The Modeling Of TBM Boring Power Based On The Measured Data

Tunnel Boring Machine is a large equipment used in tunnelling construction.Boring power is an important parameter reflecting the equipment performance.It has important significance to analyze the influence of main factors such as geological parameters and operating condition to boring power.With the development of information technology,it is very necessary to explore how to make use of engineering data to set up prediction model,which is effected by multiple variables,and has the characteristics of the nonlinear relationship.The tunneling process is analyzed in this paper firstly,the variables affecting the performance is classified,relevant variables data set is established for a specific project,and preliminary analysis of the distribution characteristics and the correlation between variables is carried on,which is prepared for modeling.And then,on the basis of the data statistical analysis,boring power model is set up with the method of linear and nonlinear regression.The advantages and disadvantages of each method as well as the applicability for the engineering problems is discussed by studying the modeling process and results.In this paper,the partial least-squares regression which can overcome the multicollinearity between the independent variables,is used to establish linear regression model of boring power.Polynomial regression combined with stepwise regression is adopted to establish the boring power model.Finally,model of boring power is set up with support vector regression.Finally,on the basis of data analysis modeling work,mechanical analysis combined with data analysis is adopted to establish the driving power of model,which can reflect equipment control law,by analyzing the mechanical characteristics of driving process and mining the relationship between geological,operating parameters and the boring power,performance.The thrust and the cutter disc driving torque are obtained by main factors affecting the tunneling process.Based on the relationship between input energy consumption and load,the approximate estimate model of boring power is established.Moreover,engineering field data is used to identify the coefficients of the model,which depends on specific conditions.Further,the applicability and effectiveness of the model are discussed with independent data which is not used in modeling process.The comparison between the measured and predicted values shows that the model reflects the influences of the key factors.Therefore,it can provide a reference for optimization control and improve the tunneling efficiency.