Detecting train driveshaft damages using accelerometer signals and Differential Convolutional Neural Networks

Resource information

Authors

Antía López Galdo, Alejandro Guerrero-López, Pablo M. Olmos, María Jesús Gómez García

Research area

Machine learning

Application domain

General AI research

AI methods

Machine learning, Deep learning, Neural networks

Research topics

cs.LG, eess.SP

arXiv record date

2022-11-15

Publication

2023. Engineering Applications of Artificial Intelligence, 126, 106840

DOI

10.1016/j.engappai.2023.106840

Upstream last modified

Feb 27, 2025

Last record update

Nov 4, 2026

Abstract

Railway axle maintenance is critical to avoid catastrophic failures. Nowadays, condition monitoring techniques are becoming more prominent in the industry to prevent enormous costs and damage to human lives. This paper proposes the development of a railway axle condition monitoring system based on advanced 2D-Convolutional Neural Network (CNN) architectures applied to time-frequency representations of vibration signals. For this purpose, several preprocessing steps and different types of Deep Learning (DL) and Machine Learning (ML) architectures are discussed to design an accurate classification system. The resultant system converts the railway axle vibration signals into time-frequency domain representations, i.e., spectrograms, and, thus, trains a two-dimensional CNN to classify them depending on their cracks. The results showed that the proposed approach outperforms several alternative methods tested. The CNN architecture has been tested in 3 different wheelset assemblies, achieving AUC scores of 0.93, 0.86, and 0.75 outperforming any other architecture and showing a high level of reliability when classifying 4 different levels of defects.