FE-DGCAN DataSet

This dataset is designed for the task of mechanical anomaly sound detection. It consists of audio recordings in .wav format, each with a duration of 10 seconds. The dataset is divided into two subsets: The recordings include various types of mechanical operation sounds, encompassing both normal and abnormal conditions. This dataset provides a reliable benchmark for developing and assessing machine learning models aimed at detecting and classifying anomalous mechanical sounds, thereby contributing to the advancement of intelligent fault diagnosis and predictive maintenance systems.