Grinding Chatter Monitoring Research Based On Generalized S Transform

CNC rail grinder,which is a common machine in the field of industrial manufacturing,is the carrier of advanced manufacturing technology and the key equipment of instrument manufacturing.Grinding chatter is one of the important reasons for the instability of grinding machine,which mainly caused by the self-excited vibration between the tool and the work-piece,will reduce the machining surface quality and accelerate damage of machine.Due to the grinding process is very complex,the grinder vibration signals are nonlinear and non-stationary signals,The signals cannot be processed effectively with the traditional signal processing method.According to the signal characteristics of the grinding chatter transition process,it is necessary to find an effective signal feature extraction method to separate the characteristic information of grinding chatter from the complex original signal,and get the exact characteristic indexes to characterize grinder vibration state.Conventional vibrating signal processing is based on Fourier,but Fourier transform is a global transform,which is unable to characterize local features of signals and difficult to extract the feature information of signals.The nonlinear and non-stationary signals which collected from the process of grinding are more difficult to identify and extract in strong background signals and noise.In view of the above problem,this thesis is supported by the Zhejiang Province Natural Science Fund Project(LZ13E050003),and the experimental signals of grinding process are collected and saved by MK7180X160 / L CNC grinding machine.The generalized S-transform is used to extract the characteristic indexes which represent the chatter state.In the end,the state evaluation model is used to find the corresponding relationship between the grinding characteristic index and the chatter state,and then put forward a practical and effective grinding chatter online detection method.Firstly,the background and significance of the topic is introduced,the thesis describes research of status identify and trends of the grinding chatter,and points out the deficiency and research trend of grinding chatter in the current methods.According to the signal characteristics of the grinding chatter,generalized S-transform,a new non-stationary and non-linear signal processing method,is introduced into the signal processing and monitoring for grinding chatter,which provides a direction for further research.Secondly,according to feature of the vibration signals in the grinding processing,the appropriate sensors and data acquisition module are chose to build the platform for grinder chatter test.Vibration acceleration sensors are arranged reasonable,by changing the grinding parameters,the signals of different vibration state are collected in real-time.Finally,principal component analysis is introduced into the grinding chatter for grinding chatter judgment and recognition.The recognition model of the chatter state evaluation is established,the real-time variance and generalized S-transform energy entropy are exacted as the characteristic indexes with the learning ability of principal component analysis,which implements the aims of the judgment and recognition for grinding chatter state.The proposed method based on generalized S-transform and principal component analysis(PCA)is verified by the signals of experiments.