MacRae et al. 2004

From Composite Systems
Jump to: navigation, search

A finite element model was developed to study the transfer of vertical bracing forces through a beam-brace-column (BBC) connection. Additionally, a series of experimental tests were conducted to verify the findings of the finite element model and quantify the performance of several different BBC connections.

Analytical Study

A variation on a typical penetrating gusset plate BBC connection was modeled using the ABAQUS nonlinear analysis program. For this connection, a gusset plate is positioned through the steel tube by means of slots on opposite sides of the tube cross section and welded in place. Thus, after the concrete is placed, force transfer between the steel and concrete can occur through a combination of bearing and friction. Beams and braces are connected to the portion of the gusset plate which is protruding from the steel tube. The three-dimensional model used eight-node brick elements. The steel was modeled as a bilinear plastic material with yield stress of 50 ksi and ultimate tensile stress of 55 ksi. The concrete was modeled as linear elastic material with a maximum stress of 10 ksi at a strain of 0.0015. For larger strains, the strength steadily decreased to zero. The steel-concrete interface was modeled taking into account normal forces, friction (the coefficient of friction varied from 0.0 to 0.4), and no cohesion. Axial forces were applied to the steel and concrete at the column ends in proportions based on their respective axial stiffness.

Five cases were examined using this model. The first case represented a typical configuration. The second and third cases were used to investigate the effect of tube and gusset plate thickness, as well as bond strength. The fourth and fifth cases were used to investigate the effect of column axial force.

Several measures of slip were reported for each of the models, with values typically less than 0.02 in. In the model, friction was found to account for as much as 30% of the total force transfer. However, the authors noted that the ABAQUS model did not simulate the BBC behavior well. In particular, the interface between the gusset plate and the concrete was not properly modeled and the model did not properly describe the bearing stresses underneath the gusset plate.

Experimental Study, Results, and Discussion

An experimental study was undertaken to determine if the analytical model was producing reasonable results and quantify the performance of different connection types. Since testing full connections would be prohibitive, a simplified configuration was examined. The lower half of the penetrating gusset plate and the corresponding portion of the CFT column (with a length of twice the diameter of the tube) were tested. The specimens were built to approximately 2/3 scale of prototype BBC connections. The circular steel tube had an outside diameter of 24 in., a thickness of 7/16 in., and a nominal yield stress of 36 ksi. The gusset plates had a nominal yield stress of 50 ksi and the concrete had a nominal strength of 7 ksi. Precautions were taken to ensure that the bottom of the specimen would bear over the entire cross section.

Fifteen different specimens were prepared with different gusset plate configurations. Gusset plates of different thicknesses, greased or sandblasted, and without concrete were tested. In addition, an assortment of configurations of the gusset plate were tested, including with welded ribs, with one large or many small holes, with different numbers of studs, and with short pieces of rebar welded perpendicular through holes in the gusset plate.

For each test, a measure of slip and percent composite were reported. The value of percent composite represented how well the load was distributed to the concrete. The largest percent composite value was obtained from the specimen with welded ribs, indicating that this configuration would result in more composite action. The smallest percent composite value was obtained from the specimen with the greased gusset plate. Shear studs were found to have little effect on the percent composite value since the amount of slip was not enough to engage the studs.


  • MacRae, G., Roeder, C. W., Gunderson, C., and Kimura, Y. (2004). “Brace-Beam-Column Connections for Concentrically Braced Frames with Concrete Filled Tube Columns,” Journal of Structural Engineering, ASCE, Vol. 130, No. 2, February, pp. 233-243. doi:10.1061/(ASCE)0733-9445(2004)130:2(233)
  • Gunderson, C. (2002). “Braced Frame Connections with Concrete-Filled Tube (CFT) Columns.” MSCE Thesis, University of Washington, Seattle, Washington.