Morphogram: A simple and useful way to determine the optimal length of flat SE used in morphological analysis for bearing fault feature extraction

Morphological analysis is a relatively new signal processing method for the extraction of bearing fault features. Its main idea is to intersect the morphological shape of a signal with a structuring element (SE). Among all different SEs, flat structuring element is popularly used in morphological analysis because of its straightforward shape and effectiveness in revealing bearing fault cyclic features. However, as of today, the optimal length selection of morphological analysis with SE is still difficult to be determined. If the length of flat SE is improperly selected, bearing fault cyclic features will be destroyed. In order to solve this problem, this paper reports an accurate method for the optimal length selection of flat SE for bearing fault diagnosis. The effectiveness of the proposed method is validated by a simulated bearing fault signal and an industrial motor fault signal. The results obtained by the proposed method illustrate that it is crucial to establish the optimal length of flat SE for enhancing bearing fault cyclic features. Besides, an empirical rule for the length selection of flat SE is not always reliable.