小波分解与EMD在变形监测应用中的比较

COMPARISON OF EMD WITH WAVELET DECOMPOSITION FOR DYNAMIC DEFORMATION MONITORING

为分析经验模式分解(Empirical Mode Decomposition,EMD)在GPS动态变形监测应用中的可行性和有效性,利用模拟数据和振动台的GPS实测数据分析比较了EMD和小波分析在信号去噪和多尺度分解上的能力。结果表明:EMD和小波均能较准确地提取结构振动信息,但当噪声较大时小波去噪和多尺度分解稳定性略优于EMD,而EMD则基于数据本身具有自适应性,不受小波基选择和分解层数的影响。

Both wavelet decomposition and empirical mode decomposition（EMD）,a new signal processing method,can be used for denoise and multi-scale decompose GPS measurements to obtain the deformation information.To analyze the feasibility and effect of this method,the simulated data and GPS data collected from the table test are used to test the ability of EMD and wavelet in signal denoising and multi-scale decomposition.From these tests,the following conclusions that both the EMD and wavelet are good for extracting the vibration of the structure.However,when the data is noised heavily,the stability of wavelet is better than EMD in denoising and multi-scale decomposition,while the EMD is an adaptive method based on the data itself,and free from the choise of wavelet base and the number of decompose layer.