Study of missing meter data impact on domestic load profiles clustering and characterization

摘要

Automated load managements and cost-effective power systems in distribution level are now becoming possible by increasing the number of installed smart meters at end-users side. Monitoring and controlling the massive datasets of demand curves require using data mining and characterizing load profiles. This paper analyses a wide range of data from a comprehensive survey of residential customers. It investigates the effect of missing load measurement in the result of clustering consumers load profiles. Different scenarios have been considered in order to detect and deal with missing or incorrect recordings. Moreover, the result of clustering for different percentage of missed values has been evaluated considering various numbers of clusters.