Seismic Behavior of 3D Steel Moment Frame with Biaxial Columns

Abstract

Three-dimensional nonlinear dynamic time-history analyses of a three-story steel moment-resisting frame designed according to the Uniform Building Code for Los Angeles seismicity were carried out. Beams in both directions had moment connections to the hollow rectangular columns. Code drift limits, rather than code considerations for bidirectional horizontal shaking, governed the member sizes. It was found that design level shaking caused the structure to exceed story yield drifts in both directions simultaneously and significant column yielding occurred above the base. Shaking in the direction orthogonal to the main shaking direction increased drifts in the main shaking direction by 46 and 64% for the design level and near-fault records, respectively, indicating that 2D analyses would not estimate the 3D response well. Also, maximum horizontal seismic components of axial force in the corner columns and side columns were 4.87 and 5.30 times the code estimations, respectively. It was shown that by increasing the column strength above the present code levels, that drifts during near-fault shaking were significantly decreased. A methodology to encourage strong-column weak-beam behavior, and to more realistically estimate the column axial forces during design level shaking is described. Building structures in seismic zones in the United States have traditionally been designed to resist lateral force with seismic frames. Other frames, referred to as gravity frames, are designed to resist gravity forces only. This design ap- proach, with separate seismic and gravity frames, requires only a few expensive moment connections, the load path is easy to follow, and 2D frame analyses may be used to model 3D frame behavior. However, as a result of the 1994 Northridge earth- quake, fracture occurred at many welded beam-column con- nections in these moment-resisting steel seismic frames (Yang and Popov 1995). The occurrence of fracture was influenced by a large number of factors, including the use of large mem- ber sizes and large welds. The lack of redundancy in these frames (Roeder 1997) and the possibility of yield in gravity columns (Mattheis 1998) are also of concern. An alternative design method in which structures are de- signed as a 3D seismic frame with beams connected to the columns, with moment connections in both directions, has some advantages. It increases the redundancy, no columns are designed to carry gravity force only, and beam and weld sizes are smaller. While many more seismic connections are re- quired, all members contribute to the frame lateral stiffness and strength allowing drift limits to be met with less material volume. However, the possibility of undesirable behavior due to significant column bidirectional loading effects should be considered. A brief summary of considerations for column bi- directional loading effects in 3D steel moment-resisting frames due to bidirectional ground shaking is given herein. The Uniform Building Code (UBC) (ICBO 1997) states that bidirectional orthogonal shaking effects must be considered if a column of a structure forms part of two or more intersecting lateral-force-resisting systems and that ''the requirement that orthogonal effects be considered may be satisfied by designing such elements for 100% of the prescribed design seismic forces in one direction plus 30% of the prescribed design seis- mic forces in the perpendicular direction. The combination re- quiring the greater component strength shall be used for de- sign. Alternatively, the effects of the two orthogonal directions may be combined on a Square Root of the Sum of the Squares (SRSS) basis.'' (ICBO 1997). The SRSS procedure (Clough and Penzien 1993) is based on the assumption that superpo- sition of the actions due to loading in each direction indepen- dently may be combined to obtain the total response. Phase incoherency of response, which may result from ground mo- tions in each direction not being in-phase or from significantly different response periods in each direction, is also assumed. The 30% rule (ICBO 1997) is based on the following con- siderations (Mattheis 1998; Tagawa 2000) following the rea- soning of Clough and Penzien (1993). If the ratio of the re- sponse due to y-direction shaking divided by that due to x-direction shaking of equal magnitude is B, and if the shaking magnitude in the y-direction is a of that in the orthogonal x-direction, then the response due to y-direction shaking, R y, is related to that due to x-direction shaking, Rx ,a s