Tunnel Surrounding Rock Stability Analysis And Prediction Based On RS-BPNN Theory

This thesis briefly introduces the tunnel development and its research status both at home and abroad, then makes an analysis of the existing problem facing China’s domestic tunnel projects, after which, it has been discovered that the surroundings and the corresponding geological conditions impose tremendous influence on tunnel surrounding rock stability. Due to the fact that tunnel surrounding rock is the complex formed by mutual combination of various geological conditions, as a result, the complicated relation between these impact factors makes it difficult to formulate a precise arithmetic expression to describe the mode and degree imposed on surrounding rock stability. So this thesis will analysis and predict the tunnel surrounding rock stability by RS-BPNN theory.This thesis gathers several groups of classification data of surrounding rock, takes rock integrity, seepage quantity of underground water, natural stress state of rock, quality indicator of rock, geological structure as the impact factors, bases on the Rough Set Theory(RS), and uses MATLAB as the calculating tool to analysis the impact factors of tunnel surrounding rock stability, then integrates classification rule as well as the extraction result of approximation set to conduct further calculation of weight of influencing ability on decision result by every attribute factor. The result indicates that some certain geological condition is relatively special, the weight of every impact factor will change correspondently as well as imposes some certain influence on surrounding rock classification. As a result, before proceeding to classify tunnel surrounding rock it is required to comprehensively consider whether the condition of this tunnel’s surrounding geological environment is special or not.Uses BPNN to establish the prediction model of surrounding rock stability on the basis of the above mentioned calculation result, take 16 groups of data as the training sample and use 3 groups of data to predict, the prediction conducts twice, the predicted value is very close to the real value.Uses the RS Theory to conduct weight analysis on impact factors of tunnel surrounding rock stability, in order to provide guidance on taking primary counter measure to keep the stability of tunnel; uses BPNN to conduct prediction on classification of tunnel surrounding rock stability, which can work out the grade of surrounding rock rapidly as well as effectively, and ensuring the approach to the real value; through the organically combination of both, To provide a theoretical basis for the design of tunnel security, but also provide strong protection to prevent tunnel accident.