ABSTRACT This study represents a pioneering leap in predictive maintenance by harnessing the power of deep learning. Focused on the MetroPT dataset and its intricate APU (Air Production Unit) metrics from a train vehicle, ...