Analyses of two connected hollow core units
To obtain background information for code prescriptions for the shear and torsion interaction in hollow core floors, finite element analyses of some basic cases were made. A simplified global model for complete floors developed in Lundgren et al. (2003) was used. A set-up with two simply supported hollow core units connected along a longitudinal joint and support of tie beams was modelled. The hollow core units were loaded with a point load situated at the longitudinal joint. Several parameters were varied; i.e. the section height, the slab length, an assumed gap in the longitudinal joint, stiffness of the tie beams, and the loading position.
When the two hollow core units are loaded, they will deflect and rotate equally much, as there is symmetry. Therefore, no shear force will be transferred in the longitudinal joint. Compressive contact forces will appear at the upper edge of the longitudinal joint, while the lower edge of the longitudinal joint will open without any force transfer. The main studied result from the analyses is the distribution of the contact forces along the upper edge of the longitudinal joint.
The results show that the contact forces in the longitudinal joint had typically a peak where the point load was applied. When no gap was assumed in the longitudinal joint and the tie beams were assumed to be infinitely stiff, compressive contact forces appeared towards the supports, almost in triangular shaped blocks. These blocks decreased and vanished when the stiffness of the tie beams decreased, and also when there was a gap in the longitudinal joint.
In the analyses where a gap was assumed in the longitudinal joint between the hollow core units, no contact is obtained for low loads. After a certain load, contact is obtained, and the contact forces between the hollow core units start to act. This load can be predicted by hand calculations.
As the analyses are non-linear, the results from various load cases can not be super positioned. This was demonstrated in one example. In the studied set-up with only two hollow core units, the deflections in horizontal bending were of the same magnitude as the deformations due to torsional rotations. Therefore, very small reductions in torsional moments were observed. The torsional moments would most likely be more reduced if floors with more hollow core units were studied.
shear and torsion interaction
transfer in joints