Research And Implementation Of Ground Settlement Prediction Model For Shield Construction Based On Industrial Big Data

In recent years,with the development of advanced manufacturing technology,such as Industry 4.0 and Industrial Internet,technological innovation and industrial development based on industrial big data has become the hotspot of research.As the continuous development of urbanization in China,the demand for subway construction is increasing.Shield tunneling technology as a relatively efficient and convenient method of tunnel excavation is being widely used."Safety First" is the requirement that the tunnel construction has consistently adhered to and carried out,but the ground settlement that affects the construction safety has always been accompanied by the shield tunneling process.Ground settlement prediction is facing many difficulties due to low frequency of ground settlement monitoring and feedback lag.Therefore,timely and accurate prediction of ground settlement is of great practical significance.At present,the increasing mass,high-dimensional and low information density shield construction data make the traditional method difficult to solve the problem of efficiency and accuracy for predicting ground settlement,which seriously affects the timely and accurate prediction of the ground settlement in the shield construction process.In view of the above problems,this paper puts forward a method named “Research and Implementation of Ground Settlement Prediction Model of Shield Construction Based on Industrial Big Data”.Through the comprehensive analysis of the key factors that affect the ground settlement of the shield construction,we can mine the cause of the ground settlement deeply.Using the big data analysis technology that makes the massive data effciently into knowledge and experience in the industry do timely and accurate forecast of a certain range of settlement,which provide the basis for scientific decision-making of construction personnel.The main contents and innovations are as follows:(1)The feature engineering method is applied to the screening of the key factors of ground settlement.In this paper,the existing business related to the ground settlement forecasting has been sorted out and the ground deformation regularity or mechanism has been analysed.In addition,the key factors of ground subsidence are selected by the method of influencing factor analysis in the feature engineering.Finally,the mechanism analysis and the key influencing factors obtained by the feature engineering are applied to the selection of modeling data.(2)Using BP neural network and support vector regression to predict ground settlement.In addition,this paper innovatively proposes a new way to predict the ground settlement through the ensembling model of BP neural network and support vector regression.Practice has proved that this method could effectively combine the two characteristics of the model and the ground settlement prediction accuracy is significantly improved;(3)Putting forward a new method innovatively to predict ground settlement of different distance monitoring points away from the shield excavation.Different from the previous single point settlement prediction,this new method is to predict ground settlement of different distance monitoring points away from the shield excavation respectively and further predict the ground settlement of shield tunneling progress in a range of distance;(4)Finally,the method in this paper has been verified based on the construction of a shield project.The results show that the method proposed in this paper has achieved ideal reliability and accuracy in predicting ground settlement of different distance monitoring points away from the shield excavation.