of the Standby Liquid Control System (SLCS) to inject neutron-absorbing poison solution would provide safe reactor shutdown. In the unlikely event of f ailure of the SLCS function, the operators are directed by procedure . to lower the water 6.2

level to the top of the core and to depressurize the reactor vessel.

Recent preliminary work at Rensselaer Polytechnic Institute (Ref. 16) suggests that the Shoreham reactor would be suberitical in this config-uration even without liquid poison injection -- that is, with the con-

, trol blades in their 25% power positions, the reactor vessel water level at the top of the core, and the reactor vessel pressure at 200 psia (or below). However, if the reactor does remain critical in this configura-tion, then the results of the BWR-LTAS code calculations discussed in Sect. 4.2 demonstrate the sluggish reaction of the core to possible positive reactivity insertions caused by uncontrolled cold-water injec-tion by the low pressure ECCS systems.

5. A shadow calculation performed for the severe accident phase of the C9D ATWS accident sequence described in the LILCO submission indi-cates that the ORNL BWRSAT program best-estimate method for calculating degraded core and reactor vessel failure events (Table 3.1) predicts a much longer time to reactor vessel pressure boundary f ailure (30.8 h vs.

10.4 h) than does the MAAP code method used in calculating the results provided by LILCO. This is because of the approach taken at ORNL that the water in the reactor vessel bottom head must be boiled away before the bottom head penetration welds can become heated to f ailure tempera-tures.

6. Fo r operation at 25% of rated power, the ve ry earliest that core degradation might begin following the initiating event of an acci-

, dent sequence is calculated to be 59. 5 min. This occurs in the large-break LOCA with total loss of injection accident sequence (AE), which .

has a much lowe r probability than the loss-of-injection or ATWS '

sequences. Failure of the reactor vessel bottom head penetrations is i not predicted to occur until 469 min (7.8 h). Additional information regarding event timing for this accident sequence is provided in Table 5.1.

I r

L 1 l 6.3 I l

7. REFERENCES

1. D. H. Cook et a 1. , Station Blackout at Browne Ferry Unit One - 1

, Accident Sequence Analysis, Vol. 1, NUREG/CR-2182, November 1981. l

2. R. P. Wichner et al. , Station Blackout at Broone Ferry Unit One - ,

Iodine and Noble Gas Matribution and Release, Vo l. 2, NUREG/CR-2182, August 1982.

1

3. W. A. Condon et a 1. , SBLOCA Outside Containment at Broone Ferry Unit One - Accident Sequence Analysie, Vol. 1, NUREG/CR-26 72, October 1982.

4. D. H. Cook et a1. , Lose of DHR Sequences at Broone Ferr"J Unit One -

- Accident Sequence Analysis, NUREG/CR-2973, May 1983.

5. R. P. Wichnet et al. , SBLOCA Outside Containment at Browns Ferry Unit One - Iodine, Cesium, and Noble Gas Matribution and Release, Vol. 2, NUREG/CR-2672, September 1983.

6. R. M. Harrington and L. J. Ott, The Effect of Santl-Capacity, High-

. Pressure Infection Systema on TQUV Sequences at Browns Ferry Unit One, NUREG/CR-3179, September 1983.

. 7. R. M. Harrington and S. A. Hodge, ANS at Browne Ferry Unit One -

Accident Sequence Analysis, NUREG/CR-3470, July 1984.

8. R. P. Wichner et a1., hble Gas, Iodine, and Cesium Transport in a Postulated Loss of Decay Heat Removal Accident at Broone Ferry ,

NUREG/CR-3617, August 1984.

9. R. M. Ha rringt on and L. C. Fuller, BWR-LTAS: A Boiling Water Reactor Long-Term Accident Simulation Code, NUREG/CR-3764, February 1985.

10. R. M. Ha rringt o n, Modifications -to BWR-LTAS Since December 1984, letter report to Dr. Thomas J. Walker, Division of Accident Evalua-tion, RES, USNRC, dated June 21, 1985.

11. R. M. Harrington, Evaluation of Operator Action Strategies for* I Mitigation of MSIV Closure Initiated ANS, letter report to Dr. Thomas J. Walker, Division of Accident Evaluation, RES, USNRC ,

dated November 11, 1985.

12. R. M. Harrington, The Effect of Reactor Vessel Pressure and Water Level on Equilibrium BWR Core Thermal Pouer During MSIV-Closure-Initiated AWS, letter report to Dr. Thomas J. Walker, Division of

. Accident Evaluation, RES, USNRC, deced January 10, 1986.

7.1

i

13. R. M. Harrington and S. A. Hodge, Loes of Control Air at Broune

, Ferry Unit One - Accident Sequence Analysis, NUREG/CR-4413, Ap ril )

1986. t i

I

14. F.. M. Harrington and S. A. Hodge, Contair9nt Venting as a Severe 1 Accident Mitigation Technique for BWR Plants with Mark I, Contain-ment, letter report to Dr. Thomas J. Walke r, Division of Accident '

Evaluation, RES. USNRC, dr.ted June. 26, 1986. ,

i

15. D. H. Cook, Codes Comparison Analysis of AZWS for the James A.

Fitzpatrick Nuclear Power Plant, letter report to Dr. Thomas J. ,

7 Walker, Accident Evaluation Branch, Division of Reactor Accident, f Analysis, RES, USNRC, dated May 11, 1987. i

16. Dr. R. T. Lahey , Jr. , Chai rma n, Department -of' Nuclear Engineering I',X and Engineering Physics, Rensselaer Polytechnic Institute, letter '

n of May 6, 1987, to Dr. Charles N. Kelber, Of fice of Nuclear Regula-  ?

tory Research. -

> Y W t) f

',I e

I[

3 l

). ),

(

1 i

'9

.f'

, -)

7.2 ,

J, ,

1,

APPENDIX Memo to: S. A. Hodge May 28, 1987 l

From: R. M. Harrington i

Subject:

Cer.t'rol Rod Reactivity Input f or BWR-LTAS l t

25% Initial Power ATWS Runs 1

9 f Three runs were gade recently to _ scope the ef fect of a 25% initial j bo,wer vs.100% initial power on the reactor response during a hypotheti-l

. cal ATWS accident without crator action: a full power run, with zero j f or parameter DKRODS and two runs with different negative control rod )

,- reactivitieu. DKRODS is the amount of reactivity in the core due to {

control rods, and is the only parameter affecting dynamic behavior that '

is diff erent f or di(f ering initial power levels. The recirculation pump speed varies with power level, but during an ATWS, the recirculation pumps are typically tripped ve ry early on high vessel pressure (o r )

slightly later on low vessel water level). Since BWR-LTAS is initial- l

, ized 50 s after the beginning of the accident, the recirculation pumps would be tripped for all cases, so there would be no need for different l i

input ruuabers expressing different recirculation flow between 100% and l 25% initial power' levels. l q To de termine what amount of reactivity to use to simulate the reactor at 25% power, the steady state version of the E.t-LTAS reactor power calculating routine was run. This program finds the steady state power level based on user input for reactor pressure, downcomer water level, jet pump pressure rise (a function of recirculation pump speed),

core inlet enthalpy, and control rod reactivity. The input parameters for voic and doppler coefficients are based on Browne Fe r r/' informa-tion. N 1'ciple runs were made to search for the combination of jet pump pressure rise (i.e. recirculation pump speed) c ' ngative control rod reactivity that would be required to match the desued 25% core thermal p owe r, and the 7% core outlet quality specified by Shoreham in Lap Cheng 's May 18, 1987 telecopy message ( Attachment 1). , Other parameters such as reactor vessel pressure were also specified by Mr. Lap Cheng, and were the same for each run. The result shows that the control rods must add a negathe 0.019 delta-K/K to reach the desired 25% powe r operating condition; the code is set up to reproduce 100% power oper-ating conditions without any control rod reactivi ty. Th e ref ore, it is concluded that a negative 0.019 delta-K/K control rod reactivity is required to reduce the reactor power f rom 100% to 25%.

The BWR-LTAS transient case with -0.017 control rod reactivity rep-

, resents a best estimate of response with'25% initial power. In order to test the sensitivity of the results to this input, the negative control rod reactivity was decreased by 20%, to -0.015 control rod reactivity.

I A.1

3 l

s As mentioned above, an input of 0.0 control rod reactivity is used for

• the 100% initial power case.

In order to test the conservatism of the BWR-LTAS based inputs for

. negative control rod reactivity at 25% power, an independent estimate l' was based on the results of detailed steady state neutronic codes, also supplied in Attachment 1. At full power the total control rod insertion has been calculated by Shoreham to be 514 notches; at 25% power the cor- i responding insertion is 1536 notches. S'as e one notch is 1/48 of the l total travel of one control rod, and there are 137 control rods, the net l total control rod insertion is 7.8% at full power and 23.3% at 25%  !

, p owe r. Th e total control rod worth is 24.968 dollars, or 0.18 delta-K/K; therefore, the extra negative reactivity required to decrease power f rom full power to 25% power is:

DKRODS = -(0.233-0.078)

• 0.18 = -0.0279 delta-K/K ,

l This shows that the -0.019 and -0.015 numbers used in the BWR-LTAS runs are well on the conservative side.

l e

1 e

a A.2

x Qx TELECoPIER 4M31S* 5-373 #1 ~F" ' 4 2770767; m 2 MY 18 '87 15:27 SNPs tED

• P' 2 kn9 Id5nd d tim hw Nuclear Enginee@ing r bepartnent 3

Til.EPHONE HEMORANDUM - -

Call Date__ f"/8- k Time]?#O Sm Incominc _

Outgeing- N Be heen_$ k $ af l k N , Y Y h e of h $lA bl Ss and / ', /, l

/ de # of HED.

Subject -

i h sood bn 8 A Nl. .

V U 5UWARY: ""YEJ AA,k IAe$ e

~

\

N 29 'kb> E e 0 snA NE E DRAlL rw & A)ne .

( 609 MW.b '

M r A k 'k k k = 32< b6 Xtab llr., Os.

$$n Ans As u 94 /pclA '

F.e}id A21sho - (,',.z HO?

zOb DWA A> rwa D W '

\

ik a.<ll~ sushN '

K w Vnkabi N~s %

=- s o7.

_. V 'l l' p 2.z.4 x d A. L , /

1

• 7MnJ % xdy,t nk = - 0 24.9/9 [Ab?~"p!!!JYbbhW Y A kb" Me - a = s.oain "reai CtdaJ Lshk if a'o+!n - t14 Lb>l l'&2Ll!% %9 f 3"iA>b le-up)o

'I to r 1 2T$

>A,

~-' /EB . kr<< or}nwo(7To 'TA. d07

= VA.)~

~ku~ ' ad .A)w Ank h sa bAbs<{

D% 9ul i M ~J .M Anail m,s hmz Y h~.

M 4 % 4 & % .dds.

cuting: I J 4 ad. %%td. &W W .dM. u.a a, L _.

u ,

T7N hl}</B

-- TAA EPS ~

_ B. djn.o (NDSb)

~

Telepho.ifenoranda File

~

s

. A.3

_