A comparative study of three fracture models in simulating austenitic stainless steel bolted lap shear connections

This paper thoroughly investigates the performances of three representative fracture models, namely the void growth model, the Lee-Wierzbicki model and the Bai-Wierzbicki model, in simulating the fracture behaviour of stainless steel bolted lap shear connections. The ultimate goal is to determine the appropriate fracture model that can provide satisfactory simulation results with the minimum calibration requirement. To calibrate the fracture models, five different types of material test were performed, with all test specimens extracted from a single austenitic stainless steel plate of Grade 304. Thereafter, five connection tests, fabricated from the same steel plate as the material tests, were performed to examine the performances of different models in simulating typical connection failure modes (i.e. net section fracture, tearout and block shear). It was concluded that the Lee-Wierzbicki model, with 2 individual model parameters and only stress triaxiality-dependency, generated the most accurate simulation results compared with the other two models. The void growth model, with only one parameter calibrated by a tensile-dominant material test, failed to simulate the initiation of shear cracking and led to unconservative simulation results. Since the critical stress states in the connections were close to plane stress, the Bai-Wierzbicki model, with four parameters and both stress triaxiality- and Lode angle-dependencies, did not show improved accuracy than the Lee-Wierzbicki model. The results of the study indicate that a more sophisticated fracture model with greater number of model parameters may not necessarily lead to more accurate simulation results for certain engineering problems. It is further recommended that the selection of fracture model and calibration procedure (including material test specimens for calibration) should be based on the prevalent stress states in the investigated problems.