论文部分内容阅读
为了在盾构机隧道掘进过程中控制土仓压力和地表变形,提出了基于遗传算法的盾构机土仓压力控制模型辨识方法。为了验证控制模型辨识方法的有效性,建立了盾构机土仓压力控制实验台。该实验台由盾构机推进系统、螺旋输送机排土系统、土仓压力与螺旋输送机转速监测系统组成。用于系统模型参数辨识的输入和输出为螺旋输送机的转速和土仓压力。系统辨识的目标函数定义为实验观测的土仓压力与模型计算值残差平方和最小,而模型计算的土仓压力是系统被辨识参数的函数。为了解决参数估计中的局部极小值问题,采用了改进的十进制遗传算法。实验结果表明,与最小二乘法和梯度搜索方法相对比,遗传算法得到的模型参数具有更高的预测精度,预测的土仓压力与观测值吻合的较好。
In order to control the pressure of silo and the deformation of the ground during tunneling of shield tunneling machine, a method of identifying the pressure control model of shield silo based on genetic algorithm is proposed. In order to verify the validity of the control model identification method, a pressure control experiment bench for shield machine is established. The experimental platform by the shield machine propulsion system, screw conveyor drainage system, soil pressure and screw conveyor speed monitoring system. Inputs and outputs for system model parameter identification are screw conveyor speed and earth pressure. The objective function of system identification is defined as the minimum square sum of residuals of experimentally observed soil reservoir pressure and model calculation, and the pressure of soil bin calculated by model is a function of the system identification parameters. In order to solve the problem of local minimum in parameter estimation, an improved decimal genetic algorithm is adopted. The experimental results show that, compared with the least square method and the gradient search method, the model parameters obtained by genetic algorithm have higher prediction accuracy, and the predicted soil bin pressure is in good agreement with the observed value.