快速连接件盾构隧道上浮机器学习预测研究

Research on Machine Learning Prediction of Floating of Shield Tunnel with Quick Connectors

摘要:

目前国内快速连接件盾构隧道较少，由于其管片上浮规律异于传统螺栓连接隧道，故需探究上浮原因以便于施工控制。基于南京某快速连接件隧道工程，对施工数据进行收集与整理，通过多种机器学习的方法对管片上浮量进行预测与缺失值填充，并采用决定系数（R2）和均方根误差（RMSE）检验模型效果。结果表明，南京某快速连接件隧道中，俯仰角、总推进力、盾尾间隙（下-上）对管片上浮的影响较大。机器学习模型可有效预测成型管片上浮量、补充上浮缺失值，为相似工程施工提供上浮预测的依据。

Abstract:

Due to the limited number of shield tunnels with quick connectors in China, the segment flotation law differs from that of traditional bolted tunnels. It is necessary to explore the reasons for segment flotation to facilitate construction control. Based on a quick connector tunnel project in Nanjing, the construction data are collected and sorted out. Various machine learning methods are employed to predict and fill missing values of segment floating. And the effectiveness of the model is evaluated by using the coefficient of determination (R2) and root mean square error (RMSE). The results indicate that in a quick connector tunnel in Nanjing, the pitch angle, total thrust force and shield tail gap (vertical) have a significant impact on the segment floating. The machine learning model can effectively predict the segment flotation and supplement missing floating values, which provides a basis for floating prediction in similar engineering constructions.