Regulating bulkhead pressure of EPB shield machines through DEM modeling and data mining

Abstract

Proper regulation of the earth pressure on the bulkhead of earth-pressure balanced (EPB) shield tunneling machines is significant to ensure safe construction. This study proposes a procedure for regulating the bulkhead pressure by combining numerical simulations and data mining, and applies the procedure to a metro line constructed in sandy pebble stratum using EPB shield. Firstly, the relationship between the bulkhead pressure and the pressure on the tunnel face is carefully obtained from discrete element modeling, and the required supporting earth pressure is derived by considering the arching effect. Secondly, aided with the machine learning method, a model is constructed for predicting the average bulkhead pressure per ring according to the operational parameters (i.e., the average driving speed and the rotation speed of the screw conveyor). Given the target value of the bulkhead pressure, the optimal values of the operational parameters are obtained from the model. In addition, an autoregressive moving average stochastic process model is developed to monitor the real-time fluctuation of the bulkhead pressure and guide the actions taken in time to avoid dramatic fluctuations. The results indicate that the pressure difference between the tunnel face and the bulkhead is considerable, and the consideration of the arching effect can avoid overestimating the bulkhead pressure. A combination of the machine learning model and the stochastic process model provides a plausible performance in regulating the bulkhead pressure around the target value without dramatic fluctuation.

Keywords

• EPB shield machine
• Bulkhead pressure
• Discrete element method
• Support vector regression
• Autoregressive moving average model