Bearing Fault Diagnostics Based on the Square of the Amplitude Gains Method

Abstract

The article presents an adaptation of a parametric diagnostic method based on the square of the amplitude gains model, which was tested in experimental studies on bearing damage detection (outer race, inner race, bearing balls damage). The described method is based on the shaft displacement signal analysis, which is affected by vibrations coming from the bearings. The diagnostic model’s parameters are determined by processing the signal from the time domain to the frequency domain in a few steps. Firstly, the recorded signal is divided into two observation periods, next the analytical autocorrelation functions are determined and approximated by a polynomial. Then, the diagnostic thresholds are adopted, and the model parameters are converted into damage maps that are easy to interpret and assess the technical condition of the bearings. The presented method shows the technical condition of bearings in a qualitative way. Depending on the received color damage maps, it is possible to determine their level of wear. Green and blue indicate poor wear or no damage, red indicates increased wear, and black clearly indicates a damaged bearing.