Many design scenarios of components made of polymer materials are concerned with notches as representative constructive details. The failure hazard assessment of these components using models based on the assumption of cracked components leads to over-conservative failure estimations. Among the different alternative approaches proposed that are based on the apparent fracture toughness, KcN is considered. In so doing, the current deterministic underlying concept must be replaced by a probabilistic one to take into account the variability observed in the failure results in order to ensure a reliable design. In this paper, an approach based on the critical distance principle is proposed for the failure assessment of notched EPOLAM 2025 CT samples with each different notch radii (ρ) including a probabilistic assessment of the failure prediction. First, each apparent fracture toughness is transformed into the equivalent fracture toughness for ρ=0 based on the critical distances theory. Then, once all results are normalized to the same basic conditions, a Weibull cumulative distribution function is fitted, allowing the probability of failure to be predicted for different notch radii. In this way, the total number of the specimens tested in the experimental campaign is reduced, whereas the reliability of the material characterization improves. Finally, the applicability of the proposed methodology is illustrated by an example using the own experimental campaign performed on EPOLAM 2025 CT specimens with different notch radii (ρ).
Keywords: failure probabilistic prediction; notched components; theory of critical distances.