Modelagem dos dados de falha de uma bomba de refrigeração de alto-forno

Abstract

The maintenance of equipment is becoming increasingly important, being responsible for sustaining the quality and competitiveness of industries in the current globalized world. In this way, the reliability study is of paramount importance for industrial maintenance, as it assists in predicting failures and, consequently, increasing the availability and productivity of industrial equipment. It is necessary to have knowledge about the equipment to perform the reliability analysis and to determine the modes and causes of failures. The present work aims at modeling the times until the failure of a blast furnace cooling pump. The times were supplied by the company Ligas de Alumínio S/A, situated in the north of Minas Gerais and analyzed using ProConf software. The Weibull distribution was the best fit for the data set. The results indicate that the refrigeration pump is at the end of the infant mortality phase, with tendency to stabilize the failure rate. That is, it is at the beginning of its useful life, indicating the use of predictive maintenance.