Adaptive forward identification and auto-tuning for motor velocity control

Abstract

This paper presents a new on-line adaptive identification of physical model parameters by adjusting the parameters in the feedforward controller and feedback PI controller for a motor drive system. The proposed scheme is characterized by new set-up of a suitable reference model which generates signals needed to identify the physical parameters of a two-mass motor system, such as inertia moments of motor and load, shaft conductance, viscous friction terms proportional to motor velocity and squared velocity, and Coulomb friction. The adaptive algorithm not only can give the model parameters but also compensate for the unknown friction effects by adjusting the feedforward controller. Stability is also investigated via the ASPR property of the controlled system. The effectiveness of the proposed scheme is examined in simulation and experiment of speed control of a two-mass motor system.