Research On The Method Of Fault Diagnosis For Large Machining Equipment Bearing

Bearing as key component of the large mechanical processing equipment, if it was fall into trouble, the entire processing and manufacturing system would be stopped and affect the production of enterprises. In this paper, large mechanical processing equipment as the research object, and the bearing fault diagnosis was the research objectives. The research contents are as follows:(1)The operation characteristics of large mechanical processing equipment were researched, analyzing the methods of obtaining bearing vibration signal. The method for the continuation of periodic extreme points was proposed to solve the problem of end effect problem based on the theory studying of local mean decomposition(LMD)algorithm. The problem of false component was solved by using the correlation analysis method. And the new method to solve problems was verified by carried out simulation experiments. The mathematical morphology was used to noise reduction the PF components obtained by LMD decomposition, and an adaptive triangular structure element was proposed. A dual LMD signal processing method of "LMD- morphological noise reduction-LMD" was proposed. This method could reduce the noise and decompose the signal. And the main PF components was selected based on the analysis of the bearing structure and was used to extract energy features.(2)The basic theory of LSSVM method was studied, and a signal kernel function was proposed based on the feature vector. The influence of LSSVM kernel parameter and adjustable parameter on the classification was studied, and the PSO method was used to optimize the kernel parameters and adjustable parameters of LSSVM. The basic theory of PSO was studied, and the chaotic method was applied to enrich the population of PSO. The dynamic parameter control was used to balance the local and global search capability of PSO. The new method was proposed by improvement the PSO-LSSVM method to pattern recognition for bearing feature vector.(3)The bearing fault diagnosis system for large mechanical processing equipment was compiled based on the theoretical study about bearing fault diagnosis of large-scale machinery processing equipment, and designed on the LabVIEW software and combined with the MATLAB software. The online diagnosis of bearing fault was realized through the system.An experiment was done based on the theory of the dual LMD signal decomposition method and the improved PSO-LSSVM fault recognition method and the results obtained were consistent with the actual situation. The experimental results showed that the method proposed in this paper was feasible. And it had important reference value for the realization of the periodical maintenance to the bearing of large mechanical processing equipment.