ML20044B593
| ML20044B593 | |
| Person / Time | |
|---|---|
| Issue date: | 12/01/1992 |
| From: | Hermann R Office of Nuclear Reactor Regulation |
| To: | Marion A NUCLEAR ENERGY INSTITUTE (FORMERLY NUCLEAR MGMT & |
| References | |
| NUDOCS 9212280279 | |
| Download: ML20044B593 (8) | |
Text
~.
t' PDR DEC 011992 Mr. Alex Marlon NUMARC 1776 I Street, N.W.
Suite 300 Washington, D.C.
20006 Re:
EROSION / CORROSION Per our telecon of November 30, 1992, I am enclosing a copy of the summary of the erosion / corrosion audits performed by the staff.
Please take a look at the summary and get back to me so that we can arrange a meeting for January 1993 to discuss the issue.
WSkned By R. A. Hermann l
Robert A. Hermann, Section Chief Chemical' Engineering and Metallurgy Section Materials and Chemical Engineering Branch Division of Engineering
Enclosure:
As Stated Distribution W.
Russell J.
Richardson B.
D. Liaw J.
Strosnider J.
Partlow Central File EMCB RF -
RHermann KParczewski JMedoff DE:EM RAHermann:adl 11/' e / 92 O
J g:\\Parczewski\\Marion G
t(
27000S 91l2.280241 hh h,
~
EROSION / CORROSION PILOT PROGRAM AUDIT / INSPECTION
SUMMARY
BACKGROUND t
Erosion / Corrosion (EC), or flow assisted corrosion, of carbon steel piping systems is a common occurrence in nuclear and fossil plants. However, the nuclear industry, as a whole, did not focus on the phenomenon until a catastrophic failure of a main feedwater elbow occurred at Virginia Power's Surry Nuclear Station in December 1986. The feedwater rupture resulted in four fatalities at the nuclear station. As a result, the NRC issued Information Notice (IN)86-106, "Feedwater Line Break," and Supplements 1, 2, and 3 to inform the industry of the event.
t In addition, the NRC issued Bulletin 87-01, " Thinning of Pipe Walls in Nuclear Power Plants," requesting licensees to take measures to address EC at their respective nuclear stations.
In 1988, the NRC conducted a series of 10 audits to assess the industry's overall response to Bulletin 87-01. NUREG 1344,
" Erosion / Corrosion-Induced Pipe Wall Thinning in U.S. Nuclear Power Plants,"
summarized the findings of the 10 audits. One of the conclusions in NUREG 1344 was that not all licensees were committed to implementing procedures and controls for long-term EC programs. The NRC then issued Generic Letter (GL) t 89-08, " Erosion / Corrosion-Induced Wall Thinning," requiring nuclear licensees to affirm that they have implemented or will implement long term EC programs designed to ensure that EC induced wall thinning would not lead to unacceptable degradation of single or two phase, high energy, carbon steel piping.
High energy, carbon steel piping and component failures are still occurring in I
spite of the attention given to EC since the feedwater line break at Surry in 1986.
The following list provides some recent examples of EC related failures or leaks which have occurred in balance of plant, carbon steel systems:
Millstone Unit 3, Dec. 1990, Moisture Separator Drain Line Millstone Unit 2, Nov.1991, Moisture Separator Reheater Drain Line Maine Yankee, July 1992, Moisture Separator Reheater Drain Line North Anna Unit 1, Aug. 1992, Class 3 Main Steam Drain Line The Pennsylvania Power and Light Company (PP&L) and the Commonwealth Edison Company recently have noted some wear in Class 1 portions of the Susquehanna Unit 1 and Dresden Unit 3 feedwater systems (BWR), respectively.
In both cases, the wear was downstream from a Class I reducing tee branching to the feedwater risers.
In the case of Susquehanna Unit 1, it was determined that the feedwater piping would not remain above the code minimum wall thickness during the current operating cycle.
In this case, the licensee decided to repair the worn area of the feedwater pipe prior to startup of the unit. The worn component in the Dresden Unit 3 feedwater system (in a reducer directly downstream of the main branch of the reducing tee) was not as severely degraded.
The Public Service Electric and Gas Company (PSE1G) and the Virginia Electric Power Company have similarly reported a significant amount of wear in Class 2 6
i
and Class 3 portions of the Salem Unit 2 feedwater system, and the North Anna Unit I main steam system, respectively.
In the case of Salem Unit 2, the worn component was located in piping just upstream of a nozzle to one of the unit's steam generators. The wear area at the North Anna Unit I was reported as a pinhole leak of a three inch, high energy, main steam line drain, inside containment (Class 3). These wear related events or determinations indicate that EC in Class 1, 2, or 3, high energy, carbon steel piping may be more prevalent in the industry than originally anticipated.
During the period January - June,1992, members of the NRR and individual Regional staffs conducted a series of audits / inspections as part of the NRC's Erosion / Corrosion (EC) Pilot Program. The pilot program was conducted to serve two objectives. The first objective was to assess the general response i
of the nuclear industry to GL 89-08.
The second objective was to test out Inspection Procedure (IP) 49001, " Inspection of Erosion / Corrosion Monitoring Programs," as a potential " Core" inspection program, or as a regional initiative.
Members of the joint NRR/ Regional teams performed an audit / inspection at a preselected site in each of the five respective NRC regions. The audits / inspections were conducted at the following plants:
Region V:
Palo Verde Nuclear Station, Units 1 and 2, l
January 13 - 17, 1992.
Region II: McGuire Nuclear Power Station, February 21 - 27, 1992.
Region I:
Susquehanna Atomic Power Station, Units 1 and 2, l
May 4 - 7, 1992.
Region IV: Cooper Nuclear Station, June 15 - 18, 1992.
Region III:
Dresden Nuclear Power Station, Units 2 and 3, June 22 - 25, 1992.
1 The results of the five pilot EC audits / inspections may be found in the individual audit reports and inspection reports referenced in Appendix A to this summary.
In addition, Region I has provided input from eight supplemental EC inspections conducted by members of the region's Engineering Branch. This summary combines the findings of the five pilot audits / inspections with the additional input provided by the Region I engineering staff.
The supplemental EC inspections conducted by the engineering staff of Region I are listed in Appendix B to this summary.
DISCUSSION OF SIGNIFICANT FINDINGS FROM AUDITS / INSPECTIONS AND RECENT EVENTS Proarammatic Findinos and Issues All of the licensees, which were audited in the EC pilot program, have implemented EC programs in accordance with the requirements of GL 89-08.
GL 89-08; however, only required licensees to provide assurances that systematic programs have been implemented to effectively monitor and repair erosion / corrosion of single phase or two phase, high energy, carbon steel systems. GL 89-08 provided leeway as to the scope and contents of erosion / corrosion programs.
2
l L
l l
l The collective NRR pilot audit and supplemental region based inspection findings indicate that the industry has spent considerable time and resources to implement long term EC programs. The majority of these programs have done a reasonable job of monitoring and assessing EC related degradation in high energy, carbon steel systems. A number of licensees have indicated, during l
discussions with members of their engineering staffs, that they are still in i
the process of refining their programs.
This indicates that the industry l
recognizes that there is room for improvement with respect to the design of EC l
programs. As a whole, a comprehensive EC program should specify the licensee's criteria for the following EC related activities:
I including / excluding carbon steel systems in the program selecting components for ultrasonic testing (UT) inspections specifying the method for marking grids on components selected for UT l
inspections, and for preparing the selected components for inspection performing UT measurements on components selected for examination l
l establishing acceptance criteria for degraded components for continued service l
repairing or replacing components which fail to meet the program's acceptance criteria, including assuring that repairs or replacements of safety-related piping is performed in accordance with the applicable requirements of the ASME Code,Section XI The findings of the five EC pilot audits and of supplemental regional inspections show that EC programs vary widely in the industry. The findings indicate that some licensees base the selection of ultrasonic testing (UT) locations on the analytical results of predictive models which rank systems and components according to their predicted EC susceptibility, while other l
licensees base the selection of UT examination locations primarily on l
engineering judgement. A few licensees hired outside contractors to aid them in the design of their EC programs.
Recent balance of plant pipe failures at i
l Millstone Units 2 and 3, in 1990 and 1991 (prior to Northeast Utilities' I
restructuring of its EC program), and at Maine Yankee in 1992, raise the question whether or not EC programs based primarily of engineering judgment are adequate to predict EC in high energy, carbon steel systems.
Region I has reported that a number of licensees have excluded high energy, t
I carbon steel components inside containment from the scope of their EC programs. This includes omission of components in safety-related portions of feedwater (BWR and PWR), steam generator blowdown (PWR), and reactor water I
cleanup (BWR) systems.
It is important for licensees to perform comprehensive engineering analyses of high energy, Class 1, 2, and 3 systems inside containment, to determine whether these systems should be included in the scope of the EC programs. Wall thinning should be closely monitored in Class 1, 2, and 3 systems inside containment, which are not justified to be excluded from the EC programs. These systems are safety-related, and are located typically in non-isolable portions of the plants.
3 l
l l
l non-isolable portions of the plants.
?
The following models have been used by some licensees to predict EC in their carbon steel systems:
i CHEC:
a computer code developed by the Electric Power Research Institute (EPRI) for predicting EC wear in single phase, high energy, carbon steel systems e
r CHECMATE:
a computer code developed by EPRI for predicting EC wear in single or two phase, high energy, carbon steel systems Massachusetts Institute of Technology (M.I.T.) method:
this method is described in NUREG CR-5007 EPRI method described in Report NP-3944, 1985:
this method predicts EC wear using Keller's equation l
From the pilot audits / inspections, the staff has found that the nuclear industry is not using EPRI's computer codes to the extent expected.
Two licensees ran CHEC or CHECMATE only as comparisons to other predictive models.
One of the licensees modeled some single phase, balance of plant systems with CHEC, but did not use results of CHEC or CHECMATE analyses as the predominant method of selecting components for ultrasonic inspection (UT) in its carbon steel systems.
In this case, the licensee selected component inspection l
locations primarily on an engineering judgment basis. Another licensee used CHECMATE for initial predictive modeling but did not incorporate field data as recommended by EPRI. If the field measurements are not incorporated into the t
predictive codes, the codes will not provide accurate predictions of wall thinning. One licensee used CHEC and CHECMATE to their full capacities as predictive models.
It should be noted that some licensees felt that the EPRI CHEC and CHECMATE computer codes were not user friendly, and preferred to use the analytical methods of other, more direct, predictive models to rank their components.
A number of licensees lacked programs to self assess their EC programs.
Most plants did not cover implementation of their EC programs, relative to safety-related components, in the scope of their quality assurance (QA) programs. No copies of QA audits, covering implementation of the EC programs relative to safety related components, were provided by licensee personnel during the progress of the NRR audits. The audit / inspection teams also found self assessments of the inputs to, and outputs from, computer generated analyses generally to be lacking.
Furthermore, some licensees provide only a minimal amount of training to pertinent engineering personnel in the implementation of their EC programs. It is important that licensees verify that computer program inputs are correct and train personnel on the details of implementing the EC program and on how to run predictive and analytical computer codes, j
i l
4 1
f f
~
l t
Imolementation Findinos and Issues i
i In general, the EC audit and inspection teams have determined that licensees
-(
have implemented their EC programs in accordance with procedures approved by j
the licensee's management review committees / organizations.
These procedures typically specify the licensee's method of performing the scope of the program, and delegate the responsibility of EC related activities among the various licensee departments and personnel.
i The audit and inspection teams have determined that licensees, in general, select components for UT inspection in accordance with the scope.of the EC program.
Licens es typically inspect components already designated for inspection in at ordance with the program's selection criteria. No errors were discovered during the NRR audits / inspections in regard to a licensee i
inspecting the wrong component.
Furthermore, licensees have performed UT examinations of components in excess of the minimum number required by the i
scope of their EC programs.
This indicates that the industry, in general, is aware of the EC problem, and is taking considerable time and effort to discover degraded piping or components prior to any failures.
Implementation of non-destructive testing activities and of welding during repairs or replacements was done with controlled procedures and certified personnel.
Licensees normally use one of the following four methods to calculate component wear rates-taking the difference between the nominal thickness of the component
- and the current minimum reading in a grid, and dividing the result by the total operating time to date l
taking the difference between the maximum reading and minimum reading in an inspection grid, and dividing the result by the last operating cycle period taking the difference between the maximum reading and minimum reading
[
in a series of bands perpendicular to flow, and dividing the result in the band showing the maximum difference' by the last operating cycle time (this is known as the BAND METHOD) t taking the difference between the reading at each inspection grid i
point taken during the previous outage and that taken at the same
[
point during the current outage, and dividing each difference by the previous operating cycle time (this method which calculates a wear rate for each grid point location is known as the POINT-TO-POINT' l
METH00)
Licensees typically calculate the minimum projected wall thickness of a component at the next refueling outage by subtracting the product of the current operating cycle time and the maximum wear rate from the current minimum wall thickness reading. Care must be taken when using the POINT-TO-i POINT METHOD or the BAND METHOD to project wall thicknesses at the next i
outage, since cases can be made when the grid point or band exhibiting the 1
5 lh
i
?
L components by dividing the difference between the current minimum measured wall thickness and the minimum allowed wall thickness by the maximum wear rate.
The number of repairs and replacements in extraction steam systems and moisture separator systems indicate that EC wear in balance of plant piping continues to be a problem. All of the licensees have performed repairs or replacements of carbon steel components which failed to meet the licensee's minimum wall thickness or alternative acceptance criteria.
Replacement materials have varied from carbon steels, to chromium-molybdenum steels or austenitic stainless steels, which are more wear resistant.
Alloying a steel with chromium, and to a lesser extent with molybdenum or copper, helps reduce the rate of EC by enabling the steel to form a wear resistant oxide layer on its surface.
Replacing worn components with austenitic stainless steels, also alleviates the EC problem, since these steels typically have a very high chromium alloying content.
Repairs or replacements with carbon steel will remove any current degraded conditions in plant piping or components, but will not alleviate the conditions leading to excessive wear.
Most of the implementation problems relative to licensee EC programs had to do with weaknesses or errors in either inputting the proper parameters into predictive models, calculating code minimum wall thickness acceptance criteria, analyzing the results of UT examinations, dispositioning of components after reviewing the results of the inspection analyses, or performing repairs or replacements of components which failed to meet the licensee's minimum wall thickness acceptance criteria.
The NRR audit teams and regional inspection teams found, for those cases where CHECMATE was used as a predictive model, that licensees sometimes made errors in entering the proper plant parameters into the computer code. The most common of these errors had to do with selecting the proper geometry code for a system's component.
Errors in entering plant parameters into CHECMATE can result in errors in the computer code's predicted wear rate.
Some inconsistencies also were noted with respect to reproducing in:pection grids during subsequent UT examinations of previously inspected components, and with finding the location of highest wear in a component. One licensee dispositioned a Class 2 feedwater component as acceptable using incorrect code i
minimum wall thickness criteria.
In this case, the licensee used the component's normal operating design pressure to calculate the code minimum wall, instead of using the component's actual design pressure.
This made the calculated minimum wall thickness thinner than it would have been using the components design pressure.
NRR audit teams and regional inspection teams have found that most plants do not perform baseline thickness measurements on new or replaced piping prior to placing the piping inservice. This can result in incomplete histories of components or piping which are included in the scope of the EC program.
A question also has arisen as to whether or not repairs or replacements of EC worn safety-related pipir.g m being done in accordance with ASME XI, Article IWA-4000/7000.
In one case, a licensee repaired a worn Class 1 feedwater component inside containment by performing a weld buildup (overlay) on the 6
exterior of the pipe. Another licensee, similarly, performed a non-code repair of a Class 2 feedwater component. Repairs or replacements of systems designed or reclassified as ASME Code Class 1, 2, or 3 systems must satisfy the requirements of 10 CFR 50.55a(g) and ASME Code,Section XI.
Reinforcement cf worn areas by means of weld overlays on the outside of high energy, safety-related, carbon steel pipe are not considered acceptable Code repairs by the NRC.
CONCLUSIONS /RECOMMENDATI0h1 The results of audits and inspections conducted by the NRR and regional engineering staffs indicate that, in general, licensees have implemented long term EC programs for the purpose of detecting excessive EC degradation in high energy, carbon steel systems. These programs typically meet the guidelines of Generic Letter 89-08.
GL 89-08 only required licensees to affirm that they have implemented or will implement long term EC programs for monitoring-EC related degradation of high energy, carbon steel piping and components.
The scope of EC programs, therefore, tends to vary from licensee to licensee.
The audit and inspection findings show that licensees have spent considerable time and resources to implement EC programs at their respective nuclear generating facilities.
The majority of these programs have done a reasonable job of monitoring and assessing EC related degradation in high energy, carbon steel systems.
However, srme programmatic and implementation weaknesses were noted during NRR audits or regional inspections of licensee EC programs.
[
?
1
)
L 7
1 i
'