Supervisory machining control: A design approach plus chatter analysis and force control components.

Advanced monitoring and control modules will be utilized in future machining operations as machine tool owners seek to increase productivity and part quality. Two key research questions regarding these modules have yet to be answered. First, how do the inherent force process nonlinearities affect machining modules and can these nonlinearities be incorporated to improve module performance? Second, when multiple modules are simultaneously implemented in complex machining operations, will they perform as desired, or will adverse interactions occur? The second question will become increasingly important as future machining systems are created with open-architecture platforms which provide the facility to easily reconfigure the machining modules and incorporate advanced functionalities. This dissertation seeks to address these questions. The effect of force process nonlinearities on the performance of common force control methodologies is investigated, and a novel, model-based force control approach which directly accounts for these nonlinearities is developed. The effect of process nonlinearities on chatter analysis is also investigated for a practical machining operation (face milling) and a stability algorithm for machining systems containing these nonlinearities is developed. A systematic design approach for constructing modular supervisory machining controllers was also developed. These controllers intelligently regulate the activity of multiple modules in machining operations such that each module performs its task properly and adverse interactions between modules do not occur. The inherent force process nonlinearities are shown to affect the transient performance of common force control techniques. The force control approach developed in this dissertation provides excellent transient performance and has a much simpler structure than previous adaptive force controllers. The inherent process nonlinearities are also shown to affect the shape of the stability lobe diagram: the stability limit is dependent on feed and the asymptotic stability borderline is not constant. The nonlinear analysis provides superior stability predictions as compared to the linear analysis in both simulations and experiments. The supervisory machining control design approach developed in this dissertation has been utilized to build two supervisory controllers. The experimental results demonstrate the supervisory controllers' ability to regulate machining modules in both job shop and large batch production environments. The modular structure of the supervisory controllers also makes these controllers ideal for implementation in future open-architecture machining systems.Ph.D.Applied SciencesMechanical engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/130284/2/9722022.pd