有载频率与深度学习在桥梁损伤识别中的应用研究  

Application of Loaded Frequency and Deep Learning in Bridge Damage Identification

作　　者：杨超 刘凯[2] 刘亚红 孙建鹏[2] YANG Chao;LIU Kai;LIU Yahong;SUN Jianpeng(Shaanxi Expressway Engineering Test and Inspection Co.,LTD.,Xi'an 710086,China;Xi'an University of Architecture and Technology,Xi'an 710086,China)

机构地区：[1]陕西高速公路工程试验检测有限公司,陕西西安710086 [2]西安建筑科技大学,陕西西安710086

出　　处：《粉煤灰综合利用》2024年第4期123-128,共6页Fly Ash Comprehensive Utilization

基　　金：陕西省自然科学基础研究计划面上项目(2020JM-475);西安市科技创新人才服务企业项目(2020KJRC0047)。

摘　　要：频率作为一种测量方法简单且精度较高的动力特性参数,在损伤识别领域得到广泛应用,但现有的频率变化参数存在非唯一性的问题。为此,提出了基于有载频率指标和深度学习理论的桥梁损伤识别网络,利用损伤前后有载频率数据,构建损伤识别指标参数来识别桥梁的损伤状态。结果表明:识别网络能够有效地进行桥梁结构的损伤定位,频率数据克服了对称结构频率变化的非唯一性限制,增强了参数的表达能力,提高了桥梁损伤定位的有效性和抗噪性能。As a dynamic characteristic parameter with simple measurement method and high accuracy,frequency has been widely used in the field of damage identification,but the existing frequency variation parameters have the problem of non-uniqueness:Therefore,a bridge damage identification network based on the on-load frequency index and deep learning theory is proposed,and the damage identification index parameters are constructed to identify the damage state of the bridge by using the on-load frequency data before and after the damage.The results show that the identification network can effectively locate the damage of bridge structure,and the frequency data overcomes the non-uniqueness limitation of the frequency change of symmetrical structure,enhances the expression ability of parameters,and improves the effectiveness and anti-noise performance of bridge damage location.

关 键 词：桥梁损伤识别 有载频率 深度学习 抗噪性 

分 类 号：U446[建筑科学—桥梁与隧道工程]