ML18026A239
| ML18026A239 | |
| Person / Time | |
|---|---|
| Site: | Susquehanna  |
| Issue date: | 06/04/1991 |
| From: | Keiser H PENNSYLVANIA POWER & LIGHT CO. |
| To: | Butler W Office of Nuclear Reactor Regulation |
| References | |
| TAC-75999, TAC-76000, NUDOCS 9106110153 | |
| Download: ML18026A239 (99) | |
Text
'll REGULATORY INFORMATION DISTRIBUTION SYSTEM (RIDS)
ACCESSION NBR:9106110153 DOC.DATE: 91/06/04 NOTARIZED: NO DOCKET FACIL:50-387 Susquehanna Steam Electric Station, Unit 1, Pennsylva 05000 50-388 Susquehanna Steam Electric Station, Unit 2, Pennsylva 050 388 AUTH. NAME AUTHOR AFFILIATION KEISER,H.W. Pennsylvania Power & Light Co.
RECIP.NAME RECIPIENT AFFILIATION R BUTLER,W.R. Project Directorate I-2
SUBJECT:
Forwards facility final response to request for addi info on PL-NF-90-001. D DISTRIBUTION CODE: A001D COPIES RECEIVED:LTR ENCL SIZE: S TITLE: OR Submittal: General Distribution NOTES:LPDR 1 cy Transcripts. 05000387 LPDR 1 cy Transcripts. 05000388 RECIPIENT COPIES RECIPIENT COPIES D ID CODE/NAME LTTR ENCL ID CODE/NAME LTTR ENCL PD1-2 LA 1 1 PD1-2 PD 1 1 RALEIGH,J 2 2 INTERNAL: ACRS 6 6 NRR/DET/ECMB 7D 1 1 NRR/DET/ESGB 1 1 NRR/DOEA/OTSB11 1 1 NRR/DST 8E2 1 1 NRR/DST/SELB 7E 1 1 NRR/DST/SICB8H7 1 1 NRR/DST/SRXB 8E 1 1 NUDOCS-ABSTRACT 1 ~ 1 OC~EHB~ 1 0 OGC/HDS2 1 0 ~TEE 1 1 1 RES/DSIR/EIB 1 1 EXTERNAL: NRC PDR 1 1 NSIC 1 1 NOTES: 2 2 D
A D
D NOTE TO ALL "RIDS" RECIPIENTS:
PLEASE HELP US TO REDUCE WASTE! CONTACT THE DOCUMENT CONTROL DESK, ROOM Pl-37 (EXT. 20079) TO ELIMINATEYOUR NAME FROM DISTRIBUTION LISIS FOR DOCUMENTS YOU DON'T NEED!
TOTAL NUMBER OF COPIES REQUIRED: LTTR 26 ENCL 24
Pennsylvania Power 8 Light Company Two North Ninth Street ~Allentown, PA 18101-1179 ~ 215/774-5151 Harold W. Keiser Senior Vice President-Nuclear 215/7744194 JUNO <59)
Director of Nuclear Reactor Regulation Attention: Dr. W. R. Butler, Project Director Project Directorate I-2 Division of Reactor Projects U.S. Nuclear Regulatory Commission Washington, DC 20555 SUSQUEHANNA STEAN ELECTRIC STATION FINAL RESPONSE TO RAI ON PL-NF-90-001 PLA-3578 FILES A7-8C A17-2 R41-2
),
References:
- 1. PLA-3566, H. W. Keiser to W. R. Butler, "In'itial Response to RAI on PL-NF-90-001 (SCU Questions)" dated April 23, 1991.
- 2. Letter, H. C. Thadani to H. W. Keiser, "Request for Additional Information - Susquehanna Steam Electric Station Units 1 and 2 (TAC Nos. 75999/76000)", dated February 15, 1991.
Dear Dr. Butler:
Reference 1 transmitted PPEL's responses to questions 1 and 2 of Enclosure 2 (questions on PL-NF-90-001) of your Request for Additional Information (Reference 2). This letter transmits PPSL's responses to the remaining questions on PL-NF-90-001.
Please note the following with regard to this response:
PPSL's response to question 17 on PL-NF-89-005 'is contained in-this submittal. Reference 3 indicated that this question would be deferred to this submittal.
Errata discovered since the original printing of PL-NF-90-001 have been provided.
Sections 2.9.2 and 2.10 have been revised based on our response to question 22.
.e,'i0bii0i53 PDR 5'i0604 AGOCK 05000387 P PDR
FILES A7-8C/A17-2,"PLA-3678 R41-2 Dr. W. R. Butler As stated in our previous responses, PPSL requests NRC feedback on a priority basis in order to support our submittal of the proposed reload license amendment for Susquehanna Unit 1 Cycle 7 (U1C7) based on PP8L methods. An understanding of any significant concerns with our responses is needed by early July, 1991 in order to avoid impacting the licensing analysis we will be preparing for the December, 1991 U1C7 submittal.
Any questions on this response should be directed to Hr. R. Sgarro at (215) 774-7916.
Very truly yours, H. W. Keiser Attachment cc: Document Control Desk (original)
NRC Region I Hr. G. S. Barber, NRC Sr. Resident Inspector - SSES Hr. J. J. Raleigh, NRC Project Manager - OWFN Hr. L. I. Kopp, NRR/SRXB - OWFN
ATTA HMENT T PLA- 7:
INALRE P ET RAI N PL-NF- I 9106110153
UESTION 3 The RETRAN conservative bias of E=-9X for the overpressurization transient is based on only three measurements. Adjust this bias in order to insure that the MCPR safety limit is not exceeded with a 95X probability and 95X confidence level.
RESPONSE 3 The mean values of the ratio of measured to calculated RCPR are 0.921 and 0.917, as presented in Section 8.3 of PL-NF-89-005, for the internal and peripheral forms of the XN-3 correlation, respectively. All values of the ratio of measured RCPR to calculated RCPR were less than 1.0. A mean value less than 1.0 indicates an inherent conservatism in the calculation of RCPR using PPKL's RETRAN model. As discussed in the responses to NRC guestions 14, 18, and 19 on PL-NF-89-005 (Reference 3-1), the conservatism of the RETRAN model for pressurization transients is the result of the RETRAN model's conservative calculation of the transmission of the pressure wave from the steam dome to the core. Due to the fact that the mean and standard deviation were derived from comparisons of code calculations with measured data, the only technically consistent way to adjust the "bias" in the model would be to revise and re-benchmark the model. PPKL does not currently plan to revise the RETRAN,model, because the current model is conservative for pressurization transients as shown in PL-NF-89-005 (Reference 3-2).
As discussed in the response to NRC guestion 2 on PL-NF-90-001 (Reference 3-3),
the goal of the PP&L SCU method is to calculate a MCPR operating limit such that the applicable Specified Acceptable Fuel Design Limit (SAFDL) will not be violated during normal, operation or Anticipated Operational Occurrences (AOOs). The applicable SAFDL is discussed in the Standard Review Plan (SRP)
Section 4.4 and is found in the Susquehanna FSAR (Section 4.4. 1. 1):
"Specifically the Minimum Critical Power Ratio (MCPR) operating limit is specified such that at least 99.9 percent of the fuel rods in the core are not expected to experience boiling transition during the most severe
r moderate (Per Regulatory Guide 1.70 Revision 2) frequency transient events" Similarly, as stated in the SER on GETAB (Reference 3-4):
"the proposed design basis (i.e., more than 99.9X of the fuel rods in the core would be expected to avoid a boiling transition caused by single operator errors or equipment malfunctions) is acceptable when applied to core-wide transients such as a turbine-trip" The PP8L SCU method calculates the MCPR operating limit to demonstrate compliance with this design basis. The approach used by GE (95X probability/
95X confidence level of not exceeding a statistically calculated MCPR value which assures compliance with the SAFDL), while an acceptable approach, is not required by the General Design Criteria (GDC) or the SRP. The appropriate SAFDL is 99.9X of the rods expected to avoid boiling transition. The PP8L SCU method calculates a MCPR operating limit consistent with this SAFDL.
References 3-1 "Response to RAI on PL-NF-89-005", Susquehanna Letter PLA-3542, March 13, 1991.
3-2 "gualification of Transient Analysis Methods for BWR Design and Analysis", PL-NF-89-005, December 1989.
3-3 "Initial Response to RAI on PL-NF-90-001(SCU guestions)", PP&L Letter PLA-3566, April 23, 1991.
3-4 "General Electric BWR Thermal Analysis Basis (GETAB): Data, Correlation and Design Application", NED0-10958-A, January 1977.
t'p 5
UESTION 4 Without the SCU methodology, the safety limit and event-specific RCPR will be calculated separately in determining the oper ating limit MCPR. Modify the technical specifications to include the safety limit MCPR.
RESPONSE 4 In the PP&L SCU methodology, the operating limit MCPR is determined directly to demonstrate compliance with the applicable SAFDL (see response to Question 3). As demonstrated by the responses to Questions 1 and 2 (Reference 4-1), PP&L believes its SCU methodology is a logical extension of currently used NRC approved SCU methods and that the PP&L method conforms to
.-the applicable regulations. Therefore, use of the SAFDL (i.e., 99.9X of the fuel rods expected to avoid boiling transition) as the "THERMAL POWER, High Pressure and High Flow" safety limit in the technical specifications is appropriate. In addition, the analyses performed to date demonstrate that the application of PP&L's SCU methodology produces more conservative MCPR operating limits than those produced by our current fuel vendor~s NRC approved licensing methods. These analyses are a further demonstration of the overall conservatism of the PP&L approach.
Reference 4-1 "Initial Response to RAI on PL-NF-90-001(SCU Questions)", Susquehanna Letter PLA-3566, April 23, 1991.
EJ gl'
'.! )
4 ~
I,
UESTION 5 Provide the description and validation for the RODDK-E code used to select the strongest worth control rod positions for the shutdown margin analyses.
RESPONSE 5 The RODDK-E code is a computer program which was developed by the Electric Power Research Institute to provide estimates of relative control rod wor ths.
The detailed methodology is presented in Reference 5-1. A brief description of RODDK-E is provided below. RODDK-E uses the FLARE methodology (Reference 5.2) to perform a 2-D neutron balance and source distribution calculation using data which is spacially collapsed from the more accurate SIMULATE-E 3-D calculation. RODDK-E is used to predict the relative order of control rod worths in order to facilitate the selection of the strongest worth rods to be used in the SIMULATE-E shutdown margin calculations. The RODDK-E predictions of relative rod worths do not replace the more detailed SIMULATE-E analyses but assist in limiting the number of rods that need to be analyzed with SIMULATE-E.
RODDK-E uses data from three-dimensional SIMULATE-E All Rods In (ARI) and All Rods Out (ARO) cases'at the point in cycle of interest to determine the collapsed two-dimensional k-infinity, migration area, and diffusion coefficient information required by the FLARE algorithm. RODDK-E reproduces the SIMULATE-E ARI two-dimensional source distribution before it performs its perturbation cases to evaluate each control rod's worth. In the perturbation cases, the control rod only affects the neutronic properties of the adjacent nodes.
In order to validate the use of the RODDK-E code, 70 separate cases were analyzed which included different reloads, cycle and core exposures,- control histories, and void histories. For each of these cases, many SIMULATE-E calculations (i.e., 602 total individual rod full out calculations) were performed to ensure that the strongest worth rod'as identified by SIMULATE-E calculations was included in the set of calculations. For 60 of the 70 cases,
J k~,
,I 4
~
the RODDK-E calculated strongest worth rod was the same rod as the SIMULATE-E calculated strongest worth rod. From the other 10 cases, the maximum RODDK-E calculated reactivity difference between the RODDK-E calculated strongest worth rod and the SIMULATE-E calculated strongest worth rod was less than
.0016. These SIMULATE-E and RODDK-E analyses form the basis for deciding which control rods will be analyzed with SIMULATE-E to ensure that the strongest worth rod is evaluated.
References 5-1 W. R. Cobb, et. al., "ARHP-02 Documentation: Part II, Chapter 8-SIMULATE-E (Mod. 3) Computer Code Manual", EPRI NP-4574-CCM, Part II, Chapter 8, Appendix C, September, 1987.
5-2 D. L. Delp, et. al., "FLARE, A Three-Dimensional Boiling Water Reactor Simulator", GEAP-4598, General Electric Company, 1964.
>.0 14 v>
~ l 1
UESTION 6 Is the cycle-specific highest worth rod used in the rod withdrawal error analysis?
RESPONSE 6 The cycle specific highest worth control rod is not always used in the rod withdrawal error analysis, since the highest reactivity worth rod is not necessarily the limiting error rod (i.e., the rod which produces the largest RCPR for the rod withdrawal error analysis).
At a given core exposure, the location of the highest worth rod changes with changes in operating conditions (e.g., power, flow, rod pattern). The worth and location of the highest worth rod is primarily dependent on the neutron flux and void profile in the core at the conditions of interest. The worth and location of the highest worth rod are of interest in various analyses. Of particular interest are the analyses of: core shutdown margin, the control rod drop accident, and the Rod Withdrawal Error (RWE). For each of these analyses, the location and worth of the limiting control rod are different, because the initial conditions for each of these events are substantially different.
For the RWE analyses, the core has a control rod density of approximately 25X or less. Twenty five percent control rod density is much less than the amount used in either the control rod drop accident analysis (approximately 40X) or the core shutdown margin analysis (>99X). In addition, the power and temperature for rated conditions (used for the RWE analysis) result in a full power void profile and in most cases, a flatter radial and axial power shape than the power shapes for either the control rod drop accident analysis or the core shutdown margin analysis. The RWE analysis is, thus, initiated from a flatter power shape than the other rod related analyses because of the effect of voids and the reduced number of rods available in the core to shift and peak the power shape. Due to the lower control rod density, the limiting rod location is closer to the center of the core where the power distribution can
be peaked more easily in order to produce a conservatively peaked power distribution for the RWE analysis.
~ ~
~
For the RWE analysis, the limiting error rod is the rod which results in the largest RCPR. This may not correspond to the highest reactivity worth rod.
The limiting error rod location is highly sensitive to the relative increase in local power, radial power, and rod block monitor response. Withdrawal of a higher worth rod may initially produce a larger local power increase, but the rod block monitor response would also have a larger increase and thereby would produce a faster rod block. Depending on the net effect of these two competing phenomena in the RWE event, the highest reactivity worth rod may not be the limiting error rod. In order to determine the limiting error rod in the RWE event, a full RWE analysis is performed for rod locations shown in Figure 2. 1-1 of PL-NF-90-001. As discussed in the response to question 7, this ensures that the most limiting error rod is considered in the analysis.
"i.l
'I
~
f,(
jt>>
b 0
. I'~y 4 '
2h'X
UESTION 7 Demonstrate that the rod withdrawal error hCPR calculated for the central region of Figure 2. 1-1 is bounding for all core locations including those with only 2 or 3 LPRH strings.
RESPONSE 7 The demonstration that the RWE hCPR calculated for the central region rods shown in Figure 2. 1-1 of PL-NF-90-001 is bounding relies on both a qualitative evaluation and sample RWE analyses. The qualitative evaluation given below focuses on the capability of the control rod pattern to put the limiting CPR bundle near the error rod location (thus producing a higher hCPR for the RWE analysis). The sample RWE analyses also described below show that error rod locations where only 2 or 3 adjacent LPRH strings exist are not limiting.
These discussions and analytical results support the assertion that the limiting rod will be located among those rods identified in Figure 2.1-1.
It should also be noted that the xenon free assumption provides significant conservatism in the development of limiting control rod patterns. Assuming a more realistic xenon concentration would eliminate most of these rod patterns from the evaluation, because the control rod density is substantially lower for increased xenon conditions, thereby providing less capability to force the limiting MCPR bundle near the error rod location. Assuming a more realistic xenon concentration would eliminate the most adverse rod patterns from the evaluation, thus reducing the calculated hCPR.
The qualitative justification that the limiting dCPR would be calculated to occur for a central region rod is based on the operational restrictions imposed on control rod withdrawal. Two fundamental restrictions are placed on control rod withdrawal: 1) Banked Position Withdrawal Sequencing constraints below the Low Power Setpoint (approximately 20X), and 2) quarter-core symmetry deeply inserted rods. Peripheral control rods which have only 2 or 3 LPRH
'or strings associated with the RBM response are of no concern in the RWE analysis, because all peripheral rods are withdrawn by approximately 20X power
J II J
(due to Banked Position Withdrawal Sequencing constraints). Even if the peripheral rods were inserted, the bundle powers near rated conditions are lower in the core periphery (due to neutron leakage) than the bundle powers in other areas of the core.
Rod patterns used for the Susquehanna units are quarter-core symmetric. This fact provides a fundamental reason why central region rods are more limiting.
Consideration of geometry, control rod density, and neutronic coupling characteristics relevant to the RWE analysis error rod location, and the quarter-core symmetry assumption lead to the conclusion that central region error rod locations produce worse dCPRs. The geometry of the control rod pattern for quarter-core symmetry provides more flexibility/capability to force the limiting HCPR bundle near the error rod location when the error rod is near the center of the core. This is particularly true for rod densities that are low (i.e., less than 30X) as in. the RWE event. Given the assumption of quarter-core symmetry, the peaked power area for rod locations outside the central area is spread over all four quadrants of the core. For a central region rod, the peak power areas of the quadrants are closer and more coupled.
This effect results in a larger bundle power rise and, hence, a larger hCPR for central region error rods.
Sample RWE analyses demonstrate this effect. Three RWE analyses, were performed with error rods located outside the central region.
Case 1: error rod location 30-07 (2 adjacent LPRM strings);
Case 2: error rod location 14-15 (3 adjacent LPRH strings);
Case 3: error rod location 46-15 (4 adjacent LPRM strings).
The results of these cases indicate that error rod locations one cell in from the periphery where only 2 or 3 LPRN strings exist are not limiting even when the failure of four LPRHs is assumed. In reality, the RBN system automatically provides a rod block when more than half of the LPRH inputs to a RBM channel fail. Hence, only 2 LPRHs would be allowed to fail in case 1, 3
LPRHs in case 2, and 4 LPRHs in case 3. Even with this conservative assessment, the RWE cases using error rod locations 30-07, 14-15, and 46-15 produced hCPRs which were less than the central region limiting case by at least 0.03.
The above discussions support the examination of only those control rods identified in Figure 2. 1-1 of PL-NF-90-001 for licensing applications.
gUESTION 8 In the RWE analysis identify the location, relative to the error rod, of the fuel bundle assumed to be on limits.
RESPONSE 8 The location of the limiting fuel bundle at the beginning of the RWE event is forced, by means of the rod pattern, to be in a control cell either diagonally adjacent or face-adjacent to the error rod control cell. Using the coordinate system in Figure 2. 1-1 of PL-NF-90-001, the initial MCPR limiting bundle (35-26) for the presented sample RWE analysis is located in a control cell diagonally adjacent to the error rod control cell (38-31).
4'r h 0l II r> '
v~
I
'I
UESTION 9 How do the LPRM and bundle power uncertainty used in the RWE analysis compare to the values approved for the ANF Susquehanna-1 and 2 safety limit calculatioris? Justify the use of smaller values.
RESPONSE 9 The SCU analysis of- the RWE event consists of two major analyses: the HCPR safety limit type analyses and the transient RCPR analysis. The bundle power and LPRM uncertainties that pertain to the RCPR calculation and those that pertain to the MCPR safety limit type calculations, however, are different and not directly comparable. The uncertainties in the absolute values of bundle power and LPRM measurement (which affects bundle power uncertainty) are of significance to the HCPR safety limit type calculation, while the RWE RCPR calculation depends on the uncertainties in SIMULATE-E's ability to predict changes in bundle power (which affects the RCPR directly) and changes in LPRH reading (which affects the time at which the RBH provides a rod block). The LPRH and bundle power uncertainties used in the PPKL RCPR analyses are discussed below in relation to the related uncertainties in the ANF MCPR safety limit calculations.
In the MCPR safety limit type analyses, the.LPRH measurement uncertainty, as it was used to determine the power distribution uncertainties, is 3.4X. The LPRH measurement uncertainty affects the bundle power uncertainty because the POWERPLEX calculated bundle power is affected by the measured LPRM feedback in the UPDATE methodology. The LPRM detector response measurement uncertainty is based on the value reported in the General Electric report NEDO-20340 (Reference 9-1). This LPRM measurement uncertainty is different in definition from the SIMULATE-E LPRM calculational uncertainty used to conservatively estimate the uncertainty on the change in RBM response. For the RWE RCPR analysis, the uncertainty in the RBH response is based on the uncertainty in SIMULATE-E's prediction of the change in LPRH response. The uncer tainty in the change in LPRM response is conservatively based on the uncertainty in the absolute value of LPRH response as described in Section 2. 1.4.2 of s~,
l' g4 1,
k,)
I
PL-NF-90-001. The LPRH calculational uncertainty of 6.24X used in the determination of the RBH response uncertainty for the RWE analysis is larger than the 3.4X LPRM measurement uncertainty used in the determination of the radial bundle power uncertainty for the MCPR safety limit type analysis, since the 6.24X contains both measurement and calculational components.
Similarly, the RCPR uncertainty is conservatively based on the uncertainty in SIMULATE-E's calculation of bundle power, as described in Section 2. 1.4.2 of PL-NF-90-001. The SIMULATE-E bundle power uncertainty is 2.77X and its only use in the RWE RCPR analysis is to conservatively determine the RCPR calculational uncertainty. In contrast, the bundle power uncertainty used in the HCPR safety limit type calculations is based on POWERPLEX calculational uncertainties and includes TIP and LPRH measurement uncertainties, which are not applicable to the SIMULATE-E uncertainty. As a result, the MCPR safety
. limit bundle power uncertainty is larger than the 2.77X SIMULATE-E calculational uncertainty.
In conclusion, the LPRM and bundle power uncertainties used in the HCPR safety limit type analyses are different from those uncertainties used in the RCPR analyses for the RWE event because of their different uses and the bases for their derivation.
References 9-1 "Process Computer Performance Evaluation Accuracy", J.F. Carew, NED0-20340, General Electric Company, June 1974.
/ah
~ i
UESTION 10 Why is the o'gyp'.032 uncertainty for the mislocated fuel bundle combined with the RCPR distribution? How is the aggpg 0.037 determined7 RESPONSE 10 The o'gyp'qual to 0.032 described in Section 2.2.3.1 of PL-NF-90-001 represents the uncertainty in the SIMULATE-E calculated RCPR for the mislocated bundle event. The RCPR distribution represents the variation in RCPR for all possible mislocations and all possible rod patterns occurring at the cycle exposure that produces the worst results (i.e., highest RCPRs). The RCPR distribution was conservatively derived from SIMULATE-E calculations using rod patterns and mislocations specifically selected to maximize RCPR.
The RCPR uncertainty was combined with the distribution of calculated RCPR for the mislocated bundle event via a Honte Carlo combination of the two distributions. The purpose of this combination is to generate a tolerance factor that, when added to the RCPR calculated with the method described in Section 2.2 of PL-NF-90-001, produces an RCPR that bounds at least 95X of the RCPRs from all possible bundle mislocations.
A Monte Carlo analysis to combine the RCPR distribution and RCPR calculational uncertainty was used because RCPR was not normally distributed. From the distribution obtained from this Honte Carlo combination, an upper bound on the, RCPR at 95X probability 95X confidence was determined (RCPR ~~). The tolerance factor (referred to as ka<<), is determined by:
ko'RcpR gag RCPReva~
RCPRyg<yg 0.037 where: RCPR,,< the calculated value of RCPR using the methodology described in Section 2.2.3 of PL-NF-90-001.
For licensing analyses, this tolerance factor will be added to the calculated value of RCPR.
'i'a II 4
)4 h
'E
\ I It t4 1~
UESTION 11 In the mislocated fuel bundle event, justify the use of the all-rods-out condition for determining hCPR.
RESPONSE 11 The conservatism of using the all-rods-out condition for the mislocated fuel bundle analysis was established by examining SIMULATE-E results for 93 separate combinations of rod pattern and mislocated bundle location for the limiting exposure (i.e., exposure that produces maximum RCPR values). All of these rod patterns and mislocations were selected to exacerbate the peaking in the area of the mislocation in order to maximize the impact of the mislocation on RCPR. Of the 93 rodded cases, only three produced slightly higher RCPRs than the RCPR calculated assuming an all-rods-out condition. The largest calculated RCPR exceeded the all-rods-out RCPR by a negligible amount (i.e.,
.0017).
Most of the rod patterns used in the mislocated bundle analysis would not be expected to occur in actual operation, because they produced excessive power peaking even in the correctly loaded core. Use of operationally realistic rod patterns would result in RCPR values which are significantly smaller than those that would be obtained from the analysis method presented in Section 2.2.3 of PL-NF-90-001. Therefore, use of the all-rods-out condition is appropriate for the bundle mislocation analysis.
~ I
+C
~ gk Ill 4W 4'
UESTION 12 In the loss-of-feedwater-heating (LFWH) event, how is the SIMULATE-E uncertainty in the change in bundle power (in addition to the core thermal power) accounted for?
RESPONSE 12 The RCPR for the LFWH event is primarily dependent on two parameters:
- 1. the change in core inlet subcooling
- 2. the change in absolute bundle power The uncertainties in core inlet subcooling are covered by the assumed 100'F decrease in feedwater. temperature. An analysis of the Susquehanna feedwater heater configuration demonstrated that the maximum expected decrease in feedwater temperature produced by a single failure or single operator error is approximately 50'F. The actual value obtained from plant startup test data is approximately 40'F. Therefore, the 100'F decrease assumed for licensing analysis provides significant conservatism.
The RCPR for the LFWH is a function of the change in,absolute bundle power.
Therefore, the uncertainties in SIMULATE-E's ability to calculate both the change in relative bundle power and the change in core thermal power are considered in deriving the value of ogpss PL-NF-90-001. The RCPR for the LFWH
~ contained in Section 2.3 of event can be expressed as:
RCPR C
- hPs chan e in RCPR change in absolute bundle power hP~ change in absolute bundle power (X rated)
Since the LFWH event is a core wide transient, the peak relative bundle power does not change significantly during the event. Therefore, the RCPR can be rewritten as:
RCPR = C
- P
- hCTP where: PR
= relative bundle power CTP = core thermal power (X rated)
The uncertainty in RCPR is:
o= C*P*v, and K<gcpg ~ = C
- Pg
- Ku~cyp The value of Ko>> ~ for the LFWH event was calculated to be 3X, based on a comparison of a SIMULATE-E calculation to startup test data for a LFWH transient. The value of C P (which is equivalent to the change in RCPR divided by the change in core thermal power) was calculated to be .005 (X ')
based on SIMULATE-E LFWH calculations. Thus, Ko'gyp' is equal to 0.015 as given in Section 2.3 of PL-NF-90-001.
Reference 12-1 "qualification of Core Steady State Physics methods for BWR Design and Analysis", PL-NF-87-001-A, July 1988.
UESTION 13 In the SIMULATE-E LFWH calculation is the xenon maintained at the initial value? If not, how is the effect on power peaking accounted fort RESPONSE 13 For the LFWH event, the xenon distribution is assumed to be equal to the initial distribution throughout the event. The LFWH event requires approximately five minutes to achieve a new steady state power level once the temperature of the feedwater entering the vessel begins to decrease. The xenon distribution does not, therefore, have sufficient time to change appreciably. Therefore, assuming no change in the xenon concentration is appropriate for the LFWH event.
1 t.
S (
UESTION 14 What is the increase in the LHGR during the LFWH event?
RESPONSE 14 For all cases analyzed, the increase in core peak LHGR was less than or equal to 21X. The LHGR during the LFWH event increases as a result of the increase in core thermal power and a shift in the axial power distribution toward the bottom of the core, due to the reduction in core inlet temperature. The net effect on the LHGR in certain nodes is an increase which is slightly larger than the accompanying increase in core thermal power. Operating within the LHGR limits and the LHGR limits for APRM Setpoints (as incorporated in the T-factor) in the Susquehanna Technical Specifications assures that the LHGR transients evaluated in Reference 14-1 bound the LFWH event.
Reference 14-1 "Generic Mechanical Design for Exxon Nuclear Jet Pump BWR Reload Fuel",
XN-NF-85-67(P)(A), Rev. 1, September 1986.
UESTION 15 Demonstrate that the SIMULATE-E steady-state analysis is bounding for the worst case transient feedwater flow, pressure and core power increase.
RESPONSE 15 The loss of feedwater heating event is a slow transient in which the core conditions vary over a period of minutes. After initiation of the event, the core thermal power increases slowly as a result of a slow, monotonic decrease in the core inlet coolant enthalpy. After about five minutes, the core thermal power reaches a new steady state power level. Due to the slow nature of the transient and the smooth transition from the pre-transient to the post-transient state, steady state core physics methods are appropriate for analyzing this event.
No rapid changes in core power, flow, pressure, or inlet enthalpy occur which would make the transient hCPR worse at intermediate times. To demonstrate this, Figures 15-1 through 15-5 show core power, core flow, dome pressure, feedwater temperature, and feedwater flow for a planned loss of feedwater heating event which occurred during the start-up test program for Susquehanna Unit 1. Core flow (Figure 15-2), dome pressure (Figure 15-3) and feedwater flow (Figure 15-5) are virtually unaffected by the change in feedwater temperature. Individual loop feedwater temperatures (Figure 15-4) vary smoothly with time. The effect on core power is shown in Figure 15-1. The fluctuations are caused by signal noise in the instrumentation and GETARS transient recording system data. RETRAN calculations of this event which are provided in Section 5.3 of the transient analysis methods topical report (PL-NF-89-005) also show a smooth monotonic increase in core thermal power.
As a result, the largest hCPR would be expected to occur at the post-event conditions as calculated by the steady-state physics methods.
FIGURE 15-1 LOFWH CORE POWER 90 85 O
O O
80 Legend
~ GETARS
'75 0 50 100 150 200 250 300 350 400 TIME (SEC)
FIGURE 15-2 LOFWH CORE FLOW 110 105
~~o 1O0 O
95 0O 90 85 Legend
~ GETARS 80 0 50 100 150 200 250 300 350 400 TIME (SEC)
0
, FIGURE t5-3 LOFWH DOME PRESSURE 1000 990 M
0 980 LLJ CL V) 0 960 Legend s GETARS
'50 50 100 150 200 250 300 350 400 TIME (SEC)
pl a
~ 44
FIGURE 15-4 LOFWH FEEDWATER LOOP TEMP 400 375 350 LJJ 325
'IJ CL 4J 300 O
0 275 ID LIJ 250 Legend 225 s LOOP A o LOOP B
~ LOOP C 200 0 50 100 150 200 250 300 350 . 400 TIME (SEC)
I V"
~ 4
FIGURE 15-5 LOFWH FEEDWATER FLOW 14 12 10 Kl 8
O w
O W 4 Legend a GUITARS 0 50 100 150 200 250 300 350 400 TIME (SEC)
g1 t>>
A y4 4II
UESTION 16 How will the shutdown margin prediction uncertainty and design criterion be determined for a specific reloads RESPONSE 16 The shutdown margin prediction uncertainty is typically calculated prior to each reload licensing analysis using the most current cold and hot k-effective data base. This data base contains the results from SIHULATE-E core follow calculations and cold in-sequence critical calculations. The shutdown margin prediction uncertainty is based on a statistical evaluation of the SIHULATE-E calculated core k-effectives for all modelled critical conditions. The SIHULATE-E calculated core k-effectives are compared to the target core k-effectives (discussed in Section 3.2 of Reference 16-1) to determine the calculational uncertainty. Tests for the normality of the error distributions are performed, and either non-normal or normal reliability factors are used, as appropriate, to determine the 95X probability/95X confidence level tolerance factor for SIHULATE-E's cold critical prediction.
This SIHULATE-E cold critical prediction tolerance factor is added to the allowance for manufacturing uncertainty associated with the control rods and the fuel to produce the shutdown margin design criterion. It should be noted that, in addition to the shutdown margin analyses, a shutdown margin demonstration is performed at the beginning of each fuel cycle in accordance with the plant Technical Specifications.
Reference 16-1 "gualification of Steady State Core Physics Hethods for BWR Design and Analysis", PL-NF-87-001-A, Pennsylvania Power 5 Light, July 1988.
L "I ai
~l
UESTION 17 What is the basis for the 165 ppm boron margin to account for imperfect mixing of boron in the evaluation of the Standby Liquid Control System (SLCS)?
RESPONSE 17 The allowance of 165 ppm is based on GE design studies for Susquehanna reported in Reference 17-1. Additional information recently presented to the BWR Owner's Group Emergency Procedures-II committee members (Reference 17-2) reported that a GE reevaluation of data reported in Reference 17-3 shows that mixing of the injected sodium pentaborate is good at core flows as low as 5X of rated. As a result, the assumption of only 660 ppm boron in the reactor coolant following SLCS injection provides significant conservatism.
References 17-1 "Design Analysis and SAR Inputs for ATWS Performance and Standby Liquid Control System: Susquehanna 1 and 2 Plant", NEDE-25458, Rev.l, April 1982.
17-2 "Information on ATWS/Stability Mechanisms", GE Letter OG91-008-62, January 29, 1991.
17-3 "Assessment of ATWS Compliance Alternatives", NEDC-30921, July 1985.
t),,
UESTION 18 Demonstrate the conservatism (relative to SIHULATE-E) of the two approximate methods for determining shutdown margin in the SLCS analysis.
RESPONSE 18 A best estimate SIHULATE-E model was developed to determine the conservatisms inherent in the two methods for evaluating the SLCS capability described in PL-NF-90-001. For this best estimate model, the inclusion of 660 ppm soluble boron in the reactor coolant was modelled by including the cross section variation as a function of both exposure and void history for all cross sections. For a UIC2 sample analysis consistent with the one presented in Section 2.5.4 of PL-NF-90-001, the best estimate SIHULATE-E model calculated that 660 ppm boron provides 0.081 dk shutdown reactivity.. This is almost twice the amount indicated by the boron worth method and approximately 0.005 hk shutdown reactivity more than the cross section modification method. (Note:
a typographical error was found in the value reported for the cross section modification method in Section 2.5.4 of PL-NF-90-001. The calculated shutdown reactivity using this method was reported as 0.070 hk but actually is 0.076 dk). The smaller difference between the best estimate and the cross section modification method is expected, since boron is a thermal absorber and primarily impacts the thermal absorption cross section. The effect of boron addition on the thermal absorption cross section is nearly constant with exposure. Therefore, the best estimate model is closely approximated by the cross section modification approach prior to the addition of the conservative factor described in Section 2.5.3 of PL-NF-90-001. Note, however, that all of these approaches assume 660 ppm boron in the reactor coolant, rather than the 825 ppm minimum value applicable to Susquehanna as discussed in the response to guestion 17. The additional 165 ppm boron would provide approximately 0.02 hk additional shutdown reactivity.
UESTION 19 Demonstrate that the end-of-cycle reactivity calculations for the ANF LOCA analyses are bounding for the entire cycle.
RESPONSE 19 The LOCA blowdown analyses for the Susquehanna units are performed by ANF with their NRC approved methodology which uses end-of-cycle scram, moderator density, and Doppler reactivities. As part of a reload licensing analysis, PP&L compares scram, moderator density, and Doppler reactivity data generated with PP&L's computer codes to the values used in the ANF LOCA analyses. Based on the results of analyses using PP&L methods to generate these reactivity inputs (described below) and our knowledge of the physical phenomena that occur during a LOCA, the combination of these reactivity components are shown to be more adverse for a LOCA evaluated at end of cycle.
PP&L has performed reactivity calculations for U2C5 Beginning-Of-Cycle (BOC),
Middle-Of-Cycle (MOC), and End-Of-Cycle (EOC) conditions. The scram, moderator density, and Doppler reactivity results are plotted in Figures 19-1, 19-2, and 19-3. All three figures are plotted with the same reactivity scale for easy comparison. Figure 19-4 shows the power versus time calculated by RETRAN for the three times in cycle. The power would decrease faster in an ANF LOCA calculation than shown in this figure because the moderator density reactivity produced by the rapid voiding during a LOCA is not simulated in the RETRAN scram reactivity calculations. Note that the power has been reduced to near decay heat levels by 3.0 seconds following scram.
As shown in Figure 19-1, the BOC and MOC scram reactivities are significantly greater than the EOC scram reactivity for times less than 3.0 seconds. This is expected because at EOC the control rods are fully withdrawn and at BOC and MOC the core is partially rodded. As shown in Figures 19-2 and 19-3, the differences in the moderator density and Doppler reactivities between BOC, MOC, and EOC are small compared to the differences in scram reactivity between BOC, MOC, and EOC. At times greater than 3.0 seconds, the core power is nearly all from decay heat (see Figure 19-4), and, hence, reactivity effects have a negligible effect on the LOCA results. Therefore, for the LOCA analysis, the overall reactivity effect at EOC is conservative compared to other times in the cycle.
FIGURE 19-1 U2C5 LOCA SCRAM CURVE COMPARISON 0
-10
-20 I
I O -30 IJJ tZ V) -40
-60 ~ EOC 0 0.5 1.5 2.5 3.5 4.5 TIME (SEC)
FIGURE 19-2 MODERATOR DENSITY REACTIVITY 0
-10
-20 I
-30 I
O IJJ CL
-40
-50 Legend
~ EOC
-60 0.5 0.55 0.6 0.65 0.7 0.75 0.8 0.85 0.9 0.95 NORMALIZED RELATIVE MODERATOR DENSITY
FIGURE 19-3 DOPPLER REACTIVITY 0
-10
-20
-40
-50 ~ w " ~ ~ Legend
~ EOC
-60 0 200 400 600 800 1000 1200 1400 1600 FUEL TEMPERATURE (F)
FIGURE i9-0 U2C5 LOCA SCRAM CURVE COMPARISON 100 80 Cl I-LL 60 O
N LIJ 40 LLI O
O 20 Legend
~ EOC 0
0 0.5 1.5 2 2.5 3 3.5 4.5 TIME(SEC)
I g
0 P
% ~
I
, jN,1 f
UESTION 20 Is the definition of reactivity (on Page 73) identical to that used by ANF in the LOCA analysis? 'If not, how will this difference be accounted for?
RESPONSE 20 PP8L's definition of reactivity used to generate data for comparison with ANF reactivities used in the LOCA analyses is the same as the definition used by ANF. For input to ANF s point kinetics LOCA blowdown model, reactivity is defined as a function of fuel temperature and moderator density using the following equation:
p~ = (hk,gg/k,sz) (1.0/P)
(20-1)
= ((k,gg 1. 0) /k,ff)// p To obtain the equation on page 73 of PL-NF-90-001, the SIHULATE-E k-effective bias is calculated by:
Bias - lb 1.0 where: lb = calculated SIHULATE-E core k-effective for the base case The calculated core k-effective for the various moderator density and fuel temperature perturbations is:
= Aq (Xb 1 0) (20-2) where: l,. = calculated SIHULATE-E core k-effective for perturbation case i Substituting the expression for k-effective in Equation 20-2 into the expression for reactivity in Equation 20-1 and rearranging the equation gives:
pi = 1. 0 '1. 0/P)
This expression is identical to the equation found on page 73 of PL-NF-90-001.
~1
,I
~ t' l;v a4,s 1a
UESTION 21 How will the maximum worth rod be determined for the control rod drop analysis?
RESPONSE 21 The maximum reactivity worth rod is determined with SIMULATE-E, using the approach described below. The determination of the location of the maximum reactivity worth rod primarily relies on two analysis assumptions about plant operation. The first assumption is that eight control rods are inoperable, which is the maximum number allowed by Technical Specifications, and are fully inserted. This conservative'assumption allows the core radial flux distribution to be shifted to maximize the worth of certain control rods. The other assumption is that adherence to the Banked Position Withdrawal Sequence (BPWS) constraints (Reference 21-1) is maintained. The BPWS constraints are enforced with the use of the Rod Sequence Control System and the Rod Worth Minimizer.
~ ~
~
Asymmetric full-core control rod patterns which include eight inoperable (i.e., fully inserted) control rods on one side of the core are assumed for the analysis. This assumption results in a peaked flux distribution in the other half of the core, which is where the maximum worth rod is located. The flux distribution is such that the flux is peaked near the periphery of the core away from the inoperable/inserted control rods. Therefore, rods near the periphery will have higher worths. A more uniform distribution of inoperable control rods throughout the core (a more likely occurrence) would reduce the worth of the maximum worth rod.
With eight inoperable control rods inserted in one-half of the core, adherence to the BPWS rules causes certain control cells to be limiting. Because BPWS rules are employed in the withdrawal and insertion of control rods, control rod patterns near the black-and-white pattern (i.e., a "checkerboard" pattern of fully inserted and fully withdrawn control rods) contain inserted rods that are surrounded by a number of unrodded cells (i.e., four to six). Control cells surrounded by six unrodded cells are potentially limiting (i.e., contain kt tg
the highest worth rods) due to the higher flux in the region, while control cells having four unrodded control cells surrounding them have lower worth rods and, hence, are not limiting. Potentially limiting control rods in each control rod sequence (i.e., A-2, B-2, A-l, and B-1) are analyzed to determine which rod has the maximum reactivity worth.
Because rod worth can change somewhat with exposure," reactivity rod worths are calculated for various core exposures during the cycle to assure that the maximum rod worth is calculated for the entire cycle.
In conclusion, the maximum reactivity rod worth is determined by establishing conservative control rod patterns and calculating control rod wor ths for the potentially limiting rod locations using SIMULATE-E. Each of the control rod sequences and various exposures are analyzed.
Reference 21-1 "Banked Position Withdrawal Sequence", NE00-21231, General Electric Company, January 1977.
I
~V I;~p yf, l
UESTION 22 What is the root-mean-square difference between the POWERPLEX" calculations with PPL and ANF nuclear data2 What increase in the ANF approved safety limit power distribution uncertainties will this ANF-to-PPL POWERPLEX" difference introducef RESPONSE 22 PP&L intends to use the POWERPLEX core monitoring system with lattice physics input from the CPH-2 code as described in Section 2.10 of PL-NF-90-001.
Currently, the lattice physics data in POWERPLEX is generated by ANF using their XFYRE code. In order to make the HCPR safety limit type analyses consistent with the use of CPH-2 data in POWERPLEX, the radial, nodal, and local power uncertainties must be determined and used in the HCPR safety limit analyses. The comparison between these uncertainties for POWERPLEX with PP&L lattice physics input and POWERPLEX with ANF lattice physics input is based on the ANF results presented in Reference 22-1 and the PP&L benchmarking analyses of Susquehanna data using CPH-2 data in POWERPLEX discussed in this response.
Reference 22-1 provides the radial and nodal predicted TIP response uncertainties used to develop the radial bundle and axial power uncertainties for ANF lattice physics input in POWERPLEX. In addition, Reference 22-1 provides the uncertainty in local power distribution applicable for ANF lattice physics data in POWERPLEX.
PP&L has reperformed the benchmarking analyses presented in Section 2. 10 of PL-NF-90-001. The new analyses incorporate an improved instantaneous void to void history correlation called VHIST13 (Reference 22-2). The VHIST13 correlation (which correlates cross section adjustments with differences between the void history and the instantaneous void level) is being used in the current Susquehanna POWERPLEX input decks developed by ANF. The original benchmarking analyses used a previous instantaneous void to void history correlation called CHPR. The incorporation of the VHIST13 correlation in PP&L's benchmarking analyses provides consistency between the benchmarking results and the current reload POWERPLEX input decks. The results presented in Table 22-1 are derived from the new analyses.
p II I I
I
<<I fg y')i gf:
Table 22-1 shows the TIP response uncertainties calculated by PP8L using CPH-2 lattice physics input in POWERPLEX. Also shown in the table are references to the equivalent uncertainties calculated by ANF using ANF lattice physics input in POWERPLEX. The ANF values are not included in the table because they are proprietary to ANF. The results presented in Table 22-1 show that the POWERPLEX TIP response uncertainties with CPH-2 data are less than or equal to the POWERPLEX TIP response uncertainties with ANF lattice physics data.
Attached are revised Sections 2.9.2 (" Uncertainties" ) and 2. 10 (" Core Monitoring System Inputs" ) of PL-NF-90-001 that reflect the new analysis results using VHIST13. These revised sections will replace the previously presented sections when the approved version of PL-NF-90-001 is issued.
The radial bundle power and axial offset uncertainties for use in the safety limit type analyses were calculated based on the TIP response uncertainties given in Table 22-1, and the local power uncertainty was derived based on the results of comparisons of CPH-2 calculations with guad Cities gamma scan data, presented in Reference 22-3. The methodology used to derive the power distribution uncertainties applicable for CPH-2 data in POWERPLEX was consistent with the ANF methodology presented in Reference 22-'1. Table 22-2 provides the resulting power distribution uncertainties which PPEL intends to use as input to the HCPR safety limit type analyses for the Susquehanna cores using POWERPLEX with CPH-2 lattice physics input for core monitoring.
References 22-1 "Exxon Nuclear Methodology for Boiling Water Reactors - Neutronic Methods for Design and Analysis", XN-NF-80-19 (P)(A), Volume 1 and XN-NF-80-19 (P)(A), Volume 1, Supplements 1 5 2, March 1983.
22-2 "Void History Correlation", Letter from R.A. Copeland (ANF) to M.W. Hodges(NRC), RAC:058:88, September 13, 1988.
22-3 "gualification of Steady State Core Physics Methods for BWR Design and Analysis", PL-NF-87-001-A, July 1988.
TABLE 22-1 Comparison of Radial and Nodal TIP Response Uncertainties for POWERPLEX With PP&L and ANF Lattice Physics Input POWERPLEX POWERPLEX Uncertainty w/ PP&L Input w/ ANF Input Calculated Radial to Measured Radial TIP 2.76X Ref 22-1, Pg 6-22 Response b Calculated Nodal to Measured Nodal TIP 5.63X Ref 22-1, Pg 6-22 Response 6<<
TABLE 22-2 Power Distribution Uncertainties for POWERPLEX w/
PP&L Lattice Physics Input Uncertainty Value Radial Bundle Power 4.0 X Local Power 2.46 X Axial Offset* 6.0 X
- ANF safety limit methodology uses the axial uncertainty in terms of axial offset, defined as the power in the top half of the core minus the power in the bottom half divided by the total power in the core.
2.9.2 Uncertainties Both the conventional safety limit and the "safety limit type" analyses utilize a Monte Carlo procedure to combine various uncertainties in order to demonstrate that 99.9X of the fuel rods in the core are not expected to experience boiling transition, in conformance with Standard Review Plan 4.4 (Reference 19). The uncertainties considered are listed in Table 2.9-1. The "system uncertainties" listed in Table 2.9-1 concern measurement uncertainties of system parameters. The values used are the same as those listed in References 17 and 18.
As described in Section 2. 10, the CPM-2 lattice physics code will be used to generate inputs to the POWERPLEX core monitoring system. Of the "fuel related uncertainties" listed in Table 2.9-1, the only ones which are potentially affected by PP8L's methodology are: 1) radial bundle power, 2) local power, and 3) axial power. The uncertainties in these three parameters using CPM-2 input in the core monitoring system are given in Section 2. 10. The uncertainties given in Table 2. 10-2 will be used in the MCPR safety limit type analyses for cycles using POWERPLEX input decks generated by PP&L using CPM-2.
42
- 2. 10 Core Monitorin S stem In uts PP8L utilizes the POWERPLEX Core Monitoring Software System (References 8 and
- 34) to perform the on-line thermal margin calculations for Susquehanna SES.
ANF is the developer of POWERPLEX, and the original application of POWERPLEX at Susquehanna SES was based on input from ANF's lattice physics code XFYRE (Reference 28). POWERPLEX input decks are fuel cycle specific and contain nuclear physics data, peaking factors, and thermal limits information. This input is used by ANF's three-dimensional reactor simulation code XTGBWR (Refer ence 28) and the other calculational routines in POWERPLEX (e.g.,-
thermal limit evaluations, Traversing In-core Probe (TIP) predictions, and isotopics determination).
PPKL developed, validated, and will use an input deck generation methodology based on results from the CPM-2 lattice physics code (References 1 and 4).
PP8L has developed modifications to the NORGE-B2 code (Reference 7) to create the lattice physics input data needed by the POWERPLEX core monitoring system (e.g., cross sections, peaking factors, etc.). The calculation flow path is shown in Figure 1-2. CPH-2 performs lattice physics calculations for specific fuel designs; NORGE-B2 formats the CPM-2 results into the POWERPLEX input deck format; and then the NORGE-82 output and other POWERPLEX inputs (i.e., thermal hydraulic data, core geometry, calculation options, neutron detector response data, etc.) are combined to complete the POWERPLEX input deck for a specific cycle.
Benchmarking analyses were performed using this process for Susquehanna SES Unit 1, Cycles 1, 2, and 3 and Unit 2 Cycles 1, 2, and 3. The results of the benchmarking analyses were evaluated against ANF TIP response uncertainties, measured TIP data, and SIMULATE-E results. The use of PPSL's input deck generation methodology for POWERPLEX (using CPM-2 lattice physics data) produced smaller TIP response uncertainties than those produced by ANF's methodology (Reference 28) as shown in Table 2. 10-1. The predicted to measured TIP response comparisons for POWERPLEX using CPH-2 input are approximately equivalent to the SIMULATE-E results (Figures 2. 10-1 through
- 2. 10-6). The k-effective calculated by POWERPLEX using PPSL's input decks is reasonable and well behaved and is compared to the SIMULATE-E calculated k-effective in Figures 2. 10-7 through 2. 10-12.
These results lead to the conclusions that for Susquehanna SES: 1) POWERPLEX using PP&L generated CPM-2 input data produces smaller TIP response uncertainties than those obtained from POWERPLEX using ANF generated inputs,
- 2) the k-effective calculated by POWERPLEX using CPM-2 input is reasonable and well behaved, and 3) POWERPLEX with CPM-2 input produces good predicted to measured TIP response comparisons. The power distribution uncertainties applicable to POWERPLEX with CPM-2 input were calculated based on the benchmarking analyses described above and on the squad Cities gamma scan comparisons presented in Reference 1. These uncertainties, which will be used in the MCPR safety limit type analyses, are presented in Table 2. 10-2.
44
TABLE 2.10-1 Com arison of Radial and Nodal TIP Res onse Uncertainties for POWERPLEX With PP&L and ANF Lattice Ph sics In ut
~PPIEI I RI PII E EE POWERPLEX Uncertaint w PP&L In ut w ANF In ut Calculated Radial to Measured Radial TIP 2.76X Ref 28, Pg 6-22 Response 6, Calculated Nodal to Heasured Nodal TIP 5.63X Ref 28, Pg 6-22 Response 6<<
TABLE 2.10-2 Power Distribution Uncertainties for POWERPLEX w/
PP&L Lattice Physics Input Uncertainty Value Radial Bundle Power 4.0 X Local Power 2.46 X Axial Offset* 6.0 X
- ANF safety limit methodology uses the axial uncertainty in terms of axial offset, defined as the power in the top half of the core minus the power in the bottom half divided by the total power in the core.
c>
FIGURE 2.10-1 SUSQUEHANNA SES UNIT 1 CYCLE 1 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0 I
19.0 ~ ~ ~ << ~ <<<< ~ ~
I <<
1B.O 17.0 Legend <<<<<<<<<<
16.0 0 SIMULATE-E ~ ~ << ~
16.0 POWERPLEX(CPM) <<
14.0 <<
13.0 M
12.0 c( 11.0 O <<
z 1O.O 9.0 I- <<
e.o ~ <<
Q 7.o B.O
- . 0 0':OQ: O 6.0 4.0 """0:.
cjoy:,~:,, G jj, 3.0 I~ << ~ ' << ~
2.0 <<<<
~
~
~
1.0 0.0 0 1 2 3 4 6 8 7 8 9 10 11 12 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-2 SUSQUEHANNA SES UNIT 1 CYCLE 2 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0 19.0 18.0-17.0 Legend g ~
1B.O 0 SIMULATE-E 16.0 POWERPLEX(CPM) 14.0 be 13.0 12.0 11.0 O
10.0 9.0 1 I- I B.o ~ 'I 4 8 7.0 B.O 0
Q 0
O A 6.0 A ~ " ~
4.0 ~ ~
3.0 2.0 1.0 0 ~ P ~
0.0 0 2 3 4 6 e 7 8 9 10 11 12 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-3 SUSQUEHANNA SES UNIT 1 CYCLE 3 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0 19.0 ,
~
18.0 17.0 Legend I I 1
'I ~
I 16.0 0 SIMULATE-E Ie-16.0 6 POWERPLEX(CPM) 14.0 13.0
', ~ I ~
12.0 11.0 a
10.0 ~ 'I ~
I ~ I 9.0 gG
":0 I-6.0 j,~" ": """ "p""0'" - *~ ':
7.0
- o: no::
6.0 5.0 4.0 3.0 2.0 ~ e ~ I ~ ~ ~
1.0 0.0 0 1 2 3 4 6 B 7 8 9 10 11 12 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-4 SUSQUEHANNA SES UNIT 2 CYCLE 1 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0 19.0 r
18.0 ~ ~ ~ ~
17.0 Legend 18.0 0 SIMULATE-E 15.0 6 POWERPLEX(CPM) r 14.0 r r r\ r r
13.0 ~ ~ ~ +
r 12.0 11.0 CI 10.0 0 9.0 r r
O~:.o:.:
1 8.0
- GD O j, .: 0 7.0 0
8.0
~~o"'~lboo oo---::----':--o-:----:- --o--- o- q----:""
6.0 4.0
~ ~ ~
3.0 r 2.0 1.0 0.0 0 1 2 3 4 6 , 8 7 8 9 10 11 12 13 CYCLE EXPOSURE (GWD/MTU)
'I
FIGURE 2.10-5 SUSQUEHANNA SES UNIT 2 CYCLE 2 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0
\ 'I 19.0 18.0 E
~ ~
17.0 Legend 1B.O 0 SIMULATE-E ~ ~ ~ ~ I' ~ ~ ~
16.0 6 POWERPLEX(CPM) 14.0 ~ ~ ~
13.0 12.0 'I 0::
11.0 Q E V
9 10.0 9.0 ....0 4 I- 0 II I
B.O 7.0 .....O..:,...........
B.O 6.0 0
- 4.0 1~
3.0 0 ~
2.0 1.0 " ."" ~
0.0 2 3 4 6 B 7 8 9 10 11 12 CYCLE EXPOSURE (GWD/MTU)
V I
FIGURE 2.10-6 SUSQUEHANNA SES UNIT 2 CYCLE 3 RELATIVE NODAL RMS OF TIP RESPONSE COMPARISONS 20.0 19.0 18.0 I' ~
17.0 Legend 18.0 0 SIMULATE-E i ~ ~
15.0 POWERPLEX(CPM) '
14.0 g I I
13.0 ~ ~ ~ I ~ I~ ~ ~ ~ (
A 12.0 11.0 O ~.
10.0 9.0 I-8.0 0
7.0 ~ 6
'gI ~ ~
Q~ \ )~ ~ ~ " ~
r
.M:
I' 8.0 'I 6.0 O Q'"''' QS" I t O.:.o...6 4.0 Oo, Q C. OO,O II 3.0 2.0 1.0 0.0 0- 1 2 3 4 6 B 7 8 9 10 11 12 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-7 SUSQUEHANNA SES UNIT 1 CYCLE 1 HOT CALCULATED K-EFFECTIVES 1.016 I 'I Legend 1.010 V ~ 0 SIMULATE-E
' ~
POWERPLEX(CPM) 1.006 I
1.000 UJ I-O LLI 0.995 I
hC 0.990 I'.985 0.980 *
~ ~ ~
0.976 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 CYCLE EXPOSURE (GWD/MTU)
1' FIGURE 2.10-8 SUSQUEHANNA SES UNIT 1 CYCLE 2 HOT CALCULATED K-EFFECTIVES 1.015 Legend 1.010 0 SIMULATE-E POWERPLEX(CPM)
I 1.006 ~ ~
'I 1.000 Lll I-O 0.996 UJ I
hC 0.990 0.985 0.980 0.976 0 4 5 6 7 8 10 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-9 SUSQUEHANNA SES UNIT 1 CYCLE 3 HOT CALCULATED K-EFFECTIVES 1.015 Legend II \
1.010 0 SIMULATE-E POWER PLEX(CPM) 1.006 * - r 1.000 LLI I-O ILI I
0.885
(@5'" """"6.- - "-.-"
hC r 0.990 1 0.885 r r 0.880 '
0.976 I 0 2 4 5 6 7 8 10
...CYCLE EXPOSURE (GWD/MTU)
gA FIGURE 2.10-10 SUSQUEHANNA SES UNIT 2 CYCLE 1 HOT CALCULATED K-EFFECTIVES 1.016 Legend ~
E 1.010 0 SIMULATE-E iAL POWERPLEX(CPM) 1.006 1.000 UJ I-O UJ 0.995 UJ I
hC 0.990
~ ~ ~
0.985 0.980
~ ~ ~ ~ ~ ~
f 0.976 I I 0 1 2 3 4 5 8 7 8 9 10 11 12 13 14 15 CYCLE EXPOSURE (GWD/MTU)
FIGURE 2.10-11 SUSQUEHANNA SES UNIT 2 CYCLE 2 HOT CALCULATED K-EFFECTIVES 1.015 Legend 1.010 \
0
~
SIMULATE-E POWERPLEX(CPM) 1.006 1.000 ill I-O 0.995 I
hC 0.990 0.986 0.980 ~
0.976 I 0 1 2 4 5 6 7 8 10 12 CYCLE EXPOSURE (GWD/MTU)
FIGUR .10-12 SUSQUEHANNA SES UNIT 2 CYCLE 3 HOT CALCULATED K-EFFECTIVES 1.016
~ ~ Legend 1.010 0 SIMULATE-E POWERPLEX(CPM) 1.006 1.000 LLI I-oUJ O.SS6
'v u LL Lll I
hC 0.990 0.986
'I II I 0.980 0.975 0, 1 4 6 6 7 8 10 CYCLE EXPOSURE (GWD/MTU)
jul f t]
UESTION 23 How is the difference between the CPH-2 and the ANF lattice calculation of the ANF fuel vault k-infinity ( 1.388 criterion accounted for?
RESPONSE 23 The ANF fuel vault criterion (k-infinity less than or equal to 1.388) was based on Honte Carlo calculations performed by ANF. These calculations were corrected for bias in the Monte Carlo calculations to produce the resultant k-infinity criterion. This assumes that the new fuel vault will be subcritical with a system k-effective of approximately 0.95 in conformance with SRP Section 9. 1. 1. No additional biases or corrections are incorporated into the criterion. Thus, any benchmarked and approved lattice physics code can be utilized to verify compliance, provided the bias and uncertainty of the selected lattice physics code is evaluated. Compliance with this criterion is verified with CPH-2, assuming the nominal U-235 enrichment in all pins is increased by 0.05 wtX U-235 and the nominal gadolinia loading is decreased by 10X. The benchmarking of PPKL's NRC approved lattice physics methods using the CPH-2 lattice physics code is documented in Reference 23-1. These evaluations show that CPH-2 slightly overestimates lattice reactivity by 0.0005 with a standard deviation of 0.0072. Because of the conservative evaluation with CPH-2 (i.e., increase in enrichment and decrease in gadolinia), the conservative KENO evaluation (i.e., 5X margin to criticality),
and the conservative bias in CPH-2, the CPH-2 bias and uncertainty are not explicitly included, and the CPH-2 calculated k-infinity is used to verify compliance with the KENO calculations.
Reference 23-1 "gualification of Steady State Core Physics Methods for BWR Design and Analysis", PL-NF-87-001-A, Pennsylvania Power 8 Light, July 1988.
UEST ION 24 How do the static analyses of Chapter 2 and the transient analyses of Chapter 3 differ from the ANF treatment of these events'ESPONSE 24 The static (core physics) analyses and transient (RETRAN) analyses described in PL-NF-90-001 are, in many cases, similar to the equivalent ANF analyses.
The core physics and transient analysis methods described in Sections 2 and 3 of PL-NF-90-001 are used for four major purposes:
- 1. To generate input to the technical specifications or demonstrate compliance with requirements of the technical specifications
- 2. To generate data as input to ANF analyses
- 3. To provide 1-D kinetics input for RETRAN
- 4. To generate lattice physics data for the POWERPLEX core monitoring system One of the principal differences between the PP&L methods and the ANF methods is in the computer codes used. The following table lists some of the main PP8L computer codes used and the corresponding computer codes used by ANF.
d~dddd 1 1 PP8tL ANF Lattice Physics CPM-2 XFYRE 3-D Nodal Core SIMULATE-E XTGBWR Physics / Thermal Hydraulics / MCPR Transient System RETRAN-02 COTRANSA (1-D)
PTSBWR (pt. kinetics)
Transient Hot Bundle RETRAN-02 XCOBRA-T Table 24-1 outlines the static and transient analyses performed by PP&L for a typical reload and compares the ANF and PP8L approaches to the analyses.
TABLE 24-1 PP8L and ANF Hethods used in Licensing Basis Analyses PP&L Hethods ANF Hethods Analysis Hethod/Output Hethod/Output Fuel Loading Error CPH-2 Calculate peaking factors for rotated bundle; verify- Not performed on a cycle-by-cycle basis (Rotated Bundle) values are less than those used for bounding analysis for C-Lattice plants (Cycle specific)
SIHULATE-E Calculate RCPR for rotated bundle using bounding peaking factors Fuel Loading Error SIHULATE E Calculate RCPR for worst mislocation XTGBMR Calculate RCPR for worst mislocation (Hislocated Bundle)
Loss of Feedwater Heating SIHULATE-E Verify applicability of generic analysis presented in PTSBHR Point kinetics calculation Section 2.3 to reload cycle Shutdown Hargin RODDK-E Determire high worth rod locations XTGBHR Calculate core k-effective with strongest worth rod withdrawn SIHULATE-E Calculate core k-effective-for one rod withdrawn for the locations identified by RNDK-E to determine shutdown margin (miniaxm)
Starxhy Liquid SIHULATE E Hethod 1: apply conservative boron worth to unborated XTGBQR Calculate borated core k-effective Control System k-effective Hethod 2: calculate borated core k-effective by modifying thermal absorption cross section.
Rod llithdrawal Error S IHULATE-E Calculate RCPR and RBH response for a RME as a fmction XTGBNR Deterministic RCPR analysis of RBH setpoint, operable RBH charael, and LPRH failure combination Honte Carlo analysis to combine relevant mcertainties LOCA Inputs SIHULATE-E Calculate void reactivity to verify ANF values used in XTGBMR same as PP&L S IHTRAN-E LOCA analysis are bounding for current cycle SIHULATE-E Calculate Doppler reactivity to verify ANF values used in XFYRE same as PP8L S IHTRAH-E LOCA analysis are bounding for current cycle RETRAN Calculate scram reactivity to verify ANF values used in COTRANSA same as PP&L SIHTRAN-E LOCA analysis are bounding for current cycle Fuel Storage Criticality CPH-2 Calculate bundle k-infinity to verify compliance with ANF XFYRE same as PP&L criterion TABLE 24-1 PAL and ANF Methods used in Licensing Basis Analyses PPSL Hethods ANF Hethods Analysis Method/Output Method/Output Control Rod Drop Inputs SINJLATE-E Calculate maxisun dropped rod worth, Doppler coefficent, XTGBNR 4 Calculate maxisus dropped rod worth and and four bundle peaking factor XFYRE four bundle peaking factor XFYRE Calculate Doppler coefficient and delayed neutron fraction SIHTRAN-E Calculate core average delayed neutron fraction HCPR Safety Limit Inputs SIWLATE-E Calculate radial bundle power distribution XTGBWR same as PPBL CPH-2 Calculate pin peaking factor distribution for each bundle XFYRE same as PPEL type Generator Load Rejection w/o RETRAN/ Calculate RCPR COT RAN SA/ Deterministic RCPR calculation Bypass CPRITER XCOBRA-T Feedwater Controller Failure RETRAN/ Calculate RCPR COTRANSA/ Deterministic RCPR calculation CPRI TER XCOBRA-T Recirculation F low Controller RE TRAN/ Deterministic RCPR calculation PTSBllR/ Deterministic RCPR calculation Failure CPRI TER XCOBRA HSIV Closure/ RET RAN Deterministic peak pressure calculation COTRANSA Deterministic peak pressure calculation ASHE Overpressure P
+ t
.Cdi 4
NOTE: The following question was asked as part of the NRC review of PL-NF-89-005. PP8L's response was deferred (
Reference:
PPEL Letter PLA-3542, Response to RAI on PL-NF-89-005, March 13, 1991), because it was determined that the question is more applicable to PL-NF-90-001.
The response is provided below.
TRANSIENT TOPICAL UESTION 17 Are the calculations performed by PPL and provided to ANF for input to safety analyses (e.g., LOCA, SLHCPR, rod drop and fuel handling) consistent with the accuracy and conservatism assumed in the approved ANF methodology'oes the ANF methodology assume that an allowance for uncertainties is included in the data provided by PPL2 How are the differences between CPH-2/SIHULATE-E and the ANF physics codes accounted forP TRANSIENT TOPICAL RESPONSE 17 The analyses performed by ANF which use PPEL results as input (described in PL-NF-90-001) are the Loss of Coolant Accident (LOCA) analysis, the Control Rod Drop Accident (CRDA) analysis, the Minimum Critical Power Ratio (HCPR) safety limit type analyses, and the Spent Fuel Storage Criticality Compliance evaluations.
In developing our licensing basis methods, PPLL had frequent discussions with ANF to assure that the data generated by PPSL for use in ANF analyses is appropriate and generated consistently with the equivalent ANF calculations.
While many conservative assumptions are used in generating data for the ANF analyses (e.g., power, time in cycle, etc.), the codes themselves are used in a "best estimate" mode, which is consistent with the way ANF generates this data.
ANF has benchmarked their physics codes to plant data to demonstrate accuracy and does not include allowances for uncertainties in their core physics models in generating input to these event analyses. Likewise, PPhL's NRC approved core physics models have been benchmar ked to Susquehanna and other plant data (e.g., Peach Bottom, guad Cities) to demonstrate their capability to accurately simulate plant measured data, and specific code uncertainties are not applied to the data supplied as input to the ANF analyses. The large inherent conservatisms of the assumptions used to generate this data and the ANF analytical methods in which the data is used (e.g., ANF's 10CFR50 Appendix K LOCA methodology) is considered sufficient to account for the small uncertainties due to the core physics methods. PPEL also uses this approach in generating data with their NRC approved core physics methods (Reference 17-1). The differences between the PPKL and ANF physics methods are negligible in comparison to the conservatisms in these analyses.
Reference 17-1 "gualification of Steady State Core Physics Methods for BWR Design and Analysis", PL-NF-87-001-A, July 1988.
ERRAT The items listed below describe typographical and other minor errors discovered in PL-NF-90-001, as well as minor changes needed to incorporate the revised Section 2. 10 described in the response to guestion 22. These changes will be made in the approved version of the report.
- 1. Pg 46, Section 2.2.3.2: After both occurrences of "less than", add the words "or equal to".
- 2. Pg 47, Section 2.2.4: "RCPR, < 0. 17" should be "RCPR, < 0. 17"
- 3. Pg 68, Section 2.5.4: The value 0.070 hk should be 0.076 hk l
- 4. Pg 113, Table 3. 1-1, Item 3: The base case input value used for the turbine control valve closure time should read: "Best estimate value based on plant data"
- 5. Pg 84: Tables 2. 10-1 and 2. 10-2 will be added and pages renumbered appropriately.
- 6. List of Tables: Table 2. 10-1 and 2. 10-2 will be added.
- 7. Pg 202: Reference 34 will be added - "Void History Correlation", Letter from R.A. Copeland (ANF) to H.W. Hodges (NRC), RAC:058:88, September 13, 1988.
- 8. Pg 200: Reference 13 will be changed to reflect the approved version of the ESCORE report (EPRI NP-5100-L-A).
,a ~
I