Predicting penetration rate of hard rock tunnel boring machine using fuzzy logic

Predicting the penetration rate of a tunnel boring machine (TBM) plays an important role in the economic and time planning of tunneling projects. In the past years, various empirical methods have been developed for the prediction of TBM penetration rates using traditional statistical analysis techniques. Soft computing techniques are now being used as an alternative statistical tool. In this study, a fuzzy logic model was developed to predict the penetration rate based on collected data from one hard rock TBM tunnel (the Queens Water Tunnel # 3, Stage 2) in New York City, USA. The model predicts the penetration rate of the TBM using rock properties such as uniaxial compressive strength, rock brittleness, distance between planes of weakness and the orientation of discontinuities in the rock mass. The results indicated that the fuzzy model can be used as a reliable predictor of TBM penetration rate for the studied tunneling project. The determination coefficient (R ²), the variance account for and the root mean square error indices of the proposed fuzzy model are 0.8930, 89.06 and 0.13, respectively.     

基于改进的距离判别分析法的南水北调西线工程TBM施工围岩分级

 对马氏距离判别法和层次分析法存在的不足进行改进,将改进的距离判别分析法应用于南水北调西线工程TBM施工围岩分级中。根据TBM施工特点和相关研究成果,将TBM施工围岩分级标准定为4级。选用岩石强度、岩组特征、结构面间距、结构面与洞轴线夹角以及石英含量5项指标作为判别因子,以南水北调西线工程杜柯河-玛柯河段实例数据作为学习样本进行训练,建立TBM施工围岩分级的改进的距离判别分析模型,利用得到的线性判别函数对待判样本进行分级。最后,将改进的距离判别分析法得到的判定结果与传统马氏距离判别法、RTBM法以及RMR方法得到的判别结果进行对比分析,验证了改进的距离判别分析法的有效性。研究结果表明,改进的距离判别分析法具有预测精度高等优点,为TBM施工围岩分级提供了一种新的有效方法。     

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.     

A methodology for evaluation and classification of rock mass quality on tunnel engineering

In this paper a new methodology for evaluation and classification of rock mass quality that can be applied to rock tunneling is presented. An evaluation model based on combing the analytic hierarchy process (AHP) and the fuzzy Delphi method (FDM) for assessing the rock mass rating is the main procedure. This research treats rock mass classification as a group decision problem, and applies the fuzzy logic theory as the criterion to calculate the weighting of factors. The main advantage of this procedure is that it can effectively change the weighting of each rating parameter with the variation of geological conditions. The proposed method was evaluated and applied to the actual cases that are the two tunnels along the Second Northern Highway around Taipei area in Taiwan, namely Mu-Zha and Hsin-Tien tunnels. It was found that the determined results were in a good agreement with the original data assessed by the RMR. Results of the analyses show that it can be provided a more quantitative measure of rock mass and hence minimize judgmental bias. The proposed method should be more feasible for future tunnel construction and for suggestions of tunnel support design in the geological area of Taiwan.     

Prediction of mode I fracture toughness of rock using linear multiple regression and gene expression programming

Prediction of mode I fracture toughness (K_IC) of rock is of significant importance in rock engineering analyses. In this study, linear multiple regression (LMR) and gene expression programming (GEP) methods were used to provide a reliable relationship to determine mode I fracture toughness of rock. The presented model was developed based on 60 datasets taken from the previous literature. To predict fracture parameters, three mechanical parameters of rock mass including uniaxial compressive strength (UCS), Brazilian tensile strength (BTS), and elastic modulus (E) have been selected as the input parameters. A cluster of data was collected and divided into two random groups of training and testing datasets. Then, different statistical linear and artificial intelligence based nonlinear analyses were conducted on the training data to provide a reliable prediction model of K_IC. These two predictive methods were then evaluated based on the testing data. To evaluate the efficiency of the proposed models for predicting the mode I fracture toughness of rock, various statistical indices including coefficient of determination (R²), root mean square error (RMSE), and mean absolute error (MAE) were utilized herein. In the case of testing datasets, the values of R², RMSE, and MAE for the GEP model were 0.87, 0.188, and 0.156, respectively, while they were 0.74, 0.473, and 0.223, respectively, for the LMR model. The results indicated that the selected GEP model delivered superior performance with a higher R² value and lower errors.     

Classification and assessment of rock mass parameters in Choghart iron mine using P-wave velocity

Engineering rock mass classification,based on empirical relations between rock mass parameters and engineering applications,is commonly used in rock engineering and forms the basis for designing rock structures.The basic data required may be obtained from visual observation and laboratory or field tests.However,owing to the discontinuous and variable nature of rock masses,it is difficult for rock engineers to directly obtain the specific design parameters needed.As an alternative,the use of geophysical methods in geomechanics such as seismography may largely address this problem.In this study,25 seismic profiles with the total length of 543 m have been scanned to determine the geomechanical properties of the rock mass in blocks Ⅰ,Ⅲ and Ⅳ-2 of the Choghart iron mine.Moreover,rock joint measurements and sampling for laboratory tests were conducted.The results show that the rock mass rating(RMR) and Q values have a close relation with P-wave velocity parameters,including P-wave velocity in field(V_(PF)).P-wave velocity in the laboratory(V_(PL)) and the ratio of V_(PF) V_(PL)(i.e.K_p = V_(PF)/V_(PL).However,Q value,totally,has greater correlation coefficient and less error than the RMR,In addition,rock mass parameters including rock quality designation(RQD),uniaxial compressive strength(UCS),joint roughness coefficient(JRC) and Schmidt number(RN) show close relationship with P-wave velocity.An equation based on these parameters was obtained to estimate the P-wave velocity in the rock mass with a correlation coefficient of 91%.The velocities in two orthogonal directions and the results of joint study show that the wave velocity anisotropy in rock mass may be used as an efficient tool to assess the strong and weak directions in rock mass.     

Seismic bearing capacity of a strip footing on rock media

The bearing capacity factors for a rough strip footing placed on rock media, which is subjected to pseudo-static horizontal earthquake body forces, have been determined using the lower bound finite element limit analysis in conjunction with the power cone programming (PCP). The rock mass is assumed to follow the generalized Hoek-Brown (GHB) yield criterion. No assumption needs to be made to smoothen the GHB yield criterion and the convergence is found to achieve quite rapidly while performing the optimization with the usage of the PCP. While incorporating the variation in horizontal earthquake acceleration coefficient (k_h), the effect of changes in unit weight of rock mass (γ), ground surcharge pressure ($q_0$) and the associated GHB material shear strength parameters (geological strength index (GSI), yield parameter (m_i), uniaxial compressive strength (σ_ci)) on the bearing capacity factors has been thoroughly assessed. Non-dimensional charts have been developed for design purpose. The accuracy of the present analysis has been duly checked by comparing the obtained results with the different solutions reported in the literature. The failure patterns have also been examined in detail.     

Predicting performance of impact hammers from rock quality designation and compressive strength properties in various rock masses

Predicting the performance of the impact hammers is one of the major subjects in determining the economics of the underground excavation projects in which they are utilized. Therefore, researchers have been attracted to developing performance prediction models for these machines. Physical and mechanical properties of rocks have been used to estimate the performance of impact hammers over the last few decades. In this study, the instantaneous breaking rate (IBR, m³/h) of an impact hammer used in construction of Levent-Hisarüstü metro tunnel (Istanbul) is recorded in detail. Sixty rock samples are obtained from tunnel route during the excavation of which the machine is employed. Physical and mechanical property tests are performed on the obtained samples. A data set including uniaxial compressive strength (UCS), rock quality designation index (RQD), Brazilian tensile strength (BTS), density (ρ), Schmidt hammer hardness (SHH), Shore scleroscope hardness (SSH), Cerchar abrasivity index (CAI), and IBR is formed. Regression analysis techniques are applied to the created data set in order to develop a performance prediction model. The investigation results in a model that can predict IBR based on UCS, RQD, and the output power of the impact hammer. The proposed model passes both F-test and t-test at 0.95 confidence level. The soundness of the model is successfully tested against two formerly developed models. Covering a wide range of application and requiring only two of the most common and versatile rock properties as input parameters are the other advantages of the suggested model.     

A real-time prediction method for tunnel boring machine cutter-head torque using bidirectional long short-term memory networks optimized by multi-algorithm

Based on data from the Jilin Water Diversion Tunnels from the Songhua River (China), an improved and real-time prediction method optimized by multi-algorithm for tunnel boring machine (TBM) cutter-head torque is presented. Firstly, a function excluding invalid and abnormal data is established to distinguish TBM operating state, and a feature selection method based on the SelectKBest algorithm is proposed. Accordingly, ten features that are most closely related to the cutter-head torque are selected as input variables, which, in descending order of influence, include the sum of motor torque, cutter-head power, sum of motor power, sum of motor current, advance rate, cutter-head pressure, total thrust force, penetration rate, cutter-head rotational velocity, and field penetration index. Secondly, a real-time cutter-head torque prediction model's structure is developed, based on the bidirectional long short-term memory (BLSTM) network integrating the dropout algorithm to prevent overfitting. Then, an algorithm to optimize hyperparameters of model based on Bayesian and cross-validation is proposed. Early stopping and checkpoint algorithms are integrated to optimize the training process. Finally, a BLSTM-based real-time cutter-head torque prediction model is developed, which fully utilizes the previous time-series tunneling information. The mean absolute percentage error (MAPE) of the model in the verification section is 7.3%, implying that the presented model is suitable for real-time cutter-head torque prediction. Furthermore, an incremental learning method based on the above base model is introduced to improve the adaptability of the model during the TBM tunneling. Comparison of the prediction performance between the base and incremental learning models in the same tunneling section shows that: (1) the MAPE of the predicted results of the BLSTM-based real-time cutter-head torque prediction model remains below 10%, and both the coefficient of determination (R²) and correlation coefficient (r) between measured and predicted values exceed 0.95; and (2) the incremental learning method is suitable for real-time cutter-head torque prediction and can effectively improve the prediction accuracy and generalization capacity of the model during the excavation process.     

Assessing the deformation response of double-track overlapped tunnels using numerical simulation and field monitoring

The unprecedented rate of metro construction has led to a highly complex network of metro lines. Tunnels are being overlapped to an ever-increasing degree. This paper investigates the deformation response of double-track overlapped tunnels in Tianjin, China using finite element analysis (FEA) and field monitoring, considering the attributes of different tunneling forms. With respect to the upper tunneling, the results of the FEA and field monitoring showed that the maximum vertical displacements of the ground surface during the tail passage were 2.06 mm, 2.25 mm and 2.39 mm obtained by the FEA, field monitoring and Peck calculation, respectively; the heaves on the vertical displacement curve were observed at 8 m (1.25D, where D is the diameter of the tunnel) away from the center of the tunnel and the curve at both sides was asymmetrical. Furthermore, the crown and bottom produce approximately 0.38 mm and 1.26 mm of contraction, respectively. The results of the FEA of the upper and lower sections demonstrated that the tunneling form has an obvious influence on the deformation response of the double-track overlapped tunnel. Compared with the upper tunneling, the lower tunneling exerted significantly less influence on the deformation response, which manifested as a smaller displacement of the strata and deformation of the existing tunnel. The results of this study on overlapped tunnels can provide a reference for similar projects in the future.     

Numerical modeling of time-dependent deformation and induced stresses in concrete pipes constructed in Queenston shale using micro-tunneling technique

Effects of time-dependent deformation (TDD) on a tunnel constructed using the micro-tunneling technique in Queenston shale (QS) are investigated employing the finite element method. The TDD and strength parameters of the QS were measured from tests conducted on QS specimens soaked in water and lubricant fluids (LFs) used in micro-tunneling such as bentonite and polymer solutions. The numerical model was verified using the results of TDD tests performed on QS samples, field measurements of some documented projects, and the closed-form solutions to circular tunnels in swelling rock. The verified model was then employed to conduct a parametric study considering important micro-tunneling design parameters, such as depth and diameter of the tunnel, in situ stress ratio (K_o), and the time lapse prior to replacing LFs with permanent cement grout around the tunnel. It was revealed that the time lapse plays a vital role in controlling deformations and associated stresses developed in the tunnel lining. The critical case of a pipe or tunnel in which the maximum tensile stress develops at its springline occurs when it is constructed at shallow depths in the QS layer. The results of the parametric study were used to suggest recommendations for the construction of tunnels in QS employing micro-tunneling.     

TBM performance estimation using rock mass classifications

Three tunnels for hydraulic purposes were excavated by tunnel-boring machines (TBM) in mostly hard metamorphic rocks in Northern Italy. A total of 14 km of tunnel was surveyed almost continually, yielding over 700 sets of data featuring rock mass characteristics and TBM performance. The empirical relations between rock mass rating and penetration rate clearly show that TBM performance reaches a maximum in the rock mass rating (RMR) range 40–70 while slower penetration is experienced in both too bad and too good rock masses. However, as different rocks gives different penetrations for the same RMR, the use of Bieniawski's classification for predictive purpose is only possible provided one uses a normalized RMR index with reference to the basic factors affecting TBM tunneling. Comparison of actual penetrations with those predicted by the Innaurato and Barton models shows poor agreement, thus highlighting the difficulties involved in TBM performance prediction.     

Machine learning-based automatic control of tunneling posture of shield machine

For a tunnel driven by a shield machine, the posture of the driving machine is essential to the construction quality and environmental impact. However, the machine posture is controlled by the experienced driver of shield machine by setting hundreds of tunneling parameters empirically. Machine learning (ML) algorithm is an alternative method that can let the computer to learn from the driver's operation and try to model the relationship between parameters automatically. Thus, in this paper, three ML algorithms, i.e. multi-layer perception (MLP), support vector machine (SVM) and gradient boosting regression (GBR), are improved by genetic algorithm (GA) and principal component analysis (PCA) to predict the tunneling posture of the shield machine. A set of the parameters for shield tunneling is extracted from the construction site of a Shanghai metro. In total, 53,785 pairwise data points are collected for about 373 d and the ratio between training set, validation set and test set is 3:1:1. Each pairwise data point includes 83 types of parameters covering the shield posture, construction parameters, and soil stratum properties at the same time. The test results show that the averaged R² of MLP, SVM and GBR based models are 0.942, 0.935 and 0.6, respectively. Then the automatic control for the posture of shield tunnel is illustrated with an application example of the proposed models. The proposed method is proved to be helpful in controlling the construction quality with optimized construction parameters.     

Evaluation of tunnel face stability by transparent soil models

Accurate estimation of tunnel face support pressure is necessary for economical and safe shield tunneling in cohesionless soils. This paper presents measurements of tunnel face support pressure and associated soil movements obtained using a transparent soil model that simulates shield tunneling in medium dense saturated sand. The use of a transparent soil surrogate permits measuring the internal soil deformations within the model soil. Soil deformations associated with various face support pressures are presented for 4 cover-to-diameter (C/D) ratios. Failure is found to be sudden with sand flowing into the tunnel leading to a prismatic wedge in front of the tunnel face and a vertical chimney of soil above. A minimum support pressure was achieved with support pressures as low as 10 ± 1% of the effective vertical stress at the tunnel axis. The stability of the tunnel face was related to the coefficient of active earth pressure with C/D ratio having a small effect on the magnitude of required pressure at collapse.     

Application of several optimization techniques for estimating TBM advance rate in granitic rocks

This study aims to develop several optimization techniques for predicting advance rate of tunnel boring machine(TBM)in different weathered zones of granite.For this purpose,extensive field and laboratory studies have been conducted along the 12,649 m of the Pahang-Selangor raw water transfer tunnel in Malaysia.Rock properties consisting of uniaxial compressive strength(UCS),Brazilian tensile strength(BTS),rock mass rating(RMR),rock quality designation(RQD),quartz content(q)and weathered zone as well as machine specifications including thrust force and revolution per minute(RPM)were measured to establish comprehensive datasets for optimization.Accordingly,to estimate the advance rate of TBM,two new hybrid optimization techniques,i.e.an artificial neural network(ANN)combined with both imperialist competitive algorithm(ICA)and particle swarm optimization(PSO),were developed for mechanical tunneling in granitic rocks.Further,the new hybrid optimization techniques were compared and the best one was chosen among them to be used for practice.To evaluate the accuracy of the proposed models for both testing and training datasets,various statistical indices including coefficient of determination(R~2),root mean square error(RMSE)and variance account for(VAF)were utilized herein.The values of R~2,RMSE,and VAF ranged in 0.939-0.961,0.022-0.036,and 93.899-96.145,respectively,with the PSO-ANN hybrid technique demonstrating the best performance.It is concluded that both the optimization techniques,i.e.PSO-ANN and ICA-ANN,could be utilized for predicting the advance rate of TBMs;however,the PSO-ANN technique is superior.     

Prediction of flyrock distance induced by mine blasting using a novel Harris Hawks optimization-based multi-layer perceptron neural network

In mining or construction projects, for exploitation of hard rock with high strength properties, blasting is frequently applied to breaking or moving them using high explosive energy. However, use of explosives may lead to the flyrock phenomenon. Flyrock can damage structures or nearby equipment in the surrounding areas and inflict harm to humans, especially workers in the working sites. Thus, prediction of flyrock is of high importance. In this investigation, examination and estimation/forecast of flyrock distance induced by blasting through the application of five artificial intelligent algorithms were carried out. One hundred and fifty-two blasting events in three open-pit granite mines in Johor, Malaysia, were monitored to collect field data. The collected data include blasting parameters and rock mass properties. Site-specific weathering index (WI), geological strength index (GSI) and rock quality designation (RQD) are rock mass properties. Multi-layer perceptron (MLP), random forest (RF), support vector machine (SVM), and hybrid models including Harris Hawks optimization-based MLP (known as HHO-MLP) and whale optimization algorithm-based MLP (known as WOA-MLP) were developed. The performance of various models was assessed through various performance indices, including a10-index, coefficient of determination (R²), root mean squared error (RMSE), mean absolute percentage error (MAPE), variance accounted for (VAF), and root squared error (RSE). The a10-index values for MLP, RF, SVM, HHO-MLP and WOA-MLP are 0.953, 0.933, 0.937, 0.991 and 0.972, respectively. R² of HHO-MLP is 0.998, which achieved the best performance among all five machine learning (ML) models.     

Influence of the Tunnel Shape on Shotcrete Lining Stresses

Tunnel excavation is frequently carried out in rock masses by the drill and blast method and the final shape of the tunnel boundary can be irregular due to overbreaks. In order to investigate the effects of overbreaks a study of the effect of tunnel boundary irregularity has been carried out. This is done developing a computational tool able to take into account fuzzy variables (i.e., thickness of the beams of the bedded spring approach used for the model). The obtained results show that irregularity effects should be considered when a shotcrete lining is used as the final tunnel lining (for the case where the tunneling procedure does not permit a smooth surface to be obtained). This is crucial to obtain a durable lining.     

Influence of rock mass fracturing on the net penetration rates of hard rock TBMs

Penetration rates during excavation using hard rock tunnel boring machines (TBMs) are significantly influenced by the degree of fracturing of the rock mass. In the NTNU prediction model for hard rock TBM performance and costs, the rock mass fracturing factor (k_s) is used to include the influence of rock mass fractures. The rock mass fracturing factor depends on the degree of fracturing, fracture type, fracture spacing, and the angle between fracture systems and the tunnel axis. In order to validate the relationship between the degree of fracturing and the net penetration rate of hard rock TBMs, field work has been carried out, consisting of geological back-mapping and analysis of performance data from a TBM tunnel. The rock mass influence on hard rock TBM performance prediction is taken into account in the NTNU model. Different correlations between net penetration rate and the fracturing factor (k_s) have been identified for a variety of k_s values.     

ANN and ANFIS performance prediction models for hydraulic impact hammers

Hydraulic impact hammers are mechanical excavators that can be used in tunneling projects economically under geologic conditions suitable for rock breakage by indentation. However, there is relatively less published material in the literature in relation to predicting the performance of that equipment employing rock properties and machine parameters. In tunnel excavation projects, there is often a need for accurate prediction the performance of such machinery. The poor prediction of machine performance can lead to very costly contractual claims. In this study, the application of soft computing methods for data analysis called artificial neural network (ANN) and adaptive neuro-fuzzy inference system (ANFIS) to predict the net breaking rate of an impact hammer is demonstrated. The prediction capabilities offered by ANN and ANFIS were shown by using field data of obtained from metro tunnel project in Istanbul, Turkey. For this purpose, two prediction models based on ANN and ANFIS were developed and the results obtained from those models were then compared to those of multiple regression-based predictions. Various statistical performance indexes were used to compare the performance of those prediction models. The results suggest that the proposed ANFIS-based prediction model outperforms both ANN model and the classical multiple regression-based prediction model, and thus can be used to produce a more accurate and reliable estimate of impact hammer performance from Schmidt hammer rebound hardness (SHRH) and rock quality designation (RQD) values obtained from the field tests.     

An evolutionary adaptive neuro-fuzzy inference system for estimating field penetration index of tunnel boring machine in rock mass

Field penetration index (FPI) is one of the representative key parameters to examine the tunnel boring machine (TBM) performance. Lack of accurate FPI prediction can be responsible for numerous disastrous incidents associated with rock mechanics and engineering. This study aims to predict TBM performance (i.e. FPI) by an efficient and improved adaptive neuro-fuzzy inference system (ANFIS) model. This was done using an evolutionary algorithm, i.e. artificial bee colony (ABC) algorithm mixed with the ANFIS model. The role of ABC algorithm in this system is to find the optimum membership functions (MFs) of ANFIS model to achieve a higher degree of accuracy. The procedure and modeling were conducted on a tunnelling database comprising of more than 150 data samples where brittleness index (BI), fracture spacing, α angle between the plane of weakness and the TBM driven direction, and field single cutter load were assigned as model inputs to approximate FPI values. According to the results obtained by performance indices, the proposed ANFIS_ABC model was able to receive the highest accuracy level in predicting FPI values compared with ANFIS model. In terms of coefficient of determination (R²), the values of 0.951 and 0.901 were obtained for training and testing stages of the proposed ANFIS_ABC model, respectively, which confirm its power and capability in solving TBM performance problem. The proposed model can be used in the other areas of rock mechanics and underground space technologies with similar conditions.