Deep Learning–Based Predictive Maintenance of Rotating Machinery Using Vibration and Acoustic Signals

Abstract

Predictive maintenance has emerged as a key enabler of intelligent manufacturing systems by reducing unplanned downtime and maintenance costs in rotating machinery. Traditional condition monitoring techniques rely heavily on handcrafted features and expert knowledge, which often fail to generalize under complex operating conditions. This paper presents a deep learning–based predictive maintenance framework for rotating machinery using vibration and acoustic signals. Multisensor data are collected from machinery under normal and faulty operating conditions, including bearing defects and imbalance faults. A hybrid deep learning model combining convolutional neural networks (CNN) and long short-term memory (LSTM) networks is employed to automatically learn spatial and temporal features from raw signals. Experimental results demonstrate that the proposed approach achieves superior fault detection accuracy and remaining useful life prediction performance compared to conventional machine learning methods. The results confirm the effectiveness of deep learning for real-time predictive maintenance in smart manufacturing environments.