"Detección de fallas de sensado en convertidores CA/CC para micro-redes híbridas usando análisis de componentes principales (PCA)"

Abstract

The present academic research work proposes a method based on Principal Component Analysis (PCA) for the detection of failures in sensors related to local control of AC / DC voltage converters in a micro-grid (MR) with the possibility of coupling to the net. In order to achieve the aforementioned, a simulation of a micro-grid is carried out with a variety of loads (linear and nonlinear) and both conventional and renewable generators, as well as storage elements that together have DC and AC systems with their respective converters. The researchers propose two failure scenarios and a normal operating scenario that serves as a reference to carry out the aforementioned analyzes. As a result of the aforementioned, an algorithm has been implemented that, based on the main components of the mentioned cases, calculates differences between the spaces obtained from a total of 195 variables collected in all the bars of the system, in addition to the sensing for local control. The observations obtained are around 145000 value takes.