Probabilistic Analysis for Characteristic Values of Fracture Toughness

The statistical significance for the method of minimum of three equivalent（MOTE） was revealed on the uncertainty of fracture toughness data, using the probability theory. Large numbers of fracture toughness data sets sampled from a parent distribution, which were fitted by 41 fracture toughness values, were used to simulate the test data. Then the characteristic values of fracture toughness for every data set were determined with the MOTE method. At last the failure probability of a typical cylinder containing surface crack was analyzed for every characteristic value, and compared with ASME low bound. The results show that the confidence level is relatively low to estimate a low quantile of parent distribution. The result of assessment will be more reliability with much more test data. However, the confidence level is only 81.5% to ensure that the characteristic value of fracture toughness from the MOTE method is lower than ASME low bound, even if the number of test data exceeds three.