Experimental study on the formation mechanism of landslides considering the spatial distribution of the locking section

The evolution processes of locked-segment landslides are difficult to determine, causing severe economic losses and casualties. Furthermore, when the locked-segment landslide is excavated, its evolution process will become more complicated. However, there have been few studies on the identification criteria for the evolution state of locked-segment landslides. In this case, this paper adopted the model test to analyze the evolution process, deformation patterns, and stress distribution of locked-segment landslides, and the identification criteria for the evolution state were proposed. Specifically, for non-locked landslides, it presented general push-type progressive failure with multiple slip surfaces. The shape, depth, and deformation amount of the slip surfaces at each section of the slope were basically the same. As for the locked-segment landslide with a lower locking ratio (15%), the push-torsion progressive failure occurred, accompanied by multiple slip surfaces and non-cooperative deformation. The locked-segment landslide with a higher locking ratio (30%) presented as shallow push-type progressive failure. The shape, depth, and deformation amount of the slip surfaces at each section were also basically the same. The research results could provide significant reference for the identification, prevention, and early warning of locked-segment landslides.

1 Introduction

Landslide disasters seriously threaten the safety of people’s lives and property, major engineering construction, and the ecological environment (Highland and Bobrowsky, 2008; Zhang et al., 2015; Juang et al., 2022). The locking section refers to the part where the sliding surface is not connected during the instability of the slope, which bears stress concentration and provides a crucial bearing function. It is a high-strength rock whose strength is significantly higher than the surrounding slope materials (Huang et al., 2017). The main inducing factors of slope instability are rainfall, earthquakes, and human activities, among which human activities mainly involve excavating slopes (Hunt, 2007; Sun et al., 2017; Yin et al., 2017a; Peng et al., 2019). The rock landslides in the locked section are the most common type of large-scale catastrophic landslides. Their development and evolution are characterized by significant concealment, suddenness, and destructiveness (Huang et al., 2017), involving the complex mechanical mechanisms of rocks (Lin et al., 2021; 2025). Large-scale rock landslides are generally accompanied by sudden brittle failure of the “locking section” along the potential sliding surface (Huang, 2012). Due to the limited detection technology, it is difficult to determine the locked segment through field investigation before the landslide failure (Zhou et al., 2022). The locked-segment landslide also has an energy accumulation effect, which makes it easy to form a high-speed and long-runout landslide. Due to the instantaneous rupture of the locked segment in the sliding source area, and the long-term rainfall softened the sliding zone and the locked segment, the Xinmo landslide was formed on 24 June 2017. This landslide destroyed Xinmo Village and caused 83 deaths (Liu et al., 2018; Yin et al., 2017b; Wang et al., 2018; Hu et al., 2018).

In the existing research, scholars generally use theoretical analysis, model testing, numerical simulation, and other methods to analyze the locked-segment landslides. Qiao et al. (2022) established a sharp point mutation prediction model for the failure of locked slopes through theoretical analysis and derived mutation characteristic values. Based on the principle of energy conversion, Yang et al. (20222), Yang et al. (2022b) proposed an equation for the radiated seismic energy of a locked-segment cracking event. Hu et al. (2023) conducted a comprehensive analysis of the instability sources, mechanisms, and processes of high and steep slopes with front-edge locked segments.

Based on real-time transmission equipment, in-situ observation equipment, and a remote sensing platform, Lu et al. (2015) constructed a spatial monitoring network to monitor the dynamic evolution characteristics of landslides and verified the reliability of the system in model tests. Zhu et al. (2015) explored the evolution process of slope reinforcement through the optical fiber sensor network embedded in the slope model. Achache et al. (1996) used D-InSAR technology to study landslides and realized all-day, all-weather, and large-scale monitoring of geological hazards.

Liu et al. (2020) and Liu et al. (2023) simultaneously tested three slope models with different types of locking sections and studied the effects of three locking section configurations on failure time and deformation behavior under the same rainfall conditions. Wang J. C. et al. (2024) conducted physical model experiments on layered landslides with different locking structures and found that their failure mode was determined by the strength of the rock mass. The high-strength landslide was destroyed in a bending form, while the low-strength rock slide experienced shear failure. Wang L. et al. (2024) used the self-developed model test device to simulate the whole process of the deformation evolution of the locked segment slope of the supporting arch.

Due to the complexity of slope stability issues in engineering construction, the application of numerical simulation is becoming increasingly widespread. Eberhardt et al. (2004) studied the mechanism of slope instability under the control of multi-stage locking segments by numerical simulation. Huang (2007) used the finite element simulation method to analyze the stress path of the locking section. Wang H. et al. (2024) used numerical simulation methods to analyze the stability impact of oblique cut rock slopes in rainfall conditions. Tang et al. (2015) established the crack formation and propagation equation based on time and space position, and revealed the temporal and spatial evolution law of the Jiweishan landslide under the action of the leading edge anti-slide block.

In summary, previous studies have conducted extensive research on the formation mechanisms as well as instability characteristics of landslides in the locked segment, but most of them ignored the identification criteria for the evolution state of landslides in the locked segment, especially for the identification criteria and stage division of the evolution state of excavation-type locked landslides. This paper adopted a model test to analyze the evolution process, formation mechanism, and stress distribution of the locking segment slope, and a specific identification criterion of the evolution state was proposed. The relevant results provide significant practical application value for the stability evaluation and early warning of locked-segment landslides.

2 Generalized geo-mechanical model

2.1 Physical modelling

The geological generalization of the slope structure was carried out, and its main characteristics were as follows: along the slope direction, the high-strength blocking body is distributed locally through the potential sliding surface in the middle and front section of the slope, which supported the deformation body of the upper slope, caused stress concentration, and produced locking effect similar to ‘retaining wall’. The upper deformation zone was relatively wide. Due to the combination of a high-strength barrier and lower limestone outburst (boundary), the shear outlet of the front edge of the slope was divided into two asymmetric sections. The deformation zone in the middle and lower part became narrow, and one side produced a locking effect similar to a ‘support arch’. In this paper, the above slope structure characteristics were identified as the locking segment slope structure. Figure 1 shows the generalized geological model of the slope structure of the locking segment.

Figure 1

Figure 1. Generalization of the geological model of locking segment slope structure.

Based on the generalized model, three groups of large-scale physical model tests were carried out to construct the relationship between the locked segment and the evolution process, determining the excavation deformation of the locked slope. Among them, the first group was mainly aimed at the slope without a locked section, which was the comparison of the other two groups of tests to reveal the differences between slopes with and without locked sections. By changing the lateral locking ratio (the proportion of the locking section length Y in the width of the front edge of the slope), of the locked section (15% and 30%, respectively), the influence of the ratio of the locked section on the evolution characteristics of the slope under the same conditions was discussed. During the test, high-definition cameras and cameras were used to obtain deformation images and surface cracks.

2.2 Experimental setup

To study the influence of the lateral locking ratio of the locking section on the deformation and failure of slope excavation, it is necessary to establish the relationship between the locking ratio of the locking section and the spatial evolution characteristics of slope instability. Three sets of model test schemes (Table 1) were designed, in which Group I had no locking section, Group II was generalized by the prototype slope, and Group III was to change the lateral locking ratio of the locking section. The influence of different locking section ratios on the evolution characteristics of slope excavation deformation under the same conditions was discussed.

Table 1

Table 1. Model test scheme design (Unit: cm).

The model of Group I was 2.0 m × 1.0 m × 0.99 m (length × width × height), and the width of the trailing edge platform was 20 cm. Three levels of slope excavation were carried out on the slope. The slope height of each level was 24 cm, the width was 30 cm, and the width of the roadway was 5 cm.

The longitudinal monitoring profiles were set up at 30 cm, 60 cm, and 90 cm, respectively. Aluminum strips of different lengths were buried on the longitudinal monitoring profiles to record the deep deformation of the slope. The deformation points from left to right and from top to bottom were recorded as deformation points 1∼12.

The d-d' and e-e' transverse monitoring profiles were set up at 87.5 cm and 52.5 cm. Along the d-d' section, three soil pressure boxes were buried at 15 cm, 30 cm, 60 cm, and 90 cm, with an interval of 6 cm. The numbers were AEP15-1∼3, AEP30-1∼3, AEP60-1∼3, and AEP90-1∼3 from top to bottom and from left to right. Along the d-d' section, three soil pressure boxes were buried at 30 cm and 60 cm, with an interval of 10 cm, numbered BEP30-1∼3 and BEP60-1∼3, respectively.

At 30 cm and 60 cm of two levels of roadway, dipmeters 1∼4 (the tilt sensor for measuring slope inclination) were placed from left to right and from top to bottom to record the change of slope inclination.

Four rows of slope deformation monitoring points (10 cm interval in the Y direction) were set on the surface of each slope to record the surface displacement during the deformation evolution process of the slope.

Figure 2a was an unlocked slope model. Figure 2b was the slope excavation model (Group II) with the locking section ratio (15%), where the locking section was located between 60 cm and 75 cm in the Y direction of the slope front edge. Based on the monitoring contents and methods of test, Group I, six strain gauges were placed at intervals of 6 cm, 8 cm, and 6 cm from top to bottom along the 70 cm axis on the back of the locking section, and six strain gauges were placed along the 65 cm in the Y-direction. According to the micro-strain situation of the strain gauge, the deformation of the soil behind the locking section and the stress state of the locking section can be analyzed.

Figure 2

Figure 2. (a) Unlocked slope model, (b) slope with locking section ratio (15%), (c) slope model with locking section ratio (30%).

Figure 2c was the slope excavation model (Group III) with a locking section ratio (30%). The monitoring contents and methods were exactly the same as Group II, but the Y-direction, 45 cm–75 cm at the front edge of the slope, was the locking section. The strain gauges on the back of the locking section were arranged on axes of 70 cm, 60 cm, and 50 cm.

The high-definition cameras were used to obtain the deformation images and surface crack information during the test. Strain gauges (Figure 3a) were pasted at different depths on the back of the locking section to measure the strain at different positions. The resistance of the strain gauges is 120.0 Ω, and their sensitivity coefficient is 2.0%. Soil pressure cells (Figure 3b) were buried at different sections and depths in the slope to measure the change of slope stress. The diameter and thickness of soil pressure cells were 12.00 mm and 4.25 mm, their testing range was from 0 to 20 kPa, and the accuracy is 0.5%. The automatic data acquisition system (DHS3821) was used to record the changes of strain and earth pressure, and the monitoring frequency is 1 Hz. To reduce the influence of the model filling process on the sensors, after the completion of the model filling, the sensors were buried at predetermined intervals (Figure 3c). After sticking the strain gauge on the back of the locking section with glue, the locking section was placed into the slope as a whole. By backfilling with fine sand, the slope mass is ensured in close contact with the locking section (Figure 3d).

Figure 3

Figure 3. Sensors for model tests. (a) Resistance strain gauge (b) Soil pressure cells (c) The place of soil pressure cells (d) The place of strain gauges.

3 Evolution characteristics of slope deformation

To compare the similarities and differences in the evolution process of slope morphology among the three models, the corresponding failure modes, failure moments, and other information of each model experiment were summarized in Table 2.

Table 2

Table 2. Comparative analysis of slope morphology evolution characteristics.

3.1 Changes in slope inclination angle

Figures 4–6 showed the changes of slope inclination angles at different times and different positions of the slope model with locking section ratios of 0, 15%, and 30%, respectively. Overall, the three models all showed that dipmeters 1 and 2 (upper part of the slope) undergo large-angle tilting deformation first, followed by dipmeters 3 and 4 (lower part of the slope) experiencing tilting deformation.

Figure 4

Figure 4. The change in inclination angle of Test I. (a) The change of inclination angle in the X-direction (b) The change of inclination angle in the Y-direction.

Figure 5

Figure 5. The change in inclination angle of Test II. (a) The change of inclination angle in the X-direction (b) The change of inclination angle in the Y-direction.

Figure 6

Figure 6. The change in the inclination angle of Test III. (a) The change of inclination angle in the X-direction (b) The change of inclination angle in the Y-direction.

For the slope without a locking section in Test I, the deformation in the X direction was very small throughout the entire process, about 0∼1.5°. The dipmeter 1 began to accelerate the deformation at 15 min, and reached the maximum range of 10° at 60 min (the corresponding X-direction inclination angle was 1.5°), and the average deformation rate was 0.22°/min. The dipmeters 3 and 4 began to accelerate the deformation at 105 min. Within 15 min, the Y-direction inclination angles increased rapidly from 1.5° to 0.5°–8° and 9.5°, respectively, and the average deformation rates were 0.43°/min and 0.6°/min, respectively.

Compared with Test I, the maximum deformation of the Test II locking segment (Locking Ratio is 15%) slope in the X direction could reach 7.0°, and the horizontal deformation increased significantly. In Test II, dipmeter 1 began to accelerate the deformation at T = 25 min, and the deformation in the X and Y directions reached 1.5° and 5°, respectively, at 120 min, accelerating the failure. The dipmeter 2 began to accelerate the deformation at T = 40 min, and the deformation in X and Y directions increased from 0° to 1° and 9.5° within 5 min, respectively. The dipmeter 3 had the largest deformation in the X direction, which could reach 5°, and the Y direction deformation reached 10° full range. Because the dipmeter 4 was close to the upper part of the locking section, the deformation in both directions was small, about 0.5° in the X direction and about 1.5° in the Y direction. Judging from the degree of damage, dipmeter 2 was larger than dipmeter 1, and dipmeter 3 was larger than dipmeter 4. Through a comprehensive comparison of Tests I and II, the locking section made the slope deformation body produce a large tilt change in the horizontal direction, forming a spatial slip-rotation failure around the flank of the locking section.

The results of Test III were similar to those of Test II. Except for the dipmeter 4 above the locking section, the other three measuring points had large inclination deformation in the X direction. The maximum inclination angles of the dipmeters 1∼3 in the X direction were 5°, 3.5°, and 2.5°, respectively. It showed that under the influence of the locking section, the upper deformation body of the slope also produced horizontal deformation during the forward sliding process, and the locking section significantly changed the spatial migration trend of the slope.

3.2 Deep deformation

From the deep deformation curve of the slope model without the locking section in Figure 7, it can be seen that the shape and buried depth of the two-stage sliding surface at three typical sections were basically the same. The burial depth of the first level sliding surface was within the range of 55∼75 cm in height, and the shear outlet was located at a height of 60 cm on the second level slope. The deformation area was above the shear outlet of the second-level slope. The buried depth of the second-level sliding surface was within the range of 35 cm–55 cm in height, and the shear outlet was located at a height of 35 cm on the first-level slope. The deformation area expanded to the upper part of the first-level slope. The maximum deformation of deep deformation measuring points 1∼4 at section a-a' was 16.6 cm, 15.8 cm, 7.0 cm, and 4.8 cm, respectively. The maximum deformation of deep deformation measuring points 5 ∼ 8 at section b-b' was 16.7 cm, 13.1 cm, 10.8 cm, and 4.1 cm, respectively. The maximum deformation of deep deformation measuring points 9∼12 at section c-c' was 16.8 cm, 11.3 cm, 8.3 cm, and 3.8 cm, respectively. The deformation and deformation trend of the three sections at the same depth were basically the same, indicating that the slope without a locking section produced progressive failure from top to bottom under load.

Figure 7

Figure 7. Deep deformation curves of slope (Test I).

Figure 8 was the deep deformation curve of the slope model of the locked section (Locking Ratio is 15%). Although the model also developed two levels of sliding surfaces, Test II had significant differences in sliding surface depth at different cross-sections due to the influence of the locking section. Based on the two levels of sliding surfaces at the section a-a', the first and second sliding surfaces in the section b-b' moved up by 5 cm and 6 cm. Compared with the section a-a', the section c-c' moved up by 9 cm and 6 cm. During the evolution of the upper deformation body, it twisted and pushed towards the left side of the model, causing the depth of the sliding surface of the section a-a' to be relatively deepest. For the monitoring point at the front edge of the first-level sliding surface, the maximum displacement of the measuring point 4 (on section a-a') was about 2.7 cm, and the maximum deformation depth was about 15.5 cm. The section b-b' was a locking section at the first-level slope, and the deformation depth and deformation were both 0. The maximum displacement of measuring point 12 (section c-c') was 0.6 cm, and the maximum deformation depth was 2.0 cm. The first-level slope on the left side of the locking section was obviously damaged, while the right side of the locking section was relatively stable, and the deformation was very small. It could be seen that the spatial slip-rotation failure mode of the locked slope was significantly different from the overall progressive failure of the unlocked section.

Figure 8

Figure 8. Deep deformation curves of slope (Test II).

Figure 9 shows the deep deformation curve of the locked segment (the locking Ratio is 30%) slope model. The slope model only formed a shallow sliding surface. The shape and depth of the sliding surface were basically the same at the sections a-a' and b-b', and the section c-c' was moved up by about 8 cm. Compared with the Test II, due to the increase of the ratio of the locking section, the overall stability of the slope was improved accordingly, and the deep sliding surface was not formed. Affected by the wings on both sides of the locking section, the deformation and failure mode of the slope were different from the overall thrust failure and the flank rotation failure of the multi-stage sliding surface, which were manifested as the shallow non-integral thrust failure with oblique cracks in different parts of the slope.

Figure 9

Figure 9. Deep deformation curves of slope (Test III).

4 Stress evolution characteristics of slope

Figure 10 was the time-history curve of soil pressure at different positions of the slope model (Test I) with the ratio of the locking section of 0. In the early stage of loading, the soil pressure values of BEP30-1 and BEP60-1 on the second-level roadway rapidly increased and were much higher than those on the first-level roadway, indicating that the load transfer path was gradually pushed down from the upper part of the slope. In the depth direction of the e-e' profile, BEP30/60–1 (Shallow) was significantly larger than BEP30/60–2/3 (Deep), and the maximum load transfer depth was between 10 and 20 cm below the surface, which was basically consistent with the shallow slip surface depth (18∼24 cm below the surface).

Figure 10

Figure 10. Time history curves of soil pressure in Test I.

The stress response characteristics of the slope without the locking section in Test I can be concluded. The shallow deformation area in the upper part of the slope was under the external loads, and the soil pressure increased accordingly. With the penetration of the shallow sliding surface, the bearing capacity of the shallow deformation body was greatly reduced, and the load was transferred to the deep layer, increasing soil pressure in the lower part of the slope. The deep sliding surface gradually formed, and a leading-edge shear outlet was formed on the surface of the first-level slope.

Figure 11 was the time-history curve of earth pressure at different positions of Test II with the length of the locking section of 15 cm. Similar to Test I, the soil pressure in the upper part of the slope increased first. With the penetration of the shallow sliding surface, the load gradually transferred to the deeper part, and the soil pressure in the lower part of the slope increased accordingly. However, due to the existence of the locking section, there were still significant differences in the variation and distribution of soil pressure compared to slopes without locking sections.

Figure 11

Figure 11. Time history curves of soil pressure in Test II.

The stress response characteristics of the locked slope in Test II were as follows: the shallow deformation area in the middle and upper part of the slope was under the external load first, and a high stress concentration along the longitudinal section of the locking section was formed, resulting in unbalanced stress distribution inside the deformation area and the formation of tension cracks. With the penetration of the shallow sliding surface, the bearing capacity of the shallow deformation area was greatly reduced, and the load was transferred to the deep layer. Affected by the locking section, differential deformation occurred on both sides of the wing, thus forming an asymmetric distribution of stress.

The evolution law of soil pressure in Test III was shown in Figure 12. Similar to Test II, the load was gradually transferred from top to bottom. The stress concentration in the longitudinal section where the locking section was located was significantly higher than that in the wing areas on both sides of the locking section.

Figure 12

Figure 12. Time history curve of soil pressure in Test III.

5 Discussion

It could be found that the identification criteria for the evolution state of the locked segment landslides could be constructed according to the deformation characteristics and stress distribution characteristics, which provided theoretical support for the identification and reasonable engineering disposal of the field locking section landslides.

The deformation differences of different locking section slopes could be compared and analyzed from three aspects: the shape, the dip angle, and the deep deformation of the slope, so as to construct the deformation identification criteria of the evolution state of the locking section slope.

5.1 The shape of the slope

For the slope without a locking section in Test I, the shallow overall thrust failure occurred first. As the bearing capacity of the upper deformation body was greatly reduced, the load was transferred to the deep soil, the sliding surface moved down, forming deep sliding, and the slope was gradually deformed from top to bottom. The results are similar to those of Mu et al. (2022). The cracks on the slope were mainly transverse through cracks, without longitudinal tensile cracks, and the shape of each part of the slope was similar.

For the locking section slope of Test II (the ratio of locking section is 15%), the shallow non-integral thrust failure occurred first. The deformation of the left wing of the locking section was large, and the deformation of the top and right wings was small. Longitudinal tensile cracks were formed on the slope surface. As the sliding surface continued to move down, the sliding of the top of the locking section and its right flank was blocked, and the differential deformation of the slope surface was further aggravated. The longitudinal tensile cracks along the left flank of the locking section were continuously extended and widened. The slope was twisted to the left flank of the locking section, showing a slip-rotation failure pattern in three-dimensional space.

Test III (the ratio of locking section is 30%) was obviously different from the first two groups of slope failure modes. Due to the increase in the length of the locking section, the overall stability of the locking section slope was significantly improved, and no deep sliding surface was formed. However, due to the anti-sliding effect of the locking section, a large number of oblique tension cracks were formed on the surface of the shallow deformation body, which cut the deformation area of the slope into a fragmented shape. At the same time, the failure phenomenon of local uplift of the leading edge was formed within a certain height range (about 8∼12 cm) at the top of the locking section.

5.2 The dip angle of the slope

The three models all showed that the middle and upper part of the slope underwent large-angle tilt deformation, and then the lower part of the slope produced tilt deformation. For the slope without a locking section in Test I, the X-direction deformation of the whole process was very small, about 0∼1.5°. The maximum deformation of the slope in the X direction in Test II could reach 7.0°, and the horizontal deformation increased significantly. The locking section made the deformation body tilt in the horizontal direction, forming a spatial slip-rotation failure around the flank of the locking section. The results of Test III were similar to those of Test II. The maximum inclination angle in the X direction was about 5°. The locking section significantly changed the spatial migration law of slope deformation.

5.3 The deep deformation of the slope

The slope without a locking section produced an overall progressive failure from top to bottom. The deformation and deformation trend of each section at the same depth, and the depth of the sliding surface, were also basically the same. In the same section, the deformation of the trailing edge of the slope was the largest, the middle was the second, and the front edge was the smallest.

The depth of the sliding surface at each longitudinal section of the locking section (the locking Ratio is 15%) was quite different. Based on the two-stage sliding surface at the a-a' section, the first and second sliding surfaces in the b-b' section moved up by 5 cm and 6 cm, respectively, compared with the a-a' section, and the c-c' section moved up by 9 cm and 6 cm. From a horizontal point of view, the deformation and deformation depth of the wing area on both sides of the locking section were also different. The maximum displacement of the left wing area was about 2.7 cm, and the maximum deformation depth was about 15.5 cm. The maximum displacement of the right flank area was about 0.6 cm, and the maximum deformation depth was about 2.0 cm. The first-level slope on the left side of the locking section had suffered significant damage, while the right side of the locking section was relatively stable. Similar to the slope without locking section, in the same section, the deformation of the trailing edge of the slope was the largest, the middle was the second, and the front edge was the smallest.

Due to the further increase in the length of the locking section (the locking ratio is 30%), the overall stability of the slope was improved, and the deep sliding surface was not formed, similar to the results of Liu et al. (2021). Only the shallow sliding on the upper part of the locking section was developed. The depth of the sliding surface of each section was basically the same, and the deformation was also gradually reduced from top to bottom.

The stress distribution of different locking section slopes could be compared and analyzed, so as to confirm the stress identification criteria of the evolution state of the locking section slope. For the slope without a locking section, the shallow deformation body in the middle and upper part of the trailing edge of the slope first bears the external load, and the soil pressure increases accordingly. With the penetration of the shallow sliding surface, the bearing capacity of the shallow deformation body was greatly reduced, and the load was transferred to the deep layer, and the soil pressure in the middle and lower part of the slope increased accordingly. On the cross-section, the soil pressure in each part was relatively balanced, and there were no obvious stress concentration sections. Taking the d-d' section (at the first-level roadway) as an example, the soil pressures of AEP15-1, AEP30-1, AEP60-1, and AEP90-1 were 1.8 kPa, 1.4 kPa, 4.0 kPa, and 2.0 kPa, respectively.

The locking section slope with a locking ratio of 15% was also subjected to the upper part of the slope in the longitudinal direction. With the penetration of the shallow sliding surface, the load was transferred to the deep layer, and the soil pressure in the middle and lower parts of the slope increased accordingly. The difference was that under the influence of the locking section, a high stress concentration was formed along the longitudinal section of the locking section, and an unbalanced stress distribution was formed inside the deformable body, resulting in differential deformation on both sides of the wing. With the aggravation of differential deformation, the asymmetric distribution characteristics of stress were further changed. Obviously, the stress level of the b-b' locking section was the largest, and the a-a' section and c-c' section on both sides were significantly smaller. The peak stresses of the three measuring points along the d-d' cross section were 1.8 kPa (AEP30), 40 kPa (AEP67) and 8.2 kPa (AEP90), respectively, and the stress ratio was 1:22.2:4.6. In addition, from the micro-strain monitoring results of the back of the locking section, it could be found that the failure side of the locking section was subjected to tensile force, and the undamaged side was subjected to extrusion force.

The variation law of soil pressure with a locking ratio of 30% was consistent with the overall law of Test II. However, due to the increase in the length of the locking section, the high stress area at the back of the locking section further developed into the deep part. The peak stresses of the measuring points on the d-d' cross section were 2.0 kPa (AEP30), 40 kPa (AEP67), and 6.0 kPa (AEP90), respectively, and the stress ratio was 1:20:3. Compared with Test II, the degree of stress concentration was slightly reduced. From the micro-strain monitoring results of the back of the locking section, it could be found that the back of the locking section was in a compressed state.

This paper mainly focuses on the case where the locking section is located at the front edge of the slope. Due to the complexity and heterogeneity of the geological conditions, the locking section can be located in the middle of the slope. The failure mode and evolution mechanism of the slope are not consistent with different slope structures. Limited by the existing research conditions, although the geometric size of the large-scale physical model test carried out in this paper is large, the results are still affected by the boundary of the model test. In future work, experiments and numerical calculations with higher locking ratios would be conducted to establish a comprehensive quantitative evaluation framework that can be promoted and applied to more complex working conditions.

6 Conclusion

Through the large-scale physical model tests, the influences of different locking section ratios on the evolution characteristics of slope excavation deformation were quantified. Specifically, the failure mode of the non-locking section landslide was the progressive failure of the multi-stage sliding surface. The failure mode of the locked segment landslide (with a locking ratio of 15%) was multi-stage sliding surface push-torsion progressive failure. And the failure mode of the locked segment landslide (with a locking ratio of 30%) was shallow progressive failure.

Moreover, the conversion thresholds of various evolution stages of the landslides were accurately captured, providing valuable reference for a better understanding of the formation mechanism of locked-segment landslides. By combining numerical calculations and comprehensive monitoring techniques, real-time warning of locked-segment landslides can be further achieved.

Data availability statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author contributions

XL: Conceptualization, Formal Analysis, Writing – review and editing, Data curation, Writing – original draft. DiL: Writing – review and editing, Methodology, Data curation. JK: Data curation, Writing – review and editing, Formal Analysis. ZZ: Writing – review and editing, Data curation, Methodology. DoL: Writing – review and editing, Methodology, Conceptualization.

Funding

The author(s) declared that financial support was not received for this work and/or its publication.

Conflict of interest

Author XL was employed by PowerChina Jiangxi Electric Power Engineering Co., Ltd.

The remaining author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Achache, J., Fruneau, B., and Delacourt, C. (1996). Applicability of SAR interferometry for monitoring of landslides. Ers Appl. 383, 165.

Google Scholar

Eberhardt, E., Stead, D., and Coggan, J. S. (2004). Numerical analysis of initiation and progressive failure in natural rock slopes-the 1991Randa rockslide. Int. J. Rock Mech. Min. Sci. 41 (1), 69–87. doi:10.1016/s1365-1609(03)00076-5

CrossRef Full Text | Google Scholar

Highland, L., and Bobrowsky, P. T. (2008). The landslide handbook - a guide to understanding landslides. Denver: Denver Publishing Service Center.

Google Scholar

Hu, K. H., Wu, C. H., Tang, J. B., Pasuto, A., Li, Y., and Yan, S. (2018). New understandings of the June 24th 2017 xinmo landslide, maoxian, Sichuan, China. Landslides 15 (12), 24665. doi:10.1007/s10346-018-1073-2

CrossRef Full Text | Google Scholar

Hu, H. H., Lu, G. Y., Chen, Y. F., Yin, X. J., Zhang, P., Lin, H., et al. (2023). Identification of instability sources and evolution process simulation for leading-edge locking canyon slopes. J. Central South Univ. Sci. Technol. 54 (11), 4461–4471. doi:10.11817/j.issn.1672-7207.2023.11.023

CrossRef Full Text | Google Scholar

Huang, R. Q. (2007). Large-scale landslides and their mechanisms in China since the 20th century. J. Rock Mech. Eng. (03), 433–454. (in Chinese). doi:10.3321/j.issn:1000-6915.2007.03.001

CrossRef Full Text | Google Scholar

Huang, R. Q. (2012). Mechanisms of large-scale landslides in China. Bull. Eng. Geol. Environ. 71, 161–170. doi:10.1007/s10064-011-0403-6

CrossRef Full Text | Google Scholar

Huang, R. Q., Chen, G. Q., and Tang, P. (2017). Study on the precursor information of locked segment rock landslide based on dynamic evolution characteristics. J. Rock Mechanics Engineering 36 (03), 521–533. (in Chinese). doi:10.13722/j.cnki.jrme.2016.1100

CrossRef Full Text | Google Scholar

Hunt, R. E. (2007). Geologic hazards: a field guide for geotechnical engineers. Boca Raton: USA: CRC Press.

CrossRef Full Text | Google Scholar

Juang, C. H., Gong, W. P., and Wasowski, J. (2022). Trending topics of significance in engineering geology. Eng. Geol. 296, 106460. doi:10.1016/j.enggeo.2021.106460

CrossRef Full Text | Google Scholar

Lin, Q., Cao, P., Wen, G., Meng, J., Cao, R., and Zhao, Z. (2021). Crack coalescence in rock-like specimens with two dissimilar layers and pre-existing double parallel joints under uniaxial compression. Int. J. Rock Mech. Min. Sci. 139, 104621. doi:10.1016/j.ijrmms.2021.104621

CrossRef Full Text | Google Scholar

Lin, Q., Zhang, S., Lin, H., Zhang, K., Fan, W., Huang, C., et al. (2025). Failure behavior of jointed rock masses containing a circular hole under compressive-shear load: insights from DIC technique. Theor. Appl. Fract. Mech. 139, 105089. doi:10.1016/j.tafmec.2025.105089

CrossRef Full Text | Google Scholar

Liu, H. D., Zhang, Y. B., and Lu, L. P. (2018). Locked landslide types in Western Henan in China. J. North China Univ. Water Resour. Electr. Power Nat. Sci. Ed. 39 (06), 1–7. (in Chinese). doi:10.3969/j.issn.1002-5634.2018.06.001

CrossRef Full Text | Google Scholar

Liu, H. D., Li, D. D., Wang, Z. F., Geng, Z., and Li, L. D. (2020). Physical modeling on failure mechanism of locked-segment landslides triggered by heavy precipitation. Landslides 17 (2), 459–469. doi:10.1007/s10346-019-01288-3

CrossRef Full Text | Google Scholar

Liu, H. D., Guo, Z. F., and Chen, J. X. (2021). Experimental study on failure mechanism of arch locked segment landslides. J. North China Univ. Water Resour. Electr. Power Natural Sci. Ed. 42 (4), 63–69. (in Chinese). doi:10.19760/j.ncwu.zk.2021051

CrossRef Full Text | Google Scholar

Liu, H. D., Liu, J. J., Chen, J. X., Guo, Z. F., and Qiu, L. (2023). Experimental study on tilting deformation and a new method for landslide prediction with retaining-wall locked segment. Sci. Rep. 13 (1), 5149. doi:10.1038/s41598-023-32477-9

PubMed Abstract | CrossRef Full Text | Google Scholar

Lu, P., Wu, H., Qiao, G., Li, W., Scaioni, M., Feng, T., et al. (2015). Model test study on monitoring dynamic process of slope failure through spatial sensor network. Environ. Earth Sci. 74 (4), 3315–3332. doi:10.1007/s12665-015-4369-8

CrossRef Full Text | Google Scholar

Mu, C. L., Pei, X. J., and Wang, R. W. (2022). Deformation and damage characteristics analysis of high slopes with multi-layered soft interlayer down-excavation based on physical model test. Chin. J. Geol. Hazards Prev. 33 (3), 61–67. doi:10.16031/j.cnki.issn.1003-8035.2022.03-07

CrossRef Full Text | Google Scholar

Peng, J. B., Wang, S. K., Wang, Q. Y., Zhuang, J., Huang, W., Zhu, X., et al. (2019). Distribution and genetic types of loess landslides in China. J. Asian Earth Sci. 170, 329–350. doi:10.1016/j.jseaes.2018.11.015

CrossRef Full Text | Google Scholar

Qiao, C., Wang, Yu., Li, C. H., Yan, B., Yang, H., and Xiao, Y. (2022). Catastrophe instability analysis of rock slopes with locked segments in open-pit mine due to freeze-thaw weathering. Bull. Eng. Geol. Environ. 81 (4), 135. doi:10.1007/s10064-022-02635-w

CrossRef Full Text | Google Scholar

Sun, P., Li, R., Jiang, H., Igwe, O., and Shi, J. (2017). Earthquake-triggered landslides by the 1718 tongwei earthquake in Gansu province, northwest China. Bull. Eng. Geol. and Environ. 76 (4), 1281–1295. doi:10.1007/s10064-016-0949-4

CrossRef Full Text | Google Scholar

Tang, H. M., Zou, Z. X., Xiong, C. R., Wu, Y., Hu, X., Wang, L., et al. (2015). An evolution model of large consequent bedding rockslides, with particular reference to the jiweishan rockslide in southwest China. Eng. Geol. 186 (24), 17–27. doi:10.1016/j.enggeo.2014.08.021

CrossRef Full Text | Google Scholar

Wang, Y., Zhao, B., and Li, J. (2018). Mechanism of the catastrophic June 2017 landslide at xinmo village, songping river, Sichuan Province, China. Landslides 15, 333–345. doi:10.1007/s10346-017-0927-3

CrossRef Full Text | Google Scholar

Wang, J. C., Chen, G. Q., and Xu, Q. (2024a). Seismic signal characteristics and precursory information of bedding landslide with locked segments. Eng. Geol. 341, 107709. doi:10.1016/j.enggeo.2024.107709

CrossRef Full Text | Google Scholar

Wang, L., Jia, H., Jiang, T., Zhang, J., Jia, Y., Li, L., et al. (2024b). Novel evaluation method based on critical arch height as instability criterion for sustaining arch locked-segment-type slopes. Sci. Rep. 14 (1), 7991. doi:10.1038/s41598-024-58737-w

PubMed Abstract | CrossRef Full Text | Google Scholar

Wang, H., Yuan, G. X., Huang, Z. Q., Dong, J., and Wei, Y. (2024c). Study on deformation and failure characteristics of oblique-cut locked rock slope under rainfall conditions. Sci. Reports 14 (1), 14159. doi:10.1038/s41598-024-64329-5

PubMed Abstract | CrossRef Full Text | Google Scholar

Yang, B. C., Bai, J. X., Duan, Y. T., and Wang, Z. (2022a). The unlocked mechanism and instability prediction of a typical locked-segment-type slope in China: the chana landslide. Bull. Eng. Geol. Environ. 81 (12), 493. doi:10.1007/s10064-022-03001-6

CrossRef Full Text | Google Scholar

Yang, B. C., Bai, J. X., Duan, Y. T., and Cheng, C. (2022b). A new method of seismic source parameter estimation of a locked-segment cracking event. Geofluids 2022 (1), 4224176–4224179. doi:10.1155/2022/4224176

CrossRef Full Text | Google Scholar

Yin, Y. P., Wang, W. P., Zhang, N., Yan, J., and Wei, Y. (2017a). The June 2017 maoxian landslide: geological disaster in an earthquake area after the wenchuan Ms 8.0 earthquake. Sci. China Technological Sci. 60 (11), 170–174. doi:10.1007/s11431-017-9148-2

CrossRef Full Text | Google Scholar

Yin, Y. P., Wang, W. P., Zhang, N., Yan, J. K., Wei, Y. J., Yang, L. M., et al. (2017b). Long runout geological disaster initiated by the ridge-top rockslide in a strong earthquake area: a case study of the xinmo landslide in maoxian county, Sichuan Province. Geol. China 44 (5), 827–841. doi:10.12029/gc20170501

CrossRef Full Text | Google Scholar

Zhang, M., Yin, Y. P., and Huang, B. L. (2015). Mechanisms of rainfall-induced landslides in gently inclined red beds in the eastern sichuan basin, SW China. Landslides 12 (5), 973–983. doi:10.1007/s10346-015-0611-4

CrossRef Full Text | Google Scholar

Zhou, Y., Zhao, X., Zhang, J., and Meng, M. (2022). Identification of a locking segment in a high-locality landslide in shidaguan, southwest China. Nat. Hazards 1-23. doi:10.1007/s11069-021-05162-1

CrossRef Full Text | Google Scholar

Zhu, H. H., Shi, B., Yan, J. F., Zhang, J., and Wang, J. (2015). Investigation of the evolutionary process of a reinforced model slope using a fiber-optic monitoring network. Eng. Geol. 186, 34–43. doi:10.1016/j.enggeo.2014.10.012

CrossRef Full Text | Google Scholar