Research on artificial intelligence recognition of 3D printer operation conditions based on acoustic signal

Monitoring the operating condition of the equipment is crucial. This study explores the feasibility of using transfer learning methods and acoustic signals to identify the operating conditions of 3D printers. A sound sensor probe was designed and fabricated to collect acoustic signals generated during the operation of the 3D printer. Sound recognition models based on ResNet34 and transfer learning were constructed to monitor the operating condition of the 3D printer. Comparative experiments demonstrated that the sound recognition model based on transfer learning showed significant improvements in convergence speed and classification accuracy, achieving a classification accuracy of 97.6%. Furthermore, the acoustic signals of the 3D printer were converted into spectrograms and Mel-Frequency Cepstral Coefficients which were then input into the transfer learning model for training. The results indicated that the overall performance of the model was optimal when the input features were spectrograms.