基于小波分析的大跨径桥梁健康监测数据预处理研究

Research on Long-span Bridge Health Monitoring Data Pre-processing Based on Wavelet Analysis

针对小波分析方法在桥梁健康监测系统中面临的小波基函数合理选择问题,采用一周期性仿真信号作为算例,考察了在其他研究领域内广泛用来进行信号去噪的Haar、Db、Sym、Coif、Bior、Rbio、Dmey等7种小波基函数,选取出去噪效果最显著的Bb8小波基函数。依托已建立的江阴大桥健康监测系统,对具有明显周期性的塔顶GPS的竖向数据进行去噪计算,并将计算结果与采用低通滤波进行去噪的结果进行比较分析。结果表明:由于小波方法采用了比低通滤波方法更为细致的处理方法,其去噪效果比低通滤波方法更有效。

How to select wavelet basis functions reasonably is still a difficult in the application of wavelet analysis in bridge health monitoring system.To address this issue this paper takes a simulated periodic signal as an example.It this example we examine seven kinds of wavelet functions,including Haar,Db,Sym,Coif,Bior,Rbio,Dmey,which are used widely in other research areas.Because the de-noising effect of Bb8 is the most significant,it is selected as the wavelet function.This method is applied to the established Jiangyin Bridge Health Monitoring System.Using wavelet analysis method to calculate and de-noising the vertical data from GPS on the top of the tower,which has a clear periodic.And comparing the results with ones from low-pass filter de-noising.The results show that because the wavelet method uses a more detailed approach,the de-noising effect of the wavelet analysis method is better than the low-pass filter method.