A novel approach to design time of use tariff using load profiling and decomposition

The connected load of consumers is known to the distribution utility but the usage pattern of them is not known without smart meters installed on the site. Furthermore, constituents of the feeders at primary distribution level are also unknown. In partially deregulated developing countries implementation of Time of Use tariff becomes a challenging task. This paper addresses this crucial issue where using data mining techniques, the load profiling at primary distribution level is carried out. Load Profiles are clustered into three major sectors commercial, industrial and residential using K-means algorithm and Silhouette analysis, then an optimization problem decomposes the load of primary feeder in terms of commercial, industrial and residential sectors. Time of Use tariff is then designed using signature profiles obtained by clustering in order to flatten the load curve and composite elasticity of demand is obtained by decomposition of feeder. The result are obtained on actual data of NYSEG system.