UBC Theses and Dissertations
Modeling and control of high speed machine tool feed drives Okwudire, Chinedum
Aerospace, die and mold, and automotive industries machine parts at high cutting speeds to reduce production cycle periods. Machine tools which carry out the cutting operations rely on either precision ball screw or linear motor direct drives to accurately position the workpiece relative to the cutting tool. However, the precise positioning capability of the drives is limited by low servo bandwidth and poor disturbance rejection resulting from structural flexibilities in ball screw drives as well as weak dynamic stiffness/robustness in direct drives. This thesis proposes modeling, parameter identification, control and online parameter estimation techniques which aim at increasing the servo bandwidth and disturbance rejection ability of high speed machine tool feed drives. A hybrid finite element methodology is used to model the structural dynamics of ball screw drives. As part of the model, two stiffness matrices are developed for connecting the finite element representation of the ball screw to the lumped-mass representation of the nut. The developed model is used to analyze the coupled axial-torsional-lateral vibration behavior of a critical structural mode that limits high bandwidth control of ball screw drives. Moreover, a method for accurately identifying the mass, damping and stiffness matrices representing the open-loop dynamics of ball screw drives is developed. The identified matrices are used to design gain-scheduled sliding mode controllers, combined with minimum tracking error filters, to effectively suppress the critical axial-torsional-lateral mode of ball screw drives thereby achieving high bandwidth control and good disturbance rejection. For direct-driven machines, a high bandwidth disturbance adaptive sliding mode controller is designed to improve the dynamic stiffness of the drive, compared to similar controller designs, without increasing the controller’s complexity. Furthermore, the cutting forces applied to the drive are estimated accurately using a disturbance recovery algorithm and used to improve the dynamic stiffness of low-frequency structural modes of direct-driven machine tools. Finally, a method for estimating the changing mass of the workpiece during machining operations with cutting forces that are periodic at spindle frequency is introduced. The techniques presented in this thesis are verified through simulations and/or experiments on single-axis ball screw and linear motor feed drives.
Item Citations and Data
Attribution-NonCommercial-NoDerivatives 4.0 International