Sistema de monitoramento da instrumentação de segurança: um estudo de caso considerando a Usina de Itaipu

Abstract

This work presents an alternative way of monitoring dams using concepts of Multivariate Analysis, Data Mining and Statistics, so it is possible, in an easy way, to identify possible anomalies that may be occurring or eminently occurring. Such anomalies may be, for example, the Slip or Tapping of one of the blocks of the dam structure. The monitoring procedure consists of knowing the maximum and minimum values of readings of each instrument within a range called Good Behavior (GB), and using a performance function that can determine if a reading indicates a Good Behavior or a Change of Behavior (CB). Since the readings of the different instruments are performed with different periodicities, before processing these values, it is necessary to standardize the data and verify their normalities, only so it is possible to perform Factorial Analysis (FA) to create factors that group the Instruments based on their characteristics, and relates them to their respective anomalies, so it is possible to analyze the factor and not instrument by instrument. To analise and results was developed na performance function used in an algorithm writen with MathLab that performs tests with simulated data and also with real data of readings of the Itaipu dam instruments to monitor changem in Dam Structure Behavior.