Date of Award
5-2016
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Legacy Department
Electrical Engineering
Committee Member
Dr. Richard E. Groff, Committee Chair
Committee Member
Dr. Konstantin Kornev
Committee Member
Dr. Ian Walker
Committee Member
Dr. Yongqiang Wang
Abstract
Cantilevered magnetic wires and fibers can be used as actuators in microfluidic applications. The actuator may be unstable in some range of displacements. Precise position control is required for actuation. The goal of this work is to develop position controllers for cantilevered magnetic wires. A simple exact model knowledge (EMK) controller can be used for position control, but the actuator needs to be modeled accurately for the EMK controller to work. Continuum models have been proposed for magnetic wires in literature. Reduced order models have also been proposed. A one degree of freedom model sufficiently describes the dynamics of a cantilevered wire in the field of one magnet over small displacements. This reduced order model is used to develop the EMK controller here. The EMK controller assumes that model parameters are known accurately. Some model parameters depend on the magnetic field. However, the effect of the magnetic field on the wire is difficult to measure in practice. Stability analysis shows that an inaccurate estimate of the magnetic field introduces parametric perturbations in the closed loop system. This makes the system less robust to disturbances. Therefore, the model parameters need to be estimated accurately for the EMK controller to work. An adaptive observer that can estimate system parameters on-line and reduce parametric perturbations is designed here. The adaptive observer only works if the system is stable. The EMK controller is not guaranteed to stabilize the system under perturbations. Precise tuning of parameters is required to stabilize the system using the EMK controller. Therefore, a controller that stabilizes the system using imprecise model parameters is required for the observer to work as intended. The adaptive observer estimates system states and parameters. These states and parameters are used here to implement an indirect adaptive controller. This indirect controller can stabilize the system using imprecise initial parameter estimates. The indirect adaptive controller overcomes the limitations of the EMK controller by stabilizing the closed loop system despite inaccurate initial parameter estimates. The experiment setup used to test the controllers is also presented. Experiments were performed to test the adaptive controller using cantilevered cobalt and nickel wires. The closed loop system using the indirect controller is stable. The wire tracks continuous desired trajectories up to 30Hz. Experiments were also performed to test the robustness of the adaptive and EMK controllers when the wire is interacting with water. The adaptive controller performs poorly when unmodeled disturbances are encountered, necessitating fall back to the EMK controller in some applications. The adaptive controller functions as an EMK controller if observer gain is set to 0. Thus, the indirect adaptive controller estimates model parameters, stabilizes the wire in the unstable region and can be switched into a non-adaptive mode for applications.
Recommended Citation
Karve, Harshwardhan, "Online Parameter Estimation and Adaptive Control of Magnetic Wire Actuators" (2016). All Dissertations. 1620.
https://open.clemson.edu/all_dissertations/1620