基于Transformer的人工神经网络城轨列车车轮磨耗预测模型构建
汤新强1
作者信息
- 1. 重庆市铁路集团产业发展有限公司,重庆 401120
- 折叠
摘要
针对目前已有地铁列车车轮磨耗预测模型,在大量车辆车轮廓型数据支撑下,分析车轮磨耗规律,如廓形参数及等效锥度等因素与踏面磨耗、轮缘磨耗之间的关系,研究各参数与车轮磨耗之间的相关性.采用数据驱动方式,构建一种基于Transformer的人工神经网络,用于建立车轮磨耗预测模型.试验结果表明,基于数据驱动的车轮磨耗预测网络能够实现车轮磨耗的准确预测,可用于预测车轮使用寿命,减少相关成本,提高经济效益.
Abstract
Based on the existing prediction models for wheel wear of the metro train,supported by a large amount of wheel profile data,this paper analyzes the law of wheel wear,such as the relationship between profile parameters and equivalent taper with tread wear and wheel flange thickness wear,and studies the correlation between each parameter and wheel wear.Using a data-driven approach,a Transformer based artificial neural network is constructed to estab-lish a wheel wear prediction model.The experimental results show that the data-driven wheel wear prediction network can achieve accurate prediction of wheel wear,which can be used to predict the service life of wheels,reduce related costs,and improve economic benefits.
关键词
地铁动车组/车轮磨耗/预测模型/Transformer/人工神经网络Key words
Metro EMU/Wheel Wear/Prediction Model/Transformer/Artificial Conicity引用本文复制引用
出版年
2024
NETL
NSTL
万方数据