一种利用改进Elman神经网络的光伏I-V特性建模方法

为了准确表征和预测光伏（PV）组件在不同工况下的电流电压（I-V）特性，本文提出一种利用改进Elman神经网络的光伏I-V曲线黑盒建模新方法。首先通过皮尔森相关系数分析影响I-V曲线的环境因素；其次使用基于电压电流的双线性插值法对实测I-V曲线进行重采样，以提高I-V曲线上数据点分布的均匀性；进而使用基于辐照度温度的网格采样法对I-V曲线数据集进行下采样，降低数据冗余度；再利用量子粒子群（QPSO）算法优化Elman神经网络的初始权值和阈值，从而构造QPSO-Elman预测模型。最后根据美国国家可再生能源实验室（NREL）提供的I-V曲线数据集进行实验验证和测试。实验结果表明，所提出的建模预测方法精度更高，稳定性和泛化能力更好。

A modeling method of photovoltaic I-V characteristics based on improved Elman neural network

In order to accurately characterize and predict the current and voltage (I-V) characteristics of photovoltaic (PV) modules under different operating conditions, this paper proposes a new method of photovoltaic I-V curve black box modeling based on improved Elman neural network. Firstly, the environmental factors affecting the I-V curve are analyzed by Pearson correlation coefficient; secondly, the bilinear interpolation method is used to resample the measured I-V curve to improve the uniformity of data points distribution on the I-V curve; then, the grid based sampling method is used to sample the I-V curve data set to reduce the data redundancy. Then, quantum particle swarm optimization (QPSO) is used to optimize the initial weights and thresholds of Elman neural network to construct the QPSO Elman prediction model. Finally, the experimental verification and testing are carried out according to the I-V curve data set provided by the National Renewable Energy Laboratory (NREL). The experimental results show that the proposed prediction model has higher accuracy, better stability and generalization ability.