新型电力AMI系统中基于Neural Prophet模型的电力负荷预测与修补研究

针对新型电力系统高级量测体系（AMI）系统中电力负荷异常值与空值的修补问题，提出一种基于Neural Prophet模型的电力负荷预测方法，用预测值修补AMI系统中出现的负荷异常值与空值。通过Neural Prophet模型的趋势模块、季节性模块、自回归模块3个模块组合进行电力负荷精准预测，可解决AMI系统的负荷数据修补问题。在电力负荷数据集上进行实验，结果表明所提方法与负荷预测经典算法相比具有较优的准确性且具有参数可解释的优势。所提方法可对AMI系统中有问题的电力负荷异常数据进行有效修改，对缺失的电力负荷数据进行填补。

Neural Prophet-based Power Load Forecasting and Repair in Novel Power AMI System

Aiming at the problem of repairing abnormal value and null value of power load in Advanced Measuring system （AMI） of new power system， Neural Prophet-based electrical load forecasting method is proposed to use the predicted values to repair abnormal and null load values in AMI systems. The combination of Neural Prophet’s trend module， seasonality module and autoregressive module for accurate power load forecasting can solve the load data repair problem in AMI systems. Experiments are conducted on the electricity load dataset， and the experimental results show that the proposed method has better accuracy and more stable prediction level compared with the classical algorithm， and has the advantage of interpretable parameters. The proposed method can effectively modify the abnormal power load data in AMI system or fill in the missing load value.