A method for linear response analysis of long bridges under multi-support seismic excitation

The study of the seismic response of long bridges under non-uniform seismic excitation date back to the early 80 s. So far, many methods have been developed to estimate the bridge responses under such excitation. Eurocode-8 is considered the first seismic design code that provided a simplified method for linear seismic analysis of long bridges under non-uniform seismic excitation. In this method, earthquake ground motion at each bridge support is decomposed into uniform and relative excitations. Any desired elastic response quantity of bridge for each input is obtained and results are combined using the SRSS combination rule. In this paper, it is shown that the SRSS method leads to very conservative and unrealistic responses. An alternative combination rule is then recommended for response estimation, and random vibration analysis is used to calculate the correlation coefficient between the two mentioned responses. Efficiency evaluation of the proposed method is investigated by Monte Carlo simulation results of nine long bridges with different geometrical and dynamic characteristics. Two simplified mathematical bridge models for longitudinal and transverse directions are developed. Then, a simple mathematical formula for the correlation coefficient is extracted, and its accuracy and performance are evaluated. Finally, it is concluded that the proposed formula is relatively accurate and appears in acceptable results.

Automated summary

The study reveals that the SRSS method leads to conservative and unrealistic responses in long bridges under non-uniform seismic excitation. An alternative combination rule is proposed, and the correlation coefficient between responses is calculated using random vibration analysis. Through Monte Carlo simulation and mathematical modeling, the proposed method shows relatively high accuracy and acceptable results.