级联H桥SVG的无差拍控制  

Deadbeat Control of Cascaded H-bridge Static Var Generator

作　　者：于雁南[1,2] 杨荣峰[1] 徐殿国[1] 武健[1] 

机构地区：[1]哈尔滨工业大学电气工程系,哈尔滨150001 [2]黑龙江科技大学电气与控制工程学院,哈尔滨150027

出　　处：《高电压技术》2017年第1期50-58,共9页High Voltage Engineering

基　　金：国家自然科学基金重点项目(51237002);国家自然科学基金青年基金(51407043);哈尔滨科技创新人才研究专项资金青年后备人才项目(2016RAQXJ035);台达电力电子科教发展计划重大项目(DREM2012001);台达电力电子科教发展计划项目(DREG2014015)~~

摘　　要：为了提高级联H桥静止无功发生器(SVG)的动态性能和跟踪精度,提出了一种改进的无差拍控制方法。首先建立了级联H桥在dq坐标系下的数学模型和离散状态方程,然后在传统的无差拍控制算法的基础上分别设计了新的状态观测器对电流和电压进行观测。在常规离散状态观测器的基础上,预先计算设置的中间变量,提前预测下一拍指令电流,从而减小控制延时。并采用了改进的重复预测算法预估输出电压以提高电压跟踪精度。最后,搭建了级联H桥SVG的MATLAB仿真模型和10 k V等级的系统试验样机,并与常规状态观测器算法进行了对比仿真实验。结果表明:与常规状态观测器相比,采用所提出的改进无差拍控制算法得到的电网电流总谐波失真(THD)数值从3.21%减小到1.18%;该算法在有效提高系统的跟踪精度和动态响应速度的同时,深度消除了采样延时所带来的跟踪滞后问题,有效提高系统的无功补偿性能。In order to enhance the dynamic performance and tracking precision of cascaded H-bridge static var generator （SVG）, we put forward an improved deadbeat controlling method. Firstly, a mathematic model and discrete state equation of cascaded H-bridge SVG are built in dq coordinate system for follow-up analysis. Secondly, two new kinds of state ob- servers are designed and blended into the traditional dead-beat controlling algorithm to carry out current and voltage observation independently. By the use of these new designed observers and combined with the conventional discrete state observer, the intermediate variables can be calculated previously, and the instructive current of next beat can be observed in advance. For the purpose of improving the voltage tracking accuracy, an upgraded repetitive prediction algorithm is adopted to predict the output voltage. Finally, the MATLAB simulating models and an experimental SVG prototype rated 10 kV are built and tested by using ordinary state observing algorithm and the controlling method presented by this pa- per, respectively. The comparative experiments shown that, compared with the traditional state observing algorithm, the improved dead-beat controlling method mentioned by this paper can obviously reduce the total harmonic distortion（THD） to 1.18% from 3.21%. The algorithm can effectively improve the SVG system＇s tracking precision and dynamic response, dramatically reduce the tracking lag caused by sample delay, and the reactive power compensating capability can be im- proved extraordinarily.

关 键 词：无差拍控制 离散状态观测器 重复预测算法 静止无功发生器 级联H桥 电流控制器 总谐波失真 

分 类 号：TM761[电气工程—电力系统及自动化]