IR 05000387/1992005
| ML17157B896 | |
| Person / Time | |
|---|---|
| Site: | Susquehanna  |
| Issue date: | 06/29/1992 |
| From: | Gray E, Mcbrearty R, James Medoff, Parczewski K NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION I) |
| To: | |
| Shared Package | |
| ML17157B895 | List: |
| References | |
| 50-387-92-05, 50-387-92-5, 50-388-92-05, 50-388-92-5, GL-89-08, GL-89-8, NUDOCS 9207100125 | |
| Download: ML17157B896 (13) | |
Text
U.S. NUCLEAR REGULATORY COMMISSION
REGION I
REPORT NOS.
50-387 2-05-388/ 2-05 DOCKET NOS.
50-387 50-388 LICENSE NOS.
h LICENSEE:
NPF-14 NPF-22 Penn lvania Power & Li ht om an 2 North Ninth Street Allent wn Penns lvania 18101 FACILITYNAME:
INSPECTION AT:
Sus uehanna Steam Electric Station nits 1&2 Allen wn Penns lvania & Berwick Penns lvani NSEC D
- eh INSPECTORS:
R. A. McBrearty, Reactor E sneer, Materials Section, Engineering Branch, DRS Date J. Medoff, Reactor Engineer, R
Date APPROVED BY:
K. Parczewski, R
P~f E. H. Gray, Chief, Materials Section, Engineering Branch, DRS
+/2 Y/9'z Date Date
<<d.:
d i
p i
d M fh
~
erosion/corrosion (E/C) inspection program to ascertain that activities relative to the long-term E/C monitoring program are being accomplished in accordance with NRC requirements (Generic Letter 89-08) and licensee commitments and procedures.
9207i00125 920702 PDR ADOCK 05000387
~Re nits:
The program and its implementation are generally good.
The engineering and inservice inspection staff responsible for E/C activities are extremely well. qualified and knowledgeable of the program activities pertinent to their individual responsibilities.
Two weaknesses in the otherwise excellent program were identified regarding the lack of quality
'ssurance and the lack of confirmation of the accuracy of the hand calculations to document the percent of erosion that has occurred in specific inspection zones.
An unresolved item was identified in regard to repairing class 1 piping in a manner not consistent with the ASME Cod.0 BACKGROUND Concern regarding erosion and corrosion in balance of plant piping systems has been-heightened as a result of the December 9, 1986 feedwater line rupture that occurred at Surry Unit 2.
This event was the subject of NRC Information Notice 86-17,Bulletin 87-01 and Generic Letter (GL) 89-08.
The GL specifies that all licensees provide assurances that a program, consisting of systematic measures to ensure that E/C does not lead to degradation of single and two phase high energy carbon steel systems has been implemented.
inspection locations were generally established in accordance with the 1985 guidelines in the Electric Power Research Institute (EPRI) Document NP-3944,'"Erosion/corrosion in Nuclear Plant Steam Piping: Causes and Inspection Program Guidelines."
Shortly after the startup of Susquehanna Unit 1 and prompted by a pipe rupture at the Oconee Unit 2 facility on June 28, 1982, Pennsylvania Power and Light Company (PP&L) reviewed Susquehanna for the potential of having two phase moisture erosion problems similar to those being reported by the nuclear industry.
A program was established to detect pipe wall thinning in both Susquehanna units.
Subsequent to the Surrey event, single-phase inspections were added to the program.
2.0 f
INSPECTION OBJECTIVE The inspection was conducted to determine whether licensee activities relative to the long-term erosion/corrosion monitoring program are being accomplished in accordance with NRC requirements and licensee commitments and procedures.
3.0 LICENSEE RESPONSE TO GENERIC LETTER (GL) 89-08 AND BULLETIN 87-01 Generic Letter 89-08 was issued on May 2, 1989 and requested licensees to provide assurances that a erosion/corrosion has been completed.
The program is to consist of systematic measures to ensure that erosion/corrosion does not lead to degradation of single phase and two phase high energy carbon steel systems.
Licensees were required to respond within 60 days of their receipt of the Generic Letter.
The Bulletin 87-01 was issued on July 9, 1987 to provide information to licensees regarding erosion/corrosion problems at nuclear plants and requested licensee responses to five specific actions defined*by the Bulletin.
Licensees were required to respond within 60 days of their receipt of the Bulletin.
The response to Bulletin 87-01 was provided by a letter dated September 11, 1987 and provided information addressing the five specific actions requested by Bulletin 87-01.
The date ol'he response letter was within the allotted 60 days required by the Bulleti.0 E/C PROGRAM The manager of Nuclear Engineering is responsible for implementation of the E/C program at PP&L.
He normally delineates the individual activities of the program between the Nuclear Technology section of the Nuclear Engineering department, the System Engineering department, and the Inservice Inspection department.
A description of the responsibilities covering the program can be found in AD-QA-547. Nuclear Technology is responsible for performing the predictive analyses of the systems included in the E/C program.
The predictive analyses are only one of the five inputs which are used to designate component locations for inspection (usually ultrasonic testing) during the upcoming refueling and inspection outage (RIO). The other inputs for defining inspection locations are by frequency (i.e., scheduling by refueling outage), industry experience, wear analyses and the results of cavitation predictive models.
It should be noted that PP&L does not distinguish between pipe thinning caused by E/C or by cavitation.
Nuclear Technology is also responsible for assisting the Systems Engineering department in performing wear analyses when inspection results do not meet PP&L's criteria for pipe wall integrity.
These acceptance criteria are specified in PP&L Specification M-1414. It should be noted that PP&L's acceptance criteria are similar in scope to those recommended by the Electric Power Research Institute (EPRI) for pipe wall thinning.
Systems Engineering is responsible for evaluation of inspection results which do not meet the acceptance criteria of M-1414.
Systems Engineering normally coordinates with Nuclear Technology in evaluating the results for wear.
The lnservice Inspection (ISI) department is responsible for incorporating the inspection results from General Electric into the E/C database file. The ISI department also is responsible for notifying the Systems Engineering department of component locations which fail to meet the criteria of M-1414.
This is accomplished by means of an Engineering Work Request.
4.
Systems Selection The following systems are included in the PP&L E/C program, as'listed in PP&L Speification M-1414: condenser demineralizer system, condenser air removal system, condensate system, feedwater system, extraction steam system, feedwater heater drains, feedwater heater vents, feedwater pump turbine drains, reactor core isolation cooling system (RCIC), high pressure coo)ant injection system (HPCI), bypass steam system, main steam system, moisture separator drains, turbine steam seals and drains, and cross over piping.
Portions of the reactor water cleanup system (RWCU) are also included in the program since most of the system's piping is made up of carbon stee The portions of the main steam, feedwater, and.reactor water cleanup systems located inside containment are among the systems included in the E/C program.
These portions serve as part of the reactor coolant pressure boundary.
It is important to inspect these portions since they are needed to help maintain the pressure boundary integrity, and typically are located in non-isolable portions of the system.
These portions are classified as Class I piping by ASME Code Section III and are safety related.
The inspectors verified that the systems listed previously were included in the E/C program.
The inspectors selected the main steam, feedwater, and RWCU systems inside the containment as the safety related systems to be reviewed.
The inspectors selected the extraction steam system as the balance of plant system for review since it is a system which has been found by industry experience as being susceptible to E/C.
4.2 Methods of Predicting Wear in Carbon Steel Piping Nuclear Technology uses four different analytical methods to predict which system and components would show wear beyond that allowed by M-1414:
I.
Kraft Werke Union (KWU) method 2.
EPRI NP-3944, 1985 method 3.
NUREG CR-5007 (developed by M.l.T.) method 4.
CHEC/CHECMATE method The first three methods have been incorporated into a PP&L computer program which was developed by Nuclear Technology,.
The predictive methods in this program are continuously being reviewed and upgraded.
The KWU, EPRI, and NUREG methods are all normally run for predicting wear rates; all components located in the line are considered in the analyses.
These components include exit and entrance nozzles, orifices, valves, 45 and 90 degree elbows, straight pipe sections, and reducers/expanders.
The inspectors verified that all of the applicable systems specified in M-1414 were analyzed by the PP&L computer program for predictive wear.
A total of 16 systems, 107 analyses, and 700+ components have been run with the program to date.
The results of the three predictive methods are compared with one another, and the highest predicted result is conservatively selected as the predicted wear for each component in the line.
CHEC/CHECMATE are also used for predicting wear rates and times to reach minimum thickness (t-min), but they are not the preferred methods of analysis.
The inspectors noted that the three primary methods of predictive analyses do not all use the same process variables for calculating predicted wear.
The NUREG method uses the system's temperature, velocity, and pH as the parameters for analysis.
This method normally is good for ranking the systems.
The EPRI NP-3944 method predicts wear using the Keller equation, which is a function of temperature, moisture content, geometry, and velocity.
CHEC/CHECMATE are computer codes developed by EPRI which predict wear as a function of temperature, moisture content, velocity, oxygen content, geometry, pH/chemistry, and material compositio The inspectors inquired whether results of the three primary analyses correlated well with one another, and also with CHEC/CHECMATE results.
The licensee showed the results of the three primary predictive analyses run for the RWCU system inside containment.
The primary predictive methods demonstrated reasonable correlation with one another for the line analyzed.
No excessive wear was predicted by the analysis for the RWCU system inside
, containment.
The licensee showed the results of a KWU analysis of the feedwater system inside the containment (from the inboard isolation valve to the feedwater nozzles), in comparison with the results of CHEC run on the same line. 'Review of the analysis indicated that the KWU predictive method correlated reasonably well with CHEC in predicting wear within the feedwater system inside containment.
Both analytical methods (July 1987 run) predicted wear at the reducing tees of the."A" feedwater loop inside containment (20-inch to 12-inch branch with branch leading to the feedwater nozzles): KWU -.62 MM/yr, CHEC -.66 MM/yrfor the tee at location X-S94; KWU -.63 MM/yr, CHEC -.66 MM/yrfor the tee at location X-l40.
Predictive wear analyses of the feedwater systems inside the containment prompted the licensee to perform ultrasonic testing (UT) inspections of both the "A" and "B" feedwater systems inside containment during refueling outages.
Predicted wear of analyses run on the n>ain steam system were less than ten percent of allowable wear.
4.3 Data Analysis of Inspection (UT) Results The Nuclear Technolo section in con'unction gy, J
with Systems Engineering, is also responsible for evaluating the inspection results of any locations which exhibit wear in excess of S0% of the allowable wear.
Nuclear Technology analyzes the inspection results of components exhibiting less than 50% of allowable wear after the RIO is finished.
Nuclear Technology uses nine different analytical methods to evaluate wear:
"1D" wear method: basically computes the difference between maximum reading in the grid and the minimum reading in the grid.
2.
"2D" band method: basically computes the difference between the maximum reading of a grid band perpendicular to the direction of flow and the minimum reading in the same band.
Results are presented as a plot of wear against band location.
"2D" range method: basically computes the difference between the minimum reading in the band from last outage and the minimum reading taken in the same band this outage.
Results are presented as a plot of range against band location.
"3D" wear method (point to point): basically computes the difference between the reading of a grid point taken last outage and the re'ading on the same grid point taken this outag,
Statistical summary: gives a summary of statistical wear, including maximum and minimum thicknesses of the grid, "1D" wear, "2D" wear and "3D" wear results, number of points below nominal wall thickness, percentage of points blow nominal wall thickness, and standard deviations of the data.
A contour map of the wall generated from the grid data.
6.
A "3D" or flat display of the wall.
7.
A cross sectional wear map of the pipe.
8.
A histogram of the number of points within a designated wear range.
9.
Raw data review (the grid data).
The nine different methods of evaluating wear results are incorporated into the same coniputer programs as are the three primary predictive methods of analysis.
This approach gives the Nuclear Technology section an easy method of comparing the predictive analyses with the actual wear results such that recommendations may be fed back to the Systems Engineering department.
4.4 Quality Assurance and Self Verification One area which was deficient, in an otherwise systematic and well organized erosion/corrosion program, was a lack of self verification of the predictive and wear analytical methods.
In addition, the computer program lacked sufficient documentation to instruct future users in how to perform the analyses.
To date, no other staff engineers are fully trained to perform the predictive and analytical E/C computer program analysis, although another engineer is currently being trained.
The Quality Assurance department at Susquehanna has not currently performed a QA audit of the E/C program.
However, an audit has been scheduled of the ISI program for the near future which will include a review of the implementation of the E/C program.
4.5 Conclusions of Corporate's Involvement in the E/C Program PP8.L has devised a comprehensive program for monitoring E/C in high energy piping systems.
The program includes the monitoring of Class I main steam, feedwater, and RWCU piping inside the containment.
The Nuclear Technology section of the Nuclear Engineering department has devised a well organized and systematic method of predicting wear in carbon steel piping systems (both in safety related and balance of plant piping), and of analyzing inspection results.
The licensee's KWU method and EPRI's CHEC method were both in
'
agreement in predicting the. wear of the feedwater system inside containment.
One area which was deficient in the program was the lack of quality assurance, self verification, and procedure control over the analytical computer program.
The licensee is currently moving to implement improvements in this area.
~ 5.0 INSPECTION TECHNIQUE Inspection for the detection of moisture erosion/corrosion at Susquehanna is controlled by Specification No. M-1414, Revision 5, "Specification for the Inspection of ASME Class 1,2,3 and B31.1 Piping for Moisture Erosion/Corrosion."
Pipe mall thickness measurements are made using ultrasonic examination techniques, and visual inspection, direct and remote, is sometimes used to help identify where ultrasonic measuring techniques should be used.
The licensee selects the piping for inspection and provides the locations to its ISI vendor, General Electric, who performs the inspections.
A grid pattern is used with spacing identitied by, M-1414 based on nominal pipe size (NPS).
The spacing ranges from 0.5" for piping 2" and less NPS to 3" for piping over 20" to 48" NPS.
Datum points to locate the grid pattern are permanently marked on the pipe using low stress stamps as spelled out by Specification M-1414.
The, thickness readings are sequentially recorded using a UDL-IIUltrasonic Data Logger which stores each reading until the data are loaded into a personal computer.
The data logger identifies the grid location that should be 'measured, allowing the technician to confirm that the ultrasonic transducer is at the correct location.
Additionally, the logger permits inspection personnel to review its data base prior to.leaving the inspection site, thereby assuring that all required inspection points have been measured.
The use of the data logger helps to assure that data are accurately recorded and, by speeding up the data acquisition process, helps reduce personnel exposure to radiation.
After loading data into the computer, the information is recorded on a computer disc and on a hard copy report which, after GE review, is submitted for review to the licensee's ISI group.
5.1 Data Review and Pr ocessing The data review process begins at the inspection site when the General Electric technician confirms the data in the data logger.
When GE prints the hard copy report the data are again reviewed and the determination is made regarding the need to issue a customer notification form (CNF).
When the results show that greater than 50% of allowable erosion has occurred, GE must submit a CNF to the licensee within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of making that determination.
All of the data packages, including CNFs when appropriate, are submitted to the ISI group who reviews the data.
When the ISI review confirms that greater than 50% of allowable erosion has occurred, an Engineering Work Request (EWR) is issued to assure that the questionable item is evaluated and dispositioned by engineering.
When an EWR is issued, it is included in the data package maintained by the ISI grou In addition to maintaining all data packages, the ISI group keeps a log of all EWRs that it has issued.
A status report of erosion/corrosion examinations is published daily by the ISI group, together with a summary of work requests and a list of erosion/corrosion locations.
With the exception of the EWR log, the reports are computer generated from various data bases maintained for tracking and trending the items included in those data bases.
During the NRC review of E/C data packages, the hard copy report for component x-175-2A/2B, a 20" x 12" te'e in the feedwater system, showed data suggesting that the E/C was 166% of allowable for inspection zone 2A, and 176% of allowable for zone 2B, for which CNFs and EWRs should have been issued and were not.
The licensee's explanation was that the questioned data was determined to represent a counterbore area and, therefore, did not require CNFs or EWRs.
Evidence of that determination was not documented, but the responsible ISI and engineering staff members independently recalled the incident, and agreed that no CNF or EWR was required.
To provide assurance that no cases existed for which the documents were not issued, but should have been, the licensee was requested to search the data packages.
On May 14, 1992, the inspector was advised by telephone that the licensee's investigation was completed, that no other cases were found where the data suggested the need for a CNF and EWR and none were issued, and that the x-175-2A/2B package was an isolated case.
The licensee's data search included the records of all of the E/C inspections that were performed during the Spring 1992 refueling outage.
5.2 Observations At tlie time of this inspection, all of the scheduled E/C inspections were completed.
Therefore, a production inspection could not be observed.
In response to the inspector's request to observe the technique used for performing an E/C inspection, the licensee located a section of extraction steam system piping that exhibited extensive wall thinning and was removed from the system.
The inspection technique was demonstrated on that test sample.
The sample characteristics are as follows:
Material Type Nominal Pipe Size Nominal Wall Thickness Minimum Wall Thickness carbon steel
] 6lt 0.375" 0.25" The test sample was prepared for the inspection by two GE technicians who.placed a grid pattern with 2" spacing on the test surface.
The ultrasonic equipment was calibrated using calibration block No. CAL-STEP-017 containing various thicknesses in 0.1" increments.
In this instance the 0.2" and 0.4" thicknesses were used to span the nominal wall thickness of the test sample.
Readings below the 0.2" calibrated thickness resulted in recalibration at 0.1" and 0.2".
When the inspection was completed and the data in the logger were confirmed, the technicians returned to their office and loaded the information into a personal computer.
The
maximum erosion was calculated and loaded into the computer data base; a hard copy report was printed.
The hand calculation of the maximum erosion appears to be a weakness in the process because it is not clear that the calculations are reviewed and checked for accuracy to, assure that piping is not being accepted based on inaccurate erosion calculations.
That concern was discussed with the licensee during the inspection and at the exit meeting.
5.3 Personal Qualification/Certification Records Qualification/certification records of the General Electric personnel who were responsible for performing the erosion/corrosion inspections at Susquehanna were examined to ascertain that the individuals were properly qualified to perform their assigned responsibilities.
The certifications were in accordance with SNT-TC-1A, the governing document, and the records confirmed that each individual was certified to the appropriate level of competence commensurate with his assigned responsibilities.
5.4 Repair of X-175 For area 3B of component X-175 (a portion of the ASME Class I feedwater system inside containment), the licensee estimated the expected wear rate to be between.06"/cycle and
.17"/cycle.
Erosion induced wear in this range could result in piping wall thickness less than the ASME Code Class I requirement prior to the end of the next operation cycle., The licensee proceeded to repair this area with an overlay weld without removing the affected area inside the pipe. While this repair was developed as a part of the engineered resolution to the projected condition in this piping by the end of the next operating cycle, this is considered to be a non-code repair by the NRC staff which required prior NRC approval in accordance with Generic Letter 90-05.
At the time of the inspection, the X-175 portion of the pipe in question, with the exception of the overlay weld, was in conformance with the ASME Code.
Tlie overlay is not expected to have any detrimental effect on functioning of the pipe and the licensee erosion/corrosion analysis was noted to be conservative.
Should the actual erosion exceed the projected level, the presence of the weld overlay would be beneficial in maintaining pipe structural integrity. This is an unresolved item pending review of the analysis submitted by the licensee, dated June 17, 1992, to justify safe operation until the portion of pipe projected to be under minimum wall is replaced ( 50-387/92-01).
5.5 Conclusion Inspections to detect erosion/corrosion at Susquehanna are effective in that pipe wall thinning is detected early enough to perniit repair or replacement prior to failure.
The inspections are performed by properly qualified examiners who are thoroughly familiar with the inspection techniqu The data acquisition process is generally strong,. but a weakness was identified regarding the lack of confirmation of the accuracy of hand calculations to document the percent of erosion that occurred in specific inspection zones.
The use of a data logger enhances the capability to accurately record inspection data and helps reduce personnel exposure to radiation by speeding up the data acquisition process.
An overlay weld repair was performed on a portion of ASME Class I feedwater system piping.
In the opinion of the NRC, this weld did not meet ASME Code Sections IIIor XI.
However, the weld is not expected to have any detrimental effects on the functioning of the-pipe.
An erosion/corrosion analysis performed by the licensee.was conservative in determining the possibility of pipe leakage before the next refuel outage.
6.0 EXIT MEETING The inspectors met with licensee representatives (denoted in Attachment 1) at the conclusion ot'he inspection on May 7, 1992, and summarized the scope and findings of the inspectio ATTACHMENT1 PERSONS CONTACTED Penns lvani Power 8c Li ht m an R. Baker, Coordination Engineer T. Dalpiaz, Manager Plant Services N. Feddler, Inservice Inspection Specialist J. Felock, Systems Engineer Group Leader E. Figard, Manager, Nuclear Maintenance J. Graham, Supervisor, Quality Control M. Hober, Senior Engineer, Nuclear Technology G. Kuczynski, Manager, Nuclear Systems Engineering G. Machalick, Systems Engineer E. Paneila, Information Specialist R. Saccone, Supervisor, Balance of Plant Systems M. Simpson, Manager, Nuclear Technology G. Stanley, Superintendent of Plant T. Steingass, Supervisor Maintenance Testing H. Webb, Supervisor Maintenance Technology R. Wehry, Compliance Engineer J. Wesner, Licensing Engineer C. Whirl, Supervisor, Quality Verification Nuclear Re ulator ommission G. S. Barber, Senior Resident Inspector J. Haughton, AEOD C. Hsu, AEOD D. Mannai, Resident Inspector