350 rub
Journal Neurocomputers №2 for 2024 г.
Article in number:
Methodology for determining critical elements of an information and measurement system based on forecasting the values of their technical characteristics
Type of article: scientific article
DOI: https://doi.org/10.18127/j19998554-202402-07
UDC: 681.518.3
Authors:

V.V. Lisitsky1, D.M. Vyaldin2

1,2 Military Space Academy named after A.F. Mozhaisky (Saint-Petersburg, Russia)

1 vka@mil.ru, 2 dmitriy_vyaldin@mail.ru

Abstract:

An analysis of the operation of the monitoring and diagnostics system of information and measurement systems has shown that data from all elements of the equipment connected to each other, as a rule, according to a hierarchical principle based on constructive division, are used to obtain information about a critical change in the values of technical characteristics. When solving the problem of forecasting technical characteristics in the interests of justifying the volume of maintenance, a large number of elements leads to a decrease in the accuracy and horizon of the forecast due to the presence of excessive information about elements that do not affect the deviation of technical characteristics from the nominal value. To ensure the necessary accuracy of the information and measurement system, it is necessary to use a large array of data on the technical characteristics of structural elements with their subsequent processing using machine learning models. To consider a methodology based on forecasting the values of the technical characteristics of the elements of an information and measurement system to determine their critical importance. Results. To solve the problem of predicting the values of the technical characteristics of an information and measurement system, it is proposed to use a gradient boosting algorithm and a linear regression method. The proposed method makes it possible to organize continuous training of the algorithm for determining critical elements in real time at the operational stage of the information and measurement system.

Pages: 70-78
For citation

Lisitsky V.V., Vyaldin D.M. Methodology for determining critical elements of an information and measurement system based on forecasting the values of their technical characteristics. Neurocomputers. 2024. V. 26. № 2. Р. 70-78. DOI: https://doi.org/10.18127/ j19998554-202402-07 (In Russian)

References
  1. Theoretical foundations of radar. Edited by V.E. Dulevich. M.: Radio engineering. 1978. 608 p. (In Russian)
  2. Boev S.F., Ignatiev S.V., Tikhonov V.B., Belyakov R.A., Linkevicius A.P. Reliability and operation of radio engineering systems: monograph. M.: Scientific book. 2018. 370 p. ISBN 978-5-9500445-7-1. (In Russian)
  3. Lisitsky V.V., Pankov B.B. Methodology for monitoring energy parameters of information and measurement systems in real time. Proceedings of the Military Space Academy named after A.F. Mozhaisky. 2021. № 679. P. 87–93. (In Russian)
  4. Vyaldin D.M., Stolbov A.V., Martynov A.M. Algorithm for developing rational options for restoring information and measurement systems in the event of emergency situations. Collection of scientific papers of VNTK "Actual scientific and technical aspects of the development, testing and operation of missile and space defense facilities". St. Petersburg: A.F. Mozhaisky VKA. 2022. P. 102–108. (In Russian)
  5. Kelleher J.D., Mac Namee B., D'Arcy A. Fundamentals of Machine Learning for Predictive Data Analytics: Algorithm Worked Examples and Case Studies. The MIT Press. 2015. ISBN 9780262029445.
  6. Alexandrovskaya L.N., Afanasyev A.P., Lisov A.A. Modern methods of ensuring reliability of complex technical systems. M.: Logos. 2001. 208 p. (In Russian)
  7. Sorjamaa A., Hao J., Reyhani N., Ji Y., Lendasse A. Methodology for long-term prediction of time series. Neurocomputing. 2007. V. 70. № 16–18. P. 2861–2869. DOI 10.1016/j.neucom.2006.06.015.
  8. Kline D.M. Methods for Multi-Step Time Series Forecasting Neural Networks. Neural Networks in Business Forecasting. 2004. P. 226–250. DOI 10.4018/978-1-59140-176-6.ch012.
  9. Cheng H., Tan P.-N., Gao J., Scripps J. Multistep-Ahead Time Series Prediction. Advances in Knowledge Discovery and Data Mining, Lecture Notes in Computer Science. 2006. V. 3918. P. 765–774. DOI 10.1007/11731139_89.
  10. Antoshina V.M., Zager I.B., Logovsky A.S., Lvov K.V., Perlov A.Yu. A simulation model of data on the technical condition of the functional systems of the radar when solving the problem of forecasting failures. Electromagnetic waves and electronic systems. 2019. V. 24. № 5. P. 11–16. DOI 10.18127/j15604128-201905-02. (In Russian)
Date of receipt: 19.02.2024
Approved after review: 11.03.2024
Accepted for publication: 26.03.2024