双有源桥变换器移动离散控制集无模型预测电压控制策略  

Model-Free Predictive Voltage Control with Moving-Discrete-Control-Set for Dual Active Bridge Converters

作　　者：尹政 邓富金 王青松[1] 詹昕 黄堃[3] Yin Zheng;Deng Fujin;Wang Qingsong;Zhan Xin;Huang Kun(School of Electric al Engineering Southeast University,Nanjing 210096 China;Yangzhou Power Supply Company of State Grid Jiangsu Electric Power Co.Ltd,Yangzhou 225000 China;State Grid NARI Technology Company Ltd.,Nanjing 211106 China)

机构地区：[1]东南大学电气工程学院,南京210096 [2]国网江苏省电力公司扬州供电公司,扬州225000 [3]国电南瑞科技股份有限公司,南京211106

出　　处：《电工技术学报》2025年第6期1853-1863,共11页Transactions of China Electrotechnical Society

基　　金：国家重点研发计划“政府间国际科技创新合作”重点专项资助项目(2022YFE0196300)。

摘　　要：传统的双有源桥(DAB)变换器模型预测控制(MPC)依赖于系统参数的准确性,针对该问题,该文提出一种DAB移动离散控制集无模型预测电压控制(MDCS-MFPVC)策略,消除了系统参数对DAB变换器电压控制的影响。首先,分析参数失配对传统MPC的影响。其次,建立DAB变换器离散数据模型,替代传统的数学模型,提高了预测电压控制的鲁棒性。再次,通过数据模型误差反馈和多输入变量的设置,结合最小二乘法对离散数据模型进行实时辨识;结合系统采样和数据模型辨识结果,实现未来时刻的电压无模型预测,并通过价值函数评估最优移相角,应用在下一个控制周期。该方法可以实时辨识DAB系统数据模型,消除了参数失配对MPC的影响,保证了输出电压的鲁棒性。最后,通过搭建DAB变换器仿真和实验系统,验证了所提方法的有效性。In power electronics,the dual active bridge(DAB)converter stands as a pivotal component in various applications,from electric vehicle charging stations to renewable energy systems.The conventional model predictive control(MPC)strategies for the DAB converter are inherently tethered to the precision of system parameters.Any deviation or uncertainty in these parameters can lead to suboptimal performance or instability in the voltage control loop.Based on a moving-discrete control set,this paper proposes a free predictive voltage control(MDCS-MFPVC)strategy to eliminate the influence of system parameters on the voltage control of the DAB converter.Firstly,the influence of parameter mismatch on conventional MPC is analyzed.Then,the discrete data model of the DAB converter is established to replace the conventional mathematical model and improve the robustness of predictive voltage control. Then, through the error feedback of the data model and the setting of multiple input variables, the discrete data model is identified in real time with the least square method. Combined with system sampling and data model identification results, model-free predictive of voltage at a future time is realized, and the optimal phase shift angle is evaluated by cost function and applied to the next control period. This method can identify the system data model of the DAB converter in real time, eliminating the influence of parameter mismatch on MPC, and ensuring the robustness of output voltage. Finally, the simulation and experiment system of the DAB converter is set up to verify the effectiveness of the proposed method. Parameter robustness simulation shows that when the output voltage reference is 50 V, the maximum output voltage error of the proposed MDCS-MFPVC method is 0.11 V, 13.89 V lower than the conventional MPC method. When the output voltage reference is 45 V, the maximum output voltage error of the conventional MPC and the proposed MDCS-MFPVC is 12 V and 0.098 V, respectively. The parameter robustness experiment shows that

关 键 词：双有源桥变换器 无模型预测 电压控制 离散移动控制集 参数鲁棒性 

分 类 号：TM464[电气工程—电器]