ML20202E691
| ML20202E691 | |
| Person / Time | |
|---|---|
| Site: | Crystal River  |
| Issue date: | 01/21/1999 |
| From: | Craig K APTECH ENGINEERING SERVICES |
| To: | |
| Shared Package | |
| ML20202E648 | List: |
| References | |
| AES-C-3543-1, AES-C-3543-1-R, AES-C-3543-1-R00, NUDOCS 9902030131 | |
| Download: ML20202E691 (28) | |
Text
'
CALCULATION COVER SHEET s
f Calculation No.:
AES-C-3543-1 Client:
Florida Power Corporation
Title:
Deterministic Operational Assessment Project No.:
AES 9810354310 for Freespan ODSCC Degradation at Crystal River, Unit 3 APTECII Omce: Sunnyvale Sheet No.
1of27 O Uncontrolled E Controlled Document Control No.:
1-2
Purpose:
The purpose of this calculation is to perform a deterministic operational assessment of ODSCC freespan steam generator tube degradation. The scope of this calculation is a cycle analysis of 1.8 EFPY duration.
Assumptions:
The major assumptions for this calculation are given in Section 3.
Results:
A summary of the results is given in Section 2. A discussion of the specific calculations and results is given in Sections 6 through 8. It is concluded that the structural performance criteria will be maintained for freespan ODSCC in the current operational cycle.
Revision Prepared Checked Venfied Approved BY BY Y
Y No.
Revision Description Date Date Date Date 0
A
/62C Initial Release
lu)% T2lul9 bluin nin/98 MEAPTECH*
amus seense senvces, w:.
REV. 8/96 nj!2oggugggg2 j P
J$APTEOl' sNoms5NNG SUMCES. MC.
Made by:
Date:kf8 Chent-Cdculation No.:
AES-C-3543 1
/S (2
FPC by:
Date:
Project No.:
Title:
Detenninistic Operational Assessment for 12[2.2/9 P AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Revison No.:
Document Control No.:
Sheet No.:
0 1-2 2 of 27 TABLE OF CONTENTS Section Title -
Pace
1.0 INTRODUCTION
3 2.0
SUMMARY
4
2.1 Background
4 2.2 Method of Approach 4
2.3 Technical Findings 5
3.0 ASSUMPTIONS 6
~ 4.0 DESIGN INFORMATION 7
5.0 TECHNICAL METHODS 8
5.1 Overall Assessment 8
5.2' Determination of Lirniting Flaw Distribution 9
5.3 Calculation of Axial Structural Limit 10 5.4 Leak Rate Model 11 5.5 Statistical Input Data 12 5.5.1 Mechanical Strength 12 5.5.2 Degradation Lengths 12 5.5.3 Probability of Detection 13 5.5.4 Degradation Growth Rate 13 5.5.5 Flaw Initiation Function 14 L
6.0 CALCULATION OF STRUCTURAL LIMIT FOR AXIAL DEGRADATION 15 7.0.
DETERMINATION OF LIMITING FLAW SIZE 17 8.0 -
OPERATIONAL ASSESSMENT 20 8.1 StructuralIntegrity 20 8.2 Accident Leakage 20 f
9.0 REFERENCES
22 QAE17 REV 8/96 4
[
GNQNSERNQ SEPMCES. DC.
Made by Date:
Chent:
Cciculation No.:
AES-C-3543-1 M
i2.[22/98 FPC Chec by:
Date:
Proicct No.:
Title:
Deterministic Operational Assessment for l 2/ 2.2./9P AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Reeson No.:
Document control No.:
sheet No.:
0 I2 3 of 27
1.0 INTRODUCTION
q 1
The degradation observed at Crystal River, Unit 3 (CR3), after 12.1 effective full power years (EFPY) of operation, included axial outside diameter stress corrosion cracking /intergranular attack (ODSCC/ IGA) at freespan areas, primary water stress corrosion cracking (PWSCC) at the roll transitions region in the upper tubesheet, and volumetric IGA in the first tube span above the lower tubesheet. An operational assessment for these tube degradation mechanisms has been completed (Ref.1). The required 90-day assessment for freespan ODSCC followed the probabilistic performance criteria of Nuclear Regulatory Commission (NRC) Draft Guide-1074 (DG-1074) (Ref. 2) for tube integrity and leakage.
To supplement the previous assessment, a deterministic analysis was performed for CR3 to the performance criteria of NRC DG-1074 for tube integrity. The purpose of this calculation is to document the deterministic evaluation. Assessment of freespan axial ODSCC/ IGA was performed according to the deterministic structural performance criteria of Section C.4.1.1 to DG-1074.
i m
QAE17 REV 8/96 l
l
$ APTEGl'
- ENQmEEMnG SEPMCES, NC.
Mad by:
Date:/22/f8 h
Chent:
C:lculation No.:
AES-C 3543-1 (2
FPC Che y:
Date:
Project No.:
Title:
Deterministic Operational Assessment for l1/22/9/
AES 98103543-10 Freespan ODSCC Degradation at Crystal River, Unit 3 ReGon No.:
Document Control No.:
sheet No.:
0 12 4 of 27 2.0
SUMMARY
2.1 Background
Operational assessments of corrosion degradation of steam generator tubing at CR3 have been performed in Ref.1 and in this calculation. Assessments of freespan axial ODSCC/ IGA and axial PWSCC at roll expansion transitions in the upper tubesheet were conducted using a probabilistic approach and the probabilistic structural performance criteria of the draft NRC document DG-1074 on tube integrity (Ref. 2). Volumetric IGA of a pit-like nature in the first tube span was evaluated according to the deterministic structural performance criteria of DG-1074. Projected leak rates at postulated accident conditions were considered as part of both the probabilistic and deterministic operational assessments.
The purpose of this calculation is to provide the evaluation of run time length for current cycle for freespan ODSCC based on a deterministic approach. This approach follows the deterministic structuralintegrity performance criteria of DG-1074, as described in Section C.4.1.1 of the draft guide.
.At the last inspection, after 12.1 EFPY of operation, an axialindication was found within the freespan region. The one freespan axialindication, interpreted as ODSCC/ IGA, was found during the last bobbin probe inspection.
2.2 Method of Approach The basic calculational technique for the deterministic evaluation of freespan ODSCC follows industry accepted methods for calculating tube integrity under a set of bounding conditions on material properties, flaw size (length and depth), growth, and detection. The structural performance criteria outlined in DG-1074 specify that the following factors of safety (SF) on differential tube pressure be satisfied at any time during the operational period:
SF 3.0 (normal operation)
=
SF 1.4 (accident conditions)
=
QAE17 c
REV 8/96
J EGAP7ECW ENGDEUUNO SUMCES rc Made by:
Date: /
j Chent:
Calculation No.:
AES-C 35431
[f4
- 2/22/98 FPC Chec-y:
Date':
Project No.:
Title:
Deterministic Operational Assessment for
( 2] t.2/9fr AES 98103543-lQ l
Freespan ODSCC Degradation at Crystal River, Unit 3 Ren s rNo.:
Document control No.:
sheet No.:
0 1-2 5 of 27 The structurallimit for axial ODSCC was based on the determination of maximum allowable flaw depth given the above safety factors, a 95%/95% minimum flow stress for the tube material, and a-95% upper limit on axial flaw length. These statistical criteria are consistent with DG-1074.
The limiting flaw depth at end-of-cycle (EOC) was established by probabilistic analysis. Monte Carlo simulation models were used to project the progression of axial corrosion degradation at CR3. The OPCON computer code was used to complete this assessment. Projected flaw numbers, depths, and growth of axial cracks allowed computation of the limiting flaw size at EOC.
The limiting flaw depth was established from extreme value distribution of largest flaw depths observed in the simulation. Again, a 95%/95% upper tolerance limit on largest depth distribution was used to determine the limiting flaw depth.
2.3 Technical Findings The deterministic analysis gave a structural limit for freespan ODSCC as 61.1% through-wall.
This limit was established for 3Ap conditions under normal operating conditions. The probabilistic measure oflimiting flaw depth at EOC was computed to be 54.9%. Therefore, the structural performance criteria of DG-1074 are satisfied for tr e current operating cycle length of 1.8 EFPY.
Projected leak rates at postulated steam line break (SLB) conditions due to freespan ODSCC are expected to be minimal and will not challenge radiological dose limits. This is based on the 95%/95% flaw size at EOC and upper bound flaw shape considerations that indicate that through-wall penetration is not likely during the current operational cycle.
i QAE17 l
REV 8/96 1
l
{
-,, ~..
b5 ENGBS5tNG SOMCEs. FIC.
Ma y:
Date:
Chent:
Calculation No.:
AES-C-3543 1
.(2/22/98 FPC C
. by:
Date:
Project No.:
Title:
Deterministic Operational Assessment for 12/ 2.2['ff AES 98103543-1Q Freespan ODSCC Degradation at Crystal River, Unit 3 R6tmon No.:
Document Control No.:
Sheet No.:
O l2 6 of 27 3.0 ASSUMPTIONS
- The following are the major assumptions used in this calculation:
- 1. Main steam line break pressure is assumed for accident conditions.
- 2. Axial corrosion degradation is modeled as a planar crack. Each axial corrosion indication is idealized as a single planar crack.
- 3. Crack growth over time is assumed to occur primarily in the through-thickness (depth) direction.
i
- 4. The strength distribution for the tubing is characterized by the normal distribution developed in the generic evaluation of once-through steam generator (OTSG) tubing j
(Ref. 3). This strength distribution is assumed to be representative of the total population..
- 5. Nominal tube dimensions are used in the assessment. This assumption is reasonable for computing allowable flaw depths given the conservative tube burst model. This assumption is conservative for computing degradation growth rates.
i e
i QAE17 REV 8/96 4
,e
. - ~
m
- - 55 b
' DONEW SfMCES. E Mad tm Date: ' j j
Chent:
Cciculation No.:
AES-C 3543-1
,2/22/f$
FPC Ch by:
Date:
Project No.:
Title:
Deterministic O# rational Assessment for 12.[2.2./1 P AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Revraion No.:
Document Control No.:
sheet No.:
O I-2 7 of 27
)
4.0 -
DESIGNINFORMATION The steam generator tubing at CR3 is fabricated from Inconel 600 with the following properties (Ref.1):
Outside diameter:
D, = 0.625 inch Nominalwall thickness:
t = 0.037 inch The design conditions for normal operating (NOP) and SLB accident conditions are summarized below (Ref.1):
Design temperature:
650 F Tube differential pressure (NOP):
1350 psig Tube differential pressure (SLB):
2500 psig The material properties per the American Society of Mechanical Engineers (ASME) Code at design temperature (650*F) are given below (Ref. 4):
S.
= 23.3 ksi S,
= 27.4 ksi S,
= 80.0 ksi l
i 4
QAE17 REV 8/96
... _ _..... - _ _ _. _ _. _ _ _.. ~. _
- m. _ _.-. _ _
h!!$
l peopspisse services, sec.
Chent:
)
M kad Dateh[22/f8 y:
Calculation No.: - AES-C 35431 t
FPC Ch y:
Date:[2.J/9fr Project No.:
Title:
' Deterministic Operational Assessment for l1 AES 98103543-!Q Freespan ODSCC Degradation at Crystal River, Unit 3 Re6cn No.:
Document control do.:
sheet No.:
0 I-2 8 of 27
- 5.0 '
TECHNICAL METHODS 5.1 Overall Assessment
? A deterministic operational assessment for freespan ODSCC must satisfy the following performance criteria: all tubes should retain margins of safety against burst consistent with the i
safety factor (SF) implicit in the stress limit criteria of ASME Section III where, SF 3.0 (normal operating conditions)
=
SF 1.4 (accident conditions)
=
Demonstration of the above structural criteria is accomplished by assuring that the limiting flaw size at the EOC does not exceed the structural limit for axial degradation. The statistical limits -
l imposed on through-wall (TW) penetration are 95% probability occurrence level at 95%
i.
confidence.
In projecting the limiting flaw size over time, the uncertainties in detection, sizing, initiation, and growth of flaws are accounted for by probabilistic methods. The limiting flaw size (%TW) is
~ determined by Monte Carlo simulation to obtain the distribution of extreme values for flaw depths.
t The OPCON computer code was used in determining the flaw distributions. The analysis is based on Monte Carlo simulation. This basic calculational technique is one of simulating the processes of crack initiation, crack growth, and detection via eddy current inspection. The Monte Carlo
. simulation model follows these processes over multiple cycles of operation. This allows
[
benchmarking of the model by comparing calculated results for past inspections with actual i
observations from past inspections and pulled-tube data. The simulation model tracks both detected and undetected populations of cracks and deals with actual crack sizes. When L
comparisons are made between calculated results and eddy current observations, an eddy current measurement error is applied to convert predicted real crack sizes to predicted eddy current
~
' observations.
Actual degradation conditions in terms of number of cracks, real crack depths, and lengths can be calculated for any selected time period. The verification of this process for CR3 is given in Ref.1.
QAE17 REV 8/96
i,
.~.
_~
~. -
I Ebhb e
- ENGBSWING SWMCES, Ber.
)
4 Made '
Date:
Chent:
j 12 [98/
Calculation No.: - AES-C-3543-1 r
FPC Ch by:
Date:
Pro}ect No.
Title:
Deterministic Operational Assessment for lif 2.2/9I AEs 98103543-1Q
' Freespan ODSCC Degradation at Crystal River, Unit 3 gegiiioll No.:
Document Control No.:
sheet No.:
0 I-2 9 of 27 5.2
- Determination of Limiting Flaw Distribution.
A probabilistic analysis is used to establish the flaw sizes during a run time cycle of 1.8 EFPY for ODSCC in freespans. The analysis requires the following input distributions:
1.' Tube strength, o + o, y
2.'- Degradation length distribution for ODSCC
- 3. Probability o'f detection (POD) for the inspection for bobbin and Plus Point probes
~ 4. Degradation growth rates for ODSCC j
- 5. Flaw initiation function for ODSCC i
A discus'sion of these data inputs is given in Section 5.5. Measurement error is implicitly accounted for in the growth rate distribution.
The OPCON computer program (Ref. 5) performs a probabilistic analysis that generates flaw size distributions as a function of time. This analysis includes many thousands of simulations that track the condition of the steam generator tubing through several past inspection periods to develop benchmark statistics. The model then projects the degradation mechanism through the current operating cycle in order to predict the structural condition of the generator as a function of cycle duration.
Each mock operating cycle and inspection event within a single steam generator simulation Lconsists of several steps that trace the initiation and development ofindividual cracks. For each potential crack site, a crack initiation time is drawn at random from a cumulative initiation
. function. A certain percentage of the crack sites will have initiated during or prior to the '
- operating cycle of interest.
QAE17 o
REV 8/96
y -
7._
EE APTECW
=
ENGME5WeG SSWICES. NC.
Made by:
Date: k22/f3 Che nt.-
Calculation No.:
AES-C 3543-1 K'
/
FPC Che e y:
Date:
Projeet No.:
Title:
Deterministic Operational Assessment for I2J2.2/ 9/-
AES 98103543-1Q Freespan ODSCC Degradation at Crystal River, Unit 3 Reson No.:
Document Control No.:
Sheet No.:
0 1-2 10 of 27 For each initiated crack, a set of descriptive parameters is drawn at random from appropriate distributions to describe the crack in detail. These parameters include the crack length, the crack form factor, and the strength properties of the tube in which the crack resides. The crack retains these particular features throughout its entire life. A growth rate is then sampled from the growth rate distribution. The growth rate is applied to the crack depth over the interval of time between inspections. The growth is assumed to be linear in time. A new growth rate is sampled after each simulated inspection and applied over the ensuing operating cycle, which accounts for potential changes in local growth environments due to startup transients. The average depth of the crack increases with time, and the maximum depth is correspondingly adjusted according to the crack form factor.
Simulated inspections are performed according to the plant-speci.fic inspection schedules. The flaw depth at the end of a completed operating cycle, together with the POD curve, determine the probability that a particular flaw will be detected during an inspection. A random number is drawn from a uniform distribution and compared to the POD. If the random draw is less than the POD, the flawis detected and removed from service. Undetected flaws are left in service and allowed to grow throughout the following operating cycle, and the process is repeated at subsequent inspections.
All flaws, whether detected or undetected, are examined at the EOC inspections to assess the probability of tube burst and leakage under SLB conditions, as well as to determine actual flaw distributions. When allinitiated flaws have been inspected over the course of prescribed past and future operating cycles, a single Monte Carlo trial of the steam generator is complete. OPCON performs many thousands of such trials to generate the distributions of flaws and the appropriate largest flaw depth for use in this calculation.
5.3 Calculation of Axial Structural Limit The axial structural limit is computed from established tube burst models. The engineering model for computing the burst pressure for an axially degraded tube is given by the following formula:
0.577 (o, + c,) t ~ 1 l(d/t)'
(5-1) p,
=
R t + 2t,
i QAE17 REV 8/96
./ EG APTSOf ENQMEPUNG SERVICES. MC.
Made bv:
Date:
Chent:
Calculation No.:
AES-C-3543-1 84 (4/22/fd!
FPC Che e y:
Date:
Project No :
Title:
Deterministic Operational Assessment for 12-/2.1/1 8' AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 RevEon No.:
Document Contro! No.:
Sheet No.:
0 I-2 11 of 27 where p,is the burst pressure, o is the yield strength, o,is the ultimate tensile strength, t is the y
wall thickness, R is the tube inner radius, e is the characteristic degradation length, and d is the i
characteristic degradation depth. This burst equation was developed by Framatome and has been selected as a reasonable conservative estimate of tube burst after a thorough review and comparison with other models and test data, as discussed in Ref.1. Equation 5-1 produces consistently conservative burst pressures for part TW degradation.
l For calculating the burst pressure for axial 100%TW cracked tubes, a separate equation is used:
t (o, + o,) [0.0613 + 0.536 e-a2m ]
(5-2) p3
=
R where R is the mean radius and 1 = f/(Rt)*8 The TW burst equation was developed by Westinghouse and is based on a regression fit to large data set of burst tests. The inverse expression for critical crack length is given by:
M [-2.224 - 3.646 Ln (p - 0.0569)]
(5-3)
I,,,,
=
where p = p R/t (o, + o ).
These equations have been verified for use in tube integrity calculations, as discussed in Ref.1.
5.4 Leak Rate Model The leak rate through a crack was developed from two-phase flow calculations. The results from j
these numerical computations were fitted to appropriate equations to provide an efficient l
determination of leak rate in Monte Carlo simulation. Friction effects, crack surface roughness, and temperature were included in the determination of the leak rate equation. The leak rate regression equation for SLB conditions is given as:
i QAE17 REV 8/96 l
l l
t
.. ~ -
~_ -.
[ !!h!
peWNEWWNS SUMCES. MC.
Made by:
Date:
j Chent:
Celculation No.: ' AES-C 35431 g.4 t 2/2298 FPC f
Ch ap y:
Date Project No.:
Title:
Deterministic Operational Assessment for 12/2Z/7 AES 98103543 lQ j'
Freespan ODSCC Degradation at Crystal River, Unit 3 Revision No.:
Document Con ut No.:
sheet No.:
O
' l.2 12 of 27 i
~ Q = {a+ b exp[c (A/t)" + d(A/t)]} A p'"
(5-4) where a through d, n, and m are regression coefficients as determined l' rom the two-phase flow
> results, p is differential tube pressure, A is crack opening area, and i is crack length. For NOP
)
. conditions, the leak rat: regression equation is:
Q = a + b exp[c(A/t)" ] A p" (5-5)
N
{1 -d exp[e A'" ]}
=
l where a through e, n, and m are again regression coefficients from the two-phase flow results for NOP pressure case.
The crack opening area is determined from classical fracture mechanics analysis, including plasticity effects. The leak rate model and crack opening area methods have been verified against other models and experimental test data, as discussed in Ref.1. It should be noted that through-wall flaws are not predicted in this assessment and the above leakage modelis not utilized.
' 5.5 StatisticalInput Data
'5.5.1 Mechanical Strength The distribution for tubing strength was taken as a normal distribution function from Ref. 3. The
. distribution of the sum of yield and ultimate tensile strength at elevated temperature is plotted in Figure 1. The mean value for o + o,is 138.7 ksi with a standard deviation of 7.24 ksi. The y
distribution is truncated at a lower value of 115 ksi and an upper value of 162 ksi, which are the limits for the data population.
5.5.2 Degradation I.engths During the eddy current examination, only one indication of ODSCC in the freespan was
- observed. Therefore, a site specific axial crack length distribution could not be developed. For l
QAE17 REv 8/96 l>
L
,w JE$Ap?tal" DONESENG SERVICES. INC, Mad bv:
Date:
Chent:
t phlt-l Calculation No.:
AES-C-3543-1 22;hh t
FPC Chec v:
Title:
Deterministic Operational Assessment for Date:
Project No.:
12[22[ff AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Revsn No.:
Document control No.:
Sheet No.;
O I2 13 of 27 this situation, the axiallength distribution from a plant of similar design was used to represent ODSCC lengths at CR3 (Ref. 3). This distribution has a log mean of-0.404 and a standard deviation of 0.52, and is shown in Figure 2.
5.5.3-Probability of Detection Several types of eddy current probes have been used at CR3, depending on the type and location of degradation. Early inspections of the roll transition and freespan regions used bobbin coil probes. The last two inspections were conducted with bobbin probes for ODSCC. A suitable POD curve for the bobbin coil probe was discussed in Ref.1 based on log-logistic regression fits of the form:
1 P
(5-6)
=
1 + exp (11.94 - 7.36x) where x is logw [100 (d./t] and d,,is the maximum depth of the crack. The log-logistic fit was used in the generic study of OTSG tubing because it is conservative at the extremes of the depth spectrum, especially for very deep flaws. A plot of the POD functions is given in Figure 3.
5.5.4 Degradation Growth Rate The axial growth rate for ODSCC in freespan regions was developed in a project described in Ref. 3. A large set of gro'wth rate data was evaluated as part of a BWOG activity fc,r freespan axial i
degradation in OTSG tubing. Depth sizing uncertainties and POD effects were considered. The j
best estimate growth rate distribution was established as a log normal distribution with a hiedian growth rate of 3.6% TW per EFPY ( = 1.28) and a standard deviation of 0.65. A plot of the freespan ODSCC growth rate is shown in Figure 4.
A certain percentr.ge of growth rates are taken to be zero. This accounts for the fact that measured growG rates,in the presence of measurement error, frequently appear to be less than zero, which is physically unrealistic. A zero growth set at 20% cumulative distribution function when comNned with the log normal distribution and assumed error distribution in growth rate analyses, produces a net result that reflects the physical growth observation (Ref. 3).
QAE17 REV 8,96
,.g.
b GGSGSWIG SUMCES. MC.
Made b :
Date:
/
Chent:
Calculation No.:
AES-C-3543-1
/*'/k/h FPC Che e Date:
Project No.:
Title:
Deterministic Operational Assessment for l 2/2 'Z/M AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 RevniorrNo.:
Document controi No.:
sheet No.:
0 I-2 14 of 27 Note that growth rates are based on change in average crack depth rather than maximum depth.
The average depth growth rate is used in the simulation model. For any particular crack, maximum depths are calculated based on a known distribution of ratios of maximum depth to structural (average) depth.
5.5.5 Flaw Initiation Function The initiation function for axial flaws due to ODSCC is based on a modified Weibull function, which requires a scale parameter and a slope parameter. The scale parameter reflects the length of time required to initiate a given percentage of all potential crack sites. This parameter may be on the order of several decades. The slope parameter is a measure of the rate ofincrease in initiated flaws over time. The scale and slope parameters are adjusted iteratively until the number ofindications produced by the simulation matches the actual number of flaws detected at recent plant inspections.
4 This calculation considers past inspections for which eddy current inspection results are available.
Since axial cracking ODSCC freespan locations was first detected in the last inspection, the parameters of the Weibullinitiation function were set by: (1) assuming a Weibull slope of 6 and (2) adjustmg the scale factor until the predicted number ofindications matched the results of the last inspection. A Weibull slope value of 6 is a good conservative estimate of the rate ofincrease of initiated cracks. This is supported by extensive studies of degradation progression rates in steam generator tubing (Ref. 6). After several trials, a scale factor of 31.0 was determined to give the same total number of detected flaws from the last inspection.
QAE17 REV 8/%
L ES APTEOF l
l
\\
/2-/22/ff Calculation No.:
AES-C-3543-1 l
Che Date. /2.'1/ff Project No.:
l
Title:
Deterministic Operational Assessment for l 2.
AES 98103543-lQ l
Freespan ODSCC Degradation at Crystal River, Unit 3 Revision No.:
Document Control No.:
Sheet No.:
0 1-2 15 of 27 6.0 CALCULATION OF STRUCTURAL LIMIT FOR AXIAL DEGRADATION l
Consistent with the criteria given in DG-1074, the structural limit for axial ODSCC in the freespan region was computed from Eq. 5-1 for a 95%/95% statistical bound on the input parameters, specifically, mechanical strength and flaw length. Rearranging Eq. 5-1 to solve for d/t, j
t,3 + 2 t
~
p R, d/t 1-(6-1)
=
t,3 0.577 S,3 t, where d/t
= Structurallimit on flaw depth = (d/t)uu R
= Tube inside radius i
t
= Tube wall thickness i
S,5
= 95%/95% statisticallower bound on o, + o, f,5
= 95%/959o statistical upper bound on flaw length p
= Tube differentialpressure = 3Ap The 95%/95% lower bound strength is determined from the normal distribution function. It is reasonably assumed that the 95%/95% tolerance limit for a very large sample is equivalent to S,5
-ko
=
where k is the 95% one-sided limit for infinite population. From the normal distribution tables, k = 1.645. Therefore, S,5 138.7 - 1.645(7.24) 126.8 ksi
=
=
The 95%/95% upper limit on axial flaw length is determined in a similar manner using a log normal distribution. Because the sample size for length measurements is small, a k factor of 2.0 is assumed to be more representative of a 95% confidence level.
QAE17 REV 8/%
ENGSEWWNG $5MCES. P0C.
Made by:
Date:
Chent:
Cciculation No.:
AES-C-3543-1 g2%
IL[22/f8 FPC f-Chec e y:
Date Project No.:
Titlei Detenninistic Operational Assessment for 1[2 t/98 AES 98103543 !Q l
Freespan ODSCC Degradation at Crystal River, Unit 3 RevsiioirNo.:
Document Control No.:
Sheet No.:
I 0
1-2 16 of 27 t,5 eXP( + kc) = exp [-0.404 + 2(0.52)]
~
l 1.89 inches t,5
=
The structural limit can now be calculated for NOP and SLB conditions. For NOP:
3Ap = 3(1.35) = 4.05 ksi p
=
R
= 0.625/2 - 0.037 0.2755 inch
=
i
-and 1.89 + 2(0.037) ' 1 -
4.05(0.2755)
~
d/t
=
1.89 0.577(126.8)(0.037),
d/t = (1.0392)(0.5878) = 0.611 Therefore, for NOP, the %TW structurallirnit on axial freespan ODSCC is (d/t)uu = 61.1%.
For SLB conditions:
)
l 1.4(2.50) = 3.50 ksi p
=
i and (3.50)(0.2755) d/t = - (1.0392) 1-(0.577)(126.8)(0.037),
l-d/t 0.669
=
Therefore, the %TW structural limit for SLB conditions is (d/t)uu = 66.9%. The minimum structural limit is set by NOP criteria.
QAE17 REV 8/96 l
lEOAPTECH' ENGDE54NG SUMCES. ING, Made v:
Date:[2/98 Chent:
Calculation No.:
AES-C 3543-1 t 2-FPC Ch y:
Date: [ 2 2/ 9 *P Project No.:
Title:
Deterministic Operational Assessment for i1 AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 RcGon No.:
Document Control No.:
Sheet No.:
0 I2 17 of 27 7.0 DETERMINATION OF LIMITING FLAW SIZE The limiting flaw size was established using a probabilistic assessment of the extreme values for the structurally significant crack depth. The crack depth distribution as a function of time has been previou. sly computed in APTECH Calculation AES C-3350-1 (Ref.1). This was accomplished by Monte Carlo simulation. The mean and standard deviation of the distribution of largest depths observed for each trial also were computed as part of that analysis. These largest depths are given below from Appendix B of Ref.1:
Largest Flaw Depth, d/t Standard Time (EFPY)
Mean. x Deviation, S 8.7 0.118 0.107 10.2 0.192 0.103 9
11.9 0.287 0.106 12.1 0.286 0.104 13.9 0.364 0.095 The above results are assumed to be the sample statistics in an extreme value distribution behavior of the largest flaw sizes from 10,000 simulations of plant operations from the beginning of service.
The limiting flaw size at each cycle end was determined from a Type I asymptotic probability function (or Gumbel's extreme value distribution) for the largest element. A Type I asymptotic distribution for maximum values is the limiting model of the maximum of n independent values from initial distribution, whose right tailis unbounded and which is exponentialin form. -
Therefore, the use of the extreme value distribution for maximum values to define the 95W95%
r limiting flaw depth will be conservative as applied to the deterministic performance criteria of DG-1074.
QAE17 REv 8/96
$NAPTECH' ENCAEDUNG SERVICES. NO.
M[ad bv:c// (
22/ 98 Date:
Client:
C:lculation No.:
AES-C-3543-1 FPC Ch y:
Date:
Project No.:
Title:
Deterministic Operational Assessment for
( 2 [22/y &
AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 RcGrsion No.:
Document Control ho.:
Sheet No.:
0 1-2 18 of 27 The parameters of the extreme value distribution from the sample mean and standard deviation from Eq. 3-55 and Eq. 56a of Ref. 7:
((6'/n]S (7-1) 6
=
x -0.577 6 (7-2)
=
The cumulative probability function of the extreme value as P is given by:
P exp[-e Y ]
(7-3)
=
- ~E y
(7-4)
=
U where z is the value for flaw depth (d/t) at the given probability, P. For P = 50%, y = 0.367 for median estimate oflargest flaw depth. For P = 95%, y = 2.97 for 95% upper bound estimate of largest flaw depth.
Solving for t in Eq. 7-4 gives:
(d/t) z =
oy +
=
Largest Flaw Depth, d/t Time (EFPY) 6 il at 50%
at 95%
8.7 0.08342 0.06986 0.1004 0.3177 10.2 0.08031 0.1457 0.1751 0.3842 11.9 0.08265 0.2393 0.2696 0.4848 12.1 0.08109 0.2392 0.2689 0.4801 13.9 0.07407 0.3213 0.3484 0.5413 QAE17 REv 8/96
$5APTEGl*
l ENGINEERN3 SUMCES. P0C.
Made Date:
/
Chent:
Cciculation No.:
AES-C-3543 1 (2 %y99 j
FPC Chec Date:
Project No.:
Title:
Deterministic Operational Assessment for
[.p:
l 2/2,2./fJ' AES 98103543-lQ p
l Freespan ODSCC Degradation at Crystal River, Unit 3 gevisr6nwo.:
Document Control No.:
Sheet No.:
0 I2 19 of 27 i
The values for d/t at 95% occurrence are taken as the limiting flaw depth for use in the deterministic operational assessment.
To verify the above results, the OPCON problem from APTECH Calc AES C-3350.-1 was rerun with the values for largest depths printed to a file. A nonparametric statistical evaluation was used to determine the 95%/95% depth value. The output for d/t for each cycle was sorted in ascending order and the 9,537th value was obtained. Based on a distribution-free theory, the ranked observation represents a 95%/95% one-sided tolerance limit for the largest flaw depth distribution. A summary and comparison of the results is given below:
Extreme Value Time (EFPY) d/t (9.537th rank) d/t @ 95%
8.7 0.326 0.318 10.2 0.379 0.384 11.9 0.480 0.485 12.1 0.469 0.480 13.9 0.549 0.541 The results are in very good agreement. Therefore, the limiting flaw depth (d/t) at EOC is defined as:
(d/t)soe 54.9%TW
=
l l
l QAE17 REV 8/96
S APTEGI'
\\
Mad v Date:
Chent:
)
/2/22/98 Calculation No.:
AES-C-3543-1 FPC Che y:
Date:
Project No.:
Title:
Deterministic Operational Assessment for I 2.[2.Z/18 AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Reviiion No.:
Document Control No.:
sheet No.:
0 I-2 20 of 27
)
8.0 OPERATIONAL ASSESSMENT 8.1 StructuralIntegrity The structural criteria of DG-1074 require the limiting flaw depth at EOC to be less than the 3 A;
- or 1.4 SLB stnicturallimit for single tube burst. The 3Ap limit is the limiting condition and (d/t)uu 61.1%TW > (d/t)coc 54.9 %TW
=
=
for a cycle length of 1.8 EFPY (EOC = 13.9 EFPY). Therefore, the structural limit for freespan ODSCC is satisfied for the current cycle length.
A plot of the deterministic evaluation for all cycles, including the current cycle,is shown in Figure 5. For comparative purposes, the growth in both mean extreme value depth and 95%
extreme value depth are shown from initial plant startup to the end of the last cycle.
8.2 Accident Leakage In consideration ofleakage under SLB conditions, the structurallimit of d/t = 61.1 %TW implies no TW cracks. To verify that the maximum depth is less than 100%TW, the shape factor is defined as:
d '"*"
F
=
d,,
where the mean value for F is 1.28, with a standard deviation of 0.12. The 95%/95% value for F is estimated from:
F,5 1.28 + 1.645(0.12) 1.48
=
=
QAE17 REV 8/%
i
._..m.
_m
-. _. - - =
ENGDESUNG SEfMCES. DIC.
Made -
Date: -
j Chent:
Calculation No.:. AES-C-3543-1
( arf 9$
FPC i
Che Date]-l2 2.M MAES 98103543-lQ Project No.-
Title:
Deterministic Operational Assessment for Freespan ODSCC Degradation at Crystal River, Unit 3 Rewsion No.:
Document Controt No.:
Sheet No.:
0 I-2 21 of 27 The maximum TW penetration is 1
d,/t F,5(d/t)uu 1.48(61.1) = 90.4 %TW
=
=
y Therefore, no leakage is expected from any flaw at EOC under SLB conditions since du,is less i
than the nominal tube wall thickness.
i 1
l QAE17 REV 8/96
I
!!!!!M
~
s.
L k omsesume semcas, sc. -
. l oste:12/22/98 M
y:
Chent:
Cciculation No.:
AES-C 35431 FPC Che Date:
Project No.:
l
Title:
Deterministic Operational Assessment for -
d.
i M22/f/
AES 98103543 lQ j
l; Freespan ODSCC Degradation at Crystal River, Unit 3 RevtFion No.:
Document Control No.:
sheet No.:
0 12 22 of 27 9.0
. REFERENCES I
l'
- "Operationci Run ilme Analysis for Crystal River, Unit 3," Aptech Engineering Services,
'D Inc., Calculation C-3350-1, APTECH Project AES 9803335010 (December 22,1998).
2.
Draft NRC Regulatory Guide DG-1074, " Steam Generator Tube Integrity", Nuclear l
Regulatory Commission (September 1997).
3.
Begley, C. J.,"A Generic Probabilistic Analysis of the Effect of Axial Freespan Corrosion on the Structural Performance and Leak Rate Behavior of OTSG Tubing," Aptech r
Engineering Services,Inc., APTECH Project AES 96102886-1Q (June 1998).
i 4.
ASME Boiler and Pressure Vessel Code, "Section III, " Appendices," 1989 Edition.
5.
"OPCON-Operational Assessment and Condition Monitoring System," Version 1.0, Aptech Engineering Services, Inc. (August 1998).
6.
Gorman, J. A., R. W. Staehle, and K. D. Stravropoulas, " Statistical Analysis of Steam Generator Tube Degradation," Electric Power Research Institute, EPRI Report NP-6967 (1991).
7.
Hahn, G. J. and S. S. Shapiro, " Statistical Models in Engineering," John Wiley (1967).
e QAE17 REY 8/96 I
9 i--
g g.
I $$ APTEG' I
5 ENWPEEmpse SGMCEs. s0C.
Mad by:
Date:
Gient:
h (2-[22/f8 Calculation No.:
AES-C-3543-1 FPC Che Project No.:
Title:
Cycle 8 Deterministic Operational Assessment for Dateh 2/ 22/f PAES 98103543-1Q Freespan ODSCC Degradation at Crystal River, Unit 3 Revdion No.:
Document Control No.:
Sheet.No.:
0 I2 23 of 27
. 0.06 0.05 0.04 5
E E 0.03 a
$e 0.02 0.01 7
0.00 110 120 130 140 150 160 170 Sum of Yleid and Ultimate Tensile Strength at Temperature, kal Figure 1 - Tubing Strength Distribution.
QAE17 REV 8/96 Y
$$APTECH' 3
s ENGNEERN3 Sf) WEE 1 FC Made y:
Date: [22;/98 Chent:
Calculation No.:
AES-C-3543 1
/1 FPC Che Date P.oject No :
Title:
Cycle 8 Deterministic Operational Assessment for
% L2/1f AES 98103543-lQ Freespan ODSCC Degradation at Crystal her, Unit 3 ReEston No.:
Docurnent Control No.;
Sheet No.:
0 I2 24 of 27 1.0 W
f-b
/
0.9
=
0.8 f
f 0.7 --
f e
tb 80.s i
=
5
=
a tt 0.5
.I a
i b
j CA 0.3 I
0.2 0.1
- Plant A 0.0 0.000 0.500 1.000 1,500 2.000 2.500 3.000 Cracklength, Inches Figure 2 - Axial Length Distributions.
QAE17 REV 8/96
's ENWNE54NG SOMCES. He, -
Made by:
Date:
Chent:
Cciculation No.:
AES-C-3543-1
/O
/2/22/98 FPC Che Date:
Protect No.:
Title:
' Cycle 8 Deterministic Operational Assessment for l Q2.2/9[
A8S98103543-lQ Freespan ODSCC Deg'radation at Crystal River, Unit 3 Revision No.:
Document contros No.:
sneet No.:
0 I-2 25 of 27 1.0 J
0.9
]
A B
0.8 l
0.7 8' 0.6 s
I i
0.5 z
V a5 j 0.4 c
0.3 "-'
O.2
-A Plus Point
-B Bobbin 0.1 r
0.0 0
20 40 60 80 100 Maximum Depth, %TW Figure 3-Probability of Detection Versus Maximum Axial Crack Depth.
i QAE17 REV 8/96
O ANG5E SGMG 50 MCg1L NC,
~
Mad a by:
Date:g22/fc9 Client:
' Calculation No.:
AES-C 35431 g&_.,
I FPC Chec'
Title:
Cycle 8 Deterministic Operational Assessment for
(
'y:
Date:
Project No.:
(2/2.?[%
AES 98103543-lQ Freespan ODSCC Degradation at Crystal River, Unit 3 Rewstfowa.:
Document Control No.:
Sheet No.:
0 1-2 26 of 27 1.0 I
- /
f 0.9 A
8 y
/
I
/
0.7 u.
00.6 i
5 50.5
.E f
5=
! 0.4
/
0 0.3 0.2
-A Ro!!Transttlen
-B Freescan 0.1 0.0 0
5 10 15 20 25 30 35 Growth Rats, %TW/EFPY Figure 4 - Axial Crack Growth Rate Distributions.
QAE17 REV 8/96
y SENcts. INC.
Mad by:
Date:[22/98 Chent:
Calculation No.:
AES-C-3543-1 N-s
/2 FPC i.
Che y:
Date:
Project No.:
l
'~
Title:
' Cycle S Deterministic Operational Assessment for
-l2/2 2.// @
AES 98103543-lQ l
Freespan ODSCC Degradation at Crystal River. Unit 3 Revision No.:
Document Controi No.:
shesc No.:
0 I-2 27 of 27 CR3 Operational Assessment 1.0 i...i.i
.i.i i.i.>>
.,,i.i.,,
NOP = 1350 si E0C P
O.9
~
SLB = 2500 psi 5
i 0.8 N
0.7 SLB Limit 3
a.
g 0.6 3ap umit 1
0.5 95/95 -
i g.
N 0.4
~@
Mean E
- 0.3 j
E i
0.2 i
g 0.1 -
i
)
0.0 0
1 2
- 3. 4 5
6 7
8 9
10 11 12 13 14 15 16 Operational Time, T (EFPY)
Figure 5 - Structural Limits for Limiting Degradation Depths for Freespan ODSCC.
l i
,a d
e i
QAE17 REV 8/96 l