Using spectral kurtosis for selection of the frequency bandwidth containing the fault signature in rolling bearings

摘要

In the naval and offshore industry, many rotating equipment use rolling bearings and any fault can cause a high impact on the costs related to a possible equipment failure, which are far superior to the cost of replacing the bearing itself. Several reasons can shorten the life of the bearings and lead them to sudden breaks. In this context, the use of fault detection techniques in rolling bearings is extremely useful. The objective of this work is to apply the spectral kurtosis technique to select the frequency bandwidth, which contains the fault signature, for fault detection in three (3) rolling bearings. Bearing # 1 has fault in the cage, bearing # 2 has a fault in the outer race and bearing # 3 has two faults, one in the rolling element (ball) and another in the cage. Based on the difference found between the theoretical fault frequencies and the fault frequencies obtained experimentally, all faults were correctly identified. This difference does not exceed 1.83% in any case.