基于机器学习的TBM破岩效率预测模型研究

Prediction Model of TBM Rock Breaking Efficiency Based on MachineLearning

在TBM施工中经常会遇到复杂地质环境,导致TBM自身设备参数受到干扰,因此有必要针对相应条件下TBM滚刀破岩效率变化规律展开分析研究。相较于传统钻爆法,TBM拥有掘进迅速,破岩时间短的优越性。因TBM的这一突出优越性,所以保障TBM破岩效率已成为研究的关键问题之一。本文依托重庆轨道交通十号线二期工程,以项目所在地采集的花岗岩为研究对象进行室内试验,基于二维节理岩体试验,引入机器学习的方法,基于优化的BP神经网络,建立TBM滚刀破岩效率预测模型,并对模型预测结果准确性进行验证。结果表明,本文预测结果准确度高,适用于TBM破岩效率预测,为TBM滚刀破岩效率预测方法研究提供了理论参考。

In the operation of TBM,the rock breaking efficiency of the hob will be interfered by the complicated geological environment and its own equipment parameters,so it is particularly critical to analyze the changing principle of the rock breaking efficiency of the hob.Compared with the general drilling and blasting method,TBM has the advantages of rapid tunneling and short rock breaking time.Therefore,one of the key problems in TBM construction is how to improve the rock breaking efficiency.Therefore,one of the key problems in TBM construction is how to improve the rock breaking efficiency.Based on the Phase II project of Chongqing Rail Transit Line 10,this paper takes the granite collected at the site of the project as the research object to carry out laboratory tests.In this paper,based on the two-dimensional hob fractured jointed rock mass test,machine learning method was introduced,the improved BP neural network was used to construct the prediction model of rock breaking efficiency of TBM hob,and the accuracy of the prediction results was tested.It is also found that the accuracy of the prediction results in this study is relatively good,which can be used in the prediction of the rock breaking efficiency of TBM.It also shows that the input parameters of the model can fully estimate the rock breaking efficiency of the TBM hob,which lays a theoretical condition for analyzing the rock breaking efficiency of the project.