Analysis And Prediction Of Group TBM Tunneling Performance

Due to its advantages of environmental protection,high efficiency,safety and reliability,TBM construction has become the first choice for many extra-long tunnel constructions.The analysis and prediction of TBM tunneling performance is a key issue in evaluating TBM equipment performance,construction period and cost.However,the analysis and prediction of tunneling performance at home and abroad mostly rely on single engineering data,and the universality of the prediction model is low.For this reason,this thesis relies on 8 group TBMs of Xinjiang YE Project and compares the data of 15 TBMs in other typical projects,analyzes the distribution law of the main tunneling performance of TBMs and the main factors affecting the tunneling performance,and establishes a multi-factor correlation based Analysis of TBM tunneling performance with empirical regression models and intelligent models.Firstly,the tunneling performance of Xinjiang YE project and other projects is compared and analyzed.The results show that equipment downtime is the main reason for the decrease of TBM’s intact rate,mechanical and belt failures are the main factors affecting the intact rate;TBM failure,tool inspection and support operations are the main factors affecting the proportion of TBM tunneling time,TBM utilization The rate and construction speed are more sensitive to lithology and surrounding rock grade;the distribution peaks of cutterhead speed,thrust and cutterhead torque gradually decrease with the surrounding rock grade from type II to type IV.Then,the study identified the main influencing factors of TBM tunneling performance.The correlation analysis between TBM tunneling performance and rock uniaxial compressive strength（UCS）and integrity coefficient（Kv）shows that the comprehensive UCS and Kv can predict the construction speed to a large extent;Geological factors are used as the correction index of the utilization rate model,and construction personnel management factors and equipment factors are directly introduced as the correction index of the TBM utilization rate prediction model;through the comparison and analysis of the cutter head speed of several relying on projects,the establishment of the diameter coefficient k₁,surrounding rock Cutterhead speed formula with coefficient k₂.Further,the regression prediction model and method of TBM penetration,tunneling speed,utilization rate and construction speed are established.By analyzing the correlation between FPI,UCS and Kv,a relational expression that can solve FPI based on UCS and Kv is established,and a TBM penetration prediction model and tunneling speed prediction method are further established;the utilization rate of TBM tunneling is studied.Influencing factors,the correction coefficients of these influencing factors were proposed,and a more perfect TBM utilization prediction model was established;based on the tunneling speed and utilization prediction model,a more universal TBM construction speed prediction method was proposed;and actual engineering data were used to carry out Validation of the predictive model.Finally,the tunneling performance intelligent model of five algorithms including random forest,support vector machine,linear regression,XGboost and Adaboost is established.Comparing the prediction results of the five intelligent models,the results show that the Adaboost algorithm performs best in both tunneling speed prediction and construction speed prediction.At the same time,the Northeast HJ tunnel data was selected to compare and verify the construction speed prediction of the intelligent model and the regression model.The results showed that the average relative prediction error of the Adaboost algorithm was 12.76%,which further verified the effectiveness of the intelligent model;and obtained the results of the intelligent and regression models.Suggestions for use:Different prediction models and methods can be selected according to the complexity of the model,the quality and ability of the user,and the pre-and post-project.Based on the group TBM data of different diameters and different surrounding rocks,the universality of the prediction model has been well verified,which provides a reference for TBM project planning,survey and design,construction,and bidding schedule prediction.