4A. 1.1. Examine the failure scenarios for each of the five PWR non-piping components (pressurizer, valves, pumps, RPV, and steam generator). For each component, list the failure scenarios that provide a minimum of 80% of the total contribution for Category 1 (leak rates > 100 gpm [380 lpm ]) LOCAs in US plants after 25, 40, and 60 years of operation. Estimate the MV contribution of these failure scenarios (> 80%). Also, provide the 90% coverage interval for the total contribution estimate of these systems.

4A. 1.2. Repeat 4A. 1.1 for Category 2 through 6 LOCAs for the non-piping PWR components.

4A. 1.3. Repeat 4A. 1.1 and 4A. 1.2 for BWR non-piping components (valves, pumps, RPV).

Notes:

a. A failure scenario is associated with a specific non-piping component, material, degradation mechanism, etc.

b. Relevant BWR and PWR non-piping failure scenarios and components are discussed in the kick-off meeting notes document (called failure modes instead of scenarios in this document). These are also summarized in the elicitation summary tables.

c. Please see EQ 1 in the Elicitation Question Development document for additional information and an example for a similar question.

4A.2.1. Choose a piping or non-piping base case which results in the most natural comparison for each of the failure scenarios described in 4A.1.1 for all five PWR non-piping component classes.

Provide a MV estimate of the ratio for the Category 1 LOCA contribution of the chosen non­piping failure scenario to the chosen base case.

4A.2.2. Repeat 4A.2.1 for Category 2 through 6 LOCAs for the non-piping PWR components.

4A.2.3. Repeat 4A.2.1 and 4A.2.2 for BWR non-piping components (valves, pumps, RPV).

Notes:

a. Please see EQ 6 in the Elicitation Question Development document for additional information and an example for a similar question.

b. Non-piping base cases are currently being quantified to determine the leaking frequencies due to all degradation mechanisms for each non-piping component listed in the kick-off meeting notes document. There will also be non-piping base cases frequencies for items that have failed such as SGTRs. Additionally, non-piping base cases can still be chosen for making relative comparisons. For instance if a panelist considers valve body failure due to cavitation erosion to be significant for Category 1 PWR LOCAs, then valve body leakage can be chosen as the base case.

Question Set 4B

4B.1.1. List the PWR non-piping failure scenarios that provide a minimum of 80% of the total

contribution for Category 1 (leak rates > 100 gpm [380 lpm]) LOCAs in US plants after 25, 40, and 60 years of operation. Now estimate the MV contribution of these failure scenarios (> 80%). Also, provide the 90% coverage interval for the total contribution estimate.

4B.1.2. Repeat 4B.1.1 for Category 2 through 6 LOCAs for the non-piping PWR failure scenarios.

4B.1.3. Repeat 4B.1.1 and 4B.1.2 for BWR non-piping failure scenarios.

Notes:

a. This question differs from Elicitation Question 4A in that only the significant failure scenarios, regardless of component, need to be considered.

b. A failure scenario is associated with a specific non-piping component, material, degradation mechanism, etc.

c. Relevant BWR and PWR non-piping failure scenarios are discussed in the kick-off meeting notes document (called failure modes instead of scenarios in this document). These are also summarized in the elicitation summary tables.

d. Please see EQ 1 in the Elicitation Question Development document for additional information and an example for a similar question.

4B.2.1. Estimate the percentage contribution for each PWR non-piping failure scenario in 4B1.1 for Category 1 LOCAs after 25, 40, and 60 years of operation.

4B.2.2. Repeat 4B.2.1 for Category 2 through 6 LOCAs for the non-piping PWR scenarios.

4B.2.3. Repeat 4B.2.1 and 4B.2.2 for BWR non-piping failure scenarios.

Notes:

a. Please see EQ 2 in the Elicitation Question Development document for additional information and an example for a similar question.

4B.3.1 Pick either a piping or a non-piping base case (or a piping reference case) for comparison with one or more of your significant non-piping failure scenarios from 4B1.1 for Category 1 LOCAs. The comparison should be natural, but if possible, should be made with one of the most significant failure scenarios that you listed. Provide a MV estimate of the ratio of the non­piping failure scenario to the chosen base case as a function of operating time (40 and 60 years). Also, provide the 90% coverage range for this ratio.

4B.3.2 Repeat 4B.3.1 for Category 2 through 6 LOCAs for the non-piping PWR failure scenarios.

4B.3.3 Repeat 4B.3.1 and 4B.3.2 for BWR non-piping failure scenarios.

Notes:

a. Base case conditions for piping systems and are defined within the kick-off meeting notes document. Base case conditions for non-piping components are being developed as discussed in the notes to Elicitation Question 4A.2

b. Please see EQ 6 in the Elicitation Question Development document for additional information and an example for a similar question.

J.2.5 Elicitation Question 5: LOCA Probabilities of Piping Components under an Emergency Faulted Load

An emergency faulted load represents an initial design consideration for a possible large transient load that was not expected to occur over any particular plant’s operating life of 40 years (rare event), or a frequency less than approximately 0.025 yr-1. These loads could be due to seismic loading or any other large pressure transients. Base cases have been developed which examine the conditional failure probability for ASME Service Level B loading. This loading level was estimated for several plants to conservatively approximate a 1*SSE event on a pipe which is flawed up to the allowable technical specification leakage rate for the given piping system and degradation mechanism. An SSE event was initially a design-level earthquake amplitude that was thought to occur once in 40 years; however, operating experience to date suggests that the frequency of an SSE event occurring is less than that.

This question will ask you to list and quantify the effect of the most significant piping systems and degradation mechanisms that contribute to each LOCA category. The quantification will be done for two emergency faulted load sizes (ASME Service Levels B and D) for three assumed damage states. The damage states will consist of tech. spec. leakage rates, the onset of leakage through a slow drip (perceptible leak), and a surface crack with a/t = 0.5. The surface crack length will be assumed by each panelist and may be a function of degradation mechanism and material. A relationship between the failure loads for a circumferential through-wall-crack and surface cracks with a/t = 0.5 and different lengths is provided in the “Piping Seismic Base Cases” document. The likelihood of each damage state will also be ascertained by each panelist relative to the operational experience data for the leak-rate frequencies corresponding to each system listed, regardless of degradation mechanism. This assessment will require nine different relative comparisons for each LOCA size category and plant type (BWR or PWR).

The appendix of this document and the “Piping Seismic Base Cases” document provide the philosophy behind the seismic piping elicitation questions and detail the seismic piping base case calculations. Both documents should be read prior to answering this elicitation question.

5A. 1.1. List the piping systems and degradation mechanisms which most significantly contribute to Category 1 LOCAs given that an assumed emergency faulted load occurs with an equivalent magnitude of an ASME Service Level B (SLB) event for PWRs. This total list should summarize at least the top 80% contributing factors to Category 1 LOCAs under assumed faulted loading conditions. Also, for each system, list the loads which may result in SLB loading and indicate if they are primary (loading-controlled) or secondary (displacement — controlled). Provide the total contribution and also the 90% coverage interval for this estimate.

5A.1.2. Repeat 5A.1.1 for ASME Service Level D (SLD) loading

5A. 1.3. Repeat 5A. 1.1 and 5A. 1.2 for each PWR LOCA size category.

5A.1.4. Repeat 5A.1.1 — 5A.1.3 for BWRs.

Notes:

a. Information on relevant piping systems, degradation mechanisms, and piping sizes is contained in the “Elicitation Meeting Notes” from the kick-off meeting.

b. In this question, pick your piping systems assuming that the pipes will completely fail. Therefore, the LOCA size category will be directly related to the pipe size.

5A.2.1. Pick a representative set of seismic base-case conditions to use for comparison for each of your significant contributors to Category 1 LOCAs in PWRs.

5A.2.2. Repeat 5A.2.1 for each PWR LOCA size category.

5A.2.3. Repeat 5A.2.1 and 5A.2.2 for BWRs.

a. A PWR and BWR base case have been defined in the “Piping Seismic Base Cases” document for a specific degradation mechanism, pipe size, and material. Additionally, figures are available which show the effects of changing materials, piping size, and service level loading with respect to the base case definitions.

b. Comparisons to the selected base cases will be made in subsequent questions.

c. A relationship between the failure loads for a circumferential through-wall-crack and surface cracks with a/t = 0.5 and different lengths is given at the end of the Piping Seismic Base Case/Background section.

5A.3.1. Consider a single piping system and degradation mechanism combination identified for

Category 1 PWR LOCAs in 5.A.1 and the associated seismic base case identified in 5A.2. Determine the ratio of the CFP for this system/degradation mechanism combination (PTSL or PTsl@slB) to the CFP for the chosen seismic piping base case (PBC). Assume that a SLB emergency faulted load occurs and that the piping system is degraded by a through-wall crack that is leaking at the technical specification limit. Also provide the 90% coverage interval of this ratio.

5A.3.2. Consider the same piping system and degradation mechanism combination identified for

Category 1 PWR LOCAs in 5.A.3.1. Next, determine the ratio of the CFP for a crack that has just formed a perceptible leak (Ppl) to the CFP for a crack leaking at the technical specification limit assuming (Ptsl) a SLB load. Also provide the 90% coverage interval of this ratio.

5A.3.3. Again, consider a single piping system and degradation mechanism combination identified for Category 1 PWR LOCAs in 5.A.3.1. Next, determine the ratio of the conditional failure probability for a crack with a maximum depth of 50% of the wall thickness (P50) to the CFP for a crack that has just formed a perceptible leak (Ppl) assuming a SLB load. Also provide the 90% coverage interval of this ratio.

5A.3.4. Repeat 5A.3.1 — 5A.3.3 for each significant piping system/degradation mechanism combination listed for PWR Category 1 LOCAs in 5A.1.

5A.3.5. Repeat 5A.3.1 — 5A.3.4 for each PWR LOCA size category.

5A.3.6. Repeat 5A.3.1 — 5A.3.5 for BWRs.

Notes:

a. The leaking crack size is a function of the degradation mechanism and is the major contributor to the differences with the base-case conditional failure probabilities.

b. A perceptible leak is a leak which has just initiated.

5A.4.1. Again consider a single piping system and degradation mechanism combination identified for Category 1 PWR LOCAs in 5.A.1. Next, determine the ratio of the CFP for a SLD event (Ptsl@sld) to the CFP for a SLB event (Ptsl@slb) assuming a crack leaking at the technical specification limit in both cases. Also provide the 90% coverage interval of this ratio.

5A.4.2. Consider the same piping system and degradation mechanism combination identified for

Category 1 PWR LOCAs in 5.A.4.1. Next, determine the ratio of the CFP for a crack that has just formed a perceptible leak (Ppl) to the CFP for a crack leaking at the technical specification limit (PTSL or Ptsl@sld) assuming a SLD load. Also provide the 90% coverage interval of this ratio.

5A.4.3. Again, consider a single piping system and degradation mechanism combination identified for Category 1 PWR LOCAs in 5.A.4.1. Next, determine the ratio of the conditional failure probability for a crack with a maximum depth of 50% of the wall thickness (P50) to the CFP for a crack that has just formed a perceptible leak (Ppl) assuming a SLD load. Also provide the 90% coverage interval of this ratio.

5A.4.4. Repeat 5A.4.1 — 5A.4.3 for each significant piping system/degradation mechanism combination listed for PWR Category 1 LOCAs in 5A.1.

5A.4.5. Repeat 5A.6.1 — 5A.6.4 for each PWR LOCA size category.

5A.4.6. Repeat 5A.6.1 — 5A.6.5 for all BWRs.

a. If your system and degradation mechanism list in 5A.1.2 for SLD loading is different from that in 5A.1.1 for SLB loading, pick a seismic base for reference in 5A.4.1 instead of referencing with respect to the SLB loading magnitude.

J.2.5.1 Estimation of Piping Damage Likelihood: Now consider the relative likelihood of the occurrence of a particular level of damage (50% through-wall, perceptible leak, tech. spec. leakage) due to the piping system/degradation mechanism combination chosen in 5A.1. All answers will be ultimately referenced to a piping base-case damage probability as in earlier questions. However, there are no numbers assigned to the base-case damage probabilities at this time, so the comparisons should be made with respect to a piping base-case damage condition. A separate piping base-case condition is defined for each piping system and LOCA size category identified in 5A.1, as the operational experience frequency of all leaks regardless of the degradation mechanism. This frequency will be determined from operational experience data.

5A.5.1 Consider a single piping system and degradation mechanism combination identified for

Category 1 PWR LOCAs in 5.A.1. Next, determine the ratio of the likelihood of a pipe having a perceptible leak due to that degradation mechanism in that piping system (Lpl) after 25 years of operation (Lpl) to the base case (LBC), which is the likelihood of a leak due to any degradation mechanism. Also provide the 90% coverage interval for this estimate.

5A.5.2 Consider the same single piping system and degradation mechanism as in 5A.5.1. Next,

determine the ratio of the likelihood of a technical specification leak (LTSL) to a perceptible leak (LPL) due to that degradation mechanism after 25 years of operation. Also provide the 90% coverage interval for this estimate.

5A.5.3 Consider the same single piping system and degradation mechanism as in 5A.5.1. Next,

determine the ratio of the likelihood of a 50% through-wall leak (L50) to a perceptible leak (LPL) due to that degradation mechanism after 25 years of operation (current-day). Also provide the 90% coverage interval for this estimate.

5A.5.4 Now determine if you believe the relative likelihood ratios in 5A.5.1 — 5A.5.3 will increase, decrease, or remain constant with continued operating time. First consider all three likelihood estimates (Lpl/Lbc, Ltsl/Lpl, and L50/Lpl) at 40 years and then 60 years of continued operation. Determine the ratio of these estimates at 40 years of operation to the current-day estimates. Next, determine the ratio these estimates at 60 years of operation to current-day estimates.

5A.5.5 Repeat 5A.5.1- 5A.5.4 for each significant piping system/degradation mechanism combination listed for PWR Category 1 LOCAs in 5A.1.

5A.5.6 Repeat 5A.5.1 — 5A.5.5 for each PWR LOCA size category.

5A.5.7 Repeat 5A.5.1 — 5A.5.6 for all BWRs.

The surge line elbow result identified as “axisymmetric nonseismic” in Table F.12 is suggested as the reference case. Table F.13 summarizes the cumulative results for the larger flow rates, which were obtained by the alternative procedure.

Two of these highly stresses elbows are considered to be present in the surge line system

F.4.3 HPI Makeup Nozzle

Probability analyses were performed with and without failure of the thermal sleeve, which has been observed to fail in service. The least favorable large leak probabilities were for a failed thermal sleeve, which immediately resulting in fatigue crack initiation, but with the same stresses as before. This is suggested as the reference case, with the column labeled cu = 0 in Table F.16 being the results of interest.

Three such locations are considered to be present in the system.

Operating experience was applied as the best method for estimating frequencies for more common failure events such as small detectable leakage and of ruptures for small diameter piping. Operating experience has the advantage of reflecting contributions from all degradation mechanisms and is not limited to a particular mechanism that can be addressed by a PFM model. For lower frequency events, for which there are little or no data from operating experience, the data were therefore supplemented by trends from PFM models. The fracture mechanics models were taken to provide relative frequencies such as for (for a given pipe size) the ratios of frequencies of for different categories of failures (in terms of leak-rates). Similarly, models can indicate ratios for one failure category of leak for differing pipe sizes. Reports of small detectable leakage (from data bases) were taken to be precursor events that can be used to calibrate estimates of frequencies for categories of larger leakage events. Another consideration was that only the larger pipe sizes contribute to the frequencies for larger leak rates. It was implicitly assumed that contributions from smaller pipe sizes dominate for the smaller categories of leak rate frequencies.

Non-nuclear experience was also used to support estimates of failure frequencies for nuclear components. Component designs, materials, construction codes, operating conditions etc. are much the same for nuclear applications and as for many non-nuclear applications. Non-nuclear experience however provides a much larger number of years of plant operation (by orders of magnitudes) than available from nuclear experience. Non-nuclear experience therefore provides additional justification for very low failure frequencies for components such as pump bodies, tube sheets, manways, etc. that imply large extrapolations from the limited years of nuclear plant operation.

D.6.4.2 Speculative LOCA Frequency at T = 40 & T = 60 — A retrospective evaluation is performed through a Bayesian update process whereby the exposure term in Equation 4.1 is modified to account for the longer exposure time. The analysis is performed by assuming that the service history at T = 40 and T = 60 years is known; zero (0) weld failures during the intervals AT = 15 years (T40 to T25) and AT = 35 years (T60 to T25). This is a purely speculative assumption implying that the ISI/NDE technologies and other piping reliability management programs remain at least as effective as at the present and that no unexpected material aging occurs. The extrapolated LOCA frequencies are summarized in Table D.20. Under the given assumptions the LOCA frequency would be expected to decrease with time.

As with many reliability databases, the completeness of the data is always an issue. While relative frequencies (e. g., percentage distribution of events by component or degradation mechanism) might be reasonably accurate, the accuracy of any absolute frequency (e. g., events per year) calculations will depend directly on how complete the data are. That is, have all events that have occurred been included in the database? In the current situation this question has two parts. First, have all relevant event been discovered? Second, of the discovered events, have they been reported (via LER)?

The completeness issue is probably more of an issue for the partial through-wall cracks than it is for the more severe failures. There are two causes for this concern. One is ambiguity in the interpretation of the LER reporting requirement (Attachment A), and the second and probably primary cause is simple lack of detection. While effort is made to make the LER reporting requirements as clear as possible, the “seriously degraded” aspect of 50.73(a)(2)(ii)(A) is difficult to quantify. How far does a crack have to extend to seriously degrade the primary pressure boundary? It is possible that some cracks are being detected and repaired (which might be considered normal plant maintenance rather than corrective action), without being reported as a LER. These events have not been captured in this search. However, detection likelihood is probably a bigger reason for coverage deficiencies of part-through wall flaws. A leak (or a non-leaking through-wall crack) is simply more likely to be detected. This issue is clearly illustrated by events in the data in which a detected leak prompted the plant to do a thorough inspection that found partial through-wall cracks. If the leak had not occurred and motivated the inspection, the partial through-wall cracks would not have been found.

Submitted by Nuclear Energy Institute (NEI)

Comment: The report should provide a discussion on probabilistic validation of the small LOCA frequency. Using a Poisson distribution with failure rate of 2.9E-03 (NUREG-1829 Category 1 LOCA frequency, excluding steam generator tube rupture events), and considering the approximately 2,500 reactor-years of operation experience, the probability of no small LOCA events (actual industry performance) is around 1 percent. This result shows an excessive conservatism in the category 1 LOCA frequency estimation of NUREG-1829.

Response: A new section (Section 7.10) has been added to the revised NUREG which compares the small break LOCA frequency estimates from the elicitation with operating experience. As discussed in this section, the elicitation estimates for BWR and PWR small break LOCA frequencies are generally consistent with the historical data. See the response to Comment 7-8. Related insights are also provided in the responses to Comments GC3, GC4, GC5, GC6, and 7-8.

Two additional points should be noted. First, the 2,500 reactor-years of operating experience referred to in the comment clearly combines BWR and PWR plants. The NUREG provides separate estimates for BWRs and PWRs and treats their operating experience separately. The number of reactor-years used in Section 7.10 is 1,023 for BWRs and 1,986 for PWRs, all as of December 2006. Second, as discussed in Section 7.10, there has been one pipe break in a PWR plant that exceeded the LOCA Category 1 flow rate threshold defined in this study. Additionally, a second PWR pipe break resulted in flows that were close to the Category 1 threshold. The comment claims that there have been no non-SGTR Category 1 events.

The total LOCA frequencies calculated above are for Category-1 LOCAs. The simple approach taken here is that the LOCA frequency is reduced by У2 order of magnitude (assuming a lognormal distribution), for each step up in size category. There are a number of reasons for this approach. Between the smallest pipe size categories (i. e., < 2 inches, and > 2 inches) there is a significant difference in the failure mechanisms. For the smallest pipes, the operating experience includes failures of compression fittings and socket welds, which are not used in larger size pipe. Also, a number of studies on crack and leak events indicate a decrease in these precursor frequencies, as pipe diameter increases. Lastly, virtually every estimate of LOCA frequencies ever made has resulted in a reduced frequency for the large LOCA sizes.

E. 6 LOCA Frequency by Degradation Mechanism and Subsystem

In addition to the LER data used to allocate the LOCA frequency between pipe and passive non-pipe components, the LOCA frequency was further allocated among the different degradation mechanisms observed and among the different RCS subsystems and components defined for this project. These allocations were based on data collected from both U. S. reactor operating experience (primarily LERs), and from foreign LWR operating experience (SLAP database). One complication to this approach is the IGSCC-related experience in U. S. BWR plants. IGSCC was an issue for BWRs in the early 1980’s.

Many U. S. BWRs implemented IGSCC mitigation programs in the mid-1980’s, which have greatly reduced the occurrence of IGSCC. To avoid unrealistically over weighting the IGSCC mechanism, the

BWR experience was segregated and only the post 1985 experience was used for allocating the relative contribution to LOCA frequency by degradation mechanism.

Lastly, although the guidance for calculating base-case frequencies for this project included estimates for 25 years, 40 years and 60 years, this particular base-case calculation assumed that the frequencies were generally independent of plant life. This is based on the IGSCC experience that demonstrated that although degradation mechanisms are at work that can result in an increase in the LOCA frequency over time, so to are mitigation programs and general performance improvement programs (e. g., more effective inspections), that can result in a decrease in the LOCA frequency. Therefore, overall these competing effects are assumed to cancel each other out for a net zero effect on LOCA frequency. That is, the current LOCA frequency (approximately 25 year life) is assumed to be application for 40 and 60 years as well.

An emergency faulted load represents an initial design consideration for a large transient load that was not expected to occur over any particular plant’s operating life of 40 years (rare event). These loads could be due to seismic loading or any other large pressure transients. Similar to the piping evaluation, base cases will be used for anchoring on the conditional failure probability. However, the actual base cases will not be developed until after the panelists’ identify the non-piping components which provide significant LOCA contributions. In the interim, each panelist should use a particular set of base case conditions for anchoring. More information on this selection will follow in Elicitation Question 6A.2.

This question will ask you to list and quantify the effect of the most significant non-piping systems and degradation mechanisms that contribute to each LOCA category. The quantification will be done for two emergency faulted load sizes (SLB and SLD) for three assumed damage states. The damage states will

consist of tech. spec. leakage rates, the onset of leakage through a slow drip (perceptible leak), and a surface crack with a/t = 0.5. The surface crack length will be assumed by each panelist and may be a function of degradation mechanism and material. The likelihood of each damage state will also be ascertained by each panelist relative to the operational experience data for the leak-rate frequencies corresponding to each non-piping component listed, regardless of degradation mechanism. This assessment will require nine different relative comparisons for each LOCA size category and plant type (BWR or PWR).

The structure of this question is almost identical to Elicitation Question 5. The appendix contains information on the philosophy behind these two questions.

6A. 1.1. List the non-piping components and degradation mechanisms (or failure scenarios) which most significantly contribute to Category 1 LOCAs given that an assumed emergency faulted load occurs with an equivalent SLB magnitude for PWRs. This total list should summarize at least the top 80% contributing factors to Category 1 LOCAs under assumed faulted loading conditions. Also, for each component, list the loads which may result in SLB loading and indicate if these loads are primary (load-controlled) or secondary (displacement-controlled). Provide the total contribution and also the 90% coverage interval for this estimate.

6A. 1.2. Repeat 6A. 1.1 for Service Level D (SLD) loading

6A. 1.3. Repeat 6A. 1.1 and 6A. 1.2 for each PWR LOCA size category.

6A. 1.4. Repeat 6A. 1.1 — 6A. 1.3 for BWR non-piping components.

Notes:

a. Information on relevant non-piping components and degradation mechanisms, are contained in the “Elicitation Meeting Notes” from the kick-off meeting and subsequent revisions to Tables B.1.13 — B.1.17 in this document.

b. In this question, pick your non-piping component assuming that it will completely fail. Therefore, the LOCA size category will be directly related to the component size.

6A.2.1. Pick a representative set of seismic base-case conditions to use for comparison for each of your significant contributors to Category 1 LOCAs in PWRs.

6A.2.2. Repeat 6A.2.1 for each PWR LOCA size category.

6A.2.3. Repeat 6A.2.1 and 6A.2.2 for BWR non-piping components.

Notes:

a. The base case conditions should correspond to at least one (or several, or all) of the significant non-piping LOCA contributors identified in 6A.1. Assume that the component is damaged with a fatigue flaw which results in technical specification leakage. Assume that the base case loading magnitude is an SLB load. Assume that absolute size of this flaw and the actual conditional failure probability to a SLB load magnitude will be quantified at a later date.

b. Comparisons to the selected base cases will be made in subsequent questions.

6A.3.1. Consider a single non-piping component and degradation mechanism combination identified for Category 1 PWR LOCAs in 6.A.1 and the associated seismic base case identified in 6A.2. Determine the ratio of the CFP for this system/degradation mechanism combination (PTSL or Ptsl@slb) to the CFP for the chosen seismic non-piping base case assuming (PBC) that an SLB emergency faulted load occurs and that the non-piping component contains a through-wall crack that is leaking at the technical specification limit. Also provide the 90% coverage interval of this ratio.

6A.3.2. Consider the same non-piping component and degradation mechanism combination identified for Category 1 PWR LOCAs in 6.A.3.1. Next, determine the ratio of the CFP for a crack that has just formed a perceptible leak (PPL) to the CFP for a crack leaking at the technical specification limit (PTSL) assuming a SLB load magnitude. Also provide the 90% coverage interval of this ratio.

6A.3.3. Again, consider the single non-piping component and degradation mechanism combination identified for Category 1 PWR LOCAs in 6.A.3.1. Next, determine the ratio of the CFP for a crack with a maximum depth of 50% of the wall thickness (P50) to the CFP for a crack that has just formed a perceptible leak (PPL) assuming a SLB load. Also provide the 90% coverage interval of this ratio.

6A.3.4. Repeat 6A.3.1 — 6A.3.3 for each significant non-piping component/degradation mechanism combination listed for PWR Category 1 LOCAs in 6A.1.

6A.3.5. Repeat 6A.3.1 — 6A.3.4 for each PWR LOCA size category.

6A.3.6. Repeat 6A.3.1 — 6A.3.5 for BWR non-piping components.

Notes:

a. The leaking crack size is a function of the degradation mechanism and is the major contributor to the differences with the base-case conditional failure probabilities.

6A.4.1. Again consider a single non-piping component and degradation mechanism combination as identified for Category 1 PWR LOCAs in 6.A. 1. Next, determine the ratio of the CFP for a SLD event (PTsl@sld) to the CFP for a SLB event (PTsl@slb). Assume that a crack exists which is leaking at the technical specification limit in both cases. Also provide the 90% coverage interval of this ratio.

6A.4.2. Consider the same non-piping component and degradation mechanism combination identified for Category 1 PWR LOCAs in 6.A.4.1. Next, determine the ratio of the CFP for a crack that has just formed a perceptible leak (PPL) to the CFP for a crack leaking at the technical specification limit (PTSL or PTSL@SLD). Assume a SLD loading magnitude in each case. Also provide the 90% coverage interval of this ratio.

6A.4.3. Again, consider the same non-piping component and degradation mechanism combination

identified for Category 1 PWR LOCAs in 6.A.4.1. Next, determine the ratio of the CFP for a crack with a maximum depth of 50% of the wall thickness (P50) to the CFP for a crack that has just formed a perceptible leak (PPL). Assume a SLD loading magnitude in each case. Also provide the 90% coverage interval of this ratio.

6A.4.4. Repeat 6A.4.1 — 6A.4.3 for each significant non-piping component/degradation mechanism combination listed for PWR Category 1 LOCAs in 5A.1.

6A.4.5. Repeat 6A.6.1 — 6A.6.4 for each PWR LOCA size category.

6A.4.6. Repeat 6A.6.1 — 6A.6.5 for all BWR non-piping components.

Notes:

a. If your system and degradation mechanism list in 6A. 1.2 for SLD loading is different from that in 6A.1.1 for SLB loading, pick a seismic base for reference in 6A.4.1 instead of referencing with respect to the SLB loading magnitude.

J.2.6.1 Estimation of Piping Damage Likelihood: Now consider the relative likelihood of the occurrence of a particular level of damage (50% through-wall, perceptible leak, tech. spec. leakage) due to the non-piping component/degradation mechanism combination chosen in 6A.1. All answers will be ultimately referenced to a non-piping base-case damage probability. However, there are no numbers assigned to the non-base-case damage probabilities at this time. Comparisons should therefore be made with respect to a non-piping base-case damage condition. A separate non-piping base-case condition is defined for each non-piping component identified in 6A.1, as the operational experience frequency of all component leaks regardless of the degradation mechanism.

6A.5.1 Consider a single non-piping component and degradation mechanism combination identified for Category 1 PWR LOCAs in 6.A. 1. Next, determine the ratio of the likelihood of the non­piping component having a perceptible leak after 25 years of operation (LPL) due to that degradation mechanism to the base case (LBC), which is the likelihood of a leak due to any degradation mechanism. Also provide the 90% coverage interval for this estimate.

6A.5.2 Consider the same non-piping component and degradation mechanism as in 6A.5.1. Next,

determine the ratio of the likelihood of a technical specification leak (Ltsl) to a perceptible leak

(Lpl) due to that degradation mechanism after 25 years of operation. Also provide the 90% coverage interval for this estimate.

6A.5.3 Consider the same single non-piping component and degradation mechanism as in 6A.5.1.

Next, determine the ratio of the likelihood of a 50% through-wall leak (L50) to a perceptible leak (LpL) due to that degradation mechanism. Also provide the 90% coverage interval for this estimate.

6A.5.4 Now determine if you believe the relative likelihood ratios in 6A.5.1 — 6A.5.3 will increase, decrease, or remain constant with continued operating time. First consider all three likelihood estimates (LpL/Lbc, LtsL/LpL, and L50/LpL) at 40 years and then 60 years of continued operation. Determine the ratio of these estimates at 40 years of operation to the current-day estimates. Next, determine the ratio these estimates at 60 years of operation to current-day estimates.

6A.5.5 Repeat 6A.5.1- 6A.5.4 for each significant non-piping component/degradation mechanism combination listed for PWR Category 1 LOCAs in 6A.1.

6A.5.6 Repeat 6A.5.1 — 6A.5.5 for each PWR LOCA size category.

6A.5.7 Repeat 6A.5.1 — 6A.5.6 for all BWR non-piping components.

Analyses were performed for this component for a range of applied stresses, with predictions compared to field experience of leaks and observed part-through cracks. Analyses were performed for no remedial action, and for a weld overlay at 20 years. The weld overlay at 20 years is considered to be the most realistic. Comparisons with experience led to an estimate of stresses that were considerably below the peak value used in the original analysis. However, when compensated for the number of weld joints involved, the system leak frequencies were nearly the same whether 49 joints with a random stress (mean cNO = 83 MPa [12 ksi]) or 2 joints with a high stress (cNO = 140 MPa [20 ksi]) were considered (see Table F.25). The case of weld overlay at 20 years with the high stress is recommended as the reference case. Table F.19 contains the cumulative results.

Two of the highly stresses joints are considered to be present in the recirculation system.

F.4.5 Recirculation Line — 28 inch

The recirculation line with no remedial action and a high stress representing the dominant joints was the only case considered, and is summarized in Table F.24.

Two such joints are considered to be present in the system.

F.4.6 Feedwater Elbow

Case C in Table F.26 is suggested as the reference case. Results for > 380 lpm (100gpm) and larger were generated by the alternative procedure, and are summarized in Table F.28.

Four such locations were considered to be present in the system.

F.4.7 Summary Table

Table F.29 provides an overall summary of the leak flow rate frequencies for the reference cases of the base case systems.

Table F.29 Summary of Results for Reference Systems

times in reactor years, 1 calendar year ~ 0.8 reactor years shaded areas are estimates based on alternative procedure leak rates in thousands of gallons per minute

cross-hatched cells are beyond maximum leak capability for that pipe size ** also applicable to > 1,900,000 lpm (500 kgpm) for hot leg if sufficient diameter

F.5 References

F.1. D. O. Harris, E. Y. Lim and D. Dedhia, Probability of Pipe Fracture in the Primary Coolant Loop of a PWR Plant, Vol. 5: Probabilistic Fracture Mechanics Analysis, U. S. Nuclear Regulatory Commission Report NUREG/CR-2189, Vol. 5, Washington, D. C., August 1981

F.2. D. O. Harris, D. Dedhia, E. D. Eason and S. D. Patterson, Probability of Failure in BWR Reactor Coolant Piping: Probabilistic Treatment of Stress Corrosion Cracking in 304 and 316NG BWR Piping Weldments, U. S. Nuclear Regulatory Commission Report NUREG/CR- 4792, Vol. 3, Washington, D. C., December 1986

F.3. D. O. Harris, D. Dedhia and S. C. Lu, Theoretical and User’s Manual for pc-PRAISE, A Probabilistic Fracture Mechanics Code for Piping Reliability Analysis, U. S. Nuclear Regulatory Commission Report NUREG/CR-5864, Washington, D. C., July 1992

F.4. D. O. Harris and D. Dedhia, WinPRAISE: PRAISE Code in Windows, Engineering Mechanics Technology, Inc. San Jose, California, Technical Report TR-98-4-1, 1998

F.5. M. A. Khaleel, F. A. Simonen, H. K. Phan, D. O. Harris and D. Dedhia, Fatigue Analysis of Components for 60-Year Plant Life, U. S. Nuclear Regulatory Commission Report NUREG/CR — 6674, Washington, D. C., June 2000

F.6. A. Deardorff, D. Harris and D. Dedhia, Materials Reliability Program: Re-Evaluation of Results in NUREG/CR-6774for Carbon and Low-Alloy Steel Components ", Electric Power Research Institute Report 1003667, Palo Alto, California, 2002

F.7. J. Keisler, O. K. Chopra and W. J. Shack, Fatigue Strain-Life behavior of Carbon, Low-Alloy Steels, Austenitic Stainless Steels, and Alloy 600 in LWR Environments, U. S. Nuclear Regulatory Commission Report NUREG/CR-6335, Washington, D. C., 1995

F.8. O. K. Chopra and W. J. Shack, Effects of LWR Coolant Environments on Fatigue Design Curves of Carbon and Low-Alloy Steels, U. S. Nuclear Regulatory Commission Report NUREG/CR-6583, Washington, D. C., March 1998

F.9. Technical Elements of Risk-Informed Inservice Inspection Programs for Piping, U. S. Nuclear Regulatory Commission Draft Report NUREG-1661, Washington, D. C., January 1999

F.10. M. A. Khaleel, O. J.V. Chapman, D. O. Harris and F. A. Simonen, “Flaw Size Distribution and Flaw Existence Frequencies in Nuclear Piping”, Probabilistic and Environmental Aspects of Fracture and Fatigue, ASME PVP-Vol. 386, 1999, pp. 127-144

F.11. ASME Boiler and Pressure Vessel Code, Section XI, Appendix C, 1992

F.12. P. Ricardella, “Probabilistic Fracture Mechanics Analysis of CRDM Nozzles”, presented at ACRS Meeting, Rockville, Maryland, June 5, 2002

F.13. e-mail from Gery Wilkowski to David Harris, “Material Property Inputs for Base Cases”, June 10, 2003

F.14. Personal communication, Art Deardorff, Structural Integrity Associates, San Jose, California, to David Harris, Engineering Mechanics Technology, Inc., San Jose, California

F.15. T. C. Chapman, et al., Assessment of Remedies for Degraded Piping, Electric Power Research Institute Report NP-5881-LD, Palo Alto, California, 1988

F.16. B. O.Y. Lydell, An Application of the Parametric Attribute/Influence Methodology to Determine Loss of Coolant Accident (LOCA) Frequency Distributions, Document No. R2003-02, May 2003, provided to members of the NRC LOCA Frequency Expert Elicitation Panel.

F.17. Attachment (Action Item 45R1.xls) to e-mail from Bengt Lydell to base case panel members, June 20, 2003

F.18. A. G. Ware, D. K. Morton and M. E. Nitzel, Application of NUREG/CR-5999 Interim Fatigue Curves to Selected Nuclear Power Plant Components, U. S. Nuclear Regulatory Commission Report NUREG/CR-6260, Washington, D. C., 1995

The piping non-seismic LOCA evaluations were conducted for PWR and BWR piping separately from a bottom-up approach using reference cases for certain pipe systems in each type of plant. The reference cases were determined from a combination of the base case results supplied to the elicitation panel. The base cases supplied consisted of two independent PFM analyses, and two independent operating — experience evaluations for certain pipe systems. The probabilistic pipe fracture mechanics analyses (PRAISE or PRODIGAL) base cases were not chosen since Dr. Wilkowski did not believe that those computer codes properly determined the probability of a long surface crack occurring, which is the actual way that a LOCA would occur, i. e., a through-wall leaking crack will be readily discovered by leakage before failing at normal operating conditions. Consequently, the two historical base cases were averaged, but only up to 25 years of operation (current time period). Dr. Wilkowski did not believe that the historical based approaches would be that good for extending the LOCA reference cases to 40 or 60 years. Consequently, his reference cases were only for 25-year time period (present), and the 40- and 60-year evaluations were adjusted depending if he thought the particular pipe system would be susceptible to some near term or long-term degradation mechanism (e. g., PWSCC), and if that mechanism could produce a large surface crack. These evaluations were done for 12 different PWR pipe systems and 13 different BWR pipe systems, with six different LOCA flow-rate categories. The uncertainties (5 and 95 percent bounds) in the predictions generally increased as the amount of time increased, i. e., the uncertainty for 25 years was less than 40 years, and the uncertainty in the 40-year predictions was less than for the 60-year predictions.

For non-piping, Dr Wilkowski felt he did not know enough about failure modes of all the different categories of non-piping components (with the exception of a few categories like CRDM nozzle ejection). He therefore chose the steam generator tube historical failure frequencies for small LOCA as a controlling PWR case, but all the other cases were governed by the piping failure probabilities.

APPENDIX L
DETAILED RESULTS