Application Of Adaptive Neuro-fuzzy Inference System In Tunnel Structure Safety Evalution

Along with the rapid development of highway tunnel transportation construction, the number of built tunnel is rapidly increasing, and occurred many kinds of diseases inevitably, which becomes a serious threat to the tunnel traffic safety and the service life of structure. The research of tunnel structure safety condition assessment in our country started late, and also mainly pauses in the qualitative stage. This paper tried to apply the adaptive neuro-fuzzy inference system in the assessment of tunnel structure safety condition, to research the quantitative evaluation method of tunnel structure safety condition.Based on the comprehensive comparison of both domestic and overseas research of the evaluation method of tunnel structure safety condition, this paper introduced the research hot spot of the assessment of structure safety condition:the basic principle and modeling points of the adaptive neuro-fuzzy inference system (ANFIS), and trial selected the suitable key parameter for the engineering application. According to the selection requirements of the evaluation indicators, this paper divided the tunnel structure common diseases into three kinds because they caused mainly by external loads、material degradation and percolating water effect, and gave the qualitative and quantitative decision criteria of disease serious level, set up the tunnel structure safety state evaluation index system. On the basis of the above, this paper set up the tunnel structure safety state evaluation management system based on the adaptive neuro-fuzzy inference system (ANFIS), set up ANFIS model. tunnel information storage database and user management interface using Fuzzy Logic Toolbox、development database interconnection (ODBC) and graphical user interface (GUI).The performance of this system is tested and verified through numerical simulation, proved that it has good learning ability and fault tolerance ability.