Neural-network-based approach applied to harmonic component estimation in microgrids

Power quality in smart microgrids must be carefully analyzed, whereas adverse consequences may harm the electrical systems without power management and appropriate measures. The main goal of this paper is to develop a 5
^th
, 7
^th
, 11
^th
, and 13
^th
voltage harmonic components identification method based on artificial neural network (ANN). This tool could provide information to the smart microgrid management and control system or be an alternative solution to the harmonic identification process of a harmonic compensator embededs into power converters. The trained algorithm can identify harmonic components amplitude and phase angle in the interfacing point between microgrid and power converters. it was possible to generate a voltage waveform with a maximum difference of 0.04 p.u. between the expected waveform and the one built with the parameters identified by ANN. The ANN method validation was performed through computer simulations.