A Frequency Adaptive Vibration Analysis Method for Variable Speed Rotating RV Reducer

Industrial robots commonly perform cyclic reciprocating motions during operation. The dynamic loads caused by frequent acceleration and deceleration make the RV reducers of the robots prone to failure and generate abnormal vibration during operation. However, the variable speed condition makes it difficult to analyze the frequency of the vibration signal. This paper proposes a frequency-adaptive method to analyze the vibration signals to identify the frequency range of vibration energy distribution. A filter bank is designed based on this frequency range, enhancing the target frequency while lowering other unnecessary frequency components of the Short-Time Fourier Transform (STFT) spectrogram. Features extracted from the enhanced spectrograms were employed to identify and classify various failure types. In experimental trials involving a six-joint serial industrial robot, the frequency-adaptive approach achieved a classification accuracy of 93.06% across 108 trials. These results suggest that the proposed method is highly effective for monitoring variable speed conditions and diagnosing faults in RV reducers.