Behaviour of steel connections under quasi-static and impact loading An experimental and numerical study

Abstract

This PhD thesis is concerned with the behaviour of steel joints and connections subjected to quasi-static and impact loading. A major part of the work involves experimental testing, which includes both component and material testing. The purposes of the component tests were to study the behaviour of the joints and connections as they were subjected to the loading, and to obtain experimental data that can be used for validation of numerical models. The material tests were mainly performed to identify material parameters employed in the numerical models, but also to support the findings in the component tests. Another major part of the thesis concerns numerical modelling. Finite element (FE) models were created and validated against the laboratory tests. The FE models enabled investigating aspects that are challenging to study experimentally such as the local strain rates in the various components of the connections. The thesis is divided into three parts, where each part covers separate but closely connected topics. The parts are linked together by a preceding synopsis, which presents the motivation, objectives, and scope of the PhD work. Moreover, the synopsis provides summaries of the different parts, and some general conclusions and suggestions for further work. In Part 1, the behaviour of beam-to-column joints subjected to quasi-static and impact loading is studied. The joints consist of H-sections beams and columns that are joined by bolted endplate connections. Both quasi-static and dynamic full-scale tests were conducted. The experiments showed that inertia effects induced a pronounced shearing action of the joints in the impact tests, which was not observed in the quasi-static test. The FE simulations were able to capture most of the behaviour observed in the tests. Moreover, the simulations demonstrated that the failure mode of the joints can be completely altered by taking the inertia introduced by notional floor slabs into account. Part 2 involves experiments and FE simulations of various bolt and nut assemblies subjected to quasi-static tension loading. The principal objective of this part was to investigate how the position of the nut along the bolt shank affects the failure mode of the assemblies. Both the tests and the simulations demonstrated that placing the nut close to the thread run-out of the bolt shank increased the chance of thread failure. The simulations revealed that when the nut was close to the thread run-out, necking of the bolt shank reduced the effective overlap between the threads of the bolt and nut, which further induced thread failure. Part 3 is concerned with fillet welds subjected to impact loading. Two different types of test specimens were designed; one with longitudinal welds and one with transverse welds. The design of the specimens allowed measuring the forces acting on the welds as well as the deformation of the welds during testing. In addition to the impact tests, corresponding quasistatic tests were performed for comparison. The test results showed that the resistance of the welds was practically unaffected by the applied displacement rate. On the other hand, the deformation capacity of the specimens with longitudinal welds was significantly reduced as the displacement rate was increased. This effect seemed to be due to thermal softening of the material. In the simulations of the tests, a comprehensive material model incorporating porous plasticity was employed. The simulations captured well the response observed in the quasistatic tests. In the impact test simulations, the force and weld deformation at failure were considerably over-predicted.