Daily typical load clustering of residential customers

Abstract

This paper presents the daily load pattern created by Fuzzy-Possibilistic C-Means (FPCM) clustering method and Fuzzy C-Mean (FCM) clustering method. Load data is used to study of customer's demand characteristic. Daily load profile is monitored from digital meters, which were installed randomly in limited locations due to budget constraint. The FCM technique assigns a degree of membership for each data set belonging to each center of all clusters. Fuclidean distance is utilized to calculate the distance between data set and each cluster center. The FPCM technique is better than the FCM technique for reducing outliers data's error. Therefore, fuzzy clustering technique was employed to determine daily load pattern of residential customer for customer's behavior analysis and customer load demand estimation. Understanding consumers' behavior is crucial for decision in power system operation and planning.