基于ANN的光伏并网逆变器SVPWM研究(英文)

Research of SVPWM Based on ANN in Photovoltaic Grid-Connected Inverter

通过分析空间矢量脉宽调制的工作原理,提出基于人工神经网络的SWPWM调制策略.这种基于ANN的光伏并网逆变器覆盖了SVPWM的欠调制和过调制两种模式,能实现从线性调制到六步模式非线性调制的平滑过渡.由于神经网络具有很强的并行处理能力以及容错能力,ANN SVPWM控制器运行速度快,能有效提高逆变器功率开关的开关频率.采用监督模式,通过固定权值分别对欠调制子网络和过调制子网络进行训练.在MATLAB仿真平台上搭建光伏并网逆变器仿真模型,由神经网络工具箱实现ANN控制器,仿真结果表明:在SVPWM过调制区用ANN控制方法简单、高效、控制效果良好.

The basic principle of space voltage vector Pulse Wide Modulation （SVPWM） is analyzed, and a novel algorithm for SVPWM based on Artificial neural network （ANN） is proposed. This ANN SVPWM controller completely covers the under-modulation and over-modulation linearly extending operation smoothly up to the six-step mode. The neural network controller has the advantage of very fast implementation by using SVPWM algorithm, and it can increase the switching frequency of power switches in the static converter. In this design, the individual training strategy with the fixed weight and supervised models have used. The simulation is developed by using MATI.AB, together with the Neural Network Toolbox for training the ANN-controller. The simulation results show that the neural network control strategy during over-modulation period features simplicity, high efficiency and excellent control effect.