Review about Control and Estimation Strategies applied to Bearingless Machines

Motors with magnetic bearings and bearingless machines have as a basic principle magnetic levitation to reduce wear on mechanical parts to increase their useful life and reduce the need for preventive and corrective maintenance, especially in equipment that operates in diffcult-toaccess environments. However, in order for them to achieve the desired performance, their respective control systems must operate effciently and synchronously on: the radial rotor position, the torque, speed, and supply currents of these machines. In this context, the first systems developed used in a broad and unrestricted way the classic controllers PI (Proportional-Integrative), PD (Proportional-Derivative), and PID (Proportional-Integrative-Derivative), allied to state estimators based on inverse models of these machines to avoid the use of costly sensors. During the research, however, it was found that the non-linearities and parametric variations inherent to these special machines could affect the performance of the control and estimation systems developed. In order to compensate for these limitations, from the 2000s onwards, new researches were developed using the main modern control and state estimation techniques successfully applied to conventional electrical machines. In this context, the development of microelectronics and new microcontrollers such as, for example, the DSP (Digital Signal Processor) was fundamental to enable the implementation of these new systems. To recover the history of these works and contribute to the development of future systems, this article presents an overview of the main control and estimation techniques applied to different types of machines without bearings and levitation systems developed so far, highlighting, critically, the advantages and limitations of each of the techniques discussed.