2CAN108107, Forwards Info Re Topical MSS-NAI-P on Reactor Physics Methods,In Response to NRC 810304 Requests & Per Util 810806 Commitments
| ML20031G500 | |
| Person / Time | |
|---|---|
| Site: | Arkansas Nuclear  |
| Issue date: | 10/16/1981 |
| From: | Trimble D ARKANSAS POWER & LIGHT CO. |
| To: | Clark R Office of Nuclear Reactor Regulation |
| References | |
| 2CAN108107, NUDOCS 8110220488 | |
| Download: ML20031G500 (64) | |
Text
3
_1_
ARKANSAS POWER & LIGHT COMPANY POST OFFICE BOX 551 LITTLE ROCK. ARKANSAS 72203 [501) 371-4000 October 16, 1981 pQf 2CAN108107 Director of Nuclear Reactor Regulation
((
l ATTN: Mr. Robert A. Clark, Chief i
L-c, 0072 2 sg g
Operating Rcoctors Branch #3 L,
Divisio~ of Licensing O M.nsa,y,,$
- C 2.$
/
U. S. Nuclear Regulatory Commission Washington, D. C.
20555 4
Q/
s
SUBJECT:
Arkansas Nuclear One - Unit 2
. J. ?
Docket No. 50-368 License No. NPF-6 Reactor Physics Methods -
MSS - NAI - P (File:
2-1510)
Gentlemen:
In response to your letter of March 4,1981 (OCNA038107) and as committed in our letter of August 6, 1981 (2CAN088102) attached is our response to your requests regarding Topical MSS - NAI - P on Reactor Physics Methods.
Very truly yours, w
David C. Trimble Manager, Licensing DCT JTE:sc
\\
8110220488 811016 PDR ADOCK 05000368 P
PDR MEMBER MIDDLE SOUTH UTIUTIES SYSTEM l
QUESTION 1 In the expression for o2 the total observed uncertainty (p. 3-5) the independence i
OBV of the calculated uncertainties for rod worth oR and the boron coefficient oB has l
not been established. Note that both quantities are calculated using PDQ-07. What independence?
is the basis for the assumption of R and B
)
ANSWER 1: While both the rod worth and boron coefficient are calculated using independence is as follows:
PDQ-07, the basis for assuming cR and B l
1.
4 ntrol rod worth is not solely a PDQ-07 calculation. The PDQ-i7 control rod worth calculation is normalized to an independent, higher order transport calculation, i.e., CPM.
l Hence, any dependence which might exist between the two com-l ponents would be minimized through the normalization process.
2.
PDQ-07 calculations have demonstrated independence of boron worth with respect to rod worth and rod worth with respect l
to boron concentration.
Specifically, a 100 PPM error in boron concentration would produce a change of approximately 0.14% in rod worth while a 10% error in rod worth wouid pro-duce a change of approximately 0.02% in the inverse boron 1
worth (IBW). Clearly, interdependence of the two components l
oR and og is negligible. Tables A.l.1 and A.l.? summarize i
the PDQ-07 results.
TABLE A.1.1 i
EFFECT OF BORON CONCENTRATION ON CONTROL R0D WORTH Boron Concentration Control Rod Worth l
(ppm)
(%an) 400 28.85 870 28.66 AbsoluteDi{ference 0.19 Coefficient 0.14 TABLE A.1.2 EFFECT OF R0D INSERTION ON IBW Normalized IBW Rod Worth (ppm /%Ao) 1.00 79.80 0.89 79.78 AbsoluteDijference 0.02 Coefficient 0.0023 IRod Worth Coefficient, i.e.,
%ao(400 ppm) - %an(870 ppm)
- 100 = Change _in Rod Worth
%Ao (870 ppm)*4. 7 100 ppm Boron Worth Coefficient, i.e.,
IBW(1.0) - IBW(0.89)
- 100 Change in Boron Worth 2
=
_.!BW(1.0)*(100-89)_.
% Change in Rod Worth
k j
l QUESTION 2
~
t On p. 3-10 and 3-11, it is stated that NAI " performed an evaluation of the compari-sons of measured and calculated ITC's... The resulting data base..."
Does this paragraph mean that some " data conditioning" or " data selection" was performed?
I Give more details on the data evaluations referred to above.
ANSWER 2: No " data conditioning" was done.
Instead, " data selection" was done in accordance with the guidelines in Table A.2.1.
ITC measurements which did not satisfy the guidelines were deemed to be unqualified.
For ANO-l Cycles 1, 2, and 3, eight measurements were performed (four in Cycle 1, two in Cycle 2, and two in Cycle 3). One Cycle 2 measure-ment was deleted because of the unavailability of the complete original reactimeter stripchart.
One Cycle 3 measurement was disqualified be-cause it failed guidelines 1 and 3.
This rejected measurement and the corresponding MSS model prediction are shown in Table A.2.2.
It is evident from Table A.2.2 that the inclusion of this rejected data from Cycle 3 has no significant impact on the determination of the reliability factor.
TABLE A.2.1 GUIDELINES FOR EVALUATING IS0 THERMAL TEMPERATURE COEFFICIENT MEASUREMENTS Parameters Guidelines 1
1.
Coolant Heatups and Cooldowns The measurement should include at least one I
heatup and one cooldown.
2.
Moderator Temperature The rate of change of moderator temperature should be less than 2 F/ min.
f 3.
Control Rod Position The control rod positions should remain un-changed at least two minutes prior to and during the measurement.
4.
RCS Boron Concentration The RCS boron concentration should remain constant.
5.
Reactimeter Trace The uncertainty in reactimeter trace slope 0
l should be less than 0.5 PCM/ F based upon successive one minute segments.
TABLE A.2.2 0
U Controlling Boron ITC(PCM/ F)
AITC (PCM/ F)
Cycle Bank Positions ppm Meas. Calc.
(Calc. - Meas.)
ANO-1 Cycle 3 5 = 75%
1066
-6.9
-7.2
-0.3
puESTION 3 0
One Kewaunee comparison for the isothermal temperature coefficient (-7.2 PCM/ F) has been deleted and the deletion was justified on the basis that the value is four standard deviations from the Kewaunce mean using the pooled estimates of the variance.
(a) Measurement cannot be rejected on 6he basis of statistical arguments; (b) at this point, the poolability of the data has not been established; and (c) the value is only about two deviations from the Kewaunee mean using the Kewaunee estimate of the variance.
Discuss the suspected measurement error.
How would this value affect the realiability factor if it was left in the pooled data?
AqSWER 3:
See answer to Question 6.
QUESTION 4 The Isothermal Temperature Coefficient data base listed in Table 3.5(a) includes comparisons from Kewaunee, Beaver Valley, ANO-1 and ANO-2. These plants repre-sent a Westinghouse 2-Loop, Westinghouse 3-Loop, B&W and Combustion plants, respectively. Address the issue of data poolability in view of the plant design diversity.
ANSWER 4:
See answer to Question 6.
4
,r
..r
QUESTION 5 The presence of a boron bias in the pooled data.for the isothermal temperature coefficient appears to indicate that spectral effects are not being fully ac-counted for by the calculational model or calculational procedures (Section 3.2,
- p. 3-10). Has this point been investigated before the bias factor was absorbed in the calculational Methodology?
ANSWER 5:
See answer to Question 6.
f l
l l
l l
l t
QUESTION 6 On p. 3-11 using the data of Table 3.5(b), a statistical equivalency test is per-fonned on Kewaunee and ANO-1 data. Given that the ANO-1 has only two entries, coninent on the validity of such a statistical test.
ANSWER 6: The ITC comparisons originally reported in MSS-NAl-P have been revised to include only those comparisons involving AN0-1 and AN0-2. Hence, Questions 3, 4, and 6 are no longer applicable.
Reanalysis of the ITC comparisons, more fully described below, demonstrates no boron dependency in the bias. Hence, Question 5 is not appli,:able.
As mentioned above, the investigation of ITC model bias and un-certainty was performed again using only ANO-1 and ANO-2 compari-sons. The highlights of the new investigation were as follows:
1.
Six ANO-1 comparisons (see Question 2) and 5 ANO-2 com-parisons were used.
2.
Due to the small size, a rigorous statistical analysis of the data was not performed.
Instead, conservative estimates of bias and uncertainty were developed, based upon engireering judgement.
3.
A new bias of 0 (prev',ously boron dependent) and a new reliability factor of a.0 PCM/ F (previous 2.5 PCM/0F) 0 are the esults of the investigation.
Revised port' ins of report MSS-NAl-P more fully describe the new investigation and are attached.
QUESTION 7 In view of the lack of measurements, demonstrate that a 10% reliability factor for the Doppler coefficient is conservative (Section 3.3, p. 3-16).
ANSWER 7_: The calculation of the Doppler coefficient is performed using the three-dimesional nodal model to first simulate the reactor conditions of interest, which represents a base condition, and a change in the power level from this base (holding the Xenon distribution and moderator temperature constant). The primary variable in this calculation is the change in fuel temperature associated with the power change which results in a reactivity change due to the change in the resonance absorption (i.e.,
Dopplereffects). The algorithm in the nodal code that deter-mines the change in reactivity due to a fuel temperature change uses data calculated by EPRI-CELL, i.e., Ap/op, The approach was to determine the accuracy of EPRI-CELL in cal-culating the change in resonance integral (RI) due to a known fuel temperature increase, and to use engineering judgement to bound this uncertainty to assure conservatism. The results of these analyses are summarized in Tables A.7.1 and A.7.2, which were taken from ARMP documentation, Part I, Chapter 4.
(Refer-ence 1) l In all cases, the calculated values from EPRI-CELL were well within the quoted measurement uncertainty (4% on the resonance integrals and 10% on the coefficients - See Reference 1, Part I, Chapter 4).
In view of this, the 10% reliability factor placed on the doppler coefficient was judged adequate and conservative.
i l
l l
l
TABLE A.7.1 COMPARIS0N OF RESONANCE _ INTEGRALS (RI)
Fuel Fuel Rod Temperature Experiment EPRI-CELL
- Relative Radius (cm)
(OK)
RI Exp.(b)
RI (b)
Difference (t) 0.52 300 21.750 21.254
-2.33 0.52 900 23.688 23.335
-1.51 1.04 300 17.019 16.916
-0.61 1.04 900 18.332 18.265
-0.37
- Relative Difference = ((RI - R!exp)/RI)
- 100 b:
barn TABLE A.7.2 COMPARIS0N 0F DOPPLER C0EFFICIENTS Fuel Relative Radius (cm)
Experiment EPRI-CELL Difference (T) 0.52
.0077 0.0083 7.23 1.04
.0067 0.0069 2.90 NOTE:
RI (T) 1
/ 6
-/T (1) Doppler Coefficient =
o RI (T )- 6 o
~
T = 900 K, T = 300 K o
6 = 1.558 barns is the (1/v) contribution to the resonance integral for neutron energies above 0.55 ev.
(2) Relative Difference = (
(EPRI-CELL) - (EXPERIMENT)
/ (EPRI-CELL))*100
QUESTION 8 Referring to Figures 3.4 - 3.6 and Table 3.6 on the isotopic comparisons, it is not clear whether or not the applicant performed the indicated calculations, hence, demonstrating his capability of using the EPRI-CELL, CPM and ARMP codes.
Did the applicant perform the calculations indicated on Figures 3.4 - 3.6?
ANSWER 8: The isotopic comparisons (Figures 3.4 - 3.6 and Table 3.6) were performed by the staff of A. B. Atomenergi, under the EPRI Research Project 118-1. Results have been published in Part I, Chapter 5 of the ARMP Documentation (Reference 1). The CPM cal-culation used the standard CPM code and the standard ARMP method-ology. MSS uses the same code and the same m?thodology.
Further-more, all of the sample problems accompanying the ARMP package were exectuted at MSS, and the results were identical with the sample results supplied by EPRI. Therefore, the results would have been the same, if MSS had performed the calculations.
P
QUESTION 9 What adjustments were made to the EPRI-CELL code to match CPM? (Section 3.4,
- p. 3-17)?
ANSWER 9: A minor correction was applied to the group 3 absorption cross section (o3/4) of Pu-240 to account for the differences between EPRI-CELL and CPM computation of post Pu-239 isotopes. This correction was on the order of 3%.
p I
l l
l l
l 1
l l
l
L i
I i
7 h
QUESTION 10 Uncertainties in the spatial nuclide inventory calculation inv^1ve uncertainties i
in the local inventory computation as well as uncertainties in the computation l
of the spatial burnup distribution (Section 3.4, p. 3-17; Section 4.3, p. 4-5).
The uncertainty in the local inventory computation has been dealt with in i
Section 3.4.
Establish the uncertainty in the computation of the spatial burnup distribution.
j 4
ANSWER 10: The spatial burnup uistribution is needed for the calculation of isotopic inventory used in both the reactor operation support and the determination of core kinetic parameters, Seff and 1*, for safety related calculations. For its application to reactor opera-tion support, such as core isotopics tracking, the best estimate
~
(i.e., c = 0) from nodal code calculations will be used.
For its application to the determination of Beff and t*, an allowance will r
be required for the calculated parameters to account for the un-certainty in the computation of spatial burnup distribution. The uncertainty in the computation of spatial burnup distribution and its effect on the calculation of core kinetic parameters, s ff and e
1*, are discussed in the answer to Question 11.
i l
l 4
.1 i
i
?
5
t i
i QUESTION 11 i
i i
Uncertainty (b) in the calculation of the spatial nuclide inventory (Section 3.5, pp. 3-22 and 3-24) is only partially addressed in Section 3.4.
What is the value of the uncertainty in the spatial nuclide inventory when the uncertainty in the 1
spatial burnup distribution is accounted for?
(
ANSWER 11: The effect of tt.: uncertainty in the spatial burnup distribution on the spatial isotopics is of interest in the manner in which I
it may impact the uncertainties of Beff and 1*.
Variations or l
uncertainties in the spatial exposure distribution are caused by variations or uncertainties in the power distribution. The 8 and 4
t* values are global (i.e., core) rather than local (i.e., nodal) and hence are impacted by global mechanisms (biases) rather than uncertainties, since the uncertainties are randomly distributed, plus or minus, and therefore their net effect is zero. Global variations in the exposure distribution are caused by power dis-i tribution biases. This mechanism for creating uncertainties in j
Seff and t* was previously neglected based on the assumption of its second or third order effect. This is illustrated in the example given in Table A.11.1.
The power distribution bias which could propagate as an exposure distribution error is shown in Figure 3.40.
This bias was used to evaluate the change in Beff for typical ANO-1 conditions. This calculation resulted on 0.014 change in s ff which is small rela-a tive to the total uncertainty of 2.8% and was therefore neglected.
i 4
i 4
i l
i 4
m_
.... ~ - -.
l TABLE A.ll.1 l
SENSITIVITIY OF 8eff DUE TO POWER DISTRIBUTION ERROR Anial Pocer Exposure Power Distribution Without Bias
- 5 1
2 3
Lo_ca tion Distribution Distribution Bias (From Figure 3.40) 8(x10-2) 1 Beff(x10-2) 8(x10-2)
I Beff(x10~2) i j
(GWD/MTU) i 12(top) 0.483 9.7 2.8 0.6195 0.483 0.6216 0.496 4
11 0.830 13.7 1.6 0.5863 0.830 0.5881 0.841 10 0.959 15.5 0.6 0.5714 0.959 0.5721 0.963 i
i 9
1.011 16.1
-0.2 0.5664 1.011 0.5661 1.007 i
8 1.093 16.3
-0.8 0.5647 1.043 0.5636 1.033 7
1.064 16.3
-1.2 0.5647 1.064 0.5631 1.049 I
j 6
1.058 16.1
-1.4 0.5664 1.058 0.5645 1.041 i
5 1.084 15.7
-1.0 0.5697 1.084 0.5684 1.093 4
1.139 15.7
-0.6 0.5697 1.139 0.5689 1.130 i
3 1.212 15.9 0.0 0.5680 1.212 0.5680 1.209 2
1.195 15.9 0.8 0.5680 1.195 0.5691 1.202 1(bottom) 0.921 12.8 1.8 0.5938 0.921 0.5957 0.936 Core 1.0 15.0 0.57300 0.57307 3
Average l
9 % error 5 =.01% <
Descriptions:
t (1) The power distribution and exposure distribution were calculated using the nodal code for a ANO-1 Cycle 3 statepoint with a core average exposure of 15.0 GWD/MTU l
i (2)
I = Power Weighting Factor i
(3) 8eff = (E(8 x I))/I(I)
(4) The biased 8 and I are calculated using the biased power and exposure distribution. The biased l
exposure is defined as -- Exposure (biased) = Exposure (unbiased) x (1 - power distribution bias)
(5) The error in Beff due to the error in burnup distribution is calculated by:
j 1
l error (%) = Beff(biased) - Beff(unbiased) / Beff(biased)
- 100 = (0.57307 - 0.5730)/0.57307 x 100 = 0.1%
1 l
1 t
i f
__._,I
QUESTION 12 Shat are the one-sided tolerance factors used when relating the uncertainties in 8 and t to the corresponding reliability factors (Section 3.5, pp. 3-24 and 3-25)
ANSWER 12: In the determination of e ff and t* reliability factors, a worst case e
scenario was constructed. Therefore, the one-sided tolerance factor is not applicable to the uncertainties calculated for e ff and E*.
e The worst case analysis was characterized by two major conservative approaches:
1.
" Worst case" errors were used as the uncertainty compocants.
For example, in the evaluation of the effect of power distri-bution uncertainty on the deteraination of _Befs, a worst case power distribution error of 7% was used.
(Act0a1 power dis-tribution error was computed to be less than 3% - See Section 3.6, p. 3-40.)
2.
These worst ;ase uncertainty components were then combined deterministically (i.e., they were summed).
i l
l l
l l
l i
QUESTION 13 It is stated on p. 3-26 for the power distribution uncertainty as measured by the Rh self powered detectors that "the signals from these detectors are corrected by the on-site process computer...these corrected signals, or reaction rates..."
(a) Usually the process computer will process the Rh detector signals in a manner which will account for detector sensitivity, depletion, leakage, etc.
(b) The process computer will determine a quantity which represents the re-action rate.
In view of the above coments, explain whether further corrections to the Rh signals have been applied and what corrections were they?
ANSWER 13: No further corrections were applied to the detector signals, with the exception of a small background currents correction applied to the AN0-2 signals.
This correction was on the order of one percent.
QUESTION 14 The reaction rate to power density conversion factors are calculated for each assembly as a function of exposure using a 2D PDQ model.
In view of the discussion of power distribution reliability factors presented in Section 3.7, p. 3-109, how are axial effects for these factors accounted for?
ANSWER 14: In the model uncertainty analysis, the power density to signal (reaction rate) conversior, factors are used to convert ti'? pre-dicted power distribution into predicted detector signals.
Axial effects are accounted for in that the local exposure (in the vicinity of the detector) is used to evaluate the factor. Any ad-ditional deviation caused by the axial variation of these power-to-signal conversion factors is absorbed in the model uncertainty.
I l
l l
l l
l
1 1
QUESTION 15 l
I l
Axial power shapes are shown at locations D10, 88, and R4 (Figures 3.12 through 3.38).
j Explain why no distributions are shown for central or near central locations. Show 1
some axial power shapes of central locations and describe the normalization to plant i
measurements and PDQ.
ANSWER 15: In question 15, "R4" probably refers to "N4" (since the location R4 is not in the core region). The presentation of two assembly locations per map was judged to be illustrative of the measured versus the predicted axial results. The choice of locations l
D10, B8, and N4 was made to show:
1.
high power assemblies, or j
2.
assemblies under the influence of control rods.
J In any event, power shapes for central and near central locations j
F8, Gil, H8, L6, and N8 (chosen randomly) are shown in Figures A15.1 - A15.6. The nodal model normalization procedure consisted of two steps:
r 1.
The radial leaka e factor (radial albedo) and the horizontal i
mixing factor (g ) were adjusted such that the radial power distribution cal ulated using the nodal code matched the ra-dial power distribution from quarter core PDQ calculations for equilibrium Xenon and Samarium conditions. Since the two-dimensional PDQ does not have thermal and hydraulic feedback, all nodal calculations were performed likewise.
4 2.
This top and bottom leakage factors and the vertical mixing factor (gy) were adjusted so that the axial power distribu-tion calculated using the nodal code matched the plant mea-surement at one time in life (B0C2). These parameters have been constrained to retain a single set of values for all of 2
the cycles.
i i
i 2_._.
.._,,._ _ ____,__..._,.-_.__.... _._.~,_.
MEASURID AND CALCU1ATED REACTION RATES
~
ANo-1 FLUX MAP 1 CYCLEI EXPOSURE 66.3 ETPD CORE LOCATION L6 i
2.0 i
i a
e i
i i
i l
- Calculated o Measured O
O l
m 4-1.0 N
2 E
U l
3 l
an E
7 oc 0
1 1
I I
I I
I 0
20 40
' 60 80 100 PERCENT OF CORE MEIGHT
MEASURED AND CALCUIATED REACTION RAT 1.S ANO-1 TLUX MAP 9 CYCLE I EXPOSURE 450.2 ErFD CORE LOCATION F8 2.0 l
Calculated o Measured e
W m
O
(
O O
O 1.0 -
t!
2
~
~
E-g i
e 0
i i
i I
I i
i 1
I 0
20 40 60 80 100 PERCDU OF CORE HEIGHT
--,-,y--
~-,m---
e w
MEASURED AND CALCU1ATED RIACTION 7.ATES AND-1 FLUX MAP I CYCLE 2 UP05URE 12.8 ETPD
~
CORE LOCATION 'HB l
- ~
l 3
l l
l 5
I I
I Calculated l.
o Measured 0
W O
g W
t O
O O
1.0 e
O W
N O
l, 2
5 2
,1
=
l\\
u u
2 b
0i i
f f
1 I
i 1
1 1
0 20 40 60 80 100 l-PERCENT OF CORE HEICHT t
(
MEASURE AND CALCULATED REACTION RATES I'
ANO-1 FLUX MAP 9 l
l CYCLE 2 EXPOSURE 243.1 ETPD l
CORE LOCATION C11 i
i I
i l
I l
j 2.0 j
i
- Calculated c Measured
\\
i i
I O
O 1.o i
l N
2 O
em U
g l
e 1
5 0
t I
I 1
I I
i 1
0 20 40 60 80 100 PERCEhi 0F CORE HIICHT l
HEASURED AND CALCUI.'.TED REACTION RATES ANO-1 F1.UX MAP 10 CYCLE 3 EXPOSURE 279.6 EFFD CORE LOCATION F8 2.0 i
i i
s i
- Calculated o Measured I
~
l l
1.0 O
~
t l
1 C
W 8
0s m
W De
~
0 8
0 20 40 60 80 100 PERCENT OF CORE HEIGHT
I M'.ASURE AhD CALCULATED REACTION RATES 00-1
. FLUX MAP 10 CYCLE 3 EXPOSURE 279.6 ETPD CORE LOCATION H8 2.0 8
3 1
1 e
i a
6
- Calculated o Measured O
O g
O O
1.0-E n
g U
a m
Nbe s
I I
I 1
1 1
I 0
O 20 40 60 80 100 PERCENT OF CORE HEIGHT
4 i
i QUESTION 16 I
' On page 3-32 it appears that "the simulation errors" are due to (a) input errors i
and (b) approximations in the representation of a given state point.
Describe in detail the errors implied by the term " simulation error".
ANSWER 16: The type of errors implied here are slight differences in rod positioning due to uncertainties in rod position indication i
(13.2 inches), and slight differences in inlet temperatures between the design value and that measured at the plant.
In addition, simulation error includes any error included because 1
the N0DE-P assumes a 3-D equilibrium condition. As explained in the report on pace 3-33, in some cases it was not possible
- to determine the cause of some small observed axial power os-cillation.
The approach takes. was to simulate the measured power distribution using the observed rod positions, power 1evel and inlet temperatures and assuming a equilibrium xenon distribution. Thus, because a small axial oscillation was observed in some of the measured plant data, a small simulation error is introduced in the equilibrium modeling assumption.
d 9
y- *y N et sy,g-gm-g.ggw ;e yr gwMY w ypg-y-71-g.*M *W rt "9WW "'-4W t t-fke$=-g'P't'1MtT9--'t:fg 't 7 r er e*-4'-p-'--h9 trde**
vt-M"v"P"'=--*b="'*'T+W*9+-*='r-*W-99 e
"'w'-*"
e rT e
-W
QUESTION 17 How are errors associated with asymmetries (e.g., asymmetric fuel burnup distribu-tions) accounted for in the model uncertainty analysis and in the core monitoring system uncertainties?
(p. 3-34)
ANSWER 17: Differences observed between symmetric detector readings were treated as measurement uncertainties. All other asymmetries are included in the model uncertainty (See Section 3.6, p. 3-35).
l l
l l
l
1 QUESTION 18 In view of the systematic pattern of errors shown in Tables 3-10, 3-11 and 3-12, how is detector intercalibration maintained? Is there a cycle or exposure de-pendent method used in correcting possible drifts away frtn intercalibration, and if so, please describe it.
ANSWER 18: All of the ANO-1 incore detectors were calibrated by the vendor before installation. Af ter initial installatien, only symmetric detectors are intercalibrated.
Using the methodology in the pro-prietary letter from K. E. Suhrke of B&W to Darrell G. Eisenhut, Assistant Director for Operational Technology, Division of Operat-ing Reactors, USNRC, dated March 4,1977, the symmetric detectors are intercalibrated about once per cycle to a symmetric ring average.
1 i
I J
3 w
l 1
1
4 1
QUESTION 19 7
It is stated on pp. 3-37 and 3-38 that hardware problems in Cycle l led to the elimination of Cycle 1 data from inclusion in estimating an axial level depend-1 ent mean difference for ANO-1 (Table 3-18, Figure 3.40, pp. 3-85 and 3-80). This suggests that there should be a large uncertainty in the bias. With respect to the data included in Table 18, on what basis is it concluded that the differences i
between Cycle 1, Cycle 2, and Cycle 3 are due te hardware problems and how is it insured thet they are specific only to Cycle l?
ANSWER 19: The hardware problem referred to on pcge 3-37 was the replacement of the original 50,0000 measur? ment resistor in the in-core detector strings with a 5,000n resistor at 24! EFPD in Cycle 1.
Investigation has revealed that:
(a) No detector cable leakage correction was applied to the Cycle 1, Map 1 in-core signals.
(b) Based on comparisons of raw detector readings before and after the resistor replacement, ar. insufficient leakage correction factor was a,mlicd to Maps 2-4 by the plant computer.
For ex-l ample, the s: ring #2 plant corputer output indicated that a leakage correction factor of ~1.409 was being applied to level 7 just before the resistor replacement. After the resistor re-placement, the correction was ~1.000.
This indicates that the,
string with the 50,0003 resistor was being increased ~40% to obtain a leakage corrected signal. However, a comparison of the uncorrected signals immediately before and inmediately af ter the resistor substituion for string 2, level 7, shows an i
increase of signal of 64%.
Since the 5,0002 resistor is be-lieved to give the best indication of power (largest signal i
and smallest leakage correction), it is concluded that the leakage correction factor was under-estimated during Maps 2-4.
Since the leakage correction increases from level 1 to level 7, Maps 2-4 signals over-estimate the power in the bottom of the reactor core.
(c) Since the leakage correction was under-estimated for Maps 2-4 as concluded in item (b), it follows that the accumulated charge for the detectors in the upper levels of the core was under-est" mated. The accumulated charge is used to calculate the rhodium depletion correction. As the accumulated charge increases, the rhodium depletion correction factor increases.
The problem in (b) above would cause the depletion correction for the upper levels to be too small. Thus, although the re-sis $or change was made at 241 EFPD, considerably more time was required for the erroneous depletion correction to become insignificant and the false shift in power to the bot-tom of the core to disappear.
9
ANSWER 19: Cont'd.
Because of this hardware modification in Cycle 1, which to some extent affected every Map in Cycle 1, the Cycle 1 measured power distributions were judged to be inappropriate for detailed model benchmarking purposes and were eliminated.
According to ANO-1 plant reccrds, no further mea-surement resistor changes were made in Cycles 2 or 3.
l l
l l
l
QUESTION 20 What is the mechanism or phenomena responsible for the axially varying reliability factor (RF)? Why does ANO-1 require an axially varying RF, while AN0-2 does not?
Is the axial variation ycle deper. dent?
ANSWER 20: The apparent axial variation of the ANO-1 reliability factor was investigated in detail. Only level 7 appears to indicate an axially wrying reliability factor. ANO-1 operated until Cycle 5 with control rods inserted in level 7.
The average power in level 7 is significantly less than any other level in ANC-1 or AN0-2. MSS investigated the possibility that small absolute differences between the small measured and predicted powers in level 7 of ANO-1 might appear as large percent differences which could be used in the calculation of model uncertainty.
Such large percent differences in '.ow power nodes could dis-tort the true uncertainty needed in the reliability factor to ensure that predicted peaks are conservative with respect to measured.
In order to verify this hypothesis, MSS reanalyzed levels 1 and 7 of AN0-1 for the 19 qualified power maps using a lower power level cutoff of.65 P/P. As expected, few comparisons were rejected at level 1 and the uncertainty changed less than.2%; however, at level 7, 260 low power comparisons were rejected.
The standard deviations of the remaining 728 level 7 compari-sons which were over.65 P/P were poolable using the Bartlett test and yielded a pooled standard deviation of 5.08%. Table A.20.1 gives the number of comparisons with P/P >.65, and the standard deviations based on those comparisons for 19 maps and level 7.
Thus, it is clear that the small differences in lower power nodes were inflating the relative percent differences and causing pooling problems. The pooled uncertainty for all relative powers over.65 gives a Fg reliability factor of 9.04%, which is consistent with what is observed for other i
i levels in ANO-1 and ANO-2 and is not overly conservative due I
to overweighed influence of unimportant low powers. Based on the preceding discussion, MSS intends to use a axially uniform bounding reliabili.j factor of 0.10 for Fg.
Revisions to pages 3-3, 3-112, and 3-113 of MSS-NAl-P are provided.
i
_ TABLE A.20.1 STANDARD DEVIATIONS FOR ANO-1 LEVEL 7 LOW POWER CUT 0FF =.65 P/P CYCLE MAP
- COMPARISONS STANDARD DEVIATION %
2 1
37 6.0 2
2 36 5.5 2
3 37 5.5 2
4 37 5.7 2
5 37 5.5 2
6 37 5.0 2
7 39 4.5 2
8 40 4.3 2
9 40 4.9 2
10 43 5.5 3
1 30 6.0 3
2 31 4.9 3
3 33 4.7 3
4 40 4.6 3
5 42 5.0 3
6 43 4.7 3
8 42 4.8 3
9 42 4.7 3
10 42 4.6 e00 LED UNCERTAINTY = 5.08%
1 l
l l
{
l l
L
I TABLE 3.1 RELIABILITY FACTORS FCR ANO-1 BENC'iMARK CALCULATIONS Parameter Reliability Factor Bias
- 0.02 to + 0.02 F
RFpg = 0.10 (refer to Figure 3.40)
F RF
=
.057 0
3g Fah Rod Worth RF
= 0.05 0
R0DS 0
Temperature RFg = 4.0 PCM/ F 0
Coefficient Doppler RFDC = 0 10 0
Coefficient
, oppler RFDD = 0.20 0
Defect Boron Worth RFB = 0.05 0
Delayed RF = 0.03 0
g Neutron Parameters RF,= 0.03 0
l 3-3
FIGURE 3.52
. l
}
F RELIABILITY FACTOR FOR AM01 N
0
=
7, -
. p.
.;g.
i 4
u
.. -7,.
..l j.
- l.
..p 7,,;.
.i;, y.
... L.
i g
j
.p i.
. ii
.:i i
...,d:
.; r.:p.. :..
I
.p..g. p.
ie
.d
..y 7
i t.
.i i
'..p-
,*)*
~
-~
s-[
t
)
N g.
e
~; -
sv
... 7.I t ;-
r
~.
m
- qs sI i
..t we 9
w
- d...
..!. i;.
.u..._.....
g
.% l
,.y 1
f i
I O
l
',j i,
'6..
i i
1, s
o li.
.e.'
- 9....
....g.
,.;. i..
. j.,
.;...t..37
. j.
.p..
4 La.
r j
f
.i i
{*
- i H
y
-+*
i.!
' #i'*
- * *'*E**
i"*
l"
- ' ' ' 'W* *EEf 1
.t
)
8"*
o.e "E* *
'h*' * * * *
" " * ' ' b
+
". 'b 5*'
..j..
E ena
_a 8*
T *" ** ""1
- j.
t:
.l -(
oe F
,8-
..e
!t
'l-EI3 pr lj
- i erC
.i.,.t
. [...
- n i
- n..
p.s a
I i :
- i g
w l
.e, a:
1
'lj dj; l'.
j.
2' l
.I f n,
.t 6
r i!
(CY
- q. p 1
G<i p-g'
..,q:g
- n.,,,..
.r-g-
4 - :.
i.
, '.. o.
- z...
r-
. q.
}
c.,-
. '....a;..
1
..i i.
j.
.;p.p...,.
, pl!. -
r r.
i i
. I
- -l.,
...j.......
c
.!i.
- , y-J-
q v.
i.
t:
. t..
l.
8 i
.r p.,,
d'
'S i
q,
- - p
- ji.t-e
.i -
- ili q-
- t. +: "t.
.p.
- i-e;i ::h t-t g
4 ;- '-.t-i+
-i - -i,: g-: -- -
- - r p:
Hl ;
- t. 6.
.a i,;
I.=
- r
!- n-
. 1
.l.
i
...i
.:o
- i. :
L,,
0 10 20 30 40 50 60 70 80 90 100
- i.#.
.-o
.m Data from Page 5 88 PERCENTAGE OF CORE HEIGHT e
e e
e e
e e
e e
TABLE 3.25 N
Fg RELIABILITY FACTORS FOR AN01 FhReliablity S
Core Model Height Uncertainty Factor
(%1
(%)
(%)
100 5.7 10.0 90 5.7 10.0 80 5.7 10.0 70 5.7 10.0 60
- 5. 7 10.0 50 5.7 10.0 40 5.7 10.0 30 5.7 10.0 20 5.7 10.0 10 5.7 10.0 0
5.7 10.0 3-113
QUESTION 21 Discuss the effects caused by the nodal code albedo selection, calibration errors and crud buildup on the axial dependence of the reliability factor.
ANSWER 21:
The reanalysis in Answer 20 has shown the Fo reliability factor to be uniform axially and thus any effects of nodal code albedo selection, calibration errors and crud buildup are a horbed in the model uncertainty.
~
i l
QUESTION :2 i
With regard to the pooling of statistics, the discussion on the application of the BartlettTest is not clear in the light of Figures 3.40 and 3.41.
Discuss the basis for pooling the statistics in Tables 3-13 tc 3-17 in view of the fact that the data failed the BartlettTest. Supply curves similar to Figure 3.41, separately for Cycles 1, 2, and 3.
ANSWER 22: MSS has analyzed the poolability of the calculational uncertainty using the conservative procedure in Table A.22.1.
The results of this poolability analysis are shown in the attached revised page 3-88.
A revised page 3-87, which is a graph of pooled (ap-1 plying the methodology in Table A.22.1) model error for Cycles 1
2 and 3, is incluied.
Cycle 1 was not included since the data had been rejected due to hardware problems as described in Answer 19. The reliability factors on Fg were recalculated and are also shown on page 3-88. These reliability factors are bounded by the values in the revised Table 3.25.
MSS will use a RF g af 0.10 and a RFFaH of.057.
F Figures A.22.1 and A.22.2 are curves of bounded model uncer-l tainties for Cycles 2 and 4 using the pooling methndology of Taole A.22.1.
Level 7 pooling is discussed in the answer to Question 20.
Y l
{
\\
FIGURE 3.41 BOUNDED UNCERTAINTIES FOR AN01 POOLED FROM CYCLES 2 AND 3 ilh.n.
e '}lik ShE.E!
.Ii:..q:
U;
.ui hi
-id i.-
i-
.22 sElh2 i.
....a u
.a a
. E ai..i.'
I I
!:., e. 'd.:
.,.Hd.
l.e!. o;;
r..
! - 1 in n.
.,il.? pi a
. l... l..
i.
i:;
_1
+
i-
- lp 1-
.!:- r:r'
.;o n..
l,
". p-. it' aulath;; G2ac
{t
, h. -.
e,.;ih a
- .1...
t.t.
. l..
. A;..; t.
. j.
.:.] '.: h:..p
! c.
.t i;.
i
._ _2
- 1 il.,1;;
""a "h' ":, s.-
,'1
- pMi li.
i-
'p-10 I-
- q N!:
!o - -
l t
u"nj...
- i
.i m
y;;,:.i:
.u.a
.ui m...
...a.
. g
.. _ a.
. a..
- i. p.
9
- a..
. 1.
i..
s
?'~?-
? ll N'=
h
- Y-=-
S 8
N-
~-
_.L l
. i.
' d-i
?
4
.u
. tJ.
. ma.
A..;
d' 'I..
..g.
MJ
. i...
L;.
L.
sn f _lla,.
5 i -
..e.r g
i 7
t...
u l
E i
u L.
n...
......L... :.!
A.L -
O
.b;
. U.. ;M 48 E
.)
1.
h gp y
I
%J
_t O
- 12;-
i i.
r; t
LAJ 5
g i
T
. [14
..l Ola.
- ..1..
a.ai..A x
. l.J.
. 31.
i.I f..
- ..l
.4..
l
'g '
' i O
J i
.I 1
gAJ i
J 4
_L i
E a:
m ua
- i..;a.:
.hd
..Ju.
o u
.a :
.u...
, ),
. k.
4 n.
..j
.(..
. l ;..
o a
~'
r:
3 i
.u.
2..
i i
8 I
2 L.l a
s
.5 e
.a. L1.. I...c a.
4..
..2.
- x...
-o n
- k
. i{;1.. id ;..co i
l-
.. a.. u
.e. _.
.h.
.;it i
..l.:
t i
.. ai.
.5 i..
.a.
l-o:. :'.
'1,
.I...
I i
i.
g
.u. ua2.;=
2..,_
..a a ;.
4 i
i 1-c i
i
.. L;
_h.. L.
i t.a 0
10 20 30 40 50 60 70 80 90' 100 i._3 _ _
._L PERCENTAGE OF CL9E HCIGHT g
N"
TABLE 3.19 P0OLED STANDARD DEVIATIONS FOR ANO-l CYCLES 2 AND 3 S
Axial Number of Model (t)
Level Comparisons
(%)
K RF+
7 728 5.1 1.742 9.0*
6 832
~ 4. 0 1.736 7.3 5
988 4.8 1.728 8.5 A
988 3.4 1.728 6.3 3
520
- 5. 7 1.762 10.1 2
988 5.1 1.728 9.0 1
988 4.4 1.728 7.9
+Using equation on Page 3-110
- See answer to Question 20 3-88
TABLE A.22.1 P0OLING METHODOLOGY (a) Using Bartlett's test, test for poolability by level for all maps.
If data passes Bartlett's test, determine opooled and K95/95 K95/95 cor-responds to the sample size of N, N=rall maps (N ).
The reliability j
If n t, go to (b).
factor will be no less than K95/95
- pooled.
(b) Using Bartlett's test, test for poolability by level for individual cycles of data.
If not poolable, go to (c).
If all cycles are individually pool--
K95/95 corresponds to the same able, determine pooled, cycle and K95/95-sample size of N, N=Ecycle (Nj). The reliability factor will be no less than the maximum of K95/95
- pooled cycle" j
(c)
If data is poolable except for one or two maps which have a very small standard deviation, eliminate these maps for that level and retest for poolability.
If all maps are poolable, then determine opooled and K95/95-K95/95 corresponds to a sample size of N, N is defined as the total number of comparisons in the reduced set.
If only poolable by cycle (for all and K The relia-cycles), then determine the cycle specific pooled 95/95 If not, bility factor will be no less than the maximum of K95/95 pooled.
i go to (d).
f (d) Determine K95/95
- maximum f r each cycle.
omaximum is either the maximum of the level standard deviation in the cycle or the pooled standard deviation l
for that cycle (if poolable).
K95/95 corresponds to the number of comparisons used in the determination of amaximum. The reliability factor will be no i
less than the maximum of K95/95
- maximum' J
t
TABLE 3.25 N
Fg RELIABILITY FACTORS FOR AN01 FhReliability S
Core Model Height Uncertainty Factor
(%1
(%)
(%)
100 5.7 10.0 90 5.7 10.0 80 5.7 10.0 70 5.7 10.0 60 5.)
10.0 50 5.7 10.0 40 5.7 10.0 30 5.7 10.0 20 5.7 10.0 10 5.7 10.0 0
5.7 10.0 3-113 o
FIGURE A.22.1 BOUNDED UNCERTAINTIES FOR ANO-1 CYCLE 2
.' r P:"! ' r! '
. 1{ Y. m.l ". - l. i! lh h;j
- ~' - -
' -~~-
~~
I l
!, F
- P. U..!)b, i. i *H jl
'l l it' l, ' l U. d i U.. '
4ll l'l
- [
- t. i j '.,.
i,h.-!!. l.!. i*li[.
11 L
,i
-~.
i
- . l i
i i
i i
i 8
ht :'
t,t i. t- 'r !' I tl[...-
i o
n!:'
qi vtj
.jf ;;
- tri,
' TP '*i--~~'
~ t** " ' ".,
" t--
-* i
. p:
'l ia
- li.
'i qp,i i
+,!-4,-
' il jI
,i ip
.i
,' l '
l,
.4, i:'
.1 8
P l.i ei
{t
- li I'
?g t*
t'
.s' e-e : 3'
'. ',.,t :
- , i-p:
ti ;l-
' ' ;;;' g*
t**
- t-t 10
- --T
--T.-
-"i
.44 a
..y 7
g..
..g.
.e 9
i I
i
!i i
- 11..
..i
.p 8
9.: q-
,.y i t i
s.
- i
- i l
- i s
4
..., 7 g
....,.; 4 g g! 5p
.I 8
vi 0
7 i
ii i
.i' y
q.
i.
g 9-t:
t;- @
j
~
tj j-
-r-*~
tf t}
{
L l!
l.
s i
w 6
. n.. e; l
..si
.i
. G.,.
4-
..i.
. r-
.p
. y.-
g
.I
.7 -...
- l
.l.
3, l
9 6
i.
1 O
i i
.t
- l'
?
l
.i...'l!
..p t
E
,.Q,_.
.I p
. ~.
q -..
.7
.o*
4 i
n J(y,
__O a.
-.q.
.2.
..i.
g
. a, 3
,.l
-n.
s 7
7 7..
7
)
e y
u
..l' 2
I i
i
._g' +;"
a
- u. g.a. a.
.g.
- r...
y
)
- i'..aj
~
i i
-... 7'
.... q.
.. p.,
l
..,i.._.. _..._._
i i
0 I
i 0
10 20 30 40 50 60 70 80 90.'
100 Percentage of Core Height I
FIGURE A.22.2 BOUNDED UNCERTAINTIES FOR ANO-1 CYCLE 3 Y{ h.'
.p;t.n m. ;.
'l ij d]1.
.dl.-%m,lI jlli 'Ij! hil lll ll III' liii P.i h'I
!]l.
i
[
Q;'
.li i
.I i.f,' l ' ;}i' au :-llilli Iii!
IU.U NlI
! b. l T!!
p r; -c.
..e
_- ;t!
h a
ki
.es iu i.e it -jj i i. %.a h.- m i
- t
.;t t
a.
i.
t c
i! 7io
.. r.
1!
h i
in
,l; i
.t.
i
.p;.,
,a..,;.
1 1 -
2:;..
- 'c'.t;
..p..
'*j
.j
.I l
i-a 6
.i..
,.;.j..
.. a
- 1
.p
- i 10
~~
^~
.I
,i i
,i
~
.l'
.4
.t..
.+
q i..
i i
i.
9
-.2
,-.i.
,' i f
.i.
..p
,t.
....y.
N 8
...l..
.!u v
j.!
'l 5
u 4.. 9 y?
g
!I t
il
- p i.
._..j,
.i..
I i
in 7.
,1 q.'. :;..
.,e
.p.
'. g #
9:.
.u.
y..
f c
i
- i u
p.
i, l
.i o
u g
_I u
- d. !L. u.
tu.
_ _ r-
.; _.q:
i.d..g.:
W q
i.
.....q
.g r
-* 5
.l.
p.
4.
...{.
l
\\O V
- l; g
.qi.'
7 "e
4 i-,
i.
i.
i
.2
' i i
._t i
O i
I O
i
'Y a.
3
,.i-t i
i
..l,i.,.
v 2
. a
- ,..._+
.g.
'. I l:
dii j
l i
id : h !
0 I'
l il i i _LO
~
t t
50 60 70 80 90 -
100 0
10 20 '
30 Percentage of Core Height 9
e
l QUESTION 23 The large differences in the statistics between Cycle 1 on one hand and Cycles 2 and 3 on the other, suggest that they are being drawn from different popula-tions. Justify, therefore, the use of a one-sided tolerance factor correspond-ing to over 1300 data points of 1.71 (p. 3-110).
ANSWER 23:
Data from Cycle 1 was aliminated from the model qualification data base as discussed in Answer 18. The reliability factors for Fg were computed using the appropriate K factor from the 4
i new Table 3.19 (see Answer 22).
t The reliability factor for FAH was recalculated from Cycles 2 1
and 3 data using the pooled model uncertainty of 2.8%.
The new value for the FaH reliability factor using the equation on page 3-110 is RFFAH =.054; however, the conservative value j
of.057 reported in Table 3.1 of MSS-NAl-P will be used.
i 4
i i
J 1
1
}
)
I i
i
,,-.-,,,.,-,,,n,..,.,,-,
n.,,,,.,,,-,,..,,,-,-,n...
..~,-ann-
i QUESTION 24 4
In the equation on page 4-4, what one-sided tolerance factor will be used with the uncertainty fraction? Does the (ANO-2) 2.2% uncertainty fraction include the uncertainties resulting from the extrapolation to unmonitored locations and other Process Computer approximations not accounted for in the model error?
W41P. is the uncertainty factor to be used in Fg?
ANSWER 24: This question addresses the reliability factor to be used in the monitoring mode by way of the equation on page 4-4.
Since the original submittal of MSS-NAl-P, additional ANO-2 power distribution comparisons have been performed. These new state-points are listed in the attached Table A.24.1 and the new com-parisons are given in Table A.24.2.
The maximum ANO-2 model uncertainties by map were compared to the model uncertainties of ANO-1 (see Answer 22). The largest rodel uncertainty of 4.6% of AN0-2 was smaller than the bounding uncertainty used in the ANO-1 analysis. MSS will, therefore, usa the conserva-tive ANO-1 bounding values of RFpg = 0.10 and RFAH = 0.057.
These quantities include all uncertainties associated with 1
the use of MSS calculational model in monitoring applications.
4 j
i a
1 l
I l
l
TABLE A.24.1 REACTOR STATE POINTS ARKANSAS NUCLEAR ONE - UNIT 2 Cycle 1 Map Boron Exposure Power Rod Position i
Number Date ppm GWD/MTU (inches withdrawn) 1 1/21/79 836
.133 20 AR0*
2 6/30/79 722
.338 50 ARO 3
1/24/80 620 2.543 100 ARO 4
8/15/80' 440 6.745 100 ARO 4
1 i
j Cycle 2 l
8/03/81 785
.639 100 AR0 2
8/31/81 730 1.459 97 ARO 4
4 i
OARO = All Control Rods Withdrawn 1
f L
i 4
2
.y y
my.,
...,.,,--.7,
_e._,
...-....~e.
TABLE A.24.2 MODEL UNCERTAIN'lIES FOR ANO-2 RMS(t) (%)
CYCLE 1
1 1
1 2
2 MAP 1
2 3
4 1
2 Level 5
2.6 1.8 3.3 4.4 1.4 4.6 4
1.8 1.3 3.4 3.1 3.7 3.2 3
1.6 1.4 2.0 2.1 1.0 2.5 2
1.7 1.4 2.5 1.8 2.1 1.6 1
1.7 1.3 2.8 2.7 3.1 3.5 Integral 2.0 2.2 1.9 1.4 0.9 1.4 f
ERRATA AND REVISIONS MSS-NAl-P 1.
A typographical error was found on Page 3-46.
Two group 5 rods were marked as group 3.
A new page 3-46 is enclosed.
2.
Revisions to Pages 3-3, 3-87, 3-88, 3-112, 3-113 and Section 3.2 of the AND-1 Benchmark Section were made in response to NRC Questions.
3.
Revisions to Pages 3-3, 3-68 and Section 3.2 of the ANO-2 Benchmark Section were made in response to NRC Questions.
l l
l l
FIGURE 3.9 CONTROL ROD IDENTIFICATION CO:: VENT 10';
AN01. CYCLE 1 I
2 3
4 5
6 7
8 9
10 11 12 13 14 15 A
I 7
4 7
- PATTERN.A g
4 7
4
- PATT E R:t B
'C 5
3 3
5 D
4 8
6 8
7 7
E 5
6 1
1 6
5 7
.7 F
8 2
2 2
8 4
4 G
3 1
5 5
1 3
H 4
6 2
7 2
6 7
7 v.
K 3
1 5
5 1'
3 L
8 2
2 2
8 4
4 M
5 6
1 1
6 5
l N
8 6
8 7
0 5
3 3
5 7
4 7
P 4
7 4
R e
3-46
,.---y_,.-,-,-._,--.,y--,-
-,-.m_ _ -,. _..,.,,.
A THE FOLLOWING PAGE REPLACES PAGE 3-3 0F THE AN0-1 BENCHMARK SECTION.
/
l l
4 rw--
.v r-n.~.-
r
-,g-.
n..
_ TABLE 3.1 RELIABILITY FACTORS FOR ANO-1 BENCHMARK CALCULATIONS Parameter Reliability Factor Bias F
RFpg = 0.10
- 0.02 to + 0.02 (refer to Figure 3.40)
F RF
= 0.057 0
AH FH Rod Worth RFR0DS = 0.05 0
0 Temperature RFM = 4.0 PCM/ F 0
Coefficient Doppler RFDC = 0.10 0
Coefficient Doppler RFDD = 0.20 0
Defect Boror. '4 rth RFB = 0.05 0
0 Delayed RF, = 0.03 0
Neutron Parameters RF, = 0.03 0
l l
l i
3-3 c
e FIGURE 3.41 BOUNDED UNCERTAINTIES FOR AN01 POOLED FROM CYCLES
? AND 3
",. ji...$
d5S.L !!E2 El[.i2 fY N
..i
.'.l
. O, !.
L.u.
.u
. L.
..a-:..
u.
.di
- p
- ;p p; ii i.N, 1-h- gli I
i ip i;
-i Jpfil!
h; i.t' n:.
.; i ii. ;
i.
1
..l..
. J.t.
. l..
.c.
- t. u!s.iliti.
@t; y, * ;i
. i. -
i:
- a. u..;a gr.. ua aa a'-
I?
-i:
- I t
e
?
1:
- 4.,.
- xl. a'.j
,i::
l.
8-
!., i.:
g i
..a
,.: s.
I
..s.
i s!. Y.
,![:
-]'!;iil 3g
't I'
l.
- l-u' 'h.14 y
- u. ' '
, "q e
a g w
.-p
- 2..a.
- !....._ a.
J.r..
4 p
rp g
i a
E_a
.a.
1:
O liL
!it ik
_. m..
. !S.
. I..~..
. l.. _.J.i
...u. a m- *
!m :"
h g
4 L
l i
9 h
F
- .-d'-
t m
_.4
'.i L. ~
.u
.J.
.U.
-.;.4 x
I-
- t;
- t c
2.
3 i
i e.
E
.11... l;.43 w,..
i C)
' I
.! M' I i.4
. u.
C;.
. ; a.
b8 E
~ 6 t...
..I
..M..
... 1.
I;.H.L L
R i
8" h
g g
g gp y
6 I'
1' A.
6
.J 42 h
., b
- n. a;.....';.
- .u, d.a
- n. ;;.,.;;. '.
a t
UJ f.
.;u C
~
O l
i v
'. (*)
.[.
a..
- .; i l
.11.l.::;
- a.ai.. ;.;.
E
.;.11
.R 1.
I
.la.
.4 L.a.
..A.
O
- I t..
,'I tl.
- 4 w
,. )
V 1d
- li... [.1.:.; d a
g 4
l i.
$: b
.l.lli; g
~
6 A.
.1
.. q
. 5{
a.:
O E
. f..
.. lu..
u i
- 3.
t n
1*-
c g,,. g 3
m l'
s
..i.Lu h.i.a a.
. L e
.:n
.a
..3.
..t 1'
a..
.a.
s
=
j.
'.aF y
.2'.. a a.
2
=
1
..22
.0 L;
.i.
a
-..; L...d...,. '1.1 4.a
.0; 8
1%
. I 1.,
.s.
.a..
l.
g
.1;.
'.I 4..
.1;.
- i. 4 I
g g.-
f 8
9 g
f i
1
.i i
....n
..-.u.
4..a a.
. aa..:..n.,J..
.1,..
i
.i
..i
.r
-1..
i n
.i i,
l t
0 10 20 30 40 50 60 70 80 90' 100 PERCENTAGE OF CORE HEIGHT 9
S" g
i i
I
_ TABLE 3.19 POOLED STANDARD DEVIATIONS i
FOR ANO-1 CYCLES 2 AND 3 i
Axial Number of SModel (t)
Level Comparisons
(%)
K RF+
7 728 5.1 1.742 9.04*
i 6
832 4.0 1.736 7.26 4
5 988 4.8 1.728 8.49 I
4 988 3.4 1.728 6.32 3
988 4.9 1.728 8.65 2
988 5.1 1.728 8.97
)
1 988 4.4 1.728 7.85
[
i 4
4 l
l
+Using equation on Page 3-110
- See answer to Question 20 4
I i
l l
l' 3-88
FIGURE 3.52 F"0 RG.IABILITY FACTOR FOR AM01
.u.
i_ l l
I
..3 -
3.
.g.
- g.
?
~.
i
~'
s-
~'
n.
~r
.1
~~
~ -~~~-
~
~
~,-- v *\\
nf r-s4 m-
- rt i
,.j t
l 3
t i
~
. '. ! 9,. 15... a
...- g..
i,
+
I
's'
'.p.-.g..
.4 9
,9
,l l
==
.y. _
0,
)
,f
.g4 I
.'..' I.
p l
4
.=
a N,
39 A..
4 g
9 m
qg..
ii.e i
p w
i
,..,f D..
d..,
6 eg
.% }
I O
., p u
F
.I.
i i
4 si.
...i.
e j
....f... *..
.p
...jg.
u.
e l
y a
r s,
's
.'I i-t.
M m-e e
I:
..g..
.* T* ** -""4
- U..
- f*
?
. J **' * *
.=.J f**
T f.
I.,**
t i.....
N 8
- e. e i
i
,{
.I i
,e
- i l'
e g
aC
- i 3
.,. _ g........
(.....-
==
+.
J l
+
m cc i.:
3
~
m 6
p.l -..
.; J y
H.; j :
l'; g [
g :,:.
V (Cr '
gl l.
p
,j l
i Q
i
.{.
.i
., g
.,. 3. '
. g r.,
\\'I.
I a..
?..
}
.- ~.
I, t
.. i l
-. 5....-..,
+y.
, y; q:
H..i'
.t.:
- ]
H...L p;
.g.
.. ~
, i:
J,,9.
- 3.,
..\\..
.4....
-i i ;.,..
~
~**
-.' r
~
J i'
r-
- ;r
- st
?*
?
r
- j h.
- r i
e L :-
r.
.l il.
,q.
- .
- q'.,
.;i p
- ,i :. 't. *1.
--- - r--.,.-
- p:---
pr 1
f.
..i
.i i
r
- ,. s...i,
.1;.
-i.,
0 10
- 20 30 40 50 60 70 80 90 100
. ~
Data from Page 3-88 PERCENTAGE OF CORE HEIGHT o
r TABLE 3.25 N
Fg RELIABILITY FACTORS FOR AN01 Fh Reliability S
Core Model Height Uncertainty Factor
(%)
(%)
(%)
100 5.7 10.0 90 5.7 10.0 80 5.7 10.0 70 5.7 10.0 60 5.7 10.0 50 5.7 10.0 40 5.7 10.0 30 5.7 10.0 20 5.7 10.0 10 5.7 10.0 0
5.7 10.0 I
l 3-113 l
h a
THE FOLLOWING TWO PAGES COMPLETELY SUPERCEDE SECTION 3.2 0F THE ANO-1 BENCHMARK SECTION.
l i
1 i
s
3.2 Temperature Coefficient This section investigates the MSS model bias and uncertainty for calculating temperature coefficients.
Measurements of the isothermal temperature coefficient at HZP can be adequately made with a reactivity computer if the rate of temperature change is slow and there is no rod motion.
Most of the ANO-1 measurements of the isothermal temperature coefficient (ITC) were obtained from Cycle 1.
There is 'nly one qualified measure-ment for Cycle 2, and one for Cycle 3.
Five ANO-2 comparisons are also included. The comparisons of the measured and calculated ITC's are pre-sented in Table 3.5.
Because insufficient data was available to provide a good estimate of bias and standard deviation, a rigorous statistical approach was not taken to determine the reliability factor.
Instead, a conservetive 0
value of 4 PCM/ F was assigned as the reliability factor and a value 0
of 0 was assigned as the bias. The conservatism of a 4 PCM/ F relia-bility factor on ITC was justified on the bas 3 that all of the aITC's D
were less than or equal to 4.0 PCM/ F and all but one did not exceed 0
2.6 PCM/ F.
Furthermore, the assigned 4.0 PCM/0F reliability factor 0
bounded the 3.1 PCM/ F reliability factor based on the AN0-1 and ANO-2 data listed in Table 3.5.
3-10 t
TABLE 3.E MEASURED AND CALCULATED IS0THER! AL TEMPERATURE C0EFFICIENTS FOR ARKANSAS NUCLEAR ONE 0
0 Controlling Rod Boron ITC(PCM/ F)
AITC(PCM/ F)
, Reactor Cycle Bank Position ppm Meas.
Calc.
(Calc. - Meas.)
Unit 1 1
7 = 73%
1566
+ 4.3
+ 1.7
- 2.6 6 = 74%
1442
+ 2.8
+ 1.8
- 1. 0 5 = 21%
1270
- 3.6
- 4.8
- 1.2 4 = 39%
1183
- 6.2
- 6.4
- 0.2 2
5 4%
1050
- /.6
- 7.2
+ 0.4 3
7 = 75*4 1350
+ 1.8
+ 0.7
- 1.1 L'ait 2 1
6 = 80%
1004
+ 0.3
- 0.5
- 0.8 2 = 65%
808
- 4.8
- 4.6
+ 0.2 8 = 80%
657
-10.7
- 9.7
+ 1.0 2
ARO 1230
+ 0.4
- 0.7
- 1.1 3 = 40%
1040
- 3.9
- 3.9 0.0 n = 11, K95/35 = 2.815 13.5 = I (AITC )2 j
i=1 o=
Nf13.5/1f = 1.1 0
RF = K*c = 2.815
- 1.1 = 3.1 PCM/ F 3-11
THE FOLLOWING PAGE REPLAC:S PAGE 3.3 0F THE ANO-2 BENCHMARK SECTION.
4
TABLE 3.1 RELIABILITY FACTORS FOR ANO-2 BENCHMARK CALCULATIONS Parameter Reliability Factor Bias N
F RFpg = 0.10 0
F RF
= 0.057 0
t2H Rod Worth RFR0DS = 0.05 0
Temperature RF =4.0KMN 0
M Coefficient Doppler RFD = 0.10,
O Coefficient Doppler Defect RF
= 0.20 0
Delayed RF = 0.03 0
6 Neutron Parameters RF
= 0.03 0
p 3-3
THE FOLLOWING TWO PAGES REPLACE SECTION 3.2 0F THE ANO-2 BENCHMARK SECTION.
b
l
?
i 3.2 Temperature Coefficient This section investigates the MSS model bias and uncertainty for calculating temperature coefficients. Analysis for Ah0-2 is identi-cal for AN0-1 and is reproduced bel u for completeness.
Measurements of the isothermal temperature coefficient at HZP can be l
adequately made with a reactivity computer if the rate of temperature change is : low and there is no rod motion.
t Most of the AN0-1 measurements of the isothermal temperature coefficient 1
J (ITC) were obtained from Cycle 1.
There is only one qualified measure-I ment for Cycle 2, and one for Cycle 3.
Five AN0-2 comparisons are also included. The comparisons of the AN0-1 measured and calculated ITC's are presented in Table 3.5.
{
Because insufficent data was available to provide a good estimate of bias and standard deviation, a rigorous statistical approach was not taken to determine the realiability factor.
Instead, a conservative value of 4 PCM/ F was assigned as the reliability factor and a value of 0
i 0 was essigned as the bias. The conservatism of a 4 PCM/ F reliability factor on ITC was justified on the basis that all of the AITC's were less than 2.6 PCM/0F, and all but one did not exceed 1.2 K? OF.
Further-0 more, the assigned 4.0 PCM/0F reliability factor bound'd the 3.1 PCM/ F reliability factor suggested by the ANO-1 and ANO-2 data listed in Table 3.5.
3-8 i
TABLE 3.5 MEASURED AND CALCULATED ISOTHERMAL TEMPERATURE C0EFFICIENTS FOR AkKANSAS NUCLEAR ONE 0
Controlling Rod Boron ITC(PCN/ F)
AITC(PCM/0F Reactor Cycle Bank Positien ppm Meas.
Calc.
(Calc. - Meas.)
Unit 1 1
7 = 73%
1566
+ 4.3
+ 1.7
- 2.6 5 = 74%
1442
+ 2.8
+ 1.8
- 1.0 5 = 21%
1270
- 3.6
- 4.8
- 1.2 4 = 39%
1183
- 6.2
- 6.4
- 0.2 2
5 = 4%
1050
- 7.6
- 7.2
+ 0.4 3
7 = 75%
1350
+ 1. 8
+ 0.7
- 1.1 Unit 2 1
6 = 80%
1004
+ 0.3
- 0.5
- 0.8 2 = 65%
808
- 4.8
- 4.6
+ 0.2 B = 80%
657
-10.7
- 9.7
+ 1.0 2
AR0 1230
+ 0.4
- 0.7
- 1.1 3 = 40%
1040
- 3.9
- 3.9 0.0 l
i n = 11, K95/95 = 2.815 13.5=$
(AITC )2 j
I i=1 o = \\/13.5/11
= 1.1 RF = K*o = 2.815
- 1.1 = 3.1 3-9
,h.
_.2..
A h__h-A.__
__su__
W
.,t
+.-m,a m*
.+
eA4
__1aa
.:E FOLLOWIt4G PAGE REPLACES PAGE 3-64 0F THE At40-2 BENCHMARK SECTION l
(
l.
i f
I i
I
Reliability factors calculated as described above (RFpg =.064) are shown as a continuous function of core height in Figure 3.25.
Nevertheless, the more conservative RFFQ = 0.10 documented in the ANO-1 benchmark analysis will be used in safety related analyses.
l l
Y'
-(
1
{
1 1
1 3-64
_....