A statistical approach for transferring fracture events across different sample shapes

The paper investigates the capability of a novel calibration method to predict accurately fracture events across different sample shapes at low temperatures. It is shown that the emergence of a threshold Weibull stress in the Weibull stress distribution is inherent in the fundamental assumptions of the Beremin model. The mathematical concept underlying the suggested calibration method is the correlation between the probability distributions of the fracture loads and the associated Weibull stresses. The calibration procedure is demonstrated using fracture data obtained in tests conducted at a test temperature of −150 °C on specimens fabricated of A533B ferritic steel. In contrast to the values found in the literature, the calibrated Weibull modulus is small and ranges from 2 to 4. The proposed methodology is straightforward to apply and yields reliable predictions of the failure probabilities of samples of different shapes.