ML082340733

Vermont Yankee July 2008 Evidentiary Hearing - Intervenor Exhibit NEC-UW_15, L. E. Hochreiter, Data Collection of Pipe Failures Occurring in Stainless Steel and Carbon Steel Piping, Nuce 597D - Project 1 (April 2005) (Corrected)

ML082340733
Person / Time
Site:	Vermont Yankee File:NorthStar Vermont Yankee icon.png
Issue date:	08/12/2008
From:	Hochreiter L Pennsylvania State Univ
To:	Sheron B Office of Nuclear Regulatory Research
SECY RAS
References
06-849-03-LR, 50-271-LR, Entergy-Intervenor-NEC-UW_15, RAS M-251
Download: ML082340733 (56)

Text

lLk-5

,4N NEC-UW_15 CORRECTED PENNSTATE n'-' - Departmcnt of Mechanical and Nuclear Engincering (814) 865-2519 College ofEngineering Fax: (814) 863-4848 The Pennsylvania State University 137 Rcber Building University Park. PA 16802-1412 DOCKETED Dr. Brian W. Sheron USNRC Associate Director for Project Licensing and Technical Analysis August 12, 2008 (11:00am)

U.S. Nuclear Regulatory Commission MS 05E7 OFFICE OF SECRETARY RULEMAKINGS AND 11555 Rockville Pike ADJUDICATIONS STAFF Rockville, MD 20852-2738

Dear Dr. Sharon:

Enclosed are the results of a project given to my Penn State Graduate Students on finding pipe failure data over a range of pipe sizes and conditions. We specifically looked for stainless steel data as well as carbon steel pipe data. Since the data is from several sources other than nuclear the pipe wall thickness may not always be comparable to reactor pipe wall thicknesses. In some of the reports the students did separate the failure and leakage data by mechanism such that we could then screen the data.

I had the students normalize the data in such a fashion that we could then compare to the break frequency spectrum curves generated by the NRC experts group. I did talk to Rob Tenoning on the best way of normalizing our data such that we would be consistent with the break frequency plots. The key findings from the students work is that the data, when plotted in the same manner as the break frequency spectrum plots from the NRC experts work, shows a much flatter behavior at the larger pipe sizes indicating a more similar probability level for failure as compared to a more significant decrease in the failure probability as given by the NRC break frequency spectrum.

I am complying all the independent sets of data in a spread sheet and will attempt a further screening. Once complete, I will send you a copy of the data. I wanted you to have these report now with all the data so you could make an independent assessment.

Please let me know if you need anything else.

Very truly yours, DoftZ No. -5 - Off ba Exhibft No.41~

OFOby: Appilicat/l itene N Other L."E. Hochreiter WWFID 04viness/Pane~l ,

Professor of Nuclear and Mechanical Engineering College of Engineering An Equal opportunity Univcrs*ity

NucE 597D - Project 1 DATA COLLECTION OF PIPE FAILURES OCCURING IN STAINLESS STEEL AND CARBON STEEL PIPING Pennsylvania State University Dr. L.E. Hochreiter April 2005 I p I

.r- ~

Executive Summary Currently the Nuclear Regulatory Commission (NRC) is contemplating changing the acceptance criteria for Emergency Core Cooling Systems (ECCS) for light-water nuclear power reactors contained in NRC Regulation 10 CFR 50.46. This regulation sets specific numerical acceptance criteria for peak cladding temperature, clad oxidation, total hydrogen generation, and core cooling under loss-of-coolant accident (LOCA) situations. Furthermore, the regulation requires that a spectrum of break sizes and locations be analyzed to determine the most severe case and to ensure the plant designcan meet the acceptance criteria under such conditions.

Currently the regulation states that breaks of pipes in the reactor coolant pressure boundary up to, and including, a break equivalent in size to the double-ended rupture of the largest pipe in the reactor coolant system must be considered. While this restricts the design, it maintains a large safety margin ensuring the plant-is covered under all LOCA situations. However, an impetus for change has resulted from materials research, analysis, and experience that indicate that the catastrophic rupture of a limiting size pipe at a nuclear power plant is a very low probability event.

If approved, the proposed change would divide the break spectrum into two categories based upon the likelihood of a break. Breaks of higher likelihood, breaks smaller than 10 inches, would need to meet the current requirements set forth in 10 CFR 50.46. Breaks of a lower likelihood, those larger than 10 inches, would only need to meet the requirements of maintaining a coolable geometry and having the capability for long term cooling.

The purpose of this project was to collect data on instances of pipe failures including cracks, leaks, and ruptures. For each instance of failure the plant type, pipe diameter, type of pipe, failure mechanism, and type of failure was recorded. The data was then collapsed based on plant type (PWR or BWR), type of pipe (carbon or stainless steel), pipe size, and failure mechanism.

Then, normalized failure frequencies were calculated as a function of both pipe size and failure mechanism per reactor year. Plots of the frequency distributions'were generated on a semi-log scale, and the frequency distributions as a function of pipe size were compared to the NRC predicted failure frequencies.

For this project our group collected two, independent sets of data. The first set was provided by the OECD Pipe Failure Data Exchange Project (OPDE), with a total of 2891 data points. The second set consists of 67 data points collected by our group from various sources. The two sets of data were not combined due to the lack of information accompanying the data presented in the OPDE database, such as plant name or exact failure size. This made it impossible to identify overlapping coverage and combine the information. Rather, within this report we have analyzed each data set individually in order to make an overall comparison of the trends observed for each data set and the NRC predictions..

The results from both the OPDE and the independent sets of data detailed in this report do not support the NRC's assertion that larger sized pipes do not break frequently enough to be used as design criteria. The overall trends of both sets of data show that the frequency of failures does not decrease as sharply with increasing pipe size as the NRC predicts.

2

Table of Contents 1.0 Detailed Introduction to the Problem ............................................................................. 6 2.0 Data Collected ............................................................................................................ 8 2.1 OECD Pipe FailureDataExchange Project....................................................... 8 2.2 Independently CollectedData............................................................................... 9 3.0 Collapsing and Analyzing the Collected Data .................................................................. 12 4.0 Results and comparisons ............................................. 15 4.1 FailureFrequencyas afunction of Pipe Size ............................... 15 4.2 FailureFrequency as afunction ofFailureMechanism ...................................... 25 5.0 Conclusions ............................................................................................................................ 31 6.0 References .............................................................................................................................. 33 Appendix A - OPDE-Light Database Appendix B - Independent Database Appendix C - Collapsed OPDE Data Appendix D - Copies of References 3

List of Figures Figure 4.1-1. Normalized pipe failure frequencies as a function of pipe group size for both carbon and stainless steel pipe failures in both BWR and PWR plants Figure 4.1-2 Normalized rupture frequencies as a function of pipe group size for both carbon and stainless steel pipe failures in both BWR and PWR plants Figure 4.1-3. Normalized Failure Frequency Distribution for PWRs Figure 4.1-4. Normalized Failure Frequency Distribution for BWRs Figure 4.1-5. Normalized pipe failure frequencies as a function of pipe size f6r PWRs Figure 4.1-6. Normalized pipe failure frequencies as a function of pipe size for BWRs Figure 4.1-7. Normalized pipe failure frequencies as a function of pipe size for PWRs using the Modified Analysis Method.

Figure 4.1-8. Normalized pipe failure frequencies as a function of pipe size for PWRs using the Modified Analysis Method.

Figure 4.2-1. Normalized pipe failurefrequency as a function of Pipe Group Size for PWRs Figure 4.2-2. Normalized pipe failure frequency as a function of Pipe Group Size for BWRs Figure 4.3-1. PWR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism Figure 4.3-2. BWR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism Figure 4.3-3. PWR and BWR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism Figure 4.3-4. Pipe Failure by Corrosion as a Function of Pipe Size (PWR & BWR)

Figure 4.3-5. Pipe Failure by Fatigue as a Function of Pipe Size (PWR & BWR)

Figure 4.3-6. Pipe Failure by Mechanical Failures as a Function of Pipe Size (PWR & BWR)

Figure 4.3-7. Pipe Failure by Stress Corrosion Cracking as a Function of Pipe Size (PWR &

BWR) 4

List of Tables Table 1-1. NRC Total Preliminary BWR and PWR Frequencies Table 2-1. Excerpt from "OPDE-Light" Database Table 2-2. Description of Plant Systems and Type of Piping Table 2-3. Definition of OPDE Pipe Size Groups Table 2-4. OPDE Pipe Failure Definitions Table 3-1. Definition of Pipe Size Groups Table 3-2. Definition of NRC LOCA Groups Table 4.1-1. OPDE Calculated, and NRC Predicted, Normalized Failure Frequencies (1/cal-yrs).

Table 4.1-2. Normalized Rupture Frequencies Table 4.1-3. Summary of PWR Pipe Failures from the OPDE Database as of 2-24-05 Tl P Table 4.1-4. Summary of BWR Pipe Failures from OPDE Database as of 2-24-05 Table 4.1-6. Summary of PWR Pipe Failures from OPDE Database as of 2-24-05, using the Modified Analysis Method.

Table 4.1-7. Summary of BWR Pipe Failures from OPDE Database as of 2-24-05, using the Modified Analysis Method.

Table 4.2-1. OPDE Calculated, NRC Predicted, and Independent Database Calculated, Normalized Failure Frequencies (1/cal-yrs)

Table 4.3-1. Failure Frequencies of Pipes for each Failure Mechanism 5

1.0 Detailed Introduction of Problem In order to ensure the safety of nuclear plants the cooling performance of the Emergency Core .

Cooling System (ECCS) must be calculated in accordance with an acceptable evaluation model, and must be calculated for a number of postulated loss-of-coolant accidents (LOCA) resulting from pipe breaks of different sizes, locations, and other properties. This is done to provide sufficient assurance that a plant can handle even the most severe postulated LOCA. LOCA's are hypothetical accidents that would result from the loss of reactor coolant, at a rate in excess of the capability of the reactor coolant makeup system. Currently, the evaluation criteria for these types of accidents state that pipe breaks in the reactor coolant pressure boundary up to and including a break equivalent in size to the double-ended rupture of the largest pipe in the reactor coolant system must be considered. In the case of such an event the NRC has set forth the following criteria that must be met for a design to be considered acceptable [37]:

a. Peak cladding temperature must not exceed 22000 F.

b. Maximum cladding oxidation must not exceed 0.17 times the total cladding thickness before oxidation.

c. Maximum hydrogen generation. The calculated total amount of hydrogen generated from the chemical reaction of the cladding with water or steam shall not exceed 0.01 times the hypothetical amount that would be generated if all of the metal in the cladding cylinders surrounding the fuel, excluding the cladding surrounding the plenum volume, were to react.

d. A coolable geometry of the core must be maintained.

e. After any calculated successful initial operation of the ECCS, the calculated core temperature shall be maintained at an acceptably low value and decay heat shall be removed for the extended period of time required by the long-lived radioactivity remaining in the core.

While requiring that all plants be analyzed in the case of a double-ended guillotine break of the largest pipe restricts the design, it does maintain a large safety margin ensuring the plant is covered in all pipe break situations. However, an impetus for change has resulted from materials research, analysis, and experience which indicate that the catastrophic rupture of a large pipe at a nuclear power plant is a very low probability event. The hypothesis that is currently being set forth is that small pipes break more frequently than large pipes. The criteria would change so that the NRC would refocus their analysis efforts because they want to make sure that the appropriateamount of time and money are being invested in the areas of most concern, Furthermore, risk analyses indicate that large break LOCA's are not significant contributors to plant risk. According to a presentation given by Dr. Brian Sheron of theNRC at Penn State in the Fall 2004, "using the double ended break of the largest pipe in the reactor coolant system as the design basis for the plant results in ECCS equipment requirements which are inconsistent with risk insights and places an unwarranted emphasis and resource expenditure on low risk 6

contributors. This also places constraints on operations which are unnecessary from a public health and safety perspective." Therefore, the proposed rule change would use the pipe size with the largest break frequency as the design basis for pipe rupture and accident analysis of the plant.

A pipe size with a 10 inch diameter is currently being suggested. [37]

The proposed change would divide the break spectrum into two categories based upon the likelihood of a break. Breaks of higher likelihood, or those smaller than 10 inches, would need to meet the current requirements set forth in 10 CFR 50.46. These include criteria (a) through (e) above. On the other hand, breaks of a lower likelihood, or those larger than 10 inches up to and including a double-ended guillotine break of the largest pipe in the reactor coolant system, would only need to meet the requirements of maintaining a coolable geometry and having the capability for long term cooling. Thus, criteria (a), (b), and (c) would be eliminated for these cases. [37]

The purpose of this project was to collect data on instances of pipe breaks, leaks, and cracking.

These failures included pipe failures from broken pipes either by splits, ruptures, or guillotines, and cracks in pipes, either circumferential or length wise. For each instance found the plant type, pipe diameter, type of pipe, failure mechanism, and type of failure was recorded. Only stainless steel and carbon steel pipes were considered. Then, normalized failure frequency distributions were developed and compared to NRC predictions.

The predicted NRC failure frequencies were taken from Table 3 on page 14 of 10 CFR 50.46, LOCA Frequency Development [38]. This table is replicated below.

Table 1-1. NRC Total Preliminary BWR and PWR Frequencies.

Plant Effective Current Day Estimates (per cal. yr)

Type Break Size 5% Median Mean 95%

1/2 3.OE-05 2.2E-04 4.7E-04 J.7E-03 1 7/8 2.2E-06 4.3E-05 1.3E-04 5.0E-04 3 1/4 2.7E-07 5.7E-06 2.4E-05 9.4E-05 7 6.6E-08 1.4E-06 6.OE-06 2.3E-05 18 1.5E-08 ].IE-07 2.2E-06 6.3E-06 41 3.5E-1 I 8.5E-10 2.3E-06 8.6E-09 1/2 7.3E-04 3.7E-03 6.3E-03 2.OE-02 1 7/8 6.9E-06 9.9E-05 2.3E-04 8.5E-04 PWR 3 1/4 1.6E-07 4.9E-06 1.6E-05 6.2E-05 7 1.1E-08 6.3E-07 2.3E-06 8.8E-06 Is 5.7E-10 7.5E-09 3.9E-09 1.5E-07 41 4.2E-11 1,4E-09 2.3E-08 7.5E-08 7

2.0 Data Collected For this project our group collected two, independent sets of data. The first set was provided by the OECD Pipe Failure Data Exchange Project (OPDE), with a total of 2891 data points. The second set consists of 67 data points collected by our group from various sources listed as references in this report. The two sets of data were not combined due to the lack of information accompanying the data presented in the OPDE database, such as plant name and exact failure size, which made identifying overlapping coverage impossible. Rather, within this report each data set was individually analyzed in order to make an overall comparison of the trends observed for each data set and the NRC predictions.

OECD Pipe FailureData Exchange Project[3]

OECD Pipe Failure Data Exchange Project (OPDE) was established in 2002 as an international forum for the exchange of pipe failure information. It is a 3-year project with participants from twelve countries, including Belgium, Canada, Czech Republic, Finland, France, Germany, Japan, Republic of Korea, Spain, Sweden, Switzerland and the United States. "The objective of OPDE is to establish a well structured, comprehensive database on pipe failure events and to make the database available to project member organizations that provide data." [3] The OPDE database evolved from what existed in the "SLAP database" at the end of 1998 [2].

OPDE covers piping in primary-side and secondary-side process systems, standby safety systems, auxiliary systems, containment systems, support systems and fire protection systems. Furthermore, ASME Code Class I through 3 and non-Code piping has been considered. At the end of 2003, the OPDE database included approximately 4,400 records on pipe failure. The database also includes an additional 450 records on water hammer events where the structural integrity of piping was challenged but did not fail.

Access to the actual OPDE database is restricted to organizations providing input data.

However, a "OPDE-Light" version of the database will be made available later this year to non-member organizations contracted by a project member to perform work or which pipe failure data is needed. This version will not include proprietary data, such as the exact pipe diameter, where failure occurred, and preclude any plant identities or dates.

Our group was fortunate enough to get a copy of this "light" version of the database for BWR and PWR pipe failures reported as of February 24, 2005. A total of 2891 failures (1536 for PWR plants and 1355 for BWR plants) were provided in this database, and considered for this project.

The database listed the plant type, reactor system, apparent cause of failure, pipe size group, number of total failures for each cause and pipe size group, and then a break down of the type of failure within the category. An excerpt from the OPDE-Light database has been provided for clarification in Table 2-1 on the following page. The database, in its entirety, has been included in Appendix A of this report.

8

However, there are a few problems with this database related to the purpose of this project. First, since the database did not provide the type of pipe (carbon or stainless) for each failure, a reasonable prediction of what type of pipe was involved in the failure based on the plant system, which was given, was made. The. type of pipe assumed for each system is also given in the following page in Table 2-2.

Additionally, as previously mentioned, no explicit pipe diameters were given for each failure due to the proprietary nature of this information. Rather, the failures were collected into group sizes before it was sent out. A total of six group sizes were utilized by OPDE. The range of pipe diameters that comprise each group is given in Table 2-3.

The main problem with these groupings, and the database in general, is that pipes larger than 10 inches in diameter are all grouped together and there is no way of determining how much larger than 10 inches they actually were. Finally, for the purpose of this analysis any crack, leak, or issue (i.e. wall thinning) with the pipe was considered to be a failure. However, the OPDE database lists the information by type of failure. The definitions of each failure type have been included in Table 2-4.

Independently Collected Data [5-36]

For the purpose of this project our group collected separate informiation on instances of piping failures and their causes. The information was collected primarily from Nuclear Regulatory Commission (NRC) bulletins, information notices, event reports, and generic letters. Our group was able to compile a total of 67 instances of piping failures. This database is provided in Appendix B. While our database is much smaller than the one compiled by the OECD Pipe Failure Exchange Project, it provides an independent check of the trends observed by that database.

A list of references is provided at the end of this report, and some of the actual references, printed from the NRC website, have been included in Appendix D.

9

Database Table 2-1. Excerpt from "OPDE-Light" PLAT PP SYSTEM APPARENT CAUSE Pan rato ANT PIPE S PIPE SIZE TOTALNO. Crack- Crack- Ddomation Large LkLeakPH.Small Rutr SeraeSml Wall TYPE TYPE GROUP GROUP OF RECORDS Full pai. LLeak eak Leak thinning BWR SS RAS Severe overloading 2 3 1 2 BWR SS RCPB external damage 3 I BWR 'SS RCPB Severe Overloading 4 1 1 BWR SS SIR Severe overloading 6 1 1 BWR CS STEAM Water Hammer 6 1 1 BWR SS RCPB IIF:Weldlng Error 3 7 1 I 1 4 13WR SS RAS TGSCC-TransgranularSCC 2 7 1 I 1 4 BWR SS SIR IGSCC - lntcrgranular SCC 4 4 1 2 1 BWR SS RAS IGSCC - Intergranular SCC 4 56 I 32 9 1 13 BWR SS SIR 0  ! 1 __1__

BWR SS RCPI3 TGSCC - Transgranular SCC 1 I I BWR SS SIR IGSCC - Intergranular SCC 2 3 1 BWR SS RCPB Overpressurization 4 2 1 BWR CS AUXC Vibration-Fatigue 5 1 1 -

Table 2-2. Description of Plant Systems and Type of Pl ing.

Plant Group Representative Plant System Names Type of Piping AUXC Service Water Systems, Raw Water Cooling Systems Carbon CS Containment Spray System Stainless EHC Electra-Hydraulic Control System Carbon EPS Emergency Diesel Generator System Stainless FPS Fire Protection System Carbon FWC Feedwater & Condensate Systems Stainless IA-SA Instrument Air & Service Air Systems Carbon PCS Power Conversion Systems (incl. Steam Extraction Carbon Lines, Heater Drain Lines, etc.)

RAS Reactor Auxiliary Systems (incl., CVCS, RWCU, Stainless CCWS, CRD)

RCPB Reactor Coolant Pressure Boundary Stainless SG Steam Generator Systems (e.g., S/G Blowdown System) Carbon SIR Safety Injection & Recirculation Systems Stainless STEAM Main Steam (from nuclear boiler/steam generator up to Carbon turbine steam admission) 10

Table 2-3. Definition of OPDE Pipe Size Grou ps.

Pipe Size Corresponding Corresponding Pipe Diameters Pipe Diameters (inches)

Group (mm)

I DN < 15 DN <0.6 2 15 < DN<25 0.6<DN < 1.0 3 25 < DN < 50 1.0 < DN<2.0 4 50 < DN < 100 2.0 < DN < 4.0 5 100<DN<250 4.0 < DN < 10.0 6 DN > 250 DN > 10.0 Table 2-4. OPDE Pipe Failure Definitions.

Type Description Crack - Part Part through-wall crack (> 10% of wall thickness)

Crack - Full Through-wall but no active leakage; leakage may be detected given a plant mode change involving cooldown and depressurization.

Wall Thinning Internal pipe wall thinning due to flow accelerated corrosion - FAC Small Leak Leak rate within Technical Specification limits Pinhole Leak Pinhole__Leak Differs from "small leak" only in terms of the geometry of the throughwall defect

_and the underlying degradation or damage mechanism Large Leak Leak rate in excess of Technical Specification limits but within the makeup capability of safety injection systems Severance Full circumferential crack - caused by external impact/force, including high-cycle mechanical fatigue - limited to small-diameter piping, typically Large flow rate and major, sudden loss of structural integrity. Invariably caused Rupture by influences of a degradation mechanism (e.g., FAC) in combination with a severe overload condition (e.g., water hammer)

3.0 Collapsing and Analyzing the Collected Data The next important step in this analysis was collapsing the collected information into a usable form by specifying pipe size groups and failure mechanisms. The data was broken into separate bins based on plant type (PWR or BWR), pipe type (carbon or stainless), failure mechanism, and pipe size. Table 3-1 below lists the pipe diameters included in each bin for this analysis.

Table 3-1. Definition of Pipe Size Groups.

OPDE Pipe ICorresponding Pipe Size Groups I Diameters (inches) I 1+2 0.0-1.0 3 1.0-2.0 4 2.0-4.0 5 4.0-10.0 6 > 10.0 Note: This grouping of piping diameters includes one less bin than used by the OPDE database.

Combination of the data from groups 1 and 2 of the OPDE database allowed the bin sizes to correspond more readily with those used by the NRC for listing predicted failure frequencies, taken from page 14 of 10 CFR 50.46, LOCA Frequency Development. The categories used for the NRC predicted failure frequencies are given in Table 3-2. [38]

Table 3-2. Definition of NRC LOCA Groups.

LOCA Effective Break Category Size (inches) 1 1/2 2 17/8 3 31/4 4 7 5 18 6 41 It can be seen that for LOCA categories I though 5 the effective break sizes fall within the ranges listed for the pipe size groups, after pipe size groups 1 and 2 from the OPDE database were combined. LOCA category 6 was not considered in this analysis since the OPDE database did not provide specific information for pipes larger than 10 inches. The effect of this on the results will be discussed later in this report.

After collapsing the data based on pipe size, the data was then collapsed further by combining some of the failure mechanisms. The following is a list of the failure mechanisms that are used to group the data. Several items have been placed into general categories for simplification purposes.

12

1. Corrosion

2. Flow Accelerated Corrosion (FAC)

3. Microbiological Induced Corrosion (MIC)

4. Erosion

5. Fatigue

a. Thermal Fatigue

b. Vibration Fatigue

6. Human Factors (already combined in the OPDE database)

a. Welding Error

b. Fabrication Error

c. Human Error

7. Mechanical Failures

a. Excessive Vibration

b. Overpressurization

c. Overstressed

d. Severe Overloading

8. Stress Corrosion Cracking

9. Water Hammer

10. Miscellaneous

a. Brittle Fracture

b. Cavitation

c. External Damage

d. Fretting

e. Freezing

f. Hot Cracking

g. Hydrogen Embrittlement

h. Unreported After collapsing the data, it needed to be normalized so that failure frequency distributions could be calculated. Failure frequencies were calculated in for carbon steel pipes, stainless steel pipes, and a composite (both carbon and stainless) pipes as a function of both pipe group size and failure mechanism, separately for PWR and BWR plants.

The number of failures in each bin was normalized by dividing by the total number of failures.

This gives the fraction of failures for each bin size. For example, when looking at carbon steel pipes in BWRs the number of failures in each pipe group size, regardless of failure mechanism, was divided by the total number of pipe failures (carbon + stainless) in BWRs. Similarly, the number of pipe failures in each failure mechanism bin, regardless of pipe size, was divided by the total number of pipe failures in BWRs.

Then, after normalizing the data, the fractional size in each bin was divided by 3390 calendar years of operation. This gives a failure frequency in l/calander-years for each bin size. The number 3390 represents the number of reactor years experience in the US (2745 years) as of the end of 2003; divided by an assumed availability factor of 0.81 to get calendar years.

13

The normalization by pipe size (regardless of failure mechanism) and failure mechanism (regardless of pipe size) was repeated for BW'R stainless steel failures, BWR composite failures, PWR carbon failures, PWR stainless steel failures, PWR composite failures, total carbon steel failures, total stainless steel'failures, and total composite failures for a total of nine situations analyzed and a total of eighteen frequency distributions developed (nine as a function of pipe size and nine as a function of failure mechanism).

Finally, the frequency distributions developed were based both on pipe size and failure mechanisms for the different types of pipes had to be plotted against the NRC's predicted frequencies. Semi-log plots of failure frequency as a function of pipe group size were used.

OPDE Database In order to use this database it had to be collapsed into a more useful form. First, after determining the type of pipe associated with each system, the plant system was no longer taken into consideration. Next, for the purpose of this project any type of failure (i.e.

crack, rupture, wall thinning) was considered to be a pipe failure. Furthermlore, as shown above several causes of failure were combined together into one failure mechanism.

category. The collapsed form of this database is provided in Appendix C.

Independent Database There were 67 incidents recorded, which in the end did not provide enough data points in each bin to come up with a good normalized frequency distribution. When the data was sorted on plant type, then pipe material and finally on pipe size, various bins of pipe sizes had zero incidents. Appendix B is a listing of all of the incidents which were found; This listing is sorted on plant type, pipe material, and finally on pipe size. The highlighted incidents throughout the Appendix represent incidents for which not enough information was given in the source to include this data in our analysis.

Failure mechanism plots were not made due to the lack of variety in failure mechanisms.

The majority of the failure mechanisms were erosion/corrosion and stress corrosion cracking.

14

4.0 Results and Comparisons 4.1 Pipe Failuresas afunction of Pipe Size from OPDEData This section of the report examines the results of pipe failures as a function of pipe size.

Normalized failure frequencies for carbon steel, stainless steel, and composite (carbon and stainless) pipes are presented individually for PWRs and BWRs. The NRC has developed their own failure frequencies for PWR and BWR plants as function of pipe size, but does not have separate frequencies for carbon and stainless steel pipes.

Table 4.1-1 lists the normalized failure frequencies for both PWR and BWR plants, regardless of pipe type, calculated from the OPDE database data and the NRC mean predictions [38].

Table 4.1-1. OPDE Calculated, and NRC Predicted, Normalized Failure Frequ encies (Ileal-) rs).

Plant Pipe Size Groups OPDE Results NRC Predictions Type (inches) 0.0-1.0 1.3E-04 6.3E-03 1.0-2.0 4AE-05 2.3E-04 PWR 2.0-4.0 2.9E-05 1.6E-05 4.0-10.0 4.6E-05 2.3E-06

> 10.0 4.2E-05 3.9E-08 0.0-1.0 8.2E-05 4.7E-04 1.0-2.0 2.3E-05 1.3E-04 BWR 2.0-4.0 5.6E-05 2.4E-05 4.0-10.0 6.2E-05 6.OE-06

> 10.0 7.2E-05 2.2E-06 Figure 4.1-1 displays this information graphically on a semi-log plot with normalized failure frequencies on the y-axis and the pipe size groups on the x-axis. The figure shows that the results of the OPDE database underestimate the failure frequency for the smaller pipe size groups and overestimate the failure frequency for the larger pipe size groups compared to the NRC predictions for both PWRs and BWRs. However, there is less disparity in the two BWR predictions than the two PWR predictions.

The NRC predicts that PWR plants are much more likely to have pipe failures in smaller pipes than larger pipes. This trend remains the same in NRC prediction for BWR plants, but is not nearly as drastic. The OPDE results for both PWR and BWR plants show a much more consistent failure frequency both over the range of pipe sizes and between PWR and BWR plants.

15

I.OOE-02 1.00E-O23

". I*OOPDE PwRResu*

1.00E03

" " 4 -NRC PWR Predicion -

• • - ** * -NRC BWR Prediction I .OOE-04

-- A Lt 1.00E-05

_ I.O0E-06 1.00E-07 1.OOE-08 0.0-1.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (inches)

Figure 4.1-1. Normalized pipe failure frequencies as a function of pipe group size for both carbon and stainless steel pipe failures in both BWR and PWR plants.

There were three issues in the data analysis that were initially thought to factor into the difference in results between the analyzed OPDE database and the NRC predictions. The first assumption was that all types of cracks, leaks, ruptures, or other issues were considered to be a complete failure in the pipe. In actuality this is not true since inspections or other indicators may catch a crack or leak before a complete failure occurs. As a result, a separate analysis considering only the pipe ruptures listed in the OPDE database was conducted. However, the calculated frequency distribution considering only ruptures did not change significantly, in either trend or magnitude, from the results obtained when considering all issues to be a failure. The results of this rupture only analysis are shown below in Figure 4.1-2.

16

1.0E-02 F- IN m,..

twm rreoucuon LL 1.0E-05 "

4, A 1.0E-07 -*,

1.0E-08 0.0-1.0 1.0-2.0 20-4.0 4.0-10.0 > 10.0 Pipe Size (inches)

Figure 4.1-2 Normalized rupture frequencies as a function of pipe group size for both carbon and stainless steel pipe failures in both BWR and PWR plants.

The data for this plot is shown in Table 4.1-2.

Table 4.1-2. Normalized Rupture Frequencies.

Normalized Plant Pipe Size instances Failure Type (inches) of Rupture Frequency (1/cal-yrs) 0.0-1.0 37 9.8E-05 1.0-2.0 14 3.7E-05 PWR 2.04.0 10 2.7E-05 4.0-10.0 29 7.7E-05

> 10.0 21 5.6E-05 Total 11l 0.0-1.0 31 8.2E-05 1.0-2.0 5 1.3E-05 2.0-4.0 6 1.6E-05 4.0-10.0 11 2.9E-05

> 10.0 7 1.9E-05 Total 60 17

i The second assumption of concern is the nature of the information contained in the OPDE database. Since the "light" version of the database did not specify the exact pipe size due to the proprietary nature of this information, all pipe failures greater than 10 inches were included in one bin for this analysis. However, for the NRC predictions there are two categories for pipes greater than 10 inches, LOCA categories 5 and 6. As a result, the OPDE calculated failure frequencies for the largest pipe group size would be expected to be larger in magnitude than the NRC's predictions since it covers a wider range of pipe sizes, and thereby a greater fraction of the total when normalized.

The final concern is the OPDE database excludes instances of steam generator tube rupture (SGTR) from consideration. By doing this the total number of failures in the smaller pipe size groups is reduced, and the calculated frequencies are lower for the smaller pipe size groups than if SGTR had been considered.

The next two plots, Figure 4.1-3 and Figure 4.1-4, present the same data as is included in Figure 4.1-1, but these figures include the ranges for the NRC prediction. It can be seen that even when the range of validity is taken into consideration, a large portion of the distribution still falls outside the boundaries for both PWRs and BWRs.

1.00E+÷O 1.OOE-01 OPDE Results -

- -NRC Mean X X NRC 05th Percentlie 1.00E 1.UE0 NRC Median

>,"1 . NRC 5th Percentle "a I.OOE-03.

1.00E-04 A - -

L6. 1.O0E .,

,,. 1.00E-06 0 . X, o10-a7 * )1 0.0-1.0 1.0-20 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (inches)

Figure 4.1-3. Normalized Failure Frequency Distribution for PWRs.

18

F NRC 5th Percentile

'a 1.00E-03 1.0OE-04 X -U..

1.00E-05 + " I-....

i. 1.00E-06 .

1.00E 0 1.00E-081 1.00E-09 1.00E-10 00.01.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (inches)

Figure 4.1-4. Normalized Failure Frequency Distribution for BWRs.

Table 4.1-3 and Table 4.1-4 serve as summaries of the information on pipe failure as a function of pipe size and pipe type from the OPDE database for PWRs and BWRs respectively. All the data contained in these tables was normalized based on the total number of failures for the given plant type (1355 for BWR and 1536 for PWR).

Table 4.1-3. Summary of PWR Pi pe Failures from OPDE Database as of 2-24-05 Both Carbon Steel and Stainless

__________Steel Pipes Carbon Steel Pipes Only Stainless Steel Pipes Only Pipe Size Normalized Failure Normalized Failure Normalized Failure (inches) Number Number Frequency Number Frequency of Failures Frequency of Failures Feun of Failures Frequeny (l/cal-yrs) (l/cal-yrs) (l/cal-yrs) 0.0-1.0 698 1.3E-04 154 3.0E-05 544 L.OE-04 1.0-2.0 228 4.4E-05 74 1.4E-05 154 3.01-05 2.0-4.0 153 2.9E-05 78 1.5E-05 75  !.4E-05 4.0-10.0 238 4.6E-05 126 2.4E-05 112 2.2E-05

> 10.0 219 4.2E-05 93 1.8E-05 126 2.4E-05 Total 1536 -- 525 -- 1011 --

19

Table 4.1-4. Summary of BWR Pipe Failures from the OPDE Database as of 2-24-05 Both Carbon Steel and Stainless Carbon Steel Pipes Only Stainless Steel Pipes Only C n e p nt e t e O Pipe Size Steel Pipes (inches)

(ine Number Numer Normalized F e u Failure ny Number of Normalized FrequenrFailure Normalized FrequencyFailure

______ of Failures

______ Frequency (l/cal-yrs) Failures Frequency (I/cal-yrs) of Failures (l/cal-yrs)

(1/cal-yrs) 0.0-1.0 375 8.2E-05 118 2.6E-05 257 5.6E-05 1.0-2.0 107 L.IE-05 32 7.0E-06 75 1.6E-05 2.0-4.0 259 2.6E-05 32 7.0E-06 227 4.9E-05 4.0-10.0 284 2.9E-05 50 1.IE-05 234 5.1E-05

> 10.0 330 3.4E-05 39 8.5E-06 291 6.3E-05 Total 1355 - 271 084 1- --

There are a few important things to note from these tables. -The first is that there have been a similar number of failures reported in BWRs as PWRs (1355 vs. 1536). Second, there were 4 times as many failures of stainless steel pipes as carbon steel pipes in BWRs (1084 vs. 271), and almost two times as many stainless steel failures than carbon steel failures in PWRs (1011 vs.

525). It was not expected to find more stainless steel failures than carbon steel failures. It should also be noted that while the number of stainless steel pipe failures is about the same for both BWRs and PWRs, but nearly twice as many carbon steel failures were observed in PWR plants than BWR plants (525 vs. 271).

Figure 4.1-5 and Figure 4.1-6 shows a more detailed representation of failure frequencies as a function of pipe size for PWR plants only, and BWR plants only, respectively. These figures present the separate failure frequency distributions for carbon steel and stainless steel pipes, where the data is normalized based on the total number of failures for each plant type. Figure 4.1-5 shows that failures of stainless steel pipes are more frequent than carbon steel pipes only for smaller pipe sizes in PWRs. Figure 4.1-6 shows that stainless steel pipe failures are much more frequent than carbon steel pipe failures at all pipe sizes in BWRs.

As previously mentioned, the data for these two figures (4.1-5 and 4.1-6) was normalized using the methodology explained in the Data Analysis Section, using the total number of failures (carbon + stainless) for each plant type. Conducting the analysis in this manner allows for relative comparisons of failure frequencies to be made between the two types of pipes, however, it does not allow for the failure frequencies to be compared to the NRC predictions. As a result, a second analysis was done where the data was normalized based on the number of failures for a given pipe type in each plant type. In other words, the BWR carbon steel failures would be normalized by the total number of carbon failures in BWRs. The results of this modified analysis are given in Figure 4.1-7 and 4.1-8 for PWRs and BWRs, respectively. The summary tables, with the recalculated frequencies, have also been included as Table 4.1-5 and Table 4.1-6.

It can be seen from these two figures that conducting the analysis in this modified manner collapses the data, meaning that the failure frequencies, based strictly on pipe size, are very similar for carbon and stainless steel pipes in both types of plants. However, the fact remains that stainless pipes are still more likely to fail than carbon pipes in both plant types, based in the relative number of failures for each. More importantly, however, conducting this modified analysis did not show any substantial improvement in matching the data to the NRC predictions.

20

0.0-1.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe SIze (inches)

Figure 4.1-5. Normalized pipe failure frequencies as a function of pipe size for PWRs.

0.0-1.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (Inches)

Figure 4.1-6. Normalized pipe failure frequencies as a function of pipe size for BWRs.

21

,J U.E-04 LL U.E-05 0

1,ED 1.OE-08 0.0-1.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (Inches)

Figure 4.1-7. Normalized pipe failure frequencies as a function of pipe size for PWRs using the Modified Analysis Method.

1.OE-02 I.--Carbon S.teel

• Stainless Steel

,, - ') '- R C S W R P r e d liction

,,L I.OE-O5 1.0E-07 i~1.02-05 0.1-1.0 1..0.2.0 220. 4.0-10.0 > 10.0 Pipe Size (inches)

Figure 4.1-8. Normalized pipe failure frequencies as a function of pipe size for BNVRs using the Modified Analysis Method.

Table 4.1-5. Summary of PWR Pipe Failures from OPDE Database as of 2-24-05, using the Modified Analysis Method.

Both Carbon Steel and Stainless Both CaboSteel Pipes Carbon Steel Pipes Only Stainless Steel Pipes Only (inches) Number Normalized Failure Number Normalized Failure Number Normalized Failure of Failures Frequency of Failures Frequency of Failures Frequency (1/cal-yrs) (1/cal-yrs) (llealyrs) 0.0-1.0 698 1.3E-04 154 8.7E-05 544 1.6E-04 1.0-2.0 228 4.4E-05 74 4.2E-05 154 4.5E-05 2.0-4.0 153 2.9E-05 78 4.4E-05 75 2.2E-05 4.0-10.0 238 4.6E-05 126 7.1E-05 112 3.3E-05

> 10.0 219 4.2E-05 93 5.2E-05 ]26 3.7E-05 Total 1536 -- 525 --- 1011 ---

Table 4.1-6. Summary of PVWR Pipe Failures from OPDE Database as of 2-24-05, using the Modified Analysis Method.

Both Carbon Steel and Stainless Carbon Steel Pipes Only Stainless Steel Pipes Only Pipe Size Steel Pipes PineSize Normalized Failure Normalized Failure Normalized Failure (inches) Number Number Feuny Number Frqey of Failures Frequency of Failures Frequeny of Failures Frequency (Fucal-yrs) (F/eal-yrs) (I/cal-yrs) 0.0-1.0 698 1.3E-04 154 3.4E-05 544 7.0E-05 1.0-2.0 228 4.4E-05 74 9.3E-06 154 2.OE-05 2.0-4.0 153 2.9E-05 78 9.3E-06 75 6.2E-05 4.0-10.0 238 4.6E-05 126 1.5E-05 112 6.4E-05

> 10.0 219 4.2E-05 93 1.IE-05 126 7.9E-05 Total 1536 -- 525 --- 1011 --

4.2 Pipe Failuresas afunction ofPipe Size from Independent Data The independent database was used primarily to confirm the OPDE database predictions, along with comparing this set of data to the NRC data. Due to the small number of incidents found in this database, some of the pipe group size data groups had values of zero. When plotted on a semi-log scale, similar to the NRC and the OPDE plots, the points do not appear on the plot for that particular pipe size group. This occurs only once for the total normalized frequency plot for BWR data.

Table 4.2-1 shows the comparison of the OPDE, NRC and the independent database frequencies.

Table 4.2-1. OPDE Calculated, NRC Predicted, and Independent Database Calculated, Normalized Failure Fre uencies (]/cal-yrs).

Plant Pipe Size OPDEData NRC Independent

-Type (inches) Prediction Database 0.0-1.0 1.3E-04 6.3E-03 3.6E-05 1.0-2.0 4.4E-05 2.3E-04 3.6E-05 PWR 2.0-4.0 2.9E-05 1.6E-05 9.4E-05 4.0-10.0 4.6E-05 2.3E-06 2.2E-05

> 10.0 4.2E-05 3.9E-08 L.IE-04 0.0-1.0 8.2E-05 4.7E-04 2.3E-05 1.0-2.0 2.3E-05 1.3E-04 O.OE+00 BWR 2.0-4.0 5.6E-05 2.4E-05 3.4E-05 4.0-10.0 6.2E-05 6.OE-06 2.3E-05

> 10.0 7.2E-05 2.2E-06 2.2E-04 The Figure 4.2-1 presents the overall normalized frequencies of PWR plants in the United States, and roughly 10 foreign plants for the independent database, the entire OPDE-light, and the NRC mean data given in reports. As seen, the NRC mean values of frequency decrease as the pipe size increases. Although in the two other independent sets of data obtained, the frequencies remain relatively the same throughout the pipe size groups. Pipe sizes which were less than roughly two inches had a lower frequency for the two independent data sets compared to the NRC data, and the pipe sizes above the two to four inches group size show a higher frequency compared to what the NRC's expert elicitation has predicted. This figure shows that the two independent data sources follow similar trends compared to what the NRC's prediction. The PWR frequency shows a vast difference at the higher pipe size groups which in turn contradicts the thinking that larger the pipe size have a smaller break frequency.

22

A

" .E-02 C e 1.E-03 I .E.04 I.E-06 1.E-07 I.E-Cl t E-08 0.0-1.0 1.0-2.0 2.04.0 4.0-10.0 > 10.0 Pipe Size (lnches)

Figure 4.2-1. Normalized pipe failure frequency as a function of Pipe Group Size for PWRs.

Figure 4.2-2 presents the overall BWR data for the independent data, the OPDE-light, and the NRC data. A similar trend for each data set can be seen in BWR's as in PWR's, except that the frequency range is much smaller for BWR's than PWR's. The independent data provided no pipe failures in the pipe size group of one to two inches, and thus on a log-scale, no data point appears on the figure. Once again the independent data and the OPDE-light data coincide throughout the pipe size groups, and contradict the NRC prediction of pipe failure frequencies; except for the range of two to four inches again they are similar. Pipes which are larger than ten inches prove to have a higher frequency in the two independent data sets when compared to that of the NRC data set provided by expert elicitation.

23

.A F.---OPDE results 1.E.03 1.E.04 I. -. *5.06 1.E.07 Z

I.E-08 1.E-09 I.E-ID 0.0-1.0 1.0-2.0 2.0-4.0 4.0-10.0 > 10.0 Pipe Size (Inches)

Figure 4.2-2. Normalized pipe failure frequency as a function of Pipe Group Size for BWRs.

Overall, the two independ1ent data sets show contradicting trends when compared to the NRC normalized frequencies. Instead of the double-ended guillotine break being analyzed for every plant for the largest pipe in that plant, the NRC is trying to make the maximum break size which needs to be analyzed ten inches. The reasoning for this is due to low frequency of breaks in pipes of larger diameter than ten inches. This data above shows that the frequency from raw data does not agree with the current NRC predictions by expert elicitation. There is a high frequency of occurrence in pipe sizes greater than ten inches according to the independent data found.

24

A 4.3 Pipe Failuresas afunction of FailureMechanism This section of the report summarizes the frequency of failure mechanisms for carbon and stainless steel pipes. The information presented in figures 4.3-1 through 4.3-3 represents the normalized failure frequencies for each failure mechanism. This data is also presented in tabular form in table 4.3-1. The data was collapsed by pipe sizes and broken apart. by steel type and plant type. The data was normalized for each type of steel based on the number of reactor years and the total amount of failures (carbon +stainless) for each plant.

Table 4.3-1. Failure Fre uencies of Pipes for each Failure Mechanism.

Plant Carbon Steel Stainless Steel Total Failure Type Failure Frequency Failure Frequency Frequency PWR Corrosion 2.04E-05 5.38E-06 2.57E-05 PWR FAC 2.29E-05 2.32E-05 4.61 E-05 PWR MIC 8.26E-06 1.92E-07 8.45E-06 PWR Erosion ).84E-05 2.30E-06 2.07E-05 PWR Fatigue 1.77E-05 9.62E-05 1.14E-04 PWR Human Factors 6.91E-06 2.42E-05 3.1 IE-05 PWR Mechanical Failures 4.23E-06 7.1 IE-06 1.13E-05 PWR SCC 9.60E-07 3.25E-05 3.34E-05 PWR Water Hammer O.OOE+00 3.84E-07 3.84E-07 PWR Misc I.15E-06 2.69E-06 3.84E-06 BWR Corrosion 6.3 IE-06 6.97E-06 1.33E-05 BWR FAC 1.26E-05 1.37E-05 2.63E-05 BWR MIC 1.3 1E-06 2.1 E-07 1.52E-06 BWR Erosion 8.71E-06 1.96E-06 1.07E-05 BWR Fatigue 1.55E-05 4.90E-05 6.44E-05 BWR Human Factors 5.22E-06 1.85E-05 2.37E-05 BWR Mechanical Failures 3.92E-06 5.44E-06 9.36E-06 BWR SCC 4.14E-06 1.36E-04 1.40E-04 BWR Water Hammer 4.35E-07 2.1SE-07 6.53E-07 BWR Misc 8.71E-07 4.14E-06 5.01E-06 25

CL

~+/-6.DE-05

~!4.OE.OS Corrosion FAC MIC Erosion Fatigue Human Mechanteal SOC Water Misc Factors Failures Hammer FailureMechanism Figure 4.3-1. PWR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism I.600E-04 1.400E-O4 0 Carbon Steel IIStailress Steel n0Carbon and Staines Steel111 1s 1.200E-04 0 .OOOE-04 E8.000E-,05 C.

6.000E-05.

4.000E-05 2.00012-05 croin FAC MIC Erosion Fatigue Human Mec~hanical SOC Water Misc Factors Factors Hammer Failure Mechanism Figure 4.3-2. BWýR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism 26

1 I nr~nrld III Stainless Steel 9.000E-05.1[ Carbon and Stainless Steel 8.000E-.05-7'.00E-.05 -

,", 6.OOOE-05 5.000E.-05 4.OOOE-05

• .3.D00E,05 Corrosion FAC MIC Erosion Fatigue Human Mechanical SCC Water Misc Factors Failures Hammer Failure Mechanism Figure 4.3-3. PWR and B'WR Failure Frequency for Carbon and Stainless Steel Pipes as a Function of Failure Mechanism From these plots it was determined that PWR plants are dominated by fatigue failures and BWR plants are dominated by stress corrosion cracking failures. However, in general the most frequent failure mechanisms for both plants are corrosion, fatigue, mechanical factors, and stress corrosion cracking* These four failure mechanisms were analyzed as a function of pipe size in figures 4.3-4 through 4.4-7.

For these plots corrosion includes general corrosion, flow accelerated corrosion, and microbiological corrosion. Stress corrosion cracking was not included with corrosion because the pipe failure method for stress corrosion cracking is different than the other corrosion types.

Though mechanical failure frequency was not the highest, mechanical failures were chosen because they appear to be independent of pipe type and plant type. Human factors were ignored because they are a factor of quality assurance as opposed to the other failure mechanisms which are primarily a factor of operation. In regards to human factors it is not known if they have decreased with reactor operating experience because the dates of failures was not included with the OPDE data.

27

1 .I.JD'I.AJ

',(0E.01 I-Carbon Steel

-- Stainless Steel 4, Carbon and Stinless Steel 0

t.OOE-02 0.

5.. 1.001-03 L~.

4, 1.OE-04 U-1.00E-06 2 3 6 6 Pipe Size Bin Figure 4.3-4. Pipe Failure by Corrosion as a Function of Pipe Size (PWR & BWR) 5-0 0.

U.

4, U.

2 3 4 5 6 Pipe Size Bin Figure 4.3-5. Pipe Failure by Fatigue as a Function of Pipe Size (PPWR & BNVR) 28

i UcE400 1.OOE-01

~a1.0OE-03 Carbon Steel

-*- Stainless Steel Carbon and Stainless Steel S1.OOE-04 1 2 3 4 6 Pipe Size Bin Figure 4.3-6. Pipe Failu re by Mechanical Failures as a Function of Pipe Size (PWR &

B3WR) 1.Q(JL+vi 1.OOE-01 1.OOE-02 1.O0E 0r 1!

1.OOE-04 I.OOE-05 1.00E-07 I 2 3 4 5 6 Pipe Size Bin Figure 4.3-7. Pipe Failure by Stress Corrosion Cracking as a Function of Pipe Size (PWVR

&BWR) 29

The frequencies of pipe failures by corrosion shown in Figure 4.3-4 are nearly independent of pipe size. With the exception of the smallest of pipe sizes (< 1.0 inches) the frequency of failure for each type of steel is relatively constant. Stainless steel has a lower frequency of failure due to corrosion than carbon steel, which is expected because stainless steel is meant to be corrosion resistant.

Figure 4.3-5 shows that carbon steel is less likely to fail by fatigue than stainless steel for all pipe sizes. The figure also shows that as the pipes increase in size they fail less frequently by fatigue.

This is more than likely due to greater movement of the pipes as they decrease in size. The amount of force required to fatigue a larger pipe is greater than that of a smaller pipe.

Figure 4.3-6 supports the information from figure 4.3-3 that shows mechanical failures being relatively equal for all pipe sizes and types. The frequencies of the different pipes in each bin are roughly the same and they stay relatively constant across the spectrum of pipe sizes. The different failures that were grouped into mechanical failures as listed in the section 3.0 are excessive vibration, overpressurization, overstressed, and severe overloading. Though the instances of these failures are low they seem to affect all pipes relatively equally.

Stress corrosion cracking appears to be much more prevalent in stainless steel pipes as opposed to carbon steel pipes as shown in Figure 4.3-7. The discontinuity in the carbon steel data is due to plotting a frequency of zero on a log scale. For both stainless and carbon pipes the frequency of failure increases for the largest pipe size (> 10 inches).

30

5.0 Conclusions from Data 5.1 Pipe Failuresas afunction of Pipe Size from OPDEData

1. The main problem with the OPDE database is it does not have any resolution beyond pipe sizes greater than 10 inches.

2. For both PWRs and BWRs the results of the OPDE database underestimate the failure frequency for the smaller pipe size groups, and overestimate the failure frequency for the larger pipe size groups, compared to the NRC predictions. In both cases the OPDE data does not predict as drastic of a difference in the frequencies for small pipes and large pipes as the NRC does.

3. The OPDE database excludes instances of steam generator tube rupture (SGTR) from consideration. By doing this the total number of failures in the smaller pipe size groups are reduced, and the calculated frequencies are lower at smaller pipe sizes than if SGTR had been considered. This may be one source of difference in the OPDE results and NRC prediction.

4. The OPDE database reports failures of stainless steel pipes are more frequent than carbon steel pipes for smaller pipe sizes in PWRs and stainless steel pipe failures are much more frequent than carbon steel pipe failures at all pipe sizes in BWRs.

5.2 Pipe Failuresas afunction ofPipe Sizefrom Independent Data

1. The data set collected independently by our group compares very well with the trends observed in the OPDE data, but does not match the results predicted by the NRC.

2. The main problem with this data set is the limited amount of data points.

3. Failure mechanism plots were not made due to the lack of variety in failure mechanisms. The majority of the failure mechanisms were erosion/corrosion and stress corrosion cracking.

5.3 Pipe Failuresas afunction of FailureMechanism

1. The failure mechanism that appears to dominate PWR plants is fatigue failure, and BWR plants are dominated by stress corrosion cracking failures. In general both plants are limited by corrosion, fatigue, and stress corrosion cracking.

2. For some failure mechanisms the frequency of failure increases as pipe size increases.

Stress corrosion cracking is one failure mechanism where this trend is seen. It should be noted that this does not necessarily contradict the NRC's assertion that larger pipes break less frequently. This conclusion only states that for some failure mechanisms large pipes fail more frequently.

31

3. Although the OPDE data does not show water hammer to be a significant failure mechanism, it should be noted that the OPDE database listed 450 separate water hammer events where structural pipe integrity was challenged but not failed. Had this data points been included as probable failures, water hammer would have become one of the leading failure mechanisms.

32

6.0 References

1) Lydell, Bengt & Mathet, Eric & Gott, Karen, PIPING SERVICE LIFE EXPERIENCE IN COMMERCIAL NUCLEAR POWER PLANTS: PROGRESS WITH THE OECD PIPE FAILURE DATA EXCHANGE PROJECT, ASME PVP-2004 Conference, La Jolla, California, USA, July 26, 2004.

2) Nyman, Ralph & Hegedus, Damir & Tomic, Bojan & Lydell, Bengt, RELIABILITY OF PIPING SYSTEM COMPONENTS - FRAMEWORK FOR ESTIMATING FAILURE PARAMETERS FROM SERVICE DATA, SKI/RA, ENCONET Consulting GesmbH, Sigma-Phase, Inc., December 1997.

3) OPDE Database Light, OECD Piping Failure Data Exchange (OPDE) Proiect, OECD/NEA (2005).

4) Choi, Sun Yeong and Choi, Young Hwan, PIPING FAILURE ANALYSIS FOR THE KOREAN NUCLEAR PIPING INCLUDING THE EFFECT OF IN-SERVICE INSPECTION, KAERI and KiNS, 2004.

5) DeYoung, Richard C., NRC - Bulletin No. 82-02: DEGRADATION OF THREADED FASTENERS IN THE REACTOR COOLANT PRESSURE BOUNDARY OF PWR PLANTS June 2, 1982.

6) Information Notice No. 82-09: CRACKING IN PIPING OF MAKEUP COOLANT LINES AT B&W PLANTS, March 31,1982

7) Jordan, Edward L., Information Notice No. 82-22: FAILURES IN TURBINE EXHAUST LINES, July 9, 1982

8) DeYoung, Richard C., NRC Bulletin N. 83-02: STRESS CORROSION CRACKING IN LARGE-DIAMETER STAINLESS STEEL RECIRCULATION SYSTEM PIPING AT BWR PLANTS, March 4, 1983

9) Jordan, Edward L., Information Notice No. 84-41: IGSCC IN BWR PLANTS, June 1, 1984.

10) Jordan, Edward L., Information Notice No. 85-34: HEAT TRACING CONTRIBUTES TO CORROSION FAILURE OF STAINLESS STEEL PIPING, April 30, 1985.

1 1)Partlow, James G., Generic Letter 89-08: EROSIONICORROSION-INDUCED PIPE WALL THINNING, May 2, 1989.

12) Marsh, Ledyard B., Information Notice 99-19: RUPTURE OF THE SHELL SIDE OF A FEEDWATER HEATER AT THE POINT BEACH NUCLEAR PLANT, June 23, 1999.

33

13) Roe, Jack W., Information Notice 97-84: RUPTURE IN EXTRACTION STEAM PIPING AS A RESULT OF FLOW-ACCELERATED CORROSION, December 11,1997.

14) Jordan, Edward L., Information Notice 86-106: FEEDWATER LINE BREAK, February 13, 1987.

15) Rossi, Charles E., Information Notice 89-53: RUPTURE OF EXTRACTION STEAM LINE ON HIGH PRESSURE TURBINE, June 13, 1989.

16) Rossi, Charles E., Information Notice 91-18: HIGH-ENERGY PIPING FAILURES CAUSED BY WALL THINNING, March 12, 1991.

17) Grimes, Brian K., Information Notice 95-11: FAILURE OF CONDENSATE PIPING

.BECAUSE OF EROSION/CORROSION AT A FLOW-STRAIGHTENING DEVICE, February 24, 1995.

18) Weaver, Brian, Event Notification Report 36016: MANUAL REACTOR TRIP DUE TO HEATER DRAIN LINE BREAK, August 12, 1999.

19) Rossi, Charles E., Information Notice 87-36: SIGNIFICANT UNEXPECTED EROSION OF FEEDWATER LINES. August 4, 1987.

20) Rossi, Charles E., Information Notice 89-07: FAILURES OF SMALL-DIAMETER TUBING IN CONTROL AIR, FUEL OIL, AND LUBE OIL SYSTEMS WHICH RENDER EMERGENCY DIESEL GENERATORS INOPERABLE, January 25, 1989.

21) Rossi, Charles E., Information Notice 88-08: THERMAL STESSES IN PIPING CONNECTED TO REACTOR COOLANT SYSTEMS, April 11,1989.

22) Rossi, Charles E., Information Notice 88-01: SAFETY. INJECTION PIPE FAILURE, January 27, 1988.

23) Martin, Thomas T., Information Notice 97-19: SAFETY INJECTION SYSTEM WELD FLAW AT SEQUOYAH NUCLEAR POWER PLANT, UNIT 2, April 18, 1997.

24) Slosson, Marylee M., Information Notice 97-46: UNISOLABLE CRACK IN HIGH-PRESSURE INJECTION PIPING, July 9, 1997.

25)Rossi, Charles E., Information Notice 91-05: INTERGRANULAR STRESS CORROSION CRACKING IN PRESSURIZED WATER REACTOR SAFETY INJECTION ACCUMULATOR NOZZLES. January 30,1991.

26) Rossi, Charles E., Information Notice 92-15: FAILURE OF PRIMARY SYSTEM COMPRESSION FITTING, February 24, 1992.

34

27) Grimes, Brian K., Information Notice 93-20: THERMAL FATIGUE CRACKING OF FEEDWATER PIPING TO STEAM GENERATORS, March 24, 1993.

28)Knapp, Malcolm R., Information Notice 94-38: RESULTS OF A SPECIAL NRC INSPECTION AT DRESDEN NUCLEAR POWER STATION UNIT I FOLLOWING A RUPTURE OF SERVICE WATER INSIDE CONTAINMENT, May 27, 1994.

29) NRC Bulletin 74-IOA: FAILURES IN 4--INCH BYPASS PIPING AT DRESDEN-2, 12/17/74.

30) Davis, John G., Information Notice 75-01: THROUGH-WALL CRACKS IN CORE SPRAY PIPING AT DRESDEN-2, January 31, 1975.

31)NRC Bulletin 76-04: CRACKS IN COLD WORKED PIPING AT BWR'S, March 30, 1976.

32) Thompson, Dudley, Circular 76-06: STRESS CORROSION CRACKS IN STAGNANT, LOW PRESSURE STAINLESS PIPING CONTAINING BORIC ACID SOLUTION AT PWR's, November 22, 1976.

33)NRC Bulletin 79-03: LONGITUDINAL WELD DEFECTS IN ASME SA -312 TYPE 304 STAINLESS STEEL, March 12, 1979.

34) NRC Bulletin 79-13: CRACKING IN FEEDWATER SYSTEM PIPING, June 25, 1979.

35) Moseley, Norman C., Information Notice 79-19: PIPE CRACKS IN STAGNANT BORATED WATER SYSTEMS AT PWR PLANTS, July 17, 1979.

36) NRC Information Notice No. 81-04: CRACKING IN MAIN STEAM LINES, February 27, 1981.

37) Sheron, Dr. Brian, Proposed Modifications to ECCS Analysis Requirements, Presentation at Penn State University, September 23, 2004.

38)NRC Document, 10 CFR 50.46 LOCA Frequency Document (Attachment).

35

PLANTlTYPd PLPETYPA SYSTEM GROUP I APPARENT CAUSE GROUP 1 ýRECOC .ck-FI. C:a-kPaj I Defonna*an LWoe Loek Leak Pf-Lea I Ruptre I S-eancs Iml ek [IwW1 l I

Is i

I 3

I I I

.A AUXC I VibraboM~ague HF.CO*STANST EC CS_

Cs I1

2 -

1 I I I 9

ThwnwaI " Sktcv("tx-vusw,4uomsv FWC I -- I I-- FWC I FW(

PWIR P -WR Pi 2

I-saw~o 1ov e.lo - iL 1v

nq -L1 2 0 f 3 4

2 I H4 - I 4 RAS I _____ Erde.Ffsckfs HAS RAS RAS RAS I

-2 2

2

-I-I I

-PWR I ss I RAS I t

J*

11 i3 44

ss - - ~ ¶ amls SOSf 3I I I - I T ¶SCCONTransauaSC 2 4 TsI"flF ý! 1 4 1 Thnr14algqW 2 I I 1Y

I 2

I 3

PWR 1 2 PWVH SR [ TGSCC.1r4Wlsgfta FCC-

-PWR -I SS SIR f TOSCO.Trwww&dWSCC SIR I TGSCC.TrMUTMIA8I 5CC SIR 7

3 I -7 I1 3- 1 I I 9

4 - a I 1ýi

PLANTTYPEI Pip :TvYp I .*AyI'EU ROUP IGROUP OPTOTATR RECORDS6. I CZ&a(-Pul I C~ack-PVa1 I D.efc~nalIo L&9.eLois I LeaS I PM-L.eis I Ru e I Seversz*. e SineSLe Iak l~n APPARENT CAUSE Conrcong Cenwosi-n _T.

common Corrosion Ccdroslon CrO.gon.a a n 6 I Eros.,-cro<Atat Eroslan-caosion Eros~onorosso Ef"o64MCncrosicr HF.CONS'IANST HFIPCONSTANST HF'Fabrm~onff&

Maouc0ctO9a~fkxw-dCc4 on Mirra"oCan k~ihd Cwioslon MUOUoio9caf Ibiced Corrosion 1 I - I-Sevee ouV Se'eove. Ca~ju, uwepbon, v~braboi-Fabqw.

E*lue fr* c~,e SWR Ccarouo HF.CONSTANST 2

105CC,k- E9.nuaSCC HFWA1b Erro -I S3WR K05rC- irdegrwia SOC 4 uWm I

-BWR I CS_

BWR I

pvvdc FWC I1 F

8 V~148I~_

3WR I-Cs F'

BWVR I 1

8WR Cs BWR 1 6s Css SS RAS 3 1 4

6 I 3

RAS

OWR I SS I I -I 5

6WR I 8 I HF:faucabon Enc I I I I BWR BYR I G 2 I1

__ ____ I '

1 I T- I- I I i~I 7LJ~7L2 iLYI I -

I SIR I1 I as7I SLR I_ 5 21 St I

.1 I

I

Appendix B Haddam Neck PWR CS 2.25 4 Erosion GL 89-08 CANDu ' PWR CS 4 4 Thermal Fatigue Korean CANDU PWR CS 4 4 Thermal Fatigue Korean CANDU PWR CS 4 4 Thermal Fatigue Korean CANDU PWR CS 4 4 Thermal Fatigue Korean Millstone Unit 3 PWR CS 6 5 EroslonlCorroslon IN 91-18 Arkansas Nuclear One Unit 2 PWR CS 14 6 Erosion IN 89-53 DC Cook Unit 2 PWR CS 16 6 Erosion Bulletin 79-13 DC Cook Unit 2 PWR CS 16 6 Erosion Bulletin 79-13 Fort Calhoun Station PWR CS 12 6 FAC IN 97-84 Surry Unit I 'PWR CS 30 6 Not yet determined IN 81-04 Surry Unit 2 PWR CS 18 6 Erosion/Corrosion IN 86-106 Trojan 1 PWR 'CS .14 6 Erosion IN 87-36 Zion 1 PWR CS 24 6 Human Factor IN 82-25 FR (Framatome Reactors) PWR CS 10 6 Corrosion Korean FR (Framalome Reactors) PWR CS 28 6 Corroslon Korean

.. .:Diabo Canyon Unit .*iPWR: CS, f - . .. ,hermal Fatigue7- - IlN,92-20*Ž,;'

.. ,..Lolsa.U"t. . .... .*$;:CS I.P.WR:. ,-'.t': s.'.Eroslon/Corrosion-;.,.-..- .IN 91-.18.:.

. -.. ,.S urry Unit .. ;: -- ;.'-..A ';CS , ;,!,,,Thx p ." "..Eros lon iCorrosgone .lN 912 8--.-',.

Wolf Creek PWR SS 0.25 1 Vibration IN 89-07 KSNP Korean Standard Nuclear Power Plant PWR SS 0.375 1 Thermal Fatigue Korean Oconee Unit 3 PWR SS 0.75 1 1 Mechanical Failure IN 92-15 WH-3 PWR SS 0.75 1 Flow Induced Vibration Korean WH-3 PWR SS 0.75 1 Flow Induced Vibration Korean H.B. Robinson Unit 2 PWR SS 2 3 SCC IN 91-05 Oconee Unit 2 PWR SS 2 3 Vibration IN 97-46 Prairie Island Unit 2 PWR SS 2 3 SCC IN 91-05 WH-3 PWR SS 2 3 Flow Induced Vibration Korean WH-3 PWR SS 2 3 Flow Induced Vibration Korean WH-3 PWR SS 2 3 Flow Induced Vibration Korean Crystal River Unit 3 PWR SS 2.5 4 Fatigue IN 82-09 Fort Calhoun Station PWR SS 3.5 4 SCC IN 82-02 Maine Yankee PWR SS 3.5 4 SCC IN 82-02 Maine Yankee PWR _S 3.5 4 SCC IN 82-02 Maine Yankee PWR SS 3.5 4 SCC IN 82-02 Maine Yankee PWR SS 3.5 4 SCC IN 82-02 Maine Yankee PWR SS 3.5 4 ScC IN 82-02 Maine Yankee PWR SS 3.5 4 SCC IN 82-02 Ginna PWR SS 8 5 SCC IE Clrcular76-06 Foreign PWR SS 8 5 Thermal Stress Bulletin 88-08 Arkansas Nuclear One Unit I PWR SS 10 6 SCC IE Circular76-06 Oconee Unit 2 PWR SS 24 6 Erosion IN 82-22 Sequoyah Unit 1 PWR SS 16 6 Fatigue IN 95-11 Sequoyah Unit 2 PWR SS 10 6 Human Factor IN 97-19 Surry Unit 2 PWR SS 10 6 SCC IE Clrcular7606

o. ;t

P.WR-* ;-:'..",.S .; i Var . * -- i - ".Human Factor .::, '. 'Bulletin.79-03:_.

,"-....San ofreunit 2,.,PWR .- ?Y .* ' --0 Human'Factor,:.'- ,.';.Bu1Ietin"79 5o.': San Vnofr.Unit 3-r':'- .* -PWR~ -r*",SS", ;4.,Yar.-.i.H i,.*. ,. '*'. -.'-",*Human'Factor:!*:., ,"Bulletin :79-03;"-

.. . . .. -j -- ,, S ;..... . . .. .°, . o

,-;-,. TMI unitElh...Ž.:,, PWR- ,SS BW, ,7  :, Y'.  : '"M*IRC<,*;.;N 79-19.*&*

Ti~~unk*It".t:.$;&;..:ý---: ,-PWWRý%.",S 01 - PA'%SC*W '<W 7I Y'

. , o e h . Y :PWR ""- ".- " ---- - " ""  : .. IN 88-01  :

-wýý'on Beachnt~'~iW; *: ~.': ____ ~ ~ .-.. ~

- ~N9-

Appendix B (cont.)

Pipe Size FalrMehns Rfrnc Plant Type Material Diameter Group Failure Mechanism Reference Dresden Unit 2 BWR CS 4 4 Human Factor Bulletin 74-10 Nine Mile Point Unit 2 BWR CS 8 5 Fatigue Event 36016 Vermont Yankee BWR CS 12 6 SCC IN 82-22 Cooper Station BWR SS 0.25 1 Vibration IN 89-07 Pilgrim BWR Ss 1 2 Corrosion IN185-34

'Browns Ferry 3 BWR. SS 4 4 .. SCIN e 84-41 Browns Ferry 3 BWR SS 4 4 SCC IN 84-41 Nine Mile Pointr Unit BWR SS 6 5 SCC Bulletin 76-04 Nie.iDreseden Unit 2 BWR SS lo 6 Thermal Fatigue IN75-01 Dreseden Unit 2 BWR SS 10 6 Thermal Fatigue IN75-01 Dreseden Unit 2 BWR SS 10 6 Thermal Fatigue IN 75-01 Dreseden Unit 2 BWR SS 10 6 Thermal Fatigue IN 75-01 Dreseden Unit 2 BWR SS 10 6 Thermal Fatigue IN 75-01 Hatch Unit I BWR SS 22 6 SCC IN83-02 Hatch Unit I BWR SS 22 6 SCC IN 83-02 Hatch Unit I BWR SS 22 6 SCC IN 83-02 Hatch Unit I BWR SS 22 6 SCC IN 83-02 Hatch Unit 1 BWR SS 22 6 SCC IN 83-02 Hatch Unit I BWR SS 20 6 SCC IN 83-02 Hatch Unit 1 BWR SS 24 6 SCC IN 83-02 Montecelloh BWR SS 22 6 SCC IN 83-02 Montecello BWR SS 12 6 SCC IN 83-02 Montecello BWR SS 12 6 SCC IN 83-02 Montecello BWR SS 12 6 SCC IN 83-02 Montecello BWR SS 12 6 SCC IN 83-02 Montecello BWR SS 12 6 SCC IN 83-02

~Desdn.U~t ~ BWR-Ž{,

ý:,'; - .1/2ý:--VtFreezing v'N943 I~;';

IHighlighted.plants-.were 'notused in thedata:analysis due'to.rmissing-information .'.-"'....

a p Appendix C. Collapsed OPDE Database Collapsed OPDE Raw Data as function of Pipe Size Plant Type Pipe Size Group Resulting Number of Failures (inches) CS SS CS+SS 0.0-1.0 154 544 698 1.0-2.0 74 154 228 2.0-4.0 78 75 153 PWR 4.0-10.0 126 112 238

> 10.0 93 126 219 Total 525 lOll 1536 0.0-1.0 118 257 375 1.0-2.0 32 75 107 BWR 2.0-4.0 32 227 259 4.0-10.0 50 234 284

> 10.0 39 291 330 Total 271 1084 1355 0.0-1.0 272 801 1073 1.0-2.0 106 229 335 2.0-4.0 110 302 412 4.0-10.0 176 346 522

> 10.0 132 417 549 Total 796 2095 2891

Collapsed OPDE Raw Data as function of Failure Mechanism Plant Type Failure Mechanism Resulting Number of Failures PlntTyealue _eas .. CS SS CS+SS Corrosion 106 28 134 FAC 119 121 240 MIC 43 1 44 Erosion 96 12 108 Fatigue 92 501 593 PWR Human Factors 36 126 162 Mechanical Failures 22 37 59 SCC 5 169 174 Water Hammer 0 2 2 Mise 6 14 20

_ _Total 525 1011 1536 Corrosion 29 32 61 FAC 58 63 121 MIC 6 1 7 Erosion 40 9 49 Fatigue 71 225 296 BWR Human Factors 24 85 109 Mechanical Failures 18 25 43 SCC 19 624 643 Water Hammer 2 1 3 Misc 4 19 23 Total 271 1084 1355 Corrosion 135 60 195 FAC 177 184 361 MIC 49 2 51 Erosion 136 21 157 Fatigue 163 726 889 PWR+BWR Human Factors 60. 211 271 Mechanical Failures 40 62 102 SCC 24 793 817 Water Hammer 2 3 5 Misc 10 33 43 Total 796 2095 2891

I v Appendix D - References

1) Lydell, Bengt & Mathet, Eric & Gott, Karen, PIPING SERVICE LIFE EXPERIENCE IN COMMERCIAL NUCLEAR POWER PLANTS: PROGRESS WITH THE OECD PIPE FAILURE DATA EXCHANGE PROJECT, ASME PVP-2004 Conference, La Jolla, California, USA, July 26, 2004.

2) Nyman, Ralph & Hegedus, Damir & Tomic, Bojan & Lydell, Bengt, RELIABILITY OF PIPING SYSTEM COMPONENTS - FRAMEWORK FOR ESTIMATING FAILURE PARAMETERS FROM SERVICE DATA, SKI/RA, ENCONET Consulting GesmbH, Sigma-Phase, Inc., December 1997.

3) OPDE Database Light, OECD Piping Failure Data Exchange (OPDE) Project, OECD/NEA (2005).

4) Choi, Sun Yeong and Choi, Young Hwan, PIPING FAILURE ANALYSIS FOR THE KOREAN NUCLEAR PIPING INCLUDING THE EFFECT OF IN-SERVICE INSPECTION, KAERI and KINS, 2004.

5) DeYoung, Richard C., NRC - Bulletin No. 82-02: DEGRADATION OF THREADED FASTENERS IN THE REACTOR COOLANT PRESSURE BOUNDARY OF PWR PLANTS, June 2, 1982.

6) Information Notice No. 82-09: CRACKING IN PIPING OF MAKEUP COOLANT LINES AT B&W PLANTS, March 31, 1982

7) Jordan, Edward L., Information Notice No. 82-22: FAILURES IN TURBINE EXHAUST LINES, July 9, 1982

8) DeYoung, Richard C., NRC Bulletin N. 83-02: STRESS CORROSION CRACKING IN LARGE-DIAMETER STAINLESS STEEL RECIRCULATION SYSTEM PIPING AT BWR PLANTS, March 4,1983

9) Jordan, Edward L., Information Notice No. 84-41: IGSCC IN BWR PLANTS, June 1, 1984.

10) Jordan, Edward L., Information Notice No. 85-34: HEAT TRACING CONTRIBUTES TO CORROSION FAILURE OF STAINLESS STEEL PIPING, April 30, 1985.

11) Partlow, James G., Generic Letter 89-08: EROSION/CORROSION-INDUCED PIPE WALL THINNING, May 2,1989.

12) Marsh, Ledyard B., Information Notice 99-19: RUPTURE OF THE SHELL SIDE OF A FEEDWATER HEATER AT THE POINT BEACH NUCLEAR PLANT, June 23,1999.

13) Roe, Jack W., Information Notice 97-84: RUPTURE IN EXTRACTION STEAM PIPING AS A RESULT OF FLOW-ACCELERATED CORROSION, December 11,1997.

0 f

14) Jordan, Edward L., Information Notice 86-106: FEEDWATER LINE BREAK, February 13, 1987.

15) Rossi, Charles E., Information Notice 89-53: RUPTURE OF EXTRACTION STEAM LINE ON HIGH PRESSURE TURBINE June 13, 1989.

16) Rossi, Charles E., Information Notice 91-18: HIGH-ENERGY PIPING FAILURES CAUSED BY WALL THINNING, March 12, 1991.

17) Grimes, Brian K., Information Notice 95-11 : FAILURE OF CONDENSATE PIPING BECAUSE OF EROSION/CORROSION AT A FLOW-STRAIGHTENING DEVICE, February 24, 1995.

18) Weaver, Brian, Event Notification Report 36016: MANUAL REACTOR TRIP DUE TO HEATER DRAIN LINE BREAK, August 12, 1999.

19) Rossi, Charles E., Information Notice 87-36: SIGNIFICANT UNEXPECTED EROSION OF FEEDWATER LINES, August 4,1987.

20) Rossi, Charles E., Information Notice 89-07: FAILURES OF SMALL-DIAMETER TUBING IN CONTROL AIR, FUEL OIL, AND LUBE OIL SYSTEMS WHICH RENDER EMERGENCY DIESEL GENERATORS INOPERABLE, January 25, 1989.

21) Rossi, Charles E., Information Notice 88-08: THERMAL STESSES IN PIPING CONNECTED TO REACTOR COOLANT SYSTEMS, April 11,1989.

22) Rossi, Charles E., Information Notice 88-01: SAFETY INJECTION PIPE FAILURE, January 27, 1988.

23) Martin, Thomas T., Information Notice 97-19: SAFETY INJECTION SYSTEM WELD FLAW AT SEQUOYAH NUCLEAR POWER PLANT, UNIT 2, April 18,1997.

24) Slosson, Marylee M., Information Notice 97-46: UNISOLABLE CRACK IN HIGH-PRESSURE INJECTION PIPING, July 9,1997.

25) Rossi, Charles E., Information Notice 91-05: INTERGRANULAR STRESS CORROSION CRACKING IN PRESSURIZED WATER REACTOR SAFETY INJECTION ACCUMULATOR NOZZLES January 30, 1991.

26) Rossi, Charles E., Information Notice 92-15: FAILURE OF PRIMARY SYSTEM COMPRESSION FITI'ING, February 24, 1992.

27) Grimes, Brian K., Information Notice 93-20: THERMAL FATIGUE CRACKING OF FEEDWATER PIPING TO STEAM GENERATORS, March 24,1993.

28) Knapp, Malcolm R., Information Notice 94-38: RESULTS OF A SPECIAL NRC INSPECTION AT DRESDEN NUCLEAR POWER STATION UNIT I FOLLOWING A RUPTURE OF SERVICE WATER INSIDE CONTAINMENT, May 27, 1994.

29)NRC Bulletin 74-IOA: FAILURES IN 4--INCH BYPASS PIPING AT DRESDEN-2, 12/17/74.

30) Davis, John G., Information Notice 75-01: THROUGH-WALL CRACKS IN CORE SPRAY PIPING AT DRESDEN-2, January 31, 1975.

31)NRC Bulletin 76-04: CRACKS IN COLD WORKED PIPING AT BWR'S, March30, 1976.

32) Thompson, Dudley, Circular 76-06: STRESS CORROSION CRACKS IN STAGNANT, LOW PRESSURE STAINLESS PIPING CONTAINING BORIC ACID SOLUTION AT PWR's, November 22, 1976.

33)NRC Bulletin 79-03: LONGITUDINAL WELD DEFECTS IN ASME SA -312 TYPE 304 STAINLESS STEEL, March 12, 1979.

34)NRC Bulletin 79-13: CRACKING IN FEEDWATER SYSTEM PIPING, June 25, 1979.

35) Moseley, Norman C., Information Notice 79-19: PIPE CRACKS IN STAGNANT BORATED WATER SYSTEMS AT PWR PLANTS, July 17, 1979.

36)NRC Information Notice No. 81-04: CRACKING IN MAIN STEAM LINES, February 27, 1981.

37) Sheron, Dr. Brian, Proposed Modifications to ECCS Analysis Requirements, Presentation at Penn State University, September 23, 2004.

38) NRC Document, 10 CFR 50.46 LOCA Frequency Document (Attachment).