Identificação e classificação de cargas através do conteúdo harmônico via redes neurais artificiais fazendo uso de um sistema de aquisição de dados baseado em arduino

Abstract

This paper presents the development of a data acquisition system based on Arduino for signal sampling. The system consists of a current transformer (CT), a voltage transformer (VT), signal conditioning circuits, and the Arduino as the analog-to-digital converter and means of communication with the computer. Data were obtained from three different pairs of lamps: incandescent, compact fluorescent and LEDs. A script created in Matlab was in charge of storing the data. The current data were input in a fast Fourier transform (FFT) algorithm in order to extract the harmonics content. It was observed that the first five odd harmonics were quite distinct among the loads. From this set of data, the use of harmonics as a way for load classification was investigated by employing the Neural Network Pattern Recognition tool on Matlab, a classification neural network, where it was found that indeed the harmonics are a source of reliable data that can be used to classify loads.