EXPERIMENTAL AND NUMERICAL INVESTIGATION OF THE T-STUB ELEMENTS WITH FOUR BOLTS IN A ROW UNTIL BOLT FRACTURE

Abstract

For the past several decades, codified design of steel connections in civil-engineering is based on the component approach. For a very common end-plate connection, tension component, named T-stub, usually dictates the connections’ behaviour. This T-stub element is greatly investigated in the configuration with two bolts in a row, but configuration with four bolts in a row is usually neglected, both in the studies and codes. This paper presents an experimental investigation of T-stub elements and important aspects of their numerical modelling. Special attention is dedicated to the material testing and modelling since all of the tests were performed until bolt fracture. Uniaxial tests of steel specimens were performed using extensometers, strain gauges, and Aramis system, while bolt material is additionally tested by microscopic examination and hardness testing. In order to obtain satisfactory calibration of numerical models developed in Abaqus, knowing material parameters including damage initiation and propagation is crucial. Several iterative numerical-experimental procedures for obtaining the true stress-strain curves are outlined and compared, along with well-known Bridgman method. The advantages of using Aramis system in calibrating numerical model, for both material and assembly are demonstrated. In the end, comparisons of numerical and experimental behaviour curves are presented and satisfactory results are obtained.