Using ANFIS to Predict Harmonic Distortion in Residential Building Loads: A case study in the Amazonian Region of Brazil

Abstract

With the increasing use of nonlinear loads in homes in Brazil comes the problem of harmonic injection in the power system and increasingly is a problem for the electric sector that needs to scale it. Knowing the loads that consume energy and inject harmonics into the system is important so that solutions are sought to make the use of the system more efficient and improve the quality of the energy that circulates in the electrical grid. This work presents simulations of DHTv and DHTi of a set of residences in order to predict the behaviour of the load over time, using previous measurements. The modelling is conducted using an ANFIS, which uses a neural network to adjust the parameters of the output that uses fuzzy rule to determine the output values of the system.