IR 05000255/1993023
| ML18059A583 | |
| Person / Time | |
|---|---|
| Site: | Palisades  |
| Issue date: | 12/23/1993 |
| From: | James Gavula, Jeffrey Jacobson, James Smith NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION III) |
| To: | |
| Shared Package | |
| ML18059A582 | List: |
| References | |
| 50-255-93-23, NUDOCS 9312300086 | |
| Download: ML18059A583 (8) | |
Text
U. S. NUCLEAR REGULATORY COMMISSION
REGION III
Report No. 50-255/93023(DRS)
Docket No. 50-255 License No. DPR-20 Licensee:
Consumers Power Company 1945 West Parnall Road Jackson, MI 49201 *
Facility Name:
Palisades Nuclear Power Plant Inspection At:
Cov~rt, MI 49043
- Inspe~tion Conducted:
September 20 through N6vember 27, 1993 Approved Inspection Summary
. Inspection on September 20 -through November 27, 1993 <Report No. 50-255/93023CDRS))
/L-"1-) -/)
Date
{ )- - 'l- ~ -i_ )
Date I :I-- l-> - 'i. )
Oa Areas Inspected:
Announced reactive inspection of activities related to a leak in the lnconel 600 safe end of the pressurizer PORV lin Results:
No violations ~ere ~dentified during this inspectio Th~ licensee established a plausible root cause for the failure, effected a temporary repair of the failed member, and inspected areas with similar materials and operating conditions for similar defect No other circumferential defects were disclosed, b~t two RTD nozzles were repaired as a result of the presence of axial crackin Radi~graphy of the* area shortly befor~ the failure disclosed the presence of an indicatio However, the accuracy of the ultra~onic technique used to characterize and size the indication was compromised by the joint geometr :'.)00086 9350012203255,....
- DR ADOCK 0 PDR
- Inspection Summary
An apparent lack *of management involvement was ~oted early in the repair effort, as evidenced by the slow response in providing a.comprehensive inspection/repair pla Also; the initial consideration of the generic aspects of this PWSCC failure was considered weak.
- ..,
DETAILS Persons Contacted *
Consumers Power Company (CPCo)
D. Bemis, Systems Engineering V. Beilfuss, Assistant Outage Manager
- J. Decker, NOT Services Supervisor
- J. Kuemin, Licensing Administration
- R. Margol,Systems Engineering Section Head J. Nordby, NECO Welding Engineer
- K. Osborne, Systems Engineering Manager
- R. Vanwagner, Systems Engineering Nuclear Regulatory Commission (NRCJ L. Ban ic ~ EMCB M. Hum, EMCB J, Jacobson, Chief, M&PS, Region* III
- M. Parker, Senior Resident Inspector
- Idaho National Enqineerin~ Laboratories (INEL} Brow~; NRC Crinsultaht, NOE
- Denotes those participating in the telephone exit meeting on November 29, 199.
Background On September 16, 1993, a leak occurred in the nozzle of the pressurizer
.Power operated relief valve (PORV) lin The pressurizer is a low alloy
~teel vessel which is clad internally with stainless steel.* The cladding extends inside the PORV nozzle to the weld* joint with the Inconel 600 safe en The safe end forms a pnysical transition from th /4" wall of the clad portion of the nozzle to the 0.438" wall of the stainle~s steel PORV pipe and is welded to the stainless steel pip *The leak occurred in the. weld heat affected zone (HAZ) on the Inconel side of this join.
Sequence of Events June 1993 A-flaw in the pressurizer relief valve nozzle safe end.wa~ found by radiography and. considered t6 be an indication from original construction, rather than a *service-induced fla NR'c inspectors onsite at the time reviewed these radiographs and informed the licensee that, in their opinion, the indication should be evaluated as a crac The licensee elected to assume the more conservative position and ev~lu2ted
the indication as a trac The evaluation consisted of performing an ultrasonic in~pection (UT) to determine size of the indication and pe~forming an engineering analysis to determine the suitability of the nozzle for future service. The evaluation was based on the assumption that the indication was a service-induced crack and on the dimensions determined by UT (1 7/8" length and less than 30% through wall dimension, originating at the inside diameter).
The results of that evaluation, summarized in internal memorandum BVV93*006, dated July 20, 1993, concluded that the flaw was not likely to grow and that the nozzle was suitable for further service. Additionally, Palisades planned to perform a re-examination of the weld during the next refueling outag The inspectors found the licensee's response to be appropriate for the data available at the tim *
September 16 Shortly after the plant r~ached hot shutdown (532 degrees F, 2060 psia), *
a 0.2 gallon per minute le~k was identified on containment sump level instrumentation and a steam leak was identified in the pressurizer shed shortly thereafte September 17-Plant achieved cold shutdown permitting direct visual and nondestructive examination of the crack, including radiography (RT).
The crack was characterized as. circumfe~ential, in the safe end, near the safe end to pipe weld, and approximately three inches in lengt A section containing the crack was cut out of the line for evaluatio The section included approximately 3/4" of Inconel saf~-end and several inches of pipe. - Root Cause Investigation Metallurgical Examination -
The nozzle-to~pipe weld section containing the crack was cut *o the longitudinal axis to form two segments of apprbximately 180 degrees, each, with the defect entirely contained in one sectio The central portion of the crack was removed and mechanically fractured at the-crac The remaining pieces were sectioned to provide metallographic cross-sections of the weld for macro and micro exami~atio The licensee's metallurgical analysis of the sections is included in internal memorandum ESM-110-93, MET Project No. 259339070140, dated September 27, 199 It stated that the failure was believed to be the result of primary water stress corrosion cracking*
(PWSCC).
Review of the samples also disclosed the following information:
.:**:*.*:**
- o The weld root had been repaired internally for a portion o the inside diameter, adjacent to the crack initiation sit o The crack initiation.site was in the HAZ of the Inconel 600 side of the safe end to pipe weld and the crack propagated straight out to the outside diameter withi~ the confines of
.the HA The cracking was intergranular and was near through-wall prior to plant startu The inspectors reviewed the metallographic evaluation and found it to be acceptabl The fractured crack sample was noted to be almost free from evidence of fresh fractur~ surface, indic~ti~g that the fracture had existed for quite some tim Most of the surface was oxidized and covered with coarse grain Revi~w of Pre~i*ous Inspection Results The failure of previous nond~structive examinations (NOE) to disclose the presence of a nearly through-wall c~ack at this *joint was a concer The licensee radiographed the joint after the leak occurred and engaged a consultant to review this and previoµs radiographs of the are The consultant indicated that the~
construction radiographs were of poor quality and interpretation in the area of concern was not feasibl He also indicated that both the June and the September radiographs showed crack An NRC NOE consultant, on site with the NRR group at the time, reviewed the radiographs and came to essentially the same conclusi6 The inspector reviewed the June and September radiographs and
~oncurred that they showed crack Based on the above observat1ons, radiography is capable of identifying this form bf failure. *However, ultrasonic sizing of the defect obviously failed to properly characterize i Review of the UT examinatfon performed to size the defect was inconclusive in that the test was performed manually and there were no equipment-generated ~ecords of the result However, it was learned that the geometry of the sample was unsuited to direct.
. inspection without removal of the weld reinforcement.. The
.
. irregular weld root geometry also contributed to.the difficulty. of interpretatio *
The licensee indicated that further evaluation of NOE techniques will be conducted as a result of this problem, and that enhanced techniques will be applied in the augmented inspection program for safe ends, beginning in the next refueling outag..,
The interpretation of radiographs is not an e~act scienc However, it is the staff's opinibn that the licensee's interpretation of the June radiograph should have been more conservative because the Combustion Engineering Owners Group (CEOG} had identified pressurizer nozzles as a most likely place for PWSCC to occu *
c~
Effect of Previous Modification of PORV Line During the investigation into the pipe crack, it was observed that the fiist elbow in the PORV line had a ~all thickness much greater than specified on the installation drawing This original construction discrepancy potentially compromised the 1989 PORV piping modification ~ecause of the stiffer properties at this elbo The modification rerouted piping from the s*econd elbow in the PORV line and significantly increased the restraint configuration to accommodate revised hydrodynamic loads.. There was a concern that th~ combined increase in stiffness had caused significant thermal stresses at the pressurizer PORV n6zzl The licensee evaluated the effect of the thicker first elbow in Engineering Analysis EA-SP-03375-0 In addition, the calculation asseised the impact of several nonconservative assumptions used in the modification analysis regarding the horizontal thermal displacement of the PORV noz.zle and temperature distribution in the PORV line during normal operitio NRC inspectors.reviewed the initial revision of the analysis, walked do~n applicable portions of the PORV piping, and had no adverse comment A minor discrepancy was noted during the review of the pressurizer thermal expansion calculation EA-SP~03375-03. However, the incoriect *
subtraction of the vessel thickness in the L2 calculation only resulted in approximately a 1 percent displacement inaccuracy and was judged by the inspector not to be significant..
'the results of the calculations indicated that the maximum PORV
- piping stress was approximately 91 percent of allowable at a location 9 feet fro~ the pressurizer nozzle.. The maximum stress level at the nozzle was calculated to be less than 50 percent of allowabl Based on these results, the NRC inspectors concluded that the 1989 PORV modification was acceptable and that even with the increased loads at the nozzle, the thermal expansion stresses did not significantly contribute to the cause of the pipe crack.
Repair of the PORV Line Weld The PORV line was repaired by re-use of the original safe end, which was modi.fied to form a new weld preparatio The replacement piping ~as divided into two lengths~ The first, which was welded to*the safe end, was short enough to allow removal of the inside weld reinforcement to enhance the weld inspectio The addition of the second length of pipe closed the syste The completed weld was inspected by RT and by PT on bot~ the inside and the outside surface *
.. Inspection of Other Pressurizer Nozzles Based on the needs identified by the failure and in response to questions received from NRR, the licensee formulated the.action plan*
transmitted to NRC by letter dated October 7, 199 The plan describes the 136 pressurizer nozzles that contain Inconel 600.. This iricludes 120 heater wells, 8 level instruments, 2 temperature instruments, 3 safety valve flanges, 1 spray line, 1 surge line ahd 1 PORV line. It also includes a description of the PORV line repai~; an identification of the inspections completed or to be completed on similar nozzles and long term corrective action The licensee was involved in frequent discussions with NRR concerning the p~rformance and results of these activities.
. Visual examination was performed on the heater well The PT, RT, and UT performed during the period of September 24 through 29 were reviewed by the NRC NOE consultant. This generally covered the PORV line, three relief valves and the* spray line. These.nozzles were of primary
interest due to the environment in the pressurizer at these location The NRC consultant.confirmed that the examinations were adequate and that they disclosed no significant indication *
The staff noted an initial lack of management involvement in formulating a.plan which coordinated the many tasks necessary to resolve the problem and restore the plant to operation.* However, work was done in a 11 the necessary techn iCa l areas of the job and the various operations progressed with minimal *delay.
Instr~ment Nozzle Inspections Among the additional penetrations inspectep as a result of the PORV line leak ~~re 2 pressurizer thermowell nozzles, 8 pressurizer level nozzles, 3 cold leg RTD nbzzles, and 10 hot leg nozzle The 3 cold leg noz~le~
penetrate the horizontal pipe at an angle of 45 de~rees (hot leg nozzles enter horizontally).
The weld joints which secure and seal these nozzles are located at the inside surface of the pipe or vesse Through the remainder of the wall, there is no physical connectio These penetrations ha~e been observed to fail at other plants.* The normal failure*mode is axial cracking adjacent to the wel The inspector obser~ed the reported -0xidation in the area below the three cold leg nozzles, and confirmed that there were no boron crystals in the area. The inspector agreed that this was not a significant indication of leakage because the piping which ran over the area was
. seen to be wet with condensation and dripping into this are When the cold leg is relatively cool, any water reaching the nozzle.area would be drawn into the crevice by capillary action and accumulate' there until heating ciused it to run out and down the path made by the coriosio The 10 hot leg RTDs showed no signs of oxidation or boron crystals.. The eight level indicators were confirmed to be acceptable by radiograph Eddy current testing of the nozzle for the vapor phase pressurizer
temperature instrument disclosed four axial indications approximately 0.4 inth~s in length ext~nding from the end.of the nozzl Because of the limitations in accessibilit~. the liquid phase pressurizer temperature instrument near the bottom of the pressurizer could not be readily inspeCted by eddy.current. However,the presence of a leak was established by boron crystals at the nozzl Confirmation of the nature of the leak at the vapor phase instrument provided strong circumstantial *
evidence of a similar leak in the liquid phase instrumen The licensee elected to repair both nozzle The repair consisted of the nozzles bei~g se~l welded at the outer surface of the pressurize Each nozzle was then cut circumferentially (from the inside) at approximately the mid-wall position of the nozzle by electrical discharge machinin The cut was made to reduce the potential for'thermal stress build-up during transients, since both inner and outer ends of the nozzle were then secured by weldin The staff considered the licensee's original ~lan t6 investigate other areas for similar cracking to be weak, because it did not include instru~entation nozzle The CEOG had previously identifi~d the tendehcy of instrumentation nozzles to fail by PWSC.
Exit Meeting The NRC inspectors held a telephone exit meetirtg with l~cen~ee representatives (denoted in Section 1) at the conclusion of the inspection on November 29, 199 The inspectors summarized the purpose, scope and findings of the inspectio The inspectors also discu~sed the ljkely informational content of the inspection report with rega~ds to documents or processes reviewed by the inspectors during the inspectio The licensee did not identify any.such documents or processes as propri etar