Classifying the Percentage of Broken Magnets in Permanent Magnet Synchronous Motors Using Combined Short-Time Fourier Transform and a Pre-Trained Convolutional Neural Network

Keyphrases

• Short-time Fourier Transform 100%
• Pre-trained Convolutional Neural Networks 100%
• Permanent Magnet Synchronous Motor 100%
• Broken Magnet 100%
• Electrical Machines 50%
• Convolutional Neural Network Model 50%
• Pre-trained AlexNet 50%
• Transfer Learning 25%
• Monitoring System 25%
• Finite Element Method 25%
• Electric Vehicle 25%
• Speed Range 25%
• System Performance 25%
• Stator 25%
• Experimental Validation 25%
• Power Density 25%
• System Failure 25%
• AlexNet 25%
• Motor Type 25%
• BOAST 25%
• Area under the Receiver Operating Characteristic Curve 25%
• Model Effectiveness 25%
• Machine Learning Technology 25%
• Training Accuracy 25%
• Magnet Structure 25%
• Training Loss 25%
• Industrial Setting 25%
• Neural Network Method 25%
• Experimental Dataset 25%
• Phase Current 25%
• Simulation Datasets 25%
• Limited Datasets 25%
• AlexNet Neural Network 25%
• New Layers 25%

Engineering

• Short-Time Fourier Transform 100%
• Permanent Magnet Synchronous Motor 100%
• Convolutional Neural Network 100%
• Network Model 50%
• Electrical Machine 50%
• Monitoring System 25%
• Systems Performance 25%
• Electric Vehicle 25%
• Receiver Operating Characteristic 25%
• Operating-Characteristic Curve 25%
• Power Density 25%
• Transfer Learning 25%
• Critical Application 25%
• Learning System 25%
• Finite Element Analysis 25%

Earth and Planetary Sciences

• Synchronous Motor 100%
• Machine Learning 25%
• Monitoring System 25%