基于原位试验的高速铁路中低压缩性土智能识别方法研究  

Research on Intelligent Identification Method of Medium-Low Compressible Soil for High-speed Railway Based on in-situ Test

作　　者：秦尚林[1] 戴张俊[1] 陈善雄[1] 余雷 王祥[3] 王亚飞[3] 胡耀芳 QIN Shanglin;DAI Zhangjun;CHEN Shanxiong;YU Lei;WANG Xiang;WANG Yafei;HU Yaofang(State Key Laboratory of Geomechanics and Geotechnical Engineering,Institute of Rock and Soil Mechanics,Chinese Academy of Sciences,Wuhan 430071,China;China Railway Economic and Planning Research Institute Co.,Ltd.,Beijing 100038,China;China Railway Siyuan Survey and Design Group Co.,Ltd.,Wuhan 430063,China)

机构地区：[1]中国科学院武汉岩土力学研究所岩土力学与工程国家重点实验室,武汉430071 [2]中国铁路经济规划研究院有限公司,北京100038 [3]中铁第四勘察设计院集团有限公司,武汉430063

出　　处：《铁道标准设计》2023年第7期12-18,共7页Railway Standard Design

基　　金：铁道部科技研究开发计划项目(2010G003-F);中铁第四勘察设计院集团有限公司科研课题(2010K29)。

摘　　要：中低压缩性土在我国广泛分布,是高铁路基主要的承载地层。其承载变形快速稳定的特征对高铁路基变形控制有积极意义,因此,中低压缩性土智能识别对高铁路基设计、施工具有重要意义。针对高速铁路路基勘察设计中中低压缩性土的快速、智能识别问题,通过大量的现场原位试验数据,择优确定了以标贯试验、静力触探、载荷试验等原位测试结果为智能判别指标,建立中低压缩性土模糊推演模式,构建高铁中低压缩性土网络预测模型,形成基于现场原位测试的中低压缩性土快速智能识别方法,并通过不同算法进行网络训练和工程预测。结果表明,网络预测结果与实际测试结果整体上吻合度均较高,共轭梯度法相对梯度下降法计算效率明显提高,实现了中低压缩性土原位快速识别与测定,为高铁中低压缩性土路基设计、施工、评估等环节提供了重要依据。Medium-low compressible soil is widely distributed in China and is the main bearing stratum of high-speed railway subgrade.Its rapid and stable bearing deformation is of positive significance to the deformation control of high-speed railway subgrade.Therefore,the intelligent identification of medium-low compressible soil is of great significance to the design and construction of high-speed railway subgrade.Aiming at the problem of rapid and intelligent identification of medium-low compressible soil in the survey and design of high-speed railway subgrade,this paper selects the in-situ test results such as standard penetration test,static cone penetration test and load test as the intelligent discrimination indexes from a large number of field in-situ test data,establishes the fuzzy deduction model of medium-low compressible soil,and constructs the network prediction model of medium-low compressible soil in high-speed railway,forms a fast and intelligent identification method of medium-low compressible soil based on in-situ test,and conducts network training and engineering prediction with different algorithms.The results show that the overall consistency between the network prediction results and the actual test results is high.Compared with the gradient descent method,the calculation efficiency of the conjugate gradient method is significantly improved,which realizes the in-situ rapid identification and measurement of medium-low compressible soil,and provides an important basis for the design,construction and evaluation of medium-low compressible soil subgrade of high-speed railway.

关 键 词：高速铁路 中低压缩性土 原位试验 神经网络 智能识别 

分 类 号：U238[交通运输工程—道路与铁道工程] U213.1