中文摘要：

框架结构是目前广泛存在的一种建筑结构形式,对其损伤识别问题进行研究具有重要的工程意义。利用FRB神经网络对框架结构损伤进行训练识别,对＂曲率模态差＂指标的框架结构损伤定位、定量识别问题进行了研究和实例分析,结果表明＂曲率模态差＂指标对框架结构的损伤敏感,适宜于进行框架结构损伤定位以及较大损伤的定量识别。

英文摘要：

It is of very important engineering significance to study the damage identification of the framework,which is a widespread building structure.Here,the training and identifying are made for framework damage by using FRB neural network,and the positioning and quantitative detection for the framework damage based on ＂the difference of curvature mode＂ are studied and analyzed.The results show that ＂the difference of curvature mode＂ is sensitive to the damage degree of framework,and suitable for the positioning of framework damage and quantifying of larger damage.