Study on relations among shield tunneling parameters based on the BP nerve network (II)

Study on relations of shield tunneling parameters in complex ground conditions is preformed using the mathematical statistics and the regression analysis method in the finished work. Results show there is no obvious rule among shield tunneling parameters basically presenting irregular distribution. More advanced method is needed to master relations among the shield tunneling parameters including the rotation of the cutter head, the advance speed, the torque of the cutter head, the total thrust of the shield and the penetration depth. In the paper, the Artificial Neural Networks is introduced to deal with the shield tunneling parameters in complex ground conditions. Being the simple structure and the reliable accommodation, the Back propagation Neural Network contains the essence part of the nerve network, and finds wide use in the practical engineering. Based on the intelligent nerve network, the nonlinear relation among shield tunneling parameters and the input and output of the multi dimensional shield tunneling parameters are realized by means of sample training and learning to the nerve network structure. Computation results show it is feasible to set up quantitative relations among tunneling parameters using BP nerve network. Especially to the shield tunneling in complex conditions, a few of tunneling parameters can be forecast effectively.