Multi-objective optimum design of dry-type air-core reactor based on improved NSGA-II algorithm

Based on the analysis of the relationship of aluminum wire weight and power loss to the average wire diameter of dry-type air-core reactor, this paper proposed that it was more applicable to the engineering practice situation to make the power loss as an objective function with the optimum design of reactor. Taking both aluminum wire weight and power loss as objective functions, a multi-objective optimization model for dry-type air-core reactor was built. NSGA-II algorithm, an algorithm with good optimization effect, was used to solve the model. In order to increase the population diversity and improve searching ability of the algorithm, it was proposed that the dominated individuals of the population were selected in a certain proportion; before non-dominated sorting, partial variables which belonged to the repeated individuals and dense individuals of combined population were mutated in a certain probability; external nondominated set was set in elitist strategy and the individuals of which participate in tournament selection. By using the improved algorithm, the optimum design results of 50 kvar reactor show that the distribution of Pareto optimal solutions is more uniform.