Issue Date
10-2024
Abstract
Electrochemical Impedance Spectroscopy (EIS) is a technique used for determining the electrical properties of a material. EIS generates impedance spectra across a frequency range that is unique in a specific material. However, the discrimination of materials is challenging as EIS produces high-dimensional data. In this study, Artificial Neural Networks (ANN) were used as a machine learning algorithm for classification. Nyquist Plots and Principal Component Analysis (PCA) were done prior classification for data visualization purposes. Results showed that the five materials have different Nyquist plot shapes, and 2D PCA is close to being successful in discriminating all materials. Only Paper and Bentonite had intersection when plotted using 2D PCA. ANN obtained a high classification accuracy of 97%, with inference time of 22.41 seconds. It is recommended to extend this research by doing regression analysis on determining intrinsic material properties. Lastly, an end-to-end working application through machine learning deployment was recommended for future use cases.
Source or Periodical Title
UP Los Baños Journal
Volume
22
Issue
1
Page
62-70
Document Type
Article
College
College of Engineering and Agro-Industrial Technology (CEAT)
Recommended Citation
Caduyac, Rob Christian M.; Ramoso, John Paolo A.; Arazas, Jose Pocholo R.; and Tomas, Rock Christian V., "Electrochemical Impedance Spectroscopy based discrimination of materials using Artificial Neural Networks" (2024). Journal Article. 6195.
https://www.ukdr.uplb.edu.ph/journal-articles/6195