ML17227A668
| ML17227A668 | |
| Person / Time | |
|---|---|
| Site: | Saint Lucie  |
| Issue date: | 12/17/1992 |
| From: | FLORIDA POWER & LIGHT CO. |
| To: | |
| Shared Package | |
| ML17227A665 | List: |
| References | |
| NUDOCS 9212210397 | |
| Download: ML17227A668 (25) | |
Text
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 1 of 22 ANALYSIS OF INPAX-II POWER PEAKING UNCERTAINTIES FOR ST. LUCIE UNIT 1 CYCLE 11 9212210397 921217
'ADOCK 05000335'DR P PDR
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 2 of 22
1.0 INTRODUCTION
This evaluation presents the results of a power peaking uncertainty analysis of the INPAX-II computer code (used to monitor the core power distribution) for St. Lucie Unit 1 Cycle 11. The analysis of the INPAX-II power peaking uncertainty is based on calculations and operational data for St. Lucie Unit 1 Cycle 11 available up to April 1992. At that time 7 (15%) of 45 detectors strings were failed. The current Technical Specifications allow up to 25% of detector strings to be failed. This analysis was performed to evaluate the INPAX-II uncertainties with an increased number of failed detector strings, up to 50%, at St. Lucie Unit 1, during Cycle 11 operation only.
The current uncertainties in the St. Lucie Unit 1 Technical Specifications are based on the results from the analysis in Reference 6.1. In cycle 6 of St. Lucie Unit 1, Florida Power Light Co. changed fuel vendors to Siemens Power Corp. (SPC),
formerly Exxon Nuclear. To ensure that INPAX-II uncertainties for peaking factors were bounded by the current Technical Specifications and Safety Analysis values, an analysis was performed (Reference 6.2) which addressed uncertainties using the INPAX-II/XTG computer codes. This analysis was submitted to the Commission as part of Amendment No. 63. The results of the evaluation show a smaller uncertainty for Fq, Fr and Fxy than those
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 3 of 22 currently being used in the Safety Analysis and Technical Specifications't that time, the decision was made not to change Technical Specifications and continue to use the more conservative values. The following Table presents a comparison of the uncertainties between Reference 6.2, Reference 6.1, and the currently used values in the Safety Analysis and Technical Specifications.
Peaking Factor Uncertainties Currently Used Reference 6.2 Reference 6.1 Fr 6.0% 4. S~o 6.0%
(Safety Analysis)
Fq 7. O~o 5.0% 6. 2~o (LHR) (Technical Specifications)
The following analysis was performed with the INPAX-II/XTGcomputer code. An extension of the Reference 6 ' calculated measurement uncertainties was performed and new measurement uncertainties were calculated. These uncertainties were compared to the uncertainties currently used in the Technical Specifications and Safety Analysis.
This analysis is only applicable to St. Lucie Unit 1 Cycle 11.
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 4 of 22 2.0
SUMMARY
OF ANALYSIS RESULTS The analysis results show that an additional 0.67% in Fr uncertainty, over that currently in use, is required if the number of failed detector strings reach 50%. For conservatism, this additional uncertainty was increased to 1.5% for Fr resulting in a total Fr uncertainty of 7.5%. Although the LHR uncertainty did not increase over the one currently being used, for conservatism it was increased by 1.0% over the one currently used in the Technical Specifications resulting in a total LHR uncertainty of 8.0%. These uncertainties were calculated based on 50% of detector strings being failed. However, for conservatism, these uncertainties will apply whenever the 25-'o detector string failure level has been exceeded. Figures F 1 and 2.2 present the total Fr and Fq uncertainty as a function of detector string failure level in percent.
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 5 oZ 22 FIGURE 2.1 Fr Peakin Factor Uncertaint FR Uncertainty Versus Percent Fal led Detector Strings 10 00 Q.OO
'LI 8 OO c
~3
+I b
- 7. OO 8
- 8. OO 5 OO 0 10 00 30 10 60 Palled Detector Strings $ 9Q
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 6 of 22 FIGURE 2.2 F Peakin Factor Uncertaint Fq (LHA) Uncertainty Versus Percent Fal led Detector Strings 10 DO
- 9. OD 8 00 c
d V
ID
- 7. 00 8.00 5 00 D 20 30 40 DD Felled Detector Str lngs (Ã)
3.0 RECOMMENDATIONS of the analysis, it is recommended that h
Based on the results operation be allowed for Cycle 11 with an increased number of failed detector strings up to 50% with an increase in the INPAX-II uncertainties as presented in Figures 2.1 and 2 '. The increase in peaking factor uncertainties will, with greater than 25% of detector strings being failed, be accommodated by increasing the LHR uncertainty included in the incore alarm setpoint and by increasing the measured Fr before comparing it to the Technical Specification Limits
ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 7 of 22
- 4. 0 ANALYSZS The fixed incore detector instrumentation pattern for St. Lucie Unit 1 Cycle 11 and the current failed detector strings are presented in Figure 4.1. The detector failures are random with no
-systematic characteristics. Note that all detectors were replaced at the beginning of Cycle 10; all of these detectors have been in the core for one full cycle plus operation to date in cycle 11. All detectors are scheduled to be replaced at the end of cycle 11, early in 1993.
4.1 MEASUREMENT UNCERTAXNTY COMPONENTS Reference 6.2 uses the following equations in the determination of peaking factor uncertainties:
Fq = FF, F, F, Fr=F,.F,F, where F: Relative power associated with the average of the detector segments in an assembly.
F,: Ratio of the assembly relative power to the average relative power of the detector segments in an assembly.
F,: Ratio of the peak planar power in an assembly
ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 8 of 22 to the assembly power.
Fy; Peak local pin power in an assembly.
Note that F can be interpreted - as the extrapolation parameter or coupling factor from an instrumented to a non-instrumented assembly and F, can be viewed as the power to reaction rate for the instrumented assembly.
Using this definition, the relative variance of Fq (LHR) and Fr can be expressed in the following equations:
S'Fq = S'F + S'F, + S'F, + S'F, S Fz' S F + S F + S Fy where S's the relative variance for the random variable and each term is assumed to be independent.
The following Table presents the standard deviation of each of the parameters presented in the previous equations with its respective number of degrees of freedom. These values were obtained from Reference 6.2.
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 9 of 22 TABLE 1 Relative Number Standard Degrees Variable Deviation Freedom F .0240 800 F .0047 514 Fz .0077 514 Fg .0135 188 Fq .0290 1180 Fr .0279 1023 Note that for Fq and Fr the number of degrees of freedom was calculated using the Satterthwaite's formula presented in Reference 6.3.
4.2 EXTRAPOLATION UNCERTAINTY The power extrapolation uncertainty is the only parameter affected by the increase in failed detector strings. Section 4.1 of Reference 6.2 provides the method utilized by SPC to calculate the extrapolation uncertainty. The method employed is to generate measured detector powers which are independent of inferred detector segment power for the same locations. The relative differences between the measured and inferred
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 10 of 22 detector segment power is an estimate of the combined error in the measured and inferred detector segment power and therefore is a conservative estimate of the error in the inferred detector segment power. To generate the required information, first the core power distribution is generated utilizing all of the operable detectors to generate a power distribution which serves to define the measured power distribution for those locations having instruments. Second, the core power distribution is calculated utilizing all operable detectors except for those locations where all four axial detectors (detector string) are assumed to be failed. This second power distribution has inferred detector powers in the assembly location where the detectors were assumed to be failed which are not equal to, and are independent of the measured detector segment power in these locations determined from the first execution for the INPAX-II code. The previous procedure was utilized by Reference 6.2 to calculate the extrapolation uncertainty using a single instrumented assembly location failed at a time The methodology employed in this analysis is similar to that used by Reference 6.2 with the exception that 50% (22 detector strings) of the incore detectors are assumed to be failed.
Specifically, 2 representative flux maps (1 at BOC and 1 at
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 11 of 22 MOC) are executed with different combinations of failed detectors. The 7 detector strings currently failed in the plant were assumed to be failed and the remaining 15 detector strings were chosen at random from the operable detectors.
The standard deviation estimate is. calculated as follows:
STD = ((Z D ND )/ (N 1) )
where N = number of paired data points representing the inferred and measured assembly power F = Measured relative assembly power associated with the average of four detector segments F'= Inferred relative assembly power associated with the average of four detector segments D = Ln (F~ / F')
D~=ZD / N
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 12 o5 22 The standard deviation for each of the BOC and MOC cases is presented in Table 4.1 and Table 4.2, respectively. Note that the scatter for the standard deviation of the BOC cases is larger than expected. A review of the data indicates that the cases where the standard deviations are highest contain several detectors that have low relative power densities such that with a small change in total power, the percent difference with respect to the base case is large.
Specifically, there are seven (7) locations with RPDs lower than 0.400 and one (1) location with RPD lower than .07. In these locations, even small differences in RPD, units cause large percentage differences. This is why the standard deviation for the individual cases appear to have more scatter than expected. This also implies that the standard deviations are not dependent upon the number of operable detector locations, but on which detectors are operable. Since it is not likely that the detectors with the smallest deviations between measured and inferred would fail more often than those detectors with the large deviations, it is concluded that using all of the available data provides the best estimate of the true standard deviation.
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 13 oZ 22 Figures 4.2 and 4.3 present a histogram of the relative differences for the BOC and MOC cases, respectively. Figure 4.4 presents a histogram of the combined BOC and MOC data.
These Figures show a reasonable normal distribution of the percentage difference between the measured and inferred power.
The data points with the large differences (-17%) are due to the fuel assembly which has an RPD lower than 0.07.
The following table presents the results of the relative standard deviation associated with the extrapolation uncertainty.
Exposure Case Relative Standard Degrees Deviation of Freedom BOC .0387 585 MOC .0338 556 Total .0363 1141
- 4. 3 PEAKING FACTOR UNCERTAINTIES The F,. parameter uncertainty was calculated in Reference 6.2 by deleting only one detector string at a time. Their results show an extrapolation standard deviation of 2.40% with 800
ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 14 oZ 22 data points. The difference between the 3.63% and the 2.40% is the increase in uncertainty by allowing 50'f the detectors strings to be failed.
Using the 3.63% instead of the 2.40%, and using the SPC calculated uncertainties for the other parameters, the new calculated standard deviation for Fq and Fr, respectively, is 3.98% and 3.90%. Using the new calculated uncertainties to account for 50% of failed string detectors, the modified Table 1 is presented below:
TABLE 1 modified Relative Number Standard Degrees Variable Deviation Freedom Fsa .0363 1141 Fr .0047 514 Fz .0077 514 Fg .0135 188 Fq .0398 1466 Fr .0390 1363
ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 15 of 22 The effective number of degrees of freedom was calculated using the Satterthwaite's formula presented in Reference 6.3.
The tolerance factors for both Fq and Fr is 1.71. The one-sided 95/95 tolerance limit for Fq and Fr using 50% of failed detector strings are therefore, 6.81% and 6.67%, respectively.
5.0 CONCLUSION
S This analysis concludes that a conservative increase in measured peaking factor uncertainties will accommodate operation with up to 50% failed incore detector strings.
For conservatism, these peaking factor uncertainties will be increased to 8.0% and 7.5% for Fq and Fr, respectively. Using these values, the following table 5.1 is constructed:
TABLE 5.1 Condition F LHR Uncertaint Fr Uncertaint
< 25% failed detectors 7. O~o 6. O~o between 25% and 50% failed 8.0% 7. 5%
ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 16 of 22 FIGURE 4 ~ 1 Cycle 11 Failed Detect;or Strings Instrument No St . Lucie Leveis 1 4 Y X V/ V T S A P N M L K J H G F E 0 C B A 21 20 18 16 14 13 12 11 10 8
8 7 17 16 15 13
Ir 5
i"J y I I'e s l v3 l,,
,7<
<.'i J'l
ATTACHMENT 3 JPN-PSL1-92-008 Rev. 0 Page 17 of 22 FIGURE 4.2 Histogram of Percent Differen'ce . Between Measured and Inferred Values BOG Data Histogram 150 0 100 I
C L
0 U
0 L
S 50 C
Z 15 11 10 -7 -5 -3 -1 I 3 5 7 9 11 13 15 17 Percent. Deviation (R)
ATTACHMENT 3 JPN-PSL1-92-008 Rev. 0 Page 18 of 22 FIGURE 4.3 Histogram of Percent Difference Between Measured and Inferred Values MOC Data Histogram 150 S
S O 100 I
L L
L 0
0 0
0 I
L.
50 E
z 15 11 10 -7 -5 -3 -1 1 3 5 7 9 11 13 15 17 Percent Deviation (KJ
ATTACHMENT 3 JPN-PSL1-92-008 Rev. 0 Page 19 of 22 FIGURE 4.4 Histogram of Percent Difference Between Measured and Inferred Values SOC and MOC Oata Histogram 300 250 0
Q 0 200 C
8 L
U 0 150 Q
I 100 E
Z 50 15 "13 -11 10 -7 -5 -3 -1 1 3 5 7 9 11 13 15 17 Percent Deviation C%$
C ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 20 of 22 TABLE 4.1 Summary of Standard Deviation Results (BOC Cases)
~Case STD 0 Variance 1 3. 556 0.0013 2 4.966 0.0025 3 4.803 0.0023 4 4 '98 0.0018=
5 3.072 0.0009 6 2.215 0.0005 7 1.532 0.0002 8 3.207 0.0010 9 2.466 0.0006 10 3.406 0.0012 11 4.690 0.0022 12 5 '48 0.0025 13 3.299 0.0011 14 2.377 0.0006 15 4.781 0.0023 16 2.522 0.0006 17 2.796 0.0008 18 2.045 0.0004 19 5.197 0.0027 20 3.475 0.0012 21 3 '56 0.0011 22 4.888 0.0024 23 3.569 0.0013 24 2.443 0.0006 25 5.171 0.0027 26 5.834 0.0034 27 5.592 0.0031 28 2.931 0.0009 29 3.709 0.0014 30 3.041 0.0009 31 4.930 0.0024 32 2.140 0.0005 33 2.482 0.0006 34 2.892 0.0008 35 5.022 0.0025 36 3.251 0.0011 37 4.846 0.0023 38 5.212 0.0027 39 3.018 0.0009 Total Standard Deviation 3.87%
-'e'1 ATTACHMENT 3 JPN-PSL1-SEFZ-92-008 Rev. 0 Page 21 of'2 TABLE 4 . 2 Summary of Standard Deviation Results (MOC Cases)
~case Variance 1 2.638 0.0007 2 4.285 0.0018 3 2.753 0.0008 4 2.315 0.0005 5 4.301 0.0018 6 2.877 0.0008 7 2.524 0.0006 8 2.455 0.0006 9 4.958 0.0025 10 1.904 0.0004 11 2.318 0.0005 12 3.202 0.0010 13 4 '23 0.0021 14 4.505 0.0020 15 2.058 0.0004 16 4. 614 0.0021 17 3.255 0.0011 18 2.252 0.0005 19 4.513 0.0020 20 2.887 0.0008 21 4.515 0.0020 22 4.465 0.0020 23 4.792 0.0023
'005'TD%
24 2.395 0.0006 25 3.154 0.0010 26 3.144 0.0010 27 4.869 0.0024 28 2 '32 0.0004 29 4.722 0.0022 30 2.984 0.0009 31 4.168 0.0017 32 1.866 0.0003 33 2.781 0.0008 34 2.236 0.0005 35 2.627 0.0007 36 4.444 0.0020 37 2.294 0 Total Standard Deviation 3.38%
ATTACHMENT 3 JPN-PSL1-SEFJ-92-008 Rev. 0 Page 22 oZ 22
6.0 REFERENCES
6.1 CENPD-153-P Revision 1-P-A, "Evaluation of Uncertainty in the Nuclear Power Peaking Measured by the Self-Powered, Fixed Incore Detector System," May 1980.
6.2 XN-NF-83-01(P), ,Exxon Nuclear Analysis of Power Distribution Measurement Uncertainty for St. Lucie Unit 1," January 1983.
6.3 F. E. Satterthwaite, "An Approximate Distribution of Estimates of Variance Components," Biometrics Bull. 2 (1946) I 110 114 ~
0