Condition Monitoring Technology Research Of Machining Center Spindle System Based On Support Vector Machine

With the coming of the Information Age,The manufacturing industry and the information technology have further integration and development.Based on the information system,the intelligent equipment and the intelligent factory are leading the change of the manufacturing way.The traditional method of CNC machine tools could no more meet the requirements of the highly effective and automatic machining process.The implementation of CNC machine tools condition and failure monitoring has great significance to improve the production efficiency,to reduce manpower cost and to raise the level of automation in production process.The VDL-1000 machining center spindle system was the monitoring object of the monitoring system.The system used the multiple sensors and the data acquisition card and other structures to build the hardware system,and it used some algorithms such as the support vector machine(SVM)to write the software system.Finally through the related examinations we could verify the feasibility of the monitoring system.The monitoring system not only could improve the automation level of the machining center,but also could provide a foundation of the subsequent failure early warning and diagnosis of the machining process.The paper firstly studied the main structure of the VDL-1000 machining center and its spindle system.And it has carried on the statistics and analysis of the common failures of the spindle system.On the basis,the monitoring parameters of the spindle condition monitoring system were determined,and the overall scheme of the monitoring system was proposed.According to the expected function of monitoring system,the paper determined the signal acquisition,the processing and the analysis technology and introduced them.The paper cheese the current transformer,the SMD type temperature sensor,the acceleration sensor and the frequency transmitter to finish the measurements of the spindle system parameters such as the current,the temperature,the vibration and the speed.It used the sensors,the data acquisition card and industrial computer to complete the building of hardware system.The software system of the monitoring system was written by the Lab VIEW programming based on MATLAB,and it had many other functions such as acquisition,displaying,filtering and analysis of the signals.The basic information management module could log in the system through password authentication,and at the same time it has the user information management,the system information management,the passwords changed management,and other functions;the data acquisition module could realize the collection of the signals through library function of the PCI8602.DLL dynamic link library;The data storage module could realize the data real-time storage and save the data in the form of a spreadsheet,and it facilitated the subsequent view and use of the data;The condition monitoring used the display controls of Lab VIEW for the real-time display of each monitoring signal;The signal processing and analysis module as the core modules of the software system,it could use the low pass filter and the wavelet noise reduction based on Birge-Massart threshold to finish the filter processing of the different signals in order to obtain more real signals and facilitate subsequent signal analysis;It could also acquire the time domain characteristics of the signals through the time-domain analysis;It used signal auto-correlation analysis to collect the cycle characteristics of the signals;And through the signal spectrum analysis and wavelet analysis it could get time and frequency domain feature of the signals.In addition,it provided the state recognition algorithm based on support vector machine(SVM)and the algorithm could be used in the failure state recognition.After completing the spindle condition monitoring system,the paper has carried on the verification experiment for the function of the system,and it proved the feasibility of the system.Finally the paper used the bearing failure monitoring case and the looseness failure monitoring case to verify the effectiveness of the failure monitoring function of the system.In this paper,the development of spindle condition monitoring system based on support vector machine(SVM)had the characteristics of simple and practical,fully functional,and it achieved the expected design goal.So it had high application value in the study of the condition monitoring of CNC machine tools.