Methods of trend removal in electrochemical noise data - Overview

In this paper we shall review popular methods of trend removal from electrochemical noise time records. The basic principles of operation of the six most popular methods are explained. The proposed methods are: high-pass filtering, Moving Average Removal, polynomial detrending, wavelet detrending, Empirical Mode Decomposition and Variational Mode Decomposition. Estimation of trend removal quality is evaluated using statistical measures like a histogram of noise voltage, power spectral density, the correlation coefficient and signal power. The advantages, disadvantages, limitations and applications of all of the methods mentioned are presented. Two examples of electrochemical noise data with a different nature of generation are used for assessing the efficiency of the presented methods. The first set of measurement data concerning electrochemical noise with a thermal drift were observed during uniform corrosion. The second one refers to noise superimposed on a curve of the discharging current of a supercapacitor. This additive noise component is generated by charge redistribution or redox reactions within porous carbon electrodes. A comparison of these methods and an indication of the most suitable one for removing the drift component from the acquired electrochemical data is summarized in this paper. (C) 2018 Elsevier Ltd. All rights reserved.