METHOD FOR PREDICTING SPECIAL CASES IN FLIGHT BASED ON EARLY DETECTION OF ANOMALOUS SEQUENCES IN THE DIAGNOSTIC DATA OF AIRCRAFT TECHNOLOGICAL EQUIPMENT
DOI:
https://doi.org/10.26906/SUNZ.2020.3.028Keywords:
flight safety, special cases in flight, parametric diagnostics, forecasting, anomalous sequence, time series, temporal pattern, hybrid stochastic modelAbstract
Modern onboard digital systems for automatic control, monitoring and diagnostics allow measuring a large number of parameters of aircraft technological equipment and receiving arrays of such information in digital form. Prediction of special cases in flight is the main task of parametric diagnostics of aircraft technological equipment. However, the existing diagnostic models based on the corresponding mathematical models do not fully use the diagnostic data arrays and do not always allow predicting the occurrence of technological equipment failures. This makes the task of predicting special cases in flight relevant. The purpose of the article is to develop a method for predicting special cases in flight based on the detection of abnormal sequences in the diagnostic data of the technological equipment of the aircraft; in order to increase flight safety. Results of the research. The paper proposes a method for predicting special cases in flight based on the early detection of anomalous sequences in the diagnostic data of the aircraft technological equipment. For the early detection of abnormal sequences, it is proposed to use a hybrid stochastic model and a method for detecting abnormal sequences in the diagnostic data of aircraft technological equipment. The input training information is provided in the form of vectors of observations of the development of the process in which the final value is especially highlighted, as a result, characterizing the facts of belonging of the vector to the class of normal or abnormal temporal patterns. Conclusion. The application of the proposed method will allow to implement the prognostic principle of flight safety management, as well as to obtain the economic effect of preventing aircraft downtime due to sudden equipment failure. Modern onboard digital systems for automatic control, monitoring and diagnostics allow measuring a large number of parameters of aircraft technological equipment and receiving arrays of such information in digital form. Prediction of special cases in flight is the main task of parametric diagnostics of aircraft technological equipment. However, the existing diagnostic models based on the corresponding mathematical models do not fully use the diagnostic data arrays and do not always allow predicting the occurrence of technological equipment failures. This makes the task of predicting special cases in flight relevant. The purpose of the article is to develop a method for predicting special cases in flight based on the detection of abnormal sequences in the diagnostic data of the technological equipment of the aircraft; in order to increase flight safety. Results of the research. The paper proposes a method for predicting special cases in flight based on the early detection of anomalous sequences in the diagnostic data of the aircraft technological equipment. For the early detection of abnormal sequences, it is proposed to use a hybrid stochastic model and a method for detecting abnormal sequences in the diagnostic data of aircraft technological equipment. The input training information is provided in the form of vectors of observations of the development of the process in which the final value is especially highlighted, as a result, characterizing the facts of belonging of the vector to the class of normal or abnormal temporal patterns. Conclusion. The application of the proposed method will allow to implement the prognostic principle of flight safety management, as well as to obtain the economic effect of preventing aircraft downtime due to sudden equipment failureDownloads
References
Yule, G.U. On a method of investigating periodicities in disturbed series, with special reference to Wolfer's sunspot numbers / G. U. Yule // Philosophical Transactions of the Royal Society of London. Series A, Containing Papers of a Mathematical or Physical Character. – 1927. – pp. 267-298.
Koskela, T. Neural network methods in analyzing and modelling time varying processes / T. Koskela et al. – Helsinki University of Technology, 2003.
Sfetsos, A. Time series forecasting with a hybrid clustering scheme and pattern recognition / A. Sfetsos, C. Siriopoulos // Systems, Man and Cybernetics, Part A: Systems and Humans, IEEE Transactions on. – 2004. – Vol. 34, No. 3. – pp. 399-405.
Esling, P. Time-series data mining / P. Esling, C. Agon // ACM Comp. Surveys (CSUR). – 2012. – Vol. 45, No. 1. – pp. 12.
Pesaran, M. H. Forecasting time series subject to multiple structural breaks / M. H. Pesaran, D. Pettenuzzo, A. Timmermann // The Review of Economic Studies. – 2006. – Vol. 73, No. 4. – pp. 1057-1084.
Wagner, N. Time series forecasting for dynamic environments: the DyFor genetic program model / N. Wagner et al. // IEEE Transactions on Evolutionary Computation – 2007. – Vol. 11, No. 4. – pp. 433-452.
Burkom, H. S. Automated time series forecasting for biosurveillance / H. S. Burkom, S. P. Murphy, G. Shmueli // Statistics in medicine. – 2007. – Vol. 26, No 22. – pp. 4202-4218.
Song, H. Tourism demand modelling and forecasting – A review of recent research / H. Song, G. Li // Tourism Management. – 2008. – Vol. 29, No. 2. – pp. 203-220.
Herrera, L. J. Recursive prediction for long term time series forecasting using advanced models / L. J. Herrera et al. // Neurocomputing. – 2007. – Vol. 70, No 16. – pp. 2870-2880.
Patcha, A. An overview of anomaly detection techniques: Existing solutions and latest technological trends / A. Patcha, J.M. Park // Computer Networks. – 2007. – Vol. 51, No. 12. – pp. 3448-3470.
Sorjamaa, A. Methodology for long-term prediction of time series / A. Sorjamaa et al. // Neurocomputing. – 2007. – Vol. 70, No. 16. – pp. 2861-2869.
Averkin, A. Time series forecasting based on hybrid neural networks and multiple regression / A. Averkin, S. Yarushev, I. Dolgy, A. Sukhanov // Proceedings of the First International Scientific Conference “Intelligent Information Technologies for Industry” (IITI’16): Volume 1. – 2016. – pp. 111-121.
Гибридная стохастическая модель обнаружения особых типов паттернов в темпоральных данных / С.М. Ковалев, А.Н. Гуда, М.А. Бутакова // Вестник Ростовского государственного университета путей сообщения. – 2013. – No3(51). – С. 36-42.
