Tweet Share

• Send by email

  Send by email

  

  Application of Artificial Neural Network for Prediction of Surface Settlement During Shield TBM

  Surface settlement induced by shield TBM tunnel excavation in shallow soft ground can threaten the stability of adjacent structures in urban areas. Controlling the ground surface settlements occurred during tunneling is one of the key issues for tunneling engineers. Numerous researches for the prediction of surface settlement have been attempted along with analytical, empirical and/or numerical analysis methods. However, these predictions have intrinsic limitations that they do not completely consider all of the key aspects of geotechnical properties, various geometric proflse and operating conditions of the TBM machine at the site. In this paper, the applicability of the artifcial neural network (ANN) to predicting the ground surface settlement was scrutinized by implementing the settlement data measured at a tunnel site in Hongkong. The tunnel was constructed by a slurry shield TBM machine and goes through an urban area. The geometric characteristic of twin tunnels is dominant at several sections along the tunnel. Thirteen different parameters representing the tunnel geometry, TBM operating conditions, and geological conditions were collected and analyzed. The optimum structure of neural network was investigated through a parametric study. Along with the optimum structure, the infuence of the twin tunnel parameter on the ground surface settlement was evaluated in this study.

  Recipient :

  Name ^* :

  E-mail address ^* :

  ^* Required fields

  Send   or  Cancel

• Print