Estudo de engenharia de confiabilidade e aplicação de ferramenta de inteligência artificial na gestão de ativos de uma unidade industrial na Aamazônia

Abstract

For a long time, Industrial Maintenance was considered a cost, naturally unwanted, by asset managers; however, today it is a strategic part and represents a high investment in the world's largest industries. This Maintenance is part of a relatively recent concept, but growing in the industrial environment, called Asset Management and some of its pillars are Reliability and prediction techniques of equipment's operational conditions. In this study, failure data were obtained during 5 (five) years of operation of 6 (six) hydrocarbon centrifugal pumps and from this history the most recent Reliability, Availability and Predictive Oil and Vibration Analyzes were studied. Evaluating several compound parameters in the universe of maintenance and reliability as well as determining which are the priorities of the responsible team, taking into account the operational perception of each component of this team, becomes a complex mission under the human point of view. This complexity is why it was important to use Artificial Intelligence, capable of relating purely subjective qualitative assessments to objective technical data, and demonstrate something feasible, tangible and directly applicable. With the correlation between the parameters, performed by the Fuzzy tool, it was possible to obtain a number that indicates the level of operational condition of the equipment from 0 (zero) to 10 (ten), and from this definition classify them in order of most able to perform the operation and x the most priority in terms of maintenance. With several equipments under his responsibility, obtaining the perception of the quality of Reliability and Availability, joining the data of Maintenance by the Condition of an equipment and relating all this with support from Artificial Intelligence, a manager can obtain which are the equipments that must have more attention and investment at that exact moment. In this study it became clear which equipment should be prioritized in terms of maintenance and which is in better operational condition at a given moment. This can be a model of resource optimization desired by any company that operates in a sustainable and responsible way.