针对低信噪比下EEMD法在分解过程中产生模态混叠导致滤波效果变差的问题,提出了一种改进的EEMD滤波方法。该方法首先设计最优带通滤波器,滤除信号中的脉冲干扰,再将滤波后的信号进行EEMD分解,然后采用SG滤波器对筛选的IMF再进行滤波,最后对滤波后IMF进行重构。最后将改进的EEMD法和EEMD法分别用于多相流测量信号滤波及含水率测量,结果表明,改进的EEMD法与EEMD法相比,信噪比提高了约2~3 dB,滤波效果更好;含水率平均测量精度提高了约3%,测量误差更小。
Aiming at the problem that EEMD causes mode mixing that leads to poor noise reduction effect under complex noise back- ground a modified EEMD de-noising method is proposed. In this method, an optimal band pass filter is firstly designed to filter out much of the pulse interference in the analyzed signal, and then the filtered signal is decomposed with EEMD. The selected IMFs are filtered a- gain with SG filter, and the filtered IMFs are reconstructed. The proposed modified EEMD method and EEMD method were applied in the measurement signal filtering of multiphase flow and the measurement of water content percentage, respectively. Experiment results show that compared with EEMD method, the proposed modified EEMD method improves the SNR by about 2-3dB, has better filtering performance; and the average measurement accuracy of water content percentage is improved by about 3% , the measurement error is smaller.