摘要
为了研究盾构机在特定地层条件正常掘进状态下的掘进参数选取和预测,以深圳地铁16号线天健花园站—龙城中路站左线土压平衡盾构施工区间的微风化灰岩地层段为背景,对其掘进参数进行数理统计、Pearson相关性分析及建立预测模型。考虑了施工工艺等因素对数据的影响;对6个主要掘进参数的相关性进行了定量分析;用多元线性回归和随机森林方法对主要掘进参数进行建模预测。研究结果表明,在该地层条件下的推进速度平均值为15.81 mm/min、总推进力平均值为12 973.69 kN等;计算出6个掘进参数间的Pearson相关系数,其中总推进力和刀盘转矩的相关系数r为0.569,存在显著的中等正相关关系;运用相关系数r、平均绝对误差(MAE)和均方根误差(RMSE)3个指标对模型进行对比评估,结果表明随机森林模型预测性能更优。
To research the selection and prediction of boring parameters of shield tunneling machine under normal boring performance in specific stratum conditions.The stratum in the micro-weathered limestone section of the left line of Tianjian Garden Station—Longcheng Middle Road Station of Shenzhen Metro Line 16 was taken as example,relevant statistical analysis and Pearson correlation analysis were conducted.The prediction model was set as well.The effect of factors like construction process to final data was taken into consideration.The research show that the average advancing speed under this stratum formation is 15.81 mm/min,and the average total thrust is 12973.69 kN.The Pearson correlation coefficients of six boring parameters had been calculated,of which the r of total thrust and cutter torque is 0.569,and there is a significant medium positive correlation.The three indexes of r,MAE and RMSE were used to compare and evaluate the models.The random forest model has the best performance.
作者
李锟
田管凤
马宏伟
杨晓林
蒲宝军
江运陈
LI Kun;TIAN Guan-feng;MA Hong-wei;YANG Xiao-lin;PU Bao-jun;JIANG Yun-chen(School of Civil Engineering,Qinghai University,Xining 810016,China;School of Mechanical Engineering,Dongguan University of Technology,Dongguan 523808,China;CREEC(Chengdu)Consulting and Supervision Co.,Ltd.,Chengdu 610031,China;Urban Rail Transit Engineering Company Limited of China Railway 15th Bureau,Luoyang 471002,China)
出处
《科学技术与工程》
北大核心
2021年第9期3814-3821,共8页
Science Technology and Engineering
基金
高等教育“创新强校工程”专项(2017KSYS009)
东莞理工学院高水平人才平台(KCYCXPT2017006)
东莞理工学院博士启动专项
广东省普通高校青年创新人才类项目(2018KQNCX254)。
关键词
土压平衡盾构
微风化灰岩地层
掘进参数
数据预处理
数理统计
Pearson相关性分析
多元线性回归
随机森林
earth pressure balancing shield
micro-weathered limestone section
boring parameters
data preprocessing
mathematical statistics
Pearson correlation analysis
multiple linear regression
random forest
