PCA-MD Planetary Gear Damage Identification Method Based on Acoustic Emission Signals

In order to study the damage identification of planetary gears in a reducer, a PCA-MD (Principal Component Analysis-Mahalanobis Distance) method for damage identification of planetary gears based on Acoustic Emission (AE) signals was proposed. In this paper, AE signals were collected using a DS5-16C Acoustic Emission instrument and RS-2A piezoelectric ceramic sensors, with a sampling frequency of 3MHz and a preamplifier gain of 40dB. First, a set of AE signals from healthy planetary gears was collected as a reference signal, and damage to the planetary gears was simulated on the experimental platform, collecting AE signals in the damaged state. Then, noise interference in the signals was removed through Empirical Mode Decomposition (EMD), and 13 feature parameters of the denoised AE signals were calculated. Through PCA, the 13-dimensional feature parameters were reduced to 2 dimensions, simplifying the complexity of data processing without losing the original valid information. Finally, the parameters obtained after dimensionality reduction were combined with Mahalanobis Distance (MD) to further establish a method for identifying damage in planetary gears of reducers. The damage condition of the planetary gears was judged by calculating the Mahalanobis distance between healthy and damaged signals. The experimental results showed that the PCA-MD method could identify early damage in planetary gears.