Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods

© 2017 By use of recurrence quantification analysis (RQA), twelve features were extracted from the electrochemical noise signals generated by three types of corrosion: uniform, pitting and passivation. Machine learning methods, i.e. linear discriminant analysis (LDA) and random forests (RF), were us...

Full description

Bibliographic Details
Main Authors: Hou, Y., Aldrich, Chris, Lepkova, Katerina, Machuca Suarez, Laura, Kinsella, Brian
Format: Journal Article
Published: Pergamon 2017
Online Access:http://hdl.handle.net/20.500.11937/57708
_version_ 1848760075773018112
author Hou, Y.
Aldrich, Chris
Lepkova, Katerina
Machuca Suarez, Laura
Kinsella, Brian
author_facet Hou, Y.
Aldrich, Chris
Lepkova, Katerina
Machuca Suarez, Laura
Kinsella, Brian
author_sort Hou, Y.
building Curtin Institutional Repository
collection Online Access
description © 2017 By use of recurrence quantification analysis (RQA), twelve features were extracted from the electrochemical noise signals generated by three types of corrosion: uniform, pitting and passivation. Machine learning methods, i.e. linear discriminant analysis (LDA) and random forests (RF), were used to identify the different corrosion types from those features. Both models gave satisfactory performance, but the RF model showed better prediction accuracy of 93% than the LDA model (88%). Furthermore, an estimation of the importance of the variables by use of the RF model suggested the RQA variables laminarity (LAM) and determinism (DET) played the most significant role with regard to identification of corrosion types. In addition, the comparison of noise resistance with the resistance obtained from EIS measurement showed that the noise resistance can be used for monitoring corrosion rate variations not only for uniform corrosion and passivation, but also for pitting.
first_indexed 2025-11-14T10:10:01Z
format Journal Article
id curtin-20.500.11937-57708
institution Curtin University Malaysia
institution_category Local University
last_indexed 2025-11-14T10:10:01Z
publishDate 2017
publisher Pergamon
recordtype eprints
repository_type Digital Repository
spelling curtin-20.500.11937-577082017-11-20T08:58:16Z Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods Hou, Y. Aldrich, Chris Lepkova, Katerina Machuca Suarez, Laura Kinsella, Brian © 2017 By use of recurrence quantification analysis (RQA), twelve features were extracted from the electrochemical noise signals generated by three types of corrosion: uniform, pitting and passivation. Machine learning methods, i.e. linear discriminant analysis (LDA) and random forests (RF), were used to identify the different corrosion types from those features. Both models gave satisfactory performance, but the RF model showed better prediction accuracy of 93% than the LDA model (88%). Furthermore, an estimation of the importance of the variables by use of the RF model suggested the RQA variables laminarity (LAM) and determinism (DET) played the most significant role with regard to identification of corrosion types. In addition, the comparison of noise resistance with the resistance obtained from EIS measurement showed that the noise resistance can be used for monitoring corrosion rate variations not only for uniform corrosion and passivation, but also for pitting. 2017 Journal Article http://hdl.handle.net/20.500.11937/57708 10.1016/j.electacta.2017.09.169 Pergamon restricted
spellingShingle Hou, Y.
Aldrich, Chris
Lepkova, Katerina
Machuca Suarez, Laura
Kinsella, Brian
Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title_full Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title_fullStr Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title_full_unstemmed Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title_short Analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
title_sort analysis of electrochemical noise data by use of recurrence quantification analysis and machine learning methods
url http://hdl.handle.net/20.500.11937/57708