Prion and McLellan 1992
The authors conducted experimental studies of through-bolt connections between steel wide-flange beams and concrete-filled hollow steel sections, and they compared their results to other types of CFT connection. Documentation of the transfer of beam shear forces (i.e., simple shear connections) to the concrete core was the focus of the study. Although through-bolt connections are capable of transmitting moment by thickening the end plate, extending the end plate past the flanges, and adding more bolts to the connection detail, this behavior was not investigated. The bolt shear capacity was the limiting factor in this connection design.
Experimental Study, Results, and Discussion
Only simple connections capable of transmitting shear forces were considered in his study. End plates welded to the beams were bolted to the hollow steel section with four bolts made of grade 4340 steel rod with an ultimate strength of 155 ksi. The columns were 12 in. square with a D/t ratio of 24 and were completely filled with 5.8 ksi concrete. The concrete should carry a large proportion of the beam’s shear force, hence the need for a connection which bears directly on the concrete core. In previous studies (Kanatani et al., 1988), columns were filled with concrete only in the vicinity of the connection detail to prevent local buckling or crushing of the hollow steel section.
Three primary modes of failure governed the through-bolt connection detail. The bolts may experience a shear failure through the shank of the bolts, or the concrete may fail as a diagonal crack develops across the hollow steel section, transverse to the longitudinal axis. Finally, splitting of the concrete may occur, pushing the concrete and hollow steel section walls outward.
Specimens were divided into two categories, bearing and non-bearing, in order to separate the effects of friction and bearing. The non-bearing specimens were fitted with plastic tubes that provided clearance holes for the bolts to avoid bearing between the bolts and the concrete and to permit relative movement between he concrete and the steel casing. The bolts were cast directly into the concrete for the bearing tests. To determine the relationship between friction and bolt tension, bearing and non-bearing specimens were tested with bolts post-tensioned to 0 kips, 20.23 kips, and 40.47 kips.
For the bearing cases, slip between the hollow steel section and the concrete core was continuous with no sudden changes in load or slip deflection. However, in the non-bearing specimens, slip between the concrete core and the hollow steel section occurred abruptly and was accompanied by a decrease in load; successive occurrence of slip took place at higher loads. The increase in slip load was a result of an increase in bolt tension.
A constant relationship existed between post-tension and slip load in non-bearing specimens. The critical region for bolt failure was at or near the interface of the concrete and steel casing. The calculated bearing loads (total loads minus slip load) were far larger than those prescribed by design codes. The moment in the bolt at the midsection was also assumed to be negligible. The bolt may generate splitting forces on the concrete, especially in the presence of thin walled steel tubes that provided little confinement to the connection area.
The authors indicated that it is most important to insure that shear loads are transferred from beams to the concrete core due to the friction between the steel and concrete and bearing of the bolt on the concrete core. Checks must also be performed on the shear capacity of the bolts and the bearing capacity of the steel plates. The bearing stress distribution affects the stress on the bolts little; this is a valuable consideration since connection detailing is usually chosen to factor ductile failure modes, such as end plate binding or beam yielding, instead of bolt failure. The authors indicated that the construction sequence initiates with the structure being erected with snug-tightened bolts. The columns are then filled with concrete and the bolts are post-tensioned after the concrete has cured.
Reference
Prion, H. G. L. and McLellan, Andrew B. (1992). “Connecting Steel Beams to Concrete-Filled Steel Columns,” Proceedings of the ASCE Tenth Structures Congress '92, Morgan, J. (ed.), San Antonio, Texas, April 13-16, 1992, American Society of Civil Engineers, New York, New York, pp. 918-921.