Rockburst Intensity Research Articles

Investigation and application of data balancing and combined discriminant model in rock burst severity prediction

In the development of intelligent rock burst prediction models, issues such as incomplete data coverage and data imbalance are frequently encountered. These issues may lead to risks of overfitting in predictive models, poor generalization capabilities, and increased bias, which in turn may result in misjudgments and unpredictable losses. To accurately predict rock burst disasters and mitigate or eliminate related threats, this paper proposes a composite prediction model that integrates Density-Based Nonlinear Resampling (DBNR)-Tomek Link data balancing algorithms with Bayesian Optimization (BO)-Multilayer Perceptron (MLP)-Random Forest (RF). Initially, this study collected and organized a total of 301 recorded rock burst disaster field observation data, covering various tectonic plates, engineering types, rock origins, rock textures, and rock burst types. Subsequently, from a data analysis perspective, we employed the PCA-SSA-K-means unsupervised clustering algorithm to delve into the underlying information contained within the data, thereby validating the rationality of categorizing rock bursts into four grades. Then, using the L2 norm to optimize the dimensionality of the indicators and supplementing with indicator importance ranking and hypothesis testing, we selected the maximum tangential stress of the surrounding rock, the ratio of the maximum tangential stress of the surrounding rock to the uniaxial compressive strength of the rock (stress coefficient), and the elastic energy index as the criteria for rock burst intensity grading. Following that, the DBNR-Tomek Link sampling method was applied to balance the sample data, optimizing the data sample ratio and ultimately expanding the sample size to 396, improving the proportion of data samples from 2:3:4:1 to 1:1:2:1, thereby enhancing the model’s generalization performance. Ultimately, a BO-MLP-RF composite prediction model was constructed based on Bayesian Optimization (BO), Multilayer Perceptron (MLP), and Random Forest (RF) algorithms, with the Bayesian Optimization method ensuring that the model fits the training data well and generalizes to the test data. The results of tenfold cross-validation demonstrated that the model’s accuracy is consistently around 92.5%, combining the training results of rock burst models with imbalanced datasets, proving that the MLP model, adept at modeling nonlinear data, and the RF model, skilled in modeling large-scale data, serve as basic classifiers. This demonstrates that the application of data balancing and combined discriminative model schemes have enhanced the model’s predictive performance and stability. The model is capable of providing high-accuracy, high-efficiency early warning monitoring services for rock burst phenomena in rock engineering, thereby ensuring engineering safety.

AdaBoost model for rockburst intensity prediction considering class differences and quantitative characterization of misclassification difference

With the resource development gradually into the deep, rock explosion phenomenon is more and more frequent. The suddenness and harmfulness of rockbursts threaten the safe development of underground resources. In order to more accurately predict the possible intensities of rockbursts in specific rock conditions and stress environments, the rock mechanical and stress parameters between different intensities of rockbursts are further explored, an AdaBoost model considering the differences in the hierarchy is established, and a Flash Hill-Climbing method is proposed to optimize the hyperparameters in the classification model. Finally, a misclassification difference index (MCDI) is defined to quantitatively characterize the severity of misclassification. The results show that the accuracy of the Improved AdaBoost model is about 2.3% higher than that of the Normal AdaBoost model, and the Misclassification Difference Index (MCDI) of the Normal AdaBoost model is 05, while the the Improved AdaBoost model is 00. The model can provide a theoretical reference for rockburst prediction.

Interpretable model for rockburst intensity prediction based on Shapley values-based Optuna-random forest

To address the limitation of traditional machine learning models in explaining the rockburst intensity prediction process, this study proposes an interpretable rockburst intensity prediction model. The model was developed using 350 sets of actual rockburst sample data to explore the impact of input metrics on the final rockburst intensity level. The collected data underwent pre-processing using the isolation forest algorithm and synthetic minority oversampling technique. The random forest model was optimized through 5-fold cross-validation and the Optuna framework, resulting in the establishment of an Optuna-random forest (Op-RF) model that generates decision rules through its internal decision tree, utilizing the properties of the random forest model. The model was further interpreted using the Shapley additive explanations algorithm, both locally and globally. The results demonstrate that the proposed model achieved an area under curve score of 0.984. In comparison to eight other machine learning models, the proposed Op-RF model demonstrated superior accuracy, precision, recall, and F1 score. The model provides a transparent explanation of the prediction process, linking impact characteristics to the final output. Additionally, a cloud deployment method for the rockburst intensity prediction model is provided and its effectiveness is demonstrated through engineering verification. The proposed model offers a new approach to the application of machine learning in rockburst intensity prediction.

Experimental investigation on the influence of weak interlayers on sandstone rockburst and associated microcracking mechanism

Weak interlayers (WI) are common in sedimentary rock masses in deep coal mines. The qualitative effect of the WI on rockbursts is widely acknowledged; however, its influence mechanism still needs further investigation. In the present study, true triaxial unloading rockburst tests of sandstone with WI and calcite veins (CV) were conducted to explore their influence mechanisms. To explore the impact of WI, the rockburst stress, failure modes, acoustic emission (AE) parameters (energy, entropy, and b-value), and spatial energy characteristics of AE events were analyzed. The influence of the area ratio of WI and their distribution patterns (centralization and dispersion) on rockburst were further investigated. The results indicate that the rockburst stress (peak of maximum principal stress) decreased by 4 % for every 1 % increase in the sandstone's dispersion WI area ratio (1.9%–9.3 %). Namely, rockburst is more likely to occur when there is appropriate WI distributed in the sandstone because WI exacerbates the microcrack activities and energy release. The CV will reduce the weakening effect of WI on rockburst stress and can enhance the rockburst intensity, especially in samples with dispersion WI. Moreover, the more considerable AE energy is released around CV for the sandstone with dispersion WI. The interface between WI and matrix is prone to rockburst for the sandstone with centralized WI because of the concentrated energy release. The results of this paper can provide a reference for the prevention and control of rockbursts in mine sedimentary rocks containing WI and CV.

Experimental study on the failure behavior of rockburst induced by a dynamic disturbance in the circular chamber for deep underground engineering

Rockbursts induced by disturbance loads have become a significant engineering issue. Block specimens featuring a circular chamber structure were designed to investigate the effects of the disturbance frequency on rockbursts. The experiment simulated a series of rockburst failure processes through a disturbance load with four different frequencies by maintaining a constant preset confining pressure. The experimental results showed that an increased disturbance frequency intensifies the shear failure response and enhances the rockburst intensity. Furthermore, the particle image velocity technique was used to calculate the ejection velocity. The results suggested that the initial ejection velocity of fragments during rockburst surged by 253 % as the disturbance frequency was increased from 0.005 Hz to 0.625 Hz. The assessment level of rockburst failure transitions ranged from slight to moderate. These results indicated that an increase in disturbance frequency exacerbated rockburst failures. Two mechanisms were proposed to explain this effect. (i) The increase in disturbance frequency elevates the internal friction angle of rock units, enhancing their ultimate energy storage capacity. (ii) Increased energy input causes the energy storage in rock units to surpass the minimum energy required for rock failure, endowing the fragmented rock mass with greater velocity and leading to more intense rockburst failures. Ultimately, recommendations for early warning of rockburst failure are also provided by discussing the microscopic failure mechanisms of rockburst experiments and the rockburst in situ.

Deep Learning in Rockburst Intensity Level Prediction: Performance Evaluation and Comparison of the NGO-CNN-BiGRU-Attention Model

Rockburst is an extremely hazardous geological disaster. In order to accurately predict the hazardous degree of rockbursts, this paper proposes eight new classification models for predicting the intensity level of rockbursts based on intelligent optimisation algorithms and deep learning techniques and collects 287 sets of real rockburst data to form a sample database, in which six quantitative indicators are selected as feature parameters. In order to validate the effectiveness of the constructed eight machine learning prediction models, the study selected Accuracy, Precision, Recall and F1 Score to evaluate the prediction performance of each model. The results show that the NGO-CNN-BiGRU-Attention model has the best prediction performance, with an accuracy of 0.98. Subsequently, engineering validation of the model is carried out using eight sets of real rockburst data from Daxiangling Tunnel, and the results show that the model has a strong generalisation ability and can satisfy the relevant engineering applications. In addition, this paper also uses SHAP technology to quantify the impact of different factors on the rockburst intensity level and found that the elastic strain energy index and stress ratio have the greatest impact on the rockburst intensity level.

Rockburst simulation tests on structural plane of deep high-stress circular tunnels: A true triaxial test study on several different hard rocks

After deep rock excavation, rockbursts can easily occur. The structural plane and rock properties play important roles in controlling rockbursts. To study the role of structural planes and rock properties in controlling rockbursts in the peripheral rocks of tunnels, five different brittle and hard rocks (Red sandstone, Gray sandstone, Granite,Marble, and Limestone) were tested, and cubic specimens with round holes (100 mm*100 mm*100 mm, Φ = 50 mm) and prefabricated fissures were used to simulate the structural plane. A high-speed camera system and an acoustic emission system were used to monitor the damage and microfracture processes of the surrounding rock. The damage process, acoustic emission characteristics, and damage patterns of five brittle rocks with and without specimens with structural planes during loading were comparatively analyzed. The main conclusions were as follows. (1) The presence of structural planes in the rock mass changed the stress and energy propagation paths in the rock mass, resulting in stress concentration; the structural planes in the five brittle rocks could not store energy, hence increasing the propensity for rockbursts. Structural planes in the Marble had the most obvious effect on the promotion of rockbursts. However, structural planes in the Gray sandstone had the weakest effect on the promotion of rockbursts. (2) The presence of structural planes significantly increased the rockburst intensity. In the Marble, regardless of the presence of structural planes, the damage mode of the surrounding rock was mainly shear flexural damage. However, the fine particle contents of specimens with structural planes were significantly higher than those without structural planes. Thus, the rockburst intensities of specimens with structural planes were greater than those without structural planes. For the Red sandstone, Gray sandstone, Granite and Limestone, the surrounding rock specimens with structural planes experienced mainly violent shear ejection damage. (3) The distribution of RA and AF and the number of ruptures in the specimens with and without structural planes were obviously changed, and the numbers of rupture events in the Red sandstone, Gray sandstone, and Granite specimens obviously increased because of the presence of structural planes. The numbers of rupture events for Marble and Limestone specimens significantly decreased due to the presence of structural planes. Gray sandstone and Granite had the most rupture events. (4) Different brittle rock specimens with structural planes had different sizes. The degrees of damage were observed to significantly increase for a period. In addition, the rock body experienced additional internal shear rupture, which intensified due to the structural planes and the tunnels between the rock columns. The ruptures were caused by the destruction of the structural planes during rockburst, and they served as predictions and early warnings, thus providing a certain reference basis. (5) In the Marble during rockbursts, there were structural planes that underwent flexural strain and slip strain. In the Red sandstone, Gray sandstone, and Granite during rockbursts, there were structural planes that underwent compression shear strain and slip strain. The disaster chain was introduced to understand and discuss the mechanism of slip strain for the development of rockburst in structural planes, which should be the final damage form. The disaster chain was a series of rockbursts caused by compression shear and slip of specimens with structural planes. The development of rockbursts in specimens with structural planes occurred sequentially. The final damage pattern was the end result of the development of a series of rockbursts in specimens with structural planes.

Effect of moisture content on the rockburst intensity of sandstones

Rockburst is a common geological hazard in deep underground engineering, and it often occurs in strata consisting of brittle rocks. In this study, the moisture content effect on the rockburst intensity of sandstones is systematically studied. A series of triaxial unloading compression tests along with the acoustic emission monitoring are performed for sandstone specimens with different moisture content levels. The mechanical properties, failure characteristics, and dilatancy behaviors of sandstone specimens are then properly compared. Comparative results reveal that the triaxial compressive strength and total strain energy of the saturated specimen decrease by about 30% and 35%, respectively, as compared to those of the dry specimen. Moreover, the magnitude of elastic strain energy tends to decrease with the increasing water content. The effect of moisture content on the rockburst intensity of sandstones is, therefore, significant. Besides, it is also found that the onset of dilatancy is generally unaffected by the water content, whereas the extent of dilatancy significantly decreases with the increasing water content. Numerical simulations for a tunnel excavation model confirm that injecting water into the surrounding rock is an effective way of reducing the rockburst intensity during tunnel excavations. These results have a guiding significance for the prevention and control of rockbursts in underground engineering.

Damage Evolution and Energy Catastrophe Behavior during Quasi-static Loading Process of Rockburst

Rockburst pose a serious hazard to deep underground engineering. The damage evolution and energy transformation of rockburst under quasi-static load are the key to reveal the mechanism of rockburst. Based on the instability theory and stiffness theory, the burst rock and surrounding rock combination specimens are designed in this study. Then a constitutive model for simulating the whole process of rockburst is established by statistical damage of rock. The law of damage deformation and energy evolution during rockburst is revealed, and the key factors affecting rockburst intensity are analyzed. The results indicate that the damage mechanism of rockburst is related to the sudden change of damage before and after the peak stress of the burst rock, and the volume ratio of surrounding rock to burst rock. The evaluation index of rockburst proneness based on energy evolution can reflect the energy conversion and transmission relationship between the surrounding rock and burst rock in the combination system, which is of great significance for the evaluation of rockburst proneness.

Rockburst prediction for deep tunneling near fault based on the PD-BEM method

A multiscale model coupling the peridynamic (PD) and boundary element method (BEM) is proposed to analyze the rockburst in deep tunnels near fault. The entire rockburst processes of the rock masses around the opening are captured by PD while the field conditions and boundary are described by BEM, which can take full advantage of their salient features and model the rockburst problems efficiently. The proposed model is verified by comparing its predictions with those from the field observations in “11.28” rockburst at the Jinping II Hydropower Station in China. Parametric analysis considering the dip angle and location of the fault as well as the ground stress is performed to reveal the rockburst mechanism of deep tunnel near fault. Numerical results indicate that the rockburst intensity of rock tunnel is a function of the distance between the fault and tunnel, the dip angle of fault and ground stress. Different failure modes, including buckling failure, slip failure and local slabbing failure are found in the simulation, which are controlled by the fault near the tunnel. Moreover, a new index for rockburst is proposed to predict the tendency and intensity of rockburst associated with deep tunneling near a fault.

Rockburst Intensity Grade Prediction Based on Data Preprocessing Techniques and Multi-model Ensemble Learning Algorithms

Rockburst Intensity Grade Prediction Based on Data Preprocessing Techniques and Multi-model Ensemble Learning Algorithms

Generalized Weighted Mahalanobis Distance Improved VIKOR Model for Rockburst Classification Evaluation

Rockbursts are hazardous phenomena of sudden and violent rock failure in deep underground excavations under high geostress conditions, which poses a serious threat to geotechnical engineering. The occurrence of rockbursts is influenced by a combination of factors. Therefore, it is necessary to find an efficient method to assess rockburst grades. In this paper, we propose a novel method that enhances the VIKOR method using a novel combination of weight and generalized weighted Mahalanobis distance. The combination weights of the evaluation indicators were calculated using game theory by combining subjective experience and objective data statistical characteristics. By introducing the generalized weighted Mahalanobis distance, the VIKOR method is improved to address the issues of inconsistent dimensions, different importance, and inconsistent correlation among indicators. The proposed method can deal with the complexity of the impact factors of rockburst evaluation and classify the rockburst intensity level. The results show that the accuracy of the improved VIKOR method with the distance formula is higher than that of the unimproved VIKOR method; the evaluation accuracy of the improved VIKOR method with the generalized weighted Mahalanobis distance is 91.67%, which outperforms the improved VIKOR methods with the Euclidean and Canberra distances. This assessment method can be easily implemented and does not depend on the discussion of the rockburst occurrence mechanism, making it widely applicable for engineering rockburst evaluation.

Probability prediction method for rockburst intensity based on rough set and multidimensional cloud model uncertainty reasoning

Probability prediction method for rockburst intensity based on rough set and multidimensional cloud model uncertainty reasoning

The application of evidence-entropy weight gray incidence theory on the risk assessment of rockburst intensity in the Daxiangling tunnel

The risk assessment of rockburst intensity is significant for tunnel construction safety. First, the depth of the rockburst (X1), the uniaxial compressive strength of the rocks (X2), the brittleness coefficient of the rocks (X3), the stress coefficients of the rocks (X4), and the elastic energy index (X5) are adopted as the evidence body, and their essential certainty and reliability is determined using the entropy-gray correlation theory. Second, the synthetic certainty reliability of other samples is calculated based on the evidence theory. Relatively to the traditional gray extension model, it can improve the predictive accuracy and determine the certainty and reliability of different evidence bodies. The difference of importance between other evidence bodies can be reflected; and an interval scale can be taken into consideration in the evaluation process, so the proposed theory can reasonably predict the grade criterion which is interval form. Conclusion demonstrated that the suggested approach is entirely consistent with the actual investigation. The proposed model not only considers the unreliability or reliability of the problem but also solves some degrees of uncertainty and ambiguity of the datum; it enhances the predictive efficiency and provides a new way and thought for future risk assessment of rockburst intensity.

Study on Catastrophe Information Characteristics of Strain-Structural Plane Slip Rockburst in Deep Tunnels

Rigid structural planes encountered during construction have an obvious influence on rockburst intensity and occurrence mechanism. The high-intensity rockburst induced by the structural plane poses a great threat to the safety construction of the tunnel. A novel 3D discrete element numerical analysis method for rockburst is proposed with the help of the bonded block model and multi-parameter rockburst energy index. According to this method, the influence of multivariate information characteristics such as stress, energy, fracture, and rockburst proneness index on the surrounding rock during the strain-structural plane slip rockburst in deep tunnels is systematically investigated. The results are drawn as follows: The results show that from the analysis of multivariate information characteristics of strain-structural plane slip rockburst, the rock between the plane and tunnel is a typical rockburst risk area. The dip angle, length, and relative distance of the structural plane have a significant influence on the multivariate catastrophe information characteristics of the surrounding rock: As the dip angle increases, the fracture propagation range within the risk rock expands, but the rockburst intensity and the occurrence range gradually decrease; as the length increases, the fracture propagation range, rockburst intensity and occurrence range within the risk rock increase slightly; as the relative distance increases, the fracture propagation range and rockburst intensity gradually weaken, and the occurrence range of rockburst first increases and then decreases. Using the “11.28” strain-structural plane slip rockburst case as a basis, engineering validation research was conducted. The simulation results are found to be essentially consistent with the rockburst condition on the field, validating the rationality and applicability of the novel rockburst analysis method proposed in this paper.

Rock Burst Intensity-Grade Prediction Based on Comprehensive Weighting Method and Bayesian Optimization Algorithm–Improved-Support Vector Machine Model

In order to accurately judge the tendency of rock burst disaster and effectively guide the prevention and control of rock burst disaster, a rock burst intensity-grade prediction model based on the comprehensive weighting of prediction indicators and Bayesian optimization algorithm–improved-support vector machine (BOA-SVM) is proposed for the first time. According to the main factors affecting the occurrence and intensity of rock burst, the rock stress coefficient (σθ/σc), brittleness coefficient (σc/σt) and elastic energy index (Wet) are selected to construct the rock burst prediction indicator system. On the basis of the research of other scholars, according to the main performance and characteristics of rock burst, rock burst is divided into four intensity levels. The collected and sorted 120 sets of rock burst case data at home and abroad are taken as learning samples, and the T-SNE algorithm is used to perform dimensionality-reduction visualization processing on the sample data, visually display the distribution of samples of different grades, evaluate the representativeness of the sample data and prejudge the feasibility of the machine learning algorithm to distinguish different rock burst intensity levels. The combined improved analytic hierarchy process (IAHP) and Delphi method determine the subjective weight of the indicators; the combined entropy weight method and CRITIC method determine the objective weight of the indicator, and use the harmonic mean criterion of information theory to synthesize the subjective weight and objective weight of the indicator to obtain the comprehensive weight of the indicators. After weighted prediction indicators, a rock burst intensity-grade prediction model is constructed based on the support vector machine, and the hyperparameters of three types of support vector machines are improved by using the Bayesian optimization algorithm. Then, the prediction accuracy of different models is calculated by the random cross-validation method, and the feasibility and effectiveness of the rock burst intensity-grade prediction model is verified. In order to evaluate the generalization and engineering applicability of the proposed model, 20 groups of rock burst case data from the Maluping mine and Daxiangling tunnel are introduced to predict the rock burst intensity grade. The results show that the accuracy of the rock burst intensity-grade prediction model based on comprehensive weighting and BOA-SVM is as high as 93.30%, which is of higher accuracy and better effect than the ordinary model, and can provide warning information with a higher fault tolerance rate, which provides a new way of thinking for rock burst intensity-grade prediction.

Experimental study on the effect of water absorption level on rockburst occurrence of sandstone

To investigate the mechanism of rockburst prevention by spraying water onto the surrounding rocks, 15 experiments are performed considering different water absorption levels on a single face. High-speed photography and acoustic emission (AE) system are used to monitor the rockburst process. The effect of water on sandstone rockburst and the prevention mechanism of water on sandstone rockburst are analyzed from the perspective of energy and failure mode. The results show that the higher the absorption degree, the lower the intensity of the rockburst after absorbing water on single side of sandstone. This is reflected in the fact that with the increase in the water absorption level, the ejection velocity of rockburst fragments is smaller, the depth of the rockburst pit is shallower, and the AE energy is smaller. Under the water absorption level of 100%, the magnitude of rockburst intensity changes from medium to slight. The prevention mechanism of water on sandstone rockburst is that water reduces the capacity of sandstone to store strain energy and accelerates the expansion of shear cracks, which is not conducive to the occurrence of plate cracking before rockburst, and destroys the conditions for rockburst incubation.

Fracture evolution characteristics of strainburst under different gradient stress

AbstractStrainburst often occurs in the loading process of tangential stress concentration from surrounding rocks after excavation and unloading of deep rock mass. The concentrated tangential stress is relatively larger on the tunnel walls, which decreases towards the interior of surrounding rocks with a certain gradient. To explore the fracture evolution characteristics of surrounding rocks under different tangential gradient stresses and understand various phenomena in the strainburst process, a large‐scale true‐triaxial gradient and hydraulic‐pneumatic composite loading rockburst test device was adopted to carry out the strainburst tests under four different gradient stresses. Based on the macroscopic failure phenomena, distribution of rockburst debris, macro‐micro morphology of failure cross section, and monitored data of acoustic emission (AE), the fracture evolution of rockburst specimens was analyzed under different stress gradient loadings. The results indicate that the stress gradient has an obvious influence on the fracture characteristics, resulting in different rockburst phenomena. With the increase of stress gradient, the failure stress of rockburst decreases, while the dynamic failure characteristics of rockburst trends to be significant. The increase of stress gradient contributes to higher fragmentation degree of debris, more uniform distribution and better continuity. The total number of cracks in the model decreases with the rise of the stress gradient, and the proportion of shear failure increases, which is the main reason for the enhancement of the rockburst intensity.

Data Preprocessing and Machine Learning Modeling for Rockburst Assessment

Rockbursts pose a significant threat to human safety and environmental stability. This paper aims to predict rockburst intensity using a machine learning model. A dataset containing 344 rockburst cases was collected, with eight inducing features as input and four rockburst grades as output. In the preprocessing stage, missing feature values were estimated using a regression imputation strategy. A novel approach, which combines feature selection (FS), t-distributed stochastic neighbor embedding (t-SNE), and Gaussian mixture model (GMM) clustering, was proposed to relabel the dataset. The effectiveness of this approach was compared with common statistical methods, and its underlying principles were analyzed. A voting ensemble strategy was used to build the machine learning model, and optimal hyperparameters were determined using the tree-structured Parzen estimator (TPE), whose efficiency and accuracy were compared with three common optimization algorithms. The best combination model was determined using performance evaluation and subsequently applied to practical rockburst prediction. Finally, feature sensitivity was studied using a relative importance analysis. The results indicate that the FS + t-SNE + GMM approach stands out as the optimum data preprocessing method, significantly improving the prediction accuracy and generalization ability of the model. TPE is the most effective optimization algorithm, characterized simultaneously by both high search capability and efficiency. Moreover, the elastic energy index Wet, the maximum circumferential stress of surrounding rock σθ, and the uniaxial compression strength of rock σc were identified as relatively important features in the rockburst prediction model.

Evaluation of the Rock Burst Intensity of a Cloud Model Based on the CRITIC Method and the Order Relation Analysis Method

Evaluation of the Rock Burst Intensity of a Cloud Model Based on the CRITIC Method and the Order Relation Analysis Method