The face bending behaviour of blind-bolted connections to concrete-filled hollow sections

Structural Hollow Sections have superior structural performance over open sections and are currently available as circular, elliptical or rectangular sections. However, the practical use of these sections is limited due to complexities involved in their connections. The lack of access to the interior of the section makes it almost impossible to use standard bolted connections. The so-called Blind Bolts are therefore used as fasteners to alleviate these complexities by allowing for bolted rather than, the-not-so-popular, welded connections to hollow sections. Lindapter’s Hollo-Bolt is one of the Blind Bolts used for hollow sections connections. However its established use is currently restricted to transferring tensile forces and vertical shear only. Filling Square Hollow Sections (SHS) with concrete, when utilising Hollo-Bolts, was found to improve the connections’ performance in resisting moments, but there is currently no guidance available for the design of such connections. Many methods are used to model connections behaviour. The so-called component method has emerged to be the most favourite and has been adopted in the Eurocode 3. In this method, the connection is divided into basic components. Each component has a contribution to the structural behaviour of the connection. For Hollo-Bolted moment resisting connections, the behaviour of two of the components, fastener in tension and concrete-filled SHS face in bending, are not available. The application of the component method is therefore not possible. This research aims to devise a model to predict the behaviour of the concrete-filled SHS face in bending. A novel analytical model of the concrete-filled SHS face bending has been proposed in this work. The model has three parts: Initial Stiffness, Yield Force and Post-Yield Stiffness. The Initial Stiffness was formulated by theoretically substituting the face of the concrete-filled SHS with a beam element. The beam is assumed to be loaded by a rigid strip and fixed at its ends. Yield line analysis was used to investigate possible failure mechanisms and associated strengths. The model adopted the mechanism which theoretically led to the critical yield force. The Post-Yield Stiffness was taken as a percentage of the Initial Stiffness in line with other work from the literature. An extensive full-scale experimental programme was undertaken to calibrate the aforementioned analytical model, and to examine the effects of varying parameters on the SHS face bending behaviour. Typical experiments involved one row of two bolts pulled out of concrete-filled SHS. A special dummy bolts were manufactured to the exact size and geometry of open Hollo-Bolts, and were used in the experimental programme to remove the influence of any deformation associated with the real Hollo-Bolts, and thus isolate the face bending behaviour. Non-contact video-based equipment was used to record the SHS face deformation. Three parameters were varied: the SHS face slenderness ratio, the bolts gauge to SHS width ratio and the concrete in-fill compressive strength. A finite element model was also developed to complement the experimental programme. The model was developed using the ANSYS Parametric Design Language (APDL) to allow for easy parametric analysis and knowledge transfer. Dimensions, parameters and materials properties could be easily altered in the fully parametric model script. The outcomes of the experimental programme and the finite element model were used to formulate design charts for two calibration factors: kis for the calculation of Initial Stiffness, kyf for the calculation of Yield Force. A chart was also formulated for the Post-Yield stiffness ratio. The proposed analytical model (semi-analytical after calibration) was compared with the results of experimental programme and finite element modelling. The model was found to capture the behaviour of concrete-filled SHS face bending with sufficient accuracy, lying between 90% prediction lines derived from the experimental results. This is considered sufficient for the proposed model to capture the concrete-filled SHS face bending component for connection design purposes