Consumption based-only load forecasting for individual households in nanogrids: a case study

Abstract

Electricity load forecasting plays an important role in planning and a vital role in the operational management of an electric power system based on smart grids. In this work, several data-driven approaches are used to forecast the individual electricity demand for the subsequent hour of three Italian households in a nanogrid context. For each user, the tested prediction models exploit only the historic series of total electricity consumption. The results show similar performances in all models implemented. Despite a widespread delay, the predictions follow the measurement trend well, while also highlighting the particular difficulty of predicting peak values.