摘要
结构的损伤会导致结构模态参数的变化,通过获取不同损伤模式下结构的模态参数并经过分析对比,可实现对结构的损伤模式识别.为此,针对咸阳至宝鸡天然气输气管道工程的渭惠渠悬索跨越结构,制作了缩尺比例为1∶8的悬索桥模型结构,针对悬索桥模型结构进行了完好状态和3种基本损伤模式状态下的输入白噪声和地震波的振动试验研究,得到了多组长输管道悬索桥模型结构在完好状态和不同损伤状态下的频率参数;且采用目前广泛应用的神经网络BP网络实现了对长输管道悬索桥完好状态和3种主要损伤模式的识别.
Structural damages may result in changes of modal parameters. And the damage patterns can be identified through analysis and comparison of structural modal parameters in different states. Natural features of structures in different damage states can be obtained with neural network even noises in existence. And the damage pattern identification of structures can be realized. In this paper, natural frequencies of cable-suspension structure model of pipelines with 1/8 reduced scale were obtained via vibration experiments. By employing the popular back-propagation neural network, the perfect status and three primary damage patterns of cable-suspension structure model of pipelines were identified.
出处
《天津大学学报》
EI
CAS
CSCD
北大核心
2010年第3期229-233,共5页
Journal of Tianjin University(Science and Technology)
基金
国家自然科学基金资助项目(50278064
50578109)
关键词
长输管道
悬索跨越结构
BP网络
损伤模式识别
自振频率
long-distance pipeline
cable-suspension structure
back-propagation neural network
damage pattern identification
natural frequency
