This work reports the results obtained by simulating numerically the whirl motions of two coupled rotors and their stabilization with rotating sensors/actuators. Various types of control laws are tested. It is found that the system can be stabilized with control forces of the size predicted theoretically only if appropriate control laws are designed; the control laws must take into account that the system is spinning rapidly while the whirl motions grow slowly. In essence, one should subtract away from the sensors/actuators their rapid rotation velocity by appropriate averaging and best-fitting, so as to recover the slow relative velocity of the whirl motions which needs to be damped. To the contrary, conventional control laws, in which relative velocities are computed simply from the difference of successive displacements, are not adequate unless control forces about a million times larger than necessary are applied. The need for huge control forces in this case derives from the fact that the slow relative velocity that should be damped is overcome by the far larger rotation velocity of the sensors/actuators; the actuators "fire" without knowing what to, and the final result is a system dominated by the active control forces themselves. The results obtained with conventional control laws are shown in Figure. A-2; the results obtained with the control law which is best in subtracting away the spin are shown in Figure B3-2. The difference is self apparent, and quite striking. This work has been carried out using the numerical code DCAP which has been developed by ALENIA SPAZIO under ESA contract.
Professor Crandall has been asked to evaluate the GG proposed experiment on the specific issue of the stabilization of whirl motions. After several months of discussions with both the GG proposing scientists and the GG reviewers at ESA, he has addressed this letter to the Chairman of the ESA panel (Maurice Jacob) and to the Principal Investigator of the GG Project (Anna M. Nobili)
Unstable whirl motions in the GG experiment are damped actively using small rotating electrostatic sensors/actuators. This work gives an estimate of the losses due to the dampers and concludes that these are negligible compared to losses in the mechanical suspensions.
GG Project Homepage
(Anna Nobili- nobili@dm.unipi.it)
Last edited, 01 March 2001