Research On Predictive Control For Vibrations Of Thin-walled Workpieces In Milling

Due to the excellent characteristics of thin-walled workpieces, such as light weight andhard structure strength, the thin-walled workpieces have been widely used in aerospace,energy, transportation and other industrial societies concerning people’s livelihood witheconomic growth and technical progress. In order to improve the productivity, manufactorieswould better marching the blanks directly and shape the workpieces at once. However, thereare many difficulties in manufacture process including practical issues, e.g., tools’ pathplanning, improvement of the metal milling rate and suppressing vibrations in milling, andtheoretical issues, e.g., how to identify the chatter from the vibrations in milling,mathematical analysis of the complex milling force. Not only could these difficultiesinfluence the life of tools and workpieces, but also affect the health of operators, thus,suppressing the vibrations of the thin-walled workpieces is of great importance in recentstudies.At present, most researches focus on suppressing vibrations with spindle-tools andwork-holders in milling, where external force is usually loaded on the system to changedynamical properties of the system. The operation to the spindles and work-holders must bevery careful since their accuracies are easily influenced such that high precision demand inmanufacture can not be satisfied. The problem on how to design an actuator and an observeris a prior consideration to that on how to suppress the vibration in milling. This thesisintroduces Active Control Experiment Flat for Milling (ACEFM), which is used to put forceon workpieces and is applied for verifying the effectiveness of the control algorithms beforebeing loaded on the CNC lathe.Model Predictive Control (MPC) is one of the advanced control algorithms and closelyrelative to industrial application, which can predict the future dynamics of system, makeactuators take action ahead and make the system easily achieve the control target. Besides,MPC can explicitly compute control sequence with physical constraints, such as outputlimitations of voice coil motor. MPC consists of Parameterized MPC and Non-parameterizedMPC. Generalized Predictive Control (GPC) belongs to Non-parameterized MPC andLaguerre Model Predictive Control (LMPC) is one of Parameterized MPC in contrast. GPCand LMPC are both applied on ACEFM in this thesis. The vibrations of the thin-walledworkpieces could be successfully suppressed when they are excited by external drivingsource.Finally, milling experiment is performed on Computer Numerical Control (CNC) lathe,MIKRON UCP800Duro, and the experimental vibrations of workpieces in milling process isobserved and investigated as well. Furthermore, simulations are given to illustrate the effectiveness of the method in this thesis.