# Design of model reference adaptive controller for electric drives

OData: XML JSON What's this?The goal of this thesis is to examine, how we can apply the model reference adaptive control for the regulation of DC motors, and to show its advantages over the traditional robust motor controllers.

In the first part the theory behind the DC motors and the power electronics supplying them is presented. In the next chapters a possible way to design a cascade speed controller is introduced. The current control is realised by a PI controller. To avoid the integral wind-up caused by saturation in the control loop, the controller is realised with FOXBORO architecture. The speed controller is also PI type. Its parameters are calculated by using the theorem of symmetrical optimum. This control system is to be used as comparison with the adaptive controller to be designed.

In the second part of the thesis the theorem of the model reference adaptive control is presented. I examine two different ways to realise the adaptation: one based on the MIT rule, and the other one based on the Lyapunov stability theory. The adaptation equations are deduced for both methods, in case the plant to be controlled was a first order system. After the comparison of the behaviours of both controllers, I decide which one is more appropriate for the control of the DC motor.

The last part of the thesis describes the design of the model reference speed controller for the DC motor. Since the adaptation equations were only deduced in case of first order systems, the speed control loop serving as reference model has to be approximated by an appropriate first order system. To avoid the oscillation in the output signal of the controller, caused by large input signals, the reference model has to be modified by a feedback of the error signal. During the simulation it has to be taken into account, that while the motor is in saturation the controller isn’t working, so for this period the adaptation mechanism has to be stopped. After determining the optimal values for the adaptation parameters, I run the simulation, and compare its results with the ones gained from the motor driven by robust controller.