1-D Residual Convolutional Neural Network coupled with Data Augmentation and Regularization for the ICPHM 2023 Data Challenge

TL;DR

This paper introduces a residual CNN model with data augmentation and regularization for fault classification in gearbox vibration data, achieving high accuracy across varied operating conditions with a compact model.

Contribution

A novel residual CNN architecture combined with data augmentation and regularization techniques for industrial fault diagnosis using raw vibration signals.

Findings

High classification accuracy on real-world data

Effective across multiple operating conditions

Model has less than 30,000 parameters

Abstract

In this article, we present our contribution to the ICPHM 2023 Data Challenge on Industrial Systems' Health Monitoring using Vibration Analysis. For the task of classifying sun gear faults in a gearbox, we propose a residual Convolutional Neural Network that operates on raw three-channel time-domain vibration signals. In conjunction with data augmentation and regularization techniques, the proposed model yields very good results in a multi-class classification scenario with real-world data despite its relatively small size, i.e., with less than 30,000 trainable parameters. Even when presented with data obtained from multiple operating conditions, the network is still capable to accurately predict the condition of the gearbox under inspection.