Study On Intelligent Discrimination Method Of Surrounding Rock Stability In TBM Tunnel Construction Stage

Hard Rock Boring Machine(TBM)is a large-scale engineering machine used for tunnel excavation.After the excavated rock mass forms the tunnel,it is necessary to judge the stability of the surrounding rock through manual observation and determine the support scheme.In the construction process,manual observation can only qualitatively evaluate the stability of surrounding rock,which is restricted by the experience of construction personnel.With the intelligent development of TBM,fast,quantitative,and intelligent become the new requirements of surrounding rock stability discrimination.In this paper,the basic quality method is used to evaluate the stability of surrounding rock,and a method for judging the stability of surrounding rock in the construction stage of TBM tunnel is proposed through image recognition and driving parameters.The specific research contents are as follows:(1)According to the construction environment of the TBM tunnel,the surrounding rock image acquisition method based on three-dimensional reconstruction and the surrounding rock strength acquisition method based on a point load instrument are proposed.The quantitative evaluation of joint and leakage is realized through image recognition,and the rock strength is recognized through driving parameters.According to the stability of the surrounding rock,the overall support scheme and local reinforcement suggestions are put forward.(2)The surface feature recognition model of surrounding rock is constructed by Mask R-CNN.Optimize the loss function to realize the identification of joints and fissures,and increase the m AP to 0.437.Replace the backbone network and shorten the detection speed of the model to 3.44 s on the premise of ensuring the accuracy of model recognition.The decreasing learning rate is adopted to shorten the number of iterations of the model to 95 epochs.(3)Preprocess the driving parameters generated in the process of TBM construction and the rock mass strength measured in the field,and construct the dataset.The empirical formula of rock mass strength is established by multi-factor fitting analysis.The mapping relationship between driving parameters and rock mass strength is constructed through three improved support vector regression(SVR)models.The fitting errors of empirical formula and support vector regression methods are compared.(4)Go to Daliangshan No.1 Tunnel of Lexi Expressway to conduct experimental research on the discrimination method of surrounding rock stability during TBM tunnel construction.The image processing method of surrounding rock is discussed.The support scheme is determined according to the rock mass quality index of underground engineering[BQ],and the local reinforcement position is determined according to the local volume joints(J_v)of the surrounding rock.