ML20236P472

Forwards Response to 870831 Request for Addl Info Re Reload Topical Repts 033 & 040

ML20236P472
Person / Time
Site:	Oyster Creek
Issue date:	11/12/1987
From:	Wilson R GENERAL PUBLIC UTILITIES CORP.
To:	NRC OFFICE OF ADMINISTRATION & RESOURCES MANAGEMENT (ARM)
References
5000-87-1414, NUDOCS 8711180020
Download: ML20236P472 (45)

Text

'

s gg. UCse M 3 GPU Nuclear Corporation iO0 interpace Parkway Parsippany, New Jersey 07054-1149 (201)263-6500 TELEX 136-482 Writer's Direct Dial Number:

November 12, 1987 5000-87-1414 U.S. Nuclear Regulatory Commission Attention: Document Control Desk Washington, D.C. 20555 Gentlemen:

Subject:

Oyster Creek Nuclear Generating Station Docket No. 50-219 Reload Topical Reports 033 and 040 1

1 Pursuant to your letter of August 31, 1987, please find attached a response to tne reauest for additional information relative to Reload Topical Reports 033 and 040. If you nave any auestions, please contact M.

W. Laggart at (201) 263-6205.

l

, S cere y, 1

8711180020 DR 871112 -

ADOCK 05000219 .

i%

PDR R . il on _

Vice President Technical Functions RFW/JDL/pa(54119 )

Att.

cc: Mr. William T. Russell, Administrator Region I U.S. Nuclear Regulatory Commission 631 Park Avenue King of Prussia, PA. 19406 l

l NRC Resident Inspector Oyster Creek Nuclear Generating Station Forked River, N.J. 08731 Mr. Alex Dromerick, Jr.

U.S. Nuclear Regulatory Commission 7920 Norfolk Avenue Phillips Building, Mail Stop 314 Bethesda, Maryland 20014 [

GPU Nuclear Corporation is a subsidiary of GeneralPubhc Utihties Corporation

r s

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT-TR-033 i

l. Please provide a list of accidents and transients GPU Nuclear intends to analyze using RETRAN-02 1-D methods and RETRAN-02 0-D methods. Justify ,

the application of the appropriate methodology to each class of accidents t and transients chosen.

RESPONSE

In general, one-dimensional kinetics will be used for transients that i involve considerable axial changes in cross-sections and power shape. 'In. '

particular, it will be used for pressurization transients, namely, turbine trip transients (with and without bypass), Generator Trip, Main Steam Isolation Valve transients, Feedwater Controller Failure in maximum demand transients and any other event that may lead to a severe over  ;

pressurization (e.g. Pressure Regulator failure in closed position). The ]

justification is due to the fact that the axial void distribution in a BWR d is highly non-uniform and the arrival'of a' pressure wave down the vessel to the upper part of the core causing a veld collapse in the top part (with the highest void fraction) will result in a considerable shift in axial power profile before the transient is terminated. The shift.in l spacial flux profile will violate the point kinetics. derivation-where the i flux shape function is normally assumed constant. The point kinetics will .

therefore be used in transients that involve small (or uniform)  !

l perturbations that would not cause a considerable axial shift,-such i

! transients are level and pressure perturbations, recirculation flow l transients (pump trip, flow increase / decrease), loss of feedwater flow, ,

stuck open relief valves and isolation condenser initiation transients >

(except ATHS type transients). The typical limiting Oyster Creek reload transients are the over pressurization transient (reference.1) namely Turbine Trip Without Bypass, Main Steam Isolation and Feedwater Controller Failure Transients where one-dimensional kinetics will be used.

l

References:

1. " General Electric Reload Fuel Application for Oyster Creek," NEDO 24195, 1983.

2999C L _ _. _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _

x s A

% q l

1 GPU NUCLEAR TOPICAL'REPO'RT TR-033'

2. Provide estimates'of uncertainties,in physics parameters-such as the ."

d Doppler reactivityLcoefficient,ithe; void reactivity. coefficient, thelscram l reactivity, the effective. delayed neutron fractionn the prompt neutron?

generation time, and the collapsed cross-section parameters listed in Table 3-1~. For each parameter, or class of parameters, cite references orf  ;

provide analyses to justify the' estimated uncertainties, Briefly describe! l how the RETRAN-02 analyses will be performed to conservatively account;forL i the uncertainties. 1 i'

RESPONSE

The uncertainties in physics parameters fall into two main categories, ,\

namely uncertainties that are attributed to the derivation of the. nuclear

~

o data and its application to the. transient model;'and uncertainties due.to .;

fuel cycle exposure where.each of the reactivity components (void, Doppler, scram and prompt reactor period which is the. ratio of the neutron; lifetime and-delayed neutron fraction) may have a different impact as ac .

function of exposure. As' discussed in Section 4.1.2 of. reference (1), the.

first category stated above is accounted for by using:aiset of- . .

conservative multipliers on the' void, Doppler and scram reactivity components. The:second category is accounted for by. calculating the above parameters at-the cycle exposure which gives the most conservative'overall effect.  !

Uncertainties used for point kinetics and their magnitudes will be  !

dependent on the type of the transient analyzed but will'be atLleast as  :

conservative as Table 4.1-2 of reference (1). The. sensitivity'to. cycle  ; '

exposure was discussed in Section 4.3 of the Topical Report.'

The uncertainty in the collapsed cross-section parameters may be estimated j by a comparison between the 1-0 flux shape'and.the 3-D average axial flux l shape. It is shown in Section 2.4 of the Topical; Report that the maximum peak node error for all cases of Cycle 1 and Cycle 10'is within 2.8% which would represent'the maximum uncertainty in the collapsed parameters.

The nuclear parameters uncertainties 1s incorporated in the RETRAN-02 transient analyses through the use of a CPR multiplier. The CPR multiplier is applied to both point and 10 kinetics' transients.

References:

1. " Generation of Vold & Doppler Reactivity Feedback for Application to BWR Design," Licensing Topical Report, NEDO-20964, December 1975.

J 2999C l

'l 4

.I J, GPU NUCLEAR TOPICAL REPORT TR-033 -,

q l

3. Has the computer code BELLEROPHON been reviewed and accepted for licensing'. 1 applications? Equation (2-14). assumes a uniform fission. cross-section- l across the' reactor core. Estimate the error introduced in B,-through this assumption-and explain'how it is accounted.for. , Explain how the- ].,

importance. factor is determined.

RESPONSE

l i

The computer code BELLEROPHON has not been reviewed and accepted for.

11 censing application to our. knowledge; BELLEROPHON is;a'small computer code that utilizes a N0DE-B power, exposure and exposure weighted-void

] 1 l

l- data for calculating Si as shown in equation '(2-14). The values:for Si, u and Fi are input parameters.

3 j ,

l The fission fraction for each isotope j is treated as.a quadratic function of nodal exposure E and weighted nodal exposure V for each

( distinct fuel type.

I i

Ff'k = Cf'" + Cf,k E n +Cf'I E,2 , ,

{

1 1

The coefficients for this equation are determined from CPM (reference 8 in TR-040) assembly depletion ~ calculations. Since delayed neutron fractions and decay constants are somewhat different for fast and thermal fission, l, fast and thermal fission in fissile isotopes are treated as two distinct isotopes. Fission fraction coefficients (C's in the above equation) and and X are input for the following isotopes: fast and thermal' group U "; fast group for U'"; fast and thermal group for PU*";land thermal group for PU 2 The error introduced in G 1 by assuming a uniform fission in equation (2-14) cross section is estimated to be 1.0%. The estimate is based on' comparisons of 4 i calculated using power squared weighting and flux squared weighting. This error is not specifically. accounted for beyond' '

the uncertainties already used with point kinetics parameters.

The imp ^rtance factor is determined from the~ delayed' neutron fraction and effect\<e delayed neutron fraction calculated in CPM. The importance.

factor is determined as a function of exposure.at core average'volds~and a core average value for the importance factor is input to.the code, 2999C 1

.m _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ __

.- .s , ,

GPU NUCLEAR TOPICAL REPORT'TR-033'

'4. Estimate the errors introduced by using a temperature independent Doppler coefficient in 0-D applications of>RETRAN-02, and neglecting the moderator.

temperature dependence of cross-sections in 1-D applications.

RESPONSE

The Doppler coefficient is calculated over'a' range of temperatures in the-3-D model. The Perturbation that is applied covers.the largest expected change in fuel temperature for transients analyzed with the, point _ kinetics model, So while a temperature independent Doppler coefficient is used, it includes spatial contributions ~over a range of temperatures.. By_ limiting the use of point' kinetics to transients where the power shape does not' change significantly, the error introduced by this; approach.is small.

Since Doppler contributes ~only 10 to 20%:to core reactivity: feedback in a transient, the error in-reactivity is estimated to'.be less'than~ 1.0L This is covered by the uncertainty associated with the~ Doppler coefficient.

Either a change in void fraction or a change in. moderator temperature:

would produce a change in moderator. density.. At' operating conditions, any, transient will result-in a change of_ void fraction, but a change'in moderator temperature may occur during a; transient'only due to a change in pressure or a change in core inlet.'subcooling. The'effect of-moderator-temperature in moderator density is much less than that of:the void-fraction. As found in the reference, it takes approximately a 15*F change-in moderator temperature to change the water density'an amount equal to a change of void fraction of 0.01. By ignoring the moderator temperature effect, a less than 1% error is estimated except for ATHS in which the system is depressurized so that temperature changes significantly.

References:

' ji

1. EPRI NP-4498, Vol. 4. "The Reactor Analysis-Support. Package (RASP):

BWR Physics Analysis Guidelines," October 1986.

2999C

_______--____-_-__-_-_______-_________=_-_____ __ ._ -- _- _ --_-

i I

GPU NUCLEAR TOPICAL REPORT TR-033.

5-. Equation -(2-21) has a typographical error. Please provide the correct equation. 1 1

RESPONSE

The correct form of the equation (2-21) is:

.j 1 '"u m a

{s,, u e

u ,,, 1 ,, ,g a

{

• O " iVu

' isk sk=

i u

u I

I 1

1 2999C

.___=__ __

., .., I l

J GPU NUCLEAR' TOPICAL REPORT TR-033

6. What is the value of radial buckling introduced into a partially )

controlled axial node in the radial leakage correction procedure?- :j l

RESPONSE

In our' applications, both SIMULATE and SIMTRAN have a 24 node core model. 1 Therefore, there is no partially controlled.. axial' node in our SIMTRAN ,I appilcations. I In SIMTRAN, the radial leakage correction terms are calculated on a planar basis from the discrete form of the'one-dimensional neutron balance equation. The resulting equation for the radial' leakage is given byL J

2 DI (k) 01 (k)B (k) =

( )(3,)2 f'1(k+1) - 2$ (k) + $g(k-1)}

f2

~I R1 t1(k). + f vrf1(k) +

02 (k) vr f 2I ")

A(az)2 g I 2M) - 2,2(k) + $ (k-1)}

2

, k=2, M-1 .-

a2")*1(k)

I Where:

B = radial buckling, ,

az - axial mesh size, J X - eigenvalue, k - axial node index; k - axial node index; K-2: bottom node of the core and k=M-l: top-node of the core. '

For a partially controlled node, the flux and cross section values would be determined from values of fully controlled node and fully uncontrolled node by linear interpolation.

1 2999C l

]

GPU NUCLEAR TOPICAL REPORT TR-033

7. Table 4-1 and Figures 4-6 through 4-19 provide 3-D/1-D comparisons of eigenvalues and axial power shapes for a set of dependent cases (control fraction varied while other independent variables held constant). Provide similar comparisons for a set of perturbed casos (e.g., void fractions and fuel temperatures varied to values approaching their transient limits) other than the ones used in collapsing 3-D to 1-D cross-sections.

RESPONSE

A comparison between RETRAN and SIMULATE power shapes for perturbed cases is presented in Figures 7-1 to 7-4. The perturbations were implemented by l pressure setpoint changes in RETRAN that would take the model from normal steady state conditions to a new steady state at pressures 20 psi and 35 psi higher. In this way, void collapse will cause a power increase with a corresponding fuel temperature change followed by power changes due to feedback effects until it reaches a new steady state at the new pressure, The new steady state conditions were used in SIMULATE to calculate a power f shape and the eigenvalue which were compared against RETRAN new power

! shape at the new steady state. The same procedure was applied to a loss I

of feedwater heating transient where a 100*F decrease in feedwater l

temperature was initiated in RETRAN and the final steady state condition were used in SIMULATE as outlined above. The pressure perturbations are typical of pressure increase close to peak power in a Turbine Trip Without Bypass transient and the average fuel temperature increase is approximately 100*F, which is closely simulated through the loss of  !

feedwater heating transient with pressure under pressure regulator l control. It is difficult to separate void and temperature effects in RETRAN. It was felt that the approach used was best suited to provide '

meaningful comparisons since in reality, void and temperature effects cannot be isolated.

As shown in Figures 7.1-7.4, the 1-0 axial power shape agrees very well with the 3-0 axial power shape. The largest difference at the peak node is 3.7%. The eigenvalue comparison is presented in Table 7.1. There is almost no change in reactivity between the RETRAN initial case and the perturbed cases. This is expected since the perturbed state conditions were defined by letting RETRAN runs reach the steady state. There is a 1-2 mk difference between eigenvalue of the SIMULATE initial and perturbed cases. Considering the thermal-hydraulic modeling difference between RETRAN and SIMULATE, this slight change in eigenvalue is quite acceptable.

The perturbed cases shown here cover the expected change of core pressure and fuel temperature at peak power during a Turbine Trip Hithout Bypass transient which is the most severe pressurization transient for OC. The results presented here demonstrate that the method used for generating 1-D kinetics parameters is adequate for reactor transient analysis.

2999C l

N , 2 4

3" O ,

2 2

S 2

% 1 6

% 2 2 RE 0

4. N 0 , 2 S RI OT

- O , 9 A

1 RI A

'A C RV E D E

, 7 8

1 V _  %

E GD AR ,

1 6

EA RA 1 5

2 ED ,

5 E NNTA 1 CEN 4 D1 O7 S

R 1 O -

R R

3 NE 1

R  :

E E

D 2 L 1

1 AG I

O E T

O N

1 XF A ,

0 A W '.

L U

K AI E

, 9 1

CE M I

PI , 8

_ S R E

3C  : W O

1

, 7 Y P

, 6 C N D 5

A E A R Z T L E A I

, 4

, 3 X RM R

i 2

A = O N

4 2 0

8 6 4 2

- a i" g

1 1 1 0 0 0 0 g

A NE I 4

2

\ I 3"

2 "m \ 2 S55 I

2

% 1 2 I 2

% 3 1 0 0'. 4. N:

0

O I

2 A

3E RI OT RI A RV I

I 9

1 8

1 E DE R 7 v_  %

4 E

G A D RA R

I i 6 1

1 "mSS 3

R

- ED NNTA i 5 E 1 O2 "o E O R

R S

i 41 3 NE C7 R

I

- E 1 RJ E D

O I

I 2 L 1

1 AG I

E E N 1 XF T

A K A

1 0 A 1

WV L U

E Pi I 9

"II M

i I 8

S R E

3  :

1 W

O

'l 7

6 P

O .

N D i 5 A E A'E R T L E A Z

I I

t 4

3 X C

_ RM R

I 2"

A Y

= O N - - - - _ -

1 i ""

C 4 1

2 1

0 1

8 0

6 0

4, g

2 0

l

A NE '\

, 3 4

2" u \

2 2

mQ2 7

% 34 2

1 2

0 0 4.

0

- O N ,

2 A

RI , 9 3-

/ OTA 1 RI 3

E RV EE

, 8 1

R v_  %

4 3

E GO AR RM D , 7 1

6 1

ou S S -

R

- EE NNTA 5 E 1

D3 aE O R

R S -

, 4 1

O7 R E

, 3 1

NE-

,1 3 E D

2 L 1

- O N

1 AG I

E 1

E T K Xi A A 0 A V

1

- L E U P! , 9 O

M y uZ I

I , 8 S R

- E . ,

7 W .

O ,

6 P

0 N E D 5

- A A 4 R Z a/ = T L E A I

Y ,

3 RM 2""

vA AC

- R '

= O a /- Y

, i N - - - - - - "

C 4 1

2 1

0 1

8 0

6 0

4 0

2 0

o, g

.1 , I

G N

N T -

i 4 2

CA i 2 3"

SE 4  %

i 2 2

- 9 i 2 1

R R  %

7. 3 0 N

, 0 2

AE R O I

9 OT ,

1 RI A 3 ~

RV , 8 V E 1

EE E

D , 7 1

GD AR 6 E ]W

% RA 1 7 ED 3 N , 5 E CEE NA 1 T

S 4 D4 O7 R 1 O

3 NE R== R ,

1 R

E , 2 L 1 AG E

E D .

1 I O XF 1

T O NS

\

A L

U N

K A

, 9 0 A 1

CS I M

S R E

E P '

, 8 3  :

W i'

, 7 O_

P O 8

, 6 0:. N E D , 5 AE A R Z T L E A I

4 X _

R M R

, 3

, 2" ACY = O N

6 4 0

8 6

, 1 C 1 1 i 0 0 4

0 2

0 g

g

Table.7.1 Elgenvalue Comparison-

.[

SIMULATE-E RETRAN Initial Core Dome Pressure'1035 PSIA 1.003 1.003 Dome Pressure 1056 PSIA 1.002 l.003 Dome Pressure 1072 PSIA' 1.002 1.003 Loss of feedwater Heating 1.001 -1.003;

.I i

l

1

-i k

i 3

I i!

2999C

__-_ _ _ - _ _ _ _ - _ _ - _ _ _ - _ _ _ _ - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ - _ _ _ _ _ _ . __ _ ._ _ _ - . _a

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040

1. Fuel Assembly Misloading i 1.1 Supply data supporting your conclusion that for a disoriented assembly the ACPR is more severe at 180* than at 90*. Since CPM is unable to calculate the 90* cases, have the 90* and 180* cases been j evaluated with the same code? Show comparisons of ACPR's for 90*- l and 180* assembly disorientation as a function of exposure and other )

pertinent data. j l

RESPONSE: 'l l

Table 1.1 below contains data demonstrating that the 180*

disorientation is more severe than the 90* disorientation. PDQ was used to analyze both the 90* and 180* disorientation. In each case, the 180* disorientation had a higher local peaking factor, R-factor, and increase in bundle power over the 90* disorientation . The ACPRs calculated for Table 1.1 assumed that the correctly loaded ,

bundle was on its CPR operating limit. The change in bundle power l for the 90* and 180* disorientation used values calculated by PDQ and 1 not an upper bounding value as used with the CPM analysis. Also, the  :

P0Q calculations did not include the axially varying water gap and )

hence the PDQ calculated ACPRs are higher than the CPM analysis. ]

Table 1.1 Comparison of 90* and 180* Disorientation Exposure ACPR GWD/MT 90* 180*

1 O 0.196 0.250 5 0.201 0.365 10 0.114 0.241 15 0.198 0.416 4

2999C 1

_ -- _ _ - - - - - a

. j l

l REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040 l

1. Fuel Assembly Misloading (Continued)  !

1.2 Describe in detail the four-assembly CPM-PDQ calculation. Is the PDQ i

calculation performed with a pin-by-pin geometry representation? How many cross section groups are used in this calculation? Have these calculations been benchmarked? l l

RESPONSE

i The four assembly PDQ calculations used a pin-by-pin geometry j representation with 44 by 44 mesh spaces and each pin having a 2 by 2 j mesh spacing. PDQ used four group cross sections obtained from CPM.  ;

For the gadolinia pins, I.., vI,. and rI<. edited ]

from CPM were adjusted for input to PDQ such that the reaction rates j in PDQ matched CPM. Also, I, and I,. from CPM were adjusted to reflect that PDQ does not consider upscatter from'the thermal to the epithermal group. Single assembly PDQ calculations, using the same mesh spacing as in the four assembly PDQ, were compared to CPM.

The agreement between PDQ and CPM was very good for K-infinity and local peaking factors.

1 i

l l

l I

2999C ]

R l

1 l

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040- )

.j

1. Fuel Assembly Misloadino (Continued) 3 1.3 What margins are applied to the ACPR calculated for the misloaded -

assembly and the disoriented assembly to account for uncertainties in the calculation of the bundle power?

RESPONSE: {

i No margins are applied to the ACPR calculated in these analyses to i account for uncertainties. 'The nature of the analysis is to provide i a conservative calculation to account for uncertainties in the analysis.

The conservatism in the GPUN methodology can be'seen in comparing the  ;

fuel bundle disorientation results to that of the current licensing i

analysis performed by the fuel vendor. Table 1.3 shows the fuel bundle disorientation results for both analyses. The fuel vendor ,

analysis includes a 0.02 adder for uncertainty in the axial Rfactor. d With the adder, the ACPR for vendor analysis is less than or equal to GPUN calculation ACPR.

]

l Table 1.3 Fuel Design Fuel Vendor

• GPUN ENC VB 0.20 0.23 i GE P8X8R 0.18 0.18 i

Reference:

1. " General Electric Reload Fuel Application for Oyster Creek," NEDO 24195, 1983.

2999C

REQUEST FOR' ADDITIONAL.INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT.TR-040

1. Fuel Assembly Misloading (Continued) 1.4 Demonstrate that the 3.2% increase in the bundle power of a disoriented assembly is an upper bound for all current and projected.

assemblies.

RESPONSE

The table below shows the calculated increase in bundle power.for current and future O.C. assemblies.

1 Table 1.4 Bundle Power' Increase for 180* Disorientation Bundle Percent Power Increase ENC VB 1.94%

GE P80RB239 2.26%

GE P80RB265H 0.70%

GE P80RB299 1.60%

GE 803218 (Cycle 12 Fuel Design) 2.30%

2999C L-__--_______------.-___---______----__ - - _ - - - _ - - - - - _

l

~ . . .

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040. j

1. Fuel Assembly Misloading (Continued) 1.5 What.is the effect of fuel misloading on the R-factor, bundle power  !

and.ACPR and how is this effect accounted for in the analysis? J

RESPONSE

l In a fuel bundle disorientation, the.R-factor and' assembly, power.

increase which reduce CPR and. increase the ACPR. The analysis. ,

accounts for both of these effects. The change in R-factor is l calculated axially, taking into account the effects of tilting.

Assembly power is assumed to increase by an upper. bounding value.

In a fuel bundle dislocation, the R-factor and bundle power will be-  ;

calculated by the core monitoring system for the correctly loaded assembly while an incorrectly loaded assembly will be in its place.

In the case where a new assembly replaces ~a highly exposed assembly, the change in R-factor will be fuel type and exposure dependent while the power will significantly increase over what is calculated,by.the  ;

core monitoring system. The ACPR between the CPR of the correctly i loaded assembly and the CPR of the mislocated assembly will increase as the difference in the assembly power increases. The analysis >

accounts for this by calculating the both the correctly loaded CPR (at a symmetric location) and the CPR for the mislocated assembly during the cycle and identifying the assembly and time in the cycle which produces the largest ACPR.

l 1

2999C L

-'~

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040

1. Fuel Assembly Misloading (Continued) 1.6 Show that replacement of an exposed assemb1'y by a fresh assembly in a control cell results in a maximum ACPR for the range of fuel. types used in Oyster Creek.

RESPONSE

The following. table shows the resulting ACPR'by (a)' replacing a.  !

fresh fuel assembly and (b) by placing an exposed fuel assembly at peak reactivity in the same core location instead of a highly exposed-fuel assembly.  ;

The control cells having tiiese core locations are the potential.

l candidates for the maximum ACPR. An ENC VB fuel assembly at peak l reactivity was used in this analysis since its peak reactivity is higher than other OC fuel designs.

Table 1.6 1  !

Comparison of Exposed and Fresh Assembly Dislocation l '

1 l

ACPR Exposed F.A. at  !

Core Location Peak Reactivity Fresh (31, 34) 0.13 0.27 (23, 38) 0.17 0.21 (15, 30) 0.14 0.26 l

l (15, 38) 0.19 0.29 The results show that the maximum ACPR is higher by replacing a highly exposed assembly with a fresh fuel assembly in a control cell.

l l

2999C q

'~

REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040 i

1. Fuel Assembly Misloading (Continued) 1.7 Demonstrate that the procedure leading to the selection of the highest cell delta-exposure and highest cell-average exposure ^

determines the fuel loading error having:the highest ACPR for present and future Oyster Creek reloads.

RESPONSE

The selection procedure for the potential candidate cells,' capable of generating the highest ACPR, was carried out for Cycle 10 core.

However, for the verification purpose, each control cell was analyzed for the fuel assembly dislocation error.

The following table shows the selected potential core locations for maximum ACPR. The table also shows the maximum ACPR by misloading different fresh fuel-types at various core locations.

Each core location corresponds to a separate control cell. The results confirm that the selected core locations generate higher " MAX ACPR".

Cell Cell Potential Calculated Maximum ACPR For Core Average Delta Candidates Location Exposure Exposure for Max. GE 2.65 -GE 2.39 EXXON VB- ,

X.Y GWO/MT GWD/MT ACPR? Fuel Fuel. Fuel 23, 30 5.83 2.93 0.16 0.16 0.07 31, 34 9.23 4.65 Yes 0.19 0.27* 0.10 23, 38 8.57 4.54 Yes 0.21 0.21 0.12 31, 42 7.27 2.95 0.16 0.19 0.14 23, 46 6.73 3.94 0.22 0.20 0.11 35, 30 9.24 4.65 Yes 0.22 0.26 0.15 ,

19, 34 9.12 4.55 Yes 0.26 0.21 0.15 19, 38 9.25 4.62 Yes 0.19 0.21 0.15 19, 42 9.12 4.62 Yes 0.18 0.22 0.14 19, 46 6.61 3.71 0.10 0.14 0.06 15, 30 8.59 4.56 Yes 0.26 0.23 0.13 15, 34 9.24 4.64 Yes 0.17 0.15 0.12 15, 38 8.95 4.53 Yes 0.29* 0.25 0.17*

39, 42 5.70 2.92 0.21 0.21 0.11 11, 22 7.29 4.31 0.22 0.17 0.14 11, 34 9.14 4.50 Yes 0.23 0.20 0.14 43, 38 5.69 2.78 0.24 0.11 0.03 11, 42 7.80 3.90 0.06 0.06 0.01 7, 30 6.77 3.93 0.24 0.18 0.08 7, 34 6.58 3.70 0.13 0.16 0.05 Maximum for Assembly Analyzed.

2999C

REQUEST FORiADDITIONAL INFORMATION REGARDING

.GPU HUCLEAR TOPICAL' REPORT TR-040' l

2. Control Rod Withdrawal Error  ;

2.1 Describe 1the core conditions assumed in the static calculations of the rod withdrawal error. (

RESPONSE:  ;

The general: core' conditions assumed for the static calculation are.

that'the core is at' hot full rated power and recirculation flow with a rod pattern that places a bundle within a 6X6 fuel assembly block ,

near thermal limits.~ The rod withdrawal error transient calculation J is performed with no xenon present at the point in the. cycle of peak reactivity. ,

I

! ')

l l J

I

l l

I 1

2999C m__-----__--_______________ _ _ _ _ _ _ _ _ _ _ Li

'.- l ., . '

REQUEST FOR' ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040

2. Control Rod Withdrawal Error-(Continued) 2.2 'Specify the core power, exposure, fuel loading, control rod pattern and flow in the rod withdrawal error analysis depicted in Figures 3-1 through 3-3 and. Table 3-1.

RESPONSE

Tables 2.2-1 and.2.2-2 and Figure 2.2-2 contain the information requested.

l l

l l

2999C

i

,. e Table 2.2 1 CONTROL R00 PATTERN FOR LFHH ANALYSIS l

-)

26 30 34- 38 42' 46 50 51.

18 24- 47  :

l 43- _j l

06 12 22~ 39 1

35 1

00 18 00* 31

)

i l 27 l 8 l

l NOTES: 1. Rod pattern is 1/4 core mirror symmetric.

~'

2. No. Indicates number of notches withdrawn out of 48. Blank is a withdrawn rod.

3. Asterisk (*) denotes the transient control rod. -)

i Core Exposure: 3.0 GN0/MT (peak cycle reactivity)

Reactor Power: 1930 MW(th)

Recirc Flow: 61.000 MLB/HR  ;

Void Fraction:' 35.4% I System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb l

1 2999C ,

l

)

Table 2.2-2 Transient Conditions in.RWE Analysis )

Control Rod l FT Withdrawn Reactor Power (MW)  !

0 1930.0 .

l 2 1959.0 )

l 4 2028.0 -

6 2066.0  ?

8 2068.0 1

.?

l 1 l l l

t I

i l

i 1

1 2999C l

P

.. .. )

I' Figure 2.2 Core Loading Pattern for Cycle,10 , l

oMMMMMo  !

- mMMMMMMMEL 1

MMMMBiMMMMMM 1 BIMMMMMMMMMEo

MMMMMMMMMMMMM '

I+iMMMMMMMMMMMM

MMMMMMMMMMMMM

MMMMMMMMMMMMM

'::MMMMMMMMMMMMM MMMMMMMMMMM5

lMMMMMMMMMMM.

,!TEMMMMMMMM*

t

i ' =MMMMM5  ;

ll l l l l l l l l 1l i

! 3 5 7 9111315171921232527293133353739414345474951 FUEL TYPE A = EXXON-VB D = EXXON-VB B = EXXON-VB E = P8DRB239 C = EXXON-VB F = P8DRB265M 2999C l.

'l

.: .. i l

REQUEST FOR ADDITIONAL INFORMATION REGARDING' i GPU NUCLEAR TOPICAL REPORT TR-040

2. Control Rod Withdrawal Error (Continued) i 2.3 What margins are assigned to the ACPR's in order to' account for f uncertainties in the calculation? l RESPONSE: l No margins are added to account for uncertainties. The calculation 1 '

of ACPR during the rod withdrawal error transient is conservative.

There are many inheren't conservatism in the analysis. The control rod patterns are not realistic, in that they force power l peaks and MCPR around a control' rod that is fully inserted; no Xenon is present to maximize power response; instrumentation is-in its worst operable l

configuration; power increase due to rod withdrawal (see response to question 2.2) is very conservative;-and-it occurs during peak cycle reactivity. These inherent conservatism are adequate to account i for uncertainties and do not require additional nargins. i l

i l

1 l

)

l i

i 2999C 1

REQUEST FOR ADDITIONAL INFORMATION.REGARDING GPU NUCLEAR TOPICAL REPORT TR-040

2. Control Rod Withdrawal Error (Continued) 2.4 In view of the large ACPR reduction that occurs during the RWE, demonstrate that the procedures _ outlined in Section 3 uniquely identify the largest undetected ACPR, when variations-in core loadings, fuel types, failed LPRM detectors and APRM channels are taken into account. That is show that the two calculations performed i in the RWE analysis (strongest rod and largest combination of.falled LPRM's and APRM channels) are bounding.

RESPONSE

The RWE analysis in general is forced to be_ bounding by performing two separate analyses. In this way, the most conservative method is used for licensing purposes. Even so, each individual analysis (Limiting Rod and Limiting Location) is conservative in the approach taken to evaluate ACPR. This is demonstrated below:

A. Limiting Rod (LR)  :

In performing the LR analysis, the control rod with the highest j worth is determined and used for the RWE analysis. A rod with  ;

the highest worth will insert the greatest amount of positive j reactivity, thereby producing higher localized bundle and core i powers. This will provide the highest ACPR since the change 1 in bundle power will be higher for a high worth rod than for a low worth rod. The worth of the rod is dependent on the fuel 1 designs near the rod and on the core loading. The location of i the highest worth rod may vary from cycle to cycle and will take j into account changes in fuel designs and core loading. The  ;

analysis also looks at the most limiting combination of l failed / bypassed LPRM and APRM channels allowed by technical specifications.

B. Limiting Location (LL)

The LL analysis is performed to ensure that conservatism is maintained by virtue of the error rods' location with respect to the instruments used to mitigate the transient. The LL analysis is conservative in that the RHE will be performed in a location which provides the slowest LPRM/APRM response with the most limiting instrument failure / bypass combination. This provides inherent conservatism since it is the detector responses to the transtent that initiates the rod block which terminates the transient. The slower the response, the longer it takes to reach the rod block setpoint, the higher the local bundle powers will be, the greater the ACPR.

2999C

l

2. Control Rod Withdrawal-Error (Continued)  ;)

I 2.4 Continued q Both methods above are designed to generate'a conservative response i to the RWE. Each analysis is performed using the same conservative 1 assumptions as explained in the answers to questions 2.1 to 2.3. .

l However, these responses are further analyzed to calculate ACPR and' :J LHGR by determining the limiting combinations of instrument failures 1 and channel bypasses as allowed by technical specifications. .The  :

calculated LPRM responses from NODE-B are used to evaluate and '

determine the limiting instrument response configuration, and-the resulting rod block point, aCPR and LHGR. j I

1 j

)

1 f

$ 4 1

i i

2999C

1

• ' I REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040- 1 i

3. Loss of Feedwater Heating ,

3.1 Specify in detail the plant operating conditions a'nd the assumptions I used in the three-dimensional-evaluation the loss of feedwater 1 heating (LFWH) event. j l

RESPONSE

. I The plant operating' conditions and assumptions used in the' D three-dimensional LOFWH event are: -!

l

a. The plant is operating at full power and full flow. Rod patterns-l 'I used for different times in the cycle.are such that K-effective is t .002 of K-critical;

b. The. loss of feedwater heating reduces feedwater temperature by.

l 100*F; 1

l c. All other systems operate as designed and continue to function throughout the event; i

No operator action is accounted for; d.

e. Feedwater flow increases to maximum level;

f. Pressure remains constant; l

g. Final porar level increases to scram point, but no scram is initiated;

h. and, Xenon levels are maintained at full power level for '

increased peaking.

l 2999C

?..

REQUEST FOR ADDITIONAL.INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040

3. Loss of Feedwater Heating (Continued)-

1 3.2 In the topical report, loss of feedwater heating is attributed to the. I closure of the steam extraction line causing a gradual cooling of the l feedwater. Loss of feedwater heating, however, can also result when feedwater is bypassed around the. heaters, causing a faster cooling of i the feedwater. How is this second feedwater heater transient 4 accounted for in the analysis?

i

RESPONSE

-l Oyster. Creek feedwater~ heaters.cannot be bypassed and only a closure i of the steam extraction line will cause a loss of feedwater. heating. '!

However, the RETRAN analysis looked at a. loss of feedwater heating j occurring in a one second and a thirty'second period. The' loss of' '

i feedwater heating in one second is.slightly more severe ^in terms of ACPR, but still bounded by the N00E-B analysis.

l l

l 2999C 1

l

I

• i f

REQUEST FOR ADDITIONAL'INFORMATION REGARDING-GPU NUCLEAR TOPICAL REPORT TR-040

3. Loss of Feedwater Heating (Continued) 3.3 What benchmarking of the.GPU analysis methods has been' performed for events of this type? Have comparisons of N0DE-P with Oyster Creek. J startup measurements been made? What uncertainty allowance has been included in the ACPR calculation to insure that the MCPR safety' limit is not violated?

RESPONSE

There is no OC data available with which the LOFWH transient can be.

directly benchmarked. The LOFWH transient.has been run with RETRAN, i which in turn was benchmarked against OC startup. test data (reference 1). Based upon RETRAN results, it was determined that the LOFWH -

i analysis could be conservatively analyzed with NODE-B. Peak power'in RETRAN for EOC 10 is 2192 MW, 41'MW (1.8%) less than the scram point .

and the power level used in N00E-8. The analysis.is run to: insure a j conservative response (see response to question 3.5) and no specific uncertainty allowance has been included.

~

1 i

I l

)

1

-1 l

l

References:

1. M. A. Alammar, et al, "BW-2 Transient Analysis Model Using the RETRAN Code," GPUN TR-045, Rev. O. September 3, 1987.

2999C

.j

I REQUEST FOR ADDITIONAL INFORMATIO'N REGARDING i GPU NUCLEAR TOPICAL REPORT TR-040  ;

i

)

3. Loss of Feedwater Heating (Continued) )

3.4 Describe the statepoints and initial conditions-(exposure, rod- .,

pattern, flow, void. fraction,' pressure, inlet subcooling) selected l I

and demonstrate that this selection bounds the maximum ACPR. How.

many N00E-B calculations are performed in a typical LFHH analysis?'

RESPONSE

]

Figures-3.4-1 to 3.4-8 describe the initial conditions at various statepoints during the core life. LFHH was analyzed for Cycle 10 at  ;

every 1.0 GWD/MT interval from BOC to EOC. j

'I The core conditions used for this analysis are based on a control rod -l stepout of the cycle to meet target power shapes. It is expected l that actual operations will result in different conditions, 1 l especially control rod patterns. However, these differences are.not significant in terms of ACPR in the transient since rod density will not vary significantly. Other core conditions may change,'but the relative effect of a 100*F loss in feedwater heating will not be impacted.

The analysis is performed at 1.0 GWD/MT intervals and the number performed will vary with cycle length.

l j

i i

1 4

2999C

,. .i l <

! 4

. . ..i,

Figure 3.4-1 CONTROL R00 PATTERN FOR LFHH ANALYSIS 2 6 10 14- 18 22 26-51 47 43 20 '28 39 35 20 36- 10 31 27 28 10 16 i

i NOTES: 1. Rod pattern is 1/4 core mirror symmetric.

~

2. No. Indicates number of notches withdrawn out of 48. Blank.is a .i withdrawn rod.

I Cycle Exposure: 0.0 GN0/MT Recirc Flow: 61.0 MLB/HR Void Fraction: 35.3%

System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb 2999C

1: s Figure 3.4-2

^

. CONTROL ROD PATTERN FOR LFWH ANALYSIS 2 6 10 14- '18 '22. 26 51 47 I 43 20 28 39' 1

\

i 35 20 36 8 i t

i 1

31

]

i 27 28 8 16 j i

NOTES: 1. Rod pattern is 1/4 core mirror symmetric. I

2. No. Indicates number of notches withdrawn out of 48. Blank is a withdrawn rod.

Cycle Exposure: 1.0 GWO/MT )

l Recirc Flow: 61.0 MLB/HR l l

Vold Fraction: 32.4%

System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb -)

x 2999C -

J 4 l a . .. .

Figure 3.4-3 CONTROL ROD PATTERN FOR LFHH ANALYSIS ,

']

1 2 6 10 14' 18 22 26-51 47 16 32 l 1

1 43-  !

1 1

1 39 32 8' 24 l 35 1 31 24 14 4 1

27 i I l i

.1 NOTES: 1. Rod pattern is 1/4 core mirror symmetric. .j I

2. No. indicates number of notches withdrawn out of 48. Blank is'a j withdrawn rod. ^

l Cycle Exposure: 2 GHD/HT l Rectre Flow: 61.0 MLB/HR j l

l Void Fraction: 35.0%

System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb' 1

l l

1 1

1 .!

! t 2999C

3 <

Figure 3.4-4

-CONTROL ROD PATTERN FOR LFNH ANALYSIS-d

.2 6 10 14 18- 22. '26 ' ,

51  :.

j 47- 24 d

{

43 1 39 18 4

)

q 35 36

1 31 24 4 14 27 j NOTES: 1. Rod pattern is 1/4 core mirror symmetric.

2. No. Indicates number of notches withdrawn out of 48. Blank is a withdrawn rod.

Cycle Exposure: 3 GND/MT Recirc Flow: 61.0 MLB/HR Void Fraction: 35.51 System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb l

j 2999C f

1

.i <

-J Figure 3.4-5 1 i

CONTROL ROD PATTERN FOR LFWH ANALYSIS 1 1

2 6 10 14 18 22 26 51 I I

47 36 I i

-l 43 20 14 1

1

.1

'39 42 i i

35 24 0 8 31 36 I

1 27 18 12 24 l lj NOTES: 1. Rod pattern is 1/4 core mirror symmetric.

2. No. Indicates number of notches withdrawn out of 48. Blank is a 'l l withdrawn rod.

Cycle Exposure: 4.0 GWO/MT  ;

Recirc Flow: 61.0 MLB/HR Void Fraction: 33.1%

System Pressure: 1050 PSIA  ;

Inlet Subcooling: 33.1 BTU /lb 2999C l l

.> 5 ..

Figure 3.4-6 CONTROL ROD PATTERN FOR LFWH ANALYSIS I 2' 6 10 14 18- 22 26 51 47 ' 38 -

43 12 24~

l 39 -40.

35 12 "18' 8-l l

31 38 l

l 27 24 -8 12 l

NOTES: 1. Rod pattern is 1/4 core mirror symmetric.

1

2. No. Indicates number of notches' withdrawn out of 48. Blank'is a-withdrawn rod.  ;

State Point Exposure: 5 GHD/MT Recirc Flow: 61.0 MLB/HR Void Fraction: 33.6%

System Pressure: 1050 PSIA Inlet Subcooling: 33.1 BTU /lb  ;

2999C a_-____-__-_--___---____ _ _ _ _ _ .

s-c- c.

..~.. .

4

. Figure 3.4 CONTROL R00 PATTERN FOR LFWH ANALYSIS 2 6 10 14 18 22 26 )

l 51 .i 47 24 42'-

]

43 1 i

l l

39 36 14, 28 l 35-31 20 8 0  ;

27 i 1

l i

l NOTES: 1. P.;d pattern is 1/4 core mirror symmetric.

2. No. indicates number of notches withdrawn out of 48. Blank is a withdrawn rod.

t Cycle Exposure: 6 GHD/MT-Recirc Flow: 55.815 MLB/HR Void Fraction: 33.8%

System Pressure: 1050 PSIA.

Inlet Subcooling: 33.128 BTU /lb i

i 2999C

'J #

es s . ', .

Figure 3.4-8 I

CONTROL R00 PATTERN FOR LFHH ANALYSIS 2 6 10 14 18 22 26 51 47 l

l 43 j 39 All Rods Out g 35 1

31 i 27 NOTES: 1. Rod pattern is 1/4 core mirror symmetric..

2. No. indicates number of notches withdrawn out of 48. Blank is a-withdrawn rod.

Cycle Exposure: 6.31 GH0/MT (EOC)

Rectre Flow: 61.0 MLB/HR Vold Fraction: 36.1%

System Pressure: 1050 PSIA 1

)

Inlet Subcooling: 33.128 BTU /lb I

l

'I l

I 2999C 1

.I 's REQUEST FOR ADDITIONAL INFORMATION REGARDING  !

GPU NUCLEAR TOPICAL REPORT TR-040

3. Loss of Feedwater Heating (Continued) 3.5 Describe the procedures used to analyze the LFWH event including the-heat balance used to relate the feedwater temperature reduction.to core inlet subcooling.

RESPONSE

The 3-0 simulator model used for this analysis employs an integrated j neutronic (NODE-B) and thermal-hydraulic (THERM-B) as described.in reference 1 of TR-040. At BOC and each~1.0 GWD/MT through EOC, a N00E-B calculation is run at full power with full" flow and feedwater flow and temperature at' full power conditions. Control rod patterns  ;

are set to maintain thermal limits and K-effective at +0.002 of  !

K-critical. The inlet subcooling is calculated by H;o!H... where l H... is an input quantity-and H,n is given the equation: ,

1 1

Hi, HT [(WT - WST - WCU)HF + HCU*HG + WFW*HFW )

i

+ HRD*WRD + QPUNP - QLOSS - QCL]

where:

H., - Core inlet coolant enthalpy, BTU /lb; i 1

WT - Total core flow, lbs/hr; WST - Steam flow leaving reactor vessel, lbs/hr; WCU = Steam carryunder in recirculation flow entering, downcomer, Ibs/hr; WFW - Feedwater flow, lbs/hr; QPUMP = Energy from recirculation pumps, BTU /hr; QLOSS = Heat loss from reactor. vessel, BTU /hr; HF = Enthalpy of saturated liquid entering downcomer (evaluate at dome pressure), BTU /lb; HG - Saturated steam enthalpy of carryunder, BTU /lb; HFW = Feedwater flow enthalpy, BTU /lb; HRD = Control rod drive flow enthalpy, BTU /lb; WRD - Control rod drive flow, lbs/hr; and QCL - Energy loss to the cleanup system, BTU /hr.

2999C

1 4

, e bA e,a " ..

I

3. Loss of Feedwater Heatino (Continued) 3.5 (Continued)

The flows (HT, WFW, and HRD) and QLOSS are' input to.N00F-B. -The steam flow is assumed to be equal to (HFH > WRD). The saturated' 1 liquid and steam enthalpies are evaluated based on the dome and core 1 pressure which are input to the code. The enthalples.for the- I feedwater flow and control rod drive flow are evaluated based on the inlet coolant temperatures. The energy from.the recirculation pumps is determined from the power drawn by each motor and the efficiency of the pumps; these values are input to the code. The steam carryunder flow is defined as a fraction of the total core flow.

1 A second NODE-B calculation is performed with the power level at the d scram point (2233 MW) and feedwater flow increased to maximum and feedwater temperature at 100*F less than the initial NODE-B core.

The inlet subcooling is recalculated using these revised inputs. 1 The maximum change in CPR for an assembly is the ACPR for the transient. The procedure to determine the maximum 6CPR includes i the following. Since the assembly having the MCPR may vary with rod pattern, the ACPR is calculated for the assemblies having CPR near the CPR limit. The ACPR for these assemblies is adjusted (as described in response to question 3.7) to reflect ACPR as if each assembly had been operating at the CPR limit. This both identifies the fuel assembly having the largest ACPR and allows for variation in the control rod pattern which may place a different assembly on the operating limit. '

If the K-effective of'the second N00E-B case (at 2233 MW) is higher than the initial case K-effective, it would be indicative that a l higher power level is possible. Therefore, a LFHH transient would be j more severe.from a lower power. In this case, power level is I increased to give a lower K-effective than occurs in the initial l case. Allowance for convergence criteria in the K-effective is considered in the analysis.

l I

2999C a

. .Y

,o.5 REQUEST FOR ADDITIONAL INFORMATION REGARDING l

{

GPU NUCLEAR TOPICAL REPORT TR-040 #

3. Loss of Feedwater Heating (Continued) 3.6 How is criticality maintained when the power and inlet subcooling are varied independently? If criticality.is not maintained between the-  !

initial and final states, what is the effect of.this approximation?- '

RESPONSE

Criticality is not maintained when the power and. inlet subcooling are varied independently. The K-effective for the final state is.less than the K-effective of the initial state indicating the final power is higher than would be possible. The effect of this approximation  ;

is to produce an artificially high power level.and therefore a conservative result.

l 1

i 2999C

.A a . om .

REQUEST FOR ADDITIONAL INFORMATION REGARDING-GPU NUCLEAR TOPICAL REPORT TR-040

3. Loss of Feedwater Heating (Continued) 3.7 How is the calculated ACPR adjusted to the. initial CPR (ICPR)'ard' what is the basis of this adjustment? Is this adjustment' conservative?

RESPONSE

The adjustment is performed as follows:

l ACPR t - ICPR - MCPR X : Operating Limit CPR-ICPR The adjustment is performed since the calculated ACPR of a fuel assembly will vary dependent-upon the initial-CPR of.the assembly.

In order to compare the ACPR from one assembly to another,-the ACPR is adjusted to reflectLthe same ICPR~ equal to the operating.

limit. While this may result in a' smaller'ACPR in some instances, the adjustment is neither conservative nor~non-conservative in that it reflects the ACPR-for an assembly at the CPR operating limit.

2999C ,

l____.___________________._______ _ _ _ _ _ _ _ _ . _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ . _ _ _ _ _ _

, . nw REQUEST FOR ADDITIONAL INFORMATION REGARDING GPU NUCLEAR TOPICAL REPORT TR-040 4

~

3. Loss of Feedwater Heating (Continued) 3.8 Are the system variables such as pressure,.feedwater. flow, steam flow, etc., assumed constant during this transient? If so, provide the basis for this assumption. Can changes in these variables result.

in a limiting ACPR during the transient making the final-state ACPR calculation not bounding?

RESPONSE

System press'ure.is assumed to remain constant during the transient. ,

Feedwater flow is' assumed to increase to maximum and remain I constant. Reactor power and steam flow increase to a constant level. FH temperature is assumed to decrease by 100*F. These.

assumptions are supported by RETRAN analyses. .Feedwater increases i with power until maximum flow is attained. Pressure drops slightly (approximately 10 psi), but the pressure regulator returns pressure back to its initial operating level. The drop in pressure is '

non-conservative in terms of CPR, however, the 10 psi change will have a minimal effect on CPR. Power level increases then drops slightly to an equilibrium level. The peak level is 2192 MN, less than the value used in the NODE-B analysis. The final state levels used in NODE-B will still provide a conservative analysis.

i

')

I 2999C