Industrial Fault Classification based on Vibration Signals: A Comparative Study between Transformers and Classical Machine Learning Models

This study investigates deep learning models for vibration-based industrial fault classification, comparing the Transformers architecture against Multilayer Perceptron (MLP), Convolutional Neural Networks (CNN), and Recurrent Neural Networks (RNN). The dataset contains time-series data from two gyroscopic sensors coupled to a rotary test bench subjected to four distinct operational configurations. Data pre-processing occurred in the frequency domain, and predictive efficacy was evaluated via the F1-Score metric. The results indicate that although the Transformers model exhibits significantly faster training convergence, the RNN architecture achieved superior overall predictive performance. This research confirms the adequacy of RNNs for lower-dimensionality structured time-series data and establishes a theoretical foundation for adopting Transformers in highly complex multivariate datasets in future research.