Product development of the BSF connections involved two series of load tests. The original BSF connections were load tested by the Norwegian Building Research Institute, Oslo, Norway, in 1988. An internal truss calculation model for design of the force transfer from the steel box unit to the concrete beam was developed and experimentally documented through full-scale tests at SINTEF Structures and Concrete, Trondheim, Norway in 1992.

Both series of load tests consisted of applying axial tension and vertical loads to concrete beams containing BSF end connections. The axial tension loads in the test were 40 to 50 percent of the vertical test loads, and represented shrinkage, creep and temperature contraction forces by sliding (friction) of the knife plate within the column box.

The first series of 51 load tests, conducted in 1988, documented the capacity of the original BSF connection units. The internal truss calculation model was developed from the first test series, and verified in 14 additional load tests carried out in 1992. This test series included measuring strains in the reinforcement at about 20 points in each test specimen.

The second test series documented the theoretical truss models reasonably well, and clarified which type of reinforcement provided the best structural behavior of the connections. The current version of the BSF connections and reinforcement were developed from the test results, including the following changes to the beam units:

• Top plate replaced by half-round steel profile or “saddle” to provide direct contact with front hanger rebar steel. 
• Use of a front plate flush with the end of the beam and welded to the beam box side plates and top half-round saddle. 
• The bottom plate reduced in width to be the same as the beam box (side plates). The bottom plate in the smaller units is replaced by the longitudinal reinforcing bar anchoring the unit for axial tension loads.

The simplified truss model for design was developed by Partek Ostspenn from the more rigorous theoretical models documented in the second load test series. The second load test series results showed the beam ends with BSF connections had greater ultimate capacity than predicted by this simplified truss model. The simplified truss calculation model neglects the additional strength contribution of: (1) the normal beam end shear reinforcement used around the BSF beam box unit; and (2) the cantilever moment capacity of the steel beam box and bottom longitudinal reinforcing bar welded to the beam box.

Some of the SINTEF test results are presented in downloadable file to the left. The results indicate the maximum loads achieved during the tests easily exceeded the ultimate capacity calculated using the hanger rebar steel (Fig. 2, Force TF1) yield capacity and the ACI 318 Code concrete shear strength limits. Figures 3 and 4 show reinforcing arrangement, strain gage locations and crack pattern after fracture for test beams B4A and B7B. The ACI 318 Code limits nominal concrete beam shear stress to 10V¥Z, regardless of the amount of shear reinforcement, to conservatively control diagonal truss mechanism concrete stresses to a value below the crushing strength of the concrete.

See more testing details in the downloadable document above.