Ratios of non-piping to piping for various category LOCAs are for 25 years only.

Pete made the point that most of the plots that Rob has shown are for 25 years, while he thought 40 and 60 years more important since 25 years is in past and the associated problems have been addressed while 40 and 60 years are for future; Rob responded that not that much difference between 25 and 40 years with some effect for certain category LOCAs at 60 years.

Sam commented that he was somewhat surprised that non-piping contribution less than piping contribution for BWRs in that piping has some active mechanisms that have been successfully mitigated in past; Karen responded that non-piping components more robust.

Bill Galyean warned about combining group distributions and panel distributions that may introduce a bias in that various members of group defined boundaries of system differently.

Much discussion on whether to chose group median or individual medians; Rob and Lee haven’t done panel distributions yet.

Comparison of BWR and PWR — effect of mitigation encompassed in results for BWRs, but not PWRs — BWRs have been doing mitigation for 15 to 20 years whereas PWRs are just starting with mitigation for PWSCC.

Inclusion of S/G tube rupture for Category 1 LOCAs will be problematic for some people in that PRA people aren’t used to accounting S/G tube failures in with rest of data; typically S/G tube rupture data is presented separately; in future Rob will present data both with S/G tube rupture data and without.

Categories 1, 2, and 3 are historically same as small, medium, and large break LOCAs respectively.

For PWR MB LOCAs, major contributor is CRDM, not S/G tube ruptures.

Discussion of whether MB LOCA was Category 2 or Category 3; some thought that MB LOCA was more in line with Category 3 LOCA.

Bengt felt we are comparing apples and oranges as we try to compare our results with historical results; NUREG/CR-5750 didn’t look at non-piping per se whereas we did, although Bill Galyean indicated that if there had been indications of TWC in non-piping components then he would have included that data in his analysis in 5750; some thought that due to apples and oranges nature of our approach with 5750 that we shouldn’t present these comparisons but Rob argued that if we don’t present these comparisons then others will; Some argued that we should present frequencies for multiple LOCA categories when we compare with small, medium, and large break LOCAs for 5750

Lee reviewed the feedback questionnaire.

Day 3: February 12, 2004

Presentation 17: Emergency and Faulted Loading: Elicitation Approach and Responses

Water hammer type loadings should be in normal operating loading history.

What we are asking panel members to estimate is only the conditional failure probability given a stress with magnitude i (PL/Si)

Ken commented that the seismic anchor motion (SAM) stress which is a secondary stress may be a bigger contributor than some primary stresses such as inertial stresses.

On VG entitled Elicitation Requirements, we are asking panel to do first bullet, we will do 2nd bullet, and plants would do 3rd and 4th bullets.

Asking them for a given system and degradation mechanism for their estimate of L50, P50, Ppl, Lpb Ltsl, and Ptsl and then we will interpolate to get entire curve.

Some people argued during their elicitations that non-piping and large piping are non contributors to LOCA frequency due to seismic; we will look at results from piping and then decide what and if we will do anything for non-piping considerations.

Bruce argued that can get some very high loads due to malfunctions of snubbers.

Pbc and Lbc are probabilities for base case and likelihood of base case