Abstract
A fault diagnosis method using improved pattern spectrum and fruit fly optimization algorithm–support vector machine is proposed. Improved pattern spectrum is introduced for feature extraction by employing morphological erosion operator. Simulation analysis is processed, and the improved pattern spectrum curves present a steady distinction feature and smaller calculating amount than pattern spectrum method. Support vector machine with fruit fly optimization algorithm which can help seeking optimal parameters is employed for pattern recognition. Experiments were conducted, and the proposed method is verified by roller bearing vibration data including different fault types. The classification accuracy of the proposed approach reaches 87.5% (21/24) in training and 91.7% (44/48) in testing, showing an acceptable diagnosis effect.
Highlights
A roller bearing is a sort of key mechanical component, and it is meaningful to propose an efficient fault diagnosis technique for its reliability
Some non-linear analytical techniques offer a feasible way in signal processing of roller bearing considering the non-linear character, such as information entropy analysis[4] and fractal dimensions.[5]
Among the non-linear analytical techniques, the mathematical morphological particle analysis is able to describe the shape changing information on different scales, and it is apt to analyze the signal essence on different levels.[6]
Summary
A roller bearing is a sort of key mechanical component, and it is meaningful to propose an efficient fault diagnosis technique for its reliability. Considering different features of morphological operators, morphological erosion operator is introduced and a definition of IPS proposed. Analyzing simulation signals with IPS proposed, the unit structuring element is set as g = [0 0 0]. Comparing with traditional PS method, three PS curves are shown in Figure 2(b) with the same g and lmax as IPS It seems that there is no monotone trend and the D-values are not consistent that presenting a bad distinction of different fault types. IPS curves with rectilinear structuring element are shown, which have no apparent distinction contrast with Figure 2(b); g =[0 0 0] is a good choice considering the calculation amount and effect. We set structuring element as g = [0 0 0], and the IPS curves are shown in Figure 5 when lmax = 80.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
