Font Size: a A A

Research On Deformation Prediction Of Tunnel Surrounding Rock Based On Multi-source Heterogeneous Information

Posted on:2021-12-30Degree:MasterType:Thesis
Country:ChinaCandidate:L H LiFull Text:PDF
GTID:2492306527463384Subject:Architecture and Civil Engineering
Abstract/Summary:
Most of the deformation diseases of tunnel surrounding rocks occur during the construction process,and effective deformation prediction can ensure construction safety and engineering quality.As one of the current research hotspots,various prediction models have emerged at the historic moment,but they also have their own limitations.This paper introduces the Gaussian Process Regression(GPR)theory,discusses the algorithm model and the feasibility of tunnel deformation,and studies the prediction of tunnel deformation.As an underground engineering,there are many factors influencing the deformation of tunnel engineering.It is difficult to ensure that other factors have the same level effect by studying the relationship between a certain type of characteristic information and the stability of surrounding rock.In this paper,one of the research sections of the Qinfeng extra-long tunnel of the Kunming(Langshan)to Chuxiong(Guangtong)highway reconstruction project is used as the engineering background,,and the relationship between the multi-source heterogeneous information such as the geological and natural properties of surrounding rock and the engineering related factors and the tunnel deformation is comprehensively considered.Combining relevant literature and actual experience to perform data mining,evaluate and screen multi-source heterogeneous information,parameter quantification,data standardization,etc.,and combine deformation monitoring data with tunnel monitoring to form sample spatial data.Enter the Matlab matrix calculation and analysis software for training,obtain the optimal hyperparameters through "kernel learning",build a GPR tunnel deformation prediction model,and continuously introduce the latest data through rolling interactive learning of data to learn the tunnel closest to the palm face Deformation development law,to achieve the purpose of reducing errors,optimizing models,and conforming to reality,so as to make predictions about the current deformation level of the palm face.By comparing the predicted value of the model output with the actual convergence of the surrounding rock deformation,analyzing the magnitude and cause of the error,and evaluating the applicability of the Gaussian process regression model in the tunnel project,the results show that the accuracy of the prediction model meets the engineering requirements.Furthermore,the model is applied to the tunnel construction permit change mechanism,and the support parameters are adjusted before the deformation and disease are enlarged to realize the dynamic construction of the tunnel project.Provides reliable theoretical data basis for change decision.
Keywords/Search Tags:Gaussian process regression, Multi-source heterogeneous information, Data mining, Deformation prediction, Dynamic construction
Related items