基于小波和 EMD 的电阻探针监测信号自适应去噪

目的从监测信号中恢复有效腐蚀信息(长期变化趋势、周期性窄带尖峰),提出一种基于经验模态分解(Empirical Mode Decomposition,EMD)和小波阈值去噪(Wavelet Threshold Denoised,WTD)相结合的自适应去噪算法(EMD-WTD)。方法先将信号进行EMD分解,利用最大信息系数(Max Information Coefficient,MIC)判断噪声主导和有效信号主导信号分量的分界点,然后将噪声主导的信号分量进行自适应小波阈值去噪。最后以人工模拟信号和电阻探针监测信号进行验证。结果 EMD-WTD算法能有效去除噪声,信噪比可提升10 d B以上。结论与多个去噪算法相比,EMD-WTD算法能够更好地保留信号中周期性窄带尖峰信息,为后续准确建立电阻探针监测信号与环境之间的数学模型奠定了基础。

Self-adaptive Denoising of Carbon Steel Corrosion Monitoring Signal Based on EMD and Wavelet

Objective To recover effective corrosion information (long-term trends and periodic narrow-band spikes) from monitoring signal, and propose an adaptive denoising algorithm (EMD-WTD) based on Empirical Mode Decomposition (EMD) and Wavelet Threshold Denoised (WTD). Methods First, the signal was decomposed by EMD. The maximum information coefficient (MIC) was used to judge the demarcation point of the noise-dominated and effective-dominated signal components. Then the noise-dominated signal was subjected to adaptive wavelet threshold denoising. Finally, Analog signal and corrosion monitoring signal were used to validate the proposed algorithm. Results EMD-WTD algorithm could effectively remove noise with improving signal to noise ratio by more than 10dB. Conclusion EMD-WTD algorithm could better preserve the periodic narrow-band spike information than multi-denoising algorithm. Moreover, EMD-WTD algorithm lays the foundation for establishment of mathematical models of between corrosion signals and environment.