具有间断事件检测和分离的经验模态分解方法

针对在经验模态分解（EMD）筛选过程中间断信号引起的模态混叠问题，提出了一种新的解决方法．该方法为根据间断过程信号与正常背景信号时间尺度的不同，利用固有模态函数（IMF）的瞬时频率特性实现间断信号的精确定位，依据定位的间断信号段经端点延拓重新做EMD分离出间断信号，从而在后续的EMD中消除该间断信号的影响．将该方法与小波法消除间断信号的结果进行了比较，显示其滤除结果的信号失真较小，并可将间断信号分解为一固有模态函数．

Empirical mode decomposition method with intermittency test and separation

Aiming at the mode mixing problem caused by intermittency signal in the sifting process of empirical mode decomposition(EMD),a new solution is presented.In this method,according to the time scale difference between intermittency signal and normal background signal,the intermittency signal was located exactly by taking advantage of instantaneous frequency characteristics of intrinsic mode function(IMF).Depend on the located intermittency signal segment,by boundary extending,redid empirical mode decomposition to separate intermittency signal,and accordingly the influence of intermittency signal was removed in the following empirical mode decomposition.The results indicate that compare with the result of removing intermittency signal,the signal distortion is less than the wavelet method,and intermittency signal can be decomposed into one intrinsic mode function.