Estimation parameters and black box model of a brushless DC motor

Abstract

The modeling of a process or a plant is vital for the design of its control system, since it allows predicting its dynamic and behavior under different circumstances, inputs, disturbances and noise. The main objective of this work is to identify which model is best for a permanent magnet brushless DC specific motor. For this, the mathematical model of a DC motor brushless PW16D, manufactured by Golden Motor, is presented and compared with its black box model; both are derived from experimental data. These data, the average applied voltage and the angular velocity, are acquired by a data acquisition card and imported to the computer. The constants of the mathematical model are estimated by a curve fitting algorithm based on non-linear least squares and pattern search using computational tool. To estimate the mathematical model constants by non-linear least square and search pattern, a goodness of fit of 84.88% and 80.48% respectively was obtained. The goodness of fit obtained by the black box model was 87.72%. The mathematical model presented slightly lower goodness of fit, but allowed to analyze the behavior of variables of interest such as the power consumption and the torque applied to the motor. Because of this, it is concluded that the mathematical model obtained by experimental data of the brushless motor PW16D, is better than its black box model.