Abstract: One of the core works of analyzing Electrochemical Impedance Spectroscopy (EIS) data is to select an appropriate equivalent circuit model to quantify the parameters of the electrochemical reaction process. However, this process often relies on human experience and judgment, which will introduce subjectivity and error. In this paper, an intelligent approach is proposed for matching EIS data to their equivalent circuits based on the Random Forest algorithm. It can automatically select the most suitable equivalent circuit model based on the characteristics and patterns of EIS data. Addressing the typical scenario of metal corrosion, an atmospheric corrosion EIS dataset of low-carbon steel is constructed in this paper, which includes five different corrosion scenarios. This dataset was used to validate and evaluate the proposed method in this paper. The contributions of this paper can be summarized in three aspects: (1) This paper proposes a method for selecting equivalent circuit models for EIS data based on the Random Forest algorithm. (2) Using authentic EIS data collected from metal atmospheric corrosion, the paper establishes a dataset encompassing five categories of metal corrosion scenarios. (3) The superiority of the proposed method is validated through the utilization of the established authentic EIS dataset. The experiment results demonstrate that, in terms of equivalent circuit matching, this method surpasses other machine learning algorithms in both precision and robustness. Furthermore, it shows strong applicability in the analysis of EIS data.

Key words: Electrochemical impedance spectroscopy, Random forest, Corrosion, Equivalent circuit model