Optimised methodology for stepper motor simulation

The paper presents the effort of the authors to minimize the computational time and resources brought in electrical machines conception, design and analyses. The proposed methodology is a mixture between Finite Element Analysis and neural network interpretation of resulting data, with important conclusions about the minimum number of inputs asked by the desired accuracy. The methodology is illustrated on stepper motor due to its important place in positioning systems, computer and peripherals architecture, electric vehicles and because its significant interest in researcher's works.