Machine learning-based classification of rock discontinuity trace:SMOTE oversampling integrated with GBT ensemble learning

This paper presents a hybrid ensemble classifier combined synthetic minority oversampling technique(SMOTE),random search(RS)hyper-parameters optimization algorithm and gradient boosting tree(GBT)to achieve efficient and accurate rock trace identification.A thirteen-dimensional database consisting of basic,vector,and discontinuity features is established from image samples.All data points are classified as either‘‘trace”or‘‘non-trace”to divide the ultimate results into candidate trace samples.It is found that the SMOTE technology can effectively improve classification performance by recommending an optimized imbalance ratio of 1:5 to 1:4.Then,sixteen classifiers generated from four basic machine learning(ML)models are applied for performance comparison.The results reveal that the proposed RS-SMOTE-GBT classifier outperforms the other fifteen hybrid ML algorithms for both trace and nontrace classifications.Finally,discussions on feature importance,generalization ability and classification error are conducted for the proposed classifier.The experimental results indicate that more critical features affecting the trace classification are primarily from the discontinuity features.Besides,cleaning up the sedimentary pumice and reducing the area of fractured rock contribute to improving the overall classification performance.The proposed method provides a new alternative approach for the identification of 3D rock trace.

Failure strength and fracture characteristics of rock with discontinuity under indirect tension

Large-scale discontinuities can significantly affect the mechanical properties of rock masses.However,the tensile behavior of rock discontinuities is often less investigated.To study the statistical characteristics of failure strength and fracture characteristics of rock discontinuities,Brazilian disc tests were conducted on limestone specimens with a single natural discontinuity at different load-discontinuity angles(β).In this study,β=0°andβ=90°correspond to the discontinuity parallel and perpendicular to loading direction,respectively.The results show that Brazilian failure strength(BFS)can reasonably represent the tensile strength of rock with discontinuities,by comparing the BFS and tensile stress in the disc center at peak force.The two-parameter Weibull distribution can capture the statistical BFS characteristics of rock discontinuities parallel to loading direction(β=0°)and at different loaddiscontinuity angles(β≠0°).All specimens with discontinuity at different load-discontinuity angles show more plastic deformational behaviour than intact rock specimen.With increasingβ,the mean BFS of limestone with discontinuity increases before reaching a plateau atβ=45°.The single plane of weakness theory best explains the BFS of fractured limestone withβ.Only a specific segment of preexisting rock discontinuity could affect the fracture process.Whenβ=0°,interfacial cracks and alternative cracks formed.Whenβ≠0°,mixed failure mode with shear and tensile failure occurred,particularly whenβ=30°andβ=60°.The findings can contribute to better understanding the failure and fracture characteristics of rock with discontinuities,particularly the interaction of pre-existing discontinuities with stress-induced fracturing.

Failure mechanism and coupled static-dynamic loading theory in deep hard rock mining: A review

Rock failure phenomena,such as rockburst,slabbing(or spalling) and zonal disintegration,related to deep underground excavation of hard rocks are frequently reported and pose a great threat to deep mining.Currently,the explanation for these failure phenomena using existing dynamic or static rock mechanics theory is not straightforward.In this study,new theory and testing method for deep underground rock mass under coupled static-dynamic loading are introduced.Two types of coupled loading modes,i.e.'critical static stress + slight disturbance' and 'elastic static stress + impact disturbance',are proposed,and associated test devices are developed.Rockburst phenomena of hard rocks under coupled static-dynamic loading are successfully reproduced in the laboratory,and the rockburst mechanism and related criteria are demonstrated.The results of true triaxial unloading compression tests on granite and red sandstone indicate that the unloading can induce slabbing when the confining pressure exceeds a certain threshold,and the slabbing failure strength is lower than the shear failure strength according to the conventional Mohr-Column criterion.Numerical results indicate that the rock unloading failure response under different in situ stresses and unloading rates can be characterized by an equivalent strain energy density.In addition,we present a new microseismic source location method without premeasuring the sound wave velocity in rock mass,which can efficiently and accurately locate the rock failure in hard rock mines.Also,a new idea for deep hard rock mining using a non-explosive continuous mining method is briefly introduced.

Deep learning-based key-block classification framework for discontinuous rock slopes

The key-blocks are the main reason accounting for structural failure in discontinuous rock slopes, and automated identification of these block types is critical for evaluating the stability conditions. This paper presents a classification framework to categorize rock blocks based on the principles of block theory. The deep convolutional neural network(CNN) procedure was utilized to analyze a total of 1240 highresolution images from 130 slope masses at the South Pars Special Zone, Assalouyeh, Southwest Iran.Based on Goodman’s theory, a recognition system has been implemented to classify three types of rock blocks, namely, key blocks, trapped blocks, and stable blocks. The proposed prediction model has been validated with the loss function, root mean square error(RMSE), and mean square error(MSE). As a justification of the model, the support vector machine(SVM), random forest(RF), Gaussian naïve Bayes(GNB), multilayer perceptron(MLP), Bernoulli naïve Bayes(BNB), and decision tree(DT) classifiers have been used to evaluate the accuracy, precision, recall, F1-score, and confusion matrix. Accuracy and precision of the proposed model are 0.95 and 0.93, respectively, in comparison with SVM(accuracy = 0.85, precision = 0.85), RF(accuracy = 0.71, precision = 0.71), GNB(accuracy = 0.75,precision = 0.65), MLP(accuracy = 0.88, precision = 0.9), BNB(accuracy = 0.75, precision = 0.69), and DT(accuracy = 0.85, precision = 0.76). In addition, the proposed model reduced the loss function to less than 0.3 and the RMSE and MSE to less than 0.2, which demonstrated a low error rate during processing.

Modeling rock failure using the numerical manifold method followed bythe discontinuous deformation analysis

A complete rock failure process usually involves opening/sliding of preexisting discontinuities as well as frac- turing in intact rock bridges to form persistent failure sur- faces and subsequent motions of the generated rock blocks. The recently developed numerical manifold method （NMM） has potential for modelling such a complete failure process. However, the NMM suffers one limitation, i.e., unexpected material domain area change occurs in rotation modelling. This problem can not be easily solved because the rigid body rotation is not represented explicitly in the NMM. The discontinuous deformation analysis （DDA） is specially de- veloped for modelling discrete block systems. The rotation- induced material area change in the DDA modelling can be avoided conveniently because the rigid body rotation is represented in an explicit form. In this paper, a transition technique is proposed and implemented to convert a NMMmodelling to a DDA modelling so as to simulate a complete rock failure process entirely by means of the two methods, in which the NMM is adopted to model the early fracturing as well as the transition from continua to discontinua, while the DDA is adopted to model the subsequent motion of the generated rock blocks. Such a numerical approach also im- proves the simulation efficiency greatly as compared with a complete NMM modelling approach. The fracturing of a rock slab with pre-existing non-persistent joints located on a slope crest and the induced rockfall process are simulated. The validity of the modelling transition from the NMM to the DDA is verified and the applicability of the proposed nu- merical approach is investigated.

Structural plane recognition from three-dimensional laser scanning points using an improved region-growing algorithm based on the robust randomized Hough transform

The staggered distribution of joints and fissures in space constitutes the weak part of any rock mass.The identification of rock mass structural planes and the extraction of characteristic parameters are the basis of rock-mass integrity evaluation,which is very important for analysis of slope stability.The laser scanning technique can be used to acquire the coordinate information pertaining to each point of the structural plane,but large amount of point cloud data,uneven density distribution,and noise point interference make the identification efficiency and accuracy of different types of structural planes limited by point cloud data analysis technology.A new point cloud identification and segmentation algorithm for rock mass structural surfaces is proposed.Based on the distribution states of the original point cloud in different neighborhoods in space,the point clouds are characterized by multi-dimensional eigenvalues and calculated by the robust randomized Hough transform(RRHT).The normal vector difference and the final eigenvalue are proposed for characteristic distinction,and the identification of rock mass structural surfaces is completed through regional growth,which strengthens the difference expression of point clouds.In addition,nearest Voxel downsampling is also introduced in the RRHT calculation,which further reduces the number of sources of neighborhood noises,thereby improving the accuracy and stability of the calculation.The advantages of the method have been verified by laboratory models.The results showed that the proposed method can better achieve the segmentation and statistics of structural planes with interfaces and sharp boundaries.The method works well in the identification of joints,fissures,and other structural planes on Mangshezhai slope in the Three Gorges Reservoir area,China.It can provide a stable and effective technique for the identification and segmentation of rock mass structural planes,which is beneficial in engineering practice.

Study on Estimation Method of Rock Mass Discontinuity Shear Strength Based on Three-Dimensional Laser Scanning and Image Technique

The estimation of shear strength of rock mass discontinuity is always a focal, but difficult, problem in the field of geotechnical engineering. Considering the disadvantages and limitation of exist- ing estimation methods, a new approach based on the shadow area percentage （SAP） that can be used to quantify surface roughness is proposed in this article. Firstly, by the help of laser scanning technique, the three-dimensional model of the surface of rock discontinuity was established. Secondly, a light source was simulated, and there would be some shadows produced on the model surface. Thirdly, to obtain the value of SAP of each specimen, the shadow detection technique was introduced for use. Fourthly, compared with the result from direct shear testing and based on statistics, an empirical for- mula was found among SAP, normal stress, and shear strength. Data of Yujian （~ River were used as an example, and the following conclusions have been made. （1） In the case of equal normal stress, the peak shear stress is positively proportional to the SAP. （2） The formula for estimating was derived, and the predictions of peak-shear strength made with this equation well agreed with the experimental re- suits obtained in laboratory tests.

Strength test of 3D printed artificial rock mass with pre-existing fracture

Discontinuities or structural planes are widely distributed in natural rock masses and significantly influence their geo-mechanical and geometric properties.Herein,a batch of rock samples,each with a single structural plane,is created using a 3D printer equipped with two robotic manipulators.One of the manipulators is connected via a nozzle to a concrete pumping truck,which can extrude brittle rock-like material to form layered intact rock masses;the rock-like material is mainly composed of cement,silica fume,sand and water.The other manipulator features a knife,which can carve discontinuities onto each layer of the printed model.By means of this system,rock masses with discontinuous joints are formed,and the failure strengths of rock masses with different joints are demonstrated via uniaxial compression tests and direct shear tests.The results thereof obtained via digital image correlation technology show that discontinuities lower the strength of the rock mass models significantly.With the increase of the angle between the fracture and horizontal plane,the uniaxial compressive strength first decreases,and then increases.During shear testing,the shear strength of the rock mass models increases with the surface roughness of the preset joint.These test results indicate that the influence of artificial joints on the mechanical properties of the models is consistent with that of natural rock mass joints.Using digital modeling and 3D printing technology,cracks hidden in a rock mass can be reproduced.