Analytical Procedure

Numerical analyses of all connection models were performed using ABAQUS standards (1998), which was developed based on Finite Element Methodology. All components for each connection were modeled using first-order eight-node C3D8 solid elements. Here, pretension forces of bolts for all connection models except models A1np, A2np, FElnp through FE4np, FE7np through FE13np, FE5 and FE6 were prescribed up to 40% of the ultimate strength of bolt, and those for connection models FE5 and FE6 were prescribed up to 20% and 60%, respectively. The numerical results for

Fig. 4. Boundary conditions of FEA model of top – and seat-angle connections.

the models without pretension forces in bolts are used to estimate prying forces due to introducing pretension forces in bolts.

Numerical analysis was performed considering real experimental setup and loading method ap­plied by Azizinamini et al. (1985), in which (1) two beams were symmetrically connected to the column flanges in a cruciform shape; (2) the ends of those beams were simply supported; and (3) the center of top plate of stub column was moved in upward direction as prescribed bending moment to be surcharged to the connection assemblages. Based on such experimental setup and considering structural symmetry, one-quarter model of connection composed of stub column, beam, top and seat angles and bolts was used for numerical analysis. Figure 4 shows a FE analysis model used for numerical analysis. The FE analyses considering pretension forces in bolts were performed in three loading steps. In the first step, a pressure equivalent to a prescribed pretension force is applied to the predefined section of bolt shank. As a result, the length of bolt shank at the pretension section changes by necessary amount to carry the prescribed load. In the second step, the prescribed load in bolt is replaced by changing the length of pretension section back into the initial length. In the third step, bending moment is introduced to the beam-to-column connection by employing vertical displacement at the middle section of plane 3-1 of the stub column (see Figure 4).

To precisely analyze the behavior of connecting members, contact model with small sliding option was applied for the contact surfaces between the vertical leg of angle and column flange, between the horizontal leg of angle and corresponding beam flange, and between the bolt and bolt hole elements. Moreover, to consider friction force occurring between sliding surfaces, Coulomb’s frictional coefficient is assumed to be 0.1.

Analysis Results and Discussions