Simulation Analysis Of 3D Full Scale TBM Tunneling Process And Study On The Variation Law Of Tunneling Parameters

Since the 21st century,the focus of China’s infrastructure construction has gradually shifted to the western region with complex geological conditions,followed by higher construction difficulty and greater disaster risk.Tunnel boring machine(TBM)has the advantages of high efficiency,economy and environmental protection.It has been widely used in the construction of deep and long tunnels in the western region.Thrust and torque are important construction parameters of TBM.Timely adjusting thrust and torque according to different geological conditions plays an important role in giving full play to the advantages of TBM.Therefore,based on the three-dimensional full-scale TBM tunneling calculation model of particle flow,this paper studies the influence of jointed rock mass parameters and tunneling parameters on TBM thrust and torque,and establishes the prediction model of TBM thrust and torque by using machine learning method.The research results can provide reference for the selection of TBM construction parameters,and the following results are obtained:(1)A three-dimensional full-scale TBM tunneling calculation model based on particle flow is established.The advantages of flat joint model and smooth joint model in simulating jointed rock mass are discussed.The establishment steps of three-dimensional full-scale TBM tunneling calculation model based on particle flow are introduced,including complete rock model,jointed rock mass model and three-dimensional full-scale cutterhead model.The validity of the model is verified from the shape of thrust and torque curves and the development and expansion law of fractures.(2)A microparameters calibration method of flat joint model in PFC3D is proposed.The calibration index system is formed through parameter simplification,and the influence relationship of macro and micro parameters is obtained by orthogonal test and sensitivity analysis.The non-homogeneous linear equations are constructed to realize the rapid acquisition of initial microparameters.Based on the priority of macroparameters,the iterative correction method of microparameters is proposed to realize the rapid correction of initial microparameters.Finally,a set of standardized microparameter calibration method of flat joint model is formed,and the correctness of the method is verified based on laboratory experiments.(3)The influence law of jointed rock mass parameters and tunneling parameters on TBM thrust and torque is explored.Based on the established three-dimensional full-scale TBM tunneling calculation model,the effects of jointed rock mass parameters and tunneling parameters on the magnitude and distribution of TBM thrust and torque are systematically studied.It is found that the thrust and torque of TBM in the stable section of excavation are roughly normal distribution,and the distribution law is less affected by parameters;The higher the rock strength,the greater the thrust and torque,and the more obvious the numerical fluctuation;In the process of joint inclination from small to large,the fluctuation of thrust and torque shows an increasing trend as a whole.The values of thrust and torque first decrease and then increase.About 60° is considered to be the most easily excavated joint inclination;With the increase of joint spacing,the fluctuation of thrust and torque shows an increasing trend as a whole,the values of thrust and torque gradually increase,and the increasing rate gradually slows down;The larger the tunneling parameters are,the fluctuation of thrust and torque shows an increasing trend as a whole,the thrust and torque values gradually increase,and the increasing rate gradually slows down.(4)The prediction model of TBM thrust and torque is established,and the accuracy of the prediction model is testedTaking joint inclination,joint spacing,cutterhead advancing speed and cutterhead rotating speed as input and thrust and torque as output,a TBM thrust torque prediction model is established by using support vector regression and bp neural network,and the accuracy of the models is verified.The relative error of the prediction results based on support vector regression remains within 10%as a whole;The relative error of the prediction results using bp neural network is mostly less than 5%,and the prediction accuracy is higher than that of support vector regression.Applying the prediction model to practical engineering can provide reference for the intelligent selection of TBM thrust and torque and improve the boring efficiency.