Phase field simulation of excavation damage zone in natural fractured rock mass
ObjectivePhase-field simulation method was applied to naturally fractured rocks of engineering sites, with the aim of gaining insights into crack propagation mechanisms. It enabled the identification of areas within the tunnel prone to fail, allowing for enhanced support measures and effective risk...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Editorial Department of Journal of Sichuan University (Engineering Science Edition)
2025-01-01
|
| Series: | 工程科学与技术 |
| Subjects: | |
| Online Access: | http://jsuese.scu.edu.cn/thesisDetails#10.12454/j.jsuese.202500076 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | ObjectivePhase-field simulation method was applied to naturally fractured rocks of engineering sites, with the aim of gaining insights into crack propagation mechanisms. It enabled the identification of areas within the tunnel prone to fail, allowing for enhanced support measures and effective risk warning. The goal is to prevent further expansion of the excavation damaged zone (EDZ), thereby ensuring the safe and reliable operation of deep underground engineering projects. MethodsThis study began by introducing the theory of brittle fracture and subsequently presented the phase-field model as a novel numerical simulation method. The model governed crack propagation through the definition of a phase-field variable. To verify the rationality and effectiveness of the phase-field approach, two-dimensional models of square specimens with both predefined and randomly distributed cracks were established. The simulation environment was based on a laboratory-scale uniaxial compression test on rock samples. Uniaxial compression tests were conducted on square granite specimens with either a single prefabricated crack or a naturally developed crack. The model validity was discussed and further validated through comparative analysis. Subsequently, uniaxial compression tests under conditions with multiple randomly distributed cracks were simulated and compared with the previously established single-crack models. This comparison provided insights and guidance for the subsequent modeling of tunnel surrounding rock. Next, the study investigated the damage zones of tunnel surrounding rock under two conditions: intact rock mass and naturally fractured rock mass. The simulation was based on the physical and mechanical properties of granite from a tunnel test site in Gansu Province. A square physical model was established with a circular tunnel of 4 m diameter placed at the center to minimize boundary effects. The bottom of the model was fixed, and the rock mass was assumed to be homogeneous before excavation, with the tunnel wall free afterward. The model was subjected to uniform loading on all four sides to simulate isotropic in-situ stress conditions. The study analyzed the distribution of in-situ stress before and after excavation, as well as the development of damage under different stress scenarios. Results and Discussion The model proposed in this study produced results consistent with previous research. For the rock specimen with a 0° prefabricated crack, failure occurred slightly earlier in this simulation, but the overall strength and deformation behavior were similar. The peak strength of 45° prefabricated crack was slightly lower, but the overall trend remained consistent. The crack propagation cloud diagrams reveal that for the horizontal (0°) prefabricated crack at 25 s, crack initiation occurred at both tips of the central crack, though the changes were relatively minor and only noticeable at the tips. The crack had extended outward in a butterfly-shaped pattern at approximately 45° at 500 s, with pronounced shear zones forming at both ends. The lighter colors around the model boundary indicated ongoing stress redistribution during crack propagation. At 1 385 s, the specimen approached failure, and the crack pattern had developed into an X-shape, indicating impending shear failure. However, the red zone representing full fracture had not yet penetrated the model. At 1 402 s, the crack fully penetrated the specimen, resulting in failure. The final crack pattern exhibited a bifurcated X-shape, and the cracks near the left and right boundaries became slightly flatter. The asymmetric angle of the X-shaped cracks across the specimen was attributed to the boundary conditions. Reversing the roller and free constraints on the left and right sides would produce an opposite asymmetry, the simulated crack pattern was consistent with the double shear failure mode observed in their tests. In the uniaxial compression test with randomly distributed cracks, the load–displacement curve exhibited significant differences compared to that of the specimen with a single prefabricated crack. The peak strength was markedly reduced, reaching only about one-third of that of the single-crack case, and the displacement at failure was also smaller. These results indicate that the compressive strength of rock significantly decreases under randomly distributed crack conditions. Due to the concentration of random cracks in the lower region of the specimen, crack propagation was initiated in this area and extended along pre-existing cracks. The damage zone connected these cracks to form a continuous failure path, ultimately leading to specimen failure. The crack evolution showed a tendency to cluster and extend within the densely fractured region.For the case of intact tunnel surrounding rock, phase-field damage cloud maps were obtained under five different lateral pressure coefficient conditions: f = 0.6, 0.8, 1.0, 1.2, and 1.4. When f = 1.4, the excavation damaged zone (EDZ) appeared at both the crown and the invert of the tunnel, with a wider distribution range. When f = 1.2, the EDZ was still present at the crown and invert but gradually extended toward the horizontal directions of the tunnel, with a reduced overall damage area. Under the condition f = 1.0, the rock mass was homogeneous, and the EDZ developed into a symmetrical and uniform circular ring. As the coefficient decreased to f = 0.8, the EDZ shifted toward the tunnel sidewalls (springlines), and the extent of the damage increased. At f = 0.6, the EDZ was concentrated at the sidewalls with a more localized and confined damage distribution.For naturally fractured surrounding rock, numerical simulations showed that cracks primarily developed above the tunnel and connected with existing fractures. Damage was more severe when fractures were located closer to the tunnel. Three models with identical fracture patterns but different spatial scales confirmed that damage increased as fracture proximity to the tunnel decreased.Focusing on the near-field fracture model, simulations under varying lateral pressure coefficients (f = 0.6–1.4) revealed that tunnel stability decreased with increasing f. At f = 0.6, damage was minimal; at f = 1.4, complete tunnel failure occurred due to full crack penetration beneath the tunnel.A quantitative damage evaluation method was proposed by integrating phase-field values along a representative damage path. Larger integral areas indicated higher damage severity. Damage zoning was classified into crushed zones (phase-field = 1), fractured zones (0.3–1), and potential damage zones (<0.3). ConclusionsPhase field crack cloud diagrams for laboratory-scale samples align closely with actual failure modes. The extent of the EDZ phase-field damage in intact tunnels is positively correlated with the lateral pressure coefficient. With a constant number of natural fractures, the distribution range of fractures and the lateral pressure coefficient significantly influence the EDZ in fractured surrounding rock tunnels. The closer the fractures are to the excavation face, the more prominent the damage cloud diagram. As the lateral pressure coefficient increases from 0.6 to 1.4, the EDZ range expands. Furthermore, a method for EDZ hazard zoning based on trace line paths and area was proposed and applied, and its effectiveness was discussed. This study provides essential methods and a basis for predicting fracture propagation in tunnel surrounding rock during excavation. |
|---|---|
| ISSN: | 2096-3246 |