基于分布式光伏发电量预测分析的运行优化策略研究

Research on operation optimization strategy based on predictive analysis of distributed photovoltaic generation

针对目前分布式光伏发电系统发电量的影响因素较多,不易预测,与其他发电系统之间运行优化策略不完善等问题。文章参考国内外光伏行业大数据应用的典型经验,基于光伏发电数据和用户的负荷需求数据,提出了一种基于RBF神经网络的光伏发电量预测和负荷预测模型,通过对数据的归一化处理和对天气因素的量化和相似度处理,对未来一段时间内的光伏用电量和负荷进行预测;采用青岛市某光伏电站的实际数据进行学习和预测,取得较好效果,从而验证了模型的可行性。此外通过对负荷的预测和对发电量的预测数据,以经济性能最优为目标制定了运行优化策略,实现了光伏发电的有效利用,使发电侧和负荷侧功率平衡,大大降低了网损和线损,提升了分布式光伏用电可靠性和经济性。

In view of the current factors affecting the generation capacity of distributed photovoltaic power generation systems,it is difficult to predict,and the operation optimization strategy between the other power generation systems is not perfect.According to the typical experience of big data applications in photovoltaic industry at home and abroad,based on photovoltaic power generation data and user load demand data,this paper proposes an RBF neural network based photovoltaic power generation forecasting and load forecasting model.The prediction of PV power consumption and load in a certain period of time is conducted through the normalization processing of data,quantification processing and similarity processing of weather factors.The actual data of a PV power plant in Qingdao and is adopted to learn and predict,and achieve better results,thus verifying the feasibility of the model.In addition,through the prediction of load and the prediction of power generation,the operation optimization strategy is formulated with the goal of economic performance optimization,which realizes the effective utilization of photovoltaic power generation,balances power between power generation side and load side,and greatly reduces network loss and line loss.The reliability and economy of distributed photovoltaic power consumption is improved.