Date of Award

Summer 1999

Document Type

Thesis - Restricted

Degree Name

Master of Science (MS)


Mechanical Engineering

First Advisor

Arkadan, A. A.

Second Advisor

Heinen, James A.

Third Advisor

Hock, Jeffery L.


Switched reluctance motors (SRMs) are of interest for high performance applications because of their reliability and high torque to weight ratio, but the major drawback is that the system identification has proven difficult because of its highly nonlinear properties. More advanced control techniques can be developed if an identifier can be utilized. The main goal of this work is to model the switched reluctance motor and drive system using the method of artificial intelligence (AI). In the first part of the work, a traditional finite element and state space model is developed. Based on the results from the simulation, in the second part, an artificial neural network (ANN) is applied in order to obtain a fast and accurate prediction of the performance of the motor and drive system at both normal and fault conditions. In further research as described in the third part, the ANN parameters are optimized using genetic algorithms (GA) that find an optimal ANN rather than train a random selected one. Thus a more systematic method is given to perform even better predictions.



Restricted Access Item

Having trouble?