基于网格空间规划功能片区的配电网负荷自动预测计算方法

Automatic Load Forecasting Calculation Method of Distribution Network Based on Functional Area of Grid Spatial Planning

常规配电网负荷自动预测计算方法的预测负荷利用率较低,为此提出基于网格空间规划功能片区的配电网负荷自动预测计算方法。收集网格化功能片区数据资料和负荷历史数据,利用皮尔逊相关系数,得到片区对负荷的月影响因子,输入月影响因子和负荷历史数据至神经网络,迭代训练后输出负荷弹性系数,利用系数平稳时间序列和生成序列,确定周期内负荷增长规律,结合当前负荷值,得到未来时刻负荷预测值。结果表明,相比常规算法,设计算法的负荷预测值更加接近实际值,提高了预测负荷利用率,避免了电力资源浪费。

The conventional automatic load prediction calculation method for distribution networks has a low utilization rate of predicted loads.To this end,an automatic load prediction calculation in the distribution network based on spatial planning functional area of the grid is proposed.Collect grid functional area data and load historical data,use Pearson correlation coefficient to obtain the monthly impact factor of the area on load,input the monthly impact factor and load historical data to the neural network,iterate and train,output the load elasticity coefficient,use coefficient stationary time series and generated series to determine the load growth pattern within the cycle,and combine the current load value to obtain the future load prediction value.Set up comparative experiments,and the results show that the designed algorithm has a load prediction value that is closer to the actual value compared to conventional algorithms,improving the predicted load utilization rate and avoiding power resource waste.