长江科学院院报 ›› 2022, Vol. 39 ›› Issue (7): 59-65.DOI: 10.11988/ckyyb.20210276

• 工程安全与灾害防治 • 上一篇    下一篇

基于变量筛选优化极限学习机的混凝土坝变形预测模型

曹恩华1,2,3, 包腾飞1,2,3,4, 胡绍沛3, 袁荣耀3, 鄢涛1,2,3   

  1. 1.河海大学 水文水资源与水利工程科学国家重点实验室,南京 210098;
    2.河海大学 水资源高效利用与工程安全国家工程研究中心,南京 210098;
    3.河海大学 水利水电学院,南京 210098;
    4.三峡大学 水利与环境学院,湖北 宜昌 443002
  • 收稿日期:2021-03-27 修回日期:2021-08-20 出版日期:2022-07-01 发布日期:2022-07-25
  • 通讯作者: 包腾飞(1974-),男,湖北黄冈人,教授,博士,博士生导师,主要从事水工结构及岩土工程的安全监控、光纤传感器在结构健康监测中的应用研究。E-mail:baotf@hhu.edu.cn
  • 作者简介:曹恩华(1993-),男,江苏东海人,博士研究生,研究方向为水工结构安全监控。E-mail:caoenhua@hhu.edu.cn
  • 基金资助:
    国家重点研发计划项目(2018YFC1508603,2016YFC0401601);国家自然科学基金项目(51579086,51739003)

A Deformation Prediction Model for Concrete Dam Based on Extreme Learning Machine Optimized by Variable Selection

CAO En-hua1,2,3, BAO Teng-fei1,2,3,4, HU Shao-pei3, YUAN Rong-yao3, YAN Tao1,2,3   

  1. 1. State Key Laboratory of Hydrology-Water Resources and Hydraulic Engineering,Hohai University,Nanjing 210098,China;
    2. National Engineering Research Center of Water Resources Efficient Utilization and Engineering Safety,Hohai University,Nanjing 210098,China;
    3. College of Water-conservancy and Hydropower,Hohai University,Nanjing 210098,China;
    4. College of Hydraulic & Environmental Engineering,Three Gorges University,Yichang 443002,China
  • Received:2021-03-27 Revised:2021-08-20 Online:2022-07-01 Published:2022-07-25

摘要: 传统的统计模型泛化能力较弱且容易引入高维变量,这将对基于神经网络预测模型的输出结果产生负面影响,同时增加了过拟合风险。因此,有必要建立一个具有适当维度的数据驱动模型,以实现对大坝变形的准确监控。选用极限学习机(ELM)作为基础预测模型,提出基于平均影响值MIV-ELM模型的变量筛选法,以消除初始变量集中的冗余信息,从而降低模型复杂度,提高预测精度。分析结果表明,与传统预测模型相比,HST-MIV-ELM不仅具有最高的预测精度和预测性能,同时也有较强的可拓展性,为大坝安全监控系统的构建提供了可靠的理论基础。

关键词: 混凝土坝变形预测, 变量筛选, 极限学习机, 平均影响值, 反向逐变量剔除法

Abstract: Traditional statistical models are of weak generalization capability and are prone to introduce high-dimensional variables,which will negatively affect the output of neural network-based prediction models and increase the risk of overfitting.It is necessary to build a data-driven model with appropriate dimensionality to accomplish accurate monitoring of dam deformation.In this paper,extreme learning machine(ELM)is selected as the base prediction model,and a variable selection method based on mean impact value(MIV)-ELM model is proposed to eliminate redundant information in the initial variable set,thus reducing the model's complexity and improving the prediction accuracy.Analysis results demonstrate that compared with traditional prediction models,HST-MIV-ELM not only has the highest prediction accuracy and robustness,but also has strong scalability.The study provides a reliable theoretical basis for the construction of dam safety monitoring system.

Key words: deformation prediction for concrete dam, variable selection, extreme learning machine, mean impact value, reverse variable-by-variable elimination method

中图分类号: