Study On Feature Extraction And Diagnosis Of Fault Vibration Signal Of CNC Machine Tool's Angle Head

CNC(computer numerical control)machines play an essential role in the top manufacturing industry including aerospace,high-end automobiles,large-scale rotating machinery and precision machining industry,all of which would be limited without the using of CNC machines.The angle head is an accessory installed on the CNC machines,which makes the machine suitable for more machining conditions and can be used to increase the machining range as well,without changing the overall structure of the machine.Therefore,The installation of the angle head can greatly improve the traditional processing methods.The early fault diagnosis and maintenance of the angle head can be realized through condition monitoring,and then its service life can be extended.It is very important for enhancing the efficiency of manufacturing structural parts of spacecrafts and shorten the development cycle.In general,the vibration signals of an angle head contains a large number of non-stationary signals in actual production,which means that the classical analysis methods including fast Fourier Transform,short-time Fourier transform and wavelet decomposition are not applicable,because these methods do not have self-adaptability and have very limited capacity to analyze complex signal.Based on the analysis above,a method for vibration signal analysis and feature extraction based on EEMD algorithm is proposed in this paper,which is used for recognizing the fault signal feature of an angle head in a CNC machine.The goals are to successfully extract the fault signal features and find the fault location accurately.The main work and research results of this project are as follows:(1)Combining the wavelet noise reduction and EMD algorithm.This joint approach is used for recognizing the vibration signal features,which can accurately identify the vibration features and suppress the noise interference at the same time through simulation and experimental analysis.This method has a faster diagnosis speed and more accurate feature recognition.It can be used to identify the features of on-site vibration signals of the angle head installed on a CNC machine and estimate its running state.(2)A method for vibration signal analysis and feature extraction based on EEMD algorithm is proposed,which overcomes the problems of modal aliasing,noise interference in EMD algorithm and poor adaptive ability,difficult threshold selection in wavelet noise reduction method.The accuracy and practical applicability of this proposed method are verified by simulations and experiments.(3)This joint method is applied to recognizing the features of fault vibration signals for an angle head.The fault features of its gear and bearing are successfully extracted,and the fault location has been predicted.By comparing with the actual condition of angle head in our experiments,the validity and reliability of the proposed method are verified.