ML20198T492
| ML20198T492 | |
| Person / Time | |
|---|---|
| Site: | Farley  |
| Issue date: | 03/31/1997 |
| From: | Pitterle T, Srinivas V WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML20198T480 | List: |
| References | |
| SG-97-01-002, SG-97-1-2, NUDOCS 9711170010 | |
| Download: ML20198T492 (19) | |
Text
_ __
SG-97-01-002 4 '
9 Farley Unit-2 Small Bobbin Probe (0.640")
Qualification Test Report March 1997 Prepared By:
T. A. Pitterle V. Srinivas Equipment Design and Regulatory Engineering Nuclear Service Division Westinghouse Electric Corporation i -
L 9711170010 971110 -
l PDR ADOCK 05000348: ,
P PDR !
- -. - - - - . . ._= . . .-. ..
.c l i
--g-
. a Farley Unit 2L >
. Small Bobbin Probe (0.640") Qualification Test Report 1.0 : INTROD UCTION This report doi uments a test conducted'at Farley Unit-2 during the fall 1996 outage -
to qualify a 0 640" bobbin probe for ARC application. The report describes the acceptance crit eria based on detectability and voltage accuracy utilized for qualifying the probe and test results. Qualification of the 0.640" diameter probe is considered
.to qualify probe sizes between 0.640" and 0.720" diameter when the probes are centered to the' nominal tube ID. Small diameter probes are required for TSP intersections that are not accessible to inspection with the 0.720" probe, such as between TSP sleeves or above tubesheet sleeves in small U-bend radius tubes which are inaccessible from the opposite tube end due to a small U-bend radius. 'This qualification effort has been requested by the NRC to supplement available data supporting use of the small diameter probe for ARC applications.
Similar small probe tests were performed in spring 1994 at D. C. Cook Unit-1 and Farley Unit-1 (Reference 1). Results from those tests showed that a 0.640" probe can meet the voltage accuracy requirement and the measured voltages tend to be biased towards conservatively high- voltages. However, additional data was deemed necessary since either the sample size or the bobbin voltage range in those tests was limited. The test performed at D.C. Cook Unit-1 had only 12 RPC confirmed indications and, therefore,it did not meet the test sample size requirement. The test performed at Farley Unit-1 had an adequate test sample size but all indications considered for data evaluation were below the 2 volts repair limit "hich distorts the results. Therefore, another small probe qualification test was performed in steam generator (SG) C ct Farley Unit-2 as part of the fall 1996 outage inspection.
This report describes, in Section 3, the acceptance criteria based on detectability and voltage accuracy for qualifying the smaller probe, the selection of the sample for the qualification test, the protocol to conduct the test as a blind test and data evaluation against the acceptance criteria. Successful qualification of the probe would permit application to other plants with 7/8" diameter tubing implementing an NRC approved
. ARC.
t-L
l I
2.0 CONCLUSION
S !
Data evaluation for the 0.640" probe qualification test at Farley Unit 2 was performed against 42 indications detected by the 0.720" probe and confirmed by RPC inspection.
Those bobbin indications ranged from 1.5 to 6.7 volts, and all 42 indications were detected by the 0.640" probe also, i.e.,100% confirmation rate. The probe detectability requirement based on Appendix H of the EPRI ISI Guidelines is 80%
detectability at 90% confidence. The Farley Unit-2 test results provide 95%
detectability at 90% confidence; thus, they exceed the Appendix H guideline of 90%
confidence. In addition to 42 RPC confirmed indications,16 other indications were also detected by both 0.720" and 0.640" probes, of which 9 were RPC NDD (b:bbin voltage between 1.9 to 3 volts) and the other 7 were not RPC tested (bobbin voltage between 0.6 to 1.4 volts) as their voltages were below the cut-off applied in this study for RPC inspection (1.5 volts from the 0.720" probe). Those 16 indications are not included in the data evaluation.
The voltages recorded by the 0.640" probe exceeded the 0.720" probe voltages for all 42 RPC confirmed indications. The voltage differences between the two probes varied from 0.05 to 2.44 volts and had an average of 0.88 volts. The standard deviation of the voltage differences was 0.50 volt which is 18.7% of the 2.67 average volts for the 0.720" probe results. The acceptance criteria for voltage accuracy was that the standard deviation should not exceed the analyst variability uncertainty by more than 10%. For a sample size of 42, the analyst variability at 90% confidence limit is 14.3%. Thus, the voltage accuracy requirement is satisfied by the 0.640" probe voltages which is biased to conservatively high voltages.
A linear regression analysis was also performed to correlate the voltage readings from the 0.720" and 0.640" probes for RPC confirmed indications and the results yielded the following relationship, Vno = 0.753 x %o (1)
Therefore, on the average 1.0 volt from 0.640" probe corresponds to 0.753 volts from the 0.720" probe. Regression analysis also showed that for indications above the 2 volt repair limit the measured 0.640" probe voltage would exceed the 0.720" probe voltage inferred from the 0.640" voltage by the regression correlation at 95% (one-sided) confidence level. Thus, the results ahow it is conservative to use 0.640" probe volts in place of the 0.720" probe volts.
It is concluded that voltage measurements obtained with the 0.640" probe can be applied for ARC applications as they meet the established acceptance criteria. Probe sizes between 0.640" and 0.720" are also considered to be acceptable as these probe sizes can be expected to even more closely match the 0.720" probe results.
2
- 1. ,
3.0 " ACCEPTANCE CRITERIA AND SA.MPLE SIZE REQUIREMENTS Acceptance criteria for the small bobbin probe are based on detectability and accuracy of the voltage measurements as described below.' The assessment utilizes indications -
detected by the 0.720" probe that are confirmed as flaws by RPC impection. RPC inspection was performed on all indications with a 0.720" probe voltage above 1.5 volts.
3.1 Detectability Requirements Detectability requirements for acceptance of the 0.640" probe used for the present test are same as those applied to tests conducted in spring 1994: the probe should satisfy-Appendix H of the EPRI ISI Guidelines which require detection at 80% probability and 90% confidence (Reference 1). The Appendix H guidelines are applied for the indications found by the 0.720" probe and confirmed by RPC inspection. Typical detection requirements are given in Table 1. As seen in Table 1, the number of indications which can be missed by the 0.640" probe is a function of the number of 0.720" indications. A minimum of eleven indications are required, and for a sample size below 17 all indications found by the larger probe must be also found by the smaller probe.
3.2 Voltage Measurement Requirements Two methods of analysis can be utilized for evaluating the voltage differences between the small and large probes, and both methods were applied in the present study. One analysis examined for a correlation between the voltage readings of the two Probes, and it showed how the two voltage readings for a given indication compare. It also provided a relationship between the two probe readings as a function of voltage, which would permit, if found necessary, an adjustment of the 0.640" probe voltage to the 0.720" probe voltage.
The second method involved comparison of the voltage differences between the two probes against the acceptance criteria. Earlier small probe qualification work (documented in Reference 1) set the acceptance limit for the voltage differences between probes at 10% based on the draft NUREG 1477 guidelines and the NRC feedback at the February 8,1994 NRC/ Industry meeting. This criterion was applied for the present small probe qualification test by requiring that the standard deviation of the voltage differences between the 0.720" and 0.640" probes should be below 10%.
However, these tests inherently include analyst variability in analyzing the data as well as tho probe to probe differences. That is, if the test was performed twice with ;
. the same 0.720" probe, a non zero standard deviation of voltage differences would be i expected. The acceptance criteria for the test should then recognize that there should l be a very high probability (>90%) of passing this test if the identical probe was used .
1 3
l
l
+
. twice for the measurements. Data presented in Reference 2 show that the analyst variability uncertainty has a standard deviation of 10.3% based on a large population size. Any small sample from the larger population may have a standard deviation differing, within bounds, from that of the larger population, this difference being a function of the sample size. For the analyst variability uncertainty of 10.3%, Table 2 shows the acceptable upper bound uncertainty (standard deviation) to obtain a two.
sided 90% confidence value. Alternately stated,90% of the samples from the large population analyst variability distribution would be expected to have a standard deviation less than the value in Table 2 and there would be a 90% probability of
. passing the probe test if identical large probes were used for both series of measurements.
The voltage measurement accuracy acceptance limit for the small probe qualification test is then established by requiring that the standard deviation from the test be within 10% of the expected difference of Table 2. This can be expressed as follows:
%o- 4, $ 10% (2) where:
%o = the standard deviation of the voltage differences between the 0.640" and 0.720" probes.
4, = the analyst variability standard deviation at 90% confidence for N samples from Table 2.
The percentage uncertainty for analyst variability was developed by dividing the standard deviation of voltage differences by the mean voltage of the population. This same method is applied to obtain the percentage standard deviation for the test results.
3.3 Sample Size Requirements To apply 80% detection probability at 90% confidence, it is desirable to have a minimum of about 25 indications found by the larger probe and confirmed by RPC.
It is desirable that the bobbin indications be > 1.0 volt but also include indications below the 2 volt repair limit; a cut-off at 1.5 volts meets these two requirements.
Also, as RPC data available is available for allindications over 1.5 volts, an adequate
, size population of RPC confirmed indications can be selected for data evaluation with the 1.5 volts cut offlimit.
To provide margins against the potential RPC confirmation rate, it was determined that at least 50 indications should be included in the small probe sample population.
For RPC confirmation rates at C0% to 80%, the 50 bobbin indications would be expected to result in 30 to 40 RPC confirmed indications for evaluating the detection probability of the smaller probe.
4
.The number of TSP intersections sampled should be much larger than the number of indications. A factor of 20 more intersections sampled thau indications found provides a reasonable margin. The intent is to apply the small probe only to the hot leg inspection where TSP indications are present. A 150 tube inspection would provide 1050 TSP intersections to be sampled by the small probe.
In summary, the inspection sample requirements for the small probe qualification are:
At least 50 bobbin indications or about 30 more than RPC confirmed bobbin indications are to be included in the sample population.
- The total sample shall include about 150 hot leg tubes inspected from the 1st to 7th TSP.
All indications with a 0.720" probe voltage above 1.5 volts are to be RPC inspected.
All bobbin and RPC data are to be analyzed to the site eddy current analysis guidelines including use ofindependent analysts and resolution of discrepancies between analysts. The resolved indications and associated bobbin voltages are the data base for evaluation against the acceptance criteria.
Only the RPC confirmed indications are to be evaluated against the acceptance requirementr.
In the present test, inspection data is available for 149 tubes and the total number ofindications found is wellin excess of the minimum 50 indications requirement.
3.4 Testing Protocol The small probe qualification test was conducted as a blind tesc. That is, personnel performing the smull probe data analysis did not have access to the large probe inspection results used to select the sample population. To achieve this objective, the following protocol was applied to this sample inspection:
The vendor lead analyst for the 0.720" probe and the RPC data evaluations (S/G selected for small probe qualification test) were not associated with nor have access to the 0.640" probe data evaluation.
- Based on the results from the large probe inspection, a plant representative selected the sample tube population to the sample size guidelines given above.
For evaluation of the 0.640" probe data, the eddy current analysts and 5
. resolution analysts did not have access to the results of the large probe inspection or a significant familiarity with the results.
The eddy current guidelines require tl.e application of the probe wear standard limits for both 0.720" and 0.640" probo inspections. This requir nent was also applied to this test.
4.0 TEST RESULTS Small Probe qualification tests at Farley Unit-2 during the fall 1996 outage examined 149 tubes and a total of 67 indications were detected by the 0.720" probe. Of these 67 indications 58 were detected by the 0.640" probe also. All indications over 1.5 volts based on the 0.720" probe reading were inspected with a RPC probe. A total 51 indications were RPC inspected and 42 of them were confirmed. All 42 RPC confirmed were detected by both probes, and their 0.720" probe voltage ranged from 1.5 to 6.7 volts. Voltages from both probes for RPC confirmed indication are shown in Table 3.
Table 4 shows indication segregation based on detection by both probes and RPC confirmation status. In addition to 42 RPC confirmed indications,16 other indications were also detected by both 0.720" and 0.640" probes, of which 9 were RPC NDD (bobbin voltage between 1.9 to 3 volts) and the other 7 were not RPC tested (bobbin voltage between 0.6 to 1.4 volts) since their voltage were below cut-offlimit for RPC inspection (1.5 volts). Those 16 indications are not included in the data evaluation.
For two of the 9 RPC NDD indications, the 0.720" probe call is PI (potential indication) but the 0.640" probe call is UOA (unusual OD phase angle). Figures 1 and 2 short the Lissajous figures from both probes for one of these indications. Both bobbin graphics are very similar with both probes showing the same large phase angle (144" vs 145"). Therefore, both of those indications were considered as detected by both pobes. For three other indications, the 0.720" probe calls are classified as either INRs (indication not reportable) or UOAs and these indications are not included in the statistics presented above or shown in Table 4.
There are 9 indications which were detected by the 0.720" probe but not by the 0.640" probe. All those indications were small with a 0.720" probe voltage below 1 volt and none of them were RPC inspected since they are below the 1.5 volts cut offlimit for j RPC inspection. On the other hand, two indications detected by the 0.640" probe were not detected by the 0.720" probe and those two were also not RPC testrd.
l 6
l l
. 4.1
- Detectability Data l
Since the implementation of ARC is inherently dependent-on confirmation of indications using a RPC probe, only RPC confirmed indications are considered for data evaluation for detectability. Since all 42 RPC confirmed indications in the Parley Unit 2 small probe qualification test were detected by both probes, the t
detectability rate is 100%. The corresponding detectability at 90% confidence is 95%.
Thus, the detectability _ requirement of 80% at 90% confidence per EPRI ISI Guidelines, Appendix H are mot.
4.2 Voltage Measurement Accuracy In addition to assessing the 0.640" probe detectability, the test results are evaluated for voltage accuracy against the requirements of Section 3, Equation 2. Voltages recorded by the 0.640" probe exceeded 0.720" probe voltages for all 42 RPC confirmed indications. The voltage differences between the 0.040"and 0.720" probes varied from 0.05 to 2.44 volta and had an average of 0.88 volta. The standard deviation of the voltage differences (cuo) was 0.50 volt which is 18.7% of the 2.67 average volts for the 0.720" probe readings. The acceptance criteria for voltage accuracy was that the standard deviation should not exceed the analyst variability uncertainty by more than 10%. For a sample size of 42, the analyst variability at 90% confidence limit is 14.3% from Table 2. Thus, the voltage accuracy requirement of Equation 2 is satisfied by the test data. Also, the 0.640" probe voltages is biased to conservatively high voltages. Results of an analysis to correlate the voltage readings from the two probes is presented in the following section.
4.3 Regression Analysis Results A linear regression analysis was carried out to examine for a correlation between the voltage readings from the 0.720" (Vuo) and 0.640" (V720) probes for RPC confirmed indications. The following form of relationship between Vuo to V 72o was considered, V 72o = bo + b, x V uo (3)
The regression analysis considered both zero intercept (bo=0) and non zero intercept cases. The index of determination (the square of the correlation coefficient) of the correlation between Vuo to V 720 for both zero intercept and non zero intercept cases were found to be the same and equal to 78.1%. Hence the data indicate that the 0.640" probe volts are strongly correlated to the 0.720" probe volts. Since exclusion ofintercept results in a simpler relationship to the fit, the final form of the regression 7
- W ~ t- e ..-9,
j ; .7 ,
. equation selected was, n .
1 V,,o = .
bxV.
3 _
(4) 4
..The above relationship yields a constant value for the average ratio between V, to V,,o and constant ration was found to be 0.753. The index of determination for the r ratio of Vm to V,,o'(against V ) was found to be.l.8% indicating tinat there is no correlation between the voltage ratio and probe voltage Vm. This is consistent with the constant voltage ratio indicated by the' correlation between the two probe voltages.
A summary of results from the regression of V,,o against Vm is provided in Table 5.
A plot of the analysis data, the fitted regression curve and the upper one sided 95% ,
prediction curve, is provided on Figure 3. The regression results show that on the
. average 1.0 volt for Vm corresponds to about 0.75 volt for Vtso or conversely 1.0 volt for Vt2o corresponds to 1.5 volts for Vm. The results in Figure 3 also show that for
- , indications above the 2 volt repair limit even the 95% one sided upper prediction >
bound value for V,,o is below Vm. Thus, the results show it is conservative to use 0.640" probe volts in place of the 0.720" probe volts.
To complete the regression analysis, a plot of the regression residuals versus the
- - predicted 0.640" probe volts is provided on Figure 4. A normal probability plot of the
. regression residuals is provided on Figure 5. Examination of these plots indicates
-that the variance of the residuals is approximately constant and the residuals are approximately normally distributed. In summary, the data do not contradict the inherent assumptions implied in performing the regression analysis.
Based on the a'uove results it is concluded that for indications detected by the 0.640"
. probo,. the 0.640" probe voltag measurements can be applied as an acceptable measurement for ARC applications, e
1 5.0 References
- 1. SG 04 08 016,"Small Probe Qualification Test Results for IPC/APC Application,"
Westinghouse Electric Corporation, August 1994.
- 2. TR 100407, "PWR Steam Generator Tube Repair Limits Technical Support Document for Outside Diameter Stress Corrosion Cracking .at Tube Support Platea," EPRI Report (Draft), Revision 2A, January 1995.
i 1
8
. - . . . - . ...- u~ o , ,e..-,.. --_,....~,--I--i. ,. , ye , . , ,--w-w , s. w ar,., .+
. Table 1
. Detection Requirements for 80% Detectability at 90% Confidence No. 720 Ind. Confirmed by RPC No. Detected by 640 Probe
- N = 11 to 16 N N = 17 to 21 N-1 N = 22 to 28 N2 N = 29 to 33 N3 N = 34 to 39 N4 N = 40 to 44 N5 N = 45 to 51 NG N = 52 to 57 N7 N = 58 to 63 N-8 N = 64 to 70 N9 l
9
. . - _ _ ___. l
Table 2 Sample Size Dependent, Analyst Variability Standard Deviation at 90% Confidence Sample Size 4, at 90% Confidence 20 16.3%
21 16.2%
22 16.0 %
23 15.9 %
24 15.8%
25 15.6%
26 15.5%
27 15.4 %
28 15.3 %
29 15.2%
30 15.1%
31 15.0 %
32 14.9 %
33 14.9 %
34 14.8 %
35 14.7 %
36 14.7 %
37 14.6 %
38 14.5 %
39 14.5 %
40 14.4 %
41 14,4 %
42 14.3 %
43- 14.3 %
44 14.2 %
45 14.2 %
46 14.1 %
47 14.1%
48 14.096 49 14.0 %
50 13.9% -
- Based on large population standard deviation of 10.3%
j 10 ;
Teble 3 Small Probe Qualifications Test at Farley-2 Comparison of 0.720" and 0.640" Probe Voltagen -
SG Row Col TSP 0.720" Probe 0.640" Probe 0.640-0.720 Volts Voltage Voltage Difference
~CC 17 6 211 2.51 3.69 1.18 14 8 111 2.89 3.99 1.1 C 18 14 til 2.41 3.3 0.89 C 14 24 211 1.94 2.88 0.94 C 7 35 311 2.24 3.37 1.13 C 34 37 til 4.37 5.64 1.27 C 38 39 211 2,72 3.87 1.15 C 44 47 Ill 2.27 2.56 0.29 C 42 48 til 1.97 2.92 0.95 C 42 48 411 1.74 1,79 0.05 C 42 48 211 3.54 4.51 0.97 C 46 51 111 2.97 3.43 0.46 C 34 53 til 6.73 7.56 0.83 C 45 56 til .B7 4.31 0.74 C 31 60 I11 2.95 3.59 0.64 C 35 66 - til 2.87 3.65 0.78 C 41 64 111 2.5 3.16 0.66 C 42 64 IH 2.16 2.86 0.7 C 39 65 til 3.36 4.45 1.09 C 30 67 Ill I.93 2.7 0.77 C 40 67 til 2.64 5.08 2.44 C 18 68 2 11 2.85 3.04 0.19 C 27 68 til 3.79 4.45 0.66 C 29 68 111 3.21 4.86 1.65 C 13 69 111 2.12 2.46 0.34 C 21 71 til 1.5} 2.64 1.12 C 33 71 111 2.28 3.43 1.15 C 20 72 l11 23 3 3.35 0.72 C 30 72 til E22 2.54 0.32 C 24 74 IH 1.91 2.48 0.57 ~
C 35 74 til 2.24 3.41 1.17 C 35 76 1 11 2.02 3.83 1.81 C 26 77 111 3.55 3.99 0.44 C 24 78 til 4.01 4.86 0.85 C 5 8I I11 1.73 2.3 0.51 C 5 81 211 1.96 2.42 0.46 C 12 82 iH 2.03 2.66 0.63 C 15 84 2 11 2.11 2.86 0.75 C 29 84 1H 1.97 4.35 2.38 C 10 88 211 2.18 2.82 0.64 C 10 90 1H 2.35 3.06 0.7i C 17 0l III 2.99 3.75 0.76 Count 42 Average 0.879 Variance 0.247 Staridard Deviation 0.497 Maximum 2.44 )
Minimum 0.05 l
REGRANLY.XL5 Table) 3497 7.24 PM 1 l
I
Table 4 -
~
Summary of Results from Tests'at Farley-2 5
RPC Conformed RPC NDD No RPC Test
'720/640 Probe Detectability No. Voltace Range No. Voltare Range No Voltare Range i Both 720 and 640 Detection 42 1.5 to 6.7 volts '7' 1.9 to 3.0 volts' 7 0.6 to 1.8 volts 720 Detection,640 NDD 0 0 8- 0.4 to 1.0 volt 720 NDD,640 Detection 0 0 2 1.6 to 1.7 volts h
h 4
12
p ,
l .
4 Table 5 Small Probe Qualification Test at Farley-2 Regression of 0.720" Probe Volts on 0.640" Probe Volts Llucar Correlation - Zero Intercept Slope - b.1 0.7535 0 Intercept - b.0 SE.bl 0.0181 - SE.b0 r^2 78.1 % 0.4333 SE.720 F 146.462 41 DoF SS. reg 27.500 7.698 SS.res F. Prob 4.08E-15 5.74E+02 SS.640 Pi .Value 3.49E-35 -
Po.Value N 42 1.024 1 + 1/N var.640 14.002 3.545 p.yo MGRAM.YM.5 Me4 M97 7 24 PM 10
l .
' M s I .
' .,e;ej[ 5
!% s e s a 8 I '
[ s ' . i t
- A a '
}1 R '
M t t
d T
t e
s s:
e E
- ,.z3
' s 7 4 e
t I[:
'T i
n g
"- e28w 9 a
- j s #' s s
s 2 a -
t e. :
e -
Y ' : : r 2 .
' eEt :
nTtt
- s. "'
j l l eNtT T. D tCa*
D E ,
, ' sSt *
- T E r '
me 5
X E
P S
2 s .
s 6 e s
a a i
( o H .
/ /
h sri ,
A G
h sr - 4 6 et eS et a 3 s s 4 e s 1 6 a
- g sC c
C a a
2 g 2 ,
R Is.
t b r 3 3
f o 2 z 3 , s.
o e s.
r y
v s b I s o
r
- K G v r
1 P e m ,g 4 r "0 / i I u 2 / f i
g 7 F 0 s:r r
- h st f
4m
- ef m
f or a
a:
nc i
s y
t m
eS s
3 C I
d .
s l n a
i n
g 8
I 't
}
LmJ r S 2 2 e s
i n r y e
s.
r .
- o. a b - w b
o B vs i I
l s:3 >]
e j ]. -
s e
a : ~ .
, ~
~
~
~
l o
n E Q-s
?.,,?n
'E 3 . r "j. r t t E
t a
g
,
- c 6 " %
7 b w . 3 i
oy e~ ~
{ .
j .
bTm E l I-k l 8E 3 m :
- 1 L- : :
'; D E I s ,
=
(g A i
s
- EE
. .. g' 3
C E lE F
N b :
i gE.
G E - _
y C
I.
- lx l *: I
.C .
g <
f, -
M $'
% n '
d g d g $ $
O d 5 ~
.c .
8 I N N f gO e \ \ '
3 fb
+
2 z V h
.9 e \
% o f i %q4
- si j !!
N E
l I
j -d -
, f, v 8 m 1 1 8 = .
.3 .- .t o _s' j " D <r__
1, _ .
a ..;
'A
^ -. -
, ae --
- I
~~
= r 2 a e m
J11h .__n EC074 01/15/97 13:07:12
O Figure 3
~
Small Probe (0.640") Qualification Test at Farley Unit 2
- Linear Regression Analysis Results -Zero Intercept -
8 .
! 7 - - - - -- - -- - - - - - - - - - - - - - - ~ -- -- -- - -
o Test Data o '
6 -- - - - - - - ------- - -- ---
- . . * = - - - -
Regression Curve ,..-
m .-
y5 -
9E% Prediction _ _._ _ . _ . _ _ _ _ _
, .~ , rf _ _ _ 1_ _.
,.~~
o .- o
.c ...-
o 4 _ _ _ _ _ _ _ _ _ . _ _ __ _ -
,, . __ o _ .. _ _
u ..-o o -
,..a oo eq
~
o N
3 - --
g - - - -
o . . . g : ' n go--
- o o o a
,...-*p c
2 - - - - - - - --
.. - -r- - - - o- c -
, . . . -o o
j,.-
1 -
d- - --- - - -- -- - - - - - - - - - - - - -
4 i
I I o
O 1 2 3 4 5 6 7 8 O.640" Probe Volts n-v on i mm,- 16
~
Figure 4 -
Small Probe Qualification Test at Farley Unit 2 .
Linear Regression Analysis - Scatter Plot of Residuals -
2.0 .
- i 1.5 _
c 1.0 o :
. .c .
20.5
t :
a : . . . .= . .
t 0.0 o - =" "%*s'.
.i u
=
- o : #
o =
-0.5 -
.a.
e -
8 -1.0 m
-1.5 .
~
-2.0 - - - - - - - - - - - - .-
0.0 0.5 1.0 1.5 ' .0 2 2.5 3.0 3.5 4.0 4.5 5.0 Predicted 0.720" Probe Volts m.mv xu . . - , m 37
~
~
Figure 5 -
Small Probe (0.640") Qualification Test at Farley Units I and 2 ,
Linear Regression Analysis - Normal Plot Of Residuals '
4.0 _
~
3.0 .
s Observed Data l
- , a 2.0 .
--- Linear Fit ,
y' '-~
a ,-
3 1.0
_E /
g f,,
.= -
~f O o0 J[p 5 -
f E.' f e -
u -
o -10 ,- ' , ' c z -
,- m
- a ,-
m -
-2.0 '
a
-3.0 -
~
_4_9
-1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 Observed Residual Values
.--m,,,-,,,,,. is