Identification of Column Bases’ Stiffnesses in Steel Moment-Resisting Frames Based on Fishbone Model and Bayesian Inference

This paper presents a method of identifying the stiffness of column bases in steel frames based on the fishbone model and the hierarchical Bayesian update method. In this study, the fishbone model considering the flexibility of column bases can be used to update the rotational and shear stiffness of the column base, which is impossible in the traditional model updating method because the previous model assumes that the column bases are fully rigid. First, the steel frame is simplified into a fishbone model considering the flexibility of the column bases, and thus the eigenvalue problem of the fishbone model is established. Then, the shear stiffness and rotational stiffness of the column base of the fishbone model are identified using incomplete modal data based on the hierarchical Bayesian model updating algorithm. The effectiveness of this method is numerically studied by using a 5-story steel frame model using the first fourth frequencies and the corresponding lateral components of the mode shapes.