Tweet Share

• Send by email

  Send by email

  

  LSTM based slurry pressure prediction model in high clogging potential ground

  The safety of shield tunneling largely depends on the choice of its control parameter, which is currently decided by human operators empirically. While easy to implement, such practice is vulnerable to human misjudgment and can cause severe consequences. From a practical perspective, it’s therefore beneficial to have a model capable of predicting the shield behavior given operation and the changing geology. In this paper, we propose such a model based on deep learning and Kriging interpolation: More specifically, long-short term memory (LSTM) is employed to conducted to extract features in the time dimension. Kriging method is employed to transform the borehole geological data into sequential geological data according to the shield machine positions. The difference of slurry pressure (SPdifference) in the working chamber and excavation chamber is taken as the output in the case study of Nanning metro, which is divided into five classes according to its value. The LSTM based prediction model of SPdifference classification can achieve a prediction accuracy of 90%. Meanwhile, the influence of tunneling parameters on model performance is also discussed.

  Recipient :

  Name ^* :

  E-mail address ^* :

  ^* Required fields

  Send   or  Cancel

• Print