Confiabilidade humana: metodologia SHERPA aplicada no setor de manutenção de uma usina hidrelétrica

Abstract

Brazil is extremely dependent on the electrical energy produced in highly automated hydroelectric plants. However, certain activities are performed in the field by employees, especially in the maintenance area. As a result, we started to pay attention to human reliability, in order to know the probability of errors occurring in any task over a period of time under certain conditions, aiming to act predictively. Among several methodologies that allow the calculation of human reliability indices, we chose the SHERPA methodology applied in a spreadsheet, where data is considered and allows the modeling and location of critical points, and the prediction and mitigation of errors. The work began with a bibliographic review and then to the construction of the tool based on the analysis of tasks as to the risk inherent to each one of them. After the initial tests, there was an improvement that allowed the application in loco in the maintenance area of a Hydroelectric Power Plant. The feedback obtained referenced the tool, proving its connection with reality and the easy application in the field. This is the first generation developed of the application, which requires the practicality, ease and reliability of the application.