ML20245L164

Application of One-Dimensional Kinetics to Vermont Yankee Transient Analysis Methods

ML20245L164
Person / Time
Site:	Vermont Yankee
Issue date:	06/30/1989
From:	Chandola V, Lefrancois M, Robichaud J VERMONT YANKEE NUCLEAR POWER CORP.
To:
Shared Package
ML20245L154	List: ML20245L150 ML20245L164 ML20245L182
References
7801R, YAEC-1693, NUDOCS 8907050393
Download: ML20245L164 (60)

Text

$

2' b

) Application of One-Dimensional Kinetics to Vermont Yankee Transient Analys4s Methods June 1969 Major Contributors: U. A. Ansari V. Chandola M. P. LeFrancois J. D. Robichaud I

l pa w 7801R I

hk P

kDO

-Prepared by: 2. " -

N.

V. Chandola ~I Transient Analysis Group

/(fate)

Prepared by: >M N IP / E M. P. LeFranc'ois ' [(Dat6)

Transient Analysis Group Prepared by: 8. P, 8

[. D. Robichaud 4[(74[)f9 Date Transient Analysis Group Ap r ve by:

P. A. Berg'eron, nager ' (Da'te)

> Transient Analy is Group

, Approved by: . . bbb h!3d [

If. C. Slifef Director (bate)

(NuclearEngineeringDepartment i

i l

Yankee Atomic Electric Company Nuclear Services Division 580 Main Street

' Bolton, Massachusetts 01740 I

-ii-7801R

DISCLAIMER OF RESPONSIBILITY This document was prepared by Yankee Atomic Electric _ Company

(" Yankee"). The use of information contained in this document by anyone other than Yankee, or the Organization for which this document was prepared under contract, is not authorized and, with respect to any unauthorized use, neither Yankee nor its officers, directors, agents, or employees assume any obligation, responsibility, or liability or make any warranty or representation as to the accuracy or completeness of the material contained in this document.

I I

l t

t f

(

I

-lii-7801R

)

ABSTRACT This report presents a description of the application of the RETRAN one-dimensional kinetics neutronic feedback option to the Vermont Yankee transient analysis methods. The ability of tne model to predict licensing basis transients is verified through comparison to results from the current point kinetics analysis, vendor calculations, and the recirculation pump trip test performed at Vermont Yankee in 1981. The intended application to generate operating limits is further supported by comparison against results of an alternate, statistical method which demonstrate the conservatism in the approach.

l l

I l

/

1 f

1

-iv-7801R o_- _ - _ - - - - _ _ _--______ _

TABLE OF CONTENTS EAER iii DISCLAIMER..................................................

iv ABSTRACT....................................................

TABLE OF CONTENTS........................................... v LIST OF FIGURES.............................................. viii ix LIST OF TABLES..............................................

ACKN0WLEDGEMENTS............................................ x 1

1.0 INTRODUCTION

2

2.0 DESCRIPTION

OF CURRENT LICENSING METH0D. . . . . . . . . . . . . . . . . . . . .

2 2.1 Genera 1................................................

2 2.2 Current Methods Qualification..........................

2.3 Licensing Analysis Methods............................. 3 7

3.0 DESCRIPTION

OF THE APPLICATION OF ONE-DIMENSIONAL KINETICS..

7 3.1 Core-Wide Model Modifications..........................

3.2 Generation of Kinetics Data............................ 9 3.3 Hot Channel Model Modifications........................ 10 l

3.3.1 Current Hot Channel Methods..................... 10 l

3.3.2 Hot Channel Model Modifications................. 11 4.0 VERIFICATION OF ONE-DIMENSIONAL KINETICS APPLICATION. . . . . . . . 15 4.1 One-Dimensional Versus Point-Kinetics Application...... 16 4.1.1 Review of Previous Qualification Work........... 16 k 4.1.2 Comparison of Vermont Yankee Point and One-Dimensional Kinetics Methods................ 17 4.1.3 Comparison to Vendor Calculations............... 18 l

4.2 Reactor Coolant Pump Trip Test Comparison.............. 19 4.3 Summary of Verification Effort......................... 20 5.0 LICENSING APPROACH AND TREATMENT OF UNCERTAINTIES. . . . . . . . . . . 37 1

1 5.1 Specific Licensing Application......................... 37 5.2 Treatment of Uncertainties............................. 38 f

) 5.3 Comparison of Vermont Yankee Licensing Approach to the Statistical Method................................. 39 l

i

-v-7801R

IABLE OF CONTENIS (Continued)

Eage 41

6.0 CONCLUSION

S.................................................

43

7.0 REFERENCES

APPENDIX A DETERMINATION OF ACPR 95 THROUGH A RESPONSE SURFACE APPR0ACH................................ A-1 APPFNDIX B MODEL UNCERTAINTY STUDY......................... B-1 i

I t

l l

1

-vi-7801R

LIST OF FIGURES Humhgr Title Eage 2.1.1 Vermont Yankee Reload Analysis Flow Chart 6 3.1.1 Vermont Yankee Core-Wide RETRAN Model Nodalization 13 Vermont Yankee Hot Channel RETRAN Model Nodalization 14 3.1.2 4.1.1 Peach Bottom-2 Turbine Trip Test 1, Neutron Power 22 Versus Time 4.1.2 Peach Bottom-2 Turbine Trip Test 2, Neutron Power 23 Versus Time 4.1.3 Peach Bottom-2 Turbine Trip Test 3, Neutron Power 24 Versus Time 4.1.4 Peach Bottom-2, TTWOBP, Comparison of Neutron 25 Power Predictions 4.1.5 One-Dimensional Verrus Point, Normal Neutron Power 26 4.1.6 One-Dimensional Versus Point, Normal Average 26 Heat Flux j 4.1.7 One-Dimensional Versus roint, Normal Core Inlet 27 Flow 4.1.8 One-Dimensional Versus Point, Steam Dome Pressure 27 l

4.1.9 One-Dimensional Versus Point, Vessel Steam Flow 28 4.1.10 One-Dimensional Versus Point, SR Valve Flow 28 l

4.1.11 Methods Comparison, Normalization Neutron Power 29 4.1.12 Methods Comparison, Normalization Average Heat Flux 29 l

4.1.13 Methods Comparison, Normalization Core Inlet Flow 30 4.1.14 Methods Comparison, Steam Dome Pressure 30 4.1.15 Methods Comparison, Vessel Steam Flow 31 I

4.2.1 RPT Test, Core Power 32 l 33 4.2.2 RPT Test, Steam Dome Pressure

-vii-7801R f

l _ _- _ -__ - _- _ _ _ _ _ _ _ _ - _ _ _ _ _ ._

u LIST OF FIGURES (Continued)

Iille f. age Humher 4.2.3 RPT Test, Core Flow 34 4.2.4 RPT Test. Steam Flow 35 4.2.5 RPT Test M-G Set Speed 36

-vili-7801R k

i LIST OF TABLES Hushar lille EAE2 2.3.1 Vermont Yankee Technical Specification Scram Speed 5 Time Limits Reactivity Coefficients Produced With Previous and 12 3.2.1 New Reactor Physics Methods Transient Model Inputs and Initial Conditions 40 5.1 Compared to Expected Values B.1 RETRAN Model Uncertainties Study Results B-2 l

i l

(

} _ix_

7801R l

ACKNOWLEDGEMENTS The authors would like to express sincere gratitude to J. T. Cronin for his technical assistance during the one-dimensional kinetics development project, to Michele Sironen for her efforts in coordinating the project, to Ron Engel and Ken Doran of Sol Levy, Inc. for assistance in statistical uncertainties methods, and to the Word Processing Center for their prompt service in preparing the report.

I t

}

r i

I 1

I

-x-l

) 7801R

L

1.0 INTRODUCTION

Yankee Atomic Electric Company (YAEC) currently uses the RETRAN computer code to analyze limiting transients and establich operating Minimum Critical Power Ratio (MCPR) limits for each Vermont Yankee Nuclear Power Station fuel reload. The analysis methods, which received NRC approval in

[1, 2), use the point-kinetics modeling option to represent transient-induced neutronic feedback. The RETRAN code, which has received generic NRC approval in (3), also contains a one-dimensional kinetics neutronic feedback model option which provides a more accurate transient power prediction than the conservative point-kinetics model. The inclusion of the one-dimensional kinetics model in the Vermont Yankee analysis methods will provide MCPR i operating margin gains which are needed to offset the anticipated operating margin reductions associated with longer operating cycles. This report documents the extension of YAEC's approved method to qualify the one-dimensional kinetics modeling option in RETRAN f or use in future Vermont Yankee transient analysis.

i I

I t

I 1

L t

l 7801R

(

h

2.0 DESCRIPTION

OF CURRENT LICENSING METHOD 2.1 General The Vermont Yankee RETRAN point-kinetics analysis method was approved by the NRC for Vermont Yankee Cycle 9. The Vermont Yankee analysis methods were submitted to the NRC for review in References [4, 5, 6] and received approval for use via References [1, 2]. The methods include a systems level calculation with point kinetics to determine the core wide power response to an anticipated event and a single fuel assembly model to calculate the limiting or hot channel response. A flow chart of the analysis process is shown in Figure 2.1.1 These methods have been used to support Cycles 9-14 reload analysis.

l 2.2 CurIent_Me.tho.ds_ Qualification The current Vermont Yankee reload licensing method utilizes the RETRAN-02 code for the core wide and hot channel analysis of the limiting events. RETRAN has received generic approval for use in licensing applications [3]. RETRAN has been used extensively by utilities and consultants to analyze a variety of transient thermal-hydraulic problems. The r base RETRAN documentation [7] contains reports on comparisons of RETRAN results to separate effects tests, system effects tests, and power reactor

! startup and special tests. Vermont Yankee has built on this base level of qualification data through application of Vermont Yankee methods to analysis of:

1 1. Vermont Yankee startup tests [4].

2. The Peach Bottom series of turbine trip tests using Vermont Yankee

[

I methods to define the RETRAN input (4].

3. The Peach Bottom representative licensing analysis [6].

4. Sensitivity studii.s on various types of transients to provide assurance that a converged solution was obtained with the chosen method technique [4].

7801R 1

i

5. Comparisons to steady state and transient boiling transition data

[5].

I The results of these evaluations as well as reload analysis for the past five Vermont Yankee fuel cycles [ References 8, 9, 10, 11, 12] have shown that ~ the Vermont Yankee RETRAN model can predict a wide variety of transients and conservatively predict MCPR operating limits for reload cycles.

2.3 Licensinc Analysis Methods Past licensing analyses show that the anticipated events which result in the minimum core thermal margins f or Vermont Yankee are:

1. Generator load rejection with complete failure of the Turbine j Bypass System (GLRWOBP).

2. Turbine trip with complete failure of the Turbine Bypass System (TTWOBP).

3. Loss of Feedwater Heating (LOWH).

i These events are typically analyzed at exposure points at End of Full Power Life (EOFPL), EOFPL-1,000 mwd /St and EOFPL-2,000 mwd /St. The loss of l feedwater heating is also analyzed at Beginning of Full Power Life (B0FPL).

Of the three thermal margin transients and exposure points, the TWOBP at EOFPL produces the largest change in critical power ratio, and results in the most limiting MCPR operating limit. For the purposes of this report, the

! Cycle 9 TTWOBP at EOFPL has been used as the demonstration case. The TWOBP i is used in all phases of demonstrating the extension of the Vermont Yankee licensing methodology.

The TWOBP (and GLRWOBP) are pressurization transients whose power surge is eventually teminated by a reactor scram. The scram times assumed in the licensing analysis directly impact the calculated MCPR. The scram times that have been used are those expressed as the Technical Specification limits. There are two scram times in the Technical Specifications commonly 7801R

i referred to as Measured Scram Time (MST) and 67B. The times to four different insertions are presented in Table 2.3.1 for each scram time. The 67B scram based MCPR limits are a conservative backup to MST based limits in the event

) the periodic plant measurements indicate test times exceed MST.

As a conservative bound to the expected plant thermal power, the current licensing analysis assumes the power level to be 104.5% of the licensed value of 1,593 MW or 1,664 MW. This value corresponds to the actual steam flow limit of the turbine generator set and bounds any expected calorimetric measurement uncertainty. The typical deterministic uncertainty assumed to bound plant calorimetric power is 2%.

The transient power response is currently calculated using the classical point-kinetics model in RETRAN, which contains six delayed neutron groups and the 11 group fission decay model. The point-kinetics model assumes

[

a fixed transient power shape for transient calculations due to its lack of spatial dependence. Briefly, the point-kinetics model, as applied in the Vermont Yankee methods, accounts for the reactivity effects of scram, moderator, and fuel temperature changes. Further details of the core power

) response model can be found in [4]. The methodology used in generating the RETRAN input kinetic parameters, scram reactivity, and feedback reactivity functions is addressed in the Vermont Yankee Core Physics Methods Report [13].

The thermal hydraulics of the reactor vessel is modeled in the core region with a Homogeneous Equilibrium Model (HEM) and 12 fluid volumes. Each fluid volume has a heat conductor with power of the heat conductor determined by the point-kinetics model. In essence, there are 12 separate point-kinetics l

nodes whose power feedback is determined by the relevant thermal hydraulic and heat conductor state properties existing at each of the axial levels. The relative contribution of each axial level is determined by the fixed axial power profile.

f L

)

7801R

}

t TABLE 2.3.1 Vermont _ Yankee Technical Specification Scram Speed Time Limits

• Fosition (% Inserted) Time to Position (sec)

MST 67B 4.51 .358 .358 25.34 .912 1.096 46.18 1.468 1.860 87.84 2.686 3.419

• These scram times are assumed in the licensing analysis.

I L

i 780lR t

RETRAN-02 CORE-WIDE MODEL CALCULATES SYSTEM RESPONSES FOR UCENSING TRANSIENT RETRAN-02 HOT-CHANNEL MODEL BOUNDARY CONDRONS FROM CORE-WIDE ANALYSIS i

1 MASSAGE CODE MASSAGES RETRAN EDITS l' j

FOR INPUT TO TCPYA01 TCPYA01 CODE BOUNDARY CONDRONS FROM HOT-CHANNEL ANALYSIS DELTA-CPR RESULTS ,

f l

5 Figure 2.1.1 VERMONT YANKEE RELOAD ANALYSIS FLOW CHART t

)

3.0 DESCRIPTION

OF THE APPLICATION OF THE ONE-DIMENSIONAL KINETICS OPTION The focus of this section is to describe the extension of the current licensing methods to the application of the one-dimensional kinetics option in RETRAN. The application of one-dimensional kinetics has resulted in other minor changes to the Vermont Yankee modeling to align the model for use of the more explicit neutronic feedback option. In particular, the algebraic water steam slip and subcooled void model options are utilized as opposed to the HEM option used in the point-kinetics application. Minor refinements in the reactor vessel model nodalization have been made to accommodate reflector regions for one-dimensional kinetics as well as to describe the actual core flow and bypass flow more precisely during the transient.

Section 3.1 describes the changes to the system level or core-wide model, including the use of one-dimensional kinetics. Section 3.1 also provides a discussion of the impact of the minor model refinements beyond the one-dimensional kinetics application. Section 3.2 is a brief summary of the methods used to generate the kinetics data for the core-wide model. Section 3.3 describes changes to the single channel or hot channel methods.

)

3.1 Care-Wide Model Modifications This section describes the application of the RETRAN one-dimensional kinetics option, the slip and subcooled void options, and the vessel nodalization refinements made to the Vermont Yankee RETRAN core-wide model.

All other aspects of this model remain unchanged from the point-kinetics version described in (4] and approved in [1, 2}.

The core-wide model nodalization for the one-dimensional kinetics option is shown in Figure 3.1.1. To accommodate the reflective boundary conditions used in the one-dimensional option, Volumes 200 and 213 have been l

addco. They represent lower and upper regions of the core adjacent to the fuel. Volumes 201 - 212 represent the active core region as in the current licensing model. Twenty-seven separate neutronic regions are modeled for consistency with the three-dimensional physics model. Each of the inactive core volumes (Volumes 200 and 213) contain one neutronic region.

7801R l

Volumes 201-211 contain two axial neutronic regions and Volume 212 contains three axial regions. Each neutronic region receives appropriate cross section fitting coefficients from the SLICK code described in [14).

The multiple control state model is used with two control states for the pressurization events. The initial control state is either all rods out for EOFPL conditions or an appropriate state representing partial insertion of various control rods for exposure states prior to EOFPL. The final state is fully rodded, modelling post-scram conditions. The initial conditions represent a critical reactor state.

Twelve neutronic radial mesh intervals are used per neutronic region.

This number of mesh intervals is based on dividing the fuel pin into intervals to approximate the mean free path of a neutron. A sensitivity analysis was i performed on this parameter to optimize the number of mesh intervals.

A sensitivity analysis was performed to evaluate the required frequency f or the axial shape f unction. The sensitivities have shown that for severe pressurization events, it is necessary to update the axial shape function

)

every timestep.

Minor modifications to the thermal hydraulic nodalization and options used in the Vermont Yankee point-kinetics model were made to make use of one-dimensional kinetics and more accurately represent the bypass flow during a transient. As discussed above, the reactor vessel nodalization has changed with the addition of Volumes 200 and 213 (see Figure 3.1.1) which represent the upper and lower inactive core regions. Individual representation of flow I

entering the fuel assembly water tubes versus flow entering the active core flow has been added to allow the fraction of flow entering the water tubes and core to vary during the transient. Junction 198 models the active core flow l plus water tube flow, Junction 199 models the fuel bundle water tube flow, and Junction 200 the active core flow. As with the other core bypass flows lumped into Junction 16, the initial flow of Junction 198 is calculated using the FIBWR code [18].

l

) 7801R

)-

The algebraic slip option is used in the core-wide model. It is the standard model used with one-dimensional kinetics for analysis of abnomal operational transients. The subcooled void option is also employed to more l accurately describe the neutronic feedback in the subecoled regions of the core. These models are consistent with the three-dimensional physics code, SIMUIATE-3, from which the one-dimensional physics data set is derived.

Sensitivity analysis have been performed on each of these minor thermal-hydraulic model refinements. The results of the sensitivity analysis demonstrates that these refinements have a negligible impact relative to current methods.

3.2 Guteration of Kineties Daj;a The details of the methodology for the generation of the l

one-dimensional cross sections and kinetics parameters are presented in Reference 14. The general approach taken in developing the kinetics data is very similar to Yankee's current methodology [13). The aim of the method is to take the neutronics effects as predicted by the three-dimensional nodal i

simulator, SIMU1 ATE-3 [19), and f unctionalize them with the f eedback variables used by RETRAN. This is achieved by running multiple SIMUIATE-3 perturbation cases for each of the two control states. The types or perturbation cases run are representative of the transient of interest and cover the range of density and fuel temperature encountered during the transient. SIMUIATE-3 also performs the spatial homogenization of the kinetics data and produces all the l

cross sections and kinetics parameters needed by RETRAN.

The Vermont Yankee SIMUIATE-3 model [19] is used to predict the neutronic response of the core. In generating the kinetics data, the EPRI i void model [20] is used. The EPRI void model is the basis of the algebraic slip and subcooled void models in RETRAN. Furthermore, it is the model used in the neutronics portion of our previous method [13). To provide some insight into the differences between the new physics methods and our previous methods, a comparison of pressure and Doppler coefficients is presented in Table 3.2.1. The coefficient data presented is representative of the perturbation cases run to produce kinetics data for a licensing basis pressurization transient. The pressure coefficient is based on a 7801R i

i three-dimensional calculation where the core pressure is perturbed with the thermal power and fuel temperature distributions fixed at the initial state condition. Thus, the pressure coefficient is a measure of void reactivity.

The Doppler coefficient is produced by varying the three-dimensional temperature distribution while the density distribution is held constant. In performing these perturbation calculations, all other variables nonna11y associated with cross sections (exposure, control history, fission product inventory, etc.) are held constant at the initial state condition. Data for two cycles at end of full power life conditions are presented. The results of the perturbation calculations.show that there is very little difference in the pressure and Doppler coefficients produced by the new and previous physics methods.

The thermal-hydraulic model used in the generation of the kinetics data is virtually identical to the actual RETRAN model. Comparisons between this model and RETRAN results have been presented in Reference 14.

3.3 lin tlhanne1Jiodal_tiadllinatisna I

3.3.1 Current Hot Channel Methoda Analysis of the limiting hot channel is performed to derive the change in Critical Power Ratio (dCPR) during the transient to derive the Operating i Limit Minimum Critical Power Ratio (OINCPR). The GEXL correlation is used in this analyses to predict transition boiling. The system level or core-wide analysis provides transient boundary conditions for the RETRAN hot channel analysis. The lower and upper plenum transient thermal hydraulic conditions I and active core region core power fractions are the boundary conditions used in the RETRAN hot channel model. The hot channel runs also result in boundary J

conditions to drive the TCPYA01 computer code (5). The results of the TCPYA01 runs are the Initial Critical Power Ratio (1CPR), the minimum CPR during the event analyzed and the decrease in CPR. The initial power level is chosen such that MCPR is the safety limit. Thus, the ICPR becomes the Operating Limit MCPR (01NCPR).

f

) 7801P

3.3.2 IloLfhannel Model Modifications The RETRAN hot channel model has been changed to maintain modelling consistency with the one-dimensional kinetics core-wide model nodalization (see Figure 3.1.1). The significant changes include the core bypass region (Volume 11, and Junctions 16, 18, and 199). Other nodalization adjustments are the addition cf the upper and lower inactive core regions as Volumes 201 and 113 which are joined to the plenums and active core region with Junctions 1, 13, 198, and 213. The core bypass region is being explicitly represented in both the core wide and hot channel models. This allows the bypass flow to vary during the transient. The bypass region models all core bypass flow paths, while the core region model is a single fuel bundle. As with the core-wide model, the steady-state flows within the lumped bypass (Junction 16), the water tubes (Junction 199), and the hot channel (Junction 1) are determined by the FIBWR computer code. For each hot channel initial power, the FIEWh code is used to determine the bypass an~d active core flow splits using the same limiting 1.4 chopped cosine axial power distribution input in the hot channel model. The algebraic steam water slip model is employed for consistency with the core-wide model.

l l

l 7801R

l TABLE 3.2.1 Reactivity Coefficients Produced With Previous and New Reactor Physics Methods Doppler Coefficient (oem/F)* Pressure Coeffitigni_Iprm/ psi)

Previous ** New*** Previous New Condition Methods Methods Methods Methods Cycle 9 E0FPL -1.46 -1.43 +8.34 +8.47 Cycle 13 EOFPL -1.57 -1.55 +8.26 +8.07 Un%2s 5

• pem = (keff-perturbation - keff-initial)/keff-perturbation

• 10 ,

• • Lattice calculation with CASMO-1; 3-D nodal calculation with SIMULATE-YA.

• • • Lattice calculation with CASMO-3; 3-D nodal calculation with SIMULATE-3.

1 I

i 7801R

e l

T O

RV-1 RV-2 3 RVad SY-1.2 O YOW Si 52 S3 54

• JUNCT10h go , i i (gp} t t

'E t

4 B

l O z1 ..-

e & .

T 2 i r

't  ;) I O

• d, i I  !

nu n u sso p Q Q l @

O  :

he, j g - - ~ ~ "- '

i6 g,,)

f! ~

'( >

i 1

u u 2{rg (

n 44 l

9 n!

l '[ Y l l

l e + ,!

l i

1 a H as b '

Ib,i b Y n,

,. in (ec) 2;c 16 iG6 l

< 2e  ! ,, '\

il/ f [g', *

' g?) 9 l \

//(C J' i

%'n'_sJ g L

/

,/ .

Figure 3.1.1 VERMONT YANKEE CORE WIDE RETRAN MODEL NODALIZATION l ' - ~

I

(

k 1

O

+

,4e ass G',)

O va- ,1 e

i l

- e. I 4

! u h

10 j

i .

r .

m

.., g i$

v ,

t

.* 4 Of

, v g 4

O 0J

l. . -

+ ._* ,

l g *1 99 9 NCC) 4 u r, e en

\

' i

\8 D /

/

/

l .~ ~.__ _

Ii UTC E 3.3.

f VERMONT YANKEE HOT CHANNEL RETRAN MODEL NODALIZATION i

I 4.0 VERIFICATION OF ONE-DIMENSIONAL KINETICS APPLICATION An extensive effort has been made to verify the application of one-dimensional kinetics in the Vermont Yankee RETRAN model. The verification consists of comparison to:

o Previous qualification of the point-kinetics method through the i Peach Bottom turbine trip tests and Peach Bottom representative licensing case.

o The current Vermont Yankee point-kinetics method.

\

o Vendor (General Electric (CE)) one-dimensional kinetics calculations.

l o Plant data via the Cycle 8 Recirculation Pump Trip (RPT) test.

l The verification effort addresses a wide-range of analysis to

- demonstrate the adequacy of the one-dimensional kinetics application. It I

is the integrated set of comparisons listed that make up the verification of the one-dimensional kinetics application.

The analysis previously completed to compare with the Peach Bottom tests demonstrates the method's ability to predict mild pressurization events and points out the conservatism in the methods. The Peach Bottom representative licensing case simulates a severe pressurization event. In comparison to others (GE and Brookhaven National Laboratory) who have also analyzed the licensing case, the Vermont Yankee point-kinetics method was shown to be a conservative predictor of integrated transient power and thus l

MCPR operating limit. The comparison between Vermont Yankee point and l one-dimensional kinetics methods demonstrates consistency in the application of the one-dimensional kinetics option. The similarity of the Vermont Yankee one-dimensional kinetics method to other approved one-dimensional methods is shown by comparison to GE licensing calculations carried out for Vermont Yankee Cycle 9. Analysis of the Vertont Yankee Cycle 8 RPT test allows comparison of the core dimensional kinetics application to actual plant data.

7801R

l 4.1 One-Dimensional Versus Point-Kinetics Application 4.1.1 Review of Previous Qualification Work A review of part of the previous analyses to qualify the Vermont Yankee point-kinetics method is presented to highlight important results relative to the extension of the Vermont Yankee methods to one-dimensional kinetics.

Reference 4 discusses the results of a comparison to the Peach Bottom series of turbine trip tests. The Peach Bottom turbine trip tests were mild pressurization events whose power transients were overturned by the increase in voiding within the core. In contrast, the power transient for a more severe licensing basis pressurization event requires core voiding and scram

[

reactivity to turn the transients. Figures 4.1.1 though 4.1.3 are comparisons of the Vermont Yankee point kinetics methods results to the Peach Bottom tests I (TT1, TT2, and TT3) for the neutron power response repeated f rom Reference 4.

The figures show that the Vermont Yankee methods overpredicted the Peach Bottom power responses. The thermal-hydraulic response of the Vermont Yankee modeling approach were in good agreement with the test data except for steam dome pressure which was higher due to the overpredicted power. Thus, the Vermont Yankee methods comparison to the Peach Bottom tests proved that the Vermont Yankee point-kinetics method would overpredict the response to a pressurization event which is a conservative treatment for licensing analysis. These comparisons demonstrate the Vermont Yankee point-kinetics method conservatively predicts the transient thermal-hydraulic response of a BWR, Reference 6 discussed results of a representative licensing case analysis of a Peach Bottom 2 Turbine Trip Without Bypass (TTWOBP) using the Vermont Yankee point-kinetics method. This analysis has become a standard analysis for those qualifying BWR methods. At the time the Vermont Yankee analysis was carried out, Brookhaven National Laboratory (BNL) and GE had also analyzed the Peach Bottom 2 TTWOBP problem. Reference 6 contains comparison of the point-kinetics response to BNL and GE. This comparison allows a more representative assessment of the neutronic feedback mechanisms and response to a severe pressurization event in line with the intended application of the method.

i 7801R

The Vemont Yankee point-kinetics analysis methods for the standard Peach Bottom licensing' problem resultEd in a more severe transient simulation' than either BNL or GE by generation.of more power before the transient is, suppressed by the. scram. The prediction of the neutron power compared'to BNL or GE is'shown in Figure 4.1.4. Note that the Vermont Yankee' methods case predicts the generation of more energy ar.d results in a more severe event.

'Overall, the Vermont Yankee point kinetics method results in a more conservative simulation of a severe pressurization event due to the generation of more energy.

4.I.2 Compatispp of VeDDQDt' Yankee Point and One-Dimensional Kinst#.cs Melhads A comparison of the Vermont Ya wee point and one-dimensional kinetic methods is shown in Figures 4.1.5 through 4.1.10. The event simulated was the E0FPL Cycle 9 TTWOBP using the MST Technical Specifications scram times. This event set the operating limit for end of Cycle 9. The one-dimensional kinetics case responds faster to the pressurization and is terminated more quickly by scram than the point-kinetics method. This difference in behavior of the point and one-dimensional kinetics is consistent with that shown in studies by EPRI in Reference (17).

In addition to the core-wide simulation of the Vermont Yankee Cycle 9 TTWWOBP, the hot channel analysis was performed to determine the differences in transient oCPR. The point-kinetics model yielded a .23 oCPR, whereas,

'the one-dimensional kinetics predicted a .18 ACPR. The gain of 0.05 is attributed to the more accurate modelling of the axial power shape change (during a severe overpressurization event like the TIV0BP) associated with one-dimensional kinetics. This gain is consistent with the observfd difference.s in response between the two methods. The predicted ACPR from the one-dimensional kinetics calculation is also consistent with vendor {

calculation as .1scussed 4 in Section 4.1.3.

L l

l 1

b

' The' conclusions that may be drawn- from the Vermont Yankee point -

kinetics.and one-dimensional kinetics comparison is:

o The differences of the neutron power response of the point and one-dimensional kinetics applications is as expected and consistent.

Other calculations in Reference (17).

-o The difference is attributed to explicit representation of the axial power. shape changes during the transient.

o The one-dimensional kinetics model and subsequent hot channel analysis shows a decrease in transient ACPR of .05 over the point-kinetics analysis for the same Cycle 9 TTWOBP at EOFPL. This is consisteric with the neutron power differences and consistent with vendor calculations (Section 4.1.3).

4.1.3 Comparison to vendor calculations Comparisons to vendor one-dimensional kinetics (GE) calculations for Vermont Yankee Cycle 9 (Reference 15) were carried out to show the similarity of one-dimensional kinetics methods and a previously approved one-dimensional kinetics method which was also compared to the Peach Bottom turbine trip tests. The transient simulated was a Vermont Yankee Cycle 9 TTWOBP. GE used 67B Technical Specification scram times, so the one-dimensional kinetics case described above was changed to the 67B scram times and run. The results of the comparison are shown in Figures 4.1.11 through 4.1.15. The neutron power responses (Figure 4.1.11) are nearly identical. The peak neutron power predicted by GE was 528%, while the Vermont Yankee one-dimensional methods resulted in 530%. The pressure response of the two simulations is compared in Figure 4.1.14. The overall Vermont Yankee one-dimensional kinetics results are in good agreement with the vendor's calculations.

A hot channel simulation was carried out with the Vermont Yankee one-dimensional method. The resulting transient ACPR was 0.22, which is similar to the vendor's calculation of 0.21. l L ,

l 7801R {

L

iThe' comparison to the vendor calculation demonstrates:

o- The neutron power and steam dome pressure response are nearly identical, thus, demonstrating the similarity of both the neutronic .I andLthermal-hydraulic modeling, o The Vermont Yankee one-dimensional kinetics method behaves similar to-an approved vendor method.

4.2 Reactor Coolant Pumo Trio Test Comparison A' simulation of the. Vermont Yankee Cycle 8 recirculation pump trip test was performed with the Vermont Yankee one-dimensional kinetics model. The simulation shows the ability of the modeling methods te predict actual plant data.

The Cycle 8 RPT test was initiated from 82.9% power and 86% core flow conditions by the simultaneous trip of the Motor Generator (MG) set drive motors from the control room. The MG set trip resulted in a power and flow coast to 43% power /32% flow under natural circulation conditions. The vermont Yankee RETRAN models were initialized at the test conditions using the plant data summarized in Reference 16 and employing the same techniques used in the licensing model. One-dimensional kinetics physics data was generated for the RPT test conditions using the same methods employed for licensing conditions via the SLICK code discussed in Section 3.2 and Reference 14.

Figures 4.2.1 through 4.2.5 show the results of the RPT test simulation. The flow response is largely a function of the inertial )

' characteristics of the Recirculation System, with a secondary influence

. resulting from the core reactivity characteristics. The MG set speed

) (Figure 4.2.5) predicted by the models is an accurate match to the test data.

The simulated core flow (Figure 4.2.3) follows the test data closely indicating the thermal-hydraulic and recirculation's control systems are properly modeled. The core power (Figure 4.2.1) matches the plant data well, verifying that the core reactivity characteristics are properly simulated.

7801R

(

'This comparison of the Vermont Yankee one-dimensional kinetics RETRAN Lepplication to plant test data clearly demonstrates the. continued accuracy of the physics and transient' analysis methods.

4'. 3 S"--arv of' Verification Effort The verification effort has demonstrated that application'of the

' Vermont Yankee one-dimensional kinetics modeling option responds accurately.

when-compared to test data and a variety of other analysis. .The comparisons

'have verified that:

o The basis for the Vermont Yankee one-dimensional kinetics RETRAN modelling method, the point-kinetics modelling method, is a conservative predictor of transient power as demonstrated in the Peach Bottom turbine trip test comparisons.

o The Vermont Yankee point kinetics method results in a conservative simulation of a severe pressurization event in comparison to either the GE or BNL result in the Peach Bottom licensing case.

o The differences in transient ACPRs are as expected, the one-dimensional kinetics methods results in a smaller ACPR (.18) than the point kinetics method (.23) in response to the typically limiting transient (TTWOBP, measured scram time).

o The Vermont Yankee one-dimensional kinetics methods compare-very well to approved vendor calculations.

o The one-dimensional kinetics methods predict the Vermont Yankee Cycle 8 RPT test data accurately.

l The verification effort shows that the one-dimensional kinetics application is appropriate for simulation of both mild and severe thermal-hydraulic transients. The use of one-dimensional kinetics option in

( severe pressurization events, used for calculating plant operating limits, has shown the expected gain in transient ACPR. Section 5.0 provides a j discussion of intended licensing approach with the one-dimensional kinetics I

7801R

= - _ _ _ _ _ - _ _ _ _ _ - _ _ _-- . - - - _ _ _ _ _ . - ._

option and the treatment of uncertainties in the one-dimensional application.

'A comparison of'the one-dimensional application to an alternate, statistical treatment of uncertainties demonstrates the continued conservatism in the Vermont Yankee method.

1 I

4 t

I l 7801R f 4

_j

h j

4C .- .. , ,  ;

.g..

o.o ** -

m e- _

e.o -

l ', -

, i i

- ,a i _

, 9 7.o --

,i l -

, i

_ ,e i, . _

, i e .o -

1 -

.g l

, i _

i

,. . I 5.o -

l  !

i l , i

, i _

, i i i

• s.o -

l l -

, , i-

, i-

_ , i -

, I f I a.c -

i -

i i

i 2.o -

/ i l 1

, s _ ,

/ \

1.0 -

l - _

o,o e i i i l i i # i l i i e i o.o .5 1.o 1.5 l- TlhE(SEC)

PB2 II1 l

l~

Figure 4.1.1 PEACH BOTTOM-2 TURBINE TRIP TEST 1 NEUTRON POWER VS. TIME lL

~ ~ ~ ~

p)p '

. .sg-

.q.

l S*0 i

. i 4 g, i .I . j . I i . .l l

DATA 7.0 -

f., _

, i i t' e.0 -

l-l

, i i t 0 l 7 _

1. 1 5.0 - ' '

l s

k. _ ,

l r s 8 l 0 1 4.0 ~^ * -

l 4

4 i

1 1

3.0 - -

o i

l _

I 2.0 - o -

s i i 1

1. l l ', -

e s

\

1.0 -

's,,._,,-

0.0 ' I ' I I I I I ' ' I I 0.0 .5 1.0 1.5 l' TIAECEC) h-PB2 II2 Figure 4.1.2 PEACH BOTTOM-2 TURBINE TRIP TEST 2 I

l

) NEUTRON POWER VS. TIME I

( i

i;: p

... . . . j i i  ; i .j . i ,

9. 0 ,-'

l- ....... m .,

l 4

e.0 - -

7.0 -

4

.I .',

i e.o --

l; q

ll

, i-1 I 5.0 -

l l . . .

i  !

l 4 t.

E 4,o l

x- , ,

2 - -

l l 3.0 - I J

l l .-

4 2.0 -

,/

1. t

- i r -

1.0- \ -

l -

~~. ~~~ ... .

I' gg I f I I l f f I l  ! I l i l o.o .5 1.0 1

j. T nE(SEC)

L PB2 113 Figure 4.1.3 PEACH BOTTOM-2 TURBINE TRIP TEST 3 NEUTRON POWER VS. TIME l

[ L.

4 I r f

12.0 y, , , , , , , , , ,, y

^11.0 -

n

. GE .

MC

.~ 10,0 -

j o.o l- -

e.0 -

Z_ -

g-7.o fs.o g

,s 1

s -

5.0 -

/,Ql \ ,t -

t i 4.o - / i

/ l \ 'i 3.o I' '

' \ 'r I

\r -

/

\gV_

'\ -

2.o- -

/- '

/

/ -

/ _/  %

. 1.0 m e: -

I I I I I I I I I I I o.0 o.o .2 .4 .e .e 1.o 1.2 l

TlhE(SEC)

L P.B.2 TTWOBP i Figure 4.1.4 PEACH BOTTOM-2, TTWOBP, COMPARISON OF NEUTRON POWER PREDICTIONS

I o

/

/

/

/

/

s 'n

- / .d

/

/

s'  %

m gl-

,# 'O 3 pI

' "", ) e W

' o & - U v

- ,' W o -

g ;s

' 'nW

~

W 4 A =

U O.,

e' g , ,,

,# y 'W CD u C/) t.D c 3 O o > >

c~.

1 O.

A B:

p g

og .

- c., .O a s -

l i

g

- o

- l z i

a 'O O

o O

. . 4

) O c o e o 0 N O P= en

- - e 6 6 Xn14.LY3H '0AY 'WBON o

n a

e o -

O , n 5.e m

- , n

/

t I e p t M W l O W 3

2

^ J q $ 8 o & - -

w o

• 8 8 i

' - 'nw w g

- ,' dv c. , g g e

w CD u > W z

i I

,. , - s o o o

t

, b .e.e

. 1 r

I <

.- O w -

~

- -s s

~-  :

c ,

{

- -6 l

)

. l

- )

1 o

i O e . > .

O O O O O O O

, d d v' d N d d I I

W3 mod NOM.tn3N 'HBON i- i j

~

1

i-

• r , q- ,

O o

O a l 6

< Q .c.e l 'O

. l O

• A , -f I i i

,O M W f-i 2 N 4

, H m I m J co. z tra 0

C O l W O 4 A M

'OW A g

W ,,

"o v a o tra M W CD $

8 I 2 O MQ

\ p >

• I A M y s o Q H

\ "N Q  ;

V'

's N

. M,

' 'O

-W O

O 9 9 b o b O g- g O O

6 O

O O

s O

n , ~

O

"

• O C

( YlSd ) 3MDSS3Bd 3HOO WY215

' ' . . , , , O

/ n

' i

/

/

- \ 'q

, 'N t

• b 3 1 M O

.O 3

' N' 3

. N N n -J f%

f 4 O a

, w O . -

,, e 9 N W s

"v 1 e I

• • a W D $ $ g o

/ c 2 O no >

' ow p > w o g i O

-* A

'o

-s p A C

• i i

, g f

1. 1 h

g.

' i

- .. g 'q  !

"O O

O 8 8 h a s y MO'Id .LT1NI 3800 *W80N

o t i a

4

' o.

- c M

'o

-4 s o s n -8 "l 5 3 o

w k.

r

.; 2 2 ow .. m

'" U w

CL m

e u $ 3 D C >

a 3 O @ s a H -s a:

ou .o "A m i O' "N H M N i

1

_.L... J o

. -a

.=

3 1

o h

o 9 b b b b o o o o o o 6 6 6 6 o o o o o o o 8

n - e o o o n o n

I I C g

( 03s/WER ) MOE 3N1YA Bs o

9

)

/

t .o

_j

()

H in

% 9 2 2 N _*

e-n -I % Q S E -: - *

.' o w O c 8 r h, 'ow w j

g

., s ' ri v W

a CD e

y >

g m

i en~' ,

- a 2 F

O W

e ai a

w r A m

~.

-e -

ae .a

-m o

'o n F m

y j'

l -% ,

1

-3

~ 'h> f

~ l I .

9

) ,

o o o o

o o 9 o o 9 i

o o o o o o o

)

o o o o o o o o o o N N N h O h i i l

( 03s/nen ) now nvais Tass 3A h

_ 5

,2

~

0

,2 L

)

P F

C O C E E

' E ,5 S P A E 1( B Y G E O M W s'

\ _ 0 I

T T T

' i _ ,1

.[! ^

,i _

s _

5

' , ,0 0

0 0 5 0 5 0 5 2 0 7 5 1 1 1 0 0 y e H $12 I >< SrOz t

0 3

C E 5

' AE ,2 YG

_ 0 L

'- ,2 P

_ ) F C O

,s E E

_ ' ' 5. S P

, ,1 ( B E O

,N IM W T T T

,0

' \tt ii' 1 ie i,s iiii 8I ,

=i,

) ,88 8'

, I , ' i l

lI 5

,0

_ - 0

_ 0 0 0 0 0 0 -

0 0 1

6 5 4 3 2 0 b*O& zO d

tz SrOz t

d e1 li, I

1 1

l1 i

, , . . O .l-j

). . 4 r) o w

_ <w i do

, o . z w

-a J g i

&< 8 i -

i i - - m

' ' o o w s w J m r

e m

m a s - e e 8 w r

s w o u-s p

> a8O<

8

=

n .

~~

o r

,o o o o o o o o 6 6 6 6 6 8-

?

8 3 2

- 8

( YlSd ) 3 BOSS 38d 3MOG WY31S g

'eu a

o Z O J

/ -W n 8 d o o -

$ b w w ~

o a.

\

- w 'qw c. < r

<w

>- o

-v m o

8 w w a

1 u y w i

P E o

=

8

.o -

g

! - e 1 -4 A

• I~ ,,*"~~, l g h

% i @

. i" .

~

.o o J> o J> o N o h a 9

" " & O o

/AOB EN! 3800 'WBON

~30-

,u 2000.0

.Y 1500.0-  ! 't -

l

,\ l n t i I

1 g i

] 1000.0- ', l j ~/

113 I i t l /

W I i

.\ l l' i 5 L 4 i g 500.0- 8, l l, l -

v '/

i l

's i I

i O a y 0.0- { ,8 .

y i I '

w -500.0- \f!

I

?

d m t i

M I

$ -1000.0- -

-15c0.0- YAEC ,

_____ GE

-2000.0 , ,

0.0 1.0 2.0 3.0 j' . .

'BME ( SEC )

!: TTWOBP EOFPL l.

Figure 4.1.15 I METHODS COMPARISON VESSEL STEAM FLOW i

q

- - _ - - - 3

=

0

- - - 0 2

A _

A T T A _

A D _

D N _

T A _

N R _ -

A T _ -

L E _ 0. '

' P R _

5 1

_ N

_ O

_ S I

) 1 R A R

- _ C 2 P E E W

_ 0. S M O 4 O

_ ,0 1

( C P E e E' r T M

B i u

g E S R O

C T

F T

P R

- 0

- ,5 -

s s

- N s

\

s s

s s

\' 0

\

0

- 0 0 0 -

0 -

0 5 5 0 7 5 2 0

1 Cmw32 wxoo

~

3 dy llt!!!!

1 o

r +

' o q l

0 t

i l

i 1

4 I

I

< F 8 i

H <

• i

< O o z l e < ,

u y e 5a.

w x

l

, -c

\

s

+ 1 i

i ,f l 1 I I t i i l I i

8 l z w O c' I

8 M O

• m v.

8 A N g M 8

g- N

< W I y A &

l 9m J B e o , -a- 0 o E s y p' E

, F

• e w H T:

y A <.

8 H W'

' A &

l x a.

I f

f f

/

/

/

/

/

/

/ o j

' ~5

) p'

/

i

) ,'

f

/

/

l ,5 I

I I

I o

e i

O o o R o

o 6 o C 8 a

g a 8 a

) (ylsd) 380SS3Bd 3 WOO NY31S I

) --

\

4I ,

4i; 0

~ - - 0

_ 2

~

~

_~

~

~

A -

A T .

T A ~

A D ~

D N ,-

T A -

N R -

A T .

L E - 0

, P R - bl N

O

_ S

-., ) 3 I s R C A W E 2. , P O

,' 0S b(

4 M O

L F

l e C

_ E r E M u T R I

T g S O i E C s' F T

,' T

- P R

~

s s'

0, N $

s N

s N

s s

N

- N s

N

\

N

- N s

0 0

0 0 0 0 0 2 0 8 6 4 1 1 mmw2* owmN2gv rod w:QoO

WI

l Ii

~

0 0

2

~

~

~

~

~

~

A m' A T w T A

• A D ~

D N .

T A -

N R -

A T ,'

L E 0

' P R . ,5 1

,~

_ ~

~

g N

% 0 S

I

- 4

) . AR W C 2 P O s

E 1 r

L

0. S 4 F g' ,0 1

( e O

C M _

s r A E T N' M I

u g S E

E T .

s T i T S s F s' T

' . P

- R s

s s'

s N

s 0

' ,5 s'

s s

.s

~

~

~

~

~ '

0 0

- 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 4 2 0 8 1 1 1 1

_ - )

AOdL )y N 2 S V 3 O N 2 U _

,i Jvt

[F ll lllL(l(lti!(1ll1L L

o e t I d

a N f

I C

f I

t a

f

< l

< s ,

H < t

< O #

O g i H < '

Z '8 b x

• o

. a. l .c .

r I

b '

i #

l #

1 l ,'

I o o

I I

Z

' O

! V)

! - n S I

, o W N ac A

m A

i O (r) E M

~

[ ~3 v W U h p W

i i 2 $ h M

)

/ F = 0

,. t +

~

l I

W i

d 1

i I

/

/

/

/

/

O

~

t "C

/

/

/

/

/

/ 1

)- e'

/

/

/

/

/

/

/

/

/

0  ?

• i

)

8 R S a 8

& 6 I I (om/m) caas ias 0-rt i

1 1

5.0 LICENSING APPROACH AND TREAINENT OF UNCERTAIN 7JES

'This section outlines the intended Vermont Yankee licensing approach' with the one-dimensional kinetics RETRAN option. It reports the resultsLof a comparison of the licem ing approach with an' alternate method which treats analysis uncertainties statistically. The comparison demonstrates that the

-Vermont Yankee licensing approach with one-dimensional kinetics continues to be a conservative method for calculating MCPR operating limits.

5.1 Specific'Licensine Application The application of.the one-dimensional option in RETRAN to perform licensing analyses will be treated in the same fashion as current approved methods. The current approved point kinetics method assumes an initial power

~

level of 104.5% for the RETRAN core-wide calculations. The initial power level for the one-dimensional kinetics method will be 102% of the rated'value, which is the typical calorimetric power uncertainty. All other initial conditions and RETRAN inputs like initial steam flow, scram position versus time (MST and 67B), etc., are set at values identical to the current analysis. .The primary difference between the proposed and the current licensing methodologies is that the one-dimensional kinetics model replaces the current point kinetics. The core-wide and hot channel calculation methods remain the same except for minor differences explained in previous sections of this report.

Many input parameters to RETRAN, including the transient initial conditions, are conservative. Table 5.1 lists the conservative input l parameters in the licensing basis model relative to the expected values.

p=

There are other inputs which have been used at their expected values, p

Overall, we have found this licensing approach to be conservative. The next h section will describe the alternate statistical approach to the treatment of uncertainties which demonstrates the conservatism in the proposed p one-dimensional kinetics licensing approach.

) 7801R f L  :

4 j

i

'5.2 Treatment of Uncertainties The statistical treatment of uncertainties in the application of the one-dimensional kinetics option is based on the use of a statistical method to derive a value of transient ACPR with a 95% probability (ACPR 95). Thus, if the limiting transient which sets the 01ECPR actually occurs, there is a 95% probability the safety limit CPR will not be exceeded.

The ACPR 95 was calculated by statistical-convolution of scram speed uncertainties with an overall model uncertainty. Convolution of statistical functions is a common practice, and the numerical technique used to perform i the convolution is described in Reference (21). The assumed scram time is the single most sensitive parameter in the severe pressurization event analysis, so it received an individual statistical treatment. The statistical function or Probability Distribution Function (PDF), in terms of ACPR, used to obtain the scram speed uncertainties was_ derived from a response surface fit. The response surface methodology employed is further described in Appendix A. The response surface fit was based on the Cycle 9 TTWOBP cases used for the verification described in Section 4.0 with plant-specific scram times plus uncertainties substituted for the MST scram times. A range of uncertainties were applied to the mean scram times which resulted in five RETRAN cases to fit the response surface.

The derivation of a model uncertainty consisted of performing uncertainty studies of sensitive inputs and combining the result of the individual uncertainty studies by the square root sum of squares method. The uncertainties study is further described in Appendix B. The square root sum

)

i of squares method assumes the listed combined uncertainties are normally i I- distributed around their mean. The value of uncertainty used in each study was assumed to represent at least a two sigma value, which is needed to obtain

) 95% probability. Many of the uncertainty studies addressed input that are considered to be conservative but did not have an explicitly applied 3 1

)

uncertainty. The most significant of these are the void and Doppler l l

uncertainty studies. Though this treatment may result in compounding of uncertainties it reinforces the statistical argument to show the conservatism f in the Vermont Yankee licensing approach.

t 7801R

Using the method discussed in Appendix A, the PDF value of ARCPR' associated with a 95% probability of the scram speed response surface is 0.127. The statistical combination of the RETRAN model uncertainties results in a mean ARCPR value of 0.04. By convoluting the two statistical functions, a ACPR of 0.18 was obtained. A 29% model unc -tainty is accommodated through this method. This value of model uncertainty is typical of other licensees. TVA in (22) reported a 25% model uncertainty.

5.3 Comparison of Vermont Yankee Licensine Anoroach to the Statistical Method A representative one-dimensional kinetics licensing case (Cycle 9 TTWOBP, 102% Power) was analyzed using the approach described in Section 5.1 to compare with the statistically derived ACPR discussed in Section 5.2.

The licensing case ACPR was 0.174. As stated previously, the statistical ACPR95 was 0.184. Due to the conservatism that remain in the licensing j model and the conservative method taken in the statistical analyses to derive ACPR95 it is concluded that the two methods result in equivalent ACPRs.

Therefore, the licensing method is a conservative approach for deriving MCPR l I

operating limits.

Further refinement of the statistically derived ACPR would result in a smaller ACPR95. Additionally, the statistical approach taken here uses a one variable response surface where others [22] have used two or more variables. Treatment of other significant parameters such as the void uncertainty, for example, in response surface would result in a smaller <

I statistically derived ACPR.

\

I i

) .

)

l

) 7801R

i rs i sy un go e ms rn

l. i c u ) oa a f( l 5 f r n n t a )) i t A odi v 00 4 n s

cem eig e

55 11

( n o

ug n

i t h pl n ) t //)) w i ,i s

a e

g t s r rt i ea l r 0005 o t c

wr ou 0007l B r oeei b e 41 31 s n l d a wmpm ad (( (( u g t e a pi r y j yt s n u msul eE . . . .l l n e i e l a pM cccce t ea v s v a rt . . oS%eeeev e vt l n i v e cacc nA1 ppppi l i s a e. t u) s ee) i + ss s st n t n c a g l 5 s p p0 a i ao V

i L

v r i n a1 v(

md ss5 e uo (

9. l t t n0 asss url t mv an l vc r

d e t e 0 0. i 5 s

mr . . .( o neeemeano ed sn t

.inl ccx hh hi wt wisui c n22 m3 xx t mosoxnnml na e o00i0 aa6il eea2i ol al aoaol o p C11 L1 MM4MATTM1 WNSFSMCMNA C x

E_

o_ .

t c

d_ .

e p

e s l

a.

p ) . d m 1 h n o 7 c a e . e u s0t -

y C_

s l d d l n a eyn n l o_

V el a a a peh i i_ d_ st t t x 1 i t

e a ) e at 1 d l t me 0) l n c_ niv 00 nsge n e nexi 05 inni E_ o p ) rr ot 1 2 ois L C x 5 eera / / ) ) ) niyn B E e1 t f pv) 00050 et ra A l u tf pr0 0503 ecar T a l < aiae3 31 21 4 wnvt i

t a( pd( s( ( ( ( ( ( t u ej,g i v l n n l l l tl ol l l l lll l l b l n a . .a .a aaac a aaaaaaal a t ai I_ n00 n xn . n n n nnnnnnnanti nr d i00i0 ai2i siei iiiiiiiuiixiu n m1 1 m2M m4 md mrm mmmmmmmnml emd a o o0 o ooooo3 oooooooaop o NiiN1 < N<NRNMN1 NNNNNNNMNS N s_

t_

u ) )

p n e n )

) k e c I_ i ) o) p e le

) s n rco s i p o t e (

d s( i ss%

o p

( p t m0 t

_ r ) e( 8) n M o e c e ca t

pr s e seoet n oD n s mmt ss e r I m / i mn De

( yTe( o i_

s r  % a e C n eun 0 y l es y a ) r s o 2 a ek man n a usi l d o( l w I

r_ o i set o e r eo i ) s srr t D ct eDd t R) p eP e eS s t e u BR( r s s c s nps c

_ NB P rn d i s mvoia n b

i %N ee eI n g ( i pro a n

r) (% r rvd o o 'v VsTC t r t R ( uuial c Li yt e e s c e sB w st V e ane Vt nsl R pp g u mmena t i N ows a Hi os m iino d L D%l oerf ti( ef coorvuudii n

( Fl r eemit st eappti PP ot t o e FP pi anoet si pt s nV Mt r C r rm ml eI M m n yl aeeo nn ee p ueaeeeet iiS eCSRC s ool S n a

s wer mTRSddTo R / sii o I

. oot oo oo pn vvvpat t rr G

- pPSCD rooRR mt ofiii opaat er x et t aeioVVVt yl l nl o SB uuol t d

o El lll t l l rSt

/aaaaal l ooc crff f ccC oa R

_ riiiiweerrS meeeee . . rr rr et t t t d sst t at biiibbii wt a l e

wiiiies snnD romlll rr cconp

- onnnneeeooR crueeeuueel oe PI I I I FVVCCCSPNRRRTTRRFCS F u

,  ! , t ,

6.0 CONCLUSION

S-In.this report, the proper use of the RETRAN one-dimensional kinetics model in the Vermont Yankee transient analysis methods has been demonstrated and the' continued conservatism in the licensing approach has been confirmed.

To recap the verification process summarized in the report:

o The Vermont Yankee point kinetics method comparison to the Peach Bottom turbine trip test demonstrate;; thnt for relatively mild pressurization events where the rawer surge .is suppressed by core voiding, the. Vermont Yankee methods overpredict the tests.

Overprediction of. pressurization events is a conservative treatment in licensing analysis. The comparison also demonstrated the accuracy of the thermal hydraulic portions of the modelling method.

o The Vermont Yankee point kinetics method analysis of'the Peach Bottom representative licensing case shows that the Vermont Yankee method overpredicted the BNL and GE results, which is a conservative treatment in licensing analysis.

o The Vermont Yankee point and one-dimensional kinetics method comparison pointed out the differences in the two methods. The one-dimensional model behaved exactly as expected when compared to the differences shown in the EPRI studies in Reference (17).

o The comparison of the Vermont Yankee one-dimensional kinetics method to GE one-dimensional kinetics calculations showed the I closeness in behavior of the two methods and that the Vermont Yankee method predicted the same transient as an approved method.

I L

}-

) 7801R

1" l

'o' The Vermont Yankee Cycle-8 RPT calculations demonstrated that;the one-dimensional method could be used.to accurately predict plant transient test data. j i

I'

' o The use of an alternate statistical method to calculate the transient ACPR for. the typically most : limiting event, the TTWOBP, has proved that adequate conservatism exists, at a 95% probability, in the Vermont Yankee licensing method to accommodate model uncertainties, o The conservatism identified by the statistical method analysis I demonstrates that the licensing approach application of'the one-dimensional kinetics option remains appropriately conservative for generating cycle-specific operating limits.

The Vermont Yankee licensing method demonstrated in this report with the Cycle.9 E0FPL TTWOBP will also be applied to the other pressurization events normally analyzed for a reload cycle as well as the LOFWH which is a subcooling event.

i i

l I.

)

l l

1 1

1 7601R l l-

)

i i

l

7.0 REFERENCES

q 1

1. USNRC Letter toLJ. B. Sinclair, SER, Acceptance for Referencing in j Licensing Actions for the Vermont Yankee Plant of Reports: YAEC-1232 YAEC-1238 YAEC-1239P. YAEC-1299P, and YAEC-1234, NVY 82-157, September 15, 1982.

2. USNRC Letter t) 'u I.. Smith, SER, Amendment No. 70 to Facility License No. DPR-28, h w x t.r 27,1981.

3. USNRC Letter to T. W. Schnatz, SER, Acceptance for Referencing of Licensing Topical Reports: EPRI CCM-5 and EPRI NP-1850-CCM, September 4, 1984.

USNRC Letter to R. Furia, Acceptance for Referencing Topical Report EPRI-NP-1850 CCM-A, Revisions 2 and 3 Regarding RETRAN02/ MOD 003 and MOD 004, dated October 19, 1988.

4. A. A. F. Ansari and J. T. Cronin, Methods for the Analysis of Boiling Water Reactors: A Systems Transient Analysis Model (RETRAN), YAEC-1233, April 1981.

'5. A. A. F. Ansari, K. J. Burns, and D. K. Beller, Methods for the Analysis of Boiling Water I,eactors: Transient Critical Power Ratio Analysis (RETRAN-TCPYA01), YAEC-1299P, March 1982.

6. J. T. Cronin, et. al., Yankee Atomic Boiling Water Reactor Analysis Methods: Analysis of a Typical BWR/4 Turbine Trip Without Bypass Transient, YAEC-1280, October 1981.

7. EPRI, RETRAN - A Program for One-Dimensional Transient Thermal-Hydraulic Analysis of Complex Fluid Flow Systems, CCM-5, December 1978.

8. A. A. F. Ansari, et al., Vermont Yankee Cycle 9 Core Performance Analysis, YAEC-1275, August 1981.

9. B. G. Baharyrejad, et al... Vermont Yankee Cycle 10 Core Performance Analysis Report, YAEC-1342, January 1983.

10. D. C. Albright, et al., Vermont Yankee Cycle 11 Core Performance Analysis, YAEC-1403, April 1984.

1 l

11. R. J. Burns, et al., Vermont Yankee Cycle 12 Core Performance Analysis Report, YAEC-1507, November 1985.

12. V. Chandola, et al., Vermont Yankee Cycle 13 Core Performance Analysis, YAEC-1600, May 1987.

1 i 13. J. M. Holzer, Methods for the Analysis of Boiling Water Reactors l Transient Core Physics, YAEC-1239P, August 1981. i

14. J. T. Cronin, Method for Generation of One-Dimensional Kinetics Data for RETRAN-02, YAEC-1694, June 1989.

I I 7801R

15. GE, Supplemental Reload Licensing Submittal for Vermont Yankee Nuclear Power Station Reload 8, Y1003J01A27, August 1981.

16. GE, Vermont Yankee Cycle 8 Stability and Recirculation Pump Trip Test Report, NEDE-25445, August 1982.

17. EPRI, EI, The Comparison of One-Dimensional and Point-Kinetics for Various LWR Transients, Proceedings: Third International RETRAN Conference, EPRI NP-3803-SR.

18. A. A. F. Ansari, Methods for Analysis of Boiling Water Reactors:

Steady-State Core Flow Distribution Code (FIBWR), YAEC-1234, December 1980.

19. R. A. Woehlke, MICBURN-3/CASMO-3/ TABLES-3/ SIMULATE-3 Benchmarking of Vermont Yankee Cycles 9 Through 13, YAEC-1683 (March 1989).

20. EPRI, Mechanistic Model for Predicting Two-Phase Void Traction for Water in Vertical Tubes, Channels, and Rod Bundles, EPRI NP-2246-SR (February 1982).

21. William H. Press, Numerical Recipes, The Art of Scientific Computing, Cambridge University Press, 1986 Edition.

22. S. L. Forkner, BWR Transient Analysis Model Utilizing the RETRAN Program, Tennessee Valley Authority TVA-TE81-01-A, April 1983.

i l

l 1

{

)

1 7801R

APPENDIX A Determination of ACPR 95 Throurh a Response Surface Approach The objective of the statistical evaluation is the develo; ment of the Probability Distribution Function (PDF) for RCPR (= ACPR/1CPR) given the statistical distribution of' the key transient input variables. The variable RCPR is a generally accepted figure of merit for deriving operating limits.

The probability distribution of ACPR is then utilized to obtain the value of RCPR which has 95% probability of not being exceeded by the operating plant if the limiting event occurs. The direct approach to developing the RCPR probability distribution would be to run trials with the RETRAN model with the key inputs selected randomly from their uncertainty distribution. However, the Monte Carlo approach requires a large number of trials to develop a precise probability distribution so direct simulation of each trial with the RETRAN model is impractical. Instead, a response surface is constructed which predicts the RETRAN m: del calculated value of RCPR as a function of the value of the key inputs. The response surface is developed by fitting model results to a polynomial with the key transient inputs as independent variables. The advantage of the response surface is that far fewer model calculations are required to develop an accurate response surface than directly develop the probability distribution on RCPR.

In the approach to demonstrate conservatism in the Vermont Yankee licensing method, a one variable response surface, which treats scram time, has been employed.

l The response surface used in the statistical approach is of the form shown in Equation A-1.

RCPR = (A9 +A y

• SS +A *SS + RSU) (A-1) 2 l

i r

A-1 7801R

i V

H ,

APPENDIX A ')

(Continued) l I

where SS = Random value of time (seconds) to 25% scram insertion minus the nominal'value.

RSU = Random response surface fitting error.

Ai = Response surface fitting coefficients.

l l In order to develop the fitting coefficients (A g ) for the response surface, five RETRAN simulations were performed. The procedure used was an orthogonal central composite design process for two parameter levels (Reference (A.1)). Table A.'1 shows the values of scram speed which were utilized. A standard least square fitting technique is utilized to determine the coefficient A g.

The trial values for the time to 25% scram insertion were assumed to be normally distributed. The mean time to 25% control rod insertion following scram solenoid de-energization was assumed to be 0.796 seconds with a standard deviation of 0.0457 seconds. These values are conservative relative to measured data for Vermont Yankee.

A standard deviation of .003 (with a 0.0 mean) for the fitting error was employed to generate the response surf ace uncertainty (variable RSU) by l randomly selecting from a normal distribution for each trial evaluation a the 1

l response surface evaluation of the response surface.

l I

I A-2

l. 7801R

y l l.

E APPENDIX A l

(Continued)

The fitted response surface (Equation A-1)'along with the uncertainty distribution.for each input variable was evaluated for 100,000 trials to cbtain the' required PDF. Each trial selects a random value for each variable in accordance with their assumed uncertainty distribution.

A sensitivity study was performed to assess the RETRAN model uncertainty (Appendix-B). Based on this study, the model uncertainty

(.04 ARCPR at two sigma) was conservatively assumed to be normally l distributed with a mean of 0.0 and a standard devietion of .02. The'effect of the.model uncertainty on the response surface PDF was determined by convoluting the response surface PDF with the model uncertainty PDF.

To obtain the RCPR value (RCPR95) which is greater than 95% of the trials, the PDF is integrated from the lowest interval up to the value at which 95% of the trials have been accumulated. A RCPR95 value of 0147 was -

obtained. The ACPR95 value is calculated from the following equation:

ACPR95 = 1.07*RCPR95 A-2 1-RCPR95 A value of 0.184 was obtained, l.

1 l

l A-3 7801R

l l

1 APPENDIX B J

I 1

Model Uncertainty Study l A sensitivity study was performed to quantify the net uncertainty in RCPR due to uncertainties associated with the key parameters in the RETRAN licensing model. Uncertainties were applied only to those significant parameters which were either not assumed at their bounding values or were not treated statistically in the response surface. The base cace for all sensitivity studies is the Cycle 9 TTWOBP (EOFPL) model at 100% power level with a Vermont Yankee-specific mean time to 25% scram insertion of

.7960 seconds. The other initial conditions were set at values identical to the licensing analysis case except for the perturbed parameter.

A summary of the sensitivity study of the TTWOBP event is presented in Table B.1. The table shows, for each parameter analyzed, the uncertainty applied. This table presents the change from the base case the maximum core average fuel rod heat flux (Aq), and the change in the ratio of transient ACPR over initial CPR (RCPR).

The various uncertainty components were combined by the root mean square method and a net uncertainty in RCPR of 0.02% was obtained. This uncertainty was-used to derive the ACPR95 value in Appendix A.

l l

L 1

B-1 7801R

TABLE B.1

+,

RETRAN Model Uncertainty Study Results-09 Q ARCPR**

Void Coef ficient More Negative (13%) '8.211 0.037*

[ Doppler Coefficient Reduced (10%) 3.27 0.006*

Prompt Moderator Heating Reduce 25%- -0.25 -0.002 Double Recirc loop Fluid Inertia -0.85 -0.004 Double Jet Pump Fluid Inertia -0.11 -0.003 Increase Jet Pump M Ratio by 8% 1.18 0.008*

Increase Steam Line Inertia 10% 1.51~ 0.010*

Reduce Steam LIne Pressure Drop 8% 0.58 0.003*

l i

l l

• Indicates items included in determination of uncertainty in model RCPR.

• • dRCPR = RCPR - RCPREASE E-2 7801R i

i

_ _ _ . _ _ _ - _ _ _ _ _ - - J