基于相似日的ARMA-LM的光伏发电功率预测

本文采用综合预测的方法,对数据进行相似日处理,大大降低了不同天气类型对光伏功率预测的影响,通过ARMA时间序列结合LM神经网络,弥补了LM神经网络在线性部分(趋势、季节变动、循环波动)的不足,大大提高了预测的精确度和稳定性。分别对晴天、阴天和雨天三种天气类型下的光伏功率进行预测,并将其与灰色预测、LM神经网络模型进行对比。结果表明,结合相似日的时间序列神经网络光伏发电功率预测模型,在光伏发电功率预测领域具有更高的精度与稳定性。

Prediction of Photovoltaic Power Generation Based on Similar Day ARMA-LM

In this paper,a comprehensive forecasting method is used to perform similar day processing on the data,which greatly reduces the impact of different weather types on photovoltaic power prediction.The ARMA time series combined with the LM neural network compensates for the linear part(trend,seasonal variation,Cyclic fluctuations),greatly improving the accuracy and stability of prediction.The photovoltaic power in three types of weather,sunny,cloudy and rainy,is forecasted,and compared with the gray forecast and LM neural network model.The results show that combined with the time series neural network photovoltaic power prediction model of similar days,it has higher accuracy and stability in the field of photovoltaic power prediction.