摘要
为对不规则框架结构的柱进行损伤识别,由位移指标对损伤初步定位后,利用均匀设计思想构造分步识别方案,通过BP神经网络利用频率和位移指标对损伤构件进行具体定位。通过6层框架数值模拟结果表明此方法克服了大型均匀设计表偏差过大,以及单一参数进行损伤识别的缺陷,能够较好地对框架结构损伤位置进行精确判定。
In order to detect the damage locations of the column of irregular frame accurately, After locating the initial positioning of damaging by the displacement index, sub -step identification scheme was used through uniform design thought, the damage component is located by frequency and displacement index through BP neural network. This method has overcome the large deviation of large uniform design table and the defect of single parameter for damage simulation, it can accurately capture the damage member identification by a six -story frame numerical of the frame structures.
作者
唐文奇
付伟庆
TANG Wenqi;FU Weiqing(Qingdao Unive. of Technology, Qingdao 266033, Chin)
出处
《低温建筑技术》
2017年第12期80-83,共4页
Low Temperature Architecture Technology
关键词
框架结构
损伤识别
BP神经网络
均匀设计实验
frame structure
damage detection
BP neural network
uniform design experiment