Detecção de anomalias e estimação da qualidade dos dados do sistema de telemetria e monitoramento hidrometeorológico

Abstract

The objective of this job is to create a computational tool that provides information about the quality of a measurement performed remotely, based on analysis of historical statistics of the behavior of measurement over time. As a practical result, this methodology was applied in the development of an application that criticizes data received by STH (Telemetry and Tracking System for Weather and Hydrology) of Cemig, based on exploratory data analysis, evaluation of their qualifiers, rates of change and historical and physical limits in order to increase the quality and confidence in data received by the data acquisition platform. The implementation of the proposed tool proved very useful for decision-making processes as it provides a quality index for the measurements and offers the ability to indirectly inform about the quality of the other information of the system of supervision and control, for example, calculated data, which originate in the measurements criticized by the validation algorithm proposed here. All this gives the method presented here a systemic comprehensiveness.