点击次数：

论文类型：会议论文

发表时间：2010-03-08

收录刊物：EI、CPCI-S、Scopus

卷号：7650

期号：PART 1

关键字：Arch bridge suspender; vibration method; suspender tension identification; neural network

摘要：The measurement of arch bridge suspender tension using vibration method is mostly considered that the suspender is idealized as taut strings. This idealization simplifies the analysis but may introduce unacceptable errors in many applications by ignoring the boundary condition and bending stiffness effects. The neural network intelligent methodology is proposed to compute suspenders tension, and design steps and optimization methods of neural network are given. In order to get correct neural network predictive model, 200 data of 20 bridges is used to train the neural network. The applicability of the proposed intelligent methodology is verified by comparison with the others method through a case. The result shows that the error isn't beyond percent of 6.