基于集合经验模态分解方法的GNSS监测数据处理

GNSS monitoring data processing based on ensemble empirical mode decomposition method

为了提取全球导航卫星系统(Global Navigation Satellite System, GNSS)监测数据中的变形信号与振动信号等有用信息,获得结构体变形特征与振动特征,本文使用集合经验模态分解(Ensemble Empirical Mode Decomposition, EEMD)方法、经验模态分解(Empirical Mode Decomposition, EMD)方法和小波分析方法处理桥梁GNSS监测数据。EEMD处理GNSS监测数据的流程是根据信号本身的尺度对其进行分解,对分解产生的本征模态函数(Intrinsic Mode Function, IMF)进行傅里叶变换,得到各个IMF分量的频谱特征,再基于频谱特征构造EEMD时空滤波器。通过对仿真信号和对监测GNSS数据进行试验对比,得到如下结果:EEMD去噪的效果比EMD方法和小波分析方法去噪的效果更好,可有效提取监测数据中变形信号与振动信号,并且避免了EMD处理信号过程中模态混叠问题和小波分析处理信号过程中小波基选择的不同造成的去噪效果的差异。本文的研究结果可为监测结构的进一步分析提供数据依据。

In order to extract useful information, such as deformation signal and vibration signal from GNSS monitoring data to obtain deformation characteristics and vibration characteristics of the structure, ensemble empirical mode decomposition(EEMD) method, empirical mode decomposition(EMD) method and wavelet analysis method are used in this paper to process bridge GNSS monitoring data. The process of EEMD processing GNSS monitoring data is to decompose the signal according to the scale of the signal itself, carry out Fourier transform on the intrinsic mode function(IMF) generated by the decomposition, get the spectrum characteristics of each IMF component, and construct EEMD spatial-temporal filter based on the spectrum characteristics. By comparing the simulated signals and monitoring GNSS data, the results are as follows: The denoising effect of EEMD is better than that of EMD method and wavelet analysis method;It can effectively extract the deformation signal and vibration signal in the monitoring data, and avoid such problems as the EMD modal aliasing in the process of signal processing and the differences in denoising effects caused by different wavelet base selections in the process of signal processing by wavelet analysis. The results of this paper can provide data basis for further analysis of monitoring structure.