ML073460046
| ML073460046 | |
| Person / Time | |
|---|---|
| Site: | McGuire  |
| Issue date: | 12/03/2007 |
| From: | Gordon Peterson Duke Energy Carolinas, Duke Power Co |
| To: | Document Control Desk, Office of Nuclear Reactor Regulation |
| References | |
| Download: ML073460046 (14) | |
Text
Duke GARY R.
PETERSON Power Vice President McGuire Nuclear Station A Duke Energy Company Duke Power MGO1 VP / 12700 Hagers Ferry Rd.
Huntersville, NC 28078-9340 704 875 5333 704 875 4809 fax grpeters@duke-energy. com December 3, 2007 U. S. Nuclear Regulatory Commission Document Control Desk Washington, DC 20555-0001
Subject:
Duke Power Company LLC d/b/a Duke Energy Carolinas, LLC (Duke)
McGuire Nuclear Station, Unit 1 Docket No. 50-369 Relief Request 07-MN-001 Response to Request for Additional Information Relief Request from Immediate ASME Code Flaw Repair of Charging Pump Discharge Line Valve 1 NV-240 By letter dated July 24, 2007, Duke submitted Relief Request (RR) 07-MN-001 to request relief from the requirement for immediate ASME Code flaw repair of Charging Pump Discharge Line Valve 1 NV-240. On October 30, 2007, Duke received a Request for Additional Information (RAI) related to this RR via electronic mail. Attached is Duke's response to this RAI.
Please direct questions pertaining to this submittal to P. T. Vu of Regulatory Compliance at (704) 875-4302.
Sincerely, G. R. Peterson Attachment www. dukepower. comr
U. S. Nuclear Regulatory Commission December 3, 2007 Page 2 xc w/attachment:
W. D. Travers, Regional Administrator U. S. Nuclear Regulatory Commission, Region II Atlanta Federal Center 61 Forsyth St., SW, Suite 23T85 Atlanta, GA 30303 J. F. Stang, Jr., Senior Project Manager U. S. Nuclear Regulatory Commission 11555 Rockville Pike Mail Stop 0-8 G9A Rockville, MD 20852-2738 J. B.-Brady, Senior Resident Inspector U. S. Nuclear Regulatory Commission McGuire Nuclear Station
ATTACHMENT RELIEF REQUEST NO. 07-MN-001 RESPONSE TO RAI
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 1 of 11 By letter dated July 24, 2007, Duke Power Company (the licensee) requested relief from the requirements of the American Society of Mechanical Engineer (ASME) Boiler and Pressure Vessel code to repair a leaking valve on the charging pump discharge line at the McGuire Nuclear Station Unit 1. The licensee submitted Relief Request (RR) 07-MN-001 to delay the repair of the leaking valve to the next refueling outage in September 2008. To complete its review, the NRC staff requests the following additional information.
- 1. On page 2 of RR 07-MN-001, the licensee stated that nondestructive examination (NDE) personnel shall conduct a best effort ultrasonic volumetric examination of the flaw location every 90 days until the degraded valve (1 NV-240) is repaired. The degraded valve is fabricated with cast austenitic stainless steel. The ultrasonic technique has not yet been qualified to examine cast austenitic stainless steel.
Therefore, the flaw and associated growth in the valve would not be detected with any certainty or accuracy. (a) Discuss whether additional volumetric examinations are needed to examine the valve. If not, justify how the UT by itself can provide reasonable assurance that the flaw growth in the valve is monitored accurately. (b)
Cite the specific sections of the ASME Code Section Xl to which the UT will be performed.
General Response 1:
Relief Request 07-MN-001 describes the combined activities of leakage monitoring, fracture mechanics, and nondestructive examination providing assurance that valve 1 NV-240 can perform its design function until replaced. A fracture mechanics evaluation determined the acceptable flaw size that could exist in the cast stainless material without compromising the structural integrity of the valve body. To assure that material in the area of the leak is free of planar flaws that exceed the acceptance limit determined by the analysis, McGuire is using ultrasonic examination at 90-day intervals. Since there is no ultrasonic examination for cast stainless steel qualified in accordance with ASME Section Xl, the McGuire exam relies on best effort methods and EPRI and industry experience to characterize the ability of the exam technique to detect flaws smaller than the acceptance limit.
Additionally, McGuire will conduct twice-daily visual observations of the leaking valve to identify changes prior to their compromising the safe operation of the plant.
McGuire will replace valve 1 NV-240 either at the next scheduled refueling outage or when predetermined monitoring thresholds are exceeded. This combination of activities provides reasonable assurance that valve 1 NV-240 can perform its design function until replaced.
Response 1 (a):
An ultrasonic examination performed every 90 days provides reasonable assurance of detecting a planar flaw in the outer 2/3 wall thickness of the cast stainless steel valve before the flaw exceeds the fracture mechanics acceptance criteria. The 90-day inspection interval was selected based on information provided in ASME Code Case N-513. In addition, the flaw growth calculation in Attachment 7 of the relief
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 2 of 11 request showed that a flaw would not grow beyond the acceptance limit in less than 90 days. The evaluation showed that with an assumed large flaw and a large number of cycles, crack growth is essentially negligible over an 18 month fuel cycle.
Additional defense in depth is provided by monitoring criteria described in the response to Question 3.
Radiography does not provide an alternative for this volumetric examination. The sensitivity for radiography of this material thickness would be expected to be about 2-2T (detectable flaw image limited to approximately 2% of the valve wall thickness).
However, given the geometry of the valve (double wall radiography with the valve internals in place), radiography is not adequate to detect a planar flaw prior to it propagating to the acceptance criteria.
A Krautkramer USN-60 was calibrated for 450, 600 and 700 beam angles. As no cast stainless calibration block was available, a SA-240 stainless steel plate with a 2 mm diameter side drilled hole at a depth of 1.0 inch was used to calibrate for the 450, 600, and 70° beam angles. The Creeping Wave probe was calibrated using a 1/16 inch diameter side drilled hole at a depth of 0.2 inch. Reference sensitivity was established using the applicable hole signal set at 80% full screen height. Scanning was performed at +12dB over reference sensitivity for the 45' probe and +6dB over reference sensitivity for 600 and 70° probes. The scanning gain for the creeping wave probe was set to reference sensitivity due to excessive front surface noise at higher gain levels.
References 1, 2, and 3 cited in Attachment 6 to the original submittal, and those cited in response to Question 4, conclude that there is reasonable assurance that the ultrasonic examination techniques used are capable of detecting planar flaws once they grow into the outer 2/3 wall of cast stainless material. During ultrasonic examination of valve 1 NV-240, the 450 beam enabled monitoring of the ID surface while all ultrasonic displays were relatively noise free. The 4 MHz straight beam search unit showed similar low noise levels. These characteristics allow the examiner to conclude that a relatively small uniform grain size is present. Since detection of a flaw in cast stainless steel material is dependent on variations in grain structure, a uniform grain size enhances the ability to detect flaws. With the 450 scan it was possible to monitor the inside surface noise level at 10% full screen height assuring that some sound energy was reaching the inner surface. There was little internal noise, indicating grain sizes approximately one wavelength or larger were not present. These characteristics are indications of an effective "best effort" exam, and follow the criteria described in the references listed in Attachment 6.
Response 1(b):
The personnel conducting the examinations were qualified under ASME Section Xl, Appendix VIII, Supplement 10, 1998 Edition through the 2000 Addenda for dissimilar metal weld examinations. This assures that highly qualified personnel were used for the examination.
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 3 of 11 The ultrasonic techniques used were consistent with the manual ultrasonic techniques used and approved in Catawba Relief Request 04-CN-001, docket Nos:
50-413 and 50-414. The techniques referenced in the relief request used large 1 MHz search units applicable to 34 inch and 36 inch diameter pipe with a thickness of approximately 2.5 inches. The focal distances chosen for the large diameter pipe were based on an area of interest extending from 1.6 inches in the pipe to within 0.1 inch of the OD. The same techniques applied to 1 NV-240 used smaller 2 MHz transducers to fit the smaller diameter of the valve and employed focal distances of 1.0 inch to within 0.1 inch of.valve OD.
- 2. On page 2 of RR 07-MN-001, the licensee stated that the valve leakage rate will be recorded once every shift. (a) Describe in detail how the valve leakage rate is recorded (such as by instrumentation or visual examination, the measurement accuracy, the lowest leak rate that can be detected and/or recorded). (b) Discuss whether insulation will be removed from the valve (if it is insulated) to record the leak rate. If insulation is not removed, discuss whether the recorded leak rate is accurate.
Response 2(a):
1 NV-240 leakage rate is checked once per shift per the Semi-Daily Surveillance procedure (PT/1/A/4600/003A). The leakage rate is quantified by visual observation and recorded in drops per minute. Leakage has remained <1 drop per minute for several months. 1 NV-240 instantaneous leakage rate can be remotely determined by the control room Reactor Operators from a live video feed. Other routine programmatic monitoring activities would allow detection of any significant change in 1 NV-240 leakage rate, as follows:
i). Auxiliary Building rounds in the general area, ii). Fluid Leak Management program periodic inspections of 1 NV-240, iii). Reactor Operators routinely monitor Volume Control Tank level trends, iv). Reactor Coolant system leakage calculations Response 2(b):
1 NV-240 and associated piping are not insulated and are readily accessible for visual examination under the Fluid Leak Management program and by Operations personnel during normal rounds.
- 3. On page 2 of RR 07-MN-001, the licensee stated that it will perform a code repair of the degraded valve during the next refueling outage in September 2008. However, the licensee did not consider a scenario where it needs to repair the degraded valve immediately due to unforeseen circumstances (e.g., severe degradation). Discuss the acceptance criteria beyond which the degraded valve will be repaired immediately and provide the technical basis for the acceptance criteria.
Response 3:
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 4 of 11 There are two separate monitoring criteria associated with 1 NV-240, which if exceeded would result in unit shutdown and valve replacement. The 1 NV-240 monitoring criteria consist of an allowed operational leak-rate, and an allowed flaw size.
1 NV-240 is part of the Chemical Volume and Control system, which is part of the Emergency Core Cooling System (ECCS) piping boundary. The dose analysis limits external leakage from the ECCS and Containment spray systems to 0.25 gpm.
Technical Specification 5.5.3 further requires that leakage from these systems be minimized. A significantly lower operational leakage threshold for 1 NV-240 is specified in the semi-daily surveillance procedure (PT/1/A/4600/003A). If 1NV-240 leakage approaches a steady stream (<<0.25 gpm), procedural actions are in-place to attempt to isolate 1 NV-240 (PT/1/A/4600/003A, Enclosure 13.3, and OP/1/A/6200/01B, Enclosure 4.12)- If a catastrophic valve structural failure were to occur, abnormal station procedures provide mitigative steps to effect 1 NV-240 isolation and achieve safe unit shutdown (AP/1/A/5500/010, NC System Leakage Within Capacity of Both NV Pumps).
Periodic (90 day frequency) UTs are also performed to ensure the flaw size does not exceed the fracture mechanics analyses acceptance criteria. The Operability Evaluation states that there is reasonable assurance that the ultrasonic examination techniques used are capable of detecting planar flaws once they grow beyond the inner 1/3 material wall thickness and have a measured length of 1.5 inches or greater. The flaw analysis in the Operability Evaluation states that for an axial crack (most limiting), up to an ID wall thinning of 33% (corresponding to being able to detect flaws in only the outer 2/3 of the wall), the allowable length of a 100% through wall flaw is 1.64 inches. Forced valve replacement will occur if the periodic UT (90 days) characterizes a flaw size greater than 1.64 inches axial or greater than 6 inches in circumferential.
Additionally, 1 NV-240 has been added to the Unit 1 Forced Outage List.
Modification ME1 01307 under Work Order 1757969 has been approved to replace 1 NV-240 during either a Unit 1 forced outage or the next refueling outage (1 EOC19).
- 4) In the second paragraph on page 6 of Enclosure 1 (Operability Evaluation), the licensee stated that UT is capable of detecting planar flaws once they grow beyond the inner 1/3 material wall thickness and have a measured length of 1.5 inches or greater. Provide the technical basis of this statement.
Response 4:
Duke Energy NDE Level III personnel conducted a thorough research of industry experience before performing the examinations.
The references cited below provide the technical basis for examining the outer 2/3t of the material and the flaw length measurement criterion:
U.,S. Nuclear regulatory Commission Attachment December 3, 2007 Page 5 of 11
- 1.
EPRI Report TR-1 07481, "Status of the Ultrasonic Examination of Reactor Coolant Loop Cast Stainless Steel Materials", March 1998
- 2.
Draft White Paper: Current Inspection Capabilities for Cast Austenitic Stainless Steel Piping, Chockie Group international, Inc., November 2005.
- 3.
Safety Evaluation for Relief Request 04-CN-001, TAC Nos. MC 2209 and MC 2210. (This SER also cites NUREG/CR-6594, "Evaluation of Ultrasonic Inspection Techniques for Coarse-Grained Materials", October 1998.)
Detection capability improves as the flaw increases in through-wall extent because the sound beam travels through less of the coarse grained material. This has been validated by studies at the EPRI NDE Center and Pacific Northwest National Laboratory. Length sizing of the flaw is dependent on variations in grain structure within the cast material and the surface condition on the outside diameter.
Ultrasonic characteristics observed during UT inspections suggested that this valve body grain structure is relatively fine and uniform. However, the surface is in the as-cast condition making constant pressure on the search unit difficult. Accurate length sizing of flaws using manual ultrasonic angle beam techniques requires a surface finish much finer than the as-cast condition. Experiments conducted on deep cracks in centrifugally cast stainless steel reactor coolant loop piping specimens show that the best performing search units were only capable of measuring 40% of crack lengths at depths of 62%, 72% and 82% through-wall. Because of these factors, Duke Energy will length-size any crack-like flaws conservatively from peak amplitude down to the baseline, and in no case less than 1.5 inches which is twice the search unit's width.
- 5) In the first line on page 6 of Enclosure 1, the licensee stated that no planar flaw indications were detected in the area of the pin holes. (a) If planar flaws were not detected, discuss any indications or flaws that were detected in the area of the pin holes and how were those indications characterized. (b) Discuss whether the pin holes were shown in the UT scan and how were they characterized.
Response 5(a):
No indications were detected in the area of the pin holes.
Response 5(b):
As expected, the pin holes were not visible in the ultrasonic scan.
- 6) Table 1 of Enclosure 1 provides the allowable through wall flaw lengths. Table 2 of provides the critical through wall flaw lengths. Tables 1 and 2 present results which were based on the assumption that NDE (i.e., UT) is capable of detecting various percentage thickness of the valve wall thickness. (a) Because the
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 6 of 11 UT is not qualified to ASME Code Section XI, Appendix VIII requirements to examine cast stainless steel, clarify the usefulness of the results in Tables 1 and 2 in evaluating the operability of the degraded valve. (b) Table 1 shows that allowable through wall flaw lengths for circumferential crack is greater than 6 inches. Table 2 shows that the critical through wall flaw lengths for circumferential crack is greater than 9 inches. Clarify whether the depth of the flaws in Tables 1 and 2 are 100%
through wall. If the depth is 100% through wall, the results imply that the length of the allowable and critical circumferential crack can be the entire circumference of the valve. Explain the results.
Response 6:
(For completeness of understanding, a narrative response is first given, then specific summary responses to parts (a) and (b) are given at the end.)
The lines in Tables 1 and 2 of Enclosure 1 correspond to various cases of UT capability. For example, Case 2, "Through-wall assuming NDE through 66% of wall thickness" corresponds to UT being able to detect flaws only in the outer 2/3 of the wall. This case is the one that corresponds to the actual capability. Attachment 6 of the Relief Request asserts that the UT techniques employed are capable of detecting flaws within the outer 2/3 of the thickness and 1.5 inches long. Thus, the, largest size undetectable circumferential-radial oriented ID connected flaw would be 1/3 wall thickness deep and 360 degrees around. The largest size undetectable axial-radial oriented ID connected flaw would be 1/3 wall thickness and infinitely long. Similarly, the largest undetectable flaw in the outer 2/3 inspectable portion would be 1.5 inches long and through-wall, either axial or circumferential.
The circumferential-radial flaws evaluated are compound flaws, to correspond to the
.worse conditions un-observable coupled with the conditions observed. There are two coincident co-planar flaws evaluated together. The first is an ID surface connected part-through-wall flaw 360' around with a uniform radial depth dimension equal to the maximum un-inspectable depth. That is, for "through wall assuming NDE through 66% of wall thickness," the flaw depth in the radial direction is equal to 1/3 thickness. The second flaw is 100% through wall, with the same length all the way through the wall. The variable solved for is the allowable or critical flaw length for the through-wall circumferential flaw in the remaining ligament assuming coincidence with the 3600 undetectable ID connected 1/3 thickness flaw. In Table 1, the maximum ASME allowable flaw length for the through-wall flaw is determined. In Table 2, the ASME critical flaw length for the through-wall flaw is determined.
The tables show that a compound flaw consisting of a 3600 ID connected flaw 1/3 of the thickness through (NDE through 66% of wall) can be accompanied by a second flaw that is 100% through wall with an allowable size of greater than 6 inches circumferential dimension, or a critical size greater than 9 inches circumferential dimension. These dimensions correspond to the potential OD surface length dimension of the detected pinhole flaw. The current size of the flaw at the surface is only pinhole dimension, and the largest undetectable size (possibly just below the
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 7 of 11 surface) is 1.5 inches. This is well below the >6 inches allowable size or >9 inches critical size. Since the largest flaws will not grow more than a negligible amount in the postulated service conditions, the current conditions (assuming circumferential-radial flaw orientation) are stable, and will remain stable, with wide margins of error.
For the axial-radial flaw, only a single flaw was evaluated, that corresponding to the second flaw above, the 100% through wall flaw, with the same length for its entire depth. The un-inspectable ID connected volume is assumed to be missing entirely in this case, instead of being assumed to contain an infinitely long part-through-wall flaw. Again, the maximum allowable and critical OD connected through-wall flaw lengths are determined. The ASME allowable and critical flaw sizes were calculated as 1.64 inches and 7.21 inches respectively for the case where NDE is effective for 66% of the wall. Again, considering that the actual flaw is pinhole sized at the surface, and will not be greater than 1.5 inches long, and considering the small amount of crack growth, this flaw should also remain stable.
The remainder of the lines in Tables 1 and 2 are provided for information only to illustrate the trend and sensitivity for other UT capabilities.
Note that Attachment 1 of the Relief Request, Figure 1 of Attachment 4 of the Relief Request, and Attachment 5 of the Relief Request all place the flaw in the valve neck.
Page 3 of Attachment 4 of the Relief Request and Page 4 of Attachment 7 of the Relief Request both give the valve body neck ID as 4.313 inches. Page 4 of of the Relief Request gives the valve body neck thickness as 0.875 inch. Page 4 of Attachment 7 of the Relief Request assumes the valve body neck thickness as 0.8125 inch (which is consistent with the thickness readings given on page 1 of Attachment 1 of the Relief Request). Attachment 7 of the Relief Request cites MCM-1205.00-0577 as the source for the ID. MCM-1205.00-0577 is the valve Stress Report, listed as Reference (e) of Enclosure 1 of the Relief Request. Using an ID of 4.313 inches and wall thickness of 0.8125 inch, the inner circumference would be 13.55 inches and the outer circumference would be 18.6 inches, both of which are larger than either the reported >6 inches allowable or >9 inches critical circumferential flaw lengths. (Please also refer to the response to Question 8.)
Summary Response 6(a):
In the response to Question 4, it is explained that the UT techniques employed are capable of detecting planar flaws once they grow beyond the inner 1/3 material wall thickness and have a measured length of 1.5 inches or greater. It is conservative to assume that this bounds the largest flaw that can be present. The fracture mechanics analysis has demonstrated that there is acceptable margin between this upper bound size and the size that would be unstable, even with flaw growth.
Summary Response 6(b):
Tables 1 and 2 list the allowable and critical lengths, respectfully, of a flaw that is 100% through-wall, and coincident and coplanar with a flaw that is 1/3 through wall
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 8 of 11 (from the ID) and either 3600 around for the circumferential-radial case, or infinitely long for the axial-radial case. The critical circumferential length of the circumferential-radial flaw is tabulated as "> 9" inches in Table 2. Nine inches is less than both the inner and outer circumference of the valye neck. It is not known precisely how much larger than 9 inches the critical size is, but it is not necessary to know this, since it is demonstrated that the actual and future sizes will not exceed 6 inches circumferential (see responses to Questions 1 (a), 3, and 4).
- 7) On. page 3 of Attachment 4, item Number 2 under the assumption section, the licensee stated that dead weight and thermal loading are assumed negligible. The dead weight and thermal loading should be obtained from the stress analysis of the Charging supply piping system performed in accordance with the ASME Code Section III. It is not clear why the dead weight and thermal loading are "assumed" to be negligible. (a) Clarify why the loads are "assumed" negligible. (b) The licensee stated that both zero and conservative bending stress of 10 ksi is used in the analysis. Clarify why a zero bending stress is assumed and why 10 ksi bending stress is conservative.
Response 7:
(For completeness of understanding, a narrative response is first given, then specific summary responses to parts (a) and (b) are given at the end.)
The stresses due to dead weight, thermal, and seismic loadings shouldtbe obtained from the Stress Report for the valve, Reference (e) of Enclosure 1 of the Relief Request, instead of the piping stress analysis. The piping analysis will provide stresses only at the.valve-to-piping interface welds, however, if the desired stresses are not available from the Stress Report, the piping analysis stresses can be used to draw conclusions regarding the magnitudes of the valve stresses.
An examination of Reference (e) to Enclosure 1 of the Relief Request reveals no particular information regarding dead weight or thermal stresses. Thus, the piping analysis problem was examined for dead weight, seismic, and thermal stresses at the valve-to-piping interface welds.
The dead weight moment stresses are 317 psi.
The thermal moment stresses are 3857 psi.
The OBE moment stresses are 1818 psi.
These stresses will be transmitted from the piping connection points into the valve neck only to a very negligible degree and are indicative of the negligible magnitude of the analogous stresses in the vicinity of the pinhole flaw.
An examination of Reference (e) to Enclosure 1 of the Relief Request reveals that the combined operational and seismic stresses, due to very high assumed accelerations, are 5402 psi. These are the stresses that would result from local mass eccentricities in the valve body itself. In a comparison to this stress
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 9 of 11 magnitude, and the piping seismic stresses, the assumed 10 ksi bending stress is clearly conservative. In Attachment 4 of the Relief Request, both 0 and 10 ksi values were used to calculate an allowable flaw length. This was intended to give the range of the allowable flaw size as the bending load was not determined at that time. The range of allowable flaw size with this assumption is shown in Table 1 (page 5 of Attachment 4). With either assumption, the allowable flaw size is very large. Note that in Attachment 7 of the Relief Request (page 4), the bending stress was more rigorously, and still conservatively, calculated as 1.14 ksi. (See also the response to Question 9(a) for further discussion regarding the reasons for the two separate calculations and their differences.)
It should also be noted that the valve body is perpendicular to the run pipe and its end is not restrained. As such, it is free to expand which would indicate that the thermal stresses will be very negligible. Furthermore, the dead weight stress at the flaw location is that associated with the weight of the valve which is relatively small.
Summary Response 7(a):
It is demonstrated by examination of valve Stress Report and piping analysis results that the actual dead weight and thermal loads are so small that they can be taken as zero.
Summary Response 7(b):
Both zero and 10 ksi bending were used in the Attachment 4 fracture mechanics analysis to give the range of the allowable flaw size as the bending load was not determined at that time. In Attachment 7, the bending stress was more rigorously, and still conservatively, calculated as 1.14 ksi. (See also the response to Question 9(a) for further discussion regarding the reasons for the two separate calculations and their differences.)
- 8) On page 7 of Attachment 7, a through wall flaw of 6 inches was used/assumed in the flaw evaluation. Explain how and why the 6-inch flaw was assumed.
Response 8:
As stated above, the 6 inches is the constant length (all the way through the wall) of the 100% through-wall circumferential flaw of the two circumferential flaws being considered concurrently. Six inches was arbitrarily chosen as an evaluation value in order to demonstrate that the allowable circumferential length is larger than the largest undetectable length, 1.5 inches. This is reflected in Table 1, where the allowable is shown as ">6" inches. The analyst could have iterated, assuming larger and larger sizes, until the safety factor in each case became just equal to the ASME Code Section XI stipulated 2.77; but this would not have been necessary.
- 9) Comparing flaw evaluations in Attachments 4 and 7, the staff has the following questions: (a) Explain why two types of flaw evaluations are performed as shown in
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 10 of 11 Attachments 4 and 7. (b) On page 7 of Attachment 4, the flaw growth analysis assumed 500 heat-up and cooldown cycles. On page 4 of Attachment 7, the flaw growth analysis assumed 100 full pressure cycles and there are no thermal transients considered. Explain why different cycle numbers and the types of loadings (heatup and cooldown vs. pressure) were used in the Attachments 4 and 7 analyses for the same valve. (c) Explain why in Attachment 4, operating pressure and temperature are used whereas in Attachment 7, design pressure and temperature are used. Since design conditions are more conservative than operating conditions, discuss why Attachment 4 analysis did not use design conditions.
Response 9:
(For completeness of understanding, a narrative response is first given, then specific summary responses to parts (a), (b), and (c) are given at the end.)
The only cyclic loadings this flaw could see include thermal and pressure. There are no cyclic design loading definitions as this is not a Class 1 component. Thermal and pressure cycling are both associated with heatup and cooldown events for the plant, or other small perturbations in temperature or pressure. The design temperature for the system and the valve is 1892F (see page 3 of Enclosure 1 of the Relief Request).
Considering this value, and the unrestrained geometry (as discussed above), it is therefore clear that temperature cycling will cause negligible stresses. This is consistent with the piping analysis stresses discussed above. Thus, temperature stresses were not considered in the calculations. uses the terminology heatup-cooldown. Attachment 7 speaks in terms of pressure cycles. In both cases, the cyclic stresses are pressure induced only.
Axial flaws were determined to be more conservative in light of their relatively small allowable and critical flaw sizes. Hence, the flaw growth analysis was performed assuming an axial flaw with the maximum pressure stress of 10 ksi, as determined on page 5 of Attachment 7 of the Relief Request. The results of the analysis indicate that crack growth is not a concern as evidenced by the relatively small crack growth in both calculations (less than 1 mil and 0.011 inch).
As stated, cyclic pressure loadings were taken as being associated with heatup and cooldown. In Attachment 4 of the Relief Request, the pressure is assumed to be cycled from zero to the operating pressure of 2500 psig. Similarly, the 110°F operating temperature is used to compute the scaling factor C in the flaw growth analysis on page 6. This is consistent with the practice for Class 1 components where design transients, reflecting operating ranges and temperatures, are defined and thus used in flaw growth computations. In Attachment 7 of the Relief Request, the pressure is assumed to be cycled from zero to the system design pressure of 2735 psig (see page 3 of Enclosure 1 of the Relief Request). Similarly, the 189°F design temperature is used to compute the scaling factor C in the flaw growth analysis on page 9. The use of design pressure and temperature is not considered to be required, but is conservative, thus was performed.
U. S. Nuclear regulatory Commission Attachment December 3, 2007 Page 11 of 11 In Attachment 4 of the Relief Request, 500 cycles were arbitrarily considered. In of the Relief Request, 100 cycles were arbitrarily considered. Any number of heatup and cooldown cycles greater than those to be experienced before the valve is replaced is sufficient. One hundred cycles is just as acceptable as 500, and still provides a wide enough margin that no special tracking is needed. As stated in the cover letter of the Relief Request, the valve is slated for replacement in approximately September 2008.
The Attachment 4 calculation was performed quickly after the flaw was discovered considering the entire valve body to be inspectable. Hence, the total thickness of the valve was used in determining the allowable flaw size. The Attachment 7 calculation was performed later, primarily to address the compound flaw cases (to match the final conclusions regarding the UT capability). Other changes regarding cycles assumed, pressure ranges, operating vs. design temperature, bending stresses, and initial flaw sizes for crack growth were also incorporated to be more conservative or more realistic, and are considered improvements.
Summary Response 9(a):
The paragraph just above explains why there are two types of flaw evaluations in the separate Attachments 4 and 7.
Summary Response 9(b): refers to 500 heatup-cooldown cycles. Attachment 7 refers to 100 pressure cycles. Both of these actually refer only to pressure cycles. Examination of valve Stress Report and piping analysis results show that the thermal loads are so small that they can be taken as zero. Any number of cycles greater than those to be experienced before the valve is replaced is sufficient. One hundred cycles is just as acceptable as 500. Both are arbitrarily large.
Summary Response 9(c):
It is considered appropriate to use operating conditions instead of design conditions for fracture mechanics evaluations, consistent with the practices for Class 1 components where cyclic loadings are defined in design specifications. See also the response to 9(a) above, regarding reasons for and the differences between the separate Attachments 4 and 7. In Attachment 7, the use of design conditions is not considered to be required, but is conservative, thus was performed.