| Keywords |
Demand response; Clustering; Load shape analysis; Building energy analysis; Smart meter data |
| Abstract |
As Demand Response (DR) programs have been widely accepted to reduce energy consumption in residential buildings, targeting appropriate customers for demand response programs is crucial to minimize losses caused by the enrollment of inappropriate customers. Previous studies have mainly used the clustering method to group and analyze usage patterns of load profiles for demand response targeting. Although the national DR program reduction request is issued according to the concentration of fine dust, selecting households with high electricity consumption for the DR program can show effective results in reducing the power load. In this paper, we proposed a two-stage clustering method to segment households’ load profiles using a machine learning-based clustering algorithm. Then, the load parameter method was used to analyze the load shape for national demand response targeting. Our findings indicate that households with high morning peak and evening peak are suitable to be targets for the national DR program. |