Aspects of multi-rowed bolted connections in pultruded fibre reinforced polymer structures

This thesis presents a comprehensive experimental and analytical investigation of multi-row bolted connections of pultruded fibre-reinforced polymer (PFRP) structures, with particular emphasis on net-tension mode of failure. Although bolted joints are indispensable for mechanically fastening PFRP assemblies, existing design provisions and semi-empirical models typically address configurations of up to three rows and three columns (3×3) of bolts. Consequently, the behaviour of larger multi-row arrangements remains insufficiently characterised, constraining safe and efficient design practice. To address this limitation, a systematic experimental programme was undertaken on 3×3, 3×4, and 4×4 plate-to-plate connections, both staggered and non-staggered, to examine joint structural response and failure mechanisms under uniaxial tension. The chosen connection geometries and bolting details were in accordance with available international design guidance, and the pultruded plate was sandwiched between two steel plates by way of finger-tightened steel bolting.

The experimental results showed that net-tension failure consistently initiated at the first bolt row, confirming the assumptions underpinning existing semi-empirical models. Introducing a fourth bolt row produced a noticeable increase in ultimate load along the longitudinal pultrusion direction, while only a marginal improvement was observed when the material was oriented transversely. Staggered bolt layouts, although compliant with current geometrical requirements, proved less efficient than non-staggered arrangements owing to shorter fracture trajectories and non-uniform stress transfer between offset bolts. Experimentally derived stress concentration factors were proposed for direct use in the net-tension design formulae specified in the CEN Technical Specification (CEN/TS 19101:2022), thereby improving the reliability of design resistance predictions for multi-row bolted connections. The laminate architecture influenced the failure crack propagation, with the failure changing from a straight fracture across the first bolt row to diagonal cracks along the ±45° fibre orientations in different PFRP materials.

Further testing on the 4×4 configurations quantified load distribution among bolt rows and the time-dependent performance of the connections. The first row attracted the highest load, whereas the final row maintained a meaningful contribution, broadly consistent with the theoretical distribution of 0.4, 0.3, 0.2, and 0.1. Sustained-load tests revealed progressive bolt-force relaxation, confirming viscoelastic effects in the polymer matrix, while repeated static loading produced partial stiffness recovery after a modest initial reduction. The study also examined installation parameters, where increasing bolt torque improved joint performance up to an optimum moderate-torque range, beyond which local crushing under the washer developed. Reducing bolt-hole clearance yielded a discernible improvement in strength and a smoother load–stroke response, whereas the standard North American clearance of 1.6 mm provided a practical balance between performance and constructability.

Overall, the research work reported in this thesis extends the applicability of current design guidance by providing new experimental evidence and validated stress concentration factors for 3×3, 3×4, and 4×4 configurations, both staggered and non-staggered. By refining resistance formulations, incorporating long-term viscoelastic behaviour, and linking laminate microstructure to joint performance, the thesis advances design-oriented procedures for bolted PFRP connections and supports their safe, durable, and efficient implementation in civil engineering practice.