改进的光伏发电系统MPPT预测算法的研究

The Reasch of an Improved GA-BP Neural Network in MPPT Predict Algorithm of Photovoltaic Generation System

在光伏发电效率预测的研究中,针对光伏供电系统受温度和光照变化影响大、太阳能利用效率低和最大功率点预测不准确等问题,提出一种改进的GA-BP神经网络的光伏系统MPPT预测算法,通过优化的BP神经网络训练光伏阵列实测数据,预测输出的最大功率。为提高算法预测精度,采用云模型云滴和遗传算法优化BP神经网络的初始权值和阀值,根据遗传算法收敛程度来调整云自适应交叉和变异算子。经Matlab仿真分析,在光照和温度变化时,改进的GA-BP神经网络比GA-BP神经网络和BP神经网络具有更好的预测效果。

For the problem of the changes of temperature and illumination having great influence on PV array output power, the low efficiency of solar energy utilization and difficult prediction of maximum power point, we propose an improved GA-BP neural network in MPPT algorithm for photovoltaic system to predict maximum power output by optimizing BP neural network to training the photovoltaic array measured data. In order to improve the prediction precision of the algorithm, the initial weights and threshold of the BP neural network are optimized by using the cloud model cloud droplets and genetic algorithm, and the cloud adaptive crossover and mutation operators are adjusted according to the convergence degree of genetic algorithm. Through Matlab simulation analysis, the improved GA-BP neural network has better tracking performance than GA-BP and BP neural networks during the change of the light and tempreture.