ML20195B615
| ML20195B615 | |
| Person / Time | |
|---|---|
| Issue date: | 10/01/1998 |
| From: | NRC (Affiliation Not Assigned) |
| To: | |
| Shared Package | |
| ML20155K227 | List: |
| References | |
| NUDOCS 9811160205 | |
| Download: ML20195B615 (5) | |
Text
_ __
SAFETY EVALUATION BY THE OFFICE OF NUCLEAR REACTOR REGULATION WCAP-14748/14749 " JUSTIFICATION OF INCREASING POSTULATED BREAK QPENING TIMES IN WESTINGHOUSE PRESSURIZED WATER REACTORS" INTRODUCTION In accordance with Westinghouse topical report WCAP-8082-P-A, " Pipe B eaks for LOCA Analysis of Westinghouse Primary Coolant Loop," January 1975, postulated piping break opening times of one millisecond have been used in the analysis of LOCA-induced loads in the reactor coolant system. The Westinghouse Owners Group (WOG) transmitted topical report WCAP-14748/14749, " Justification of increasing Postulated Break Opening Time" to the U.S.
Nuclear Regulatory Commission (NRC) for staff review in January 1997. The report was developed in conjunction with WOG's program for inspection and possible replacement of degraded Baffle Barrel Bolts (BBB)in Westinghouse reactors. The purpose of this report was to provide information to justify the use of larger break-opening times (BOTs) in Westinghouse calculations of LOCA-induced component loads and to request NRC's approval of this change.
The WOG has determined that individual licensees should be able to replace degraded BBB following the provisions of 10 CFR 50.59. However if the WCAP-14748/14749 methodology is acceptable, licensees will have the option of replacing selected degraded BBB in a prescribed pattern, thus reducing personal exposure and economic impact. This approach would be acceptable due to the lower loading on the reactor vessel internals under conditions of a postulated loss of coolant accident (LOCA).
The WOG believes that the topical report WCAP-14748/14749 provides a more realistic BOT design basis which includes a break-size dependent methodology for calculating BOT. The suppo-ting information, provided in References 1 and 2, which justifies this design basis, consists of experimental data on pipe BOTs, records or data relating to established practice by Combustion Engineering, Babcock & Wilcox and foreign nuclear suppliers, and comparisons of Westinghouse and NRC large break LOCA load calculations. Analytical estimates have been made to demonstrate that a BOT greater than one millisecond provides a conservative representation of BOTs for piping at Westinghouse pressurized water reactors (PWRs).
DISCUSS:3N The BOT for circumferential breaks consists of two components. The first component is the time it takes for a crack to propagate around the circumference of the pipe. The second component consists of the time it takes for the severed ends of the pipe to move away from each other. Westinghouse has performed a number of analyses to determine calculated break opening areas for 2,3 and 4 loop plants at various postulated break locations on the primary loop. The break opening areas as a function of time are determined on the basis of the primary loop snalyses, taking into account the stiffness characteristics of the piping system, typical restraints and vessel motiot.s generated from reactor pressure vessel (RPV) system analyses.
9811160205 981001
{DR TOPRPENVWg ENCLOSURE
l The circumferential rupture flow area increases with the displacement of the broken ends. The displacement of the ends of the broken pipe is computed on the basis of a detailed dynamic analysis, initially, the various time-history hydraulic loads acting on the system are computed by assuming a break opening area and break opening time. When these loads are applied, an area-time history can be determined for use in new force calculations, replacing the original assumed flow area and assumed break opening time. This iterative process is repeated until the resultant area converges to a stable value. The peak values of calculated break areas and the corresponding break opening times for a variety of Westinghouse plants are provided in Table 5-1 of the subject topical report. Based on its review of the results of these analyses, the report shows that it takes approximately 20 milliseconds (msecs) to attain the maximum break size for typical breaks such as the inlet and outlet nozzle breaks on the RPV, reactor coolant pump and the steam generators. Since the inclusion of the BOT component due to crack propagation around the circumference will always yield a total BOT greater than 20 msecs, the l
staff finds the proposed BOT conservative and acceptable.
Cold-springing loads are imposed on piping during installation to connect slightly misaligned pipe ends. Since the cold-springing loads can affect the rate of displacement of the severed ends during postulated pipe breaks, these loads can potentially affect the BOT, in response to the staff's concerns related to the effect of cold-springing of the pipe segments on BOT during postulated pipe breaks, WOG has provided information to support the conclusions that cold-springing loads are negligible. Cold-springing loads fall into one of two categones. In the first category are cold-springing loads due to installation misfits and misalignments. Installation procedures limit misfits and misalignment loads. A typical procedure for Class 1 piping will, for instance, provide relationships between the pipe size, the observed misfit displacement, and the minimum piping length perpendicular to the misfit which is required to accommodate the misfit. If the required piping length is not available, modifications to the piping must be made.
Using a procedure of this type, maximum cold-springing loads for a 10-inch accumulator line were estimated by Westinghouse it was found that the maximum cold-springing load allowed would be less than 0.1% of the dominant break-opening load, which is the pressure load acting on the piping following the break. Therefore, WOG concluded that the loads due to misfit correction can be considered negligible in the calculation of break-opening dynamics. The staff concurs with this assessment. The second category includes cold-springing loads which are engineered into the system. This is done to compensate for thermal growth, which may otherwise induce undesirable loads into the piping system. While such loads are readily quantifiable, they remain insignificant in the calculation of break-opening dynamics for PWR applications. For instance, the instantaneous pressure load acting on a 10-inch line, following a breek, is on the order of 60 tons. For a rupture in the main coolant piping, the instantaneous load is on the order of 600 tons. WOG asserts that cold-springing loads of this magnitude are never imposed on piping systems in Westinghouse PWRs. The staff finds the estimates reasonable and acceptable.
For the reasons discussed abova, it can be concluded that cold-springing loads are negligible in the calculation of bre ak-opening dynamics in the primary and secondary piping systems of Westinghouse PWfw. Therefore, the 20-msec BOT estimate for circumferential breaks in the primary piping remains conservative and acceptable.
, Lonaitudinal Breaks The BOT for longitudinal flaws is dependent primarily on crack propagation speed, unlike circumferential breaks in which the rate of separation of the severed ends is the dominant factor. WOG has proposed a correlation model relating break opening area with time based on test data obtained by Battelle Memorial Institute in the 1960s for determining BOTs for longitudinal breaks (Reference 4). Experiments on pressurized carbon and stainless steel pipes with artificial, longitudinal flaws were conducted. These tests measured crack propagation speeds along and perpendicular to the crack and permitted the development of a correlation for a break opening area as a function of time.
The data provided in the subject topical report was obtained with test specimens which had artificially-induced defects. Since the fracture speeds in pipes with artificially-induced defects are likely to be different from fracture speeds in pipes with actual defects, the staff requested information to demonstrate that the proposed fracture speeds represented bounding values. In addition, the staff was concerned that the ductility of pipes in service may be lower than the test specimens. The decreased ductility of pipe material due to the thermal aging and irradiation effects could lead to higher fracture speeds and decreased BOTs. While satisfactory responses have been provided by the WOG (Reference 2), to address these concerns, the staff has determined that the available applicsble data is insufficient to conclusively validate the proposed crack propagation speeds. The staff, therefore, did not fully investigate the bases of the values cited and the acceptance of the proposed BOT for longitudinal breaks is based on the argument that longitudinal breaks are typically less limiting than circumferential breaks.
Specifically, WOG has determined and the staff concurs, that longitudinal breaks are non-limiting in tha analyses to obtain acceptable BBB distributions.
The only longitudinal break that has been identified for Westinghouse plants is at the steam generator elbow intrade' and is snown as item 7 in Figure 5-1 of the subject topical report. The postulated locations and type of breaks in the primary piping system were determined as a result of detailed three dimensional finite element stress and fatigue analyses during most severe normal and upset system transieMs. The type of break (i.e., circumferential or longitudinal) was determined by a comparison of the maximum primary plus secondary stress intensity in the axial and hoop directione. The opening area for this break is described as the
" cross-sectional flow area of the loop pipe," as are all the primary coolant piping breaks. Since a longitudinal break flow area equal to a single pipe flow aree vill not discharge as much flow as does a double-ended guillotine break, the steam generator intrados longitudinal break will be less severe than the limiting double-ended guillotine break at the hot leg outlet. Thus, for circumferential and longitudinal breaks of approximately equal flow area, dynamic analyses justify a BOT of approximately 20 msecs or more.
Since the staff did not fully investigate the bases for crack propagation speeds cited in the subject topical report for reasons stated earlier, this SE does not constitute an endorsement of any particular crack propagation speed. If crack propagation speeds are to be used in other analyses, consideration will have to be given to the nature of the application and the appropriate level of required conservatism.
._.m.
. l Auxiliarv Pioing Systems Experimental and analytical data for determining BOT in auxiliary piping systems is very limited.
l
\\ NOG is proposing that the methodology based on an analysis performed by M. Schramm, Kraftwerk Union (RSK), Germany (Reference 3) be used for calculation of circumferential breaks in auxiliary piping systems. Schramm's methodology neglects the contribution of crack propagation time around the circumference. The break opening is determined by the movement of two pipe ends from each other similar to the pipe dynamics analyses by Westinghouse described earlier for calculation of circumferential breaks in primary systems.
The results of Schramm's analysis indicated a quadratic dependence of break opening area with time. However, Westinghouse has modified this relationship slightly as shown in Equation (8) of the subject topical report and is proposing to use this relationship for calculation of circumferential breaks in auxiliary piping systems. While Equation (8) yields somewhat higher BOT values than do the German RSK guidelines over some ranges of piping diameters, WOG asserts that Equation (8) provides a better correlation with available smaller diameter pipe data and is therefore considered more suitable for auxiliary piping system application. The inclusion of the crack propagation time, or a fraction thereof, in the overall BOT will always yield higher values than does Equation (8). The staff, therefore, concludes that the calculation of BOTs for circumferential breaks based on the Schramm's methodology is conservative and acceptable for the awliary piping systems.
CONCLUSION WOG is requesting NRC approval of this WCAP for generic application to Westinghouse PWRs with the immediate need to support the BBB program.
Based on its review as discussed above, the staff concludes that an increased break-opening time of 20 msecs for the calculation of Westinghouse primary coolant piping LOCA loads is acceptable for application to the BBB program. In addition, the staff finds the methodology for estimating break opening times of circumferential breaks in auxiliary piping acceptable.
i However, this acceptance does not constitute an endorsement of any particular crack propagation speed. If crack propagation speeds are to be used in other analyses, consideration will have to be given to the nature of the application and the appropriate level of required conservatism. The staff finds the use of the increased BOTs in calculations of LOCA-induced component loads acceptable for application to the BBB program. Application to other issues involving different phenomena outside the BBB program will require staff review and approval on a case-by-case basis.
Principal Contributor: J. Rajan, NRR/EMEB l
e e
-.-,--v--,
,, -- -~
5-REFERENCES 1.
" Justification for increasing Postulated Break Opening Times in Westinghouse i
Pressurized Water Reactors," WCAP-14748/ Proprietary Class 2C, and WCAP-14749 (Class 3).
2.
Letter, ' Louis F. Liberatori, Jr., Westinghouse Owners Group, to NRC dated March 2, 1998, Westinghouse response to request for additional information by the NRC related to WCAP-14748/9.
3.
M. Schramm, " Analysis of Break Opening times for Circumferential Break in Piping Systems," Transaction of the Sixth International Conference on Structural Mechanics in Reactor Technology, August 1981.
4.
G. Wilkowski, et al., " Analysis of Experiments on Stainless Steel Flux Welds," Battelle Memorial Report BMI-2151, RF,RS and NUREG/CR-4878, April 1987.
l l
t i'
e w
,, 1 t
,w-w pw
,