CAJ | 학술논문

针对风电功率的预测，从神经网络的训练仿真入手，综合运用BP神经网络、平均绝对误差MAE等多种方法，建立风电功率预测模型，运用M atlab和Excel软件编程，得到了后续7天中时隔5min和15min的风电功率预测趋势和时隔5min样本数据的预测误差水平相对于15min来说降低了6.349％等结果。
침대풍전공솔적예측，종신경망락적훈련방진입수，종합운용BP신경망락、평균절대오차MAE등다충방법，건립풍전공솔예측모형，운용M atlab화Excel연건편정，득도료후속7천중시격5min화15min적풍전공솔예측추세화시격5min양본수거적예측오차수평상대우15min래설강저료6.349％등결과。
Aimed at the wind power forecasting ,the wind power forecasting model is established based on the training simulation of the neural net-work via means of the BP neural network algorithm ,MAE ,etc. Through such a model and using Matlab and Excell,we get the wind power fore-casting trend at the intervals of 5minutes and 15minutes of the following 7 days and the result that the forecasting error of the sample data at the interval of 5minutes is 6.349% lower than that of 15minutes .