Reliability models of electric machines with structural defects

Particular features of solving the problem of forecasting the reliability of electric machines with long mean time between failures by means of taking into account the state of their basic structural units are considered. The prospect of using mathematical variable-structure reliability models based on joint use of fuzzy logic and neural network provisions is grounded for solution of this problem. It allows taking into consideration the influence of informative parameters characterizing the state of structural units and the environmental conditions.