Research On Intelligent Fault Diagnosis Of NC Machine Tools Based On Industrial Internet

In the context of "China Manufacturing 2025",the manufacturing industry is in full swing toward the intelligent,digitized direction.As the core of manufacturing,CNC machine tools play an important role in the field of medical devices,space,automobiles.Therefore,once the CNC machine tool has failed,the consequences of the generated are self-evident.In order to ensure the efficient and safe operation of the machine,it is a top priority to the failure of the CNC machine tool.In this paper,the X-axis feed system drives the X-axis feed system as the research object,collects CNC machine current signals,extracts the time domain and frequency domain feature vector,and feeds to the RBF neural network for troubleshooting.The main study of thesis:(1)Analysis of the fault mechanism of linear motor drive feed system.In this paper,the feed system of the CNC camshaft grinding machine is an example,and the composition of the linear motor drive system is first introduced to analyze the fault mechanism;then analyze the failure mechanism of the static pressure rail.Finally,in the absence of the fault of the throttle,the relationship between the motor output current and time is provided by the MATLAB,which provides theoretical basis for the problem of fault diagnosis.(2)Analysis of signal acquisition and communication protocol of CNC machine tools.Data collection is carried out in the experiment using industrial Internet and CNC machine tool acquisition system.The physical layer uses the Wire Shark captain software to obtain the correct packet via the hub,analyzing the address and data of the motor driver,NCU,PLC parameters,through Ethernet,with EBOX wireless gateway and CNC machine tools to acquire motor drive,NCU,Data such as PLC;Network layer development cloud server module,with cloud server for background,through the Modbus protocol,TCP protocol to send cloud server database;application layer is created by the cloud platform,programming the E wireless gateway by PLC programming software,and controls CCC The data of the machine is read from EBOX and then transmitted to the cloud.(3)Feature extraction of the signal.Use the energy loss function to make a wavelet packet basis function,select the DB3 wavelet packet base that is the smallest energy loss.Use it to decompose the current signal to extract energy characteristics.Finally,select the peak,variance,root mean square and wavelet package energy value(low frequency band)of the time domain and frequency domain,as a characteristic vector sample of the diagnostic fault.(4)Creation of neural network fault model.Determine the network structure and network initial parameters,select the fundamental function center and the least squares method with the algorithm.Building a RBF network fault diagnosis model,and finally enter the network with the time domain and frequency domain feature of current signals,training and testing the network model,and the results show that the RBF network can realize the fault diagnosis of CNC machine tools.