Bearing Fault Diagnosis with Quantum Machine Learning: Experimental Case Study on a Test Bench

This work investigates the application of Quantum Machine Learning (QML) techniques for the diagnosis of failures in rotating machines, focusing on the classification of the state of integrity of bearings from vibration signals. For this, an experimental bench was used that simulates a mechanical transmission system, operating under controlled conditions that represent healthy bearings, with mild damage and severe damage. Vibration data were collected by piezoelectric accelerometers in constant rotation. The resulting dataset was used in the training and evaluation of quantum neural networks, including Quantum Convolutional Neural Networks (QCNN), built with parameterized quantum circuits (PQC). The experiments showed that QCNN achieved an accuracy of 74.17% with eight input characteristics, presenting competitive performance compared to other quantum architectures evaluated.