Deterministic fracture mechanics analysis often involves computing critical crack size or remaining life of a component subjected to cyclic or steady state stresses. Since many of the inputs needed to perform the analysis have considerable scatter, conservative input values are employed to estimate critical crack size or remaining life. The final results that are obtained using such methods are necessarily conservative and sometimes overly conservative.
Probabilistic Fracture Mechanics (PFM) overcomes this limitation by considering the variables with scatter as distributed random variables. Rather than pass/fail, the result is a probability of certain events occurring; for example, the probability of the critical crack size being reached.
As plants continue to age and units consider long-term operation, opportunities to apply probabilistic techniques are increasingly promising, especially in light of recent regulatory acceptance.