复杂地质条件下的隧道大变形组合预测研究  

Study on mechanism analysis and combine dprediction of larged eformation of tunnel under complex geological conditions

作　　者：贺华刚[1] He Huagang(Chongqing Technology and Business Institute,Chongqing 401520,China)

机构地区：[1]重庆工商职业学院,重庆401520

出　　处：《山西建筑》2020年第6期124-126,共3页Shanxi Architecture

基　　金：重庆工商职业学院基金项目(NDYB2019-18)。

摘　　要：隧道大变形是隧道工程中的常见病害之一,严重威胁施工安全及工程进度,对其研究具有重要意义。以BP神经网络和支持向量机为基础,利用遗传算法和粒子群算法优化其结构参数,以实现单项预测模型优化,再构建线性组合模型和非线性组合模型,实现了隧道大变形的组合预测。实例分析表明:组合预测模型较单项预测模型具有更高的预测精度和稳定性,且以BP神经网络权值法的组合效果最优。通过研究,为大变形防治及变形规律评价提供了借鉴和参考,具有一定的实用价值。Large deformation of tunnel is one of the common diseases in tunnel engineering.It seriously threatens construction safety and engineering progress,and has great significance for its research.Based on the BP neural network and support vector machine,the genetic algorithm and particle swarm optimization are used to optimize the structural parameters,so as to achieve the optimization of the single prediction model,and then to build a linear combination model and nonlinear model.Combined model,the combined prediction of tunnel large deformation is realized.The case analysis shows that the combined prediction model has a higher prediction accuracy and stability than the single prediction model,which validates the effectiveness of the combined prediction method in this paper and the weight value of BP neural network.The combination of the method is the best.This study provides reference and reference for the prevention and control of large deformation and the evaluation of deformation law,and has certain practical value.

关 键 词：隧道大变形 变形机理 BP神经网络 支持向量机 

分 类 号：U457.2[建筑科学—桥梁与隧道工程]