A Regional Monthly Electricity Consumption Forecast Method Based on HP Filter and Seq2Seq

With the continuous popularity of temperature sensitive electrical equipment such as urban cooling/heating, the variations in regional electricity demand are increasingly affected by meteorological parameters such as temperature and humidity. This puts forward higher requirements for monthly electricity forecasting technology. Therefore, a regional monthly electricity forecasting method based on HP filtering and Seq2Seq is proposed. Firstly, a monthly variation trend characterization model of regional meteorological parameters based on HP filtering was established based on years of historical observation data, providing reliable meteorological background information for subsequent power consumption modeling. Then, a regional monthly electricity consumption prediction neural network model based on Seq2Seq was established, which predicted the regional monthly electricity consumption by integrating static meteorological background information and dynamic historical electricity consumption information. An example analysis based on data from a certain location in Zhejiang shows that the proposed method has a much higher prediction accuracy.