Aplikace umělých neuronových sítí v řízení elektrických pohonů

Abstract

The thesis deals with using of artificial neuron network in control of electrical drives when a sensorless control of rotation by means of the velocity estimator and controlling of electrical drive with predicative controlling is being solved. Generating and training of particular neuron networks is implemented in the programming environment Matlab-Simuling by means of the Neural Network Toolbox. The objective of the thesis is a description of controlling principles by using neuron networks and comparison of the theoretical knowledge and simulation results.