This article aims to identify the best
distribution functions of random variables: Time Between Failures (TBF) and Time To Repair (TTR) for hydro generator groups equipped with Francis turbines. Distribution functions tested in this study are those commonly used in the energy field: the exponential
distribution,
Weibull distribution and normal
distribution. The
distributions identification is done using a
hybrid algorithm (abbreviated BB-BC-CG) which results from combining Big Bang Big Crunch (BB-BC) algorithm, with Conjugate Gradient algorithm (CG) existing in the optimization tools of MathCAD software. After the
distributions identification a statistic testing is performed using the Kolmogorov-Smirnov test. In the final, the paper includes a case study on parameter estimation of deemed
distributions based on data collected over an operating period of ten years for hydro generator groups from a hydro power plant.
The identified distributions can be used to analyze the predictive reliability.