Open AccessVibration Anomaly Detection Using Multivariate Time Series
Author(s) -
Gicu Călin Deac,
Gicu Călin Deac,
Radu Constantin Parpală,
Cosmin Popa,
Constantin-Adrian Popescu
Publication year - 2022
Publication title -
international journal of modeling and optimization
Language(s) - English
Resource type - Journals
ISSN - 2010-3697
DOI - 10.7763/ijmo.2022.v12.801
Subject(s) - multivariate statistics , series (stratigraphy) , computer science , set (abstract data type) , anomaly (physics) , anomaly detection , recurrent neural network , vibration , data set , time series , pattern recognition (psychology) , artificial intelligence , algorithm , artificial neural network , data mining , machine learning , geology , acoustics , paleontology , physics , condensed matter physics , programming language
The paper presents a set of deep learning algorithms for detecting vibration anomalies in bearings using multivariate time series on datasets provided by Case Western Reserve University. The study considers a problem of multiclassification of the condition of the bearings depending on the type of defect, but also on the degree of defect, considering only punctual defects in an incipient phase. Once the data sets are correctly labeled and the algorithms are trained on this data, they can accurately predict the type and the size of defect. The model with the best results in the set is RNN - CNN (Recurrent Neural Network with Convolutions) giving an accuracy greater than 97% in all (load) cases.