Proceedings of the 3 rd ICEENG Conference, 14-16 May, 2002 1 Military Technical College 3 3 rd International Conference Kobry ElKobbah On Electrical Engineering Cairo, Egypt ICEENG 2002 GENETIC DESIGN OF A SWITCHED RELUCTANCE MOTOR BASED ON MINIMUM OUTER VOLUME Amged El-Wakeel * and A.C. Smith ** ABSTRACT This paper presents a new method for improving the design of a Switched Reluctance Motor. The main disadvantages of classical methods of design are that they are generally not optimised and don't offer a global view of the feasible design surface. The proposed method optimises the outer volume of a Switched Reluctance Motor (SRM) using Genetic Algorithms and gives at the same time a strong view about the feasible surface of the motor design. The approach combines a full analytical procedure for motor performance estimation with a Genetic Algorithm to achieve an optimal design. The motor performance analysis combines a set of non-linear equations and a time stepping algorithm to represent the static and steady state operation of the motor. A comparison between different genetic optimisation parameters has been done in order to explore their effects on the optimisation procedure. Also, a comparison between the genetic design and the classical design method is made in order to explore the advantages and disadvantages of each technique. Finally, a FEA analysis for the optimised motor has been done to verify the final design of the motor. KEY WORDS Switched Reluctance Motors, Genetic Algorithms, Optimisation methods, Finite Element Analysis. 1.INTRODUCTON A variety of publications on the design of SRMs are available and provide a wealth of information regarding the design and construction [1:6]. Although the design of these motors is well documented, there are only a few papers that deal with design and construction optimisation [7:10]. This can be associated to the high nonlinearly and difficulties encountered in accurately estimating the performance of these motors. Some of these papers apply classical hill - climbing methods for shape optimisation [7,8] and others apply an analysis program to estimate the motor performance under a wide range of parameter variations (Exhaustive method) from which, the optimum value of certain parameters can be estimated [9,10]. *PhD Student, Dept of Electrical Engineering and Electronics, UMIST, UK. **Associate Professor, Dept of Electrical Engineering and Electronics, UMIST, UK. ICEENG 2002