基于DRNN自整定准PR控制的光伏并网系统研究  

RESEARCH ON QUASI-PR CONTROL BASED ON DRNN OF SELFTUNING FOR PHOTOVOLTAIC GRID-CONNECTED SYSTEM

作　　者：周振雄[1] 石硕 

机构地区：[1]北华大学电气信息工程学院,吉林132021 [2]国网湖北省电力公司荆州供电公司,荆州434000

出　　处：《太阳能学报》2017年第11期2932-2940,共9页Acta Energiae Solaris Sinica

基　　金：国家自然科学基金(61179012);吉林省科技创新团队项目(20150519023JH);吉林省教育厅项目(2014167);吉林市科技计划(2015131009)

摘　　要：以配置LCL滤波器的单相光伏并网逆变系统为研究对象,针对其非线性、时变特性,提出一种基于准比例谐振(quasi proportional resonant,QPR)控制的直流电压外环、并网电流内环和电容电流内环三闭环控制策略,并给出准PR控制器的离散化方法。针对准PR控制器参数固定无法适应电网参数变化,难以达到系统控制性能最优的不足,提出一种基于动态对角递归神经网络(diagonal recurrent neual network,DRNN)自整定准PR控制器并给出其控制算法,DRNN采用基于二阶梯度的递推预报误差算法(recursive prediction error,RPE),其较一阶梯度反向传播算法(propagation,BP)具有更快收敛速度。仿真对比分析和实验结果表明,采用上述控制策略的单相光伏并网逆变系统具有良好的输出电流质量和快速的动态响应性能,可有效避免并网电流谐振并实现并网电流高功率因数。The single-phase photovoltaic grid-connected inverter system with LCL filter is studied as the research objectin this paper. Based on the nonlinear and time-varying characteristics of PV grid-connected system,the three-closed loopcontrol strategy which consist of DC side voltage outer loop,grid-connected current inner loop and capacitive currentinner loop based on quasi-PR control is presented,and the discretization method of quasi-PR control is given. For thequasi-PR controller of fixed parameters is unable to adapt to changes in power network parameters,and the systemcontrol performance cannot be optimized. A quasi-PR control method based on DRNN of self-tuning is presented and itscontrol algorithm is given. DRNN is based on the two-step RPE algorithm,which has a faster convergence rate than theBP algorithm using the first-order gradient algorithm. The simulation and experiment results proved that the gridconnected inverter with the above control strategy has a good quality of the output current and fast performance indynamic response,the resonance of grid current is effectively avoided and high in-grid power factor is achieved.

关 键 词：光伏并网逆变器 LCL滤波器 准PR控制 对角递归神经网络 无静差跟踪 

分 类 号：TM615[电气工程—电力系统及自动化]