High-performance machine tool controllers: A control theoretic study and a PC-based realization

High-speed machining is recognized as one of the most important areas in manufacturing. The need for high speed operations in machining is driven not only by the quest for higher productivity but also by other production benefits such as burr-free improved surface finish, thin web machining potential, and reduced structural stiffness requirements. However, machining at high-speed faces many challenges which require further research effort. One important problem is the design of advanced motion controllers for high-speed/high-accuracy operations in the presence of disturbances, nonlinearities, and unmodeled dynamics. This dissertation focuses on the development of such motion control algorithms and their implementations on a vertical machining center. An open architecture PC-based controller is built as a part of this study to replace the original controller of the machining center. This controller allows easy and feasible implementation of motion control algorithms as well as higher level functions such as machining process monitoring and control.; First, two general robust controllers, a disturbance-observer-based (DOB) and an adaptive-robust-controller-based (ARC), are tested using the machine tool X-Y table. The goal is to achieve nominal tracking errors near the measurement resolution, 1 {dollar}mu m,{dollar} including during all stages of the operation transients. The resulting closed-loop system should have stability robustness and performance robustness. The experimental results confirmed that both controllers have the ability to achieve near-perfect tracking of a variety of trajectories and to reduce the effects of discontinuous disturbances such as coulomb friction when motion changes direction. However, the ARC had superior disturbance rejection and better tracking performance with a tracking error profile within {dollar}{lcub}pm{rcub}2 mu m{dollar} at a tracking speed of 7 m/min. The DOB performance degraded significantly when the acceleration profile of the desired trajectory contained large discontinuities, while the ARC handled such discontinuities effectively. The performance of the two controllers remained the same during cutting experiments. When large parameter variations were introduced, the DOB performed inadequately in one test where its stability robustness condition was not violated. When the parameter variations were large enough to violate such condition, the DOB became unstable. On the other hand, the ARC performance was not affected. The performance superiority of the ARC over-the DOB can be attributed to the following facts: (1) the ARC can handle large parameter variations through parameter adaptations, (2) extra nonlinear robust terms can easily be added to the ARC to improve transient performance, and (3) the ARC has an anti-integration windup mechanism which alleviates the control saturation problem.; Finally, a modified design of the DOB is presented and implemented on the machine tool vertical axis to address the issues of time delay and position-dependent friction. The modifications include the addition of a Smith-predictor in the feedback loop and the consideration of time delay in the nominal model. Several experimental and simulation tests were conducted and the results confirmed the higher performance of this control...