ICEEMDAN and LSTM Spindle System Fault Diagnosis Method

Research on fault diagnosis of rotating machinery is crucial as it serves as a core component. In order to diagnose spindle system fault types quickly and accurately, Improved Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (ICEEMDAN) and Long Short-Term Memory (LSTM) fault diagnosis method is presented. Firstly, collect the fault vibration, then ICEEMDAN is used for vibration signals to extract seven times domain features and four frequency domain features. Finally, the time domain and frequency domain features’ parameters are input into the LSTM network for fault classification. The method is verified for the practical industrial CNC machine center spindle system.