Detección y diagnóstico de fallas en motores mediante el análisis de vibraciones aplicando técnicas de inteligencia artificial.

Currently, the detection and diagnosis of motor faults in industry is essential for reducing both machinery (motor) maintenance costs and mitigating production interruptions. Vibration analysis establishes a competent technique for detecting faults or malfunctions in motors, and its level of accurac...

Full description

Saved in:
Bibliographic Details
Main Authors: Jair Elías Araujo Vargas, Dilan Yesid Franklin Coronel, Victor Manuel Arias Ruiz
Format: Article
Language:English
Published: Fundación de Estudios Superiores Comfanorte 2023-01-01
Series:Mundo Fesc
Subjects:
Online Access:https://www.fesc.edu.co/Revistas/OJS/index.php/mundofesc/article/view/1652
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Currently, the detection and diagnosis of motor faults in industry is essential for reducing both machinery (motor) maintenance costs and mitigating production interruptions. Vibration analysis establishes a competent technique for detecting faults or malfunctions in motors, and its level of accuracy has increased thanks to the combined application of artificial intelligence. Therefore, the performance of different artificial intelligence algorithms in the field of fault detection and diagnosis using vibration analysis in motors was evaluated. An exhaustive comparison of the results obtained with each algorithm was carried out using metrics and evaluation techniques in order to determine which is the most effective. In this specific task, indicators such as the precision, sensitivity and specificity of the algorithms or aspects such as vibration signal conditioning techniques, extraction methods, selection of key features, training of artificial intelligence models, neural networks and support vector machines were taken into account. Thus, artificial intelligence algorithms demonstrated high accuracy in fault detection and resolution, identifying various types of problems in a timely manner; the models contributed significantly to the overall analysis, offering a more reliable approach to predictive industrial maintenance, paving the way for future improvements and the adoption of new, more robust and adaptable algorithms.
ISSN:2216-0353
2216-0388