Intelligence Improvement of a "Prosumer" Node through the Predictive Concept

This paper describes a new predictive algorithm that can be used to improve the intelligence of a prosumer node. Prosumers - which means the entities that are consumers and producers at the same time - play an important, active role within the context of smart grids. On the other hand, a smart grid is truly "smart" if all its nodes are smart, including prosumer nodes. The algorithm is based on predictive functions that allow to perform optimized choices in advance, on the basis of information acquired from the field, from the examined building and from on-line data banks. The main functions performed by the algorithm are: the creation of an internal data bank, a learning procedure, and the decisions to be activated. These functions are also continuously upgraded. The main results supplied by the algorithm, at each established time interval (normally, a quarter of an hour), consist in the definition of the optimal amount of energy to be consumed, stored and locally generated. In this way a substantial increase in efficiency is reached with immediate, significant economic returns.