基于人工神经网络方法的钢筋混凝土梁式桥损伤鉴识研究

摘要

越来越多的桥梁需要进行损伤维护，面对的首要问题即是如何快速高效地进行损伤鉴定。本文探讨了人工神经网络方法在损伤鉴识方面的可行性，通过对68座钢筋混凝土梁式桥破坏信息的收集整理，建立仿真样本库，训练网络。经过比较选择泛化能力优越的RBF网络，开发了“桥梁工程鉴识专家系统”。利用该系统对一座损伤桥梁进行鉴识分析，良好的效果说明了该系统的有效性以及人工神经网络方法在解决工程鉴识问题方面的卓越能力。%More and more damaged bridges need maintenance. The quick and efficient damage forensic became the most important issue. In this paper, the feasibility of artificial neural network method applicated to the damage forensics was investigated. 68 reinforced concrete girder bridges' damage information was collected for establishment the simulation sample library and training multiple networks. Basing on the compare, the RBF network, because of its generalization, was selected for developing Bridge Engineering Forensic Expert System. The network was used for a damage bridges' forensic. The excellent results demonstrate the effectiveness of the network, and the outstanding ability of artificial neural network method in solving engineering forensic problems.