电工技术学报2025，Vol.40Issue(6)：1695-1706,12.DOI:10.19595/j.cnki.1000-6753.tces.240349
一种低损耗的感应滤波变压器结构参数分层设计方法
A Low Loss Hierarchical Design Method for Structural Parameters of Inductive Filtering Transformer
摘要
Abstract
The industrial rectifier system is responsible for converting AC and DC electrical energy,and the electrical operating environment is characterized by high power,high current,and low voltage.It faces significant challenges,such as severe harmonic pollution and power loss.Transformer loss contributes to 40%of the total loss in the industrial rectifier system.Therefore,reducing transformer loss is key to managing the operational environment of the industrial rectifier system.In recent years,the research on the reduction of transformer loss has primarily focused on three areas:improving the material process,optimizing the physical structure,and managing the electromagnetic environment.Previous literature has rarely researched the zero-impedance matching design and performance optimization of inductive filtering transformers. This paper aims to develop a new hierarchical optimization strategy for the multi-objective optimization of the inductive filtering rectifier transformer,considering winding impedance matching.The transformer's structural parameters reduce operational loss by establishing hierarchical architecture through Sobol sensitivity analysis of parameters.Firstly,the filtering mechanism is revealed based on the topological structure of the inductive filtering rectifier transformer.According to the relationship between voltage and current,impedance-matching conditions for implementing inductive filtering technology are obtained.The correlation between harmonic filtering performance and matching conditions is analyzed.A harmonic filtering rate is proposed,and a constraint of approximately zero equivalent leakage impedance is given for different harmonic filtering rates.Secondly,regarding the high dimension of the transformer structure parameters and the strong interaction effect between the parameters,the Sobol sensitivity analysis algorithm is used to stratify the optimization transformer parameters.Finally,a hierarchical optimization design method is proposed based on an improved differential evolution algorithm.The optimized parameters are used as the initial values of the unoptimized parameters in the local optimization.Then,the parameters of each level are optimized successively,which updates the optimization parameters of the next level.It can jump out of local convergence by reducing dimensions to find a better solution.This optimization method ensures a good induction filtering effect and reduces the operating loss of the transformer. Through case analysis,it can be seen that under the same number of iterations,the target average of total loss obtained by hierarchical optimization is 44.88 kW,and the optimal value is 42.37 kW.The target average of total loss obtained by control optimization is 47.21 kW,and the optimal value is 47.15 kW.The optimal and mean values of hierarchical optimization are better than control optimization.Compared with control optimization,hierarchical optimization can effectively improve the optimization ability and stability.The finite element simulation results show that the current harmonic distortion rate at the high voltage network side is reduced from 8.84%to 4.29%,which meets the design requirements of filtering performance.Compared with the global optimization method,the hierarchical optimization method can further reduce the transformer loss by 11.3%,which verifies the feasibility of the optimal design.关键词
整流变压器/灵敏度分析/阻抗匹配/分层优化设计Key words
Rectifier transformer/sensitivity analysis/impedance matching/hierarchical optimization design分类
信息技术与安全科学引用本文复制引用
刘乾易,杨伊,刘芳,刘建峰,王傲..一种低损耗的感应滤波变压器结构参数分层设计方法[J].电工技术学报,2025,40(6):1695-1706,12.基金项目
国家自然科学基金(52207146)、湖南省重点领域研发计划(2023GK2007)和江苏省能源动力高端装备工程研究中心开放课题基金(JSNYDL-202401)资助项目. （52207146）
