The second main bullet on VG7 should be decrease with “decreasing” piping size.

Dave Harris disagreed with comment that PFM models had problems modeling mitigation.

There was a problem with VG16 with interpreting results for Panelist L.

A number of the panelists were surprised with VG17 that surge line results for Cat 5 are comparable to that of cold leg.

There is a discrepancy in maximums for Category 6 LOCAs at 25 years between VG19 and 20. VG20 shows participant L having the maximum value while VG19 shows participant B as having the maximum value.

Participant J showed the most impact of age on the LOCA frequencies; really obvious in VG21.

Bruce pointed out that for higher Category LOCAs that there was less uncertainty; may be an artifact that there are less systems than can contribute (only large diameter); also these larger systems are better inspected, i. e., better controlled.

Every plot shown is for 25 years unless specifically stated.

The Base Case LOCA frequency models are based on three Excel files entitled “PlantBWelds” (BWR Base Case), “PlantA. aWelds” (PWR RC-HL and Surge Line), and “PlantA. bWelds” (PWR HPI/NMU). These Excel files are part of the plant design information supplied by members of the Expert Elicitation Panel. Adding information on weld type by location and weld failure parameters to respective Excel files provides the basis for calculating LOCA frequencies including uncertainty distributions.

D.6.1 LOCA Frequency Models

For the BWR Base Case, the original “PlantBWelds” includes two spreadsheets, one for the feedwater system and one for the reactor recirculation system. In a first step to create a LOCA frequency model, each spreadsheet was split into two, one for Loop A and one for Loop B of respective system. Next a new column was added to each new spreadsheet to include information on weld type by location. A review of isometric drawings provided the input to the new columns.

The statistical information that is summarized in Section D.5 of this report is included in five separate spreadsheets of the modified Excel file. The posterior weld failure rates are included on Tab “Parameters,” each parameter assigned a unique variable name. The calculation of LOCA frequency, including Monte Carlo merge operations are performed on Tabs “Intermediate FW_A”, “Intermediate FW_B”, “Intermediate RR_A”, and “Intermediate RR_B.” These intermediate calculation sheets are linked to the design information for FW Loop A, FW Loop B, RR Loop A and RR Loop B, respectively. Using the variable names as defined on Tab “Parameters,” each weld is assigned an appropriate failure rate (including uncertainty distribution as defined using a Crystal Ball “assumption”). Finally, an integrated calculation of LOCA frequency by leak threshold category (0 through 5) is performed on a separate spreadsheet, which is linked to the intermediate calculations. Each integrated LOCA frequency calculation is defined as a Crystal Ball “forecast.”

For the PWR Base Case, the original “PlantAWelds” includes a single spreadsheet with the ASME XI Class 1 Category B-F/B-J welds. In a first step to create a LOCA frequency model, each spreadsheet was split into two, one for the RC-HL and one for the pressurizer surge line. A third spreadsheet with the HPI/NMU weld listed was added to form a new Excel-file corresponding to the PWR Base Case LOCA frequency model.

D.6.2 BWR Base Case LOCA Frequencies at T = 25

Summarized in Table D.16 are the LOCA frequency uncertainty distributions that are derived for BWR Base Case Cat0 through Cat6 LOCA. The results are representative of Plant B after 25 years of operation (T = 25).

D.6.3 PWR Base Case LOCA Frequencies at T = 25

The PWR Base Case Cat0 through Cat6 LOCA frequencies including uncertainty distributions, are summarized in Table D.17. These results are representative of Plant A. a/A. b after 25 years of operation (T = 25).

The base cases LOCA frequencies will provide absolute LOCA estimates that each panel member will use to anchor the relative likelihood of LOCAs in other (non-base case) piping systems. Each panel member will also determine how well the base cases depict expected current and future LOCA performance in the piping systems that they model. It is therefore not important that a panel member agree with the modeling assumptions, approach, and results provided by the base case team. However, it is imperative that the base case development is completely understood by each panel member. Each panel member will be able to correct perceived deficiencies in the calculated base case frequencies during the elicitation. Each panel member will also determine, relative to the base case results, the LOCA contributions of other (non base case) piping systems, and the contributions and uncertainty induced by the primary piping system variables.

It would have helped to have had Gery’s presentation earlier, before the panel tried to answer seismic question.

Would have been nice to have a video of plants showing various systems as one tours plants with video camera.

Amount of information available was overwhelming; try to do a division of labor so one or two people review something and provide a tutorial to others so everyone is working from same basis; otherwise everyone is inventing the wheel themselves; maybe have a meeting to review these tutorials.

Need a roadmap of where information can be found.

Periodic/weekly update of changes made to ftp site; alternatively an alert message when something added to site; maybe a readme file when something added and what was added and when.

NRC management must make sure that staff are available to panel members during the process; Rob getting pulled off for Davis Besse was a problem; delayed things and then panel members only had a few weeks to respond at the end.

Bruce would like time at meetings to do actual work on elicitations because once they get back home they will get pulled off on to other things and won’t be able to get back to answering questions for a long time.

Mr. Galyean has over 25 years experience in performing PRA on commercial nuclear power plants. After earning his Bachelor of Science degree in Physics at Millersville State College in 1976, he went to the University of New Mexico where he obtained his Master of Science degree in Nuclear Engineering in 1978.

Mr. Galyean is a senior PRA analyst at the Idaho National Engineering and Environmental Laboratory (INEEL), which he joined in 1986. In recent years, Mr. Galyean was the Principal Investigator of two large NRC-sponsored PRA-related programs. The latest entailed performing detailed statistical analyses on nuclear power plant reliability data collected from operating experience. This data was then used to estimate both the historical system reliability for operational missions actually performed and the expected reliability for postulated risk-significant missions. The earlier program was a research program aimed at supporting the resolution of USNRC Generic Issue 105, “Interfacing System Loss-of-Coolant Accidents at Light Water Reactors.” This program integrated many disciplines in the area of risk analysis, including: human reliability, thermal-hydraulics, consequences, external events, and stress analysis. In addition a number of innovative techniques were developed for generating human error probabilities and fluid system rupture probabilities. Mr. Galyean has also developed and teaches 1-week courses on PRA modeling techniques and on Level-2 PRA, and developed low-power and shutdown (LP/SD) specific PRAs as part of the Standardized Plant Analysis Risk (SPAR) model program. Other activities include serving on an NRC-sponsored expert panel that was formed to produce updated estimates of loss of coolant accident frequencies, supporting the independent validation of a PFM computer code (FAVOR), participating as the INEEL project manager on a multi-company, multi-lab project funded through the DOE NERI program performing a risk-informed assessment of new reactor design requirements. He has also provided significant contributions to a number of other risk/reliability related programs, including NUCLARR (Nuclear Computerized Library for Assessing Reactor Reliability), a PC-based databank of reliability data for both hardware and human actions; a reliability analysis of the INEEL site power distribution system; and an analysis of the risk significance of possible operator actions for managing severe accidents at a commercial nuclear power plant.

Prior to joining INEEL Mr. Galyean worked for Falcon Research and Development Company, Science Application International Corporation, and the NUS Corporation.

Mr. Galyean is a member of the American Nuclear Society, and has served on an International Atomic Energy Agency (IAEA) review teams (International Peer Review Service — IPERS) that reviewed a Level-1 PRA on the Borssele NPP (Dutch) and a Level-2 PRA on the Krsko NPP (Slovenian).

Prior to this presentation, Rob noted that this presentation was not included in the handout, but will be put on the ftp site. Rob also asked that the base case team provide him in an electronic format with any references that they used so that the references can be put on the ftp site.

It was first noted that we have not been successful in locating additional failure data for several of the components where we were lacking data. Fred Simonen had spoken with Spencer Bush and was not able to locate additional data. Rob and others were also somewhat unsuccessful.

Based on the leak rate versus opening area from one of the prior presentations, Pete Riccardella questioned if we needed multiple failures for the heater sleeves as shown in slide 3 in Rob’s presentation. Bruce Bishop indicated that thermal fatigue needed to be added to the degradation mechanisms for vessel

head bolts (slide 4). Rob indicated that these tables are not filled in completely at the present time. These tables are neither comprehensive nor as complete as the piping component table. We focused more on the piping issues at the kick off meeting than we did on the non-piping issues.

For all of the major groups, we initially listed at least one component (bolded item in the original non­piping tables in presentation and kick-off meeting notes document) where the panel thought that failure (leaks or cracks) data is available. It was thought that these bolded items represented potential non­piping base cases for anchoring. If these non-piping base cases can’t be develop, then we will have to use a piping base case for anchoring. As Bruce Bishop indicated earlier that would be an unnatural comparison, making it somewhat difficult. The question was asked how we develop these non-piping base cases. It was thought that areas where we could come up with passive-system failure data, e. g., steam generator tubes, would be logical first choices. It was thought that data on steam generator tube failures should be easy developed.

Bruce Bishop indicated that there was an INEEL NUREG report by Vic Shah that has piping and non­piping failure data that could possibly be used for base cases. Obviously, this would be a good place to start. It was noted though that this INEEL report is for PWRs only, and the steam generator tube failure data will be included in this report. Bill Galyean and Rob Tregoning are to locate and distribute copies of this report to the panel members.

Pete Riccardella volunteered to run a base case analysis for feedwater nozzles and the belt line region for the RPV due to LTOP. Pete thought he could have some analysis results by the end of the month. These would be for BWRs and will be done using predetermined flaw distributions. Note, to date there have only been cracks, there have been no leaks to date.

Another potential non-piping base case for anchoring is the PTS study for the belt line region of the RPV. This would be for PWRs. Rob Tregoning will extract this data. Gery Wilkowski noted that we should use caution in using the PTS results and should only use the contribution from non-pipe break transients to ensure that the comparison is consistent. In a related action, Bruce Bishop volunteered to provide a Westinghouse nozzle study for PWRs and will also provide PWR vessel failure probability for areas outside the beltline region covered by the PTS study.

Pete Riccardella suggested classifying the alloy 600 penetrations as non-piping failures for consistency with other nozzles. A suggestion was made to move the in-cores and CRDMs to the non-piping category. Fred Simonen suggested putting them all under a separate category called vessel penetrations, with a separate bin for RPVs and pressurizers. Karen Gott and Pete Riccardella are to create a base case for penetrations using the CRDM data based on a prior MRP study. Rob Tregoning will move all the vessel penetrations from piping to the vessel bin and supply updated tables.

Bengt Lydell has a non-piping data base that he will query by end of the month. He will also query the IRS data base (Incident Reporting System by INEA) by end of the month. Note the IRS database only includes data countries chose to include. It was also noted that MITI and NUPEC (both in Japan) have a data base for non-piping components. The NRC supposedly has a copy of this database. Rob will check into relevancy and availability. Bill Galyean volunteered to examine his database to see any relevant non­piping events, i. e., stream generator tube rupture statistics, incidents of bolting connections, etc., by the end of the month. Rob will be the conduit for getting the results from the various individuals searching the databases out to the rest of the group. Results will be posted to the ftp site and more important items will be bulk emailed to the panel.

The question was asked if the panel members could take home the modified tables for the pressurizer, RPV, pumps, valves, and steam generators and fill them out and return them back to Rob within two weeks. Rob will modify these non-piping tables (Tables B.1.7 and B.1.8 from the First Elicitation Minutes) with the additions made during the second meeting discussions and send them out to the panel members. Each panel member is to modify these non-piping tables and get them back to Rob Tregoning.

Fred Simonen has an electronic version of the GALL report (Generic Aging Lessons Learned) which identifies the key degradation mechanisms that could be used to help fill out these tables. Fred will extract the relevant tables to be used in identifying the key degradation mechanisms.

Base Case Report 2 develops BWR and PWR LOCA frequency distributions using a ‘bottom-up approach.’ Statistical analysis of relevant service experience data is used to quantify the weld failure rate and rupture frequency of individual welds. Next the failure rate and rupture frequency (= LOCA frequency) for an entire system is calculated by concatenating the individual weld failure rates and rupture frequencies. Markov model theory is used to evaluate the influence of alternate strategies for in-service inspection and leak detection on the frequency of leaks and ruptures.

D.2.1 Overview of Analysis Steps

Different approaches have been applied to estimating pipe failure rates and rupture frequencies; from probabilistic fracture mechanics, via direct statistical estimation to expert judgment. The most straightforward approach is to obtain statistical estimates of piping component failure rates based on data collected from field experience. A variation of this approach is to augment statistical estimates of pipe failure parameters with simple correlations that express the problem in terms of a failure rate and a conditional probability for each failure mode of interest such as the approach used in NUREG/CR-5750 [D.17].

A limitation of the statistical analysis approach is that attempts to segregate the service data to isolate the impact of key design parameters and properties of various degradation and damage mechanisms often leads to subdividing a database into very sparse data sets. If not optimized properly, this approach may introduce large uncertainties in the failure rate estimates. In addition, historical data may reflect the influence of no longer relevant inspection programs. If changes to these programs have been implemented, such changes may render the failure rate estimates no longer relevant. In risk-informed applications, the failure data and analysis methods need to provide future predictions of piping system reliability that can account for changes in the inspection strategy or improvement in the NDE technology.

An objective of the work documented in this report is to demonstrate the utility of a pipe failure data collection. Time-dependent LOCA frequencies are calculated by making full use of the PIPExp database in combination with Markov model theory [D.14]. The LOCA frequency calculation in this report is structured to support the Expert Elicitation and consists of four steps; each step is addressed in a separate report section:

• Section D.3. The service experience that is applicable to the five bases cases is summarized in this section. The data summaries correspond to queries in the PIPExp database.

• Section D.4. The approach to calculating time-dependent LOCA frequencies is presented. A Bayesian update process is used to derive failure parameters that reflect the attributes of respective base case definition. The results of this analysis step are in the form of generic weld failure rate distributions. These distributions represent the industry-wide service experience prior to the implementation of the specific pipe failure mitigation programs that are currently in place.

• Section D.5. In this section current state-of-knowledge (or base case specific) weld failure rate distributions are develop. The chosen estimation approach includes a formal uncertainty analysis that accounts for uncertainty in the failure data and exposure data. Engineering judgment and insights from the review of service data are used to address the conditional probability of pipe failure given presence of through-wall flaws.

• Section D.6. An Excel spreadsheet format is used to develop LOCA frequency models corresponding to each of the five base cases. These models generate LOCA frequency distributions at T = 25 years. A Markov model is used to investigate the time-dependency of LOCA frequencies. The output of this model consists of LOCA frequencies at T = 40 years and T = 60 years.

Rob Tregoning then made a presentation on LOCA frequency determination using expert elicitation. The objectives of this presentation are to motivate the expert elicitation effort, discuss the limitations of relying solely on past operating experience; present ongoing NRC-sponsored research in this area; define the objective of the expert elicitation; outline the approach; and discuss the structure of the kick-off meeting. Some of the specific key points from his presentation and subsequent discussion are outlined below:

• The LB LOCA design basis size will be determined by considering all relevant LOCA sources.

A PFM-based model is under development for predicting LOCA contributions as a function of break size. This code will also account for LOCAs from non-piping sources, and will include contributions from future unknown failure mechanisms. Expert elicitation input from this panel will be used throughout program development.

• The objectives of the elicitation are to

• Develop future SB, MB, and LB LOCA frequency estimates extending up through the end of the plant-license-renewal period (approximately 35 years).

• Develop benchmark problems and standardized inputs for conducting PFM simulations of LB LOCA events in important BWR and PWR systems.

• The elicitation approach will construct base cases. Quantitative LOCA estimates as a function of break size will be developed for these base cases. Then, important variables and issues will be discussed within the relative framework of the base cases.

Discussion: It was stressed that the base cases will just provide a reference point. Adjustments to the base cases will account for the impact of those issues which contribute significantly to the LOCA frequencies. These adjustments must consider their effect on current LOCA frequencies and their time dependence up through the end of the plant license-renewal period (~ 35 years).

• The programmatic approach for the elicitation was presented. It consists of the following important areas.

• Conduct the kick-off meeting

• Develop elicitation questions

• Allow individual study to develop answers for elicitation questions

• Conduct the individual elicitations

• Analyze results (facilitation team)

• Conduct a wrap-up meeting

Discussion: It was conveyed that when the facilitation team queries each elicitation panel member, best estimate answers will be sought as well as the uncertainty in the estimates. It was also stressed that each panel member need not answer all questions, but only those that they feel that they are qualified and comfortable with answering. However, people will be encouraged to answer all questions, even in areas outside of their specific expertise. Uncertain knowledge should be reflected in the uncertainty estimates. Rob Tregoning indicated that one job of the facilitation team will be to filter out responses which are not well-founded, and exclude them from the final analysis.

• The principal components of the kick-off meeting were reviewed again. There are several major objectives which need to be accomplished during this kick-off meeting:

• Present the elicitation objectives and define fundamental terms to ensure common understanding.

• Undergo elicitation training to understand the process and approach.

• Construct methodology for developing baseline LOCA estimates. Develop a classification scheme and approach for issues which could affect the baseline LOCA estimates.

• Identify and classify issues for consideration. Discuss issues as necessary for clarification.

• Agree on significant issues to include in the elicitation.

• Determine the structure of the elicitation questions.

Similar degradation mechanisms as with PWR mechanisms: thermal fatigue, mechanical fatigue,

Sam commented with regards to Slide #4 that analysis of BWR feedwater says it should crack, but don’t find these cracks in service.

For Slide 8 it was suggested to change “risk” to “high” in last bullet.

Maximum diameter of RWCU system is 6 inch; not 24 inch as shown in VG17.

In order to decompose the elicitation topics further, the group determined that it would be useful to further decompose non base case piping systems and variables. This was accomplished by defining a set of reference conditions for each LOCA-sensitive piping system identified in Tables B.1.7 and B.1.8. The reference conditions are similar to the base cases in that they define a unique set of conditions (materials, geometric variables, mitigation and maintenance procedures, and degradation mechanisms) that can be analyzed. They are different from the base cases in that absolute LOCA frequencies will not be developed for the reference cases by the base case team. The reference cases for various systems will be compared to determine the relative LOCA-severity among piping systems. LOCA-severity variability within any specific system can then be gauged with respect to the reference case for that system. It will be up to each panel member to determine the method for developing these relative comparisons.

The BWR reference cases (Table B.1.11) represent specific combinations of the possible BWR piping as previously listed in Table B.1.7. In general, one material and degradation mechanism has been chosen is ideally representative of each piping system. The effect of fabrication defects and repair (discussed earlier) should be considered for its effect on the other degradation mechanism in all cases. Nominal pressure, thermal, RS, and DW loading should be considered for all cases. One loading transient was identified for each system. All the transients should be fairly well identified based on past discussion, but the overload transient for the control rod drive (CRD) piping needs to be better defined. In all cases, the snubber is considered to be functional.

Table B.1.11 BWR Reference Case Conditions

Specific piping, safe end, and weld materials were not determined by the group. Table B.1.11 represents an initial attempt to select these materials based on the general discussion. However, the group did decide to consider uncrevised Type 304 where proper in the reference cases. The mitigation and maintenance for all systems should assume NWC. Standard ISI with technical specification leakage detection should be considered along with augmented inspection as defined in generic letter 88-01. The mitigation and maintenance also has listed alloy 182 and stress improved, but these concepts need to be better defined and summarized.

The PWR reference case conditions are provided in Table B.1.12. The reference cases were again distilled from the LOCA-sensitive PWR piping systems (Table B.1.8). The philosophy behind this table was consistent with the BWR reference case development (Table B.1.11) with a few notable exceptions.

The effects of fabrication defects and repair on the other listed degradation mechanisms should again be considered in all piping systems. However, several PWR reference cases list multiple other degradation mechanisms which is a departure from the BWR approach (Table B.1.11). The PWR piping reference cases also account for nominal loading supplied by pressure, thermal, residual stress, and dead weight loading to each system. However, very few PWR systems have associated loading transients while all BWR systems do. It may be necessary to add associated PWR transients for consistency. The PWR piping, safe end and weld materials were also not specified. Some initial choices have been made in Table B.1.12, but feedback from the group is required in order to finalize selection. The mitigation and maintenance to be considered for each reference case consists of ISI with TSL. No other special mitigation procedures were identified.

Table B.1.12 PWR Reference Case Conditions

It is important that the baseline and reference case conditions be clearly defined prior to the start of the elicitation so that each panel member understands the general attributes of each of these cases. As mentioned previously, the elicitation questions will be structured to query the variability and uncertainty associated with each piping system with respect to the reference cases. Each panel member will compare the reference and base cases to assess the relative importance of each piping system to the total LOCA frequencies. Every effort will be made to accommodate all requests and information will be shared among the group. Additionally, any areas or issues which are not clear to a panel member should be raised to Rob Tregoning as soon as it arises.