Multi-objective Optimal Design of Dual Stator Winding Induction Generators Based on Genetic Algorithm and Finite Element Analysis

Abstract

This paper proposes a multi-objective optimized design procedure for Dual Stator Winding Induction Generators (DSWIGs) applied for wind energy harvesting. The proposed design procedure concentrates on efficiency optimization as well as the power density optimization of a 15kW DSWIG. In this regard, the optimization problem is formulated based on the DSWIG geometry, magnetic and electric equations, first. Then the multi-objective optimization is fulfilled implementing the Genetic Algorithm (GA), which has numerous advantages in comparison with other evolutionary algorithms. A novel fitness function is introduced which determines the priority of the objective functions depending on the expected terms. The proposed fitness function, includes two variables that are weighted by $a$ and $b$ coefficients. Increasing each of these coefficients toward each other, further improvement is achieved. Now by plotting the objective function variations toward different amounts of a/b ratio, according to the application requirements and the desired objectives, the best a/b ratio is selected. Finally, optimization results are evaluated through the two-dimensional finite element based method simulation in ANSYS/MAXWELL environment. The simulation results confirm the effectiveness of the proposed optimization procedure.