The effect of random error, which inevitably occurs in measurement of stable crack growth, on estimation of fracture toughness obtained by means of the multispecimen technique, is analyzed. Using the Monte Carlo simulation method, statistical distributions of the toughness parameters are generated, as well as the statistical distribution of the lower bounds as estimated by the model of simple regression. Two examples are analyzed, which are representative of the majority of cases which occur in practice, pointing out the dispersions which occur in toughness values as a consequence of errors in measurement of stable crack growth. © 1992.

Peer Reviewed