Bearing Fault Diagnosis Based on Flat Attenuated Triangle Structuring Element

The performance of morphological filtering is highly determined by structuring elements (SEs). In real applications, flat linear SEs are quite often applied. However, they are little similar to the impulsive signals sampled from fault bearings. In order to tackle this situation, flat attenuated triangle SE is presented. It is validated that the proposed SE is effective to extract the impulsive features of the bearings with an outer and an inner race faults respectively. At the same time, conventional flat linear SE is used for further comparisons. The results show that the morphological filtering can extract much more impulses due to applying the flat attenuated triangle SEs.