Calculation And Prediction Method Of Shield Machine Cutter Torque

The shield machine is the core equipment for the tunnel excavation by the shield method.It can complete various mechanical operations such as breaking the soil on the face,discharging the muck,and ventilating,and at the same time protecting the people and equipment in the underground operation.The cutter head is the main mechanism for cutting the soil during the tunneling process of the shield machine,and also serves to support the working surface and ensure the stability of the cutting.The cutter torque is an important parameter for designing the cutterhead drive and manufacturing the economical shield.It is also the key parameter to ensure the normal advancement of the shield.In this paper,the cutterhead torque is taken as the research object,the torque calculation theory formula,the torque real-time prediction feature selection and prediction model construction,and the shield machine state monitoring are studied.The specific research content and work are as follows:(1)Aiming at the problem of the low calculation accuracy of the cutter head torque of the shield machine,the existing cutter head torque calculation methods are analyzed and summarized.For the main component of the cutter head torque-the front friction torque,a new method considering the actual shape of the cutter head and the distribution of earth pressure is proposed.The above integral formula is discretized into an easy-to-calculate sum formula by means of image processing.The proposed algorithm is compared with the theoretical algorithm and the actual construction torque for the actual cutter head.The comparison results prove that the actual shape of the cutter head is a key factor affecting the cutter head torque and cannot be ignored.The proposed algorithm reduces the calculation error by 39.6% under the working condition in this paper,and can calculate the torque of cutter head more accurately.(2)Aiming at the problem that the state machine has too many state variables in the torque forecasting of the shield machine,on the basis of removing the non-stationary working state,using the variance method and the peak method to screen out the characteristics of the distribution concentration in the shield machine data from the feature selection angle,Linear and nonlinear angles are used to analyze the correlation between feature and torque by correlation coefficient and KL divergence respectively.Finally,a set of shield machine state characteristics that can be used to accurately predict realtime torque is obtained.(3)Aiming at the problem that the existing torque prediction is mostly the average value,a real-time torque prediction model based on Long-Short Term Memory(LSTM)network is proposed.Firstly,the characteristics and difficulties of shield cutter torque prediction compared with other prediction problems are summarized.The limitations and disadvantages of common prediction algorithms for cutter torque prediction are summarized.Then the LSTM torque prediction model is established and used to normalize the model.The actual data set is validated and compared with the BP network model.The results show that the model can well fit the nonlinear relationship between key state parameters and cutter head torque,and has good prediction ability and generalization performance.(4)For the problem of how to implement stable and long-term monitoring of the critical state of the shield machine,analyze the software requirements from multiple angles,select the high-performance state monitoring software and hardware platform,build a simple and efficient software structure,and write a simple and easy-to-use UI interface.