The subject of the study in the article is the modes of operation of the aircraft engine TV3-117 and methods of their control and diagnostics. The purpose of the work is to develop methods of control and diagnostics of the technical condition of the aircraft engine TV3-117 on the basis of neural network technologies in real time. The following tasks are solved: substantiation of the preconditions of the use of neural networks in the task of control and diagnostics of the technical condition of the aircraft engine TV3-117, construction of the generalized neural network and the choice of the algorithm for its training, the solution of the task of controlling the parameters of the technical condition of the aircraft engine TV3-117 with the use of neural networks. The following methods are used: methods of probability theory and mathematical statistics, methods of neuroinformatics, methods of the theory of information systems and data processing. The following results were obtained: The feasibility of using neural networks in the task of controlling and diagnosing the technical condition of the aircraft engine TV3-117 was substantiated. The expediency of developing neural networks based on the NN Predictive Controller. The expediency of using the gradient method of teaching neural networks is substantiated, as well as the method of training a neuro-regulator based on a neuro-modulator with the use of the method of reverse error propagation. The expediency of using the gradient method of teaching neural networks is substantiated, as well as the method of training a neuro-regulator based on a neuromodulator with the use of the method of reverse error propagation. The solution of the task of controlling the parameters of the technical condition of the aircraft engine ТВ3-117, which confirms the expediency of using neural networks in the task of control and diagnostics of the technical condition of the aircraft engine TV3-117, is obtained. Conclusions: The application of neural network technologies is effective in solving a wide range of poorly formalized tasks, one of which is the task of controlling the technical condition of the aircraft engine TV3-117. The advantage of neural networks in their application in the tasks of control and diagnostics of the technical condition of the aircraft engine TV3-117 is the possibility of working with small training samples, the appointment of soft tolerances, using the experience of experts to assess the technical condition of the aircraft engine TV3-117, which is important in the condition's information incompleteness.

Purpose. Construction of a mathematical model of the aircraft engine TV3-117 based on the results of observations of its reaction to environmental disturbances. The solution of the problem of identifying a dynamic model of TV3-117 engine in onboard conditions by classical methods, including the method of least squares and approximation by cubic splines, and neural network, by building a neural network according to source data. Methodology. The work is based on the methods of probability theory and mathematical statistics, neuroinformatics, information systems theory and data processing. In this paper, the Elman recurrent network with one hidden layer with a sigmoid neuron activation function with feedback was applied. Results. A method was developed for determining the optimal structure of a neural network, which consists in determining the neural network architecture, choosing the optimal algorithm for finding weights of neurons and teaching a neural network, analyzing the effectiveness of various neural network training algorithms, determining the structure of a neural network, which consists in finding the minimum error of neural network training depending on the number of neurons in the hidden layer, as well as in the analysis of the effectiveness of the results. The ability of the developed neural network to smooth out white noise was proved by determining the identification error of the rotational speed of the turbocharger's rotor, which was 0,005 % and did not exceed the limit-permissible value of 0,5 %. Originality. The scientific novelty of the results obtained is as follows: For the first time, a method was developed for determining the optimal structure of a neural network, which made it possible to solve the problem of identifying a dynamic model of an aircraft engine, the TV3-117, in onboard conditions with minimal errors. The method of identifying the technical condition of the TV3-117 aircraft engine in onboard conditions, which differs from the existing ones due to the use of neural network technologies, makes it possible to increase the reliability of monitoring and diagnostics of the technical condition of the TV3-117 aircraft engine under its flight conditions. Practical value.The developed neural network can be one of the units of the expert system that can automatically make decisions regarding the technical condition of the aviation engine TV3-117 in flight modes and provide information to the crew about the possibility of further safe movement of the aircraft. The task of developing an expert system can be effectively solved using the mathematical apparatus of neural networks, since its use increases the reliability and accuracy of classification of modes, identification, control, diagnostics, time series analysis (forecasting), debugging of engine parameters, etc., which will increase reliability of obtaining the necessary results. References 10, table 1, figure 4.

The subject matter in the article is TV3-117 aircraft engine and methods of identification of its technical condition. The goal of the work is to develop methods for identifying the technical state of the aircraft engine TV3-117 on the basis of real-time neural network technologies. The following tasks were solved in the article: the task of identifying the reverse multi-mode model of the aircraft engine TV3-117 using neural networks. The following methods used aremethods of probability theory and mathematical statistics, methods of neuroinformatics, methods of the theory of information systems and data processing. The following results were obtained-The application of the neural network apparatus is effective in solving a large range of tasks: identifying the mathematical model of the aircraft engine TV3-117, diagnosing the condition, analyzing the trends, forecasting the parameters, etc., despite the fact that these tasks usually relate to the class difficultly formalized (poorly structured), neural networks are adequate and effective in the process of their solution. In the process of solving the task of identifying the mathematical model of the aircraft engine TV3-117 on the basis of neural networks, it was established that neural networks solve the problem of identification more precisely classical methods. Conclusions: It was established that the error of identification of the aircraft engine TV3-117 with the help of a neural network of type perceptron did not exceed 1.8 %; For the neural network of radial-basic function (RBF)-4.6 %, whereas for the classical method (LSM) it makes about 5.7 % in the considered range of changes in engine operation modes. It was found that neural network methods are more robust to external perturbations: for noise level ζ = 0.01, the error of identification of aircraft engine TV3-117 with the use of perceptron has increased from 1.8 to 3.8 %; for the neural network RBFfrom 4.6 to 5.7 %, and for the least squares methodfrom 5.7 to 13.93 %. In the process of solving the task of identifying the inverse multi-mode model of the aviation engine TV3-117 on its parameters on the basis of neural networks (perceptron and RBF) it was shown that their use allows for indirect measurement of the parameters of the flowing part of the engine at different modes of its operation: in the absence of noisewith an error of not more than 1,8 and 4,6 % respectively; in the presence of noise (ζ = 0,01)with an error of not more than 3,8 and 5,7 % respectively. Application in these conditions of the least squares method (polynomial regression model of the 8th order) allows us to obtain the error value: in the absence of noiseno more than 5,7 %; in the presence of noiseno more than 13,93 %.

No abstract