Document Type
Thesis
Date of Award
Spring 5-31-1999
Degree Name
Master of Science in Electrical Engineering - (M.S.)
Department
Electrical and Computer Engineering
First Advisor
Timothy Nam Chang
Second Advisor
Andrew Ulrich Meyer
Third Advisor
Zhiming Ji
Fourth Advisor
Reggie J. Caudill
Abstract
Ultra-high precision positioning is of strategic importance to modern industrial processes such as semiconductor manufacturing. Traditional drives with mechanical transmission elements exhibit nonlinearities such as friction, backlash and hysteresis which limit the system performance significantly. The magnetic leadscrew in this work belongs to the class of contactless drives which overcome the above mentioned limitations of contact-type drives. The operation is based on leadscrew/nut coupling but unlike mechanical Ieadscrews, the threads of the nut and the leadscrew are aligned magnetically and do not come in contact. Thus, "hard" nonlinearities are substantially reduced resulting in high precision and high resolution.
The dynamics of the system are, however, lightly damped and result in vibration of the nut upto tens of microns peak-to-peak. Due to the high frequency of the modes, typically a few hundred Hz, the dynamics are difficult to control using conventional techniques, limited actuator bandwidth being one of the reasons. Active control must therefore be employed. This work develops a passband control scheme based on the Hilbert Transform which gives the orthogonal components of the oscillating modes. The components are extracted using a neural network to enhance the robustness of the controller.
Performance of the controller is evaluated under self-resonance, forced oscillation and transient response. Self-resonance is shown to be completely eliminated while for forced oscillation, the axial gain is shown to be reduced. Stabilization time of the transient response is also significantly reduced, thereby confirming the vibration suppression capabilities of the controller.
Recommended Citation
Dani, Bhaskar Vinayak, "Vibration control of ultra-high precision magnetic leadscrew using recurrent neural network" (1999). Theses. 845.
https://digitalcommons.njit.edu/theses/845
