On Detection and Classification of State Changes in Physical Processes by Signal Processing Techniques

Dorel Aiordachioaie

2022

Abstract

The paper presents a comparation of some methods based on signal processing, for detection and classification of the changes in the states of physical processes. The investigated processes generate mechanical vibrations, which are properly managed for computer-based processing. As a case study, the incipient faults in bearings are considered. Five signal processing methods are promoted, which are based on statistical processing, signal modelling, spectral analysis, time-frequency image processing, and information modelling. Statistical processing method considers ten features based on statistical moments of various orders. The method of signal modelling involves models with parameters estimated by an Kalman estimator. The method of spectral analysis considers the power spectrum of the vibration’s signals. The method based on time-frequency analysis considers the short-time Fourier transform and Choi-Williams transform for feature extraction. The information-based method is based on information source identification and processing. The classifier is based on similarity comparation using a distance-based classifiers. The novelty/contribution of the paper is the evaluation of the methods for change detection and diagnosis, all based on signal processing paradigm. Each of the considered method has advantages and disadvantages and depends on available data. The involved techniques could be applied also in process monitoring and conditioning.

Paper Citation

in Harvard Style

Aiordachioaie D. (2022). On Detection and Classification of State Changes in Physical Processes by Signal Processing Techniques. In Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC; ISBN 978-989-758-622-4, SciTePress, pages 440-445. DOI: 10.5220/0011950800003612

in Bibtex Style

@conference{isaic22,
author={Dorel Aiordachioaie},
title={On Detection and Classification of State Changes in Physical Processes by Signal Processing Techniques},
booktitle={Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC},
year={2022},
pages={440-445},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011950800003612},
isbn={978-989-758-622-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Symposium on Automation, Information and Computing - Volume 1: ISAIC
TI - On Detection and Classification of State Changes in Physical Processes by Signal Processing Techniques
SN - 978-989-758-622-4
AU - Aiordachioaie D.
PY - 2022
SP - 440
EP - 445
DO - 10.5220/0011950800003612
PB - SciTePress