Strong ground motions for bridge design and non-linear dynamic response analysis of bridges

The ground motion records of the 1999 Chi-Chi, Taiwan earthquake are used to develop a methodology to characterize ground motions in terms of site and directivity effects. A methodology is developed to estimate the site classification (rock vs. soil) of Taiwan stations. The method is based on the analysis of the shape of 5% damped pseudo-acceleration spectrum of horizontal ground motion component normalized with respect to PGA from 87 stations of known site conditions. The results obtained from the proposed methodology correlate well with the available known site information data. Developed methodology is applicable to estimate site classification at other regions. Directivity effects are quantified by the residuals or model error of the median ground motion relationship expressed as a function of selected geometrical parameters related to the fault. The geometrical parameters include the fraction of the fault rupture that lies between the hypocenter and the site, and the angle between the fault plane and the path from the hypocenter to the site. Analysis indicates that directivity effects are significant for peak ground acceleration and pseudo-acceleration at all periods.; Advanced Ritz vector methods for inelastic response analysis of bridge structures are developed. The methods are then implemented in NEABS program. The Ritz algorithms are the Lanczos method and the WYD method, wherein both methods consider the spatial distribution of the earthquake loading. Two inelastic methods considered are pseudo-force method and tangent spectrum method. The pseudo-force method is efficient in terms of computation effort. This method provides an accurate numerical result if the non-linearities are not severe. It is found that the computational time to calculate the time history response of bridge systems using the derived Ritz vectors is approximately 25% to 75% of the time required by using direct step-by-step integration. The tangent spectrum method is more expensive in terms of computation effort. New set of base vectors is computed if second order norm of stiffness matrix is larger than a specified tolerance. A cut-off criterion based on the concept of effective mass participation factor is used to determine the number of derived Ritz vectors needed to be included in the time history analysis. (Abstract shortened by UMI.)...