MODIFIED D-VALUE METHOD TO PREDICT SEISMIC STRENGTH OF MULTI-STORY STEEL MOMENT RESISTING FRAMES WITH MID-STORY PIN COLUMN BASE SYSTEM

Abstract

 With an embedded type column base, or a fixed column base, column bases of the first story columns for frames designed following current design codes may yield simultaneously when the second story beams yield, because it cannot be avoided that the bending moment become substantially large at the base. Kimura et. al [4] has proposed the concept to realize beam yielding mechanism of a steel moment-resisting frame by applying a pin-support column base system to midpoints of the first story columns. With a proposed column support system, or mid-story pin column base system, it can be reliably designed that the first story columns remain in elastic until the maximum story drift of a frame exceeds 0.03 rad. While seismic performance of the frame is significantly improved with this new column base system, yielding of columns must be prevented or collapse mechanism may be formed.

 In practice, structural design requires multiple trials to optimize selection of structural components whether their combinations meet seismic demands on a frame. The D-value method [3] is a design method based on fundamental structural mechanics to predict story drift distribution of a moment-resisting frame under designated lateral force distribution. The D-value method has advantage to predict structural performance directly with respect to structural properties.

 This paper evaluates seismic performance, especially flexural demand on columns for a moment-resisting steel frame adopting a mid-story pin column base to maintain them elastic until a frame reaches the ultimate limit state. Because the original D-value method is only applicable to a frame with linear structural behavior and with a conventional column base, it is extended to predict elasto-plastic behavior of frames. In the proposed modified D-value method, replacement of column base to a mid-story pin column base is simply considered as shift of the location of an inflection point in the first story column. In addition, rotational restriction of the column by plastic beams is neglected to calculate moment and displacement distribution of the plastic stories in the incremental steps, assuming a beam has the perfect-plastic behavior.

 The major findings of this paper can be summarized as follows:

 1) The proposed modified D-value method successfully predicts mechanical behavior calculated by static analysis.

 2) Series of static analysis shows that beam yielding mechanism is formed by reaching a story drift ratio of larger than 0.03 rad, and that required flexural demand on the column reaches twice as large as the full-plastic moment of the beams for a 3-story frame and exceeds for 6- and 9-story frames.

 3) Seismic analysis shows that column moments calculated by the modified D-value method sufficiently agree with those calculated by simulation while response of frames remains in elastic range. Once the response of frames goes into the plastic range, calculational errors are increased up to 21%, 31% and 18% for 2-, 6- and 9-story frames. Seismic analysis tends to yield larger column moment, but still the global trend such as in the relation between the maximum moment of the column and the maximum story drift is sufficiently reproduced by the modified D-value method.

 4) The larger maximum moment of the column calculated by simulation is thought to be caused by two reasons: one is that story shear distribution at recording the maximum moment of the column becomes closer to triangle story shear distribution rather than Ai distribution, which is adopted in calculation by the modified D-value method, and the other is that sequence of beam plasticity of the frames are substantially changed from those assumed in the D-value method.