Smart Meter Data Taxonomy for Demand Side Management in Smart Grids

Emerging Smart Grids have given rise to deployment of advanced metering infrastructure (AMI) which produces voluminous load data with respect to time. The information extracted from this data can potentially be used to design Demand Side Management (DSM) policies. The major impediment in the deployment of DSM is the lack of knowledge of individual consumers' load demand to determine target consumers for the incorporation in DSM. This paper proposes an innovative probabilistic approach to generate typical load profiles of consumers using smart meter data. The method incorporates clustering and then re-clustering the individual clusters using the k-means clustering algorithm until reliable load profiles showing true load-time characteristics of consumers are extracted. Next, the multi layered re-clustering methodology is applied to generate typical load profiles for consumers on an alternative scale by segregating them into different classes of load levels in such a way that they represent typical smart meter consumers of the population. The approach enables the utility to determine the consumers whom can be applied the DSM actions for an effective operation of a smart distribution network and to reduce the load peak and rebound effects.