基于遗传算法优化的BP神经网络方法在光伏发电功率预测中的应用OA

In order to solve the issue of the fluctuation and intermittence of output performance of photovoltaic power station under the influence of climatic conditions,which seriously affect the stable output of electric energy and the safe operation of power station,firstly,a photovoltaic power station in Inner Mongolia Autonomous Region is taken as the research object,and the operating data such as total solar irradiance,ambient temperature,ambient humidity,wind speed,wind direction and component inclination angle are collected.Secondly,the back propagation(BP)neural network is improved,and the genetic algorithm(GA)is introduced to construct the improved-back propagation(Imp-BP)and GA-Imp-BP neural network prediction models respectively.Finally,Imp-BP and GA-Imp-BP neural networks are trained and the output power prediction effects are compared.The results show that compared with the traditional BP neural network,Imp-BP and GA-Imp-BP neural networks have significant advantages.The mean absolute percentage error(MAPE)is calculated by using the actual power and the installed capacity as the denominator,respectively.The MAPE values of GA-Imp-BP neural network are 10.96%and 7.17%respectively,and the MAPE values of Imp-BP neural network are 14.21%and 7.52%respectively,which are lower than those of traditional BP neural network.The prediction accuracy of GA-Imp-BP neural network is higher in four seasons(especially in spring and autumn),and the prediction errors in summer and winter are significantly lower than that of traditional BP and Imp-BP neural network.