基于布谷鸟算法的双向直流电子变压器谐振参数优化

OPTIMIZATION OF RESONANT PARAMETERS OF BIDIRECTIONAL DC ELECTRONIC TRANSFORMER BASED ON CUCKOO ALGORITHM

目前通常使用基波分析法(FHA)结合试验迭代法对谐振参数进行设计。针对FHA法不够直观等问题,通过分析和研究不同谐振点区域内直流增益、直流输入阻抗、谐振变换器损耗与损耗率等谐振网络参数等与谐振变换器谐振频率特性的关系,进而得到变换器性能影响因数。针对试验法优化效果有限,利用布谷鸟搜索算法进行优化。该算法具有莱维(Lévy)飞行和巢寄生行为,可对存在不可导点的目标函数进行优化,具有更强的全局搜索能力,可避免陷入局部最优,且求解精度高。最后,搭建一台实验样机,通过实验验证了在宽输入电压下,可稳定输出电压;在全负载范围下,该优化设计方法优于FHA法。

In order to improve the conversion efficiency of isolated DC-DC in bidirectional DC electronic transformer,soft switching resonance technology is generally used.The selection of resonance parameters directly affects the performance of electronic transformer.At present,fundamental wave analysis(FHA)combined with experimental iteration method is usually used to design resonance parameters.Aiming at the problem that FHA method is not intuitive enough,this paper analyzes and studies the relationship between resonant network parameters and resonant frequency characteristics of resonant converter,such as DC gain,DC input impedance,resonant converter loss and loss rate in different resonant point regions,and then obtains the performance influence factor of resonant converter.In view of the limited optimization effect of the experimental method,this paper uses cuckoo search algorithm to optimize.The algorithm has the advantage of lévy(It has stronger global search ability,can avoid falling into local optimum,and has high precision.Finally,an experimental prototype is built to verify that the output voltage can be stabilized under wide input voltage;In the full load range,the optimal design method is better than FHA method.