Gradient method as a tool for mathematical modeling in earthquake engineering

The goal of this paper is to investigate the role of viscodamper behavior in the identification of frame models from dynamic response data caused by seismic forcing functions. Including nonlinear damping of viscodampers in the mathematical model for even simple structures significantly affects the distribution of damping, and the accuracy with which response can be predicted. A number of different mathematical models of these structures are evaluated using system identification. Each mathematical model depends on a number of parameters related to the characteristics of the structure. The gradient method is applied to calculate the values of these parameters which best reproduce the measured response of the structure. This paper presents the mathematical model formulation of a five-story steel frame model using the parameter system identification technique and shaking table experiments. The base isolate system consisting of helical springs and viscodampers was manufactured at GERB, Germany. The experimental work was conducted using the shaking table of the Institute of Earthquake Engineering, University of Skopje, Macedonia.