Hybrid Quantum Convolutional Neural Network for CNC Machine Bearing Fault Detection Using Vibration and Acoustic Signals

Flexible manufacturing systems (FMS) rely heavily on CNC machine tools, and the machines' failure can be attributed to bearing failure. Bearing fault detection is critical in avoiding machine downtime and expediting expensive repair work. To enhance the precision of CNC machine bearing failure detection via vibration and sound signals, the present research suggests a Hybrid Quantum Convolutional Neural Network with Skill Optimization Algorithm (QCNN- SOA). For enhanced defect classification, the method integrates a skill optimization technique with quantum convolutional networks. Preprocessing of signals is performed using the SWVO-RKF to eliminate noise and outliers without distorting fault-related patterns. The Inception Convolutional Vision Transformer (ICVT) model is used for feature extraction to capture local and temporal dependencies. Hybrid QCNN is employed to classify features that are extracted. A classical fully connected layer is employed for classification after employing quantum gates for convolution and encoding of the signal. With an error rate of 0.8%, the proposed method achieves 99.2% accuracy, 99.6% recall, 98.7% precision, and 99.1% F1-score.