Text

Responds to NRC Re Violations Noted in Insp Repts 50-327/99-04 & 50-328/99-04.Corrective Actions:Risk Determination Evaluation Was Performed & Licensee Concluded That Event Is in Green Regulatory Response Band
	ML20216J935
Person / Time
Site:	Sequoyah
Issue date:	09/27/1999
From:	Salas P; TENNESSEE VALLEY AUTHORITY
To:	; NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM)
References
	50-327-99-04, 50-328-99-04, NUDOCS 9910070044
	Download: ML20216J935 (12)
	v • d • e

o L. ,

Tennessee Vaity Authonty Post Ofhce Box 2000. Soddy-Daisy. Tennessee 37379 September 27, 1999 U. S. Nuclear Regulatory Commission ATTN: Document Control Desk Washington, D.C. 20555 Gentlemen:

In the Matter of ) Docket Nos. 50-327 Tennessee Valley Authority ) 50-328 SEQUOYAH NUCLEAR PLANT (SQN) - NRC INTEGRATED INSPECTION REPORT 50-327, 50-328/99 RESPONSE TO 3FQUOYAH FLOODING EVENT RISK DETERMINATION

Reference:

NRC letter to TVA dated Septem:ur 7, 1999, "Sequoyah Flooding Event Risk Determination (NRC Integrated Inspection Report No. 50-327, 328/99-04)"

This letter is to provide our response to the risk determination contained in the referenced letter. As discussed with Mr. Paul Fredrickson of the NRC staff on September 21, response to the referenced letter was extended to September 28, 1999. We requested the additional time to address questions raised by Mr. Rudolph Bernhard of the NRC staff on September 16, 1999, during a telephone conference call. We appreciate the opportunity to provide NRC with input on the risk 6etermination of the SON flooding event as the issue continues in Phase III of NRC's Significance Determination Process (SDP) . Your risk determination concluded that the June 30, 1999 flooding event is in the l White regulatory response band under the SDP.

O'70009 .

9910070044 990927 PDR ADOCK 05000327 G PPR

a >

. U.S. Nuclear Regulatory Commission

. Page 2 September 27, 1999 We have performed a risk determination evaluation of this event and concluded that the event is in the Green regulatory response band. A conference call was conducted with Mr. Bernhard, the staff's probabilistic safety assessment analyst on September 16, 1999, and a review of the NRC risk determination evaluation contained in the referenced letter was performed to identify evaluation differences. The difference in risk significance of the event, resulting in the NRC White versus the TVA Green regulatory respr7se bands, stems from risk attributes for:

1) Identification and operator action of a flood potential condition (event identification), and

2) Recovery time for offsite power through restoration of flood-damaged electrical panels.

Each of these areas are addressed in the enclosed evaluation.

The enclosed risk determination evaluation, including evaluation methodology and assumptions, is for your use. We believe that the evaluation is conservative and appropriately considers overall risk of the flooding event. If you do not agree with our conclusions after reviewing our evaluation and the resulting regulatory response band, we request a meeting with you to discuss the risk significance evaluation in order to ensure that we fully understand the staff's evaluation.

Considering the evolving state of the SDP process the disposition of this item may require policy-level decisions.

Therefore, we request the participation of the NRC personnel entrusted with responsibility for those decisions.

If you have any questions regarding this response, please contact me at (423) 843-7170 or James D. Smith at (423) ;

843-6672. '

Si erely,l' oh t -

/ l Salas Enclosure cc: See page 3 il

f?'

'~'

I I

. U.S.' Nuclear Regulatory Commission

- Page 3' September 27, 1999' ,

cc (Enclosure) :

Mr. R..W. Hernan, Project Manager Nuclear Regulatory Commission '

One White Flint, North 11555 Rockville Pike Rockville, Maryland 20852-2739 i

NRC Resident Inspector Sequoyah Nuclear Plant 2600 Igou-Ferry Road ,

Soddy-Daisy, Tennessee 37379-3624 l I

Regional Administrator U.S. Nuclear Regulatory Commission Region II Atlanta Federal Center l 61 Forsyth Street, SW, Suite 23T85 Atlanta, Georgia 30303-3415 Mr. Steve Floyd Nuclear Energy Institute 1776 I Street, NW , Suite 400 Washington, D.C. 20006-3708 I

l l

I L l

}

ENCLOSURE j l

TENNESSEE VALLEY AUTHORITY {

SEQUOYAH NUCLEAR PLANT UNITS 1 AND 2 l i

i INSPECTION REPORT NUMBER 50-327, 328/99-04 RESPONSE TO SEQUOYAH FLOODING EVENT RISK DETERMINATION l

FLOODING EVENT OVERVIEW )

Event:

On June 30, 1999, rainfall from an isolated thunderstorm resulted in water buildup on the floor of the turbine building railroad bay (TBRB). The water level in the railroad bay in the vicinity of the 6.9-kV unit boards is estimated to have reached 2 inches.

The TBRB experienced a similar event in 1994. Some of the corrective actions taken following the 1994 event were to clean the storm drain system and to store cand bags in the vicinity of the railroad bay doorway to be used to limit water flow into the TBRB through the doorway.

Objective e'vidence indicates that these corrective actions were effective. For example in 1995, one rainfall event had 1.38 inches in 15 minutes and did not result in any flooding in the TBRB, whereas the June 30, 1999, event only had 0.67 inches of rainfall recorded in 15 minutes. These rainfall intensities and durations were recorded at the meteorological data station e located approximately 3/4 of a mile away from the turbine building.

Plant Layout

A schematic of '.he TBRB and immediate vicinity is shown in Figure 1.

The storm drains outside the railroad bay doorway are designed to ,

accept rainfall from roughly a 25 year maximum precipitation event. A temporary change was made in April 1999 to route the discharge of the cooling water flow from the bus duct coolers I (800 gallons per minute (gpm]) through 2, 6-inch fire hoses to 1 of the 2 storm drains near the railroad bay doorway.

El-1 i

i

y U-

~The railroad bay area runs the width of the turbine building.

TheJfloor area-in1the vicinity of the tracks is approximately

-2 inchesLbelow the'high pointrof the bay floor area, which is lalong the main ~ support - column' line -(located between the railroad track and the unit boards).

The railroad :b'ay area is 322 feet long and 54 feet ~ wide (17,388 square teet). There are 3, 6-inch drains running along the railroad 1 track; area and 4, 3--by 9-inch drains along the sump drain areasbehind the unit boards. These drains are routed to the turbine building sump, which is located several floors be1<

the' railroad. bay. These drains are in addition to the 2, 3-inc.2 drain. lines that run along'the outside of the railroad bay doorway.

The above described. design means that excess runoff will flow to the railroad track area and drains, eventually filling this low-lying area along its entire 322 feet length before the water approaches the 6.9-kV unit boards. This volume of water is

-calculated to be approximately 6,000 gallon. by:

((1/2)- (29 ft.) (322 ft.) (2/12 ~ f t. ) (7.48 gal./cu. f t. ) ] . Once the storm runoff reaches the elevation of the base of the unit

' boards, approximately 10,000 gallons is required to raise the water level in the railroad bay by 1 inch-as calculated by:

-(7.48' gal./cu. f t. ) (17,388 sq. f t. ) (1/12 f t. ) .

The 8. plant 6.9-kV-unit boards are laid out along.the width of the' turbine building with unit board 1A nearest the railroad bay doorway. Located next to the 1A unit board are the 1B, 1C, and 1D-boards, running approximately half the width of the turbine building. ,The 2A, 2B, 2C, and 2D unit boards are located in the remaining half of the railroad bay.

Discussions.with the plant electrical staff confirm that the unit

' boards would not be at risk until the water level reaches more than 6. inches at the unit board itself. This would put each of

the' floor drains under about 8 inches of water. As the water

levelin.the' railroad bay rises, flow through the doorway would ,

be expected to decrease due to loss of driving head.

. H 4 !

i y.

A.

f El-2 e

I l

Figure 1: Turbine EWJ;ng Railroad Bay l l Transformer Yard l.

u l I

l Unit 2,6.9kV Unit Boards l l Unit 1,6.9kV Unit Boards l l Doorway - 29 fl. wide l )

l54 ft- l <

l l Railroad Tracks l l32211. l .

l l

RISK DETERMINATION EVALUATION Event Tree:

I The following event tree summarizes the risk evaluation for the I flooding of the TBRB. In this representation, the downward branch indicates a failed or degraded condition.

l A B C D E F G )

0.0033 -

OK (unit boards remain operable) l 1.0 OK (unit boards are flooded and fail / at least l one EDG operable) l 2.09E-4 OK (AC power restored prior to core l l uncovery at 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months ) l l 0.4 l CD (no AC power and a seal LOCA results l in core uncovery at 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months ) ,

l 0.072 l OK (AC power restored prior to core l uncovery at 1.7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months ) l 1.0 CD (no AC power and TDAFW inoperable results in core uncovery at 1.7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months )

Total CD = (0.0033)(1.0)(2.58E-4)

= 8.5E-7/ year El-3 l'

Event Description A. Various precipitation rates and durations can result in the capacity of the storm drain system in the vicinity of the railroad bay doorway being exceeded (initiating event / precursor frequency). For a given precipitation rate (inches / hour) and duration (hours or minutes), the frequency for the rainfall event can be predicted based on historical meteorological data contained in Attachment 2 I of Calculation SCG1S506. The storm drain system is l designed to accommodate precipitation rates / durations expected at a frequency of 0.04/ year (i.e., a1 in 25 year storm). When the discharge of the bus duct coolers was routed to 1 of the 2 storm drains near the railroad bay l doorway, the capacity of the storm drain system was l reduced. The 800 gpm discharged to 1 of the storm drains in the vicinity of the railroad bay doorway is only a l fraction of the total capacity of the storm drain system and based on the 800 gpm flow rate would not be expected to significantly degrade the capacity of these storm drains. However, other hydraulic phenomena introduced by the 800 gpm discharge reduced the capacity of the storm drains in the vicinity of the railroad bay doorway so that it could accommodate precipitation rates / durations expected at a frequency of 0.1/ year (i.e., a 1 in 10-year storm). This degraded performance is based on a qualitative estimate of the degraded storm drain system capacity. Due to the passive capacity (floor area) and drain capacity (through the floor drains) of the railroad bay, a precipitation rate / duration equal to the design or degraded capacity of the storm drain system would not result in a significant accumulation of water in the railroad bay. That is, a higher precipitation rate / duration is required than that associated with a 1 in 25-year storm for the design condition and a 1 in 10-year storm for the degraded condition to recult in significant water accumulation in the railroad bay. Based on historical precipitation rates / durations, it is estimated that it would take a precipitation event with a frequency of 0.02/ year and 0.04/ year for the design and degraded conditions to result in significant water accumulation in the railroad bay (these estimates are based on historical meteorological data from Attachment 2 of Calculation SCG1S506) .

The precipitation events with the potential to flood the railroad bay, are the high intensity short duration events, that is, those lasting less than 60 minutes. The longer duration lower intensity events do not have the potential to flood the railroad bay since the runoff from these events is low enough to be accommodated by a storm El-4

o 1 b ~

l drain system. The potential to overwhelm the storm drains during the longer duration events would only occur due to

)

a general rise in the river water elevation that could only result from a very large scale precipitation event.

This form of precipitation event would be expected to 1 occur in conjunction with other known extreme weather I patterns, such as the remnants of a hurricane passing over the site. In the event of such a severe weather pattern, j the plant would be expected to have sufficient time to implement compensatory actions to protect the 6.9-kV unit boards, including the ability to start emergency diesel generators (EDGs) if warranted. Assuming the distribution of precipitation events is logarithmically distributed toward the short duration high intensity events as shown in Attachment 2 of Calculation SCG1S506, the fraction of storms lasting less than 60 minutes is no more than 0.50.

The increase in initiating event frequency for the degraded storm drain condition is calculated as: the difference in the precipitation event frequency for the I normal and degraded conditions, multiplied by the fraction of high intensity /short duration precipitation events, which could potentially result in flooding of the railroad bay and multiplied by the fraction of a year the degraded storm drain condition existed. The degraded storm drain conditions existed for approximately 4 months; therefore, the change in initiating event frequency to be used in l determining the increase in core damage frequency (CDF) is: (0.04 - 0.02) (1/2) (4 /12) = 0.0033.

B. This event tree branch is associated with the operator failing to sandoag or seal the TBRB doorway. Note that there are two assistant unit operators (AUOs), whom are by virtue of their duty station, cognizant of weather conditions: the turbine building AUO and the outside AUO (whose duty station is the essential raw cooling water (ERCW) pumping station, EDG buildings, component cooling water (CCW) pumping station and switch yard). In addition, as a matter of course, severe weather conditions are monitored on a daily basis by Operations (via National Oceanographic Atmospheric Administration weather radios in the main control room) and by the TVA transmission load dispatch organization which notifies Operations of severe weather conditions. In addition, Operations was aware of the potential for flooding the TBRB and knew to close/ seal the railroad bay doorway. Based on this information, the probability that the operator fails te close/ seal the railroad bay doorway is judged to have an upper bound of no more than 0.1 with a lower bound of approximate.' y 0.01.

However, in this rick evaluation, no credit is taken for El-5

4 operator action (i.e , guaranteed failure of operator action).

C. Not used.

D. A station blackout (SBO) occurs due to loss of all 6.9-kV unit boards and a failure of any of the 4 EDGs to operate within the first hour. A single EDG can power any of the 4 shutdown boards (2 shutdown boards / unit) by use of the 6.9-kV shutdown utility bus. The procedure to establish these cross connections is emergency procedure EA-202-4.

The major equipment necessary to establish seal cooling, i reactor coolant system (RCS) inventory control, and (secondary side) shutdown heat removal is 1 component {

cooling system (CCP) and 1 auxiliary feedwater (AFW) pump i per unit, 1 CCS pump and 1 ERCW pump, which have a combined electrical load of: 2*418 + 2*477 + 282+ 571 =

2643 kW (reference Appendix C of AOP-P.01). This electrical load is much less than the continuous rated j capacity of 4400 kW of a single EDG. 1 E. The turbine-driven auxiliary feedwater pump (TDAFWP) fails to operate. Without alternating current (AC) power .

available, shutdown heat removal is accomplished by the I TDAFWP providing AFW to the steam generators which are depressurized to reduce RCS pressure to minimize RCS inventory loss through the reactor coolant pump (RCP) seals. The procedure for mitigating a loss of all AC power is ECA-0.0, which also directs operators to EA-250-1 (Load Shed of Vital Loads After Station Blackout) to extend battery life to 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months . Depressuri zation of the RCS allows the cold leg accumulators to inject, which increases the time until core uncovery. The RCP seal leak rates used in the SQN probabilistic safety analysis (PSA) are based on NUREG-5116. Using these RCP seal leak rates, it is estimated that core uncovery occurs no earlier than ,

approximately 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months given that the TDAFWP is operable I and the RCS is depressurized as shown in the Table below:

El-6

r

,c -

Description Time RCS RCP RCP Remaining RCS Inventory (hours) pressure Leak leakage above the Top of the Core +

(psia) Rate (gal) CLAs (gal)

(gpm)

Seal cooling lost 0 2250 84 0 93353.85 (note 1)

Seals Fail 1 1162 84 5040 88314 (note 2)

RCS 1.216 650 635 8225 80089 depressurizes - (note 3) (note 4) to CLA pressure CLAs depleted 2.44 175 428 31445 48644 (note 5) )

Core Uncovery 5.35 175 279 48694 -49 {

(1) This is the total RCS liquid inventory above the top of the core plus the water volume of four cold leg accumulators (CLAs) minus the liquid volume that 14e RCS decreases in going from Tave = 578 F (power operations) to Tave = 280 F (minimum temperature to avoid i pressurized thermal shock (PTS) concerns, ref. ECA-0.0).

(2) This is the RCS pressure based on inventory loss due to seal leakage (3) This is the RCS pressure at which the CLAs begin to inject (4) This le** ;'s is calculated as:

elCP leak rate = 1000*sqrt(Pave /2250)

Pave - average RCS pressure during the time interval (5) This is the RCS pressure at the end of depressurization (per ref. ECA-0.0).

It is important to note that the seal leak rates used in the SON PSA are based on the RCP seals utilizing the old style 0-rings. Currently, all the RCPs utilize the high temperature 0-rin, , which are not expected to suffer significant degradation under SBO conditions (WCAP-10541, Revision 2, Supplement 1). Under nominal seal leakage conditions and with the TDAFW pump operable and the RCS depressurized, the time to core uncovery can be increased j to days. After depletion of the station batteries at 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months , the TDAFWP can be operated locally without direct current control power per EA-3-7 (Local Operation of TDAFW Pump After Loss of DC Control Power). In addition, the steam generator atmospheric relief valves can be operated locally (by reach rod, see 0-GO-8) to maintain the steam generators in a depressurized condition. Although there will be no steam generator level indication, given the relatively slow change in shutdown heat after 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months , AFW flow should not require significant adjustment after 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months .

Failure of the TDAFWP resulcs in the steam generators boiling dry followed by loss of RCS inventory through the pressurizer safety relief valves. In this case, the SQN PSA predicts core uncovery in 1.7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months .

ri-7

I i

~

l

' l T. .AC: power is restored prior to.2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months . Without the TDAFWP '

. operable,. core uncovery~ occurs at 1.7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months . Therefore,

.AC power must-be restored prior to 1.7 hours8.101852e-5 days 0.00194 hours 1.157407e-5 weeks 2.6635e-6 months to prevent i core damagt. In this event, AC power is restored by either returning a single. unit board or EDG to operation.

(A single' unit board.has a sufficient capacity to support I the expected. power. requirements for two units, see [D] l

.above.) -In'the SON-PSA, for a'SBO, the failure to recover AC power prior to core melt is assigned a probability of 0.06472 when the TDAFWP is not-operable (split fraction REC 6) and a probability of 0.04690 when the TDAFWP is l operable (split fraction REC 5) . These recovery split fractions account for both recovery of a EDG or a recovery of'offsite power. Since it is not convenient to determine which portion of the split fraction REC 6 is due to the recovery of a EDG or recovery of offsite power, in this evaluation-it has been set to guaranteed fail. Note that this over predicts the increase in core damage since by setting ~ REC 6 to guaranteed fail, the probability of returning an EDG to operation within 1.7' hours has been assigned a probability of 0.0. !

'G. AC power'is restored within 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months by either repairing an inoperable diesel generator or by returning 1 unit board '

to. service; As discussed under (E) above, when the TDAFW pump operates, the steam generators /RCS are depressurized and the loads on the vital batteries are shed, then core uncovery does not occur.for 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months . Recall from (A) that the: event. precursor for this scenario is a.relatively short duration, high intensity rain event that overwhelms Lthe capacity-of the storm drain systes, which results in the excess _ rain water flowing into the TBRB. Once the rainfall subsides, the accumulated water will quickly drain from the railroad bay.through the storm drain system and~ turbine building drains. Therefore, it is likely that the unit boards would be accessible and available to be reenergized within 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months . This leaves about 4 hours4.62963e-5 days 0.00111 hours 6.613757e-6 weeks 1.522e-6 months to

' return a. unit board to service. Discussion with the system engineers and electrical craft indicate that a unit board that has been submerged to the level of the main bus could be returned to service in about 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months . In addition,.the SON PSA indicates that the probability of recovering an EDG in 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months is greater than 0.60.

-Therefore, it is considered almost a guaranteed success that AC' power can be restored within 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months . However, in this evaluation, REC 5 has been assigned a probability

.ofE0.40, which means that no credit has been taken for

, recovery'of the unit boards within 5 hours5.787037e-5 days 0.00139 hours 8.267196e-6 weeks 1.9025e-6 months , only recovery of-an EDG'is considered.

El-8

SQN PSA Model Results:

Revision 1 of the SQN PSA was used to determine the increase in core damage frequency (CDF) for the above described scenario. In the model, the AC recovery split fractions RECS and REC 6 were replaced with RECA and RECB, respectively, for SBO with and .

without AFW available, respectively. As discussed under (F) above, RECA was assigned a value of 0.4 and RECB was assigned a value of 1.0 (i.e., no possibility of recovering AC power in 2 hours2.314815e-5 days 5.555556e-4 hours 3.306878e-6 weeks 7.61e-7 months ). The model predicts a loss of offsite power (LOSP) induced CDF of 1.25E-5. The LOSP initiating event frequency in the model is 0.0485 so the conditional core damage frequency for a flood induced LOSP is 1.25E-5/0.0485 = 2.58E-4. Using the ,

above event tree, the increase in CDF for this event is:

(0. 0033) (1. 0) (2. 58E-4 ) = 8.5E-7/ year.

As discussed in the Risk Determination Section under (E), SQN RCPs use high temperature O-rings which are not expected to suffer significant degradation under SBO conditions. Under nominal seal leakage conditions and with the TDAFWP operable and J the RCS depressurized, the time to core uncovery can be increased to at least a day. This amount of time guarantees the recovery of AC power prior to core uncovery and the above event tree reduces to core damage only occurring when the TDAFWP fails to operate. In addition, as discussed in the Risk Determination '

Section under (1B ) , operator action to seal the railroad bay doorway is expected to reduce the potential for flooding of the unit boards by at least a factor of 0.1. When full credit is taken for these conservatisms in this risk evaluation, the increase in CDF for this best estimate scenario is 5.6E-8/ year.

Conclusion:

Based on this evaluation, the increase in core damage frequency due to this event is less than 1E-6 and is classified as " Green" per the NRC Significance Determination Process.

This evaluation is conservative in that:

a logarithmic distribution for precipitation events which could potentially result in flooding of the railroad bay has been used in determining the initiating event frequency, -

no credit has been taken for operator action to mitigate the flooding event by sealing the railroad bay doorway, and

the analysis does not credit the near certain recovery of the unit boards prior to core uncovery.

El-9

.

ML20216J935

Text

Reference:

Conclusion:

Navigation menu

Search