AINeutralarXiv – CS AI · Jun 195/10
🧠
Leveraging systems' non-linearity to tackle the scarcity of data in the design of Intelligent Fault Diagnosis Systems
Researchers propose a novel Deep Transfer Learning approach for Intelligent Fault Diagnosis Systems that addresses data scarcity by leveraging system non-linearities and multi-excitation vibration analysis. The method combines pre-trained CNNs with a new data visualization and augmentation technique, validated on railway pantograph structures.
