Time-Frequency Multi-resolution Analysis Applied To Electrochemical Technique Of Metal Corrosion

The research of metal corrosion has important meaning in the aspect of marine project, the petroleum platform of well drilling, the structural thing of reinforced concrete and so on. To do this job, the mensuration of corrosion speed is very important. Along with the development of electronic industry, the corrosion measure technique using electrochemical principle has progressed at a rapid rate. The coulostatic technique induces an electrical discharge on a cell under test. By analyzing slopes and intersects in the logarithmic graph of the cell time response, the electrical parameters of some predefined network models can be found.However, the signals obtained by compute are affected by noise drastically since the electrode response signal induced by the coulostatic perturbations is faint. The noise degrades the accuracy and precision of the analysis and detection. Signal denoising is therefore highly desirable. In this paper, based on time-frequency multi-resolution analysis theory, wavelet and EMD method are respectively proposed to separate noise from coulostatically induced transients.In this paper, coulostatically induced transient in time domain and its spectrum are proposed first. It is possible to derive the equation for the cell potential response and impedance of electrode using network analysis in the Laplace domain. These equations can also contribute to interpretting coulostatically induced transients of different equivalent circuit model. Then, a method of denoising based on wavelet thresholding for coulostatically induced transients is presented. In this method, noisy data could be extracted by reducing the wavelet coefficients under a quantity equal to the threshold value. Some key factors such as decomposition level and thresholding policy are further discussed. Also, another method based on empirical mode decomposition (EMD) to denoise coulostatically induced transient signal is presented.