ANN FUNCTIONS FOR HIGH FREQUENCY PMSM DESIGN

Abstract

This paper proposes ANN functions for high fundamental frequency permanent magnet synchronous machine general design with suitable machine materials. The final design offers significant reductions in both weight and volume in a power range of 5:500 Kw. A rotor length to diameter ratio is used as an important design parameter. The results are well depicted by figures for a number of machines sizing. The simulation of generators sizing is performed using MATLAB. Then, this paper introduces three neural network units. Later, deduction of algebraic nonlinear function which, connects between inputs and outputs for neural network is presented to be used without training each time. Finally, a complete analytical design example is introduced with study of THD and Rotor losses.