Font Size: a A A

Study On Abrasive Signal Processing And Feature Extraction In Lubricating Oil

Posted on:2020-03-28Degree:MasterType:Thesis
Country:ChinaCandidate:J GuoFull Text:PDF
GTID:2392330599460221Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The monitoring of metal particles in lubricating oil of mechanical system can help to make maintenance plan and design reliable preventive maintenance system.However,it is difficult to extract the effective information of the output signal of the oil sensor.In order to solve this problem,first on the basis of the principle of electromagnetic,the field distribution and change of particle sensor have been studied,the mathematical model of tri-power between the abrasive particle size radius and the output signal voltage of the sensor is established.And the output voltage characteristics of the sensor are analyzed,and a method for solving the information of abrasive size,quantity and magnetism based on the characteristics of the output electrical signal,such as peak value and phase,is obtained.Then the information processing method of sensor output signal is analyzed.An improved integrated modal decomposition method(EEMD)was proposed by improving modal decomposition method(EMD).The optimized EEMD parameter values are discussed,and an adaptive method to set the EEMD parameter is proposed to solve the problem of setting the EEMD parameter,EEMD while reduce the EMD some shortcomings,but also brought some side effects to the decomposition results.The adverse reactions were analyzed and the post processing method was adopted.Then by comparing EMD method with EEMD method through simulation data and real data test,the superiority of the improved EEMD method is verified.Finally,by comparing the residual error of the wavelet threshold filtering method and the improved EEMD method with the accuracy of the range of abrasive particle size detection,the accuracy of the improved EEMD method was demonstrated,and the signal extraction ability of the improved EEMD method was verified.At the same time,the model curves of abrasive grain size and the peak value,trough value and peak-to-trough value of output signal wave were fitted,which proved the accuracy of the established abrasive grain size model.The subsection study on the size of abrasive particles is put forward and its feasibility is proved by experimental data.
Keywords/Search Tags:Improved EEMD method, Signal processing, Feature extraction, Wear particle identification, Fault diagnosis
PDF Full Text Request
Related items