三电平并网逆变器的占空比优化MPC定频算法

MPC Fixed Frequency Algorithm for Duty Cycle Optimization of Three-level Grid-connected Inverter

经典的三电平并网逆变器模型预测算法具有计算量大、开关频率不固定、权重因子难以选择等缺点,这限制了其在并网逆变器领域的发展。为了解决以上问题,此处提出一种基于占空比优化的三电平并网逆变器的有限控制集模型预测定频算法,该算法采用基于无差拍原理的扇区分区方法,选择出每个控制周期需遍历的7个真实开关矢量,然后依据代价函数遴选出最优开关矢量和其绑定的占空比优化矢量,并分别以最优矢量作用时间依次作用于逆变器。该算法实现了能在尽量保持较小计算量的情况下,还具有固定的开关频率与并网电流纹波显著减少等特点。最后,实验与仿真结果表明,此处所提算法较之经典的模型预测方法有较大改善,且具有较大应用价值。

The classical three-level grid-connected inverter model prediction algorithm has the disadvantages of large amount of calculation,switching frequency is not fixed,and weighting factor is difficult to choose,which limits its development in the field of grid-connected inverter.In order to solve the above problems,a finite control set model predictive fixed frequency algorithm is proposed based on duty cycle optimization of a three-level grid-connected inverter.The algorithm uses a sector partition method based on principle of deadbeat,and selects each 7 real switch vectors to be traversed in each control cycle,then select optimal switch vector and its bound duty cycle optimization vector according to the cost function,and act on inverter in turn with the optimal vector action time.It can achieve the characteristics of fixed switching frequency and significantly reduce grid-connected current ripple while keeping the calculation amount as small as possible.Finally,the experimental and simulation results show that the algorithm proposed is greatly improved compared with the classical model prediction method and has greater application value.