延拓小波变换识别的桥梁模态参数研究

振动系统模态识别是当今桥梁结构动力特性研究的热点之一。从复模态理论的一般阻尼系统的模态参数分析入手,利用径向神经网络插值技术,对含有噪声的振动信号进行信号预测延拓降噪处理,借助连续的Morlet小波变换,识别出了振动结构系统的模态。以重庆大佛寺长江大桥为研究背景,使用模态叠加法和Morlet小波分析识别结构,二者吻合程度较高。研究结果表明,基于径向神经网络的延拓预测的信号降噪效果好;Morlet小波变换识别模态参数精度满足工程要求。

Bridge structural modal parameter identification based on wavelet transform with symmetric boundary-extension

Modal identification of a vibrating system is one of the research hotspots in studying the structural dynamic characteristics of a bridge. The modal of a vibrating system was identified with the help of continuous Morlet wavelet transform by first analyzing the modal parameters of a general damped system with eomplex modal analytieal theory and then using RBF neural network interpolation technique to predict,extend and denoise the vibration signal containing noise. This paper takes Chongqing Dafosi Yangtze River Bridge as the researeh object and identifies its vibrating system mode with respectively the mode superposition method and the Morlet wavelet analytical method, of which the results show high conformity. The research findings indicate that signals extended and predicted with RBF neural network can be denoised effectively and the parameter accuracy of modal identification with Morlet wavelet transform meets the engineering requirement.