基于小波核极限学习机的隧道围岩变形预报  

Study on Tunnel Rock Deformation Prediction Based on Wavelet Kernel Extreme Learning Machine

作　　者：蒋方媛 王艳 汪辉洋 汪磊 高延超[2] JIANG Fangyuan;WANG Yan;WANG Huiyang;WANG Lei;GAO Yanchao(Shenzhen Survey and Research Institute Co.,Ltd.,Shenzhen 518000,China;Chengdu Geological Survey Center,China Geological Survey Bureau,Chengdu 610081,China)

机构地区：[1]深圳市勘察研究院有限公司,广东深圳518000 [2]中国地质调查局成都地质调查中心,四川成都610081

出　　处：《城市勘测》2023年第5期187-191,共5页Urban Geotechnical Investigation & Surveying

基　　金：四川省自然科学基金(23NSFSC0297)。

摘　　要：隧道围岩变形预报可视为一时间序列问题处理,由于围岩变形序列的非线性复杂程度较高,针对该问题,提出一种适用于隧道围岩变形预报的遗传小波核极限学习机模型(GA-ELM-k)。联用核方法和极限学习机算法,保证了极限学习机算法的稳健性,又融合了小波方法所具有的处理非平稳信号的优势。针对小波核函数参数的确定,将遗传算法用于参数寻优。通过黄榜岭隧道围岩变形预报表明,提出的新模型预报准确度优于ARIMA模型和灰色新陈代谢模型,可作为实际工程的一种参考方法。Tunnel surrounding rock deformation forecast can be treated as time series problem,as a result of the deformation sequence of nonlinear complexity is higher,aiming at the problem,put forward a kind of applicable to tunnel surrounding rock deformation forecast model(GA-ELM-k).Combining extreme learning machine and kernel method,the robustness of machine learning algorithm can be ensure,and combines the wavelet method have the advantage of the non-stationary signal processing.For the determination of the wavelet kernel function parameters,the Genetic Algorithm for parameter optimization.Through Huang Bang tunnel surrounding rock deformation prediction show that the prediction accuracy of the new model is better than that of ARIMA model and grey metabolism model,which can be used as a reference method in practical engineering.

关 键 词：隧道工程 变形预测 极限学习机 小波核 遗传算法 

分 类 号：P642[天文地球—工程地质学] TU457[天文地球—地质矿产勘探]