Computationally Efficient DMPC for Three-Level NPC Back-to-Back Converters in Wind Turbine Systems With PMSG
Computationally Efficient DMPC for Three-Level NPC Back-to-Back Converters in Wind Turbine Systems With PMSG
复制标题
DOI：
10.1109/tpel.2016.2637081
复制
发表时间：
2017-10
影响因子：
6.7
通讯作者：
Zhenbin Zhang;C. Hackl;R. Kennel
中科院分区：
文献类型：
--
作者：
Zhenbin Zhang;C. Hackl;R. Kennel
直接模型预测控制（DMPC）是一种极具吸引力的控制方法，尤其适用于多电平变流器。然而，经典DMPC方案的计算时间随着切换状态的数量呈指数增长，与基于调制的控制器相比，实时实现可能不可行。针对永磁同步发电机风力发电系统中三电平中性点夹持背靠背变流器的转矩和功率控制问题，提出了具有六边形候选区域（HCR）和三角形候选区域（TCR）两种计算效率较高的DMPC控制方案。通过对候选区域的适当选择，可以大大减少合理切换状态的数量，HCR和TCR分别节省了55%和83%的计算时间。通过现场可编程门阵列（FPGA）系统的实时实现和实验室搭建的测试平台的测量结果，比较并验证了所提出的DMPC方案的计算效率和控制性能的提高。所提方法的控制性能与经典DMPC相当，同时大大减少了计算时间。
Direct model predictive control (DMPC) is an attractive control method, in particular, for multilevel converters. However, the computation time for classical DMPC schemes increases exponentially with the number of switching states and, compared to modulation-based controllers, real-time implementation may not be feasible. In this paper, two computationally efficient DMPC schemes with hexagon candidate region (HCR) and triangle candidate region (TCR) for torque and power control of three-level neutral-point clamped back-to-back converters in wind turbine systems with permanent-magnet synchronous generator are proposed. By an appropriate selection of the candidate regions, the number of reasonable switching states is drastically reduced which saves computation time up to 55% for HCR and up to 83% for TCR, respectively. The computational efficiency improvements and the control performances of the proposed DMPC schemes are compared and validated by real-time implementations on an field programmable gate array (FPGA) system and by measurement results at a lab-constructed test bench. The achieved control performance of the proposed methods is comparable with that of the classical DMPC, while the computation times are drastically reduced.