Fault Diagnosis for Road Heading Bearings Based on a Multiscale Enhanced Cascaded Difference Filter

Abstract

In this paper, a novel multiscale morphological filter (MMF), called multiscale enhanced cascaded difference filter (MECDF), is proposed for the fault detection of road heading bearings. First, the cascaded morphological operators are established based on the cascade of the basic morphological operators with similar properties, and then the morphological difference operation is introduced to propose the cascaded difference operators. Subsequently, the enhanced cascaded difference operator (ECDO) is constructed through the convolution of cascaded difference operators. Moreover, since the scale range of structure element (SE) also determines the filtering performance of multiscale morphological filter, an improved multiscale analysis method is presented to select the optimal scale range. Finally, the bearing experimental signals are implemented to validate the effectiveness of MECDF. Experimental results testify that the scale range determined by the MECDF is better than other multiscale morphological filters. Meanwhile, the feature extraction capability of ECDO is also better than other existing morphological difference operators.