Technical state prediction of complex technical systems by the Berg algorithm and Bayesian networks

DOI: 10.34759/trd-2020-113-14

Аuthors

Dorozhko I. V.^*, Osipov N. A.^**, Ivanov O. A.^***

Mlitary spaсe Aсademy named after A.F. Mozhaisky, Saint Petersburg, Russia

*e-mail: Doroghko-Igor@yandex.ru
**e-mail: bayes@mail.ru
***e-mail: kompa4ka@yandex.ru

Abstract

The purpose of the presented work consists in predicting the diagnostic characters changes and correlating them with possible type of technical state for undertaking preemptive measures while complex technical systems diagnosing.

The approach being proposed and models are based on the basic concepts and relations of reliability theory and technical diagnostics of systems. The source data represents information on the reliability (failure rate of elements, structural and logical schemes) of complex technical structures and diagnostic models linking the types of technical states and diagnostic features. The sets of conditional probabilities or distributions density are being pointed for the cause-and-effect relationships between types of technical states and diagnostic features, depending of continuous or discrete type of the diagnostic character.

The algorithm for predicting the values of diagnostic features proposed in the article is based on the Berg linear prediction algorithm. It is adapted and studied herewith for the main functional dependencies of diagnostic features. The results of diagnostic characters predicting are considered on the specific examples, and estimates of the forecast accuracy are given. The model based on dynamic hybrid Bayesian trust networks includes discrete and continuous variables, which describe the cause-and-effect relationships of types of technical states and diagnostic features, as well as the relationship of blocks (elements) in terms of reliability. The results of the logical-probabilistic inference indicate that the decision on the type of technical condition of a complex technical structure changes significantly when accounting for the forecast estimates of diagnostic characters.

The proposed scientific and methodological approach can be employed to create diagnostic software for modern complex technical structures with artificial intelligence elements.

Mathematical models of diagnostics considered in the article account for the connections types, elements reliability, as well as dynamics of technical states types and their relationship with diagnostic features, which can be both continuous and discrete. the Technical condition predicting allows implementing the proactive management concept, parry possible failures, changing the operating modes in advance, switching to backup elements, etc.

Keywords:

prediction, technical state, diagnostic character, Berg algorithm, Bayesian network

References

1. GOST 27.002-15. Nadezhnost’ v tekhnike. Osnovnye ponyatiya. Terminy i opredeleniya. (State Standard 27.002-15. Reliability in engineering. The basic concepts. Terms and definitions), Moscow, Standartinform, 2016, 24 p.

2. Klyuev V.V. et al. Tekhnicheskie sredstva diagnostirovaniya (Technical means of diagnostics), Moscow, Mashinostroenie, 1989, 671 p.

3. Gusev P.Yu., Gusev K.Yu. Trudy MAI, 2020, no. 110. URL: http://trudymai.ru/eng/published.php?ID=112933. DOI: 10.34759/trd-2020-110-20

4. Kosinskii M.Yu., Shatskii M.A. Trudy MAI, 2014, no. 74. URL: http://trudymai.ru/eng/published.php?ID=49315

5. Shevtsov S.N., Sibirskii V.V., Chigrinets E.G. Trudy MAI, 2016, no. 91. URL: http://trudymai.ru/eng/published.php?ID=75572

6. Arsen'ev V.N. Otsenivanie kharakteristik sistem upravleniya po ogranichennomu chislu naturnykh ispytanii (Characteristics evaluation of control systems according to the specified number of full-scale tests), Moscow, Izd-vo “RESTART”, 2013, 126 p.

7. Burg J.P. Maximum Entropy Spectral Analysis, PhD thesis, Department of Geophysics, Stanford University, Stanford, CA, 1975.

8. Cedrick Collomb. Burg's Method, Algorithm and Recursion, 2009, URL: http://www.emptyloop.com/technotes/a%20tutorial%20on%20burg's%20method,%20algorithm%20and%20recursion.pdf

9. Tulup'ev A.L., Nikolenko S.I., Sirotkin A.V. Osnovy teorii baiesovskikh setei (Fundamentals of the Bayesian Theory), Saint Petersburg, Izd-vo Sankt-Peterburgskogo universiteta, 2019, 399 p.

10. Cowell R.G., Dawid A.P., Lauritzen S.L., Spiegelhalter D.J. Probabilistic Networks and Expert Systems, Springer-Verlag, 1999.

11. Jensen F.V. Bayesian Networks and Decision Graphs, New York, Springer-Verlag, 2001, 457 p.

12. Pearl J. Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference, New York, Morgan Kaufman Publ., 1991. DOI: 10.1016/0004-3702(91)90084-W

13. Dorozhko I.V., Kochanov I.A., Osipov N.A. et al. Trudy Voenno-kosmicheskoi akademii im. A.F. Mozhaiskogo, 2016, no. 652, pp 137 – 146.

14. Dorozhko I.V., Zakharova E.A., Osipov N.A. Trudy Voenno-kosmicheskoi akademii im. A.F. Mozhaiskogo, 2019, no. 669, pp. 216 – 223.

15. GeNIe & SMILE. Decisions systems laboratory. School of Information Sciences. University of Pittsburg. URL: http://genie.sis.pitt.edu/

16. Nadezhnost’ v tekhnike. Osnovnye ponyatiya. Terminy i opredeleniya. GOST 27.002-15. (Reliability in Engineering. The basic concepts. Terms and definitions. State Sstandard 27.002-15.), Moscow, Standartinform, 2016, 24 p.

17. Dmitriev A.K. Modeli i metody analiza tekhnicheskogo sostoyaniya bortovykh sistem (Models and methods of onboard systems technical condition analysis), Saint Petersburg, VIKU imeni A.F. Mozhaiskogo, 1999, 171 p.

18. Dmitriev A.K., Kopkin E.V. Izvestiya vuzov. Priborostroenie, 1999, no. 9, vol. 42, pp. 3 - 10.

19. Dmitriev A.K, Yusupov R.M. Identifikatsiya i tekhnicheskaya diagnostika (Identification and technical diagnostics), Leningad, MO SSSR, 1987, 521 p.

20. Kopkin E.V., Kravtsov A.N., Myshko V.V. Analiz tekhnicheskogo sostoyaniya kosmicheskikh sredstv (Technical condition analysis of space means), Saint-Petersburg, VKA imeni A.F. Mozhaiskogo, 2016, 189 p.

21. Kopkin E.V., Kravtsov A.N., Myshko V.V. Kontrol’ i diagnostika kosmicheskikh sredstv (Control and diagnostics of space means), Saint-Petersburg, VKA imeni A.F. Mozhaiskogo, 2016, 198 p.

22. Okhtilev M.Yu., Sokolov B.V., Yusupov R.M. Izvestiya YuFU. Tekhnicheskie nauki, 2015, no. 1(162), pp. 162 - 174.

23. Okhtilev M.Yu., Mustafin N.G., Miller V.E., Sokolov B.V. Izvestiya Vuzov. Priborostroenie, 2014, vol. 57, no. 11, pp. 7 – 15.

24. Motienko A., Basov O., Dorozhko I., Tarasov A. Proactive Robotic Systems For Effective Rescuing Sufferers. GmbH: Springer-Verlag, Lecture Notes In Computer Science, 2016, pp. 172 - 180. DOI: 10.1007/978-3-319-43955-6_21

Download