Font Size: a A A

Multi-source Data Fusion Analysis And Tunnel Safety Evaluation For Ground Settlement In Shield Construction

Posted on:2022-04-18Degree:MasterType:Thesis
Country:ChinaCandidate:M L ShenFull Text:PDF
GTID:2492306602965049Subject:Master of Engineering
Abstract/Summary:
In order to alleviate the increasing traffic pressure in our country,many large and medium-sized cities are vigorously developing subways,which have become an important factor in measuring urban modernization.However,in the face of complex construction conditions and harsh external environments,there are still problems in shield construction that ground settlement cannot be effectively analyzed,and shield construction tunnels cannot be accurately and comprehensively evaluated for safety,so many serious construction accidents Generated from this.In view of the existing problems in the project and the urgent need for intelligent improvement,this thesis comprehensively uses industrial big data technology and intelligent algorithm modeling to propose a multi-source data fusion analysis and tunnel safety evaluation method for shield construction surface settlement.The research content is as follows:(1)Through systematic analysis of the related theories of shield construction surface settlement and safety evaluation,combined with data mining technology,the overall technical framework of this article is constructed,which specifically includes multi-source data collection and preprocessing,extraction of key influencing factors,Construct a ground settlement analysis model and a safety evaluation analysis model.(2)Extraction of key factors affecting surface settlement during shield construction.In order to ensure data quality,the collected multi-source data was preprocessed first.On this basis,this article combines the analysis of the surface subsidence mechanism,innovatively integrates meteorological data into the analysis of surface subsidence,and constructs an association rule model based on the FP-Growth algorithm to qualitatively extract the association that meets the minimum support and minimum confidence The rules reversely verify the correctness of the theoretical analysis.At the same time,in order to reduce the data dimension and the complexity of model training,the RF random forest and PCA algorithm are used to extract the key influencing factors of the multi-source data set,and the key feature set is constructed according to the feature importance.(3)Construct a ground settlement analysis model for shield construction.Different analysis data packages are formed according to the key feature set,and then the IFOREST algorithm is used to build a real-time anomaly detection model for surface subsidence to solve the problem of the inability to track the subsidence status in real time and the low frequency of monitoring in the project;use the GRU deep neural network to build the surface subsidence warning model,Solve the problem of dynamic warning of settlement trend in shield tunneling,and use genetic algorithm to optimize the parameters in the network,thereby reducing the prediction error of the model and improving the reliability of settlement warning results.(4)Construct a safety evaluation analysis model for shield construction tunnels.In this paper,the ground settlement analysis is closely linked with the tunnel safety evaluation.Aiming at the tunnel accident risk caused by the ground settlement,the AdaBoost algorithm is used to build a shield tunnel accident classification prediction model to obtain the types of possible accidents and their corresponding probability of occurrence;Secondly,in order to prevent or reduce the occurrence of accidents,comprehensive consideration of human,machine,material,environment and other risk factors in the shielding process,and combined with the analytic hierarchy process to build a three-tier indicator system of target level,criterion level and index level,using random The comprehensive weight of each index is calculated by the forest and entropy weight method,and finally the security level of the shield tunnel is calculated based on the fuzzy comprehensive evaluation theory.This thesis uses a certain unit shield engineering project as the background to implement and verify the application of the method proposed above.The analysis results are highly consistent with the actual project,which proves the effectiveness of the method,which has important reference value for the safe construction of shield tunnels.
Keywords/Search Tags:Shield construction, ground settlement, abnormal detection, GRU neural network, accident classification, safety evaluation
Related items