ML20056F686
| ML20056F686 | |
| Person / Time | |
|---|---|
| Site: | San Onofre  |
| Issue date: | 06/30/1993 |
| From: | Nazareth V, Swoope S, Thomsen O SOUTHERN CALIFORNIA EDISON CO. |
| To: | |
| Shared Package | |
| ML19310D647 | List: |
| References | |
| NUDOCS 9308300262 | |
| Download: ML20056F686 (80) | |
Text
{{#Wiki_filter:. . . . - . . . . . . -. . . . .-, - , p 4 l l l t ROOT CAUSE EVALUATION FOR CPC AXIAL SHAPE ANOMALY . FOR ; i SONGS UNITS 2 AND 3
}
i i JUNE 1993 1 l l i l r Authored by: V.F. Nazareth S.W. Swoope ! 0.J. Thomsen l i SCE NUCLEAR FUEL MANAGEMENT JggB3OO262930719 y g ADOC} 05000361 h = . - - PDR l
188LE OF CONTENTS SECTION PAGE 1.0.
SUMMARY
. . ........ .....................5 4
2.0. INTRODUCTION ............ ............... 8 3.0 BACKGROUNO .................... ........ 9 3.1. Reload Power Ascension Measurement ............. 9 3.2. SAM Acceptance Cri teri a . . . . . . . . . . . . . . . . . . . 9 3.3. CPC Overall Uncertainty Analysis .............. 10 4.0. SYMPTOMS OF THE PROBLEM . . . . . . . . . . . . . . . . . . . . . . . 11 4.1. BOC SAM Measurement . . . . . ................ 11 4.2. M0C to E0C Axial Shape Variations . . . . . . . . . . . . . . 12 1 a , 5.0. POSSIBLE ROOT CAUSES INVESTIGATED . . . . . . . . . . . . . . . . . 14 I 5.1. Excore Detector Errors .................... 14 5.2. Excore Signal Variability . . . . . . . . . . . . . . . . . . 15 5.3. Signal Precision ...................... 17 5.4. Relationship between Signals . . . . . . . . . . . . .. . . . 18 5.5. Effect of Number of Cases . . . . . . . . . . . . . . . . . . 18 5.6. CECOR Peripheral Power Calculation ............. 20 5.7. Possibility of CEFAST Ccde Problems . . . . . . . . . . . . . 20 5.8. ROCS Modeling Effects . . . . . . . . . . . . . . . . . . . . 20 5.9. Comparisons for Past Cycle Operation ............ 22 5.10. New Plant Computer Deadband Effect ............. 23 5.11. Fuel Management Effects . . . . . . . . . . . . . . . . . . . 26 6.0. ROOT CAUSE CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . 28 7.0. CORRECTIVE ACTIONS ........................ 29 8.0. REFERENCES ................ ............ 31 9.0. APPENDIX (CPC OPERABILITY LETTER) . . . . . . . . . . . . . . . . . 74
--Page 1 of 79--
1 1 l l i LIST OF TABLES I i i PAGE l TABLE
- 1. Comparison of Unit 2 Cycle 6 Measured SAMs ............. 32 j
- 2. Comparison of Unit 3 Cycle 6 Measured SAMs ............. 33 l 1
- 3. Maximum Excore Detector Error During FPA .............. 34 I
/
- 4. Normalized Excore Detector Signal Change with Burnup ........ 35 i
- 5. Excore Detector Signal Variability ................. 36 l
Comparison of Measured and Calculated SAMs 37 ;
- 6. .............
- 7. Data Set Size Impact on " Poor" SAM ................. 38 j t
- 8. Data Set Size Impact on " Good" SAM ................. 39 j ;
- 9. Comparison of Measured SAMs (Unit 2 -- CPC Channel A) . . . . . . . . 40 l i
- 10. Comparison of Measured SAMs (Unit 2 -- CPC Channel B) . . . . . . . . 41
[
- 11. Comparison of Measured SAMs (Unit 2 -- CPC Channel C) . . . . . . . . 42 .fi
- 12. Comparison of Measured TAMS (Unit 2 -- CPC Channel D) . . . . . . . . 43
- 13. Comparison of Measured SAMs (Unit 3 -- CPC Channel A) . . . . . . . . 44 )
- 14. Comparison of Measured SAMs (Unit 3 -- CPC Channel B) . . . . . . . . 45 (
/
- 15. Comparison of Measured SAMs (Unit 3 -- CPC Channel C) . . . . . . . . 46
- 16. Comparison of Measured SAMs (Unit 3 -- CPC Channel D) . . . . .. . . . 47 :
- 17. Qualitative Criteria Symptomatic of the CPC Axial Shape Anomaly for 30C Measured SAMs . . . . . . . . . . . . . . . . . 48 l i
l
- 18. Qualitative Acceptance of Measured SAMs for Various Cycles in Unit 2 ...................... 49 l t
- 19. Qualitative Acceptance of Measured SAMs for l Various Cycles in Unit 3 ...................... 50 l i
--Page 2 of 79-- i I
i l
i LIST OF TABLES : I i PAGE ' TABLE t
- 20. Acceptability of Measured SAMs Using MOC to E0C l Variation Criteria for Various Cycles in Unit 2 . . . . . . . . . . . 51
- 21. Acceptability of Measured SAMs Using MOC to E0C Variation Criteria for various cycles in unit 3 . . . . . . . . . . . 52
- 22. Overall Qualitative Acceptability of CPC Shape Synthesis for Various Cycles ......................... 53 l
- 23. Comparison of Neutronics Related Parameters for ABB-CE Digi tal Plants . . . . . . . . . . . . . . . . . . . . . . 54 l A
i
}
r l i r i 4 I l : l i 2
--Page 3 of 79--
l 4
LIST OF FIGURES FIGURE PAGE
- 1. Axial RMS Change Study (CPC-D) ................... 55
- 2. Core Average Axial Shape Comparison, Unit 2 Cy-le 6 (16 GWD/T) ... 56
- 3. Core Average Axial Shape Comparison, Unit 3 Cycle 6 (10 GWD/T) ... 57
- 4. Core Axial Shape Synthesis Comparison, Unit 2 Cycle 6 CPC-D (16 GWD/T) .......................... 58
- 5. Excore vs. Peripheral Relationship, U2C6 BOC FPA (CPC-D) ...... 59
- 6. Excore vs. Peripheral Relationship, U3C6 BOC FPA (CPC-D) ...... 60
- 7. Excore vs. Peripheral Relationship, U2C6 M0C Oscillation ...... 61
- 8. Excore vs. Peripheral Relationship, U3C6 MOC Oscillation ...... 62
- 9. ROCS Modeling Comparisons, SONGS-2 Cycle 6 (420 EFPD) . . . . . . . . 63
- 10. Comparison of SAM Determinants, SONGS Unit 2 ............ -
- 11. Comparison of SAM Determinants, SONGS Unit 3 ............ 65
- 12. Axial Shape RMS Error Change Study, Unit 2 ............. 66
- 13. Axial Shape RMS Error Change Study, Unit 3 ............. 67
- 14. Incore and Peripheral Power Correlation, U3C3 FPA (CPC-D) . . . . . . 68
- 15. Incore and Peripheral Power Correlation, U3C4 FPA (CPC-D) . . . . . . 69
- 16. Incore and Peripheral Power Correlation, U3C6 FPA (CPC-D) . . . . . . 70
- 17. Comparison of Unit 3 FPAs, CPC-C Middle Peripherals . . . . . . . . . 71 l
- 18. Comparison of Unit 3 FPAs, CPC-D Middle Peripherals . . . . . . . . . 72
. 19. Relative Excore Signal, Units 2 & 3 . . . . . . . . . . . . . . . . . 73 l l
--Page 4 of 79--
l 1
1.0
SUMMARY
In February 1993 during extension of the core follow process for SONGS Units 2 i and 3, an anomaly was discovered in the Core Protection Calculator (CPC) ; The CPC channels were i synthesized axial power shapes for both units. i generating " cosine type" axial shapes from middle of cycle (MOC) to the time ; in cycle of discovery while the predicted The axialobjective shapes were of this" saddle-type" report is to j axial shapes for these times in cycle. ; i describe the symptoms of the anomaly, the potential root causes investigated, ' 3 the conclusion as to the root cauce of the problem and the corrective actions i that will be implemented. ! The purpose of the CPC system is the Anticipated Operational Occurrences to To A00). tripdo(the this, reactor each CPCin the event of o channel l synthesizes the core average axial power shape based on three levels of excore detector signals. The relative excore detector readings are subsequently i l i adjusted constants. within the CPCs by a set of channel dependent shape l l the reload startup power ascension and installed into the TheCPC FPA channels by program involves i l means of a Fast Power Ascension (FPA) test program. ' taking incore and excore signal data at regular intervals during the initial The incore data i startup power ascension from approximately 20% to 68% power. ! is subsequently processed through a core analysis computer code, CECOR, to l determine the relative power at the core periphery. An automated data -
)
reduction computer code, CEFAST, then verifies the data, calculates the SAM i constants accentance criteriaand determines whether the measured to .iustify its implementation into the CPC SAM channels meets 3 [ a set o
*P i 3 provide assurance that at a 95'/95 ~ !
probability / confidence lev , the CPC DNBR and LHR will remain conservative and produce a reactor trip for a CPC protected design basis event. ; The functional aspects of the CPC software, the related plant instrumentation J and the procedures used to calibrate the instruments were reviewedToand support the FPA ' measured SAM was identified as the immediate cause of the anomaly. l full power operation, a SAM measurement was re-performed for both units using l a xenon oscillation initiated during the power reduction from 100% to The 85% e power at the time wher. the CPC axial shape anomaly was d! the type of CPC axial shapes consistent with the overall uncertainty analysis. l l A comparison of the BOC and MOC measured SAMs revealed four symptoms They that w } characteristic of the CPC axial shape anpmaly in the BOC measured SAMs. i are a low value of SAM middle element, negative values in the SAM inverse ! matrix, large magnitude off-diagonal corner elements and low value of the SAM ; determinant. Also two other characteristics in the CPC axial shape synthesis They are a large increase in i from MOC to EOC were symptomatic of the anomaly. l axial shape RMS error with depletion and a cosine type CPC synthesized axial shape. These six characteristics guided the root cause investigation. The root cause summary table shown beiow describes the areas of investigation pursued, the results of. the investigation and the root cause conclusions reached from each area of investigation.
-Page 5 of 79--
- Proprietary
ROOT CAUSE
SUMMARY
TABLE l 1 l Report Areas Results Conclusion Section Investigated 5.1 Excore a) Excore signals not Did not cause anomaly Detector mi scali brated Errors b) No unrecognized detector problem c) Excore signal change with i burnup sufficient for proper evolution of CPC synthesized i family of shapes 5.2 Excore Signal Excore signal variability larger Did not cause anomaly i Variability than at other ABB-CE digital plants 5.3 Signal Adequate signal precision in BDC Did not cause anomaly l t Preci si on SAM measurement ; 5.4 Relationship Relative change in normalizeri Lack of coherent relationship l tetween middle peripheral power integral is between nomalized middle l Signals comparable to random signal excore signal and normalized > variations during BOC SAM middle peripheral integral measurement caused anomaly 5.5 Effect of Number of cases did not Did not cause anomaly Number of significantly change SAM , Cases 5.6 CECOR Code calculation of peripheral Did not cause anomaly Peri pheral power was correct for FPA Power Calcul ation 1 5.7 CEFAST Code Code calculates SAM correctly Did not cause anomaly Probl ems ; 5.8 ROCS Modeling ROCS overpredicts axial shape Did not cause anomaly but may , Effects change during FPA explain why design codes did ! not predict anomaly , 5.9 Comparisons Quality of SAMs deteriorated from a) Anomaly first occurred on a 3 for Past Cycle cycle 3 to cycle 4 and remained generic basis in cycle 4 i Operation poor thereafter for both units b) Anomaly not caused solely by FPA 5.10 New Plant Lack of coherent relationship a) Incore Deadband in new , Computer between normalized middle plant computer during the Deadband peripheral integral and normalized FPA could have caused middle excore signal could have anomaly , i been produced by b) Manual data collection in L a) Incore deadband FFA for unit 2 cycle 4 b) Manual data collection could have caused anomaly in unit 2 cycle 4 ) 5.11 Fuel Insufficient change in the middle The listed fuel management j Management of the core during the FPA could effects individually did not Effects have been caused by but collectively could have a) Near Zero ITC at BOC, caused anomaly b) Relatively flat axial shape at BOC : c) Low leakage f uel management , ' pattern l 1 l t l --Page 6 of 79-- ! 1 7..
i
\
l l The fundamental root cause of the SONGS CPC axial shape anomaly is a lack of coherent relationship between the normalized middle peripheral power integral i and the normalized middle excore detector signal during the Fast Power ' Ascension SAM measurement at BOC. The breakdown in coherent relationship during the FPA measurement is concluded to be caused either individually or collectively by (1) the incore deadband implemented in the software of the ' i plant computer, (2) manual collection of excore signal data and (3) a combination of fuel management effects of near zero ITC at BOC, relatively . flat axial peak at BOC and low leakage fuel management. l The corrective actions that will be implemented for unit 2 cycle 7 to prevent :' reoccurrence of this anomaly are (1) reduction or removal of the incore detector deadband in the plant computer software for the fast power ascension, (2) modification of the reload startup SAM measurement procedure such that if the automated data collection mechanism is not available, manual data ; collection will be designed to eliminate time lag effects for the FPA program, .i , (3) changing ABB-CE design code modeling for the FPA simulation, (4) adding a : second level of criteria at BOC to determine the need for a "later-in-cycle" ; measurement and (5) monitoring the CPC synthesized axial shapes on a routine l basis to determine the acceptability of the CPC axial shape synthesis and, if i needed, the time in cycle for the later-in-cycle measurement. Based on the , results of the unit 2 cycle 7 startup SAM measurement and subsequent axial ! 3 shape monitoring during the cycle, these corrective actions may be modified ! for future cycles. ; I f f l i . F f : 1 i l I t [ 4 t i E i
--Page 7 of 79-- l
, 1
.-,---,r-
- . . . . - - - - ~ -- .. - - .
A 2.0. INTRODUCTION In February 1993, during extension of the physics core follow process, an j anomaly was discovered in the Core Protection Calcelator (CPC) synthesized axial power shapes for both SONGS units 2 and 3. From cycle burnups after approximately middle of cycle, the CPCs continued to generate cosine-type ' axial power shapes instead of the predicted saddle-type shapes. This was ' inconsistent with the type of response assumed for the overall CPC uncertainty analysis methodology used to generate penalty factors installed in the CPC : channels. On 2/25/93 and 2/26/93, a 15% penalty was applied to the DNBR calculation for both units after the units were downpowered to approximately i 85% power. Thereafter, new CPC algorithm constants were measured and -! implemented, the 15% penalty was removed and the unit was restored to full ! power on 3/5/93. The new measurement for cycle 6 will henceforth be referred to as the middle of cycle or "M0C SAM Measurement". This report was generated to describe the symptoms of the anomaly, the ! potential root causes investigated, the conclusion as to the root cause of the problem and the corrective actions that are to be implemented. The assessment of the operability of the CPCs with the anomalous CPC axial shape synthesis l i for cycles 3 through 6 for units 2 and 3 is enclosed in the Appendix. ; t i 5 i a i i l I ) i ! . i a e i ! i ! l
--Page 8 of 79-- l l
i
I l 1 3.0. BACKGROUND 3 .1. Reload Power Ascension Heasurement The CPC system is designed to protect the plant by tripping the reactor to ensure that the Specified Acceptable Fuel Design Limits (SAFDLs) on minimum ! Departure from Nucleate Boiling Ratio (DNBR) and peak Linear Heat Rate (LHR) r are not violated in the event of an Anticipated Operational Occurrence (A00). , As part of the calculation of minimum DNBR and peak LHR, the CPC algorithm l synthesizes axial power shapes using the excore detector signals. ' l t The current methodology to assure the accuracy of this axial shape synthesis l in the CPCs requires the measurement of Shape Annealing Matrix ' SAM) , addressable constants for each CPC channel during the initial startup after ' each reload. This measurement is referred to as the Beginning of Cycle or "BOC SAM Measurement". The methodology used to detemine the SAM constants is called the Fast Power Ascension (FPA) Program *. The FPA program involves taking snapshots of incore and excore signal data - during the initial power ascension for each cycle from approximately 20% power t to approximately 68% power. This snapshot data is taken approximately once , every 30 stinutes with a power ascension rate of approximately 3% per hour. : The snapshot data is then processed through the core analysis computer code CECOR , to determine the core peripheral power distribution in the region of E l the core that affects the excores. CECOR subsequently integrates the peripheral power distribution into three axial regions in the core, resulting in one third core peripheral power integrals. The relationship between the one third peripheral power integrals and the normalized three axial level excoresignals,determinestheSAM.{[ , . *? <
][The CPC cfiainel depenifint SAM measured ~usTng tliis'~tRiinf(ue is ; then used in the axial shape synthesis for that channel for the entire reload cycle.
j 3.2. SAM Accertance Criteria . 3
!. The SAM cr.iculation process is automated using a computer code called CEFAST .
This code verifies the acceptability of the plant data, calculates the SAM , i values and determines the acceptability of the measured SAM by comparing it to ! several criteria. There are three sets of review criteria and one acceptance criteria to judge the adequacy of the SAM as follows:
~(a) Excore Signal Check -- This review criterion checks the consistency of the excore signals during ,
the FPA measurement by comparing the. , predicted and measured normali: red excore l signal. Thecriterionis} between J measured and predicted signals in the top , andbottomofthecoreand{ between
--Page 9 of 79-- ;
4
- Proprietary l
l l l i the measured and predicted signals.in the ! middle of the core. If any of the criteria are not met, then the detector signals are classified as suspect and may be deleted from the SAM calculation, as appropriate. This review criterion tests the quality of (b) Test Matrix Criteria -- the SAM. The test value is a measure of the ". goodness" of the SAM. A. test value ofTetween[ *p l}meansthattheSAM . measured is acceptable. (c) Peripheral Power Criteria-- This review criterion determines the capability of the SAM to regenerate the peripheral power integrals for the set of cases used in the FPA measurement. The top, middle and bottom third core peripheral power integrals using the excore detector signals and the measured ~ SAM are compared to the-actual CECOR peripheral integrals. A standard each of , deviation the integralsof of the less difference _fo4] means i than[*p - the SAM is acceptable. (d) Axial Shape RMS Error -- This is the acceptance criterion used to demonstrate that the CPC synthesized axial shape using the measured SAM is consistent . with the axial shape synthesized fror; the incore detectors using CECOR. An axial power shape Root _Mean Square (RMS)_ error of less thanl[ *p l 1' '] demonstrates that the CPC will synthesize acceptable shapes for the set of cases used and.the SAM is acceptable for installation into the CPCs. 3.3. CPC Overall Uncertainty Analysis . The reload pcwer ascension measurement process is simulated as input into the CPC overall uncertainty analysis'. The uncertainty associated with the measurement of the SAM is combined with the other uncertainties in the CPC system to detennine overall DNBR and LHR uncertainty penalties to be installed in the CPCs. The uncertainti~es along with the simulation methodology, provide assurance that at a 95/95 probability / confidence level, the CPC DNBR and LHR will remain conservative and produce a reactor trip for a CPC protected design basis event.
--Page 10 of 79--
- Proprietary
4.0, SYMPTOMS OF THE PROBLEM The CPC axial shape anomaly was caused by the CPC synthesized axial power ! shapes being outside the assumptions of the overall CPC uncertainty analysis i methodology. The uncertainty analysis based on a FPA simulation predicted l that the CPCs would generate " flat" or " saddle" type of axial power shapes at nominal conditions from HOC to EOC while the axial shapes synthesized by the CPC channels continued to be " cosine-type" shapes during these times in life. ; The discussion in this section will focus on the inconsistency between the i synthesized axial power shapes in the CPC channels and the nominal expected j CPC synthesized axial shapes for the CPC uncertainty analysis for both units. l The CPC axial shape anomaly was discovered by recognizing that the CPCs were synthesizing cosine-type axial shapes while the predicted shapes at the time i in cycle where saddle-type shapes. On further investigation it was realized that certain characteristics of the BOC SAM were symptomatic of the anomaly. i The following sections are characteristics of the symptoms of the problem and i are not in the order of discovery. ; 4.1. BOC SAM Measurement The BOC SAM measurements at startup for cycle 6, at both units 2 and 3, were [ performed in accordance with the Fast Power Ascension Program guidelines and i test procedures. Although the measured SAMs met most of the review l i criteria and also met the acceptance criteria for the test, investigation into - the results identified the following four characteristics of the measured SAMs ; that signify the likelihood of the existence of this anomaly: ! t (a) Low Middle SAM Element 1 ] Tables 1 and 2 show the SAM elements for the B0C SAM measurement for ! cycle 6 using the Fast Power Ascension technique and compare iL to the i ! MOC SAM measurement for units 2 and 3 respectively. One consistent i l difference between the BOC and MOC measured SAMs is in ;he middle l element of the matrix (i .e. S2 ,) . For the BOC SAM measurement this ; element is lower by a factor of between 3 and 7 times the value i determined for the MOC SAM measurement. This signifies that the ! relationship between the normalized center peripheral power integral and i
- normalized center excore detector signal is significantly weaker for the BOC SAM measurement than the MOC SAM measurement. ,
(b) Negative Inverse Matrix Elements ; Tables 1 and 2 also list the inverse matrix elements (S"n) of the BOC
- and MOC measurement SAMs for units 2 and 3, cycle 6. A consistent ;
difference between the two matrices is that the BOC SAM measurements , have negative elements in their inverse matrices whereas the MOC SAM ) measurement have no negative inverse matrix elements. This signifies a l lack of sufficient physical relationship between the normalized 4 peripheral power integrals and excore detector signals for the BOC SAM measurement when compared to the MOC SAM measurement. ]
--Page 11 of 79--
1 i i i
I (c) Large Magnitude Off-Diagonal Corner Elements t From Tables 1 and 2 it can be seen that the off-diagonal corner elements 5 33 and 5 3 (i .e. that in the M0C SAM measurement are considerably smaller j than for tb)e BOC SAM measurement for cycle 6 for both units. For 4 the MOC SAM measurement these values _ ranged between -1.2 and +1.2 while , for the BOC SAM measurement the values varied between -3.5 and -1.5. l Typically, off-diagonal corner elements should be small since there ' should be very little correlation between the top peripheral power , integral and bottom detector signal or the bottom peripheral power integral and top detector signal. The larger off-diagonal corner l' 2 elements signify that the SAM from the BOC measurement is less physically meaningful than from the MOC measurement. ; ! (d) Low Value of Determinant Tables 1 and 2 also list the determinant of the measured SAMs' for units 2 and 3 cycle 6 respectively. The value of the determinants for the HOC : SAM measurements were in the neighborhood of 200 while for the BOC SAM , measurement none of the determinants were above 50. In mathematical l terms a determinant of zero defines a singular matrix. As the t determinant nears zero, the calculation of the matrix becomes.less i physically meaningful. The low determinant for the BOC SAM measurement { . thus signifies that the correlation between the peripheral power '! integrals and excore signals is very weak and the SAM generated is not i able to calculate reasonable axial shapes when the family of axial : j shapes in the core are different from those used to calculate the SAM. [ l 4.2. MOC to EOC Axial Shape Variations Investigation into the axial shape variation during cycle 6 revealed two specific characteristics of the CPC axial shape synthesis that could identify , a problem with the measured SAMs. These two characteristics are as follows: ; (a) Significant Increase in Axial Shape kMS Error with Depletion l,
- The axial shape RMS error is determined by comparing the CPC channel l synthesized axial shape and the incore based axial shape as synthesized ;
through CECOR, on a node by node basis. After deleting end node l 1 effects, a RMS of the errors is performed to detemine a measure of the ; i agreement between the shapes. Figure 1 shows the change in axial shape j i RMS error with cycle burnup for units 2 and 3 cycle 6 for CPC Channel D. l ) Channel D was chosen as a representative channel. l For unit 2 cycle 6, Figure 1 shows the axial shape RMS error varying . from approximately 3.5% at BOC to approximately 12% at 16 GWD/T. At the ] End of Cycle (EOC) this axial shape RMS error change could have reached i j 15%. For unit 3 cycle 6, Figure 1 shows the axial shape RMS error ! varying from approximately 3% at BOC to approximately 5% at 10 GWD/T. i At EOC this axial shape RMS error change could have reached 10%.
--Page 12 of 79--
_ __ . - - , ._. -r . -
Typically the axial shape RMS error changes by approximately 3% from the BOC value throughout the cycle (see Section 5.9 and Figures 12 and 13 for units 2 and 3 cycle 3). A larger axial shape RMS error during
, the cycle may signify that the CPC synthesized axial shape with the measured SAM is not adequately representing the axial shape change
- occurring in the core. Note that a high axial shape RMS error could be , caused by other factors in the CPC axial shape synthesis process, such
- as equipment failure, and is not necessarily due to the SAM.
(b) Cosine Type Shapes Instead of Predicted Saddle shapes ; During normal cycle depletion the true axial power shapes in the core , ! change from the " cosine" type shape to a " flat" type shape and i subsequently to a " saddle" type shape. If the measured SAM is based on : 4 a good physical correlation between the peripheral part of the core and j l the excore detector signals, the CPC synthesized axial shapes will track ; this evolution of type of shapes through the cycle. Figures 2 and 3 l show the axial power shape synthesized by the CPC channels in cycle 6 , for units 2 and 3 respectively. In all the channels for both units, the ; CPCs continued to synthesize " cosine" type shapes instead of the " flat" l or "" saddle" shapes predicted by CECOR. This signifies that the CPCs ! l are not tracking the evolution in type of axial shapes through the j depletion from M0C to EOC. Since this occurs in all CPC channels, the j cause of this phenomena is consistent with a problem with the measured : SAMs. Note that as in the case of the axial shape RMS error discussed j above, this non-evolution in the type of shapes in CPC could also be l l caused by other factors in the CPC axial shape synthesis process, such l t as equipment failure, and is not necessarily due to the SAM. I 1 i i t-l } ! 1 i
?
l l
--Page 13 of 79--
1 i 5.0, POSSIBLE ROOT CAUSES INVESTIGATED j This section describes the process used by SCE, with assistance from ABB-CE, j to investigate the root cause for the CPC axial shape anomaly. The initial l part of the investigation focussed on cycle 6 for units 2 and 3. Thereafter, i the focus changed to the previous cycles for each unit to determine the cycle j at which the problem first occurred. The root cause investigation was l performed by a process of elimination starting from the more obvious and , likely causes of the anomaly to the more subtle causes. The basic root cause i investigation focussed on the following major areas , ' i 5.1. Excore Detector Errors ! 5.2. Excore Signal Variability j 5.3. Signal Precision : 5.4. Relationship between Signals i 5.5. Effect of Number of Cases
- 5.6. CECOR Peripheral Power Calculation j 5.7. Possibility of CEFAST Code Problems :
5.8. ROCS Modeling Effects ! 5.9. Comparisons for Past Cycle Operation ! 5.10. New Plant Computer Deadband Effect 5.11. Fuel Management Effects . I Each of these areas will be addressed separately in the remaining part of this : j discussion. , 1 5.1. Excore Detector Errors ! i One potential cause of the " cosine" type axial shapes in the later part-of the ; cycle is errors in the excore detector processing or readings. There were i three major areas investigated as part of the root cause under this topic I (a) Possible Miscalibration of Excore Detectors ! The CPC neutron flux power calibration is checked on a shiftly basis. ! If the neutron flux power is not within the acceptance criteria, the , excere gains are adjusted as appropriate. Typically these adjustments ! are performed based on the total detector signal and do not impact the relative axial level reiationship between the top, middle and bottom i excores. However, to verify that the relative excore signals were : consistent with the installed SAM, the excore drawers were checked to ! l ensure that the excore signal gain adjustments for each of the three _ ! levels remained unchanged from the BOC values at the time the CPC shape : anomaly was discovered. The excore signal adjustments were verified to l ~ be performing according to the instrumentation and control procedures , and showed no generic problem that could affect all four channels for i both units. Typically, prior to the startup of each cycle, excore signal gain ; settings are adjusted so that no linear subchannel calibration is ; required to be performed at power. This adjustment was made for both i I1 l 1 --Page 14 of 79-- r ? , l
- i i !
i t
. . - - - -- . . = - .
units prior to cycle 6. During the collection of the data for the fast powe. ascension, the relative excore signals are compared to the' I predicted excore signals from CECOR. If the signals do not meet the respective criteri3, the linear subchannel calibration can be re- l perfomed and appropriate adjustments made to the SAM constants prior to l their implementation. Table 3 shows the maximum excore detector errors during the BOC FPA SAM measurement for cycle 6 for both units. In both l unit 2 and unit 3, the linear subchannel calibration met its respective l criteria thus eliminating the linear subchsnnel calibration as a cause of the anomalous CPC axial shapes. . , (b) Unrecognized Detector Problem ' Since the excore detector signal is a key ingredient in the CPC axial shape synthesis, one possible root cause of the CPC ' axial shape anomaly is that the signals were not changing sufficiently, particularly in the middle of the core, to indicate a transition from the cosine type axial l shape to the saddle type shape. This could have been caused by some { unrecognized problem with the excore detectors. Table 4 shows the representative change in measured excore detector signals as a function l i of burnup for cycle 6 for both units. These signals were processed -t through the BOC measured SAMs and the M0C neasured SAMs. For the MOC measured SAMs the signal changes were sufficient to facilitate the expected CPC synthesized evolution in shapes through the cycle while the BOC measured SAMs continued to synthesize cosine type shapes thr w h the cycle. Therefore, excore detector performance was not a cause oT the , CPC axial shape anomaly. i - (c) Artificial Adjustment of Excore Signals J Another avenue that was investigated to determine whether an excore detector problem during the cycle could have caused the CPC axial shape anomaly was artificially decreasing the normalized middle excore signal and consequently increasing the normalized top and bottom signals that are used in the CPC channels. For the unit 2 cycle 616 GWD/T cycle ~, burnup case, the normalized middle excore signal was reduced by an additional 15's and the top and bottom excores were each increased by 8's t along with slight nomalization adjustments. The purpose of this-exercise was to simulate whether an obvious deep saddle shape seen by l; the excores would produce a CPC synthesized saddle shape. The results listed in Table 4 and shown pictorially in Figure 4 demonstrate that ' even for this drastic change in excore signals, the CPCs using the BOC measured SAM would continue to synthesize a cosine type shape. , For these reasons it was' concluded that the anomalous CPC axial shapes were
> not caused by excore detector errors. i l
i 5.2. Excore Signal Variability i Fast Power Ascension and xenon oscillation data from all currently operating CPC/COLSS units were examined for signal variability. The examination was _ carried out using a data smoothing method callet[ *P
] ^ !
--Page 15 of 79.-
- Proprietary !
I e a ! i ,
~
Cubic Spline Smoothing . 5 This technique uses cubic splines to fit the data for each detector segment with the least RMS error while simultaneously the " roughness" in the fit (i.e the mean squared second derivative minimizing i[ *P _j The method provides " smoothed" data and RMS estimates for the ~~gnal si variabilitj [ *p The CEFAST history files used in the examination contained " raw" ex-core signals (unnormalized) and normalized third-core peripheral power integrals. The ex-core data was examined both before and after Because therethe normalization were substantial which is
~
performed prior to the SAM determination. differences in the variability between normalized and raw ex-core signals. selected sets of data were examined for correlated " noise" signatures. ;[
*P
_ _ __ 3 _ ___ Table 5 summarizes the signal variability data that was reviewed. Significant results from this examination are as follows: a) The raw ex-core variability for the BOC FPA data for SONGS Unit 2 Cycle 5, Unit 2 Cycle 6, and Unit 3 Cycle 5 was significantly higher (by a f actor of 4-8) than that for all other data considered. This increased variability could diminish the accuracy of the SAM calculation during the FPA. b) Raw ex-core signal variability data that was examined for cross-correlation between detector segment signals showed very high levels of positive correlation \[ *p 3etween all levels and all ex-core channels. This correlation te]nds to decrease the impact of the signal variability on the SAM measurement since normalization of the ex-core signals tends to eliminate correlated variations. c) As a consequence of (b), the normalized ex-core signals used'in the SAM calculation showed significantly smaller levels of signal variability than the raw signals d xcept for the SONGS data, valuesaretypicallylessthan{+pgTheSONGSUnit2 Cycle 6BOC FPA data examined resulted in signal variability levels of roughly
*P ]for_ nonnalized ex-core data. Although this value is higher
[hanthel[*P]!typicallyusedfortheotherCEdigitalplants,it t l
--Page 16 of 79--
- Proprietary f
't is consistent with the noise level used for the SONGS uncertainty analysis.- It is worth noting that, while increased noise tends to , diminish the accuracy of the SAM determination for a given set of l data, there is a strong saturation effect which limits the impact i of the noise. The difference in the CPC uncertainty analysis 4 l between using [+p ] noise and[ *p ] noise is typically minimal. i Based on these analyses, the following conclusion towards the root cause of the CPC axial shape anomaly was developed: The excore signal variability for SONGS during the BOC SAM measurement ' was higher than that observed for the other CE digital protection plants. However, this did not cause the anomaly. 5.3. Signal Precision . , , I The BOC FPA SAM simulation SCE attem performed d to duplicate thisby ABB-C s in the CPC overall uncertainty analysis.The . FPA simulation using simulation using a comp . SIMULATE-3 methodology determined the calculated use in the SAM calculation. Table 6 shows the results of this simulation. In reviewing the normalized peripheral power integral and normalized excore detector signal response, the peripheral detector signal was Therefore, observed to be a check flat and unchanging at 3 decimal place precision. calculation with SIMULATE-3 was performed in which the data precis ! changed to 4 decimal places. The SAM calculation performed with for CPC channel D of SONGS Unit 3 Cycle 6.this 4 decimal place p characteristics as the MOC SAM. " These results indicate that normalized.one-third peripheral power integral al normalized excore detector signal precision ofdata The startup at least 4 decimal was thus examinedplaces to is !
.important to the BOC SAM calculation. It was ;
determine the amount of precision used in the BOC SAM calculation. , confirmed that 4 place precision in the normalized per , SAM for the BOC Fast Power Ascension SAM measurement. , l Based on this analysis signal precision was eliminated as a possible cause of the " poor" quality SAM measured and consequently ruled out as a possible : factor in the root cause of the CPC axial shape anomaly. .- 1
--Page 17 of 79--
.~ '
- Proprietary
1 5.4. Relationship between Signals Given the need for an identifiable pattern of change to the excere signals and i the peripheral integrals during the FPA, as described in sections 5.2 and 5.3, , a detailed review of the cycle 6 measured SAMs was performed. This review showed significant weakness in the relationship between the normalized middle excore signals and normalized peripheral power integrals for the BOC SAM d measurement relative to the MOC SAM measurement (see elements S 23 and S 22 in Tables 1 and 2). Figures 5 and 6 show the change in normalized middle excore ;
; detector signal during the BOC SAM measurement for units 2 and 3 respectively. !
Figures 7 and 8 show the same change in normalized middle excore detector j signal during the MOC SAM measurement for units 2 and 3 respectively. ] The BOC SAM measurement shows no identifiable pattern between the normalized ! . middle excore detector signal and normalized middle peripheral power integral. l Both the peripheral power integral and normalized excore signal show random ; ] variations in the signal instead of a definite pattern as expected for this ; i test. Although the specific relationship between the normalized middle excore 3 detector signal and the normalized middle peripheral power integral for unit 2 . ] is different than that from unit 3 for the MOC SAM measurements, an , identi'iable pattern exists between the middle excore signals and middle l
! peripheral power integrals. The randomness that existed in both the i
normalized middle peripheral power integrals and excore signals for the BOC
' SAM measurement was significantly reduced for the MOC SAM measurement.
I l From this study several conclusions regarding root cause can be developed as ; follows. , l (a) In order to have an acceptable SAM measurement, a coherent relationship l' must exist between the normalized middle excore detector signal and the normalized middle peripheral power integral. l ,~ (b) One cause of the inadequacy in BOC SAM measurements is the apparent ! randemness in the normalized peripheral power integrals during the power ! ascension. (c) The relative change in the middle of the core is not sufficiently large to overpower the cause of the randomness effect in the normalized middle j peripheral power integral and provide a coherent pattern for the BOC FPA l 1 SAM measurement. l i f t ) 5.5. Effect of Number of Cases ; i l During BOC FPA SAM measurement at SONGS, data snapshots are recorded at 1 approximately 30 minute intervals. Two sets of test data were used to j evaluate the possible impact of the frequency of data taking on the CPC power shape anomaly. These were unit 2 BOC SAM measurement FPA data (27 data points) which resulted in a " poor" SAM and the MOC SAM measurement data on
--Page 18 of 79--
#-v --- .,- , ,
l i l Unit 3 (107 data points) which resulted in a " good" SAM. For each . set, the ' data from CPC Channel D was used as a whole and was divided into sub-sets ; consisting of the odd numbered cases and the even numbered cases. The Unit 3 j data was additionally divided into 3 sub-sets of about one third of the cases each, a subset of about one fifth of the data and a subset of about one seventh of the data. All sub-sets of data were used to generate SAMs which ! J were compared to the SAM calculated using all the data to estimate the effect i of the quality of the SAM as a function of the number of data points. Tables ; l 7 ar.d 8 summarize the results of this study. ; Table 7 shows the results for Unit 2 BOC 6 CPC channel D. The effect of using i half the data (either the " odd" or the "even" data set) instead of all the data was generally small even though the number of cases in each of the " half < data cases was only 13 to 14 points. The ends of the main diagonal (Sn and ! Sn) showed some effects while the center element (Szz), which was suspected of being a significant cause of the CPC anomalous power shapes, showed little ; variation among the cases presented. This indicates that reducing the number . of data points in a set of data that produces a " poor" SAM does not make the SAM significantly worse. l Table 8 shows the results for the mid-cycle 6 oscillation on Unit 3. Again, the effect of using half the data instead of all the data was generally small 1
- but not insignificant (slight variations were seen depending on whether the
" odd" or "even" set of cases were chosen). The ends of the main diagonal (S u I
j and Sq ) showed some effect while the center element (5,,) showed little ! variation. The one-third data cases showed similar characteristics to the i half-data cases. That is, there were relatively modest effects on the S j 2 ! element and somewhat larger effects on the ends of the main diagonal. Tfie ; l one-fifth and one-seventh data cases continue this trend. The effects on the 5 22 element are larger than for the cases with more data sets. j The results presented in Table 8 show that, even when the number of cases used !'
- to generate a SAM was reduced to values similar to those in Table 7, the '
general character of the generated SAM did not' change. That is, by itself, the number of cases will not change a " good" SAM measurement to a_" poor" one ; , or vice versa. In theory, in a noise free environment, 3 data sets are enough ! to give a good SAM if they are sufficiently different in shape. That is, if ) data were measured for shapes with widely spaced ASI values (e.g. -0.2, 0.0, and +0.2), a good SAM can be expected. From this study the following conclusion can be made regarding the root cause ; of the CPC axial shape anomaly: 3 ! The degradation in the quality of the SAM generated during the BOC SAM l measurement that resulted in the CPC axial shape anomaly was not caused i by the number of cases used. I
--Page 19 of 79--
i a a
5.6. CECOR Peripheral Power Calculation CECOR infers full core power distributions from a finite set of in-core I detector signals. This process uses precalculated coefficients to relate the detector signals to the assembly power over the length of the detectors, and to relate the power in uninstrumented assemblies to power in instrumented l assemblies. These coefficients are determined from design codes under full l power operating conditions only, yet during the fast power ascension they are i used at various power levels. Errors in the peripheral axial power l distributions resulting from operation under changing power levels are very r small for the following reasons. ; a) The purpose of the CECOR runs in the SAM measurement is to : determine the peripheral power in three axial zones so that they ! can be compared to the excore signals. The determination of the ! peripheral power involves the use of coupling coefficients generated at full power. An evaluation of calculated power ! distributions during the power ascension indicated that they are ; very coupled radially and that coupling coefficients of peripheral ' assemblies are a very weak function of the power level. b) Another finding is that the relative variation of assembly-wise ! axial shape index is not power dependent, although the shape indices themselves are power dependent. This means that the ; change in axial shape with power level is approximately the same in all assemblies, and that the very weak power dependence of the coupling coefficients is approximately the same at all levels. These conclusions validate the use of CECOR coefficients at reduced power l levels for the purpose of determining peripheral axial power distributions. ; 5.7. Pcasibility of CEFAST Code Problems ! In the early stages of the evaluation of the CPC axial shape anomaly, the question was raised as to whether there were any differences between the SCE , and ABB-CE versions of the CEFAST code which could be the cause of any part of ' the effect. Comparisons of SCE and ABB-CE CEFAST runs of the same data set ; with the same selection of cases used gave identical answers. In addition, the same version of CEFAST has been used successfully by other ABB-CE digital plant utilities. l For these reasons it is concluded that CEFAST code differences was not a l factor in the cause of the CPC axial shape anomaly. 5.8. ROCS Modeling Effects Site data collected during the MOC SAM measurement power reduction induced j oscillation indicated that the core is more stable to axial oscillations than l is predicted by the ABB-CE design code (ROCS7 ) model. Based on analytical l ROCS simulations, it had been predicted that a wide range of axial shapes j
--Page 20 of 79--
would be obtained, whereas the plant had a more restricted range. This tends to generate a better quality SAM from ROCS than can be measured at startup making it increasingly difficult to assess the impact of the proposed fuel management changes on the FPA measurement. Three factors that potentially could affect axial stability have been evaluated as follows:
.._______.'[
+P 3
Modified ROCS models which included a larger [ *P ] pere generated and found to give a better match to the measured axial shape cnanges than the design models during the MOC SAM measurements at both SONGS units. Figure 9 illustrates the improvement obtained in the prediction of the SONGS Unit 2 Cycle 6 mid-cycle oscillation. The measured and calculated amplitudes are in i good agreement. The difference in phase is not significant because a very small change in the initial conditions could have reversed the phase. The l modified models are expected to provide a better simulation of the fast power i ascension and will be included in future design models and reflected in the < CPC overall uncertainty analysis constants, as appropriate. Based on this evaluation, it is concluded that the ROCS modeling does not 4 significantly ' contribute to the root cause of the CPC axial shape anomaly but , 1
--Page 21 of 79--
J
- Proprietary i
would be beneficial to better design modeling of the BOC SAM measurement in , the CPC overall uncertainty analysis and for assessing the impact of fuel i management changes on the FPA program. l l
)
l 5.9. Comparisons for Past Cycle Operation One important aspect of the root cause investigation focused on a comparison of the BOC measured SAM for cycle 6 with previous cycle calculated SAMs. The purpose of this investigation was .to identify the time in plant operation that ' the problem first occurred. The previous cycle data was analyzed to determine which cycle exhibited the same symptoms that were discovered in cycle 6. The ! past cycle analysis used the CPC axial shape anomaly symptom description discussed in Section 4.0.
)
(a) B0C SAM Measurement Tables 9, 10, 11 and 12 show the results of the BOC SAM measurements for channels A, B, C and D respectively for unit 2 cycles 1 through 6 and also the M0C SAM measurements for comparison. Tables 13, 14, 15 and 16 ! r show the same results for unit 3 cycles 1 through 6 and the MOC SAM ! ! measurements. For the BOC SAM measurement the symptoms of the problem, ' as discussed in Section 4.0 are low middle detector element, high off-diagonal corner elements, negative values in the inverse matrix elements and low values in the determinant. It should be noted that the BOC , measured SAMs for all channels met the startup test acceptance criteria described in Section 3.2. Table 17 shows a list of qualitative criteria, derived from Tables 9 ! through 16, to determine whether the BOC measured SAM is exhibiting the ! characteristics of this problem. Using this qualitative criteria, !
, Tables 18 and 19 were generated to show a summary of the assessment of I the quality of the BOC measured SAMs for the various cycles (the MOC SAM i data for cycle 6 was also included for comparison). Figures 10 and 11 l pictorially show the change in the determinant of the measured SAMs for the various cycles for enits 2 and 3 respectively. This evaluation i 1 shows that on a generic basis for cycles 1, 2 and 3 for both units a ;
j " good" quality SAM was calculated during the BOC startup whereas for } cycles 4 through 6 a " poor" quality SAM was calculated at the B0C ; startup. Also Tables 18 and 19 show that the MOC SAM for cycle 6 could 'j be characterized as a good quality SAM. i (b) M0C to EOC Axial Shape Variations I As described in Section 4.2 there are two significant characteristics of , the CPC synthesized axial shapes that are symptomatic of this anomalous l behavior. These characteristics are a large monotonically increasing l axial shape RMS error through the cycle and a cosine type axial shape ; synthesized by CPC from MOC to EOC instead of a flat or saddle type - shape. From cycle 3 onwards all the data available was analyzed from the perspective of these two criteria. Cycle 1 and 2 data was not used since the CPC algorithm was changed prior to the cycle 3 startup and the , CEFAST code reflects this new algorithm. Figures 12 and 13 show the l i ! i --Page 22 of 79-- ; l 6
-. - _ )
change in axial shape RMS errors for a representative channel through previous cycles for units 2 and 3 respectively. Tables 20 and 21 summarize the assessment of the whether the CPC axial shape for all channels exhibit the symptoms of this behavior for previous cycles for units 2 and 3 respectively. The results show that the cosine-type CPC . axial shape synthesis from MOC to EOC existed for all channels from ! cycle 4 and for two of the channels in cycle 3 for unit 2. In unit 3 the cosine-type CPC axial shape synthesis from MOC to E0C existed for , all channels from cycle 4 onwards and did not exist in cycle 3. l Table 22 is an overall summary of the BOC SAM measurement, MOC to E0C axial shape variations and the kind of changes made during the past cycles that could be related to the quality of the SAM measurement. Based on the results of these evaluations several conclusions can be made regarding the root cause of the CPC anomaly as follows: (i) From the perspective of the symptoms of the CPC axial shape anomaly discussed in section 4.0, the quality of the BOC SAM deteriorated significantly from cycle 3 to cycle 4 and thereafter became even worse for cycles 5 and 6 for both units. (ii) Since the CPC axial shape anomaly existed from cycle 4 onwards and the Fast Power Ascension program was introduced in cycle 3, the FPA by itself is not the root cause of the anomaly. However, the reduced range of ASI values induced by the FPA relative to the xenon oscillation measurement method can be expected to increase the sensitivity of the SAM measurement to effects causing variations in peripheral power integrals or excore signals. 5.10. New Plant Computer Deadband Effect One change that was made in unit 2 cycle 4 after the SAM measurement and in unit 3 cycle 4 prior to the SAM measurement was that a new plant computer was installed. All subsequent plant startups used this new computer. The next stage of the root cause investigation thus led to any changes that could have been made when installing the new plant computer. Inquiry into the changes in the new plant computer software led to the realization that a deadband on incore detector signal had been introduced with the new computer. The effect of the incore deadband on the SAM simulation was not included in the CPC overall uncertainty analysis and thus not reflected in the uncertainty penalty constants installed in the CPC channels for either unit. The remaining part of the discussion here will describe the working of the incore deadband and its potential effect on the BOC SAM calculation. (a) Operation of the Incore Deadband Incore detector signals are read in for each of 5 detector levels in 56 detector strings in the core. The plant computer reads the detector signals at regular intervals and performs an engineering units conversion. If the change in detector reading is less than the pre-
--Page 23 of 79--
)
1 a
, designated deadband in the software the detector reading is not updated. }
i As soon as the incore detector reading exceeds the deadband the incore !
, reading is updated and the new value is stored until the next update. i ; This mode of processing incore detector signals did not exist in the i previous plant computer. l (b) Potential Impact of the Deadband on the SAM calculation I In the fast Power Ascension program incore and excore signal data is collected during the reload startup power ascension from approximately >
20% power to 68% power. The rate of power increase (average of 3% per i
! hour) and the frequency of snapshots (2 per hour) gave a power change l i' per snapshot of 1.5% per snapshot. The maximum full power detector signal for the incore detector near the periphery of the core ranges from 0.6 to 1.3 volts. The deadband was set at 0.01 volts. Therefore, i j on the average at full power the deadband masked approximately 1% of the !
i signal. During the fast power ascension, the average deadband error was ! j approximately 5% of the signal at 20% power and reduced to approximately I s 1.5% of the signal at 68% power.
- i l Figures 5 and 6 show the normalized middle peripheral power integral and !
- normalized middle excore detector signal changes for the units 2 and 3 i BOC SAM measurement in cycle 6 respectively (also see Section 5.4 for ,
i further discussion of these figures). One important symptom of the B0C : measured SAMs identifiable in these figures is the random variation in f the middle peripheral power integrals during the test. This random !
- variation could have been caused by the incore deadband in this region i l of the core. If the actual change in incore signal between snapshots !
were near 0.01 volts, then there could be a time lag of between a few l i minutes to an hour between the updating of the top and bottom signals l
- and the middle signal. The subsequent normalization process could mask -
4 the change and make the peripheral power integrals appear random in the j middle of the core relative to the excore detector signals, as seen in ; i Figures 5 and 6. This could be a significant contributing factor to the ! I weakness in the BOC SAM measurement. j j (c) Impact of the Deadband in Past Cycles l s The calculation of the peripheral power integrals in CECOR for the SAM l calculation is primarily based on the incore detector signals that are in closest proximity to the excore detector. Coupling coefficients are l then used in CECOR to determine the contributions of the neighboring , assemblies to the excore detector. Figures 14, 15 and 16 show a j
- comparison of the change in middle incore signal closest in proximity to
channel D and the change in middle raw peripheral power for that area ; . for unit 3 cycles 3, 4 and 6 respectively. This study was done on unit ) 3 since all the data was readily available for unit 3. The conclusions ! shculd be the same for unit 2 except as noted below for unit 2 cycle 4.
- In Figures 15 and 16 for unit 3 cycles 4 and 6 it can be seen that the l l change in incore detector signals hover in the 0.01 volt range of the l deadband and there are significant amount of cases where no change is seen from the previous snapshots. In Figure 14 for unit 3 cycle 3 the j middle raw peripheral power closely follows the middle incore detector
--Page 24 of 79--
) g - w -
--r-- - ,
t changes and there are very few cases where no change is seen either in ( the incore or peripheral power. This coupled with the effect of ; normalization makes the peripheral power integral show less of a ! randomness in .the signal change during the FPA for cycle 3 than for , cycles 4 and 6. Figures 17 and 18 show the change in normalized ; peripheral power signal during the fast power ascension for unit 3 l cycles 3, 4, 5 and 6 for CPC channels C and D respectively. These ' trends show that the deadband is producing a random variation in the l middle peripheral power integrals during the FPA for cycles 4 through 6 i while there is almost no random type variation for cycle 3 for both ! units. (d) MOC SAM measurements ! It should be noted that for the Cycle 6 MOC SAM measurements " good" ! quality SAMs were calculated inspite of the deadband being present. The 2 reasons for this were as follows: j i (i) The MOC measurement was perfonned during the power reduction from l 100% power to 85% power where the impact of the deadband relative i to the signal change would be reduced, l (ii) a xenon oscillation and not a fast power ascension test was ! 1 performed thus using a wider range of axial shape changes for the ! measurement, and ! (iii) the type of axial shapes are more akin to a saddle type shape - where the maximum axial powers are in either the top or bottom i regions of the core and not the middle region of the core. ! Therefore, although the MDC SAM measurement produced " good" quality : SAMs, the test was sufficiently different that it provides no specific root cause conclusion as to the impact of the deadband on the BOC SAM measurement. It is however, significant that it is possible to generate ,
- a " good" quH ity SAM for a MOC type xenon oscillation inspite of the ;
incore deadoand in the event that future M0C measurements need to be performed. j . j j (e) Special case of Unit 2 Cycle 4 s Unit 2 cycle 4 BOC SAM measurement was a special case since the new
, plant computer was not ready and the SAM was performed using the COLSS
- backup computer. The incore detector deadband was thus not present in the peripheral power integrals but this was the first cycle using low ,
^ leakage fuel management scheme. From Table 18 it can be seen that the SAM appears marginally better for unit 2 cycle 4 than cycle 5 and cycle i 6 but the M0C axial shape RMS errors and cosine type axial shape still occur for this cycle from MOC to EOC (see Table 20). Also no
- discernible difference in characteristics can be seen from unit 2 cycle
- 4 (i.e. no deadband) to unit 3 cycle 4-(i.e. with deadband) (see Tables
- 19 and 21 for unit 3 data). Therefore, although unit 2 cycle 4 did not l have the deadband there were several variances between the unit 2 cycle
, 4 test and the ether fast power ascensions that could have factored in the deterioration of the SAM calculation as follows: i
--Page 25 of 79-- )
a g --. ~
(i) The excore signal data was taken manually from the CPC console since the new plant computer was in the process of being i italled at this time. A time lag of up to 5 minutes was possible between when the CECOR snapshot was taken and the excore detector signals were read off the CPC consoles. The basis for the timing requirements employed during manual data collection did not anticipate the small shifts in normalized middle peripheral power integral that were characteristic of the low leakage fuel management design. - J (ii) Rods were inserted at low power during the power ascension. The dat3 was thus taken only in the 35% to 65% power range instead of ; the 20% to 68% power range for the other fast power ascension startups. l from this study several conclusions can be made regarding the root cause of the CPC axial shape anomaly: ! (i) The deadband associated with the new plant computer coupled with the f normalization process in the peripheral power integrals can create the , kind of random variation in the nonnalized middle peripheral power - integrals relative to the normalized middle excore signal that could ' significantly contribute to the deterioration of the BOC SAM calculation for low leakage fuel management cores. (ii) The poor quality of the unit 2 cycle 4 BOC measured SAMs, inspite of the ; non-existence of the deadband at the BOC measurement, demonstrates that . the incore deadband may not be the sole cause of the deterioration in ' the quality of the SAM and the CPC axial shape anomaly. 4 (iii) There is no clear SONGS cycle measurement with the low leakage fuel , pattern, without the deadband and with automated excore detector ! readings to constructively conclude whether removal or reduction of the ; deadband by itself will result in " good" quality SAMs at BOC. 1 5.11. Fuel Management Effects i In evaluating the cause of the CPC power shape anomaly, it was thought that i the neutronic characteristics of the SONGS units from cycle 4 onwards might be , contributory. A study was perfonned to determine whether the SONGS units i showed less axial shape variation during power ascension which would impact the generation of an acceptable SAM. Also since the effect was first seen on , 4 a generic basis for cycle 4 and a low leakage fuel management pattern was : first introduced in cycle 4, the possibility that the low leakage design was ! having a negative impact on the ability to determine the excore detector . signals was evaluated. (a) Near zero ITC at BOC For SONGS-2/3, the Isothermal Temperature Coefficient (ITC) is predicted i to be essentially zero at Hot Zero Power (HZP), B0C, no xenon i conditions. During the initial power ascension from 20 to 70 percent ) i ] --Page 26 of 79-- l i I
k i power the ITC, originally near zero, becomes slightly negative. Table 23 shows representative values of the ITC at hot zero power in SONGS ' when compared to that of other ABB-CE digital plant cores. While near zero ITC will reduce the change in axial power distribution during the ; FPA, the SONGS cores are not significantly different from other digital , plants to make this a root cause. Therefore, it is concluded that the near zero ITC at BOC, by itself, did not cause the anomaly. (b) Relatively flat axial shape at BOC ) i The axial power shape, in its index value (ASI) and peak (Fz), also ! shows small variations over the initial power ascension. The ASI ranges , from -0.18 to -0.11 (slightly top-peaked to a flatten cosine shape) i during the initial power ascension of 20 to 70 percent power, and the ; maximum axial power peak decreases from 124 to 1.18 over the same power ! history. The predicted variation, while being flatter in the central L 50% of core height, is similar to other cores of comparable C-E NSSS design. Table 23 shows the BOC axial peaks for SONGS along with the location of the peaks at SONGS when compared to other ABB-CE digital plant cores. The axial peak at SONGS was the lowest and the peak was more towards the middle of the core in SONGS than the other CE digital ! plants. During the FPA this could have contributed to the reduced axial I power shift in the middle core periphery. l (c) Reduced Radial Leakage Since cycle 4, the SONGS-2/3 cores have employed a low leakage fuel l management strategy where the fresh fuel is located predominately in the core interior in a checkerboard loading with burned fuel. Cores that , use low leakaga fuel managements have, as a natural consequence of the > loading pattern, reduced excore detector signals. Figure 19 shows the thange in relative ercore signal for SONGS from cycles 1 through 6. The ! reduced radial leakage at SONGS-2/3 Cycle 6, while less than that of prior SONGS cycles, is not as low as currently observed in other reactors (see Table 23} For these reasons, it is concluded that , reduced radial leake ey ^ 1 elf, did not cause the anomaly. l Based on this evaluation, a f .iowing conclusion can be made regarding the f root cause of the CPC axial shape anomaly: (i) The near zero ITC at BOC or the reduced radial leakage individually did not cause the anomaly. (ii) The SONGS fuel management design that resulted both in lower axial peaks - and less shift in the axial power in the middle of the core during the FPA, when coupled with the near zero ITC and reduced radial leakage : could have caused the anomaly. ! t
--Page 27 of 79--
h t 6.0. ROOT CAUSE CONCLUSION The fundamental root cause of the SONGS CPC axial shape anomaly is a lack of coherent relationship between the normalized middle peripheral power integral : and the normalized middle excore detector signal during the Fast Power Ascension Program SAM measurement at BOC. The breakdown in coherent , relationship is concluded to be caused either individually or by a combination ; of the following factors: i a) Plant Computer Deadband -- The axial power changes during the BOC Fast Power Ascension program were of approximately the same magnitude per ! CECOR snapshot as the deadband. Due to the time lag between the incore i detector signal and excore detector signal data, the relationship could ; have been inconsistent between the top, middle and bottom levels of peripheral power integrals and the excore signals. Depending on the i exact time lag, this effect significantly deteriorated any existing correlation between the peripheral power integrals and the excore detector signals and consequently reduced the quality of the SAM i calculated during the B0C measurement. However, the deadband did not affect the MOC SAM measurement since the shift in the normalized incore i signal was large compared to random deadband effects. j l b) Manual Collection of Data -- For unit 2 cycle 4, the excore signal data was taken manually from the CPC console. This process created a time , lag between the taking of the incore signal data and excore signal data and contributed to the deterioration of the SAM for this cycle. c) Fuel Management Effects -- Other factors that individually did not but ! collectively could have contributed to the cause of the loss of coherent relationship in the middle of the core are as follows: 1 (i) Near Zero ITC at B0C -- One characteristic of the reload fuel ; design from cycle 4 onwards is a near zero ITC at beginning of , cycle. Due to this ITC, the axial shape changes during the FPA may not have been sufficiently large to effect a significant t movement of power on the core periphery. - i (ii) Relatively Flat Axial Peak at B0C -- The axial peak for SONGS is ! the lowest of the ABB-CE designed digital plants and the location j of the axial peak is more near the center than for the other pl ants . Both these factors could adversely affect the SAM l calculation. ; (iii) Low Leakage Fuel Management -- The leakage of neutrons detected by i peripheral incores and excores reduced significantly from cycle 3 l to cycle 4, the first low leakage fuel cycle (see Figure 19). The ; SAMs for all cycles for cycle 4 onwards were of poor quality. l I i !' --Page 28 of 79-- ! l I l l i
t 7.0. CORRECTIVE ACTIONS '
+
In order to eliminate or significantly reduce the likelihood of the , reoccurrence of the CPC axial shape anomaly, several corrective actions will l be employed for the unit 2 cycle 7 reload startup SAM measurement. Based on 4 the results of the unit 2 cycle 7 startup measurement and subsequent axial ! ! shape monitoring during the cycle, these corrective actions could be modified - ! for future startups. The corrective actions are as follows: ; ! i a) Reduction or Removal of Plant Computer Deadband i The presence of the incore detector deadband in the plant computer as described in Section 5.10 is concluded to be a significant contributor ( to the CPC axial power shape anomaly. As such, SCE will eliminate the [ effect of this deadband on the SAM measurement by setting it l functionally equivalent to zero for startup testing. Evaluation of the : need for the deadband is in progress. The current value is believed to , have been based on the concern that continuous conversion of digital sensor voltages to engineering units would excessively load the plant - computer possibly interfering with the required timing of various i functions. This is being reviewed to determine if the deadband can be ! eliminated entirely, only during startup, or if a small value needs to jl 4 be retained. l I b) Data Collection for Fast Power Ascension y The manual collection of data in the BOC FPA SAM measurement is identified as a cause of the anomaly for unit 2 cycle 4 as described in ! l Section 5.10. Therefore, SCE will change the BOC SAM measurement ! a procedure such that if the automated data collection process is not ' available, timing of manual data will be designed to eliminate negative ! - time lag effects, for the FPA startup from 20% power to 68% power. l c) ROCS Modeling Modifications : 5 Section 5.8 discusses improvements to the computer code ROCS to better i simulate the fast ponC_ Ascension. ABB-CE will incorporate the discussed changes ing *p !
]in the ROCS model to better j simulate the fast power ascension measurement. This simulation will !
then be incorporated into the CPC overall uncertainty constants as j appropriate. ; d) Modification of SAM Measurement Acceptance- Criteria The FPA startup test methodology described in Section 3.1 and the review and acceptance criteria described in Section 3.2 will continue to be
- used for the acceptability of the measured SAMs at BOC. If the SAM :
passes the acceptance criteria it will be installed and subsequent power l
--Page 29 of 79-- l
)
i ! l i i
- Proprietary !
- . .-- .. - - .- ~_. _ _ - .
a i ascension to full power will continue as in the current methodology. l The current FPA SAM measurement methodology, the review and acceptance > criteria, and the associated uncertainty constants are adequate for plant operation between BOC and near MOC. l j An additional second level of SAM criteria will be used by SCE at BOC to j determine the need for a "later-in-cycle" measurement of the SAM values. i i This new set of criteria is listed in Table 17. If the FPA measured SAM '
! fails this new criteria, the time in cycle for the "later-in-cycle" l measurement will be determined based on the axial shape monitoring ,
performed during the cycle (see below). The CPC overall uncertainty ' 3 analysis will also be confirmed commensurate with the "later-in-cycle" i
- test procedure and criteria and new CPC uncertainty addressable '
- constants may be installed as appropriate.
i e) Monitoring of CPC Axial Shapes , SCE has already commenced routin; monitoring of CPC axial shapes. This monitoring will identify any significant problens with the CPC axial
- shape synthesis mechanism. This program will be continued regardless of !
d whether the SAM acceptance criteria is met at BOC. ) i l :' i e i 1 a ) l, i ) i a i l i i l
--Page 30 of 79--
1 1
,w-t- - -s-- w
-I 4
l 8.0. REFERENCES
- 1. ABB-CE Topical Report #CE-NPSD-369, Rev.0, " Fast Power Ascension Generic Test Guidelines", August 1986.
- 2. ABB-CE Topical Report #CE-NPSD-103-P, Rev 1-P, "CECOR 2.0 General !
Description, Methods and Algorithms", August 1984.
- 3. ABB-CE Topical Report #CE-NPSD-365-P, Rev. 0-P, "CEFAST Users Guide", -
August 1986.
- 4. ABB-CE Topical Report #CEN-283(S)-P, " Statistical combination of Uncertainties, Parts I, II and III", October 1984. ;
E
*p 3
- 6. SCE Topical Report #SCE-9001-A, " Southern California Edison Company PWR Reactor Physics Methodology Using CASM0-3/SIMULhTE-3", September,1992. _
ABB-CE Topical Report #CENPD-266-P-A, "The ROCS & DIT Computer Codes for [ 7. Nuclear Design", April 1983. , i i i 1
--Page 31 of 79--
- Proprietary -
f 1 r----- ,
TABLE 1 COMPARISON OF UNIT 2 CYCLE 6 MEASURED SAMs l BOC SAM MEASUREMENT MOC SAM MEASUREMENT CPC-A CPC-B CPC-C CPC-D CPC-A CPC-B CPC-C CPC-D j S,3 4.6780 4.0550 3.8605 4.1452 5.7953 5.5905 5.7931 5.7657 l Sy 0.3781 1.5933 1.0394 0.4759 -1.3019 -1.4509 -2.0230 -2.1593 Sn -2.2414 -3.1179 -2.3207 -1.8809 -0.9547 -0.7415 0.2440 -0.0319 l ) S,, 0.7345 0.8268 0.2306 1.4017 -3.2790 -3.5809 -3.1843 -2.5836 l 5,2 2.1975 1.6155 2.4402 1.3329 7.8051 8.7851 8.1982 7.3404 . S,3 0.1258 0.8387 0.3935 0.7010 -3.7760 -4.1473 -3.7467 -3.2321
- Sn -2.4126 -1.8818 -1.0911 -2.5469 0.4837 0.9905 0.3912 -0.1821 1
S .,, 0.4244 -0.2088 -0.4796 1.1913 -3.5032 -4.3342 -3.1752 -2.1811 { Su 5.1157 5.2790 4.9272 4.1799 7.7307 7.8887 6.9907 6.2639 !
- ummmm m asammme mmmmmmmmmeammmmmmmmmemummmmmmmme summmmmmmesummmmmmmme emmmmmmmm
! Deter- 38.2219 16.8826 39.5954 6.4689 235.035 246.866 222.709 188.021 . l minant t S - 1, , 0.2927 0.5155 0.3084 0.7322 0.2004 0.2079 0.2039 0.2071
- l
,I S ,,~. -0.0755 -0.4596 -0.1012 -0.6539 0.0571 0.0594 0.0600 0.0723 i l S-2 3 0.1301 0.3775 0.1534 0.4391 0.0526 0.0508 0.0251 0.0384 j S ,, -0.1062 -0.3520 -0.0395 -1.1817 0.1001 0.0978 0.0934 0.0892 l S ,, 0.4846 0.9204 0.4164 1.9379 0.1926 0.1816 0.1814 0.1921 i ! S-2,3 -0.0585 -0.3541 -0.0519 -0.8568 0.1064 0.1047 0.0940 0.0996 ! ] , S-1, 3 0.1469 0.1698 0.0644 0.7829 0.0328 0.0276 0.0310 0.0371 f S - 1,, -0.0758 -0.1274 0.0181 -0.9507 0.0837 0.0923 'O.0790 0.0690 l i a l S-l u 0.2617 0.3100 0.2319 0.7510 0.1743 0.1779 0.1843 0.1954 : I i l
--Page 32 of 79--
r- - , , . ...-, ~ . . _ . . . . _ . . . _ _ _.-_ _ - _ . _ _ _
l TABLE 2 COMPARISON OF UNIT 3 CYCLE 6 MEASURED SAMs ! i ! BOC SAM MEASUREMENT ! MOC SAM MEASUREMENT l PARAMETER 1 CPC-A CPC-B CPC-C CPC-D CPC-A CPC-B CPC-C CPC-D i S,, 3.9798 4.0556 5.2176 4.1682 5.5514 5.1114 4.6116 4.6966 Sg 1.0071 1.0924 -0.7660 0.6639 -1.6550 -0.8004 -0.5801 -0.6956 1
- S,3 -2.4569 -2.6152 -1.5712 -2.3173 -0.5897 -1.1155 -1.0896 -1.0056
) S,, 0.5516 1.1629 1.0872 0.9376 -2.6855 -3.2377 -2.7699 -2.6220 , S,, 2.5648 1.3593 1.1071 2.0212 7.5876 8.3202 8.0286 7.7302 S,, -0.0750 0.8327 1.2521 0.2068 -3.4173 -3.9083 -3.4338 -3.5790 Sy -1.5313 -2.2185 -3.3048 -2.1058 0.1342 1.1263 1.1583 0.9254 S _,,
-0.5719 0.5483 2.6590 0.3149 -2.9326 -4.5198 -4.4486 -4.0346 S,, 5.5318 4.7824 3.3191 5.1106 7.0070 8.0239 7.5234 7.5846 r ; Deter- 44.4631 6.8653 -2.5554 28.7666 205.086 227.807 195.032 1!?.572 ; minant S-2,, --
0.3181 0.8804 -0.1351 0.3568 0.2104 0.2155 0.2314 0.2295 ,
, s j S~'n -0.0937 -0.9698 0.6400 -0.1433 0.0650 0.0503 0.0472 0.0485 1 S-l y 0.1400 0.6503 -0.3054 0.1676 0.0494 0.0545 0.0551 0.0533 l S-3,,~^ -0.0660 -1.0792 3.0314 -0.1817 0.0895 0.0947 0.0865 0.0861 l 1 !
i S-2 ,, 0.4105 1.9801 -4.7449 0.5709 0.1901 0.1856 0.1844 0. " 1 ,
; S-2,3 -0.0238 -0.9349 3.2250 -0.1055 0.1002 0.1035 0.0967 0.1u10 i l S-l y 0.0812 0.5321 -2.5630 0.1582 0.0334 0.0231 0.0155 0.0178 l
! S~ls , 0.0165 -0.6769 4.4385 -0.0942 0.0783 0.0975 0.1017 0.0951 i i i l S-2,3 ~ 0.2171 0.6179 -2.5864 0.2712 0.1837 0.1753 0.1816 0.1791 i j ! 1 l i;
--Page 33 of 79--
d I G i
i TABLE 3 i MAXIMUM EXCORE DETECTOR ERROR
- DURING FPA i
, i ! PLANT BOTTOM DETECTOR MIDDLE DETECTOR TOP DETECTOR OVERALL i & CPC ACCEPTANCE ' CYCLE CHANNEL MAX. ACCEPT. MAX. ACCEPT. MAX. ACCEPT. (YES/NO) . ERROR CRITERIA ERROR CRITERIA ERROR CRITERIA (%) (%) (%) (%) (%) (%) , A 3.5 3.8 3.3 Yes ! SONGS 2 B 2.7 1.8 3.9 Yes I CYCLE C 2.1 6.0 1.2 4.0 2.3 6.0 Yes ! 6
- D 1.6 1.4 1.4 Yes ;
A 1.3 2.3 3.3 Yes SONGS 3 B 2.3 0.8 1.5 Yes ; CYCLE 1.8 6.0 2.8 4.0 5.0 6.0 Yes ! ] C 4 6 ; D 3.7 1.3 4.4 Yes
- Error (%) = [(Da / Dp) - 1]
- 100
! i
, where, !
Da = Actual normalized detector signal O = Predicted normalized detector signal (from CECOR) g I i
- I I 1 W
i l 1 i ) --Page 34 of 79-- i 1
l TABLE 4 NORMALIZED EXCORE DETECTOR SIGNAL CHANGE WITH BURNUP i l PLANT BURNUP NORMALIZED CHANNEL D BOC SAM MOC SAM **
& (GWD/T) EXCORE SIGNAL CYCLE
] BOTTOM MIDDLE TOP AXIAL TYPE OF AXIAL RMS TYPE OF I RMS CPC ERROR
- CPC l ERROR
- SHAPE SHAPE .
3 0 0.2934 0.3939 0.3127 3.6 Cosine 4.1 Cosine ! 3 0.2993 0.3932 0.3075 4.2 Cosine 4.3 Cosine SONGS 6 0.3006 0.3917 0.3077 5.4 Cosine 4.1 Cosine 2 CYCLE 9 0.2995 0.3897 0.3108 6.9 Cosine 3.7 Flat 6 12 0.2995 0.3875 0.3129 8.3 Cosine 3.2 Flat j 16 0.3018 0.3836 0.3146 12.0 Cosine 2.7 Flat mmmmmmmmme-- summmmmmmmmmmme j 0 0.2945 0.3889 0.3166 3.3 Cosine 2.9 Cosine i i SONGS 3 0.2975 0.3883 0.3162 3.6 Cosine 3.0 Cosine l 3 6 0.2969 0.3863 0.3167 4.4 Cosine 2.7 Flat
- 6 9 0.2981 0.3852 0.3167 5.4 Cosine 2.5 Flat !
r 10 0.2977 0.3849 0.3173 5.2 Cosine 2.4 Flat j SONGS i
- 2 16 .3268 .3336 .3396 Cosine * * '
- Saddle !
4 CYCLE
- 6
> ..+
i !
- CPC vs. CECOR synthesized shape f
** Normal evolution of axial shapes and RMS error expected l
*** This data set has artificially induced excore signals with a 15% reduction in the i middle signal and approximately a 8% increase in top and bottom signal in order to i l attempt to force a saddle shape
- **** Axial RMS not possible since the excore signals were artificial for this case '
i ! ^ l 4
--Page 35 of 79--
t I l
- - ~
s 1 l u i P d i
?
1 i ~ 1 t
't v
l h 6 e i 1 k e f I k f I [ *P ] i i i i. l f 6 i t a l I 5 i 4 d f I i. 4 i, r i i. t. t r
--Page 36 of 79-- !
~ f t
- Proprietary U
rr 1
. . , _ . . _ i
l 1 a i TABLE 6 COMPARISON OF MEASURED AND CALCULATED SAMs SONGS Unit 3 Cycle 6 Channel D SAM Elements BOC FPA 3 Place
- 4 Place
- M0C Osc. >
Item Meas. < Meas. SIMULATE-3 SIMULATE-3 Sn 4.1682 4.3269 4.9995 4.6966
- Su 0.6639 -0.2091 -1.2677 -0.6956 Sn -2.3173 -1.1037 -0.3966 -1.0056 S 73 0.9376 .5030 -1.8258 -2.6220 S 77 2.0212 2.2060 5.7293 7.7302 S 73 0.2068 .4972 -1.7534 -3.5790 Sy -2.1058 -1.8299 -0.1737 0.9254 Sy 0.3149 1.0031 -1.4616 -4.0346 Sy 5.1106 3.6065 5.1500 7.5846 Deter- 28.767 27.824 120.943 192.572 minant S-2 3
.357 .268 .223 .230 S-2g .143 .013 .059 .048 l
S-2 3
.168 .084 .037 .053 Sj) .182 .098 .080 .086 l 4
i S-277 .571 .488 .212 .190 ; S13 .106 .097 .078 .101 t S-233 .158 .163 .030 .018 l S"y .094 .142 .062 .095 ] 1 S-233 .271 .347 .218 .179 i )
- This is the number of decimal places of precision used for the excore and peripheral power signals calculated by SIMULATE-3. ;
d i a i e i i J 1 I
--Page 37 of 79-- l
- j. .
4 i i
TABLE 7 DATA SET SIZE IMPACT ON " POOR" SAM SONGS Unit 2 BOC Cycle 6 CPC Channel D Method of SAM Generatica SCE ABB-CE Odd Data Even Data Startup Verified Cases 27 27 14 13 u < S 13 4.15 4.15 3.79 4.69 5, 1 0.48 0.48 1.04 -0.36 l S 13 -1.88 -1.88 -2.25 -1.33 S 22 1.40 1.40 1.39 1.43 S 32 1.33 1.33 1.35 1.29 S z3 0.70 0.70 0.68 0.74 S 31 -2.55 -2.55 -2.18 -3.11 S 32 1.19 1.19 0.61 2.08 . 5 33 4.18 4.18 4.57 3.59 Test Value 5.82 5.82 5.80 6.44
--Page 38 of 79--
l l i TABLE 8 DATA SET SIZE IMPACT ON " GOOD" SAM ; i SONGS Unit 3 MOC Cycle 6 CPC Channel D Method of SAM Generation SCE A8B-CE 2i 1+2i ! Installed Verified Even Odd 31 1+3i 2+31 Si 7i l Cases 107 107 53 55 37 36 35 22 15
)
S 33 4.67 4.67 4.01 5.36 4.39 4.99 4.54 4.36 4.28 , S 32 -0.66 -0.66 0.46 -1.81 -0.18 -1.18 -0.43 -0.13 -0.00 S 13 -1.03 -1.03 -1.76 -0.28 -1.34 -0.70 -1.17 -1.37 -1.45 i S 23 -2.63 -2.63 -2.57 -2.69 -2.94 -0.52 -2.43 -2.14 -1.98 S 22 7.74 7.74 7.65 7.83 8.26 7.55 7.41 6.93 6.67
}
S 23 -3.59 -3.59 -3.53 -3.64 -3.92 -3.46 -3.38 -3.06 -2.89 S 33 0.96 0.96 1.56 0.33 1.55 0.53 0.89 0.78 0.71 ; i 1 S 32 -4.09 -4.09 -5.11 -3.03 -5.08 -3.38 -3.98 -3.79 -3.66 [ 5 33 7.62 7.62 8.29 6.92 8.26 7.16 7.55 7.43 7.34 l Test Value 4.99 4.99 5.40 4.71 5.30 4.70 4.93 4.83 4.77 ; l l . t I ! I i I l i I l i i I i
--Page 39 of 79--
i i i l TABLE 9 I COMPARISON OF MEASURED SAMs (UNIT 2 -- CPC Channel A) l COMPARISON BOC STARTUP l MOC CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6
- Type of Xenon Xenon FPA FPA FPA FPA Xenon test Osc. Osc. Osc.
1 Fuel Load In-Out In-Out In-Out low Low Low Low l Pattern Leakage Leakage Leakage Leakage t New Plant No No No No Yes Yes Yes Computer ; Sn 6.4097 5.9962 4.8843 4.4788 4.2408 4.6780 5.7953 4 S,, -2.6174 -2.0093 -0.0796 0.8202 1.0039 0.3781 -1.3019 Sn 0.1951 -0.4172 -1.9863 -2.5297 -2.5120 -2.2414 -0.9547 i Sy -2.8431 -2.7979 -1.5267 0.1964 1.1893 0.7345 -3.2790 Sn 7.2757 7.3922 6.1294 3.0898 1.9836 2.1975 7.8051 S,3 -2.9387 -3.0086 -2.7418 -0.4498 -0.1121 0.1258 -3.7760 ; l S 33 -0.5666 -0.1682 -0.3577 -1.6753 -2.4301 -2.4126 0.4837 I S ,, -1.6583 -2.3829 -3.0497 -0.9100 0.0124 0.4244 -3.5032 i Sn 5.7436 6.4258 7.7281 5.9795 5.6241 5.1157 7.7307 ! Deter- 191.242 201.395 175.901 67.9273 28.7295 38.2219 235.035 ' minant S- 1, , 0.1930 0.2003 0.2218 0.2660 0.3884 0.2927 0.2004 S- 2, , 0.0769 0.0690 0.0379 -0.0383 -0.1976 -0.0755 0.0571 S- 2,3 0.0328 0.0453 0.0705 0.1096 0.1695 0.1301 0.0526 l ! S-1,, 0.0941 0.0918 0.0726 -0.0062 -0.2233 -0.1062 0.1001 l ll S ,, 0.1931 0.1910 0.2105 0.3319 0.6177 0.4846 0.1926 S-2,3 0.0956 0.0954 0.0934 0.0223 -0.0874 -0.0585 0.1064 S-23 , 0.0462 0.0393 0.0389 0.0736 0.1683 0.1469 0.0328 ! S-1,,-- 0.0633 0.0726 0.0848 0.0398 -0.0867 -0.0758 0.0837 , s ' S-l y 0.2049 0.1922 0.1695 0.2014 0.2512 0.2617 0.1743 l i i
--Page 40 of 79-- j 1
3 1 :
i 'l 1 TABLE 10 . ] COMPARISON OF MEASURED SAMs l (UNIT 2 -- CPC Channel B) i i i i COMPARISON BOC STARTUP l MOC l , PARAMETE'tS i i ] CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6
- J
! Type of Xenon Xenon FPA FPA FPA FPA Xenon j ' test Osc. Osc. Osc. Fuel Load In-Out in-Out 'n-Out Low Low Low- Low i j Pattern Leakage Leakage Leakage Leakage i
! New Plant No No No No Yes Yes Yes !
Computer ! I Sn 5.9241 5.8165 4.9938 3.7403 3.7842 4.0550 5.5905 q Su -1.9559 -1.8152 -0.3572 1.7711 1.3248 1.5933 -1.4509 ! Sy -0.2198 -0.4811 -1.7100 -2.9991 -2.5015 -3.1179 -0.7415 ? Sn -2.5036 -3.0735 -2.2149 -0.9613 0.3856 0.8268 --3.5809 1 1 3 Sn 6.8472 7.9035 7.1327 4.5360 2.5690 1.6155 B.7851 l S,3 -2.6901 -3.3903 -3.1562 -1.0988 -0.0548 0.8387 -4.1473 l ; l S ,, -0.4204 0.2570 0.2211 0.2211 -1.1697 -1.8818 0.9905 i 4 r S,, -1.8912 -3.0882 -3.7755 -3.3071 -0.8937 -0.2088 -4.3342 ! I S ,, 5.9100 6.8714 7.8662 7.0979 5.5563 5.2790 7.8887 + l j Oeter- 176.766 214.642 203.104 111.959 44.4221 16.8826 246.8663 minant l i S-*n 0.2001 0.2042 0.2176 0.2551 0.3202 0.5155 0.2079 : S'2 , 3 0.0677 0.0650 0.0456 -0.0237 -0.1154 -0.4596 0.0594 ! S-2 n 0.0383 0.0464 0.0656 0.1041 0.1430 0.3775 0.0508 ; S-1 n 0.0901 0.0943 0.0823 0.0588 -0.0468 -0.3520 0.0978 I i S'!,, 0.1975 0.1868 0.1953 0.2430 0.4075 0.9204 0.1816 l S'13 , 0.0933 0.0988 0.0963 0.0625 -0.0170 -0.3541 0.1047 ! S'l ,, 0.0431 0.0348 0.0334 0.0194 0.0599 0.1698 0.0276 I l S-2,, 0.0680 0.0815 0.0924 0.1140 0.0412 -0.1274 0.0923 $ S'2 y 0.2018 0.1882 0.1715 0.1667 0.2073 0.3100 0.1779 ; 4 1
--Page 41 of 79--
i l l
TABLE 11 COMPARISON OF MEASURED SAMs (UNIT 2 -- CPC Channel C) COMPARISON BOC STARTUP MOC PARAMETERS CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 Type of Xenon Xenon FPA FPA FPA FPA Xenon l test Osc. Osc. Osc. Fuel Load In-Out In-Out In-Out Low Low Low Low . Pattern Leakage Leakage Leakage Leakage l New Plant No No No No Yes Yes Yes Computer Sn 6.1953 5.4237 4.6317 3.7298 4.2078 3.8605 5.7931 , t Sg -2.6008 -1.5518 -0.0089 1.7712 0.6264 1.0394 -2.0230 l Sn 0.3718 -0.4493 -1.7811 -2.9840 -2.0550 -2.3207 0.2440 l Sn -2.7740 -2.2352 -1.8369 -0.5298 -0.7372 0.2306 -3.1843 - S,3 7.2030 6.2925 5.9713 3.6286 2.4365 2.4402 8.1982 f S,3 -2.9487 -2.1558 -2.0472 -0.3944 -0.1995 0.3935 -3.7467 l Sy -0.4214 -0.1885 0.2052 -0.2000 -1.9450 -1.0911 0.3912 l S,,
-1.6021 -1.7407 -2.9625 -2.3997 -0.0629 -0.4796 -3.1752 l Sy 5.5769 5.6051 6.8283 6.3784 5.2544 4.9272 6.9907 l l
t ) Deter- 178.915 148.588 153.144 82.9611 46.6523 39.5954 222.709 i d minant l t i ! S-2 ,3 0.1981 0.2121 0.2266 0.2676 0.2742 0.3084 0.2039 S~2,7 0.0777 0.0638 0.0349 -0.0499 -0.0678 -0.1012 0.0600 f i S-2 g 0.0279 0.0415 0.0696 0.1221 0.1046 0.1534 0.0251 l i S-2 n 0.0934 0.0871 0.0792 0.0417 0.0913 -0.0395 0.0934 ! l. S- 2,, 0.1940 0.2040 0.2089 0.2796 0.3882 0.4164 0.1814 S ,, 0.0963 0.0854 0.0833 0.0368 0.0505 -0.0519 0.0940 l S-1,,- 0.0418 0.0342 0.0275 0.0241 0.1026 0.0644 0.0310 i 3 ; S-2 g 0.0616 0.0655 0.0896 0.1036 -0.0204 0.0181 0.0790 l l S-2 33 0.2091 0.2063 0.1805 0.1744 0.2297 0.2319 0.1843 I i 4 i ! j --Page 42 of 79-- l a l ! )
1 i i l I TABLE 12 l COMPARISON OF MEASURED SAMs ('JNIT 2 -- CPC Channel D) i l COMPARISON BOC STARTUP MOC PARAMETERS CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 Type of Xenon Xenon FPA FPA FPA FPA Xenon , test Osc. Osc. Osc. Fuel Load in-Out In-Out In-Out Low Low Low Low . Pattern Leakage Leakage Leakage Leakage { New Plant No No No No Yes Yes Yes ! Computer Sn 5.4562 5.3514 4.3697 4.3312 4.0150 4.1452 5.7657 Sn -1.7307 -1.6165 0.0897 1.0482 1.0961 0.4759 -2.1593 ; Sn -0.1129 -0.3298 -1.6859 -2.6844 -2.4226 -1.8809 -0.0319 l Sy -2.5749 -2.3629 0.0035 1.7195 1.3246 1.4017 -2.5836 l i S;; 7.1363 6.9248 4.4432 1.0556 1.5644 1.3329 7.3404 ! I j 5,3 -3.0213 -2.8233 -1.8527 0.7470 0.3405 0.7010 -3.2321 l t Sy 0.1188 0.0115 -1.3732 -3.0507 -2.3396 -2.5469 -0.1821 l t S 37 -2.4056 -2.3083 -1.5329 0.8961 -0.3393 1.1913 -2.1811 ! . Sy 6.1342 6.1531 6.5386 4.9373 5.0822 4.1799 6.2639 Deter- 171.873 167.920 104.489 -4.3943 16.3552 6.4689 188.0205 : minant j Sy 0.2124 0.2149 0.2509 -1.0337 0.4932 0.7322 0.2071 l < i j S-l n 0.0633 0.0638 0.0191 1.7251 -0.2903 -0.6539 0.0723 l j S-2,3 0.0351 0.0408 0.0701 -0.8230 0.2545 0.4391 0.0384 j S-2,, 0.0898 0.0864 0.0241 2.4506 -0.4603 -1.1817 0.0892 [ i S-37 , 0.1948 0.1961 0.2513 -3.0028 0.9011 1.9379 0.1921 i S-1,3 0.0976 0.0946 0.0774 1.7867 -0.2798 -0.8568 0.0996 I S-2 y 0.0311 0.0320 0.0583 -1.0835 0.1963 0.7829 0.0371 i 1 , S-2,, 0.0752 0.0735 0.0629 1.6109 -0.0735 -0.9507 0.0690 l S-l u 0.2006 0.1979 0.1858 -0.6303 0.2953 0.7510 0.1954 i
! --Page 43 of 79--
i I i .i
I l l 6 TABLE 13 ! COMPARISON OF MEASURED SAMs i (UNIT 3 -- CPC Channel A) l 1 COMPARISON BOC STARTUP MOC , CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 Type of Xenon Xenon FPA FPA FPA FPA Xenon i test Osc. Osc. Osc. , Fuel Load In-Out In-Out In-Out Low Low Low Low 4 Pattern Leakage Leakage Leakage Leakage l New Plant No No No Yes Yes Yes Yes ; Computer ! Sn 4.8788 5.0226 5.3813 4.0399 3.1800 3.9798 5.5514 ] Su -0.5671 -0.6402 -0.8993 0.8179 1.7330 1.0071 -1.6550 [ Sn -1.2374 -1.2238 -1.3052 -2.1877 -2.5250 -2.4569 -0.5897 Sy -1.7644 -2.4415 -2.2300 -0.0955 1.7663 0.5516 -2.6855 j Sn 5.8834 6.8927 6.6906 3.9707 1.5463 2.5648 7.5876 ] , ) Sn -2.1011 -2.7498 -2.6168 -1.3124 -0.1401 -0.0750 -3.4173 f l S ,,- -0.1144 0.4189 -0.1513 -0.9445 -1.9463 -1.5313 0.1342 i 4 ! S ,, -2.3163 -3.2525 -2.7913 -1.7886 -0.2793 -0.5719 -2.9326 ! S 33 6.3386 6.9736 6.9221 6.5001 5.6651 5.5318 7.0070 , Oeter- 145.830 180.153 186.234 87.7296 4.5101 44.4631 205.086 ; minant l i S-2,, 0.2224 0.2172 0.2095 0.2674 1.9336 0.3181 0.2104 { S - 1, , 0.0443 0.0469 0.0530 -0.0160 -2.0204 -0.0937 0.0650 f Ii S-l,, 0.0581 0.0566 0.0595 0.0868 0.8119 0.1400 0.0494 [ l S-2,, 0.0783 0.0881 0.0850 0.0212 -2.1582 -0.0660 0.0895 f S-1,, 0.2111 0.1973 0.1990 0.2758 2.9047 0.4105 0.1901 f I S-1,3 0.0853 0.0932 0.0912 0.0628 -0.8901 -0.0238 0.1002 ; S ,, 0.0326 0.0281 0.0389 0.0447 0.5579 0.0812 0.0334 S-2,, 0.0779 0.0892 0.0814 0.0736 -0.5509 0.0165 0.0783 j ___ S-1 33 0.1900 , 0.1835 0.1826 0.1837 0.4116 0.2171 0.1837 i i
--Page 44 of 79--
. f
TABLE 14 COMPARISON OF MEASURED SAMs (UNIT 3 -- CPC Channel B) COMPARISON BOC STARTUP MOC CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 Type of Xenon Xenon FPA FPA FPA FPA Xenon I test Osc. Osc. Osc. Fuel Load in-Out In-Out In-Out Low Low Low Low i I Pattern Leakage Leakage Leakage Leakage
- 1 New Plant No No No Yes Yes Yes Yes
- Computer l 1
S,1 5.8811 5.7887 5.7697 4.3915 3.8962 4.0556 5.1114 5 13 -2.0886 -1.8562 -1.2187 0.8557 1.3118 1.0924 -0.8004 Sy -0.1727 -0.3952 -1.2749 -2.5979 -2.6937 -2.6152 -1.1155 l Sy -2.7659 -2.9423 -3.1233 -0.7704 1.5260 1-1629 -3.2377 ! S,, 7.3899 7.6418 7.6429 4.5017 1.3065 !. 3593 8.3202 . Sy -3.1435 -3.2119 -3.0235 -1.2658 0.4468 0.8327 -3.9083 1 Sy -0.1152 0.1536 0.3537 -0.6211 -2.4222 -2.2185 1.1263 ; Sy -2.3013 -2.7856 -3.4243 -2.3574 0.3818 0.5483 -4.5198 , l 6.3162 6.6071 7.2984 6.8637 5.2460 4.7824 8.0239 ! Su I i i Deter- 193.472 202.536 225.437 115.801 4.02:r3 6.8653 227.8073 ! l minant l i S-2 33 0.2039 0.2051 0.2015 0.2411 1.1598 2 0.8804 0.2155 [ S-2 u 0.0702 0.0660 0.0588 0.0022 -1.9644 -0.9698 0.0503 { S-I n 0.0405 0.0443 0.0596 0.0916 ". 0194 0.6503 0.0545 S ,, 0.0922 0.0935 0.0964 0.0525 -2.2568 -1.0792 0.0947 ; S-2,, 0.1919 0.1891 0.1888 0.2464 3.4560 1.9801 0.1856 i 5 S-1,, 0.0980 0.0975 0.0950 0.0653 -1.4529 -0.9349 0.1035 l S-1,3 0.0373 0.0347 0.0355 0.0398 0.9304 0.5321 0.0231 i t Sy 0.0712 0.0782 0.0857 0.0848 -1.1583 -0.6769 0.0975 l l S-2 y 0.1948 0.1914 0.1787 0.1764 0.7669 0.6179 0.1753 l i i !
! --Page 45 of 79-- !
i i 1 TABLE 15 COMPARISON OF MEASURED SAMs . (UNIT 3 -- CPC Channel C) l j i COMPARISON BOC STARTUP MOC i PARAMETERS i i CYCLE I CYCLE 2 CYCLE 3 l CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 r Type of Xenon Xenon FPA FPA FPA FPA Xenon ; test Osc. Osc. Osc. ; 1 In-Out In-Out in-Out t Fuel Load Low Low Low Low i Pattern Leakage Leakage Leakage Leakage l ) New Plant No No No Yes Yes Yes Yes l Computer i i ! S,, 6.1261 6.1777 5.3247 3.2779 4.3938 5.2176 4.6116 ] l S,, -2.7151 -2.8075 -1.0162 1.9245 0.4503 -0.7660 -0.5801 ! S,3 0.3702 0.3556 -1.2797 -2.8512 -2.1476 -1.5712 -1.0896 I Sn -2.5209 -3.0279 -3.0439 -1.0722 -0.8445 1.0872 -2.7699 f
- S, 7.0349 7.8446 7.5053 4.5720 4.3127 1.1071 8.0286 3
f 5,3 -2.9041 -3.3584 -2.6328 -0.9233 -0.8496 1.2521 -3.4338 l t . Sy -0.6053 -0.1497 0.7193 0.7944 -0.5494 -3.3048 1.1583 l S ,, -1.3198 -2.0372 -3.4891 -3.4965 -1.7629 2.6590 -4.4486 ! l Sy 5.5339 6.0027 6.8825 6.7745 5.9971 3.3191 7.5234 Oeter- 175.169 198.806 200.088 103.178 101.264 -2.5554 195.032 minant .1
)
S-l y 0.2004 0.2024 0.2123 0.2689 0.2406 -0.1351 0.2314 - S,, 0.0830 0.0811 0.0573 -0.0297 0.0107 0.6400 0.0472 S-l o 0.0301 0.0334 0.0614 0.1091 0.0877 -0.3054 0.0551 S-1,, 0.0897 0.0940 0.0952 0.0633 0.0546 3.0314 0.0865 j l S-1,, 0.1948 0.1868 0.1878 0.2372 0.2486 -4.7449 0.1844 j S-1,3 0.0962 0.0989 0.0895 0.0590 0.0548 3.2250 0.0967 2 S-2 n 0.0433 0.0369 0.0261 0.0011 0.0381 -2.5630 0.0155 1 S~ly , 0.0555 0.0654 0.0892 0.1259 0.0740 4.4385 0.1017 l l S-333 0.2070 0.2010 0.1843 0.1652 0.1909 -2.5864 0.1816 i i
- --Page 46 of 79--
i l l
l i i TABLE 16
. COMPARISON OF MEASURE 0 SAMs
! (UNIT 3 -- CPC Channel D) 1 ) 5 1 i
- COMPARISON BOC STARTUP MOC ll PARAMETERS j CYCLE 1 CYCLE 2 CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6 CYCLE 6 t
Type of Xenon Xenon FPA FPA FPA FPA Xenon j test Osc. Osc. Osc. ! i Fuel Load In-Out In-Out In-Out Low Low Low Low j Pattern Leakage Leakage Leakage Leakage l 1 New Plant No No No Yes Yes Yes Yes l Computer 4 Sn 5.6770 5.6492 5.3514 3.8160 3.4889 4.1682 4.6966 f.
-2.1531 -2.1016 -1.5891 0.9694 1.4686 -0.6956 S,7 0.6639 S,3 0.1008 0.0075 -0.3989 -2.1972 -2.5241 -2.3173 -1.0056 j Sn -2.5900 -2.7848 -2.6922 0.3350 1.5818 0.9376 -2.6220 ;
l S,," 7.3198 7.6565 7.7285 3.6226 1.6606 2.0212 7.7302 : L S,, -3.2228 -3.3864 -3.5529 -1.2522 -0.1368 0.2068 -3.5790
-0.0870 0.1356 0.3408 -1.1510 -2.0707 -2.1058 S ,,-- 0.9254 l s
- l. S ,,~~
-2.1067 -2.5549 -3.1394 -1.5920 -0.1292 0.3149 -4.0346 l 1 i j Su 6.1220 6.3789 6.9518 6.4464 5.6609 5.1106 7.5846 -!
! Deter- 180.642 190.708 197.688 72.8208 11.8379 28.7666 192.5722 j j minant ( 5 i j S~2,, 0.2094 0.2107 0.2154 0.2933 0.7926 0.3568 0.2295 l l S*2,, 0.0718 0.0702 0.0622 -0.0378 -0.6747 -0.1433 0.0485 l 4 l S-2 n 0.0343 0.0370 0.0442 0.0926 0.3371 0.1676 0.0533 f 4 i
- S-2 n 0.0893 0.0907 0.0885 -0.0099 -0.7325 -0.1817 0.0861 i
- I
! S'2 n 0.1924 0.1890 0.1889 0.3031 1.2269 0.5709 0.1898 j f S-2,3 0.0998 0.1002 0.1016 0.0555 -0.2970 -0.1055 0.1010 l S - 1,, 0.0346 0.0319 0.0294 0.0499 0.2732 0.1582 0.0178 : ! 1 - S-2,, 0.0691 0.0742 0.0822 0.0681 -0.2188 -0.0942 0.0951
- =
g S~l y 0.1992 0.1961 0.1876 0.1854 0.2932 0.2712 0.1791 1 l
- --Page 47 of 79--
i j 1 l 2 I
1 TABLE 17 OUALITATIVE CRITERIA SYMPTOMATIC OF THE CPC AXIAL SHAPE ANOMALY FOR BOC MEASURED SAMs l l i QUALITA11VE CRITERIA
- i PARAMETERS !
GOOD ** MARGINAL BAD l i Middle SAM t 5.5 4.0 to 5.5 s 4.0 ! Element i Inverse Matrix All eleaents Both off-diagonal Any off-diagonal l Elements e 0.0 corner elements corner element ; a -0.03 < -0.03 j and or ; all other any other ; elements a 0.0 element < 0.0 : Off-Diagonal Both elements Both elements Any element ! Corner SAM -2.0 to 1.5 -2.5 to 1.5 s -2.5 l Elements or and a not GOOD 2 1.5 l l Determinant of a 140 100 to 140 s 100 l l SAM > 3 l l These criteria are symptomatic of the type of problem that is the ; j subject of this evaluation and are not actual SAM acceptance criteria , j used at startup. [ ~ The criteria listed in this column will be used during the BOC SAM ! j measurement to determine the need for a later-in-cycle SAM measurement. l 1 : I h 1 1 l a 1 I l l i t a ) I
--Page 48 of 79--
l l i 4 l i l
TABLE 18 00ALITATIVE ACCEPTABILITY OF MEASURED SAMs FOR VARIOUS CYCLES IN UNIT 2 CPC QUALITATIVE BOC STARTUP MOC CHANNEL CRITERIA ; PARAMETERS CYCLE CYCLE CYCLE CYCLE CYCLE CYCLE CYCLE 1 2 3 4 5 6 6 Middle SAM Element G G G B B B G {' Inverse Matrix G G G B B B G Elements A SAM Off-Diagonal G G G B B M G l Corner Elements _ l Determinant of SAM G G G B B B G OVERALL CHANNEL G G G B B B G Middle SAM Element G G G M B B G i Inverse Matrix G G G B B B G Elements i B SAM Off-Diagonal G G G B B B G ' Corner Elements Determinant of SAM G G G H B B G OVERALL CHANNEL G G G M B B G ; 1 Middle SAM Element G G G B B B G l i' Inverse Matrix G G G B B B G Elements j C SAM Off-Diagonal G G G B M M G f Corner Elements i l Determinant of SAM G G G B B B G OVERALL CHANNEL G G G B B B G Middle SAM Element G G M B B B G > 1 1 Inverse Matrix G G G B B B G ) l Elements ; i 0 SAM Off-Diagonal G G G B M B G , Corner Elements ; l Determinant of SAM G G H B B B G f i
- OVERALL CHANNEL G G M B B B G !
! i
- Acceptability legend (see Table 17) --> G = Good; M = Marginal; B = Bad l l .
, 1 d
--Page 49 of 79--
l 4 a
TABLE 19 1 00ALITATIVE ACCEPTABILITY OF MEASURED SAMs j FOR VARIOUS CYCLES IN UNIT 3 i
; BOC STARTUP MOC ;
CPC QUALITATIVE
} CHANNEL CRITERIA i i PARAMETERS CYCLE CYCLE CYCLE CYCLE CYCLE CYCLE CYCLE 1 2 3 4 5 6 6 :
Middle SAM Element G G G B B B G l 4 Inverse Matrix G G G B B B G Elements l A SAM Off-Diagonal G G G M B M G Corner Elements , l Detenninant of SAM G G G B B B G
- OVERALL CHANNEL G G G B B B G j Middle SAM Element ' G G G M B B G I Inverse Matrix G G G G B B G [
l Elements ! i j B SAM Off-Diagonal G G G B ' B B G l Corner Elements
; Determinant of SAM G G G M B B G
{ t j 0/ERALL CHANNEL G G G M B B G ! i Middle SAM Element G G G M M B G ! J Inverse Matrix G G G B G B G !
; Elements ,
C l SAM Off-Diagonal G G G B M B G j Corner Elements j d Determinant of SAM G G G M M B G ! j ! OVERALL CHANNEL G G G B M B G Middle SAM Element G G G B B B G ! Inverse Matrix G G G B B B G ; l Elements j D SAM Off-Diagonal G G G M B M G f Corner Elements i Determinant of SAM G G G B B B G h OVERALL CHANNEL G G G B B B G >
)
- Acceptability legend (see Table 17) --> G = Good; M = Marginal; B = Bad ,
l 4
--Page 50 of 79--
r f a' t
TABLE 20 ACCEPTABILITY
- OF MEASURED SAMs USING MOC TO EOC VARIATION CRITERIA FOR VARIOUS CYCLES IN UNIT 2 l
i BOC SAM MOC SAM i CPC TIME CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6** CYCLE 6** I CHANNEL IN LIFE Axial Type Axial Type Axial Type Axial Type Axial Type RMS of RMS of RMS of RMS of RMS of Error Error Error l 1 Error Shape Error Shape Shape Shape Shape . (%) (%) (%) (%) (%) l t i MOC 7.5 Cosine 6.4 Cosine 4.3 Cosine 6.0 Cosine 3.0 Flat l 11.0 Cosine 12.9 Cosine 10.0 Cosine 10.5 Cosine 2.4 Saddle ! A EOC f Accept- B B B B B B B B G G ability l MOC 3.9 Flat 4.7 Cosine 3.4 Flat 5.5 Cosine 3.1 Flat B EOC 5.0 Saddle 9.0 Cosine 8.4 Cosine 10.1 Cosine 1.8 Saddle Accept-G G B B B B B B G G ability ! MOC 4.2 Fl at 5.0 Cosine 4.4 Cosine 5.2 Cosine 3.0 Flat ! i
- C EOC 5.1 Saddle 8.5 Cosine 9.5 Cosine 9.1 Cosine 1.9 Saddle !
l Accept- G G B B B B B B G G j ability L J j MOC 6.7 Cosine 6.6 Cosine 5.0 Cosine 7.1 Cosine 3.6 Flat ! i D EOC 8.8 Cosine 14.8 Cosine 10.8 Cosine 12.0 Cosine 2.4 Saddle j l Accept- B B B B B B B B G G j ability \ . i Acceptability criteria is as follows: { PARAMETER GOOD (G) MARGINAL (tjl BAD (B) i I Axial RMS 55.5% <7.0% 27.0*5 : i Error MOC & EOC (M0C & EOC) (MOC or EOC) Type of Saddle or Flat Cosine (MOC) & Cosine i . Shape (MDC & EOC) Saddle or Flat (EOC) (MOC & EOC) l t
** Cycle 6 "EOC" data was at approximately 16 GWD/T and not at EOC r
I i 1 d I 1 l
- --Page 51 of 79--
i
l TABLE 21 ACCEPTABILITY
- OF MEASURED SAMs USING MOC TO EOC VARIATION CRITERIA FOR VARIOUS CYCLES IN UNIT 3 BOC SAM M0C SAM CPC TIME CYCLE 3 CYCLE 4 CYCLE 5 CYCLE 6* CYCLE 6*
CHANNEL IN LIFE Axial Type Axial Type Axial' Type Axial Type Axial Type RMS of RMS of RMS of RMS of RMS of Error Shape Error Shape Error Shape Error Shape Error Shape (%) (%) (%) (%) (%) MOC 2.8 Flat 3.4 Cosine 4.0 Cosine 6.1 Cosine 3.4 Flat A EOC 4.2 Saddle 7.7 Flat 10.1 Cosine ** ** ** ** Accept- G G M ** ** ** ** B B B ability MOC 3.2 Flat 4.4 Cosine 4.0 Cosine 5.3 Cosine 2.0 Flat B EOC 4.7 Saddle 8.8 Flat 9.8 Cosine ** ** ** ** Accept- G G B M B B , ability MOC 3.1 Saddle 3.0 Cosine 2.4 Flat 5.6 Cosine 2.1 Flat C EOC 4.8 Saddle 6.9 Saddle 4.8 Saddle ** ** ** ** Accept- G G M M G G ** ** ** ** ability MOC 2.1 Flat 4.7 Cosine 2.9 Flat 5.2 Cosine 2.4 Flat D EOC 3.2 Saddle 9.3 Fl at 8.7 Cosine ** ** ** ** Accept- G G B M B B ability Acceptability criteria is as follows: PARAMETER GOOD (G) MARGINAL (M) BAD (B) Axial RMS 55.5% <7.0% 27.0% Error MOC & EOC (MOC & E0C) (MDC or E0C) Type of Saddle or Flat Cosine (MDC) & Cosine Shape (MOC & EOC) Saddle or Flat (E0C) (MOC & EOC)
** No EOC data for this cycle available. This criteria results are thus inconclusive.
--Page 52 of 79--
TABLE 22 OVERALL QUALITATIVE ACCEPTABILITY ** OF CPC SHAPE SYNTHESIS FOR VARIOUS CYCLES UNIT CPC QUALITATIVE BOC STARTUP MOC CHANNEL PARAMETERS CYCLE CYCLE CYCLE CYLLE CYCLE CYCLE CYCLE ' 1 2 3 4 5 6 6 B0C STARTUP G G G B B B G , M0C-EOC B B B B G ! BOC STARTUP G G G M B B G MOC-EOC G B B B G B0C STARTUP G G G B B B G MOC-EOC G B B B G 2 BOC STARTUP G G H B B B G MOC-E0C B B B B G Overall G G G B B B G Type of Test Xenon Xenon FPA FPA FPA FPA Xenon Fuel Loading In- In- In- Low Low Low Low ALL Pattern out Out Out Leak Leak Leak Leak Computer No No No No Yes Yes Yes Update BOC STARTUP G G G B B B G MOC-E0C G B B BOC STARTUP G G G H B B G MOC-EOC G B B BOC STARTUP G G G B M B G MOC-EOC G H G 3 BOC STARTUP G G G B B B G M0C-EOC G B B Overall G G G B B B G Type of Test Xenon Xenon FPA FPA FPA FPA Xenon fuel Loading In- In- In- Low Low Low Low ALL Pattern out Out Out Leak Leak Leak Leak Computer No No No Yes Yes Yes Yes Update Data not available or no conclusion possible (see Section 5.9(b) ana Tables 20 & 21}
- Acceptability legend --> G = Good; M = Marginal; B = Bad
--Page 53 of 79--
L a TABLE 23 COMPARISON OF NEUTRONICS RELATED PARAMETERS i FOR ABB-CE DIGITAL PLANTS i i + REPRESENTATIVE AXIAL PEAK LOCATION OF FZ . PLANT RADIAL ISOTHERMAL (F,) (% Core height)
& LEAKAGE
- TEMPERATURE 20% =70% 20% =70% l CYCLE (% Ap) COEFFICIENT
- l (x10"Ao/*F) Power Power Power Power SONGS 2 1.8 0 1.240 1.177 72.6 62.4 C)CLE 6
~ Waterford 3 1.2 -0.05 1.194 67.4 Cycle 6 i 1.220 72.6 67.4 Palo Verde 1.0 0 1.316 3 Cycle 4 l AND 2 1.8 -0.10 1.212 72.6 Cycle 10 l
- At BOC, hot zero power conditions
** Data not available -
t a i i e
, l l !
i
- l
--Page 54 of 79--
l
FIGURE 1 : AXIAL RMS ERROR CHANGE STUDY l (CPC-D) j 15 . . . . . 14 - , 13 -
. i 12 - ' % . e' , ' '
o 11 . , m 4 - x 10 - t , W _ m 9 -
- E t 8 - '#
w ~
+++ r k **
y 6 - d,p 4 . ; OM
- a. 5 -
*e _
2
- - e ,,* p -
3-2 . _ s 2 - a . 1 - 0 ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '- 0 1 2 3 4 5 6 7 8 9 1011 1~213141516171819 20 l BURNUP (GWD/MTU) I
- Unit 2 Cycle 6 o Unit 3 Cycle 8 i
i L t F F
--Page 55 of 79-- .
l
FIGURE 2 CORE AVERAGE AXIAL SHAPE COMPARISON UNIT 2 CYCLE 6 (16 GWD/T) 12 s3 e 3 o 3
/
e
-AAAA A m
~ %A 0
w 0.8m18 1^9 % b! a u f 0.7 x
@ 0.6_ i d L 0.5 . .- . . . . . . . . . .......
1 2 3 4 5 6 7 8 9 1011121314151617181920 NODE (Bottom To Top)
-et- CECOR CPC-A Z CPC-B
- CPC-C -M- CPC-D i
J
--Page 56 of 79--
4 i
i l
~
FIGURE 3 i ' i CORE AVERAGE AXIAL SHAPE COMPARISON UNIT 3 CYCLE 6 (10 GWD/T) :, 1.2 r- ' C a g1 O 1
~
_ A A A A A -
[ ! 4 a oe w 0.8 N gf g. i
~3 < 0.Y 1
i 2 1 ? a m I O L 1 z 0.6 4:e i 0.5 . . . . . . . . . . . . . . . . . . . ; 1 2 3 4 5 6 7 8 9 1011121314151617181920 ! NODE (Bottom To Top) !
-de- CECOR CPC-A Z CPC-8 CPC-C CPC-D r
4 4 i l
--Page 57 of 79-- ,
l {
l
)
FIGURE 4 CPC AXIAL SHAPE SYNTHESIS COMPARISON UNIT 2 CYCLE 6 CPC-D (16 GWD/T) 1.8 i 16 1.4 v g'12 y . 0.8 s 06 0.4 _ m 1 2 $4 $$ $ 9 1O 11 1213141516171819 20 NODE (Bottom to Top) BOC SAM (Base) : SOC SAM (15% low) - MOC SAM (Base) -M- MOC SAM (15% low)
--Page 58 of 79--
l
i l l FIGURE 5 EXCORE vs. PERIPHERAL RELATIONSHIP U2C6 BOC FPA (CPC-D) 0.401 -0.399 a a 0.4 C ,
~
z 0.399 W cn 0.398 ] \]\ z
-0.395 ]
c 0.397 $ O O396
\ \k -0.393 $
b 0395 I ' *- = ] O'3942
- M O D
< 0.393 \ -- }= ~ N} Nt -0.389 s 2
C O 0.392 [ 4 3 z 0.387 2 0.391 g 0.39 , , , . , , 0.385 Z 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 ' TIME (Fraction of Days) Z Middle Excore ' Middle Peripheral i
--Page 59 of 79--
l l l
l i FIGURE 6 EXCORE vs. PERIPHERAL RELATIONSHIP USC6 BOC FPA (CPC-D) 0.396 0.392 a a 0.395 cc
< .0.39 0 W
I Z 0.394 i 52 to 0.393 z
-0.388 ] t
$ 0.392 *# $ i
! O 0391- * !\ N \+ -0.386 $ 0 39 *^ -" ^-- - g 0.389 o se z. - vn o
< 0.388 s -0.382 % .
a : C O 0.387 <
~
l 2 -0.38 2 . 0.386 I !' O O.385 0.378 Z l 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 j TIME (Fraction of Days) [ t Z Middle Excore Middle Peripheral -[ i i ! i
--Page 60 of 79-- :
)
i l l 1 l l FIGURE 7 EXCORE vs. PERIPHERAL RELATIONSHIP U2C6 MOC OSCILLATION 0.389 -0.347 a ! a 0.388 C
< -0.345 0 z 0.387 W G z cn 0.386 -0.343 ] $ 0.385 *#'" " A s $ !
b x 0 .384
+'
- +4 -0.341 $
n_ uJ #~ - E" "=c-*- E o 0383-ggm" e pf s -0.339 g N 0.382 gs g d 2 0.381c[ -0.337 % : [ 3 r O 0.38 < z 0.379 -0.335 2 C O 0.378 , , , , , , 0.333 Z 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 TIME (Fraction of Days)
~: Middle Excore '
Middle Peripheral i i l I
--Page 61 of 79--
i
1 l l l 1 FIGURE 8 EXCORE vs. PERIPHERAL RELATIONSHIP ~ U3C6 MOC OSCILLATION 0.389 0.364 a a 0.388 cc z 0'387 -0.362 @ p*+ sa t t co 0.386 ? t
-0.36 g
m 0 385
* *g -
a O 0 .384 N -0.358 w 0 383 b
'E
- 0.382 o
a
< 0.381 \ % -0.354 W 0.38 ! k 2
0.379 1k ny -0.352 2 c - o O.378 Z
. , , , , , 0.35 0 0.2 0.4 0.6 0.8 1 1.2 1.4 TIME (Fraction of Days) 9 Z Middle Excore : Middle Peripheral
- The different response direction of the nomalized middle peripheral integral when compared to the nomalized middle excore signals as depicted in this figure is expected since the middle excore detects neutrons from the bottom and top of the core in addition to the middle of the core.
..Page 62 of 79--
t I
~
FIGURE 9 ! ROCS MODELING COMPARISONS l k SONGS-2 CYCLE 6 (420 EFPD)
' 3 0 500 i 0 400 N !
f 1 0 300 ~ I q ; 0.200 -/ D G 0 100 - n . a M g < l
$ 0 000 I --
/ n 0 100 N '
O - 0.200 !
- 0.300 .,
u O
-0.400 S
-0.500 8 i !
-0.600
- 0. 0T 0.400 0.000 1.200 1 600 2.000 2. 00 2.900 CTexe92)
TIWE IN WIMJIES FW 00 00 FEB.25,1993 , D M + SITE lediflec ;DCS ! I l
. l l
l l
--Page 63 of 79-- l l
i i
_. - ~ t I ' FIGURE 10 ' COMPARISON OF SAM DETERMINANTS : SONGS UNIT 2 ' i } 250 ' r i 225 -F - 2 200 ""E" I
$ 175 V \ 'If 8 330 NN AN //
i )
- 1 \ \\ N // !
5 '25 N N \~ // . I j 77 N AN // E H 50 !! w C 25 \ NN ' E 0 "
-25 . , ,
1 2 3 4 5 6 7 i
- CYCLE i l
-*- CPC-A :
CPC-B CPC-C CPC-D
?
)
I l
--Page 64 of 79-- 1 l
l 1 l l
l 1 i FIGURE 11 , COMPARISON OF SAM DETERMINANTS
. SONGS UNIT 3 250 225 2 2 00 ==*
g y; -gy wasue m 7 i U5150 D // z 125 \\\. //
$ 100 \\ \b- T // - !
5 75 \ \ /// E 50 \ \ /8 b 25 N- d !
- 0 \l
-25 . , ,
1 2 3 4 $ $ 7 CYCLE ;
-*-- CPC-A .
CPC-8 CPC-C CPC-D 8
--Page 65 of 79--
l
l FIGURE 12 AXIAL SHAPE RMS ERROR CHANGE STUDY UNIT 2 15 . 7 14 _ 13 - f' ~ e 12 - g/ g _ 11 g3 m 10 - P**
#~
m / _' ; 9 -
. l 2 / -
e 8 -
,,aa .
5
~
W 7 -
'W S pe**
~
l
& ,+* y ,
W d 4-- e ,+,,
**v . .
3 l- , 2' : 1 - 0 ' - . , , , , , , , , , , , t 0 1 2 3 4 _5 67 8 9 1011121314151617181920 BURNUP (GWD/MTU) ; o CYCLE 3 CPC-C ' v CYCLE 4 CPC-D l c CYCLE 5 CPC-D .
+ CYCLE.6 CPC-D ;
i F
--Page 66 of 79--
i FIGURE 13 1 f AXIAL SHAPE RMS ERROR CHANGE STUDY ; UNIT 3 ! 15 . , 14 - 13 ,- , 4 12 - i x - > O 11 -
$ 10 -
m 9 - 1 : t 8 - t avv' e f
- y !
i t.a o_ 7 - o - g/ y-5 m
's[
h4
< 5 3<- .
8v' l 29 - i 1 - i 0 ' ' ' '~ ' ' ' ' i 0 l 1 2 3 4 5 6 7 8 9 1011121.314151617181920 t BURNUP (GWD/MTU) o CYCLE 3 CPC-D { v CYCLE 4 CPC-D t a CYCLE 5 CPC-D {'
+ CYCLE 6 CPC-D
- i l
i I
--Page 67 of 79- . j i
I
l i i
~
l l i 4 FIGURE 14 :
't INCORE AND PERIPHERAL POWER CORRELATION l U3C3 FPA CHANNEL D !
0.1 0.0025 i 3 0.08 C : o [] 2}' O.002 $i C v e en 0.06 .
~ ~
0.0015 r t E { -- h [ 2!
+ a S u
g ug sune v
- 1
- 0. 5 3 -0.02 f -0.0005 c.
il !
-0.04 1
x ;c -0.001 y
-0.06 , , . .
-0.0015 !
0 0.5 1 1.5 2 2.5 Time (Fraction of Days) !
- Middle incore #37 Peripheral Power l
l I i l
--Page 68 of 79--
FIGURE 15 INCORE AND PERIPHERAL POWER CORRELATION U3C4 FPA CHANNEL D 0.025 0.0012 0.02 -- I r -0.001 i g e ! -0.0008 E a 0.015 3 g c
.c 0 * * " -
- 0. 05 i 0.0 2 0 -. _
-0 ,e
$ t -0.0002c.~h 3 -0.005 i
-0.0004 g
-0.01 #+g/u
-0.0006 $
-0.015 . . . . . . . . . -0.0008 0 0.3 0.6 0.9 Time (Fraction of Days)
- Middle incore #37 Peripheral Power
--Page 69 of 79--
1 1 FIGURE 16 INCORE AND PERIPHERAL POWER CORRELATION U3C6 FPA CHANNEL D 0.04 0.001 0.035 -0.0009 g g 0.03-
-0
- 08 3 j 0.025 1 -
- 7 5
0~02 li - mW p - eg n tf ru 5 lb M l -o w0s ; i J A MA J/ #1#V%l A -o o*4 i f
- ll l Tt 17
~
B IC ~- 2 0.005 6 -
+ K71
+ i
-0.0002 0 1+ - -- - 1 g
1 '
--0.0001 CC -0.005 . .
1.1 1.2 0 1 1.3 1.4 1.5 1.6 1.7 1.8 Time (Fraction of Days)
- Middle incore #37 .
Peripheral Power
--Page 70 of 79--
l l l
l FIGURE 17 ' COMPARISON OF UNIT 3 FPA'S CPC-C MIDDLE PERIPHERALS 0.415 s g 0.41 g
@ 0.405 s.. $ 0.4 i $ 0.395 W
E 0.39 a 4 W
" sM M"x A m m -
W J 0.38 gg dmGlh 0.37 Amm P 7 0.365 0 0.'2 0.'4 0.'6 0.8 1 TIME (Fraction of Days)
. CYCM 3 CYCd 4 - CYCE S ~ CYCG 6
--Page 71 of 79--
i
1 i 1 FIGURE 18 i COMPARISON OF UNIT 3 FPA'S CPC-D MIDDLE PERIPHERALS 0 42 3 g 0.415 % - - 0 m 0.41 , H . z 0.405 5 0.4 w I 0.395 8 '* W n ,~ 0.385 wa,iqg- _-
%._ - z 0.
i ' 0.37 . , . . O 0.2 0.4 0.6 0.8 1 TIME (Fraction of Days) CYCG 3 CYCd 4 CYCd 5 Z CYCW 6
--Page 72 of 79--
l
{ i I
+
FIGURE 19 l l. RELATIVE EXCORE SIGNAL l UNITS 2 & 3 . 0.9 . 0.85 /\ -
-J j 0.8 i
s2 O /
/ -{
i w 0.75 i I C) - 0 0.7 l X w w i
> 0.65 !
i H i fw 0.6 : 1 [ l 0.55 } i
- 0.5 1
2 3 4 5 6 l CYCLE I
- ! ~
j. l I l l
--Page 73 of 79 .
4
i I e t i i i I APPENDIX ; LETTER REGARDING OPERABILITY OF CPC CHANNELS , FOR
- SONGS UNITS 2 AND 3 CYCLES 3. 4. 5 AND 6 ,
r l I i
--Page 74 of 79--
i l
ABB ASEA BAOWN BOVERi April 30, 1993 Mr. Paul D. Myers Southern California Edison 23 Parker Street . Irvine CA 92718
Subject:
SONG 8 Units 2 and 3 CPC Opera.bility for Cycles 3, 4, 5, and 6
Dear Mr. Myers:
At the request of SCE, CE has evaluated the operability of the SONGS Units 2 and 3 CPCs during Cycles 3, 4, 5, and 5. During portions of these cycles, the CPCs were recently discovered to be doing a relatively poor job of representing the actual core cxial shape. In particular, the CPCs were calculating a " cosine" type axial shape throughout the cycle rather than the expected evolution !
-from " flattened cosine" to a " saddle" shape as predicted from MOC through EGC times in life. The as-found condition of the CPCs were evaluated to determine whether or not they met the appropriate criteria for operability during the period in question.
CE has concluded as part of the assessment that the SONGS Units 2 and 3 CPCs were operable during cycles 3, 4, 5, and 6. The overall operation of the SONGS CPCs was deemed conservative and within the current SONGS analysis of record for UFSAR events. Additional details of the evaluation and supporting information for these conclusions are contained in the attachment to this letter. Please note that the information contained in the attachment has been formally Quality Assured by CE. I' f you have questions or require additional information on the SONGS CPC operability evaluation, please_ contact either Mr. Mike Book at (203) 5285-5072 or me. Sincerely, COMBUSTION ENGINEER , INC.
~
R. . Land Pro ct Manager, SCE Nuclear Fuel l ABB Combustion Engineering Nuclear Fuel
--Page 75 of 79--
i SONGS Units 2 and 3 CPC Operability With Unexpected Axial Shapes ' i Observations at both SONGS Unit's 2 and 3 indicated that when the CPCs use I installed Shape Annealing Matrices (SAMs) derived from the BOC Fast Power ' Ascension (FPA) measurements they did a poor job of representing the actual
- core axial shape from approximately the middle to the end of each operating cycle. :
In particular, the CPCs calculated a " cosine" type axial shape !
' throughout the cycle rather than the expected evolution from " flattened -
cosine" to a " saddle" shape as predicted for NOC through EOC times in life. ' This unexpected behaviour has been shown to exist for the latter part of all of the previous operating cycles for both units beginning with Cycle 4, the first 24 month low leakage cycle. There are indications that one or two Unit , 2 CPC channels also showed similar behavior during Cycle 3, the first cycle that used the FPA program. ; These observations have led to the question of whether or not the CPCs were
" operable" during the latter parts of cycles 4, 5, 'and 6 for both units and perhaps Unit 2 Cycle 3. The purpose of this letter is'to answer this -
i
" operability" ~ question by presenting the results of analyses of the basis for the SONGS Units 2 and 3 CPC setpoints and then identifying any potential non- i conservatisms in these setpoints which may have existed during the.past and current cycle operation. (Note that -
the newly measured, non-FPA based SAMs at both Unit 2 near end of Cycle 6, and Unit 3 near middle of 1 Cycle 6, do not exhibit the observed anomalous behaviour when installed in the CPCs.) , An evaluation of the inact of the unexpected axial shape behavior on CPC-related analyses determed that only the calculation of CPC [
*P i
] are impacted by shape annealing i matrix and axial shape variations of this type. Therefore, in order to assess ;
the operability*P of the CPCs when the unexpected axial shape behavior occurred, i CPC\[ ]were perfomed in order to simulate the symptoms ' observed at SONGS. ' l The characteristics and results of CPC [ *P ] already performed ! for SONGS.and those perfomed specifically for this evaluation were compared ! in' order to determine bounding values for potential non-conservatises in the : CP.C calculations. . The [ *P .] to'be evaluated were chosen to cover the range fros standard analysis methodogy to a methodology which was
, created specifically to model the symptoms observed at SONGS. These analyses included: ,
i
- 1. A review of the Analysis of Pecord for SONGS (up to the time at which '
the new measured SAMs were installed durino Cycle 6 of each Unit) - ~the extended cycles progran[ *P ]l that were performed to cover Units 2 and 3 Cycies.3 Inrougn 6 and equilibrium 24 month cycles. 2.
~
{ Analyses of the end of Cycle 6,[ ;
*P '
]lmethodolog uncertainty values [y *Pusing typtcal, excore detector and peripheral power -
~) and increased excore detector and peripheral ,
--Page 76 of 79-- I e
- Proprietary i
i power uncertainty values [ *P 1 The[ *P] excore detector and peripheral power uncertainty vaHue was itffsen since it is the valge used ' in typical analyses based on the current [ *P J methodology. The *P <alue was chosen since it would emphasize the
}[ymotoms s
*P of the un[expec]ted axial shape behavior a
]} which were at one time believed to be a root cause of the unexpected axial shape behavior.
- 3. Analyses af the end of Cycle 6}[ *P I
]l modified to emphasize the symptoms of cosine "axH1 ~p~ower shapes, at end of cycle, also using typical @0.l!.fj and increased [ *P ]excore detector and peripheral power uncertainty values. -
, 4. An EoC T *P ] based on a single simulated SM (correspondingtonoexcor]edetectorandperipheralpoweruncer and a single BOC6 FPA. measured SM. re used for comparison to the AOR and new;[ Thesecasesw)casesinthe
*P comparison of axial shape characteristics.
A qualitative comparison of CPC [ *P ] axial shape characteristics, including axial peak values and locations, was performed for the evaluated cases and determined that the EOC [
*P
- ]yjeld similar CPC liiial shape characteristics, including predominance of center -
peaked (cosine) shapes. These shapes are also similar to the characteristics of the shapes produced when a BOC6 FPA-measured SM ,is used. In addition, the base SM from the specially [ *P ] produces an axial shape which is similar to the axial shape produced when the SOC 6 FPA-measured SAM is used at EOC 100 % power nominal conditions. On the other hand, the base SAM - from the)[ *P -] yields a' saddle-type CPC shape, significantly '
~
different trom tne others. , A review of the Analysis of Record (AOR) determined that the spectrum of shape annealing matrices used to calculate the installed CPC uncertainty factors at the time of the unexpected axial shape behavior would yield axial shapes which are even more center ceaked than the new analyses based onl[
*P ] and the typical measured SM. These axial shapes would als.o be significant1v sore center peaked than the axial shape that is obtained when the[ *P ] base SM (without excore-detector and peripheral power uncertainty) is used. Therefore, it can be judged that the axial shape characteristics and results of the AOR were similar to those for the new analyses.
From this comparison of CPC [ *P ] axial shape characteristics (which is summarized in Tablie 1~below), it can be concluded that the unexpected axial shape symptoms observed at SONGS Units 2 Lnd 3 between middig, , and end of Cycles 3 through 6 are bracketed by the set ofl -
"P J !
run for end of Cycle 6 based ort the [ *P l l J excore detector and peripheral. power uncertainty values. In addition, previous analyses have shown that the
*P j results tend to be most limiting. The results from these){ *P . cases can therefore be used to determine bounding
--Page 77 of 79--
- Proprietary
crc unt.ertainty times ractors (BERRO of the unexpected axial shape through symptoms. BERR4) for these cycles during the' The CPC uncertainty factors for the analysis of record which forced the basis i for the installed factors for Units 2 and 3 Cycles 3 through 6 wer ! withtheEOC6CPCuncertaintyanalysiscalculationsbasedonthefcompared
~
- - IL. - . . _ _ . . - -
l and peripheral power uncertainty values. This ! installed CPC BERR values may have beenindicates non-con.compari_ that the s; [. *P cpera ti on.-
)) between the MOC and EOC portions of previous cycleser!
[ . _ _ _ _ . _ . _ . . - t
*P ;
J !The potential non-conservatisi in[*P]is .iudged._to..b7 ' ' insignificant since there is sufficient margin to,fhej_[ :
*P !
"-~ ]j Therefore, it is reasonable to conclude.that even tfiough the CfCs may have been inaccurate in their axial shape ' calculations for [ *P ] as required during anf CPC design basis event:j they would havebeprovided a trip : Therefore, it can concluded ! that the CPCs were operable during the periods of unexpected axial shape : symptoms. - When COLSS is out of ' service, one of the C'PCs may be used to perforia the ONBR and LHR monitoring fenettons. During_ actual monitorinot.any_ potential non-conservatism in thel [
*P
. service monitoring. -
] !
The evaluation supporting the conclusions of this report has been quality assured to ABB CENF's qua11,ty assurance procedures. i This evaluation did not attempt to determine the reasons for the observed anomolous behaviour of the FPA based SAMs at SONGS. This issue will be
~a ddressed in separate documents. .
. \
--Page 78 of 79--
- Proprietary l
i
.! J 4 .;
, Table 1 i ' Summary of Axial Shape Analyses I
. i l
Type of i[ Nominal 100% Power Predominant ! Analysis ARO.EOC Shape Type- OUA Shades I
-t AOR Saddle - I i Costne i FPA Saddle Cosine d
I
- FPA *P Saddle -
Cosine j {
- 1
. .FPA 4 Cosine Cosine ! FPA Cosine ' Cosine 1 Measured . ] BOC 6 3 Cosine . Cosine FPA
- 9 p
l
~
- ,i 4
6 9 e 4
- 4 9
o - j l
)
l 6 e 0
' ~ '
.--Page 79 of 79-- -
e e #
" Proprietary}}