Efficient Management of Demand in a Power Distribution System with Smart Meter Data

This paper presents a novel approach for demand side management (DSM) in a power distribution system by incorporating smart meter data. The approach is aimed at savings maximization by minimizing the energy consumption cost of electricity consumers. The core of the approach consists of data clustering in order to forecast demand for the benefit of DSM decisions by incorporating alternate profiles through extended kmeans algorithm, Taylor series linearization and particle swarm optimization. Two cases including integration of PV generation are simulated using the Irish data of more than 5000 smart meters. Different demand flexibility levels are considered in different Scenarios. The paper argues that inherent non-linearity of raw profiles, is likely to provide suboptimal DSM solutions against electricity consumer cost savings, however the uniformity and smoothness of reshaped alternate profiles are more likely to provide optimal DSM solutions, providing electricity consumers a true benefit for their participation in the DSM process.