ML19344F418
| ML19344F418 | |
| Person / Time | |
|---|---|
| Site: | Calvert Cliffs  |
| Issue date: | 09/10/1980 |
| From: | Lundvall A BALTIMORE GAS & ELECTRIC CO. |
| To: | Clark R Office of Nuclear Reactor Regulation |
| References | |
| NUDOCS 8009150250 | |
| Download: ML19344F418 (17) | |
Text
{{#Wiki_filter:.-.c-- (-
- A.
B ALTIMORE G AS AND ELECTRIC' COMPANY P.O. B O X 14 7 5 B ALTIM O R E. M A R Y L AN D 2120 3 Antwum E. LUNDt ALL,JR. v.cc pas s.oc, s-. September 10, 1980 Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission l Washington, DC 20555 l l Attn: Mr. R. A. Clark, Chief Operating Reactors Branch #3 Division of Licensing
SUBJECT:
Calvert Cliffs Nuclear Power Plant Units No.1 and 2, Docket Nos. 50-317 & 50-318 Margin Improvement Topical Reports Reference (a) : A. E. Lundvall to R. A. Clark letter, dated June 26, 1980, same subject.
Dear Mr. Conner:
Reference (a) submitted for your review and approval several Combustion Engineering (CE) topical reports which we propose to reference in future . cycles of both Calvert Cliffs Units. At a meeting with NRC staff in -Bethesda on July 17, 1980, CE made a summary presentation of those topical reports. NRC Staff posed several questions and also suggested that CE review the SER relative to the Westinghouse methodology for statistically combining uncertainties in state and system parameters. I is CE's response to those questions and Attachment 2 is a comparison of the CE and Westinghouse methodologies. These documents should assist the staff in completing their review of the Statistical Combination of Uncertainties Topicals and the FIESTA Topical. BALTIMORE GAS AND EL TRIC COMPANY By: /J. edd . E. Lundy 'll, Jr. Attachments (1) and (2) - 40 Copies dc: J. A. Biddison, Esq. w/o Attachments G. F. Trowbridge, Esq. w/o Attachments Messrs: E. L. Conner, Jr, NRC P. W. Kruse, CE - 8009150350 /
F ATTACHMENT 1 RESPONSE TO QUESTIONS POSED BY NRC STAFF ON MARGIN IMPROVEMENT PROGRAM BG&E/CE/NRC MEETING BETHESDA, MD July 17, 1980 i .i
p FIESTA-1. QUESTION -Is the'use of FIESTA-calculated scram data consistent with the type of DBE analysis investigathd (e.g. is the effect of loss of flow on the scram curves taken into account in the LOF analysis)? ANSWER - Yes, the use of the FIESTA calculated scram reactivities is consistent with the DBE analysis.. Reactivities are calculated as a function of the ASI (Axial Shape ' fex) of the axial power distribution prior to scram' rod motion. The reactivity insertion during rod motion is computed based on the space-time varying flux shape resulting from rod motion. Feedback need not be included because the scram only requires 3 seconds to complete. Feedback and flow effects are considered separately in the analysis of each DBE. -Changes in rod worth after the scram is completed are con-servatively included in the reactivity data used in the analysis of the DBE's. The effect of feedback on the time variation of the flux space during a reactor scram can be neglected. Rod motion has the dominant influence on the changing neutrcn flux shape during a . scram. The time constants of the coolant flow and the fuel and moderator thermal-hydraulic feedbacks are too large to have significant effect on the flux shape during a scram. For example,.in a Loss of Flow accident the minimum DNBR is achieved in less than 1.0 second following reactor scram rod motion. Significant flux shape redistributions due to flow and feedback effects are not possible for such a small time interval. At the time of minimum DNBR, the scram rods are inserted approximately 30%. Rod motion of this extent causes substantial changes in the axial power distribution, and this transient power distribution is calculated more accurately with space-time methods than with static methods.
J= y m 2. -QUESTION ~- Do we use' scram curves that are consistent with the initial axial shape at the beginning of the transient / accident? ANSWER - The FIESTA generated scram reactivities have been ^ calculated over a range of axial shape indices (ASI). (.6 to +.6) which span'the expected initial conditions from which a scram could be initiated. However, ASI is not a unique measure of either power shape or scram reactivity. For example, a cosine B0C shape or a saddle BOC shape may have the same ASI but they will have different scram reactivity characteristics. In order to cover all expected power shapes in future'evcles, -the scram reactivities were calculated for a range of power shapes _ including the most cosined and the most saddled shaped expected-in reload cycles. A lower bound envelope of reactivity . insertion.versus initial ASI was developed to cover all expected shapes for future cycles. Each curve was developed on the basis of.50 to 100 different burnup-dependent power shapes in the ASI range of .6 to +.6. A separate curve was generated for each 5 percent rod insertion. Examples of these curves are illustrated -in Figures 5.3 through 5.6 of Reference 1. The applicability of this generic scram reactivity data to the reload cycle is evaluated for each cycle. REFERENCES 1. " FIESTA: A one dimensional., two group space-time kinetics code for calculating PWR scram reactivities", U. Decher, CEN-122(B1 November,1979. e - ~,, v
E r '.. 3. . QUESTION NRC expressed disbelief that delayed neutrons could have as large an effect on scram curves as FIESTA calculated. ANSWER - The difference between the space-time and the static calculation of reactivity is related directly to the space-time method accounts for the delayed neutrons which remain distributed according-to the initial power shape. Since static methods do not accoudt for these delayed neutr_ons, the flux shape change during a scram is overpredicted and hence the reactivity insertion .is underpredicted. The methods agree whenever the change in rod insertion has little effect on the flux shape. This is the case for initial rod motion (up to 5% rod' insertion) or final rod motion (from 95% to 100% rod insertion), or for flux shapes that are not affected by rod motion (extreme top or bottom peaked shapes). At intermediate rod inser-tions for symmetric flux shapes, the methods yield different results because the static method overpredicts the flux shape change caused by the rod mation. The results of a specific static and space-time comnarison are shown in Figure 1. This figure shows the overprediction in the flux shape change derived from static methods which ignore the effects of delayed neutrons. It shows that the delayed neutron source in the rodded portion of the core results in a higher frection of neutrons in the region and ~ ence a larger reactivity insertion than predictec by critical methods. The case described here examines the insertion of a rod bank worth 6% ap into an initially symmetric shape which has about half (47%) the neutron population in the top half of the core. If the flux shape were_ unchanged as a result of the rod motion, the inserted reactivity would be about 3% ap. However, the e#fect of rod motion can substantially alter the flux shape. Static methods predict that at 50% rod insertion, the neutron population in the top half of the-core.is decreased to 5.9% of the total population and the reactivity inserted is 0.352% ap. Space-time calculations properly account for delayed neutron effects and predict a decrease in the neutron population to 13.9%.of the total population with a corres-ponding reactivity insertion of 0.798% ap. The primary effect of either method is a loss in rod worth caused by the shift of the flux shape away from the rods as the. rods are 'being inserted. When either method is compared to the linear insertion of reactivity from the example above, only 11.7% of that reactivity is inserted with the static method, and only 26.6% of that reactivity is inserted with the space-tims method. Therefore, although a.;ubstantial increase in reactivity. insertion can be obtained for some axial flux shapes, the actual magnitude of the increase is small. j j u
. h .h h h* ~ - 7 __ _ t.. _.q ___. .i.I l j... L . I k... a,.. + g!..+ .. ! ~ l I
- g m --
q'- - - n -.---- i. --. m Figure'l. i r-- p r-- -- -- - -i-- - 4 r.- --- I-i 7 ~ d(,,d.t. p yc -l computed _by-space-time. methods at. a_scPam rod positioln .. t Comparison, of a static power : shape a'nd,a power shape - ', ' i ' ' of 50%. i l O
- q. g.~
~ . (; 6; 4p ull: insertion)@th)' I j!,t .._..._,-_.__7.. . ~. .... J)...
- l.
e t-- ..} .s. ..i~.. . j - i a-t. e, - _ _ _ p., x
- 3...
~ i .p.3-.- c ._.:.._ u u... g.. ___ . l'. 7... ; .p: q. e a { . l. - . p: .q_ l
- -.i
~i: ~. .i .[ q_
- .]i.
..+u .2 [ j l. .]..l : [j,a. . 2.5-. .. i y. ._.9 _ i 1 i _...._.l..... L_ . L_ _. p _. 3 .; p ..i.__...... _ _, 3_=,.. 6 i....- t . n. l .,..-... +.. _ ; :g g' .1 i J.-p_ J.:._{ l - ! - u,. :_ y. -_ i (.. ! ._.} / .a.. 22.0- _ gp.1 j.;.. .1 i .g. p 3 l:y g, ,~_._I.._._ . _ __. I... _.l - ~ I, -_ l + - 3 i + i -l f l 7 L l iSpace-tine transient ] i t . 3 _.-W power shape at 50% u .. p._.q _ 24_... _L , _ g... 4
- I.
. ! ' I scram rod insertion i l ~ "J L_... 1_._s_. I._ k__ j.
- .. L _ ;_.. j.... l.. _. I p; ;.,_
_._1__..;_,__ i I .J i ! + i _l.5-4 i I t i-4 4' g m Static power shape I -i e i .. _. t.. ... _q _.a 3 i Initial unrodded shape [- at 50% scram rod ~ - ~ ~ ~~- ~t l'. .3 t i . i nsertion .I. ..,.._...._..___._,..r_._..j I f I 1 i i ],Q .. M ... f......_... _ j...._. - j i t i
- 1 s
I l t i ' _J._ ..._.....a. ._.f______ i 5 I g _.._.p.. .._.t._...}_ , ___ u .. _ q._ _. .___u._
- 7.....
0.5-j i -l-l i l 4 [ I. ..t. 1 I I l i L- "- - - -- i a -[ t i -- 0 - i 3 i - --i i 80i -- - 90 100 0 10 20 30 40 50 60 70 Top Percent from the Top nf fho Arf.ive fore llo f inq
zSCU 1 _ UESTION - NRC has not given Westinghouse credit for statistical 1. Q treatment of.the LHR LCO-related uncerta~inties. This is because . Westinghouse included credit for a decay heat uncertainty in its analysis. Does C-E take credit for this type of: uncertainty in 'its analysis?. Is there anything -in Appendix K which i ould preclude the NRC's approval.of C-E's treatment of the LHR LC0? l ANSWER -- No,'C-E does not incorporate this uncertainty in LC0 . margin uncertainty calculations, The statistical combination of uncertainties for the LHR-LCO, as described in SCU, Part 3, includes the following uncertainties: 1. power measurement, 2. axial shape index: separability, shape annealing, and calibration, 3. peaking factors, and 4. equipment processing. Decay heat uncertainty is not included. Any assessment of decay heat uncertainty is performed in setting the LOCA limit and would be. independent of the SCU analysis. C-E's-proposed statistical combination of uncertainties has no effect on the'ECCS evaluation models. The loss of coolant accident is evaluated as in the past. SCU impacts only the monitoring of the LOCA LHR limit with excore detectors. REFERENCES 1. " Statistical Combination 0; Uncertainties Methodology, Part 3," CE Report'CEN-124-(B)-P, March, 1980-e O._- 1 i'
0 ~ 1 2. LQUESTION - Is it conservative to statistically accout for varying setpoint potentiometer settings in the electronic processing uncertainty? ANSWER - Yes, the electronic processing uncertainty values used in the SCU analysis are conservative with respect to plant conditions. The codes usr.d to estimate electronic processing uncertainties calculate a standard deviation in catput signal for a particular set of potentiometer settings and input signals. During the SCU analysis, it was decided that the results should depend as little as possible upon equipment calibration and settings during any particular cycle. Thus, the uncertainty in processed protection signals was calculated for 2000 randomly selected sets of plant conditions and potentiometer settings. The uncertainty value used represents a 95/95 probability confidence limit on signal standard deviation. Randomly chosen combinations of potentiometer settings .do not preserve the functional correlations usually set in the equipment. Signal uncertainties were found to be much greater when typical plant potentiometer settings were used. Since impact on margin was small, the conservative processing uncertainty was used to ensure that regardless of any equipment setting changes required during the cycle, the results of the SCU analysis would still be valid. 6-
[" f 3. QUESTION _ NRC asked that_a comparison between old methods and new methods be made'for the-limiting A00's. Part 3 of SCU already gives examples of new methods.for LOFA a,nd CEA drop. ANSWER - The methods used to arealyze the events -are essentially .the same as the FSAR methods or methods approved by NRC for reload -license submittals. Since the methods used are similar to previously approved methods, Appendix C of Reference 1 presents the. impact of statistically combining uncertainties on the selection of initial conditions used in_ transient analyses and on the magnitude of the
- variation -in margin degradation ' attributable to the uncertainties.
The result of the analysis performed in Appendix C is that the calculation of_R0PM is not affected by-the statistical combination of uncertainties. REFERENCES 1. " Statistical Combination of Uncertainties Methodology, Part 3", CE Report.CEN-124 (B)-P, March, 1980 i d i L;:
y' ~ [4. -~ QUESTION Show that' system parameter uncertainties are independent . ~ ~ of state parameter uncertainties'. Show that they can be evaluated separately.in a conservative manner. -ANSWER'- The effects of system parameter uncertainites on MDNBR -are not independent of state parameters. Section 3.1 of. Reference 1 established.the most adverse set of state parameters. This set of state parameters was used to generate the MDNBR response surface that'was used in the SIGMA code to combine system parameter probability ~ distribution: functions (pdf's). The most adverse set of state parameters was that set which maximized the sensitivity of DNBR to variations in system parameters. With this sensitivity maximized, the DNBR penalty incurred because of system parameter uncertainties-was also maximized. Since the maximum penalty in DNBR was calculated in the statistical combination of system parameters with the.most adverse set of state parameters, it is conservative to apply'the resultant MDNBR limit over the state parameter ranges of Table 3-1 in Reference 1. State parameter uncertainties are treated in 1 . References 2 and 3. .:e REFERENCES 1. " Statistical Combination of Uncertainties Methodology, Part 2: Combination of System Parameter Uncertainties in Thermal Margin Analyses for Calvert Cliffs Units 1 & 2, CE Report CEN-124-(B)-P, January,1980. -2. " Statistical Combination of Uncertainties Methodology, Part 1: Combination.of System Parameter Uncertainties in Thermal Margin Anal Calvert Cliffs Units 1 & 2", CE Report -CEN-124-(B)yses ft. -P, December, 1979. 3. " Statistical Combination of Uncertainties Methodology, Part 3: ' Combination of System Parameter Uncertainties in Thermal Margin Analyses for Calvert Cliffs' Units 1 & 2", CE Report CEN-124-(B)-P, March,1980. i O
f., b - 5. QUESTION -' Has CE statistically convoluted system parameter uncertainties with CHF correlation uncertainties? Westinghouse had not been allowed to convolute these uncertainties since the DNB correlation uncertainty was not indebendent of the uncertain-ties Westinghouse addressed. ANSWER - A review of. Reference 1 has shown that Westinghouse made no attempt to convolute-CHF correlation uncertainties with the uncertainties addressed in this reference, CE has combined system parameter uncertainties statistically with the CHF correlation uncertainty in Reference 2. The result was an increased MDNBR limit. If DNBR values calculated with a design model having nominal values of system parameters are greater than.or equal to the increased MDNBR limit, there is at least 95% probability with at least 95% confidence that the hot pin will not experience DNB. The system parameters uncertainties considered by C-E in Reference 2 were: - inlet flow distribution (3.3) fuel rod pitch (3.8) fuel clad 0.D. (3.7) 235 - -fuel rod U content (3.5) - fuel. enrichment and fuel pellet 0.D. (3.6) - fuel rod bow (3.9) .These uncertainties were combined statistically with the CE-1 CHF correlation uncertainty to obtain an increased MDNBR limit. The new DNBR limit is 1.23, 3.4% above the older limit of 1.19. REFERENCES 1. C. B. Brinkman, "WCAP-8567 SER", July 25, 1980 2. " Statistical Combination of Uncertainties Methodology, Part 2: Combination' of System Parameter Uncertainties in Thermal Margin Analyses for Calvert Cliffs Units 1 & 2, CE Report CEN-124-(B)-P, January,-1980.
~ b /CTACHMENT 2 Comparison-of SER for WCAP-8567 and C-E Statistical Combination of Uncertainties Topicals j. Introduction-L The ' Statistical Combination of Uncertainties (SCU) topicals, Parts 1 and l '3 (Reference la and ic) jitstify the methods used to statistically combine l-l the state parameter uncertainties for the Limiting Safety System Settings l (LSSSs).and Limiting Conditions for Operations ('COs), respectively. Part 2 of the SCU topical-(Reference Ib) justifies statistically combining system parameter uncertainties to obtain a new DIGR limit. Tne Westinghouse report WCAP-8567. (Reference 2) dealt mainly with the methods used by W to obtain a new DNBR limit. The Safety Evaluation Report (SER) (Reference 3) .for:the W report therefore emphasi::es the methods used to calculate the new W DNBR limit. Many of the aspects of the SER can be best compared in detail with the methods used in Reference lb. However, some comparison of staff positions and ecommendaticns outlined in the SER can be made with the methods documented in References la and Ic. This should facilitate the review of these topicals. - Discussion The W design bases, accepted by the HRC with some restrictions, are equivalent l l to the C-E treatment of state parameter uncertainties for determining LSSS's and LCO's. However, the methods used by the two vendors are different. C-E used Monte Carlo simulation of state parameter uncertainties to determine penalty factors on overpower. (or equivalently on DI3R). W used analytical i-methods to combine probability distribution functions to arrive at a penalty on DNBR. W used Monte Carlo simulation to verify the validity of the results [' obtained. In the SER, NRC staff suggested the use of Monte Carlo simulation - technique to statistically combine tncertainties. In combining the system parameter uncertainties, C-E used an Orthogonal L Central Co=posite (OCC) experiment design to generate the 'DUBR data base I for the response surface. The OCC design consists of a full factorial L design, augmented by additional extreme points. In the SER, NRC staff suggested.that the'use of factorial design in the evaluation of these uncertainties vould be more appropriate than using the functional relation-ship between DNBR and " input parameters (equation 1 of SER) derived by W. C-E statistically combined system pa'rameter uncertainties with the Critical . Heat Flux (CHF) : correlation uncertainties; W did not. C-E oelieves this to be a valid and justifiable approach because the uncertainty associated with the CHF correlation is independent of system parameters. Statistically - combining the CHF correlation uncertainties with system uncer a nt i ties e resulted in the increase of DUBR limit from 1.19 to 1.23. The use of. the new D13R limit in combination with the design model having nominal values of system parameters which yields minimum DNBR's equal to or greater than p- ' the new limit' assures that there -is; at-least 955 probability with at least - 95%. confidence that the hot' pin vill not experience DNB. ( e
wa v, In the SER, NRC staff i= posed a'k% additive penalty on DUER since E did not include THINC-IV code uncertainties nor quantify the offsetting code ~ .conservetisms. Similarly the staff imposed a 1% additive DUER penalty for the transport code uncertainties. The h% pedalty on DUER was imposed because E did not consider the follov r a incertainties: 1.. Uncertainty in pressure drop due to uncertainty in the two-phase friction factor correlation. 2. Error-in normalization of axial shape. U e of inlet flow boundary condition. 3. s 4. Crossflow resistance independent of void fraction. 5. Age effects on grids. 6. Effects of non-uniform clad surface roughness (crud effects). T. Use of 3% reduction in inlet flov in place of true flow and pressure drep calculations. C-E has -accounted for items 1, 2, 3, h, and 7 either inplicitly, explicitly-or by using bounding data. TORC predicted pressure drops were shown to be in excellent agreement with measured values in Reference h. The uncertainty in the pressure drop calculations 'is therefore small and will have a negligible effect on MDIGR. Since the uncertainty in the two phase friction factor is only a contributor to the uncertainty in the. pressure drop, it must also be small. As discussed in Reference h, DNBR is relatively insensitive to ite=s 2 and h. Items 3 and.7 are treated explicitly in Reference Ib. The C-E design' methods - explicitly account for'non-uniform inlet flow and exit pressure distributions and their associated uncertainties. The new limit was generated using a 3-pump operation inlet flow distribution with corresponding uncertainties. Thic is conservative for 4-pump operation analyses since the 3-pu=p inlet flow distri-bution and corresponding uncertainties are more, adverse than their h-pump counterparts. -In Reference Ib, it is demonstrated that the exit pressure -distribution and its uncertainty does not affect the DUER. 'The proposed:new CE-1 DNBR limit of 1.23 includes 'the 5% interim renalty - imposed by'the staff pending review of the CHF data from non-uniform axial heat flux experiments.' These. nata were submitted to the NRC in ' Reference 5 'and are under staff review. R ) I h T .A. 1 f r N r a vv -
.- The staff imposed the 15 DUER penalty since W did not cuantify the conserva-tists in the transient analyses methods. Appendix C of Refrence le quanti-fies the conservatisns in the transient analyses of limiting Anticipated Operational Occurrences. The Reference identifi'es at least 25 to 35 over-power units or equivalently 55 to 75 DNBR units conservatists in the transient analyses methods. Conclusion, The NRC staff approved the W topical with the restrictions given in Table 1. The applicability of these restrictions to the C-E topical is also given in this tcble. The comparison of the SER for UCAP-8567 and C-E's topicals on Statistical Combination of Uncertainties reveals the following salient points: 1. The W design bases accepted by I RC is equivalent to C-E treatrent of state parameter uncertainties. 2. C-E used Monte Carlo simulation technique to statistically ccchine state parareter uncertainties. 3 C-E used conservative representations of state parameter uncer-tainties vaich vere verified by cctparison to plant data. 4. C-E used an Orthogonal Central Co=posite experiment design to generate the DNER data base to statistically combine syster parameter uncertainties. 5 C-E explicitly treated non-uniform inlet flow distribution, exit pressure distribution and their associated uncertainties. 6. C-E quantified the conservatists in its transient analysis methods. 7 C-E included the NRC-icposed 55 interim CHF correlation penalty in generating the new DNER lindt. 8. C-E treattent of rod bov penalty is consistent with HEC recor-mended methods (References 6 and 7). 9 C-E analyses are performed on a plant specific basis. e e
TABLE 1 ..a o NRC RESTRICTIONS ON WCAP-8567 AND THEIR APPLICABILITY TO C-E TOPICALS ON STATISTICAL COMBINATION OF UNCERTAINTIES ITEM NRC RESTRICTIONS ON WCAP-8567 APPLICABILITY OF RESTRICTIONS TO C-E SCU TOPICALS 1 Seasitivity factors used in WCAP-8567 are C-E realizes that the results documented in. acceptable if the W-3 DNB correlation is used and . References la, ib, and Ic are applicable only for specified parameter ranges are not exceeded. the CE-1 correlation and approved thermal. hydraulic design code. Any substantive changes in the 2. Changes in DNB correlation, THINC-IV correlations, correlation or the design code will. require a or use of p,arameters outside specified ranges re-evaluation of the sensitivity factors and the will require a re-evaluation of the sensitivity uncertainty allowance documented in'the topicals. factors and of the assumed relationship between DNB and the parameters. 3 If the sensitivity factors change, the use of an allowance and the validity of the uncertainty'ssumptions must be re-evaluated. linearity a '4 Variances and distributions ~ for input parameters This restriction is not applicable, since C-E's must be justified on a t ant by plant basis, analysis is performed on~a plant by plant basis.- il g pending generic approval. / 5 Rod bow cannot be treated as described in This restriction is not applicable. C-E treated WCAP-8567 rod bow as recommended by NRC in their June 12, 1978 letter. 6 Nominal initial conditions may be used only in As stated in~ Appendix C of Reference 1C, nominal the DNBR analyses which use the new method. initial conditions will be used only for transients Other analyses, such as overpressure calculations, with DNBR as a criterion. Other transients with require the appropriate conservative initial criteria different than DNBR will use appropriate conditio1 assumptions. conservative initial condition as' umptions. s 7 Nominal conditions considered for. use in C-E. analyses documented in References.la, lb and ic analyses should bound all permitted methods of bound all permitted methods of four pump plant plant operation. operation.
J TABLE 1 (CONTINUED) ITEM NRC RESTRICTIONS ON WCAP-8567 APPLICABILITY OF RESTRICTIONS TO C-E SCU TOPICALS 8 DNBR uncertainties of +4% for the THINC-IV code The code uncertainties imposed on W are not and +1% for transient analyses mUst be included applicable for C-E SCU topicals because: in DNBR analyses with the new method. i) C-E design methods included non-uniform inlet flow distribution and exit pressure distribution and their associated uncertainties. ii) 'C-E quantified the conse'rvatisms in the N transient analyses methods. iii) C-E included the NRC-imposed interim CHF correlation penalty in determining the new limit. 9 Part loop operation is beyond spe-ified C-E's reports are not intended to be applicable for pt ameter ranges, therefore, the imw DNBR part loop analyses. limits may not be applied to part loop ~ analyses. D e
. f t%' References i la. " Statistical Combination of Uncertainties 11ethodology; Part 1: 'C-E Calculated Local Power Density and Thermal "argin/ Low Pressure LSSS for Calvert-Cliffs Units I and II", CEN-124(B)-P, December,1979. lb. " Statistical Conbination of Uncertaintie's Methodology: Part 2: l Combination of System Parameter Uncertainties in Thermal Margin Analyses for. Calvert Cliffs Units I and II", CEN-124(B)-P, January,1980. ic. " Statistical Combination.of Uncertainties Methodology; Part 3: C-E Calculated Local Power Density and. Departure from Nucleate Boiling Limiting Conditions'for Operation for Calvert Cliffs Units I and II", -CEN-124(B)-P, March,1980. H. Chelemer, et 21., " Improved Thermal Design Procedure", Westinghouse 2.- Topical Report WCAP-2567 (Proprietary) and WCAP-8568 (non-Proprietary), July, 1975. 3. Safety Evaluation Report for WCAP-8567 4. " TORC Code: A Computer Code for Determining the Thermal Margin of a Reactor Core", CENPD-161-P, July, 1975. 5. "C-E Critical Heat Flux Correlation for C-E fuel Assemblies with Standard Spacer Grids, Part 2: Non-Uniform Axial Power Distribution", CENPD-207-P, June, 1976. -6. Letter from D. B. Vassello (NRC) to A. E. Scherer (C-E), June 12, 1978. 7.- " Fuel and Poison Rod Bowing, Supplement 3", CENPD-225-9, Supplement 3-P, June, 1979. - k S 9 = e =}}