高铁路基动力稳定性能评价方法与应用研究  

Evaluation Method for Dynamic Stability of High-speed Railway Subgrade and Its Application

作　　者：邓志兴 谢康 王武斌[2] 徐林荣[1] 李泰灃 苏谦[4] 王迅[4] 刘宝 DENG Zhixing;XIE Kang;WANG Wubin;XU Linrong;LI Taifeng;SU Qian;WANG Xun;LIU Bao(School of Civil Engineering,Central South University,Changsha 410075,China;National Engineering Research Center of Geological Disaster Prevention Technology in Land Transportation,Southwest Jiaotong University,Chengdu 611731,China;China Railway Engineering Research Institute,China Academy of Railway Sciences Corporation Limited,Beijing 100081,China;School of Civil Engineering,Southwest Jiaotong University,Chengdu 610000,China;China Railway Design Corporation,Tianjin 300308,China)

机构地区：[1]中南大学土木工程学院,湖南长沙410075 [2]西南交通大学陆地交通地质灾害防治技术国家工程研究中心,四川成都611731 [3]中国铁道科学研究院集团有限公司铁道建筑研究所,北京100081 [4]西南交通大学土木工程学院,四川成都610000 [5]中国铁路设计集团有限公司,天津300308

出　　处：《铁道学报》2025年第3期126-134,共9页Journal of the China Railway Society

基　　金：国家自然科学基金(U2268213,52478473);中国国家铁路集团有限公司科技研究开发计划(J2021G002)。

摘　　要：目前缺少一种快速、超前、综合的高速铁路路基动力稳定性能评价方法,限制了路基动响应和累积变形数据的利用程度,易导致对路基的服役性能评价不充分、甚至在运营期间劣化等问题。因此,基于激振测试和机器学习(ML)提出一种高铁路基动力稳定性能评价方法。首先,利用大型激振器模拟高速列车荷载,以快速获得高速铁路路基动力稳定性能评价指标,包括动响应(动应力、振动加速度)和累积变形;其次,基于优选的最佳ML模型对累积变形进行超前预测,探索累积变形的未来发展趋势;最后,综合路基动响应评估和路基累积变形评估,实现高速铁路路基动力稳定性能评价,并依托杭绍台高速铁路路基试验段验证该方法的有效性。结果表明:激振试验过程中,路基最大动应力、振动加速度和累积变形分别为42.5 kPa、8.496 m/s^(2)、3.26 mm,均小于各自阈值50 kPa、10 m/s^(2)和15 mm;优选得到性能最佳的LSTM模型,并基于LSTM模型得到累积变形预测值稳定于3.27 mm,表明累积变形满足超前评估要求。综合路基动响应和路基累积变形评估结果,得到该高速铁路路基动力稳定性能较好。相关研究成果为高速铁路路基结构动力稳定性能评价提供新的思路,也可为高速铁路智能运维提供超前参考。The current scarcity of a rapid,advanced,and comprehensive assessment method for the dynamic stability performance of high-speed railway subgrades hampers the utilization of dynamic response and cumulative deformation data of the subgrade,often resulting in inadequate service performance assessment and,at times,subgrade deterioration during the operational phase.Therefore,a method for assessing the dynamic stability performance of high speed railway subgrades was proposed based on excitation test and machine learning(ML).Firstly,large exciters were used to simulate high-speed train loads to quickly obtain the assessment metrics for the dynamic stability performance of high-speed railway subgrades,including the dynamic response(dynamic stress,vibration acceleration)and cumulative deformation.Secondly,advanced predictions were made for the cumulative deformation based on the selected optimal ML model,to explore the future development trend of the cumulative deformation.Finally,the assessment of dynamic response and the assessment of cumulative deformation were integrated to achieve the assessment of the dynamic stability performance of the high-speed railway subgrades.The effectiveness of the method was verified by relying on the subgrade test section of Hangzhou—Shaoxing—Taizhou High-speed Railway.The results show that during the excitation test,the maximum dynamic stress,vibration acceleration and cumulative deformation of the subgrade are 42.5 kPa,8.496 m/s^(2) and 3.26 mm,respectively,all within the respective limits of 50 kPa,10 m/s^(2) and 15mm.The LSTM model obtained from the preferred selection displays the best performance.The predicted value of cumulative deformation obtained based on the LSTM model is stabilized at 3.27 mm,showing that the cumulative deformation meets the requirements of advanced assessment.The results of the comprehensive assessment of the dynamic response and the cumulative deformation of the subgrade show the good dynamic stability of the subgrade of this high-speed railway.The re

关 键 词：高速铁路路基 激振测试 机器学习 动力响应 累积变形 

分 类 号：TU413.5[建筑科学—岩土工程] U213.1[建筑科学—土工工程]