Publication detail
Utilization of Machine Learning in Vibrodiagnostics
ZUTH, D. MARADA, T.
English title
Utilization of Machine Learning in Vibrodiagnostics
Type
journal article in Scopus
Language
en
Original abstract
The article deals with possibilities of use machine learning in vibrodiagnostics to determine a fault type of the rotary machine. Sample data are simulated according to the expected vibration velocity waveform signal at a specific fault. Then the data are pre-processed and reduced for using Matlab Classification Learner which creates a model for identifying faults in the new data samples. The model is finally tested on a new sample data. The article serves to verify the possibility of this method for later use on a real machine. In this phase is tested data preprocessing and a suitable classification method.
English abstract
The article deals with possibilities of use machine learning in vibrodiagnostics to determine a fault type of the rotary machine. Sample data are simulated according to the expected vibration velocity waveform signal at a specific fault. Then the data are pre-processed and reduced for using Matlab Classification Learner which creates a model for identifying faults in the new data samples. The model is finally tested on a new sample data. The article serves to verify the possibility of this method for later use on a real machine. In this phase is tested data preprocessing and a suitable classification method.
Keywords in English
Classification learner, Classification method, Machine learning, Matlab, Neuron network, Parallel Misalignment, PCA, Static unbalance, SVN, Vibrodiagnostics
Released
05.08.2018
Publisher
Springer Verlag
Location
Cham
ISBN
978-3-319-97887-1
ISSN
2194-5357
Book
Recent Advances in Soft Computing
Number
2017
Pages from–to
271–278
Pages count
8
BIBTEX
@article{BUT149453,
author="Daniel {Zuth} and Tomáš {Marada},
title="Utilization of Machine Learning in Vibrodiagnostics",
booktitle="Recent Advances in Soft Computing",
year="2018",
number="2017",
month="August",
pages="271--278",
publisher="Springer Verlag",
address="Cham",
isbn="978-3-319-97887-1",
issn="2194-5357"
}