一种减小有限控制集模型预测控制开关状态计算时间的方法

A method for finite-control-set model predictive control to reduce computation time of switch states

有限控制集模型预测控制(finite-control-set model predictive control,FCS-MPC)依靠被控逆变器所具有开关状态对控制目标滚动寻优计算。为减小多电平逆变器中滚动计算开关状态所需时间,该文对经典有限控制集模型预测电流控制进行改进。首先,利用控制系统的离散数学模型和参考电流求得被控逆变器在此参考电流下所应输出的参考电压值。此时,引进SVPWM中的分扇区计算概念,判断参考电压所处扇区,最后将该参考电压所处扇区内所包含的开关状态来循环寻优计算。最后Matlab/Simulink仿真平台搭建二极管钳位型五电平并网逆变器系统模型,通过仿真结果,得以验证改进算法的正确性和实用性。

Finite-control-set model predictive control （FCS-MPC） can realize rolling optimizationcalculation for control objectives based on the switching state of controlled inverter. To minimize thetime of rolling calculation of switching state of multilevel inverter, predictive current control ofclassical finite control set model was improved in this paper. First of all, reference voltage thatshould be output by controlled inverter under reference current was calculated based on thedispersed mathematical model of control system and reference current. Then, the concept of per-sector in SVPWM was introduced to determine the sector of reference voltage. Finally, cyclingoptimizing calculation was carried out based on the switching state contained in the referencecurrent sector. Finally, the Matlab/Simulink simulation platform was provided with a diode clampingtype five-level grid inverter system model, which verified the correctness and practicability of thealgorithm proposed in this paper.