ML20072S097

Forwards Comments on Preliminary Accident Sequence Precursor Analysis of Operational Event & Related Condition for Licensee Peer Review at Plant Unit 2

ML20072S097
Person / Time
Site:	Quad Cities
Issue date:	09/08/1994
From:	Schrage J COMMONWEALTH EDISON CO.
To:	Russell W Office of Nuclear Reactor Regulation
References
NUDOCS 9409130420
Download: ML20072S097 (15)

Text

.- - .- -

Comm:nwrith Edi;on

. ~ 1400 Opus Place

'. As Downers Grove, Ilknois 60515 September 8,1994 i

Mr. William T. Russell, Director Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington D.C. 20555 Attn.: Document Control Desk

Subject:

Quad Cities Station Unit 2 Comments on Preliminary Accident Sequence Precursor Analysis of OperationalEvent and Related Conoition for Licensee Peer Review at l Quad Cities, Unit 2.  ;

NRC Docket Nos. 50-265

Reference:

C.P. Patel to D.L. Farrar letter dated August 9,1994 i

Mr. Russell, l l

In the referenced letter, the NRC Staff transmitted the Preliminary Accident  ;

i Sequence Precursor Analysis of Operational Event and Related Condition for Licensee Peer j Review at Quad Cities, Unit 2, for review'and comment by Commonwealth Edison  !

(Comed). The NRC Staff requested comments within 30 days from receipt of the referenced letter. This letter transmits Comed's comments in Attachments A and B. j If there are any questions, please contact John L. Schrage at 708-663-7283. I Very truly your , i f ohn L. Schrage 9409130420 940908 Nuclear Licensing Administrator j PDR ADOCK 05000265 P

PDR 1

i Attachment A: Comments on Preliminary Accident Sequence Precursor (ASP) Analysis of Operational Event ,

and Related Condition for Licensee Peer Review '

at Quad Cities Unit 2 in 1993 ' i Attachment B: Probabilities of Failing to Recover Offsite Power for Quad Cities Station cc: J. Martin, Regional Administrator - Region Ill C. Miller, Senior Resident inspector - Quad Cities Station 1 R. Pulsifer, Project Manager - NRR g Office of Nuclear Facility Safety - IDNS 0 t

k:nla quadzletters: asp 1993.wpf:1 I b0120 m j

Attachment A Comments on Preliminary Accident Sequence Precursor (ASP) Aralysis of e i Operational Event at Quad Cities Unit 2 in 1993 i

The review guidance provided with the preliminary analysis stated that comments  !

regarding the analysis should address four areas. Those areas and Comed comments I are given below.

1. Characterization of possible plant response.

Comments: The discussion of possible plant response in the preliminary analysis appears accurate. The discussion of the Safe Shutdown Makeup Pump (SSMP) is ,

appropriate as this system is not included in the ASP event trees for BWRs.

The preliminary analysis addresses the degraded condition of the Unit 2 Diesel Generator Cooling Water Pump (DGCWP). The discussion points out that, had a loss of offsite power (LOOP) occurred during the 7-month period of interest, the Unit 2 Emergency Diesel Generator (EDG) could have operated for some time before failure of the Unit 2 l DGCWP. The discussion also mentions monthly runs (of approximately 2.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> each) l of the Unit 2 DGCWP. The preliminary analysis concludes that the average lifetime of l the Unit 2 DGCWP during the 7-month period was 7.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. This is discussed further l in item 3 below.

2. Representation of expected plant response used in the analytical models.

Comments: Modifications of sequence probability calculations were included in the preliminary analysis to account for high pressure injection sources that were not modeled in the ASP event trees. The availability of the SSMP to provide high pressure injection was included in the analysis of ASP Sequences 67 and 69. Also, the preliminary analysis included the Reactor Core Isolation Cooling (RCIC) system in ASP Sequence 69.

Nevertheless, ASP Sequences 67 and 69 appear to be invalid despite these modifications. These sequences show that failure of high pressure injection leads directly to a core damage state; the possibility of using the Automatic Depressurization System (ADS) followed by Low Pressure Core injection (LPCI) is omitted, presumably based on i an implicit assumption that AC power is not available for LPCI because the successful long-term recovery of AC power occurs too late to prevent core damage. The expected plant response assumed by the preliminary analysis, however, is that the Unit 2 EDG could possibly have been available for an average of 7.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, and use of ADS and  ;

LPCI as an altemative to high pressure injection would be possible. I A-1 l

l i

.--. .- - -- - . - - - =- -- - -.

3. Representation of plant safety equipment configuration and capabilities l at the time of the event.

Comments: The preliminary analysis uses an average lifetime of 7.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> for the Unit 2 DGCWP during the 7-month period based on the assumption that the pump was failed after the last surveillance run. This assumption is conservative because the pump may '

have been capable of continuing to run, but an accurate estimate of the pump's remaining life following the last surveillance run does not appear to be feasible. '

1

4. Assumptions regarding equipment recovery probabilities. l Recovery of Offsite Power Comments: The preliminary analysis assumes for the dominant sequence (Case 2b, Sequence 83) that recovery of offsite power must occur before battery depletion (at 11.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />) to prevent core damage. The preliminary analysis for.the dominant sequence uses probability estimates of 0.66 and 0.23, respectively, for failing to recover offsite power in ,

the short-term and prior to battery depletion. Based on a review of the calculation of the 1 core damage probability for this sequence, these values appear to have been multiplied in the calculation to give a total probability of 0.15 for failing to recover offsite power prior '

to battery depletion. This value appears to be overly conservative, however, when I compared with the value calculated using site-specific data for the Quad Cities IPE.

The analysis for the Quad Cities IPE, based on NUREG 1032, gives 0.03 as the  :

probability of failing to restore offsite power within 11.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. (See Attachment B for  ;

details.) ,

Furthermore, the assumption that recovery of offsite power must occur before battery  ;

depletion to prevent core damage is conservative. At 11.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, decay heat would be reduced. A Comed analysis using the MAAP code (assuming that RCIC was used to maintain level until battery depletion) predicts that the reactor water level would not drop  !

to the top of active fuel until approximately 14.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> after the initiation of the event and that core damage would not begin untillater. Restoration of offsite power prior to 14.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> would allow the use of RCIC, SSMP, or ADS /LPCI to prevent core damage, and the analysis in Attachment B gives 0.024 as the probability of failing to restore offsite power within 14.6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />. Therefore, use of the value of 0.03 (for 11.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />) would be conservative.

Use of the value of 0.03 for the probability of failing to restore offsite power would appear to significantly reduce (by approximately a factor of 5) the calculated core damage probability for the dominant sequence. L A-2

i Recovery of 1/2 DGCWP j Comments: The preliminary analysis assumes a probability of 1.0 for operator failure to ,

restore cooling water to the 1/2 EDG. This is based on a cited paper

• that gives an l estimated probability of 0.26 that an auxiliary operator would report to the diesel  !

4 generator room within 10 minutes of an EDG trip while running or starting.  !

The model cited does not appear to be completely applicable to the problem with the 1/2 i DGCWP failing to start, however. The cited paper states that among delays included in  ;

development of the probability distribution used are " delays for the control room to contact  ;

the operator and describe the problem." The cited paper also states that "the control  ;

room operators...may attempt to manually restart the affected engine from its control room ,

switch." The cited paper also states that "It is estimated that the operator's response time for going to the turbine building from any of his normal duty locations is approximately 5  ;

to 10 minutes after notification." i i

l in the case of the 1/2 EDG starting and loading to its safety bus but the 1/2 DGCWP failing to start, no manual restart attempt would be made by the control room operator.  !

Section 0.1.2 of the preliminary analysis report mentions that loss of EDG cooling is not specifically annunciated in the main control room and would result in a " Diesel Generator 1/2 Trouble" annunciator. Nevertheless, main control room panel meters would show the output of the 1/2 EDG and main control room panel indicating lights would show that the 1/2 DGCWP was not running. (The main control room has run indicating lights but no manual switch for the 1/2 DGCWP.) Therefore, no attempt would be made by the control room operator to start the already running 1/2 EDG. As stated in the Licensee Event Report, " Simulator lesson plans include this scenario. Operators, as part of training, dispatch personnel to the diesel generator whenever it is autostarted. This dispatch increases the likelihood that the inoperable DGCWP condition would have been promptly corrected." For these reasons, the probability of an operator reporting to the 1/2 EDG room within 10 minutes is estimated to be greater than the value of 0.26 given in the cited paper.

In summary, the preliminary analysis assumption of a probability of 1.0 for operator failure to restore cooling water to the 1/2 EDG appears overly conservative. Nevertheless, this assumption appears to have a much lesser impact on quantification of the dominant sequence than does the assumption concerning recovery of offsite power discussed above.

'J. W. Road and K. N. Fleming, " Electric Power Recovery Models," Proceedings of the International Topical Meeting on Probabilistic Safety Assessment, PSA '93, January 26-29,1993.

A-3

l 1

l l

'l Attachment B '

Probabilities of Falling to Recover Offsite Power for Quad Cities Station Source: Quad Cities IPE Calculation Note OC-CN-93-003,

" Power Recovery Probabilities for Quad Cities IPE," approved 11/22/93.

l l

l l TABLE OF CONTENTS ,

l l

1 Section Page 1.0 Introduction B-2 2.0 Determination of Offsite Power Cluster Subgroups B-2 2.1 Grid Reliability / Recovery Group B-4 2.2 Severe Weather / Recovery Group B-5 2.3 Extremely Severe Weather Loss of Offsite Power B-5 Frequency Group 3.0 Determination of Offsite Power Cluster Group B-6 4.0 Probability of Not Recovering Power at Time X B-6 5.0 References B-7 Relevant Tables B-8 B-1

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

I l

1

1.0 INTRODUCTION

l The purpose of this calculation is to document the source of the probabilities used for two nodes in the Quad Cities Loss of Offsite Power and Station Blackout Plant Response Tree j notebooks (References 1 and 2). The two nodes in question are:

I ROP 1 -

Recovery of Offsite Power in Time to Preclude Core Damage (LOOP)

ROP 2 -

Recovery of Offsite Power in Time to Preclude Core Damage (SBO)

The methodology used to determine these probabilities is contained in this document and is based on the information found in NUREG-1032 " Evaluation of Station Blackout Accidents- ,

at Nuclear Power Plants" (Reference 3). The steps to determine these probabilities include:

i Determine the "Offsite Power Cluster Group" that Quad Cities should be included in by implementing the selection criteria found in NUREG-1032. '

I Determine the probability of recovering power in time to prevent core damage l

~

(ROP 1/ ROP 2) using the frequency distributions contained in NUREG-1032.

Each of these steps will be discussed in detail in the following sections.

2.0 DETERMINATION OF POWER CLUSTER SUBGROUPS i i

The Offsite Power Cluster Grouping is an attempt to account for any relationship between switchyard design characteristics, local weather, power recovery procedures, and the duration  !

of loss of offsite power events at a given plant. The methodology used to determine the appropriate Offsite Power Cluster Group is based on the selection criteria found in Tables ,

l A.2, A.3, A.6 and Tables A.8 through A.11 of NUREG-1032. The Offsite Power Cluster Group is determined by the unique combination of four subgroups. These subgroups, defined by grid design and local weather, are shown below:

B-2

i I

1. Switchyard Configuration Group (11, 12, 13)

. 2. Grid Reliability / Recovery Group (G1, G2, G3, G4)

3. Severe Weather-Induced Loss of Offsite Power Frequency / Recovery Group (SR1, SR2, SR3, SR4, SR5, SR6, SR7, SR8, SR9, SR10)

4. Extremely Severe Weather-Induced Loss of Offsite Power Frequency Group (SS1, SS2, SS3, SS4, SSS)

Two factors, grid and switchyard design, are potentially significant with regard to frequency and duration of loss of offsite power events. The impact of these design factors is determined by the blend of responses (yes/no) to the following statements. The unique blend of yes and no responses define the impact of these features and the subsequent Switchyard Configuration Group to which the plant belongs.

A. Independence of offsite power sources to the nuclear plant.

1. All offsite power sources are connected to the plant through one switchyard.

YES i

2. All offsite power sources are connected to the plant through two or more I switchyards, and the switchyards are electrically connected. NO

3. All offsite power sources are connected to the plant through two or more switchyards or separate incoming transmission lines, but at least one of the AC sources is electrically independent of the others. NO B. Automatic and manual transfer schemes for the Class 1E buses when the normal source of AC power fails and when the backup sources of offsite power fa;l.

1. If the normal source of AC power fails, there are no automatic transfers and there is one or more manual transfers to preferred or alternate offsite power sources. NO

2. If the normal source of AC power fails, there is one automatic transfer but no manual transfers to preferred or alternate offsite power sources. NO

a. All of the Class 1E buses in a unit are connected to the same preferred power source after the automatic transfer of power sources. YES B-3

i . l

b. The Class 1E buses in a unn connected to separate offsite power sources after the automatic tran w Of power sources. NO

3. After loss of the normal AC power source, there is one automatic transfer. If this source fails, there may be one or more manual transfers of power sources to preferred or alternate offsite power sources. YES

a. All of the Class 1E buses in a unit are connected to one preferred power source after the first automatic transfer. YES

b. The Class 1E buses in a unit are connected to separate offsite power sources after the first automatic transfer. NO

4. If the normal source of AC power fails, there is an automatic transfer to a preferred j source of power. If this preferred source of power fails, there is an automatic transfer to another source of offsite power. NO l l

a. All of the Class 1E buses in a unit are connected to the same preferred 1 power source after the first automatic transfer. NO

b. The Class 1E buses in a unit are connected to separate offsite power l sources after the first automatic transfer of power source. NO The responses to the above statements are based on information contained in the Electric Power Systems Notebook (Reference 4) and show that Quad Cities falls into Switchyard Configuration Group 13. All group designations based on design factors are shown in Table B-1 below. I 2.1 Grid Reliability / Recovery Group I The Grid Reliability / Recovery Group combines into a single factor the inherent reliability of the local power grid and the ability of the plant to rapidly recover from the loss of power.

From the Quad Cities Initiating Events Notebook (Reference 5), the frequency of grid related losses is 2.7E-3 per year. Quad Cities has never experienced a grid related loss of offsite power. NUREG-1032 implies use of a grid loss frequency of 1E-2 per year if no loss of power events have occurred at the individual site. However, use of either frequency will B-4

place Quad Cities in Grid Group G1. Table B-2 below shows the relationship of grid loss frequency to Grid Group.

• The next step in determining the Grid Reliability / Recovery Group is to identify the recovery group. The recovery group qualitatively identifies the plant's ability to recover power within 1/2 nour following a grid blackout. The plant must have the capability and procedures to recover offsito (non-emergency) AC power to the site within 1/2 hour following a grid blackout to be considered in the R1 group. By default, all other plants not in the R1 group are

! contained in the R2 group. Quad Cities does not have specific procedures in place for recovering power in this time frame and therefore falls into the R2 recovery group. This combination of factors leads to a Grid Reliability / Recovery Group of GR5 as identified by Table B-3 below.  :

2.2 Severe Weather / Recovery Group The severe weather / recovery group combines into a single factor the likelihood of loss of offsite power due to severe weather events with the ability of the plant to recover from the  ;

r event in a rapid manner. From the Quad Cities initiating Events Notebook, the frequency of  ;

I severe weather related loss of offsite power events at the Quad Cities station is 8.1E-3 per year. This frequency, in combination with the recovery group R2 identified earlier in section 2.1, defines a Severe Weather / Recovery Group of SR7. Tables B-4 and B-5 below show the l 1

manner in which severe weather frequency and plant recovery ability are grouped to arrive >

l at the SR7 group designation.

l 2.3 Extremely Severe Weather Loss of Offsite Power Frequency Group I

l This group is determined strictly by the frequency of extremely severe weather, postulated in this case. This event consists of losses of offsite power caused by extreme weather such as hurricanes, very high winds (greater than 125 mph) and major damage to switchyards due to tornado strikes. Restoration of offsite power following these events is assumed to require J B-5  !

1

at least 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. The Quad Cities initiating Events Notebook gives a frequency of 2E-4 per year for this type of event. The group designators associated with each occurrence frequency range are shown in Table B-6 below. With an occurrence frequency of 2E-4, Quad Cities is considered to be in group SS1.

3.0 DETERMINATION OF OFFSITr oOWER CLUSTER GROUP l ,

The subgroups previously defined in section two permit determination of the offsite power cluster group. These subgroups 13, G1, SR7 and SS1 can be inserted into the matrix shown in Table B-7 below to determine the proper cluster group. The results of this process show that Quad Cities should be included in Offsite Power Cluster Group 2.

4.0 PROBABILITY OF NOT RECOVERING POWER AT TIME X NUREG-1032 gives frequency distributions for durations of loss of offsite power events for each of the cluster groups (table A.11 of Reference 3). The probability of not recovering power at each hour was derived using the median values of the frequency distribution data contained in this table. The frequency at each time interval was divided by the frequency at time = 0 hours0 days <br />0 hours <br />0 weeks <br />0 months <br /> to normalize the values and thus render probabilities. Since the information contained in NUREG-1032 did not have values for every hour, values for each missing hour through 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> were obtained by using log extrapolations, which provide a good fit to the to the loss of offsite power frequency duration curves presented in Figure A.15 of NUREG-1032. The values for each of the intermediate hours not given was estimated by the following equation:

LOGPO+LOGM P=10 where x = probability at previous hour given and y = probability at the next succeeding hour given B-6

. = _ . .. - - , _ . - . - _ _.

This eradon gives a value for a point midway between two known times x and y. The results of applying this equation can then be used again to determine a new intermediate value and the equation reapplied until all the unknown values are determined. This information is summarized in the following table for Offsite Power Cluster Group 2 for events of up to 16 hours1.851852e-4 days <br />0.00444 hours <br />2.645503e-5 weeks <br />6.088e-6 months <br /> duration.

i DURATION Loss of offsits power frequency Probability of not (HR) (Table A.11 of NUREG-1032) recovering power Eag Normalized  :

0 0.1040 1.000E+00 1.000E+00 1/2 N/A N/A 6.068E-01 1 N/A N/A 3.682E-01 2 0.0141 1.356E-01 1.356E-01

, 3 N/A N/A 9.553E-02 4 0.0070 6.731 E-02 6.731 E-02 5 N/A N/A 5.852E-02  ;

6 N/A N/A 5.088E-02 7 N/A N/A 4.424E-02 8 0.0040 3.846E-02 3.846E-02 9 N/A N/A 3.569E-02 10 N/A N/A 3.312E-02 11 N/A N/A 3.074E-02 j 12 N/A N/A 2.852E-02 13 N/A N/A 2.647E-02 14 N/A N/A 2.456E-02 15 N/A N/A 2.280E-02 16 0.0022 2.115E-02 2.115E-02

5.0 REFERENCES

1. Loss of Offsite Power (LOOP) Plant Response Tree Notebook, Quad Cities Nuclear Power  !

Station Units 1 and 2, prepared by IPEP, October 1993, Rev. O.

2. Station Blackout (SBO) Plant Response Tree Notebook, Quad Cities Nuclear Power Station Units 1 and 2, prepared by IPEP, October 1993, Rev. O.

3. Baranowsky, P.W., et. al., " Evaluation of Station Blackout Events at Nuclear Power Plants", U.S. NRC Report NUREG-1032, June 1988.

B-7

~

l l

4. System Notebook for Electric Power System, Quad Cities Nuclear Power Station Units 1 and 2, prepared by IPEP, October 1993, Rev. O.

~

5. Initiating Events Notebook, Quad Cities Nuclear Power Station Units 1 and 2, prepared by IPEP, October 1993, Rev. O.

TABLE B-1 DEFINITION OF SWITCHYARD CONFIGURATION GROUPS GROUP FACTOR 11 A1, A2, or A3 and B4 12 A1 or A2 and B2b or B3 13 A1 or A2 and 81 or B2a TABLE B-2 DEFINITION OF FREQUENCY OF GRID GROUPS GROUP FREQUENCY OF GRID LOSS PER SITE YEAR G1 less than 1.67E-2 G2 1.67E-2 to 5.0E-2 G3 5.0E-2 to 0.167 G4 equal to or greater than 0.167 B-8

TABLE B-3 DEFINITION OF GR GROUPS FREQUENCY GROUP RECOVERY GROUP GRID l RELIABILITY / RECOVERY GROUP (GR)

G1 R1 GR1 G2 R1 GR2 G3 R1 GR3 l

G4 R1 GR4 G1 R2 GR5 G2 R2 GR6 G3 R2 GR7 l

TABLE B-4 l

DEFINITION OF FREQUENCY OF SEVERE-WEATHER GROUPS l

! GROUP FREQUENCY PER SITE YEAR S1 less than 3.0E-3 S2 3.0E-3 to 1.0E-2 l S3 1.0E-2 to 3.0E-2 l

S4 3.0E-2 to 0.1 l

SS 0.1 to 0.33 l

I i

B-9 l

i

{

TABLE B-5 DEFINITION OF SR GROUPS FREQUENCY GROUP RECOVERY SEVERE-WEATHER / RECOVERY GROUP GROUP l

S1 R1 SR1 S2 R1 SR2 S3 R1 SR3 S4 R1 SR4

, 1 l S5 R1 SR5 l S1 R2 SR6 j S2 R2 SR7 S3 R2 SR8 i S4 R2 SR9 ,

l l S5 R2 SR10 l TABLE B-6 DEFINITION OF EXTREMELY SEVERE WEATHER-INDUCED GROUPS l GROUP FREQUENCY PER SITE YEAR SS1 less than 3.0E-4 SS2 3.04E-4 to 8.3E-4 SS3 8.3E-4 to 3.0E-3 SS4 3.0E-3 to 1.0E-2 SS5 greater than or equal to 1.0E-2 B-10

TABLE B-7 l

CLASSIFICATION OF OFFSITE POWER CLUSTER GROUPS CLUSTER I GR SR SS l GROUP l l

1 1,2 1,3,5 1,2,6,7 1,2 l 1,2 1,3,5 1,6 3 l 1,2 1,3,5 3 1,2 l 2 1,2 1,3,5 8 1,2,3 1,2 1,3,5 4 1-4 1,2 1,3,5 2,3,7 3,4 1,2 1,3,5 1,6 4 3 1,3,5 1,2,6,7 1-4 3 1,3,5 3,8 1,2 3 1,3,5 3 3,4 3 1,3,5 4 1-4 3 same as cluster 7 same as cluster same as cluster 2 and 1 2 and 1 2 and 1 4 1,2,3 1,3,5,7 10 1-5 i

B-11