模块化多电平变换器的电容电压卡尔曼滤波预测方法

Kalman Filter Predictive Method for Capacitor Voltage of Modular Multilevel Converter

为避免子模块电压互感器发生故障对直流配电网模块化多电平变换器(modular multilevel converter,MMC)安全可靠运行造成威胁,需为MMC配置备用测量装置。工程上通常为所有子模块配置备用电压互感器以保证高可靠测量,但使得硬件结构较为复杂。为此,文中提出一种适用于中高压MMC测量系统备用模式的电容电压卡尔曼滤波预测方法,并融合快速模型预测控制(fast model predictive control,FMPC)策略提升电能变换稳定性。首先,为各桥臂仅配置一个电流互感器,无需任何电压互感器,并构建卡尔曼滤波观测器,提出卡尔曼滤波预测方法;进而根据电容电压预测值建立FMPC目标优化函数,确定桥臂最优电平数和子模块投切情况,在保证MMC输出电流和环流跟踪性能及电容电压预测准确性的前提下,显著简化测量系统硬件结构。最后,搭建21电平仿真模型和三电平实验平台,通过多个算例和工况验证所提方法的可行性和实用性。

To avoid the threat of the submodule(SM)voltage sensor failure to the safe and reliable operation of the modular multilevel converter(MMC)in DC distribution grid,the standby measurement equipment should be configured.Generally,the SMs are equipped with standby voltage sensors to ensure the highly reliable measurement,while complicating the hardware structure in engineering.Therefore,a Kalman filter predictive method for capacitor voltages of the standby mode of MMC measurement system applied to medium and high voltage applications is proposed,integrating the fast model predictive control(FMPC)strategy simultaneously for improving the stability of power conversion.Firstly,each arm is equipped with only one current sensor without any voltage sensors,and the Kalman filter observer is constructed.Then,the Kalman filter predictive method is proposed.Furthermore,in terms of the predicted values of the capacitor voltages,the objective optimization cost function of FMPC is established to determine the optimal arm level number and the switching conditions of SMs.With the proposed method,the hardware structure of the measurement system could be significantly simplified,on the premise of ensuring the tracking performance for the output currents and circulating currents,and the prediction accuracy for the capacitance voltages.Finally,the 21-level simulation model and 3-level experimental setup of MMC are built.The feasibility and practicability of theproposed method are validated by the simulation examples and experiment conditions.