When estimating RR-PRODIGAL leak rates through the final through wall defect in a pipe weld, evaluations were made using an elastic crack opening displacement (COD) analysis. However, it was felt that the uncertainties associated with assessing both the defect length around the pipe circumference as well as the COD needed for estimating the flow rate through the crack, were too great and too subject to ongoing development, to allow a suitable analysis of the leak rate. Thus, RR — PRODIGAL does not contain, within itself, a routine for evaluating the flow rate from the final defect size.

Instead, it was concluded that the leak rate from a through wall defect could be considered independently of the probability of the breach, i. e. the leak rate from the defect is not dependent on the probability of the defect cracking through the pipe wall. Note, however, that the COD, crack length, and hence leak rate is not independent of the mechanism that led to the failure, only the probability of the failure itself.

Within the Naval Nuclear program, computer programs have been developed to assess the leak rate from different defects based primarily on the ‘SQUIRT’ model. However, for consistency within this program, the data on leak rate against defect area provided by the USNRC were used, as shown in Figure G.1.

G. 3 Procedure

The procedures used to develop the base case numbers are as follows:

1 Evaluate the basic fatigue failure probability using RR-PRODIGAL code using the transient data supplied[16].

2 Evaluate an elastic COD as a function of defect size.

3 Use expert judgement to extend this COD beyond the elastic limit.

4 Evaluate a mean defect cross-sectional area for a given defect size using its associated COD.

5 Evaluate the mean leak rate from a given defect size using the data supplied by the USNRC, see Figure G.1.

[Note for Steps 2, 3, 4 and 5 above a defect length is given. Thus, Steps 2, 3, 4 and 5 provide a

mapping from a given defect size at failure to the mean leak rate in gpm, given this defect exists.]

6 Use expert judgement to assess the distribution of the defect length at failure.

7 Combine Steps 5 and 6 to obtain the conditional probability of a leak rate greater than the given leak rates for Categories 1 through 6. These categories being as follows;

Table G.1 Leak Category Leak Rates

8 Combining the conditional probability of Step 7 with the basic fatigue failure probability in Step 1 gives the required final probability of a leak greater than each of the categories.