ML20101B661
| ML20101B661 | |
| Person / Time | |
|---|---|
| Site: | Wolf Creek  |
| Issue date: | 05/29/1992 |
| From: | Rhodes F WOLF CREEK NUCLEAR OPERATING CORP. |
| To: | NRC OFFICE OF INFORMATION RESOURCES MANAGEMENT (IRM) |
| References | |
| ET-92-0103, ET-92-103, NUDOCS 9206040260 | |
| Download: ML20101B661 (16) | |
Text
_ ___ .. __ _-- _
)
[ .
t t-
- t .
6 L~-
I W@ NUCLEAR LF CREEK OPERATING C ;
s j Forrest T. Rhodes May 29, 1992 l Vice President j rnpneering & Tt<hNcal Service.
! ET 92-0103 j U. S. Nuclear Regulatory Commission j ATTN: Document Control Desk l Mail Station F1-137 '
) Washington, D. C. 20555 I
1 References 1) Letter dated December 27, 1991 from W. D. Reckley, l USNRC, to B. D. Withers, WCNOC
., Subjectc Docket No. 50-482: Response to Request for Additiona) j Information on the Core Thermal-Hydraulic Analysis
- Methodology for the Wolf Creek Generating Station ,
i Gentlemen l
The purpose of this letter is to submit Volf Creek Nuclear Operating i Corporation's (WCNOC) response to a Request for Additional Information (RAI)
, provided in Reference 1. -The RAI concerns WCNOC's " Core Thermal Hydraulic-
! Analysis Methodology for the Wolf Creek Generating Station
- which was submitted in Reference 2.
At a meeting en Janua_ry _2 8, 1992. . WCNOC _and_ the Nuclear Regulatory-Commission (NRC) staff discussed the questions in the RAI and agreed-that l WCNOC would provide a drart response to the-_ staff and- submit- a formal response- after further discussion with the staff. Following a telephone conservation on February 21 -.1992 . between Mr. W. D. Reckley, Project Manager, NRC and Mr. S. G. Wideman,. WCNOC, a draft response to the RAI-was provided by WCNOC. On April:28, 1992, Mr. Reckley indicated in a telephone conversation- with _Mr. Wideman that WCNOC should submit _the formal respense i to the RAI. The_ Attachment provides -WCNOC's_ formal response- to the j questions provided in-Reference 1.
,6 wm 9206040260 920529 f g
>o 80,4ii , eornnoton. xs e8839, enone: <3is> 384 883 gDa ADoCKo500@2 .
An Equal opportunity Empiover M F HC VU
ET 92-0103 Page 2 of 2 If you have any questions concerning this matter, please contact me or Mr. S. G. Wideman of my staff.
Very truly yours,
. f, u Forrest T. Rhodes Vice President Engineering & Technical Services ,
FTR/jra Attachment cc: A. T. Howell (NRC), w/a R. D. Martin (NRC), w/a G. A. Pick (N?^', w/a W. D. Reckley taRC), w/a p
Attachment to ET 92-0103 Page 1 of 16 Response to Request for Additional Information for the Core Thermal Hydraulic Analysis Methodology for the Wolf Creek venerating Station Wolf Creek Nuclear Operating Corporation (WCNOC) submitted " Core Thermal Hydraulic Analysis Methodology for the Wolf Creek Generating Station" Topical on August 21, 1990. Letter dated December 27, 1991, from the Nuclear Regulatory Commission (NRC) requested WCNOC provide additional information on the topical. At a meeting on January 28, 1992, WCNOC and the NRC staff discussed these questions with WCNOC agreeing to submit a drait response and to submit a formal response af ter further discussion with the staff. Following a telephone conversation on February 21, 1992, between Mr. W. D. Reckley, NRC Project Manager, and Mr. S. G.
Wideman, WCHOC, a draft response to the questions was providad to the NRC staff. During a telephone conversation on April 28, 1992, between Mr. W. D. Reckley and Mr. S. G. Wideman, Mr. Reckley indicated th .st WCNOC should transmit the formal response to the request for additional information. Provided below is the NRC requested information with WCNOC's respot?se immediately followin s.
Reauest la Provide justification for the assumption that the three-step radial power distribution and the modified Core statepoint-1 used in determining the core-wide protection limits are bounding. What evaluation will be performed to confirm these assumptions for a specific reload cycle?
Response
The analysis of a number of "real" power distributions and a bounding three-step distribution were used to demonstrate that the hot pin statistical design limit (SDL) provides a conservative means of demonstrating core-wide protection. The power distributions used for this evaluation were chosen to be representative _ of limiting actual power shapes but were not considered to be bou. ding. The three ster powe; distribution was chosen to be bounding'in terms of the number of pins with peaking factors at high enough levels to be . limiting. The evaluation of this power distribution demonstrates that, even with one-third of the core at the design peaking limit, the statistical design limit derived from the hot pin protection criteria is limiting.
The conservative nature of the core-wide protection results obtained ;
with the modified Core Statepoint-1 was assured by examination of each of the core states described in Table 4-13, Th-90-0025 for which the mean Departure from Nuclente Boiling Ration (DNBR) represents an allowable core state. For example, the mean DNBR at core state 5, Table 4-13 was approximately 1.2 while the corresponding SDL is 1.33. The use of unallowable core states for the determination of the hot pin SDL is acceptable since the relevant parameter is . the uncertainty on DNBR, not the absolute magnitude of DNBR. However, for core wide protection, core peaking distributions must be analyzed only for core states that feature a mean DNBR that is greater than or equal to the hot pin statistical design limit. Results of the analysis of these additional core states indicate that the modified core statepoint-1 yields a limiting core-wide SDL.
_- _ _ _ _ _ _ _ _ - \
Attachment to ET 92-0103
- Page-2 of 14 Inherent in the selection of the core state points for use in establishing both the hot-pin and core-wide SDL is the definition of the nddpoi *. or " nominal" condition state point. Small changes in nominal or dcmign values of the core state variables included in the determination of the statistical design limit are accounted for in the parameter ranges used to define the a points on the response surface i model (RSM) and by the large variations in core states used in !
g establishing the DNB protection criteria. Evaluations of the design values of the core state variables along with their associated uncertainty ranges are made by WCNOC during the reload safety evaluation for each cycle design. Should a change in a plant parameter be made which would significantly af fect the definition of the center point on the RSM, (i.e., power uprate, .large reduction in thermal design flow, large increase in design FAH), it would be recessary to perform an analysis to demonstrate that a core wide statistical design limit determined based upon a RSM referenced to some previous "nomintl" core state point remains valid or to develop a new RSM that' represents the new nddpoint values for the core state variable and subsequently, to establish a new statistical design limit.
Reauest 24 How will it be ensured that the power distribution assumsd in the 17-channel Model (of section-3) used to determine the DNBR safety limit lines are bounding for a specific reload cycle?
ResDonses s
The design power distribution described in the development of the base thermal-hydraulic model was derived from scrutiny of actual power distributions for che Wolf Creek Generating Station- (WCGS). The bounding nature of this distribution will be verified for each reload cycle through three mechanisms. First, the maximum allowable peaking ,
(MAP) limits are utilized in a physics maneuvering analysis in which predicted three dimensional peaking distributions throughout cycle life are compared to the peaking limits defined by the MAP's. Core power distributions are influsnced by ' cycle specific design parameters and ,
operational conoitions. During cycle operation the core power distribution is dependent primarily upon f uel depletion, power level, control rod. position, and xenon distribution. In the maneuvering analysis, peaking distributions resulting from power operation over the entire expected range of these parameters are calculated. The peaking distributions which challenge the maximum allowable peaE.ng limita are i
identified and core thermal-hydraulic check cases using the predicted power distributions are analyzed to confirm that positive margin to peaking limits is maintained. Secondly, a check is made to guarantee that the peak predicted FAH is less than the design limit (i.e., 1.55) assumed in the design distribution. Finally, a check is made on the pin-to-box toctors for predicted peaking distributions for each reload to insure that the 1.05 peak-to-average ' f rom the design distribution remains bounding.
1
Attachment to ET 92-0103 Page 3 of 14 Iteauest 3: How will cycle-to-cycle variations in fuel design and core loading be accounted for in the VIPRE-01 model?
Resoonse Typically, a reload design will feature fuel which is functionally identical to the fuel resident in the core. The design parameters for the fuel for a proposed reload are verified during the reload safety evaluation process. Should a fuel design change cccur, such as the addition of intermediate flow mixing grids, modifications to the base thermal-hydraulic model would be made which reflect the desi1n change.
Appropriate sensitivity studies would be made to demonstrate that the model yielded a conservative thermal-hydraulic environment for use in the core thermal-hydraulac design. It should be noted that significant model enanges would require that analyses be perfonned to either confirm the statistic design limit or which establish a new SDL. This is because the RSM is optimized to reflect the DNB response for a particular VIPRE-01 model. When this model is changed, the DNB response of the VIPRE-01 model would be expected to also change.
The effect of changes in core loading primarily affects peaking distributions in the core. These effects are considering in the physics maneuvering analysis (see response to request 1).
Recue'11_1: How are assembly rod-wise power distributions that are not octant symmetric, due either to fuel design or global core power distribution, accounted for in the VIPRE-01 model?
Resoonse:
The effects of assembly rod-wise power distributions which are not octant symmetric are accounted for in the plant physics maneuvering analysis. Penalty factors are applied on calculated power distributions to account for core tilt. The augmented power peaking is then compared to the maximum allowable peaking limits established _in the core thermal-hydraulic analysis. It should be noted that. core loading patterns for-WCCS are designed such that 1/8 core symmetry will be maintained.
t Peauest 5: Does the VIPRE-01 axial representation assume that the minimum DNBR (MDNBR) occurs between the 68 and 130 inch elevations a ad, if so, how are situations where the MDNBR occurs outside this region treated?
l l Resoonse The axial noding schema examined in the development of the base thermal-I hydraulic does assume that the minimum departure from nucleate boiling ratio will occur within the span from 68 to 130 inches. The combination i
of core thermal conditions and power distributions typical of most core 1 thermal-hydraulfc evaluations will place the point of .MDNBR within this axial range. However, certain trans!ents may feature unusual thermal-conditions or power distributions which could cause the elevation _ of MDNBR to fall outside the range in which the axial node size in base thermal-hydraulic model is small. Example of transients of this nature-l
s Attachment to ET 52-0103 Page 4 of 14
.would include the er.eam line break event and the rod withdrawal from suberitical events. For core thermal-hydraulic analysis of events in which the point of MLNBR is determined to - f all outside the 68 to 130 inch elevation, modifecations to the axial noding in base thernal-hydraulic model will le made to insure accurate resolution of the solution. These modifications will be made using the same guidelines used to develop the base thermal-hydraulic model as described in the l WCNOC submitta'i.
Recuest 6: Provide the b.vuis for the uncertainties and the assumed (normal and uniform) distritutions for the variables given in Table 4-12.
ResDonse:
The uncertainties propagated through the RSM to establish the statistical design limit were selected to reflect or bound plant specific uncertainties for WCGS. The distributions assumed for these uncertainties were either normal or uniform. When a clear basis for use of the normal distribution was unctrtain, the more conservative uniform distribution was used.
The 2% of rated thermal power uncettainty applied - to _ the core power accounts for the error associated with a normalization of the reactor primary side power indication to a iecondary side calorimetric _ heat balance and the error can be she,wn to . onform to a normal distribution (Reference 6.1)
Reactor coolant flow is measured by elbow meters located in the cold leg of each primary system loop. 7 he - elbow meters are normalized to a precision calorimetric measurement of flov at the beginning of each-cycle. The reactor coolant flowuncertainty is determined from the elbow mrater dif ferential pressur-e measurement uncertainty,- the precision calorimetric measurement unce'. tainty , and the Reactor Coolant System (RCS) pressure and temperatt e measurement uncertainty. Evaluatione of.
the uncerta.inties asse.- i.4ted with the measurement of reactor coolant flow have conservatively established 2.5% as the flow uncertainty for WCGs. 'A uniform distribution was used for f.he propagation _of the reactor coolant flow uncertainty to insure conservative results.
The bypass flow uncertainty results .from uncertainties in _the calcula. tion of_ mass flow rate through each bypass _ flow path. Five flow paths isxist for coolant flow to bypass ti e core. The outlet nozzle leakage _ describes the flow that goes directly from.the_ inlet nozzle in to the outlet nozzle through the gap between .the vescsel and the outlet nozzle. The baffle-barrel region bypass flew is an . intentional bypass '
flow path which provides for. cooling of the baffle, formers and , core barrel.. The upper head cooling spray provide.s a third path for coolant flow to bypass the core. The upper head cooling spray anjects flow into the upper head to reduce relative motion of the upper head at vencul fitup and to minimize wear on the spring seal. Thimble cooling bypass flow describes the coolant flow that bypasses the ccre to provide cooling for core cemponent rods. The final bypass: flov. path is - the cavity flow. Cavity flow is that portion of the total . coolant flow a _ _ _ _ _ _ _ _ _ _ _ _ __ _ _ _ _ _ J
Attachment to ET 92-0103
. Page 5 of 14'
.which passes between the outer fuel assemblies and the baffle (cav.ty gap). The uncertainty on bypass flow arises from the uncertainties in calculating the flow through each of these paths (i.e., uncertainty in form loss coefficients, etc.). Plant specific analyses of the bypass flow uncertainty for WCGS indicate that the 1.5% used to establish the statistical design limit will conservatively bound the actual bypass flow uncertainty. A uniform distribution was used for the propagation of the bypass flow uncertainty to insure conservative results.
Pressurizer pressure is controlled by comparison of the measured vapor space pressure and a reference value. The uncertainty in the RCS pressure includes allowances for the pressure transmitters, process racks, and controller and pressure overshoot /undershoot due to the interaction and thermal inertia of the heaters and spray. A plant specific evaluation of the uncertainties associated with the kCS pressure control indicates that. the 30 psi used to establish the statistical design limit conservatively bounds the actual uncertainty for WCas (Reference 6.1). A uniform probability distribution was applied to the pressure uncertainty for conservatism.
The average coolant temperature is controlled by a u m-that compares the auctioneered high T avg from the four coolant l_ f., with a reference value. The reference is derived from the turbine . impulse chamber pressure. Allowances for uncertainties are made for the Resistant Temperature Detectors (RTDs), prot ass racks,- and terbine pressure transmitter. A plant spec!fic-analysis for WCGS has established 4.85 "F as the uncertainty on the - average temperatur,. control for the RCS.
Since the uncertainty on coolant temperature treated in the determination of the statistical design limit only accounted for a 4.0
- F uncertainty, a 0.85 *F penalty will be compounded in all core thermal-hydraulic analyses. Note this - discussion applies only to the errors associated with the T avg allowance for controller deadband and ;
measurement allowance and does not include the margin alloccion for steam generator tube fouling. An additional _ l.65 "F penalty on RCS coolant temperature must also be compounded in all core thermal-hydraulic analyses to account for the steam generator tube fouling margin. The' uncertainty on Tnvg was applied assuming a uniform probability distribution for conservatiam.
The calculational and measurement uncertainties on radial peaking, axial peaking and elevation of the axial peak (i.e., R, A, & Z) were selected
- t. bound the actual uncertainties f or WCGS . A total uncertainty on-
. total peak of 7.3% (5.0% applied to . the radial component and 2.5%- 3 applied to the axial component) conservatively bounds- the actual uncertainties for WCGS (Reference 6.2). These uncertainty parameters have been shown to conform to a normal distribution probability function f.
(Reference 6.3) The 3.0% hot channel factor uncertainty . and .1.$% }
initial bundle spacing penalty are l plant specific - values applicable to the 17-x 17 fuel assembly geometry curreatly in.used at WCGS (Reference 6.4). The hot channel factors have been shown to be well represented by a normal probability ' distribution function (Reference 6.5), while the penalty due to -initial bundle spacing was represented with a unif ortn probability- distribution- function for conservatisn. The 8 inch uncertainty on the location of the axial. peak is a consequence of the j
Attachment to ET 92-0103 Page 6 of 14 nodal codes used to perform neutronic calculations for WCGS. These codes typically use a node size of 4 inches. Thus an eight inch uncertainty was applied to anservatively bound the nodal uncertainty.
This uncertainty has been shown to be represented by a normal probability distribution function (Reference 6.3).
The uncertainty applied .or the WRB-1 c tical heat flux correlation was derived from the statistical performance of the correlation in the VIPRE-01 code. The calculation of the level of uncertainty for the WRD-1 correlation is given in TR-90-0025. This uncertainty has been shown to be represented by a normal probability distribution function (see response to item #14). The uncertainty on the VIPRE-01 code could be considered to be included in the WRB-1 correlation uncertainty and is therefore accounted for twice in the current application. However, a 5.0% uncertainty was retained for the VIPRE-01 code as an additional conservatisn. using a normal probability distribution. The RSM to VIPRE-01 fit uncertainty was derived from the statistical performance of the optimized RSM. Calculation of the level of uncertainty for
.. corporation in the statistical design limit is given in TR-90-0025.
This error has been shown to be represented by a normal probability distribution functio-Egferences:
, 6.1 Carroll, M. E., "WONOC Nuclear Safety Analysis Setpoint M e thodc .. >gy for the Reactor Protection System", TR-89-001, Wolf Cresa Nucle.tr Operating Corporation, June 1989.
6.2 Jackson, E. W, et al., " Qualification of Steady State Core Physics Methodology for Wolf Creek Design and Analysis", TR-91-0018, Wolf Creek Nuclear Operating Corporation, December 1991.
6.3 Hassan, H. A., et al., " Power Peaking Nuclear Reliability Factor",
BAW-10119-P-A, Babcock & Wilcox, Lynchburg, Virginia, February 1979.
6.4 Olson, C. A., " Hot Channal Factors", Weetinghouse Electric Corporations, August 1991.
6.5 Chelemer, H, boman, L. H, and Sharp, D. R., " Improved Thennal Design Proc edu re , NCAP-9567, Westinnhouse Electric Corporation,
( July 1975.
E.ecuest 7: How do the Wolf Creek Nuclear Operating Corporation (WCNOC) statistical core design 0;CM sud response surface motbodology differ from the methods described in .% f e rence-16 of the topical report?
Explain any differences.
ResDonse:
The WCNOC statistical core design and RSM are based entirely upon the i methods described in BAW-10170-P-A. The f orn of the RSM equation used by WCNOC is identical to that given in the reference. This equation, which contains linear, quadratic, and cross product terms, is comprised
Attachment to E'" 92-0103 Page 7 of 14 ,
of a maximum of 36 coefficients and involves each of the core state variables. Identical criteria, based upon a central composite design methodology, are utilized in the selection of points for the optimization of the RSM coefficients. The propagation of the uncertainties for the hot pin prutection is accompliched usino the F,& W developed SDLHOT code as wacribed in BAW-10170-P-A. The propsjation of uncertainties for core wide protection 'an acccuplished using a modified ,
version of the FDLCORE code.
As was done in BAW-10170-P-A, the WCNOC base thermal-hydraulic model f1.u , VIPRE-01 model) utilized an intrabundle peaking distribution for the hot bundle ud then a lumped representation of the remainder of the core in a single pin. For the core wide analysis, WCNOC also applied the hot bundle intrabundle peaking distribution to the predicted radial '
peak for each fuel assembly in the core. However, as listed in the reference, the B&W core protection code, SDLCORE, does not conservatively account for the highly peaked internal peaking distributions typically found in assemblies with lower powers.
Specifically, as it is listed in the reference, the B&W core protection code establishes estimates of pin powers by multiplying pin radial local factors from the design intrabundle peaking distribution with each assembly average power. In the B&W implementation, this results in a peak pin power of 1.05 times the assembly average power. However, as stated in BAW-10170-P-A, the " colder" assemblies typically have internal peaking distributions which are more highly peaked than the design distribution. To conservatively predict the numbcr of pins that could experience DNB in the core wide protection analysis, a Mdification was made to the SDLCORE core protection code which scales the pin-by-pin radial local peaking factors for the design distribution with the pin-to-box factor for each assembly power. Thus, in the WCNOC application of the core wide protection methodology to actual peaking distributions for WCGS, the number of pina that might contribution to the total number of ' pins in DNB is established by applying the design hot bundle ,
intrabundle peaking distr? " tion to each assembly in the core while also making the assumption that the highest power rod in each assembly serves as the reference for predicting the number of pins in DNB. 1 Recuest 8: In the selection of the random variables from the normal and ;
uniform dirtributions, are values greater than the 95 percent points o selected? If not, how is this simplification accounted for? q 821D2n1R ]
1 The normal and uniform distribution models which were used in the Monte !
Carlo propagation of uncertainties. on the core state variables are identical to the models used in BAW-10170-P-A. . These models, based upon l discrete 11 point model generators, have been shown to exhibit excellent-attributes for normality and uniformity. Ah compered to the ANSI standard on normality (Reference 8.1), the sample distributions obtained from the normal distribution model conform to standards of normality even at the most restrict.ive level (i.e., 20%). Examination of Table 3-3 in BAW-10170-P-A indicates that the 11 point distribution models will indeed result in the selection of pointe greater than the 95% level, i
. .-. . - , . ,~.
Attachment to YT 92-0103 Page 8 of 14 Referene n:
9.1 American National St.tndard, " Assessment of the Assumption of Normality (Employing Individual Observed Values)", AHS 15.lb-1974, Ameri National Standards Institute, 1994.
Reauest 9 r..,<ide justification for the use of the K=1.724 95/95 upper toleranen factor for the RSM fitting error. What error is introduced by this assustption?
Resdonse The use of 1.724 for the Owen's one-rided tolerance factor was based upon the number of experimental points in the WRB-1 correlation database. This was an added conservatism in the WCNOC application of the B&W Statistical Core Design methodology in that in the reference application, an Owen's one-rided tolerance factor of 1.645, w respondirg to an infinite pcpulation, was used to establish the error on the thermal-hydraulic code to RSH fit.
Argument can be made for the use of a t71rt 1.ce f acto corresponding to the number of mints used to darive vxN to thermal-hydraulic code fit. This would imply the use of a K factor equal to 1.982 for the 73 points in the current application. This would yield a RSM to VIPRE-01 error of 4.054. Since a 4% error was used to define the error frra the RSM to VIPRE-01 fit in the WCNOC application, use of the larger K factor would have negligible effect on the hot-pin statistical design limit.
11nouest 10 What evaluation will be performed to ensure tnat the statepoints used in determining the hot-pin protection statistical design linit are bounding for a specific., reload cycle?
Resoonse The state points used in the detarmination of the hot-pin statistical design limit were chosen to cover the anticipated range of operating limits f or WCr.S. Both high and low pressure m actor protectjin system limit state points were analyzed at design overpower condition *.
Nominal condition state points were also examined with both synnetric and outlet peaked axial power distributions, additional cases were also examined in which the core power was incrossed beyond the operating range and inlet temperatures rere artificially increaseo to yield miniwum DNB ratios near the design limie. Finally, the miopoint, or nominal state point from the development of the RSM was analyzed at a low flow condition and at a high power ccndition.
T's purpose of the analysis of many different state points in esitablishing hot-pin protection was to maximize the coefficient of variation resulting from the error propagation. Results shown in Table 4-14 ef the topical indicate that the overall coefficient of variation shows little change compared to the large changes in core states examined. It is therefore concluded that reasonable maximization of the coefficient of variatien, and subsequent maximization of the hot-pin statistical design limit, was achieved.
Attschmeit to ET 92-0103 page 9 of 14 Inherent in the selection of the core state points for use in establishing the hot-pin SnL is the definition of the midpoint or
" nominal" condition state point, small changes in nominal or Gesign values of the core stato variables included in the determination of the statistical design limit are accounted for in the parameter ranges used to define the it points on the REM snd by the large variations in core st c es used la ents.blishing the hob pin SDL. Eva. nations of the design values of the core state variables along with their associated u' ertainty ranges is made by WCNOC during the reload safety evaluation for each cycle eiesign. Should a change in a plant parameter be made which would significatly affect the definition of the center point on the RSM, (i.e., pcVet uprate, large reduction in thernal design flow, large increane in design FAH), it would be necessary to perform an analysis to demonstrate that a hot-pin statistical design limit determined based upon e previous " nominal" core state point remains valic' or to develop a new RSM which represents the new midpoint values for the core state variable and subsequently, to establish a new hot-pin statistical design limit.
Reauest 11: Are the WCGS fuel designs to which the WRB-1 correlation $
will be applied included in the presently approved applications of WRB-1/THINC7 Resoonse WCNOC will apply the WRB-1 critic %l heat flux correlation only to the analysis of fuel designs approved by the conunis sion in the safety Evaluation Report of WCAP-8762. For WCGS these fuel types might include the 17 x 17 standard and 17 x 17 optimized fuel assemblies (OFA).
Recuest 12: Why is the correlation design liuit to ha used with WRB-1 in VIPRE-01 proprietary? The design limit value is given in WCAP-8567, as well as in the MR SER, without proprietary brackets, and therefore should not be considered proprietuy in this topical report.
Responses Proprietary brackets will be removed from both the design limit and standard deviation in the WCNOC topical report.
Reauest 13 Why does the number of dM,a points and test series given in WCAP-8762 differ from the number given in Table 2-47 Please justify the use of the smaller number.
Response
In response to the Commission's request for additional information during the review of WRB-1 topical report, Westinghouse submitted Supplenient 1 to WCAP-8762. This supplement documented additional critical heat flux test data obtained by Westinghouse at Columbia's Heat Transfer Research Facility. This new data was obtained to extend the WRB-1 correlation to 14x14 OFA fuel assembly designs. As stated in the supplement, "DNB testing of the 14x14 OFA typical cell geometry has l
Attachment to ET 92-0103 Page 10 of 14 shovn that the WRB-1 correlation correctly accounts for the geonatry changes from the reference design, using the same performance factor as used for the 0.422 inch rod evaluations. The 14x14 ora data can be added to the WRB-1 R-grid database without changing the DNDR limit of 1.17." Table 3 of this supplement summarizes the WRD-1 database after the addition of the new 14x14 test data. As indicated, the new database consists of e total of 1108 test runs with a mean M/P ratio of 1.0079 and a sample standard deviation of 0.0859. This is the database used in the qualification of the WRB-1 critical heat flux correlation for use with the VIPR2-01 code.
Reauest 14e What tests have been perfomed to enoure that the M/P date is normal?
Basoonse The D' test, based upon the development by D'Agostino (Reference 14.1),
was used to ensure that the distribution of the measured to predicted critical heat flux data would approximate a normal distribution.
Referring to ANSI N15.15-1974 (Reference 14.2), the critical values for a population of 1108 members at the 5% level of significance are 10344 and 10454. The D' value calculated for the M/P data obtained from the quald fication of WRB-1 in the VIPRE-0) code was 10377, indicating that assumption of nomality may be accepted.
Refereggt 14.1 D'Agostino, R. B. "An Omnibus Test of Normality for Moderate and Large size samples", Biometrika, Volume 58, 1071, pp. 341-348.
Reauest 11: Is the procedure used to determine the steam generator safety valve (sosv) line (Equation 5-2) the same as is presently used?
Response
The formation of the equation defining the steam generator safety value line given in the Core Thermal-Hydraulic Analysis Methodology topical report is the same as is presently used to establish the core thermal limit line for WCGS (Reference 15.1). The steam generator safety valve line is an integral part of the definition of the core thermal limit lines due to the physical limit on reactor power and temperature placed on the plant due to the action of the steam generator safety valves.
The temperature drop from the steam generator primary to secondary is proportional to the power transferred. Since the maximum secondary side, temperature is constant at the saturation temperature corresponding to the lift pressure of the steam generator safety valves, the primary temperature cannot rise above this saturation temperature plus the temperature drop across the generator tubes. Therefore, the steam generator safety valve line define- one of the boundaries ou core power and temperatv ~ % the core thermal limits, i
1 i
- - . _ . - . - . . - . - - . - . - . . - _ ~ - - - . - . - ~ _ - - - . ~ . _ - . - -.- -
Attachment to E? 92 0105 Page 11 of 14 P.ef erences 15.1 Ellenbe ger, S. L., et al, " Design Bases for the Thermal Overpower AT and Thermal overte:tperature AT Trip runctions", WCAP-8745-P-A, Westinghouse Electric Corporation, September, 1986.
Recuest 16 Are the WCDC procedures for determining the MAP curves the same as the Babcock and Wilcox ruel company methods? If not, discuss any differences.
Responsen WCNOC utilizes the same procedures as BWrc to produce the Maximum Allowable Peaking curves for use in the plant maneuvering analysis.
Recuest 17: Is the part-power multiplier used below 75 percent power?
If so, provide the basis.
Resoonse WCNOc has performed analyses which indicate increasing DNB margin at powers below 75% as compared to the part power multiplier line as shown in Figure 6-14 of TR-90-0025. Below 75% power, limitations on the hot channel exit quality imposea by the range of applicability of the critical heat flux correlation place limits on the total peaking allow however, studies show that sufficient margin exists at all powers below 100% to support the use of a part power multiplier coefficient of 0.3.
Recuest 18e Do the three points on the coro safety limit lines used to determine the MAP curves provide the most limiting MAPa? For example, since the low pressure MAPS are mqre restrictive, why wasn't the SGSV limit line MAP calculated on the lot < pressure DNDR limit line?
Resoonses The three points on the core safety limit lines used to determine the MAP curves have been shown to yield .a limiting set of Maximum Allowable Peaking limits for WCGS. Examination of rigure 5-2, TR-90-0025 provides insight for not calculating a set of MAP limits at the intersection of the steam generator - saf ety val"e limit line and the low pressure DNB limi line. As shown, this intersection occurs well down on the 1860 psis. limit line in the region where the plant is vessel exit boiling-linu .ad . The cure AT at this intersection corresponds _to a core power significantly less than 100% of rated thermal power. Since it has been shown that increased DNB rnargin exists with respect to peaking at powers below 100%, calculation of MAP limits at these conditions trould be less restrictive that the MAP limits computed at the design overpower conditions.
Reauest 19: Why are the curves in rigures 6-13 and 6-15 different?
Attachment to ET 02-0103 Page 12 of 14 Resoonses The MAP limits shown in Figure 6-13, TR-90-0025 were adjusted, in the more restrictive dieaction, as a result of the analysis to establish the part power multiplier. As shown in Figure 19-1 below, use of a part power multiplier constant of 0,3 would reLJ1t in slightly non-conservative results when the ratio of the power allowable peaks is compared to the unadjusted safety limit MAP's. To account for this, the limiting safety limit MAP's were reduced by 1.6%. As shown in Figure 6-14, TR-90-0025, this insures that the peaking restrictions imposed by the safety limit MAP's are conservative at all powers. Figure 6-15, TR-90-0025 is simply the maximum allowable peaking limits shown in Figure 6-13, TR-90-0025 reduced by 1.6%.
WCNOC Response to RAI for TR-90-0025 Figure 19-1 1.18 -
- 1.17 -
] 1.16 - . 1.1 Axiol J 1.15 - + 1.5 Axiol 1 1 14 -
5a 0 1.9 Axiol 1 13 - A Multiplier Line
.9 1.12 -
a 1.11 -
0 1.1 -
1.09 -
$ 1.08 - '
E 1.07 -
1.06 -
1.05 -
o 1.04 -
.9 1.03 -
1o $ NN "
1 0.74 i i 0.78 i i 0.82 i ,
0.86 i i 0.9 i , -i m. i .
0.94 0.98 Fraction ci Nominal Power Recuest 20: The Chen heat transfer correlation does not result in the j highest fuel and clad temperatures in Table 3-59. hiow will conservative maximum fuel temperatures be calculated in specific transients?
Resoonset The chen heat transfer correlation results in the highest fuel and clad temperatures when compared to the cases which do not enter a post-chf heat transfer regime. comparison of the heat transfer coefficients for i
Attachment to ET 92-0103 Paga 13 of 14 case 1, Table 3-59, TR-90-0025, ir41 cates that the Chen correlation does in fact yield the omallest heat transfer coefficient, 28% leso than the Than correlation and 32% less than Thom plus single phase. As a result, ,
the fuel and cladding temperatures are highest for the Chen correlation for this case.
As would be expected, the Chen correlation does not yield the highest fuel and clad temperatures when compared to corralation sets that enter a post-cht heat transfer regime. spee.Lfically, examination of the results from the 120% power case (i.e., case 2. Table 3-59, TR-95 0025) indicates that while the Chen correl ation yields the highest fuel and clad temperature for the correlation sets that remain in the saturated boiling heat transfer regime, case HEATI enters the post-cht heat transfer regime and the resulting heat transfer coef ficient, obtained with the G5.7 correlation, is significantly less than the Chen coefficient. The reason for the switch to post-cht heat transfer in case HEAT 1 is the use of the Electrical Power Research Institute (EPRI) critical heat flux correlat'.on to define the peak of the boiling curve (see table 3 ~8, TR-90-0025). As noted, the heat transfer mode was switched to the post-cht regi.me when the HDNBk calculated reach 1.50.
Since the WRB-1 correlation get. orally yields higher DNB ratios than the EPRI correlation, cases HEAT 2 through HEAT 6 did not enter the poet-cht heat transfer regime.
The Safety Evaluation Report issued for the VIPRE-01 code limits use of the code to heat transfer modes up to the point of critical heat flux.
Thus, WCHOC will not attempt to predict fuel and cladding temperature when conditions result in heat transfer modes in the transition or film boiling regions.
Finally, the choice of an appropriate heat transfer correlation set in VIPRE-01 must be selected to yield conservative results for the parameter of interest. For departure from nucleate boiling ratio evaluations, heat transfer in the single phase region will ha defined by the EPRI correlation, the subcooled and saturated nue . ste boiling regions will be characterized by the Thom plus single phase correlations, and the peak of the boiling curved will be definca with the WRB-1 critical heat flux correlation. For fuel and cladding temperature evaluations, heat transfer in the single phase forced convection regime will again be characterized by the EPRI correlation while the subcooled and naturated nucleate boiling regime wili be defined by the Chen correlation. Again, the point af critical heat flux will be determined using the Westinghouse WRB-1 correlation. No attempt will be made to use the VIPRE-01 code in any analysis in which the heat transfer mode is predicted to enter the transition or film boiling regions of the boiling curve.
Recuest 213 Certain combinations of fluid correlations have not been included in the section-3.3.4 comparisons. How will these cases compare to f ae base-case ther nal-hydraulic stodel?
u Attachment to ET 92-0103
+
Page 14 of 14 Resoonses The selections of correlation combinationo for use in the flow correlation sensitivity study were made such that only correlations with consistent and complementary bases were examined. The objective of the fluid correlation study was to demonstrate that the calculation of a minimum departure from nucleate boiling ratio is relatively insensitive to the fluid :orrelation set used; not to_ examine every possible combination and permutation of the fluid correlations available in VIPRE-01. This was accomplished since the variation in calculated minimum DNB ratios was less than 3% for all correlation combinations with the exception- of the case which utilized the Beattie two-phase friction multiplier correlation. The reasons tor the variation observed with the Beattie correlation are well documented in the topical report.
The fluid sensitivity study found that the EPRI subcooled void, EPRI bulk void, and EPRI two-phase multiplier correlations are aufficient to yield accurate predictions of; fluid conditions. Further, __ - t he qualification of the WRB-1 critical heat flux correlation using this combination of fluid correlations- and the generation of acceptable statistical results as compared to the experin. ental data in the WRB-1 database, serves to confirm the selection of fluid correlations for use in tne base thermal-hydraulic model.
I
__..m_