经验模态分解在管道缺陷漏磁检测信号处理中的应用研究

Application of EMD in the Signal Process of Pipeline Defect Magnetic Flux Leakage Inspection

针对管道缺陷漏磁检测信号中存在严重噪声干扰的问题,将经验模态分解方法用于漏磁检测信号的噪声分离和有效信号提取,对实际测试的与输油管道材质相同且具有人为模拟缺陷的漏磁信号进行处理,结果表明,该方法可以很好地抑制噪声从而得到清晰的、表征缺陷特征的有用信号,达到与小波变换相同的处理效果,同时克服了小波方法中基函数选择困难的问题。

To the bad noise interference problem occurred in the signal of pipeline defect magnetic flux leakage inspection, the empirical mode decomposition（EMD） method developed recently was used to the noise suppressing and the effective signal extraction for the signals of magnetic flux leakage inspection. After introducing to the EMD algorithm simply, the actual magnetic flux leakage signals, measured by the sensors of hall and gaint magneto--resistance from the pipes which had simulative defects, were processed by the EMD method. The results show this method can extract the vivid defect signals by suppress noise effectively, and it has the same effect from the wavelet transform, but it overcomes the difficulty of selecting primary functions, which are vital for the wavelet method.