中国安全科学学报 ›› 2024, Vol. 34 ›› Issue (6): 57-64.doi: 10.16265/j.cnki.issn1003-3033.2024.06.1734

• 安全工程技术 • 上一篇    下一篇

基于CatBoost-MOEAD的大直径泥水盾构施工多目标预测优化

吴贤国1(), 刘俊1, 苏飞鸣1,**(), 陈虹宇2, 冯宗宝1   

  1. 1 华中科技大学 土木与水利工程学院,湖北 武汉 430074
    2 香港理工大学 建筑与房地产学部,香港 999077
  • 收稿日期:2023-12-08 修回日期:2024-03-21 出版日期:2024-06-28
  • 通讯作者:
    **苏飞鸣(1997—),广西南宁人,男,博士研究生,研究方向为土木工程建造与管理。E-mail:
  • 作者简介:

    吴贤国 (1964—),女,湖北武汉人,博士,教授,主要从事数字工程集成建设关键技术及应用、隧道工程施工与运营安全监控等方面的研究。E-mail:

  • 基金资助:
    国家自然科学基金资助(51378235); 国家自然科学基金资助(71571078); 国家自然科学基金资助(51308240); 国家重点研发计划项目(2016YFC0800208)

Multi-objective prediction optimization for large-diameter slurry shield tunneling construction based on CatBoost-MOEAD

WU Xianguo1(), LIU Jun1, SU Feiming1,**(), CHEN Hongyu2, FENG Zongbao1   

  1. 1 School of Civil and Hydraulic Engineering, Huazhong University of Science and Technology, Wuhan Hubei 430074, China
    2 Department of Building and Real Estate, The Hong Kong Polytechnic University, Hong Kong 999077, China
  • Received:2023-12-08 Revised:2024-03-21 Published:2024-06-28

摘要:

为有效优化盾构施工参数,实现在大直径泥水盾构掘进过程中安全、高效和节能的目标,提出分类助推(CatBoost)和基于分解的多目标进化算法(MOEAD)相结合的混合智能算法;综合考虑盾构施工参数与地质条件,以主要的盾构施工参数为研究对象,选择地表沉降、贯入度和掘进比能为预测和控制目标;优化调控选择的盾构施工参数,并以武汉市轨道交通某号线为例,验证该混合算法的有效性。结果表明:采用CatBoost算法建立的预测模型在大直径泥水盾构上表现出来的预测性能良好,对3个控制目标的拟合精度(R2)均达到0.9以上;预测模型的重要性排序表明:大直径泥水盾构的总推进力和推进速度对地表沉降、贯入度和掘进比能有显著影响;所提出的CatBoost-MOEAD混合智能算法对3个控制目标的优化效果明显,地表沉降、贯入度和掘进比能分别达到12.35%、7.47%和10.70% 的优化幅度,并给出相应盾构施工参数的控制范围。

关键词: 大直径泥水盾构, 分类助推(CatBoost), 基于分解的多目标进化算法(MOEAD), 多目标优化, 地表沉降

Abstract:

To effectively optimize the shield construction parameters and achieve the goals of safety, efficiency, and energy-saving in the large-diameter slurry shield tunneling process, a hybrid intelligent algorithm combining categorical boosting (CatBoost) and decomposition was proposed based on a multi-objective evolutionary algorithm (MOEAD). The main shield construction parameters were set as the major research objects considering shield construction parameters and geological conditions, and the surface settlement, penetration rate, and tunneling-specific energy were determined as the prediction and control objectives. Moreover, the selected shield construction parameters were optimized, and a line of Wuhan rail transit was used to validate the hybrid algorithm performance. The results showed that the proposed CatBoost algorithm had great prediction performance for large-diameter slurry shields with the fitting accuracy (R2) of the three control objectives more than 0.9. The model's importance rank indicated that the total propulsion force and propulsion speed of the large-diameter slurry shield had significant influences on surface settlement, penetration, and tunneling-specific energy. The proposed CatBoost-MOEAD hybrid intelligent algorithm had an obvious optimization effect on the three control objectives, and the optimization ranges of surface settlement, penetration rate, and tunneling-specific energy reached 12.35%, 7.47%, and 10.70%, respectively. Moreover, the control ranges of corresponding shield construction parameters were presented.

Key words: large-diameter slurry shield, categorical boosting (CatBoost), multi-objective evolutionary algorithm based on decomposition (MOEAD), multi-objective optimization, surface settlement

中图分类号: