System of Complex Identification of Technical State of Aviation Gas-turbine Engines

It is shown the groundlessness of application of stochastic methods, especially at the early stage of diagnosing the technical state of aviation gas-turbine engines, when the body of information is fuzzy, limited and uncertain. The application of a new technology "Soft Computing" at these stages of diagnosing is shown to be efficient because of using methods of fuzzy logic and neuro networks which allow one to conduct training of the multi-dimensional linear and non-linear models (regression equations) with a high precision on the basis of fuzzy stochastic data. To construct more adequate model of the gas-turbine engine state, the dynamics of change of asymmetry and excess factors is analyzed. Results of the analysis show that laws of distribution of operational parameters of the gas-turbine engines are of fuzzy nature. The dynamics of change of correlation factors considered and its fuzzy nature is indicated. When the body of information is adequate it is proposed to use the recurrent algorithm of identification of technical state of aviation gas-turbine engines (technology "Hard Computing" is applied). The algorithm uses measurements of input and output parameters of multi-dimensional linear and non-linear generalized models taking into account the noise of measurements. As an example, the technical state of the engine D30KU-154, newly introduced into operation, is estimated.