Análise de Parâmetros para Desagregação de Sinais de Consumo de Energia Elétrica

In this article, we address the problem of energy disaggregation that can be handled through dictionaries techni- ques where we synthesize the most representative characteristics of the signals involved to carry out the disaggregation. In our proposal, we can highlight two contributions. In the first one, we modified the steady-state identification algorithm to handle signals of different dimensions and then performed an analysis of parameters that influence the formation of the dictionaries and, consequently, produce different disaggregation performances. The experiments were carried out using the Reference Energy Disaggregation Data Set (REDD) database. The results show that the proposal produces good disaggregation performance.