基于PSO-LSSVM算法的隧道掘进机掘进参数预测方法

Tunnel Boring Machine Tunneling Parameters Prediction Method Based on PSO-LSSVM Algorithm

为了规避隧道掘进机(tunnel boring machine,TBM)掘进参数人为设定的主观性,提出了一种基于粒子群-最小二乘支持向量机算法(PSO-LSSVM)的TBM掘进参数预测方法。通过从海量TBM工程掘进数据中探寻参数变化规律,降低了TBM主司机设定掘进参数的主观性,辅助其合理选择掘进参数,有利于提高掘进效率、规避工程风险,经实验和工程数据验证,PSO-LSSVM算法通过对样本粒子全局迭代寻优来优化参数,提升了预测算法泛化能力和预测精度,对推力、扭矩和推进速度参数预测数值偏差满足要求,可辅助指导主司机设定掘进参数。

In order to avoid the subjectivity of manual setting of TBM(tunnel boring machine)tunneling parameters,a TBM tunneling parameter prediction method based on PSO-LSSVM algorithm was proposed.By exploring the rule of parameter change from the massive TBM engineering tunneling data,the subjectivity of TBM main driver in setting tunneling parameters was reduced,and the reasonable selection of tunneling parameters was assisted,which is conducive to improving tunneling efficiency and avoiding engineering risks.Through experiments and engineering data verification,PSO-LSSVM algorithm optimizes parameters through global iterative optimization of sample particles,which improves the generalization ability and prediction accuracy of prediction algorithm The deviation of the predicted values of torque and propulsion speed parameters meets the requirements,which can assist the main driver in setting tunneling parameters.