The existing standards to evaluate the technical state of overhead railway crane tracks, although important, are precision-limited and do not consider the complex interconnection between various defects occurring during the operation. The study aims to overcome these defects by developing an intellectual decision support system to help experts evaluate the technical state of crane tracks and make substantiated decisions on their potential further operation.
Crane tracks are exposed to multifaceted wear during operation that manifests in various defects, such as rail deterioration, joint deformation, railway geometry deviation, corrosion of elements, etc. The combination of these defects and their interaction can significantly impact the bearing capacity and the safety of operation of a crane track. The complexity of the issue requires considering the practices of operation, expert knowledge, and the analysis of large amounts of data.
The study aims to develop an intellectual decision support system in order to evaluate the technical state of overhead railway crane tracks based on the statistical database, information on defects, their combinations and the dynamics of development, the results of operation of crane tracks, formalized knowledge and experience of the experts, and recommendations of the existing standards that will help experts with no vast practical experience in inspection of crane tracks make decisions on their potential further operation.
The neural network is proposed for use as a computational core that considers various combinations of wear and deviations of geometrical parameters of a crane track with different degrees of deterioration expressed in percentage of limiting values. It can analyze, in particular, such parameters as rail deterioration (vertical, lateral), deviations from straightness and horizontality, size of junction gaps, as well as the degree of corrosion and deformation of fixing elements. At the same time, each parameter can be assigned a weight coefficient indicating its impact on the general operability of the crane track.
The resulting evaluation obtained by the artificial neural network will help classify the state of a crate track by categories (for example, operable, requiring repair, inoperable) and formulate recommendations regarding required repairs or limitations in operation.
2. Korotkiy A.A., Panfilov A.V., Khvan R.V., Yusupov A.R. Integral method of assessing defects on the operability of steel rope using artificial neural networks. Transportnoe, gornoe i stroitelnoe mashinostroenie: nauka i proizvodstvo = Transport, mining and construction engineering: science and production. 2023. № 18. pp. 73–79. (In Russ.). DOI:10.26160/2658-3305-2023-18-73-79
3. Egelskiy V.V., Nikolaev N.N., Egelskaya E.V., Korotkiy A.A. Influence of the Competencies of Lifting Crane Specialists on the Probability of Emergencies. Bezopasnost tekhnogennykh i prirodnykh system = Safety of Technogenic and Natural Systems. 2023. № 2. pp. 70–79. (In Russ.). DOI:10.23947/2541-9129-2023-7-2-70-79
4. Khvan R.V. Use of artificial neural networks for solving the problem of residual resource estimation of hoisting cranes. E3S Web of Conferences. 2024. Vol. 515. DOI:10.1051/e3sconf/202451504015
5. Dennig D., Bureick J., Link J., Diener D., Hesse C., Neumann I. Comprehensive and Highly Accurate Measurements of Crane Runways, Profiles and Fastenings. Sensors. 2017. Vol. 17. Iss. 5. DOI: 10.3390/s17051118
6. Seliverstov G.V., Antsev V.Yu., Voblikova Yu.O. Building an optimal algorithm for crane metal structure diagnostics. Stroitelnye i dorozhnye mashiny = Construction and road machinery. 2013. № 7. pp. 23–24. (In Russ.).
7. Kraus M., Mämpel S. Kennwerte neuer und abgenutzter Kranschienen für die Bemessung von Kranbahnträgern. Stahlbau. 2017. Vol. 86. Iss. 1. pp. 36–44. DOI: 10.1002/stab.201710445
8. Kettler M., Derler C., Schörghofer A., Macho C., Unterwegeret H. Laboratory and numerical tests on real crane runway girder with box section. Journal of Constructional Steel Research. 2019. Vol. 160. pp. 540–558. DOI: 10.1016/j.jcsr.2019.06.002
9. Khvan R.V. Comparative Analysis of the Performance of Artificial Neural Networks in Assessing the Technical Condition of Steel Ropes. Bezopasnost tekhnogennykh i prirodnykh sistem = Safety of Technogenic Natural Systems. 2024. № 2. pp. 68–77. (In Russ.). DOI: 10.23947/2541-9129-2024-8-2-68-77
10. Panasiuk J.A., Adamus A. A study on Elements of Machine Vision and Machine Learning in Industrial Vision Systems. Problems of Mechatronics Armament Aviation, Safety Engineering. 2024. Vol. 15. Iss. 3. pp. 87–106. DOI: 10.5604/01.3001.0054.7513