Design and control methods for high-accuracy variable reluctance actuators
Author(s)
MacKenzie, Ian (Ross Ian)
DownloadFull printable version (67.86Mb)
Other Contributors
Massachusetts Institute of Technology. Department of Mechanical Engineering.
Advisor
David L. Trumper.
Terms of use
Metadata
Show full item recordAbstract
This thesis presents the design and control techniques of a variable reluctance actuator for driving a reticle motion stage in photolithography scanners. The primary thesis contributions include the design and experimental demonstration of a magnetic flux controller that uses a sense coil measurement, the design and experimental demonstration of a novel method to estimate actuator hysteresis in real-time, and the development of an actuator model that incorporates the effects of eddy currents. The reticle stage in a scanning lithography machine requires high accelerations combined with sub-nanometer position accuracy. Reluctance actuators are capable of providing high force densities (force per moving mass) and lower power values relative to the present state-of-the-art Lorentz actuators that are used to drive the reticle stage. However, reluctance actuators are highly nonlinear with both current and air gap. They also display other nonlinear behavior from hysteresis and eddy currents. Linearizing the reluctance actuator is required for the high force accuracy required in the scanning stage. In this thesis, we present a way to linearize the reluctance actuator with flux control using a sense coil as the feedback measurement. Because the sense coil is AC-coupled, we design a low-frequency estimate of the magnetic flux based upon the actuator current and air gap measurements. We combine the low-frequency estimate with the sense coil measurement using a complementary filter pair that provides an estimate of the flux from DC to frequencies of several kHz. For the low-frequency estimate, we develop a novel method for estimating the actuator hysteresis in realtime. For this flux estimator, we use an observer to model the actuator flux which treats the changing air gap as a disturbance to the plant model. The use of an observer allows the identification of a single-variable hysteresis model of actuator current rather than a two-variable hysteresis model of current and air gap. We also introduce a novel way for expressing the actuator hysteresis, whereby we incorporate the linearizing effect of the air gap directly into a Preisach hysteresis model via a change of variables. We demonstrate experimentally that this method is numerically stable in the presence of a dynamically changing gap, in contrast to some alternative methods. We designed and built a reluctance actuator prototype and 1-DoF motion testbed to demonstrate the accuracy of the actuator models and control techniques. We experimentally demonstrated that we can achieve a flux control bandwidth of 4 kHz that is capable of reducing the stiffness of the reluctance actuator to less than 0.012 N/[mu]m for frequencies up to 100 Hz. This results in a force error of less than 0.03% of the full-scale force for a 10 [mu]m air gap disturbance at this frequency. We also demonstrate that the actuator hysteresis model is capable of estimating the actuator flux accurately in the presence of dynamic gap disturbances of at least 35 1m peak-to-peak and with a static offset from the nominal air gap of at least 50 [mu]m.
Description
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Mechanical Engineering, 2015. Cataloged from PDF version of thesis. Includes bibliographical references (pages 421-427).
Date issued
2015Department
Massachusetts Institute of Technology. Department of Mechanical EngineeringPublisher
Massachusetts Institute of Technology
Keywords
Mechanical Engineering.