期刊名称：Advances in Science and Technology Research Journal
印刷版ISSN：2080-4075
电子版ISSN：2299-8624
出版年度：2019
卷号：13
期号：3
页码：217-228
DOI：10.12913/22998624/111705
语种：English
出版社：Society of Polish Mechanical Engineers and Technicians
摘要：Chatter is a series of unwanted and extreme vibrations which frequently happens during different machining processes and impose variety of adverse effects on the machine-tool and surface fnish. Chatter has two main typesnamely forced-chatter and self-existed chatter. The forced-chatter has an external cause; however, self-exited chatter has no external stimuli, rather it is created due to the phase difference between the previous and current waveson the surface of the workpiece. Due to the self-generative nature of this type of chatter, its recognition and prevention is much more difcult. For preventing self-exited chatter its model should be available frst. The chatteris usually simulated as a one degree of freedom mass-spring-damper model with unknown parameters that theyshould be determined somehow. In this paper, the parameters of the tool equation of motion i.e. mass, damping,and stiffness coefcients of the system are predicted through a wavelet-based method online, and then based on theachieved parameters, the system is controlled via Model Predictive Control (MPC) approach. For the validation,the algorithm is applied to 25 different experimental tests in which the acceleration of the tool and cutting forceare measured via an accelerometer and a dynamometer. By investigation of the SLDs generated by the predictedparameters, the presented system identifcation method is validated. Also, it is shown that the chatter vibration iscompletely restrained by means of MPC. For investigation of the MPC performance, MPC algorithm is comparedwith PID controller and simulations has indicated a much stronger performance of MPC rather than PID controllerin terms of vibration attenuation and control effort.
关键词：turning process; chatter phenomenon; system identifcation; discrete wavelet transform; and model
predictive control
