ML20054K382

Safety Evaluation Supporting Amend 71 to License DPR-53

ML20054K382
Person / Time
Site:	Calvert Cliffs
Issue date:	06/24/1982
From:	Office of Nuclear Reactor Regulation
To:
Shared Package
ML20054K380	List: ML20054K379 ML20054K382 ML20054K383
References
NUDOCS 8207010519
Download: ML20054K382 (75)

Text

.

~%r:. v:.

o, UNITED STATES l

8 3.-

(,;

NUCLEAR REGULATORY COMMISSION b p gg

'" /

E W ASHINGTON, D. C. 20555

.%; v(/

DESIGITATED ORIGIllAL Cortified By b

07%Ld!

U SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION SUPPORTING AMENDMENT NO. 71 TO FACILITY OPERATING LICENSE NO. DPR-53 CALVERT CLIFFS NUCLEAR POWER PLANT UNIT N0. 1 DOCKET NO. 50-317 I

l P

o-

-Table of Contents Page 1.0 Introduction 1

2.0 Analytic and Test Methods 2

2.1 Thermal Hydraulic Evaluation 5

2.1.1 Design Methodology Review 6

2.1.2 CETOP-D Thermal Margin Design Analysis Code 6

2.1.3 CE-1 Correlation (Generic Limit) 8 2.1.4 Fuel Rod Bow 8

2.1.5 SCO Review 8

2.1.6 Comparison of Thermal Hydraulic Design 9

Conditions 2.1.7 Evaluation Summary 12 2.2 Fuel System Design Evaluation 13 2.2.1 Cladding Creep Collapse 13 2.2.2 CladdingSweilingandRuptureDuringLOCA 14 2.2.3 Fuel Assembly Shoulder Gap 15 2.2.4 Guide Tube and CEA Integrity 16 2.2.5 Thermal Performance Analysis 18 2.2.6 Other Experimental or Demonstration Fuel 20 Assemblies 2.3 Nuclear Design Evaluation 21 2.3.1 Nuclear Parameters 22 3.0 Safety Analysis 24 3.1 CEA Withdrawal-24 3.2 Full Length CEA Drop Event 24 3.3 Fuel Misloading Event 25 3.4 CEA Ejecticn Event 26 I

3.5 Boron Dilution Event 30 3.6 Startup of an Inactive Reactor Coolant Pump 30 Event (" Cold Water Accident")

3.7 Loss of Load Event 30 3.8 Excess Load Event 31 3.9 Loss of Feedwater Event 32 3.10 Excess Heat Removal 32 3.11 Reactor Coolant System Depressurization 33 Event

Page 3.12 Loss of Coolant Flow (LOCF) Event 33 3.13 Anticipated Operational Occurrences Resulting from the Malfunction of One Steam Generator 34 3.14 Loss of All Non-emergency AC Power Event 35 3.15 Steam Line Rupture Event 36 3.16 Steam Generator Tube Rupture (SGTR) Event 37 3.17 Seized Rotor Event 39 3.18 Loss of Coolant Accident (LOCA) 39 3.19 Fuel Handling Accident 40 4.0 Radiological Consequences of Accidents for Extended 41 Burnup 5.0 Technical Specifications 42 5.1 Thermal Margin Safety Limits 42 5.2 Peripheral Axial Shape Index 43 5.3 Minimum DNBR and Power Uncertainty 43 5.4 Use of Excore Detectors for Linear Heat Rate 43 Monitoring 5.5 Implementation of BASSS 44 5.6 Shutdown Margin 44 5.7 MSIV Closure Time 44 5.8 RTD Response Time 45 5.9 fuel Enrichment 45 5.10 Pressure Transmitters 45 6.0 Conclusions 47 6.1 Environmental Considerations 47 l

6.2 Safety Conclusions 47 l

7.0 References 48 Appendix A - Safety Evaluation of CEN-124(B)-P, Parts 1, 2 53 and 3, " Statistical Combination of Uncertain-ties (SCU)"

Appendix B - Safety Evaluation of CEN-182(B), " Statistical 67 Approach to Analyzing Creep Collapse of Oval Fuel Rod Cladding Using CEPAN" Appendix C - Safety Evaluation of CENPD-183(B)-P, "Applica-70 tion of CENPD-198 to Zircaloy Component Dimensional Changes" l

l

1.0 INTRODUCTION

By application dated February 17, 1982, Baltimore Gas and Electric Company (BG&E), the licensee, requested changes to Technical Specifications (TS) for Calvert Cliffs Unit 1.

The proposed changes to the TS are required to allow Cycle 6 operation. The February 17, 1982 application was sup-plemented by letter dated April 28, 1982.

The BG&E application and supporting analyses are unusual in that (1) they

~

incorporated a number of previously unreferenced analytic techniques, l

and (2) the analyses assume that a longer fuel cycle (extended burn-up) l will be utilized.

Section 2.0, herein, addresses the analytic methods l

and test programs which have been used to support the Cycle 6 applica-tion.

Previously unreferenced analytic techniques are further addressed in Appendices A, B, and C, herein.

Section 3.0 presents the results of accident and transient analyses.

Section 4.0 provides an evaluation of the radiological consequences of various accidents and transients and, specifically, the effects attributable to the. extended burn-up Cycle 6.

The proposed Technical Specifications are addressed in Section 5.0 while our conclusions are presented in Sections 6.0 and 7.0, respectively.

i 4

l l

f 1

L

~i 2.0 ANALYTIC AND TEST METHODS We have reviewed the request by Baltimore Gas and Electric Company to reload and operate Unit 1 of the Calvert Cliffs Nuclear Power Plant for Cycle 6, with regard to the acceptability of the analytic methodology.

The thermal-hydraulics evaluation is presented in Section 2.1, the fuels evaluation in Section 2.2, and the physics evaluation in Section 2.3.

In addition, three appendices are included. Appendix A is an evaluation of the Statistical Combination of Uncertainties (SCU) topical reports, CEN-124(B)-P, Parts 1, 2, and 3, by the Thermal-Hydraulics Section. Ap-pendix B is the Fuels Section Evaluation of the Statistical Approach to Analyzing Creep Collapse of Oval Fuel Rod Cladding Using CEPAN, CEN-182(B).

Appendix C is the Fuels Section evaluation of the Application of CENPD-198 to Zircaloy Component Dimensional Changes, CEN-183(B)-P.

The analytic methods described herein were used by BG&E to investigate the behavior of Calvert Cliffs Unit 1 (CC-1) during Cycle 6 operation.

The fuel management pattern was developed to accommodate a Cycle-5 endpoint exposure up to 13,000 MWD /MTU, thus making the core-average end-of-cycle (E0C) 5 exposure about 21,900 MWD /MTV.

The Cycle-6 core will be comprised of 217 fuel assemblies that were manufactured by C-E, the original NSSS vendor. After the reload, the core-average beginning-of-cycle (B0C) 6 exposure will be about 10,900 ffWD/fiTU, thus making the pre-dicted core-average EOC 6 exposure about 24,900 f1WD/?1TU.

The Cycle-6 core loading inventory is given in Table 2-1.

The major changes to the core for Cycle-6 are the removal of 1 Batch-D assembly, j

28 Batch-F assemblies, and 52 Batch-E assemblies. These assemblies will be replaced with 40 Batch-H (high enrichment) assemblies,32 Batch-H/ assemblies, and 9 previously irradiated Batch-D assemblies.

To accommodate extended burnup cycles, each of the Batch-G/ and Batch-H/ fuel assemblies employs 8 burn-able poison pins.

l l

2

1 As in the previous cycle, the Cycle-6 core will contain 3 unrodded assemblies with Inconel tubes placed inside the guide tubes. These assemblies will reside in high flux regions. The intent of this experimental program is to provide irradiated Inconel-625 for material property testing.

Also, the SCOUT demonstra-tion assembly, a Batch-F test assembly, will remain in the core for a third cycle of irradiation.

Of the Batch-G assemblies, 4 are designated as PROTOTYPE demonstration assemblies.

These assemblies are considered as " experimental test assemblies" and will be undergoing their second cycle of irradiation. The PROTOTYPE program is an extension of the SCOUT program and includes unique fuel designs.

The intent of the PROTOTYPE program is to provide a statistical basis for interpreting new fuel design performance.

1 3

J Table 2-1 Calvert Cliffs, Unit 1, Cycle 6 Core Loading Inventory Initial BOC E0C Assembly Number of Enrichment Burnup Average Burnup Average l

Designation Assemblies (w/oU235)

(MWD /HTU)

(MWD /MTU)

D042 1

3.03 32,200 42,800 D

8 3.03 19,900 33,000 F*

44

-3.03 23,000 35,900 G**

40 3.65 10,200

.26,200 G/

52 3.03 14,000 29,400 H

40 4.00 0

12,100 H/

32 3.55 0

16,800

~

Includes one SCOUT demonstration assenbly.

Includes four PROTOTYPE assemblies.

l l

1 1

4 c

)

2.1 THERMAL-HYDRAULIC EVALUATION By letter dated November 19,1981 (Ref.1) Baltimore Gas and Electric (BG&E),

the licensee, proposed modifications to Technical Specifications and provided reports in support of Calvert Cliffs, Unit 1 (CC-1), Cycle 6 Reload. These reports include the safety analyses for those, transients which required reanalysis, a comparison of the Cycle 6 themal hydraulic parameters at full power with those of Cycle 5, and the proposed modifications to the Technical Specifications due to changes in methodology.

In addition, the following reports describe the methodology changes implemented for the Cycle 6 themal hydraulic analyses to show that acceptable themal margin is maintained at the full power.

(a) The CETOP-D Core Themal Margin Code (Ref. 3) 4 This code replaces the TORC code (Refs. 8, 9, 10) used in Calvert Cliffs-1 Cycle 5 analysis.

(b)

Effects of fuel rod bow on DNBR margin (Ref. 4)

Proposed modifications on the effects of fuel rod bow on DNBR to Calvert Cliffs-1 Cycle 6 are described in this report.

This report is under review by the staff and is scheduled for completion in June 1982.

(c) Statistical Combina tion of Uncertainties (Refs. 5, 6, and 7)

The themal margin methodology for Calvert Cliffs-1 C cle 6 has been modi-f fied by the application of statistical methods instead of the application of deteministic methods applied in Calvert Cliffs-1 Cycle 5 for the treat-ment of uncertainties.

5

i The objective of this review is to confim that the themal hydraulic design of the reload core has been accomplished using acceptable methods, and provides acceptable margin of safety from conditions which could lead to fuel damage during nomal operation and anticipated operational transients.

2.1.1 Design Methodology Review The Calvert Cliffs-1 (CC-1) Cycle 6 design methodology involves several changes over Cycle 5.

The TORC /CE-1 (Refs. 8, 9, 10, 11, and 12) thermal design code has been replaced by the CETOP/CE-1 code. The treatment of plant system paran-eter uncertainties has been changed from the deteministic approach to statis-tical combination of uncertainties (SCU) and incorporates the system parameter uncertainties directly in the DNBR limit (Refs. 5, 6, and 7). The rod bow compensation for the proposed DNBR limit is also calculated using a method (Ref. 4) which is under review but not yet approved. Therefore, the Cycle 6 themal design is a major change from the original Cycle 5 design methodology.

2.1.2 CETOP-D Themal fiargin Design Analysis Code The CETOP-D computer code is used as a core themal margin design analysis tool for the CC-1 Cycle 6 reload. CETOP-D is an open-lattice themal hydraulic code which solves the same conservation equations and uses the same constitutive equations as in the TORC code (Refs. 8, 9).

TORC, derived from COBRA-III C (Ref.13), is a multi-stage themal margin code. The detemination of hot channel coolant conditions and minimum DNBR are perfomed through three sequen-tial steps, i.e., core-wide, hot fuel assembly and hot subchannel DNBR calculations.

A simplified TORC design modeling method was developed and described in CENPD-206P (Ref.10).

In simplified TORC, two sequential calcu-lations are made for themal margin analysis, i.e., a core-wide analysis detemining lateral boundary conditions for hot assembly; and a hot assembly The analysis detemining hot subchannel coolant conditions and mininum DNBR.

CETOP-D design code simplifies one step further by simply using a one step calculation for the core themal margin analysis.

The modeling uses a four-channel core representation with a lumped-channel technique.

It uses " transport coeffic:ents" serving as weighting factors for the treatment of diversion 6

crossflow and turbulent mixing between adjoinir; channels.

Furthemore, a

" prediction-correction" method is used to solve the conservation equations, replacing the iterative method used in the TORC code.

The magnitude of the changes, therefore, requires that the CETOP-D code be totally reviewed for acceptability as a themal design tool.

The licensee has submitted the Calvert Cliffs eersion of the CETOP-D topical report (Ref. 3).

The CC-1 CETOP-D topical repart is essentially identical to the Arkansas Nuclear One Unit-2 (ANO-2) C: TOP-) topical report (Ref.14) except for the plant-specific data base constants.

Tie staff has previously reviewed extensively the CETOP-D code for ANO-2 and found it &cceptable (Ref. 15).

The review of this code included the lumped subchannel modeling; the use of the prediction-correction method to solve the contervation equations; the use of transport coefficients for the treatment of diversion crossflow and turbulent mixing between adjacent channels; and the use of a hot assembly flow starvation factor to ensure the conservatism of CETOP-D with respect to TORC.

The details of the review on the ANO-2 CETOP-D program are addressed in the ANO-2 Cycle 2 reload SER (Ref.15).

The staff has reviewed the CC-1 CETOP-D topical and discovered many l

typographical errors which also existed in the ANO-2 CETOP-D topical and'were l

identified as non-consequential during the ANO-2 reload review.

The licensee has subsequently submitted an errata to correct these errors (Refs. 2 and 3).

In order to avoid duplication of efforts, the staff review on the Calve'rt Cliffs CETOP program is concentrated on the plant-specific constants. TN values of the flow starvation factors given in the data base document (Ref. 3) are detemined by comparison between the CETOP-D and TORC results of the themal margin analysis using Calvert Cliffs plant data. This comparison is described ir. Section 5 of Ref. 3 and it covers the plant operating ranges of inlet temperature, system pressure, primary flow rate, and axial shape 'index.

In all the cases provided, the CETOP-D calculates minimum DNBR lower than the TORC calculations.

Since the TODC code has been approved for use.in Combustion Engineering themal margin design, the staff concludes, based on the conserva-tism of CETOP-D relative to TORC, that the CETOP-D code is acceptable for Calvert Cliffs themal margin calculations.

7

l t

2.1.3 CE-1 Correlation (Generic 1.imit) i The CE-1-correlation has previously been approved for interim plant-specific j

applications with a minimum DNBR limit of 1.19.

However, our generic evalua-j tion has now been completed; our findings will be discussed in detail in the Safety Evaluation Report of CENPD-207-P (Ref.12).

The proposed limit of 1.19 for the CE-1 correlation is conservative in comparison to 14x14 CHF test data and is, therefore, acceptable.

l 2.1.4 Fuel Rod Bow j

I The licensee has proposed a rod bow compensation of 0.6 percent on DNBR using the method described in Supplement 3P to CENPD-225-p (Ref. 4) which is not an approved document.

Results of this document will be pemitted for use af ter l

the approval if the licensee desires. Accordingly, it is the staff position that the approved interim method of the rod bow compensation described in 1

Reference 17 shall be applied for Cycle 6 (Refs. I and 2) operation. A total of 137 fuel assemblies will exceed the NRC specified DNB penalty threshold burnup of 24 GWD/T (Ref. 17) during Cycle 6, the maximum assembly burnup reaching 42.8 GWD/T by the end of cycle.

For those assemblies which will experience a burnup of between 24 and 28.3 GWD/T at any time during Cycle 6, I

the minimum best estimate margin available relative to more limiting peaking values present in other assemblies is greater than 10 percent.

The DNB rod bow penalty for this burnup range, as detemined from Reference 2, varies from j

i 0 to 1.4 percent.

For assemblies which experience burnups in excess of 28.3 GWD/T, up to a maxinun of 42.8 GWD at E0C-6 for one assembly, the minimum best l

i estimate margin available is considerably greater than 20 percent. The DNB rod bow penalty for this latter burnup range varies from 1.4 to 6.3 percent.

t In summary, for both burnup ranges, the magnitude of the nargin available is I

considerably in excess of the corresponding DNB rod bow penalty. Therefore.

l no power penalty for fuel rod bowing is required in Cycle 6.

I l

2.1 5 SCU Review j

i The staff, in conjunction with our contractor, Battelle Pacific Northwest l

Laboratories has reviewed the SCU methodology presented in CEN-124(B)-P; cur I

8 l

l

o evaluation is described in Appendix A.

We have concluded that the SCU is acceptable with the following exceptions:

1.

code uncertainties of 5 percent 2nuld be included in SCU analysis; 2.

pending approval of CENPD-225-P, the currently approved interim nethod for rod bow should be used for rod bow compensation calculation; 3.

the new equivalent DNBR limit is 1.23 including SCU for system parameters and excluding rod bow compensation on DNBR; 4.

any changes in codes or correlations used in the analysis will require a reevaluation of the SCU; 5.

there are errors in Table 3-1 of the reports (Refs. 5 and 7).

We require that the corrected values pr.ovided in Reference 18 be used in future analyses; and 6.

nominal initial conditions chosen for use in analysis should bound all pemitted methods of plant operation in subsequent cycles.

2.1.6 Comparison of Thermal Hydraulic Design Conditions A comparison of the thermal hydraulic design conditions for CC-1 Cycle 5 and 6 is provided in Table 2-2.

Cycle 6 is characterized by a higher total reactor coolant flow, higher coolant flow through the core, higher average mass velo-city, higher pressure drop across the core, higher total pressure drop across l

the vessel, and higher film coefficient at average conditions. Other differences I

for Cycle 6 compared to Cycle 5 are lower inlet temperature, lower average film temperature difference, and lower average core enthalpy rise. These dif-ferences between Cycle 5 and Cycle 6 values are due to the application of the SCU methods to Cycle 6.

The actual plant values of these parameters have not changed.

The SCU methods combine measurement and other uncertainties, statis-tically, to obtain a penalty on power that accounts for each of the component uncertainties.

Consequently, no allowance is needed for uncertainties in the individual values of the parameters (Ref. 2).

f 9

Table 2-2 Calvert Cliffs Unit 1 Themal-Hydraulic Parameters at Full Power Reference General Characteristics Unit Cycle 5*

Cycle 6**

2700 2700 Total Heat Output (core only)

MW{ Btu /hr 10 9215 9215 Fraction of Heat Generated in Fuel Rod

.975

.975 Primary System Pressure Nominal psia 2250 2250 Minimum in steady state psia 2200 fiaximum in steady state psia 2300 Inlet Temperature

• F 550 548 Total Reactor Coolant Flow gpg 370.000 381,600 (steady state) 10 lb/hr 139.0 143.8 6

Coolant Flow Through Core 10 1b/hr 133.9 138.5 Hydraulic Diameter ft 0.044 0.044 (nominal channel) 6 2

Average fiass Velocity 10 lb/hr-ft 2.51 2.61 Pressure Drop Across Core psi 10.4 11.1 (minimum steady state flow irreversible Up over entire fuel assembly)

Total Pressure Drop Across psi 32.4 34.4 Vessel (based on nominal dimensions and minimum steady state flow) 2 Core Average Heat Flux (accounts Btu /hr-ft 186,435***

184,266 for above fraction of heat generated in fuel rod and axial densification factor) 2 Total Heat Transfer Area ft 48,192***

48,748 (Accounts for axial densifi-cation factor) 2 Film Coefficient at Average 8tu/hr-ft 5765***

5930***

Conditions Average Film Temperature

• F 32-31 Difference See footnote (s) last page of table.

(cont'd) 10 l

i Table 2 (Continued)

Reference General Characteristics Unit Cycle 5*

Cycle 6**

Average Linear Heat Rate of kW/ft 6.23***

6.15*****

Undensified Fuel Rod (accounts for above fraction of heat generated in fuel rod)

Average Core Enthalpy Rise Btu /lb 68.8 66.5 Maximum Clad Surface Temperature

'F 657 657 Calculational Factors Engineering Heat Flux Factor 1.03 1.03****

Engineering Factor on Hot Channel 1.02 1.02****

Heat Input Rod Pitch and Clad Diameter 1.065 1.065****

Factor Fuel Densification Factor (axial) 1.01 1.01 Total Planar Radial Peaking Factors For DNB Margin Analyses (Fr) 1.620 1.700 For kW/f t Limit Analyses (Fxy) 1.620 1.700 Peak Allowable Linear Heat Generation Rate (kW/ft) 15.5 15.5 Limiting Transient Minimum DNBR CEA Drop 1.195 1.23 Loss of Flow 1.195 1.23 Minimum Allowable DNBR 1.195 1.23 NOTES

• Design inlet temperature and nominal primary system pressure were used to calculate these parameters

• • Due to the statistical combination of uncertainties described in References 5, 6 and 7, the noninal inlet temperature and nominal primary system pressure were used to calculate some of these parameters.

i Based on a generic value of 1100 shims.

• • • • These factors have been canbined statistically with other uncertainty factors at 95/95 confidence / probability level (Ref. 6) to define a new design limit on CE-l' minimum DNBR when iterating on power as discussed

~

in Reference 6.

• • • • • Based on Cycle 6 specific value of 672 shims.

11

2.1.7 Evaluation Summary We have reviewed Calvert Cliffs-1 Phase I Cycle 6 Reload thermal design methodology and safety analyses as summarized below:

(a) The CETOP code is acceptable for use in Calvert Cliffs Unit I and Unit 2 safety analyses as a substitute for TORC.

(b)

The CE-1 DNBR limit for Calvert Cliffs Unit I and Unit 2 has been evaluated.

The proposed limit of 1.19 for the CE-1 correlation is conservative in comparison to 14x14 CHF test data and is, therefore, acceptable.

(c) Our review of'SCU is complete. We have found the SCU methodology acceptable.

However, a correlation cross-validation uncertainty and a 5 percent code uncertainty must be included resulting in an increase of DNBR limit by 0.015. The approved DNBR limit is 1.23 excluding rod bow compensa tion.

(d) According to Section 2.1.4, no rod bow compensation is required for Calvert Cliffs-1 Cycle 6.

Results of Supplement 3P to CENPD-225 (Ref. 4) will be permitted for use after the approval if the licensee desires.

l 12

1 2.2 FUEL SYSTEM DESIGN EVALUATION The objectives of the fuel system safety review are to provide assurance that (a) the fuel systen is not damaged as a result of normal operation and anticipated operational occurrences, (b) fuel systen damage is never so severe as to prevent control rod insertion when it is required, (c) the number of fuel rod failures is not underestimated for postulated accidents, and (d) coolability is always maintained. We have reviewed the information provided in support of Calvert Cliffs, Unit 1, Cycle-6 operation to determine if these objectives have been met.

Our evaluation of the fuel design is based on engineering analyses, tests, and a substantial amount of in-reactor operating experience.

In addition, the performance of the design is subject to continuing surveillance of operating reactors by C-E and licensees having C-E NSSS plants.

These programs continually provide confirmation and current design performance information.

2.2.1 Cladding Creep Collapse Ca,bustion Engineering has written a computer code that calculates time-to-collapse of Zircaloy cladding in a pressurized water reactor environment.

This code is described in the report CENPD-187, "CEPAN Method of Analyzing Creep Collapse of Oval Cladding" (Ref. 26). We have reviewed this code and found it ' acceptable as described in our safety evaluation, which is contained in Appendix B, herein.

For Cycle-6 operation, C-E has performed time-to-cladding-collapse calculations using (a) CEPAN, (b) a new statistical method (see Appendix B) of establishing data input to initialize the calculation, and (c) and new criterion (see Appendix B) for the occurrence of collapse.

The input data include internal rod pressure, cladding dimensions, cladding temperature, and neutron flux.

13

The results of this analysis showed that the minimum time-to-collapse is in excess of the design batch-average discharge lifetime of the fuel, which will not be exceeded during Cycle-6 operation. We, therefore, conclude that the fuel rod cladding in the Calvert Cliffs, Unit 1 Cycle-6 core will not collapse and is acceptable in this regard.

2.2.2 Cladding Swelling and Rupture During LOCA The NRC staff has been generically evaluating three materials models that are used in ECCS evaluation models (EM). Those models are for cladding rupture temperature, cladding burst strain, and fuel assembly flow blockage. We have previously concluded (NUREG-0630, Ref. 27) that these three materials models in the C-E ECCS Eli were non-conservative over some regions of applicability.

Although C-E has subnitted a new ECCS Efi that incorporates revised materials models, the NRC review of the new ECCS Et1 has not been completed and the new ECCS Eli has not been used for the Calvert Cliffs, Unit 1, Cycle-6 LOCA analysis.

Hence, supplemental ECCS calculations are needed to confim that Calvert Cliffs, Unit 1, would continue to be in confomance with the ECCS criteria of 10 CFR 50.46 if NRC staff materials models (NUREG-0630) were substituted for those models of the C-E ECCS EM.

Baltimore Gas and Electric has reviewed the major Cycle-6 parameters (i.e.,

reflood rates, peak rod power, fuel rod pressure, stored energy) affecting ECCS perfomance for the limiting large-break LOCA analysis and concluded that, in this respect, the Cycle-6 fuel performance (e.g., time of hot rod rupture) during a LOCA would be very similar to that of Cycle 5.

Consequently, BG&E has not perfomed supplemental ECCS calculations for Cycle-6 operation, but instead has drawn upon (a) a previous generic calculation submitted for Calvert Cliffs operation and (b) inferences from CE calculations submitted for l

new operating license applications.

The previous generic calculation submitted for Calvert Cliffs operation pre-dicted that the peak fuel cladding temperature will be lowered with the use of the NUREG-0630 ramp-rate-dependent strain and flow blockage models, provided that offsetting margins are allowed for the use of the new steam cooling models in the C-E revised ECCS evaluation model.

This infomation, however, 14

i i

did not address the impact that the use of the NUREG-0630 rupture temperature model would have on the Calvert Cliffs LOCA analysis.

In the stress region of application to the Calvert Cliffs analysis, the NUREG-0630 rupture temperature model underpredicts (i.e., is more conservative than) the C-E rupture temper-ature model.

However, we believe that the impact of this omission is offset by C-E's conservative use of only peak strain and flow blockage values that are given in NUREG-0630, irrespective of the specific Calvert Cliffs cladding failure stress and temperature conditions. We therefore conclude that BG&E has provided an acceptable justification that Calvert Cliffs, Unit 1, will remain in compliance with 10 CFR 50.46 criteria during Cycle 6 (which involves operation at a peak linear heat generation rate of 15.5 kW/f t).

We therefore conclude that the concerns related to LOCA-induced cladding swelling and rupture are satisfied for Cycle 6 operation.

i 2.2.3 Fuel Assembly Shoulder Gap During irradiation, fuel rods and fuel assembly guide tubes undergo axial growth at different rates.

If this differential growth progresses to the point of consuming all of the available shoulder gap, then mechanical inter-ference will occur between the fuel rod end caps and the fuel assembly s tructu re.

To ensure that an adequate design shoulder gap exists for the fuel assemblies that will comprise the Cycle-6 core, C-E has perforned calculations on all Cycle-6 fuel.

These calculations were performed with the methods described in the C-E topical report, CENPD-198, (Ref. 28) its 2 supplements, (Refs. 29, 30) and CEN-183 (Ref. 31, See Appendix C). We have reviewed these topical reports and approved then for referencing (see Appendix C).

Fram these calculations, C-E concluded that all clearances will be adequate during Cycle-6 operation.

Therefore, we conclude that the concern of adequate fuel assembly shoulder gap has been satisfied for Cycle-6 operation.

15

2.2.4 Guide Tube and CEA Integrity Fretting wear in operating C-E reactors has been observed in irradiated fuel assemblies and control element assemblies (Nf. 32). Observations have revealed unexpected degradation of fuel assunbly guide tubes that were under CEAs and of the cladding on these CEAs.

It was concluded that coolant turbulence was responsible for inducing vibratory motions in the nonnally fully withdrawn control rods and, when these vibrating rods were in contact with the inner surface of the guide tubes, wearing of the guide tube wall and CEA cladding has taken place.

The wear of the guide tubes had been more severe than the wear occurring on CEAs because the guide tubes are constructed of relatively sof t Zircaloy-4 whereas the Inconel-625 cladding on the control rods provides a relatively hard wear surface.

The extent of the observed wear has appeared to be plant dependent and has in some cases extended completely through the guide tube wall.

As an interim fix, BG&E installed stainless steel sleeves in new and old fuel assembly guide tubes that are to be used in CEA positions.

Other guide tubes have been modified by reducing the number and/or size of the flow holes, thus reducing the turbulence by reducing the coolant mass flow which passes through the guide tubes.

Our review of sleeving programs has been documented in previous safety evaluations; for example see the Millstone-2 Cycle 3 reload safety evaluation in Reference 33 which concluded that guide tube sleeves will perfonn their function of reducing guide tube stresses to acceptably low values in worn assemblies and that sleeves are satisfactory for mitigating further fretting l

wear in irradiated or fresh fuel assemblies.

Our previous approvals of Cycles 4 and 5 also permitted operation with unsleeved, reduced-flow fuel assemblies, which were placed in CEA positions.

Those approvals were based on C-E out-of-pile flow tests that indicated that the resulting decrease in guide tube flow is accompanied by less CEA flow-induced vibration and, therefore, less guide tube wear.

Thirty-two reduced-i flow fuel assemblies from Cycle 5 will be reused during Cycle-6 operation in rodded positions.

16

During the Cycle-6 outage, wear measurements with bobbin and azimuthal eddy current probes were perfomed on all 160 guide tubes in those 32 reduced-flow fuel assemblies.

Preliminary results (interpretations of the eddy current probe voltages) are that 9 guide tubes had wear, but that the wear "was minor and does not affect the integrity of the guide tube."

Ninety days after returning to power, BG&E should submit a fomal report for NRC review on the present fretting wear inspections. We anticipate that these fuel assembly inspections will not reveal any unacceptable fretting wear and that the observed degree of wear would serve as a basis upon which to detemine whether surveillance should be perfomed durir.9 the Cycle-7 outage.

It is.

l however, desirable that some surveillance should be perfomed during the Cycle-7 outage on reduced-flow fuel assemblies that are to be rodded again in Cycle-7. Therefore, a submittal to specify all plans for continuing guide tube fretting wear examinations should be submitted 90 days prior to the Cycle-6 shutdown for refueling.

We therefore conclude that (a) the guide tubes in the C-E sleeved fuel assemblies will continue to meet their design functions, and (b) the guide tubes in the C-E reduced-flow fuel assemblies should be acceptably resistant to wear; however, if they fail to perfom as predicted, the overall degradation to the core is restricted to a total of 32 fuel assemblies.

During the Cycle-6 outage, there were no CEA fretting wear measurements perfomed.

On the basis of previous testing conducted on CEAs from Unit 1 at the end of Cycles 3 and 4 and from Unit 2 at the end of Cycles 2 and 3, BG&E concluded that the rate of CEA wear degradation would allow continued operation for several nore cycles.

The rate of CEA fretting wear and the ability to maintain hemiticity and hence to achieve design lifetime of CEAs deserves further discussion between f

NRC and BG&E. Therefore, we recommend that the BG&E fretting wear report l

discussed above should also describe continuing CEA surveillance or the justification in detail for discontinuing these examinations.

At a minimum, such a proposal should be submitted 90 days prior to the Cycle-6 shutdown for refueling.

l 1

17

2.2.5 Thermal Performance Anal d The perfomance of the fuel in the Calvert Cliffs, Unit 1, Cycle-6 core has been enalyzed using a revised version (Ref. 34) of the Combustion Engineering fuel perfomance code (Ref. 35), called FATES-3.

The Cycle-6 reload safety analysis is the first time FATES-3 has been used in a licensing application.

The code is used in e number of areas in the safety analysis, including fuel rod initial conditions for the analysis of the LOCA and other transients and accidents, the power-to-centerline melt limit, minimum and maximum core-average gap conductance, fuel stored energy for containment analysis, maximum end-of-life rod pressure, and fuel mechanical design limits.

We have not yet completed our review of the FATES-3 code and the review to date has resulted in several unresolved issues, notably code conservatism and fission gas release.

Becayse the reload schedule for Cycle-6 has not pemitted resolution of these issues, we have obtained Cycle-6 specific conditions from l

the licensee (Ref. 36) and have reproduced a number of the fuel perfomance calculations described in the reload report with a staff audit code (Ref. 37).

These calculations, which include LOCA initial conditions, power-to-conterline melt, maximum fuel average temperatures, and end-of-life rod pressure, were expected to be most limitirg in the reload safety analysis and most affected by those issues identified previously.

A comparison of the results of our audit calculations with those obtained from the FATES-3 code shows a number of differences.

The maximum fuel-averace temperatures calculated by the staff for low and moderate power levels _(i.e.,

near core-average conditions) were lower than those calculated with FATES-3, whereas the power-to-centerline melt limit calculated by the staff was higher than that calculated with FATES-3.

In both of these cases, the analyses l

presented by BG&E are more limiting than those produced by the staff. As a consequence, we find the core-average temperature conditions and power-to-centerline melt limits in the Cycle-6 reload report acceptable.

18

In the remaining two areas, LOCA initial conditions and end-of-life rod pressure, the analyses presented by BG&E are less limiting than those produced by the staff. The fact that the conditions specified in the reload report are less severe than those predicted by our own audit code does not invalidate the FATES-3 results.

However, we must question these results pending completion of the FATES-3 review.

In response to our concern about the LOCA initial conditions calculated with FATES-3, BG&E has reported (Ref. 38) the results of a supplemental calculation with the Combustion Engineering energency core cooling system (ECCS) perform-ance code, STRIKIN-II.

The input gap conductance to this code was reduced until a volume-average initial fuel temperature higher than that calfulated by the staff was obtained.

The ECCS transient was then run at a peak linear heat rate of 15.5 kW/ft, the Cycle-6 Technical Specification limit. The calculated peak cladding temperature, peak local, and core-average cladding oxidation levels renained below the 10 CFR 50.46 acceptance limits. On this basis, we find the Cycle-6 Technical Specification limit on peak linear heat generation rate acceptable without further review of FATES-3.

With regard to the end-of-life rod pressure limit, we have concluded that this limit will be met for assembly-average burnups below approximately 38,000 MWD /MTU.

This value is based on our own audit calculation as well as one produced by the previous version of the FATES code (Ref. 35) using fission gas release burnup enhancement factor supplied by the staff (Ref. 39). Based on l

information contained in the Cycle-6 reload report, the end-of-cycle burnups are predicted to be below this limit for all assemblies in the Cycle 6 core except assembly 0042 in the center of the core. We have excepted this assembly from the burnup limitation because (1) only a single assembly in the core is involved and (2) the assembly-average powe.r density for this assembly is con-siderably less (<76%) than the core-average value and is, therefore, not expected to be limiting -for transient, accident, or fuel mechanical design analysis.

19

In conclusion, we have examined the fuel themal perfomance analyses submitted in support of the Calvert Cliffs, Unit 1 Cycle-6 reload and conclude that the application is acceptable without generic approval of methodology used (i.e.,

FATES-3).

2.2.6 Other Experimental or Demonstration Fuel Assemblies In addition to the PROTOTYPE assemblies, other non-standard fuel assemblies to be used in Cycle-6 include 3 assemblies that will contain Inconel tubes in the center guide tubes and 1 SCOUT demonstration assembly. We find the use of these 4 non-standard fuel assemblies acceptable since (a) they are few in number and constitute a small portion of the Cycle 6 core, (b) they are to be placed in non-limiting positions, (c) they have been evaluated and approved for irradiation in previous cycles, (d) they have undergone outage inspections that confimed the acceptability of their continued use, and/or (e) we have evaluated and approved specific analyses discussed previously herein (e.g., cladding creep collapse).

l l

20

2.3 NUCLEAR DESIGN EVALUATION Most of the nuclear design analyses used in the previous cycle (reference cycle) have been used for Cycle 6 in the same manner and with the same methods.

One exception is that the DIT assembly spectrum code (Ref. 40), which is based on integral transport theory, was used to generate neutron cross sections for both the ROCS and PDQ codes. Local power peaking is calculated with PDQ using few-group fine mesh cross sections from DIT multigroup transport theory calcul-ations.

Increased pin peaking near water holes is accounted for, as in previous cores employing 14x14 fuel assemblies, by imposing a bias factor derived from the difference between transport (DIT) and diffusion theory (PDQ) calculated local peaking.

ROCS and DIT are presently under review by the NRC staff and have been used in the Arkansas Nuclear One Unit 2 Cycle 2 reload analysis (Ref 41).

These codes use state-of-the-art techniques and provide agreement with measurements on reactivity and power distribution that' is substantially improved from previous methods. Therefore, pending our final review of these codes, we find them acceptable for use in the nuclear design of Cycle 6.

The ROCS computer code was used to calculate the following safety parameters:

l 1

Fuel Tenperature Coefficients Moderator Temperature Coefficients Boron llorths Critical Boron Concentrations Scram Reactivity Llorths and Allowances Reactivity Worth of CEA Regulating Banks l

CEA Ejection and CEA Drop Reactivity Worths i

21

2.3.1 Nuclear Parameters The Cycle 6 burnup is expected to be between 13,200 MWD /MTV and 13,800 MWD /MTU, depending on the final Cycle 5 temination point.

The Cycle 6 core char-acteristics were calculated for Cycle 5 tenninations between 12,000 and 13,000 ffWD/f1TU and the loading pattern presented is applicable to any Cycle 5 temination point within this band. We find the core characteristics reason-able and acceptable.

In the Cycla 3 Safety Evaluation Report, we found that the incorporation of stainless steel sleeves into the CEA guide tube had minimal effect on reactor physics. The operation of the Calvert Cliffs Unit 1 for the previous three cycles with these sleeves has borne out this conclusion.

The Cycle 6 noderator temperature coefficient is calculated to be -0.2 x 10-4

~

c.p/*Fforbeginningofcycleand-2.1x10 ap /*F for end of cycle.

These values are bounded by the values used in the safety analyses for the reference cycle (-2.5 x 10-4 to +0.5 x 10~4).

The Doppler coefficient for Cycle 6 is a best estimate value expected to be accurate to within 15%.

In order to assure that a conservative value was used in the safety analysis, a value 15% greater or:eless than this was used, depending upon whether a more negative or a less negative coefficient was conservative. We find the values of the moderator temperature coefficients and Doppler coefficients to be acceptable.

The zero power steam line break accident occuring at end-of-cycle is the most l

limiting and provides the basis for establishing the Technical Specification required shutdown margin which for Cycle 6 is 5.3%4f. At the end of cycle 6, the calculated hot zero power reactivity worth of all CEAs inserted assuming the highest worth CEA is stuck out of the core is 7.6%sp.

The CEA bite, which accounts for the possibility of the CEAs being slightly inserted rather than fully withdrawn, reduces the worth by an additional 1.7%ap, resulting in a calculated scram worth of 5.9%4f.

Assuming a 10% calculational uncertainty, 22

the net available calculated hot zero power scran worth at end of cycle is 5.3%aj.

Since this is equal to the Technical Specification shutdown margin a

(and includes a 10% uncertainty for the physics calculations) the shutdown margin is acceptable.

The limiting parameters of dropped CEA reactivity worth and maximun increase in radial peaking factor and the augmentation factors (used to account for the power density spikes due to axial gaps caused by fuel densification) for Cycle 6 are identical to the values used in the previous cycle and are, there-fore, acceptable.

Incore detector measurements are used to compute the core peaking factors l

using the INCA code (Ref. 42).

The coefficients required to perform this data reduction are obtained using the methodology described in the topical report.

For Cycle 6 operation, the power distribution measurement uncertainties used will be 6% for the total integrated radial peaking factor (F ) and 7% for the r

total power peaking factor (F ).

These are identical to those approved and q

applied in the Reference Cycle and are, therefore, acceptable.

1 l

l l

I 23 l

3.0 SAFETY ANALYSIS We have reviewed the response of Calvert Cliffs Unit 1 to previously reviewed accidents and transients for Cycle 6 operation.

The results of our review are presented herein.

3.1 CEA Withdrawal Event The CEA withdrawal event was reanalyzed for Cycle 6 to detennine the initial margins that must be maintained such that the DNBR and fuel centerline ~to melt (CTM) design limits will not be exceeded.

NRC approval of the CEAW topical report (Ref. 43) now allows the CEA withdrawal event to be classified as one for which the acceptable DNBR and CTM limits are not violated by virtue of sufficient initial steady state thermal margin provided by the DNBR and Linear Heat Rate (LHR) related Limiting Conditions for Operation (LCO).

Reliance is also now placed on the Variable High Power Level Trip or the Axial Flux Offset Trip rather than on the Thermal Margin / Low Pressure Trip.

The event was reanalyzed for reactor initial conditions of zero power and full power. The methods used to detennine the peak fuel rod response and the input to that analysis such as reactivity insertion rate, moderator and fuel temperature feedback effects, and initial axial power distribution, have been examined. The results of the analysis show that the 'NB and CTM design limits D

will not be exceeded during a CEA withdrawal event.

The staff concludes that the calculations contain sufficient conservatism, in both assumptions and models, to assure that fuel danage will not result from CEA withdrawal transients.

3.2 Full Length CEA Drop Event The full length CEA drop event was reanalyzed for Cycle 6 to determine the initial thermal margins that must be maintained by the Liniting Conditions for Operation such that the DNBR and fuel centerline melt design limit will 24

not be exceeded.

The methods used to detennine the peak fuel rod response and the input to that analysis such as power distribution changes, CEA reactivities and reactivity feedback effects due to moderator and fuel temperature changes, have been examined.

The resulting extreme conditions of fuel power, temperature, and DNB have been compared to the acceptance criteria for fuel integrity and the analyses have shown that these limits are not exceeded.

The staff concludes that the calculations contain sufficient conservatism, in both input assumption and models, to assure that fuel damage will not result from a full length CEA drop.

3.3 Fuel fiisloadino Event The analysis of the fuel misloading event.was perfomed in two steps. The first step was to detemine which fuel loading errors would be detectable by symmetry checks.

The second step was the evaluation of the consequences of normal operation with an undetectable fuel loading error.

The most adverse loading error which would be undetectable is the interchange of a fresh shimmed assembly with a once-burned assembly.

The maximum radial pin peaking factor in this case was 10% above the Technical Specification limit including appropriate uncertainties.

Since the limiting conditions for operation provide 17% margin on DNB and 35% margin on peak linear heat generation rate, this increase in radial peaking above the Technical Specification limit does not cause the fuel safety limits to be exceeded.

l The staff has evaluated the consequences of a spectrum of postulated fuel l

loading errors. We conclude that the analyses provided by thc licensee have shown for each case considered that either the error is detectable by the available instru' mentation (and hence remediable) or the error is undetectable l

I but there are no adverse offsite consequences since there is no fuel damage.

The licensee affims that the available instrumentation will be used during the hot zero power testing prior to Cycle 6 startup to search for fuel loading errors.

25 l

3.4 CEA Ejection Event The CEA ejection event was reanalyzed for Cycle 6 to determine that the NRC peak enthalpy limiting criterion of 280 cal /gm is not exceeded and to detennine the number of fuel pins that experience DNB. The analytical method used in the reanalysis is consistent with the reference cycle analysis except that CETOP/CE-1 with a DNB limit of 1.23 was used instead of TORC /CE-1 to calculate DNBR.

The most limiting key safety parameters in Cycle 6 were used to bound the most adverse conditions.

These included the least negative Doppler coefficient, the most positive moderator temperature coefficient, and an end of cycle delayed neutron fraction to produce the highest power rise during the event.

The licensee's analysis shows that 11.0% of the fuel pins experience DNB for the ejection from full power and 6.3% experience DNB for ejection from zero power.

The analysis also shows that both the zero power and full power cases result in peak fuel enthalpies less than the NRC limiting criterion of 280 cal /gm.

Therefore, prompt fuel rupture with consequent rapid heat transfer to was. assumed not to occur.

the coolant from finely dispersed molten UO2 We conclude that the initial assumptions and analytical models used ensure that primary system integrity will be maintained in the event of a CEA ejection.

The rod ejection accident was reanalyzed using the staff's evaluation of the l

clad failures (10% versus 11% estimated by BG&E). Further, the staff reviewed the assumption of 10% of the noble gas and radioiodines j

assumed to be in the gaps of rods that suffer clad failure. BG&E stated i

that the increased burnup fuel assemblies were in non-limiting locations and the staff agrees, estimating (from data presented in the cycle reload documen-tation) that the limiting location and time is in first cycle fuel from the middle to the end of the cycle. A model based on the ANS 5.4 proposed standard for gas release from LWR fuel was used to account for linear heat generation rates (LHGR) and burnup. This model is a "best estimate" model and suitable conservatism was retained t'y using the Calvert Cliff's TS limit LHGR,15.5 26 l

= _ _ _.

i kw/ft. For fuel at the end of the first cycle, the model predicts release I

of about 17% of the I-131, and about 10% for 1-133. Since these two make up the majority of the dose equivalent I-131, the average factor of 13.5% was used for the iodine gap release.

It was assumed that it would take 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> to isolate the release through the secondary side (see the discussion under the steam generator tube rupture accident). Other pertinent assumptions are given in Table 3-1 and the doses, both at the Exclusion Area Boundary (EAB) and Low Population Zone (LPZ), in Table 3-2.

The thyroid dose at the LPZ (79 Rem) is marginally above the Standard Review Plan (SRP) 15.4.8 dose guideline (75 Rem) and, therefore, the design of the plant for mitigating the consequences of the accident is acceptable.

TABLE 3-1 ASSUMPTIONS USED IN DESIGN BASIS ANALYSES Meteorological conditions: Duration and X/Q (sec/cu meter)~

Exclusion Area Boundary 0-2 hr 0.00033 Low Populatior Zone 0-8 hr 0.00006 8-24 hr 0.000042 i

24-96 hr 0.000019 96-720 hr 0.000006 Power Level 2700 MW J

I.

Control Rod Ejection Accident Fraction of clad failed 0.1 f

Gap activity 13.5% of iodines l

Fraction of fuel melted 0.0 Peaking factor 1.0 Containment leak rate 0.2%/ day for 1 day 0.1%/ day thereafter Primary coolant volume 57,000 gal.

Primary / secondary leak rate 1.0 gal / min Steam generator decontamination factor 10 Secondary side emission duration 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> Cantainment leakage auration 30 days 27

TABLE 3-1 (Continued)

!!. Fuel Handling Accident Clad failure 1 module (of 217)

Peaking factor 2.5 Gap activity 17% of iodines Time after shutdown 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> Pool decontamination factor 100 Filter efficiency for elemental iodine 90%

Filter efficiency for organic iodine 70%

III.

Steam Generator Tube Rupture Accident (Simplified Calculation)

Frimary/ secondary leakage 200,000 lbs Flashing fraction 10% (D.F. = 1.0)

Steam generator decontamination factor 10 Average coolant activity (Case 1) 30 micro Ci/gm dose equivalent I-131 Duration of secondary side emissions (Case 1) 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br /> Coolant activity (Case 2) 60 micro Ci/gm dose equivalent I-131 28

-,.-----n.

-, -.. ~. -,.

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

TABLE 3-2 Thyroid Doses * (Rem) from Design Basis Accidents Exclusion Area Low Population Boundary Zone I.

Control Rod Ejection 1.

Containment leakage 46 79 2.

Secondary Side Emission 59 11 II. Fuel Handling Accident 28 5

III. Steam Generator Tube Rupture 1.

From equilibrium coolant activity limit 10 2.

Pre-existing iodine spike 40

• All whole body doses are below 1 Rem.

29

3.5 BORON DILUTION EVENT The boron dilution event for the Calvert Cliffs plant is still under review by the staff with regard to the need for additional prctective instrumentation. The consequences of the boron dilution event, however, are not expected to be more serious than previously evaluated and are therefore acceptable.

3.6 STARTUP 0F AN INACTIVE REACTOR COOLANT PUMP EVENT (" Cold Water Accident")

A " cold water" accident at the Calvert Cliffs Nuclear Power Plant is precluded by Technical Specifications. TS 3.4.1.1 states that during Modes 1 and 2 both reactor coolant loops and both reactor coolant pumps in each loop shall be in operation.

Since these are the only modes for which Keff can equal or exceed 0.99, any segregated cold water will be dispersed throughout the reactor coolant system prior to taking the reactor critical. Therefore, the positive reactivity associated with segregated cold water cannot cause a reactivity excursion as long as this TS is complied with. We find this approach acceptable.

3.7 LOSS OF LOAD EVENT The Loss of Load (LOL) event is an undercooling transient that results from the sudden closure of the turbine stop valves without a simultaneous reactor trip.

The methods used by the licensee to analyze this event are consistent with those used for Cycle 5 which was used as the reference cycle for this event (Reference 44), except that CETOP/CE-1 computer code was used instead of TORC /CE-1 to calculate the DNBR. This code was previously reviewed by the staff and was found acceptable (Reference 45).

Conservative assumptions were used in the Loss of Load transient analysis as shown in the following:

a) To maximize the RCS presture during the transient, the steam dump and bypass valves were assumed not to be operable; also, the pressurizer spray and relief valves were assumed to be closed.

b) To maximize the rate of change of heat flux and the pressure at the time of reactor trip, the moderator temperature coefficient (MTC) was assumed to l

30 l

l l

be +.5x10-4Ao/oF. c) To minimize the negative reactivity inserted during the initial portion of the scram following a reactor trip and to maximize the time required to mitigate the pressure and heat flux increase an initial core average axial power distribution for this transient was assumed to be a bottom peaked shape.

The Loss of Load event resulted in a minimum transient DNBR of 1.38 and a peak reactor coolant pressure of 2550 psia. These values have safety margins of 12% on DNBR and 200 psia as compared to the limiting criteria of a DNB of 1.23 and 2750 psia, respectively.

The analysis results for this transient showed that the peak RCS pressure and the minimum DNBR do not exceed their respective design limits. This meets the acceptance criteria of SRP Section 15.2.1, and since the LOL event is limiting at B0C then extended burnup has no adverse impact during Cycle 6 operation. We find this analysis acceptable.

3.8 EXCESS LOAD EVENT The excess load event is the res0lt of any rapid increase in steam gener-ator steam flow other than a steam line rupture. The event is an over-cooling transient caused either by rapid opening of the turbine admission valves at power or the opening of all steam dump valves and bypass valves at full power or hot standby conditions. Such rapid increases in steam flow would result in a power mismatch between core power and steam generator load demand. Consequently, a decrease in reactor coolant tem-perature and pressure follows. Under such conditions a negative moderator temperature coefficient of reactivity causes an increase in core power.

Cycle 5 was also used as the reference cycle for this event (Reference 44).

The excess load event was analyzed for full power and hot standby condi-tions with the CETOP computer code (Ref. 45). The most limiting load increase transients at full power and hot standby conditions are due to the complete opening of the steam dump and bypass valves (Ref 46).

In the analysis, the licensee used conservative assumptions to account for (a) auxiliary feedwater flow rate, (b) End of Cycle Moderator Temperature Coefficient (c) minimum CEA worth and (d) Beginning of Cycle Fuel Tempera-ture Coefficient.

31

The results of the analysis show that for a full power excess load event the DNBR is 1.48 compared to the design limit of 1.23.

The maximum local linear heat generation rate for the event is 16.1 Kw/ft compared to the steady state value of 2.3 Kw/ft. The minimum DNBR calculated for the zero '

power excess load (hot standby) event is 2.92, and the linear heat genera-tion rate is 14.4 Kw/ft. Because of the large safety margins that exist between the above values, we conclude that the analysis results for this transient are acceptable and meet the acceptance criteria of SRP Section 15.1.1.

3.9 LOSS OF FEE 0 WATER FLOW EVENT A loss of feedwater flow event could be caused by main feed pump failure, feed control valve malfunction or loss of offsite power.

Loss of feedwater flow would result in decrease in steam generator water level, increase in primary system pressure and temperature, and reduction in the secondary system capability to remove the heat generator in the reactor core.

The event is a heatup transient and is more limiting at BOC. The last detailed analyses for this event was performed for Cycle 2 reload which was used as the reference cycle for this event (Reference 47). The licensee indicated that the reference cycle analysis for the loss of feedwater flow event was re-evaluated for Cycle 6 operation.

It is justifiable to do so because the key parameters used for the reference cycle analysis remain unchanged. The licensee further stated that the loss of,_edwater flow event would result in a less severe transient than the loss of load (LOL) event. Therefore, this event is bounded by the LOL transient. The conclu-sions reached for the LOL event are hence applicable for the loss of feed-water flow event, and thus we find them acceptable.

3.10 EXCESS HEAT REMOVAL EVENT Cycle 2 was used as the reference cycle for this event (Ref. 47). The excess heat removal event could be caused by decrease in feedwater temperature, excess feedwater flow, or excess steam flow.

Decrease in feedwater temperature because of the loss of high pressure feedwater heatup is the most adverse event in terms of cooling effects on the RCS.

This event is similar to the excess load event in that it is more limiting at E0C.

It also has the same effect on the primary system as a small increase, approximately 9%, in turbine demand, which is not 32

matched by an increase in core power. Hence. the DNBR degradation asso-ciated with this event is less severe than fcc the excess load event where a longer effective increase in turbine c'emand, i.e., 45%, is analyzed.

The excess heat removal event is therefore bounded by the excess load event. The conclusions reached for the excess load event are therefore applicable for this event, and thus we find them acceptable.

3.11 REACTOR COOLANT SYSTEM DEPRESSURIZATION EVENT The reactor coolant system depressurization event is postulated to occur due to an inadvertent opening of both pressurizer relief valves while operating at full power. Rapid depressurization while at full power causes a corresponding rapid decrease in DNBR.

The analytical method used in the reanalysis of this event is consistent with Cycle 4 which is used as the reference cycle for this event (Reference 48).

The Thermal Margin / Low Pressure (TM/LP) trip provides protection to pre-vent the DNBR SAFDL from being exceeded during the transient.

Conservative assumptions were used to maximize the rate of pressure de-crease and consequently, the fastest approach to DNBR SAFDLs. These i

l assumptions include:

(a) bottom peaked initial axial power shape; this power distribution maximizes the time required to terminate the decrease in DNBR following a trip.

(b) It was also assumed that the charging pumps, the pressurizer heaters and the pressurizer backup heaters were inoperable; this maximizes the rate of pressure decrease and, therefore, the rate of approach to DNBR SAFDL.

f The key transient parameters for this event as seen from the above discussion are independent of burnup and hence extended burnup has no impact on this event. None of the key transient parameters to determine the pressure bias factor for this event are outside the range of the reference cycle analysis (Cycle 4, Reference 48).

Hence, the'results and conclusiens reached in the reference cycle analyses are applicable for Cycle 6, and thus we find them acceptable.

33

3.12 LOSS OF COOLANT FLOW (LOCF) EVENT The loss of coolant flow event was reanalyzed for Cycle 6 to determine the minimum initial margin that must be maintained by the limiting conditions for operation (LCOs) such that in conjunction with the reactor protection system low flow trip, the DNBR limit will not be exceeded. The methods used to analyze this event are consistent with those discussed in Reference 7 which were found acceptable. The only difference from Reference 7 is that the CETOP/CE-1 computer code (Reference 45) was used instead of TORC /CE-1 to calculate the DNBR.

The 4-pump LOCF produces a rapid approach to the DNBR limit due to the rapid decrease in the core coolant flow. Automatic reactor trip on low reactor coolant flow and initial steady state thermal. margin provide j

protection against exceeding the DNBR limit.

The moderator temperature coefficient (MTC) and the fuel temperature i

coefficient (FTC) are the only key parameters which are impacted by extended burnup. Since this transient is more limiting at B0C, corre-sponding MTC and FTC values were assumed in the analysis. Hence, extended burnup has no adverse impact on this event.

The analysis assumed a loss of flow to the four reactor coolant pumps at a 0.0 axial shape index. An initial value of 0.0 axial shape index was chosen since it results-in a lower initial steady state DNBR and a more conservative LOCF type of event as was proven in the Cycle 5 analysis (Reference 44) which is considered as the reference cycle for this transient.

l The analysis of this transient resulted in a minimum DNBR of 1.23 and an RCS pressure of 2308 psia as compared to the safety criteria cf 1.23 and 2750 psia respectively. We conclude that the results of this analysis are acceptable and meet the acceptance criteria of the SRP Section 15.3.1.

3.13 ANTICIPATED OPERATIONAL OCCURRENCES RESULTING FROM THE MALFUNCTION OF ONE STEAM. GENERATOR The transients resulting from the malfunction of one steam generator were analyzed for Cycle 6 to ensure the DNBR and fuel centerline temperature design limits are not exceeded.

34

L I

The methods used to analyze these events are consistent with those reported in Cycle 5 which was used as the reference cycle (Reference 44), except that CETOP/CE-1 was used instead of TORC /CE-1 to calculate the DNBR.

The four events which affect a single generator are:

loss of load to one steam generator (SG); excess load to one SG; loss of feedwater to one-SG; or, excess feedwater to one SG. Of these four events, it has been detenr.ir.ed that the loss of load to one steam generator (LL/ISG) transient is the limiting single SG event. The event is initiated by.the inadvertent closure of a single main steam isolation valve.

The loss of load to the single steam generator increases its pressure and temperature to the opening pressure of the secondary safety valves. The intact steam generator temperature and pressure decrease due to the loss of load in the other steam generator. The cold leg asymmetry causes an inlet temperature tilt which results in an azimuthal power tilt, an increase in the peak linear heat generation rate (PLHGR), and a degraded DNBR.

The moderator temperature coefficient (MTC) is the only key parameter which is adversely impacted by extended burnup. The analysis assumed an E0C MTC of -2.5x10-4 ap/0F which is more conservative than the TS limit of 2.2x10-4 ap/0F; therefore, we conclude that the effects of extended I

burnup have been conservatively included in the analysis.

l The minimum transient DNBR calculated for the loss of load to one steam generator event is 1.43 which is conservative compared to the minimum t

acceptable DNBR of 1.23. The linear heat generation rate of 21.3 Kw/ft was not exceeded. We conclude that the rnEjlts of the analysis are

' 9its are not exceeded, no fuel acceptable since the DNBR and CTM destin pins are predicted to f ail, and r At me-ernup has no adverse impact during this event.

3.14 LOSS OF ALL NON-EMERGENCY AC POWER EVENT The loss of all non-emergency A-C power incident (LOAC) r65ults in un-l availability of electric power to the reactor coolant pumps and the main circulating water pumps. Under such circumstances, the plant would expe-rience a simultaneous loss of load, loss of feedwater flow, and loss of forced reactor coolant flow.

The LGAC is followed by an automatic reactor trip.

In the absence of forced reactor coolant flow, convective heat transfer through the core is maintained by natural circulation. Following the automatic startup of the emergency diesel generators, the auxiliary feedwater is manually initiated and plant cooldown is controlled via remotely-operated atmos-pheric steam dump valves.

Cycle 2 (Reference 47) was used as the reference cycle to re-evaluate the LOAC for Cycle 6 operation at extended burnup to determine that the DNBR design limit is not exceeded and to verify that the site boundary doses are within those reported in Cycle 2 analysis.

For the first few seconds of the transient, the loss of all non-emergency AC power behaves like a' loss of flow event. Therefore, the transient minimum DNBR of 1.23 that was calculated for the loss of flow event is applicable for this event and thus is found acceptable.

For the remainder of the transient, the DNBR remains within required limits.

3.15 STEAM LINE RUPTURE EVENT The licensee has reanalyzed the event for cycle 6 to verify that the critical heat flux is not exceeded during the event.

The analysis assumed that the event is initiated by a circumferential cupture of a 34-inch steam line at the steam generator nozzle. This break i

is the most limiting, since it causes the greatest rate of cooldown of the reactor coolant. With a negative moderator coefficient of reactivity, a

the cooldown will produce a positive reactivity addition.

Following a steam line rupture, reactor trip is initiated by low steam generator pressure. The analysis assumed that the auxiliary feedwater is initiated in three minutes from the initiation of a low steam pressure trip, an MSIV closure time of 12 seconds and a manual trip of the reactor coolant pumps on safety injection actuation signal due to low pressurizer pressure.

In addition, the analysis conservatively assumed that all the auxiliary feedwater flow is fed only to the steam generator associated with the steam line rupture and the control element assembly is stuck in the fully withdrawn position which yields the most severe combination of scram worth and reactivity insertion.

36

+

l t

The results of the analyses for both the 2754 MWt (102% rated power) and no-load cases show the minimum DNBRs stay above 1.3 and the reactor coolant system pressures stay below the initial RCS pressure which is below 110% of the design pressure and are therefore acceptable.

A C-E report, CEN-199, " Effects of Vessel Head Voiding During Transients and Accidents in C-E NSSS's" (Ref. 49), has been submitted to NRC for staff review. The report indicates that the impact of void formation in the reactor vessel upper head region upon the post-trip return to power can be significant for steam line breaks.

Implementation of any modifica-tions to Calvert Cliffs Unit I will be determined by the results of NRC staff review of the CE report, CEN-199.

3.16 STEAM GENERATOR TUBE RUPTURE (SGTR) EVENT The licensee has re-evaluated the event for cycle 6 operation at extended burnup to verify that the radiological consequences of the event are within those reported in cycle 5 analysis. The licensee stated that of these key parameters which determine the site boundary doses, only primary and secondary coolant activity is, in principle, burnup dependent. Since the TS limits on primary and secondary activities will remain at the cycle 5 values and since none of the other key transient parameters have changed, the licensee concluded that the results and conclusions reported for cycle 5 analysis are valid for cycle 6 operation at extended burnup. We find these results acceptable.

The licensee's cycle 6 evaluation of the steam generator tube rupture concluded that no change to the analysis of record was warranted specifically for cycle 6 since parameters of importance were not changed. However, for several reasons the staff performed a highly simplified calculation of the SGTR. The staff's estimate of the consequences of SGTR accidents indicate that it would not be possible to isolate emissions from the secondary side within 30 minutes as assumed by the licensee. Since no final evaluation is available to assess the actual time, the staff assumed that the secondary side releases continued for the full time over which exclusion area boundary (EAB) doses are evaluated (2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />).

During this time, more primary / secondary tube 37

leakage would occur and the iodine released to the coolant as a result of the spike assumed to be caused by the SGTR would raise the primary coolant concentration. The amount of coolant leakage was assumed to be 200,000 lbs. This results from two factors:

an evaluation by C-E (which has not been reviewed or accepted by BG&E) indicating that inclusion of vessel head voiding would increase the primary / secondary leakage and the secondary side emission, and the increased time of seco.Hary side emission.

The average concentration of the primary coolant leaked to the secondary side was assumed to be 30 micro-C1/gm dose equivalent I-131, based on evaluations of concentration versus time for other plants, i

The second case of SGTR reviewed by the staff was from the Technical Specification shutdown coolant concentration, 60 micro-Ci/gm. This simplified calculation used the same primary / secondary leakage as for the first case.

Pertinent assumptions are given in Table 3-1 and doses in Table 3-2.

Since the thyroid dose at the EAB is normally limiting, other doses were not evaluated in this simplified review. The EAB thyroid doses for both cases are within the Standard Review Plan 15.6.4 guidelines.

It should be noted that, as a result of the recent Ginna steam generator tube rupture accident, the staff's design basis assumptions are being reviewed. The review is being undertaken on a generic basis.

Reference 7 indicates that the integrated primary to secondary leakage in the first 30 minutes increased about 10% with void formation in the reactor vessel upper head region. This will lead to more severe radiological consequences for the SGTR event.

Implementation of any modifications to Calvert Cliffs Unit I will be determined by the results of NRC staff review i

of the CE report, CEN 199.

38

3.17 SEIZED RCP ROTOR EVENT The licensee has reanalyzed this event for cycle 6 to verify that the RCS peak pressure will not exceed 110% of the design pressure, and that only a small fraction of fuel pins are predicted to fail during this event. This event was analyzed to show that the site boundary doses are within the limits of 10 CFR 100.

The single RCP shaft seizure is postulated to occur as a consequence of a mechanical failure. The event results in a rapid reduction in the reactor coolant pump flow to the three-pump value. A reactor trip is i

initiated by a low coolant flow rate as determined by a reduction in the sum of the steam generator hot to cold leg pressure drops. The pressurizer pressure reached a maximum value of 2313 psia at 3.5 seconds. The licensee stated that no more than 3.0% of fuel pins are predicted to experience DNB.

The resultant site boundary doses are within 10 CFR 100 limits and are therefore acceptable.

3.18 LOSS OF COOLANT ACCIDENT (LOCA)

An ECCS performance analysis was performed for Calvert Cliffs Unit 1 cycle 6 to demonstrate compliance with 10 CFR 50.46. NRC approved models 4

and codes were used for the analysis; e.g., DES /PD model, and STRIKIN-Il code.

The FATES 3 model was also used.

4 The results of the evaluation confirm that 15.5 Kw/ft is an acceptable value for the allowable peak linear heat generation rate (PLHGR) in cycle 6.

It should be noted that this value was the same in cycle 5.

The overall results for cycle 6 are very similar to those predicted for cycle 5.

The most limiting case results in a peak clad temperature of i

20380F, which is well below the acceptance limit of 2200 F, The maximum 0

local and core wide values for zirconium oxidation percentages are below the acceptance limit values of 17%, and 1%, respectively. Therefore, operation of Unit 1 cycle 6 at a PLHGR of 15.5 Kw/ft and a power level of 2754 MWT (102% of 2700 MWT) results in compliance with the 10 CFR 50.46 acceptance criteria, and is acceptable.

i 39

3.19 FUEL HANDLING ACCIDENT The fuel handling accident was reanalyzed by the staff using the assump-tions in Table 3-1.

The doses for the EAB and LPZ are given in Table 3-2; all are well within the guidelines. The review was in accordance with Ragulatory Guide 1.25 except for the following. The peaking factor of 2.5 was used to scale the core average power to the Technical Specifica-tion LHGR of 15.5 kw/ft. Since the accident is assumed to take place 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> af ter shutdown, then the majority of the dose equivalent I-131 is I-131; 17% of the fuel assembly content of iodine is assumed to be re-leased from the gap. We find these results acceptable.

}

l l

l l

l 40

4.0 RADIOLOGICAL CONSEQUENCES OF ACCIDENTS FOR EXTENDED BURNUP In Reference 50, the licensee provided information regarding the use of extended fuel cycles for Calvert Cliffs Unit 1.

We have reviewed this information, and the assessment provided by DOE (Reference 51) to determine the potential environmental consequences of extended burnup for the fuel cycle for Calvert Cliffs Unit 1.

Increases in burnup (or fuel enrichment, though this is not an issue in this reload review) beyond the traditional range covered in the Regulatory Guides and Standard Review Plan could affect the radiological consequences of accidents by changes in the fuel failure rate, changes in the total inventory and mix of radioisotopes in the fuel, the fraction of isotopes accumulated in the fuel-clad gap, iodine spiking behavior, and the decontamination factor for fuel handling accidents due to in-creased gas pressure within the fuel rod.

Traditionally, the staff has considered radioactive noble gases and radiciodine in design basis accidents, the latter a surrogate for all other fission products (that is, non-noble gas fission products), due to its high volatility. However, for extended burnup, the continued suitability of this practice must be examined.

Cesium-137 is one such particulate; it is often considered to be released in conjunction with the radiciodines, and it is also volatile (though less so than iodine).

Preliminary calculations of the core content of cesium show that for extended burnups the core content is increased, while the iodine remains quite constant. Also the fraction of the rod content of CS-137 which is in the gap is higher than for I-131, but by a very small amount for the Calvert Cliffs TS limit of 15.5 kw/ft. Dose conversion factors for Cs-137 are about a factor of 20 (or better) less than the thyroid dose conversion factor for I-131. This balances the factor of about 10 in the 10 CFR 100 guideline doses for thyroid versus whole body.

For other radionuclides, volatilities are lower and margins are larger.

Therefore, we conclude that for this modest increase in burnup, iodine may still be used as a surrogate for non-noble gas nuclides.

I 41 l

a The conclusion in Reference 50 as well as Reference 51 is that there is no significant incremental adverse impact on the environment from the radiological effects of accidents.

This conclusion is based on considera-tion of Design Basis Accidents (DBA) by comparison with their safety analyses and severe accidents by comparison with the Reactor Safety Study, WASH-1400.

The NRC staff's evaluation of the radiological consequences of DBAs for Cycle 6 concludes that they are acceptable compared to 10 CFR 100 guidelines (see Sections 3.4, 3.16, 3.17 and 3.19, herein).

For severe accidents, the potential increase in radiological consequences due to an increase in long-lived nuclides is small in comparison to the uncertainties in the state-of-the-art of probabilistic risk assess-ment.

These uncertainty bounds could be well over a factor of 10, but are not likely to be so large as a factor of 100.

5.0 TECHNICAL SPECIFICATIONS Draft Technical Specifications are present'ed in the February 17, 1982 application and in the April 29, 1982 supplement. These changes are addressed herein.

5.1 Thermal Margin Safety Limits TS Figure 2.1-1, " Reactor Core Thermal Margin Safety Limit", has been changed to reflect higher radial peaking factors and implementation of the margin recovery programs.* Table 2-2, herein, shows the increase in radial peaking compared to the Cycle 5 (reference) analysis. The corresponding Basis in B2.1.1 has been changed to reflect an increase in the maximum steady state peak linear heat generation rate (centerline fuel melt) from 21 to 21.3 kw/ft. We find these changes acceptable.

• 0n July 17, 1980 the licensee met with the NRC staff to discuss the margin recovery program. This program involves'the use of SCU and other analytic techniques which were necessary to " recover" the decrease in operating margins that would have resulted from the generally less advantageous parameters (such as peaking factors) associated with the extended burn-up fuel cycle, i

42 I

t 5.2 Peripheral Axial Shape Index TS Figure 2.2-1 has been changed to reflect the increase in maximum steady state peak linear heat generation rate (centerline fuel melt) as indicated above. We find this change acceptable.

1

'i.., !!ir.imum DNBR and Power Uncertainty The minimum DNBR reflected in TS Figure 3.2-4 and as stated in Basu B.2.1.1, B.2.2.1, and B 3/4.2.5 has been changed from 1.195 to 1.23 as a result of application of SCU.

In addition, the TM/LP trip description in Bases B.2.2.1 has been revised to reflect the SCU methodology. The NRC review of the SCU methodology is contained in Section 2.1.5 herein.

Another change associated with SCU involves the 2% power uncertainty I

which is no longer required. This change is reflected in Bases B.2.1.1 and B.2.2.1.

We find these changes accepta!le.

5.4 Use of Excore Detectors'for Linear Heat Rate M'onitoring A modification to the use of excore detectors used to monitor linear heat rate (LHR) has been made. This technique will avoid unnecessary power level changes resulting from temporary on-line computer outages.

The computer is required to interpret the in-core detectors which had previously been the sole means of monitoring the peak LHR to justify operation at full power. The following changes are associated with this l

use of excore detectors:

l (1) Figure 3.2-2 is changed. This change also reflects the margin j

recovery programs and the increase in radial peaking.

(2) TS 4.2.1.3 is revised to provide credit for the calculated value ofFx}whenmonitoringtheLHRLimitingConditionforOperation with the excore detectors.

(3) TS 3/4 2.2 has now been divided into 3/4 2.2.1 and 3/4 2.2.2 to reflect the option of using either incore or excore detectors to monitor LHR at full power. Figure 3.2-3 has been revised and re-numbered 3.2-3a and Figure 3.2-3b has been added. The above changes reflect the use of excore detectors to monitor LHR for short periods of time at full power and also an increase in radial peaking.

We find these changes acceptable.

43

5.5 Implementation of BASSS The Better Axial Shape Selection System (BASSS) uses the fixed rhodium incore detector system to monitor the Departure from Nucleate Boiling-Limiting Condition for Operation (DNB-LCO) rather than the excore de-tectors. Control Element Assembly (CEA) position and core average axial shape index are monitored to provide an alarm on power when the DNB-LC0 is exceeded.

Specifically, the BASSS determines allowable power level from knowledge of CEA position, total integrated radial peaking factor, and core average axial shape'index using the on-line computer code PSINCA. The BASSS provides an alarm on power if the measured power level exceeds this allowable level. The BASSS method-ology was previously approved for use at Calvert Cliffs (Reference 52).

The following TS changes involve imi>lementation of BASSS:

(1) Figure 3.1.2, the power dependent insertion limit (PDIL) is modified to indicate the BASSS operating region.

(2) TS 3.2.6, "DNB Parameters" and Table 3.2-1 are modified to require the axial shape index, core power,to be maintained via BASSS.

(3) TS 3.2.3 is modified to incorporate BASSS and reflect an increase in radial peaking.

We find the above changes acceptable.

5.6 Shutdown Margin The shutdown margin given in TS 3.1.1.1 is increased from 4.3 to 5.3%

AK/K. This change is' also reflected in the PDIL in Figure 3.1.2.

The change in the shutdown margin results from the end of cycle (E0C),

hot zero power (HZP), steam line break analysis (see Section 3.15).

Bases B 3/4 1.1.1 and 3/4 1.1.2 have been changed to be consistent with TS 3.1.1.1.

We find these changes acceptable.

i 5.7 MSIV Closure Time The main steam isolation valve (MSIV) response time has been increased from 6.9 to 12.9 seconds and the MSIV test closure time has been increased from 3.6 sec. to 4.0 sec.

The licensee explained in a telephone t

44

conversation that the reason for this change is plant availability advantages and the revised time was assumed in the licensee's new main steam line break accident analysis. The following changes in the TS relate to this issue:

(1) TS Table 3.3-5, " Engineered Safety Features Response Times" has been changed to increase the MSIV response time from 5 6.9 to 1

5 2.9 seconds.

(2) TS 4.7.1.5, " Main Steam Line Isolation Valves", has been changed to increase the MSIV test closure time from 3.6 to 4.0 seconds.

We find these changes to be acceptable as reflected in the safety analysis for Cycle 6.

5.8 RTO Response Time The resist'nce temperature detector (RTD) response time has been in-a creased from 1 8.0 to 1 12.0 seconds.

These response times are reflected in the following entries in Table 3.3-2, " Reactor Protective Instru-mentation Response Times":

(1) Power Level-High (2) Axial Flux Offset j

(3) Thermal Margin / Low Pressure We find these changes acceptable as reflected in the Cycle 6 analysis.

5.9 Fuel Enrichment The maximum enrichment for reload fuel, specified in TS 5.3.1, " Fuel Assemblies", has been increased from 3.7 to 4.1 weight percent U235 We find this change acceptable.

5.10 Pressure Transmitters By letter dated April 29, 1982 the licensee informed the NRC that certain pressure transmitters located inside containment had been replaced by environmentally qualified transmitters, manufactured by Barton, in order to satisfy NRC concerns on environmental qualifications of 45

electrical equipment. The following changes are made to the TS in order t

to insure that the safety analysis for Cycle 6 operation remains valid l

for the use of the Barton pressure transmitters:

(1) TS Table 2.2-1, " Reactor Protective Instrumentation Setpoint Limits" has been changed. The Steam Generator Pressure-Low Trip l

Setpoint has been changed from < 570 to < 635 psia to reflect the uncertainty associated with Barton pressure transmitters during the Main Steam Line Break Event.

In addition, the note (2) in Table 2.2-1, relating to the Steam Generator-Low Trip Bypass has been changed from bypass below 685 to bypass below 710 psia to reflect the change in the Trip Setpoint.

This same change was incorporated into Table 3.3-1, " Reactor Protective Instrumentation" and Table 3.3-1, " Engineered Safety Feature Actuation System Instrumentation".

(2) TS Bases B.2.2.1 is revised to describe the change in the Steam Generator Pressure-Low Trip Setpoint, addressed above.

In addition, the low pressure for the TM/LP Trip Setpoint has been changed from 1750 to 1875 psia to reflect the uncertainty associated with the Barton pressure transmitters during a LOCA.

(3)

TS Table 3.3-4, " Engineered Safety Feature Actuation System Instru-mentation Trip Values" has been changed.

The SIAS Pressurizer Pressure-Low Trip Setpoint has been changed from > 1578 to > 1725 psia to reflect the uncertainty associated with the Barton trans-mitters during a LOCA.

In addition, Table 3.3-3, the SIAS Pressure-Low Bypass, has been changed from < 1700 to < 1800 psia to ref. lect the change in the actuation setpoint.

1 (4) TS Table 3.3-4, the SGIS Setpoint, has been changed from > 570 to > 635 psia to reflect the uncertainty associated with Barton pressure transmitters during the Main Steam Line Break Event.

r We find the above changes to the TS to be acceptable and consistent with the Cycle 6 analyses.

46

6.0 CONCLUSION

The environmental conclusions presented in Section 6.1 are based upon the consideration of Cycle 6 operation of Calvert Cliffs Unit 1 and, specifically, the effects of the extended fuel cycle as presented in Section 4.0.

The safety conclusions presented in Section 6.? are based upon the NRC evaluation of accidents and transients present.,' il Section 3.0 and the proposed TS changes presented in Section # ?

6.1 Environmental Considerations We have determined that the amendment does not authorize a change in cffluent types or total amounts nor an increase in power level and will not result in any significant environmental impact.

Having made this determination, we have further concluded that the amendment involves an action which is insignificant from the standpoint of environmental impact and, pursuant to 10 CFR 651.5(d)(4), that an environmental impact statement or negative declaration and environ-mental impact appraisal need not be prepared in connection with the issuance of this amendment.

6.2 Conclusion We have concluded, based on the considerations discussed above, that:

(1) there is reasonable assurance that the health and safety of the public will not be endangered by operation in the proposed manner, and (2) such activities will be conducted in compliance with the Commission's regulations and the issuance of this amendment will not inimical to the common defense and security or to the health and safety of the public.

Date:

June 24, 1982 Principal Contributors:

L. Kopp i

G. Schwenk D. Powers J. Vogelwede A. Gill J. Mitchell D. Jaffe 47 l

l

].0 References 1.

Letter from A. E. LundvaH (BG&E) to R. A. Clark (NRC) with Calvert Cliffs Unit 1, Phase I Cycle 6 Reload Report submittal, November 19, 1981.

2.

Letter from A. E. Lundvall (BG&E) to R. A. Clark (NRC), "Calvert Cliffs Unit 1, Docket No. 50-317, Responses to Questions on Phase I Cycle 6 Reload Application," January 5,1982.

3.

CEN-191(B)-P, "CETOP-D Code Structure and Modeling Methods for Calvert Cliffs Units 1 and 2," December 1981.

4.

Supplement 3-P (Proprietary) to CENPD-225P, " Fuel and Poison Rod Bowing,"

June 1979.

5.

CEN-124(B)-P, " Statistical Combination of Uncertainties, Part 1," January 1980.

6.

CEN-124(B)-P, " Statistical Combination of Uncertainties, Part 2," January 1980.

7.

CEN-124(B)-P, " Statistical Combination of Uncertainties, Part 3," March 1980.

8.

CENPD-101-P, " TORC Code, A Computer Code for Detemining the Themal fiargin of a Reactor Core," dated July 1975.

9.

Letter from K. Kniel (flRC) to A. E. Scherer (CE), " Evaluation of Topical Report CENPD-161-P," dated September 14, 1976.

10.

CENPD-206-P, " TORC Code, Verification and Simplification fiethods,"

January 1977.

1 I

48 i

11. CENPD-162-P-A (Proprietary) and CENPD-162-A (non-Proprietary), " Critical Heat Flux Correlation for C-E Fuel Assemblies with Standard Spacer Grids Part 1, Unifom Axial Power Distribution," dated September 1976.

12.

CENPD-207-P, "C-E critical heat flux:

Critical Heat Flux Correlation for C-E Fuel Assemblies with Standard Spacer Grids Part 2, Nonunifom Axial Power Distribution," June 1976.

13.

D. S. Rowe, " COBRA III C: A Digital Computer Program for Steady State and Transient Thermal-Hydraulic Analysis of Rod Bundle Nuclear Fuel Elements," BNWL-1695, March 1973.

14.

Letter, D. C. Trimble (AP&L) to Director, NRR, "CETOP-D Code Structure and flodeling Methods, Response to First Round Questions on the Statistical Combination of Uncertainties Program (CEN-139( A)-P)," July 15, 1981.

15.

Final Safety Evaluation Report Supporting Facility Operating License Amendment No. 26 on Docket No. 50-368 and Operation of AN0-2 During Cycle 2, July 21,1981.

16.

Letter from G. ft. Hesson to S. C. Gupta, dated January 12, 1982.

i I

17.

Letter, D. F. Ross and D. G. Eisenhut (NRC) to D. B. Vassalo and K. R.

Goller (NRC), " Revised Interim Safety. Evaluation Report on the Effects of l

Fuel Rod Bowing in Thennal Margin Calculation for Light Water Reactors,"

February 16, 1977.

18.

CEN-123(B)-P, "C-E Response to NRC Second Round Questions on the Statistical Combination of Uncertainties Program," August 1981.

19.

Combustion Engineering, " Response to NRC First Round Questions on the Statistical Combination of Uncertainties Part 2 of CENPD-124(B)-P,"

l February 9, 1981.

49

20.

F. J. Berte, F. L. Filstein, and R. Goldstein, " Representative Sampling in Reactor Data Analysis," TIS-6532, Combustion Engineering, Windsor, Connecticut.

21.

Combustion Engineering, " Responses to Second Round Questions on Part 2 of CEN-124(B)-P", September 1981.

22.

Letter from G. fi. Hesson to H. Balukjian dated April 9,1981.

23.

CENPD-210-A, Combustion Engineering, " Quality Assurance Progran,"

Revision 3, dated November 1977.

24.

Cell-119(B)-P, "BASSS C-E Topical Report," dated November 1979.

25.

CENPD-199-P, "C-E Setpoint tiethodology," dated April 1976.

i 26.

"CEPAN fiethod of Analyzing Creep Collapse of Oval Cladding," CE Report CENPD-187-A, March 1976, 27.

D. A. Powers and R. O. fieyer, " Cladding Swelling and Rupture flodels for LOCA Analysis," NRC Report NUREG-0630, April 1980.

28.

"Zircaloy Growth:

In-Reactor Dimensional Changes in Zircaloy-4 Fuel Assemblies," CE Report CENPD-198, December 1975.

29.

"Zircaloy Growth: Application of Zircaloy Irradiation Growth Correlations for the Calculations of Fuel Assembly and Fuel Rod Growth Allowances,"

CE Report CENPD-198, Supplement 1, December 1977.

30.

" Response to Request for Additional Information on CENPD-198-P, Supplement 1,"

CE Report CENPD-198, Supplement 2-P, November 1,1978.

31.

" Application of CENPD-198 to Zircaloy Component Dimensional Changes,"

CE Report CEN-183(B)-P, September 1981.

50 l

32.

W. J. Bailey, et al., " Fuel Perfomance Annual Report for 1980," NRC Report NUREG/CR-2410, December 1981.

33. Millstone-2, Cycle-3 Reload SER.

34.

" Improvement to Fuel Evaluation Model," Combustion Engineering Topical Report CEN-161(B)-P, July 1981.

j 35.

" Fuel Evaluation Model," Combustion Engineering Topical Report CENPD-139-P-A, July 1974.

36.

" Partial Response to NRC Questions on CEN-161(B)-P, Improvements to Fuel Evaluation Model," Combustion Engineering Topical Report CEN-193(B)-P, Supplement 2-P, March 21, 1982.

37.

C. E. Beyer et al., "GAPCON-THERMAL-2: A Computer Program for Calculating the Themal Behavior of an Oxide Fuel Rod,"

Battelle Pacific Northwest Laboratories Report BNWL-1898, November 1975.

38.

A. E. Lundvall (BG&E) letter to R. A. Clark (NRC) dated May 19, 1982.

39.

R. O. Meyer et al, " Fission Gas Release from Fuel at High Burnup,"

U.S. Nuclear Regulatory Commission Report NUREG-0418, March 1978.

l f

40.

CENPD-266-P, "The ROCS and DIT Computer Codes for Nuclear Design,"

l December 1981.

41. Letter from D. C. Trimble to R. A. Clark, " Arkansas Nuclear One Unit 2 cycle 2 Reload Report," February 20, 1981.

i

42. CENPD-145, " INCA: Method of Analyzing In-Core Detector Data in l

Power Reactor," April 1975.

1

43. CEN-121(B)-P, "CEAW, Method of Analyzing Sequential Control Element Assembly Group Withdrawal Event for Analog Protected Systems,"

November 1979.

51

44.

Letter from A. E. Lundvall to R. A. Clark, Supplement 1 to 5th cycle license application, dated November 4, 1980.

45.

Letter from R. A. Clark to A. E. Lundvall dated March 15, 1982.

46. Letter from A. E. Lundvall to R. A. Clark, dated February 17, 1982.

]

47.

Letter from A. E. Lundvall to B. C. Rusche, second cycle license applica-tion, dated October 1, 1976.

48.

Letter from A. E. Lundvall, Jr., to R. W. Reid, 4th cycle license applica-tion, dated February 23, 1979.

49. CEN-199, CE report titled " Effects of Vessel Head Voiding Ouring Tran-

+

sients and Accident in C-E NSSS's, dated March, 1982.

50. CEN-122(B)-NP and P, " Environmental Impact of Extended Burnup Fuel i

Cycles in Calvert Cliffs, Units 1 and 2," March 23, 1981.

i 51.

00E/EA-0118, " Environmental Assessment-D0E Program to Improve Uranium Utilization in Light Water Reactors," August 1980.

I 52.

Letter R. A. Clark to A. E. Lundvall dated May 20, 1981.

52

APPENDIX A Safety Evaluation of CEN-124(B)-P, Parts 1, 2, and 3

" Statistical Canbination of Uncertainties" (SCU)

The licensee has defined the input data required for a detailed thermal-hydraulic analysis by type:

(1) system parameters which describe the physical system and are not monitored during reactor operation and (2) state parameters, which describe the operational state of the reactor and are monitored during operation.

There is a degree of uncertainty in the value used for each of the input para-meters used in the design safety analyses.

This uncertainty has been handled in the past by assuming that each variable affecting DNB is at its extreme most adverse limit of its uncertainty range.

The assumption that all factors are simultaneously at their most adverse values leads to conservative restrictions in reactor operation. The licensee has proposed in three parts of the CEN-124(B)-P (Refs. 5, 6, and 7) a new methodology to statistically combine uncertainties in the calculation of new limits for Calvert Cliffs-1. These limits will ensure with at least 95 percent probability and 95 percent con-fidence level that neither DNB nor fuel centerline melt will occur. Part I describes the application of the SCU to the development of the local power density (LPD) and thermal margin / low pressure (TM/LP)' limiting safety system f

settings (LSSSs).

These are used in the analog reactor protection system to protect against fuel centerline melt and DNB, respectively.

Part 2 uses SCU methods to develop a new DNB limit.

Part 3 uses SCU methods to define limiting conditions for operations (LCOs).

A.1 PART ONE Part 1 of the report (Ref. 5) defines the methods used to statistically combine uncertainties applicable to the LSSSs and evaluates the aggregate of these uncertainties as they determine the reactor protection against DNB and fuel centerline melt.

The report further defines those uncertainties that have to be considered and evaluates their probability distributions.

53

A.1.1 Thermal-Hydraulic Summary and Evaluation of Part 1 The metnods by which the licensee determines the setpoints in the Calvert Cliffs-1 reactor protection system are given in CENPD-199-P (Ref. 25). The statistical combination of variables does not alter these methods. The same variables are considered and, once the uncertainties have been identified, statistically canbined, and applied to the setpoint variables, the development of the setpoints proceeds as has been done in the past to develop the LSSSs.

Basically, ordered pairs of values of the peripheral axial shape index and the power to the specified fuel design limit are plotted. A lower bound is drawn under the " flyspeck" data such that all the core power distributions analyzed are accommodated.

This in itself retains much of the conservatism of the past practices, since all of the data points lie above the lower bound and must lie well above. The lower bound is then reduced by uncertainties derived from the statistical combination and the generation of the trips proceeds much as has been the past practice.

The variables considered in the LSSS determination are listed in Table 3-1 of Part 1 of the report (Ref. 5) together with values of their uncertainties.

There are errors in Table 3-1 of the report (Ref. 5).

The values for the primary coolant mass flow and the power distribution monitoring system processing uncer-tainties are not the most recent values.

Corrected values have been supplied (Ref. 18).

Subsequent reloads will require that the corrected values provided in Reference 18 will be used in future calculations.

The bases of the uncertainty values of Table 3-1-are given in Appendix A of Reference 5.

More infonnation (Ref.18) has been provided in response to a request for more detailed justification.

The source and magnitude of the uncertainty estimates were reviewed and found to be acceptable. The method of combining the various uncertainties of a single variable will produce valid estimates of the total.

The calculations were spot-checked and found to be Correct.

l l

54

A.1.2 Statistical Summary and Evaluation of Part 1 Uncertainties associated with DNB and LPD limiting system safety settings are combined statistically. A stochastic simulation technique is used to estimate the probability distribution function (pdf) of DNB overpower (p/fdn) and power to fuel design limit on linear heat rate (P/fdf) for a specific axial power distribution.

The simulations are carried out for a number of axial power distributions characterized by peaking factors and nomalized axial shapes.

For each axial shape, the pdf's of P/fdn and P/fd/ are estimated.

For each pdf the ratio of the mean value to the lower 95/95 probability / confidence limit is computed.

The statistically combined uncertainy is taken as the maximum ra'tio over all axial shapes used.

Evaluation of the statistical validity of the uncertainty combination methodolgy requires examination of the following points:

1.

Sampling liethod design of the simulation experiment number of samples (simulation runs) random number generator 2.

Uncertainty distributions of independent variables l.

distribution fom, e.g., Gaussian, unifom l

statistical analysis method These points will be discussed in order.

1.

Sampling Method For the TM/LP LSSS the input parameters subject to uncertainty are:

primary coolant inlet temperature pressurizer pressure primary coolant flow l

55

T/ flux power radial peaking factor ASI correction tems The simulation is carried out by selecting a peripheral axial shape index and the corresponding axial shape.

For the selected axial shape at least 500 simu-lation trials are carried out, with each trial using one sampled value from each input parameter distribution.

The sampling is carried out using the SIGMA code and a Latin Hypercube Sampling (LHS) design.

The LHS design with 500 trials will produce acceptable estimates of the distribution of P/fdn.

The SIGMA code is described in Section 4.4.1.1 and CE's response to the first round questions (Ref.19).

The sample generation procedures depart somewhat from standard statistical practice.

For example, the sample mean from a Gaussian distribution when the standard deviation is estimated from the same sample follows a student's t-distribution.

SIGMA handles this by sampling a 2

variance from a X distribution and then sampling from a Gaussian distributio..

using the sampled variance.

As a second example, SIGMA generates nomal deviates using an approximation to the inverse Gaussian distribution function.

Standard statistical methodology produces nomal deviates by a transfomation of unifom deviates.

However, in the instances where SIGMA does not use standard techniques, the methods used will produce similar or more conservative results.

The random number generator used in the simulation trials was identified (Ref.19) and test of autocorrelation, length of monotonic runs, and runs above and below mean were given.

Since some random number generators can introduce inadvertent correlation, the use of a thoroughly tested generator is essential.

The tests indicate that the generator is satisfactory.

The method l

used to select axial power distributions is described in Berte, Filstein and Goldstein (Ref. 20).

The method is divided into two parts.

The first part is an algorithm for summarizing the distribution of axial shapes as a frequency distribution of hypercubes. The second part is a method of sample selection called Least Discrepancy Sampling (LDS), used to select a sample from the frequency distribution of hypercubes.

The sampling procedure LDS does not preserve statistical properties of the sampled population and is, therefore, 56

I not acceptable.

However, LDS was not used in selecting axial shapes.

Instead, the sample was selected using simple random sampling or stratified sampling.

Either of these methods is acceptable.

2.

Uncertainty Distributions For the most part, the methodology used to obtain uncertainty distributions on the independent parameters is acceptable.

Distributions were not assumed to be Gaussian without being tested, and where data from several sources could not be pooled, conservative variance estimates were used.

A signal processing system is approximated by a first order Taylor series and the Central Limit Theorem (CLT) is applied to the approximation. The application of the CLT in Appendix A3 (Ref. 5) is justified by stating that the variances of the independent variables are small in relation to their overall ranges.

However, the criterion that is necessary is that the variances be small relative to the size of the region of adequate approximation.

Our review concluded that the necessary criterion is satisfied.

The error analysis perfomed on the shape annealing factor data has no statistical validity.

Inspection of the data in Table 4 of Appendix A3 (Ref. 5) shows that the data from Calvert Cliffs-1 is from a different population than the data from the other reactors in the table.

Both the mean and the variance, af ter correction for cycle and channel effects, are larger for the Calvert Cliffs-1 data. The incorrect error analysis attempted to account for the larger variance by using a multiplicative error structure.

However, the standard deviation apparently increases faster than the mean, so the multi-plicative structure does not remove the systematic component of the error.

Additional data on shape annealing factors for Calvert Cliffs-1 was provided and analyzed in Reference 18.

The analysis concluded that the existing uncertainty estimate was conservative for Calvert Cliffs-1.

This analysis of the Calvert Cliffs-1 data has some statistical faults.

However, these faults lead to an overestimate of the uncertainty so that the conclusion remains valid.

Thus, the existing stochastic simulation of the axial shape index uncertainty is acceptable.

57

A.2 PART TWO The licensee's approach for SCU is to adopt a single set of "most adverse state parameters" and generate a MDNBR response surface of the system parameters, which is, in turn, applied in Monte Carlo methods to combine numerically the system parameter probability distribution functions with the CHF correlation uncertainty.

Our review of the SCU methodology includes the selection of the most adverse state parameters, the eliminaton of some system parameters from the response surface, the uncertainties of system parameters in the response surface and the statistical method used in calculating the final equivalent MDNBR 1imit.

A.2.1 Most Adverse State parameters Generation of the actual response surface simultaneously relating MDNBR to both system and state variables would require an inordinate number of detailed TORC analyses.

The licensee's solution to this problem is to select one single set of state parameters for use in developing the system variable response surface.

The problem then becomes one of selecting a single set of state parameters, termed the most adverse state parameter set, that leads to conservatism in the' system parameter response surface; i.e., the resultant MDNBR uncertainty is maximized.

Calculations are performed with the detailed TORC code to detennine the sensitivity of the system parameters at several sets of operating conditions (state parameters).

By tabulating the results of l

the sensitivity studies and through an examination of tables and exercise of engineering judgment, the "most adverse"is listed in Section 3.1.5 of the CEN-124(B)-P report (Ref. 6).

Our review has found that the values of these parameters, such as system pressure, inlet coolant temperature and primary flow rate, are very likely at their most adverse values. However, the conclusion is not valid for the axial l

shape index (ASI).

I In Section 1.1 of Reference 6 it is stated that the MDNBR is a smoothly varying function of the state parameters.

This is not the case for the ASI.

The ASI enters the calculation of MDNBR by the selection of a value of ASI from a l

58 l

T finite collection of axial shapes and corresponding ASIS.

Because the correspondence between ASI and axial shape is a multi-valued relationship, MDNBR cannot be a continuous function of ASI.

Thus, a relatively small per-turbation in ASI could lead to a large change in MDNBR.

The data presented in CEN-124(B)-P indicate the possibility of an ASI that is considerably more adverse than the ASI selected as most adverse.

In response (Ref. 21) to our question (Ref. 22) the licensee provided additional evaluations of the sensi-(

tivity of f1DNBR near the most' adverse ASI. With this additional infonnation, l

the ASI selected as most adverse can be accepted as leading to conservative estimates of the sensitivity of f1DNBR to system parameter variation. We, therefore, conclude that the licensee has achieved the goal of finding the most adverse set of state parameters.

A.2.2 System Parameter Uncertainties The CEN-124(B)-P report lists each of the system variables and then either provides the rationale for eliminating the variable from the statistical combination or provides the appropriate uncertainty value.

Our review of these variables follows:

(i)

Radial Power Distribution Conservatism in the thermal margin modeling is listed as a reason that uncertainty in the radial power distribution need not be considered.

A subsequent response to questions (Ref. 21) outlined the proprietary calculational technique currently being used to maintain the conservatism.

The technique was reviewed and found to be satisfactory.

The elimination of the radial power distribution uncertainty is justified.

(ii)

Inlet Flow Distribution The sensitivity studies in CEN-124(B)-P (Ref. 6) has shown that MDNBR in the limiting hot assembly is unaffected by changes in the inlet flow of assemblies which are diagonally adjacent to the hot assembly.

Therefore, only the inlet flow to the hot assembly and its contiguous neighbors are included in the analysis. We find this approach acceptable.

59

(iii)

Exit Pressure Distribution The sensitivity study provided in Table 3.10, CEN-124(B)-P (Ref. 6) has shown the insensitivity of MDNBR with respect to the variation in exit pressure distribution.

Therefore, we conclude the elimination of the exit pressure distribution uncertainty from the MDNBR response surface acceptable.

(iv)

Enthalpy Rise Factor Enthalpy rise factor is used to account for the effect on hot channel enthalpy rise of the fuel manufacturing deviation from nominal values of fuel dimension, density, enrichment, etc.

The enthalpy rise factor is determined in accordance with an approved quality assurance procedure (Ref. 23).

This involves a 100 percent recording of the relevant data which are then collected into a histogram.

The mean and standard devia-tion are detennined with 95 percent confidence. We find this procedure and the uncertainty listed in Table 5.1 (Ref. 6) acceptable.

i (v) Heat Flux Factors Manufacturing tolerance limits and fuel specifications which conservatively define the probability distribution function of the heat flux factor are used. We find the mean and the standard deviation of heat flux factor used in the analysis are conservative and, therefore, acceptable.

(vi)

Clad 0.D.

Proprietary measured clad diameter mean and standard deviations are given-based on as-built data.

The minimum systematic clad 0.D. and its standard l

deviation are used in the development of the heat flux factor since this l

gives the most adverse effect on DNB. The (ninimum clad 0.D. and its standard deviation are used in wetted perimeter calculations which pena-lizes the MDNBR.

This double accounting of the clad 0.D. uncertainty introduces conservatism in the analysis and is acceptable.

60

(vii) Systematic Pitch Reduction As-built data are used to detennine proprietary mean and standard deviations of gap width.

The minimum mean and its standard deviation are chosen for combination with maximum clad 0.D. to give the minimum pitch.

The use of the minimum gap width is a conservative approach and is acceptable.

l (viii) Fuel Rod Bow The methodology for calculating rod bow compensation is discussed in Section 2.1.4, herein.

The rod bow compensation is applied directly as a multiplier to the t1DNBR limit and the approach is acceptable.

(ix) CHF Correlation The DNBR limit associated with the CE-1 correlation as discussed in Section 2.1.3 is imposed to account for only the uncertainty of the corre-lation.

Other uncertainties associated with plant system parameters and measurenents of operating state parameters are accounted for, separately, through accompanying uncertainty factors.

In our review of the correlation prediction uncertainty, we also applied a cross-validation technique, where the test data are divided into two equal portions.

The parameters of the correlation are estimated separately on each half.

The estimated correlation from one half is then used to predict the data from the other half.

Based on results of the cross-validation technique, we conclude that the standard deviation of the measured to predicted CHF ratio should be increased by 5 percent.

This increase in correlation uncertainty should be included in the derivation of the DNBR limit, (x) Code Uncertainty j

Uncertainty exists in all subchannel codes.

Our evaluation result of the CE-1 DNBR limit using the COBRA IV code differs slightly from the licensee's 61 I

analysis using the TORC code. This is, to a great extent, a result of the inherent calculational uncertainties in the two codes.

The licensee contends that since the same TORC code is used for both CHF test data analysis and CHF calculations in the reactor, the code uncertainty is implicitly included in the minimum DNBR limit that is used for reactor application.

However, we find the argument not valid since the CHF test section, being a small number of representative pins, differs from the reactor fuel assemblies in the large reactor core.

Even though the heated shrouds are used in test assembly, the two-phase frictional pressure drop and diversion cross flow phenomena, etc., result in uncertainties in thermal hydraulic conditions predicted in the test assembly and reactor core.

Information to quantify these uncertainties are not easily obtained and have not been provided.

Therefore, consistent with past practice, we have imposed a 4 percent uncertainty for the subchannel codes and 1 percent uncertainty for transi.ent codes which predict conservatively against data.

These code uncertainties are imposed only when SCU is used for design analysis.

The code uncertainties should be included in the SCU to assess the effect of the uncertainties on DNBR limit.

A.2.3 Response Surface of System Parameters The use of a response surface to represent a complicated, multi-variate function is an established statistical method.

A response surface relating MDNBR to system parameters is created.

Conservatism is achieved by selecting the "most adverse set" of state parameters that maximizes the sensitivity of MDNBR to system parameter variations.

The response surface includes linear, cross-product, and quadratic terms in the system parameters.

Data to estimate the coefficients of the response surface are generated in an orthogonal central composite design using the TORC code with the CE-1 CHF correlation. The resulting MDNBR response surface is described in Table 4-2 of CEN-124(B)-P (Ref. 6).

The licensee has calculated the coefficient of detennination associated with 4

the response surface to be 0.9995 and the standard error of.003396. We conclude that the response surface prediction of MDNBR is acceptable.

62

A.2.4 Derivation of Equivalent MDNBR Limit The probability distribution function (pdf) of MDNBR is estimated using the response surface in a lionte Carlo simulation. The simulation also accounts for uncertainty in the CHF correl 1 tion.

The estimated MDNBR pdf is approxi-mately nomal, ano a 95/95 probability / confidence limit is assigned using normal theory.

The SIG!iA code is used in a simulation to estimate the distribution of MDNBR.

SIGMA is reviewed in the statistical, evaluation of Part 1 of CENPD-124(B)-P (Ref. 5).

The results of the simulation were compared to results obtained using an analytical propagation of variance.

The two methods are in close agreement.

Therefore, we conclude the use of lionte Carlo simulation and SIGMA code acceptable.

In our review of the statistical methodology used in deriving the final equivalent MDNBR limit (Section 6.1, Reference 6), we discovered that an incorrect number of degrees of freedom is used in calculating the error asso-ciated with the response surface at 95 percent confidence level.

However, since the error associated with the response surface is very small, the error results in minimal effect on DNBR limit.

The derivation of the SCU-equivalent MDNBR limit is generally acceptable except for the omissions of the CE-1 correlation cross-validation uncertainty and code uncertainty.

As described in Item A.2.2-ix, the standard deviation of l

the measured / predicted CHF ratio should be increased by 5 percent resulting from cross-validation of the test data.

This increased uncertainty results in an increase of MDNBR by 0.005.

Secondly as described in Item A.2.2-x, a 5 percent code uncertainty should be included in the response surface.

Assuming this uncertainty equal to two standard deviations, and combining the j

standard deviation with the standard deviation of the response surface by root l

sum square method, the 11DNBR limit will increase by a factor of 1.008 (Ref.16),

i.e., an increase of 0.01 in MDNBR limit. With the generic MDNBR limit of 1.19 for the CE-1 correlation, the SCU-equivalent MDNBR becomes 1.234. As was explained in Section 2.1.4, no rod bow DNBR compensation is required for Cycle 6, therefore, l

63 L

the licensee's proposed final fiDNBR limit value of 1.23 is correct and is i

acceptable to the staff.

A.3 PART THREE Part 3 of the report describes the method for statistically canbining the uncertainties involved in the calculation of the limits for DNB, linear heat rate (LHR), and limitirg condition for operation (LCO).

The methods outlined parallel those given in Part I to develop the statistical combination method for LSSSs.

For this reason the comments on the discussion for Part 1 of this review also apply to Part 3.

The differences between Part I and Part 3 of this report arise in the develop-ment of those distributions which impact LCOs differently than they impacted the LSSSs, in particular, the determination of whether statistically cambining 4

uncertainties affects the selection of initial conditions for the transient analyses. Also it is necessary to examine the sensitivity of the required over power nargin (R0Pli) to the initial condition to determine the magnitude of variations of R0Pfi within the range of the uncertainties.

A.3.1 Thermal-Hydraulic Evaluation, Part 3 The uncertainty distributions which are different for the LC0 determinations described in Part 3 from the LSSS determinations described in Part I have to do with the Axial Shape Index (ASI).

The LC0 determinations for Calvert Cliffs uses two sets of instruments, an in-core set and an ex-core set. The ex-core detectors are used in the power ratio recorder monitoring system.

These detectors are located in symmetrical positions to those used in the LSSS safety channels and therefore, except uncertainties in instrument circuitry, have the same uncertainties as those given in Part 1 of this report.

The circuitry uncertainty was calculated with standard techniques.

The uncertain-ties of this system are compiled in Table A-1 (Ref. 7) and are satisfactory.

The in-core detector uncertainties are used to calculate the core average axial shape index.

The system is described in the BASSS report (Ref. 24).

The uncertainties of this system are given in Table A-2 (Ref. 24) and are satisfactory.

64

The licensee has detemined that the reactor coolant system (RCS) depressurization event gives the maximum pressure bias tem for the entire range of system parameters allowed by the Technical Specifications LCO. The methods and initial conditions used in this analysis are selected in the same manner as is currently done (Ref. 25). No changes in the detemination of the Tfi/LP trip for protection against design basis events is required as a result of the change of combining uncertainties from deteministic to statistical.

l The licensee has also detemined that none of the design basis events has a margin degradation from time of trip signal to time of peak kw/ft greater than the bias already included in the LPD trip system.

Therefore, the method of i

combining uncertainties, statistical or deteministic, has no impact on the initial conditions selected for analysis, The four pump loss-of-flow event (LOF) and the control element assembly (CEA) i drop events characterize those events for which RPS trips or sufficient initial steady-state margin is necessary.

For both events, the maximum variation in the R0Pf1 was detemined.

This margin variation is added to the cycle specified R0Pfi calculated for nominal conditions to establish the LCO.

The analysis of these events contains several conservative assumptions.

For the four pump LOF events they are:

I 1.

The magnetic flux decay in the holding coils was assumed to be 0.5 second.

Field tests show a more realistic 0.4 second.

2.

A low flow response time of 0.5 second was assumed.

Field tests show that this is conservative by at least 0.1 second.

3.

CEA drop time of 3.1 seconds was assumed.

A more realistic value would be 2.9 seconds.

4.

The flow coastdown did not take credit for the coastdown assist feature.

l l

65 l

l

i-l l

For the CEA drop event the conservative assumptions are:

i 1.

A bounding valte of the integrated radial peaking factor was assumed which was conservative by 2 percent.

The analysis also assumed a minimum CEA drop worth which does not produce the maximum radial peaking factor change.

2.

No credit was taken for the lowering of the margin requirement for increasing pressurizer pressure which would occur.

i 3.

The moderator temperature coefficient assumed was the most negative allowed by Technical Specifications.

Best estimate calculations were made for both cases which showed that the conservatism is considerable, r

There are errors in Table 3-1 of the report (Ref. 7).

The values for the i

primary coolant pressure and axial shape monitoring system processing are not the most recent values.

Corrected values have been supplied (Ref.18).

Subse-quent reloads will require that the corrected values provided in Reference 18 j

f be used in calculations.

l l

66

[

i

APPENDIX B Safety Evaluation of CEN-182(B)

" Statistical Approach to Analyzing Creep Collapse of Oval Fuel Rod Cladding Using CEPAN" Introduction The fuel cladding in light water reactors is usually under an external hydrostatic force for all of the irradiation time in the reactor.

The resultant compressive stress induced in the cladding wall causes the cladding tube to creepdown.

The cladding creepdown process results in a decrease of the average diameter, an increase of the average wall thickness, and an increase of the ovality.

For severe fuel duty conditions, plastic instability may occur wherein the cladding will collapse into axial gaps present in the fuel pellet column due to densification, missing pellets, etc.

(The approved C-E cladding collapse analytical methods employs an infinite gap length model which does not rely on the presence of fuel pellets to support the cladding.)

Because of the large local strains that would ensue with collapse, the cladding is assumed to fail if collapse is predicted.

Therefore, it has become a general, industry-accepted design criterion that cladding collapse be precluded throughout the fuel li fe time.

l l

Summary of Topical Report l

l The topical report describes modifications to the C-E generic method for predicting collapse of Zircaloy-4 cladding.

The major computer code used for calculating collapse is called CEPAN.

This code, which is described in the generic topical report.CENPD-187-P-A, (Ref. 26), will remain unchanged; however, the revisions of CEN-182(B) will supersede portions of CENPD-187-A (viz.,

Section 4.0).

67 l

Specifically, the new changes to the creep-collapse analysis are (a) a revision to the method used for establishing uncertainties in cladding geometrical parameters that are used in the collapse analysis and (b) a new criterion for the occurrence of collapse.

The fomer change introduces the use of the SIGMA computer code to statistically detemir.e probabilities for the cladding outer diameter, wall thickness, and initial ovality.

Previously, these 3 cladding geometrical parameters were individually selected on a 95% probability at a 95% confidence level basis.

Four remaining parameters that are also needed to initialize CEPAN are nominal primary system pressure, fuel rod internal pressure, cladding temperature, and fast neutron flux.

The conservative manner by which values of these parameters are chosen is unchanged from the original method as described in the report CENPD-187-P-A.

The SIGt1A code uses random generation (tionte Carlo) and stratified sampling (Latin Hypercube) techniques. The. code is first used to generate random com-binations of cladding dimensions that are derived from probability distri-butions of each of the dimensions.

For each set of combinations, a CEPAN run is made to detemine a unique collapse time. Subsequently, the SIGt1A code is again used to organize all of the collapse times into a probability histogram.

If the one-sided lower 95/95 tolerance limit for collapse time shown on the histogram is less than the fuel lifetime, then collapse is predicted.

Summary of Staff Evaluation We have reviewed the subject report including the fundamental assumptions, limiting criteria, and use of the analytical tool SIGliA.

Our review of the latter was cursory because its use is approved elsewhere in licensing cal-l culations (i.e., CEN-124(P), " Statistical Combination of tincertainties").

Inasmuch as the revisions to the creep-collapse analysis are prinarily changes to input parameters, we have not performed nor required audit calculations.

68

Wa recognize the conservative manner used by C-E in selecting cladding geometrical parameters for the original creep-collapse method. As discussed in Section 4 of the CEPAN report, selection is accomplished for each parameter value by picking a value that either (a) coincides with the extreme allowable manufacturing limit or (b) is equal to the mean plus-or-minus two times the standard deviation as measured from cladding production lots. Obviously, this selection process unnecessarily stacks conservatisms upon each other.

In fact, the use of the largest cladding diameter with the least cladding wall l

thickness is the most conservative aspect of the original C-E collapse analysis.

The new statistical method for calculating the ninimum collapse time will reduce the unnecessary degree of conservatism associated with the original deteministic method of combining adverse cladding dimensions, i

These methods as described in the report are generally accepted as standard engineering practices.

Regulatory Position The subject report provides (a) an acceptable method for statistically establishing cladding geometrical parameters that are used for input to the collapse analysis and (b) an appropriate new criterion for the time to collapse for C-E Zircaloy fuel cladding under operating reactor conditions. These revisions reduce some of the previous calculational uncertainties, but should still result in con-servative creep collapse assessments.

The report may be referenced in future licensing applications employing C-E fuel cladding.

l

1 i

l l

69

APPEtlDIX C Safety Evaluation of CEN-183(B)-P

" Application of CEtiPD-198 to Zircaloy Component i

Dimensional Changes" i

introduction f

For in-reactor service, the dimensional behavior of Zircaloy core components is governed by metallurgical condition, mechanical interference, creep, and growth. The axial dimensional changes in currently designed C-E fuel rods and assemblies accrue predominately as a result of irradiation-induced stress-free growth with the growth of the fuel rods exceeding that of the assemblies. The ability to quantitatively predict these changes is important not only for the detemination of core operational tolerances, but also for optimized fuel utilization and management. fiost notably, to preclude fuel rod bowing that could result from nechanical interference between fuel rods and fuel assembly end fittings, a fuel assembly shoulder gap must be maintained.

Likewise, to prevent the collapse of fuel assembly holddown springs, adequate clearance with respect to the vessel internals nust be maintained.

Summary of Topical Report The topical report describes a modification to the previously approved method (see CEllPD-198, and its 2 Supplements, Refs. 28-30) that is used for the calculation of allowances for (a) axial growth of fuel a~ssemblies and (b) differential axial growth between fuel rods and their fuel assembly end structures.

The modification increases the precision of the previous cal-culation by tracking dimensional changes of various components throughout their lifetime and accounting for feedback to other components.

For instance, as a fuel assembly guide tube grows axially with accumulated fluence, the resulting increase in holddown spring force that is transmitted to the guide tube counteracts the growth by inducing compressive creep in the guide tube.

70

I The new analytical model for Zircaloy growth calculations have been coded into the computer program SIGREEP. The execution process of SIGREEP is similar to t

that described in CENPD-198 and its supplements.

Specifically, a Monte Carlo technique is used to generate joint probability density functions for random combinations of (a) fuel rod growth coefficient, (b) guide tube growth coefficient, and (c) component tolerance variables. A value for each parameter is then randomly selected and SIGREEP is used to incrementally analyze the fuel throughout its period of operation. The results of each time history

]'

calculation is a single value for a specific component dimension.

This prncess is then repeated for typically thousands of sets of new parameters, thus generating a probability histogram.

Finally, the upper or lower (as appropriate) 95% probability value is then selected from the resultant probability histogram to determine whether confomance to the growth criteria will be attained.

The report also provides measurements on guide tube length, fuel assembly length, and shoulder gap spacing that was taken from Maine Yankee and Calvert Cliffs, Unit I fuel assemblies at various refueling outages.

These measure-i ments are compared with SIGREEP predictions for upper and lower 95% probability limit predictions.

Summary of Staff Evaluation We have reviewed 'the subject report including the in-reactor data and data j

predictions provided in support of the new methodology described in SIGREEP.

i The new revision to the previous calculational methodology improves the f

predictive capability of the analysis, yet it is not a major alteration, and conservatism is retained in the new model.

The theoretical bases employed in SIGREEP are used extensively throughout the nuclear industry and are con-I sidered as standard engineering practices.

The SIGREEP predictions are favorably supported by power reactor data to exposures approxinating axially averaged fast fluences of 9x10 /cm2 (E > 0.821 Mev); an equivalent assembly average 21 burnup is about 50 GWd/MTU.

1 71

Stress-free irradiation growth of zirconium-bearing alloys is dependent not only on fast neutron flux, service temperature, and time, but also on texture (preferred crystallographic orientation) and retained cold work.

These latter two variables are strongly dependent on the specific fabrication techniques employed. Therefore, the SIGREEP model depends strongly on the data base from l

which it was benchmarked and, consequently, is applicable only to fuel assembly canponents that are metallurgically equivalent to those in the data base.

Regulatory Position The report describes an acceptable time-history modification to the previously approved method that is used for the calculation of (a) axial growth of fuel assemblies and (b) differential axial growth between fuel rods and fuel assembly end structures.

This modification reduces the degree of conservatism previously obtained with the original methods.

The report also provides in-reactor data and data predictions that verify the new analytical model called SIGREEP.

The report may be referenced in future licensing applications employing C-E fuel assemblies.

However, because the growth characteristics of Zircaloy I

components are sensitive to the fabrication process, future applications of CEN-183(B)-P nust be accompanied with descriptions of the metallurgical state of the components being analyzed.

If the metallurgical condition of t'hese components does not differ significantly from those used in the development and verification of CEN-183(B)-P, then the description may be brief; otherwise l

the use of CEN-183(B)-P must be justified.

l

(

72 i

-.. _ - -