NPC型三电平并网逆变器自适应模型预测控制  

Adaptive Model Predictive Control for NPC-type Three-level Grid-connected Inverter

作　　者：刘春喜[1] 田宝奇 刘志乐 赵昱诚 张天琦 LIU Chunxi;TIAN Baoqi;LIU Zhile;ZHAO Yucheng;ZHANG Tianqi(School of Electrical and Control Engineering,Liaoning Technical University,Huludao125105,China)

机构地区：[1]辽宁工程技术大学电气与控制工程学院,葫芦岛125105

出　　处：《电力系统及其自动化学报》2023年第1期143-151,共9页Proceedings of the CSU-EPSA

摘　　要：针对中点箝位NPC(neutral point clamped)型三电平并网逆变器模型预测控制中模型参数扰动导致的控制精度低的问题,研究了一种改进的自适应模型预测控制策略。通过在遍历寻优过程中迭代控制参数以估计误差最小值,识别模型参数扰动量来更新系统模型参数,同时对控制集的作用区域进行优化,并结合两步预测法来减小遍历寻优过程的计算量。通过仿真对所提方法进行了验证,结果表明该方法提高了系统的控制精度和响应速度。Aimed at the problem of low control accuracy caused by the perturbation of model parameters in the model predictive control(MPC)of a neutral point clamped(NPC)-type three-level grid-connected inverter,an improved adaptive MPC strategy is studied. Under this strategy,the minimum error value is estimated by iterating the control parameters in the ergodic optimization process,and the perturbation of model parameters is identified to update the system model parameters. At the same time,the action area of the control set is optimized,and the calculation amount of the ergodic optimization process is reduced by combining a two-step prediction method. The proposed method is verified by simulations,and results show that it improves the control accuracy and response speed of the system.

关 键 词：三电平并网逆变器 模型预测控制 参数扰动 自适应 遍历寻优 

分 类 号：TM464[电气工程—电器]