多电平级联H桥逆变器的模型预测控制策略

针对多电平级联H桥三相逆变器的控制性能提高问题,设计了一种新型的模型预测控制(model predictive control,MPC)策略。MPC控制器基于预测模型对系统未来状态进行预测,然后将计算得到使所设计成本函数最小化的开关矢量进行输出。不同于传统两电平逆变器,多电平级联H桥逆变器的数学模型较为复杂,包含有大量电压矢量,导致计算负担较重。为此,通过设置和选择有效电压矢量子集,显著减少了得到最优电压矢量所需的计算量,同时不影响控制性能。搭建了五电平和九电平三相级联H桥逆变器实验平台开展了相关实验,实验结果验证了新型MPC控制器具有较优的控制性能。

Model Predictive Control Strategy for Multi-level Cascaded H-bridge Inverter

In view of the issue of control performance improvement for multilevel three?phase cascaded H?bridge inverter,a kind of new model predictive control(MPC)strategy has been designed.The MPC controller predicts the future condition of the system based on the prediction model,and then calculates the switching vector that minimizes the designed cost function for output.However,unlike the traditional two?level inverter,the mathematical model of multilevel cascaded H?bridge inverter is complex and contains a large number of voltage vectors,resulting in a heavy computational burden.For this reason,by setting and selecting effective voltage vectorsubset,the amount of calculation required to obtain the optimal voltage vector is significantly reduced without affecting the control performance simultaneously.The experimental platform of five?level and nine?level three?phase cascade H?bridge inverter is set up to develop related experiments.The superior control performance of the new MPC controller is verified by the experimental result.