In this paper, it is shown that the use of probability‐statistic methods, especially at the early stage of diagnosing the technical condition of aviation gas turbine engines (GTE) when the flight information has fuzzy and limitation and uncertainty properties, is unfounded. Hence the efficiency of the use of Soft Computing methods‐fuzzy logic and neural networks at these diagnostic stages is considered. Training with high accuracy of fuzzy multiple linear and non‐linear models (fuzzy regression equations) which received on the statistical fuzzy data basis is made. Thus, for to make a more adequate model of the technical condition of GTE, the dynamics changes of skewness and kurtosis coefficients are analysed. Research of skewness and kurtasis coefficients shows, that the statistical distributions of the work parameters of GTE have a fuzzy character. Hence, consideration of fuzzy skewness and kurtosis coefficients is expedient.
Investigation of the basic characteristics of the changes in the dynamics of the work parameters of GTE allows to draw the conclusion that it is necessary to use fuzzy statistical analysis during the preliminary identification of the technical condition of engines.
Research of changes in the values of correlation coefficients also demonstrates their fuzzy character. Therefore for models choice the application of the Fuzzy Correlation Analysis results is offered. The fuzzy multiple correlation coefficient of fuzzy multiple regression is considered for checking the adequacy of models.
At the information sufficiency is offered to use recurrent algorithm of aviation GTE technical condition identification (hard computing technology is used) on measurements of input and output parameters of the multiple linear and nonlinear generalised models at presence of noise measured (the new recursive Least Squares Method (LSM)).
The system that is developed to monitor the condition of GTE provides stage‐by‐stage estimation of the technical condition of an engine. As an application of this technique, an estimation of the new operating aviation engine temperature condition was made.
Santrauka
Straipsnyje atskleidžiamas tikimybinio-statistinio metodo nepagrįstumas diagnozuojant dujų turbininius variklius, kai informacija yra netiksli, ribota ir neapibrėžta. Parodytas technologijos Soft Computing taikymo efektyvumas. Taikant netikslios statistikos, netikslios logikos ir neuroninių tinklų tikslius metodus dujų turbininių variklių diagnozavimui atliekamas daugiamačių tiesinių ir netiesinių modelių (regresijos lygčių), gautų iš netikslių statistinių duomenų, apmokymas. Taikant aprašytą metodą buvo atlikta pradėto eksploatuoti turbininio variklio šiluminės būsenos analizė.
Symbolic Dynamic Analysis of Transient Time Series for Fault Detection in Gas Turbine Engines
