针对光伏发电波动性与不确定性对电力系统稳定产生的影响,对多情景光伏发电功率的多步预测进行研究.首先通过密度峰值算法根据太阳辐射量、温度、湿度等气象数据对天气状况进行精确分类.其次,为了使模型表现出更好的性能,建立了鹈鹕算法优化随机森林(POA-RF)的因素筛选特征变量,模型用鹈鹕算法对随机森林的决策树数目和深度两个参数进行寻优,加强了因素筛选的有效性.最后,基于Informer模型对不同天气状况的光伏功率进行多步预测.实例计算结果验证了所提模型预测精准度的有效性与精准性.
In view of the fluctuation and uncertainty influence of photovoltaic power generation on the stability of power system,the paper makes a study on the multi-step forecasting of multi-scenario photovoltaic power.Firstly,the density peak algorithm is used to accurately classify weather conditions according to solar radiation,temperature,humidity and other meteorological data.Then in order to make the model show better performance,the pelican algorithm is used to optimize the factor screening characteristic variables of random forest.The model uses the pelican algorithm to optimize the number and depth parameters of decision tree for the random forest,enhancing the effectiveness of the factor screening.Finally,the Informer model is used to implement the multi-step forecasting of the photovoltaic power in different weather conditions.The result verifies the validity and accuracy of the proposed model.