基于IFATA-KELM软岩隧道变形预测方法研究  

Deformation Prediction of Soft Rock Tunnel Based on IFATA-KELM Model

作　　者：郭栋梁 GUO Dongliang(The Third Engineering Co.,Ltd of China Railway 18th Bureau Group,Zhuozhou 072750,Hebei,China)

机构地区：[1]中铁十八局集团第三工程有限公司,河北涿州072750

出　　处：《广东交通职业技术学院学报》2025年第2期33-38,共6页Journal of Guangdong Communication Polytechnic

摘　　要：为实现对软岩隧道变形的准确预测,提出了一种改进海市蜃楼优化算法(IFATA)优化核极限学习机(KELM)的软岩隧道变形预测方法。针对海市蜃楼优化算法(FATA)存在局部停滞现象,融合佳点集初始化、自适应莱维飞行传播和量子t分布变异策略提出了IFATA算法,提高了算法的寻优性能,进一步将IFATA应用于优化KELM模型参数,以此构建了IFATA-KELM软岩隧道变形预测模型。以在建某软岩隧道工程为例,验证了IFATA-KELM模型的有效性和适用性,结果表明:对于软岩隧道变形问题,IFATA-KELM模型预测精度优于BP、KELM、FATA-KELM模型,这也表明了IFATA-KELM预测模型具有较强的工程适用性,为隧道变形预测提供了新的途径。To achieve accurate prediction of deformation in soft rock tunnels,an improved fata morgana algorithm(IFATA)is proposed to optimize the kernel extreme learning machine(KELM)for predicting deformation in soft rock tunnels.The IFATA is proposed to address the issue of local stagnation in the fata morgana algorithm(FATA).It incorporates superior point set initialization,adaptive Levy flight propagation,and a quantum t-distribution mutation strategy.This algorithm enhances optimization performance and further applies IFATA to optimize KELM model parameters,thereby constructing an IFATA-KELM soft rock tunnel deformation prediction model.Taking a soft rock tunnel engineering project under construction as an example,the effectiveness and applicability of the IFATA-KELM model were verified.The results showed that for deformation problems in soft rock tunnels,the prediction accuracy of the IFATA-KELM model was superior to that of the BP,KELM and FATA-KELM models.This also indicates that the IFATA-KELM prediction model has strong engineering applicability and provides a new approach for predicting tunnel deformations.

关 键 词：软岩隧道 变形预测 海市蜃楼优化算法 核极限学习机 

分 类 号：U456.3[建筑科学—桥梁与隧道工程]