Vol 27, No 4 (2019)

The effectiveness of the binary artificial neural network at low volume of training samples was investigated. The studies were conducted with the MATLAB software’s Neural Network Toolbox module used for the classification task. Feed-forward artificial neural network was thoroughly studied. It was discovered, that the number of neurons in the inner layer in the range of changes had little effect on efficiency. The volume of training samples reduces efficiency when the value is very small. The influence the of the input training samples dimension on the efficiency was studied; the dimension was reduced from 64 to 16 by eliminating a different set of input parameters. It is established that such changes have a different impact on efficiency, and under certain conditions, the efficiency increases. For a cascade binary network of feed forward propagation and a cascade binary network with reverse propagation of error, it was established that the efficiency changes slightly depending on the number of neurons in the hidden layer in the studied range. A two-layer cascade network with reverse error propagation with a certain combination of neurons has a higher efficiency compared to a single-layer one. The results allow us to outline ways to improve efficiency.

The paper provides the methodology for ground testing of an aircraft flight safety management system. The technique proposed by the author allows us to evaluate the operability of the software and hardware of the system in the normal mode of its operation, as well as in the conditions of failure of interacting equipment. Particular attention is paid to the description of the test bench for the system, which contains simulators of the aircraft avionics complex and the model of the aircraft, which provides ground-based tests of the system under various operating conditions. The flight conditions of the aircraft and the condition of its on-board equipment are set by the operator of the ground test bench of the system using specialized software and auxiliary test hardware. In the process of testing, the input and output variables of the system are registered with their display at the operator's workplace. Using the registered variables, it is possible to evaluate the operability of the software and hardware of the aircraft flight safety management system. The presented methodology for ground tests of the aircraft flight safety control system can be applied in the process of developing systems of this class at the stage of their preliminary tests

The paper presents the results of the analysis of different approaches to risk management in the production activities of enterprises, risk classifications. Several definitions of the concept of risk, its types and assessment possibilities are considered. The formation of production plans, as a system, is focused on the partners of the enterprise in the production chain, in particular, on the consumers of its products. Therefore, an important aspect is taking into account the customers interests, conditions and options for interaction with them, reputation and financial stability of the consumer. All this determines the degree of reliability of the buyer as a partner in the planning of production for the enterprise. The importance of assessing the degree of reliability of consumers is emphasized by the fact that it is a factor that determines the direction and results of a comprehensive analysis of orders of the production enterprise, predetermines the development of possible alternative production plans, their choice, then-participates in the organization of production and sale of finished products to consumers. This view determined the most preferable classification of risks, and as a priority direction in planning the portfolio of orders of the production enterprise - the approach to the assessment of consumers in terms of reliability factors of interaction between them. As a result, the mechanism of multi-criteria evaluation of the production enterprise customers on the basis of hierarchy analysis method is proposed. Key reliability criteria, such as delays in payment of orders, payment terms, frequency and volume of purchases, financial stability of the consumer etc., were selected as criteria for comparing consumers and forming an aggregate reliability assessment of each of them, and a system for the formalized description of this assessment mechanism was developed. On the basis of the received estimates it is possible to define the most reliable and, consequently, priority partners, and also to plan their share in the orders portfolio of the production enterprise as a larger one in comparison with the others.

The paper discusses the problem of the computer equipment distribution between units of a large engineering enterprise. It is shown that automated accounting of the state of a large number of computer facilities is essential for organizing effective resource management. In this case, it is necessary to develop optimal control algorithms for the computational problems distribution and related computer equipment. The task of optimal use of computer equipment in industrial enterprises is one of the most important in modern information technologies. This is due to the fact that the operation, maintenance and modernization of computing equipment require significant financial costs. The organizational structure of the space industry engineering company is considered. It is shown that a decisive role for the efficient distribution of computational resources is keeping the whole range of tasks performed. The correspondence of the task’s types to the types of computer equipment and the production orientation of units is given. A system model of an evolving information system based on lattice algebraic models is described. It represents a complex of two models: structural and behavior models depending on the set of tasks being solved. The statement and solution of the problem of optimizing the computer technology distribution taking into account the types and configurations of computers is the main scientific result of the article. The problem of optimizing the of computer equipment distribution taking into account the computers types and configurations is formulated. The cost minimization problem with the corresponding restrictions relates to discrete programming problems with Boolean variables. Constraint equations relate the distribution of tasks to units, as well as computer hardware parameters. A graph model of the variables relationship for the discrete optimization problem is proposed. The results of experimental studies in the distribution of computer resources based on the solution of the optimization problem are considered. The proposed methodology allowed reducing the amount of purchased computer equipment and operating costs.

The paper deals with the development of wavelet-transform computation basis which allows to restore the original signal wavelet-coefficients from the coefficients given by decomposition of the original signal with Chebyshev-Hermite functions. To form the basis in question, the wavelet transform of the Chebyshev-Hermite functions is analytically calculated. Derivatives of the Gauss function are used as wavelets. The paper considers the formation of transition bases from the coefficients of signal expansion by Chebyshev-Hermite functions to wavelet transforms using the Gauss functions of the 1st and m-th order as analyzing wavelets. As an example, the basis of the transition from the expansion coefficients of the original signal in terms of the Chebyshev-Hermite functions to the wavelet-transform using the MHAT wavelet is given. In this case, the wavelet transform of the signal is carried out in two stages. At the first stage, the decomposition of the initial signal is obtained in the form of a weighted sum of the Chebyshev-Hermite basis functions. At the second stage, knowing the weight factors of the functions obtained at the first stage, as well as the analytical expression of the continuous wavelet-transform for specific basis functions and the wavelet, it is possible to restore the wavelet-transform of the original signal. The examples of wavelet-transforms of the bases formed are given. Using the obtained formulas for calculating wavelet coefficients, it is possible to construct fast computational processing algorithms. For calculations and graphical representations of the simulation results, the Wolfram Mathematica 11.3 computer algebra system was used.

The paper is devoted to the development and investigation of the automatic correction system of rotor radial position updating of the micro electric drive, used in high-precision processing systems. A new approach of solving the problem of stabilizing the rotating element of motor, based on active current updating of the spatial position is offered. Automatic stabilization is achieved by organizing an electromagnetic effect of reversible nature in the electromechanical system. It is proposed to use the paired groups of electromagnets operating on the control object. The development of the automatic correction algorithm is done by the signal-adaptive inverse model method. The approach is able to provide the desired quality indicators for the control process of a nonlinear object, and also to achieve insensitivity of the correction system for external and internal signal and parametric disturbances. The paper describes the way of constructive implementation of this method, structural synthesis methodology of the control law. Structural and functional diagrams of the automatic correction system for radial deviations are presented in the article, as well as the results of simulation, confirming potential of the conducted research.

Currently, submersible induction motors are widely used in electric centrifugal pump units, which provide the bulk of oil production in Russia. To maintain a given dynamic fluid level in the well, submersible pump control stations are equipped with closed systems with frequency converters. For the correct synthesis of the regulators of the dynamic level stabilization system, it is necessary to know the transfer function of the induction motor as a control object. It should be borne in mind that for matching a high-voltage submersible motor with a frequency converter, a step-up transformer and scalar frequency control of an asynchronous machine are used. Existing linearized mathematical models of an induction motor with this control method are approximate. Therefore, the purpose of this study is to obtain an updated model of a submersible induction motor. The article provides a well-known system of equations of an asynchronous machine with scalar frequency control. It is shown that this system is non-linear since a number of variables are multiplied in it. The linearization of the main nonlinearities was carried out by the method of their expansion into a Taylor series. In the linearized system, the frequency of the supply voltage of the stator is taken as the input coordinate, and the rotational speed of the rotor of the induction motor is taken as the output coordinate. Getting rid of the intermediate variables by the method of sequential substitution, we obtained the transfer function of a submersible induction motor with scalar frequency control. It is shown that the characteristic polynomial of the found transfer function is of the fifteenth order. Analytical expressions are found that connect the coefficients of the transfer function with the parameters of the induction motor and the coordinates of the starting point. The adequacy of the obtained mathematical model is estimated by comparing the modeling results of step response in a nonlinear system of equations and in a linearized one. It is shown that the discrepancy between the results does not exceed 0.54%.