Quantitative Fault Diagnosis of Railway Wheel Polygonization Based on Fractional-Order-Enhanced Transform-Domain Multi-Instance Learning and FiLM Modulation
4 affiliations
- 1School of Transportation and Electrical Engineering, Hunan University of Technology, Zhuzhou 412007, China
- 2National Key Laboratory for Heavy-Duty High-Power Electric Locomotives Zhuzhou, Zhuzhou 412001, China
- 3CRRC Zhuzhou Locomotive Co., Ltd., Zhuzhou 412001, China
- 4Zhuzhou Times New Material Technology Co., Ltd., Zhuzhou 412007, China
Abstract
Wheel polygonization is a typical out-of-roundness fault in high-speed trains, and multi-order roughness spectra can quantitatively describe its fault condition. Variations in operating speed, wheel diameter, and reprofiling state cause differences in axle-box vibration responses, making the diagnosis of multi-order roughness amplitudes difficult. A fractional-order-enhanced transform-domain multi-instance learning method with FiLM modulation is proposed for quantitative fault diagnosis of wheel polygonization. First, axle-box vibration acceleration, operating speed, and measured roughness data are jointly processed to construct multi-instance samples containing angular-domain time series, order-domain spectra, vibration statistics, and operating-condition information. Second, a model integrating transform-domain multi-instance learning and FiLM modulation is constructed, combining a state-adaptive multi-regression mechanism with order-aware correlation mapping to achieve quantitative diagnosis of roughness amplitudes across orders 1–32. Finally, the Grünwald–Letnikov fractional-order difference is introduced into the loss function to enhance the model’s ability to learn inter-order correlations and overall spectral-shape features, thereby reducing diagnostic errors in multi-order roughness amplitudes. Experimental results show that the proposed method achieves an RMSE of 1.157 dB, 60.5% lower than that of CNN-LSTM, the best-performing comparison method, providing a quantitative diagnostic method for high-speed train wheel condition assessment and reprofiling evaluation.
Related
- Sampling and Density Theorems for the Fractional Opdam-Cherednik Transform
- A positive-definite Fourier transform, with applications to fractional Brownian motions and Bayesian probabilistic numerics
- Solving Fractional Partial Differential Equations Using the Double Natural Transform Decomposition Method
- Connections Between Quadratic Transform for Fractional Programming and Schur Complement
- Quantitative Stability for Fractional Yamabe minimizers
BibTeX
@article{li2026quantitative,
title = {{Quantitative Fault Diagnosis of Railway Wheel Polygonization Based on Fractional-Order-Enhanced Transform-Domain Multi-Instance Learning and FiLM Modulation}},
author = {Tao Li and Xuyuan Gao and Zhoncheng Jiang and Wei Liu and Dengke Li and Xuancheng Yuan and Bowen Yao},
journal = {Fractal and Fractional},
year = {2026},
month = nov,
doi = {10.3390/fractalfract10110774},
url = {https://www.mdpi.com/2504-3110/10/11/774},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Quantitative Fault Diagnosis of Railway Wheel Polygonization Based on Fractional-Order-Enhanced Transform-Domain Multi-Instance Learning and FiLM Modulation
AU - Tao Li
AU - Xuyuan Gao
AU - Zhoncheng Jiang
AU - Wei Liu
AU - Dengke Li
AU - Xuancheng Yuan
AU - Bowen Yao
PY - 2026
DA - 2026/11/05
JO - Fractal and Fractional
DO - 10.3390/fractalfract10110774
UR - https://www.mdpi.com/2504-3110/10/11/774
AB - Wheel polygonization is a typical out-of-roundness fault in high-speed trains, and multi-order roughness spectra can quantitatively describe its fault condition. Variations in operating speed, wheel diameter, and reprofiling state cause differences in axle-box vibration responses, making the diagnosis of multi-order roughness amplitudes difficult. A fractional-order-enhanced transform-domain multi-instance learning method with FiLM modulation is proposed for quantitative fault diagnosis of wheel polygonization. First, axle-box vibration acceleration, operating speed, and measured roughness data are jointly processed to construct multi-instance samples containing angular-domain time series, order-domain spectra, vibration statistics, and operating-condition information. Second, a model integrating transform-domain multi-instance learning and FiLM modulation is constructed, combining a state-adaptive multi-regression mechanism with order-aware correlation mapping to achieve quantitative diagnosis of roughness amplitudes across orders 1–32. Finally, the Grünwald–Letnikov fractional-order difference is introduced into the loss function to enhance the model’s ability to learn inter-order correlations and overall spectral-shape features, thereby reducing diagnostic errors in multi-order roughness amplitudes. Experimental results show that the proposed method achieves an RMSE of 1.157 dB, 60.5% lower than that of CNN-LSTM, the best-performing comparison method, providing a quantitative diagnostic method for high-speed train wheel condition assessment and reprofiling evaluation.
SN - 2504-3110
ER -