Experimental verification of calculation methods for hollow core section fin-plate joints with varying strength grades

Abstract

This research studies the experimental and numerical behaviour of a simple beam to column joint using a fin-plate welded to an rectangular hollow core section. The following aspects of the joint are investigated:
The resistance of the welded connection.
Defining the force distribution inside the joint.
Improvement of the design method for resistance of the fin plate connection.

Four series of controlled experiments are conducted to test the joints until failure. The series combine a S355 plate with S690/S355 column grade and matching and overmatching weld material. With the data from these experiments and coupon tests, a finite element model is created using DSS Abaqus. Results are validated to accurately follow the force-displacements obtained in the experiments.
Key findings from the experiment:
The resistance of the welds was found to be much greater than predicted using the nominal values from the Eurocode. The utilization of the nominal design resistance compared to the failure load in experiments was 19% following EN1993-1-8:2005 and 25% with EN1993-1-8:2020. The main reason for this was the high penetration depth and increased weld throat when welding a 3mm weld according to standard procedure, accounting for a 100% increase in resistance. The second reason was that the failure stress was 50% higher than nominal for both the matching and overmatching infused weld material.
Insights from the finite element models:
It was proven that for certain boundary conditions it is possible to only account for shear transfer through the weld. However it was found that when expanding the Abaqus model to a full building size the boundary conditions change such that also a moment needs to be transferred through the weld and the bolts. The ratio between the stiffness of the beam and the column face determines the magnitude of these moments. The addition to the stress in the start and end of the weld can result in a 75% lower design resistance when comparing to the same weld with only a shear load. The bolt group transfers the moment through horizontal forces, again depending on this ratio. The magnitude of this horizontal force component in the outer bolts can be equal to the vertical component.
The following calculation methods were proposed:
The moment transfer through the weld from bridging the bolts eccentricity can be reduced according to the stiffness ratio R between the beam and the column face. There are four stiffness factors which influence the force distribution; The column stiffness K_c, the column face stiffness K_cf, the plate bearing stiffness K_p and the beam rotation stiffness K_b. It was found that the governing influence comes from the face stiffness and the beam stiffness. The face stiffness can be calculated by integrating the resistance over the effective length of the column h_p.
If there is no yielding in the beam cross-section then the problem can be simplified and welds should be calculated with taking into account the shear force V_ed and the bending moment M_ed which follows directly from the stiffness ratio R.