基于改进的二代非支配排序遗传算法对电子变压器多目标优化  

Multi-objective Optimization of Electronic Transformer Based on an Improved NSGA-Ⅱ Algorithm

作　　者：杨慧娜[1] 张永帅[1] 刘钢[1] 

机构地区：[1]华北电力大学电气与电子工程学院,北京102206

出　　处：《科学技术与工程》2015年第19期139-145,共7页Science Technology and Engineering

摘　　要：对多目标、多变量优化方法研究的基础上,提出了改进的二代非支配排序遗传算法。在该算法中,通过增加种群多样性和提高个体竞争力,有效地减少了早熟收敛现象的发生;同时,通过种群分割操作,大大减少了交叉运算的计算量。依据这一改进算法,建立了三维优化模型,对电子变压器进行了多目标优化设计,获得了电子变压器优化设计参数,使其体积更小、效率更高,更容易找到全局最优解。与非支配排序遗传算法(NSGA)和二代非支配排序遗传算法(NSGA-Ⅱ)相比,改进的二代非支配排序遗传算法在电子变压器优化设计方面具有明显的优势。最后,依据优化结果,制作了一台磁芯材料为超微晶合金的高频变压器,温度校核结果表明了此优化方法的可行性。Multi-objective and multi-variable optimization method has been studied and an improved NSGA-Ⅱ algorithm is proposed. In this algorithm, the prematurity phenomena were depressed by enlarging the diversity of population and strengthening the competitiveness and the amount of calculation was reduced by population division. With this algorithm, the multi-objective optimal design of electronic transformer was carried out. And the three-di- mension optimization models were established to determine the optimum design parameters of the transformer, which has smaller volume and higher efficiency. The predominance of the Improved NSGA- Ⅱ algorithm in optimal design of electronic transformer was obvious compared with the NSGA and NSGA-Ⅱ. An electronic transformer, which had a core of nanocrystalline alloy, was made based on the optimal results and the temperature check showed the possibility of optimization.

关 键 词：多目标优化 早熟收敛 改进的二代非支配排序遗传算法 电子变压器 

分 类 号：TM43[电气工程—电器]