Smart grid demand-side management by machine learning

Abstract

The smart grid is a complex electrical network system comprising of different subsystems at different levels of aggregation. It facilitates a bidirectional information flow among all the actors such as producers of electricity, end users of electrical energy, transmission and distribution system operators (TSO/DSO), and demand response (DR) aggregators. Smart grid contains smart meters that send user statistics to the server. Accurate forecasting of the electricity usage is required in order to take controlled actions to balance the supply and demand of electricity. This forecasting can be achieved using machine learning-based predictive models. This paper deals with the forecasting of short-term and mid-term load for the grid entity using machine learning. A predictive system is designed using machine learning techniques in order to process the smart meter data which in turn is used as the training data for the model. The outcomes are then shown with data and results to make it more understandable by the reader.