LCL电压型整流器模型预测优化控制

Model Predictive Optimal Control of LCL Voltage Source Rectifier

传统的LCL整流器模型预测控制需要附加有源阻尼算法MPC-AD(model predictive control-active admping)来消除LCL滤波器引起的谐振问题,此时由于网侧电流由整流器侧电流间接控制,导致其控制性能不佳。为此提出了一种基于LCL滤波的电压型整流器VSR(voltage source rectifier)有限控制集模型预测优化控制策略。利用模型预测控制中多变量控制的特性,通过构建单个代价函数同时控制网侧电流矢量、滤波电容电压矢量以及整流器侧电流矢量(MPC-i_(1)i_(2)u_(c)),无需状态变量反馈有源阻尼,然后通过所构造的代价函数J来选取下一时刻最优开关状态作用于系统。为验证所提出方法的可行性,进行仿真实验并搭建测试样机。通过与传统的MPC-AD控制方法进行比较,验证了所提优化控制策略的可行性与优越性。

The conventional model predictive control of an LCL rectifier model requires an additional active damping(MPC-AD) algorithm to eliminate the resonance problem caused by the LCL filter,which results in a poor control performance since the grid-side current is indirectly controlled by the rectifier-side current.To this end,a finite control set model predictive optimal control strategy for LCL-filtered voltage source rectifier(VSR) is proposed.The multivariable control feature of MPC is exploited,and a single cost function is constructed to simultaneously control the grid-side current vector,filter capacitor voltage vector and rectifier-side current vector(MPC-i_1i_2u_c),which does not need state variables to feed back the active damping.Then,the constructed cost function J is used to select the optimal switching state that will be applied to the system at the next moment.To verify the feasibility of the proposed method,a simulation experiment was carried out and a test prototype was built.The comparison with the conventional MPC-AD control method verified the feasibility and superiority of the proposed optimal control strategy.