摘要
斜拉桥拉索损伤识别中不仅需要拉索发生损伤时自身索力的变化,还需要考虑拉索发生损伤引起的其他拉索索力变化,以单索索力变化为指标的方法难以适应斜拉桥的拉索损伤识别。以不同损伤下全桥索力变化率作为输入向量,拉索损伤位置和损伤程度工况为输出向量,建立了较为精准的BP神经网络模型。对斜拉桥模型在发生单根和2根拉索损伤工况时进行损伤识别,并随机选取不同损伤工况对BP神经网络模型的准确性进行验证,拉索损伤识别误差在6.0%以内。结果表明:以拉索索力变化率为指标,基于BP神经网络模型能够有效地识别斜拉桥拉索的损伤位置和损伤程度,为桥梁安全运营提供保障。
For cable-stayed bridges,the damage of a single cable will cause the stress redistribution of the whole cables.Although using frequency of single cable as the damage index will detect abnormalities in multiple cords,it is impossible to determine the location and degree of damage accurately.To address this challenge,an effective method was explored to recognize cable damage location and its degree through BP neural network in this study.In total,a benchmark finite element model of a double-tower double-cable semi-buoy continuous composite beam cable-stayed bridge was established.The damage conditions of single cable and two cables were simulated through the change of the elastic modulus.A BP neural network model was established based on cable force gradient as identification index of cable damage and its accuracy was verified.A damage identification method for long-span cable-stayed bridges was established.The results showed that the established neural network model identify the damage location and damage degree of stay cables with high accuracy.The identification error was within 6.0%,so the BP neural network can be used for cable damage identification of cable-stayed bridges.
作者
林友勤
郑学善
余印根
王志俸
LIN Youqin;ZHENG Xueshan;YU Yingen;WANG Zhifeng(College of Civil Engineering,Fuzhou University,Fuzhou 350116,China;Fujian Gaohua Construction Engineering Co.,Ltd.,Fuqing 350301,Fujian,China;Fujian Yongzhen Construction Quality Inspection Co.,Ltd.,Fuzhou 350001,China)
出处
《南昌大学学报（工科版）》
CAS
2023年第4期370-376,共7页
Journal of Nanchang University(Engineering & Technology)
基金
国家自然科学基金项目(51378112)
福建省建设科技项目(2017-K-59)。
关键词
斜拉桥
拉索
神经网络
索力
损伤识别
cable-stayed bridge
cable
neural network
cable force
damage identification
