Fault feature extraction based on artificial hydrocarbon network for sealed deep groove ball bearings of in-wheel motor

Abstract

Sealed deep groove ball bearings (SDGBBs) are employed to perform the relevant duties of in-wheel motor. However, the unique construction and complex operating environment of in-wheel motor may aggravate the occurrence of SDGBB faults. Therefore, this paper proposed a novel feature extraction from vibration signals which performance should owe to artificial hydrocarbon networks (AHNs). AHNs are a novel machine learning method which inspiration follows from chemical rules of organic chemistry. When a signal is properly divided into small parts for simulating organic structure, and the vibration information of each part are used to represent the behavior of each molecule or compound, the interested information can be packaged perfectly. All packages can retain the nature of the signal. In that sense, a AHNs-based filtering is established to cancel the noise and extract the feature from vibration signals. The presented method in this article has been applied to perform the feature extraction of in-wheel motor SDGBBs' faults, and practical examples have verified that it is feasible and effective to extract the features of vibration signals by AHNs-based filtering.