COMPARISON OF FRAGILITY CURVES BASED ON SOFT COMPUTING AND ANALYTICAL METHODS ON HORIZONTALLY CURVED BRIDGES

Today, a wide range of tools and software are available for nonlinear dynamic analysis as a comprehensive method for estimation of seismic demand in highway bridges. Generally, nonlinear structural analysis methods based on the structural mechanic and mathematical methods are using in the estimation of the structural vulnerabilities. The Soft Computing (SC) components include multiple main branches and many techniques inspired by the activity of the human brain that has proven useful in solving complex problems. Among SC methods, the Artificial Neural Network (ANN) is widely is employed in engineering optimization problems. In the recent decades, a significant number of horizontally-curved steel bridges have been built in urban areas. The comparison of these bridges with straight bridges indicates fewer studies on the seismic vulnerability of these bridges. Figure 1 represents the structure of neural network using structural characteristics, earthquake ground motion records and engineering demand parameters.To evaluate the seismic vulnerability of most representative of bridges, a typical layout from a single curve of curved bridges is selected. The bridge superstructure of two-span bridge consists of a concrete deck slab that rests on six I-shape steel girders, equally spaced at an interval of 1.46 m. Due to the presence of lateral restraints, the girders distortion is neglected. The modeling of curved bridge conducted in OpenSEES software.

کلیدواژه ها:

Seismic fragility curve ، Neural Network ، Statistical Method ، Bridge