Abstract. Modeling fracture is important for predicting the actual behavior of structural steel in fire, where fracture of steel members is often the failure mode that affects the overall strength and deformation capacity of steel members. Therefore, in this paper, recent research on modeling steel fracture at elevated temperature is presented. Extended Finite Element (XFEM) together with fracture criterion of principle logarithmic strain is used to simulate the fracture behavior of three different types of tensile test specimens named MA, MB and MC made from American structural steel ASTM A992 at elevated temperature up to 1000°C.The simulation results are then compared with test results to validate the accuracy of simulations. In addition, study is conducted to evaluate how sensitive the predicted engineering stress-strain curves and fracture initiation point on the curves to the selected value of principle logarithmic strain at fracture. Finally, the generalized principle logarithmic strains at each temperature case are proposed for fracture simulation of ASTM A992 steel specimens.
IntroductionFracture behavior of steel at elevated temperature is of great interest because it can significantly influence the response of steel structure in fire. One example of collapse of steel building mainly due to connection fracture is WTC building collapse in US in fire event after severe attacks. It was also observed that steel fracture in tension was a common failure mode during heating or cooling stage in fire events. However, to accurately predict fracture behavior of steel in fire is still difficult due to lack of enough test data and studies. Therefore, in order to propose an approach to predict fracture behavior of steel in fires with reasonable accuracy, standard steel specimens under unaxial tension at room temperature and elevated temperature up to 1000°C with 100°C intervals are investigated and simulated using advanced computational method such as finite element method. The obtained engineering stress-strain curves and failure mechanism of structural steel are compared with several tests data to verify the proposed approach.