A study for clustering method to generate Typical Load Profiles for Smart Grid

Abstract

Interests in green growth for environmental protection are recently increased and encourage research on Smart Grid to use power efficiently. Among interesting issues in this research, the methodology of data mining is an emerging issue which stands for utilizing power usage data collected every 15 minutes from customers for the computation of electricity rates. Load analysis method based on VLP (Virtual Load Profile) is used to create virtual 15 minutes power usage data for non-AMR (Automatic Meter Reading) customers with 15 minutes power usage data from AMR customers. In this paper, TLP (Typical Load Profile) generation method, hierarchical clustering, k-means clustering, fuzzy c-means clustering, and two-stage fuzzy clustering are investigated and their performance are also analyzed.