基于数据挖掘的光伏发电预测

Forecast of PV power generation based on data mining

外界环境和光照强度的不确定性决定了光伏发电出力的非平稳性和随机性,有效的光伏发电量预测不仅能保护接入电网的稳健运行,还有助于电网的调度安排和光伏电站的运维决策。文章提出以周天气特征因子和气象因子为特征,结合历史发电数据,建立起一个最小二乘支持向量机预测模型,进行光伏发电量的超短期预测。通过实验对训练好的模型进行预测精度的评估,结果表明,与未结合周天气特征的预测模型相比,该模型的平均预测精度提高了30%左右。

The uncertainty of the external environment and sunlight intensity determines the non-stationary and randomness of the photovoltaic power generation,and the effective PV generation prediction can not only protect the stable operation of the access grid,but also help the dispatching of the power grid,and the operation and maintenance of the photovoltaic power station.In this paper,a characteristic factor of the week weather is proposed,and the prediction model of least square support vector machine is established by combining the meteorological factors and historical power generation data.The ultra-short term prediction of photovoltaic power generation is carried out,and the prediction accuracy of the trained model is evaluated through experiment.The results show that the average prediction accuracy of the model is 30 percentages higher than those of without characteristic factor of the week weather.