Key Technologies Study On Assistant Decision-making System Of Tunnel Advanced Geology Prediction

As an important part of tunnel informatization construction,the tunnel advanced geology prediction has made great progress recently and has been recognized in guaranteeing the safety of the tunnel construction.However,as a complex open system,it involves a large number of tunnel geological knowledge,interpretation of geophysical methods and the field operation methods,which make it face many challenges to realize the expected performance.Theoretical analysis,simulation experiment,field experiment,numerical computation and computer programming are adopted to research the common metallic interference principle and its' elimination of tunnel TEM,the attribute extraction of GPR,the target recognition technology of the typical karst geological anomalies,the development of expert system of tunnel macro-geological prediction and the multi-source information fusion of tunnel geological prediction based on evidence theory,which can be attributed to the software for assistant decision-making system of tunnel advanced geology prediction.The research achievements can perfect the theory system of tunnel comprehensive geological prediction as well as provide theoretical basis and technical support to realize "informatization and intellectualization" in the process of geological prediction.The main achievements of this paper are as follow.(1)Laboratory physical simulation and field detection test are combined to research the influence law on TEM results caused by tunnel excavation carts,steel arches and other metallic bodies.Resistive limits are then introduced to propose a fast approximate inversion method for metallic interference elimination in tunnel and low resistance geological anomalies.The inversion results of physical simulation and field practice show that the method can effectively eliminate metallic interference and improve the interpretation accuracy of TEM(the transient electromagnetic method).(2)After the signal interpretation technology of GPR(ground penetrating radar),which is based on the attribute analysis,is introduced.Physical simulation test and a huge amount of data of field detection are used as the base to research the corresponding relations between typical karst geological anomalies and three types of GPR attribute,namely,time domain,frequency domain and time-frequency domain and propose a computational method for attenuation parameter of radar wave absorption,which is based on general S-transform and wavelet spectrum simulation.And the target recognition model of the typical karst geological anomalies based on Gauss machine learning method is established,which can not only recognize the category of typical karst geological anomalies but also obtain specific probability of the classification results and realize the quantitative description of uncertainty evaluation of GPR prediction results.(3)By starting from the formation conditions of the common geological anomalies in tunnel and using the theory of engineering geological mechanics,system engineering,statistical analysis and expert consultation to analyze comprehensively all kinds of geological factors involved in the process of macroscopic geological prediction of tunnel,the macroscopic prediction index systems is established for three types of common geological anomalies,i.e.,fault fracture zones,water-bearing situation and karst development situation.Meanwhile,the expert system of macroscopic geological forecast based on Clips is developed.The reliability of the systems are verified by successful application in practical engineering.(4)According to the characteristics of the current tunnel geology comprehensive prediction method,D-S evidence theory is introduced to construct the multi-source information fusion model of tunnel geological comprehensive prediction.The evidence obtained in each step is superposed layer by layer following the comprehensive prediction process that from macroscopic prediction to the geological analysis in tunnel and then comprehensive geophysical prospecting(long distance and short distance).With the evidence more sufficient,the unfavorable geological information is much more clear for predicting the unfavorable geology of tunnel barrel,which can be narrowed down to a specific location.The comprehensive geological prediction of the left line in the LingJiao tunnel can be used as a practice to show the application effect of the model.(5)Based on the previous research,software for the assistant decision-making system of the tunnel advanced geology prediction is developed,which includes five modules as the macro-geological prediction,the metallic interference elimination of TEM and the fast inversion of the low-resistance geological anomalies,the attribute analysis of GPR and automatic recognition of typical karst geological anomalies,the tunnel geological comprehensive prediction based on information fusion technology and the data management system.Being a kind of prediction-query software systems,the software has a good man-machine interactive interface and is helpful in tunnel advanced geological prediction by prediction workers.