Research On Sparse Fault Diagnosis Method Of Automatic Tool Changer System In Machining Center

As the representative products of CNC machine tools,machining centers have the characteristics of high efficiency,high speed,high precision and multi-function.Compared with CNC milling machines,the machining center is equipped with an automatic tool changer system(ATCS),which can continuously complete multiple processes such as milling,drilling,tapping and carving after the workpiece is clamped.It can greatly reduce the positioning error introduced by multiple tool installations and save the non-cutting time occupied by the manual tool changing.However,the ATCS is a complex electromechanical system,which often needs to complete hundreds of thousands or even millions of automatic tool-changing cycles in the whole life cycle.Because of the frequent tool changing,the working performance of the ATCS will decline inevitably,and even leads serious faults such as tool dropping and tool stuck.These failures not only lead to the shutdown of the machining center,make the efficiency of production lower and increase the maintenance cost,but also sometimes break the cutting tools and damage the workpieces,causing economic losses to the enterprise.Mechanical fault diagnosis technology plays an important role in improving the reliability of the ATCS.The failure is found and eliminated in time,which prolongs the normal working time of the ATCS without failure,that is,the reliability is improved.Therefore,the research on fault diagnosis of the ATCS has significances to improve the reliability of the ATCS and ensure the saft and reliable operation of the machining center.Based on the vibration data of the tool-changing process,this dissertation focuses on the application of sparse signal processing method in the fault diagnosis of the ATCS.The main work of this paper is as follows:(1)There are many failure modes of the ATCS,which are closely related to its mechanical structure and working principle,it is necessary to first analyze the type,composition and working principle of the ATCS.According to the statistical analysis results of the field failure data,four typical faults of the ATCS are identified.Through establishing the geometric model and mechanical model,the causes of three typical faults are analyzed in detail,and the evolution law of typical faults and corresponding early faults are mastered.(2)The tool-changing vibration signal has the characteristics of transient nature and sparsity nature,a fault diagnosis method of the ATCS based on sparse representation of parametric impulse dictionary is proposed.The proposed method includes two important aspects:constructing the parametric impulse dictionary and solving the sparse coefficient.Based on the characteristics of multi-order modes presented by the transient components in the tool-changing vibration signal,the parametric impulse dictionary is constructed using multi-dimensional real Laplace wavelet.At the same time,the improved state space method is used to optimize the atomic waveform parameters of the dictionary,which greatly reduces the redundancy of the dictionary.In the aspect of sparse coefficient solving,the split variable augmented Lagrangian shrinkage algorithm is used to solve the objective function,which greatly reduce the execution time of the algorithm.The simulation result shows that the proposed method can estimate the occurrence time of transients in tool-changing vibration signal with high accuracy.The proposed method is applied to the typical fault diagnosis of the ATCS,and satisfactory diagnosis results are obtained.(3)Aiming at the problem that the signal sparse representation method needs to construct a dictionary and may lose the weak fault information contained in the signal,a fault diagnosis method of the ATCS based on the overlapping group sparse collaborative filtering is proposed.The overlapping group sparse collaborative filtering(OGSCF)which combines the overlapping group sparsity denoising and the collaborative filtering can simultaneously filter out the additive noise and compositional noise of a group of similar signals,and better retain the important weak feature information.Aiming at the problem that it is difficult to optimize appropriate the regularization parameter and the group size of the overlapping group sparse denoising method in practical applications,the collaborative filtering is used to construct an estimation signal to replace the real signal.Finally,the criteria for optimizing the parameters is identified.Because the differential operation in collaborative filtering can amplify additive noise,the estimation accuracy of the estimation signal becomes worse.To handle this problem,the overlapping group sparsity denoising is used to filter the additive noise.Under the iterative optimization strategy,the group sparsity denoising and the collaborative filtering are performed alternately until convergence.The effectiveness of OGSCF is verified by the numerical simulation.In order to test the noise reduction performance of OGSCF,the simulation tests under different noise level were also carried out.Finally,based on this method,the estimated signals of the measured vibration signals under normal condition and fault condition are obtained,respectively.In order to quantitatively evaluate the severity of the early faults,the template signal corresponding to the normal condition is used as the reference signal and the statistical distribution model of standard error of the reference signal is also established.The p value corresponding to the standard error of the template signal in the fault condition is used to evaluate the severity of early faults.In the end,three typical early faults of the ATCS are diagnosed by he proposed method,and satisfactory results are obtained.(4)The transient component of the tool-changing vibration signal are mostly multi-component signals and the distribution of fault information on each component is sparse,a fault diagnosis method of the ATCS based on non-convex overlapping group sparse collaborative filtering and the adaptive variational mode decomposition is proposed in this thesis.The l_2,1 norm underestimates the high amplitude component of the signal,which affects the estimation accuracy of the denoised signal.So,the non-convex overlapping group sparse collaborative filtering is used to suppress the additive noise and compositional noise.In order to extract the mode components containing fault information,the adaptive variational mode decomposition is used to decompose the denoised tool-changing vibration signal.By eliminating the mode components without fault information,the fault mode components are highlighted and the weak fault information can be further purified and concentrated.The effectiveness of the proposed method is validated by the simulation tests.For the sake of assessing the noise reduction performance of the proposed method,the simulation tests under different noise level were also carried out.Finally,the proposed method is applied to three typical early fault diagnosis of the ATCS and achieve satisfactory results.