L-83-567, Forwards Addl Info Re Loss of Offsite Power,In Response to NRC & 830926 Telcon

Forwards Addl Info Re Loss of Offsite Power,In Response to NRC & 830926 Telcon

ML17214A565
Person / Time
Site:	Saint Lucie
Issue date:	11/21/1983
From:	Williams J FLORIDA POWER & LIGHT CO.
To:	John Miller Office of Nuclear Reactor Regulation
References
L-83-567, NUDOCS 8311280278
Download: ML17214A565 (25)

Text

I REGULATORY DEFORMATION DISTRIBUTION S EM (RIDS)

ACCESSION NBR:83]1280278 DOC.DATE~ 83/11/21 NOTARIZED:- No F4CIL:50 335 St,'Lucie "Planti Unit 1i Flor ida Power 8 Light Co.

,AUTH ~ NAME AUTHOR AF F ILIATION WILLIAMS~J~H ~

F l or ida* Power 8 Light Co,

,'RECIP ~ NAME RECIPIENT AFFIL'IATION ILLEReJ,RE Operating Reactors Branch 3

SUBJECT:

Forwards addi info re loss of offsite poweriin response

-to NRC 831012 ltr 8 830926 telcon

~

DISTRIBUTION CODE ~

A015S COPIES 'RECEIVED!LTR.Q ENCL, Q

'SIZE+

TITLEs OR Submittal:

Onsite Emergency Power System NOTES'OCKET

¹ 05000335 RECIPIENT ID CODE/NAME NRR ORB3 BC 01 INTERNAL NRR/DL/ORAB 1?

• -NRR/DSI/PSB 14 RGA2 COPIES t.TTR ENCL 7

1 1

1 1

1 RECIPIENT "ID CODE/NAME ICSB 09 0Q RM/DDAM'I/MI8 18 COPIES LTTR ENCL' 1

1 1

1 EXTERNAL; ACRS NRC PDR NT/S 16 02 6

6, 1

1 1

LPDR NSIC 03 05 1

1 1

1

.TOTAL NUMBER OF COPIES REQUIRED'TTR 23 ENCL 23

t jj ff

~ 1PNjNN 1I N

~ Ij J[

N j

(N (

r N

~

~

jI ffff

]

J

~

~

N Q

N ff ff I lt 1

P OX 14000, JUNO BEACH, FL 33408 QC.

FLORIDAPOWER & LIGHTCOMPANY November 2I, I 983 L-83-567 Office of Nuclear Reactor Regulation Attention:

Mr. James R. Miller,Chief Operating Reactors Branch 7/3 Division of Licensing U. S. Nuclear Regulatory Commission Washington, D.C. 20555

Dear Mr. Miller:

Re:

St. Lucie Unit I

Docket No. 50-335 Loss of Offsite Power-Additional Information Re uest In response to your letter of October l2, l983, attached please find Florida Power 8c Light Company's answers to questions asked during a

telecom with NRC on September 26, I 983.

Very truly yours,

~ J. W. Williams, Jr.

Vice President Nuclear Energy Department JWW/RJS/cab Enclosures cc:

J. P. O'Reilly, Region II Harold F. Reis, Esquire 8311280278 831121 I t PDR ADOCK 05000335 p

PDR PEOPLE... SERVING PEOPLE

ATTACHMENT l HISTORICAL BACKGROUND

Historical Pers ective A

licable to All Questions Two sets of analyses (each of which contained an analysis for both the Seized Rotor (SR) and the Loss of Offsite Power (LOOP) events) are germane to the present review of the LOOP at the NRC.

The first is the SLl Uprating submittal, which was submitted as Reference (7) and requested an increase in licensed operating power from 2560 to 2700 MWTThe second

set, Reference (8), contains the LOOP analysis currently under review by the NRC.

The second set was submitted in satisfaction of a commitment by Florida Power and Light (FPL) to provide a reassessment of the SR and the LOOP events to include additional assumptions which were not a part of the original SLl design basis from the Final Safety Analysis Report (FSAR) but which were included in Post-FSAR Standard Review Plans (SRPs)

This section provides some historical background to aid in understanding the differences in assumptions used.

Many of the points which the NRC has asked FPL to clarify appear to arise from the assumption differences between the two analyses.

Cycle 4 operation at a licensed power of 2560 MWT began on May 10, 1980.

The-licensing basis for the Reload Safety Evaluation (RSE) supporting Cycle 4 operation at this power level included analyses for the SR and LOOP events which were consistent with the design basis approved in the FSAR in that no Worst Single Failure (WSF) assumption was made and Auxiliary Feedwater (AFW) was assumed to be manually initiated.

Reference (7) is the 2700 MWT Uprating submittal forwarded to NRC in November 1980.

The various analyses contained in it used Cycle 4 parameters.

It was both FPL's and the NRC's expectation that the review of Reference (7) would be completed sometime during the scheduled Cycle 4 and that the uprating would be accomplished "mid-cycle" ~

Consistent with all pr'ior 2560 MWT, analyses, the new 2700 MWT SR and LOOP analyses in the stretch submittal did not account for a

WSF, and both analyses assumed that auxiliary feedwater was manually initiated.

The accident at Three Mile Island (TMI) resulted in delays for all non-essential licensing activities by NRC.

Review of FPL's Uprating submittal was consequently not completed until after Cycle 4 shutdown which occured on September 9,

1981.

Because it was recognized that the first cycle of 2700 MWT operation would be Cycle 5, three analyses from Reference (7)

(Main Steam Line Break, Excess

Load, and Steam Generator Tube

Rupture) were reanalyzed incorporating Cycle 5 parameters and Post-TMI hardware modifications, notably the interim automatically initiated auxiliary feedwater system which had, by this time, been installed at the plant.

These analyses were submitted to NRC by References (9) and (10).

(The interim system provides auxiliary feedwater to both steam generators after a three minute time delay. ) An assessment of the applicability of the other Cycle 4 analyses to Cycle 5

operations was submitted as Reference (ll).

Prior to NRC approval for operation at 2700 MWT, FPL commited to submit two additional reanalyses (the SR and the LOOP) with additional assumptions not contained in the Unit 1 design basis but consistent with SRPs in use at that time.

Cycle 5 operation at 2700 MWT commenced on December 2,

1981.

Reference (8) forwarded these reanalyses for SR and LOOP.

The LOOP analysis contained in Reference (8) is the subject of the current review.

These analyses differed from all prior SL1 submittals in that NRC had requested assumptions in accordance with recently updated Standard Review Plans (SRPs),

i.e., the Seized Rotor also assummed a

LOOP and a

WSF; and the LOOP event assumes a

WSF.

In both cases the WSF was identified to be the failure of the ADV in the open position.

Since Cycle 5 was already in progress, Cycle 5 parameters were used.

Additionally, in order to envelope the effects of an anticipated plant hardware

change, the upgraded "SMART" automatic AFW system was assummed for Reference (8), rather than the interim automatic AFW system currently installed.

ATTACHMENT 2 SPECIFIC QUESTIONS CONCERNING THE LOOP ANALYSIS FROM TELECON WITH NRC (Paraphrased in some cases for clarity)

NRC Question What is the Worst Single Failure (WSF) for the LOOP events Answer The WSF for the LOOP event has been determined to be the failure of an Atmospheric Dump VALVE (ADV) in the open position.

Such a failure could be postulated based on mechanical failure, electrical failure, failure of an automatic ADV control system (should one be used) or operator error.

Failure of the ADV in the open position maximizes steam release from the secondary side and therefore also maximizes radiological dose.

The failure of the diesel generator to start was also considered as a candidate for WSF but was rejected since no equipment supplied with power by the diesel (which would then be unavailable if it failed) increases the radiological dose as significantly as failure of the ADV.

Overpressure is not a concern from a loss of the diesel since overpressure criteria have been shown to be met by both SLl and SL2 for complete station blackout (including diesel generator failures),.as part of licensing actions for SLY

NRC Questions Are ADV's in "AUTO" mode or not7 Why2 Tech Specs?

What is "AUTO" mode for ADV's?

What is the setpoint7 Why does the unaffected ADV open at F 08 seconds2 Answer This analysis assumes automatically controlled ADVs, even though no such system is in use at SLl.

ADVs at SLl are only operated manually.

(An automatic ADV system was included in the SLl design but has never been used).

The mode of operation for ADVs is not mentioned in Technical Specifications.

The reason for assuming a

mode which is not used was to provide an analysis which conservatively enveloped both existing plant operation and anti-cipated future changes.

At the time of the LOOP submittal, hardware and procedural changes to make an automatic ADV system at SLl were under serious consideration for the upcoming Cycle 6.

For this analysis, one of SLl's two ADVs was assumed to fail in the open position at t=0. This failure is a postulated initial condition chosen to maximize the potential for fuel damage and radiological release and is not a predicted consequence of the progression of the events during the transient.

This assumption conservatively envelopes the effects of either the

'failure of an automatically operated ADV system, an operator error, or a mechanical failure since the excessive cooldown starts immediately.

(An automatic

system, actuated by an increasing secondary pressure would not function (or fail) until later in the transient.)

The second ADV was allowed to function in accordance with the assumed automatic program, wherein the ADV will open at 970 psi increasing and reseat at 970 psia decreasing.

The Reference (8) analysis predicts the opening of the second ADV at 2.08 seconds.

Assuming automatic control of the second ADV is also conservative.

It provides for greater and earlier cooldown than the Main Steam Safety valves (which have a higher opening setpoint).

NRC Questions The Reference (8) analysis assumes that the ADV block valve is shut at 30 minutes.

Reference (7) assumed 15 minutes.

Why is there a difference2 What happens if the block valve is not closed at 30 minutes2 Answer The assumed operator response time to manually block the failed ADV was conservatively increased from 15 to 30 minutes to demonstrate safe plant operation despite additional delays and also to be consistent with other submittals of the same time period.

FPL considers that 30 minutes is sufficient time for operators to recognize and block a failed ADV.

Consideration of effects of failing to block the ADV within this period are thus unnecessary.

NRC Question What are the assumptions concerning the Auxiliary Feedwater System

?

Reference (7) says manual while Reference (8) says automatic.

Answer For AFW, it was decided that the finalized design for the upgraded "SMART" AFW system would be assumed, even though its installation was not scheduled until Cycle 6.

This was based on analysis results for the sister unit, SL2, which had a

functionally similar "SMART" AFW system.

The SL2 results showed that the radiological doses from events in which an ADV is assumed to fail are maximized with upgraded "SMART" AFW systems.

The designs of the interim AFW system currently installed and the proposed "SMART" system differ significantly.

An Auxiliary Feed Actuation Signal (AFAS) will be generated on low steam generator level.

The interim system delivers full AFW to both steam generators with a 3 minute time delay.

The upgraded "SMART" system has additional logic which senses high differential pressures between either the two steam headers or the two feedline headers, indicative of a steamline or feedline break, respectively.

If the differential exceeds a setpoint, the steam generator on the lower pressure is considered "faulted" and receives no AFW.

The important effect of the new system to both of the Reference (8) analyses (SR and LOOP) is that the failed ADV behaves similar to a steamline break, in that the differential between the two steam header pressures exceeds the setpoint, and AFW to the steam generator (SG) on the affected side is shut off. Subsequent dryout of the steam generator occurs and the dilution of assumed primary to secondary leakage (as accounted for by the iodine partition factor) is lost.

When the partition factor rises from 0.1 to 1.0, the event's predicted offsite doses are maximized.

The setpoints and delays for the upgraded "SMART" AFW system and the associated timing for AFW flows evident in the sequence of events table in Reference (8) were based on the vendor's setpoints projected at the time of the submittal'.

Hardware issues unrelated to the analysis coupled with FPL's desire to review performance of the now operational SL2 "SMART" AFW system have resulted in a postponement of the SLl installation until Cycle 7 (1985),

and a reevaluation of all setpoints and delay times for the system.

F The Reference (8) submittal remains conservative,

however, because simply having the upgraded "SMART" system, with the setpoints

used, results in predictions of steam generator dryout on one side, and the resulting loss of benefit from the partition factor.

The interim system now installed would be predicted to supply AFW to both sides, without shutoff, and without the loss of the partition factor benefit.

A functionally similar system to the one used for this analysis is described in Section 10.4.9 of the St. Lucie Unit, 2 (SL2)

FSAR, Reference (6).

NRC Question The NRC SER for Amendment 48 to the SLl Operating License agreed that an initial SG pressure of 909 psia was conservative for the LOOP because it maximized the initial steam release.

The Reference (8) analysis uses 900 psia.

Is this still conservatives The steam release numbers for the Reference (7) and Reference (8) analyses (shown below) are different.

Explain.

Reference (2)

Reference (3)

Release thru MSSV's 0-2 Hrs (ibm]

163000 16341

-89.9%

Release thru ADV' 0-2 Hrs Llbmg 590000 742712

+25.9%

Total Release 0-2 Hrs Llbm3 753000 759053

+ 0.8%

Total Release in 903000 Cooldown to 325 F [lbmj 913780 Answer For Reference (8), an ADV is assummed to fail open at t=0.

This assumption results in significantly greater steam release in the early seconds of the transient over that which was seen for Reference (7), in which Main Steam Safety Valves (MSSVs) cycled only after steam pressures increased above their opening setpoint and ADV's were opened under operator control much later in the transient.

Despite the greater steam release earlier in the transient, DNB limits are not breached, no fuel damage

occurs, and radiological release remains determined by maximum Technical Specification limits on primary coolant activity.

The 9 psi difference in the steam generator pressure therefore has no significance.

Differences do exist in reported mass integrated

flows, primarily because the assumption of a failed ADV redistributes the flow path of secondary fluid out of the system from the MSSV's to the ADV's and also causes

greater mass outflow in the early seconds.

However, the reported total outflow (from both MSSV's and ADV's combined) for the first two hours and for the entire event (down to 325 F) are only about 1% or less different, which is well within expected agreement for the Reference (7) versus Reference (8) cases.

Both with and without ADV failure, the same energy must be removed from the system for cooldown over the long term.

The slight differences in mass outflows can be considered to be due to the differences in the timing of the mass 'outflows and slight differences in enthalpies at those times.

Since no fuel damage is predicted for either case, Technical Specification activities are assumed and radiological doses are effectively equivalent,.

NRC Question tlhy did CEA worth change from 5.3% 5 k/k in Reference (7) to 5 '% b k/k in Reference (8)7 Answer Cycle 4 parameters were used for Reference (7) vis-a-vis Cycle 5 parameters for Reference (8), since Cycle 4 was over and Cycle 5 was already in progress when the second submittal was made.

NRC Question Why was a Doppler Coefficient multiplier of 0.85 used in Reference (8) vis-a-vis the 1.15 multiplier used for Reference (7)?

Answer The value of the Doppler Coefficient was inadvertantly changed between the two analyses.

However the vendor has indicated that for the LOOP event for SLl, the transient is not sensitive to this value.

Results of a reanalysis would not be expected to change by more than at most a few percent.

The ADV failure assumption has a far more significant effect than the Doppler Multiplier.

The change is not considered significant.

Justification for this assessment may be seen by considering that the Doppler coefficient is only important if its effects (1) adversely affect the functioning of the RPS, or (2) significantly degrade margins to the SAFDL's.

For this event, reactor trip occurs as the result of the low flow trip, sensed by comparing loop flows measured by Differential Pressure Cells to the minimum setpoint.

The RPS function in this event is therefore clearly independent of Doppler.

The Doppler coefficient will affect the magnitude of the slight power excursion after trip, but it can be seen from both Reference (7) and Reference (8) that this excursion is small.

No SAFDL is breached.

The multiplier used is therefore of no significant consequence.

NRC Question Why did the system f(ow change to 138.3 x 10 ibm/hr 6

vis-a-vis 133.8 x 10 ibm/hr for the LOOP event Answer This is a typographical errgr.

The correct value for the LOOP analysis is 133.8 x 10 ibm/hr.

A new Table 1 for this event is provided on the following page and should replace the Table 1 for the LOOP from Reference (8).

The value of 138.3 x 10 ibm/hr is correct for the SR analysis

since, consistent with the Reference (7)

SR analysis, Statistical Combination of Uncertainty (SCU) methods were used.

TABLE 1 LOSS OF AC AND STUCK OPEN ADV KEY PARAMETERS ASSUMED IN STEAM RELEASE CALCULATIONS Parameter Initial Core Power Level Units Mwt

~Cele 5

2754 Initial Coolant Inlet Temperature Initial Core Flow Rate Initial Reactor Coolant System Pressure Initial Steam Generator Pressure Initial Steam Generator Level 10 ibm/hr psia psia 551 133.8 2300 900 36.2 above tube sheet Low Flow Analysis Trip Setpoint of initial flow 93.0 Moderator Temperature Coefficient Doppler Coefficient Multiplier CEA Worth on Trip Reactor Regulating System Steam Bypass System Auxiliary Feedwater System x10 5p/

F

+0.5 0'5

-5.6 Manual Mode Inoperative Automatic

NRC Question As shown below, differences in both the values and the timing of peak system pressures were noted between the Reference (7) and (8) analyses.

Can this be explained

'?

LOOP L Ref (7) 3 w/o ADV faiure Peak S/G Pres.

1034 PSIA 9 6.4 sec Peak RCS Pres

~

2534 PSIA 8 7.4 sec LOOP L Ref(8) w/ ADV failure 2457 PSIA 8 3.8 sec 1093.5 PSIA 9 13.38 sec Answer Because several assumptions were changed between the two submittals, (e.g.

ADV failure, physics parameters, etc.) it is difficult to attribute the above effects to a specific cause.

However, note that the failure of an ADV at t=0 will result in a greater energy removal from the primary system earlier in the transient which would tend to cause both the lesser peak primary pressure and the occurence of the peak earlier in the transient.

The behavior of the steam generator pressure can be explained in part by noting that the peak pressure occurs on the unaffected side, i.e. the side with the normal (closed)

ADV.

The steam flow from the affected side (with the open ADV) from the Reference (8) submittal would tend to create larger pressure and temperature imbalances between the two steam generators.

The times for peaks would also be affected.

Changes in physics inputs (Cycle 5 versus Cycle 4) would also result in changes to both peak values and timing.

In both submittals, the values were derived from the vendor's NRC-approved CESEC code using models which have been previously used and reviewed.

Additionally, since overpressure limits are not breached, since DNB limits are not

breached, and since the major parameter of interest then becomes the offsite dose, dictated primarily by the failed ADV assumption, the behavior of primary and secondary pressure peaks is not of primary interest for this transient.

NRC Question Reference (8) reports the actuation of Safety Injection.

Reference (7) does not.

What is happening to the primary at this time Answer The Reference (8) submittal reports Safety Injection actuation at its correct trip point as a normal consequence of the cooldown process.

Reference (7) omitted mention of Safety Injection, but identical behavior would be expected.

In both cases, Safety Injection occurs late in the transient and has no effect on predicted accident consequences.

NRC Questions What are the references for Reference (8) 2 Answer The list of references for Reference (8) was inadvertently omitted from the submittal.

The list of references from letter, following as Attachment (3) may be used.

The numbering convention is ident, ical.

Mention is made in the Reference (8) submittal of present References (l) through (5).

NRC has not been forwarded a copy of Reference (l), a draft Reload Safety Evaluation (RSE) for Cycle 5 operation supplied to FPL by the fuel vendor, Combustion Engineering (CE).

In lieu of this document, FPL had submitted Reference (ll).

The important point is that the Reference (8) analysis was performed using parameters of the Cycle 5 reload, and not those of Cycle 4.

Table l of Reference (8) correctly summarizes the parameters used.

References

l. Letter R.

R. Mills (CE) to C.

G

. O'Farrill (FPL),

10/1/81, F-CE-7567 2.

CENPD-107, "CESEC Digital Simulation of a C-E Nuclear Steam Supply Syst.em", April 1974 3.

CENPD-135-P, "STRIKIN II A Cylindrical Geometry Fuel Rod Heat Transfer Program",

August 1974 4 ~ CEN-123(F)-P, "Statistical Combination of Uncertainties Methodology", February 1980 5.

CENPD-183, "C-E Methods for Loss of Flow Analysis", July 1975

6. St. Lucie Unit 2 Final Safety Analysis Report, Section 10.4 '

("SMART" AFW system description)

7. Letter R.

E. Uhrig to D.

G. Eisenhut, 11/14/80, L-80-381 (Original stretch power submittal)

8. Letter R.

E. Uhrig to D.

G. Eisenhut, 8/31/82, L-82-381 (Current submittal under review)

9. Letter R.

E. Uhrig to D. G. Eisenhut, 7/23/81, L-81-306 (Post-TMI MSLB analysis)

10. Letter R.

E. Uhrig to R. A. Clark, 9/4/81, L-81-388 (Post-TMI Excess Load and SGTR) ll. Letter R.

E. Uhrig to R. A. Clark, 10/7/81, L-81-439 (Applicability of prior analyses to Cycle 5)

12. Letter J.

R. Miller to R.

E. Uhrig, 10/12/83

• References (1) through (5) are also the missing references for Reference (8).

JV