采用自然激励技术提取自由衰减振动信号的桥梁船撞报警研究  

Research on bridge ship collision alarms using natural excitation technique to extract the free decay vibration signals

作　　者：苏彬建 张欣 魏斌 许肇峰 张建[2] SU Binjian;ZHANG Xin;WEI Bin;XU Zhaofeng;ZHANG Jian(Guangdong Jiaoke Testing Co.,Ltd.,Guangzhou 510440,China;School of Civil Engineering,Southeast University,Nanjing 214135,China)

机构地区：[1]广东交科检测有限公司,广州510440 [2]东南大学土木工程学院,南京214135

出　　处：《振动与冲击》2025年第8期297-308,共12页Journal of Vibration and Shock

摘　　要：一般采用振动阈值算法和振动时域特征向量算法难以从船舶撞击状态下桥梁的振动加速度实测数据中识别出船撞信号。为解决采用加速度实测数据进行桥梁船撞异常报警,存在阈值设定不同、信号类型识别困难、识别结果易受环境噪声影响的问题。研究采用自然激励技术提取加速度自由振动衰减响应信号进行桥梁船撞报警的方法。首先,采用滑动窗算法对加速度实测数据进行截取得到对应的加速度子序列。接着,采用自然激励技术对加速度子序列进行振动响应曲线提取,并利用希尔伯特变化提取振动响应信号的包络,再对包络线进行衰减函数判断,通过衰减函数判定信号类型,并在此基础上,计算不同状态下的马氏距离均值比,利用肖维勒方法对不同桥墩顶处加速度振动响应的马氏距离均值比的异常值进行判断,通过出现异常值的时间和大小来判断船撞时刻。最后,以九江大桥两次船撞桥事件的加速度实测数据为例,分析了不同墩顶位置处的加速度振动衰减信号和不同状态下的马氏距离均值比。结果表明,该方法能够识别出船撞时刻,能够有效进行船撞报警,克服了一般基于振动时域参数特征向量算法难以识别的缺点。同时,该算法通过冲击振动衰减特性,进行衰减函数判断,解决了一般桥梁船撞算法存在信号类型识别困难的问题。Traditional methods using vibration threshold algorithms and vibration time-domain feature vector algorithms face difficulties in identifying ship collision signals from the measured vibration acceleration data of bridges under ship impact conditions.To address the problems encountered in bridge ship collision anomaly alarms using measured acceleration data,including different threshold settings,difficulties in signal type recognition,and identification results susceptibility to environmental noise,a method for bridge ship collision alarms was proposed in this study,using natural excitation technique to extract the acceleration free vibration decay response signal.Firstly,a sliding window algorithm was used to extract corresponding acceleration subsequences from the measured data.Then,the natural excitation technique was applied to extract the vibration response curve from the acceleration subsequences,and the Hilbert transform was used to extract the envelope of the vibration response signal.The attenuation function was applied to determine the signal type based on the envelope.On this basis,the Mahalanobis distance mean ratio was calculated for different states.The Chauvenet method was used to judge the abnormal values of the Mahalanobis distance mean ratio of the acceleration vibration response at different bridge pier tops.The occurrence time and magnitude of the abnormal values were used to determine the timing of the ship collision.Finally,taking the measured acceleration data from two ship-bridge collision events on the Jiujiang Bridge as an example,the acceleration vibration decay signals at different pier top positions and the Mahalanobis distance mean ratio under different states were analyzed.The results demonstrate that the proposed method is capable of identifying the timing of ship collisions and can effectively perform ship collision alarms,overcoming the limitations of traditional vibration time-domain feature vector algorithms in recognition.Moreover,the algorithm utilizes the characteristics of

关 键 词：桥梁船撞报警 自然激励技术 马氏距离均值比 自由衰减 包络 异常检测 

分 类 号：TU313.2[建筑科学—结构工程]