基于自适应差分进化和BP神经网络的光伏功率预测

针对光伏功率的波动性和间歇性,通过分析光伏发电的影响因素,建立了基于自适应差分进化和BP神经网络的光伏功率预测模型.该模型利用自适应差分进化算法优化BP神经网络的权重阈值,克服了BP算法收敛速度慢、容易陷入局部极值的缺点.利用光伏电站的历史数据和气象观测站的气象数据,对预测模型进行训练和光伏功率预测.结果表明,基于自适应差分进化和BP神经网络的模型预测精度高于BP神经网络模型,验证了所提模型和算法的有效性和可行性.

Photovoltaic Power Prediction Based on Adaptive Differential Evolution and BP Neural Network

In the light of the volatility and intermittent of photovoltaic power,a prediction model based on adaptive differential evolution and BP neural network is proposed by analyzing its factors.The model optimizes the weights and thresholds of BP neural network using adaptive differential evolution algorithm,which can overcome the defect of traditional algorithm,such as fall into local minimum easily,slow convergence and so on.According to the historical data of PV power plant and meteorological data form meteorological stations,the prediction model is trained and PV power is predicted.The forecast results show that the prediction accuracy of proposed model is higher than that of model based on BP neural network,which verifies the effectiveness of the proposed model and algorithm.