Adaptive tuning of a PI speed controller for a brushless DC motor: optimum speed control using a neural network

This paper presents a method of adjusting the gains of a PI speed controller adaptively to give the minimum integral absolute error between the speed demand and the output response for a brushless DC drive. A feedforward neural network, trained using a genetic algorithm, adjusts the controller gains to improve the speed response to changes in speed demand and load. Experimental results demonstrate how the gain values vary with speed demand and load set point.