ML20116P324
| ML20116P324 | |
| Person / Time | |
|---|---|
| Site: | Fort Calhoun  |
| Issue date: | 06/12/1983 |
| From: | Biffer J, Jonsson A, Terney W POWER SYSTEMS, INC. |
| To: | |
| Shared Package | |
| ML20116P302 | List: |
| References | |
| TIS-7405, NUDOCS 9608230249 | |
| Download: ML20116P324 (32) | |
Text
...
{,010,8'0001:50 ID:SPF-301 Series FAX:
PAGE 2
i ~
LIC-96-010a l
Attachment ',
i l
THE C-E CECOR FIXED IN-CORE 1
\\
l DETECTOR ANALYSIS SYSTEM l
l l
i W. B. TERNEY i
J. L. BIFFER C.O.DECHAND j
A.JONSSON l
R. M. VERSLUIS l
Nuclear Power Systems
{
Cornbustion Engineering. loc,
{
Windsor. Connecticut 1
i 1
l' Presented at AMERICAN NUCLEAR SOCIETY ANNUAL SUMMER MEETING June 12-17,1983 Detroit, Michigan 1
H POWER SYSTEMS
)
i
(
9608230249 960819 1
1 PDR ADOCK 0500 5
Tis 7405 i
i 01 08 '00 01:50 ID:SPF-301 Series FAX:
PAGE 3
l l
l l
THE C-E CECOR FIXED IN-CORE DETECTOR ANALYSIS SYSTEM
- l. INTRODUCTlON distributions are obtained by a few mode Fourier expan.
sion which matches the assembly powers at each detec-The Combustion Engineering (C-E) CECOR System tor level and uses axial boundary conditions derived from provides a method of synthesizing detailed three dimen.
3.D coarse mesh calculations. Subsequently, peak pin-1 sional assembly and peak pin power distributions from powers in each assembly as a function of height are ob.
)
the signals of a hmited number of fixed, self powered tained by multiplying the assembly power at each height neutron sensitive in core detectors. The actual synthesis by assembly normalized local peaking factors,(F ).These p
is done in the CECOR program using libraries of pre-pin to-box factors are obtained from fine-mesh, diffu.
calculated coefficients generated by the standard C-E dif.
sion theory calculations with the appropriate transport fusion and transport methods and codes. The system is theory corrections. The CECOR algorithms and required used to fulfill the required startup testing, monitoring and libra.y are described in detail in Section 111.
surveillance functions as well as to provide the base of The accuracy of the system has been evaluated by com-j measurement information for core follow and methods bining the uncertainty in measured power for in-verification.
strumented locations obtained from a large data base with The fixed in core detectors consist of self-powered the synthesis uncertainty for extrapolation to uninstru-rhodium neutron detectors. Each detector segment is 40 mented portions of the core.(6) This analysis is aho em long and either four or five segments are equally described in Section III. The overall uncertainty for the spaced axially in each instrumented assembly. Approxi-CECOR system, which has been approved for C-E licen.
mately 25% of the assemblies arc instrumented. A sing submittals, has been quantified on the basis that there description of the instrumentation hardware is given in is a 95% probability that at least 95% of the true F.
q the following section, together with information on the Fxy, and Fr values will be less than the values obtained reliability and reproducibility of the measured data.
from CECOR plus 6.2, 5.3 and 6.0%, respectively.
The synthem of signals into full-core three dimensional A basic function of the fixed in core system is to satisfy power distributions is carried out by the off-line CECOR monitoring and surveillance requirements. 'lbrough computer program. CECOR has evolved from the earlier CFCOR. it regularly provides both full-cole.
INCA program (3 5) CECOR calculates the assembly 3-dimensional power distributions and thc limiting values of peak F,, Fxy, F, as well as evaluations of the core power in cach instrumented fuel assembly through the q
use of prefit power-to-signal assembly conversion factors tilt and axial shape mdices. On current C L plants it is (W') The W' are provided for all instrumented also used to verify the operation of the COLSS and CPC assembbes in the core. They are obtained from fine mesh, limii monitoring and protection systems.W in addition, multi group diffusion theory calculations, with ap.
the sys:cm is used to produce measured power distribu-propriate transport theory corrections. Full-core planar tions for comparison with calculated power distributions power distributions at each detector level arc obtained calculated in core follow analyses. Core follow analyses through the use of precalculated avetage coupling coef-are perlormed as a couomer service to determine if the ficients, < ec >, which are the inverse ratio of the power core is operating as expected, and if not, to provide key in an uninstrumented assembly to tiie average power in diagnostic input to deictmining the causes of the dif.
its neighboring assembhen. The coupling coefficients arc ference. Measured power distributions are also used io obtained from two or three-dimensional dif fusion theory vahdale methods and quantify calculation and measure.
l calculations. The resulting set of equations is solved for ment uncertainties. These various applications are the power in uninstrumented assemblics. Axial power di> cussed further in Section IV.
t 1
01l08'0001:51
[D:SPF-301 Series FAX:
PAGE d
- 11. SYSTEM HARDWARii DiiSCRil'1 ION AND rhodium emitters 40 cm long and 0.045 cm in diameter.
OPERATIONAL EXPERIENCE Fach detector is an integral segment of a piece of coaxial, l
mineral insulated cabic having an in.conel shcath and a l
A. Harosare Description 0.16 em diameter central conductor. A string of 4 or 5 The fixed in-core instrument system hardware includes SPNDs is contained in an instrument assembly, which the instrument assemblies, guide paths and other suppor.
mechanically protects the detectors by means or an outer ting uructures, field cabling and connectors, and signal sheath, and positions them accurately. The assembly ter-conditioning and processing equipment. An overview is minates electrically in a multi-pin electrical connector, shown in liigure I for C.E's bottom. mounted, fixed plus which mates with the field cable connector. Mechanical.
moveable instrumented system, which is part of the 3800 ly it terminates in a seal plug which brings out the clec.
MWt System 80. Figure 2 shows the axial position of the trical leads through the primary pressure boundary. This fne detectors in a System 80 instrument assembly relative is shown in Figure 3 for C-E's bottom mounted system.
to the active core as well as the locations of the in.
Also shown is the calibration tube. It is a dry thimble strumented fuel anemblies in the core, Approximately which provides access for a movab!c detector in the com.
25% of the assemblies are instrumented. Each instrument bined fixed / movable in-core detector systems supplied on assembly is inserted in a guide tube at the center of the C E's 3400 and 3800 MWt NSSS designs. Individual fuel assembly. The bottom-mounted system permits in-detectors are wrapped around the calibration tube within strumentation of both rodded and unrodded fuel the outer sheath, assemblics. Earlier plants have a top mounted design.
Rhodium.103 is the neutron. sensitive nuclide. It is
't he C-E fixed iacore instrumentation system employs mostly a thermal neutron detector (it has a theimal-the so-called self powered neutron detector (SPNDS) with neutron cross section of about 100 barns) with an epither-Plant Computer N*
C::s Oo Printer c::::3
%g%
Orive Machine Instrument Path Assembly Transfer D a V9 4/
i, n
0 Fuel 0 {,
@y
)g
,y
"""d'*-
tiectricai
/
Connector
'b 0
V-w q
0 Operator Core Pressure intedace
]
Seal
/
Containment Watt
/
l Fig 1: Overusew ofIsxed/ movable in core metrument system 2
, 01l09 '00 01t52 1D:SPF-301 Series FAX:
PAGE 5
I
=
mal e ntribution of about 15%. After absorbing a r r q ,1 3, L. 1 neutron, the rhodium 104 nucleus decays, cmitting gam-
'i 8
,;r-l p,P
~4 ;
.{' --[* l ma and beta radiation. A fraction of the energetic betas 7,"
.1 i
(electrons) escape the emitter and are coliceted in the
~
, ! *. p.
detector sheath, thus creating an "clectron pump". If the z j
' T., *, f.
l.
circuit between the lead wire and the shcath is closed, a j
current which is proportional to the neutron absos ption
. J..... -
w.
~
.l -
- l,
" a
- s,--t-rate is measured by the data acquisition system. About a _1 J.f;i.~ -"Ti [i i.a i.
6% of the betas are emitted promptly after neutron ab.
sorption; the remainder decap with a half life of 42
,j--.. --%'1,:-'
- " MiW/.'00"'"
- seconds. For npplications requiring a prompt response,
- i.,.
1 t.r
- c,1glfg;,a,*.5"a C E has constructed a dynamic compensation filter n
T~h N~.,
.i, (Reference 8).
..e Cr Al l'
The small electrical current (about 2 micro 3mps at Outlet ThormocouDie nominal conditions)is conducted through well-shielded 0
Cahbration cabe the plant computer data acquisition system.
s 70%
Tube with Specially designed, low impedance, high gain current to-
~
voltage amplifiers, which also perform noise filtering, et et clo, 2
produce a 010 Volt output. This is multiplexed, and con-(c re
{
verted from an analog to digital signal. The signalis then p
Height l
{
f T pical further processed in the plant computer to correct for
/
background, emitter burnup effects, and if needed, Axlal \\
N3)"F o.
dynamically compensated for beta decay effects.
Flux istributio
t0%
Further processing consists of signal to power conver-e.____..
sion, power distribution synthesis, and calculation of Local Flux margins to core operating limits. These calculations are Detectors performed partially on the plant computer, on line, and partially off line in CECOR. with the split depending on 17g. 2. System 80 fused sn core instrumentation NSSS vintage.
loca tions Electrical Connector Detector Seal Plug Signal Lead Wires -
/ _
3, m,._
f
/r
~ ~
'/ ahbration Tube
-C Outer Sheath _.Zm Bullet Nose Dry Detector Tube 0 270 0 D.
-Rhodium Detector Signal Cable Sheath Signal Cable Lead Wire 0 062 Dia Ref Incone!600
-- - Emitt er 0 0t0 0ia Sheath O.D.-
bg3gM 1h1U Liny A
}
3 0 062 Nom Dia 1
Outer Sheath Tube Khg XQu.rt11p i
Dackground
'"O 44 t O D Ref Ottector.
Emitter insulation 0.062 Dia Ref
- Filler Wire 0 062 Dia
.. Epoxy Seal A I,0 '
+ 40 Cm Nom - --
Fig. 3: In core detector anembly 3
j g ~, g {.cy ~~ 3 b 5 F-305 Shies FAX:
PAGE 6
~
~
[
1 The mechanical layout of th; instrument system and C.F's experience with the system spans 34 cycles of its support structure is shown in Figure 4 for C E's operation to date. Some of the early experience is found botiom. mounted system. During refueling, while the in Rclerence (9). As is to be expected, a number of design water is at the vessel flange, the entire support frame in-and manufacturing improvements were made to benefit cluding the moveable detector drive and transfer machine-from operating experience The primary causes of detec-ty is inoved out of the way, after disconnecting the dry tor failure are breah in the lead wire and loss of clec-thimbic>. This allows access to the seal table to disman-trical resistance between lead wire and sheath due to tie the seals and withdraw the instrument assemblies by moisture.
cbout 25 feet so the connectors and thimbles may be kept Since the detector assembhes must be inserted and dry when the water levelis raised to near the operating removed through tightly curved guide paths in the top-floor. This wet seal table arrangement during refueling mounted systems employed in the pre-ANO-2 plants, they permits the removal and disposal of instrument awemblics must be flexible and yet strong. This has resulted in a i
under water.
design for these plants with a permeable, flexible outer 1
sheath and an extra internal member to increase tensile s
B. Operational Experience strength. This design has overcome earlier problems of 1
In terms of its most important performance para.
breaking instrument assemblies during removal, 1
meters, accuracy and availability, the C-E fixed in-core Because the sheath is permeable, the individual detec-instrument system has an excellent record. By analyzing tors are in contact with the coolant, which increases the i].
a sieable data base of measured power distributions, C E chances of moisture ingress through corrosion. While l
has licensed the following 95/95 uncertaintics(6);
moisture ingress still can occur it has been greatly re.
duced by removing any welds, brazes or splices from the F
6.2Ve detector external surface (" integral" detector) and careful q
l'r or Fm C.0%
materials management and QA by the vendor.
j F
5.3 %
Starting with Arkansas Nuclear One Unit 2 (ANO 2),
j xY all top mounted in-core systems have combination fixed-This analysis is described in more detail in Section moveable in core detector assemblies and less complex III.C As for availability, the in core detector system has guide paths, The solid sheath instrument assembly which not limited core power performance in any plant. When is less flexible than the permeable sheath instrument l
Judging this availability performance, it must bc borne assembly design has shown good performance in three
]
in mmd that m C E reactors, fixed incores not only fulfill cycles of operation at ANO-2.
l the power distribution surveillance function, but also Another improvement initiated in ANO-2 is to ac.
2 monitor the kw/ft limit and in doing so, provide con.
complish the current to voltage conversion required for
]
siderably more margin to th LOCA limit than an ex-core signal multiplexing by active emplifiers, rather than drop-monitoring system could obtain, ping resistors. Dropping resistors are simple devices and i
free from drift, but they offer a high terminating im.
p er dgn h an @
pc a e e
Movable Detector Drive Machines (2) lower impedance, thereby reducing leakage currents that j
/
may occur in older detectors, and better noise filtering l
- 4
]
capabilities.
j Transfer Machines f
111. DESCRIPTION OF CECOR ANALYTICAL l
[h METHODS l
instrument t
A. Introduction Seal Tablo y [
-+v l
A description of C E's originalin core detector signal i
analysis system, INCA, was published in 1%9.(l) Since then, many improvements have been made and the code has evolved to the present CECOR system (2 5), it pro-vides a method to synthesize a limited number of fixed, self-powered, in core detector readings into detailed full-g t.
s core radial and axial distributions for assembly and peak
-instrument pin powers.
Gude Basically, CECOR uses prefit data from detailed, two-hb s dimensional, multi group diffusion theory calculations eg f
to convert detector readings to local box-power values through the use of power to signal (W') factors. These Fig. 4: Mechanicallaynut ofinstrument system are provided (or all instrumented assemblics in the core.
4
01/08 '00 01:53 ID:SPF-301 Series FAX:
PAGE 7
The planar power distributions at each detector level are where obtamed through the use of pre calculated average cou-Q accumulated charge pling coefficients, < cc >, which relate the power in an So the initial sensitivity for the detector,
=
uninstrumented assembly to the average power in its supplied by the vendor.
neighboring auemblics. The axial power distributions are Q(t) the fraction of the total available charge 1
=
obtained by a few mode Tourier expansion which matches Q3" for the detector at time t. where Q. is the box powers at each detector level. Subsequently, the supplied by the vendor.
pcak pin powers in each assembly are obtained by the n
=
the empirically determined fitting application of pin to box (F ) factors which are pre.
parameter, p
criculated using fine mesh, two-dimensional transport.
theory corrected diffusion theory.
The W' are calculated for each assembly with a detec.
The first section of this chapter describes the analytical tor as a function of life from transport cortected fine-j methods in detail, starting with the conversion of signal mesh,2-D planar depletion calculations using multi. group to power, followed by a dir'ission of the types of edited diffusion theory. T he W' are then fit as a function of I
output quantities. The se6ond section of the chapter assembly burnup for each detector location and take in.
discuues the content and construction of the CECOR to account whether an assembly is rodded or not. 'Ibe library. The final section describes the analysis used to behavior of W' with burnup is illustrated in Figuie 5 for quantify the accuracy of the system.
typical auemblics in the different batches of a first cycle Corc.
- 11. CECOR Algorithms UO 1.
Signal to Box-Power Conversion g
Tbc power integral Pin over a detector level (n)in an a
1.08 instrumented assembly (i)is obtained by multiplyingthe vi background corrected, integrated detector signal (I,h )
h 1.06 Typicaf Batch A Assembly n
by constants measured and supplied by the vendor, and l
constants calculated from fine mesh, two dimensional, 1 1 04 Typical Batch B Assembly multi-group diffusion theory calculations. The signal to-
}
box power conversion for an instrumented assembly i at
$ 102 Typical Batch detector level n is expressed by:
~
LOO
/ C Assembly P. = laf W,.
(l) 0 2
5 6
8 10 12 ff~i6 Assembly Average Burnup.10$ MWD /MTU CALIB WL Fig. S Power to rhodsum utgnal conversion (artars vs W,.
=
(2) assembly average burnup 2.
Planar Power Distributions W', r P,JJ J o.** (N)f(r..E)dr dE (3)
At each detector level the powers in the uru.nstrumented g y
assemblies are determined from the measured powers in instrumented assemblies. This is accomplished through the use of radial coupling coefficients, and where CAllB is a factor to convert from surface flux The average coupling coefficient of an assembly at to Rhodium activation and the double integral represents detector level n is defined as the ratio of the escrage power the Rhodium activation in the detector.
in the surrounding boxes to the power in the assembly The initial dctector sensitivity, S, relates the current, itself. For the case of a cell surrounded by N, boxes, this is I, produced by the detector to the incident neutron flux at the surface of the detector. The initial sensitivity and N,
relative calibration of the sensitivity are supplied by the I
vendor. The sensitivity as a function of depletion for a j.1 g',
given detector is determined by
<cc>
=
(5)
N P,.
S"'(t) = S. (1 - Q(t))
(4) i
%ere P jn denotes the power in boxes neighboring cell i.
5
01 08 '00 01:5d 1D:SPF-301 Seties FP.X:
PAGE 8
T he average coupling coefficient given by Fq. (5) is Knowing the average coupling coefficient values at a inwnutive to mild power tilts acrow the core, in addi-given detector level, the power in any uninstrumented box tion it has been found that local perturbation effects in f may be obtained from:
a cell (e g., from todd cucntially affect the coupling coef.
Inient for only the subject box: i.e.,
Ng I
j=1 Py
.h*
p,=
(7) d <cc>,.
=
dk (6)
N <CC#
- g I
f N, M 8 Originally (l). INCA worked on an octant or quarter core basis, where symmetry was assumed and all instru.
as can be der.ived from the modified one group diffus. ion ment readings were reflected into the octant of interest.
equation.(5) Eq. (6) has,in fact, been tested and found As a consequence, each uninstrumented box was sur.
to accurately predict such perturbations m < cc >in-rounded by four instrumented boxes and Eq. (7) itself -
In practice, the coefficients are pre-calculated from Eq.
could be solved directly for each uninstrumented box. In (5) using bux power distributions obtained from either the actual CECOR full core calculation, each 2 D fine mesh or 3 D coarse mesh depletion calculations uninstrumented box is not surrounded by four in-using multi group diffusion theory, and are then fit as strumented neighbors, in fact, many boxes have only one i
a function of burnup for each assembly. Explicit sets of or two instrumented neighbors; and some have none. As burnup fitted coefficients are prepared for each control
. a result Eq. (7) cannot be solved directly for each rod configuration or rod bank region. The actual values uninstrumented box alone, since some of the neighbor.
l of < cc > n und in CECOR are obtained from these sets ing powers are also unknown. However, at each level n.
of fitted coefficients using average assembly burnup the equation may be written for every uninstrumented values obtained by integration over detector level n. A box, and then rearranged with the unknown powers on typical example for an assembly in a rodded and unrodd.
one side of the equation and the known powers on the cd condition is shown in Figure 6.
other. When all the equations for each uninstrumented -
box are written and grouped, a matrix equation results:
A P = S_
(8)
I8 The A matrix is symmetric and very sparsc. In each 8
8 row the diagonal element is the product of the number of neighbors times the coupling coefficient. The off-diagonal elements are all zero except for the 3 4 positions 16 of the uninstrumented neighbors of the assembly, when the value is 1. If uninstrumented assemblies are sur.
-f gj [
N rounded by four instrumented assemblies, this new 1
CECOR formulation reverts back to the originalINCA scheme.
]
The solution for the unknown powers is obtained by solv-F 2
ing Eq- (8). Symbolically:
P = A-8 S, (9) g g to Unrodded This system is solved at each detector level.
~
^
Since the matrix can be large (on the order of 200x200).
0.8 cfficient solution routines are needed. Because of speed and memory requirements, a modification of the con-jugate gradient method (10)which takes advantage of the unique nature of the A matrix is used.
A feat:are of this formulation is that if an instrument 06
^
~
0 20 0
6O 80 fails, the box is simply treated as uninstrumented and the Bumup.10' MWD /MTU resulting larger set of equations is solved. For this to work, average coupling coefficients are precalculated and Fig. & Typical accrage coupling coc//scieni behopior provided for all boxes, uninstrumented or instrumented.
6
01b8'0001:5d ID:SPF-301Serie[
Ff4X:
PFE 9
This concept also provides a w:y of checking detectors, An important parameter in the Fourier expansion is since the actual detector power may be compared to the the extrapolatior distance (d) or fitting parameter H. It deduced power obtained by treating the box as is a function of core life and is radial core zone depen.
uninstrumented if the difference exceeds a predetermined dent. B is obtained from nominal 3 D design depletion criterion, it is an indication that either the detector has studies. it is chosen for each zone so as to minimize the failed or a local perturbation may have occurred.
error in fitting the axial peak to average power ratio in each assembly. It is a well-behaved function of burnup, 3.
Axial Power Distribution as illustrated in Figure 7.
thing the power in the assembly at each detector level, the detailed axial shape can be synthestred by using a Fourict expansion. This axial fitting technique is des.
cribed in detailin Reference 3. Basically, the axial power 0 87 distribution in an assembly is represented as the sum of the first N Fourier modes 0 86 0 85 N
o gg P,(e) -
I
- a. sin nnB,z (10) g n-I g 0 83 80 0.82 where P;(z)is the power per unit length for assembly i.
k a are the unknown combining coefficients, z is the axi.
2 0 81 Outer FuelZone n
al elevation in fraction of the core height, H, and B; is ch 080-the fitting parameter given by
[ O 79 j
E ti O 0.78 11 B. -
(II)
H + 2d, E77 ~ lnnerFuelZone 0.76 0 75 Note that d is essentially an extrapolation distance. The 0
1 2
3 4
5 6
7 8
9 N combining coefficients, and thus the axial power Core Average Burnup.108 MWD /MTU distributions are obtained by matching the box power at i
cach level to the integral of Eq. (10) over the axial extent of each of the N detectors. Thus, Fig. 7: Typical cycle 2 CECOR axial boundary conditions N
P,. c f
- de ( I a. sin nnB,z), n = 1,... N (12) 4 Total Power, Pin Peaks, and Alarm Limit Ratios n=1 Determination of radial and axial power distributions for the assemblies is completed using the methodology j
described in the previous sections. The core r.,wer is nor.
is used to solve for Ihe a salues for assembly i, where malized to the calorimetric power. The nodal burnups are n
and o are the elevations for the bottom and top of updated by integrating the energy production ovcr the n
n detector level n, respectively. For instrumented clapsed time since the previous time point. This infor-assembhes, the Pin values are obtained directly from mation is used to update the assembly, batch, and core detector measurements. For uninstrumented assemblies, exposure counters. These, in turn, provide information the values are obtained from planar power distribution advantageous to fuel management, and are used to solutions at each level, using ihe coupling coefficient for.
evaluate the fits for CECOR coeIficients, mutation described in I!!.A.2. For all C.E cores, an n-Detailed threc. dimensional peak pin power distribu.
mode Fourier expansion is used to matet, all n detector tions are obtained axially at each node for each assembly powers simultaneously, where n is either 4 or 5.
through the use of precalculated pin to box (F ) factors, p
7
01 08 '00 01:55 ID:$PF-301 Series FAXt PCGE 10 4
1 l
which are fit to the assembly burnup. The pin to box fac.
C. CECOR Libraries l
tor is defined as the ratio of the masimum 1 pin power 1.
Formahsm l
m the hos to the aserage pin power in the box. The precalculated values are obtained from detailed two-
'I he INCA /CECOR system rehes on the use of a large dimensional depletion calculations, u>ing multi. group, hbrary of predetermined coefficients which are used to transport theory corrected diffusion theory, for each rod convert the measured signals into detailed, threc.
bank configuration or region. Typical examples of the dimensional assembly and peak pin power distributions.
burnup behavior of F for different assemblies are given The values of these coefficients depend on the anembly i
in Figure 8. The peak pin power at any axiallocation is and axial location, core configuration, and the local then obia ned using assembly or nodal burnup.
The library coefficients <CC), W', and F dermed i
p above are currently obtained from detailed, two-dimensional depletion calculations (II) using multi group P,.(r) - I,.(z) P,(z)
(13) diffusion theory. Each assembly, instrument, and pin cell is exphcitly represented in these calculations for each nor-mal control rod configuration. The coefficient values are w here I p;(r)is taken from the appropriate coefficient set calculated for each assembly and instrument, and are fit and evaluated with the local assembly burnup at that as functions of the local burnup for each assembly. 'Ihe cles ation.
two-dimensional calculations are taken to represent ex-1 plicit axial regions or planes in the core.
For first cycles, which generally have separable power distributions, a core average plane is used with the various control rod configurations for cores with either four or 7 0 --,
f ve detector levels. For later cycles, which generally do not have separable power distributions, a plane repre-U 08 C _-
senting the axial mid region of the core is currently us-O
/
ed. This leads to acceptable uncertainties, on the order i
p 16 Typical Batch C Asssembly of 5 6%, as shown in Reference 6. However, the use of several different planes to represent the different axial i
I 1.4 regions of the core would lead to lower uncertainties, j
Typical Batch B Assembly since the axial variation of the coefficients would be in.
- 12
/
cluded. The transition to multi-level coefficients obtained Y
directly from the normal, reload 3 D calculations is
/
underway for reload cores with either four or five detec-1.0 Typical Batch A Assembly 1
TW d M m le Whis uncehin 0
1 2
3 4
5 6
7 8
than those quoted in Reference 6, and will remove con.
Assembly Average Burnup,108 MWD /MTU servatism as is shown in section iv.
2.
Generation and Quality Assurance of CECOR Fw. & Sirmte pin peaking factor variurion with burnup Libraries it is apparent from the previous sections that the CECOR libraries are both large and compicx. In a typical application, approximately 5000 coefficients must be stored and retrieved upon demand. An automated com.
The 3 D pin power distributions can then he examin-puter code system has been developed to f acilitate both ed to find the peak pin power (Fqi) for each assembly the generation and quality assurance of these libraries.
and for the core (F ). Axial integrals can be formed to The sequence of steps required to produce a typicallibrary q
lind the peak pin channel power (F ) for each assembly is shown schematically in Figure 9.
ri and for the tore (F ). l'orther, the peak pin power to the The generation of the basic data required for CECOR, r
avesage pin power can be obtained for cach axial eleva-i.e., the pointwise powers, fluxes, and concentrations, is l
tion, Fxy(r), and for the coic as a whole,(Fxy). Conver-currently performed through standard fine mesh diffu.
sions are also made to give peak linear heat rates in terms sion theory calculations using the PDQ-7 computer ot actual powers rather than as ratios, and also as alarm code.(II) Approximately 15 such calculations are re-lumt ratms. The code also provides most of the quired. Primary editing of the PDQ 7 files is performed paiametcis required io cstimate DND ratios. The DNBR by the CERISE code. This operation transforms the caleulation is not perfoimed in CFCOR, but through the pointwise PDQ-7 data into assembly wise data required l
link in peripheral codes.
by CFCOR. Summary files are written for use in Paths 8
l l
{
01 08 '00 01:55 1D:SPF-301 series FAX:
PCGE 11 A Genoration of Dasic Data Base Depletion
(?D PDO)
Disk Files r
~
~
x RS Rodded Cases Point Powers I2D PDOI Concentrabons Qsk Filesj
(
j Ubrary 1
Y***'D D Generation of CECOn Library PDO Assembly Libra F tng qDisk Fsley Y
Q brary F p Detector Segnals
(
N Y
C Simulated Execution I
h Simulated l1 j
I Depletion Disk Filey I
(CECOR)
CECOR D Ouahty Assurance (Assembly Powersj Comparison Routines O A' (CERISE)
StaWs Ft. 9: Generation and quality assurance of CECOR libraries D D in Figure 9. Detector signals consistent with the struments, (2) box power synthesis uncertainty for the ex-library are also generated for testing the completed trapolation to uninstrumented locations, (3) pin peaking librtry.
synthesis uncertainty, and (4) pin peaking calculative The generation of the CECOR library is indicated in uncertainty. These are then statistically combined to yicld P th H of Figure 9 and is the work of the CEFIT code, overall uncertainties. Details are provided in Reference 6.
i At this stage libraries are available for either quality The basic measurement uncertainty is the error I
assurance testing or actual use at the reactor site.
associated with the measurement of assembly average All the codes (ll 16) used in generating the libraries and power in instrumented locations. It includes uncertain-their functions are given in Table 1. These codes are quali.
ties in raw detector signal, background correction, initial i
ty assured and maintained in accordance with internal calibrated detector sensitivity, sensitivity depletion, and quality assurance procedures for computer codes. To ade-signal to power conversion. The measurement uncertainty quately test the library, the CECOR code is executed was evaluated by comparing the large data base of using the file created in Step B and the signals in Step rneasured and calculated box powers with the calcula.
A in a fashion identical to the manner in which the PDQ-7 tional error removed. The uncertainties are evaluated for cases were executed. The results are stored a'id compared the local peak power to core average power ratio (F ),
q to the original CERISE data (Step D).
planar peak power to planar average power ratio (liy),
Final testing of the CECOR program and libraries and axially integrated peak power to core average power must be performed at the reactor site. Sample test cases ratio (F )-
r from the above effort serve to verify the coding and the An alternative formulation uses the powers in sym-data. Test results and procedures are properly recorded metric instramented assemblies to provide an estimate of and documented for future reference, the F,Fxy, and F basic measurement uncertainties. The q
r advantages of this approach are that the uncertainty com-ponents are available at the same time as other CECOR D. Accuracy of the System results. Further the calculation does not require 3-D l
The error inherent in the use of the CECOR program ROCS results for comparison purposes. The results of I
can be broken into four components. These are: (1) basic the two n.cthods are essentially identical as shown in box power measurement uncertainty due to the in-Table 2.
I 9
~
^^
01/0800 01iS6 ID:SPF-301 Series FAX:
PAGE if TABLE 1
SUMMARY
OF COMPUTER CODES USED IN OENERATION OF CECOR LIBRARIES i
Typical Number Code Type Description of Ceses DIT/CEPAK Cross section standard design depletion 5 depletions (of (14, 15, 16) generation calculations 10 time steps each)
PDO-7 2 D fme mesh standard unrodded design 15 cases (11) 2 or 4 group 2-D depletion plus several diffusion rodded cases theory calculations ROCS 3-D coarse standard unrodded design 10 cases (12, 13) mesh 2-group 3-D depletion diffusion theory calculations CERISE PDO-editor provides edited information one for each
)
(i.e., detector signals.
PD O-case assembly power, etc.)
CEFIT CECOR library fits all CECOR coefficients one per core generator to burnup and provides properly formatted library 1
i l
TABLE 2 The pin peaking synthesis uncertainty is the error COMPARISON Os-assxisted with the axial synthesis of maximum pin CALCULATION-MEASUREMENT AND powers using CECOR pin-to-box coefficient libraries con-SYMMETRIC INSTRUMENT STATISTICS RESULTS structed from single plane depletion calculations. It is evaluated by comparing pin-to-box factors from 2 D Core Follow Sym. Inst.
planar calculations representing a mid planc and a top.
Quantity Method Method plane.
Fxy
.0189
.0183 The pin peaking calculative uncertainty is associated Fr 0157
.0139 with the basic calculational error in the pin to box fac-Fq 0189
.0183 tor used in the CECOR synthesis, which is obtained from transport theory corrected diffusion calculations. It is evaluated by comparing calculated pin powers against pin powers from critical assemblies.
The box power synthesis uncertainty is the error The components of these four uncertainties are then associated with the radial and axial extrapolation of combined statistically to obtain overall uncertainties for measured box powers to uninstrumented locations. These F,F and F. As shown in Reference 6, the overall crrors arise from radial coupling by means of the CECOR CkCOk, uncertainty is such that there is a 95% probabili-x r
coefficient library and the Fourier expansion from 4 or and F values ty that at least 95% of the true F, Fxy,from CECOR q
r 5 axial detector locations to all axial points. The uncer-will be less than the values inferred tainty components for F, Fxy,ference 3 D coarse rnesh measurements plus 6.2%,5.3%, and 6.0%, respective-and F are obtained by q
r comparing box powers from re ly. This is summarized ir Table 3.
diffusion thcory calculations (ROCS) to box powers from CECOR synthesis calculations using test signals derived IV. APPL.lCATIONS OF THE CECOR SYSTEM from the reference 3 D depletions and single level CECOR is used in a numect of applications, by both CECOR coefficients obtained from consistent 2 D planar utility and C-E personnel. It is the source of operating depletions, data for routine monitoring and surveillance functions 10 l
01/08, '00 01:57 ID:SPF-301 Series FAX:
PAGE 13
+
TABLE 3 1 26 APPROVED CECOR UNCERTAINTY VALUES
~
& 1.22 aROCS Quentity 95/95 Uncertelnty (%)
8 o CECOR l'I8 0 Fq 6 24 1
)8 0
Fxy 5 29 h 34 3
Fr 6.04 I 1 10 e,
l A 106 i
end is used to verify the operation of the on line COI.SS V
l monitoring and CPC programs (7) on newer C.E plants.
1.02 i
it is especially important during startup test phases fo, O
e' 8000 12000 16000 j
..a new plants and new cycles. Normal core follow activities Cycle Average Exposure. MWD /T involve comparisons of predictions and CECOR j
measured data. On a few occasions, core follow activities l
hsve revealed core or plant computer problems. Also, Fig.10: ROCS and CECOR axialpeaks (F2>
)
design code and method verifications are made possible with the use of CECOR measured power distributions.
This has led to improvements in codes and methods. A 0.15 x
more recent activity has been the verification of the sen-8 a ROCS sitivity depletion law given in Equation 4.
5 010
)
oCECOR 0.05 A. Core Monitoring and Surveillance non CECOR is used to provide operating plant data for
.e comparison with limiting conditions of operation (LCOs)
-f-005 I
and for verification that the plant is operating within the E
Technical Specifications limits. Among the data recorded
- 0.10
' b ' b ' b ' b-0 are the planar radial peaking factor (Fxy), integrated radial perking factor (F ). 3-D l pin peak (F ), axial Cycle Average Exposure (MWD /T) r q
i shape index, axial peak to average power azimuthal tilt magnitude, as well as radial and axial power distributions.
Fig.II: ROCS and CECOR internalaxial shape indices For the newer C E plants, CECOR is used to verify l
the operation of the Core Operating Limit Supervisory System (CPC).(gCOLSS) and the Core Protective Calculators
)
In addition to the normal surveillance functions, Diagnostics CECOR is used during plant and cycle startup tests to When basic core follow and instrument follow pro.
l verify key core parameters, grams uncover disparities between measured and calculated data, it is necessary to determine the cause.
Sometimes this reveals a core anomaly or some error in l
B.
Core Follow the analysis systems. The CECOR system has proved useful in this process as the following examples Core follow activities encompass several of the demonstrate.
CECOR functions. Monitoring data and detailed 3 D In the past, core follow activities pointed out the power distributions are recorded as a function of core possibility of shim failures in the first cycle of an early lifetime to determine trends in important core parameters.
plant. This was discovered during core follow of the start-In this way, the proper functioning of the core is verified.
up phase for the plant. There was a 20% increase in the Figures 10 through 12 are examples of data trended in measured core average axial peak within one month of typical core follow efforts. These include the CECOR startup, as shown in Figure 13, rather than the normal measured values and the values calculated by the ROCS flattening of the distribution. Normal standard hviations design method. When the core is not operating as ex.
for radial comparisons are in the range oi 1% to 2%,
pected, the reason rnay be an actual core anomaly, site versus the 4% seen in Figure 14. Also, an increa,ing reac-problems or design model deficiencies. Examples are livity error with burnup was noted indicating a loss of
]
discussed brictly in the following sections.
poison. Of the possib!c causes for this behavior, it was
0168'0001:57 ID:SPF-301 Series FAX:
PAGE u
12 -
1.2 oooo 1
1 g
5
{' O 8 08
] 06 k 06
'4 a
261 MWD /T E O4 1019 MWD /T g04 CECOR z
CECOR z
O nocs 0.2 U _
-ROCS 0.2 i
0 t
i i
i f
8 1
0 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percent Core Height Percent Core Height
'E~
12
~
_m m n._
1 1
m 08 08 06 06
.a w
g04 3084 MWD /T
{ 04 2021 MWD /T Z
CECOR Z
CECOR O
ROCS 0.2 02 0
ROCS 0-8 i
1 8
i 1
1 0-O 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percent Core Height Percent Core Height Fig. I2: Comparison of ROCS and CFCOR axialpotver distribution l
l concluded that some form of shim failure could explain
' ~" "
1.5 -
the observed effects with the proper magnitudes. Subsc.
quent studies of the removed shims confirmed this I
j (Reference 17).
Crud deposits have occurred in a number of plants both in the U. S. and Europe. A recent case was detected t 1.0 by noting that measured power distributions were shifted g
in towards the core center and down towards the core a
bottom. This is thought to result from crud preferentially i
3 depositing in regions of elevated temperature in the core.
E 05 A 21 50 52976 170 MWD /MTU Figure 15 shows the ROCS and CECOR values for the l
B 15:70 6-12-76 425 core average axial shape index as a function of burnup.
C 13 80 63076 620 The largest discrepancy is near 3000 MWD /t, and the D
7 40, 9-76 760 large measured positive value (.015) means the power moved to the core bottom. Figure 16 shows the calculated 0-and measured values of the core average peak to average 10 20 30 40 50 60 70 60 90
% of Core Height power. Fg, as functions of burnup. The large difference Bottom Top between ROCS and CECOR indicated a problem with the core axial power distributions after the initial good c
Fia.13 ArinI porver shapes for four carly snap 4 hots agreement. This is confirmed by Figure 17, which shows ahuwing c//cci e/ shim failure core average axial power profiles for four time points in 12
01r08 '00 01:58 ID:SPF-301 Series FAX:
PAGE 15 I
l l l l l 1.26 4 ROCS
.a sa em 1.22 oCECOR e..
i,.
o
$ i 1B en
'o I
.u
.u l
~
e.n sa 3
33o 6
.a.
s
.u 1 06
/.".
em n.
1.02 a
8" OK 2K 4K 6K 8K 10K 12K 14K 16K 18K
- a in Cycle Average Exposure (MWD /r)
.u Fig 16. Axial peaks (Fe) vs cycle average exposure showmg effect of crud IIIII affect the power distributions inferred from in core in.
Standard Deviation of Differences 4 06%
strument data. Recently. during a plant computer restart, excessive incore detector depletion was inadvertently Iig.14: Relative power density in instrumented esem.
entered. This was detected in the course of normal bhes (percent difference berneen calculation and instrument follow activity (Figure 19). The sensitivity is measurement) showing anomalous shim beha vior depleting linearly in a reasonable fashion. up through about 4500 MWD /T. The discontinuity in the lines oc.
0 15 curred after a plant outage when the plant computer was updated to account for the computer outage. The downtime was mistakenly entered as 10 days instead of E o10 4 ROCS
- 1. Following the break in the lines, the depletion continues
]
1 o
o CECOR w th the same proper slope as the original lines, but start.
g ing from the incorrect value, it was possible to locate the 2 0.05 time of the problem to within the exact day, as seen in the figure, f 00 a
4K 6K 8K 10K 12K 14K 16K --
During a later cycle of one of the five-detector plants.
a --+- -.
i core follow comparisons of measured and calculated g
_2 h -0.05 power distributions showed considerable disparity be.
tween instrumented and uninstrumented symmetric
~
bundles. A clue to the nature of the problem was the 0 10 detector-level dependent nature of the discrepancies. The Cycle Average Exposure. MWD /T problem was traced to the fact that a single level Fig. // Internal a rial shupe index (ASI) es cycle (midplane) set of CECOR radial coupling coefficients was l
accruge exposure showing c//ect o/ crud being used to represent ull axiallevels of this taller. five.
deteam' core. Thus, the substantially different burnup profiles and reactivity at the different levels in new and the cycle. The shift to the core bottom by 2854 MWD /t old fuel were not explicitly accounted for axially. Since is easily seen. This was accompanied by a radial power the CECOR system of programs was explicitly designed roll to the core center as shown in Figure 18 by com.
to produce and utilize multilevel CECOR libraries, a parisons of ROCS and CECOR powers at the four axial 5 level coupling library was produced. When the level detector levels (20,40,60, and 80% of core height. respec.
dependent library was used, the discrepancies disap-tieely). Proceeding up the core from level I to level 4, peared. It is planned to use multi level coefficients as a a greater fraction of the level power is shifted into the matter of course in future reloads.
core center. The formation of crud was attributed to air An additional b:nefit of the multi 1cvel approach is leakage into the makeup water systems; when this was the reduction in over-conservatism for peak monitoring corrected and the primary system treated, the power data. The use of single-levei coefficients gives highly con.
distribution returned to normal.
servative values of peaking towards the top of the core.
Of course, there need not be actual core problems to This is illustrated in Figure 20 which shows a comparison 13
01 08 '00 01:58 ID:SPF-301 Series FAX:
PAGE 16 l
826 MWO4 1460 M W O/T 1.4 1.4 1.2 12
} 10 f 1.0 s
O O8 0.8 E
1 06
\\
j 0.6 s
--- ROCS
~~~
04 0.4 CECOR CeCoR 0.2
~
00 i
^l
- I i1 - '
i
_i>
00 t - i - I 1
t I
i 1
t i 0
10 20 30 40 50 60 70 80 90 100 0
10 20 30 40 50 60 70 80 90 100 Percent Core Height Percent Core Height 2112 MWD /r 2864 MWD /T 1.4 1.4 12 1.2 i
f :.-
\\
$ 1.0
/,._ _.
---,N 10 08 08 1 b
@ 06 /
$0.6 -[
k O4 -
--- ROCS
~~
04 CECOR 02 CECOR 02
~
i
~
'*I
' - I I
1
'=
0.0
- ' - 1 I
00 0
10 20 30 40 50 60 70 80 90 100 0
10 20 30 40 50 60 70 80 90 100 Percent Core Height Percent Core Height
[
Fig.17: Axialpowerdistributiorscomparisors showing effect of crud of the assembly planar peaking factors from a 3-D ROCS fast diffusion constant which forced the agreement of dif.
l calculation and a CECOR simulation. The CECOR fusion theory and transport theory albedos. The work wa.t simulation was run using single level coefficients and extended for both 2 and 4 group and fine and course mesh i
signals from the 3-D ROCS. The use of multi. level coef.
diffusion theory calculations. The use of the formalism l
ficients removes this over. conservatism, as seen in Figures leads to good power distribution agreement between l
21 and 22.
measurement and calculations.
By the late 1970's coarse mesh 3 D neutronics calcula.
D. Verification of Design Codes and Methods t ons were assuming a major role in fuel management and Over the years, many improvements have been made core follow of reload cores.This occurred because of the in C.E design codes and methods. These changes grew degree of Spatialinseparability of the neutron flux within l
in response to benchmarking against both critical ex.
reload cores. It was found necessary to account for these i
l periments and the ever increasing data base of CECOR 3 D cffcets in the design and follow of cach succeeding results from operating power reactors. A few of these cycle. The ROCS code, which had been developed to pro-resulting improvements are summarized here.
vide such solutions while maintaining a realistic balance in the mid 1970's, comparison of measured and between speed and accuracy, was ori inally a modified calculated CECOR (INCA) power distributions showed one. group (I 1/2 energy group) code.(I ) For reload cores that 4 group diffusion theory models overpredicted power in order to gain acceptable accuracy it was necessary to distributions in the center of the core by a few percent, extend the ROCS formalism to 2-groups. When this was and underpredicted the dhtributions at the core periphery done errors were reduced to leu than half their original j
by a similai amount. The problem was traced to the values.0 3) This use of coarse mesh codes required tech-l reflector treatment in diffusion thcory calculations. Dif-niques to produce accurate homogenited cross sections.
fusion theory was underpredicting the albedos of the The DIT code based on integral transport theory has been reflector regions in the fast group. An analytic study was developed and provides both fine and coarse mesh cross carried outO 8) which led to the derivation of a corrected sections.(1415) In addition, it is necessary to have a l
14 r
i
01/08 '00 01:59 ID:SPF-301 Series AX:
PME 17 0 000 0 016 0 020 0 02t f,f, 0 024 00N 0 036 0 006 0 00 P O CD6 0 08' 0 040 nais a 0is 0 028 0 040 0 027 c ot t 4 002 0 002 0 040 0 06' 0 037 00's 0 003 0 008 l
00'S 0 032 0 014 O cue C ott 4 096 40,7 0 029 0 042 0 026 0 01%
00i)
-00i4 0 016 l
0 028 00M 0 016 0 004 0 002 0 000 40:3 4 02t 0 013 0 o40 0 02$
0 00e 0 003 0 012 0 026 4 029 0 022 0 011 0 004 4 000 4 000 4 022 4 021 4 041 0 038 0 010 0 012 4 002 4 007 4 03g 0 031 4 04$
0 017 0 000 0 000 0 004 4 016 4 022 4 036
-0 032 0 022 0 000 0 018 6 000 4 006 404 0 064 4 046 4 00t 0 006 4 000 4 m2 4 007 4 016 4 017 4 042 4 040 0 096 0 004 0005 4W 0 021 0 027 0M 4 070 4 006 0 080 0 0t9 0 010 0 007 0 00s 4 014 4 038 4 033 0 056 4 000 00tF 0 006 0 004 4 029 4 06) 4 043 4 t01 0 070 t...I i i.,to, t o., 2 i
l 0 025 0 029 0 093 0 004 US 0 046 00M 0 0$4 0 022 0 027 0 036 0 088 0 053 00 0 0 006 0 062 6 061 0 063 0 030 0 010 0 000 0 066 0 049 00r3 Coss 0 026 4 020 0 031 0 046 0 037 0 029 4 019 4 010 4 050 0 014 0 046 0 034 0 044 0 015 0 C02 4 020 00JS 0 019 0 037 0 010 0 005 4 009 00M 0CD8 0 030 00t3 0 000 0 020 0 020 0 007 4 M3 4 030 00.M 0 024 0 029 0 002 4 007 4 054 4 04$
4 toe 0 01%
0 016 0 041 0 014 0 012 40's0 0 CDS 4 119 0 030 0 030 0 035 4 002 4 066 4 060 4 104 0 076 0 012 0 048 0 062 0.02:
409 40:2 4 toe 4 074 0 017 0 001 0 013 0 022 0 006 4 02s 0 0e6 4 107
-0 604 4 843 4 000 0 035 00S3 40s0 4 040 4 516 4 100 4 146 00/1 4 001 0 030 0 006 0 011 4 033 0 107 4 077 4 144 4 000 00'3 0 004 4 015 0 022 0 117 4 073 4 150
-0 000 tws t..i a rroe l
i I
Fw. IR-Levelwise ROCS-CECOR power comparisorse showing effects of crud 15
01 08 '00 01:59 ID:SPF-301 Series FAX:
PAGE 18 35-method for obtaining local pin powers from coarse mesh
^ l' I results. This has led to the development of the MC pro-o Level 2 gram using the imbedded method for recovering local j
E**
34 o Level 3 a
- 11. Verification of Detector Sensitivity Depletion o Level 4 l
Behavior (33
~
detector sensitivity depletion behavior. The current pro-A more recent application has been to examine the 5
j duced by a self powered rhodium detector is proportional to the product of the neutron absorptions in the rhodium 7
p'3'2 and the beta escape probability neglecting the small frac-I tion (1-2%) produced directly by gammas, i c.
i 31 I-I' N,. f f a t dE dV (14)
VE where 3
_._ t,..
i 3000 4000 5000 6000 e
is the electron charge Cycle Burnup (MWD /T)
P is the average beta escape probability o
is the Rhodium cross section Fsct.19. Sensitivity d pletion of typicalinstrument
+
is the flux in the detector NRh is the average Rhodium number density V
is the volume of the Rhodium emitter.
14 i
i i
i i
i i
i i
i i
i i
i O
O o
O O
O O
O O O
O O
2 A A A
A A o A
A A A l
O A
b 1.3 cL E
A i
k 1.2 Eo
- u o CECOR E
ct A ROCS l
n I 1.1 E
10 6
7 8
9 10 11 12 13 14 15 16 17 18 19
/
ROCS Plane l
Fig.20. Cvmpurison of ROCS and CKCORvinnorreck in pinne nowrace box posver (3 d multilevel coefficients) l l
l 16 l
01/08 '00 02:00 ID:SPF-301 Serles FAX:
PAGE 19 14 i
i i
i i
i i
i i
i g
o o
a 2 8
8 Ba e
a g o f1.3 E
t<
Fn 1.2
~
E
~
B g
O CECOR E
a a nocs 1.1 1.0 I
8 8
I I
i 6
7 8 9 to 11 12 13 14 15 16 17 18 19 ROCS Plane Fw. 21: Comparison ofROCS and CECORplanarpeak to plane average box power (2 d single level coefficients) 10 i
i e
i i
g Five-Detector String
=
g
=
E 7 - Monlioring Limits j
15%
85%
56 b
g5 v54 m
3 c.
j2 1
t A
i i
i i
f f
0 10 20 30 40 50 60 70 80 90 100 Percent Core Height Fig. 22: Comparison of ROCS CECOR planar box power synthesis cerors using sing!c. and multi level coefficients (near B O C) 17
01/08 '00 02:00 ID:SPF-301 Series FAX:
PAGE 20 From Equation (14) the beta escape probability is deter-mined as 1 00 r
1 I
E" N,. J f o (dE dV (15) 75 yg If the detector sensitivity is based on currrent per rhodium j(
activation, i.e.
j E
I 3
@)
S- * "
a f f o 4 dE dV 25 vE it can be shown that the sensitivity follows a linear deple-tion law if the beta escape probability is constant with 0 00 ~ '
depletion 0
100 200 300 400 Accumulated Charge (Coulombs)
~
Fin.23: Relative sensitivity cies appear. This is shown in Figure 24. The resulting Q.
where is essentially the same with a value of 316 coulombs.
So is the initial sensitivity The linear behavior is not surprising since the beta Q
is the accumulated charge escape probabliity was found to be essentially constant eFVNo is the theoretica! total charge Q.
r-or flat over the cycle length. A typical example is shown Further, if the beta escape probability is constant with in Figure 25 for an instrument that remained for the en-depletion, the accumulated charge also follows a linear tire 3 cycles. The scatter during a cycle is on the order of 12%. There were also variations between cycles for law Q = s i' V N. (1 -
)
(18) 18 where the slope is proportional to I'.
incores use a linear depletion law.(21)g fixed rhodium('e Currently. C E and others employin An initial study has been performed at C E to evaluate this using I.75 operating data. Measured currents were compared to the 5
S rhodium activations and reaction rates calculated by the j
ep,o newly developed and verified imbedded fine mesh g.50
.%'e, standard C E coarse mesh ROCS code coupled with the g
assembly MC program.(20). The calculations were car-
]
ried out in a core follow mode over the first three cycles y
of a typical C E 2800 MW design in which detectors were y
depicted to about Iwo. thirds of their theoretical charge, g.25 Iigure 23 shows the results of the sensitivity calcula-O tion.There is quite a bit of scatter but the entire ensem-ble demonstrates a linear depletion law with Q of 308 OMg' gy
' y - u yoo coulombs. The scatter is caused to a large degree by the difference in the box power calculated by the ROCS core-Accumulated Charge (Coulombs) follow models from that observed by CECOR. When the ROCS MC activations are adjusted for these power dif-ferences, the scatter decreases greatly but cycle dependen-As. 2t Corrected Relative Senaitiviry 18
01 0,8 '00 02:00 ED:SPF-301 Series
^
i-k PCGE 21 any instrument on the order of I-2%. These are apparent-sity for each cycle and accumulated charge is shown in ly caused by modeling effects rather than actual physical figure 26. The behavior is piece-wise linear with the in.
effects. When the values are not adjusted for the dif.
tercept showing a Q. of about 308 coulombs. The slope ferences hetween ROCS and CECOR box powers the implies a beta escape probability of.403, consistent with scatter was larger, but the basic behavior was the same, the above observations.
i The average value of the beta escape probability was This study tends to confirm the use of linear deple.
about 0.40 0.42. This is consistent with values reported tion law at least to about two thirds depletion. Further in the literaturc(22) for this diameter instrument, work with more depleted instruments is necessary to con.
The comparison of calculated rhodium number den-firm the law for higher depletions.
.50
- *$ ' Basic
.40 1.00
- 's c
[.30-3 p.75 h.50-
~
a E
R us E
ve q
0 9
O j.50 b
c3 Corrected
...w o
.40 Cycle 1 Cycle Cycle 3
)
l 2
cc a,
j.25
.30
]
b 0 00
--a
- - + " ' '
0,00
'8
+
0 100 200 300 400 0
100 200 300 400 Accumulated Charge (Coulombs)
Accumulated Charge (Coulomb 6)
Fug. 25' Reta escape probabstity for a typical Fig. 26: Rhodium depletion instrument 1
i I
1 19
01 08 '00 02:01 ID:SPF-301 Series
' FAX:
PAGE 22 REFERENCES 1.
R. l. licilens, T. G. Ober, R. D. Ober "A Method 12.
T. G. Ober, J. C. Stork. I. C. Hiekard, J. K.
of Analyzing in. Core Detectur Data in Power Reac.
Gasper, " Theory, Capabilities, and U.sc of the tois," Trans ANS. 12, 820 (1969).
Threc. Dimensional Reactor Operation and Control Simulator (ROCS)," Nucl. Sci. and Eng,64 (605),
2.
T. G. Ober, P. II. Gavin, "Use of In Core in-3977-strumentation in Combustion Enginecting Power Reactors," Tram ANS, 19, 218 (1974).
13.
T. G. Ober, J. C, Stork, R. P. Bandera, W. B.
Terney, " Extension of the ROCS Coarse Mesh 3.
W. B. Tcrney. G. H. Marks, E. A. Williamson, Jr.
Physics Simulator to Two Energy Groups," Trans.
and T. G. Ober, " Axial Power Distribution from
^*'""#
- " 8 63).IM8.
Fourier Fitting of Fixed in-Core Detector Powers,"
Trans ANS, 22, 683 (1975).
14.
A. J nss n, et. alu " Discrete Integry Transport Theory Extended to the Case with Surface Sources " Atomkernenergie, Bd. 24. 1974 4.
W. D. ') erney T. G. Ober, and E. A. Williamson, Jr. "Three Dimensional Calculational Verification 15.
A. Jonsson, et. al., " Verification of a Fuel of C-E's in. Core Instrumentation System with Assembly Spectrum Code Based on Integral Lifetime," Trans ANS, 22, 683 (1975).
Transport Theory," Trans. Am. Nucl. Soc., 28 (778), 1978.
3.
W. B. Terney C. A. Williamson. Jr. and T. G.
Ober, "CalcuMon Verification of the Combustion 16.
Combustion Engineering Standard Safety Analysis Engineering Full Core Instrumentation Analysis Report (CESSAR), Chapter 4.3.
System CECOR." Trans ANS, 24, 429 (1976).
17 Y. D. Harker, et. al., " Reactivity Measurements 6.
A. Jonnon.
W. B. Terney, M. W. Crump, of St. Lucie Shims at the Advanced Reactivity
" Evaluation of Uncertainty in the Nueicar Power Measurement Facility," Trans Am. Nucl. Soc. 26, Peakmg Measured by the Self. Powered Fixed in-(Supplement 1), 25 (1977)
Core Detector System," CENPD 153-Rev. lA, Combustion linginecting. Inc. (1980).
18.
W. B. Terney, " Albedo Adjusted Reflector Fast Diffusion Coefficient," Trans. Am. Nucl. Soc.,18, 7
R. W. Knapp, C. R. Musick, " Digital Core Monitoring and Protection Systems," presented at 19.
- 1. C. Rickard N. R. Gomm, T. G. Ober, "Calcula-IAEA Working Group on Nuclear Power Plant tional and Experimental Verification of the Com-Control and instrumentation, Speciahsts' Meeting, bustion Engineering Coarse Mesh Physics Cadarache France. January 26 27,1977 (C-E TIS.5ill).
S mulator," Trans. Am. Nucl. Soc.,24,340(1976).
20.
S. F. Grill, A. Jonsson, J. R. Rec, "A Nodal im-8.
L. A. Banda, B.1. Nappi, " Dynamic Compensa-bedded Method to Recover Local Power Peaking tion of Rhodium Self. Powered Neutron Detectors,"
from Coarse Mesh Reactor Calculations," Trans.
lEEE Transactions on Nuclear Science, Vol. NS-23 Am. Nucl. Soc., 35, 580 (1980).
1975, Ps. 311.
- 21. " Rhodium In-Core Detector Sensitivity Deple-9.
L. A. Banda " Operational Experience in Rhodium tion " Interim Report NP-1405, EPRI, Palo Alto, Self Powered Detectors," IEEE Transaction on California (May 1980).
Nuclear Science, Vol. NS.26, February 1979, Pg.
910.
22.
T. Lanksonen, J. SaastemMnen, ' "alculational Studies of Sensitivity Characteristics and Their 10.
A.
Ratsion, "A First Course in Numerical Burnup Behavior for Rhodium Self Powered Analysis," McGraw Hill Book Co.,1967.
Neutron Detectors," Proceeding of the IAEA Specialists Meeting on in Core Instrumentation and 11.
W. R. Caldwe'.1, "PDQ.7 Reference Manual,"
Failed Fuel Detection and I,ocation. AECI. 5124 WAPD-TM.Md, January,1969.
111 (May 1974).
20
.CTv17,.'95 14:37 1D:SPF-301 Series FAX:
pc(;E 2
LIC-96-0108 i
June 17,1996 FC-FFe0057 Rev. I Ms. Jan Bostelman Omaha Public Power District P.O. Box 399 Ft. Calhoun, NE 68023-0399
Subject:
Fort Calhoun Cycle 16 Excore I,HR LCO Reanalysis to Support Plant Operation Above 90% Power
Dear Ms. Bostelman:
ABB has completed the rWysis of Fort Calhoun Cycle 16 excom LHR LCO to support plant operation above 90% power. ne purpose of the reanalysis was to determine an acceptable technical approach for operating the plant above 90% power until the completion of Cycle 16 in the event that the liscore detector monitoring system is declared inoperabic.
This letter transmits the results of the reanalysis, which employed ABB's current setpoint analysis process for a variable (power level dependent) calorimetric power measurement uncertainty, in summary, operation at 100% power is allowed in the ASI range of -0.12 to
+0.05 asiu provided the core is unrodded above 90% power. This result is applicable to Cycle 16 only.
De details of the analysis are provided in the Attachment.
De recommended approach does not change the results of the safety analysis or other setpoints for Cycle 16. De recorded calculation will be transmitted separately. The LHR LCO reanalysis was quality assured per the ABB CENO Quality Assurance Program.
If you have any questions, please call me at (860) 285 5512.
Sincerely, COMBUSTION ENGINEERING, INC.
7e llacasa Su rvisor, Setpoint Analysis ABB Cornbustion Engineering Nuclear Operations certvsio-l ag. nee'q Mr PO Bos 500
'oiorr.ccc (mi) eM to
- Uino Pmanasi e Ras 1 Fan 18C4 295 951?
wecsor Conne:t c.i %09'rWW
. 06 17 '95 14:37 ID:SPF-301 Series FCet:
PCGE 3
FC FPA057 Rev.1 Page 2 Distribution:
M, J. Guinn OPPD R. Jaworski OPPD (w/o Attachment)
R. C. Whipple ABB (w/o Attachment)
R. O. Doney ABB (w/o Attachment)
J. A. Noyes ABB (w/o Attachment)
R. T. Pearcc ABB (w/o Attachment)
R. S. Spies ABB (w/o Attachmerx)
QR(2) l
,03 17,'95 14:38 ID:SPF-301 Series FAX:
PE 4
I l
Attachment to FC-FE-0057 Rev. I Fort Calhoun Cycle 16 Excore LHR LCO Reanalysis
,. 06q7 '95 14:38 1D:SPF-301 Series FAX:
PC/A 5
PC FIM)057 Rev.1 Page A-1 4-Purname of Analvsis l
OPPD has been experiencing a high rate of incore detector failures during the operation of Ft. Calhoun Cycle 16. If the observed failure rate continues, the incore detector i
monitoring system may have to be declared inopemble before the end of Cycle 16. If j
this situation arises, Pt. Calhoun's procedures will allow the plant to continue to operate provided that the WR and DNB LCOs are monitored using the excore detector monitoring system. However, the plant will not be allowed to operate at full power i
because the maximum power level permitted by the excore LHR LCO tent is presently 90% power. The excore DNB LCO tent supports full power operation.
The excore LHR LCO was reanalyzed for Ft. Calhoun Cycle 16 to justify operation at
]
a power level greater than 90% until the end of Cycle 16 if the incore detectors are declared inoperab!c. This was accomplished by trading off LHR margin with imposed j
restrictions on plant operation above 90% power.
l Analysis Summary SCU penalty factors were calculated for use in the CESMAP runs based on the input uncertainties tmnsmitted by OPPD in References 1 and 2. 'Ihe CBSMAP WR LCO codes j
for the variable power measurement uncertainty process were utilized to calculate a new IRR LCO tent boundary and its corresponding N factor versus F[ tradeoff curve. The I
physics, fuel performance, and transient analysis inputs for the CBSMAP runs were based on the numerical data in the INPUT 3.tfO electronic file transmitted to ABB through Reference 1 and confinned in Reference 2.
)
'Ihe CBSMAP LHR ILO analysis took credit for plant operation in an umodded configuration for power levels above 90%.
An unrodded configuration has been j
defined as a configuration which allows for exercising of the CEAs, but not allowing insertion for prolonged periods of time beyond the 126 inches withdrawn position. An improvement in the maximum allowed power level was accomplished by restricting increased radial peaking at high power levels by requiring that the lead bank be fully j
withdrawn. ABB's anal the N factor versus F/ysis supports the excore LHR LCO tent shown in F
{
tmdeoff curve shown in Figure 2 for Cycle 16 with the j
following operating limitations, which must be implemented administratively in the i
7 event that the LHR LCO has to be monitored with the ex-core system, such as when the incore detector system is declared inoperable:
- 1. The reactor core must be completely unrodded before the power level is allowed to increase above 90% power.
06 17 '95 14:38 ID:SPF-301 Series Fax:
PAGE 6
FC FB-0057 Rev.1 Page A-2
- 2. The CEAs cannot be inserted while the power level is greater than 90% power and any power reduction must be accomplished through changes in the soluble boron concentration or some other reactivity adjustment. Otherwise, the reactor power level must be reduced to a value less than or equal to 90%.
- 3. CEA insertion is pennitted up to the PDIL lang Tenn Insertion Limit (25%
insertion of CEA group 4) when the power level is less than or equal to 90%.
It should be noted that the N factor versus F
- tmdeoff curve is intended for situations y
where the incore detector system can still be utilized (disectly or indirectly) to measure F/. If the incore detectors are declared inoperable, it would be prudent to monitor with the excort LHR LCO tent provided that the predicted maximum unrodded F ' for y
the remaining part of the core Cycle (including appropriate calculative uncertainties) does not exceed the COLR limit of 1.86. This situation is equivalent to having an N factor of unity (i.e.,100%) for the remaining part of the core cycle as show in Figure 2.
T
'Ite recommended changes to the excore UfR LCO tent and the N factor versus Fy tradeoff curve does not change the results of the safety analysis or other setpoints for Cycle 16.
ABB recommends that the following note be added to the COut Figure for the UDL LCO tent as in Figure 1:
" Note: Rodded operation is not allowed above 90% of Rated 7hennal Power."
References
- 1. Letter, J. L. Bostelman (OPPD) to D. Bollacasa (ABB), " Data Transmittal for Support of ASI Tent Expansion at Low Powers," PED DEN-96-0268, May 6, 1996.
i 2.14tter, J. L. Bostelman (OPPD) to D. Bollacasa (ABB), " Confirmation of Uncertainties for the LHR LCO Tent Reanalysis," PED-DEN-96-0306, June 4, 1996.
)
I i
8
.A
]
Attachment to FC-FE-0057 Rev. 1 Figure 1 Page A-3 Fo.at Calhoun-Cycle 16 Ex-Core LHP LCO Tent 5
(Note:
Rodded operation is prohibited above 90% of Rated Thermal Power.)
_iisigiiiigiiiiiiiiijiiiigiiiigsiinji iig,ii jiisi gs gs i
i i i i6,
100
(.12)too)
{ **8T)100)
Z f
w I
g O,
A 1
8 90
( *IV)90)
(+. i y ) 9,)
~
n o
e g
u R.
c E
80
(,. 4 Ro)
(-t. 2.(,3 ro) 3 f
c.
o n.
c O
o 70 v
~
l i!ni lr le lr
{n nl#
}re ir{t elt iitl
~
n r e t e e ei i
t i i e ier iie a t e
e i g
-O 6
-0.4
-0 2 0.0 02 0.4 0.6 y,
Perinneral Axlal Shape Index
81
. -~
Attacncene to Ft-re-co37 gy. 1 Figure 2 Page A-4 j
ort Calhoun Cycle 16 N Factor Curve
- E c
M INote:
Rodded operation 2s prohibited above 90% of Rated Thermal-Power.)
110 i
i i
i i
i i
i i
i i
_i i
g g
g g
g g_
i i
i i
i i
i i
k i
g v>
100 (IT4 3 800 )
e-o n
(l. F4 9 0) 90 3
y a
~
m g
R g
( 3 9 A afro) e 80 u
z (2 OSS>75) 70 2
Z i
i I~
~
60 i 85 1.90 1.95 2 00 2 05 2.10 m.
Planar Radial Peaking Factor
~_ d_ s llc-96-0108 ISTC TechnicalProcess Review
.3 PageIof5
)
EXECUTIVE
SUMMARY
ISTC TECHNICAL PROCESS REVIEW (Fort Calhoun First-Cycle ICIFailures)
June 3-7,1996 Team:
R. Driscoll (ABB)
P. Hellandbrand (ABB)
C. Hoffmann (ABB)
P. Kasik (MPR Associates)
K. Margotta (ABB)
D. Sylvain (ABB)
ABB CE NUCLEAR OPERATIONS June 19,1996
d'
- g ISTC Technical Process Review
~
Page 2 of 5 i
i NOTICE The information, assessments and conclusions set forth herein are preliminary, and should not be used or construed as final.
The root cause analysis review is ongoing, and the content hereofis subject to further review and revision.
1
)
ISTC Technical Process Review Page 3 of 5 1.0 EXECUTIVE
SUMMARY
As a result of the recent increase in In-Core Instrument (ICI) field failures, particularly first-cycle failures at Fort Calhoun (OPPD), ABB CENO performed a Technical Process Review at the ICI manufacturer, Image and Sensing Technology, Canada (ISTC). This review was part of a Root Cause Analysis (RCA) being performed by ABB to investigate not only potential causal factors related to ICI Materials and Manufacturing, but also all other potential causal factors in the areas of Core Operations, Maintenance, and ICI Installation. The RCA is being pursued in parallel with the Materials and Manufacturing investigation, and a final RCA report will be issued upon its completion (targeted for late summer,1996).
This report summarizes the results of the ISTC Technical Process Review. It was performed during the week of June 3 '
A6 by a team of five ABB personnel and one consultant from MPR Associates (bases i Alexandria, Virginia). The objective was to collect data for the sole purpose of ideutifying materials and manufacturing - related potential causes of the OPPD first-cycle ICI failures. The approach was a technically-oriented review and involved observing work in process, evaluating hardware, talking with operators, and reviewing actual product records including material certifications, inspection reports, radiographs, photomicrographs and metallographic mounts.
Although the failure mechanism has not yet been confirmed by examination of a failed ICI, field failure data from Ft. Calhoun indicates the most probable failure mechanism to be a breach of the individual detector sheath, allowing moisture to get into the instrument and cause a short circuit. A preliminary review of available information suggests that stress corrosion cracking (SCC) may be a contributing factor in causing the breach.
Correlation of failures with manufacturing information indicates that certain manufacturing batches have a higher frequency of first-cycle failures. This information was used to focus the technical process review on certain key areas:
ICI's manufactured for Fort Calhoun in particular. Although information was collected that will assist in evaluating failures at other plants, the primary focus was on those instruments experiencing high first-cycle failures at Fort Calhoun.
The individual detector material procurement and fabrication process, rather than the entire instrument fabrication process.
Three specific production time periods, as represented by certain batches of l
detectors. These periods represented " normal field performance", "high first-l cycle failure field performance", and " current detector production"; i.e., those detectors fabricated but not yet assembled into instruments.
I
t ISTC Technical Process Review Page 4 of 5 1.0 EXECUTIVE
SUMMARY
(Continued)
Changes between, and problems during, each of these three production time periods. This was based on th: premise that something has changed, since the high frequency of first cycle failures is a recent phenomenon (starting in Cycle 16 at Fort Calhoun).
The areas of ISTC design, materials, fabrication, and inspection (results and trends). These areas would cover the full range of materials and manufacturing activities.
This review identified a total of twelve changes that occurred between the " normal field performance" production period and the "high first cycle failure" production period. One of these changes is believed to be significant relative to increasing the susceptibility of detectors to stress corrosion cracking. A change was made in the detector tubing starting material (Inconel 600) which, beginning around 1990, included a nominal 10% cold work as opposed to the previous fully annealed condition. This was done to eliminate problems being encountered with the fully annealed tubing during manufacturing (e.g.,
kinking during loading of insulators).
Review of mechanical properties data from materials certification reports indicates that the initial cold work levels in the starting material could be up to 25%.
This change and the variability of the starting material in the cold worked condition is reflected in a variation in the grain structure. The microstructure of some detector tubing exhibits relatively uniform grain size through the detector wall. Grain size is estimated to be ASTM ~ 7-8. Microstructures of the other batches of detectors exhibit a much finer grain size, typically ASTM 10, with considerable variation of the grain size through the wall thickness of the detector. In general, larger grairis were observed on the outside diameter surface with increasingly finer grains near the inside surface. The net effect of this is that there is a noticeable variation in the final microstructures between batches of detectors. The variations in microstructure may result in some material being more susceptible to stress corrosion cracking than others. At this time, the magnitude of this increase in susceptibility is not known.
o
,.=
ISTC TechnicalProcess Review
.g-Page3of5 1.0 EXECUTIVE
SUMMARY
(Continued)
One other change was noted in the context ofincreased susceptibility to stress corrosion cracking. This change was a reduction in the final annealing temperature from ~ 1900 F to ~ 1650 F, which occurred around 1981-82, prior to all three of the " specific" time periods covered during the review. Heat treatment of Inconel 600 at a temperature of 1650 F does not produce a complete anneal of the material. This temperature is high enough to produce recrystallization in the cold worked material but too low to solutionize the carbides present in the material. As a result, when recrystallization occurs, the resultant location of the carbides is intragranular. The heat treatment results in a fine grain structure, with few carbides on the grain boundaries, some residual cold work, and moderate to high strength. All three of these conditions are considered to make Inconel 600 tubing more susceptible to stress corrosion cracking when exposed to PWR environments.
As noted above, a comparison was also made between the "high first cycle failure" production period and " current detector" production period. Only three changes were noted and none of them are considered significant relative to increasing material susceptibility to SCC.
Based on the above, the preliminary findings of the Root Cause Analysis can be summarized as follows:
There may be an increased susceptibility to SCC in the material used to manufacture the ICI's that experienced first-cycle failures at Fort Calhoun. Similar material may also be present in other ICI's installed at other CE plants (i.e., Calvert Cliffs 1 and 2; St. Lucie 1 and 2; and Millstone 2)'.
The magnitude of this increase is unknown at this time. It is unclear, at this time, whether these conditions by themselves, or in combination with currently unknown factors, are sufficient to cause the failures at Fort Calhoun, and other CE Plants that recently began experiencing high failures rates. It is conceivable that multiple causal factors are involved in these failures (i.e., increased material susceptibility to SCC along with a more severe temperature and/or stress environment and/or instrument and maintenance handling issues). If this is true, failures may occur more readily at some plants than at others. ABB CENO is continuing its investigation in an attempt to more completely identify and assess these potential causal factors. The results will be reported in the final root cause analysis report, targeted for completion late this summer.
Note that the failures at other CE plants have occurred more frequently in the 2nd and 3rd cycles, not the first cycle as has occurred at Fort Calhoun