MMC的新型模型预测控制与子模块优化均压

The Improved Predictive Control Research on Sub-module Voltages Balancing for MMC

模块化多电平换流器(MMC)的模型预测控制是一种有效的非线性控制方法,但电容电压平衡控制效果不理想,当控制系统中存在多个控制目标时,系统最优控制目标函数选取存在一定困难。针对以上问题,提出了一种外环采用新型神经元比例-积分-微分(PID)控制、内环采用电流模型预测控制的双闭环控制策略。首先推导了考虑电容充放电过程的MMC离散状态方程;其次将综合考虑控制目标电流跟踪、子模块(SM)均压和器件开关频率建立了代价方程,针对控制目的侧重点不同,同时保证其他指标满足系统要求,设计了3种控制参数方案以供选择。最后通过搭建9电平MMC仿真控制系统和小功率实验平台对所述理论进行验证。结果表明了此处所提控制方案的有效性和可行性。

The model predictive control of modular multilevel converter（MMC） is an effective nonlinear control method, but in the field of the capacitor voltage balancing, the method is not ideal.Moreover, it is difficult to establish the op- timal cost function when there exist multiple control objectives.To solve the above problems,an improved dual close- loop control structure is proposed,which adopts current model predictive control as the inner loop and uses single neuron proportion integration differentiation （PID） control for the outer-loop.In the specific,firstly, the discrete state eq- uation is derived considering charging and discharging process of capacitors.Secondly, taking the current tracking, sub- module （SM） voltage balancing and switching frequency into consideration, the multi-objective function is established. To achieve different control objectives and ensure other indicators reach the request, three control schemes for param- eters are designed for adoption.Finally,a nine level MMC simulation control system and a small power experimental test platform is built to verify the effectiveness and feasibility of the proposed strategy.