Research On Incipient Fault Diagnosis Of Rolling Element Bearing Based On Improved Autogram For Computer Numerical Control Machine Tools

As a key supporting part widely used in the spindle system of CNC machine tools,the running state of rolling element bearings closely affects the overall performance and processing quality of mechanical equipment.Therefore,it is necessary to monitor and diagnose the health status of rolling element bearings in real time.The ultimate goal of this article is to develop a software for early fault diagnosis of rolling bearings based on Linux operating system for Huazhong No.9 CNC system,the core diagnostic method based on software is the Envelope Analysis,however,the selection of demodulation frequency band is a difficult problem which has been concerned and discussed by many scholars,and this article also focuses on this topic.The main research contents of this article are as follows:Two mature demodulation frequency band selection method is studied,one is the Fast Kurtogram,the other is the latest Autogram,the limitations of two algorithms and the reasons are explored and explained.In order to overcome the shortcoming of the Spectral Kurtosis in extracting narrow-band fault features from cyclostationary signals,combined with the principle of Spectral Negentropy,an Improved Weighted Spectral Negentropy is created.The narrow band which possesses both the impulsive in time domain and the cyclostationary in frequency domain can be evaluated quickly by the new indicators.To overcome the shortcomings of the two algorithms,an Improved Autogram algorithm is proposed by combining the Improved Weighted Spectral Negentropy.The new algorithm is faster than the original Autogram algorithm and has a higher success rate,at the same time,a more refined narrow-band selection is provided,which makes the new algorithm more suitable for fault diagnosis of rolling element bearings in industrial field.Based on the public authoritative vibration data sets of rolling element bearings with faults and the real vibration signals collected from the experimental platform,compared with the Fast Kurtogram and the Autogram,the superior diagnostic ability of the new algorithm in several kinds of strong noise disturbance is verified.Finally,a software for early fault diagnosis of rolling element bearings under Linux operating system was developed for Huazhong No.9 CNC system.Both the function and beautification are taken into account in the design of the software which is easy to install and use,and the professional knowledge requirements of the operators are reduced and simplified,some technical support and basic reference are provided for building a complete intelligent fault diagnosis system for domestic high-end precision CNC machine tools in the future.