ML20058N116

Forwards Responses to Discussion Items from 900516-17 Meetings,Including Drywell Head Failure,Containment Overpressure Protection,Source Term & Fire & Seismic Risk

ML20058N116
Person / Time
Site:	05000605
Issue date:	08/09/1990
From:	Marriott P GENERAL ELECTRIC CO.
To:	Chris Miller NRC, NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
EEN-9046, MFN-099-90, MFN-99-90, NUDOCS 9008130232
Download: ML20058N116 (12)

Text

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

... . ,O  !

(f) t-

' ~

CE Nuclear Energy

,.x

. &rw i+:t*,: Cc-wr 175 Came t ru W .ksr CA 95175 August 9,1990 MFN No.099 90 1 Docket No. STN 50 605  !

EEN 905 l Document Control Desk l U.S. Nuclear Regulatory Commission I Warmgton, D.C. 20555 j Attention: - Charles L. Miller, Director 1' Standardization and Non Power Reactor Project Directorate

Subject:

Rssponse to NRC/CE May 1617,1990 Meeting Discussion Toples

Reference:

Dino C. Scaletti," Summary of Meeting with General Electric on ABWR", (May 1617,1990) June 8,1990 Enclosed are thirty four (34) copies of our responses to the discussion topics of the reference meeting. These discussion topics include: drywell head failure; containment overpressure protection; source term; shutdown risk; fire and seismic risk; and_ lower drywell flooder, it is intended that GE will amend the SSAR, as appropriate, in a future amendment.

. Sincerely, i

j <

E P, W. Marriott, Manager Regulatory and Analysis Services

' M/C 382, (408) 925 6948 oc: F. A, Ross (DOE)

D. C. Scaletti (NRC) -

D. R. Wilkins (GE) 7 ,

J.F. Quirk (GE) {

~ \

9009130232 900809 .

PDR ADOCK 05000605 1 x A- PDC

-.- - - . - . ~ .

1 ,  :

e DISCUSSION TOPICS AND RESPONSES i

NRC/GE May 1617,1990 Meeting Discussion Tonic 1 Dnwell Head Failure a Drywell Seal Leakage The staff was concerned about the GE analyses which assumed that s ificant Sealleakage will occur at 52 psig (d: sign pressure 45psig). GE stated that the assumed i le kage was based on a bounding analysis. However, realistically they would not expect any -

leakage at 52 psig. GE will get back to us on this issue and the issue of the proposed use of silicone )'

rubber seals.

b) Structural Failure GE's analyse; assumed drywell head failure at 90 psig; the calculated failure  !

pressure was 100 psig. De staff sees this as significant problem in view of the previous findings '

that most of the sirnilarly designed (i.e. 45 psig design pressure) BWR containments et n witV.and higher ultimate presures. OE agreed to increase the drywell head pressure capability. As would  :

not be a major modibcation and would probably result in an ultimate containment fadure pressure l p' in excess of 120130lisig. .

Resnonse to Discuss on Tonic 1 a) Drywell Seal 1.eakige A pres $ure of 52 psig was assumed to be the pressure level required to result in an initiation of separation of sealing surfaces of the drywell head closure. It is 1.15 times y design pressure - .f a. consistent with the Sandia's findin p (SAND 88 0331C) that the separation i pressure for ;arge, operable penetrations typically ranges f rom .1 to 1.5 times design pressure.

s Pressure,in excess of the separation pressure can lead to leakage if and oniv if the seals are also '

e degra/ed due to high temperatures. %e seal degradation temprature was assumed to be 500*F in accordance wi'h Sandia's test results (SAND 891631C), Lea (age was assumed to occur only for '

those severe accidt.nt scenarios which could lead to pressures in excess of 52 psig and temperatures in excess of 500*F. This is a very small fraction of severe accident cases. A steady state heat transfer analysis shows that if the containment temp,erature reaches 700*F ! representative for '

l those accic'ent in whbh the temperature exceeds 500 F), the difference between the containment l and seal temperatures is cri sh: uder of 10 F. De seals in the drywell head closure are essentially subjected to the same temperatures at the containment. De leakage area estimates were made in -

a manner consister.t with the assumptions for the upper bound estimate in NUREG 1037 in that ,

no credit is taken for the degradeci seals and that flange separation is uniform around the xrimeter of the sealing surfaces. On :he issue of the use of silicone rubber seals, GE conducted a  :

l .iterature review of seal performance which indicates that silicone rubber is a common gasket l- material used for containment netrations in existing nuclear power plants. In fact, the Sandia's seal tests (SAND 891631C) cemerstrated that silicone rubber (including thermal aging only or thermal plus radiation aging)is one of the better materials to resist high temperatures. Any seal material, including silicone rubbn that is environmentally qualified, is acceptable. -

b) Structural Failure GE would like to emphasize that the drywell head pressure capabilitg is temperature dependent. The 90 psig pressure is the calculated capability asgociated with 700 F.

The 100 psig pressure is the calculated capability associated with 500 F which is a typical temperature for most severe accident sequences. Nevertheless, the capability has been increased by changing the head thickness from l' :o 11/4*. The 1 1/d* thick drywell head has a pressure l capability of 134 psig at 500 F, and 120 psig at 700*F. The pressure capability is higher than 2.5 l times doign aressure. The ABWR SSAR will be updated to reflect this increased pressure L capabilit) in a future amendment.

1 .

1 i

l L

e DISCUSSION TOPICS AND RESPONSES NRC/GE May 1617,1990 Meeting

1. Discuulon Tonic 2 Containment Overoressure Protection a)It appears that GE has done sufficient analysis to conclude that 1) pool swell and rapid depressurization are not a problem, OE has determined that depressurizt. tion will take apprnimately one half hour; 2) a demister is not necessary; and 3) the period prior to the vent actuation :.h he extended (probably beyond 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />) through optimized use of the fire water

,- system.

L

_ OE also agreed not to use limestone concrete in the drywell cavity area in order to reduce the

_ generation of non condensable gasses; thereby minimizing early challenge to containment and the overpressure protection system.

w b) Manual operation of the containment overpiessure device: OE has agreed to reconsider the i need for manual control (provision of a by () over the vent process.

- Rey will get back to the staff on this issue. pass valve around a rupture dis y

Resoonse to Discussion Topic 2 (a)(1) Pool Swell and Rapid Depressurization If the pool swell were sufficient to raise the water level to the elevation of t 2 rupture disk there would be potential damage of the piping due to the dynamic loads. In order to determine the swollen level of the pool the mass flow rate of steam from the containment was determined. Using this flow rate, the void fraction of the pool was calculated using a drift flux model. A maximum level =well nf about 2 meters was calculated. De connection of the overpressure protection piping will be well above this elevation, therefore, there is no risk to the piping due to level swell.

(a)(2) Demister There is a potential for the release of the fission products dissolved in the su?pression pool as a result of entrainment through the overpressure protection piping. A ca culation of the carryover of wa;er through the rupture disk was performed to assess this potential. An uppej bound for the release fraction associated with this mechanism was found to be on the order of 10 . Therefore, there is no significant increase in fission prodoet release due to entrainment of water through ihe rupture disk and a demister is not warranted.

(a)(3) Extending the Period Prior to the Vent Actuation - Analysis has shown that optimized use

, of the firewater addition system can provide this capability for non ATWS events. De details of

_ this analysis wiU de provided in a future amendment of the ABWR SSAR, Exclusion of Limestone Concrete Exclusion of limestone concrete from the drywell cavity area l- (drywell floor) will be covered in ABWR SSAR Subsection 3.8.3.6. His change will be made in a I future amendment of the ABWR SSAR, b) On July 17,1990, representatives of GE (Quirk, Sawyer) met with members of NRC staff (Thadani, Kudrick, Scaletti) to report on the results of GE's reconsideration of the need for manual control (provision of a bypass valve around a rupture disk) of the ABWR containment

, overpressure protection system. OE informed the staff that the results reaffirmed the desirability of retaining the passive only functioning of the overpressure protection feature. OE stated it did not view adding a provision for bypassing the rupture disk as a desirable addition since it could be misused leading to unplanned, premature release. OE views the downside risks of bypassing the rupture disk as compromising the mitigation benefits of the system. OE has sponsored the 2-

DISCUSSION TOPICS AND RESPONSES .

NRC/GE May 16-17,1990 Meeting Re :nonse to Discussion Topic 2 (Continued) ,

addition of an overpressure protection feature for the ABWR on the merits of its simple design which is not predicated on the need for operator action. OE characterized some of the criticism of the overpressure protection device as 'proceduralizing the bypassing of containment

• and, as a  !

result, potentially giving rise to political intervention. OE offset such criticism with a design that i has passive.only actuation occurring prior to postulated gross failure of the containment resulting  !

In a preferred failure location and thus insuring a scrubbed, filtered pathway precluding large i release, i

The ABWR containment desij;n does not rely on the overpressure protection device to mititate l ,

design bases accidents, or to x the first line of defense for postulated severe accidents..a ro3ust ABWR containment design meets the Utility /EPRI ALWR Requirements and NRC requirements in such cases. Rather, the overpressure protection device provides a fall safe containment given postulated end of spectrum events that threaten containment.

Discumaton Topic 3. Source Term GE agreed to expedite their submittal to justify taking credit for plate.out and holdup of fission I products in the non seismically designed steam lines and condenser and to track the BWR Owners l Grou 3 on this issue. However, GE wishes to maintain their existing SSAR analysis concluding comp.iance with Part 100 should the effort to give credit for non seismic equipment be delayed, i Resnonse io Discussion Tonic 3

! {

! nis continues to be unresolved for the BWR Owners Group. No GE action required at this time.

L Discussion Tople 4 Shutdown R_isk GE is convinced that the risk during shutdown condition is low for BWRs. They will submit a qualitative discussion supporting their beliefs.

Response to Discussion Tonic 4 i A shutdown risk study has been initiated for the ABWR design. Thus far, the shutdown risk I

events / accidents and their corresponding causes and preventions have been identified in the attached Topic 4 Shutdown Risk table. nis shutdown risk study will be transmitted the NRC by I mid September,1990. The ABWR SSAR will be modified to reflect this study in a future L amendment. '

Discussion Topic 5. Other PRA Tonics as Annropriate a) Fire and Seismic Risk GE is following the EPRI lead (no PRA needed for fire but PRA to be done for seismic events). There appears to be two problems with this approach.1) Probjbilistic anglysis of seismic events to show that these events are not significant at a frequency of 10 /RY to 10' /RY is only going to lead to unnecessary debate in view of the mostly subjective nature of the 3-

DISCUSSION TOPICS AND RESPONSES NRC/GE May 1617,1990 Meeting Discussion Tople 5. Other PRA Topics as Anprocriate (Continued) methodology. (Note, we are talking about earthquakes well in excess of 10 peak ground acceleration). A margins type approach might be more sensible. 2) The approach for fire and  !

seismic events should be consistent. The staff plans to meet whh EPRI in the near future to discuss this matter. ,

b) 1.mt Drywell Flooder OE wiu provide 1) the lower drywell flooder design information and the testing to be conducted,2) estimated steaming rate when substantial amount of core debris is  ;

ejected into a pool of water in the lower dnwell, and 3) Identification of the events when the lower drywe3 is Gooded before the core debris is expected to reach the drywell.

Reanonse to discussion Toole 5 a) No action aquired by GE. The staff has not reported to OE on their meeting with EPRI  !

pertaining to ti current PRA approach.

(b)(1) Design and Testing of the Lower Drywell Flooder . The function of the lower drywell  :

flooder (LDF) is to Good the lower drywell with water from the suppression pool during severe accidents where core melting and subsequent vessel failure occur. The LDF consists of 10100mm pipes that run from the vertical pedestal vents into the lower dnwell. Each pipe contains a fusible plug valve connected to the end of the pipe that extends into the 100 Jnwell by a flange. In the unhkely event that molten corium Dows to the lower drywell floor and is not covered with water, i the lower drywell atmosphere will rapidly heat up, ne fusible plug valves open when the dowell air space (and subsequently the fusible valve is mounted in the verticalwith position, plug the fusible valve) metal facing temperature reaches downward, to facilitate the 260 C. De opening of the valve when the fusible metal mehing temperature is reached. When the fusible plug -

valves open, a minimum of 105 liters /sec of suppression pool water will be supplied through the I system to the lower dr>well to quench the corium, cover the corium and remove corium decay heat, which is estimated at 1% of rated thermal power. He result will be a reduced drywell temperature and pressure from noncondensible gas generation. Dere will be less chance of overpressurizing the containment and increasing leakage, ne LDF is a passive injection system and is maintained in an operable state whenever the reactor is critical. No operator action is required.

Qualification of the proposed fusible plug material will be completed before certification. These qualification tests wi] confirm,that the plug will maintain zero leakage at design temperature (171*C) and pressure (1.lkg/cm d) and will be fully open at 260 C.

i No testing of the LDF system will be required during normal plant operation. During refueling outages, the following surveiUance activ!tles would be required:

(1) During cach refueling outage, verify that there is no leakage from the fusible plug valve flange or outlet when the suppression pool is at its maximum level.

4-

l DISCUSSION TOPICS AND RESPONSES t NRC/OE May 1617,1990 Meetirig Resonnse to Discussion Tonic 5 (Continued) i (2) Once every two refueling outages, lower suppression pool water level or plug the flooder pl x inlet and replace two fusible plug valves. Test t 1e valves that were removed to confirm their function. His practice follows the precedent set for in service testing of standby liquid control system (SLCS) explosive valves m earlier BWRs.

(b)(2) GE was also asked to estimate the steaming rate when a substantial amount of core debris is ejected into a ml of water in the lower dr>weU. *)ese calculations are in progress, and the results will be transm tied to the NRC by mid September 1990.

(b)(3) He NRC asked GE to identify the everits in which the lower drywell is flooded before core debris is expected to reach the dryweu. His event sequence is expected to occur only when core coolant is lost through a LOCA in the lower head. All the penetrations in the lower head are small and any loss of coolant accident through them is classified a small break LOCA. A conservative estimate of the core Jamage frequency resulting from LOCAs in the lower head is the total CDF associated with all small break LOCAs for the ABWR, nis value.pssumin incorporation of a gas turbine generator with an unavauability of 5 percent, is 5.09X10' events car as shown in Ta ale 19.3 7 of the SSAR. Comparing this value to the total CDF reveals that a smrl! break LOCA's contribute just 3% of the total. The actual fractio's of events which would lead to water in the lower dryweu is smaller than this estimate; however, a more accurate determination would require '

a significant amount of effort and the 3% value should lead to acceptable results.

L i

l.

I

.i

. I

.j DISCUSSION TOPICS AND RESPONSES NRC/OE May 1617,1990 Meeting t.

t t

i i

?=

t EVENTS IMPACTING ABWR SHUTDOWN RISK (TOPlc 4) 1

.g ._ .

i

, S!!UTDOWN RISK TABLE. TOPIC 4 l

. EW;b'I CAUSE PREVENTION i r

r i

1.- REACTIVITY ACCIDENTS

1. Control Rod Drop A) Stuck control rod 1) Positiu bayonet coupling l Accident. plus separation between the control rod and of control rod the drive with which the '

and/or drive. connection between the  ;

control rod and the drive cannot be separated unless .!

they are rotated 45 degrees.

2) Class 1E device to detect separation of control rod .

from the drive (when such separation is detected, further ' .

control rod withdrawal will be prevented.) i

3) Hollow iston latch which detects if the hollow piston is j separated from the ball nut and rest of the drive due to stuck rod and limit subsequent rod drop to a -

distance of 8 inches.

3

4) Procedural couplingchecks to assure proper coupling. ,

B) Control rod 1) Class 1E aration f installed without -detection 7evice.  ;

couphng or c.

structural failure 2) Hollow piston latch. ,

of coupling plus stuck rod in F Procedural coupling checks. ,

the same FMCRD, l

2. Control Rod A) Major break in Shoptout restraints Ejection Accident the FMCRD i housing, weld between housing and vessel, drive mounting bolts or outer tube.

• B) Break in the 1) FMCRD Electro mechanical drive insert pipe brake

2) FMCRD check valve 1

4

3. -  !

\

SilUTDOWN RISK TABLE TOPIC 4 i EVENT CAMSE PREVENTION 3 i

3. Control Rod Operator Error Refueling interlock causes l Withdrawal Error rod block to assure only one -l (Coincident rod can be withdrawn wi:hdrawalof a '

control rod with another control .

rod) J

4. Continuous contr01 Operator error or 1) Rod Control and '

rod withdrawal error malfunction of the Information System (RCIS) during startup automatic rod preventswithdrawalof any ,

movement control out of sequence rods. "

system.

2) The startup range neutron  :

monitor (SRNM) initiates rod block or scram depending upon other parameters. ,

5. Fullloading error A fuel bundle loaded Core verification (expected (misplaced fuel in wrong location and frequency of this error is buncle) the fue! bundle for 0.002 events per year. May that location is also iesult in an undetected located incorrectly or reduction in thermal mar is discharged. during power operation) gin

6. Fuelloading A) Unload all fuel No incentive to remove all  :

(criticality) . bundles fuel bundles during refueling because veiy few FMCRD's B) Remove two need to b: maintained during adjacent control refueling outage, blades Refueling interlock prevents C) Load fuelin the second control blade from first being withdrawn when one is

• uncontrolled cell removed. Note also that (i.e. cell without Technical Specifications control blade) tequire that refuel position one rod out interlock shall D) Load fuelin the be operable during shutdown adjacent et refueling. If the interlock i uncontrolled cell is inoperable, the required action is to suspend control l- rod withdrawal and fully l Insert allinsertable control rods in core cells containing one or more fuel assemblies. ,

2-i.

I I

l 1.

E' SHUTDOWN RISK TABLE TOPIC 4 l

. EVENT CAUSE PREVENTION l L i 1

11. RPV DRAINING ACCIDENT 3 l

1. Recirculation - A) Human error during 1) Metal to metal seal internal pump (RIP) RIP motor removal maintenance 2) Inflatable seal ]

i B) Human error during 1) Metal to metal seal  ;

RIP shaft J H

2) Inflatable seal l

3) Plate is bolted at the bottom ,

to prevent accidentalleakage  !

2. FMCRD Human error during Control blade is back seated maintenance (drive drive replacement on the guide tube to provide a  ;

removal) metal to metal seal his is i similar to the feature in BWR (

6 plants.

3. Drain line Operator error 1) Procedure

2) ECCS instrumentation (low wate: level)

3) RPS instrumentation (scram on low water level alerts operator

4. Incore monitoring Operator error 1) nreaded sleeve inserted >

lines from below before incore instrumentation is ren:wed i or replaced

2) Ifleak startu m:ause the ~

sleeve is not inserted, inserting the incore mstrumentaion will stop the leak.

5. RPV differential Operator error during 1) Procedure

)ressure sensing maintenance oflines l

.ine and Core 6P 2) Orifice at nozzle restricts flow -i

,ressure sensing

.ine 3) Operating ECCS Systems

6. Note: No mode of RHR operation can drain the RPV accidentally 3

b

~ . _ - _v.

l~g. .

SilUTDOWN RISK TABLE TOPlc 4 EVENT CA1)$E PREVENTION III. CORE COOLING ACCIDENTS

1. Loss of Coolant Operator error (also Many potential sources of Accident or RPV see Section 11) core cooling systems: 2 draining event HPCF, CR ) System, ADS and 3 low pressure ECCS or condensate pum ps or fine water systems. Technical Specifications require that at -

least two ECCS subsystems shou.d be in operation.

2. less of offsite Er.ernal 1) Site has two offsite power and power one gas turbine and three diesel penerstors. Technical Specifications require that at least one circuit between off site transmission netowrk and the on. site Class IE distributiot; system and one diesel generator be operable during refueling.

L

2) Because of the low decay _ heat rate, CRD pumps can

- successfully prevent core damage, Also, there is sufficient time available to recover offsite power, diesel generator, failed core cooling equipment or arrange for a fire truck.

IV. CONTAINMENT HEAT REMOVAL ACCIDENT

1. .less of heat removal Equipment failure and Three heat removal systems operator errors (RHR) available, any one of which is sufficient to remove the decay heat from the -

suppression pool. Technical Specifications requires that at least two RllR systems be available during refueling shutdown.

-4 gr

~ _ . . _ __ _ _ -_

i j ,,

c SHUTDOWN RISKTABLE TOPIC 4 l EVENT CAUSE PREVENTION l

2. I.oss of offsite Exter..al 1) Site has two offsite power and I power one gas turbine and three  !

diesel generators. Technical i Specifications require that at  ;

least one c ' ult between  ;

c off site trai .nission network and the on site Class IE  ;

distribution system and one diesel generator be operable  ;

during refueling.  ;

2) Because of the low decay heat .

rate, a significant amount of time is available for recovering offsite power. i diesel generator, gas turbine or any other equipment that t may have broken down.

Failure of all decay heat removal paths isjudged to i have a low occurrence frequency,  ;

V. EXTERNAL EVENTS

1. Seismic PRA External low core damage frequency.

based on results of seismic .

PSA conducted for plant at full power.

2. Fire, Flood PRA External Expected negligible CDF . I because of good separation j criteria used in layout design. ,

3. Other External External Negligible CDF ex;cted.

Events .

/ ,

L l

L u '

, ., .. . .. -. ._ - _ _ . ____ _ _ .__ _ __ - - _.