ML20081E815
| ML20081E815 | |
| Person / Time | |
|---|---|
| Site: | Comanche Peak  |
| Issue date: | 10/31/1983 |
| From: | Furchi E, Mcinerney J, Swamy S WESTINGHOUSE ELECTRIC COMPANY, DIV OF CBS CORP. |
| To: | |
| Shared Package | |
| ML19268E343 | List: |
| References | |
| MT-SME-3136-(NP, MT-SME-3136-(NP), NUDOCS 8311020248 | |
| Download: ML20081E815 (30) | |
Text
.
11T-SiiE 3136 TECHNICAL BASES FOR ELIMINATING LARGE PRIMARY LOOP PIPE RUPTURES AS THE STRUCTURAL DESIGN BASIS FOR COMANCHE PEAK UNITS 1 AND 2 Prepared by: E. L. Furchi S. A. Swamy J. J. McInerney October,1983 i
APPROVED: .
D [ J APPROVED: 1kk7s' J) N. Chirigos, Managdr E. R. Johnson, Manager Structural Materials Structural and Seismic Engineering Development APPROVED: /2J[b W. S. Brown, Manager Mechanical Equipment and Systems Licensing 831102024e 831031 PDR ADOCK 05000 A .
Section Title Page
1.0 INTRODUCTION
2 i
2.0 OPERATION AND CHEMICAL STABILITY OF THE PRIMARY 5 4
COOLANT SYSTEM 3.0 PIPE GE0 METRY AND LOADING 7 4.0 FRACTURE MECHANICS EVALUATION 9 5.0 LEAK RATE PREDICTIONS 13 6.0 FATIGUE CRACK GROWTH ANALYSIS 14
7.0 CONCLUSION
S 16
8.0 REFERENCES
17 APPENDIX A 19 i
1
1.0 INTRODUCTION
1.1 Puroose The current structural design basis for the reactor coolant system (RCS) primary loop requires that pipe breaks be postulated as defined in the approved Westinghouse Topical Report WCAP 8082, Reference 1. In addition, protective measures for the dynamic effects associated with RCS primary loop pipe breaks have been incorporated in the Comanche Peak plant design.
However, Westinghouse has demonstrated on a generic basis that RCS primary loop pipe breaks are highly unlikely and should not be included in the structural design basis of Westinghouse plants (see Reference 2). The purpose of this report is to demonstrate that the generic evaluations performed byi Westinghouse are applicable to the Comanche Peak plant. In order to demonstrate this applicability, Westinghouse has performed a comparison of the loads and geometry for the Comanche Peak plant with envelope parameters used in the generic analyses (S.ection 3.0); fracture mechanics evaluation (Section 4.0); determination of leak rates from a through-wall crack (Section 5.0),
f atigue crack growth evaluation (Section 6.0); and conclusions (Section 7.0).
1.2 Scope This report applies to the Comanche Peak plant reactor coolant system primary loop piping. It is intended to demonstrate that specific parameters for the Comanche Peak plant are enveloped by the generic analysis performed by Westinghouse in WCAP-9570 (Reference 3) and accepted by the NRC as noted in a letter from Harold Centon dated May 2,1983 (Reference 4).
1.3 Objectives The conclusions of this report (Reference 3) support the elimination of RCS primary loop pipe breaks for the Comanche Peak plant. In order to validate this conclusion the following objectives must be achieved.
0450e:1/102683 .
2 9
- a. Demonstrate that Comanche Peak plant parameters are enveloped by generic Westinghouse studies.
- b. Demonstrate that margin exists between the critical crack size and a postulated crack which yields a detectable leak rate.
- c. Demonstrate that there is sufficient margin between the leakage through a postulated crack and the leak detection capability of the Comanche Peak plant.
- d. Demonstrate that fatigue crack growth is negligible.
1.4 Background Information i
Westinghouse has performed considerable testing and analysis to demonstrate that RCS primary loop pipe breaks can be eliminated from the structural design basis of all Westinghouse plants. The concept of eliminating pipe breaks in the RCS. primary loop was f.irst presented to the NRC in 1978 in WCAP 9283 (Reference 5). This Topical Report employed a deterministic fracture mechanics evaluation and a probabilistic analysis to support the elimination of RCS primary loop pipe breaks.
This approach was then used as a means of addressing Generic Issue A-2 and Asymmetric LOCA Loads. Westinghouse performed additional testing and analysis to justify the elimination of RCS primary loop pipe breaks. As a result of this effort, WCAP 9570 was submitted to the NRC. The NRC evaluated the technical merits of this concept and prepared a draft SER in late 1981 endorsing this concept. Additionally, both Harold Denton and the ACRS have endorsed the technical acceptability of the Westinghouse evaluations.
Specifically, in a May 2,1983 letter (Reference 4) Harold Denton states that
. . . it is technically satisfied with Westinghouse Topical Report 9570 Rev.
2 . . . ." Additionally, the ACRS stated in a June 14, 1983 letter (Reference
- 6) that "... there is no known mechanism in PWR primary piping material for developing a large break without going through an extended period during which the crack would leak copiously." ,
W0rsh.A/DViWM9
- The NRC funded research through Lawrence Livermore National Laboratory (LLNL) to address this same issue using a probabilistic approach. As part of the LLNL research effort, Westinghouse performed extensive evaluations of specific plant loads, material properties, transients, and system geometries to demonstrate that the analysis and testing previously performed by Westinghouse and the research performed by LLNL applied to all Westinghouse plants including Comanche Peak (References 7 and 8 ). The results from the LLNL stuay were released at a March 28, 1983 ACRS Subcommittee meeting. These studies which are applicable to all Westinghouse plants east of the Rocky Mountains, determined the mean probability of a direct LOCA (RCS primary loop pipe break) to be 10-10 per reactor year and the mean probability of an indirect LOCA to b e 10-7 per reactor year. Thus, the results previously obtained by Westinghouse (Reference 5) were confirmed by an independent NRC research study.
The above studies establish the technical acceptaoility for eliminating pipe breaks from the Westinghouse RCS primary loop. The LLNL study has been shown applicable to Comanche Peak plant by inclusion of plant specific data. This report will demonstrate the applicability of the Westinghouse generic evaluations to the Comanch'e Peak plant.
a j
't I
t- fiYN.Tihd/L5kT@Widt .A
, 2.0 OPERATION AND CHEMICAL STABILITY OF THE PRIMARY COOLANT SYSTEM The Westinghouse reactor coolant system primary loop has an operating history (over 400 reactor years) which demonstrates its inherent stability characteristics. Additionally, there is no history of cracking in RCS primary loop piping. In addition to the fracture resistant materials used in the piping system, the chemistry of the reactor coolant is tightly con-trolled and variations in temperatures, pressure and flow during normal operating conditions are insignificant.
As stated above, the reactor coolant chemistry is maintained within very specific limits. For example, during nonnal operation oxygen in the coiolant
_.+
i is limited to less than This stringent oxygen limit is achieved +a,c,e by controlling charging flow chemistry and maintaining hydrogen in the reac-tor coolant at a concentration of The oxygen concentra +a,c.e
, tion in the reactor coolant is verified by routine sampling and chemical analy-sis. Halogen concentrations are also stringently controlled gmaintaining concentrations of chlorides and flourides at or below . This concen +a,c e tration is assured by controlling charging flow chemistry and specifying
! proper wetted surface materials. Halogen concentrations are also verified by routine chemical sampling and analysis.
In order to ensure dynamic system stability, reactor coolant parameters are stringently controlled. Temperature during normal operation is maintained .
within by control rod position. Pressure is controlled by pressurizer +a,c.e
~
heaters and hessurizer spray, to a variation of less than for steady +a,c,e;
~ -
state conditions. The flow characteristics of the system remain constant during a fuel cycle because the only governing parameters, namely system re-l sistance and the reactor coolant pump characteristics are controlled in the design process. Additionally, Westinghouse has instrumented typical reactor coolant i systems to verify the flow characteristics of the system.
1 i
1 l
l l
t 5
l
The reactor coolant system, including piping and primary components, is designed for normal, upset, emergency and faulted condition transients.
The design requirements are conservative relative to both the number of transients and their severity. '
i I
f i
i 3
f 4
l 6
i
3.0 PIPE GE0 METRY AND LOADING A segment of the primary coolant hot leg pipe is shown in Figure 1. This segment is postulated to contain a circumferential through-wall flaw. The i
inside diameter and wall thickness of the pipe are 29.0 and 2.45 inches, respectively. The pipe is subjected to a normal operating pressure of [ ] +a ,c.e i
psi. The design calculations indicate that the junction of the
[- ]+ is most highly stressed. At +a,c.e this location the axial load, F, and the total bending moment M, are [ ]+ +a,c.e kips (including the axial force due to pressure) and [ ]+1n-kips, +a,c.e respectively. Figure 2 identifies the loop weld locations. The material properties and the loads at these locations resulting from Deadweight, Thermal Expansion and Safe Shutdown Earthquake are indicated in Table 1. The metho'd of obtaining these loads can be briefly summarized as follows:
4 j The axial force F and transverse bending moments, My and M , are chosen z
for each static load (pressure, deadweight and thermal) based on elastic-static analyses for each of these load cases. These pipe load
! components are combined algebraically to define the equivalent pipe static
- loads Fs ' Mys, and Mzs. Based on elastic SSE response spectra analyses, amplified pipe seismic loads, Fd ' bd' Mzd are obtained. The maximum pipe loads are obtained by combining the static and dynamic load components as follows
1 F= F + F s d M. M 2g 2 '
where h* ys
+
yd
+
M z" zs zd I
b
The corresponding geometry and loads used in the reference report (Reference
- 3) are as follows: inside diameter and wall thickness are 29.0 and 2.5 inches; axial load and bending moment are [ ]+ inch +a,c,e kips. The outer fiber stress for Comanche Peak is [ ]+ ksi, while for the +a,c.e reference report it is [ ]+ksi. This demonstrates conservatism in the +a,c.e reference report which makes it more severe than the Comanche Peak project.
0450e:1/102683 8
4.0 FRACTURE MECHANICS EVALUATION 4.1 Global Failure Mechanism Determination of the conditions which lead to failure in stainless steel must be done with plastic fracture methodology because of the large amount of deformation accompanying fracture. A conservative method for predicting the failure of ductile material is the
[
i
]+. This methodology has been shown to be applicable to +a,c.e ductile piping through a large number of experiments, and will be used here to predict the critical flaw size in the primary coolant piping. The failure criterion has been obtained by requiring
[ ]+ (Figure 3) when loads are +a,c,e aoplied. The detailed development is provided in Appendix A, for through-wall circumferential flaw in a pipe with internal pressure, axial force and imposed bending moments. The [ ]+for such a pipe is given by: +a,c,e
+a,c,e i nn2 Ant 9
+a,c.e The analytical model described above accurately accounts for the piping internal pressure as well as imposed axial force as they affect i
[ ]+. Good agreement was found between the analytical +a,c.e predictions and experimental results [9].
4.2 Local Failure-Mechanism The local mechanism of failure is primarily dominated by the crack tip behavior in terms of crack-tip blunting, initiation, extension and finally crack instability. Depending on the material properties and geometry of the pipe, flaw size, shape and loading, the local failure mechanisms may or may not govern the ultimate failure.
The stability will be assured if the crack does not initiate at all.
It has been accepted that the initiation toughness, measured in terms of J gg from a J-integral resistance curve is a material parameter defining the crack initiation. If, for a given lead, the calculated J-integral value is shown to be less than J yg of the material, then the crack will not initiate. If the initiation criterion is not met, one can calculate the tearing modulus as defined by the following relation: -
dJ E T
app. da 2
'f 10 rueindwnwre _ _ , _ _
- wh;re T,pp = applied tearing modulus E = modulus of elasticity
]+ (flow stress) #' '*
of = [
a = crack length
[ ]+, +a,c,e In summary, the local crack stability will be established by the two step criteria:
J<J IN T,pp < Tmat if J > J IN 4.3 Results of Crack Stability Evaluation Figure 4 shows a plot of the [ ]+ as a function of through- +a,c,e wall circumferential flaw length in the [ ]+ of the main coolant piping. +a,c,e This[ ]+ was calculated for Comanche Peak data of a pressurized pipe +a,c,e at [ ]+ with ASME +a,c,e Codeminimum[ ]+ properties. The maximum applied bending moment of +a,c.e
[ ]+ in-kips can be plotted on this figure, and used to determine the +a,c e critical flaw length, which is shown to be [ ]+ inches. This is +a,c.e considerably larger than the [ ]+ inch reference flaw used in Reference 3. +a,c,e l
11 L 0450e:1/102683
[
+a,c.e
]+. Therefore, it can be concluded that a postulated [ ]+ inch +a,c.e through-wall flaw in the Comanche Peak loop piping will remain stable from both a local and global stability standpoint.
i 12
5.0 LEAK RATE PREDICTIONS Leak rate calculations were perfonned in Reference 3 using an initial through-wallcrack[ ]+. The +a,c,e computed leak rate was [ ]+basedonthenormaloperatingpressureof +a,c.e
[ ]+ psi.
+a,c.e
]+. This computed leak rate [ ]+
significantly exceeds the smallest detectable leak rate for the plant. The s
Comanche Peak plant has a RCS pressure boundary leak detection system which is consistent with the requirements of Regulatory Guide 1.45 and can detect leakage of 1 gpm in one hour. There is a factor of [ ]+ between the i +a,c.e calculated leak rate and the Comanche Peak plant leak detection systems.
t k
13 I .-
6.0 FATIGUE CRACK GROWTH ANALYSTS 1
To determine the sensitivi.ty of the primary coolant system to the prcaence of small cracks, a fatigue cra;ck growth analysis was carried out for the
[ s ]+ region of a typical system. This region was +a,c e selected because it is typically one of the highest stressed cross sections, and crack growth calculated here will be conservative' for. application to the entire primary coolant system.
s
~
A finite element stress analysis was carried out for the
[ ]+ of a plaat typical in geometry and ' operational +a,c.e characteristics to any Westinghouse PWR System. [
\
]+. All normal, upset and test conditions were considered, and +a,c.e circumferential1y oriented. surface flaws were postulated in the region, -
assuming the flaw was located in three different locations, as shown in
- Figure 5. Specifically, these were
Cross Section A:
Cross Section B: +a,c.e Cross Section C:
Fatigue crack growth rate laws were used [
+
]+. The law for stainless steel was derived from +a,c,o Reference 11, with a very conservative correction for R ratio. The ratio of minimum to maximum stress during a transient is:
14 0450e:1/102683 _ _
4.48 h = (5.4 x 10-12) g inches / cycle where keff " Kmax (I-NI i
R-Kmin/Kmax s,
i
+a,c.e s
The calculated fatigue crack growth for semi-elliptic surface flaws of circumferential orientation and various depths is summarized in Table 2, and shows that the crack growth is very small re.sardless [
]+ +a,,c e
.6 15
7.0 CONCl.USION S This report has established the applicability of the generic Westinghouse evaluations which justify the elimination of RCS primary loop pipe breaks for the Comanche Peak plant as follows:
- a. The loads, material properties, transients and geometry relative to the Comanche Peak RCS primary loop are enveloped by the parameters of WCAP 9570.
b, The critical . crack length at the worst location in the RCS primary loopis[ ]+. This is significantly greater than the +a,c.e
[. ]+ inches stable crack used as a basis for' calculating leak i +a,c.e rates in WCAP 9570.
]+ crack in the RCS primary loop is [ ]+ +a,c,e
- c. The leakage through a'[
~ based on WCAP 9570. The Comanche Peak plant has a RCS pressure boundary leak detection' system which is consistent with the requirements of Regulatory Guide 1.45 and can detect leakage of 1 gpm Fa,c e in one hour. Thus,-there is a factor of ( ]+betweenthe calculated leak rate and the Comanche Peak plant leak' detection systems.
- d. Fatigue crack growth was determined for postulated flaws and was found to be extremely small over plant life and, therefore, is considered insignificant.
Based on the above, it is concluded that RCS primary loop pipe breaks should not be considered in the structural design basis of the Comanche Peak plant.
}
16
8.0 REFERENCES
- 1. WCAP 8082 P-A, " Pipe Ereaks for the LOCA Analysis of the Westinghouse Primary Coolant Loop", Class 2, January 1975.
- 2. Letter from Westinghouse (E.P. Rahe) to NRC (R. H. Vollmer) dated May ll, 1983.
- 3. WCAP 9570, Rev. 2, " Mechanistic Fracture Evaluation of Reactor Coolant Pipe Containing a Postulated Circumferential Through-Wall Crack", Class
- 3. June 1981.
- 4. Letter from NRC (H. R. Denton) to AIF (M. Edelman) dated May 2,1983.
i
- 5. WCAP 9283, "The Integrity of Primary Piping Systems of Westinghouse Nuclear Power Plants During Postulated Seismic Events", Class 2, March,1978.
- 6. Letter from ACRS (J.J. Ray) to NRC (W.J. Dircks) dated June 14, 1983.
- 7. Letter From Westinghouse (E.P. Rahe) to NRC (W. V. vohnston) dated April 25, 1983.
- 8. Letter From Westinghouse (E.P. Rahe) to NRC (W. V. Johnston) dated July 25, 1983.
- 9. Kanninen, M. F. , et. al., " Mechanical Fracture Predictions for Sensitized Stainless Steel Piping with Circumferential Cracks" EPRI NP-192, September 1976.
- 10. Bush, A. J. Stooffee, R. 8., " Fracture Toughness of Cast 316SS Piping Material Heat No. 156576, at 600*F", W R&D Memo No. 83-5P6EVMTL-M1, March 7, 1983, Westinghouse Prop. Class 2. ,
- 11. Bamford, W. H. , " Fatigue Crack Growth of Stainless Steel' Piping a Pressurized Water Reactor Environment" Trans. ASME Journal of Pressurized Vessel Techno-logy Vol .101, Feb.1979.
17
+-'-
- M '-4YW 9 r-m*-"* e*+f 4 *w^e 'h W e"---T W " W v m-- " M ~ '-- -
{
+a,c.e J+
18 0450e:1/102683
APPENDIX A
+a,c,e 2
19 0450e:1/102683
TABLE I COMANCHE PEAK PRIMARY LOOP DATA
__ M fa,C,e i
o
(
G wi
-% f n
TABLE 2'
- + +a,c,e FATIGUE CRACK GROWTH AT (40 YEARS)
FINAL FLAW (IN)
INITIAL FLAW (IN)
~
D '~ l* '+
L . i _ .
0.292 0.31097 0.30107 0,30698 0.300 0.31949 0.30953 0,31626 0.375 O.39940 0.38948 0.40763 0.425 0.-45271 0,4435 O.47421 l
I 21
4 l
i i
4 4
Crack I 2.45-D
! _ 4 .
\
~
i i
1 M
d M
) -
q t a.c.e I 3 i
i i
u i
I 1
I i
i 4
i I
FIGURE 1 REACTOR COOLANT PIPE
' N.,
CD = A 2
3 2 ; &
- .L' Reactor Coolant Pump Steam Gererator l l o A G @
+a,c.e FIGURE 2 SCHEMATIC DIAGRAM 0F PRIMARY LOOP SHOWING WELD LOCATIONS 23
._ , . - - _ _ _ _ ~ _ _ _ _ _ . . . . _ - . . . _ _ - . . . _ _ ~ . . . , _ . _ _ _ _ . . _ _ _ _ . . _ _ _ , _ _ _ _ - _ . - _ . _
^
l
)
l
+a,C,0 ;
- -? \
////////
l 2a Neutral Axis Nj
+ a,C,e FIGURE 3 [ ] STRESS DISTP.IBUTION 24
l **
- _ PREDiciED[ ]
+a,c.e M
, h I
+
Flaw GE0*f727
- U i
l 1
i ,
l i
~
CRACK LENGTH , INCHES i' FIGURE 4 i-CRITICAL FLAW SIZE PREDICTION a
l
_-,m..
- i . -
l 1
4 i
l I
1 b
ra
.{
O #
i l
i i
i l
1
(
i i
1 l
! J i .
+.
e, j ..
1 .o I "
.!4 i
i j FIGURE 5 TYPICAL CROSS-SECTION OF [ ]+5*C*e -
Ta,c,e
- ~
E d
M.
E O
c -
+
2 2
k, 4
m I
E C
C C
lu:
W .
m u
STRESS INTENSITY FACTOR RANGE (agt (KSIG.)
FIGURE 6 REFERENCE FATIGUE CRACK GROWTH CURVES FOR 3+a,c e 27
- .- . - - _ . . - .~. _ . - , ,
.+,S,C,e e
j W $
,(
e I
eO t
i 4
e o
d 4
0 28 e E e
- q r ve---,-,,,_ I' FF T New W m q._9 "N-'*'W1my p q m.yg ,_ _
Y"NP F w'W W"- ng n 3, _ , _ - _ "" MM " eve p p~n.c,_w_,__W"+'7'w,--y.m,,
2 a w - e_ EF A. A a -
'h
+ a,C,e e
4 i
f I
- 6 Y
4 e
l
.i i
+
l i
1 k
M 29
-- -, . - + . . , v.,n, ,,ne,e-,-._.w_.,,_...--.. .w.,n.,,,w., ,,an.,_,--, nw,,,,- ,_ .wnn.,,,--m.-._mn~,,,,,.,. , - ,