Intelligent diagnosis method using probability density distribution and principal component analysis — Application on gear rotating machinery

摘要

This paper proposed a novel fault diagnosis method by combining statistics filter (SF) and probability density functions (PDFs). First, the vibration signals are processed using SF to reduce the noise automatically. Second, PDFs is introduced to reflect the features of the vibration signals. The segment values of the PDFs (SVPDFs) are integrated into symptom parameters (ISPs) through multivariable analysis methods and used for recognize signal states instead of the conventional symptom parameters (SPs). In the condition survey step, the parameter of SF, which is named selecting discrimination index (SDI) is optimized according to the accuracy rate of the identification. In the precise diagnosis, the optimized SDI and ISPs are used for diagnosing the state of the signal. The efficacy of this method was confirmed by the results of the condition diagnosis for gears on the experimental device.