ABSTRACT
Machinery fault diagnostics is an essential part of condition-based maintenance activity in industry. Many techniques are available, and among these vibration analysis is widely used. In this work, different sizes of artificial defect were seeded on the inner race of a rolling element bearing. The vibration signatures obtained were analyzed and comparisons made between healthy and faulty bearings. All the readings were taken with different loading conditions and different sampling rates. Bearings are loaded with hydraulic cylinders. Vibration signatures were analyzed for pure radial loading condition. The vibration signals obtained were analyzed in the time domain and the frequency domain. Various statistical parameters were measured, namely Root Mean Square (RMS), peak to peak, kurtosis and crest factor. As the fault is present on the inner race, vibration signatures are combined with the side bands and it becomes difficult to identify the defect frequency. In this work, side bands are due to the rotation of the shaft with rotating inner race. In order to remove the side bands, a band-pass filter is used, and a kurtogram is used to select the proper bandwidth. Further, an envelope analysis technique using the Hilbert transform is used for removing side bands.
