基于神经网络的光伏阵列局部阴影建模研究

摘要

在局部阴影条件下，光伏阵列的输出特性将发生变化。利用神经网络具有逼近任意复杂非线性函数的能力，采用粒子群算法( PSO)来优化BP神经网络的内部连接权值和阈值，以改善神经网络的预测性能，并基于这种改进的神经网络结构搭建局部阴影下的光伏阵列模型。仿真结果表明，此算法泛化能力强、收敛速度快，能够对阴影下的光伏阵列进行建模。%The output characteristics of PV array changes under the partial y shaded conditions. The neural network with the ability of ap-proximating any complicated nonlinear function and the particle swarm optimization (PSO) are used to optimize the BP neural network's in-ternal connection weights and threshold value in order to improve the neural network prediction performance and to establish the model of PV system under the partial y shaded conditions. The simulation results show that, through the establishment of the PV array local shadow model identification precision and fast convergence rate,it perfectly verifies the effectiveness of the proposed method.