Abstract
Speed limits for special self-propelled rolling stock are usually set from design documents, while the actual stress state of an aged frame may differ. This thesis shows how a strength-based speed limit can be obtained directly from running-test data. Maximum stresses measured by 16 strain-gauge channels on the frame of an ADM-1 railcar at 10, 30, 55 and 60 km/h were fitted by linear regression. The models explain 79–99% of the variance (median R² = 0.92). The most loaded zone, under the diesel-engine beam, reaches the permissible stress of 152.75 MPa at about 54 km/h. The same regression structure can serve as a first surrogate model for machine-learning-based monitoring of electric rolling stock.
References
1. Rakhmiddinov I.O. Modernization and service life extension of special self-propelled rolling stock operated in the Republic of Uzbekistan. PhD thesis abstract, specialty 05.08.05. Tashkent State Transport University, Tashkent, 2025.
2. Rakhmiddinov I., Fayzibaev S., Kasimov O., Kosimov K., Panchenko V. Improving the structural reliability of self-propelled railcar frames through modernization techniques. AIP Conference Proceedings, 2026, 3447, 060011. https://doi.org/10.1063/12.0044293
3. GOST 31846-2012. Special railway rolling stock. Requirements for strength of supporting structures and dynamic qualities. Moscow: Standartinform, 2013.
4. Chen T., Guestrin C. XGBoost: A scalable tree boosting system. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 2016, pp. 785–794.