Analysis and improvement of local cleavage fracture models at elevated loading rates

Abstract

Brittle fracture has to be excluded in safety-relevant components such as nuclear reactor pressure vessels. Recent work has indicated that brittle failure under dynamic loading cannot be assessed reliably by macroscopic methods such as the Master Curve concept (ASTM E1921). This finding is strongly connected to adiabatic heating processes and local crack arrest events in the crack tip region. However, a detailed study involving the assessment at elevated loading rates with local cleavage fracture models is not available to this point. The present work deals with the application of a local model to various dynamic loading conditions, as well as a proposed model adjustment to improve the analysis results. It was found that existing local concepts are not capable of describing fracture behavior adequately. Though the explicit numerical consideration of heat generation and conduction at elevated loading rates allows a better estimation of failure compared to the Master Curve concept, this still poor assessment quality can also be obtained by simply utilizing the adjusted “dynamic” Master Curve with an increased exponent of p = 0.030 /°C. It is therefore concluded that this exponent adjustment approximates the impact of heat effects quite well. The shortcomings of the numerical model were observed to correlate strongly with the documented amount of local crack arrest events on the fracture surface. This leads to a micromechanically-motivated adjustment of a local approach model to incorporate this mechanism. This modification ensures a more satisfying physical background of the local approach model. Moreover, it significantly improves model analysis accuracy for all examined testing conditions.