ML17347A798

Radiological Data Prepared for Resolution of USI A-46

ML17347A798
Person / Time
Site:	Saint Lucie, Turkey Point, 05000000
Issue date:	06/16/1988
From:	EBASCO SERVICES, INC.
To:
Shared Package
ML17347A796	List: ML17347A795 ML17347A797 ML17347A798 ML17347A800
References
REF-GTECI-A-46, REF-GTECI-AC, REF-GTECI-SC, TASK-A-46, TASK-OR NUDOCS 8807050479
Download: ML17347A798 (57)

Text

~

~

I RADIOLOGICALDATA PREPARED FOR THE RESOLUTION OF USI A-46 ST LVCIE UNIT 1 AND TURKEY POINT UNITS 3 AND 4 NUCLEAR POWER PLANT SITES Prepared By Ebasco Services inc.

Greensboro, North Carolina 88070S0P7q 8850000250 PDR ADOCK 0 PDR P

I ~

I I

1'

I

~ ~

UNRESOLVED SAFETY ISSUE A-45 OFFSXTE IMPACT ANALYSIS ST+

LUCIE UNIT 1

AND TURKEY POINT UNITS 3

& 4 1~

INTRODUCTION The issue of Seismic Qxalification of Mechanical and Electrical Equipment in Operating Nuclear Power Plants was designated by the Nuclear Regulatory Commission (NRC) as Unresolved Safety Issue A-46 (USI A-46) in December, 1980.

In order to address USI A-46, the NRC determined that each licensee of an operating plant not previously reviewed to current seismic criteria would be required to perform a seismic verification review of that plant.

Furthermore, the NRC determined that this requirement is a backfit as defined by 10 CFR 50.109 and must therefore be )ustified by a value/impact analysis.

As required by the backfit rule, a generic value/impact analysis was performed by the NRC.

The results of this analysis, which are presented in NUREG-1211

[1]*, show that a seismic review will result in a significant safety benefit.

The value/impact analysis consists of a qualitative assessment and quantitative examples'ne of the quantitative examples discussed in the NUREG is the analysis of the failure of an electrical cabinet anchorage following a safe shutdown earthquake (SSE) and the subsequent release of radioactive material.

Based on the conclusions drawn in NUREG-1211, the NRC published a list of power plants that are required to perform a seismic verification review.

Florida Power and Light has three reactors on the list:

St. Lucie Unit 1 and Turkey Point Units 3

& 4 ~

To help determine whether a seismic review is cost-effective for St. Lucie Unit 1 and Turkey Point Units 3

& 4, plant-specific analyses of the ten hypothetical PWR accidents described in WASH-1400 were performed for the two sites.

Since the Turkey Point reactors have the same specifications, a single set of analyses were performed for the two units, which will bc hereafter referred to simply as Turkey Point.

• Numbers in square brackets are reference

numbers, followed (in some cases) by page or table number

I I

b

2 ~

METHODOLOGY The Calculation of Reactor Accident Consequences.,

Version 2

(CRAC2) computer code was.used to assess releases of radioactive materials from hypothetical accidents involving St. Lucie 1 or Turkey Point.

The CRAG computer code was originally developed by Sandia National Laboratories to support the Reactor Safety Study, WASH-1400.

Version 2 of the code contains corrections and improvements over the original code used in the preparation of WASH-1400.

2.1 Input Data The code includes all necessary data to perform generic analyses.

To perform a plane-specific analysis, certain of these data sets had to be modified.

These included the radionuclide core inventory of each reactor, the population distribution in the area surrounding the site, and the local meteorology.

2.1.1 Radionuclide core inventories WASH-1400 [2:3-3] lists a radionuclide inventory for a typical Westinghouse four-loop PWR with a power level of 3200 Mwt.

These data are included with the CRAC2 code.

The most recent available data on the St. Lucie 1 core inventory is contained in an Exxon Nuclear report [3:Table 3.3].

That report, however, lists values only for radioisotopes of iodine, cesium, tellurium and the noble gases; CRAC2 requires a much longer list of radionuclides.

Since the buildup of radionuclides in the core depends on the power level, it is reasonable to calculate the core inventory of a given reactor by multiplying the WASH-1400 inventories by the ratio of the power level of that reactor to the power level of the WASH-1400 reactor.

St. Lucie 1 has a stretch power level of 2700 MWt.

To model accidents involving this plant, each of the WASH-1400 core inventory was multiplied by the factor 2700/3200

~ 0.84.

The power levels of the Turkey Points Units 3

4 are rated at 2200 MWt.

The Turkey Point inventory was calculated by multiplying the WASH-1400 activities by the ratio 2200/3200

~.69.

The validity of this approximation for St. Lucie 1 was verified by comparing the inventories of three radiologically significant radionuclides listed in the Exxon report to the values calculated for the CRAC2 analyses.

These calculated invetories are shown in Table 1, below.

~ I ~

s

Table 1

COMPARISON OF CORE INVE¹ORY CALCULATIONS FOR ST.

LUCIE 1

Nuclide I-1 31 I-1 33 Xe-133 8'E+7 1.7E+8 1 ~ 7E+8 7 '4E+7 1.47E+8 1.47E+8 Core Inventor (curries)

WASH-1400 Exxon CRAC2 7.17E+7 1.43E+8 1.43E+8 See ref. [2:3-3]

bSee ref. [3:Table 3.3] ~

The inventories of all three nuclides, as calculated for use by the CRAC2

code, are close to those in the Exxon reporto 2.1.2 Meteorological data Site>>specific meteorological data was used for both the St. Lucie 1 and Turkey Point analyses'he St. Lucie Unit 2 Updated FSAR [6] contains comprehensive meteorological data for the St. Lucie site, which is occupied by both units.

This data, which is in the form of a joint frequency distribution, is shown in Table 2, below.

Since CRAC2 only recognizes six Pasquill stability classes, the frequencies for class G were added to those of class F in the input data file.

Table 2

ST+ LUCIE SITE PERCENT JOINT FREQUENCY OF OCCURRENCE OF WIND SPEED AND STABILITY CLASS Stability Wind Speed Range (Meters per Second)

Class 0.0-1.5 1.5-3.0 3.0-5.0 5.0-7.5 7.5-10.0

)10.0

~ 12

~ 09

~ 15 1 ~ 58 7 ~ 33

1. 54

.69 2'4 1 ~ 18 1 ~ 33

10. 28 19'6

2. 29

.63 7'6 2.21 2.16 14.62 15.25

.32

.01 1 59

~ 32

~ 34 2'5 2 70 F 02 F 00

.01

.04

~ 02

~ 24

~ 14

.00 F 00

.00 F 00 F 00 F 01

~ 02

.00

.00

I I ~

0

A similar table was compiled for Turkey Point by utilizing data collected at the FP&L South Dade site, located about 6 1/2 miles SSW of the Turkey Point control room.

Hourly data for the years 1985 and 1987 (the most recent years for which complete data were available) vere used to produce a single set of joint frequencies, shown in Table 3, below.

Table 3

TURKEY POINT PERCENT JOINT FREQUENCY OF OCCURRENCE OF WIND SPEED AND STABILITY CLASS Stability Class 0 - 3 Wind Speed Range (Miles per Hour) 3 7 7 - 12 12 15 15 - 18 18 - 24

> 24 0.03

0. 02

0. 06 0+89 4 ~ 06 4.20 1.22 1.10 0.89 1.45 7'5 17.62 6'6 3.43 6.05 3.78

3. 68 12'2 11.83 0.23 0'9 2.78 1.16 0 93 3.28 1 49 0 F 01 0'7 0.43 0.24 0'8 0.85

0. 37 0.01 0.02 0.07 0.02 0.05 0.21 0'3 0 F 00 0.00 0.01 0.01 0.01 0.02 0.00 0.00 Wind direction frequency data (i.e., the data vhich is used in constructing a

wind rose) must be entered separately into the CRAC2 input files.

Since this data is used in conjunction vith non~niform population distributions, it is listed along vith the population distributions around the two sites in Tables 4 and 5, belov.

2.1.3 Population distribution St.

Lucie 1

Tvo sets of analyses for were performed for St. Lucie 1 ~

One set models an accident occurring at the present time. It utilized the estimated population distribution in the St. Lucie site vicinity for the year 1983, the most recent year for vhich such data was available.

The other analyses predict the impact of a hypothetical accident occuring near the end of the plant s life, and utilized the projected distribution for the year 2030.

The data for both distributions vere taken from the St. Lucie 2 Updated FSAR [6:Figs. 2.1-5

For the purpose of calculating off-site impacts, the area surrounding the site is divided into five concentric rings.

Since there are no inhabitants within one mile of the center of the site, the first ring lies between one and ten miles of the center.

All subsequent rings are ten miles wide, as measured from the inner to the outer radius.

Each of the rings is further subdivided into 16 angular

segments, one for each ma)or direction of the compass, yeilding a total of 80 sectors The populations of each sector are listed ia Table 4, below.

~Tntke Point Two sets of analyses were also performed for Turkey Point:

one utilized a distribution computed from the actual 1980 census data, while the other used the pro]ected distribution for 2020, which was based on the 1980 data [4].

Since there are no inhabitants within three miles of the center of the site, the first ring lies between three and ten miles of the center.

Otherwise, the population sectors were constructed in the same manner as those for St.

Lucie.

The 1980 and pro]ected 2020 populations of each sector are listed below.

~

~

Table 4

POPULATION DISTRIBUTION AND WIND FREQUENCY IN THE ST, LUCIE SITE VICINITY Directiop from site

)0 124 380 20 30 Distance From Site to Outer Ed e of Sector (miles):

50 40 Wind b

Frequency

5. 43 6.32 6.20 4.21 SSE SSW SW WSW W

130 295 4491

)1303 11859 25969 39216 28741 10977 59677 15582 118839 2791 45031 7979 26311 36014 38055 3610 17842 0

0 11916 29764 28387 70904 4250 26989 2808 46175 0

44601 2580 3975 695 1169 1620 32743 4749 15152 0

0 0

0 0

0 24633 55348 761 1653 5389 13460 2581 6446 0

0 1491 7082 5550 96380 0

0 75324 184360 8902 21948 685 1643 152 379 11888 26113 10898 23936 0

0 3050 14472 4199 30301 0

0 153388 272151 91074 223866 5478 11705 14380 30727 724 1628 3205 6370 172 411 0

0 2871 36025 3'3 4.21 6.87 6.48 5.86

4. 05 5.25 6.40 9.78 9')

8.82 The first value for each sector is the estimated 1983 population, the value under it is the pro)ected 2030 population.

Wind frequency in percent, calculated from data in the ref. [6:Table 2.3-28]

g ~

~

Table 5

POPULATION DISTRIBUTION AND WIND FREQUENCY IN THE TURKEY POINT SITE VICINITY Direction from site 20 30 10 Distance From Site to Outer Edre of Sector (miles):

40 50 Wind Frequency NNE NE 7367 10091 195780 265232 8945 11277 349855 469171 420427 565446 237122 339939 330024 446225 285429 529725 182125 283476

4. 12 3.37 3.44 ENE 2.97 3.64 ESE 3.52 Sc 5.25 SSE SSW SW 649 1029 2161 3457 196 313 5515 8808 4997 7937 1369 2157 9.09

6. 13 2.67 5.98 WSW NNM 12 16 7505 10260 33795 46066 23859 32587 11061 15155 29 40 2570 3467 11057 14843 8797 12029 o9005 96065 0

0 10773 14867

8. 51 15.91 12.30 7.68 5.31 Data are from reference

[4].

The first value for each sector is the 1980 census population, the value under it is the pro]ected 2020 population.

Wind frequencies in percent, site (see text).

calculated from data collected at South Dade

t

~

~

~

~

~

~

e 3 'ERIFICATION he val cabinet anchorage.

The object of this analysis was to compare the CRAC2 calculated dose to that stated in the NUREG.

To check t idity using CRAC2 to perform the value/impact analyses, the code was used to calculate the conditional public dose resulting from a hypothetical PWR 2 release from the Byron plant, using the assumptions listed in NUREG-1211 in the discussion of the seismic failure of an electrical 3.1 Input Data For this case the data used are from NUREG-1211 (i.e.,

a power level of 1120 MWe and a uniform population density of 340 persons per square mile).

When data were not specified in NUREG-1211, data from WASH-1400 were used.

The latter include the core inventory t'2:3-3) and ]oint frequency meteorological data for a central midwest plains site [2:5-4], listed in Table 6, below.

Table 6

CENTRAL MIDWEST PLAIN SITE PERCENT JOINT FREQUENCY OF OCCURRENCE

~ OF WIND SPEED AND STABILITY CLASS Stability Wind Speed Range (Meters per Second)

Class 0 -

1 1 2 2 3 3 - 4 4 5 5 - 6 6 7 7

A No Rain Rain

~ 41 1.43 1 ~ 92 F 05 2.08 1.24

.75

~ 98

.01

.03

~ 06

.Ol

~ 03

~ 05

~ 03

~ 03 B No Rain Rain

~ 35 78

.65

~ 76

.63 20 18

.02

.00

.01

.00 F 00

~ 03

.02

.26

.00 C No Rain Rain

~ 59 1 ~ 13 1 ~ 00

~ 82

~ 02

~ 02

~ 06

.07

~ 94

.42

~ 05

.02

~ 25

~ 35 F 00

.01 D No Rain Rain 3 ~ 87 5.95 6 ~ 22 5 ~ 57 5 ~ 22 2 ~ 05 1.78 2 ~ 16

.18

~ 50

~ 56

.71

.51

~ 31 32

.33 E No Rain F No Rain Rain 9o46 3'7 1.77

~ 38 o15

.07

~ 02

~ 03

~ 09 01

~ Ol

.00 5 ~ 48 5 ~ 75 4 ~ 81

3. 18

2. 13 1 ~ 13 o 24

~ 20

~ 40

~ 37

~ 18

~ 06

~ 04 1 ~ 1L

~ 02 00 00

.03

.00

.00

'a

3.2 Result of Verification The conditional public dose calculated by CRAC2 was 4.4E+6 man-rem.

This compares favorably with the NUREG-1211 value of 5E+6 man-rem, particularly if, as is likely, the NUREG-1211 value is the NUREG/CR-3568 [5] value of 4.8E+6 man-rem rounded to one significant figure.

The difference in the two values "might be explained if the NUREG/CR-3568 analysis utilized hourly averaged data from the Byron site rather than the )oint frequency table shown above.

In any

case, the values are sufficiently close to )ustify the use of the CRAC2 methodology for the site-specific value/impact analysis.

4 RESULTS The calculated collective radiological absorbed dose-equivalents to the populations within 50 miles of the St. Lucie and the Turkey Point sites, respectively, due to the ten postulated reactor accidents are listed in Table 7, below.

Also listed is the average dose-equivalent, calculated by weighting each of the ten values by the probability of occurrence of that accident.

A complete description of each of the accident scenarios can be found in ref. [2:2-1] (attached)

~

10 Table 7

COLLECTIVE ABSORBED DOSE-EQUIVALENTS WITHIN 50 MILES (MAN-REM)

Release Site:

St. Lucie Category Year:

1983 2030 Turke Point 2020 1980 PWR 1A PWR 1B PWR 2 PWR 3

PWR 4 PWR 5

PWR 6

PWR 7

PWR 8 PWR 9 2.5E+6 1 ~ 2E+6 1.7E+6 2, 1E+6 9.9E+5 4.7E+5 7.5E+4 1.3E+3 4 'E+4 7.4E+1 7.5E+6 3.4E+6 4.8E+6 5.8E+6 2.8E+6 1.4E+6 2.0E+5 3 'E+3 1 'E+5 1 'E+2 3.2E+6 2 'E+6 2 5E+6 2 'E+6 1.4E+6 5.5E+5 7.2E+4 1 ~ 3E+3 3 'E+4 7

OE+1 4.4E+6 2.9E+6 3.3E+6 3 'E+6 1.9E+6 7.4E+5 9,1E+4

1. 6E+3 4.8E+4 8.2E+1 Weighted Average 3.6E+4 1 'E+5 4 'E+4 6.6E+4 See ref. [2:2-1]

5 ~

CONCLUSIONS The calculation presented in NUREG-1211 [5:19], used as a quantitative example of potential value/impact, employed a conditional public dose of 5E+6 man-rem for a PWR 2 release, All four PWR 2 releases calculated in the present study result in doses that are less than or equal to this value.

\\

~ ~

~ ~

~

REFERENCES 1.

Chang, T. Y. and N. R. Anderson, 1987:

Regulatory Analysis for Resolution of Unresolved Safety Issue A-46, Seismic Qualification of Equipment in Operating Plants.

%KEG-1211, U.S. Nuclear Regulatory Commission.

2.

Reactor Safety Study, An Assessment of Accident Risks in U.S. Commercial Nuclear Power Plants, App. VI.

WASH-1400, U.S. Nuclear Regulatory Commission, 1975.

3.

Exxon Nuclear Company, Inc.:

St. Lucie Unit 1 Radiological Assessment of Postulated Accidents.

XN-NF-84-85(P), Rev.

1 ~

4.

Bolch, W. E.,

1986:

Population Density Within a Fifty Nile Radius of the Turkey Point Nuclear Facility.

(Unpublished).

5.

Battelle Pacific Northwest Laboratory:

Handbook for Value-Impact Assessment.

NUREG/CR-3568, U.S. Nuclear Regulatory Commission, 1983.

6 ~

Florida Power

& Light Company:

St. Lucie Plant Unit No. 2, Updated Final. Safety Analysis Report.

~

~

~

~

~

~

I

~

ANALYSIS AND CONCLUSIONS A.

Gene ic Letter 81-14 Seismic uglification of Auxiliar Feedwater S stems Before discussing Generic Letter 87-02, it is instructive to review the NRC staff's use of risk analysis technology and Value/Impact

studies, under the umbrella of 10CFR50.109, to resolve Generic Letter 81-14 at Oconee Units 1, 2,

and 3,

on a

plant specific basis.

A copy of the analysis prepared by the Reliability and Risk Assessment Branch is enclosed in this section.

(1)

The NRC staff accepted the use of plant specific data in response to the Generic Letter.

(2)

The plant specific data, using risk analysis technology, and Value/Impact studies performed by the NRC, provided an upper bound estimate of the risk worth as being

$540,000 worth of safety improvement.

(3)

The NRC staff agreed to conclude the issue without requiring the Oconee units to seismically qualify their two seismically unqualified AFWS pumps because to do so would exceed

$540,000.

Instead, a

course of action based upon risk analysis technology and unrelated to the seismic qualification of the AFWS pumps but costing less than

$540;000 was agreed

upon, and a Safety Evaluation Report was written.

B.

U eso ved Safet Issue USI A-45 Shutdown Deca Heat removal It is also instructive to review the use of risk analysis technology and Value/Impact studies under the umbrella of Page 1

~ l ~

I J

I

I ~

~

10CFR50.109 to resolve USI A-45.

Studies to date show that it would not be prudent management of the nuclear industry's resources if the "dedicated system" or other proposed physical changes were to be imposed on the industry.

FPL participated with Sandia National Laboratories in its A-45 evaluations at both St.

Lucie Unit 1

(NUREG/CR-4710) and at Turkey Point Units 3/4 (NUREG/CR-4762).

Many similarities exist between (the seismic portion of) A-45 and A-46 and the NRC has also recognized this by providing many references from A-45 in NUREG 1211.

In the referenced

NUREGs, Sandia prepared seismic risk assessments for St. Lucie Unit 1 and Turkey Point Units 3/4, and performed extensive seismic walkdowns in 1986.

The Sandia conclusions may be summarized as follows:

NUREG/CR-4 7 10 St. Lucie Unit 1 SSE=0.10 (1)

The region is seismically benign (2)

Appendix A to 10CFR. 100 mandates a

minimum SSE of

0. 10g, otherwise the SSE might be

0. 09g as calculated by Sandia.

(3)

All components examined had median failure acceleration levels above the SSE.

(4)

The components with the minimum seismic margin were:

(a)

Refuelin Water Stora e Tank, Buckling with Anchor Bolt Yielding at 0.24g Page 2

~

I ~

I

~

~ I

/

e I

(b)

Condensate Stora e Tank Buckling with Anchor Bolt Yielding at 0.26g (5)

The present potential radiation exposure to the public due to seismic hazard is 1.83 man-rem/year.

At a value of

$ 1,000 per man-rem, this is

$ 1830 at a seismic level of 0.24g.

Sandia calculated this value using the probabilistically generated sum of all WASH 1400 release categories.

NUREG/CR-4762 Turke Point Units 3

4 SSE

= 0. 15 (1)

The region is seismically benign (2) If the procedures in Appendix A to 10CFR100 were applied to Turkey

Point, a

value of peak ground acceleration based on earthquake history of less than 0.10g would be obtained and hence the SSE would be set at the prescribed minimum of 0. 10g.

The actual value calculated by Sandia was 0.06g.

(It is FPL's intention to make application for a SSE of 0. 10g at Turkey Point at a future date.)

(3)

All components examined had median failure acceleration levels above the SSE.

(4)

The components with the minimum seismic margins were:

(a)

Refuelin Water Stora e Tank, Buckling with Anchor Bolt Yielding at 0.24g Page 3

~

~ ~

I ~

(b)

Condensate Stora e Tank Buckling with Anchor Bolt Yielding at 0.26g (c)

Co onent Coo in Water Heat Exchan ers Support Failure at 0.39g (d)

Diesel Generator Control Panels Support Failure at 1.0g (5)

The potential radiation exposure to the public due to seismic hazard is l. 31 man-rem/year.

At a value of

$ 1,000 per man-rem this is

$ 1,310 at a seismic level of

0. 24g.

Sandia calculated this value using the probabilistically generated sum of all WASH 1400 release categories.

From the foregoing, it can be clearly seen that St.

Lucie Unit 1

and Turkey Point Units 3/4 do not pose undue risk to the public health and safety at this time.

C.

NUREG 1211 As mentioned previously, in October of 1986 the CRGR instructed the NRC staff to prepare a

generic Value/Impact analysis to justify the information request contained in Generic Letter 87-02.

This generic analysis may be found in Section V of NUREG 1211.

Page 4

j

~

~

General Comments (1)

The generic analysis only addresses the information request, and does not apply to the correction of any "deficiencies" for which case by case additional analyses will be required.

(Section V, Page 16)

(2)

The generic analysis "is based on the cost effectiveness, considering how much risk reduction is achieved for the money spent".

(Section V, Page 16)

(3)

The generic analysis adopts USI A-45 data as being appropriate.

(Section V, Page 18)

Comments on Generic uantitative Exam le Section V

Part A

(1)

The generic initiating event, P (i)," is a

SSE with a return frequency of 2.5E 4/RY.

Although not stated, the source, was from work done by Dr.

M.

P.

Bohn of Sandia using averages for SSE's east of the Rocky Mountains.

Using the,EPRI/SOG approved methodology, FPL has calculated the site specific values for P (i) to be 1.27E - 5/RY at St.

Lucie Unit 1

and 4.97 E-6/RY at Turkey Point Units 3/4.

These values are substantially lower than the generic value and reflect the seismically benign nature of FPL's sites.

(2)

Using CRAC 2,

the NRC staff has calculated the generic conditional public dose "R(D/CR)" to be 5E+6 man-rem, based on a

WASH 1400 Category 2 release, a

1120 MWe PWR and Byron (mid-west) meteorology and demographics.

Using CRAC 2,

FPL has calculated, the site specific conditional public dose to be 1.7 E +

6 for St. Lucie Unit 1 and 3.6E

+

6 for Turkey Point Units 3/4.

FPL also used a

Page 5

WASH 1400 Category 2 release to allow comparability, however it is FPL's position that the implicit assumption of a

probability of 1.0 by the NRC staff that there would be a

Category 2 release should be reviewed,

because, for example, there is no postulated failure of the containment spray system which would be condition precedent for a Category 2

release.

One alternative would be to use the probabilistically generated sum of all WASH 1400 release categories as was done by Sandia in NUREG/CR-4710 for St.

Lucie Unit 1

and in NUREG/CR-4762 for Turkey Point Units 3/4.

Another alternative would be to use a

Category 5

release which is suitable for the postulated leakage type failure mode.

The result would be substantially reduced conditional public doses for both the NRC and FPL analyses.

(3)

The generic risk potential was calculated at 940 man-rem per reactor and the safety benefit from the A-46 verification program was assumed by the NRC staff to reduce the seismic risk potential by more than an order of magnitude.

This means that 94 man-rem of risk would remain and that the reduction would be 846 man-rem, as opposed to the full 940 man-rem used by the NRC staff to justify (at

$ 1,000 per man-rem) an expenditure of up to

$940,000 per reactor to implement the information request contained in Generic Letter 87-02.

(4)

The probabilities of failures of equipment anchorages "P(AF)"; of electrical cabinet failure "P(C/AF)"; cabinet relationship to hot shutdown "P(HF/C)"'ore melt "P(CM/HF)";

and containment failure "P(CR/CM)"

are subjectively based and should be reviewed.

For

example, when the first four are combined with P(i), the core melt probability is over one third of the total core melt probability from all causes calculated for some plants.

The result of the

review, using more realistic assumptions, Page 6

would be substantially reduced conditional public doses for both the NRC and FPL analyses.

Summarizing the foregoing, at the present time FPL has only substituted site specific values for P(i); fission product inventory; meteorology; demographics and remaining reactor life.

With these substitutions, and using the balance of the NRC assumptions and methodology, and considering the risk worth of safety improvement, no more than

$ 15,000 and

$ 8,000 can be prudently justified to implement the information request at St.

Lucie Unit 1

and Turkey Point Units 3/4, respectively.

In addition, further

~ reductions are most likely if the potential HASH 1400 release categories are reviewed and if the assumed probabilities are reviewed.

Comme ts on Exam le 1

Section V

Part B

Combustion En ineerin PWR with SSE

= 0.10 Although not stated in NUREG 1211, this is St. Lucie Unit 1, and the NUREG 1211, using information from NUREG/CR 4710, states that St.

Lucie Unit 1 is not expected to experience any seismic problems up to 0.24g which is 2404 of the level of the SSE.

The SSE is the highest seismic level of concern to A-46.

NUREG/CR 4710 stated that stiffening the Refueling Water Storage Tank and Condensate Storage Tank (present seismic margins 0.24g and 0.26g) could reduce the conditional public dose from 1.83 man-rem/year

($ 1,830) to 0.13 man-rem/year

($ 130).

The NRC staff, in NUREG 1211, performed a Value/Impact study to illustrate that up to

$50,000 could be prudently justified to make these changes.

Page 7

I I 1 I

Since only

$50,000 could be prudently justified by the NRc to reduce potential risk at the 0.24g 0,26g level, it follows that less if anything could be prudently justified at the SSE (0.10g) level.

Additionally, it is difficult to understand the rationale of using examples of seismic levels beyond the SSE to justify taking action in an issue whose upper seismic level is the

SSE, particularly when the regulatory analysis states that no problems are expected at the seismic level of the SSE.

Even at the 0.24g 0.26g level, the regulatory analysis states that St. Lucie Unit 1 is "a plant where few if any modifications would be required" (Section V, Page 24).

Conclusions 1.

St. Lucie Unit 1 and Turkey Point Units 3/4 provide adequate protection of the public health and safety at this time.

2 ~

St.

Lucie Unit 1 and Turkey Point Units 3/4 have, in 1986, undergone extensive seismic

scrutiny, including walkdowns and risk assessments whose results have been published in NUREG/CR-4710 and NUREG/CR-4762, respectively.

These results indicate that the sites are seismically benign.

3 4 St.

Lucie Unit 1 and Turkey Point Units 3/4 would have to spend at least

$750,000 per plant to implement the information request in Generic Letter 87-02, which would involve a self assessment to identify "deficiencies" against new generic seismic criteria.

Using the NRC staff's methodology and site specific data, FPL has calculated that no more than

$15,000 and

$8,000

.could be prudently justified.

Page 8

1 I I

4.

St.

Lucie Unit 1

and Turkey Point Units 3/4 do not fall within the "envelope" of plants covered by the NRC staff's Value/Impact analysis.

5.

The NRC staff should affirm that satisfactory resolution of Generic Letter 87-02 and USI A-46 has been achieved for St.

Lucie Unit 1 and Turkey Point Units 3/4.

Page 9

I

~

~ ~

I

APPLICATION OF EARTHQUAKE EXPERIENCE TO LOW SEISMICITY SITES THE INTENT OF THIS PRESENTATION IS TO DFMONSTRATE THAT EVEN IF A MAXIMUM EVALUATION EARTHQUAKE WERE TO OCCUR AT EITHER THE ST.

LUCIE OR THE TURKEY POINT SITES A COMPARISON WITH TklE EXPERIENCE DATA FOR'TRONG MOTION EARTHQUAKE SHOWS THAT NO SEISMIC INDUCED DAMAGE WOULD RESULT EVEN FOR EQUIPMENT AND DISTRIBUTION SYSTEMS WHICH ARE NOT DESIGNED TO BE EARTHQUAKE RFSISTANT.

1

~

~

~

IN FIGURE l IS SHONN THE FOUR ACTUAL RESPONSE SPECTRA ON NHICH THE SSRAP EXPERIENCE DATA BOUNDING SPECTRUM IS BASED.

I

~

~

I I

2.0 1.d 1.6 c

1A X.

O.

1 02 CC LQ 1.0 U

os V

LLI o.e ee00

~ ~

~

~

~

~

0

~

0

~

0 0

0 0

0

~

0 0

0 0

0

~

~

~

0

~

~

~

~

0

~

~

0 0

r

~ e

~

0 ~

~

~ ~

~

~ ~

~

~ ~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

~

Ii V

5% DAMPlNG

~re ~

~

~

~

~

~

~

ei '0

/

~

'g.r'----

COALINGA

EL CEMTRC 8YLMAR

---

LLOLLEO

~ REFERENCE SPECTRUM f1,5 x BOONDlNG SPECTRUMj

~~

era~

~ ~

,.h,w ~

~ ~

~ 0 op~

~ 0 ~ ~

~~

~ ~

~ e ere ~ ~ ~ ~ 0 ~ ~ ~ ~ 0 ~ ~ ~ ~ ~ ~ ~ ~ ~ ~

~ eI0 e ~ ~ ~ ee e 0000 ~ eeea o~ ~ orna,

~ em

~

0.0 10 20 25 2S FgcQUENCY (H>)

FIGURE 1

HORIZONTAL RESPONSE SPECTRA

IN FIGURE 2 IS SHOWN THE SSRAP DEVELOPED "DRAFT" BOUNDING SPECTRUM.

THIS BOUNDING SPECTRUM REPRESENTS APPROXIMATELY TWO-THIRDS OF THE ESTIMATED AVERAGE FREE-FIELD ACCELERATION TO WHICH EQUIPMFNT WAS ACTUALLY EXPOSED AT SITES WITH ESTIMATED MEAN PEAK GROUND ACCELERATION IN EXCESS OF ABOUT 0.4g.

IN RESPONSF TO EARTHQUAKES OF THIS INTENSITY THERE WERE A VERY LIMITED AMOUNT OF DAMAGE AND FAILURES (LESS THAN 0.01 PERCENT OF COMPONENTS AT RISK)

OR A CONDITIONAL PROBABILITY OF FAILURE, Pf ~ 10-4.

ALL SUCH FAILURES WOULD HAVE BEEN ELIMINATED IF SSRAP CAVEATS WERE FOLLOWED.

$.D 5'ic DAMPING BOUNDING SPECTRUM GROLIND ACCELERATION = 0.33 g~

0.0 52 46 FREQUENCY (Hz) 20 28 FIGURE 2 SEISMIC MOTION BOUNDING SPECTRUM HORIZONTAL GROUND MOTION

J

~

I k

IN FIGURE 3 IS SHOWN A SPECTRUM WITH A SAFETY FACTOR OF 1.5 WHICH DEFINES THE THRESHOLD OF DAMAGE TO INDUSTRIAL EQUIPMENT (USING THE BEHAVIOR OF PIPING AS A

LOWER BOUND) IN EARTHQUAKES BASED ON THE 0.2g OBSERVATION PRESENTFD IN NUREG 1061, VOL. 2 ADDENDUM.

THIS CURVE WAS DEVELOPED BY TAKING THE SSRAP BOUNDING SPECTRA AND NORMALIZING IT TO A 0.13g ZPGA.

WITHIN THE BOUNDS OF THIS SPECTRA NO DAMAGE TO CONVENTIONALLY CONSTRUCTED INDUSTRIAL EQUIPMENT IS ANTICIPATED WITHOUT RECOURSE TO CAVEAT LIMITATIONS.

E f

~

1

~

0.5 0 4 R0 0.3 K

Ul LLJO 0 2 K

0.1 V

LLJ 0(0 5% DAMPING THRESHOLD EARTHQUAKE BOUNDING SPECTRUM GROUND ACCELERATION=.133 G

0.0 0

12 16 FREQUENCY (HZ) 20 24 28 FIGURE 3 THRESHOLD EARTHQUAKE DAMAGE SPECTRUM HORIZONTAL G'ROUND MOTION

8

~

9

~

e

IN FIGURE 4 IS SHONN A COMPARISON OF THE SSRAP BOUNDING AND THE THRESHOLD EARTHQUAKE DAMAGE SPECTRA TO THE MAXIMUM EVALUATION SPECTRA USED IN THE SEISMIC QUALIFICATION OF THE ST.

LUCIE AND TURKEY POINT PLANTS.

II

< ~

~ 1.0 0.8 CO R0 0.6 LLlV 0.4 CC 0.2 UJ CO MAX.: EVALUATION SPECTRUM TURKEY POINT 5% DAMPING SSRAP BOUNDING SPECTRUM THRESHOLD EARTHQUAKE DAMAGE SPECTRUM MAX.: EVALUATION SPECTRUM--ST. LUCIE 0.0

. 0 12 16 mEQUENCY (HZ) 20 24 28 FIGURE, 4 COMPARISON OF MAXlMUMEVALUATlON SPECTRA FOR THE ST.'UCIE AND TURKEY POlNT SlTES TO B'OUNDING AND THRESHOLD DAMAGE SPECTRA

e+

C

OBSERVATION:

2.

THE BOUNDING SPECTRA DEVELOPED BY SSRAP OBVIOUSLY ENVEl OPS BY A SIGNIFICANT MARGIN THE EVALUATION BASIS AND RFALISTIC SPECTRA DEFINED FOR ST.

LUCIE AND TURKEY POINT SITES.

THE THRESHOLD DAMAGE SPECTRUM ENVELOPS THE EVALUATION BASIS SPECTRA AT ST.

LUCIE AT ALL POINTS.

IN ADDITION, IF CURRENT 10CFR 100 APPENDIX A REQUIREMENTS WERE USED THE EQUIVALENT SSE SPECTRA AT TURKEY POINT WOULD BE ANCHORED AT A O.lg ZPGA AND ALL POINTS OF A TURKEY POINT SSE SPECTRUM WOULD BE ENVELOPED AS WELL.

CONCLUSION:

BECAUSE THE DESIGN BASIS (SSE EQUIVALENT) SPECTRAFOR ST.

LUCIE AND TURKEY POINT PLANTS FALL WITHIN THE THRESHOLD DAMAGE SPECTRUM BASE ON EARTHQUAKE EXPERIENCE, THERE SHOULD BE NO NEED (WITH THE POSSIBLE EXCEPTION OF LARGE FLAT BOTTOM VERTICAL TANKS) TO REEVALUATE SEISMIC DESIGN ADEQUACY OF MECHANICAL AND ELECTRICAL EQUIPMENT WIlHIN THE CURRENT SCOPE OF A-46 FOR THESE VERY LOW SEISMICITY PLANT SITES.

r

~ l

~