Abstract
This paper describes the implementation of a vibration analysis in an industrial servo motor for anomaly detection. For this, a test bench was built with the purpose of simulating an industrial process. The vibration analysis was performed with an accelerometer which took the acceleration data from a running engine. For the detection of anomalies, an Autoencoder was used which was trained with samples of the normal operation of the motor in order to reconstruct a “normal operation” signal. Once the model was trained, the MAE (Mean Absolute Error) is used to see the differences between the analyzed signal and the one reconstructed by the Autoencoder, if the difference is greater than a threshold, the signal is classified as an anomaly. The proposed methodology represents an alternative to perform vibration analysis in rotative machines and can be used to conduct predictive maintenance in several industrial processes.
Original language | English |
---|---|
Title of host publication | Advances in Computational Intelligence - 21st Mexican International Conference on Artificial Intelligence, MICAI 2022, Proceedings |
Editors | Obdulia Pichardo Lagunas, Bella Martínez Seis, Juan Martínez-Miranda |
Publisher | Springer Science and Business Media Deutschland GmbH |
Pages | 252-265 |
Number of pages | 14 |
ISBN (Print) | 9783031194955 |
DOIs | |
State | Published - 2022 |
Event | 21st Mexican International Conference on Artificial Intelligence, MICAI 2022 - Monterrey, Mexico Duration: 24 Oct 2022 → 29 Oct 2022 |
Publication series
Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
---|---|
Volume | 13613 LNAI |
ISSN (Print) | 0302-9743 |
ISSN (Electronic) | 1611-3349 |
Conference
Conference | 21st Mexican International Conference on Artificial Intelligence, MICAI 2022 |
---|---|
Country/Territory | Mexico |
City | Monterrey |
Period | 24/10/22 → 29/10/22 |
Bibliographical note
Publisher Copyright:© 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.
Keywords
- Anomaly detection
- Artificial intelligence
- Autoenconder
- Industry 4.0
- Machine learning
- Neural networks
- Predictive maintenance