.png){kind=icon}
.png){kind=icon}
Return
Intelligent algorithm-driven optimization for turnout rail grinding strategy and carbody vibration suppression
.png){kind=icon}
PRE
AI
DOI：10.1080/00423114.2025.2570879.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
High-speed trains face excessive carbody lateral acceleration (CLA) and swaying when traversing worn turnouts, while rail grinding is effective yet lacks a systematic strategy for precise operations. To address this, worn turnout rail profiles are collected and analysed, and target grinding profiles are optimised via Genetic Algorithm-Back-Propagation-Non-dominated Sorting Genetic Algorithm III (GA-BP-NSGA III), a vehicle-turnout dynamic model, Non-Uniform Rational B-Spline (NURBS) theory, and grinding vehicle efficiency. A Rail Grinding Strategy Model (RGSM) is developed, driven by Newton-Raphson-based Optimizer (NRBO), Transformer, Bidirectional Gated Recurrent Unit (BiGRU), and Kernel Extreme Learning Machine (KELM), providing real-time precise strategies to guide field work. Research findings indicate that rail wear causes varying degrees of profile deviation at different measurement points, resulting in uneven turnout rail surfaces and significant fluctuations in nominal equivalent conicity. These deviations may lead to excessive CLA and swaying. The optimised rail profiles not only reduce grinding depth effectively but also maintain favourable wheel–rail contact interaction. The RGSM demonstrates high accuracy and strong generalisation capabilities, achieving 96% training accuracy and 97% testing accuracy. After application, rail profile deviations are controlled within 0.2 mm of the optimised profile, resolving CLA issues, enhancing grinding efficiency and vehicle operational stability.
Keywords:
Railway turnout maintenance
rail grinding strategy
rail wear
hybrid intelligent algorithms
wheel-rail contact optimisation
Journal
V
IF:
3.9
Papers:
3.1K
Citations:
8.9K