Tunneling Control Of EPB Shield Machine Based On Model Predictive Control

Shield tunneling is becoming the main method of tunnel excavation and underground engi-neering construction.Under the background of rapid economic growth and urbanization in China,it is of great significance to study the relevant technology of shield machine.Tunneling process is an important part of shield construction.This paper focused on the control of shield tunneling process.There are many problems in shield tunneling,for instance,prediction of earth pressure in the chamber is inaccurate,fracture accident of driving shafts caused by the unbalanced distri-bution of torque on driving shafts and earth pressure instability in the chamber caused by random disturbance.Therefore,the study object of this paper was rock cutting with cutter head and earth pressure balance of chamber,which were the two important components of driving tunneling pro-cess.Achievements have been made in prediction of earth pressure in chamber under driving and non-driving conditions,balance control of torque on driving shafts in cutter head driving system and earth pressure control of chamber under random disturbance.It is expected that this paper will be helpful for studying automatic construction of shield machine.The main contents and innovations of this dissertation are as follows:1.Prediction model for earth pressure in chamber under driving condition was established based on the multi-model fusion method.Based on the analysis of the meaning of each field in the data set of shield tunneling,the prediction target and input features were selected,and then the data were preprocessed.Based on the prediction number and the characteristics of data set,the method of leaving one was used for cross-validation and the mean square error(MSE)was used to evaluate the prediction effect.Based on LASSO regression model,SVR model,RFR model and GBRT model,the earth pressure pre-diction model of chamber was established.The prediction results of each prediction model were fused by the model fusion method based on average method.Finally,features importance was obtained by RFR model and GBRT model to analyze the relationship between input features and prediction targets.2.The earth pressure attenuation model of chamber under non-driving condition was estab-lished based on particle filter method.Firstly,the earth pressure attenuation fragments of chamber under non-driving condition were selected from the shield construction data,and the model structure of the earth pressure attenuation model of chamber was determined as an exponential model by curve fitting method.A particle filter-based modeling method for earth pressure attenuation model of chamber was proposed.The first step of this method is to obtain the prior probability distribution of the parameters of the earth pressure attenuation model by statistical analysis of all earth pressure attenuation fragments of chamber and the second step is to update the estimation of parameters of the earth pressure attenuation model based on particle filter method.The predictive performance of the proposed method was validated by the experimental results.3.The driving shaft torque balance control law was designed based on model predictive control method.Based on the characteristic of one-way rotation of cutter head system in shield tunneling,the state-space model of cutter head drive system was obtained by linearizing the model of cutter head drive system with non-linear backlash links.An augmented state space model of the cutter head drive system was established by introducing angular acceleration into the state vector.A method for estimating the torque of the drive shaft was designed based on the augmented state space model.The state observer of the augmented model was also designed.Based on the method of model predictive control,the driving shaft torque balance control law for cutter head drive system was designed by introducing the balanced allocation item of driving shaft torque into the control target.The simulation results showed that the proposed method could achieve the target of tracking cutter disc speed and the target of balanced distribution of driving shaft torque.4.The earth pressure control law of chamber was designed based on stochastic model predic-tive control method.Firstly,the state space model of the earth pressure control of chamber was established based on the mechanism model of the earth pressure of a chamber,and then the input and output con-straints were given based on the physical background of the shield machine system.A state space model based on working point was established considering the working point of earth pressure and disturbance.In order to eliminate the steady-state control error,an augmented state space model was established by adding integral term in the input.Considering the stochastic statistical charac-teristics of disturbances such as propulsion speed,the earth pressure control law of chamber was established based on stochastic model predictive control.An extended state observer was designed to estimate disturbance.The simulation results showed that the proposed method could improve the control effect of the earth pressure of chamber.