Sistema de detecção de anomalias associado a armazenadores para redução de picos de consumo de energia elétrica.

Abstract

dentification of anomalous behaviors in the electric power supply can indicate defects, unplanned consumption and opportunities for use of alternative sources and energy storage. The analysis of consumption is the first step to identify actions that result in efficiency, quality of energy and reduction of the consumption of electric energy provided by the concessionaire. This work presents an application of artificial immunological systems, focusing on the negative selection algorithm, to detect anomalies in the consumption of electric energy. In this algorithm we verified the anomalous consumption of electric energy based on the standard load curves. The data of energy consumption were provided by a smart meter installed in a building at the Federal University of Campina Grande (UFCG). A model of storage system of energy by lead-acid batteries with different capacities is simulated and thus provided the state of charge behavior of the batteries for currents demanded by detected power consumption peaks. In an interface, the anomaly detection algorithm is associated with the commercial battery model of different capacities. The results obtained in the simulations demonstrate the ability of the algorithm to detect several types of consumption anomalies such as peaks, valleys, short circuits, and high efficiency. Its association to a battery-powered energy storage system can mitigate power consumption peaks by evaluating the response of each battery system as to the ability to reduce spikes, cost of investment and lifetime for the proposed application.