基于模型预测控制的固态变压器控制策略研究

摘要

When converter in Solid State Transformer (SST) operates in rectifier mode, the AC-side current predictive control is applied to achieve unity power factor operation and smooth DC voltage. Output voltage prediction is adopted to converter operating in inverter mode, improving the ability against load disturbances. In this paper, two nonzero vectors and two zero vectors are adopted in fixed switching frequency model predictive control. According to the relationship between duty cycle and cost function, the optimal duty cycles of each vector are applied and rational switching patterns are chosen. This strategy needs no pulse width modulation controller, which can avoid instability caused by complex parameters settings of proportional integral controller and reduce high sampling frequency of single nonzero vector model predictive control. The simulation results show that the voltage and current of each level have good dynamic response and the SST is robust against load disturbances in different typical operation conditions, meeting the performance requirements of microgrid for SST.%固态变压器(SST)中的变流器在整流状态时,通过交流侧电流预测控制,使其工作于单位功率因数且直流侧电压保持稳定;逆变状态时,采用输出电压预测控制,以提高输出电压稳态精度和动态性能.提出了固定开关频率模型预测控制选取双非零矢量和零矢量进行预测,根据各矢量占空比与优化函数的关系,确定一个控制周期内各矢量的作用时间和开关序列.固定开关频率模型预测控制无需脉宽调制控制器,避免了比例积分控制器参数整定引起的不稳定问题,且有效降低了单非零矢量预测控制中过高的采样频率.仿真研究表明,固定开关频率模型预测控制的SST 在各典型工况下,各级电压、电流具有良好的动态响应和抗负载扰动能力,可满足微网对SST 的控制要求.