EMD方法在消除桥梁振动信号局部强干扰中的应用

APPLICATION OF EMD(EMPIRICAL MODE DECOMPOSITION) METHOD TO ELIMINATE LOCAL STRONGER DISTURBANCE FROM VIBRATION SIGNAL OF A BRIDGEZZZ

在进行桥梁的健康监测和状态评估时 ,外界环境的影响常常会在采集的振动信号中形成局部强干扰 ,导致分析结果的严重失真。为解决这一问题 ,文中基于经验模态分解 (empiricalmodedecomposition ,EMD) ,提出一种信号强干扰的消除方法。首先利用EMD把一个时间序列的信号分解成不同时间尺度的本征模函数 (intrinsicmodefunction ,IMF)和残余项 ,然后采用合适的带通滤波器对前几个IMF进行滤波 ,在存在强干扰的区段 ,用滤波后的数据代替滤波前的数据 ,并使后几个IMF在相应区段的幅值为零 ,最后将所有的IMF及趋势项重新进行叠加 ,即得到消除强干扰后的信号。将该信号再次进行EMD分解 ,可得到一系列新的IMF ,它与未消除干扰时信号的分解结果有显著差别。通过对实测南京桥有对讲机干扰的应变信号进行分析 ,结果表明该方法可行。

When carrying out bridge health monitoring and state assessment, environment influence usually causes local stronger disturbance in acquired vibration signals, consequently brings analysis results serious distortion. For this reason, a method based on the empirical mode decomposition (EMD) is introduced to eliminate local stronger disturbance from signals. The first step is to decompose a time series signal into several intrinsic mode functions (IMFs) of different time scale and residue by EMD technique. Then, the fore several IMFs are filtered through appropriate band-pass filter, replacing the filtered data for the original ones in abnormal sections and taking the corresponding values of the rest IMFs as zero. Finally, a new signal with the stronger disturbance eliminated can be reconstructed by superposing all the IMFs and the residue. Decomposing the new signal by EMD technique again and a set of new IMFs, which are quite different from the original ones, are obtained. Results of analysis of measured strain signal of the Nanjing Yangtse River Bridge with the disturbance of interphone show the feasibility and validity of the method.