Simulação computacional utilizando o método de Monte Carlo para determinar a disponibilidade inerente de uma bomba centrífuga

Abstract

Some indicators of maintenance such as average availability, punctual availability, expected failures, reliability for the period, and average time between failures are critical information for maintenance managers to develop efficient strategic plans. Most of them can be determined analytically. However, to predict the inherent future mean availability and the number of failures it is necessary to develop some kind of simulation. In order to find these indicators, the present work proposes an algorithm developed in MATLAB to simulate the fault and repair behavior of a multi-stage centrifugal pump, using the Monte Carlo methodology. To ensure accurate results, an examination of the history of repair and failure times is performed in order to find values that may interfere with the simulation responses.