Detection Of Cutting Chatter For Machine Tool

With the rapid development of global industry,machine tool and other large equipment gradually develop in the direction of automation,high-speed and intelligence.However,the wear and chatter of the machine tool can occur due to the complicated working environment of the machine tool.The machining accuracy of the machine tool will be seriously affected by this situation.It is important to improve the technology so that judgments can be made timely by the equipment to ensure the safe operation of the equipment.In the process of machine tool flutter diagnosis,it is critical for the analysis of cutting chatter,the extraction and identification of chatter characteristics.In this paper,the detection and identification of cutting chatter is investigated based on the improved VNCMD algorithm.And it is organized as follows.Firstly,the mechanism of machine tool cutting chatter is briefly described by means of dynamic modeling.The generation of cutting chatter is verified by designing machine tool cutting experiments and cutting chatter detection systems.The variation of cutting vibration under different cutting parameters is collected,which lays a foundation for the subsequent analysis of cutting chatter.During machining process,the cutting chatter may be generated with the change of cutting state from smooth to unsmooth.The impact effect will be induced in this process,which is harmful to machining precision.In order to effectively avoid the influence of cutting chatter,the improved Variational Nonlinear Chirp Mode Decomposition(VNCMD)algorithm is presented for the detection and identification of cutting chatter.The wideband and weak characteristic of the signal are considered by the improved algorithm,which can not only overcome the modal aliasing and pseudo-component problems of the Empirical Mode Decomposition(EMD)algorithm,but also overcome the drawbacks of the wideband signal extraction.To eliminate the influence of the number of intrinsic mode components on the decomposition of the VNCMD algorithm,the correlation coefficient method is used to determine the optimal intrinsic mode components number.And the improved algorithm is further applied to the detection of the cutting chatter.Although the noise of extracted intrinsic mode components has been filtered out,it is a little weak in the description of chatter characteristics.Next,the different types of simulation signals are studied before the fourth-order cumulant is used to extract the operating state of the machine cutting process.And then the fourth-order cumulant is used to analyze the numerical characteristics of each intrinsic mode component of the target signal,which are taken as a characteristic of cutting chatter.The different states of the cutting chatter are clearly reflected by the extraction results.It fully demonstrates that this method can effectively determine the change in cutting chatter.Based on the research of permutation entropy,the effectiveness of detecting abrupt signals was verified by different types of simulated signals.Next,the second numerical characteristics of intrinsic mode component of the target signal are analyzed by the permutation entropy.And then it is taken as features of the cut chatter.The extraction results clearly reflect the different running state of machine tool cutting.The different states of the cutting chatter are clearly reflected by the extraction results.It fully demonstrates that this method can effectively determine the change in cutting chatter.When the data to be processed is large,our visual observation on the data is somewhat difficult.So this paper proposes a fault feature recognition method based on combined feature vectors and optimized depth belief network,which can achieve more accurate recognition by accumulating cutting data and training rural network continuously.The test results show that this method can achieve higher fault state recognition rate.Finally,the paper was summarized and the future research work was made a prospect.