摘要
电力电子变压器电磁暂态详细模型存在矩阵维数高、仿真步长小、速度慢的问题。基于半隐式延迟解耦电磁暂态仿真方法,该文将其应用于级联H桥型电力电子变压器的电磁暂态快速仿真研究中,提出一种解耦电路简单、导纳矩阵恒定、易于并行、仿真效率高的解耦和快速仿真方法。该文首先从状态方程出发,利用矩阵分裂和延迟技术,建立双有源桥的半隐式延迟解耦模型;然后给出外端口及级联H桥采用不同串、并联组合时的等值解耦电路,实现了各变换级以及不同模块间的细粒度解耦,给出了计算流程,并分析了方法的特点。最后,通过算例验证了该文方法的准确性和有效性。
The detailed electromagnetic transient model of the power electronic transformer(PET)has the problems of high matrix dimension,small simulation step and slow speed.Based on the semi-implicit delay decoupling electromagnetic transient simulation,this paper applies it to the electromagnetic transient fast simulation research of the cascaded H-bridge power electronic transformer(CHB-PET),and proposes a decoupling and fast simulation with simple decoupling circuit,constant admittance matrix,and easy parallel and high simulation efficiency.In this paper,first,based on the state equation,a semi-implicit delay decoupling model of dual active bridge(DAB)was established by using the matrix splitting and the delay technique.Then,the equivalent decoupling value path was given when the outer port and the cascaded H-bridge adopted different series and parallel combinations.The fine-grained decoupling of each transformation stage and between different modules were realized.The calculation flow was given,and the characteristics of the method were analyzed.Finally,an example is given to verify the accuracy and effectiveness of the proposed method.
作者
许明旺
马嘉昊
李蕴红
王潇
姚蜀军
汪燕
韩民晓
XU Mingwang;MA Jiahao;LI Yunhong;WANG Xiao;YAO Shujun;WANG Yan;HAN Minxiao(School of Electrical and Electronic Engineering,North China Electric Power University,Changping District,Beijing 102206,China;North China Electric Power Research Institute Co.,Ltd.,Xicheng District,Beijing 100045,China)
出处
《电网技术》
EI
CSCD
北大核心
2023年第6期2503-2511,共9页
Power System Technology
基金
国家自然科学基金项目(52077077)
冀北电力科学研究院科技项目(KJZ2021029)。
关键词
电磁暂态仿真
半隐式延迟解耦
级联H桥型电力电子变压器
细粒度解耦
并行计算
electromagnetic transient simulation
semi-implicit delay decoupling
cascaded H-bridge power electronic transformer
fine-grained decoupling
parallel computing