ML18142A153

Nonproprietary Version of Revised VEPCO Evaluation of Control Rod Ejection Transient.

ML18142A153
Person / Time
Site:	Surry, North Anna, 05000000
Issue date:	12/31/1984
From:	Cross R, Erb J, John Miller VIRGINIA POWER (VIRGINIA ELECTRIC & POWER CO.)
To:
Shared Package
ML18142A152	List: ML18142A151 ML18142A153
References
VEP-NFE-2-A(NP), VEP-NFE-2-A-(NP, VEP-NFE-2-A-(NP), VEP-NFE-2-A-NP), NUDOCS 8501020260
Download: ML18142A153 (171)

Text

-~~ --~,-

VEP-NFE-2-A DECEMBER, 1984 Vepco VEPCO EVALUATION OF THE CONTROL ROD EJECTION TRANSIENT

- NOTICE -

THE ATTACHED FILES ARE OFFICIAL RECORDS OF THE DIVISION OF DOCUMENT CONTROL. THEY HAVE BEEN CHARGED TO YOU FOR A LIMITED TIME PERIOD AND MUST BE RETURNED TO THE RECORDS FACILITY BRANCH 016. PLEASE DO NOT SEND DOCUMENTS CHARGED OUT THROUGH THE MAIL. REMOVAL OF ANY PAGE(S) FROM DOCUMENT FOR REPRODUCTION MUST BE REFERRED TO FILE PERSONNEL.

DEADLINE RETURN DATE 8501020260 841219 PDR ADOCK 05000280 P PDR POWER STATION E NUCLEAR F RECORDS FACILITY BRANCH 1

Virignia Electric

VEPCO EVALUATION OF THE CONTROL ROD EJECTION TRANSIENT BY J. G. MILLER AND J. 0. ERB NUCLEAR .FUEL ENGINEERING GROUP POWER STATION ENGINEERING DEPARTMENT VIRGINIA ELECTRIC AND PO\.JER COMPANY I RICHMOND, VIRGINIA I OCTOBER 1983 I

I RECOMMENDED FOR APPROVAL:

I Supervisor, Nuclear Fuel Engineering I APPROVED:

I Director, Nuclear Fuel Engineering I

UNITED STATES NUCLEAR REGULATORY COMMISSION WASHINGTON, D. C. 20555 September 26, 1984 Mr. W. L. Stewart, Vice President Virginia Electric and Power Company Richmond, Virginia 23261

Subject:

Acceptance for Referencing of Licensing Topical Report VEP-NFE-2 (P/NP), "Vepco Evaluation of the Control Rod Ejection Transient 11 We have completed our review of the subject topical report submitted November 23, 1983 by Virginia Electric and Power Company letter No. 657.

We find that this report describes an acceptable method for analyzing the control rod ejection accident in Vepco's Surry and North Anna reactors and this method may be referenced in future Vepco license applications to the extent specified and under the limitations delineated in the report and the associated NRC evaluation, which is enclosed. The evaluation defines the bases for acceptance of the report.

We do not intend to repeat our review of the matters described in the report and found acceptable when the report appears as a reference in license applications, except to assure that the material presented is applicable to the specific plant involved. Our acceptance applies only to the matters described in the report.

In accordance with procedures established in NUREG-0390, it is requested that Vepco publish accepted versions of this report, proprietary and non-proprietary, within three months of receipt of this letter. The accepted versions shall .

incorporate this letter_ and the enclosed evaluation between the title page and the abstract. The accepted versions shall include an -A (designating accepted) following the report identification symbol.

Should our criteria or regulations change such that our conclusions as to the acceptability of the report are fnv,lidated, Vepco will be expected to revise

• 1r. W. L. Stewart September 26, 1984 and resubmit the respective documentation, or submft justfficatfon for the continued effective applfcabilfty of the topical report without revision of the respective documentation.

Sincerely,

~o.~---

cec11 O. Thomas, Chief Standardization and Special Projects Branch Divfsfon of Licensing

Enclosure:

As stated

ENCLOSURE CORE PERFORi'tANCE BRANCH SAFETY EVALUATION OF VEPCO EVALUATION OF THE CONTROL ROD EJECTION TRANSIENT VEP-NFE-2

Enclosure CORE PERFORMANCE BRANCH SAFETY EVALUATION OF VEPCO EVALUATION OF THE CONTROL ROD EJECTION TRANSIENT VEP-NFE-2 I

1.0 Slt1MARY OF TOPICAL REPORT I

This report describes the methods developed by the Virginia Electric and Power Canpany (Vepco) for the analysis of a PQStulated control rod ejection transient 1n the North Anna and/or Surry tticlear Power Stations.

' A description of the rod ejection transient and a discussion of the acceptance criteria which must be met to assure the safe operation of I the plant in the event of sucn a transient is also presented.

The RETRAN computer program is used for the analysis which is performed in two parts. First, a point kinetics analysis is used to calculate the average core nuclear power history. Next, a hot spot thermal-hydraulic calculatjon is used to determine the hot spot enthalpy and temperature transients from which the amount of fuel damage and radio-logical consequences are assessed. Detailed descriptions of the calculational model and the techniques employed in the analysis are presented.

The results of sensitivity studies used to quantify the effect of uncertainties in important core parameters and modeling assumptions on the model's predictions are shown. Also presented are comparisons of the results of the fuel .vendor (Westinghouse) methodology with the Vepc:o methodology as well as ccrnparisons of point kinetics results with three-dimensional space-time kinetics model results.

2.0 STAFF EVALUATION We have reviewed the subject report, including the mathematical ~models and analytical procedures and methods. The RETRAN computer program is the principal calculational tool. A point kinetics analysis is used

• to calculate the average core nuclear power during a rod ejection transient. The hot spot (hottest fuel p;n) enthalpy and temperature transients are detennined fran a hot spot therynal-hydraulic calcu-lation. This is the usual procedure used by the nuclear industry to analyze the spa ti ally dependent transient with a point kinetics model and has been found to be acceptable and usually conservative.

Although the RETRAN program 1s used to analyze the rod ejection transient, the report emphasizes the implementation of the RETRAN models since the detailed code description is available 1n a separate*

document. Therefore, we did not review the RETRAN program,*per se, but rather the qualification of its use in detennining the consequences of a rod ejection transient.

The staff position, as well as that of most of the reactor vendors and licensees, has been to limit the average fuel pellet enthalpy at the hot spot following a rod ejection transient to 280 calories per gram (cal/gm). This was based primarily on the results of the SPERT tests which showed that, in general, fuel failure con~~quences for uo have been insignificant below 300 cal/gm for both irradiated 2

and unirradiated fuel rods as far as rapid fragmentation and dispersal of fuel and cladding into the coolant are concerned. In this report, Vepco has chosen more stringent design limits. The limiting fuel failure criterion has been reduced to 225 cal/gm for unirradiated rods and 200 cal/gm for irradiated rods. Since this is a conservative revision, the staff finds these criteria acceptable.

Vepco proposes a clad temperature limitation of 2700°F as the temper-ature above which clad embrittlement may be expected. Although this is several tundred degrees above the maximum clad temperature limitation proposed in the ECCS criteria, the staff feels this is adequate in view of the relatively short ti~e at temperature and the highly localized

,. effect of this transient. The staff has no limiting temperature criterion for rod ejection transients.

• The neutronics model including the reactivity insertion. neutron kinetics parameters. fuel and moderator temperature feedback, and reactor trip assumptions described in the report are in conformance with the recOfllJlendations of USNRC Regulatory Guide 1.77 and are acceptable. Since RETRAN uses a point kinetics neutronics model, the effect of locally peaked core flux shapes due to the rod ejection is not included in the Doppler reactivity calculation. Therefore. a power weighting factor (PWF) is used to modify the Doppler reactivity versus core average fuel temperature data in order to better approximate the local reactivity effects. The lWINKLE computer program is used by Vepco to justify the use of a PWF. Since TWINKLE is a three-dimensional spatial kinetics code, it accounts for the highly localized reactivity effects of a control rod ejection accident more realistically than a point kinetics calculation and is. therefore. an acceptable method for verifying the PWF. For cases initiated fran both hot zero power and hot full power initial conditions, the RETRAN poin~ kinetics predictions are

• shown to be conservative compared to the TWINKLE three-dimensional predictions and, therefore, the described RETRAN model is acceptable.

The radiological effects of a rod ejection transient have been addressed generically in the Surry 1 and 2 and the North Anna 1 and 2 Updated Final Safety Analysis Reports as well as in the Westinghouse *Evaluation of the Rod Ejection Accident in Westinghouse Pressurized water Reactors Using Spatial Kinetics Methods" (WCAP-7588). In this latter report, a departure from nucleate boiling (DNB) analysis was perfonned using the transient THINC-111 code for a worst-case three-dimensional rod ejection transient. The results indicated that less than 10% of the fuel rods enter DNB. Since one of the-NRC acceptance criteria for this transient requires the assumption that all of the fuel rods which experience DNB release their entire gap inventory of fission products to the coolant,

-the generic result indicates that less than 101 of the core will release fission products. The so~rce tenn for the radiological release'.calcu-lations for these and other Westinghouse designed plants is. therefore.

based on 10% of the fuel experiencing clad failure and releasing their

• entire gap activity. This is in confonnance with Regulatory Guide 1.77 and, hence, acceptable.

Since there is no fuel dispersal into the coolant, the pressure surge during the transient may be calculated based on conventional heat transfer methods. Generic pressure surge calculations for the most severe excess addition of energy to the coolant indicate that any system overpressurization due to a rod ejection transient will meet the NRC criterion of being less than that pressure which would cause stresses to exceed the Service Limit C as defined in Section III of the ASME Boiler and Pressure Vessel Code.

Vepco has analyzed the rod ejection transient using Westinghouse codes and techniques and comparea their results to the results obtained by Westinghouse. We have reviewed these comparisons and concur that Vepco can adequately replicate Westinghouse results. We have also reviewed comparisons between results obtained with t~e NRC approved Westinghouse methodology with those obtained with the Vepco methodology presented in the topical report using RETRAN. We concur that these comparisons demonstrate the acceptability of the Vepco methodology.

A comparison is also presented between the Vepco calculated results using both Westinghouse and Vepco methodologies and results obtained by Vepco using the Westinghouse three-dimensional space-time kinetics code, TWINKLE. The staff finds that this comparison adequately demonstrates the conservatism of the Vepco rod ejection calculational

• method.

3.0 EVALUATION PROCEDURE The staff has reviewed the report within the guidelines provided by Sections 4.3 and 15.4.8 of the Standard Review Plan (NUREG-75/08?)

and by Regulatory Guide l. 77. Part of our review was based on our familiarity with and comparison of similar analyses for control rod

• ejection transients provided in topical reports by other PWR vendors.

The staff al so *had the__ benefit of a meeti_ng with VEPCO concerning the report.

4.0 REGULATORY POSITION The subject report (VEP-NFE-2) provides an acceptable method for analyzing the control rod eJection event for Vepco's Surry and North Anna Nuclear Power Stations. This methodology may be referenced in future Vepco license applications .

PAGE 2 CLASSIFICATION/DISCLAIMER

• The data, information, analytical techniques, and conclusions in this report have been prepared solely for use by the Virginia Electric and Power Company (the Company), and they may not be appropriate for use in situations other than those for which they were specifically prepared.

The Company therefore makes no claim or warranty whatsoever, express or implied, as to their accuracy, usefulness, or applicability. In particular, THE COMPANY MAKES NO WARRANTY OF MERCHANTABILITY OR FITNESS FOR A PARTtCULAR PURPOSE, NOR SHALL ANY WARRANTY BE DEEMED TO ARISE FROM COURSE OF DEALING oa USAGE OF TRADE, with respect to this report or any of the data, information, analytical techniques, or conclusions in it.

By making this report available, the Company does not authorize its use others, and any such use is expressly forbidden except with the prior written approval of the Company. Any such written approval shall itself be deemed to incorporate the disclaimers of liability and disclaimers of warranties provided herein. In no event shall the Company be liable, under any legal theory whatsoever (whether contract, tort, warranty, or strict or absolute liability), for any property damage, mental or physical injury or death, loss of use of property, or other damage resulting from or arising out of the use, authorized or unauthorized, of this report or the data, information, and analytical techniques, or conclusions in it .

PAGE 3 ABSTRACT This report describes the methods developed by the Virginia Electric and Power Company (Vepco) for the analysis of the postulated control rod ejection transient. The principle calculational tool used in this development has been the RETRAN transient thermal-hydraulics code; a point kinetics core physics option is used for predicting the system response of a plant undergoing a rod ejection event. The local fuel temperature response is calculated using a separate RETRAN model of the core hot spot. Comparison with the results from vendor methodologies are included to demonstrate the acceptability of the Vepco methodology for use in the reload core safety analysis and licensing process.

PAGE 4 ACKNOWLEDGEMENTS The authors would like to express their thanks to Messrs. N. A. Smith, N. P. Wolfhope, R, w. Cross and s. M. Bowman for their technical assistance in the development and preparation of this report. T~e authors would also like to express their appreciation to a number of people who reviewed and provided comments on this report.

\

I I

PAGE 5 TABLE OF CONTENTS Title Page CLASSIFICATION/DISCLAIMER 2 ABSTRACT . . . . . . . . . , , , , , , . . . . . . . . . . . . . . . . . * . . . . . . . . . . .. ... 3 ACKNOWLEDGEMENTS . , , , , , * , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 TABLE OF CONTENTS , , , , , * , . . . . . . . . . . . .* . . . . . . . . . . . . . . , . . . 5 LIST OF TABLES ... , , , , , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 LIST OF FIGURES .. , , , , , , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 SECTION 1 - INTRODUCTION - . . . . . . . . . . . . . ~. . . . . . . . . . . . . . . 11 1.1 Pu:cpose and O:cganization of the Repo:ct . . . . . . 11

1. 2 Desc:ciption of the T:cansient . . . . . . . . . . . . . . . . 1.4

1. 3 Accept~noe C:ci te:cia ..... *. . . . . . . . . . . . . * . . . . . . 16 ECTION 2 - METHODOLOG~ 18 2.1 Calculational Model -- The RETRAN Code . . . . . . 18

2. 1. 1 RE TRAN Single Loop Model . . . . . . . . . . . . . 23 2.1.2 RETRAN Hot Spot Model . . . . . . *.......... 37 2.2 Calculational Technique . . . . . . . . . . . . . . . . . . . . . 47 2.2.1 Steady State Physics Analysis . . . . . . . . 48 2.2.2 Co:ce Ave:cage T:cansient Analysis .~.... 52 2.2.3 Powe:c Weighting Facto:c 57 2.2.4 Hot Spot T:cansient Analysis . . . . . . . . . . 63 2.2.5 System Ove:cp:cessu:ce Analysis . . . . . . . . . 67 2.2.6 Radiological Conce:cns . . . . . . . . . . . . . . . . 68 SECTION 3 - SENSITIVITY STUDIES . . . . . . . . . . . . . . . . . . . . . . . 69

3. 1 Int:coduction *********************** 0.000 ... 0.0 69

)

PAGE 6

• 3.2 Sensitivity Study Result~

3.~.1 3.2.2 Point Kinetics Neutronics Parameters Point Kinetics Model Thermal 70 70 Hyd%aulic Parameters 95 3.2.3 Hot Spot Parameters 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0' 0 101 SECTION 4 - VERIFICATION COMPARISONS 11 0

4. 1 Introduction . * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 0

4. 2 Vendor Lioensing M~thodology . . . . . . . . . . . . . . . . . 112 4.3 Verification with Licensing Analysis ........ 119 4.3.1 Core Average Power History . . . . . . . . . . . 121 4.3.2 Hot Spot Analysis . . . . . . . . . * . . . . . . . . . . 130 4.3.3 ConQlusions . . . . . . . . . . . . . . . . . . . . . . . . . . 139 4.4 Comparison to Three-Dimensional Space-Time Kinetics , , , **........ : . . . . . . . . . . . . . . . . . . . . . . 140

4. 4. 1 Thr~e-Dimensional Model . . . . . . . . . . . . . . 141

4. 4. 2 Comparison Results 149 SECTION 5 - SUMMAR~ AND CONCLUSIONS 160 SECTION 6 - REFERENCES * * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161

PAGE 7 LIST OF TABLES Table Title Page 2-1 Thermal-Hydraulic Design Parameters . . . . . . . . . . . 31 2-2 Single LO?P Model Cont~ol Volume Descziption .. 32 2-3 Single Loop "odel Junction Description . . . . . . . . 33 2-4 Single Loop Model Trip Description . . . . . . . . . . . . 35 2-5 Hot Spot Model Heat Transfer Correlations .... . 42 2-6 Hot Spot Avezage Fuel Tempezature and Enthalpy 46 2-7 Hot Channel Fuel Melt Fraction Table . . . . . . . . . . 66 3-1 Point Kinetics Neutzonics Sensitivity Study .. . 79 3-2 Point Kinetica Thezmal Hydraulic Sensitivity Study . . . . , , * , , . * . . . . . . . . . . . . . . . . . * . . . . . . . . . . . . 97

• 3-3 Hot Spot Sensitivity study 106 4-1 Comparison of Vendor/Vepco Licensing Methodologies , * * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116 4-2 Ve-ndor/Vepoo Analysis Results Using Vendor Methodologies , . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 4-3 Verification Comparison Cases . . . . . . . . . . . . . . . . . 120 4-4 FACTRAN/RETRAN Hot Spot Model Comparisons *.... 138 4-5 3-D Compazison Cases .; . . . . . . . . . . . . . . . . . . . . . . . .

• 148 4-6 3-D Hot Spot Model -Comparison Results ..*...... 153 I

I I.

PAGE 8

• Figure 2-1 LIST OF FIGURES Title Vepco RETRAN Single Loop Model . . . . . . . . . . . . . . . .

Page 30 2-2 Vepc<? RETRAN Hot Spot Model . . . . . . . . . . . . . . . . . . . 41 2-3 RETRAN Control Model for Thom Correlation ..... 44 2-4 RETRAN Control Model for Bishop-Sandberg-Tong Correlation * , , , * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 2-5 Normalized Trip Reactivity Curve . . . . . . . . . . . . . . 56 2-6 Power Weighting Factor . . . . . . . . . . . . . . . . . . . . . . . . . 62 3-1 Sensitivity study -- HZP Doppler Reactivity Feedback (Power History) 83 3-2 Sensitivitf study -- HZP Doppler Reactivity Feedback (Ene~gy Release) ..... . . . . . . . . . . . . . . . . 84 3-3 Sensitivity Study-~ HFP Doppler Reactivity Feedback . , , , , * , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 3-4 Sensitivity Study -- Moderator Reactivity Feedback . , , , * , * . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86 3-5 Sensitivity Study Delayed Neutron Fraction. 87 3-6 Sen~itivitp Study Ejected Rod Worth 88 3-7 Sensitivity Study Decreased Time 0£ Ejection 89 3-8 Sensitivity study Increased Time 0£ Ejeqtion 90 3-9 Sensitivity Study Trip Delay Time . . . . . . . . . . 91 3-10 Sensitivity Study Trip Worth 92 3-11 Sensiti~ity Study Initial Zero Power Level . 93 3-12 Sensitivity Study Beta Yield Fractions ..... 94

PAGE *9 3-13 Sensitivity Study~~ Gap Heat Transfer Coefficient ,,1************0*******0000,;,0000000 99 3-14 Sensitivity Study -- Point Kinetics Fuel Pin Geometry . , , * , , , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 0 3-15 Sensitivity Study Hot Spot Fuel Pin Geometry 108 3-16 Sensitivity Study Hot Spot Pellet Tempezature Distribution , , . .. . . . . . . . . * . . * . . . . . . . . . . . . . . . . . . 10 9 4-1 Nuclear Power Transient S1CS BOL HZP 124 4-2 Nuclear Power Transient S1CS BOL HFP 125 4-3 Nuclear Power Transient S1C5 EOL HZP 126 4-4 Nuclear Power Transient S1CS EOL HFP 127 4-5 Nuclear Power Transient S+MTC BOL HZP 128

• 4-6. Nuclear Powe~ Transient S+MTC BOL HFP 129

• 4-7 Hot Spot Fuel Centerline Temperature Transients

-- S1CS BOl, Cases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 132 4-8 Hot Spot Fuel Centerline Temperature Transients

-- S1C5 EOL Cases . . . . . . . . . . . . . . . . . . * . . . . . . . . . . 133 4-9 Hot Spot Fuel Centerline Temperature Transients

-- S+MTC Cases , , . . . *. . . . *. . . . . . . . . . . . . . ** . . . .

• 134 4-10 Hot Spot Fuel outer Clad Temperature Transients

-- S1CS BOL Cases . . . . . . . . . . . . . . . . . . . . . . . * . . . . . 135 4-11 Hot Spot Fuel outer Clad Temperature Transients

-- S1C5 EO?, Cases . * . . . . . * . . . . * . * . * . . . . . . . . . . . . 136 4-12 Hot Spot Fuel Outer Clad Temperature Transients

-- S+MTC Cases . *. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137 4-13 Surry Unit 1 Cycle 1 Core Loading Plan ....... . 144

PAGE 10 4-14 Ridial Geometry for 3-D TWINKLE . . . . . . . . . . . . . . . 145 4-15 Steady State Radial Power Distributions ...... . 146 4-16 Nuclear Power Transient -- 3-D Benchmarks .... . 154 4-17 Total Energy Release -- 3-D Benchmarks . . . . . . . . 155 4-18 Hot Spot Powe:r History -- 3-D Benchmarks .. *... . 156 4-19 Hot Spot Fuel Centerline Temperature Transients

-- 3-D Benchmarks . . . * . . . . . . . . . . . . . . . . . . . . . . *. . . 157 4-20 Hot Spot Fuel Outer Cla~ Temperature Transients

-- 3-D Benchmarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158 I

PAGE 11 SECTION 1 - INTRODUCTION 1.1 Purpose and Organization of the Report This report documents the methodology developed by the Virginia Electric and Power Company (Vepco) for the*analysis of the postulated control rod ejection transient for the North Anna and Surry Nuclear Power Stations.

The intent of this methodology is to provide Vepco with the capability of performing the licensing analysis required for the Condition IV rod ejection transient addressed in the Final Safety Analysis Report.

The purpose of this report is to provide an acceptable reference for safety and licensing analysis of the rod ejection transient for a reload core. This analysis is performe~ in order to demonstrate that an ccurrence of the transient for the core loading in question will neither i'nterfere with the core cooling capability nor result in fuel damage which will lead to an unacceptabie radiation release within the guidelines of 10 CFR Part 100, "Reactor Site Criteria." The approach adopted for verification of this transient analysis method is to compare results obtained with the Vepco methodology to results obtained by Vepco using an aecepted vendor methodology. These results are demonstrated to be conservative comparison to Vepco results based on a three-dimensional ~pace-kinetics model and a detailed hot spot thermal hydraulic model.

The Introduction Section of this report presents a statement as to the purpose of the report, a description of the rod ejection transient, and a discussion of the acceptance criteria which must be met by an analysis

PAGE 12 o insure the safe operation of the plant in the event such a transient occurs.

Section 2 provides detailed descriptions of the calculational model used and the methods by which it is employed to analyze the transient in a conservative manner. The licensing analysis is performed in two parts:

1. a point kinetics analysis to calculate the average core nuclear power history, and

2. a hot spot thermal-hydraulic calculation to determine the hot spot enthalpy and temperature transients from which the amount of £uel damage and radiological consequences of the accident map be assessed.

of the analyses are performed with the RETRAN transient hermal-hydraulic analysis code, but with different code models for each part. Description of the two RETRAN models, henceforth referred to as the RETRAN Single Loop Model for the average core nuclear power history calculation, and the RETRAH Hot Spot Model for the hot spot transient calculation, are provided in Section 2.

The actual technique~ employed in. the analysis are described in the latter part of Section 2, including the additional concerns of providing core physics parameters for the RETRAN models for the speci~ic core loading to be analyzed, and the investigation of the system overpressure and radiological concerns.

The use of point reactor kinetics instead of spatial kinetics in the I core average power history analysis allows for application of a eighting f~ctor to the Doppler reactivity feedback model used in the I

I

PAGE 13 oint kinetics analysis. This weighting £actor., henceforth referred to as the Power Weighting Factor, provides a more accurate, although conservative, estimate of the actual Doppler reactivity feedback to be expected during the transient. A discussion of the derivation and use of this weighting factor is also presented in Section 2.

The results of a se~ies 0£ sensitivity studies performed with the RETRAN models used £or the rod ~jection study are provided in Section 3. These sensitivity studies are used to quanti£y the impact of uncertainties in important core parameters and modeling assumptions on the models' predictions.

9omparisons of the results of a vendor methodology for standard FSAR and eload core licensin~ analyses with those of the Vepco developed methodology are provided in Section 4 to further quantify the acceptability of the Vepco approach. In addition, a comparison of power history calculations between the point kinetics RETRAN model and a three-dimensional space-time kinetics model is presented to demonstrate the conservatism of the point kinetics approach.

The report's conclusions and references are provided in Sections 5 and 6 respectively.

PAGE 14

.2 Description of the Transient The rod ejection transient is a postulated Condition IV event initiated by the mechanical failu~e of a control rod mechanism pressure housing.

Following such a failure, the action of the coolant pressure is assumed to result in a rapid ejection of a rod cluster control assembly and drive shaft from ~he core to a fully withdrawn position within a time interval on the order of 0.1 seconds. This in turn leads to a fast reactivity insertion and may cause a severe asymmetric core power distribution, possibly leading to fuel rod damage.

Reactor protection for the transient is provided by negative reactivity feedback effects and by reactor trips on high neutron flux levels. Trip etpoints for the Surry and North Anna nuclear units are 118% of full ower (including setpoint and instrument errors) for ~eactor operations at or above 10% full power, and 35% of full power (including setpoint and instrument errors) for reactor operations beiow 10% full power.

After an appropriate delay, the rod cluster control assemblies assumed available (less the ejected rod) are assumed to drop into the core.

However, before such rod movement is initiated, the large *power surge of the core is turned around by the negat~ve Dopple~ reactivity feedback resulting from the quick rise in the core's fuel temperature, particularly in the vicinity of the ejected rod. This proves to be the dominant effect in limiting the consequences of the transient.

In general, the core loading for each reload cycle is designed so as to limit the amount of ~eactivity insertion which would result from the jection of a rod cluster control assembly from any of the locations I

PAGE 15 ere one ~ould be inserted during normal power operation. The consequence of such a rod ejection event would be a very rapid increase in core average power level, with an accompanying core pressure surge and a particularly severe temperature transient in the vicinity of the ejected rod. This localized temperature transient may be severe enough to cause some 0£ the fuel to experience DNB (Departure from Nucleate Boiling) and localized fuel damage. The degree of potential damage will be mainly governed by the worth of the ejeqted rod.

Additional details on the postulated transient's description, available protection mechanisms, and consequences may. be found in the appropriate Updated FSARs CUFSARs) for the Surry and North Anna Nuclear Power Stations (Refs. 1 and 2.)

PAGE 16 Acceptance Criteria As detailed in USNRC Regulatory Guide 1.77 (Ref. 3), the acceptance criteria to be used in evaluating the results of a rod ejection transient analysis are1

1. Reactivit9 eHcursions will not result in a radial average fuel enthalpy greater than 280 cal/g (504 Btu/lb) at any axial location in any fuel rod.

2. Maximum reactor pressure during any portion of the assumed transient will be less than the value that will cause stresses to exceed the Emergency Condition stress limits as defined in Section III of the ASME Boiler and Pressure Vessel Code.

3. Offsite dose consequences will be well within the guidelines of 10 CFR Par*t 10 0, "Reactor Site Criteria."

It should be noted that the

• 280

• cal/g criteria applies to both irradiated and unirradiated fuel in providing a conservative maximum limit to limit fuel damage from the prompt pin burst phenomenon and not impair significantl9 the core cooling capabilities.

Additional in-house design limits for acceptability of the transient analysis for a reload have been imposed by Vepco. These limits are:

1. peak clad temperature less than or equal to 2700 °F,

2. fuel centerline melting less than or equal to 10?o at the hot spot,

3. average hot spot fuel enthalpy less than 225 cal/.gm (405 Btu/l~) £or unirradiated fuel, and

PAGE 17

4. average hot spot fuel enthalpy less than 200 cal/gm (360 Btu/lb) for irradiated fuel.

The peak clad temperature limit of 2700 °F reflects a conservative estimate of the temperature at which clad embrittlement may occur. These additional criteria reflect more severe limits than those delineated in Regulatory Guide 1,77, and are consistent with the limits applied in the design of both the Surry and North Anna Nuclear Power Stations.

The predicted results 0£ an analysis which fall within the acceptability limits of these criteria demonstrates that the *consequences of a rod ejection transient for the specific core reload design will be acceptable to the safety of the general public and will maintain the integrity of the co;e cooling capability.

PAGE 18 SECTION 2 - METHODOLOGY 2.1 Calculational Model -- The RETRAN Code The principal calculational tool used by Vepco for the rod ejection transient analysis is the RETRAN computer code. The RETRAN code has been developed by Energy lncorporated CEI) under the auspices of the Electric Power Research Institute CEPRI) as a generalized and versatile computer code for the analysis of *light water react9r systems. The theory and numerical algorithms~ programming details, user's input manual, and examples of its applications to Nuclear Steam Supply Systems CNSSS) are thoroughly documented for the RETRAN code in Volumes 1 through' 4 of Ref.

4*

e RETRAN computer code is based upon t~e RELAP4/003 Update 85 code which was released by the United State Nuclear Regulatory Commission as part of the Water Reactor Evaluation Model CWREM), (Ref. 5).

The RETRAN fluid differential and state equations used to represent homogeneous equilibrium flow in one dimension are described in Ref. 4.

The representations used in previous RELAP codes for control volumes and junctions are also used in RETRAH and allow the analyst to model a system in as much detail as desired. The modeling flexibility 0£ the code is important and will be described in more detail in Sections 2.1.1 and 2.1.2 below.

The equation systems, which describe the flow conditions within the channels, are obtained from the local fluid conservation equations of ass, momentum and energy by use of mathematical integral-averaging

PAGE 19 chniques. Forms 0£ the momentum equatiQn are available for both compressible and. incompressible flow.

The heat conduction representation capabilities of RETRAN have been increased over previous RELAP versions. The principal augmentation to RETRAN is the capability to more accurately calculate two-sided heat transfer .. The appropriate heat transfer correlation is selected based on thermodynamic conditions in each of two flow streams, on either side of a heat conducting solid. Consequently, representations of the heat transfer processes occurring in the steam generator, for example, are more accurate than previously possible.

Reactor kinetics are represented in RETRAN using a point kinetics model ith reactivity feed~ack. The reactivity feedback can be represented by onstant coefficients or in tabular form and accounts for explicit control actions Ce.g., rod scram) and changes in fuel temperature, moderator temperature and density, and soluble boron concentration.

The system component models utilized in RETRAN include a pump model that describes the interaction between the centrifugal pump and the primary system fluid, and valve models that represent either simple valves, check valves or inertial valves. The flexibility of the valve representation and their configuration is important in allowing a wide variety of options *to the user for the modeling of system dynamics.

Several representations of heat exhangers can be modeled by the code.

These include the previously discusssed two-sided heat transfer and one-sided heat transfer in conjunction with user specified boundary onditions. A non-equilibrium pressurizer can be modeled in which the

PAGE 20 ermodynamic state solutions 0£ the ~iquid and vapor regions 0£ the pressurizer are determined from a distinct mass and energy balance £or each region.

As in RELAP, a variety 0£ trip £unctions can be modeled in the RETRAN code to represent various reactor protection system actions. A refinement 0£ the RETRAN code over the RELAP4 code is the addition 0£ a reactor control system modeling capability. Consequently, the dynamics 0£ linear and non-linear control systems are represented with RETRAN models 0£ the more common analog computer elements. This additional capability is necessary £or both best-estimate and licensing analysis, since the response 0£ various control and protection systems may have a significant e££ect on the overall system response.

he analysis 0£ the rod ejection transient has been performed by Vepco with the RETRAN-02 version of the code. The second major code version 0£ RE~RAN, RETRAN-02 was released to allow £or additional capability in modeling certain BWR transients, small break LOCAs, balance 0£ plant modeling and anticipated transients without scrams. Volume 1 0£ Re£. 4 provides a detailed summary 0£ the changes and new capabilities*

available with RETRAN-02.

One of these new capabilities is a one-dimensional kinetics model which all~ws a spatial kinetics modeling 0£ the core in the axial direction (that is, along the axis of fluid flow through the core) to be used in place 0£ the simpler point kinetics model. Since this is the dimension through which the %od is ejected, the one-dimensional kinetics model

~ppears to offer the advantage 0£ a more accurate modeling 0£ the rod

PAGE 21

~jection event in flac~ of the point kinetics model. However, since the primary power redistribution effect which contributes to the negative rectivity feedback in the transient is in the radial plane, no significant advantage over using the point kinetics model is obtained by using the one-dimensional model. On the other hand, the simplicity of use and the flexibility of the point kinetics model compared to the one-dimensional kinetics model make the former the preferential choice for modeling this transient.

The licensing analysis for the rod ejection event as performed by Vepco is divided into two parts--a point kinetics analysis to calculate a conservative core average power history, and a prediction of the enthalpy and temperature transients in the hot spot of the core base4 on

~hat power history. The core average power history calculation is performed using the point kinetics option of the RETRAN-02 code employing the standard Vepco Single Loop Model. This model, as described below in Section 2.1,1, is further documented in Ref. 6 for performing

-0ther non-LOCA transient analyses as well as best estimate analyses for plant operational support.

The core average power history *s calculated with the RETRAN Single Loop Model is weighted by the maximum total power peaking factor for the transient to conservatively estimate the power history of the hot spot.

This data is input to a detailed fuel heat transfer model of the hot spot location, denoted as the RETRAN Hot Spot Model, to determine the hot spot temperature and enthalpy history as a function of time rod ejectio~. Results from this analysis are used to a~certain

PAGE 22 e extent of the fuel damage, if any, and the radiological consequences of the transient.

These two RETRAN models will be described in detail below, followed by a discussion of their application to the actual transient analysis .

PAGE 23 1.1 RETRAN Single Loop Model For the rod ejection transient, the system thermal-hydraulic response of all reactor coolant loops is essentially identical. Hence, a single loop representation of the NSSS suffices. Furthermore, the quickness of the core response during the transient is such that the major changes in core parameters have all taken place in the time interval of about 10 seconds, approximately the time for the coolant to make one complete pass. through the primary coolant loops. Therefore, little impact on the NSSS outside of the core is expected during the rod ejection transient.

Essentially, only a modeling of the reactor core need be performed in order to predict the core average power history with sufficient conservatism for input to the, Hot Spot Model. However, the ready ailability and e,ctensive documentation of the Vepc'o. RETRAN Single Loop Model make it an ideal candidate for use in performing the first part of the transient analysis. Therefore, this model is used as part of the standard Vepco methodology for the rod ejection transient.

Vepco RETRAN Single Loop Models for either the Surry or North Anna Nuclear Power Stations are similarly constructed. Both stations consist of two identical operational nuclear units. All four units are Westinghouse designed three coolant loop pressurized wate~ reactors with core thermal ratings of 2441 Mwt for the Surry units and 2775 Mwt for the North Anna units. The three similar heat transfer loops are connected in parallel to the reactor vessel with each loop containing a centrifugal pump, loop stop valves and a steam generator. The system a pressurizer and the associated control system and

PAGE 24 strumentation necessary for operational control and protection.

The reactor vessel encloses the reactor core consisting of 157 fuel assemblies with each Surry assembly having 204 fuel rods and 21 thimble tubes arranged in a 15 K 15 array while each North Anna assembly has 264 fuel rods and 25 thimble tubes arranged in a 17 K 17 array. The fuel for both station*s

• consists of slightly enriched uranium dioxide fuel. pellets contained within a Zircaloy-4 cladding. General thermal and hydraulic design parameters for the reactor systems are listed in Table 2-1:

The RETRAN thermal hydraulic model is formulated* by representing individual portions of the hydraulic system as nodes or control volumes.

Control volumes are specified by the thermodynamic state of the tluid ithin the volume and basic geometric data such as volume, flow area, quivalent diameter and elevation. The flow paths connecting the volumes or boundary conditions associated with a volume are designated as junctions. Junctions are described by specifying the flow, flow area, elevation, effective geometric inertia, form loss coefficient and flow equation specification for that particular *flow path. Thermal interactions with system metal in the NSSS are modeled with heat conductors. Heat conductors may represent heat transfer from passive sources such as the metal of the reactor coolant system piping or the steam generator tubes. In addition, the internal generation of heat in the core may be represented by active heat conductors designated as powered conductors. Heat conductors are primarily specified by providing the heat transfer area, volume, hydraulic diameter, heated equivalent diameter and channel length of the particular part of the system being

PAGE 25 deled. Temperat,ire dependent material properties (specific heat, thermal conductivity and linear thermal expansion coefficient) are also input. In general, the basic NSSS model is formulated with the code capabilities discussed above. An extensive research effort was conducted to determine the appropriate input required £or the ~odels of the Surry and North Anna units. Information was obtained from plant drawings, the Final Safety Analysis Reports, Vepco internal operating documents, equipment technical manuals *and specific information requested from the NSSS vendor. Specific c:ontrol capabilities and constitutive models of system components will be discussed in the following paragraphs.

The Vepco RETRAN Single Loop Models of the North Anna and Surry nuclear units represent the three actual reactor coolant'loops as one loop. The esulting geomet~y is provided in Figure 2-1 and consis~ of 19 volumes, 29 junctions and 7 heat conductors. While the specific model input for the Surry and North Anna plants is different, the basic model description is the ~ame for the single loop models of both plants. The reactor vessel includes representation of the downc~mer, upper and lower plenums, core bypass and reactor core. The steam generator is represented by four volumes on the pr;mary side, one volume on the secondary side and four heat conductors representing the tubes. Single volumes represent the hot leg piping, steam generator inlet plenum, pump suction piping, .reactor coolant pump, cold leg piping, pressurizer, and pressurizer surge line. Primary system boundary conditions are specified with junctions representing the pressurizer relief and safety valves.

Junctions representing the feedwater inlet, steam outlet, atmospheric team relief and steam line safety valves provide secondary system

PAGE 26

~undary conditions, Specific aspects of the basic model will be discussed below.

All control volumes in the model are homogeneous with the exceptions 0£ volumes 17 and 19. the pressurizer and secondary side of the steam*

generator, which contain two-phase mixtures. Volumes modeling the loop piping use the RETRAN "temperature transport delay" option to represent fluid temperatur~ change movements in the loop as a front, (that is, fluid entering a pipe does -not mix with the fluid present but instead displaces it.) All junctions specify single-stream, compressible flow except .for junction 21, the pressurizer surge line connection to the loop, for which incompressible flow with no momentum flux is specified.

Extended Henry Csubaooled) and Moody (saturated) choking is assigned for

~l junctions. All junctions use Baroczy two-phase multipliers with Fanning friction to define the wall friction except for the four junctions on the secondary side of the steam generator (volume 19) which use a homogenous two-phase multiplier with Fanning friction. All heat conductors use the Dougall Rohsenow heat transfer correlation to describe post-DNB heat transfer. Tables 2-2 and 2-3 summarize the volume and junction descriptions~

The RETRAN code contains several system component models which are used in the Vepco Single Loop Model. These include pump models which describe the interaction between the centrifugal pump and the primary system fluid. These models calculate pump behavior through the use of empirically developed pump characteristic curves which uniquely define the head and torque response of the pump as functions of volumetric flow

PAGE 27 d pump speed. RtTRAN includes "built-in" pump characteristics which are representative of pumps supplied by the major reactor coolant pump manufacturers. These ourves may be modified, as appropriate, by the user to more realistically represent a specific pump design. Although the built-in data are not appreciably different from Vepco's plant-specific curves, Vepco's Single Loop Models incorporate the specific head versus flow response for first quadrant operation found in the units' UFSARs (Refs. 1 and 2).

The Single Loop Model incorporates the RETRAN pressurizer model which defines two separate thermodynamic regions that are not required to be in thermal equilibrium. A non-equilibrium capability is particularly necessary when the transient involves a surge of subcooled liquid into e pressurizer. In addition, the Single Loop Model represents the effects of subcooled spray, electrical immersion heaters, liquid droplet rainout and vapor rise in the pressurizer.

The reactor systems trip logic is modeled to the detail required for a specific analysis. RETRAN trip functions are used to model 1) protective functions, such as the overtemperature delta-T trip, which result in reactor scram, 2) control system bistable element logic, such as coincidence trips which model "majority" logic and 3) general problem control (e.g., problem termination, etc.)

The protective function trips modeled in the standaid Single Loop Model include:

1. High flux

2. Overtemperature delta-T

PAGE 28

3. Overpower delt~-T

4. Low/high pressurizer pressure

5. High pressurizer level

6. Low coolant £low

7. Loss of power to the reactor coolant pumps The Single Loop Model also incorporates the RETRAM control system capability to model the following MSSS control and protection features:

1. Overtemperature delta-T setpoint

2. Overpower delta-T setpoint

3. Pressure controller

4. Lead/lag compensation of the low pressure trip signal.

Tabl-e 2-4 presents a summary of the trips in the standard Single Loop odels.

The core power response is determined by the point kinetics model. in conjunction with explicit reactivity forcing functions and thermal feedback effects from moderator and fuel in the three core regions~ The point kinetics modei specified for the Single Loop Model incorporates one prompt neutron group, six delayed neutron groups with decay *heat represented by 11 gamma emitters, and the important radioactive actinides U-239 and Mp-239. Explicit reactivity forcing functions represent scram and reactivity insertion due to control rod withdrawal in the Single Loop Model as the particular analysis requires. c6nstant temperature coefficients or reactivity tables as a function of temperature (fuel), density (moderator) or power represent feedback Core power is distributed axially among the three core

PAGE 29

~nductors approximating a symmetric cosine shape. Three core materials regions are used to ~epresent uranium dioxide fuel pellets, the helium filled gap and the Zircaloy cladding. Direct moderator heating is appropriately accounted £or in'the model. The transient fuel and clad temperatures are calculated based on* temperature-dependent thermal properties, which are input in tabular form. Additional details on the point kinetics model as specifically used to analyze the rod ejection transient are provided in Section 2.2.2 .

PAGE 30 FIGURE 2-1 VEPCO RETRAN. SINGLE LOOP MODEL STEAM STEAMLLNE RELIEF SAFETY VALVES VALVES STEAM 28 SAFETY 26 29 23 VALVES POW.ER

@ OPERATED RELIEF 24 VALVES 25 22 HEATERS

~ ~- -FEEDWATER riJ3 2 21 2

8 JO 9 18 t SPRAY INTAKE LEGEND

...._DIRECTION OF FLOW

.Q VOLUMES

~ HEAT CONDUCTORS

~AGE 31 TABLE 2-1 THERMAL-HYDRAULIC DESIGN PARAMETERS SURRY PLANT:

TotaL core heat output, Mwt 2441 Heat generat~d in fuel, % 97.4 System operating pressure, psi 2250 Total coolant flow rate, gpm 265500 Coolant temperatures, °F ca 1007. power)

Nominal inlet 543 Average rise in core 65.3 Average rise in vessel 62.6 Core aver_age 577.0 Vessel average 574.3 Average linear power density, Kw/ft 6.2 NORTH ANNA PLANT:

core heat output, Mwt 2775 eat generated in fuel, % 97.4 ystem operating pressure, psi 2250 Total coolant flow :cate, gpm 278400 Coolant temperatures, °F ca 100% power)

Nominal inlet 549.5 Average rise in core 69.4 Average :cise in vessel 66.6 Core average 586.1 Vessel average 582.8 Average linear powe:c density, Kw/ft 5.4

PAGE 32

• TABLE 2-2 SINGLE LOOP MODEL CONTROL VOLUME DESCRIPTION Volume Mixtu:re Tempe:ratu:re ID Desc:ription Type T:ranspo:rt Delay 1 Vessel upper plenum H No 2 Reacto:r hot leg H Yes 3 S/G inlet plenum H No 4 S/G tube volume 1 H No 5 S/G tube volume 2 H No 6 S/G tube volume 3 H No 7 S/G tube volume 4 H No 8 Pump suction piping H Yes 9 Reacto:r coolant pump H No 10 Reacto:r cold leg H Yes 11 Downcomer H Yes 12 Vessel lowe::c plenum H No 13 14 .

Co::ce bypass Core section 1 H

H Yes No 15 Core section 2 H No 16 Co:re section 3 H No /

17 P:ressu:rize:r N No 18 P:ressu:rize:c su:rge line H Yes 19 S/G seconda::cy side T No Abb:ceviations:

S/G - steam generato:r H - homogeneous equilib:rium N - two-phase non-equilib::cium T - two-phase equilib:rium 1

• PAGE 33

• TABLE 2-3 SINGLE LOOP MODEL JUNCTION DESCRIPTION Two-Phase Fanning Junction F:ciction Valve ID Desc:ciption Type Multiplie:c Index H/V 1 Vessel outlet nozzle No:cmal Ba:coczy No V 2 Hot leg outlet No:cmal Ba:coczy Yes H 3 S/G inlet plenum No:i:::mal Ba:coczy No H 4 S/G tubes No:i:::mal Ba:i:::oczy No H 5 S/G tubes No:i:::mal Ba:coczy No V 6 S/G tubes No:cmal Ba:coczy No H 7 S/G-pump suction No:i:::mal Ba:coczy No H 8 Pump intake No:i:::mal Ba:coczy No H 9 Pump discha:i:::ge No:cmal Ba:coczy No V 10 Vessel inlet nozzle No:i:::mal Ba:coczy No V

• ~13 14 15 Downcome:c outlet Bypass inlet Lowe:c plenum Co:ce inte:cnal Co:r:e inte:i:::nal

- co:ce No:i:::mal Ho:i:::mal No:i:::mal No:i:::mal Ho:rmal Ba:coczy Ba:coczy B*a:coczy Ba:coczy Ba:coczy

. No No No No No H

H H

H H

16 Co:re - uppe:c plenum Ho:rmal Ba:coczy Ho H 17 Bypass outlet Ho:i:::mal Ba:r:oczy No H 18 Cold leg sp:ray intake Fill Ba:coczy No V 19 P:i:::z:c. sp:ray Sp:ray Ba:roczy No H 20 P:i:::z:r. - su:rge line No:i:::mal Ba:coczy No H 21 Su:cge line - hot leg No:rmal Ba:coczy No H 22 Feedwate:r fill Fill Ba:coczy No* V 23 S/G, outlet Fill Homog. Yes H 24 PORV 1 Fill Ba:r:o~zy No H 25 PORV 2 Fill Ba:r:oczy No H

PAGE 34

• TABLE 2-3 (cont.)

SINGLE LOOP MODEL JUNCTION DESCRIPTION Two-Phase Fanning Junction F:tiction Valve ID Description Type Multiplie:r Index H/V

---

~- ------ -------- -----

26 S/G atm. steam :relief Fill Homog. No H 27 P:rz:r. safety valve Fill Ba:roczy No H 28 Steamline safety valve 1 Fill Homog. No H 29 Steamline safety valve 2 Fill Homog. No V Notes:

All junctions have single-st:team comp:ressible flow except junction 21 which is incomp:ressible flow.

Abb:reviations:

PORV - powe:t ope:tated :relief valve atm. - atmosphe:rio S/G - s~eam gene:rato:r P:rz:r. - p:tessu:tize:r Homog. - homogeneous V - ve:ttically dist:tibuted junction a:tea H - ho:tizontally dist:ributed junction a:rea

PAGE 35

• TABLE 2-4 SINGLE LOOP MODEL TRIP DESCRIPTION Trip ID Cause of Trip Activation Trip Action End of transient time End calculation 2 High flux (normalized power) Scram 3 Overtemperature delta-T Scr:am 4 Overpower delta-T Scram 5 High pressuri2er pressure Scram 6 Low pressurizer pressure Scram 7 High pressurizer level Scram 8 Low coolant flow Scram 9 User specified time* Close loop isolation valves 10 Low backup heater setpoint Turn pressurizer heaters on 11 High backup heater setpoint Turn pressurizer heaters off 12 User specified time* Shut off reactor coolant pumps 13 Transient time= 0 sec Trip initialization 14 User specified time* Uncontrolled rod withdrawal

• 15 User specified time* Scram 16 High pressurizer pressure Open PORV i 1 17 Low pressurizer pressure Close PORV # 1 18 High spray setpoint Open PORV i 2 19 Low spray setpoint Close PORV i 2 20 High S/G pressure Open atm. steam relief valve 21 Low S/G pressure Close atm. steam relief valve 22 High S/G pressure Open S/G safety valves 23 Low S/G pressure Close S/G safety valves 24 High pressurizer pressure Open pressurizer safety valves 25 Low pressurizer pressure Close pressurizer safety valves

PAGE 36

• TABLE 2-4 (cont.)

SINGLE LOOP MODEL TRIP DESCRIPTION Trip ID Cause of Trip Activation Trip Action 26 User specified time

• Turbine trip 27 Low power End calculation 28 Low-low steam generator mass Sc:ram 29 Low-low steam generator mass Auxiliar'Y feedwater on 30 Scram Tu:rbine trip Notes:

• Hot applicable ,for most transients.

Abbreviations:

PORV - power operated relief valve atm. - atmospheric S/G - steam generatpr

PAGE 37

~1.2 RETRAN Hot Spot Model With a conservative prediction of the core average power history during a postulated rod ejeotion event obtained from the RETRAN Single Loop Model, a RETRAN Hot Spot Model is used to predict the thermal-hydraulic response of the fuel £or the hot spot location of the core. The Hot Spot Model describes a segment of a single fuel rod at the location where the peak core power occurs during the transient. Using the core average power history as a basis for determining a conservative peak power history for the hot spot location, this information is ~nput to the Hot S~ot Model as a driving function. The resulting fuel and clad temperatures and enthalpy history fo~ the hot spot as predicted by the model are then used to ensure that the fuel melt and clad embrittlement imits have not been exceeded for the transient.

The Vepco Hot Spot Model consists of 2 control volumes, 2 flow junctions and a single heat conductor. The volumes are stacked one on top of the other as shown in Figure 2-2. The cross sectional area of each volume equals that of a single channel. The lower volume models a one foot high hot channel section of the core surrounding the one foot high heat conductor. Upstream from this hot volume is a one-foot high unheated, time-dependent volume which serves as a reservoir for the flow from the hot channel. The hot channel's volume, flow area and hydraulic diameter are based on the nominal unit cell dimensions fo.r either the Surry or Horth Anna nuclear cores, depending on which unit is undergoing analysis. Both control volumes are homogenous. Pressure and enthalpy are I specified for both volumes for the initial pre-transient core average i.

PAGE 38

~nditions £or the ~ppropriate unit and initial power level.

Both junctions specify single-stream, compressible flow and use an isoenthalpic model for choking, (should the model predict choked flow at the junction.) The junctions have a horizontally distributed junction area and assume no two phase flow wall friction multipliers. Junction 1, which feeds into the bottom of the hot volume, is a fill junction which supplies flow to the hot volume and has the same £low area. Junction 2, which connects the hot volume to the time-dependent volume, is a normal junction.

Fluid properties tables (dynamic viscosity, specific heat and thermal conductivity) are specified for the coolant based on a pressure of 2250 sia which is the nominal operating pressure for both the Surry and orth Anna cores. These tahles are ~sed by the control system models to calculate post-DNB heat transfer as discussed below. The £luid properties are derived from Ref. 7. Material properties for the fuel, fuel-clad gap, and Zircaloy clad along with an enthalpy table for the

£uel as a function 0£ temperature are built into the model. These prop~rty tables are essentially the same as those provided in the RETRAN Single Loop Model with the exception of covering a higher temperature range and allowing for changes which occur upon melting of the fuel. The mateiial properties are derived from Ref. 8.

Three core materials regions are used to ~epresent the fuel pellet, the helium filled gap and the zircaloy cladding. The fuel pellet consists of concentric mesh spacings 0£ equal radial size, the gap of a single spacing and the cladding of 3 mesh spacings of equal size. A

PAGE 39

~rabolic power distribution is assumed through the pellet (Ref. 2, Section 15.4.6.2.1.2) based on a fuel enrichment which is conservative (high) for the fuel in the reload core loading under analysis.

The heat transfer* correlations used for the surface of the heat conductor are modeled with the RETRAN code's control system. Initially the Thom correlation for nucleate boiling (Ref. 4, Vol. 1) is used.

However, at 0.1 seconds into the transient, departure from nucleate boiling CDNB) is assumed to occur and a trip control in the model switches the heat transfer correlation to the Bishop-Sandberg-Tong film boiling heat transfer correlation, (Ref. 10.) These two heat transfer correlations are summarized in Table 2-5 while Figures 2-3 and 2-4 present an outline of the control block strategy used to model the

~orrelati~ns. Additional control blocks are used to calculate a fuel average temperature and enthalpy for the hot spot as described in Table 2-6.

Initially the values of the gap heat transfer coefficient and nucleate boiling heat transfer coefficient are adjusted to yield the desired steady state tempe~ature distribution throughout the fuel pin. At 0.1 seconds into the transient the value of the hot spot gap heat transfer coefficient is changed to a very large value (10000 Btu/ftZ-hr-°F) to model a closure of the gap at the hot spot location, (Ref. 17.)

The RETRAN option to calculate the heat generated from an exothermic Zircaloy-water reaction between the coolant and the cladding is included in the model. Since this reaction only occurs in the presence of steam, inlet enthalpy of the fill junction emptying into the hot control

PAGE 40

~lume is deliberately ramped to 1000 Btu/lb over a 0.2 second interval to induce steam quality in the volume. (Additional details 0£ this metal-water reaction option are presented in Vol. 1 of Ref. 4.)

PAGE 41

• FIGURE 2-2 VEPCO RETRAN HOT SPOT MODEL 1.0 ft.

2 Fuel 'E Legend: Junction (flow direction shown)

Control Volume 0

Heat Conductor--Fuel Gap Clad

PAGE 42

• TABLE 2-5 HOT SPOT MODEL HEAT TRANSFER CORRELATIONS I. Thom Subcooled Boiling Correlation (Transient time 0,0 to 0.1 seconds.)

hnb = qw/(Tw - To) hnb = nucleate boiling heat transfer coefficient,* Btu/ftZ-hr-°F qw = wall heat flux, Btu/ftZ-hr Tw = wall temperature, 0 f

1/2 where Tw = Tsat + 0.072*exp(-P/1260)*qw Tsat = coolant saturation temperature, °F P = coolant pressure, psia Tc= coolant temperature, 0 f

PAGE 43 TABLE 2-5 *(cont.)

HOT SPOT MODEL HEAT TRANSFER CORRELATIONS II. Bishop-Sandberg-Tong Film Boiling Correlation (Transient time> 0.1 seconds.)

o.~ 1.23 o.68 0.068 ChD/k ).f = . 0. 0 193 CDG/,4 )r ( Cp µ /k 1 C; / ~ ) cs; / f, )

h = film boiling heat transfer coefficient, Btu/ft 2 -hr-°F D = equivalent diameter, ft k = thermal conductivity, Btu/ft-hr-°F Cp = specific heat, Btu/lb-°F J,l = viscosity, lb/ft-hr G = mass velocity, l*b/ft2-hr; r, = saturated steam density, lb/ft3

~ = saturated liquid density, lb/ft3

~ = bulk density, lb/ft 3 Subscript f refers to properties at the film temperature, Tfilm, which is defined as O.SCTw + Tsat) where Tw = wall temperature, °F Tsat = coolant saturation temperature, °F

PAGE 44

• FIGURE 2-3 RETRAH CONTROL MODEL FOR THOM CORRELATION qw----,. MAX ~ CON S C=0.5) Ensure heat flux> 0.5 BTU/hr ft~

e

,1,

• I XPO XPO 0.5 -P/1260 ~ -P/1260 K 0.072 e r MULT Calc¥late wall superheat Tsat ~ SUM Calculate Tw Tw

• I,

-Tc--+- SUM Note: See Table 2-5 for parameter de£initions.

qw~ DIV 1---~:. hnb

PAGE 45

• FIGURE 2-1.f RETRAN CONTROL MODEL FOR BISHOP-SANDBERG-TONG CORRELATION Tsat *---..+ SUM ~ T~ Calculate X 0.5 Tfilm Tfilm FNG FNG FNG Calculate k vs. Cp vs. f'. vs. Fluid Tfilm Tfilm Tfilm P:cope:cties k Cp DIV MULT

~ MULT DIV

~

XPO XPO DIV G

1. 23 0.8 X D

,1, XPO XPO

-0.68 0.068 MULT X .0193

• ~

MULT I

' MULT I

~

Note: See table 2-5 for parameter definitions.

k MULT h X 1/D

.>AGE 46

• TABLE 2-6 HOT SPOT AVERAGE FUEL TEMPERATURE AND ENTHALPY I. Ave%age Fuel Tempe%atu%e Assume the £uel pellet is modeled in the %adial dimension by n equally spaced concent%ic %ings bounded by n+1 mesh points (nodes). Let the tempe%atu%e at node i be given by Ti, whe%e i = 1,2 ***. ,n+1. CThe%e£o%e, T1 is the £uel cente%line tempe%atu%e and Tn+1 is the pellet su%£ace tempe%atu%e.) The a%ea weighted ave%age £uel tempe%atu%e Tis given by 2 n T = (1/n) [(1/2)CT1) + l::cTi)(2i - 2) + (1/2)(Tn+1)(2n - 1) J i=2 0% the RETRAN.Hot Spot Model, n = 10. There£o%e, 10 T = (1/100) [C1/2)CT1) + 1::CTi)C2i - 2) + C19/2)CT11) J i=2 I~. Average Fuel Enthalpy F%om a table 0£ enthalpy versus fuel tempe%ature, the value of the fuel enthalpy at node i, hi, is found £0% each co%responding value 0£ Ti.

The same area weighting as given above £0% T is then used to derive the ave%age £uel enthalpy h. That is, 10 h = (1/100) [C1/2)Ch1) + 1::_chi)C2i - 2) + C19/2)Ch11) J i=2 .

PAGE 47

• • 2 Calculational Technique The analysis of the rod ejection transient by Vepco may be grouped into three major phases as follows:

1. Generation 0£ physics data from steady state physics calculational methods for use by the RETRAN models,

2. An analysis of the reactor coolant system response to the transient using the RETRAN Single Loop Model with the point kinetics option to predict the core average power history, and

3. A thermal-hydraulic analysis of the core hot spot location using the RETRAN Hot Spot Model.

All three phases incorporate assumptions in their analytical t~chniques o provide a conservative prediction of the results of a rod ejection transient .

PAGE 48 2.1 Steady State Physics Analysis In order to model the core physics conditions for a specific reload design with the RETRAN Single Loop Model, the following core physics parameters are required:

1. ejected rod worth,

2. maximum total power peaking factor (Fq) before and after the rod is ejected,

3. Doppler power defect,

4. delayed neutron fraction,

5. moderator temperature coefficient,

6. total trip worth less the most reactive stuck rod, and

7. prompt neutron generation time.

~

~e first £our parameters are the most critical in determining the severity 0£ the transient. The transient is minimally sensitive to the remaining three parameters as will be shown in Section 3.

The maximum total power peaking factor is not used directly in the point kinetics analysis performed with the RETRAN Single Loop Model, but is important in formulating a reasonable description 0£ the Doppler reactivity feedback to be expected during the transient (see Section 2.2.3 below), and is 0£ critical importance in predicting the thermal-hydraulic response of the hot spot.

The use 0£ the term "rod" refers to a single rod cluster control assembly CRCCA); that is, a RCCA consists of the entire cluster of individual rodlets which move as a single unit within a single assembly .

PAGE 49 or Westinghouse ~lants, a rod insertion limit for the various control rod banks is defined as a function of core power level. This limit is set such that operation of the plant with the control banks inserted above their respective limits insures that the core maintains an adequate shutdown margin and acceptable power distribution. A search is performed with the physics codes to determine the location of the highest worth ejected rod with the banks at their appropriate insertion limits for the desired core conditions. For example at hot zero power CHZP) core conditions only the D and C banks will typically be inserted into the core while at hot full power CHFP) only the D bank will be inserted. (See Refs, 1 and 2 for a description of the control rod bank nomenclature and geometry for Vepco's nuclear units.) Rod worths are

~alculate~ uith £rozeq £eedhack as described belou.

The parameters mentioned above are calculated £or each reload core using Vepco's steady state physics codes as described in Refs. 11, 12 and 13.

Due to the low sensitivity of the transient to the prompt neutron lifetime (see 3.2,1,4), a generic value is used in the RETRAH point kinetics analysis although this value is checked with the calculated reload value to assure that no large deviation has occurred. Generic values 0£ the minimum total trip reactivity are also used in the RETRAN point kinetics model, but these values are always compared to the minimum available reload calculated values of the total trip reactivity.

to ensure that the generic values are conservative.

The remaining parameters when used by the RETRAN models are modified by their appropriate nuclear reliability £actors in a direction which is

PAGE 50 nservative for the rod ejection transient. For example, the ejected rod worth is increased by 10%, which leads to a higher core reactivity insertion and thus produces a more severe transient than would otherwise be expected. These nuclear reliability factors are documented in Ref.

14 .

Additional conservatism is insured by calculating all physics parameters at steady state conditions using the "adiabatic assumption." The adiabatic assumption asserts that any fuel damage which might occur during the transient takes place in a small time interval immediately following the ejection of the rod and before the thermal-hydraulic feedback effects of the core become important. This freezing of the core's fee~back leads to larger values of the total power peaking factor nd ejected rod worth than would otherwise be expected in the transient.

An analytical assessment of this effect is presented in Section 4.4 0£ this report where a comparison is presented between steady state peaking factors calculated without thermal-hydraulic feedback and transient peaking factors calculated with thermal-hydraulic feedback using a three-dimensional kinetics code.

For each core reload analysis, key steady state physics parameters are used in determining the severity of each transient. Values of these parameters derived for the reload core are compared to those used in a reference safety analysis. When any reload parameter is not botinded by the value used in the reference analysis, the affected transientCs) must be reevaluated based on .the new key parameter values.

ddi~ional details on the use of the Vepco physics design codes in

PAGE 51

~alyzing the sa£et~ 0£ a reload cycle and in their application to the rod ejection transient in particular are provided in Ref. 15.

PAGE 52

~.2.2 Core Average Transient Analysis The core average transient analysis is performed using the RETRAN Single Loop Model with* the point kinetics physics option. Core power is distributed among three equal size control volumes stacked axially one upon the other. Each contains a single heat conductor. Core power is distributed axially among the core volumes with an approximate cosine shape; that is, 25?. of the power and reactivity feedback effects occur in the upper and lower core volumes while the remaining SOX occurs in the middle core volume. The fuel rod is modeled with two radial concentric fuel regions, a single gap region, and a single clad region.

Initial reactivity insertion due to the ejection of the control rod is

.illlllmplemented by modeling a linearly increasing reactivity insertion into

-he core from ze:co to the total worth of the ejected rod over a time interval of 0.1 seconds, (Ref. 17.) The value used for the total integral worth of the ejected rod is the value provided by the physics design codes increased by the Nuclear Reliability Factor of 10% for conservatism.

Negative reactivity insertion from reactor trip is initiated by the high power trips of the RETRAN Single Loop Model. For HFP cases the trip setpoint is 118% of* full power (including setpoint and instrument errors) while for ~ZP cases the trip setpoint is 35% of full power (again including setpoint and instrument errors). A 0.5 second trip delay is assumed between the time the trip setpoint is reached and the time trip reactivity insertion begins, (Refs. 1 and 2). A conservative

~stimate of the total trip reactivity integral rod worth is provided for

PAGE 53 e appropriate core condition. This is typically 4000 pcm for a HFP condition and 2000 pcm for a HZP condition. CA "pcm" is equal to a percent mille of reactivity.) This total integral trip worth is multiplied by a conserv~tive curve of normalized trip worth versus time to provide an estimate of the actual negative reactivity inserted at a particular time following trip, (see Figure 2-5).

Reactivity feedback effects due to a change in the coolant density are modeled by a moderator temperature coeffici~nt. Reactivity feedback effects due to a change in the fuel temperature are modeled by a table of Doppler defect versus fuel temperature. Both the moderator temperature coefficient and Doppler defect values are provided from the physics design codes and are modified by the appropriate Nuclear eliability Factors in a conservative direction, (i.e., +3 pcm/°F for the moderator temperature coefficient and less 10% for the Doppler defect.) In addition the Doppler defect values are multiplied by a Power Weighting Factor CPWF) to more accurately model the additional feedback effect obtained from actual three-dimensional geometry versus the point kinetics geometry of the RETRAN code. A description of the derivation and use of the PWF is provided in Section 2.2.3 below.

All reactivity parameters input to the RETRAN code are in units of dollars ($). Hence, the values obtained from the physics design codes are converted to dollars by dividing by the value of the delaye4 neutron fraction. The delaye4 neutron fraction is provided by the physics design codes for the appropriate core donditions and modified by a Nuclear

• Reliability Factor in a conservative direction .

PAGE 54 I

e value 0£ the gap heat transfer coefficient CHGAP) used in the fuel pin model is held constant throughout the point kinetics calculation.

Since the correlation between core average fuel temperature and power level (necessary to construct a table 0£ Doppler defect values as a

£unction of the fuel temperature) is dependent on HGAP, the problem arises 0£ an appropriate value 0£ HGAP to use. Steady state calculations were performed with the RETRAN Single Loop Model at various core power levels while varying the va-lue of HGAP. Correlations were derived between power level ~nd core average fuel temperatures as a £unction 0£ HGAP. Heat transfer from the fuel increases or decreases with increasing or decreasing values of HGAP, respectively. Lower values 0£ HGAP result in a higher relative core average fuel temperature £or a given power evel. Since higher fuel temperatures result in greater negative Doppler reactivity feedback, a high value 0£ HGAP appears to be conservative £or the :cod ejection t:r:ansieht. This effect is examined in Section 3 where the RETRAN point kinetics model is shown to be only moderately sensitive to the value of HGAP used. A value 0£ 1000 Btu/£tZ-h:r:-°F is used for HGAP in the point kinetics calculation.

In addition to a prediction of the core average power history £or the transient, the RETRAN point kinetics model also calculates the normalized core energy release. This parameter is obtained by integrating the no:i:malized power as a £unction 0£ time using an integrator control block.

The transient is analyzed at £our different core conditions £or a given eload cycle. These conditions are:

PAGE 55

• 1.

2.

3.

Beginning-of-life CBOL) , hot zero power CHZP)

Beginning-of-life CBOL), hot full power CHFP)

End-of-life CEOL), hot zero power CHZP)

4. End-of-li:fe (EOL), hot full power CHFP)

These conditions are the current licensing basis for the Surry and North Anna nuclear units, (Refs. 1 and 2.)

FIGURE 2-5 PAGE 56 NORMAL1ZED TR1P REACT1V1TY CURVE

• . 0 .9

• o .a N

0 R *o. 7 M

A ..

L 1

Z 0.6 E

D

.R E 0.4 A

C T

1

. V ().3 1

T y

0.2 a.a 0.4 a.a 1 *2 1 .6 2.0 2. 4 TJME AFTER TRJP lSECl

PAGE 57

~ 2 . 3

• Power Weighting Factor The RETRAN point kinetics model calculates the Doppler reactivity feedback during the transient from a table of negative reactivity insertion versus core average fuel temperature*. This table is constr.ucted from the results of Doppler reactivity calculations performed with a steady state diffusion code such as PD207. The table provides a reasonable set of values for the core during a pre-transient core condition; that is, with the ejected rod still inserted into the core.

Upon the ejection of the rod, an extreme skewness results in the radial flux and temperature shapes about the position of the ejected rod. This ondition violates the basic ~ssumption behind the applicability of the oint kinetic~ approach--i.e., that the spatial flux shape throughout the core does not vary.appreciably with time. Use of the typical point kinetics models produces results for this transient that can be *very conservative. Accurate modeling of the core's actual reactivity feedback for such a case requires that the spatial kinetics effects, especially in the radial direction be included. However, the spatial kinetics approach has the disadvantage of requiring considerably greater computer resources than the point kinetics approach.

An acceptable solution to simulating the radial kinetics effects with the point kinetics model may be found in the application of a Power Weighting Factor CPWF) to the Doppler reactivity feedback table used by the point kinetics model. To understand the concept behind the use of a consider the case where one actually knows the Doppler temperature

PAGE S8

~efficient for the post-ejected rod core condition. Let this value be denoted as DTC_OUT. This is the value of the Doppler reactivity fee*dback which should actually be reflected in the point kinetics Doppler reactivity table. Likewise, assume the value of the Doppler temperature coefficient for the core with the ejected rod still inserted, (i.e .* the pre-transient value), is known. Let this value be denoted as DTC_IN.

In general, a Doppler reactivity coefficient is the change in qore reactivity, DELRHO, resulting from some unit change in the core average fuel temperature, DELTF. That is, DTC = DELRHO/DELTF For the case of the ejected r°od being fully withdrawn from the core, e

his equation may be written as:

DTC_OUT = DELRHO_OUT/DELTF_OUT Likewise, £or the pre-transient case, we have:

DTC_IN = DELRHO_IN/DELTF_IH Since the unit changes ~n fuel temperature for the two cases are normally equal, (i.e., DELTF_OUT equals DELTF_IH), the ratio of the Doppler temperature coefficients the two different rodded configuration may be written as:

DTC_OUT/DTC_IN = DELRHO_OUT/DELRHO_IN The Power Weighting Factor is defined as:

PAGE 59

• where the PWF = DELRHO_OUT/DELRHO_IN assumption is made that the change in core average fuel temperatures over which the resulting changes- in reactivities are calculate-d is the same for both cases.

Given a method for calculating DELRHO_OUT and DELRHO_IN, and thereby the PWF, -for various rod ejection event conditions, it would be advantageous to find a correlation between the PWF and some other parameter of the rod ejection transient. If such a correlation exists, the value of the PWF for, a particular rod ejection case may be determined based on the corresponding value of the parameter to which the PWF is correlated.

Since the value of DELRHO_IH, or more correctly, the Doppler reactivity

~ ~able for the pre-transient condition, is readily available from steady state diffusion code calculations, the.value of the PWF may be applied to the table to more correctly predict the true Doppler ieactivity feedback to be used by point kinetics in modeling the transient. This approach has the advantage that DELRHO_OUT need not be explicitly calculated for each rod ejection analysis.

Two-group neutron perturbation theory was used* to calculate the value of the PWF for various post-ejected rod core conditions for* bot~ Surry and North Anna core loadings. Changes in the core average reactivity due to a perturbation of the fuel temperature were calculated for the various radial fuel temperature distributions. Values of nuclear macroscopic cross sections and normal and adjoint flux shapes required for the

.alculations were_provided by radial full core, coarse mesh calculations

PAGE 60

~rformed by the PD207 One-Zone Model, (Ref .. 12.) Details on the use of neutron perturbation theory and its application to calculating point kinetics parameters are provided in Ref. 16.

The resulting values of PWFs- were then correlated with the maximum values of the post-ejection radial power peaking factors CFxy) as calculated with* the two-dimensional PD207 Discrete Model, (Ref. 11.) The resulting least squares fit to the points is presented in Figure 2-6.

The value of the weighting factor increases with increasing Fxy due to the stronger feedback effect. The value of Fxy is in turn strongly correlated with the worth of the ejected rod. This curve is used to derive the value of the PWF to be used for a specific rod ejection analysis. The Doppler defect curve used in the point kinetics RETRAN del is then weighted by this value of the PWF to provide a more reasonable estimate of the Doppler reactivity feedback effect during the transient.

Since the value of the post-ejection Fq is-routinely provided for the transient, (suitably increased by a Nuclear Reliability Factor,NRF), the value of the post-ejection Fxy used to derive the appropriate.PWF to be*

used is calculated bys Fxy = Fq/(NRF x 1.55) where NRF removes the additional conservatism of the Nuclear Reliability Factor (1.075 for the Vepco models) and the facto~ of 1.55 is a generic value of the axial peaking factor for a typical cosine axial power distribution shape.

~ m e additional points of clarification in the use of the PWF follow.

PAGE 61

~enefit is taken for negative Doppler reactivity feedback effects which result from the change in the axial flux shape during the ejection of the rod. Hence, the application of the PWF is conservative from this standpoint.

For a HFP case, the worth of the ejected rod is relatively low. Hence, the value of the post-ejection rad1al peak power is also relatively low and typically yields a valu~ of the PWF near unity. As will be demonstrated in the sensitvity studies of Section 3, such a low value of PWF has little effect on the transient results. Therefore, the PWF is 0£ less importance for a HFP case.

For a HZP case, the importance of the PWF is increased. The relatively

• gh ejected rod worth usually encountered results in a large value of Fxy and hence a PWF value which has a more significant impact on limiting the initial power excursion following the rod ejection.

The application of a PWF derived from perturbation theory to the point kinetics model is an improvement in the prediction of what the* actual Doppler reactivity feedback would be for the transient; however, best estimate analysis 0£ the Doppler reactivity feedback is not implied. To demonstrate that th~ use of the PWF produces a result that is still conservative relative to a more accurat~ prediction of the transient, comparisons with results from a three-dimensional space-time kinetics model are provided in Section 4.4.2. These comparisons will not only verify the conservatism of the use of ~he PWF, but of the point kinetics

~ode1 in gene,a1 £0, the aod ejection ana1ysis 0£ Vepco nuc1ea< units.

FIGURE 2-* 6. PAGE 62 PDWER WEJGHJTNG .FACTOR

• 4.00 3.75 X

3 ,50, 3.25 p

0 X 1-J 3.00 E

R w 2,75 E

I G

H 2,50 T

• ~F A

C 2.25 2.00 T

0 R

1 . 75 l . 50 1

• 25*

0 2 3 4 5 6 7 8 9 10 11 12 RADIAL PEAK1NG FACTOR

PAGE 63

~.2.4 Hot Spot Transient Analysis As described in Section 2.1.2, in order to calculate the thermal-hydraulic response of the hot spot core location to the ejection of the rod, the power history of the hot spot is required. Ideally this would be modeled by multiplying the total power peaking factor CFq) history of the hot spot by the core average power history calculated with the point kinetics RETRAN model.

In place of the hot spot Fq history, the pre- and post-ejection values of Fq as calculated by the physics design codes are available. Since these values have been calculated using the "adiabatic assumption", they will be higher Ci.e,, conservative) compared to the actual values of Fq xpected to occur throughout the initial part of the.transient. A conservative hat spot Fq history is constructed by assuming that initially the value of Fq is the steady state pre-ejection value provided by the physics codes. This value is then linearly increased to the post-ejection value of Fq over a time interval of 0.1 seconds and held there for the remainder of the transient. (This is the same time interval assumed for the complete ejection of the rod from the core.)

This Fq power history is then multiplied by the core average power history to provide the hot channel power history for the Hot Spot model.

Upon reactor trip, the insertion of the scram banks into the core will*

cause additional perturbations to the core power distributions both radially and axially. I~ is possible for the value of Fq at a later time in the transient to actually exceed the post-ejection value of Fq which s assumed to occur at 0.1 seconds into the transient. However, because

PAGE 64

• the the the first relatively resulting part hot of low value of the core average power at such a time, channel power will be appreciably lower than during the transient when the core average power is at a maximum. In addition, the location of the hot channel in the core will change during the transient whereas the hot spot analysis assumes a single location throughout the transient. An actual plot of the hot channel power during the transient as predicted by a three-dimensional space-time kinetics model is provided in Section 4.4.2 and shows the power in the hot channel to be steadily decreasing following the early peak due to the ejection of the rod.

In summary, the power history driving function input to the RETRAN Hot Spot Model consists of two components: ~ conservatively predicted core erage power histo~y and a conservative total peakin~*factor power history.

At 0.1 seconds into the transient, the hot channel is forced into DNB by switching from a suboooled nucleate boiling surface heat transfer coefficient, (Thom correlation, Ref 4. Vol. 1), to a film boiling surface heat transfer coefficient (Bishop-Sandberg-Tong correlation, Ref. 10.) Conservatisms applied to the latter correlation are the use of a safety factor to reduce the calculated heat flux and the assumption of a constant bulk coolant density.

The large increase in core power level is expected to lead to a rapid expan~ion of the £uel pellet. This reduces the pellet-clad gap size, resulting in an increase in the gap heat transfer coefficent. At 0.1

.econds into the transient, the value of HGAP in the Hot Spot Model is C.

PAGE 65

~craased from its initial value to a relatively high value of 10000 Btu/ftZ-hr-°F to reflect this closure of the gap.

From the hot channel calculation, values of the fuel temperatures at the boundaries of each of the 10 fuel pellet regions, the clad temperatures at the inner and oute: clad surfaces and the pellet average enthalpy are obtained as a function of time. These parameters are used to assess the fuel response against the acceptance criteria presented in Section 1.3.

Typically, if the maximum amount of fuel melt at the hot spot is less than 10%, the pellet would be expected to maintain its configuration and, therefore, no unacceptable radiological consequences or core damage are expected to occur. An upper bound on the percentage of fuel melt is from the ratio of the cross sectional area of the portion of the where. melting occurs to the total pellet cross sectional area.

The cross s~ctional area for melting is determined by noting those radial concentric fuel nodes whose temperature at some point in the transient exceeds the assumed temperature for fuel melt. At BOL case this melting temperature is assumed to be 4900 °F-while for an-EOL case a temperature of 4800 °Fis assumed. Table 2-7 presents the table used to compute the maximum fraction of fuel melt for the RETRAN Hot Spot Model .

PAGE 66

~ TABLE 2-7 HOT CHANNEL FUEL MELT FRACTION TABLE Highest Numbered Node n Maximum melt for which Tn < Tmelt fraction(%)

1 0 2 1 3 4 4 9 5 16' Notes:

Tn = fuel temperature of node n Tmelt = fuel melting temperature

= 4900 op for BOL case

= 4800 op £or EOL case RETRAN Hot Spot Mddel contains 11 fuel nodes, (10 concentric fuel rings.)

PAGE 67

• . 2.5 System Overpressure Analysis Included in the acceptance criteria is the limitation that the maximum reactor pressure qu:ring the transient will be less than the value that will cause the stress to exceed conservatively de£ined stress limits" This problem is resolved by a pressure stress calculation and has been addressed genericall1 £or the North Anna and Surry nuclear units as described in Refs. 1, 2 and 17. Since it was concluded that no violations will occur £or Vepco nuclear units due to the rod ejection transient, the system overpressure analysis is not per£ormed as part 0£ the Vepco methodology .

PAGE 68

.2.6 As Radiological Concerns with the case of the system overpressure analysis, the radiological concerns for the rod ejection transient for the Surry and North Anna nuclear units have been addressed generically, (Refs. 1, 2 and 17.) In asses~ing the fission product release, it is assumed that all of the rods which experience DNB release ~heir entire gap inventory ~i fission products to the coolant. Vepco's additional acceptance criteria that the temperature for clad embrittlement will not be exceeded and that the fuel pellet configuration will be maintained guarantees that the condition of the fuel rod losing its integrity from entering DNB is a very conservative assumption.

-* nature occur immediate of the r:od ejection event causes a large hot channel factor only in a very localized regio*n of the core, i.e. , :i,.n the vicinity of the ejected rod. Fuel census analysis as provided in Raz. 17 showed that for the worst case investigated, the average fuel rod failed to reach DNB, and even for those fuel rods reaching DNB no excessive ielease of ~ission products was to be expected.

In summary, the localized nature of the event coupled with the small number of rods eKpected to reach DNB and the large conservatisms inherent in the analysis ensures that meeting the Vepco imposed limits of allowable fuel melt and clad embrittlement temperature for the hot channel assures that the radiological limits for the event as' specified

~n Regulatory Guide 1,77 (Ref. 3) will be met .

PAGE 69

• SECTION 3 - SENSITIVITY STUDIES 3.1 Introduction A sensitivity study was performed for both the RETRAN Single Loop Model and RETRAN Hot Spot Model to quantify the impact of uncertainties in core parameters and modeling assumptions on the models' predictions for the rod ejection transient. This section provides a summary of the study's results.

The study is divided in three parts: (1) neutronics parameter sensitivities for the point kinetics calculation, (2) thermal hydraulic sensitivities for the point kinetics calculation, and (3) thermal hydraulic sensitivities for the hot spot calculation. The first two

~rts were per.formed with the RETRAN Single Loop Model for a .Surry plant. For these two parts, the four bounding cycle conditions were analyzed for each ~ensitivity, (i.e., BOL HZP, BOL HFP, EOL HZP and EOL HFP.) The sensitivities for the hot spot calculation were analyzed with the RETRAN Hot Spot Model for typical HZP and HFP hot spot power histories only at BOL, since for this calculation, with the exception of the assumed temperature for fuel melt, the cycle lifetime is not explicitly reflected in the input .

PAGE 70

~2 Sensitivity Study Results 3.2.1 Point Kinetics Heutronics Parameters Sensitivity studies were performed £or the point kinetics calculation to assess the impact of various neutronics parameters on* the predicted core average power history, For the rod ejection event, the core average power history may b~ divided £or sensitivity analysis purposes into_ two major time domains. The first of these domains runs from the beginning of the* transient until approximately 0.5 seconds into the transient.

During this time the core average power rises rapidly, *obtains its maximum value, and is turned around by the negative reactivity insertion due to Doppler feedback. The transient results are particularly

• ensitive to the core average power histories in this first domain. as a ignificant increase. in the power value will lead to higher transient hot spot temperatures and thus produce more severe transient results.

The second domain, starting at approximately 0.5 seconds into the transient. models the reactor shutdown with the insertion of the control rod banks and the effects of moderator r~activity feedback. During this time interval. the core average power level is ~ignificantly lower than during the first domain.

Table 3-1 presents a summary of the results of the sensitivity study for the neutronics parameters £or the £our different cycle conditions. The peak normalized core power level (where unity is equal to full power) is presented £or each case along with the total energy ~eleased during the five seconds o:f the transient. This latter quantity is of

PAGE 71

.portance for HZP cases in demonstrating the sensitivity of a parameter. This i~ due to the narrowness of the initial core average power peak which occurs during the first time domain for HZP cases. What is in reality a small impact on the transient results may show a difference in the peak normalized power between the nominal and sensitivity cases 0£ several full power levels. Hence, the total energy release (which is the integral 0£ the core power history) at some later point in the transient is used as a more representative quantity £or demonstrating the magnitude 0£ the sensitivity £or HZP cases. The total energy release is also useful in assessing the sensitivity 0£ a parameter whose effects are not noticable until the second time domain, (e.g., trip worth.)

~ e first entry in the table £0~ each cycle condition is for the nomin~l case with no perturbation to any of the neutronics parameters.

Comparison of the sensitivity values for the peak normalized power and energy release for each case with those of the nominal case provides a concise summary of the relative sensitivity of each case.

3.2.1.1 Doppler Reactivity Feedback The first sensitivity investigated is that of the Doppler reactivity feedback. This is modeled in the RETRAN point kinetics calculation through the input of a table 0£ Doppler defect as a function of fuel temperature. Since the Doppler defect values in the table are multiplied by an appropriate Powe~ Weighting Factor CPWF), comparisons are shown between the PWF used £or the nominal case and a PWF of unity .

PAGE 72 the two HZP cases, the necessity of the use of the PWF becomes obvious. Without the PWF the total energy release is tripled for the BOL case and nearly doubled for the EOL case. Figure 3-1 presents plots of core average power histories for the two HZP cases with and without the PWF applied. Corresponding plots of the total energy releases are presented in Figure 3-2. For comparison, additional curves are plotted with the Doppler* reactivity feedback curve used in the nominal cases decreased by 10% (typical of the uncertainty assumed in a Doppler feedback calculation by steady state physics models.) The 10%

uncertainty in the Doppler defect shows little sensitiiity compared to that resulting ftom deletion of the PWF.

or the HFP cases as summarized in Table 3-1, reducing the PWF to unity, lthough it increases the peak normalized power and total energy release, shows a much lower sensitivity than the HZP cases. This is as would be expected, since for a HFP case the ejected rod worth, and, therefore, the resulting power peaking and flux redistribution, are smaller than for a HZP case. Hence, PWF values which are near unity have little impact on the transient predictions. Figure 3-3 presents the normalized power history.comparisons for~the two HFP cases.

3.2.1.2 Moderator Reactivity Feedback Moderator reactivity feedback effects in the RETRAN point kinetics calculation are modeled by inputing a moderator temperature coefficient CMTC) . The assumed uncertainty in the value of the MTC calculated with steady state physics codes is presently +3 pcm/°F. The value for

PAGE 73

~ e MTC input to the nominal cases was increased by this uncertainty for the sensitivity study. The result was a minor increase in the core average power prediction as would be expected for the insertion of an additional positive reactivity feedback effect. Figure 3-4 presents a comparison of the normalized power histories for the two BOL cases. The effect of the MTC is shown to be most prominent early on in the transient where the greatest change in core power level, and therefore, the greatest change in moderator temperature, occurs. A lag occurs between the time of peak power and the onset of the effect of the moderator feedback since the moderator temperature responds more slowly to changes in core vower than the fuel temperature.

Delayed Neutron Fraction The delayed neutron fraction is used in estimating the dollar worth of the reactivity effects in the solution of the point kinetics equation.

The value of the delayed neutron fraction decreases with core burnup.

Decreasing the delayed neutron fraction value by 5%, (the assumed design uncertainty), represents a reduction in the percentage of delayed neutrons in the core a~d thereby implies a core with faster response to changing conditions. In addition, the dollar worth of the two dominant reactivity effects £or the transient, the ejected rod worth and the Doppler reactivity ;eedback, are both modified by perturbing the value of the delayed neutron fraction. In effect, reducing the value of the delayed neutron fraction by 5% increaaes the dollar worth of both parameters by 5%. A~ described in Section 3.2.1.1 above, the transient

PAGE 74 ows little sensitivity to an increase in the Doppler reactivity feedback of the magnitude of 5%. However, as documented in Section 3.2.1.5 below, a 5% increase in the ejected rod worth is significant.

The ejected rod worth is therefore the stronger of the two competing reactivity effects, This results in a higher predicted core average power with decreasing delayed neutron fraction. As is demonstrated in a comparison of BOL and EOL cases with similar core parameters, the EOL cases with signifioantly lower delayed neutron fractions show higher predicted core average powers.* The same effect appears in the sensitivity results which show slightly higher predicted powers than the nominal cases, ( Table 3-1) . Figure 3-5 presents a comparison of the power histories for the two BOL cases .

• . 2.1.4 Prompt Neutron Generation Time An increase in the mean value of the prompt neutron lifetime is expected to slow the rate Qf initial inc~ease in core average power during the transient since, on the average, each prompt neutron will now survive longer in the core before it is absorbed. As shown in Table 3-1, increasing the val~e of this parameter by 5E-6 seconds (from a nominal value of 18E-6 seconds) typically decreases the predicted peak normalized power although the total energy release shows little sensitivity.

Likewise, a decrease in the mean value of the prompt neutron lifetime is expected to hasten the. rate of initial increase in the core average power during the transient, with each neutron being absorbed sooner. As

PAGE 75 in Table 3-1, decreasing the prompt neutron lifetime by SE-6 (from a nominal value 0£ 18E-6 seconds) increases the predicted peak normalized power µhile the total energy release again shows little sensitivity. The e££ects of these changes is more pronounced £or the HZP cases than the HFP cases.

Since the assumed design uncertainty in this parameter is only 5%, the transient can be eHpected to show little sensitivity to even large variations in the value used by the point kinetics model. Therefore, a generic value of 18E-6 seconds is used for all analyses.

3.2.1.5 Ejected Rod Worth The assumed design uncertainty for rod worth is 10%. Increasing the

• jected rod worth by 10% shows the greatest impact on the power history of all the sensitivities investigated. As expected, the increased reactivity insertion due to the withdrawal of a greater rod worth from the core results in a higher peak normalized power. Figure 3-6 presents the power history comparisons for the BOL cases.

3.2.1.6 Rod Ejection Time In the nominal cases, the rod is ~jected by modeling a linear reactivity insertion into the core over a time interval of 0.1 seconds. Two sensitivity cases were investigated for both a shorter rod ejection time C0.05 seconds) and a longer rod ejection time (0.2 seconds.) Little impact resulted from either sensitivity. Typically, for the shorter rod

~jection time, the peak core average power occurred earlier in the

PAGE 76 ansient (see Figu~e 3-7 for a BOL comparison), while for the longer rod ejection time the peak power occurred correspondingly later in the transient, (Figure 3-8). The lower peak powers resulting for the loriger transient time fo% the HFP cases is a result of the negative feedback effects having more time to mitigate the effects of the positive reactivity insertion from the ejection of the rod. No significant impact on the integrated en~%gp release was observed.

• 3.2.1.7 Trip Delay Time A trip delay time of 0.5 seconds is assumed in the nominal cases.

Sensitivity studies were performed by increasing this value'to 1.5 seconds. Since the peak normalized power for the rod ejection transient ypically occurs before the scram banks begin movement, increasing the delay time has no impact on the peak normalized power as shown in Table 3-1. However, the delay in.negative reactivity insertion slows the rate of decrease in the core power level leading to a greater energy release throughout most of the period through which the scram banks are inserting. Figure 3-9 presents the BOL powe~ history* comparisons £or this sensitivity.

3.2.1.8 Trip W~rth As with the ejected rod worth, the assumed design uncertainty 0£ 10% was applied to the trip worth for the sensitivity study. In this case, the worth 0£ the trip tanks were decreased by 10% in order to cause an increase in the core power level. As with the "trip delay time"

PAGE 77

~nsitivity study, the power history predictions of the sensitivity cases were identical to those of the nominal cases during the first time domain. As shown in Table 3-1, the sensitivity of changing the trip worth is minor. Figure 3-10 presents comparisons of the power histories

• for the BOL cases.

3.2*.1.9 Initial Zero Power Level For the nominal zerp power cases, an initial normalized core power level of 1.0E-9 was as::;umed, This p*ower level was increased to a value of 1.0E-3 (i.e., 0.1% full power) for the sensitivity study. The transient showed little sensitivity to this change as shown in Table 3-1. Figure 3-11 presents the power history comparisons for the two HZP cases. For he sensitivity case, the peak normalized power is reached earlier than the nominal case, This would be as expected since the sensitivity case is initially starting at a higher power level.

3.2.1.10 Time Step stze The time step sizes used in obtaining the time-dependent numerical solutions to the equations in the RETRAN model are chosen according to the expected rate Qf change of core conditions during the transient.

Thus, early in the transient where the most rapid changes in core power and temperatures are taking place, the time step size is chosen relatively small compared to later in the transient where the rate of change of core parameters is slower. Ideally, the smaller the time step size, the greater the confidence in the accuracy of the solution.

. I PAGE 78 e* sensitivity 0£ the time step size was investigated by decreasing it by a factor of approximately 0.2 over the nominal value used throughout the transient. The results as presented in Table 3-1 show little sensitivity to this magnitude of reduction.

3 *2

• 1* 11 Beta Yield fractions The RETRAN point ~inetics model has a built-in standard set of beta yield fractions for the reactor kinetics calculations. The sensitivity of this parameter was checked by inputting a set of beta yield fractions representative of Surry reloads in the RETRAN model. The results, as shown in Table 3-1, show little sensitivity. Figure 3-12 presents a comparison of the power histories for the BOL cases.

PAGE 79 TABLE 3-1 POINT_ KINETICS NEUTRONICS SENSITIVITY STUDY I. BOL HZP studies!

Peak Energy Norm. Release Parameter Sensitivity Power Ct=5 sec)

Nominal case values Not applicable 51.1 2.23 Power weighting factor CPWF) Changed from 2.4 to 1.0 127. 6.69 Doppler reactivity feedback Decreased 10% CPWF=2.4) 56.5 2.52 Moderator reactivity feedback Increased +3 pcm/°F 51.6 2.34 Delayed neutron fraction Decreased 5% 58.8 2.30 Prompt neutron generation time Increased 5.0E-6 sec 41.5 2.23 rompt neutron generation time Decreased 5.0E-6 sec 73.5" 2.24 jected rod worth Increased .10% _81. 8 2*61 Rod ejection time Changed to 0.05 sec 52.5 2.24 Rod ejection time Changed to 0.2 sec 53.2 2.23 Trip delay time Increased to 1.5 sec 51. 1 2.59 Trip worth Decreased 10% 51. 1 2.26 Initial zero power. level Changed to 1.0E-3 52.2 2.24 Time step size Multiplied by 0.2 49.9 2. 22 .

Beta yield fr~ctions Surry Reload Values 51. 3 2.28 Notes:

Peak Norm~ Power= peak normalize~ power occurring during transient, (1.0 is equivalent to full power level.)

Energy Release Ct=5 sec) = total core energy release up to a transient time of 5.0 seconds, (units of full-power-seconds.)

PAGE 80 TABLE 3-1 (cont.)

POINT KINETlCS NEUTRONICS SENSITIVITY STUDY II. BOL HFP Studies; Peak Energy Norm. Release Parameter Sensitivity Power Ct=5 sec)

---

~----------- ----------------------

Nominal case values Not applicable 2.02 3.50 Power weighting facto:r Changed from 1. 2 to 1. 0 2.04 3.63 Moderator reactivity feedback Increased +3 pcm/°F 2.02 3.58 Delayed neutron fraction Decreased 5% 2. 13 3.52 Prompt neutron gene:i:ation time Increased 5.0E-6 sec 2.02 3.50 Prompt neutron gene:ration time Decreased 5.0E-6 sec 2.02 3.50 jected :cod worth Increased 10% 2.25 3. 67° d ejection time Changed to 0. 05, !?eC 2.04 3.49 Rod ejection time Changed to 0.2 sec 1 . 94 3.52 Trip delay time Increased to 1. 5 sec 2.02 4.80 Trip worth Decreased 10% 2.02 3.57 Time step size* Multiplied by 0.2 2.02 3.49 Beta yield f:raction~ Surry Reload Values 2.03 3.51 Notes:

Peak Norm. Powe~~ peak normalized power occurring during transient.

(1.0 is equivalent to full power l~vel.)

Energy Release Ct=S sec) = total core energy release up to a transient time of 5.0 seconds, (units of 'full-power-seconds.)

PAGE 81 TABLE 3-1 (cont.)

POINT KINETICS NEUTRONICS SENSITIVITY STUDY III. EOL HZP Studies:

Peak Ene:rgy No:rm. Release Pa:ramete:r Sensitivity Powe:r (t=5 sec)

Nominal case values Not applicable 80.2 1.87 Power weighting faoto~ CPWF) Changed from 2.3 to 1.0 139. 3.65 Doppler reactivity feedback Dec:reased 10% CPWF=2.3) 87.2 2.02 Moderator reactivity feedback Increased +3 pcm/°F 81.2 1.91 Delayed neutron fraction Decreased 5% 87.0 1.91 Prompt neutron generation time Increased 5.0E-6 sec 62.7 1. 87 Prompt neutron gene1:ation ti.me Dec:reased 5.0E-6 sec 11 0. lJ 1. 88

• jected rod worth Inc:reased 10% 116

• 2. 14 Rod ejection time Changed to 0.05 sec 78.4 1. 87 Rod ejection time Changed to 0.2 sec 80. 1 1. 87 Trip delay time Increased to 1.5 sec 80.2 2.06 Trip worth Dec:reased 10% 80.2 1. 89 Initial zero power level Changed to 1.0E-3 75.9 1. 88 Time step size Multi~lied by 0.2 ~ 77.9 1 . 86 Beta yield fractions Surry Reload Values 80.3 1 . 89 Notes:

Peak Norm. Power= peak no~malized powe:r occu:r:ring during t:ransient, (1.0 is equivalent to full power level.)

Energy Release Ct=S sec) = total core energy :release up to a transient time of 5.0 seconds, (units of full-power-seconds.)

PAGE 82 TABLE 3-1 (cont.)

POINT KINETICS NEUTRONICS SENSITIVITY STUDY IV. EOL HFP Studies; Peak Ene:r:gy No:z:m. Release Pa:z:amete:z: Sensitivity Powe:z: Ct=S sec)

---

~----------- ---------------------- ----- ---------

Nominal case values Not applicable 2.60 3.02 Powe:z: weighting £actor Changed £:z:om 1.2 to 1.0 2.63 3.05 Mode~ato:z: :reactivity £eedback Inc:z:eased +3 pcm/°F 2.60 3.03 Delayed neut:z:on £:r:action Dec:z:eased 5% 2.79 2.99 P:z:ompt neut:z:on gene:z:ation time Inc:r:eased 5.0E-6 sec 2.57 3.02 P:z:ompt neut:r:on gene:r:ation time Dec:z:eased 5.0E-6 sec 2. 6 1 3*01 E~ected :r:od wo:r:th Inc:r:eased 10% 3.01.J 3. 1 0

• od ejection, time Changed to 0.05 sec 2.67 3.00 Rod ejection time Changed to 0.2 sec 2.I.J3 3.03 T:r:ip delay time Inc:r:eased to 1.5 sec 2.60 '+. 0 6 T:r:ip wo:r:th Dec:reased 10% 2.60 3.07 Time step size Multiplied by 0.2 2.59 3.00 Beta yield £:r:actions Su:r:ry Reload*Values 2.60 3 .01 Notes: "

Peak No:r:m. Power~ peak no:rmalized powe:r: occu:z::r:ing during t:r:ansient, (1.0 is equivalent to £ull powe:r: level.)

Ene:z:gy Release Ct=5 sec) = total core energy :release up to a t:ransient time of 5.0 seconds, (units 0£ full-power-seconds.)

FIG ._E 3-1 SENSITIVITY STUDY - HZP DOPPLErt 11E./\CTIVITY FEF.OAACK (ENERGY r.ELEJ\SE)

END OF "LIFE BEGINNING OF LIFE 1000.00 100.00 N

N .o 0 R 10,00 R 10.00 n n A A L L I I z z E E 0 0

p p 0 I .oo 0 I .oo w w E E R R

0.10 0,10 O.O 0,5 I ,0 I .5 2,0 2,5 3,0 1,5 4,0 4,5 5,0 0,0 0.5 1.0 I ,5 2.0 2,5 3,0 3.5 4.0 4,5 5.0 TIME I SEC I TIHE 16ECI STAR: NOHINAL CASE IPMF:2.31 STAR: hOHINAL CASE IPWF:2.41 SQUARE: SENSITIVITY CASE IPMf=I ,01 SQUARE: SENSITIVITY CASE lPMf=I .01 TRIANGLE: NOHINRL CASE ILESS 107.1 TRIANGLE: NOMINAL CASE !LESS 107.1 flOURE 3-IB FIGURE 3-IA

FIGURE 3-2 SENSITIVITY STUDY - HZP DOPPLER REACTIVITY FEEDBACK (ENEP-GY RELEASE)

END OF LIFE BEGINNING OF LIFE 1 3,6 6

3,0 N

0 2,1 N R 0 5 II R A II L 2,4 A I L

I z

E z D 2,1 E 4 D E N

E E I ,8 N R E G R y G 3 I ,5 y

R E

R L E E I ,2 L A E 5 A 2 E s 0,9 E

0,3 0,0 0,5 I ,0 I ,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 0,6 1,8 2,4 3,0 3,6 TIIIE l&ECI TIIIE l&EC I STAR: NOIIINAL CASE IPWF:2,31 SQUARE: SENSITIVITY CASE IPNF=l ,01 STAR: NOIIINAL CASE IPNF:2.41 TRIANGLE: NOIIJNAL CASE !LESS IOZI SQUARE: SENSITIVITY CASE IPWF:1 ,01 TRIANGLE: NOIIJNAL CASE ILESS IOZI FIGURE 3-28 FIGURE 3-2A

FIGL'RE 3-3 SEtlSITIVITY STLIOY - IIFP DOPPL!=:R P.E,'l.CTIVITY FEE11BACK BEGINNING OF LIFE END OF LIFE 2.0 2.75 2.50

, .a 2.26 I .6 2.00 I .4 N D I. 75 N II D

R I .2 A

L , .so A

L

.I z

I E z D I .25 E , .o D r D

p N , .oo D E N o.e R E

R 0.1s o.&

0-50 o., 0 .25

'CJ 0.2 ,j)

Q.O 0.5 I .O I .5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 rl"l TlnE l&ECI o:>

u, O.O 0.5 I .O 1.5 2.0 2.6 3.0 3.5 4.0 4.5 5.0 TlnE I &EC I STAR: NOnJNAL CASE IPNf:I .21 SQUARE: SENSITIVITY CASE IPNf:\.01 STAR: NDnJNAL CASE IPNf:I .21 0

flGURE 3-3A SQUARE: SENSITIVITY CASE IPNf~I .QI FIGURE 3-36

F.3-4

• SErlSITI VITY STL'DY - M0!1ERATOP. 11.E/\CTI VITY FEEDBACK

• BEGINNING OF LIFE, HOT ZERO POWER . BEGINNING OF LIFE, HOT FULL POWER

!DD.DD ID DD 0

"0 R

0 R

N N A A L L I I z I -DD z E E D D p P' 0 0 N N D 8 0

E E R R 0.6 D ID 0

• u.

):,,

G)

ITl Q:)

0)

D.D 0-5 l D 0

1°5 2.D 2.5 3.D 3.& 4.D 4.5 5-D 0 -0 0 -5 I 0 0

I 5 0

2 .o 2 .5 3 .Q 3 .5 4 .O 4 .5 5 .o TINE I SEC I TINE I SEC I

. STAR: NOMINAL CASE STAR: NONINAL CASE SQUARE: SENSITIVITY CASE I* 3 PCN/DEG fl SQUARE: SENSITIVITY CASE I+ 3 PCN/DEG fl FIGURE 3-4A FIGURE 3-48

• SENSITIVITY ST~DY DELAYED NEUTRON FRACTl()[J BEGINNING OF LIFE, HOT ZERO POWER BEGINNING OF LIFE, HOT FULL POWER 2.25

i. 2.00 I .75 I .50 N N 0 0 R R A "LA I .25 L I I z z , .oo E E I .DO D D p p 0 0 w w E E o.,s R R o.,o o.o 0,5 1,D 1.5 2.0 2.5 3,0 3.5 4.0 4.5 s.o TIHE I SEC I o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 T111E I SEC I STAR: NOHINAL CASE SQUARE: SENSITIVITY CASE ILES6 51.1 fJGURE 3-5A STAR: NOHINAL CASE SQUARE: SENSITIVITY CASE ILESS 51.1 FIGURE 3-5B

• SENSITIVITY snmv - EJECTED ROD \*JORTH

• BEGINNING OF* LIFE, HOT ZERO POWER BEGINNING OF LIFE, HOT FULL POWER 100.00 10.00 N N 0 0 R R L

A "

A L

I I z 1.00 z 1.25 E E D 0 p p 0 0 1.00 N N E E R R 0.75 0.10 0.50 0.25

""C (Tl co 0.0 0.5 I .o I .5 2.0 2.5 3.0 3.5 4.0 4,5 5.0 00 o.o 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 TIHE I SEC I TJHE I SEC I STRR: NOHINRL CASE SQUARE: SENSITIVITY CASE IPLUS 101.1 STAR: NOHINRL CASE SQUARE: SENSITIVITY CASE IPLUS 101.1 l'IGURE 3-6A FIGURE 3-68

F 3-7 SENSITIVITY STUDY - DECREASED TIME OF EJECTION BEGINNING OF LIFE, HOT ZERO POWER 100.00 BEGINNING OF LIFE, HOT FULL POWER 2.0 I .ii 10.00 I .6 N 1.4 o*

R N "LA 0 R

z I

1.00 A

L E J D z p E D

0 N p E 0 R N E

R

o. ,o O.O 0.5 I .o 1.5 2.0 2.5 3.0 3.5 4.0 4.5 S.O TIHE ISECI O.O D.5 I .o I .5 2 .o 2 .5 3 .o 3 °5 4 .O 4 .5 5 °0 STAR: NOMINAL CASE 10.t SECI TIHE I SEC I 5QUARE: 5EN5JTJVJTY 5TUDY 10.05 5ECI STAR: NOHINRL CASE ID.I SECI FIGURE 3-7A SQUARE: SENSITIVITY CASE ID.OS SECI FIGURE 3-7B

• F-3-8 e SENSITIVITY STUDY - INCREASED TIME OF EJECTIO~

BEGINNING OF LIFE, HOT ZERO POWER BEGINNING OF LIFE, HOT FULL POWER.

2.0 1-8 I .5 I .4 N

N 0 0 R R

"R "l

A I -2

.l I J l z ,.oo E I .o E D D

p 0

0 H o.e H E E R R

0-6 o.,o o.,

0.2 0-0 0.5 I .O I .6 2.0 2-5 3.0 -3.6 4.0 4.5 5.D o.o o.s ,.o 1.5 2.0 2.5 3.0 3.5 **o ,.5 s.o TIHE I SEC I TIHE I SEC I STAR: NOHJNAL CASE 10.1 5ECI STAR: NOHJNAL CASE 10.1 SECI SQUARE: SENSITIVITY CASE 10-2 SECI 60UARE: SENSITIVITY CASE 10.2 SECI FIGURE 3-8A FIGURE 3-8B

• SENSITIVITY STUDY - TRIP DELAY TI~E BEGINNING OF LIFE, HOT ZERO POWER BEGINNING OF LIFE, HOT FULL POWER 2.0 1-8 I -6 I .4 N H 0 0 R R 11 11 I -2 R R L L I I z I .00 z E E 1.0 0 0

.p p 0 0 N N 0 .11 E E R R 0-6 0.10 O.O 0.5 I .o I .5 2.0 2-5 3-0 3.5 4.0 4.5 5.0 0-0 0.5 I .o I .5 2-0 2,5 3.0 3.5 4,0 4.5 5,0 Tll1E ISECI

• Tl11E ISEC I STRR: NOl1JNRL CRSE 10-5 SECI STRR: NOl1JNRL CR5E 10-5 SECI SDURRE: SENSJTIVJTY CRSE 11 .5 SECI SDURRE: SEN6JTJVITY CRSE II .5 SECI FIGURE 3-9R FIGURE 3-9B

• SENSITIVITY STUDY - TRIP HORTH BEGINNING OF LIFE, HOT ZERO POWER BEGINNING OF LIFE, HOT FULL POWER 100.00 2.0 1.e I ,6 10.00

. I 1 .4 N

N D 0 R R 11 1 .2 11 A A L L I I z

.z 1 .OO E 1 .o E

0 o.

p p 0 D w w o.e E E R R 0.6 0 .,o 0,4

.I 0.2 0,0 0,0 0,5 1 ,0 I .5 2.0 2.5 3.0 3,5 4.0 4,5 5,0 o.o 0,5 1 ,0 1 ,5 2.0 2.5 3,0 3.5 ~-0 4.5 5.0 Tll1E I SEC I Tl11E I &EC I STAR: NOl11NAL CASE STAR: NDl11NAL CASE SQUARE: SENSITIVITY CASE ILESS 107.1 SQUARE: SENSITIVITY CASE ILESS 107.1 FIGURE 3-108 FIGURE 3-!0R

E 3-11 SENSITIVITY STL1DY - l:IITI/\L ZERO POI.ff!"{ LEVEL BEGINNING OF LIFE, HOT ZERO POWER END OF LIFE, HOT ZERO POWER 100.00 EOLZP 100.00 10.00 10.00 N

0 R

"L A

I z 1.00 t.00 E

0 p

D M

E R

0.10 0.10

• -o J;:,

rr,

.,"")

lO w

o.o 0.5 1.0 1.6 2,0 2,5 3.0 3,5. ,.o 4.5 5-0 0 .O O.5 I .O I .5 2 .O 2 ,5 3 ,0 3 .5 4 .O 4 .5 5 ,0 TIHE I SEC I TIHE I SEC I STAR: NDHJNAL CASE II .OE-91 STAR: NDHINAL CASE II .OE-91 SQUARE: SENSITIVITY CASE 10.001 I SQUARE: SENSITIVITY CASE 10.001 I FIGURE 3-IIA FIGURE 3-IIB

-~-~-------------------------------------------

• FIGURE _3-12 SENSITIVITY STUDY - BETA YIELO FRACTI0NS BEGINNING OF LIFE, HOT ZERO POWER BEGINNING OF LIFE. HOT FULL POWER 2,D I ,8 I ,6 I ,4 H

0 N R 0 R

"L A n A

-I L z I ,DD I E z 0 E I*

0 p

I 0 p N 0 E w R E R

0,10 0.0-1..,......................."""'...,............,...........""T'......."'Tf'"""'"'"'l'"""'"'"'l'"""'"'"'l'~""T o.o 0,5 ,.o 1,6 2,0 2,5 3,0 3,6 4,0 4,6 6,0

• o.o 0,5 I ,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 llHE 16EC I llHE ISECI 6lAR: NONINAL CASE IRElRAN CODE VALUES I STAR: NONINRL CASE IRElRAN CODE VRLUESI SOUARE: SENSITIV!lY CASE !Surry Reload Values) SQUARE-: SENS IT IV In CASE (SUrr.)I Reload Values) ..

flGURE 3-1211 flGURE 3-\2B

PAGE 95 2.2 Point Kinetics Model Thermal Hydraulic Parameters Four parameters were evaluated in assessing the sensitivity of thermal hydraulic parameters on the RETRAN point kinetics calculation. These we:re:

1. gap heat transfer coefficient

2. fuel pin geometry desc:ription

3. co:re inlet temperatu:re

4. system_pre~sure Table 3-2 presents a summa:ry of the results in the format of Table 3-1 discussed previously, 3.2.2.1 Gap Heat Transfe:r Coefficient the RETRAN point kinetics model, the fuel-clad gap heat txansfer coefficient (HGAP) is modeled by using a constant conductivity fo:r the gap mate:rial th:roughout the rod ejection t:ransient. Fo:r sensitivity study purposes, the value of HGAP was dec:reased by 50% f:rom its nominal value of 1000 Btu/ftZ-hr-°F to 500 Btu/ft 2 -h:r-°F. The results as p:resented in Table 3~2 show little sensitivity.

As described in Section 2, a change in the value, of HGAP is expected to result in off-setting effects on the .t:ransient. A reduction in HGAP leads to a dec:rease in the :rate of heat t:ransfer f:rom the fuel to the coolant with a resultant increase in the rate of fuel heatup. This increased rise in fuel temperature in tu:rn results in a greater negative Doppl~r reactivity feedback effect which leads to a mo:re rapid reduction

' ~ n core power which in turn tends to reduce the fuel temperature.

PAGE 96 gure 3-13 presents a comparison of the power histories for the two BOL cases.

3 2. 2.2 0 Fuel Pin Geometry Description The standard RETRAN Single Loop Model describes the fuel pin with two concentric fuel pin regions, a gap region and a cladding region. For the purpose of the sensitivity study, the number of concentric fuel pin regions was increased from two to six. As shown in Table 3-2, this change had little impact on the transient predictions. Figure 3-14 presents a comparison of the power histories for the two BOL cases.

.2.2.3 Core Inlet Temperature The initial core inlet temperature was increased 4 °F from the nominal values. (4 °F is typical of the uncertainty assumed in the initial coolant temperature for a safety analysis.) A change of this magnitude showed little effeot as seen in Table 3-2.

3.2.2.4 System Pressure Typical uncertainty ass~med for system pressure in a safety analysis is 30 psia. The initial system pressure in the point kinetics analysis was reduced by this magnitude to investigate the sensitivity to system pressure. As with the core inlet temperature, no noticeable impact on the transient result ~as found, (Table 3-2.)

PAGE 97 TABLE 3-2 POINT KINETICS THERMAL HYDRAULIC SENSITIVITY STUDY I. Bai HZP studies, Peak Ene:rgy No:rm. Release Pa:ramete:r Sensitivity Powe:r Ct=5 sec)

Nominal case values Not applicable 51. 1 2.23 Gap heat t:ransfe:r coefficient Reduced 50% 51. 0 2.02 of fuel pin meshes Changed £:rom 2 to 6 51. 1 2, 15 Co:re inlet tempe:ratu:re Inc:reased 4 Of 51. 1 2.23 System p:ressu:re Inc:reased 30 psia 51. 1 2.23 II. BOL HfP Studiesi Peak Ene:rgy No:rm. Release Pa:ramete:r Sensitivity Powe:r Ct=S sec)

---

~-~--------- ---------------------- ----- ---------

Nominal case values Not applicable 2.02 3.50 Gap heat t:ransfe:r coefficient Reduced 50% 2.02 3.45

• of fuel pin meshes Changed £:rom 2 to 6 2.02 3.45 Co:re inlet tempe:rat~:re Inc:reased 4 Of 2.02 3.50 System pressu:re Ino:reased 30 psia 2.02 3.50 Notes:

PeaK No:rm. Powe:r = peak no:rmalized powe:r occu:r:ring du:ring t:ransient, (1.0 is equivalent to full powe:r level.)

Ene:rgy Release Ct=S sec) = total co:re ene:rgy :release up to a t:ransient time o~ 5.0 seconds, (units of full-powe:r-seconds.l

PAGE 98

• TABLE 3-2 (cont.)

POINT KINEiICS THERMAL HYDRAULIC SENSITIVITY STUDY I. EOL HZP Studies; Peak Energy Norm. Release Parameter Sensitivity Power Ct=S sec)

Nominal case values Not applicable 80.2 1. 87 Gap heat transfe:r ooe:f:ficient Reduced 50% 80. 1 1. 86

1. of fuel pin meshes Changed f:rom 2 to 6 80.2 1. 87 Co:re inlet temperature Inc:reased 4 OF 8 0. 1 1. 87 System p:ressu:re Increased 30 psia 80.3 1. 88 II. EOL HFP Studies, Peak Energy No:rm., Release Pa:ram~ter Sensitivity Power Ct=S sec)

Nominal case values Not applicable 2.60 3.02 Gap heat transfer coe:fficient Reduced 50% 2.60 3.06

• o:f fuel pin meshes Changed from 2 to 6 2.60 3.05 Core inlet temperatu:re Increased 4 OF 2.60 3.02 System pressu:re Increased 30 psia 2.60 3.02 Notes:

Peak Norm. Power= peak no:rmalized power occurring during transient, (1.0 is equivalent to full power level.)

Energy Release Ct=S sec) = total co:re energy release up to a transient time of 5.0 seconds, (units of full-power-seconds.)

• SENSITIVITY STUDY RE 3-13 GAP HEAT TRANSFER BEGINNING OF LIFE, HOT ZERO. POWER BEGINNING OF LIFE, HOT FULL POWER*

2.0 I .8 I .6 N

0 N R 0 11 R A 11 L A I L z 1.00 I f

0 z

f p D 0 p w 0 E w R f R

0-10 o.o 0.5 , .o 1.5 2.0 2.5 3.0 3.5 4.0 4 .5 5 .o o.o 1...... . . . . . . . .....,...,~r<J"""'"........r"'""'......,.,.........,"'T'"'°......'""".'""'"""':'1:"'"'~5~51O I .5 2.0 2.5 3.0 3.5 4.0 ~. .

o.o 0-5 I .O TINE ISECI TINE I SEC I STAR: NONlhAL CASE ST~R: NOMINAL CASE SOUARE: SENSIJIVITY CASE 1LES6 507.1 SOUAR[: SENSl!!VITY CASE IL[SS 507.8 FIGURE 3-13R flGURE 3-138

FI GURF. 3-14 SENSITIVITY STUDY - POEfi KHIETI CS FUEL PIN GEOMETRY BEGINNING OF LIFE, HOT FULL POWER BEGINNING OF LIFE, HOT ZERO POWER 100.00 2.0

} .e I .6 10.00 I.*

N N 0 0 R R 11 I ,2 A

"LA L I

I t z I .oo E I .O E D D

r p 0 0 w o.e H E E ,11 R

Q.6 0 .10 o.*

0.2 O.O 0.5 I .O 1-5 2.0 2.5 TIHE I SEC I 3.0 3.5 *. 0 4.5 5.0 0.0.,............,...""""...,...........,...................................,.................................................

O.O 0.5 I .o I .5 2.0 2.5 TINE I SEC I STAR: NOHINAL CASE ITWO FUEL REGIONS!

3.0 3.5 4.0 4,5 5-0 0

0

• ' STAR: NOHINAL CASE ITWO fUEl REGIONS! SQUARE: SENSITIVITY CASE ISIX FUEL REGIONS!

SQUARE: S[NSIT6V!TY CASE !SIX FUEL AEGIONSI FIGURE 3-148 ll'IGURE 3-14A

PAGE 101

.2.3 Hot Spot Parameters Sensitivity studies were performed with the RETRAN Hot Spot Model to evaluate the impact of various thermal hydraulic and neutronics parameters on the temperature and enthalpy history of the h?t spot location. For a given nuclear unit at a given initial power levelt the only differences in the RETRAN Hot Spot Model between a BOL and EOL case are the assumed temperature of fuel melt and the input hot spot power history. Therefore, the sensitivity studies were performed only at BOL conditions for both a zero power and full power case.

The sensitivities investigated were:

1. total power peaking factor

2. power historp shape

3. gap heat transfer coefficient

4. fuel pin geometry description

5. inlet temperature

6. system pressure

7. mass flow rate

8. metal-wate; reaction option Table 3-3 presents a summary of the sensitivity study results for the Hot Spot Model. The percentage 0£ fuel melted, maximum fuel centerline temperature and clad temperature, and maximum fuel average enthalpy are compared £or each sensitivity case with those predicted £or a nominal case since these are the key parameters in assessing the potential £or fuel damage and radiological release for the transient .

PAGE 102

~2.3.1 Power Peaking Factor The core average power history predicted with the point kinetics model is weighted by a hot spot total power peaking factor CFq) shape before input to the hot spot calculation. The design uncertai~ty on Fq as predicted by Vepco's steady state physics models is presently 7.5%. The power histories for the two sensitivity study cases were increased by this factor (7.5%). The results as presented in Table 3-3 show an increase in the temperatures and average enthalpy for both cases.

3.2.3.2 Power History Shape One of the point kinetics neutronics sensitivies investigated was the of r~d ejection which is nominally set at 0.1 seconds for the rod be completely ejected from the core.* To moa.'el a similar type of sensitivity with the Hot Spot Modei, the hot spot power history was displaced 0.1 seconds later in time; i.e., the time of the peak power and all later powers were increased by 0.1 seconds. As seen in Table 3-3, the sensitivity was minimal.

3.2.3.3 Gap Heat ?ransfer Coefficient For a HFP case, the gap heat transfer coefficient CHGAP) is initially adjusted to produce a target initial radial temperature distribution for the fuel pin. At 0,1 seconds into the transient, this value is increased to 10000 Btu/ft 2 -hr-°F and held there for the remainder of the transient to model the expected closure of the gap. Since the initial value of has no impact on the pin's initial radial temperature distribution.

PAGE 103

~~ a HZP case, the value of HGAP is maintained at 10000 Btu/ft 2 -hr-°F for th~ entire transient in order to simplify the analysis. (Additional sensitivity studies showed this change to have no significant impact on the Hot Spot Model's predictions for HZP cases.)

For the sensitivity investigation, the values of HGAP used in the nominal analysis were reduced by 10%. This would be expected to lead to higher fuel temperatures since the ability of the fuel to transfer heat to the coolant has been reduced. The results presented in Table 3-3 show little significant impact from such a reduction.

3.2.3.4 Fuel Pin Geometry Description fuel pin geometry of the Hot Spot Model is described with ten oncentric fuel regions, a single gap region, and three concentric cladding regions. To test the sensitivity of this geometry, the number of fuel regions was reduced from ten to six. The results shown in Table 3-3 show little impact on the model's predictions.

An additional. note is required on the percentage of fuel melt which is shown for the HFP case in Table 3-3. Fox the sensi~ivity cas.e this is listed as <11.1  ?. while for the nominal case it is given as <9 %.

Although the acceptance criteria for percentage of fuel melt is <10 %,

the six node sensitivity case value of <11.1  % does not necessarily imply that the acceptance criteria was violated. Due to the closeness of the comparisons of the average fuel enthalpy and fuel temperature histories between the nominal and sensitivity cases, the percentage of melt for the sensitivity case is indeed <9 %, (see Figure 3-15.)

PAGE 104

~ditional confirmation is presented in Figure 3-16 which provides a comparison of the radial fuel temperature distribution through the pellet for the sensitivity and nominal cases at ~he time of maximum fuel melt.

The simple technique used for determining percentage of fuel melt with the RETRAN Hot Spot Model merely puts ~pper and lower bounds on the percentages. As shown in Table 2-6, the estimate of maximum fuel melt being <9  % for the ten fuel region nominal case is derived from the fact that the highest numbered node £or which the transient fuel temperature did not exceed the assumed melting point was node 4. For the six fuel region sensitivity case, this table is invalid. The highest numbered fuel node for the si~ fuel region case which did not exceed the assumed lting temperatura was node 3. Applying the methodology used to derive Table 2-6 to the six region case gives a percentage of fuel melt somewhere between 2.8 ~ and 11.1 %.

3.2.3.5 Inlet Temperature The initial inlet temperature to the hot spot volume was increased by 5

~F for the sensitivity study. As expected the maximum temperatures for both cases showed a slight rise over the nominal values (Table 3-3), but overall the effect was insignificant.

3.2.3.6 System Pressure The initial system pressure was decreased by 30 psia. This produced a

~light increase in the maximum clad temperatures for the sensitivity

PAGE 105 ses (Table 3-3), but again the overall effect was minimal.

3.2.3.7 Mass Flow Rate Reducing the mass flow ~ate throughout the Hot Spot Model by 5?. produced the most impact on the clad temperature where it would be expected to show the most notica*bie effect (Table 3-3). The impact on the pellet centerline temperature, as shown in the HZP case results, was insignificant.

3.2.3.8 Metal-Water Reaction The RETRAN Hot Spot Model includes a calculation of the additional release due to ieaction between the coolant the the Zircaloy of the cladding, Turning off this option reduced the temperatures and enthalpy predicted by the model as expected (Table 3-3). Again the sensitivity was more pronounced in the clad than in the fuel .

PAGE 106

• TABLE 3-3 HOT SPOT SENSITIVITY STUDY I. HZP Studies:

% Max. Max. Max.

Fuel Tel Tclad Enthalpy Pa:r:amete:r: Sensitivity Melt (OF) (OF) (Btu/lb)

---

~- ---------------------

Nominal Case Values Not Applicabl~ 0 4017 2460 266 Powe:r: Peaking Factoz Inc:r:eased 7.5% 0 4205 2609 285 Powe:r: Histo:r:y Shape 0.1 sec delay of peak 0 4018 2463 266 Gap Heat T:r:ansfe:r: Reduced 10% 0 4018 2456 266 Coefficient i of Fuel Pin Meshes Changed £:r:om 10 to 6 0 4012 2471 266 Inlet Tem~e:r:atu:r:e +nc:r:eased 5 °F 0 4020 2462 266 ystem P:r:essu:r:e Reducgd. 30 psia 0 4018 2474 266 Mass Flow Rate Reduced 5% 0 4019 2496 267 Metal-Wate:r: Reaction Tu:r:ned off 0 4014 2372 263 Abb:r:eviations:

% Fuel Melt= Maximum% pellet melting at hot spot Max. Tel = Maximum pellet cente:r:line tempe:r:atu:r:e at hot spot Max~ Tclad = Maximum cladding tempe:r:atu:r:e at hot spot

PAGE 107

• TABLE 3-3 (cont.)

~OT SPOT SENSITIVITY STUDY II. HFP Studies:

% Max. Max. Max.

Fuel Tel Tclad Enthalpy Parameter Sensitivity Melt (OF) C° F) (Btu/lb)

Nominal Case Values Not Applicable <9 4903 2285 317 Power Peaking Factor Increased 7.5% <9 4904 2358 328 Power History Shape 0,1 sec delay of peak <9 4903 2297 318 Gap Heat Transfer Reduced 10%. <9 4904 2304 322 Coefficient i of Fuel Pin Meshes Changed from 10 to 6

<11.1 4902 2296 316 Inlet Temperature Increased 5 °F <9 4903 2288 317

~ystem Pre~~ure Reduced 30 psia <9 4903 2298 317 Mass Flow Rate Reduced 5% <9 4903 2319 318 Metal-Water Reaction Turned off <9 4903 2239 315 Abbreviations:

% Fuel Melt= MaHimum % pellet melting at hot spot Max. Tel = naximum pellet centerline temperature at hot spot Max. Tclad = Maximum cladding temperature at hot spot

• As described in Section 3.2.3.4, the actual% of fuel melt for the sensitivity case is <9%.

FULL POWER HOT SPOT CENTERLINE TEMPERATURE E 3-15 SENSITIVITY STUDY -

• T SPOT FUEL PIN GEOMETRY FULL POWER HOT SPOT AVERAGE ENTHALPY 5000 320 310 4900 300 4600 C 290 E

. N ..

T 4100 E 260 ft L E I N N 4600 T 210 E H A

T L E P 260 H 4500 .y p

E 6 ft T 250 A u T *HOO I u L R B 240 E

D 4300 E 230 G

f 4200 220 210 4100 200 4000 0 2 3 4 5 6 '1 B . 9 10 0 2 3 4 5 1 8 9 10 TJtlE I SEC I TlnE ISECI STAii: NOHINAL TEN fUEL REGION CASE STAR: NOHINAL TEN FUEL REGION CASE SQUARE: SENSITIVITY SIX fUEL REGION CASE SQUARE: S~NSITIV!JY SIX FUEL REGJON CASE ......

0 FIGURE 3-ISA FIGURE 3-158 OJ

PAGE 109

• FIGURE 3-16 SENSITIVITY STUDY -- HOT SPOT PELLET TEMPERATURE DISTRIBUTION 4750 4500 4250 F

u E 4000 L

T E 3750 11 p

E R 3500 3250 E

D E 3000 G

F 2750 2500 2250 a.a 0.1 0.2 o.3 o.4 o.s- o.s o.7 a.a o.s 1.0 NORMALIZED PELLET RADIUS STAR= 10 NODE PELLET GEOMETRY SQUARE= 6 NODE PELLET GEOMETRY

PAGE 110

~ SECTION 4 - VERIFICATION COMPARISONS 4.1 Introduction The purpose of this report is to document Vepco's analytical capability for performing reload core safety and licensing analysis for the rod ejection transient. As .verification of this capability, appropriate results and comparisons are provided for a representative series of analyses of licensing transients.

A description of the licensing methodology and codes presently used by the vendor (Westinghouse) is provided below. These methods are currently approved for the licensing of Vepco reload cores and were implemented by to analyze six specific cases of the rod ejection transient--the cases reported for the reload analysis of Surry 1 Cycle 5 CBOL HZP, BOL HFP, EOL HZP and EOL HFP), (Ref. 19), and two BOL cases CHZP and HFP) documented in the Surry positive moderator coefficient submittal (Ref 22).

Results from the Vepco analyses performed with the vendor's codes and techniques are compa~ed to the vendor's results for the same analyses to demonstrate Vepco's ability to replicate vendor results. A comparison is then made between the licensing results obtained with the vendor methodology (using vendor codes) to those obtained with the Vepco methodology using the RETRAN code.

A ~omparison is also presented between the Vepco calculated results the v~ndor and Vepco methodologies and results obtained by Vepco

~sing

PAGE 111

• ing the vendor's three-dimensional space-time kinetics code. This comparison demonstrates the conservatism of the Vepco approach to the rod ejection analysis,

PAGE 112 2 Vendo% Licensing Methodology The vendo% licensing analysis of the %Od ejection t%ansient pa%allels that of Vepco in that the analysis may be b%oken into two main phases:

1. a co%e ave%age nuclea% powe% t%ansient calculation, and

2. a t%ansient the%mal-hyd%aulics analysis of the co%e's hot spot.

Unlike Vepco, whicn uses a point kinetics model to de%ive a co%e ave%age nuclea% powe% histo~y, the vendo% uses a one-dimensional (1-D) space-time kinetics model, the TWINKLE code, (Ref. 18.) TWINKLE is a space-time neut%on diffusion code which uses nuclea% mac%oscopic C%oss sections gene%ated by the vendo%'s steady state physics design codes to solve the neut%on diffusion equations fo% two ene%gy g%oups. The code's eomet%y may be specified in one, two 0% th%ee dimensions with up to 2000 spatial points. Six delayed neut%on g%oups a%e assumed. A detailed multi%egion, t%ansient fuel-clad-coolant heat t%ansfe% model is included fo% p%edicting pointwise Dopple% and mode%ato% feedback effects.

The vendo% uses the 1-D axial. geomet%y configu%ation fo% the %od ejection analysis (Ref. 17). Input fo% t~e model is specified to bound the conditions fo: a given nuclea% powe% plant. This app%~ach has been accepted fo% p%oviding the co%e ave%age powe% fo% the t%ansient. The delayed neut%on £:action, Dopple% %eactivity feedback, t%ip inse%tion cha%acte%istics, t%ip setpoints, and total t%ip %eactivity wo%th a%e conse%vatively adjusted to fit the co%e conditions and plant cha%acte%istics fo% the specific analysis unde% conside%ation. The

~ o d e s a to, feedback effects a%e adjusted by changing the value of the

PAGE 113

~luble boron concentration. As with the Vepco methodology, a weighting factor is applied to the Doppler reactivity calculation to more accurately model the effects of severe flux redistribution in a three-dimensional geometry.

Despite having an explicit representation of the axial core geometry, the ejection of the rod is not modeled by perturbing the nuclear cross sections in a space-time dependent manner. Instead the code's eigenvalue is ramped over a 0,1 second time interval by a value equivalent to the assumed ejected rod worth. Upon reactor trip, (with an appropriate trip time delay), a modeling of the trip banks being inserted into the core is performed. This accounting of the axial effects is the major difference between the TWINKLE model and the RETRAN point kinetics model here in the latter the scram is based on a generic normalized trip reactivity table.

The vendor thermal-hydraulics anal~sis of the core hot spot is performed with a detailed fuel and clad transient heat transfer code, FACTRAN, (Ref. 9.) This co~e computes the transient temperature distribution in a cross section of a m~tal clad uranium dioxide fuel rod, and the heat flux at the surfaoe of the rod, using as input the normalized core average power history predicted by the TWINKLE code and the local coolant conditions £or the actual plant and core conditions under consideration. The tuel rod geometry is input for the particular plant being analyzed. A Zircaloy-water reaction is modeled, and all material properties are represented as functions of temperature. The fuel rod re I . assumes an initial parabolic radial power generation.

PAGE 114 code uses the Dittus-Boelter or Jens-Lottes heat transfer correlations to determine the surface heat transfer characteristics before the onset of DNB. AT 0.1 seconds into the transient, the code is forced into DNB by specifying a conservative DNB heat flux, and the Bishop-Sandberg-Tong correlation is used to determine the film boiling heat transfer coefficient. The gap heat transfer coefficient is adjusted in order to provide agreement of the full power steady state temperature distribution with the distribution predicted by Westinghouse design fuel heat transfer codes, As the transient progresses, the gap heat transfer coefficient is ramped from ~ts initial value to a very high value to model the gap closure expected to *occur during the temperature transient (Ref. *17).

with the Vepco methodology, the pre~ and post-ejection design maximum total power peaking factor (Fq) is input. The value of Fq used by the code is assumed to increase from the pre-ejection value to the post-ejection value over a time interval of 0.1 seconds, and remain there £or the duration of the transient.

In -

summary, the analytical techniques in modeling the transient of both Vepco and the vendor are s~milar. Table 4-1 presents an overview of the two methodologies. Table *4-2 presents the results obtai.ned £or the six comparison cases using the TWINKLE/FACTRAH codes with vendor methodology performed by both the vendor and by Vepco. The results from the vendor analysis are provided in Ref. 19 for the Surry 1 Cycle .5 reanalysis cases and Ref. 22 for the Surry +MTC cases. The close agreement bet~een results demonstrates Vepco's ability to perform th~ rod ejection l

PAGE 115

. a . l y s i s using the TWINKI.E/FACTRAN codes and vendor methodology .

PAGE 116 TABLE 4-1 COMPARISON OF VENDOR/VEPCO LICENSING METHODOLOGIES I. Core Average Po~~r History Calculation Item Vepco Vendor Principal code RETRAN TWINKLE Physics model Point kinetics 1-D space-time kinetics Rod ejection Reactivity ramp Reactivity ramp Doppler feedback Reactivity vs. temp. Calculated from nuclear table cross sections Moderator feedback Moderator temp. Calculated from nuclear coefficient cross sections Trip reactivity insertion Reactivity vs. time Calculated from nuclear table cross sections el~yed neutron groups 6 6 Doppler reactivity Yes Yes weighting factor Fuel heat transfer.model Yes Yes HGAP Constant Variable t

PAGE 117 TABLE 4-1 (cont.)

COMPARISON OF VENDOR/VEPCO LICENSING METHODOLOGIES Ir~ Hot Spot Thermal-hydraulics Calculation Item Vepco Vendor Principal code RETRAN FACTRAN Number of fuel pellet 10 6 regions Fq history Ramped over 0.1 sec Ramped over 0.1 sec Zircaloy-water reaction Yes Yes Pre-DNB heat transfer Thom Dittus-Boelter or correlation Jens Lottes Post-DNB heat transfer Bishop-Sandberg-Tong Bishop-Sandberg-T9ng orrel~tion

• orced DNB at 0.1 sec Yes Yes Material properties Function of temp. Function of temp.

Gap closure modeled Yes Yes

PAGE 118 TABLE 4-2 VENDOR/VEPCO ANALYSIS RESULTS USING VENDOR METHODOLOGIES (TWINKLE-FACTRAN CODES)

The vendor calculated value is followed by the Vepco calculated value separated by a slash(/).

Surry 1 Cycle 5 Values Parameter I BOL HZP BOL HFP EOL HZP EOL HFPI Fuel Pellet Melting (,,.) 0/0 <10/<10 0/0 <10/<10 Max. Fuel Aver.age Temp. (op) 3430/3432 4185/4210 3373/2949 3844/3794 Max. Clad Average Temp. (Op) 2500/2520 2488/2482 2317/2169 2151/2167 Max. Fuel Enthalpy (cal/g) 145/145 185/186 142/121 166/164

~ Surry +MTC Values Parameter I BOL HZP BOL HFP I Max. Fuel Average Temp, C°F) 2883/3394 3639/3630 Max. Fuel Center Temp, C°F) 3353/3918 4958/4874 Max. Clad Average Temp, C°F) 2123/2488 2013/2079 1*.

PAGE 119 3 Verification With Licensing Analyses Six cases of the rod ejection transient were compared for benchmarking the Vepco methodology with that of the vendor. Assumptions for each comparison calculation were matched as closely as possible between the two methodologies.

Due to the greater flexibility of specifying reactivity effects in the Vepco RETRAN model, the steady state values of the Doppler defect (without any additional weighting factor)~ moderator temperature coefficient, and total trip rod worth were first calculated by Vepco using the TWINKLE code; these calculations used nuclear cross section input chosen to give Gonservative values for these parameters. These were then ~sed in the corresponding RETRAN. point ~inetics*

alculation. An exception was the reactivity feedback weighting factor used for each. methodology. For the Vepco calculation, the weighting factor CPWF) was obtained from Figure 2-6 based on the value of Fq assumed for each analysis. The corresponding weighting factor used in TWINKLE was derived from the data in Ref. 19 for the Surry 1 Cycle 5 reanalyses cases and from Ref. 22 for the Surry +MTC cases. Table 4-3 presents a summary of the input conditions for the six cases .

PAGE 120

• TABLE 4-3 VERIFICATION COMPARISON CASES S1C5 S1C5 S1C5 S1C5 S+MTC S+MTC BOL BOL EOL EOL BOL BOL Parameter HZP HFP HZP HFP HZP HFP Ejected rod worth Cpcml 900 300 840 300 920 160 Delayed neutron fraction .0055 .0055 .0044 .0044 .0059 .0059 Pre-ejection Fq NA 2.55 NA 2.55 NA 2.55 Post-ejection Fq 15.2 5.46 15. 2 5.46 15.2 4.8 Zero to full power Doppler -1088 -1088 -850 -850 -1088 -1088 defect (pcm)

RETRAN moderator temperature 7.7 5.4 -20.0 -30.8 7.7 5.4 coefficient C:ecm/°F) mber of operating pumps 2 3 2 3 2 3 TWINKLE reactivity feedback 2.725 1. 3 2.32 1. 3 2.725 1. 2 weighting £actor RETRAN Power Weighting Factor 2. 95* 1. 4 2.95 1. 4 2.95 1. 25 Notes: S1C5 = Su:i::ry , Cycle 5 S+MTC = Su:i::ry Positive Moderator Coefficient NA = not applicable pcm = percent mille

PAGE 121

~3.~ Core Average Power History Results from the nuclear power transient analyses for the siK cases are graphed in Figures 4-1 through 4-6. Each ~igure presents a normalized power history comparison from the TWINKLE 1-D and RETRAN point kinetics calculations and the corresponding normalized energy release curves. As with the sensitivity study results presented in Section 3, a value of 1.0 on the "normalized power" aKis represents a value of 100% full power while a value of 1.0 on the "energy release" axis represents one-full-power second of energy. For the HZP cases, both the TWINKLE and RETRAN models assumed an initial normalized full power level of 1.0E-9.

The three HZP cases, (Figures 4-1, 4-3 and 4-S), show excellent greement between the power histories. The minor differences occurring fte~ the initiation of the -trip can be explained by the different methods used in modeling the trip reactivity insertion. Whereas RETRAN uses a table of normalized trip worth insertion versus time which is identical for each case, TWINKLE calculates the change in core reactivity through an axial variation of nuclear cross sections with time. This is a more realistic modeling of the insertion of the control banks and takes into account the impact of the axial flux shape on the rate of change of reactivity.

Although the ~aximum normalized core power predicted by the two methods may differ for a apecific case, the effect is not ~ignificant due to sharpness of the power peaking portion of the curve. This is verified in the plot of total energy release where both models track reasonably After the trip insertion begins, the normalized power curve

PAGE 122

~edicted by the TWINKLE code asymptotically approaches a lower power level relative to that predicted by RETRAN and this accounts for the eventual divergence of the energy release curves.

For the three HFP cases. (Figures 4-2, 4-4 and 4-6), the divergence between the two models is more pronounced. Typically, TWINKLE predicts a higher peak power level and maintains the power at a relatively high value longer than t~e RETRAN model. As with the HZP cases, .a deviation in the predictions is noticable during the trip portion of the transient with TWINKLE once again settling out at a lower relative power level.

The anergy release ourves show reasonable agreement at the most critical part of the transient between the two models with a larger divergence in the energy release predictions occurring only after the. power level of e core has dropped to a relatively'low value.

Comparison of the trends inherent in the three HFP cases show the effect of the delayed neutron fraction and ejected rod worth on the transient.

The BOL and EOL cases for the Surry 1 Cycle 5 reanalysis, (Figures 4-2 and 4-4), both assumed the same ejected rod worth. The only significant difference betwee~ the two cases was in the delayed neutron fractions, C0.0055 £or BOL versus 0.0044 for EOL), with the higher value causing the power curve to hang, (this is due more to delay in the core response than to changes in core conditions.) That this effect is due entirely to the delayed neutron fraction is _especially obvious from the RETRAN analysis where it was the only parameter which changed between ~he two calculations .

• comparison between the two BOL HFP cases, (Figures 4-2 and 4-6),

PAGE 123

~monstrate a similar trend. Here, the difference in the values of the delayed neutron fractions is less pronounced than in the previous example, but the S+MTC case has a significantly lower ejected rod worth than the Surry 1 Cycle 5 case. This results in less reactivity insertion into the core which in turn causes less of a rise in fuel temperature and therefore a less pronounced negative Doppler reactivity feedback effect to turn the ttansient's power excursion around.

Considering the differences in the two models used, the power history predictions compare well. Both show the same characterics for the transients and demonstrate reasonably close agreement in actual magnitude .

NORMALIZED POWER HISTORY NUCLEAR PO\ffR TR/\ T ~ $1C5 BOL HZP NORMALIZED ENERGY RELEASE 1 .6 I .6 1,4 N

o*

II +(a,c)

• "A I .2 L

I z

[

N D I .0 0

R [

"LA N

[

R O.B I G z I .OO y*

r 0 R

• r 0*.6 r L 0 [

H A

.r . 5 R E 0.4 0 .10 0.2 o.o o.s, 1.0 1,5 z.o 2.5 1.0 3,5 4,0 4.5 5.o TIME ISECI.

WESTINGHOUSE PROPRIETARY *cLASS 2 O.O 0.5 1.0 1.s LO z.s J.O J.S 4.0 4.5 s.o

~- TIME 1src1 , ..

SIAR: TWINKLE 1-0 MODEL STAR: THINKLt 1-0 HODEL SQUARE: ~EIRAN fOINI IIINEJICS MODEL SQUARI,..* REIRAN POINT IIINEIICS MODEL f IOUR[ 4-Hi .

FIGURE 4-IA

RE 4-2 NUCLEAR POWER TRANSIENT sics BOL HFP.

NORMALlZED POWER HISJ'ORY NORMALIZED ENERGY RELEASE 2-25 3-6

, +(a,c).

2.00 ].3 3.0 I .75 2.1 H

I .SO 0 N R 0 11 2 **

R II n L A I L z 2-1 I E z D

[

0 [ I .e

,. N

[

0 I\

M G y

I .s

[

R I\

[ I .2 L

[

R*

5 0.9

[

0°6 O.]

O.O 0.5 1.0 1.5 2.0 2.5 J.O J.5 LO 4.5 5.0 Mn£ 16CC I o.o 0.5 ,.o 1-5 2.0 2.5 ,.o 3.5 4.0 ~-5 5.0

.... 1 IHE I SEC I WEST.IN~HOUSE PROPRIETARY CLASS 2 SIAR: TMINKL[ 1-0 noD[L N

.SQUARE: REJRRN rDINT K!N[t!CS noO[L STAR: TWINKLE 1-D nODEL u, SQUARE: RE1RRN rolNT KIN[IIC& HODEL nGURE 4-1!11 flGURE 4-26

~UCLEAR POWER TR IIT S1C5 EOL IIZP

.. NORMALIZED POWER HISTORY NORMALJZED ENERGY RELEASE 100.00 I *6 .

I .6

I I *4 ff 10.00 0 R

ti

• A I .2 L'

I H l

[ .

0 R D I .o I .

ti A t L H

[

I l I .oo R .0 .e

[ C ..

. ** .y 0

II .

.r [ 0.6 0

II L

[ [

R A 5

[ 0. 4 0 .10 o.z 1.5 z.o Z,5 J.O *3.5 4.0* 4.5 5.0 I Int I &EC I 0.01 I *

• I B' I I

• I H~STINGHOUSE PROPRIETARY CLASS 2 o.o o.s 1.0 _,. 1.5 i.o z.s

• J.o* J.5 Lo 4.5 s.o

. T Int I src I

~*

!.1,1/1: TMIHKL[ 1-0 HOD(L s1n11: JIIJHKL[ 1-0 noDCL SOUAR[: RCIRAN roJNr KIN[ilC6 nootL 60UAR~; R[IIIAN ro!NJ KIN[IICS HODEL lr!CUIIIE ~-:llR f!OUII[ 4-Jlil

FIGURE- 4-4 NUCLEAP. PO~IEP. TP.A;*IS I S1C5 EO~ HFP

\i/ESTINGHOUSE 'PROPRIETARY CLASS 2

. NORMALIZED POWER HISTORY NORMALIZED ENERGY. RELEASE

].00 ],25 1.15

• I I I

],00 I!\ I z.so 2,15 z.zs 2,50 +(a, c)

• Z.00 N

0 2,25 .

R 11 R

,I I. 75 l l

I 2.00

[.

'.so D I ,15

. E

~ H .

• 1.25 *r I ,50 0 fl M. G

. T

• '.oo- I -25 R

E l

0.1s E I ,DO R

5 E

o.so 0,15 o.so I 0.25 *1 I

0,25 I'

i 0,0 0.5 1,0 1.5 i.o 2,5 l,O ].5 4,0 4,5 5,0  !.

• 0,00 TlnE 16ECI 0,0 0,5 I ,0 I ,5 2,0 Z,5 ],0 ].5 4,0 4 ,5 5,0

~- fl11E l&EC I

-0 SlRfl: TMINKL[ l*D HODEL  :,,*

S1RR: iMBHKl[ a-o nOO[l SOURRE: REJRRH rDIHT KINE1!CS noocL Gl SOURRE: REIRRN rDIHV KINETIC& HDDEl m ' .

lrDGll.ll\E ~-llllil II-'

flGURE 4*4R '

-~ -~-~-------- N

...... i,'

FIGURE 4-5 NUCLEAR PmlER T Elf( - *s+MTC BOL HZP-

.. NORMALIZED POWER HISTORY NORMALIZED ENERGY RELEASE I *0 I ,6 I *4 N

0 R

M A I ,Z L +(a,c)

I z

[

N D 1,0 0

R E M N A [

L I " o.o 0

l l .oo y E

D R

p E O ,6 0 L

[

If

• E R

• s

" E O ,4 0,10 o.z 0,0 0,5 I .O 1,5 Z.O 2,5 3,0 3,5 4,0 4,5 5,b TIME ISEC I l*JESTINGHOUSE PROPRIETARY CLASS 2 0,0 0,5 \,D 1,5: 2,0 :Z.5 3,0 3,5 4.0 4.5 5,0 I

TIME !SEC I *!

~- 61 RR: TM I NKLE 1-0. MODEL STAR: THINKLE 1-0 MODEL SQUHhi: RETRRN POINT KJNETIC6 MODEL SQUARE:- RE TRAN POINT Kl"NET IC& MODEL FIGURE 4-!iR FIGURE 4.:.59

NUCLEAR POWER HU\N - *s+MTC BOL HFP

.HESTINGHOUS_E PROPRIE1:AR_Y CLASS 2 NO;RMALIZED POWER HISTORY .NORMALIZED ENERGY RELEASE

].00 I .4 1111 ,,

I .3 I JIJ;'.*1 j Z .15 11,.1,1

, 111.':

I .z ' ...

.z .so I .1 +(a,c) z.zs I .o If 0

• 0 N

R 0.9 "

n z.oo A

"R . o.e L I

L l I .15 I [

Z 0.1 0

[

0 .[ I .so o.& H r E 0 R&

. M o.s c;* I .zs

£" y R

0. 4 II E I .OO L

o.J E A

5 D .15 o.z [

o., o.so 0.25 o.o o.s ,.o 1.5 z*.o z.5 :,.o l.5 4.0 4.5 s.o-1111E lfiECI.

o:o o.5 1.0 1.s z.o 1.s J.o l,& 4,0 c.5 s.o 1111£ UCCI

-~*

...

• SIAR: TMINKLE 1-0 MODEL .

... . S rAII : TM I NKLIE i :.o 110Dtl . &QUAIi[! RCIRRN rDINW K!N[l!t5 nOD[L SQUAR[I IIEIRAN ro!Ni KINETIC& 110D[L r!GURE 4-H FIGURE 4-6R

PAGE 130

• 3. 2 Hot Spot Anal!a"sis Results of the comparisons between the FACTRAN and RETRAN Hot Spot Model predictions for the hot spot transient are presented-in Figures 4-7 through 4-12 and Table 4-4. All FACTRAN calculations were performed based on the appropriate TWINKLE 1-D model predicted core average power histories, while RETRAN Hot Spot Model calculations were performed based on the predicted RETRAN pdin.t kinetics power histories. Both hot spot models used identical assumptions as to the Fq history curve input; i.e., between a time of zero and 0.1 seconds the Fq value was linearly ramped from the pre-ejection value to the post-ejection value and held there for the remainde~ of the transient.

igures 4-7 through f 4-9 plot the fuel centerline transients at the hot pot location. The curves for the HZP cases* demonstrate good agreement and follow the trends predicted by the TWINKLE/RETRAN core average power history predictions--i.e., RETRA~ predicts a higher fuel temperature than FACTRAN later into the transient just as it predicts a higher power level than TWINKLE at that point. In none of the HZP cases did any fuel melt occur.

The corresponding HFP cases (Figures 4-7 through 4-9) demonstrate more deviation, especially in the two Surry 1 Cycle 5 cases where fuel melt occurred. For the RETRAN cases, upon reaching the assumed fuel melting temperature, (4900 °r for the BOL case and 4800 °F for the EOL case),

the temperature remained constant for some time due to the high heat of fusion required to melt the fuel, (Ref. 8). Although FACTRAN assumed the ame fuel melting ~emperatures a~ RETRAN, it apparently has a smaller

PAGE 131 lue £or the heat 0£ fusion 0£ uranium dioxide. Thus the FACTRAN temperature can at first level 0££ at the melting point, and resume its rise sooner than the RETRAN Hot Spot model temperature. The main differences between the two predictions appear to be in the material properties tables used by the two codes. To check this, the TWINKLE power history used by FACTRAN was input to the RETRAN Hot Spot Model.

The RETRAN predicted temperature transient characteristics were consistent with those predicted by the RETRAN Hot Spot Model using a RETRAN point kinetics power history input.

Figures 4-10 through 4-12 are comparisons 0£ the outer clad temperature transients predicted by the vendor and Vepco models. Again agreement is better £or the HZP cases, although £or all cases the curves show

~ceilent convergenca late in the transient. The slight deviations between the predictions at the very start 0£ the transient (where the FACTRAN predicted clad temperature actually decreases in value) are due to modeling differences between the two models 0£ the gap and surface heat trans£e~ coefficients before the onset of film boiling.

Table 4-4 is a summary of the maximum values of the hot spot temperatures, average enthalpies and percentage of fuel melt for the two models. As de~cribed ab~ve, two Surry 1 Cycle 5 HFP cases exp~rience fuel melt. For these cases, the FACTRAN te~perature rises above the assumed melting temperature while the RETRAN temperature has yet to overcome the heat of fusion input into that model's material properties tables. The FACTRAN code predicts an actual percentage of fuel melt while RETRAN predicts only an upper boundary based on Table 2-6.

• FIGURE 4-7

• HOT SPOT FUEL CENTERLINE TEf1PEPJ\TURE TRANSIENTS - S1C5 BOL CASES HOT FULL POWER HOT ZERO POWER 5000 uoo 3900 UDO 1600

,1100 t

,100 £ t

[

H "t noo

[

I 1000 II i [ l

1. II l .

I N *600 I 2100 [

N

£ 1

[

uoo '

[

,",soo

", 2100

£

[

II fl II 1 uoo fl u U

II 11100 *

[

[ D 000

[

D 1500 Ii

[

Ii f' uoo r 1200 900 uoo 600

. 10001.--...---,.-.......---,r---i---r---r---,--...---,.

o.o o.s !.O

• I .S 2.0 2.s ,.o ,.s e.o ,.s s.o I-'

w IIN[ I lite I l'v o.o o.s , .o , .s 2.0 2.s 3.0 1.s ,.o a.s s.o 1!ft£ 16EC\ SIAII. rAtlllAN COO[

Slflll: f'IICTRIIN COO[ SOUR~[. ~llllAN HOT s~o, ftOO[L SQUAii[: ft[IRflN HOI &POI NODfl f"!liUIIC 8-111 f'J Glllllt 4-111

• FI r,u~ : 4-8 HOT SPOT FUEL CENTERLrnE TEMPERATURE TRANSIENTS - S1C5 EOL CASES HOT ZERO POWER HOT FULL . ' POWER uoo UDO

&850 3900 HOO 3600

,1so l300 t C [

[ N N T *100 T lOOO [

[ II II L L I *&SO 1 noo N

[

N

[

T *&DO T HOO [

E n

n,.

[ *SSO E 2100 II II A A

1 U 11100 u *soo II

[

II

[

DUSO D !SOD [

[ G G

r uoo r 1200 USO 100 noo 600 e2so-l..~......~....,.........-.i,--~~~...-~"T"""~-r-..........,..~...,.~--r 300 l...~....-~...-~--r~--r--""""1,.....~,.....--"T""~'"'T"'~....,..---....,-

.0.0 0,5 I .O I .S 2.0* 2,5 3.0 3.5 *.O 4.5 5,0 o.o o.s I .Q I .S 2,0 Z.S 3.0 J.5 4.0 &.5 5.0 fin[ UCCI unt 1&Et, SIAII. fACIIIAN COO[

SQUAii[, -(IIIAN HOI SP'OT HODEL StAII: FACTRAN CODE 50UAR[: R[JRAN HOT SP01 HODEL f!GUlll a-1111 f JGURE 6-!JA

FIGURE 4-9 HOT SPOT FUEL CENTERLINE TEMPERATURE TRANSIENTS - S+MTC Cl\SES HOT ZERO POWER HOT FULL POWER UDO UDO

&150 3900

&100 3600

&'150

noo C t E

£ N N I noo I 3000 E

£ II II l l I 4650 I 2100 H N E

£ I 4600 I 2&00 E

[

"Er r,oo "[r 4550 II II A A 1 I U tSOO U 1100 II II [

£ DUSO 0 1500 [

£ c; c;

r uoo r 1200

&350 900

&300 600 w

o.o o.5 1.0* ,.s 2.0 2.s ,.o 3,s ,.o ,.s s.o -+=>

0.0 0,5 I .O* I ,5 2,0 2,5 3.0 3,5 4.0 &.5 5,0 11"£ I S[C I TIHC l&ECI SIAR: F~CTIIAN CODE*

SIAR: FACTIIAN CODE* SQUARE: REIRAN HOT 6Ptl*HOD[L SQUAR[; ll[IRAN HOT*srot f'J GUii[ 4-!lll'l flGUIIE &-98

F-~

11

.E 4-10

~1.. URE TR/1.NSIE!'ITS - SlC5_ BIJL CASES HOT SPOT FUEL OU_TEr., CLAD TE HOT ZERO POWER HOT FULL POWER 2400 2400 2200 2200 2000 2000 0 D u u T T E E 1600

!BOO "C "

C L L R A 1600 D 1600 D T 'T E E N 1400 "E 1400 p p E

R R R R T T 1200 u 1200 u R .

R E E.

D D 1000 E 1000 E G G t . r BOO BOO 600 400 o.o 0,5 1,0 1,5 2.0 2,5 3,0 3,5 4,0 4,5 5,0 o.o 0,5 I ,0 I ,5

• 2 ,0 2 ,5 3 ,0 3 ,5 4 ,0 4,5 5,0

-0 T.IHE I SEC I ):>

TINE 16EC I Ci>

rTl STAR: FACTRAN CODE STAR: FACTRAN c*oor

-SQUARE: RETRAN HOJ 6POT NODEL SQUARE: RETRAN HOT &PUT HODEL ....

w FIGURE 4-IOA . 'flOURIE 4-!0B u,

FIC'~-11 HOT SPOT FUEL OUTER CLAD TEf1P-E TRANSIENTS - SlC5 EOL CASES HOT ZERO POWER HOT FULL POWER 2200 2100 2000 1900 1800 0 0 u u T T 1100 E E R 1150 II 1600 C C L L R R 1500 D 1500 D T T 1400 E E n

p 1250 E

"P'E 1300 R It R A 1200 T T u 1000 u R It 1100 E E D D 1000 E 150 E G G 900 f f.

500 800 100 250 600 0 I .O I .5 2 .O 2 .5 3 .O 3 .5 4 .O 4 .5 5.0 o.o o.5 o.o o.5 I .o I .5 2.0 2.5 3.0 3.5 ... o 4.5 5.o TJnE I SECI TJnE I SEC I STAR: FACTRRN CODE STAR: fACTRAN CODE SQUARE: RETRAN HOT SPOT nODEL SUURRE: RETRRN HOT SPOT nODEL FIGURE 4-118 FIGURE 4-1111

FIGURE 4-12 HOT SPOT FUEL OUTER CLAD TE.iURE TRANSIENTS - s+mc CASES HOT* ZERO POWER ,HOT FULL POWER 2100 2,00 2000 1900 2200 1800 0

U 1700 0

T u E II 1600 T

E 1800 C II L 1500 A

C D L 1400 A 1600 T D E' T

E N 1300 E

"p E

R 1200 A

T R U 1100 A II T E u 1000 R D E

• E G 900 f

800 f

700 600 600 500 0,0 D-5 1-0 1,5 2.0 2.5 3.0 S,5 ,.o ,.5 5-0

. ,oo TINE I &EC I o.o 0.5 1.0 1-5 2.0 2-5 3.0 3.5 ,.o ,.5 s.o JIHE l&ECI STAR: fACTRAN CODE STAR: fACTRAN CODE &DURRE: RETRAN HOT SPOT HODEL

&DURRE: RETRAN HOT &POT nODEL f I GURE 4-12B fJGURE *-t2A

PAGE 138 TABLE 4-4 FACTRAN/RETRAN Hot Spot Model Comparisons The FACTRAN calculated value is followed by the RETRAN calculated value separated by a slash(/).

Surry 1 Cycle 5 Values Parameter I BOL HZP BOL HFP EOL HZP EOL HFPI Fuel Pellet Melting (~) o~o <10/<10 0/0 <10/<10 Max. Fuel Center Temp, C°F) 3954/4003 5017/4903 3390/4096 4889/4802 Max. Clad Temp. C°F) 2520/2484 2482/2292 2169/2568 2228/2172 Max. Fuel Enthalpy (Btu/lb) 261/267 334/318 219/278 301/299

__ Surry +MTC Values_

Parameter I BOL. HZP BOL HFP I

---

~--------- --------- ---------

Fuel Pellet Melting C?.) 0/0 0/0 Max. Fuel Center Temp, C°F) 3918/3990 4874/4733 Max. Clad Temp. C°F) 2488/2471 2079/2036 Max. Fuel Enthalpy (Btu/lb) 258/266 -280/275

PAGE 139 3.3 Conclusions Comparisons between results predicted with the Vepco and vendor methodologies for the analysis of the rod ejection event for Vepco nuclear units showed similar trends and close agreement despite differences in the two models. For the cases analyzed in this report, both methods predicted a comparable degree of fuel melting for the cases in which it occurred, but for all cases it was under the 10% limit imposed by the acceptance criteria. Likewise, neither method predicted a violation of the acceptance criteria for the peak clad temper~ture or peak average fuel enthalpy.

PAGE 140 4 Comparison ~o Three-Dimensional Space-Time Kinetics A three-dimensional space-time kinetics code was used to verify the acceptability and conservatism of the point kinetics approach to predict the core average power history. As noted by Yasinsky, (Ref. 20), the accuracy and acceptability of the point kinetics approach is dependent both on the method used and the particulars of the nuclear qore undet analysis. The use 0£ a weighting factor to increase the magnitude of the Doppler reactivity feedback used in the point ~inetics model Casis done with the Vepco methdology) is appropriate as indicated by the verification of the conservatism of the approach.

The TWINKLE code (Ref. 18) was use~ to analyze two rod ejection cases ing three dimensional geometry:

1. BOL, HZP, ejected rod worth of 1024 pcm

2. BOL, HFP, ejected rod worth of 200 pcm The results from these analyses were compared to the results from the RETRAN point kinetics method and the TWINKLE 1-D method for similar transient conditions in order to verify the conservatism of the latter two methods .

PAGE 141

~4.1 Three-Dimensional Model A three-dimensional (3-D) TWINKLE model was constructed for an initial Surry unit fuel loading (Cycle 1) for a core consisting of all fresh fuel assemblies. A quarter core loading map of the cycle is presented in Figure 4-13, and gives the initial fuel enrichment of the three batches Cw/o U235) and number of fresh burnable_poison rodlets Ci BP rods) present in each assembly. Appropriate nuclear cross section data describing the core loading was derived from vendor steady state core physics design codes, and. input to the TWINKLE code following the procedures outlined in Ref. 18. A three-dimensional geometry was specified with one radial mesh point per assembly, 11 axial mesh points per assembly, one mesh point per top and bottom axial reflectors and a

~ d i a l reflector region one mesh point deep at the nearest approach of a peripheral fuel assambly. A.half core geometry model was *used with the full core being split along the vertical axis giving a total of -1989 mesh points. The location of the ejected rod is on the vertical axis through the center 0£ the core and produces a radially symmetric power and flux distribution about the vertical axis. Therefore a half core geometry is sufficient for modeling the transient in three dimensions.

(See Figuie 4-14.) The patterns used for withdrawing the control rod banks from the core from HZP to HFP conditions cause the.ejected rod to be located on an aKis of symmetry.

The steady state radial power distributions at HZP and HFP for an all rods withdrawn core configuration were normalized to those calculated by the vendor for the initial FSAR for the Surry units, (Re£. 21.) This I.

PAGE 142

.rmalization was aarried out by adjusting the fast diffusion coefficient £or the radial reflector region and the macroscopic absorption cross sections £or the various fuel batches until reasonable agreement was obtained with the radial power distributions published in Re£. 21. The resulting power distribution comparisons £or initial steady state, all rods withdrawn core conditions are presented in Figure 4-15.

In a similar fashion, the remaining critical core parameters as predicted by the steady state 3-D TWINKLE model were normalized to values typical of those used in a rod ejection analysis. For example, the zero to £till power Doppler defect was normalized to a value of ~1164 pcm by adjusting a multiplier to the fuel temperature component of the macroscopic fast absorption cross section, Typical values of the derat~r temperature coefficient predicted by the model Mere obtained by adjustin~ the soluble boron concentration input to the code.

Control rod worths were normalized by adjusting the thermal absorption cross section used to model the presence of control rods. Using the values provided by the steady state physics code £or these cross sections and ejecting the rod from the rod inseition limits specified

£or the initial Surry-cycle resulted in ~jected rod worths of such small magnitude that only minor power excursions were produced. To model transients which were more typical of those analyzed £or reload cores, the insertion limits were deepened and the worth of the ejected rod increased by increasing the thermal absorption cross section of the D bank. This resulted in a HZP ejected rod worth of 1024 pcm (percent mille) £or an initial core configuration of the D bank fully inserted

PAGE 143 d the C bank inserted 43.5% into the core. For the HFP case, an ejected rod worth 0£ 200 pcm resulted with the D bank inserted 56.5%

into the core. All other control and shutdown banks were initially out 0£ the core.

For simplicity 0£ modeling, reactor scrams were modeled by inserting the control banks (banks D, C, Band A) less the ejected ~o~. The thermal absorption cross section £or banks A,B and C was modified to yield conservative total trip worths _typical of those used in the rod ejection analysis, (i.e., approximately 2000 pcm for the HZP case* and 4000 pcm

£or the HFP case.)

Table 4-5 presents a summary 0£ the steady state core physics parameters or the final 3-D TWINKLE model used. for the benchmark cases.

The two transient cases were ~nitiated by ejecting the cho~en rod from the core at a constant velocity over a 0.1 seconds time interval.

Reactor trip was initiated on the same trip setpoints assumed in the RETRAN point kinetics analysis and assumed a 0.5 seconds trip delay between activation of the trip and the start of rod motion. The trip rods entered the core using the same rod insertion model as is assumed in the standard TWINKLE 1-D analysis 0£ the transient. Since the 3-D model predicts the effects of three-dimensional flux redistribution during the transient, no weighting factor was applied to the calculation of the Doppler reactivity feedback .

PAGE 144

• • FIGURE 4-13 SURRY UNIT 1 CYCLE 1 CORE LOADING PLAN CEighth Core Geometry)

H G F E D 08 1 09 2 1 12 :ft: Fuel t,J/ 0 Batch Assys U235 10 1 2 1 1 53 1. 85 12 2 52 2.55 3 52 3. 1 0 11 2 1 2 1 12 12 I

12 1 2 1 2 1 I 12 12 .I I

I 13 2 1 2 3 3 I 12 12 12 I I

14 1 3 3 3 12 Legend:

15 3 3 X ==> Batch :ft:

xx ==> :ft: of BP Rods

PAGE 145

• FIGURE 4-14 RADIAL GEOMETRY FOR 3-D TWINKLE H G F E l) C B A

-- --: . ... I ' ' * . . . . . . . . . . . : . . . :

01 I I  ; . ':

02 I_I_

I D I A I_I_  : ... ...: ... :

03 I SA I_ ... : . ..:

04 I B C I_ . .. : . .. :

05 I SB I_ ... : ... :

06 I C D B A Legend:

I_  : ...:

05 I SB SA I I_ _I . . .: I I ==> Fuel Mesh Point I C D I I_I

--* _I . . .:

09 SB SA I

_I . . .: ==> Reflector Mesh 10 C D B A I Point

_I . . . : . . . :

11 12 B SBI I

_I .. . : . . . : Control Banks:

I C D 13 14 X SA A

-'- I I

C B

A SB 15

. .

• t,,,1 . * .

~

. 0 SA X = Location o:f Ejected Rod o**',,,: ... : ... :000:000:

PAGE 146

• FIGURE 4-15 STEADY STATE RADIAL POWER DISTRIBUTIONS Beginning of Life, Hot Ze:ro Powe:r, All Rods Withd:rawn H G F E D 11.130 08 r1*.210

-6.6 1 . 18 0 1 . 12 0 09 1.220 1. 190

-3.3 -5.9 1 . 120 1 . 160 1. 090 10 1

• 17 0 1 . 18 0 1 . 13 0

-4.3 -1. 7 -3,5

1. 150 1 . 100 1. 11 0 1. 0 0 0 11 1.1{50 1. 12 0 1. 090 1. 00 0

0. 0 -1. 8 1,8 0. 0 1 . 100 1 . 130 1. 0 30 0.970 0.7801 12 1. 0 9 0 1. 0 9 0 1.010 0.910 0.7201 0.9 3.7 2,0 6.6 8.3 I I

1 . 140 1. 080 1.04:0 0.920 0.6301 13 1. 07 0 1. 0 40 0.970 0. 910 0.6201 6.5 3.8 7.2 1. 1 1. 6 I I

1.060 1. 12 0 1.000 0.650 14 1. 0 0 0 1 . 120 1 . 08 0 0.670

6. 0 0.0 -7.4 -3.0 Legend:

0. 9'SO 0.710 lx.xxxl ==> FSAR Relative Powe:r 15 0.940 0.760 ly.yyyl ==> TWINKLE 3-D Relative Powe:r

-1. 1 -6.6 I z.z I ==> % Diffe:rence I I

I~

PAGE 147 FIGURE 4-15 (cont.)

STEADY STATE RADIAL POWER DISTRIBUTIONS Beginning of Life, Hot Ful.l Powe:c, All Rods Withdrawn H G F E D 1 . 190 08 1. 240

-4.0

1. 2 30 1.1701 09 1. 240 1.2201

-0.8 -4. 1 I I

1.160 1. 200 1.1301 10 1.200 1. 200 1.1601

-3.3 0.0 -2.6 I I

1 . 180 1 . 120 1.13011.020 1.178 1. 150 1.11011.020

.1 0.9 -2.6 1. 8 I 0.0 I

1 . 11 0 1 . 13 0 1.04010.970 0.7901 12 1. 10 0 1 . 10 0 1.03010.930 0.7501

0. 9 2.7 1. 0 I 4.3 5.3 I I I 1 . 12 0 1.060 1.01010,900 0.6301 13 1 . 0 6 0 1. 0 40 0.960j0.910 0.6401 5.7 1. 9 5,2 I -1 . 1 -1. 6 I I I

.11 . 0 10 1. 060 0.95010,630 14 10.970 1.070 1.02010.660 I 4. 1 -0.9 -6.9 1-4.5 Legend:

I I 10.870 0.670 lx.xxxl ==> FSAR Powe:c 15 10.890 0.730 ly.yyyl ==> TWINKLE 3-D Powe:c 1-2.2 -8.2 I z.z I ==> % Diffe:cence I I I

PAGE 148

• TABLE 4-5 3-D COMPARISON CASES BOL BOL Pa::camete::c HZP HFP Ejected :cod wo::cth (pcm) 1024 200 Delayed neut::con f::caction .0059 .0059 P::ce-ejection Fq NA 2.55 Post-ejection Fq

• 10. 6 4.33 T::cip :cod wo::cth Cpcml 2210 4069 Ze::co to full powe::c Dopple::c -1164 -1164 defect (pcm)

TWINKLE 3-D soluble bo::con 2-260 1700 oncent::cation (ppm)

ETRAN mode::cator tempe::cature 2.9 1. 5 coefficient Cpcm/°F)

Numbe::c of operating pumps 2 3 TWINKLE 1-D ::ceactivity ** 1. 74 1. 2 feedback weighting factor RETRAN Power Weighting Facto::c 2.3 1.25 Notes:

• Peak 3~D steady state nodal powe::c weighted by a generic pin-to-box ::catio and uncertainty factor

• • Weighting facto::c fo::c 3-D TWINKLE= 1.0 NA= not applicable ppm= pa::cts per million pcm= pe::ccent mille

PAGE 149 4.2 Comparison Results In order to select the proper power weighting factor CPWF) for the RETRAN point kinetics analysis for comparison to the 3-D analysis, a hot sp~t total power peaking factor CFq) for the ~ondition of the ro~ being ejected must first be known. Since the core conditions being analyzed are not comparable to those which would typically be predicted for Surry Unit 1, Cycle 1, (i.e., the ejected rod worths have been arbitrarily inc~eased), values of Fq used in the comparison analysis were derived from the 3-D TWINKLE model. steady state values of peak ~ore nodal power at initial control ~od core configurations less the ejected rod were calculated with the 3-D TWINKLE model for both HZP and HFP conditions.

(The arP calculation assumed frozen thermal hydraulic feedback.) These lues were.increased by a 20% generic pin-to-box ratio to convert.them from peak nodal powers to hot spot total power peaking factors. An additional 10% was added for uncertainty. As presented in Table 4-5 above, the assumed post-ejection total power peaking factors were therefore 1_0.6 for the HZP case and 4.33 for the HFP case. The PWFs for the RETRAN analysis were derived from these values of Fq based on Figure

~-6. Likewise, the reactivity feedback weighting factors used in the 1-D TWINKLE analysis were based on these values of Fq.

Typically, the higher the value of Fq assumed for the analysis, the less severe will be* the power history predicted by the point kinetics calculation since a higher value of Fq implies a higher PWF which in turn causes a lowe~ power history curve to be predicted. The derivation of the Fq values was based on steady state 3-D TWINKLE calculations

PAGE 150

( *:. *stead of the transient calculations for this reason, since the transient calculation derives some benefit from thermal hydraulic feedback effects and therefore predicts lower Fq values than the steady state calculation, This is the method used for core reload analysis where the Fq values are ~btained with steady state physics codes assuming frozen thermal hydraulic feedback. The peak nodal powers predicted by the TWINKLE 3-D -odel with the rod ejected from the core Cand before the initiation of the scram) were as follows:

HZP HFP Post-ejection steady state 8.04 3.28 Transient 8.39 2.59 In essence, the use of conservatively high values of Fq will yield lower ETRAN point kinetics power history predictions and therefore lead to a closer power history comparison with the TWINKLE 3-D prediction than might otherwise be expected. From the standpoint of the hot spot calculation, the lowez core average power history will tend to lower the predicted peak temperature and enthalpy predictions. But this is offset by the higher Fq values used to weight the core average power his~ory which will in turn cause the peak temperature and enthalpy predictions to be more conservative. The latter effect is the more dominant, especially £or the HfP case where the PWF has little effect on the core average power history prediction.

Figure 4-16 pre~ents the core average power history comparis~ns. For both the HZP and ttfP cases, the RETRAN point kinetics predictions are conservative compared to the TWINKLE 3-D prediction thus verifying the

PAGE 151 r--*-**ceptabili ty of the point kinetics methodology in general, and the usage of the PWF in pa:rticula:r. The conservatism of the point kinetics approach is even more apparent in Figure 4-17 which compares the total energy :release of the two models. Both figures show :reasonably close agreement between the RETRAN po~nt kinetics predictions and the TWINKLE 1-D predictions.

The hot spot power histories input to the RETRAN Hot Spot Model were calculated using the method outlined in Section 2 based on the Fq values provided in Table 4-5. A similar method was used fo:r the FACTRAN calculations based on the TWINKLE 1-D predicted core average power histories. That is, the value of Fq as a function of time ~as linearly

ramped* fo:rm its pre-ejection value to its post-ejection value over a interval of 0,1 seconds and maintained at the post-ejection value the :remainder of the transient. The core average power history is then multiplied by this Fq curve to derive a hot spot power history.

A different approach was used to derive the hot spot power histories fo:r the FACTRAN calculations fo:r the 3-D cases. Fo:r these cases, the hot spot power fo:r a specific time was found by multiplying the 3-D core average normalized power and the 3-D peak nodal power (increased by a 20% pin-to-box ratio and a 10% uncertainty factor.) This method was used to :reflect the fluctuation of Fq throughout the transient. Typically, the Fq value at any time will be less* than that assumed fo:r the point kinetics methodology, However, because of the flux redistr1bution which results from the movement of the scram banks into the core, it is possible for the peak normalized nodal power in the core to exceed that

PAGE 152 sumed for the post-ejection condition. Use of a constant generic pin-to-box ratio to convert the peak normalized nodal power to a Fq value may result in a 3-D Fq prediction later in the transient which exceeds that calculated for a post-ejection condition by steady state methods. The actual impact of this on the transient is mitigated by two effects: (1) the actual core location of the hot spot in the core will be expected to shift throughout the transient, and (2) the actual power level of the core during the tr~p is relatively l~w compared to that at the transient peak. Therefore, the Fq transient shape input to the point kinetics analysis would still yield a conservative hot sp~t power history compared to what would actually transpire in a three-dimensional core. Figure 4-18 presents a comparison of the hot spot power histories

..illllllllinput to the RETRAN Hot Spot Model and the FACTRAN code for the three

~ethodologies unde~ comparison. Since the 3-D TWINKLE derived hot spo~

power history is appreciably lower than that of the RETRAN point kinetics calculati.on, it may be reasonably predicted that the results of the RETRAN hot spot analysis based on the RETRAN hot spot power history will be conservative compared to the FACTRAN analysis results based on the 3-D TWINKLE power history.

Table 4-6 presents.the results of these analyses based on the hot spot power histories presented in Figure 4-18. Plots of the fuel centerline and outer clad tempe%ature transients for the three models are presented in Figures 4-19 and 4-20. As expected, the RETRAN point kinetics and TWINKLE 1-D based analyses are conservative compared to the 3-D analysis .

PAGE 153

• TABLE 4-6 3-D HOT SPOT MODEL COMPARISON RESULTS I. HZP Case:

1-D TWINKLE 3-D TWINKLE Parameter /FACTRAN RETRAN /FACTRAN Fuel Pellet Melting C?.) 0 0 0 Max. Fuel Center Temp. (°F) 4346 3872 2659 Max. Clad Temp. C°F) 2869 2387 1703 Max. Fuel Enthalpy (Btu/lb) 296 255 16 1 II. HFP Case:

1-D TWINKLE 3-D TWINKLE

- ~ ______ Parameter _________ _ /FACTRAN RETRAN /FACTRAN Fuel Pellet Melting (?.) 0 0 0 Max. Fuel Center Temp. C°F) 4819 4685 4419 Max. Clad Temp. C°F) 2037 1990 1851 Max. Fuel Enthalpy (Btu/lb) 274 268 235

FIGURE 4-16

• NUCLE/1.R POHER IE~TS *~* 3-0 BEt-lf.H11/\RKS HOT FULL POWER HOT ZERO POWER 1 ,5 100.00 I i* I *4 I

I .J I

I .z 1.1

• 10 .oo

,. N ..

0

'I. R 0.9 Ii  !

0 R

"L*.R

• 0 .e 11 I R l-L . E 0.1 I .0 z I .OD E  : j*.. :ro* iJ.6 0

1. . M

.r0. E 0.5 i R M

E R

,. 0 .4 *.*-

I i: 0,3 0 .10 I 0,2 i*I 0.,

i o.o i o.o Q.5 I .O I *5 2

• O* Z *5 3 ,0 3 *5 4 *O 4 .S 5 .*

I o*.01 I* TIME I 5EC I o.o o.s 1.0 ,.s z.o z.5 J.o* J.s co c5 s.o TIME I SEC I S TRF. : Hl JIIKlt* I -0 t11.l0tl STAR: TMJNKLt 1-D MOOtL SDUAR[: n[lRHN rolNl KIN[TICS ~OO[L SOUAR(: R[TRAN ro1u1* KIH[TICS MODEL T~IAHGL[: TWINKL[ 3-D MDUCL TRIANGLE: THIHKLE 3-0 MOD(L FIGURE 'i-J61J flGl/RE 4-J'fl

~U.RE 4-17 WESTINGHOUSE PRO~RIETARY CLASS 2

• liOT ZERO POWER TOT/\L ENEP.GY.AS~,.- 3:-.D OENCIH1ARKS WESTINGHOUSE PROPRIETARY CLASS 2

• HOT FULL POWER 2 .15 3.00 2.so 2.15 (a;c) +(a,c) '

2.25 2.50

.. N 0 2*.00 R 2.25 H N A 0 L I .15 R I H 2.00 l A

[ L 0 I .50 I l I, 15 E E N D E I .25 R E I .50 G N y [

R

. I .OO

• R:

G I .zs E y L

.[ 0.15 .R A E I .oo s l E E o.so A s *0.15 E

o.zs o.so o.zs o.o o.s 1.0 1.s 2.0 *2.s 3.0 3.s 4.0 ,.s s.o TInf I SEC I o.o o.s ,.o 1.5 z.o z.5 3.0. 3.5 4.0 4.5 5,0 T!HE I SEC I STAR: JWINHLE 1-0 HODEL SQUARE: RETRRN POINT KINETICS HODEL STAR: TWINKLE 1-0 HODEL SQUARE: RITRAN POINT KINETICS HODEL I-'

TRIANGLE: TH IHHLE 3-0 HODEL \J1 TRIANGLE: IWINHLE 3-D HODEL I.JI f.lCUR[ 4-HA FIGURE 4-.178

FIGURE 4-18 HOT *ZERO POWER HOT SPOT POl~E.ORY D BENCHMARKS HOT FULL POWER 6

100.00

• N 0 6 R

N 0 "LA R I "LA z E 4 I 10.00 0 z H E 0 D T H S 3 0 p T 0 s T p 1.00 p 0 0 T W2 p E 0 R w

E R

D.10 o.o 0-6 I .2 I .9 2-4 3.0 3.6 4.2 TINEI SEC I o.o 0.5 1.0 1.6 2.0 2-5 3.0 3.5 4.0 4.5 5.0 TINEI SEC I STAR: RETRAN POINT KINETIC&

SQUARE: TWINKLE 1-D STAR: RETRAN POINT KINETICS TRIANGLE: TWINKEL 3-D SQUARE~ TWINKLE 1-D TRIANGLE: TWJNKEL 3-D FIGURE 4-l8B FIGURE 4-l8A

• . -.E 4-19 HOT SPOT FUEL CE~ITE~LINE TEr1°ER.l\TURE TR/'.MSIENTS 0 BE!*!CW1i\RKS HOT ZERO POWER HOT FULL POWER 4500 4600 4000 4700 C 4600 E 3500 N

T C E E 4500 R N L T I 3000 E N R 4400 E L I

T N E E 4300 11 2500 p T E E R 11 4200 A p T E U 2000 R R R 4100 E T u

D R E E 4000 G 1500 D

f E G 3900 f

3600 3700 o.o 0.5 1.0 I ,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0 Tl11E I &EC I 3600 l...~.......~...............,......~..............'1""........,........'"T"~'""T'~....,.-:--~

o.o 0,5 1.0 1 ,5 2.0 2,5 3,0 3,5 4,0 4.5 5,0 T111E I SEC I .....

u,

&TAR: FRCTRRN CODE 11-D POWER HISTORY! ""-J SQUARE: RETRRN HOT &POT IPOINT KINETICS POWER Hl&TORYI STAR* FRCTRAN CODE 11-0 POWER HISTORY!

TRIANGLE: FRCTRRN CODE 13-0 POWER HISTORY! SQUARE* RETRRN HOT SPOT !POINT KINETICS POWER Hl&TORYI TRIANGLE: FACTRRN CODE 13-D POWER HISTORY!

FIGURE 4-19R FIGURE 4-196

FIGURE 4-20 HOT .SPOT FUEL OUTE r. CL PER.lHURI: TRNISIPffS D BPl<W1!\~.KS HOT FULL POWER HOT ZERO POWER 2000 1900 1800 1700 0

2500 u 0

T 1600 E

u R IT 2250 L C

1500 A 1,00 D

T 1300 E

T E

p 1200 E

n 1750 R p A 1100 E T R u A R 1000 T 1500 E u

R D 900 E E G

800 f

f 700 600 500 o.o 0.5 I .O I .5 2.0 2.5 3.0 3.5 4.0 4.5 5.o TINE I SEC I 3.0 ll.5 ,.o 4.5 5.o TJnE 16ECI STAR: fACTRAN CODE 11-D POWER Hl6TORYI SQUARE: iJ~:ANF~~JR:~0coDE 11-D POWER HJSTORYI SQUARE: RETRAN HOT 6POT IPOINT KJNETIC6 POWER Hl5TORYI TRIANGLE: fACTRAN CODE 13-D POWER Hl6TORYI .

TRJANGLE: fACTRA~ !~D0EINTl3KD1NETIC6 POWER Hl6TORYI

- POWER HJ6TORYI FIGURE ,-10B f IGURE ,-20A

PAGE 159 SECTION 5 -

SUMMARY

AND CONCLUSIONS The Virginia Electric and Power Company CVepco) has developed a methodology using the RETRAN transient thermal. hydraulics code to analyze the control rod ejection event £or Vepco's Surry and North Anna Nuclear Power Stations. The analysis is performed in two steps: (1) a point ~inetics calculation to determine the core average power history, and (2) a thermal hydraulics analysis, based on the core average power history from step (1), to predict the fuel and cladding temperatures and average fuel enthalpy history of the core hot spot. Results from step (2) are used to confirm that safe plant operation acceptance criteria are met. These criteria are those in USNRC Reg. Guide 1.77 (Ref. 3) and the additional in-~ouse limits discussed in Section 1.3 .

• cceptability of the methodology-has been demonstrated by comparison of selected analytical results obtained with the Vepco methodology to corresponding results obtained with a NRC approved vendor methodology.

The conservatism 0£ the Vepco methodology, specifically the point kinetics calculation and its use of a Doppler reactivity feedback weighting factor with the very conservative hot spot static peaking, was demonstrated through the comparison of the Vepco results with those based on a three-dimensional space-time kinetics model and a detailed hot spot thermal hydraulic model. In Section 3, a comprehensive sensitivity analysis was performed and reported for neutronic and thermal hydraulic parameters to verify the range of applicability. The overall good agreement between the two methodologies affirms that the Vepco methodology can be used for performing the reload safety analysis

PAGE 160

~~ the cont~ol ~od ejection analysis fo~ Vepco's nuclea~ powe~ plants .

PAGE 161

,. SECTION 6 - REFERENCES

1. Surry Power Station Units 1 and 2, "Updated Final Safety Analysis Report," Virginia Electric and Power Company, 1982.

2. North Anna Power Station Units 1 and 2, "Updated Final Safety Analysis Report," Virginia Electric and Power Company, 1982.

3. "Assumptions Used for Evaluating a Control Rod Ejection Accident for Pressurized Water Reactors," Regulatory Guide 1.77, USAEC, May 1974.

4. J. H. McFadden et al., "RETRAN-02: A Program for Transient Thermal-Hydraulic Analysis of Complex Fluid Flow Systems," EPRI NP-1850, April 1981 (Vol. 1-3) and January 1983 (Vol. 4).

5. "WREM: Water Reactor Evaluation Model," NUREG-75/0S6, Revision 1, May 1975.

6. N. A. Smith, "Reactor System Transient Analyses Using the RETRAN Computer Code," VEP-FRD-41, Virginia Electric and Power Company, March 1981.

J. H. Keenan, et al., "Steam Tables: Thermodynamic Properties of Water Including Vapor, Liquid, and Solid Phases," John Wiley & Sons, Inc. , 19 6 9.

8. "MATPRO - Version 11 (Revision 1), A Handbook of Materials Properties for Use in the Analysis of Light Water Reactor Fuel Rod Behavior," NUREG/CR-0497, TREE-1280, Rev 1, USNRC, Feb. 1980.

9. C. Hunin, "FACTRAN - A FORTRAN IV Code for Thermal Transients in a U02 Fuel Rod,".WCAP-7337, June 1972.

10. L. S. Tong and J. Weisman, "Thermal Analysis of Pressurized Water Reactors," American Nuclear Society, 1970.

11. M. L. Smith, "T~e PDQ07 Discrete Model," VEP-FRD-19A, Virginia Electric and Power Company, July 1981.

12. J. R. Rodes, "The PD207 One Zone Model," VEP-FRD-20A, Virginia Electric and Power Company, July 1981.

13. W. C. Beck, "The Vepco FLAME Model," VEP-FRD-24A, Virginia Electric and Power Company, July 1981.

14. J. G. Miller, "Nuclear Design Reliability Factors," VEP-FRD-45A, Virginia Electric and Power Company, October 1982.

15. S. A. Ahmed el al., "Reload Nuclear Design Methodology," VEP-FRD-42, Virginia Electric and Power Company, April 1981.

PAGE 162 17.

Roger A. Rydin, "Nuclear Reactor Theory and Design," University Publications, 1977.

D. H. Risher, "An Evaluation of the Rod Ejection Accident in Westinghouse Pressurized Water Reactors Using Spatial Kinetics Methods," WCAP-7588, Revision 1-A, January 1975.

18.

• R. F. Barry and D. ,H. Risher, "TWINKLE - A Multi-Dimensional Neutron Kinetics Computer Code," WCAP-7979, December 1972.

19

• Letter from C, M, Stallings, Vepco, to E. G. Case, NRC, Serial N~. 108, Docket Nos. 50-280 and 50-281, March 15,1978.

20. J. B. Yasinsky, "On the Use of Point Kinetics for the Analysis of *Rod-Ejection Accidents," Nuclear Science and Engineering: 39, 241-256, 1970.

21. D. B. Waters, "The Core Physics Characteristics of the Surry Unit 1 Nuclear Power Station," WCAP-7534, Rev. 1, July 1970.

22. Letter from C, M, Stallings, Vepco, to K. R. Goller, NRC, Serial No. 553, June S,1975 (Surry Positive Moderator Coefficient).