supervised anomaly detection of electric motors based on acoustic signals in noisy environments

In this study, the sound signals generated by motors in noisy environments were usedfor fault detection. to extract relevant features from these signals, techniques such as short-time fourier transform (stft), and mel-frequency cepstral coefficients (mfcc) were employed. for analysis and fault detection, two advanced deep learning methods including dnn and 1d cnn were used. based on the results, a dnn with stft feature achieved accuracy of 96% and with mfccs feature obtained an accuracy of 99% in fault detection of electric motors in noisy environments.