基于神经网络的电力并网逆变器控制技术研究

摘要

Proportion resonant (PR) control technology of power grid-connected inverter of current model control was researched, and the model of power grid-connected inverter of LCL filter of double current circles was analyzed. Aiming at the difficulty of realization with digital controller for new-style PR controller, this paper put forward a incremental PR control technology based on BP neural network. On the MATLAB platform, a mathematical model simulation was built. Simulation experiment confirmed that the output current of gird-connected inverter controlled by incremental PR control technology based on BP neural network had better dynamic and static performance on condition of current mutation. Control results of grid-connected current trained by neural network turned out to be good.%研究了电流模式控制的电力并网逆变器的PR( proportion resonant)控制策略,针对提高并网供电功率优化逆变换器设计,分析了含LCL滤波器的电流双环控制电力并网逆变器模型.目前采用的比例谐振控制器难以用数字控制器实现的问题,提出了BP神经网络的增量式PR控制技术.在MATLAB平台上,建立数学模型仿真.仿真结果证明,在电流发生突变情况下,采用BP神经网络的电力并网逆变器的增量式PR控制,电流波形具有更好的动态性能与静态性能.对神经网络训练进行仿真,结果表明,并网供电控制取得良好的供电效果,为设计提供了参考依据.