|
romac |
|
University of Virginia Rotating Machinery and Controls Laboratory P.O. Box 400746 122 Engineer’s Way Charlottesville, VA 22904 (434)924-3292 (434)982-2246(fax) |
|
Rotating Machinery and Controls Laboratory |
|
High Speed Rotors
Student: Guoxin Li Advisors: Paul Allaire, Zongli Lin Funding: ROMAC Project Objectives: The control of a strongly gyroscopic rotor over a wide speed range using magnetic bearings is a challenging problem. Two dominant issues must be addressed in the control design. First, high bandwidth controllers are often required to stabilize the forward conical mode. This may compromise the high robustness requirements. Second, the dynamics of the rotor vary with the operating speed; a single linear time invariant controller often cannot stabilize the system over the entire operating range. Advanced multi-variable robust optimal controls, H∞ and m-synthesis, provide systematic tools to achieve system robustness. To address the speed dependent dynamics, gain scheduled robust control including the switching control techniques are being investigated. Linear parameter varying (LPV) approach was employed to design gain scheduled H∞ controllers. The gainscheduling problem was formulated in the context of convex semidefinite programming by linear matrix inequalities (LMIs). The LPV controller was obtained by solving the LMIs using interior point methods. A piecewise m-synthesis design was also adopted. Overall control was implemented by switching between controllers as speed varies from one region to another. A bumpless transfer scheme was implemented to guarantee smooth transition between controllers. Progress in the Past year: A test rig was constructed as a platform for investigating of different controllers. First, an accurate nominal model including the substructure modes was developed from physical laws and refined by experimental data. An uncertainty representation and a performance criterion were developed for the model AMB system. The influence of gyroscopic effects on the stability and performance of AMB system under a MIMO controller was analyzed.
|