Tunnel Collapse Risk Prediction And Control

Cave-in is the most frequent accident during the construction of a tunnel. Once it happens, not only the schedule will be interrupted, but also the construction cost will be increased greatly and the life of the builders and technicians will be threatened. Therefore, a scientific prediction and control of tunnel cave-in is of great importance. This thesis aims at the study of this scourge. The main research contents and findings are shown as follows:(1) The data from 300 road, railway and subway cave-ins are collected and sorted out for a systematical classified statistical research on all the factors concerned with cave-in before and in the middle of the construction of a tunnel and on the basis of the comprehensive analysis, a relational tree of runnel cave-in factors is put forward.(2) Considering there are not enough data on tunnel cave-in in our country, the research findings of some predecessors are used here for the selection and quantitative study of the influencing factors of cave-in and, on the basis of SVM neural network theory, a tunnel cave-in SVM prediction model is built while related programs are developed by using a Libsvm tool box. Also suggested is a creative tunnel cave-in network prediction system, by which a tunnel cave-in can be predicted rapidly and scientifically.(3) On the basis of the predecessors'research findings, and analysis and prediction of cave-in causes done in this thesis, the research on the control of tunnel cave-in is divided into two phases:before and during the construction of a tunnel, while countermeasures are narrated in detail accordingly.The quantitative prediction of tunnel cave-in can be effected and the risks under systematic and scientific control after the prediction index system is created, some prediction methods are suggested, the prediction model is built, some calculation programs are developed, and some regulatory measures are found out. Therefore, references can be given to the theoretical and methodological tunnel cave-in security management and its use and development in practical engineering.