Research On Optimization Decision Of Excavation Parameters In TBM Construction Based On Surrounding Rock Perception

With the rapid development of my country’s infrastructure construction,more and more deep and long tunnels are under construction or proposed.Compared with the drill and blast method,TBM construction has been widely used in various tunnel projects due to its advantages of fast speed,high quality,low cost,and construction safety,and is recognized as the preferred construction method for deep and long tunnels.However,due to the constraints of construction progress,construction personnel level,technical equipment level and other factors,it is difficult to accurately obtain the surrounding rock conditions.The adjustment of TBM parameters mainly depends on the subjective experience of operators.It is often difficult to determine the optimal value of tunneling parameters under different surrounding rock conditions.As a result,the TBM tunneling speed is low and the construction cost increases.In order to solve the above problems,based on the measured data of a certain project in Xinjiang,this paper combines support vector machine,Bayesian,K-means clustering,BP neural network,long short-term memory neural network,extreme learning machine and other algorithms to establish a TBM tunneling model.The TBM tunneling parameter optimization decision-making model aimed at the highest efficiency forms an intelligent tunneling system including TBM stable section tunneling performance prediction,surrounding rock classification and stratum perception,and TBM tunneling parameter optimization decisionmaking.The main research contents of this paper are as follows:(1)The tunneling performance prediction of the TBM stable section based on the tunneling data of the ascending section.Collect on-site TBM excavation data,divide the excavation cycle into empty push section,ascending section,stable section and descending section,and process the data to establish a sample database.Taking thrust,torque,penetration,average cutter head speed and propulsion speed of the stabilizing section as input,and thrust and torque of stabilizing section as output,a predictive model is established using the support vector regression algorithm.The Bayesian optimization algorithm is used to optimize the kernel function of the support vector regression model to realize the prediction of the tunneling performance of the stable section of the TBM through the tunneling data of the ascending section.(2)Establish a TBM tunnel surrounding rock classification and stratum identification model based on tunneling parameters.Based on the thrust and torque of the stable section of the TBM,combined with the propulsion speed and cutterhead rotation speed,a classification model from tunneling parameters to formation grades is established through k-means clustering.The BP and LSTM algorithms were used to verify the classification of surrounding rocks,and compared with the grade of surrounding rocks predicted by the BP neural network.The comparison results show that the classification method obtained by K-clustering analysis has more accurate feature extraction and is easier to be recognized by the model,learning,and prediction,can better distinguish the tunneling parameters under different surrounding rock conditions,and lay a foundation for the next step of predicting the tunneling speed based on the classification of each surrounding rock.(3)Research on optimal decision-making of TBM tunneling speed based on formation perception.According to the stratum classification model established in Chapter 4,using thrust,torque,penetration,rotational speed and field depth-of-cut index as independent variables,and advancing speed as the target optimization variable,the TBM advancing speed based on the tunneling parameters under each surrounding rock grade was established predictive model.Taking the maximum value of the advancing speed under each surrounding rock grade in the past as the expected value,the corresponding optimal combination of tunneling parameters is found through exhaustive enumeration,and the functional relationship between the tunneling speed and each tunneling parameter is further fitted to evaluate the influence of each tunneling parameter on the tunneling speed,to establish the adjustment direction of tunneling parameters under different surrounding rock conditions.Finally,taking three types of surrounding rocks as examples,the simulated tunneling test is completed,which verifies the feasibility of the theory in this chapter.