Modeling, analysis and tracking control of brushless DC motors for robotic applications

The problems of modeling, analysis, and control associated with Brushless DC Motors (BLDCM) are considered. In the modeling process, three classes of BLDCMs have been distinguished. The first class of BLDCM includes motors whose magnetic structure may be assumed to be linear, and have uniform air gaps. The second class constitutes the BLDCM with linear magnetic structure and non-uniform air gaps. For these classes of BLDCM, the governing differential equations are presented and used toward the analysis of their electromechanical properties. It is demonstrated that the torque producing capabilities of BLDCMs, both with uniform and non-uniform air gaps, can be enhanced by appropriately supplying optimal voltage inputs across the stator phase windings.;BLDCM constitutes a multivariable nonlinear dynamical system. Accordingly, to guarantee the high performance operation of BLDCM in a motion control system, its complete dynamical characteristics are accounted for. Based on the transformation theory of nonlinear control systems, a nonlinear control law is designed which compensates for the nonlinearities of BLDCM, transforming it to a linear controllable system description. A direct drive robotic arm actuated by a BLDCM is chosen to study the effectiveness of the control law. The control scheme exhibits satisfactory performance when accurate measurements of position, velocity, and acceleration are available. To guarantee high accuracy in the performance when the system is subject to measurement errors and modeling uncertainties, a robust control law is added to the nonlinear controller. Using computer simulation techniques, it is demonstrated that the robust controller performs effectively, capable of compensating for modeling uncertainties, external disturbances, and measurement errors.;The third class of BLDCM considered includes motors with non-uniform air gaps which operate in the range where magnetic saturation exists. It is demonstrated that the modeling problem for this class of BLDCM can be formulated in terms of mathematically modeling a set of four-dimensional surfaces, corresponding to the electromagnetic torque function and the flux linkages associated with the phase windings. The mathematical modeling problem is reduced to that of identifying a two-dimensional electromagnetic torque surface. An experimental identification procedure is designed and satisfactorily implemented. By obtaining torque and current measurements, a mathematical model for a BLDCM is constructed and compared with data obtained from independent laboratory experiments.