基于改进萤火虫算法的5G基站光伏功率预测
Photovoltaic Power Prediction of 5G Base Station Based on Improved Firefly Algorithm

为保证5G基站光伏供电系统的稳定性，需要精确地预测光伏发电输出功率，因此，本文提出了一种基于改进萤火虫算法(Firefly Algorithm, FA)来优化反向传播(Back Propagation, BP)神经网络的光伏功率预测模型(IFA-BP)，首先采用灰色关联分析法分析多种气象因素对光伏功率预测的影响程度，然后利用Circle混沌映射使萤火虫种群分布更加均匀，并对寻得的萤火虫最优解加入非线性突变扰动来避免陷入局部最优解，最后建立IFA-BP光伏功率预测模型。通过MATLAB进行仿真，将本文提出的IFA-BP预测模型与FA-BP预测模型、BP预测模型的光伏功率预测结果对比分析，结果表明，提出的IFA-BP预测模型在不同天气下均具有最佳的预测精度。
In order to ensure the stability of 5G base station photovoltaic power generation system, it is necessary to accurately predict the photovoltaic power generation output. Therefore, this paper puts forward an algorithm based on improved firefly to optimize the photovoltaic power of the back propagation neural network prediction model (IFA-BP), first using the method of grey correlation analysis of many kinds of meteorological factors to predict the impact of photovoltaic power. Then, Circle chaos mapping was used to make the firefly population distribution more uniform, and nonlinear mutation disturbance was added to the optimal firefly solution to avoid falling into the local optimal solution. Finally, the IFA-BP photovoltaic power prediction model was established. Simulation experiments were carried out by MATLAB, and the results of photovoltaic power prediction were compared with those of FA-BP prediction model and BP prediction model. The results show that the proposed FA-BP prediction model has the best prediction accuracy under different weather conditions.