Abstract

The seismic response of a nonlinear, large-span cable-stayed bridge is a key problem in the assessment of structural safety. We propose an optimized deep learning network to study the random vibration of an uncertain bridge subjected to an earthquake. The dynamic formulas of the uncertain system were produced by transient analysis in the time domain to construct a numerical model comprising two functional modules: convolutional neural network (CNN) training with input rail irregularities and a long short-term memory (LSTM) layer for the bridge response time-history prediction. The LSTM cell was simulated by introducing the randomness of excitation and uncertain characteristics of parameters of the system into a portion of the cell, enabling the numerical model to convey the uncertainties of the bridge and obtained its stochastic response. The seismic response of railroad cable-stayed bridges under ground motion was calculated using the proposed method, taking into account the geometric nonlinearity of the bridges, and its accuracy and validity were verified. The effects of the distribution state of the seismic response samples, traveling wave effects, and intensity of the seismic spectrum were investigated, and the extremes of responses were analyzed based on Poisson's crossover assumption.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call