Imagen de Portada

Classification of Bearing Fault Signals in Rotating Machinery Using Neural Networks

Guardado en:
Publicado en:Journal Europeen des Systemes Automatises vol. 58, no. 1 (Jan 2025), p. 89-97
Autor principal: Retz Mahima Devarapalli
Otros Autores: Kontham, Raja Kumar, Kontham, John David Christopher
Publicado:
International Information and Engineering Technology Association (IIETA)
Materias:
Predictive maintenance
Accuracy
Datasets
Deep learning
Wavelet transforms
Fourier transforms
Fault diagnosis
Bearings
Machinery
Sensors
Neural networks
Classification
Breakdowns
Vibration
Acceso en línea:Citation/Abstract
Full Text - PDF
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
Resumen:Rotating Machinery is a vital component in the manufacturing process. Its health conditions directly affect production, and any failure of the Machinery may reduce production and cause accidents. Condition-based monitoring detects faults in the early stages, which, in turn, reduces machine failures. Machine learning condition monitoring has made remarkable achievements in fault detection, but it requires various feature calculations and is a time-consuming process. Recently, deep learning-based models outperformed traditional machine learning techniques as they automatically identify features through the learning process. This paper proposes a deep-learning model to classify bearing faults, specifically a convolution Neural Network Model (CNN) and Convolution Invariant Neural Network (CINN). The bearing dataset from Case Western Reserve University (CWRU) is used for training and testing the proposed CNN and CINN Models. The performance of model is evaluated on different working conditions of the bearing faults with varying loads, demonstrating 99% and above accuracy.
ISSN:1269-6935
2116-7087
DOI:10.18280/jesa.580110
Fuente:ABI/INFORM Global