Intelligent monitoring of the transportation processes based on the dynamic method of principal components

The paper considers a new technology for intelligent monitoring of the railway transportation processes using the dynamic method of principal components. It includes a recursive principal feature calculation algorithm and three statistical criteria used in the decision engine. The application of the proposed scheme demonstrates the feasibility and efficiency of recursive algorithms for adaptive monitoring of complex poorly formalized processes in online mode.

Whereas the most technological processes undergo slow, evolving changes, such as aging of floor equipment, sensor drifts, periodic maintenance and modernization of technical equipment, it is expected that the adaptive monitoring scheme proposed in the article will be widely used in railway transport.

1. Analytical review of the proceedings of the IITI’19 conference / S. M. Kovalev, V. Snashel, A. N. Guda [et al.] // Vestnik Rostovskogo Gosudarstvennogo Universiteta Putey Soobshcheniya. – 2020. – No. 3. – P. 86–105. – DOI: 10.46973/0201–727X_2020_3_86.

2. Dolgiy, A. Intelligent Models for State Assessment and Behavior Prediction in Railway Processes Based on Descriptive Analytics and Soft Computing / A. Dolgiy, A. Khramtsov, S. Kovalev // Proceedings of the Sixth International Scientific Conference “Intelligent Information Technologies for Industry”(IITI’22). – Cham: Springer International Publishing, 2022. – P. 358–368. – ISBN 978-3-031-19619-5.

3. Khatlamadzhiyan, A. E. Cognitive Measurements and Predictive Analytics for Railway Infrastructure Components / A. E. Khatlamadzhiyan, S. M. Kovalev, V. B. Tarassov // International Conference on Intelligent Information Technologies for Industry. – Springer, Cham, 2021. – P. 513–526. – ISSN 2367-3370.

4. Laiton-Bonadiez, C. Industry 4.0 technologies applied to the rail transportation industry: A systematic review / C. Laiton-Bonadiez // Sensors. – 2022. – Т. 22. – No. 7. – P. 2491. – DOI: 10.3390/s22072491.

5. Levin, D. Yu. Dispatch control of car flows / D. Yu. Levin // Intelligent control systems in railway transport. Computer and mathematical modeling (ISUZhT–2019). – 2019. – P. 51–58. – ISSN 1992-3252.

6. PCA-SVM-based automated fault detection and diagnosis (AFDD) for vapor-compression refrigeration systems / H. Han, Z. Cao, B. Gu, N. Ren // HVAC&R Res. 16 (2010). – P. 295–313. – DOI: 10.1080/10789669.2010.10390906.

7. Wang, J. A new subspace identification approach based on principle component analysis / J. Wang and S. J. Qin // Journal of Process Control. – 2002. – Vol. 12, No. 8. – P. 841–855. – DOI: 10.1016/S0959-1524(02)00016-1.

8. Moving window kernel PCA for adaptive monitoring of nonlinear processes / X. Liu, U. Kruger, T. Littler [et al.] // Chemometrics and Intelligent Laboratory Systems. – 2009. – Vol. 96, No. 2. – P. 132–143. – DOI: 10.1016/j.chemolab.2009.01.002.

9. Rannar, S. Adaptive batch monitoring using hierarchical PCA / S. Rannar, J. MacGregor and S. Wold // Chemometrics and Intelligent Laboratory Systems. – 1998. – Vol. 41, No. 1. – P. 73–81. – DOI: 10.1002/cem.678.

10. Qin, S. J. Recursive PLS algorithms for adaptive data modeling / S. J. Qin // Computers & Chemical Engineering. – 1998. – Vol. 22, No. 4–5. – P. 503–514. – DOI: 10.1016/S00981354(97)00262-7.

11. Recursive PCA for adaptive process monitoring / W. Li, H. Yue, S. Valle-Cervantes and S. Qin. Journal of Process Control. – 2000. – Vol. 10, No. 5. – P. 471–486. – DOI: 10.1016/S09591524(00)00022-6.

12. Efficient recursive PCA algorithms for process monitoring / L. Elshenawy, S. Yin, A. Naik and S. Ding // Industrial Engineering Research. – 2010. – Vol. 49, No. 1. – P. 252–259. – DOI: 10.1021/ie900720w.

13. A simplified recursive dynamic PCA based monitoring scheme for imperial smelting process / Z. Hu [et al.] // International Journal of Innovative Computing, Information and Control. – 2012. – V. 8, No. 4. – P. 2551–2561 – ISSN 1349-4198.

14. Kovalev, S. Incremental Structure-Evolving Intelligent Systems with Advanced Interpretational Properties / S. Kovalev, A. Kolodenkova, A. Sukhanov // Artificial Intelligence. RCAI 2020. Lecture Notes in Computer Science, Vol 12412. – 2020. – Springer, Cham. – ISSN 0302-9743.