基于改进自编码解码网络的轨道不平顺评价方法  

Evaluation Method for Track Irregularity Based on Improved Deep Autoencoder and Decoder Network

作　　者：杨建伟[1] 王小慧 刘佩珊 杨飞 王金海[1] 孙培文 YANG Jianwei;WANG Xiaohui;LIU Peishan;YANG Fei;WANG Jinhai;SUN Peiwen(Beijing Key Laboratory of Performance Guarantee on Urban Rail Transit Vehicles,Beijing University of Civil Engineering and Architecture,Beijing 100044,China;School of Civil Engineering and Transportation,Guangzhou University,Guangzhou 510006,China;Infrastructure Inspection Research Institute,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China)

机构地区：[1]北京建筑大学城市轨道交通车辆服役性能保障北京市重点实验室,北京100044 [2]广州大学土木与交通工程学院,广东广州510006 [3]中国铁道科学研究院集团有限公司基础设施检测研究所,北京100081

出　　处：《铁道学报》2025年第2期131-144,共14页Journal of the China Railway Society

基　　金：国家自然科学基金(52272385,52205083);中国博士后科学基金(2024M750619);博士研究生科研能力提升项目(DG2023019)。

摘　　要：目前无砟轨道平顺性的评价方法只适用于中短波长,难以满足大范围动态复杂线路的变化需求。如果仅依靠单一波长轨检数据评价轨道状态,会导致其他波长局部波动或潜在病害不能被有效识别的现象。通过挖掘不同波长下的轨道不平顺信息,将深度学习方法与现有评价方法相融合,提出改进的自编码解码高斯生成对抗神经网络(DAGGAN)及自适应轨道质量指数方法(DAE_TQI),实现对轨道状态的实时监控。为验证方法的准确性和有效性,将其与其他方法进行对比,验证DAGGAN模型在识别单几何不平顺时的有效性,同时DAE_TQI方法可综合评价无砟轨道状态,结果可作为高速铁路轨道养护维修的理论依据。Current evaluation methods for ballastless track irregularity are only applicable to medium-short wavelengths,making it difficult to meet the changing needs of large-scale and dynamically complex track networks.Relying soly on track inspection data of a single wavelength to evaluate track status can lead to the phenomenon where local fluctuations in other wavelengths or potential defects may not be effectively identified.In this paper,combining deep learning methods with existing evaluation methods,a deep autoencoder and decoder Gaussian generative adversarial network(DAGGAN)and an adaptive track quality index method(DAE_TQI)were proposed by mining data information at different wavelengths to achieve real-time monitoring of track status.To verify the accuracy and effectiveness of the method,a comparison was made with other methods.The results show the effectiveness of the DAGGAN model in identifying local irregularities.Simultaneously,the DAE_TQI method can comprehensively evaluate the status of ballastless tracks,The results can be used as the theoretical basis for high-speed rail track maintenance.

关 键 词：轨道不平顺 轨道质量指数 自编码解码网络 高斯对抗网络 自适应轨道质量指数 

分 类 号：U216[交通运输工程—道路与铁道工程]