Abstract
In order to guarantee the presence of adequate margin against ultimate fracture, it is essential to establish a reliable methodology for estimating ductile fracture behavior of structures. Following the development of equations for describing true stress–strain relation and their application to notched bar and grooved plate specimens made of four materials used in nuclear power plants without or with the effects of damage depending on stress triaxiality factor, incorporation of the effect of the Lode angle was explored for improving the consistency with the test results. Dependency of true rupture strain on the Lode angle as well as on the stress triaxiality factor was introduced in the new approach. Damage calculated by summing a ratio of equivalent plastic strain increment to the true rupture strain was used for determining failure point and reduction of preceding deformation resistance. Furthermore the effect of the Lode angle on the resistance against plastic deformation was introduced in a separate way to address some situation. After examining the behavior of these parameters in each specimen type, related material constants for each combination of material and temperature were determined by iterating calculations with different constants and comparisons of predicted load–displacement relations with experimental ones. Although constants needed to be adjusted according to materials and temperatures, reasonable agreements were achieved eventually between the experimental and predicted load–displacement curves for all the specimen types investigated. In particular, the maximum load and the displacement at failure could be predicted within an error of about 5% and 10%, respectively, providing compelling evidence for the soundness of the approach.
