A performance-based hybrid deep learning model for predicting TBM advance rate using Attention-ResNet-LSTM

The technology of tunnel boring machine(TBM)has been widely applied for underground construction worldwide;however,how to ensure the TBM tunneling process safe and efficient remains a major concern.Advance rate is a key parameter of TBM operation and reflects the TBM-ground interaction,for which a reliable prediction helps optimize the TBM performance.Here,we develop a hybrid neural network model,called Attention-ResNet-LSTM,for accurate prediction of the TBM advance rate.A database including geological properties and TBM operational parameters from the Yangtze River Natural Gas Pipeline Project is used to train and test this deep learning model.The evolutionary polynomial regression method is adopted to aid the selection of input parameters.The results of numerical exper-iments show that our Attention-ResNet-LSTM model outperforms other commonly-used intelligent models with a lower root mean square error and a lower mean absolute percentage error.Further,parametric analyses are conducted to explore the effects of the sequence length of historical data and the model architecture on the prediction accuracy.A correlation analysis between the input and output parameters is also implemented to provide guidance for adjusting relevant TBM operational parameters.The performance of our hybrid intelligent model is demonstrated in a case study of TBM tunneling through a complex ground with variable strata.Finally,data collected from the Baimang River Tunnel Project in Shenzhen of China are used to further test the generalization of our model.The results indicate that,compared to the conventional ResNet-LSTM model,our model has a better predictive capability for scenarios with unknown datasets due to its self-adaptive characteristic.

An Optimized System of Random Forest Model by Global Harmony Search with Generalized Opposition-Based Learning for Forecasting TBM Advance Rate

As massive underground projects have become popular in dense urban cities,a problem has arisen:which model predicts the best for Tunnel Boring Machine(TBM)performance in these tunneling projects?However,performance level of TBMs in complex geological conditions is still a great challenge for practitioners and researchers.On the other hand,a reliable and accurate prediction of TBM performance is essential to planning an applicable tunnel construction schedule.The performance of TBM is very difficult to estimate due to various geotechnical and geological factors and machine specifications.The previously-proposed intelligent techniques in this field are mostly based on a single or base model with a low level of accuracy.Hence,this study aims to introduce a hybrid randomforest(RF)technique optimized by global harmony search with generalized oppositionbased learning(GOGHS)for forecasting TBM advance rate(AR).Optimizing the RF hyper-parameters in terms of,e.g.,tree number and maximum tree depth is the main objective of using the GOGHS-RF model.In the modelling of this study,a comprehensive databasewith themost influential parameters onTBMtogetherwithTBM AR were used as input and output variables,respectively.To examine the capability and power of the GOGHSRF model,three more hybrid models of particle swarm optimization-RF,genetic algorithm-RF and artificial bee colony-RF were also constructed to forecast TBM AR.Evaluation of the developed models was performed by calculating several performance indices,including determination coefficient(R2),root-mean-square-error(RMSE),and mean-absolute-percentage-error(MAPE).The results showed that theGOGHS-RF is a more accurate technique for estimatingTBMAR compared to the other applied models.The newly-developedGOGHS-RFmodel enjoyed R2=0.9937 and 0.9844,respectively,for train and test stages,which are higher than a pre-developed RF.Also,the importance of the input parameters was interpreted through the SHapley Additive exPlanations(SHAP)method,and it was found that thrust force per cutter is the most important variable on TBMAR.The GOGHS-RF model can be used in mechanized tunnel projects for predicting and checking performance.

开挖补偿法防控深部地下岩爆灾害——引汉济渭工程秦岭输水隧洞案例分析

Rockburst disasters occur frequently during deep underground excavation,yet traditional concepts and methods can hardly meet the requirements for support under high geo-stress conditions.Consequently,rockburst control remains challenging in the engineering field.In this study,the mechanism of excavation-induced rockburst was briefly described,and it was proposed to apply the excavation compensation method(ECM)to rockburst control.Moreover,a field test was carried out on the Qinling Water Conveyance Tunnel.The following beneficial findings were obtained:Excavation leads to changes in the engineering stress state of surrounding rock and results in the generation of excess energy DE,which is the fundamental cause of rockburst.The ECM,which aims to offset the deep excavation effect and lower the risk of rockburst,is an active support strategy based on high pre-stress compensation.The new negative Poisson’s ratio(NPR)bolt developed has the mechanical characteristics of high strength,high toughness,and impact resistance,serving as the material basis for the ECM.The field test results reveal that the ECM and the NPR bolt succeed in controlling rockburst disasters effectively.The research results are expected to provide guidance for rockburst support in deep underground projects such as Sichuan-Xizang Railway.

Effects of data smoothing and recurrent neural network(RNN)algorithms for real-time forecasting of tunnel boring machine(TBM)performance

Tunnel boring machines(TBMs)have been widely utilised in tunnel construction due to their high efficiency and reliability.Accurately predicting TBM performance can improve project time management,cost control,and risk management.This study aims to use deep learning to develop real-time models for predicting the penetration rate(PR).The models are built using data from the Changsha metro project,and their performances are evaluated using unseen data from the Zhengzhou Metro project.In one-step forecast,the predicted penetration rate follows the trend of the measured penetration rate in both training and testing.The autoregressive integrated moving average(ARIMA)model is compared with the recurrent neural network(RNN)model.The results show that univariate models,which only consider historical penetration rate itself,perform better than multivariate models that take into account multiple geological and operational parameters(GEO and OP).Next,an RNN variant combining time series of penetration rate with the last-step geological and operational parameters is developed,and it performs better than other models.A sensitivity analysis shows that the penetration rate is the most important parameter,while other parameters have a smaller impact on time series forecasting.It is also found that smoothed data are easier to predict with high accuracy.Nevertheless,over-simplified data can lose real characteristics in time series.In conclusion,the RNN variant can accurately predict the next-step penetration rate,and data smoothing is crucial in time series forecasting.This study provides practical guidance for TBM performance forecasting in practical engineering.

Studies on the evolution process of rockbursts in deep tunnels

This paper focuses on the evolution processes of different types of rockbursts occurring in deep tunnels. A series of laboratory tests and in-situ monitoring in deep tunnels excavated by tunnel boring machine （TBM） and drill-and-blast （D＆B） method have been conducted to understand the mechanisms and processes of the evolution of different types of rockbursts, including strain rockburst, strain-structure slip rockburst, immediate rockburst and time-delayed rockburst. Three different risk assessment methods are proposed to evaluate the intensity and potential failure depth of rockbursts. These methods can be applied before excavation and the results can be updated according to the real-time information during excavation. Two micro-seismicity based real-time warning systems have been established for predicting various intensities ofrockbursts, such as slight, moderate, intensive and extremely intensive rockbursts. Meanwhile, the probability and intensity of the rockburst are also given. The strategy for excavation and support design has been suggested for various intensities of rockbursts before excavation. The strategy for dynamic control of the rockburst evolution process is also proposed according to the monitoring results. The methodology has been successfully applied to rockburst risk reduction for deep tunnels at Jinping II hydropower project. The results have illustrated the applicability of the proposed methodology and techniques concerning rockbursts.

Reducing risk in long deep tunnels by using TBM and drill-and-blast methods in the same project-the hybrid solution

There are many examples of TBM tunnels through mountains, or in mountainous terrain, which have suffered the ultimate fate of abandonment, due to insufficient pre-investigation. Depth-of-drilling limitations are inevitable when depths approach or even exceed l or 2 km. Uncertainties about the geology, hydro-geology, rock stresses and rock strengths go hand-in-hand with deep or ultra-deep tunnels. Unfortunately, unexpected conditions tend to have a much bigger impact on TBM projects than on drill-and-blast projects. There are two obvious reasons. Firstly the circular excavation maximizes the tangential stress, making the relation to rock strength a higher source of potential risk. Secondly, the TBM may have been progressing fast enough to make probe-drilling seem to be unnecessary. If the stress-to-strength ratio becomes too high, or if faulted rock with high water pressure is unexpectedly encountered, the ＂unexpected events＂ may have a remarkable delaying effect on TBM. A simple equation explains this phenomenon, via the adverse local Q-value that links directly to utilization. One may witness dramatic reductions in utilization, meaning ultra-steep deceleration-of-the-TBM gradients in a log-log plot of advance rate versus time. Some delays can be avoided or reduced with new TBM designs, where belief in the need for probe-drilling and sometimes also pre-injection, have been fully appreciated. Drill-and-blast tunneling, inevitably involving numerous ＂probe-holes＂ prior to each advance, should be used instead, if investigations have been too limited. TBM should be used where there is lower cover and where more is known about the rock and structural conditions. The advantages of the superior speed of TBM may then be fully realized. Choosing TBM because a tunnel is very long increases risk due to the law of deceleration with increased length, especially if there is limited pre-investigation because of tunnel depth.

Energy release process of surrounding rocks of deep tunnels with two excavation methods

Numerical analysis of the total energy release of surrounding rocks excavated by drill-and-blast （D＆B） method and tunnel boring machine （TBM） method is presented in the paper. The stability of deep tunnels during excavation in terms of energy release is also discussed. The simulation results reveal that energy release during blasting excavation is a dynamic process. An intense dynamic effect is captured at large excavation footage. The magnitude of energy release during full-face excavation with D＆B method is higher than that with TBM method under the same conditions. The energy release rate （ERR） and speed （ERS） also have similar trends. Therefore, the rockbursts in tunnels excavated by D＆B method are frequently encountered and more intensive than those by TBM method. Since the space after tunnel face is occupied by the backup system of TBM, prevention and control of rockbursts are more difficult. Thus, rockbursts in tunnels excavated by TBM method with the same intensity are more harmful than those in tunnels by D＆B method. Reducing tunneling rate of TBM seems to be a good means to decrease ERR and risk of rockburst. The rockbursts observed during excavation of headrace tunnels at Jinping II hydropower station in West China confirm the analytical results obtained in this paper.

Vibrations induced by tunnel boring machine in urban areas: In situ measurements and methodology of analysis

Excavation with tunnel boring machine(TBM)can generate vibrations,causing damages to neighbouring buildings and disturbing the residents or the equipment.This problem is particularly challenging in urban areas,where TBMs are increasingly large in diameter and shallow in depth.In response to this problem,four experimental campaigns were carried out in different geotechnical contexts in France.The vibration measurements were acquired on the surface and inside the TBMs.These measurements are also complemented by few data in the literature.An original methodology of signal processing is pro-posed to characterize the amplitude of the particle velocities,as well as the frequency content of the signals to highlight the most energetic bands.The levels of vibrations are also compared with the thresholds existing in various European regulations concerning the impact on neighbouring structures and the disturbance to local residents.

基于PSO-LSSVM算法的隧道掘进机掘进参数预测方法

为了规避隧道掘进机(tunnel boring machine,TBM)掘进参数人为设定的主观性,提出了一种基于粒子群-最小二乘支持向量机算法(PSO-LSSVM)的TBM掘进参数预测方法。通过从海量TBM工程掘进数据中探寻参数变化规律,降低了TBM主司机设定掘进参数的主观性,辅助其合理选择掘进参数,有利于提高掘进效率、规避工程风险,经实验和工程数据验证,PSO-LSSVM算法通过对样本粒子全局迭代寻优来优化参数,提升了预测算法泛化能力和预测精度,对推力、扭矩和推进速度参数预测数值偏差满足要求,可辅助指导主司机设定掘进参数。

复合地层小直径隧道掘进机掘进速度区间预测

合理准确预测隧道掘进机(tunnel boring machine,TBM)的掘进速度是实现TBM智能化控制的关键问题之一,复合地层小直径TBM施工的不确定性较常规地质条件更强,而传统预测方法对施工过程的不确定性考虑不足。在此通过引入区间预测方法,提出基于4种不同Bootstrap方法结合KELM-ANN模型的TBM掘进速度区间预测模型,并以南水北调安阳输水隧洞工程为例,选取142组工程实测数据验证区间预测模型的有效性。研究结果表明:基于Rademacher分布建立的模型预测结果优于其他3种方法,不仅可以得到较好的点预测结果,还可以构造出较为清晰可靠的区间将掘进速度实测值完全包络在内;随着置信水平的提高,区间可容纳的不确定性和风险也逐渐上升,通过变化区间宽度,能较好地量化和解释TBM施工过程中的不确定性因素对掘进速度的影响。研究结果可为TBM掘进性能预测和掘进参数优化提供参考。

Soft ground tunnel lithology classification using clustering-guided light gradient boosting machine

During tunnel boring machine(TBM)excavation,lithology identification is an important issue to understand tunnelling performance and avoid time-consuming excavation.However,site investigation generally lacks ground samples and the information is subjective,heterogeneous,and imbalanced due to mixed ground conditions.In this study,an unsupervised(K-means)and synthetic minority oversampling technique(SMOTE)-guided light-gradient boosting machine(LightGBM)classifier is proposed to identify the soft ground tunnel classification and determine the imbalanced issue of tunnelling data.During the tunnel excavation,an earth pressure balance(EPB)TBM recorded 18 different operational parameters along with the three main tunnel lithologies.The proposed model is applied using Python low-code PyCaret library.Next,four decision tree-based classifiers were obtained in a short time period with automatic hyperparameter tuning to determine the best model for clustering-guided SMOTE application.In addition,the Shapley additive explanation(SHAP)was implemented to avoid the model black box problem.The proposed model was evaluated using different metrics such as accuracy,F1 score,precision,recall,and receiver operating characteristics(ROC)curve to obtain a reasonable outcome for the minority class.It shows that the proposed model can provide significant tunnel lithology identification based on the operational parameters of EPB-TBM.The proposed method can be applied to heterogeneous tunnel formations with several TBM operational parameters to describe the tunnel lithologies for efficient tunnelling.

Technical innovation in the “Beishan No. 1” hard rock tunnel boring machine for high-gradient spiral tunnels

1 Project overview The“Beishan No.1”,the world’s first-ever hard rock Tunnel Boring Machine(TBM)tailored for high-gradient spiral tunnels,constitutes a pivotal element within the Beishan Underground Research Laboratory(URL)initiative in China.Beishan URL,the first URL for geological disposal of high level radioactive waste(HLW)in China,is a national key construction project listed in the“13th Five-Year Plan”.In 2019,subsequent to receiving approval from the China Atomic Energy Authority,the Beijing Research Institute of Uranium Geology(BRIUG),serving as the project’s owner,initiated its construction.This underground facility,categorized as a“third generation”URL for HLW disposal,i.e.,area-specific URL,was located in Beishan,Jiuquan City,Gansu Province,China,following more than three decades of rigorous research on site selection.

The influence of microwave irradiation on rocks for microwaveassisted underground excavation

Demand is growing for explosive-free rock breakage systems for civil and mining engineering, and space industry applications. This paper highlights the work being undertaken in the Geomechanics Laboratory of Mc Gill University to make a real application of microwave-assisted mechanical rock breakage to fullface tunneling machines and drilling. Comprehensive laboratory tests investigated the effect of microwave radiation on temperature profiles and strength reduction in hard rocks(norite, granite, and basalt)for a range of exposure times and microwave power levels. The heating rate on the surface of the rock specimens linearly decreased with distance between the sample and the microwave antenna, regardless of microwave power level and exposure time. Tensile and uniaxial compressive strengths were reduced with increasing exposure time and power level. Scanning electron micrographs(SEMs) highlighted fracture development in treated basalt. It was concluded that the microwave power level has a strong positive influence on the amount of heat damage induced to the rock surface. Numerical simulations of electric field intensity and wave propagation conducted with COMSOL Multiphysics~ software generated temperature profiles that were in close agreement with experimental results.

Wear Analysis of Disc Cutters of Full Face Rock Tunnel Boring Machine

Wear is a major factor of disc cutters’ failure. No current theory offers a standard for the prediction of disc cutter wear yet. In the field the wear prediction method commonly used is based on the excavation length of tunnel boring machine（TBM） to predict the disc cutter wear and its wear law, considering the location number of each disc cutter on the cutterhead（radius for installation）; in theory, there is a prediction method of using arc wear coefficient. However, the preceding two methods have their own errors, with their accuracy being 40% or so and largely relying on the technicians’ experience. Therefore, radial wear coefficient, axial wear coefficient and trajectory wear coefficient are defined on the basis of the operating characteristics of TBM. With reference to the installation and characteristics of disc cutters, those coefficients are modified according to penetration, which gives rise to the presentation of comprehensive axial wear coefficient, comprehensive radial wear coefficient and comprehensive trajectory wear coefficient. Calculation and determination of wear coefficients are made with consideration of data from a segment of TBM project（excavation length 173 m）. The resulting wear coefficient values, after modification, are adopted to predict the disc cutter wear in the follow-up segment of the TBM project（excavation length of 5621 m）. The prediction results show that the disc cutter wear predicted with comprehensive radial wear coefficient and comprehensive trajectory wear coefficient are not only accurate（accuracy 16.12%） but also highly congruous, whereas there is a larger deviation in the prediction with comprehensive axial wear coefficient（accuracy 41%, which is in agreement with the prediction of disc cutters’ life in the field）. This paper puts forth a new method concerning prediction of life span and wear of TBM disc cutters as well as timing for replacing disc cutters.

Design Theory of Full Face Rock Tunnel Boring Machine Transition Cutter Edge Angle and Its Application

At present, the inner cutters of a full face rock tunnel boring machine (TBM) and transition cutter edge angles are designed on the basis of indentation test or linear grooving test. The inner and outer edge angles of disc cutters are characterized as symmetric to each other with respect to the cutter edge plane. This design has some practical defects, such as severe eccentric wear and tipping, etc. In this paper, the current design theory of disc cutter edge angle is analyzed, and the characteristics of the rock-breaking movement of disc cutters are studied. The researching results show that the rotational motion of disc cutters with the cutterhead gives rise to the difference between the interactions of inner rock and outer rock with the contact area of disc cutters, with shearing and extrusion on the inner rock and attrition on the outer rock. The wear of disc cutters at the contact area is unbalanced, among which the wear in the largest normal stress area is most apparent. Therefore, a three-dimensional model theory of rock breaking and an edge angle design theory of transition disc cutter are proposed to overcome the flaws of the currently used TBM cutter heads, such as short life span, camber wearing, tipping. And a corresponding equation is established. With reference to a specific construction case, the edge angle of the transition disc cutter has been designed based on the theory. The application of TBM in some practical project proves that the theory has obvious advantages in enhancing disc cutter life, decreasing replacement frequency, and making economic benefits. The proposed research provides a theoretical basis for the design of TBM three-dimensional disc cutters whose rock-breaking operation time can be effectively increased.

Real-time prediction of rock mass classification based on TBM operation big data and stacking technique of ensemble learning

Real-time prediction of the rock mass class in front of the tunnel face is essential for the adaptive adjustment of tunnel boring machines(TBMs).During the TBM tunnelling process,a large number of operation data are generated,reflecting the interaction between the TBM system and surrounding rock,and these data can be used to evaluate the rock mass quality.This study proposed a stacking ensemble classifier for the real-time prediction of the rock mass classification using TBM operation data.Based on the Songhua River water conveyance project,a total of 7538 TBM tunnelling cycles and the corresponding rock mass classes are obtained after data preprocessing.Then,through the tree-based feature selection method,10 key TBM operation parameters are selected,and the mean values of the 10 selected features in the stable phase after removing outliers are calculated as the inputs of classifiers.The preprocessed data are randomly divided into the training set(90%)and test set(10%)using simple random sampling.Besides stacking ensemble classifier,seven individual classifiers are established as the comparison.These classifiers include support vector machine(SVM),k-nearest neighbors(KNN),random forest(RF),gradient boosting decision tree(GBDT),decision tree(DT),logistic regression(LR)and multilayer perceptron(MLP),where the hyper-parameters of each classifier are optimised using the grid search method.The prediction results show that the stacking ensemble classifier has a better performance than individual classifiers,and it shows a more powerful learning and generalisation ability for small and imbalanced samples.Additionally,a relative balance training set is obtained by the synthetic minority oversampling technique(SMOTE),and the influence of sample imbalance on the prediction performance is discussed.

Examining the effect of adverse geological conditions on jamming of a single shielded TBM in Uluabat tunnel using numerical modeling

Severe shield jamming events have been reported during excavation of Uluabat tunnel through adverse geological conditions, which resulted in several stoppages at advancing a single shielded tunnel boring machine(TBM). To study the jamming mechanism, three-dimensional(3D) simulation of the machine and surrounding ground was implemented using the finite difference code FLAC3D. Numerical analyses were performed for three sections along the tunnel with a higher risk for entrapment due to the combination of overburden and geological conditions. The computational results including longitudinal displacement contours and ground pressure profiles around the shield allow a better understanding of ground behavior within the excavation. Furthermore, they allow realistically assessing the impact of adverse geological conditions on shield jamming. The calculated thrust forces, which are required to move the machine forward, are in good agreement with field observations and measurements. It also proves that the numerical analysis can effectively be used for evaluating the effect of adverse geological environment on TBM entrapments and can be applied to prediction of loads on the shield and preestimating of the required thrust force during excavation through adverse ground conditions.

Real-time rock mass condition prediction with TBM tunneling big data using a novel rock-machine mutual feedback perception method

Real-time perception of rock mass information is of great importance to efficient tunneling and hazard prevention in tunnel boring machines(TBMs).In this study,a TBM-rock mutual feedback perception method based on data mining(DM) is proposed,which takes 10 tunneling parameters related to surrounding rock conditions as input features.For implementation,first,the database of TBM tunneling parameters was established,in which 10,807 tunneling cycles from the Songhua River water conveyance tunnel were accommodated.Then,the spectral clustering(SC) algorithm based on graph theory was introduced to cluster the TBM tunneling data.According to the clustering results and rock mass boreability index,the rock mass conditions were classified into four classes,and the reasonable distribution intervals of the main tunneling parameters corresponding to each class were presented.Meanwhile,based on the deep neural network(DNN),the real-time prediction model regarding different rock conditions was established.Finally,the rationality and adaptability of the proposed method were validated via analyzing the tunneling specific energy,feature importance,and training dataset size.The proposed TBM-rock mutual feedback perception method enables the automatic identification of rock mass conditions and the dynamic adjustment of tunneling parameters during TBM driving.Furthermore,in terms of the prediction performance,the method can predict the rock mass conditions ahead of the tunnel face in real time more accurately than the traditional machine learning prediction methods.

Excavation of underground research laboratory ramp in granite using tunnel boring machine: Feasibility study

Underground research laboratory(URL)plays an important role in safe disposal of high-level radioactive waste(HLW).At present,the Xinchang site,located in Gansu Province of China,has been selected as the final site for China’s first URL,named Beishan URL.For this,a preliminary design of the Beishan URL has been proposed,including one spiral ramp,three shafts and two experimental levels.With advantages of fast advancing and limited disturbance to surrounding rock mass,the tunnel boring machine(TBM)method could be one of the excavation methods considered for the URL ramp.This paper introduces the feasibility study on using TBM to excavation of the Beishan URL ramp.The technical challenges for using TBM in Beishan URL are identified on the base of geological condition and specific layout of the spiral ramp.Then,the technical feasibility study on the specific issues,i.e.extremely hard rock mass,high abrasiveness,TBM operation,muck transportation,water drainage and material transportation,is investigated.This study demonstrates that TBM technology is a feasible method for the Beishan URL excavation.The results can also provide a reference for the design and construction of HLW disposal engineering in similar geological conditions.2020 Institute of Rock and Soil Mechanics,Chinese Academy of Sciences.Production and hosting by Elsevier B.V.This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/).

Thrust and Torque Characteristics Based on a New cutter-head Load Model

Full face rock tunnel boring machine（TBM） has been widely used in hard rock tunnels, however, there are few published theory about cutter-head design, and the design criteria of cutter-head under complex geological is not clear yet. To deal with the complex relationship among geological parameters, cutter parameters, and operating parameters during tunneling processes, a cutter-head load model is established by using CSM（Colorado school of mines） prediction model. Force distribution on cutter-head under a certain geology is calculated with the new established load model, and result shows that inner cutters bear more force than outer cutters, combining with disc cutters abrasion; a general principle of disc cutters＇ layout design is proposed. Within the model, the relationship among rock uniaxial compressive strength（UCS）, penetration and thrust on cutter-head are analyzed, and the results shows that with increasing penetration, cutter thrust increases, but the growth rate slows and higher penetration makes lower special energy（SE）. Finally, a fitting mathematical model of ZT（ratio of cutter-head torque and thrust） and penetration is established, and verified by TB880 E, which can be used to direct how to set thrust and torque on cutter-head. When penetration is small, the cutter-head thrust is the main limiting factor in tunneling; when the penetration is large, cutter-head torque is the major limiting factor in tunneling. Based on the new cutter-head load model, thrust and torque characteristics of TBM further are researched and a new way for cutter-head layout design and TBM tunneling operations is proposed.