摘要
在不平衡电网电压下,分析了风电并网变流器直流侧电压存在二倍频波动的原因,针对该母线电压波动及其引发的输出功率波动、并网电流谐波等问题,建立了基于正向dq+坐标系下并网变流器的正负序动态模型,并提出了一种无锁相环综合控制策略。系统外环引入了谐振积分器和谐振滤波器对直流系统二倍频波动进行抑制,降低了电网不平衡运行对并网电流的影响,内环基于准比例谐振控制器(Quasi-PR)对正负序电流进行综合控制,无需分解正负序分量。为了提高系统抗干扰能力,采用径向基函数(RBF)神经网络在线调整PI控制器参数。Matlab仿真和对比分析表明:该综合控制策略不仅能抑制直流侧电压及输出功率的二倍频波动,并且降低了并网电流的低次谐波含量,验证了所提方案的可行性及有效性。
Under unbalanced grid voltage,the reasons for the double frequency fluctuations in DC side voltage of the wind power grid-connected converter under the unbalanced grid voltage faults are analyzed.In view of the DC voltage fluctuations,the output power fluctuations and grid-connected current harmonics,a positive-negative sequence dynamic model of the grid-connected converter based on positive synchronously rotating reference dq+frame is established,and an integrated control strategy without phase-locked loop is proposed.The resonant integrator and resonant filter are used to suppress the double frequency fluctuations of the DC system in outer loop.The inner loop is based on quasi proportional resonant(Quasi-PR)controller to control the positive and negative sequence current without decomposing the positive and negative sequence components.In order to improve the anti-jamming ability,a radial basis function(RBF)neural network is used to adjust the PI controller parameters online.Simulation and comparative analysis show that the integrated control strategy can not only eliminate the double frequency fluctuations in DC side voltage,restrain the double frequency fluctuations in output power,but also reduce the low-order harmonics in grid-connected current,which verifies the feasibility and effectiveness of the proposed scheme.
作者
谢宙桦
赵晨聪
黄万启
XIE Zhou-hua;ZHAO Chen-cong;HUANG Wan-qi(Xi'an Thermal Power Research Institute CO.LTD,Xi'an 710054,China;School of Automation and Information Engineering,Xi’an University of Technology,Xi’an 710048,China)
出处
《控制工程》
CSCD
北大核心
2020年第9期1573-1581,共9页
Control Engineering of China
关键词
不平衡电网
并网变流器
准比例谐振控制器
谐振滤波器
RBF神经网络
Unbalanced grid voltage
grid-connected converter
quasi-proportional resonant(Quasi-PR)
resonant filter
RBF neural network