EVOLUTIONARY BASED OPTIMAL DESIGN OF SR MOTORS VIA NEUROFUZZY MODELING OF NATURAL FREQUENCIES OF CYLINDRICAL SHELLS

Analysis of dynamic behavior of cylindrical shells is essential in design wherever it is used. Equations of shell vibrations are partial differential equations of order eight which their exact solution is possible only in special cases with a few known boundary conditions and with a lot of simplified assumptions. On the other hand finite element method does not yield a lumped model or a general solution for natural frequencies of cylindrical shells. In this paper natural frequencies of cylindrical shells in a wide range of dimensions are obtained with either exact solution or finite element method and they are applied to training of a Locally Linear Neurofuzzy Network. Finally a general model for calculation of natural frequencies of cylindrical shells has been proposed. Then the model has been applied for optimal design of a Switched Reluctance motor with the evolutionary algorithms as optimization method.

Keywords:

Natural Frequency , Cylindrical Shells , Finite Element Analysis , NeuroFuzzy Model , Evolutionary Algorithms