An Outliers Processing Module Based on Artificial Intelligence for Substations Metering System

Abstract

One of the main problems of the data acquired by the power utilities is the presence of outliers affecting the measurements database throughout the electrical system damaging the distribution scenario analysis. This work proposes a new module to complement the measurements made by the metering systems. A detection technique and three outliers correction techniques were developed, based on fuzzy logic, artificial neural networks and the ARIMA model. The first technique, with a fuzzy approach, develops an inference system based on the variations of the previous 3 measurements to determine the future variation. In the second algorithm developed using ANN, the outliers were corrected using a prediction model with 10 previous samples. The last correction technique was based on the ARIMA model with 96 previous measurements. In order to demonstrate the applicability of the developed methods, a case study was carried out on a substation in a city of Paraíba, a Brazilian state. The three techniques of correction of the outliers presented mean relative error less than 5% for all the test scenarios.