Aplicação de método de deslocamento de carga para sistemas de gerenciamento de energia

Abstract

In the context of expected advances with the implementation of smart grids and considering the modernization of the forms of the electricity billing system in the relationship between user and manager of the electric system goes through a differentiated relationship. The user becomes an active agent and his behavior becomes of great importance in the behavior of the electric network. The cost factors for the user side and the overload on the distribution network are extremely important factors for these involved and the use of the load shifting technique is an alternative for the balance between the factors. Given this challenge, this dissertation presents methodology for load scheduling through an objective function that considers the cost and peak power to average power (PAPR) metrics, the model considers a convex linear combination between the two metrics. In the modeling is considered periods of restrictions for scheduling equipment that will meet the interests of the user and all loads are scheduled. As a solution to the scheduling and aiming to obtain a quick response to the scheduling, 3 methods were implemented, being a unique solution method in Greedy Search style, the method Particle Swarm Otimization - PSO and Evolutionary Particle Swarm Otimization - EPSO. To validate the implementation of the methods in the proposed modeling, simulations are performed considering scenarios with different loads and two different tariff models, for different coefficients of convex linear combination. This work presents a comparison between the results in the 3 implemented methods and concludes that EPSO was more efficient for this modeling.