Multiple-point Earth Pressure Balance Control Strategy For Shield Tunneling Based On Neural Network

With the increasing urban intensity, the ground space for use is reduced sharply. It is an urgent priority to create and make use of the underground space. Shield technology has become the best way for tunnel building due to its all kinds of advantages.During tunneling, poor earth pressure balance in the capsule will directly lead to the instability of the excavated surface and ground surface accidents, such as subsidence and swell. Problems emerge in the complex construction process, for example, the work condition and geological condition fluctuate significantly, and the relationships between tunneling parameters cannot be showed in one accurate mathematical model due to the strong coupling. The excavation construction has to be accomplished manually through human experience and the accuracy and quality of construction remains uncertain. So in order to achieve the automatic control of the pressure, it's of great theoretical and practical significance to set up the earth pressure balance control model with artificial intelligence methods. The extant control models don't take the spatial distribution of soil pressure into consideration and can only achieve the pressure control of a single point. This makes them hard to guide the actual construction. Based on the above content, this thesis did the following researches:First, based on the extensive researches on the international shield earth pressure balance control theories, this paper provided the theoretical basis for the earth pressure balance control according to the mechanism analysis. Second, this paper came up with an optimal control strategy based on BP neural network to achieve the shield capsule multiple-point earth pressure balance control by adjusting the screw conveyor speed. The control model took the pressure at current time, the driving force, the driving speed, the screw conveyor's speed at current time and the screw conveyor's speed at next time as the inputs, and took the pressure at next time as the output. The control method took the minimum error between the multiple-point pressures and the settled ones as the goal to find out the optimal value of the speed of the screw conveyor. Finally, a simulation was carried out and the results showed that this control strategy can meet the requirement of the earth pressure balance control. It is more reliable than the human experience and has higher accuracy compared with the traditional single-point model. The strategy can be put into practical use as a reference and guide for tunneling.