一种改进的经验模态分解算法与测井信号去噪

An Improved EMD Algorithm with Logging Signal Denoising

经验模态分解中的端点效应导致了虚假分量的产生和边界信号的失真,为此基于边界局部特征延拓法,通过判断原始信号的端点能否作为极值点插入到新的上下包络中,有效抑制了端点效应。将改进的经验模态分解算法应用于核磁共振油气井探测回波信号处理中,通过Hilbert谱分析可知,得到了良好的去噪效果。

Considering the fact that endpoint effect can lead to false component generation and boundary signal distortion,and basing on local boundary feature continuation method,to judge whether the original signal end-point can be used as extreme point inserted into the new upper and lower envelope to restrain the endpoint effect effectively was proposed.Hilbert spectrum analysis shows that applying the empirical mode decomposition (EMD)algorithm improved to processing echo signals from nuclear magnetic resonance (NMR)oil and gas wells can bring about a good de-noising effect.