优化极限学习机在城市轨道交通地表沉降预测中的应用  

Application of Optimized Extreme Learning Machine in Prediction of Surface Subsidence of Urban Rail Transit

作　　者：黎冶 陈铮 董振川 李浩标 张献州 LI Ye;CHEN Zheng;DONG Zhenchuan;LI Haobiao;ZHANG Xianzhou(Faculty of Geosciences and Environmental Engineering,Southwest Jiaotong University,Chengdu 611756,China;Chengdu Institute of Surveying and Mapping,Chengdu 610081,China;National&Local Joint Engineering Laboratory of Safe Space Information Technology for High-Speed Railway Operation,Chengdu 611756,China)

机构地区：[1]西南交通大学地球科学与环境工程学院,四川成都611756 [2]成都市勘察测绘研究院,四川成都610081 [3]高速铁路运营安全空间信息技术国家地方联合工程实验室,四川成都611756

出　　处：《测绘地理信息》2021年第4期60-64,共5页Journal of Geomatics

基　　金：成都市科技项目城市轨道交通基础设备服务期变形多传感器监测大数据处理方法研究(2015-RK00-00218-2F)。

摘　　要：为了确保轨道交通隧道施工作业的正常进行及地表建构筑物的稳定安全,采用遗传算法(genetic algorithm,GA)对极限学习机(extreme learning machine,ELM)模型的输入权值和隐含层偏置进行优化,通过构建GA-ELM模型,实现对轨道交通施工隧道地表断面监测点实测数据进行变形预测研究,并将GA-ELM模型与ELM模型、传统BP(back progation)模型的变形预测结果进行对比分析。实验研究结果表明:优化后的ELM模型预测效果得到很好的改善,证明了GA-ELM预测模型在施工隧道地表沉降预测中的可靠性和适用性。In order to ensure the normal construction of rail transit tunnel and the stability of the ground surface structures,genetic algorithm(GA)is used to optimize the Extreme Learning Machine(ELM)model’s input-weight and bias of hidden layer,and then GA-ELM model is constructed to predict the deformation based on the measured data of the surface section monitoring points of the rail transit construction tunnel,and lastly GA-ELM model,traditional ELM model and BP neural network model are compared and analyzed for deformation prediction results.The result of study shows that the prediction effect of the optimized ELM model is well improved,which proves the reliability and applicability of the GA-ELM model in the prediction of surface subsidence of tunnel construction.

关 键 词：城市轨道交通 施工隧道 极限学习机 遗传算法优化 沉降预测 

分 类 号：P258[天文地球—测绘科学与技术]