Research On Chatter Identification And Dynamic Reconstruction Error Compensation Of Thin-walled Parts Based On Vision Inspection

With the development of aerospace,automobile and other fields,thin-walled parts with high strength and light weight are widely used in engineering.Due to the poor stiffness of this kind of parts,it is easy to chatter in the milling process,which not only worsens the machining surface quality,but also affects its machining efficiency.In order to improve the intelligence and automation of mechanical manufacturing,aiming at the problem of chatter,the state detection of milling chatter of thin-walled parts and the non-contact dynamic reconstruction of its surface are realized by using advanced vision detection technology,which is of great significance to guide the processing of thin-walled parts,avoid the influence of chatter,and improve the processing efficiency and processing stability of thin-walled parts.Firstly,the theory of vision detection technology of milling chatter is deeply studied.The correlation characteristics between milling chatter surface and image of thin-walled parts are analyzed,and then the image preprocessing scheme is designed according to the influence of processing environment.At the same time,according to the different characteristics of images in different processing states,the image feature extraction algorithm of local binary mode and gray level co-occurrence matrix is studied and improved.Secondly,on the basis of milling chatter image preprocessing and feature extraction,a chatter feature classification model based on K-nearest neighbor algorithm is proposed.Through the chatter milling of thin-walled parts on the machine tool and the visual identification experiment,it is obtained that the accuracy of the chatter visual identification method proposed in this paper is 95.5% and the average running time of the algorithm is 0.069 seconds,The real-time identification and prediction of milling chatter are realized.Thirdly,in order to eliminate the influence of milling chatter of thin-walled parts on the error of structured light reconstruction measurement,the theoretical model of structured light projection measurement is systematically studied,and the influence mechanism of the dynamic characteristics of thin-walled parts chatter on structured light3 D vision detection is analyzed.In order to improve the reconstruction measurement accuracy of thin-walled parts and avoid the ripple phenomenon of the measured point cloud,a phase principal value solution correction model considering phase error is established based on the phase shift method.Finally,the dynamic reconstruction experiment is carried out by using the structured light vision detection system and the chatter simulation experimental platform of thin-walled parts.The experimental process and scheme of dynamic reconstruction of thin-walled parts are designed as a whole.In the case of typical slight chatter and serious chatter,the principal value of structural light phase is compensated and corrected,and the surface measurement results of thin-walled parts before and after compensation are analyzed.It is proved that the method proposed in this paper can effectively reduce the reconstruction ripple error of chatter thin-walled parts.