Some recent experimental studies have been concerned with the stabilization of rotating machinery through control action applied at critical points on the rotor. These studies have tended to ignore fundamental constraints on the positioning of active control inputs and measurement transducers. Such constraints must be considered to ensure both a cost effective design and system integrity in the event of component failure. This paper presents the first stage of a theoretical examination of active control by investigating a mathematical model of a three-mass flexible rotor symmetrically supported on flexibly mounted journal bearings. Three questions are posed: how many control sources are required, what measurement transducers are required and what information about the system dynamics is required in order to implement active control? The multivariable nature of the model and uncertainty about the oil-film dynamics makes general solutions to these problems difficult. However some solutions are proposed based upon the structural properties of the model. It is shown that the general requirements for system identification and control are conflicting. The practical implications of this conflict are discussed.

This content is only available via PDF.
You do not currently have access to this content.