Stored waveform adaptive motor control

Abstract

This study investigates an adaptive control scheme designed to maintain accurate motor speed control in spite of high-frequency periodic variations in load torque, load inertia, and motor parameters. The controller adapts, stores and replays a schedule of torques to be delivered at discrete points throughout the periodic load cycle. The controller also adapts to non-periodic changes in load conditions which occur over several load cycles and contains inherent integrator control action to drive speed error to zero. Using computer simulations, the control method was successfully applied to a 3Φ synchronous motor and a permanent magnet D.C. motor. The D.C. motor (or A.C. servo-motor) controller has superior characteristics and this system performance was compared to P, PI and PID control for two severe load cases - a periodic step load and a four-bar linkage load. Simulation studies showed the schedule control method to be stable and in comparison to the PID controller to have 1) nearly the same speed of response but without the overshoot found in PID control, 2) nearly the same mean speed error (~ O), 3) 12-50 times better reduction in speed fluctuation, and 4) the schedule controller gains were much easier to find than PID gains for this low-order, highly responsive system.