ML20204G837

Safety Evaluation Accepting Topical Rept 151, Haddem Neck Plant Non-LOCA Transient Analysis, Except for Issue of Feedwater Event

ML20204G837
Person / Time
Site:	Haddam Neck File:Connecticut Yankee Atomic Power Co icon.png
Issue date:	10/18/1988
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20204G826	List: ML20204G824 ML20204G837 ML20204G864
References
NUDOCS 8810240245
Download: ML20204G837 (5)

Text

V

. i.

1. 1. p aeog'o, ,

UNITED STATES

[ g NUCLEAR REGULATORY COMMISSION s p WASMG TON, D. C. 20665

%,...../

ENCLOSURE 1 SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION NORTHEAST UTILITIES TOPICAL REPORT 151 FOR NON LOCA TRANSIENT ANALYSIS CONNECTICUT YANKEE ATOMIC POWER PLANT DOCKET NO. 50-213

1.0 INTRODUCTION

By letter dated June 30, 1986, ConnecticutYankeeAtomicPowerCompany(CYAPCO) submitted NUSCO Topical Report 151, entitled "Haddam Neck Plant Non-LOCA Transient Analysis." This was prepared to demonstrate the licensee's ability to perform in-house safety analyses and to provide an updated set of design basis transients. The report contains a reanalysis of all 12 design basis-events described in chapter 10 of the Haddam Neck Facility Description and Safety Analysis (FDSA) report and an additional analysis (RCP rotor seizure / shaft break) as a result of the Systematic Evaluation Program (SEP) review. This report constitutes a complete and consistent set of design basis analyses.

The analysis for the 13 events used the RETRAN02/M00003 and VIPRE01 computer codes. The staff has evaluated the licensee's qualifications to use the RETRAN code, per Generic Letter 83-11, in reference 1. The VIPRE methodology was previously reviewed and accepted by the NRC staff (Reference 2).

To expedite support of the Cycle 15 reload, the NRC staff previously reviewed five of these events separately. They are:

8810240245 63101G PDR ADOCK 05000213 P PDC

<c

< t 4

g Uncontrolled rod withdrawal i

Boron dilution '

RCCA ejection Droppe.' rod Loss of flow l

The five events were reviewed by the NRC staff and were found to have been i reanalyzed by the licensee in a conservative manner. They were found

~;

acceptableforsupportofreloadforCycle15 operation (Reference 3). The '

., staff further noted that the steamline break accident becomes the limiting .

case for shutdown margin af ter approximately mid-cycle. Thus, approval of a revised main steamline break analysis would be required for plant operation '

beyond the mid-point of Cycle 15. Of the original 13 events, the following '

eight events, addressed in the subject topical report, remained for staff  ;

review : i l

Isolated loop startup (SRP15.4.4)

Excess feedwater (SRP15.1.2)  !

! Excessive load increase (SRP15.1.3)  !

j Steanline rupture (SRP15.1.5)

Stean generator tube rupture (SRP15.6.3) l Loss of load (SRP15.2.1)

{

loss of feedwater flow (SRP15.2.7) -

I Reactor coolant pump rotor seizure / shaft break (SRP15.3.3/15.3.3)

InternationalTechnicalServices,Inc.(ITS),consultantstotheNRCassisted '

j the NRC staff in its review of the topical report and CVAPCO's response to the staff's request for additional information. The associated transient and '

accident analyses were reviewed in accordance with Section 15 of the Standard Review Plan (NUREG-0800).  !

l I

i I

i *

}

i

. a 3

2.0 EVALUATION The licensee uset methods and transient assumptions that conservatively bounded the consequence', of events.

, t The minimum departure from nucleate boiling ratio (MDNBR) was calculated using 4 the W-3L or the MacBeth correlation in the VIPRE01 code. A conservative value of 1.3 for the MDNBR was used in all cases as the threshold for fuel failure, I

In addition, the review examined the application of the RETRAN code for each  :

! event.

I

, It was assumed that the reactor protection instrumentation response times i in the topical report are correct. The Instrumentation and Control Systems j Branch (ICSB) is reviewing the upgrade of the Haddam Neck reactor protection systems. If ICSB finds any significant change in response time, the NRC staff l

would require a reanalysis of the affected event (s).

The results of the review by ITS of the isolated loop startup, excess i feedwater, steamline rupture, steam generator tube rupture, excessive load i j increase, loss of load, loss of feedwater flow, and RCP rotor seizure /shaf t I break are given in the attached technical evaluation report (TER). On the  ;

] basis of their review, ITS concluded that l i

(1) Pressure in the reactor coolant and main steam

, systens does not exceed the acceptance criterion of 110 j percent of the design values I i (2) Fuel cladding integrity is maintained by ensuring that '

the minimum DNBR remains above the 95/95 DNBR linit ,

! discussed in SRP Section 4.4  ;

l (3) The time allowed for operator action during the excessive feedwater event may not be adequate

)

1 l

e a

. s 4

4 1

l The NRC staff is in agreement with the ITS conclusions relating to the f i

thermal-hydraulic analyses. The radiological consequences of the steam

(

! generator tube rupture event are under review by the Radiation Protection ,

. Branch. .

With the exception of the excessive feedwaiir event, the staff finds that l l l the results of the. analyses performed meet the acceptance criteria for each l l event accordino to the Standard Review Plan. The operator action time for {

, the excessive feedwater event, however, does meet the original design basis  !

and is therefore acceptable. I t

l The short operator response time for the excessive feedwater event was f

] identified in the SEP review, Topic XV-2. The topical reanalysis has not . [

4 <

j resolved this issue. . l l

1

3.0 CONCLUSION

The NRC staff concludes that the thermal hydraulic analyses of the events

are acceptable since the analyses indicates that the system. pressures do not  ;

exceed the acceptance criterion of 110 percent of the design criteria and the l I

, MNDBR does not fall below the minimum limit of 1.3.

i The staff also finds that the analysis of the exceesive feedwater event does f

! not satisify the operator response time of the SRP although it meets the

! original design basis. The licensee has proposed to review this issue by e 1  :

initiating a new Integrated Safety Assessment Program (ISAP) topic. The staff l

, agrees that the ISAP approach is reasonable to resolve this item.

l J

4.0 REFERENCES

i 1

l 1. Letter from Alan B. Wang (NRC) to Edward J. Mroczka (CYAPCO), "Safety j Evaluation for Northeast Utilities Topical Report 140-1,'Nusco Thermal ,

l Hydraulic Model Qualification, Volume 1 (Retran)'," July 26, 1988.

i

. s

's t

2. Letter from Francis M. Akstulewicz (NRC) to to John F. Opeka (CYAPCO),

" 'Nusco Thermal Hydraulic Model Qualification, Volume !! (Vipre),'  ;

Topical Report Nusco 140-2." October 16, 1986.

3. Letter from Alan B. Wang (NRC) to Edward J. Mroczka (CYAPCO), "Revised l '

j Safety Evaluation and Technical Specifications Supporting amendment No. ,

97," April 28, 1988. i

[

I

]

i ,

1 l I i

i I I, '

i j l i

i ,.

l i  !

i f

I

ITS/NRC/88 4 August 1988 Technical Evalusij,.qq, Q.i NUSCo Tooical Reoort 151 "Haddam Neck Non-LOCA Transient Analysis"

1.0 INTRODUCTION

NUSCo's objective in submitting of NUSCo 151 (1) was w <ie a- .r.

• e ability tu perform in-house safety analyses, using the RCRA, a t. ethodology, constituting a complete and consistent set of design btsis analyses. Hence, NUSCo reanalyzed all twelve Chapter 10 design basis events, and the ,

l additional analyses requirad as a result of the SEP (2) review, using RETRAN02/M0000) and VIPRE 01. These analyses were contained in NUSCo 151. l j However, of 13 transient analyses submitted in NUSCO 151, sev(n have since been reanalyzed (3). Six transients (steam line break, lose of flow, dropped rod, RCCA ejection, boron dilution and ur, controlled rod withdrawal) were  !

reanalyzed due to changes in the physics parameters and Technical Spacification (Tech Spec) changes associated with the Cycle 15 reload 4

(4,5,6,7,8), and the loss of normal feedwater transient was resubmitted because of a revised minimum expected auxiliary feedwater flowrate i

, (9,10,11,12).

For the reload analyses, RETRAN02/M00003, VIPRE 01 and a NUSCo adaptation of [

Westinghouse physics methodologies were employed. Although RETRAN02/M00003 l has not yet been approved by the NRC, it is reportedly a corrected version of i RETRAN02/M00002 which was reviewed and approved by the NRC, subject to

certain restrictions (13). Although RETRAN02/M00003 has not been formally l reviewed by the NRC, we find that there are reasonable assurances that  !

analysis using this version of RETRAN is acceptable as .ed by NUSCo for I

these transients, unless future review of that code determines to the  !

l 1 fjQtOD

.r 4

. contrary. The VIPRE-01 methodology has been previously approved by the NRC l (14). The NUSCo physics methodology, which has its base in the Westinghouse  !

methodology, was used to compute physics parameters for the Cycle 15 reload '

core. This methodology has also been approved by the NRC (15).

2.0 TECHNICAL EVALUATION

' l 2.1 Transient Analysis Assumotions and Accentance Criteria l I

The accident analyses for Haddam Neck were reviewed in accordance with j

! Standard Review Plan (SRP) Chapter 15 (NUREG 0800) (16] to assure conformity  ;

with the acceptanct criteria, except as noted for each of the sections. In addition, this review examined the acceptability of t',e use of the RETRAN l 3

code for each of the analysis submitted.  :

)

The NUSCo model qualification using RETRAN for Haddam Neck, submitted as  !

j NUSCo 1401 (17) and supplementary documents, has been previously reviewed ,

l and found to be generally acceptable (18), although only a selected set of f transients were analyzed in that report. The transients analyzed therein,  !

! however, are similar to the others which must be analyzed for the purposes on ,

j a complete Chapter 10 but were not analyzed in NUSCo 140-1. We found that  ;

j the models developed during the qualification effort were acceptable in the j

] context of demonstration that such models were reasonable representations of

the plant. In this review, the models were further examined to assure

acceptability in the SRP context; i.e., for adequate conservatism. .

I

i l The licensee investigated a broad spectrum of related events to determine the l l bounding case for each event, including the worst single active failure. l

) Table 1 lists those which were analyzed in NUSCo 151 vs Final Design Safety i j Analysis (FDSA) Chapter 10 transients (19) vs. SRP Chapter 15 transients. I Sensitivity studies were perfomed by NUSCo to identify parameters for j

initial conditions, and appropriate credit for systems an' their performance

l during the limiting events in terms of protection of various barriers. l Unless otherwise noted, full power (HFP) level is assumed to be 102% of

! 2 l i l l l i ,

i <

i 1 ,

o

. j design power level at 1861.5 MWt in 4-loop operation and 67% at 1222.75 MWt l in 3-loop operation. These values include a 2% allowance to account for powermeasurementuncertainties.(1)

Other system parameters used in the analyses also include conservatisms to account for uncertaintles; for example, scram characteristics used in the analysis assumed the most reactive rod stuck out of the core, and the reactor protection system septpoints and delay times used account for margin of error.

Transients analyzed are protected by the following reactor trips: core power; high pressure; low pressure; pressurizer level, high; pressurizer level, low; low coolant flow; high steam flow; steam / feed flow mismatch coincident with low steam generator (SG) level (1). During the current outage. Haddam Neck's

• reactor protection and control system was significantly changed to reflect state-of the art technology in instrumentation. Those instrumentations changed which may impact the safety analysis are; pressurizer oressure, pressurizer level, reactor coolant system (RCS) delta temperature and RCS

average temperature, RCS wide range temperature, high pressure steam dump and charging flow control (20). It was assumed for the purposes of this eview l that the delay times contained in Table 5 of Reference S which are those used in the NUSCo 151 analyses and reanalyses, were conservatively determined by the licensee.

The core burnup was selected to yield the most limiting combination of i i moderator temperature coefficient and Doppler coefficient in light of the j revised shut down margins as presented on Tables 1 through 4 of Reference 8 and Table 1 of Reference 6. Power profile and radial power distributions I were selected to yield conservative results and assumed to remain constant  !

throughout the transient.

For

• transient and accidents, the licensee utilized a mothed that conservatively bounds the consequences of the event by accounting for uncertainties directly in the calculations by way of introducing margins to

the safety setpoints. l

, 3 i

l The ultimate acceptance criterion is that there be acceptable minimum departure from nucleate boiling ratio (MDNBR). DNBRs were calculated using the W 3L correlation in the VIPRE-01 code and the MacBeth correlation for steam line break analysis (when the thermodynamic conditions are outside the range of the W 3L correlation), with a minimum DNBR of 1.3 used in all cases as the threshold for fuel failure. (the NRC review of the W-3L DNBR limit is attached as Attachment 1 of this report.) The 95/95 DNBR limit was determir.ed, based upon a set of experimental data, to be 1.16 (10). Thus use of 1.3 was conservative. ,

i

. 2.2 Non-LOCA Tr'ansient Analyses The number in parentheses following each NUSCo report section number listed in the headings indicates the corresponding section in the SRP Chapter 15.

Unless otherwise noted the nodalization used in the RETRAN analysis was a four-looprepresentation(6). l l

l 2.2.1 (15.1) Increase in Heat Removal by the Secondary System i

Three events in this category were analyzed by the licensee; Excess j Feedwater. Excessive Load increase and Steam Line Break (SLB). <

J ,

2.2.1.1 (15.1.2) Excessive Feedwater l

4 An initiator of this event is a feedwater control system malfunction. The consequences which must be mitigated are fuel damage due to increased power and structural problems associated with steam generator overfill from excess feedwater.

1 Two transient analyses were performed: a failure at the full open position of

one feedwater regulating valve (1) at HFP for both 4-loop and 3-loop operation and (2) at 40 % power for potential steam generator (SG) overfill.

The 40% power is the minimum power level at which two feedwater pumps are operating, which causes the failure of the feedwater regulating valve to the 4

l 1

l l

- - - . - - - - - - r --- - - -

y ,r - - - - r

l I

, , full open position to result in the greatest stearq/ feed mismatch and hence l the fastest rise in SG 1evel. l l

Among the conservative assumptions used as initi',1 conditions were; the most negative moderator temperature coefficient (STC) and the least negative Doppler coefficient were used to minimize power reduction due to fuel heatup.

In addition, to maximize the cool down, a lower value for feedwater (FW) temperature was used. No credit was taken for the heat stored in the RCS and steam generator metal mass which would hau reduced the resulting plant  ;

cooldown. It was further assumed by NUSCo that the core inlet temperature ,

was initially at its m:ximum and the pressurizer pressure was at its minimum.

The normal reactor control system and engineered safety systems are not required to function. The aggregate effect of these assumptions was to constitute the worst cases in this category.

The transient was terminated by tripping of the main feedwater pumps (FWPs) and the reactor by the operator following the high-high SG water level alarm from the faulted loop.

In 4-loop operation from 40% power, the high-high SG water level alarm was actuated 23 seconds into the transient and it was assumed that the operator took 37 seconds to trip the main FWPs to prevent the overfill. For 3 loop operation, the alarm was actuated at 25 seconds into the event and 45 seconds later the operator was assumed to trip the FWPs. The times assumed for

operator actions were obtained through a series of computations by which j NUSCo determined the times by which the operators must act to avoid steam generator overfill, and differ due to the difference in initial SG water inventories. Failure of the operators to act by such times would, therefore, result in steam generator overfilling. Conservative estimates of mass inventories were different for 4-loop and 3 loop operation (57,000 lbm vs 53,000lbm). The thermal hydraulic analysis review assumed that operator action times were as stated in the NUSCo analysis.

Although the feedwater flow to one SG was increased to 150% of its full load by the full opening of one feedwater regulating valve during 4-loop HFP 5

operation, and to 183 % for 3-loop operation, the computed decrease in RCS cold leg temperature of the affected loop was less than 5' in each case.

Thus, the NUSCo results support their conclusion that a negligible change in DNBR is indicated. The results further support NUSCo's conclusions that they have examined the worst case events in this category and that this transient

, is bounded by both the excessive load increase and the main SLB events.

Provided that the assumed operator action times are acceptable from a human factor perspective, we have adequate assurances that the system safety limits are not challenged by events in this category.

2.2.1.2 (15.1.3) Excessive load increase The limiting excess load increase transient is initiated by a sudden failure to full open position of all 10 steam bypass to condenser valves.

Conservative assumptions made for initial conditions include maximum power l

1evel at which the opening of all bypass valves will not result in an  ;

imediate reactor trip from the excessive steam flow for high power signals.

83% full power (FP) for 4-loop and 47% FP for 3-loop operation were assumed.

The most negative MTC and the least negative Deppler coefficient were assumed to maximize the reactivity insertion. The core inlet temperature was assumed at its maximum value and the primary pressure was assumed at its minimum l value. It is further assumed that the reactor was in manual control and no credit was taken for pressurizcr pressure control. The FW flow was assumed

to increase to match the increase steam flow during the event. Nu other engineered safety systems were required to function. These assumptions constitute the worst case excess load transient. '

The transient was terminated when the power to flow mismatch disappeared by equilibration of the system, which occurred at approximately 80 seconds for l j 4-loop and 120 seconds for 3-loop operation. l The 4 loop analysis results indicate an increase of roughly 45% in power and

! a decrease of 3% in core inlet temperature during the first 40 seconds. The MDNB of 1.4 was computed to occur at roughly 35 seconds into the event. For 6

-_ _ _ - - - - , a_-,. , , . , . - - _ _ , -,a - - - - - _ --

l

. . , 3-loop operation, the core power increased 76% and the core inlet temperature decrease about 4% resulting in MONBR of 1.4 at 40 seconds into the transient.

Both of these analyses indicate that the assumptions made for this transient are conservative and the MDNBR is above the minimum allowable value. We therefore have reasonable assurances that the system safety limits are not challenged by this event.

i These results also support NUSCo position that this transient bounds the excessive feedwater event.

2.2.1.3 (15.1.5) Steam Line Break The original analysis contained in NUSCo 151 was replaced with reanalysis contained in a submittal dated November 19, 1987 (6).

, This transient was analyzed at HFP for both 4-loop and 3-loop operation and at the hot zero power (HZP) level which was assumed to be initially at 1%

power. The HZP cases with offsite power available were determined to t,e most limiting from the point of view of the degree of return to power since the primary side had the least stored energy and the SG secondary side contained the maximum inventory, nximizing the potential cooldwn and therefore maximizing the return ,to power. However, for DNBR and fuel centerline  !

, evaluations, the HFP case led to a greater cha11ence to fuel integrity because it had a higher power level during the transient, j An extensive series of sensitivity and parametric analysis (17) were performed to assure that the NUSCo plant nodalization and computer code model  !

] selection would provide reasonably accurate and conservative results. These l

! studies included sensitivity studies of the nodalization of the steam .

generator (SG), surge line and reactor vessel (including the eventual choice of a split core model in the vessel).

The SLB aanalysis used a split core nodalization and computed moderator feedback assuming the entire core fluid to be at the temperature of the broken side coli leg. Thus, the core fluid inlet temperature, for the l \

7 l

, purposes of computation of moderator reactivity feedback, was conservatively assumed to be the temperature of the cold leg fluid on the affected side.

The unique feature of this analysis for Haddam Neck is the assumption, based upon the plant's operation, that 2 of the 4 reactor coolant pumps (RCPs) trip on reactor trip. This causes the flow through the core to reduce to approximately 1/2 of the full power flow, resulting in the coolant exiting the core at a higher average temperature, causing less moderator overcooling than in a plant in which all 4 RCPs continue to run. Thus, the actual moderator feedback is considerably smaller than would occur in a plant in i which all four RCPs continu to run.

i Parametric studies were also performed to demonstrate that the particular choice of splitting the reactor core 1/4 to 3/4 did not produce substantially different results from splitting the core 50/50 (6,7). Based on the i foregoing, we have reasonable assurances that the SLB reanalysis reactor vessel nodalization and moderator feedback computation is conservative.

The computation of upper head temperature is important in a SLB transient because after the pressurizer empties the upper head fluid saturates and the i upper head acts like a pressurizer, controlling RCS pressure and thereby l impacting both safety injection and the DNBR calculation. The upper head was modeled as a hT ot plant and the nodalization was carefully adjusted to obtain circulttien flows which are in rearonable agreement with flows computed based l upon experiments, and therefore we have reasonable assurances that the upper head fluid temperature is reasonably computed. Modeling the upper head as a j Th ot plant maximizes the pressure and thereby delays the safety injection

, beyond the time of the minimum DNBR, and results in boron reaching the core only after the time of minimum DNBR, Therefore the particular modeling of the upper head flow is not of strong importance. We have reasonable assurances that the modeling of the plant as a T hat upper head plant produces conservative results.

Through use of the control system in the code, the boron transport was 1 modeled in a best estimate approach. Although a best estimate approach would ordinarily be unacceptable for licensing computations because it does not 1

8

4 4

.. , incorporate adequately conservative assumptions, the particular modeling of

! boron transport for this transient does not impact th$ conservative estimate l 4

of the DNB or the return to power, because the safety' injection was delayed i until after the minimum DNBR. Thus we have reasonable assurances that the  !

nonconservative boron transport modeling does not alter the otherwise i conservative results of the SLB analysis. l

} l The pressurizer was modeled as a non equilibrium pressurizer coupled by a 4 single node surge line to the hot leg. This model was used to compute a l I

plant transient during which the pressurizer outsurged and found to under-l predict the RCS pressure, which tends to cause a worse DNBR and is therefore I conservative. i The feedwater temperature was assumed to be at its minimum for conservatism.  ;

In addition, a series of other conservative assumptions used in the analysis thclude; and of life shutdown margins, no initial boron, most reactive rod stuck out, maximum SG initial inventory, perfect SG steam separation.

conservative initial powers, most negative moderator temperature coefficient, f break outside containment to delay safety injection actuation signal, l

! diversion of all auxiliary feedwater to the affected SG and continued reactor L j coolant pump operation. The results of these analyses indicated that no fuel j rod exceeded the DNB limit. l L

y a

Based on the forgoing, we have reasonable assurances that the DNB limit n uld r j not exceeded in the worst case SLB accident during Cycle 15. Reanalysis of ,

j the worst case SLB transients were found to be conservative and to yield i

! reasonable assurances that the safety margins would not be violated. The -

2 assumptions and conditions used in the analyses were found to be consistant L j with the proposed Tech Specs. Although NUSCo used a best estimate boron

) transport methodology with the RETRAN computer code, this was found to be

] acceptable for this SLB analysis since it does not impact the DNBR. However, i

if, in the future, boron transport modeling is a significant aspect of the I safety margin, this methodology will be reviewed for acceptable

conservatisms.

j S i

J i

_ , _ _ _ - . . _ . _ _ . _ . , _ _ _ _ __ -.,__ _ . _ _ - - _ ~ _ _ _ . _ _ _ _ , , _ _ _ _ _ _ _ . _ _ _ _ . . - _ _ _ , . . . , . . _

2.2.2 (15.2) Decrease in Heat Removal by the Secondary System Two analyses were performed under this category and are: Loss of Load and Loss of Normal Feedwater Flow.

2.2.2.1 (15.2.1) Loss of Load ,

A complete loss of steam load was evaluated from HFP 4-loop and 3 loop  ;

operation. Closure of both turbine stop valves is assumed since this will result in a more, severe transient than will closure of the turbine control  !

valves to terminate steam flow to the turbine. Reactor trip on turbine trip is not credited in order to show the adequacy of the pressure rclieving devices and to demonstrate margin to DNB. This assumption is conservative l since it will delay reactor trip until other conditions result in a trip. ,

Similarly assumptions made in the analysis are designed to (1) minimize DNB,

. (2) minimize decay heat removal, (3) maximize peak pressure, and/or da'ay j reactor trip. In order to minimize margin to DNB at the initiation of the '

transient for conservatism, maximum values for initial reactor power and RCS temperature and the minimum values for initial RCS pressure were assumed (to the extent consistent with steady state full power operation, including allowances for calibration and instrument errors). The least negative Doppler and most positive moderator temperature reactivity coefficients were l

used, minimizing the pre-trip negative reactivity insertion and maximizing  !

the core power and RCS pressure excursion prior to reactor trip. No credit l was taken for automatic reactor control. To minimize and delay heat removal and maximize peak pressure, no credit was taken for the operation of the l atmospheric dump valve or manual initiation of auxiliary feedwater flow or.

low low steam generator water level. Similarly primary and secondary safety valves were modeled with minimum flow capacities and with the opening setpoints increased from the nominal values by the setpoint uncertainties.

The aggregate effect of these assumptions is that NUSCo has considered the worst case transients in this category.

l The results indicate that the PONBRs occurred at about 13 second into the i

10 i

. . , transient for 4-loop operation and at roughly 16.5 second for 3 loop operattor. and are 1.8 and 1.9 respectively. We therefore have reasonable assurances that the system safety limits would not be exceeded by events in this category.

+

2.2.2.2 (15.2.7) Loss of Normal Feedwater Flow This Lnalysis was replaced with reanalysis submitted on March 10, 1987 (5),

due to the use of non conservatively high auxiliary feedwater (AFW) flow assumed in the analysis submitted in NUSCo-151. Lower assumed AFW flow (8) was used in the reanalysis, i

In order to minimize the heat removing capability, offsite AC power was assumed available since operation with fewer SGs available is pos'sible if offsite power is available, thus reducing the available secondary side originated heat removal after the loss of feedwater; 2 SGs are available and ,

2 down in 4-loop operation and 1 SG available and 2 down for 3-loop ,

operation, e en compared with all SGs available for natural circulation if a loss of offsite power is r.sumed. Hence the limiting loss of normal i feedwater event is initiated with offsite AC power available and with the limiting single failure. It is assumed that in 3-loop operation, only one of j the RCPs, on loop four, is running. The analyses did not credit the '

pressurizer PORVs. l After performing sensitivity studies, the limiting single failure in 4-loop opero. ion was determined to be failure of one of the two steam driven AFW pumps to start using the criterion to maximize the pressurizer level. For 3-laop operation the limiting failure was found to be failure to open of the AFW injection valve to the SG on the same loop as the running RCP, These assumptions constitute the worst case transients in this category.

The analysis of the 4-loop operation case indicates that the RCPs in loops one and three trip roughly 1 minute after the generator trips, which results in reverse flow in those two loops. As the two remaining SGs remove decay heat, the SG pressure reached the safety valve set point and those SGs began

! 11

}

to lose inventory. Two minutes after the SG Iow level AFW setpoint was reached, at least one AFW pump was started, reducing the rate of water level decrease. The AFW pump capacity is su:h that the water level in the SG is maintained above the lowest level at which sufficient heat transfer area is available to remove heat and to prevent water relief from the RCS relief or safety valves.

During this event the peak pressurizer pressure was computed to be 2527 psia (and, except for the peak, maintained at the safety setpoint of 2525 psia),

which is below the limit of 110% of the design pressure or 2750 psia. Since the pressurizer did not become water solid, it was assumed that no water was discharged from the valves.

The FJNBR for this case was computed to be 1.75 at 10 seconds.

In the 3 loop operation analysis, since the most of the decay heat load was placed on the single SG on the loop with the running RCP, the SG rapidly lost inventory through the safety valves and dried out at roughly 1200 seconds into the event. Loss of inventory for the other 2 SGs from the safety valves was much more gradual. Since the automatic actuation of AFW pump used the two out of three logic, automatic actuation could not occur; thus the ,

operator was assumed to actuate the pump at 10 minutes into the event.

Because the single failure assumption for this analysis was to prevent the AFW flow into the active SG, once the peak pressurizer pressure reached 2528 Lsia (slightly above the safety valve setpoint) it remained thure for roughly l 1800 seconds. The evert was eventually turned around when the decay heat generation was finally compensated for by the heat removal by the AFW flow.

The MDNBR was computed to be 2.52 and occurred at 11 seconds.

In both cases (1) the peak pressurizer pressure reached the safety valve setpoint of 2525 psia, including 1% for uncertainty, and lifted the first stage of safety valves, and (2) the MDNBR was well above the established minimum acceptable level. Therefore we have reasonable assurances that 1he j system safety liaits are not challenged by events in this category.

12 l

l l

l

2.2.3 (15.3) Decrease in Reactor Coolant System Flow Rate Three analyses in this category were submitted by NUSCo; Loss of Forced Reactor Coolant Flow (L0r), Reactor Coolant Pump Rotor Seizure and Shaft Break.

2.2.3.1 (15.3.1) Loss of Forced Reactor Coolant Flow The analysis contained in NUSCo 151 was replaced with reanalysis submitted with Tech Spec change request [3).

The licensee reanalyzed this event to make certain that the previous Tech Spec limit for low RCS flow rate for 4 loop operation would be met, since the prior analysis indicated very little available margin and some degree of steam generator tube plugging was anticipated. Because of the potential tube i plugging, the subn'ltted LOF analysis assumed a lower flow rate. In addition, the reanalysis included a change of the 3-loop low flow trip setpoint in the governing Tech Spec when going from 4-loop operation to three operation.

Flow rate requirements were based upon a steam generator plugging level i consistent with 500 equivalent plugged tubes par steam generator', and an i evaluation had been performed with the core physics by-pass flow fraction ,

reduced to 4.5% from 9% which indicated that although the reactor vessel flow rate decreased due to the plugging, the core flow increased due to the reduced bypass flow. The net effect of these was that core flow remained the 1 same when the minimum RCS flow rate Tech Spec was reduced to 246,000 gpm (4-l loop) and 194,000 gpm (3-loop) and that the core flow rate used in the VIPRE calculations was the same as that used in the RETRAN calculations. Since i core flow remained the same, the transient was not reanalyzed. Core flow was l compared to the FDSA computation for a 4 RCP coastdown (17) in which l

excellent agreement was obtained, thus lending credibility to the choice of I those parameters and to the overall core flows computed by the RETRAN model for this transient. Although no experimental data were presented to verify the flow coastdown curves, for the purpose of this transient, good comparison i 13

to FDSA gives adequate assurances of accuracy of the RETRAN model to permit  !

its use in this event. l The worst case LOF event analyzed is a complete loss of flow from full power. [

which results in the most severe power to flow ratio and therefore the lowest l DNB. LOF sensitivity studies (17) were presented to show the impact of l variation of RCP inertia, junction inertia, rod insertion time, delay of scram signal and reduction of RCS flow loss coefficients. It was concluded ,

that variation of these parameters had a minimal impact upon DNBR calculated [

using the W 3L correlation, and therefore that normal values could be used. l t

From the perspective of the acceptability of the RETRAN computer model of the plant for this transient, acceptability depends only upon the accuracy of the f RETRAN computation of core flow inlet temperature and flow and primary l pressure during the short period of time out through the minimus DNBR, anc f the accuracy of the VIPRE 01 computation using them. We have adequate  !

assurance of the accuracy of the RETRAN computation of those parameters for j short transient of this type. '

i The transient was conservatively assumed to commence from HFP for both of the (

4-loop and 3 loop transients, and the RCP inertia was assumed to be reduced I by 10% to cause the flow to reduce more rapidly. In addition, minimum initial RCS pressure and flow, and maximum core inlet temperature were  ;

assumed in the analysis. Each of these assumptions is conservative, tending i to lessen the computed minimum DNBR. In addition, the computed RCS pressure rise was minimized by assuming minimum initial pressurizer level, maximum initial SG 1evel and maximum turbine stop valve closure ties, pressurizer ,

heaters off and maximum pressurizer spray flow, and charging isolated with f letdown flow available, Finally, reactivity insertion was maximized through  !

the use of the least negative Doppler coefficient and the most positive MTC

]

and reactor trip computation included instrument response and delay times, i The aggregate effect of these assumptions is that NUSCo has conservatively )

analyzed the worst case transient for this category.

The minimum DNBR was computed to be approximately 1.4 in the worst case LOF 14

transient. We therefore have reasonable assurances that the system safety i limits are not exceeded by transients in this category, f

. 2.2.3.2 (15.3.2) Reactor Coolant Pumo Rotor Seizure i

The Reactor Coolant Pump Rotor Seizure / Shaft Break event was not in the originaldesignbasis(19)andidentifiedasopenitemsrequiringanalysisas a result of the Systematic Evaluation Program review (2).

The purpose of the RCP rotor seizure and shaft break analyses was to determined both the MDNBR and peak RCS pressure for 4 loop and 3-loop operation. The RCP rotor seizure accident resulted in the more limiting DNBR and RCS pressure than the RCP shaft break. l l

t The rotor seizure causes reactor trip due to low loop flow, which trips the turbine. Two cases were analyzed; one to maximize the peak RCS pressure and l the other to minimize the DNBR.

{

. Assumptions included; 3 loop and 4-loop HFP operation; reactor trip on low ,

j loop flow after assuming the maximum instrument response time and trip delay

[

time; a least negative Doppler and most positive MTC was used; a minimum initial RCS flow rate; a loss of AC power coincident with tae low flow f reactor trip signal; initial RCS pressure minimized in.the DNB analysis and maximized in the peak RCS pressure analysis; and core inlet temperature  !

maximized in the DNS analysis and minimized in the peak RCS pressure  !

t analysis. Other assumptions on the initial physical parameters are j consistently made conservative to achieve either the limiting MONBR or '

maximum peak RCS pressure. The net effect of these assumptions was that NUSCo conservatively analyzed the worst case events in this category.

Rotor seizure at 100% power was more severe for both IONBR and peak pressure than rotor seizure at 74% power. The results indicated that the MDN8R for 4 i loop operation at 100% power is roughly 1.4 at 2.7 seconds and for HFP 3 loop operation the MONBR was about 2.0 at 2.8 seconds. The peak p.esturizer pressure occurred for HFP 4-loop operation at 2550 psia, which is '1 belev 15 l

2750 psia, 110% of design pressure. Therefore it was concluded that for each analysis, the HFP 4-loop operation produced more limiting results than that from HFP 3 loop operation. Based on the foregoing, we have reasonable assurances that the system safety limits are not exceeded by events in this category.

2.2.3.3 (15.3.3) Reactor Coolant Pumo Shaft Break This event assumes an instantaneous separation of the RCP impeller and shaft from the RCP motor / flywheel assembly. Although flow through the affected i loop is rapidly reduced, the initial rate of reduction of coolant flow is less than the RCP rotor seizure event. Therefore, NUSCo concluded that the shaft break event is bounded by the RCP rotor seizure with respect to the minimum DNBR and peak clad temperature. We concur with this conclusion.

l 2.2.4 (15.4) Reactivity and Power Distribution Anomalies Six events in this category were analyzed by NUSCo: Uncontrolled Rod Withdrawal from Suberitical and at Power, Startup of an Isolated or Idled Loop, Boron Dilation, Dropped Rod Cluster Control Assembly, Rod Cluster Control Assembly Ejection.

2.2.4.1 (15.4.1 & 2) Uncontrolled Control Rod Groun Withdrawal The analysis contained in NUSCo 151 was replaced with reanalysis submitted in supportofTechSpecChanges(3).

Due to the Cycle 15 reload, the differential rod worth has increased substantially beyond the current design basis assumption (4). Therefore, Tech Specs were altered accordingly. In addition, Tech Specs were proposed to require a different (more restrictive) number of operating RCS loops during subcritical conditions, j Uncontrolled rod withdrawal transient analyses were performed for 4 loop l

operation comencing at 100% power and 65% power and comencing at 65% power 16 i

i

for 3 loop operation and at subcritical with and without rod stop functioning. In all suberitical transient cases, it was reasonably concluded that such transient would be terminated by the start-up rate trip or by operator action before a significant power level was reached and therefore that fuel thermal limits would not be challenged nor would DNBR limits be reached.

Parametric studies were made for those transient cases started from power, using both positive and negative axial offsets, varying the reactivity insertion rates ,and using both maximum and minimum feedback. The following specific reactivity assumptions were use: (1) least negative Doppler and most positive MTC vs most negative Doppler and MTC; (ii) highest RCCA stuck out; and (iii) reactivity insertion rates up to 22.5 pcm/sec (which was the ,

maximum obtainable from any single or combination of two banks).

Conservatism were introduced by the use of limiting values of fuel rod conductivity, maximum core inlet temperature and minimum RCS pressure and ,  !

flow. Based on the foregoing, we have reasonable assurances that NUSCo has conservatively analyzed the worst case events in this category.

l No paramctric computations were done to verify the accuracy of the RETRAN nodalization for this analysis. Nevertheless, since (i) this is .a particularly short transient in which only the primary pressure, core flow, j and inlet temperature and power are important, and since (ii) comparison by l NUSCo of the pressure and temperature computed by RETRAN to actual plant

, pressure and temperature data for a 30% load rejection and a partial loss of I feedwater event was good, and since (iii) as stated above, we found the core l

flow computation acceptable for short transient such as this, we have 4

reasonable assurances that computation of core flow, core inlet temperature and primary pressure are accurately predicted by the RETRAN model.

Furthermore, we have reasonable assurances th.c core power was conservatively computed in this analysis. Thus we find that we have reasonable assurances that the input parameters to VIPRE 01 which were derived from RETRAN are

, adequate for the purpose of the 4 RCP loss of flow analysis. Based upon the j foregoing and upon the review of the VIPRE-01 methodology contained in j Attachment I hereto, we have reasonable assurances that the system safety 17 4

,. - - - - . - - - , , - - . ~ - , - - ,

. . , limits are not exceeded by events in this category.

2.2.4.2 (15.4.3) Control Rod Assemb1v Misooeration Droceed Rod Accidgni The analysis contained in NUSCo 151 was replaced with reanalysis submitted in support of Tech Spec Changes (3).

The analysis for this event used the same methodology as described for the RCCA ejection and rod withdrawal accidents described above.

Parametric analyses were conducted varying the dropped rod worth from 0 to 180 pcm (minimum to maximum expected values) and with turbine load runback in manual and in automatic. The RCS was, in contrast to the reactivity insertion transient discussed above, assumed to be initially in conditions with maximum core inlet temperature, but with minimum primary pressure and core flow, in each case intended to produce Oe minimum computed DNBR. In addition, the most negative Doppler and MTC are used (except in the analysis of the transient with the automatic turbine runback, since there is no trip in that case) to maximize the power thus also tending to produce a lower ,

ONBR. Based upon the above, we have reasonable assurances that NUSCo has analyzed the worst case events in this category.

The minimum DNBR was obtained for the full power 4 loop operation without taking credit for either turbine runback or rod stop protective features. In that case, NUSCo concluded that the minimum DNBR remained well above 1.3. On the basis of the foregoing, we conclude that there are adequate assurances that the dropped rod accident for Haddam Neck does not exceed the system safety limits. l 1

2.2.4.3 (15.4.4) Startue of an Isolated or an Idled toen  !

Technical Specifications were established to prevent an increase in core reactivity and power when an isolated loop is brought back into service due 18

. - , to mismatching of the temperature and/or the boron concentration of the isolated loop. An idled loop is allowed to be brought back into service at a maximum power level of 60% power. Thus in the analysis the transient was initiated at maximum permissible power of 60 % and the maximum allowable temperature difference of 30*F (20'F difference plus 10'F for uncertainties) between the isolated loop and the rest of the system.

l The effect of imperfect thermal mixing was introduced into the analysis by the use of a split core model. For this model, the downcomer, plena and core volumes were divided into 3/4 and 1/4 parts and scaled proportionally. The moderator feedback was teken from the cold leg side and applied to both sections of the core. This approach results in a conservative simulation of this event and therefore is more limiting with respect to MDNBR. ,

, L j

, Although the RCPs are equipped with an interlock which prevent starting the i 4

pump if the cold leg isolation valve is open, the analysis does not take i credit for this.

1 l

Because technical specifications prohibit startup of an isolated or idled

! loop at a lower boron concentration than the remainder of the RCS, analysis j assumes that there is no boron concentration mismatch.

1 Sensitivity studies were performed to determined combinations of assumptions i which result in lower MONBR: the most negative MTC, most positive Doppler coefficient, an initially cold idle loop and starting the RCP just after opening the cold leg isolation valves. Based upon the foregoing, we have ,

i reasonable assurances that NUSCo has conservatively analyzed the worst case event in this category, r

Since the computed minimum DNBR was computed to be roughly 1.55 at about 9.2 ,

seconds into the event, we have reasonable assurances that the system safety ,

limits are not exceeded by events in this category..

5 19 1

I

l ,

.. , 2.2.4.4 (15.4.6) Baron Dilution The analysis contatned in NUSCo 151 was replaced with reanalysis submitted in ,

support of Tech Spec Changes [3).

The maximum possible dilution rate is 180 gpm. NUSCo performed a simple :

computation of the reactivity insertion rate for such dilution rate and reached the reasonable conclusion that such rate is "well within the reactivity insertion rates of the uncontrolled RCCA withdrawal analysis." On that basis, no ,RETRAN and no VIPRE-01 DNBR computations were explicitly performed for the boron dilution accident. We concur with NUSCo's conclusion and have adequate assurances that the minimum DNBR will not be challenged by the boron dilutioa accident.

2.2.4.5 (15.4.8) Rod Eiection pCCA tiection The analysis contained in NUSCo 151 was replaced with reanalysis submitted in

] support of Tech Spec Changes [3).

The RCCA Ejection represents the most rapid potential reactivity insertion accident. The analytical methodology for this accident is similar to that i

described above for the uncontrolled rod withdrawal accidents, including the following conservatisms

(1) the limiting burnup parameters were combined to a

generate the most severe system response, and (2) point kinetics was used i with no Doppler weighting multiplier. More specifically, burnup parameters (

used were: no credit for flux flattening effects of reactivity feedback; j maximum bank insertion at such power level; adverse menont margins added to ejected rod worth to account for calculational uncertainties fuel (

temperature feedback assumed to be at its minimum value over the entire burnup range; use of the most positive MTC; use of the smallest delayed neutron fraction over the entire burnup range (to minimize time to prompt criticality). In addition, trip reactivity was computed without considering

]

the ejected red and using trip and trip response delays. Based upon the l j

]

l l

J

1 foregoing, we have adequate assurances that NUSCo has conservatively analyzed the worst case events in this category.

Computations were performed for 4-loop HFP and HZP and 3 loop HFP and HZP operation.

With an assumption of failure of all rods which reached DNBR of less than 1.3, 18% of the fuel rods failed and no rod had fuel melting at the centerline in the 4 loop HZP case and none in the HFP cases. The VIPRE methodology used to make this assessment has been reviewed (see Appendix 1) and found acceptable. Thus we have adequate assurances that this computation is reasonable.

2.2.5 (15.5) Increase in Reactor Coolant Inventerv -

1

] The licensee did not analyze transients in this category since none of these transients were analyzed in original Chapter 10.

2.2.6 (15.6) Decrease in Reactor Coolant Inventerv l

Because of the objective of the topical to analyze non l.0CA transient:, most i of the transients in this category were not analyzed; the only loss of j coolant related transient analyzed was the Steam Generator Tube Rupture transient.

2.2.6.1 (15.6.3) Steam Generator Tube Failure A guillotine rupture of a single steam generator tube is assumed.

The worst case design basis event analysis would be one which maximizes the radiological releases to the atmosphere by maximizing break flow during the period prior to loop isolation, and assuming a conservatively long time for operator action to isolate. Toward that end, the primary to secondary leakage and the primary to secondary heat transfer were maximized by assuming no efforts to cool down (or depressurir') the primary side prior to loop 21

I isolation. Under the proper operating conditions, loop isolation valves can {

be closed to isolate the afft;ted steam generator which would terminate the  !

break flow. NUSCo assumed that the operator isolated the affected loop at 30 minutes.

In order to maximize the radiological consequences, reactor trip was assumed to occur virtually at event initiation, and loss of offsite power was assumod to occur coincident with reactor trip.

Operator actions, modeled during the first 30 minutes of this event were: (1) manual trip of the reactor at 10 seconds so that the iodine release (therefore the radiological release) is from the steam generator to the 3

environment rather than through the condenser; (2) starting both charging pumps 1 minute after reactor trip to keep the RCS pressure hight (3) i

isolation of AFW to the affected SG 10 minutes after reactor trip to keep the i iodine concentration in SG high; and (4) isolation of the affected loop at 30 minutes to terminate the break flow.

. Operator actions which would tend to reduce the primary pressure, and

! therefore reduce the break flow, were not modeled during these first 30 ,

minutes. Instead, in order to maximize the releases, it was assumed that the operator restarted the charging pumps which keeps the pressure in the primary

side et the steam generator high tending to maximize the break flow. In i j addition, it was assumed that offsite power was lost at reactor trip, causing j steam dump to the condenser to be lost, and resulting in steam being released 4

directly to the atmosphere.

I Although the SGTR Emergency Operating Procedures (EOPs) would have the l operator initiate the plant cooldown and depressurization to terminate the .

leakage before closure of the loop isolation valve, in order to maximize the computed release, these actions were not modeled. It is expected that after  ;

loop isolation the E0Ps would be followed to take the plant to the RNR system l entry conditions.

{ Based upon the foregoing, including specifically the fact that the primary j 22 i

I loop pressure remained high during the entire period of this computation and therefore the upper head remained full and no two phase flow conditions were encountered in the primary loop, we have reasonable assurances that the NUSCo RETRAN model is acceptable for this particular analysis and that the worst  !

case event was conservatively analyzed.

i

3.0 CONCLUSION

S AND RECOMMENDATIONS l

The licensee evaluated the ability of the Haddam Neck plant to withstand l anticipated operational occurrences and a broad spectrum of postulated l accidents without undue hazard to the health and safety of the public. The i results of these analyses were used to show conformity with General Design '

Criteria (GDC) 10, 15 and 26 of 10CFR Part50. ,

For each event analyzed, the worst operating conditions were assumed (the nuclear feedback coefficients were conservatively chosen to produce the most l

adverse core response. The reactivity insertion curve, used to represent the l

l control insertion, accounts for a stuck rod, in accordance with GDC 26), and cradit was taken for minimum engineered safeguards response.

The acceptance criteria specified in the SRP which are relevant for evaluation of the consequences of the postulated accidents were satisfied. I These include:

(1) Pressure in the reactor coolant and main steam systems should be maintained below 110% of the design pressure, except that calculated pressure of 120% of design may be permitted for very low l probability events. I (2) The potential for core damage should be evaluated on the basis that I it is acceptable if the minimum DNBR remains above the 95/95 limit l discussed in SRP 4.4. If the DN8R falls below these values, fuel damage (rod perforation) should be assumed unless it can be shown.

If fuel damage is calculated to occur, it should be of sufficiently limited extent so that the core will remain in place and geometrically intact with no loss of core cooling capability.

23

The licensee aceratt ior /ariations in initial conditions used in the analyses submitted u ' 01 cal Report NUSCo 151 and its attachments and '

adequately demonstrated that these assumptions were appropriate for the event being considered. The assumptions for initial conditions were found to be  !

acceptable because they were conservatively applied to produce the most adverse effects.

Since VIPRE 01 has been previously reviewed and found acceptable by the NRC, we have adequate assurances that the computation of DNBR by NUSCo is acceptably accurate. In addition, NUSCo has imposed sufficient degree of }

]

j conservatism to the appropriate parameters such that we have adequate l assurances of conservative results.

i NUSCo concluded that the minimum DNBR was greater than 1.3 for all transients i I except the rod ejection event (for which the fuel thermal limits were computed not to be exceeded) and therefore that fuel thermal limits would not  ;

be exceeded for any examined event. Based upon the foregoing, we have reasonable assurances that thkse conclusion are correct.

l l The timing of operator actions during the excessive feedwater transient analysis appears to be short. Acceptability of use of such short operator j action times should be further investigated.

Q Finally, since the majority of the analyses in Chapter 10 were reanalyzed and  ;

there was a considerable amount of information which was provided in a j supplementary fashion over an extended period of time after the original submittal, we recommend that all of these materials be incorporated into one  ;

document and the whole chapter be reorganized so as to conform to the current  !

Regulatory Guide 1.70 Standard Format. This would facilitate easier future

, reference of this document.

L l

f 24 r .

4 o ,' o.

]

. . . . t

4.0 REFERENCES

i t

i 1. "Haddam Neck Plant Non LOCA Transient Analysis," NUSCo 151, June .

j 30, 1986. l

[

] 2. "Integrated Plant safety Assessment Report, Systematic Evaluation ,

l Program Haddam Neck," NUREG 0826. June 1983.

l "Haddam Neck Plant Revisions to Reanalysis of Non LOCA Design Basis 1

! 3.

l Accidents," E.J. Nrocaka (CYAPC) letter to USNRC, May 8, I M7. i

)

i 4. "Haddam Neck Plant Cycle 15 Reload. Technical Specification Change  !

j Requests and Reload Report " dated June 1, 1987.

' l J  ;

i 5. 'Haddam Neck Plant - Cycle 15 Reload and Change of Technical  !

i specifications,' M.W. Hodges to C.0. Thomas, September 23, 1987 i i

i 6. "Haddam Neck Plant Additional Information Reanalysis of Non LOCA l i Design Basis Accidents," E.J. Mrocaka (CYAK) letter to USNRC, i l November 19, 1987.

l i

{ 7. 'Haddam Neck Plant Additional Information Reanalysis of Non LOCA i i Design Basis Accidents TAC #41990)," E.J. Mrocaka (CYAK) to U.S.  !

Nuclear Regulatory Come ssion, dated February 11, 1948.

l j j 8. "Haddam Neck Plant Additional Information - Reanalysis of Non LOCA l 1 Design Basis Accidents,' E.J. Mrocaka (CVAK) to USNRC, dated  !

l September 2, 1987.

)

g. "Haddam Neck Plant Revised Non LOCA Design Basis Accident Analysis, Loss of Normal Feedwater Flow," E.J. Mrocaka (CVAK) letter to i

USNRC, March 10, 1987.

10. 'Haddam Neck Plant Auxiliary Feedwater System 5tatus", J.F. Opeka j (CYAPC) letter to J.A. Zwolinski (USNRC) September 20, IME. j

11. 'Haddam Neck Plant Response to Request for Additional Information f J Concerning the Auxiliary Foodwater System ' J.F. Opeka (USNRC)

! letter to C.I. Grimes (USNRC), June 30, 1984, i

! 12. 'Haddam Neck Plant Auxiliary Feedwater Flowrates,' J.F. Opeka l letter to C.I. Srimes, dated October 14, 1986.

I

13. ' Acceptance for Referencing sf Licenstop Topical Report EPRI CCM 5, 1 'RETRAN A Program far One dimensional ' ransient Thermal -Hydraulic Analysis of Complex Fluid Flow systems,' and EPRI NP 1850 CCM,

'RETRAN 02 A Program for Transient Thorsal Hydraulic Analysis of [

Complex Fluid Flow Systems," Letter from C.O. Thomas (NRC) to T,  :

! W. Schnata (UGRA), September 4, 1984. j

! I

14. "Safety Evaluation by the Office of Nuclear Reactor Regulation I L

j 5 l

! i 1  !

l i

.. . . i

. Regarding NUSco Topical Report 140 2 VIPRE 01 Connecticut Yankee t Atomic Power Company Docket No.50 213 Haddam Neck Plant," October 1986.

i

15. "Safety Evaluation by tha Office of Nuclear Reactor Regulat'.'on Relating to NUSCo Topical Report on Physics Methodology for PWR

• Reload Design (NUSCO 152) Northeast Utilities Service Company Haddam Neck Plant Docket No. 50 213,' July 24, 1987. .

! 16. ' Standard Review Plan," NUREG 0800 Rev.2, 1981, i

17. "NUSCo Thermal Hydraulic Hodel Qualification Volume I (RETRAN),'

NUSCo 140 1, July 30, 1984.

18. "Technical Evaluation Report on NU$C0 140 1,' !TS/NRC/88 2, !

International Technical Services, Inc., dated May 16, 1988 '

19. Chapter 10 - Incidents and Potential Hazaeds. Facility Description  !

and Safety Analysis, Haddam Neck Plant May 1966.  ;

20. "Modernize the Reactor Protection and Control System,' Plant Design I Change Record, Ccnnecticut Yankee, November 21, 1986. i l

I t

f I

i l

t f

i 26 l

l I

l l

. e i

fehle 1 Casperison of WSte's Chapter 10 tents with SAP Definitions ,

t wet W be, mutte Chapt. 10 ImJ5Ce 151

1. Incatatt la staf MuovaL et int MCtacAar litt il.'

L Decreees in f eoeter imperstwee 19.9 1 '

a leuroese in Poesheter flew il f.I 10.2.5 4.4

{

latentiw treresse in Steam flow if.1.3 10.3.2 4.5 Ireevertemt 0peatg of e steen Gereretor 1 .1.4 tellef vol w or lefety weln i

i ttese Line B4twe it.1.5 10.3.3 4.9

• II. DECMatt la staf MashAL Bf 14 34CCb4Alf $lti il, )

Lees of taterrut Lead 15.8.1 10.3.5 4.11 3

Twbf re Trip 15,2.2 1  ;

Less of Cerutersee veccun 15.2.3 lesererteet Cleews of mein Steen leeletten vetwe 11.3.4 l Steen Prese w e toptater Fellw e 15.2.5 i

Less of evemergerwy Pouer to the Stellen AJallieries 19.2.6 I l Lane of hereal Poehter Flow 15.2.7 10.3.4 4.12 Ill. MCMall le GLAC10R Cpf little FLthi RAf t 1$.3 l t

linete & httlple teettee Castem P W Trips 15.3.1 4.8 teettee Coolent Pwp D ef* Seitw o 1$.3.3 4.13 teettee seulemt Pep Imeft treet 15.3.3 4.13 f

l I

i i

f i

1

e, 1at>le 1 (sentinued)

(vent SAP be, tutte Chapt.10 LtIlce 151 IV. ItActivitt Aka PCn48 OllitlIUtlCh Aas:snatigt 11.4 Uncentrolled Centret Bad Grae Witheremet 15.4.1 10.2.1.2 4.1.1 free a thritical Ceditim t,msernretled Centret Bad Grae Withermeet et Pouer 15.4.2 f..I.1.3 4.1.2 Cetret Red Aweady miseperetten 11.4.3 10.t.6 & F 4.7 Start e of inacti n Sou ter testent Piage 15.4.4 10.2.2 4.2 metf a tten of a adt Leap flow Centrollee 15.4.5 5/4 t/4 Ed es md Pw ifisetten I ntes P3tfinction 15.4.4 10.2.3 4.3 ,

tredrertent Leedire ord Deerette of 15.4.7 e rust Awese>ty in en leproper peeltlen Spectrum of ted Ijectlen Addicente 15.4.8 10.2.7 4.4 L

l v. Istatant la stAtton C00LAnt luvtutost 15.3 1

Innerettent Operetten of ICCS Dwirg Power Operette 15.5.1 menee arti Pwificetten Intes malf 6metten 15.1.2 Castrg increceed teacter Cootent trmtery v!. HCatast is etAtten Coolant leNtatont 15.6 treevertent Osentrg of a Presowlser lef aty 15.6.1 er belief Detre f allwe of teelt Liree Carryirg 11.6.2 Priesey Leet et Chteles Centelreent l Steen Gerwroter inee feltwe 15.4.3 4.10 Spectrw of Det Pipe feltwee 15.6.4 t/4 t/A l

Lees of Costant Acclamt 15.6.1 10.3.3

^

l 28 1 \

l \

l l

1 l

' ~~

'^

.a ,., ,

Attachment 1 W 3L DNBR Limit In the staff safety evaluation report (1) accepting the licensee's use of the VIPRE-01 code for licensing application, it was required that the licensee perform an analysis to demonstrate that the ENBR limit of 1.3 for the W 3 -

critical heat flux correlation (with the spacer grid convection factor) ut.1 in VIPRE-01 can predict its data base of DNB occurrence with at least a 95%

probability at a 95% cor.fidence level. In Appendix A, the licensee perfo.wed an analysis of the W-3L CHF correlation using the EPRI CHF test data [2]. -

! These' data are from the test assemblies similar to the Haddam Neck Plant ,

fuel. For each test run, the CHF is calculated using the W-3L correlation, ,

and a ratio of the predicted CHF to the measured CHF is obtained. A distribution of these ratios is obtained for all the test data. A statistical analysis is then performed to determine a minimum DNBR limit '

which ensures with a 95% probability at a 95% confidence level that departure from nucleate boiling will not occur for a higher DNBR. The staff has reviewed the analysis and confirmed that the DNBR limit of 1.3 used in the HilP analysis is conservative and acceptable.

References:

1. Letter from F.M. Akstulewicz (NRC) to J.F. Opaka (CYAPCo), "NUSCo Thermal H.ydraulic Model Qualification, Volume II (VIPRE). Topical Report NU.Co 140 2," October 16, 1986.

2. EPRI.HP.2609, "Parametric Study of CHF Data. Vols. I through 3,"

September 1982 29 l

l

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