Litcius/Paper detail

Intelligent fault diagnosis method for rolling bearing using WMNRS and LSSVM

Xuezong Bai, Shilong Zeng, Qiang Ma, Zihao Feng, Zongwen An

2023Measurement Science and Technology17 citationsDOI

Abstract

Abstract A weighted multi-neighborhood rough set (WMNRS) algorithm is designed to resolve the issue in which the neighborhood radius must be adjusted iteratively and cannot be automatically determined in the neighborhood rough set. This algorithm combined with the least squares support vector machines (LSSVM) is used for analyzing rolling bearing condition monitoring data; consequently, an intelligent fault diagnosis method is proposed. Specifically, the time-domain and frequency-domain features are extracted from the collected vibration signals to construct an original feature set. The WMNRS algorithm is then applied to screen the primary sensitive components from the constructed feature set. Finally, an optimized LSSVM is utilized to recognize the fault types. The developed method is validated on a public dataset and a measured rolling bearing dataset. The results demonstrate that the method achieves excellent diagnostic performance. Furthermore, the proposed method has some supremacy regarding running time.

Topics & Concepts

Fault (geology)Computer scienceBearing (navigation)Rough setSet (abstract data type)Support vector machineFeature (linguistics)Pattern recognition (psychology)VibrationDomain (mathematical analysis)Data miningConstruct (python library)Time domainFeature vectorArtificial intelligenceAlgorithmMathematicsComputer visionQuantum mechanicsProgramming languageLinguisticsPhilosophyGeologyMathematical analysisPhysicsSeismologyMachine Fault Diagnosis TechniquesGear and Bearing Dynamics AnalysisRough Sets and Fuzzy Logic