الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 198 |  |  |
| | | from | 198 | | |
Abstract
Switched Reluctance Machines (SRMs) are receiving significant attention from industries in the last decade. SRMs are rather inexpensive, reliable and weigh less than other machines of comparable power outputs. Although the design principles of the machine are available as a concatenation of many different sources, the need for a unified, stepbystep design procedure from first principles of electromagnetic is an absolute requirement. The thesis presents an integrated computer program (SRMcad), written in Visual Fortran. The program takes over the design problems of the switched reluctance motor. The program simulates an integrated algorithm, which is a combination of a proposed design algorithm and a machine analysis algorithm.The computation processing is executed in two stages. In the first stage, the program determines a preliminary design according to the given rating and design specification. All the magnetization characteristics are calculated in tabulated form according to dimensions of the preliminary design. In the second stage, the expected performance is determined for each couple, switching on (?on) and conduction period (?cond), picked from the switching condition list. The machine performance is represented by the average torque, the rms current, and the efficiency. If this performance satisfies the machine rating as a motor, it will be stored in a permanent list. Having the expected performance for all couples included in the switching conditions list, a search process starts within the permanent list to pick up the best performance in this design iteration cycle. The best performance is that one which ensures the required motor rating at highest efficiency. If the efficiency of the best performance is close to the given expected efficiency the design iteration stops. The program algorithm and logics are then tested using a reference design example. Results are verified and are found very close to reference example results.