The soft-hard interbedded rocks are often encountered in rock engineering such as mining engineering, tunnel engineering, and natural gas engineering. In this study, the horizontal soft-hard interbedded composite rock-like material (HSHC) specimens were prepared with different thickness ratios of rock-like layers, and a series of uniaxial compression tests were performed by using a self-designed loading mold. A parameter R was defined for the HSHC specimens, which is related to the thickness ratio of the soft and hard rock-like layers, the length-width ratio, and the mechanical parameters of the soft and hard rock-like materials. R was regarded to characterize the strength properties of the HSHC specimens. The test results showed that the axial strength of the HSHC specimens decreases first and then tends to be constant with increasing of R. Digital image correlation (DIC) technique was used to investigate the strain evolution and crack propagation process of the HSHC specimens. The displacement variation trend of the specimen can be roughly divided into three stages: stable deformation stage, rapid crack growth stage and instability failure stage. The failure mode of the HSHC specimens was mainly splitting tensile failure along the bedding plane. Meanwhile, a simplified mechanical model of laminated rock mass under uniaxial compression is established. The strength variation trend predicted by this model is consistent with the experimental results. The present study can reveal the failure mechanism of soft-hard interbedded rock mass at the bottom of the tunnel under the action of large horizontal stress after excavation.
The uniaxial compression tests of plate-shape granite samples containing two prefabricated holes on each side,prepared from Iddefjord granite blocks in Norway,are carried out by Instron hydraulic servo control testing machine in laboratory.The axial stress,axial and lateral strain,acoustic emission(AE) counts and failure modes of the samples are monitored during the tests.The splitting cracks occur near the holes' vertical surface and parallel to the loading direction when the loading stress increases to a certain value.Some characteristics of rockburst phenomena such as rock block ejection and slabbing failure can be observed around the prefabricated holes under high compressive stress.The failure modes of samples always initiate from the splitting fractures around the holes under uniaxial compression.By observing the AE counts rate curve,it is shown that more jumping mutation points appear in the curve than that of intact rock samples.Based on the experimental studies,numerical models are validated by FLAC3D(3 dimensions fast lagrangian analysis of continua).The principal stresses distributions in the samples are obtained by the elastic model analysis,while the failure process is studied by the Mohr-Coulomb strain softening model.The modes of failure zones are monitored and characterized by the AE count events.Most failure zones are failed in tension and connected together to form as splitting fractures parallel to the vertical boundary of the holes.The numerical results are almost the same as the laboratory tests.The research results,to some degree,reveal the formation mechanism of slabbing fractures,which are usually parallel to the excavation boundary surrounding underground openings in hard rocks under high in-situ stresses.
The occurrence of rockburst can cause significant disasters in underground rock engineering. It is crucial to predict and prevent rockburst in deep tunnels and mines. In this paper, the deficiencies of ensemble learning algorithms in rockburst prediction were investigated. Aiming at these shortages, a novel machine learning model, deep forest, was proposed to predict rockburst risk. The deep forest combines the characteristics of deep learning and ensemble models, which can solve complex problems. To develop the deep forest model for rockburst prediction, 329 real rockburst cases were collected to build a comprehensive database for intelligent analysis. Bayesian optimization was proposed to tune the hyperparameters of the deep forest. As a result, the deep forest model achieved 100% training accuracy and 92.4% testing accuracy, and it has more outstanding capability to forecast rockburst disasters compared to other widely used models (i.e., random forest, boosting tree models, neural network, support vector machine, etc.). The results of sensitivity analysis revealed the impact of variables on rockburst levels and the applicability of deep forest with a few input parameters. Eventually, real cases of rockburst in two gold mines, China, were used for validation purposes while the needed data sets were prepared by field observations and laboratory tests. The promoting results of the developed model during the validation phase confirm that it can be used with a high level of accuracy by practicing engineers for predicting rockburst occurrences.
The validity of the dynamic tensile strength in rock materials is problematic in dynamic Brazilian splitting tests. The purpose of this study is to understand the failure law and effectiveness of the dynamic Brazilian disc (BD) test for five different rock types under three typical loading modes, including platform loading (mode-I), steel bar loading (mode-II) and arc loading (mode-III). Based on the digital image correlation (DIC) technique and high-speed camera, the deformation localization and rupture evolution of the BD specimen under impact loading were obtained, and two rupture characteristics of the specimen were summarized. The loading mode significantly affects the crack initiation location of the disc in the dynamic BD test. The experimental results show that loading mode-III is the most consistent with Brazilian test theory in the dynamic BD test, while loading mode-II violates the test principle. At the same incident energy, the loading rate and dynamic tensile strength are functions of loading mode and lithology, which are highest in loading mode-III, followed by loading mode-I, and the lowest in loading mode-II. The strain and displacement field after peak load were extracted by DIC, and it was found that there were four typical failure modes in dynamic BD tests.
Pronounced tensile anisotropy in interbedded rock masses critically affects damage initiation in slopes and tunnels. In this study, Brazilian tests were conducted on soft-hard interbedded rock-like specimens with seven orientation angles (θ: 0°–90°) at a loading rate of 0.2 mm/min to investigate the effect of θ on tensile mechanical properties and fracture behavior, and further examine three loading-axis positions at θ = 90°, with real-time monitoring using acoustic emission (AE) and digital image correlation (DIC). The results indicate that Brazilian tensile strength (BTS), absorbed energy(U), and energy absorption rate decrease, while the degree of anisotropy (γ) increases with θ. BTS is positively correlated with U and negatively with γ. The loading-axis position significantly affects fracture morphology, which is governed by interface effects and the mechanical contrast between soft and hard layers. Fracture typically initiates at specimen ends and evolves in three stages, with failure shifting from central to non-central zones as θ increases. Interfaces play a crucial role in specimen fracture with larger θ promoting activation. Unlike single-interface rocks, interbedded rocks exhibit synergistic interactions among multiple interfaces; Stress tends to concentrate at the interfaces near the ends of the loading zone, triggering competitive crack propagation. Moreover, specimens with fewer interfaces exhibit higher BTS. Beyond 45°, the primary factors governing failure transition from hard matrix thickness and interface number to interface strength, leading to a convergence in tensile strength. These findings provide new insights into the complex anisotropic fracture behavior of soft-hard interbedded rock masses under tensile stress.
Chongqing is a mountain city in western China with a complex geological environment, which brings many difficulties to the construction of large sections of underground projects. In order to avoid serious problems in the construction of Guobo Station with large sections, intensive coring and laboratory tests were carried out to obtain the rock mass properties in detail. The boreholes reveal that the lithology of the rock mass in the project area consists mainly of mudstone and a small area of sandstone. The uniaxial compressive strength (UCS) and the Brazilian tensile strength (BTS) of mudstone are highly scattered, which are 4.43~42.43 MPa and 0.66~4.30 Mpa, respectively, but they all show a lognormal distribution. The average UCS of sandstone is 65.31 MPa, and the BTS is 3.06 MPa. Meanwhile, the softening coefficients of mudstone and sandstone are 0.33 and 0.73, which indicate that the water content has a great influence on the mechanical properties of the rock. For the in situ stress field, the relationship between the three principal stresses is as follows: the maximum horizontal principal stress > minimum horizontal principal stress > vertical principal stress, and the lateral pressure coefficient is 2.11~5.71. Based on the present experimental results, it is proposed that the potential risks of excavation include: (1) bias pressure (2) high in situ stress; (3) surrounding rock deterioration.
Abstract In the field of deep mining engineering, it poses a challenge to promptly determine the mechanical properties of rocks under poor geological conditions through in-situ tests. However, the indirect determination of uniaxial compressive strength (UCS) of rocks can be achieved through the point load strength index (PLSI) test on irregular samples. In the present study, laboratory uniaxial compressive and field PLSI tests were carried out on irregular ore and rock blocks extracted through mechanical mining methods from a stope at a lead–zinc mine in Yunnan Province, China, with a depth of approximately 1000 m. The effects of mechanical excavation and drilling-blasting methods on the PLSI of rocks and ores are compared. It is found that there are significant differences in the point load strength indexes obtained by different excavation methods, and the I s (50) obtained after the mechanical excavation method approximates the actual value of ore and surrounding rocks. Two correction methods were utilized to obtain the point load strength indexes I s (50)-1 and I s (50)-2 of irregular rock samples. The correlation factors ( k ) linking I s (50) to UCS and Brazilian splitting strength are derived based on testing results. The findings indicate that using the conversion factor ( k ) recommended by ISRM to predict the UCS of rocks may significantly underestimate the actual strength of rocks in intricate mining environments. This study can serve as a benchmark for analogous deep mining projects.
