Research On Prediction And Identification Technology Of Milling Stability Of Thin-walled Parts

Thin-walled frame parts occupy an important position in the aerospace field,but due to their low stiffness,weak damping and complex structure,chatter is easy to occur in the cutting process,resulting in the machining efficiency and surface quality of the workpiece can not meet the use requirements.Therefore,chatter has always been an important research topic at home and abroad.In this paper,chatter simulation,chatter identification and state identification are studied.In the milling process of thin-walled parts,the process damping will occur due to the interference between the tool flank and the machining surface,which has a certain impact on the milling stability,and the feed speed and vibration speed will have an impact on the direction of the actual cutting speed.In order to solve these problems,this paper establishes a dynamic model of thin-walled workpiece milling considering process damping and speed effects,and uses the dynamic model to predict the milling stability of thin-walled workpiece.The proposed dynamic model and chatter identification algorithm are integrated in Simulink to form a chatter system simulation.The accuracy of stability prediction is verified through the combination of simulation and experiment.In order to solve the problem that it is difficult to identify the chatter in the milling process of thin-walled parts,a parameter adaptive optimal variational mode decomposition algorithm and a multi-scale sample entropy based on chatter feature extraction method for thin-walled parts are proposed.Firstly,in order to solve the problem that the parameters of the variational modal decomposition algorithm are difficult to select,a parameter adaptive method for optimizing the variational modal decomposition algorithm is proposed.Secondly,the sub components with high energy ratio are selected for signal reconstruction.Finally,in order to solve the problem that the single scale sample entropy can not well reflect the characteristics of milling force signal when chatter occurs,multi-scale sample entropy is introduced to detect milling chatter.Aiming at the problem that the traditional machining state recognition model can not adapt to the changing working conditions,which leads to the reduction of its recognition accuracy and generalization ability,an evolutionary recognition model of thin-walled workpiece milling state based on optimized ICELM-FCM is proposed.Firstly,the improved particle swarm optimization algorithm is used to optimize the parameters of the limit learning machine;Secondly,the incremental limit learning machine algorithm is proposed to solve the problem that the limit learning machine can not adapt to the changing working conditions;Then,aiming at the fact that the incremental limit learning machine relies too much on its own model to identify the processing state and does not judge the correctness of the classification results,an evolutionary recognition model of processing state is proposed,which combines supervised learning with unsupervised learning.Finally,the proposed machining state recognition model is verified by experiments.Finally,the above methods are verified by machining thin-walled frame parts.In view of the different stiffness of thin-walled structure caused by the structure of thinwalled frame parts,modal analysis was carried out for different positions of thinwalled structure by using modal experiment,and the stability leaf maps of different positions were constructed.Initial processing parameters were selected by using the leaf maps,and chatter identification algorithm and state identification algorithm were used to detect chatter,it is verified that the method proposed in this paper can accurately predict and identify machining chatter of thin-walled frame parts.