Introduction

At present, in China, most mines have already entered the stage of deep mining. Numerous roadways situated in deep soft rock are confronted with many engineering problems. Floor heave is a prevalent engineering issue in deep soft rock roadways. the deformation of soft rock roadway caused by floor heave is a major challenge for coal mines in China western mining areas1. Floor heave seriously affects the safety and efficient mining of the deep mines2,3. In order to solve the problem of floor heave in soft rock roadways, many experts and scholars have carried out a lot of research work to study the mechanism of floor heave in deep soft rock tunnel.

For the soft rock roadways with good lithology and obvious layered structurecharacteristics in the bottom rock layer, some scholars believe that the floor heave are caused by the release of energy release accumulated in the bottom rock layer due to the redistribution of surrounding rock stress caused by tunnel excavation4,5,6. Some other scholars believe that the mechanism of bottom plate uplift is due to the joint action of roof pressure transferring to the bottom through two sides and upward movement of the bottom layer. The floor layer of roadway undergoes bending deformation towards the free surface, when the bending deformation of floor layer reaches its maximum value, the floor layer breaks and rises upwards7,8,9,10,11,12,13,14. Based on the stress characteristics and structural characteristics of the layered soft rock roadway, a thin plate mechanical model15, a beam mechanical model16,17 and a three-hinged arch-spring model18 have been established.

For the soft rock roadways with low rock strength and incomplete structure of the floor layer, some experts believe that the floor heave is mainly caused by the compression flow and shear displacement of the floor layer under horizontal stress, which leads to the uplift of the floor layer towards free surface of the roadway19,20,21,22,23. Based on the soil pressure theory, a mechanical model for the floor heave of soft rock tunnels has been established24,25. Some experts have also found that the floor heave of soft rock roadways is caused by water absorption expansion26,27,28, creep deformation29,30 advanced support pressure of the working face31.To summarize, floor heave in layered soft rock roadway primarily occurs due to the bending, fracturing, compression, and shear displacement of the floor rock under the influence of surrounding rock stress. The suction, softening, expansion and mining-induced disturbance pressure exacerbate the floor heave. In general, the floor heave of layered soft rock roadways is a gradual development process32 and the deformation is relatively large.

The influence of rock structure elements on the floor heave of soft rock roadways also can not be ignored, such as rock type, thickness, and inclination angle33,34,35.In order to explore the mechanical mechanism of floor heave that occurs in thin layered soft rock tunnels, a physical model experiment as well as a numerical simulation test were carried out so as to reproduce the process of floor heave in the tunnels36,37,38.

Model test

Prototype of physical model

A roadway was chosen as the prototype of which was located in Shi Yakou Coal Mine, in Yunnan province, east of China. The depth of roadway is about 750 m. The angle of stratum was near 10°. A detailed stratum histogram is showed in Fig. 1. The physical and mechanical parameters of rocks were shown in Table 1. There are some problems in this roadway, such as floor heave, serious shrinkage of sidesandroof. Floor heave is a main problem of this roadway.

Fig. 1
Fig. 1The alternative text for this image may have been generated using AI.
Full size image

Stratum histogram.

Table 1 Physical and mechanical parameters of the rocks.
Full size table

Model building

Simulation material and similarity analysis are very important for model tests. A good rock’s simulation material was used to build the physical model. The simulation material is made by quartz powder, plaster, barite powder, water and talcum powder. Different artificial rocks can be made by the simulation material. In this case, it was used to make three types artificial rocks, namely siltstone, coal and mudstone. The ratio of artificial siltstone is 0.609 (barite powder):0.156 (quartz powder):0.097 (plaster): 0.010 (talcum powder): 0.128 (water) in mass. The ratio of artificial coal is 0.489 (barite powder): 0.146 (quartz powder): 0.085 (plaster): 0.089 (talcum powder): 0.191 (water) in mass. The ratio of artificial mudstone is 0.548 (barite powder): 0.168 (quartz powder): 0.068 (plaster): 0.069 (talcum powder): 0.147 (water) in mass. The physic-mechanical parameters of artificial rock are listed in Table 2.

Table 2 Physical and mechanical parameters for the artificial rocks.
Full size table

According to some studies, we found that stress similarity constant Cσ, geometric dimension similarity constant Cl and body force similarity constant Cγ are key similar parameters for physical model experiments. Besides, the following similarity index requirement should be met:

$$\frac{{{C_\sigma }}}{{{C_l }{C_\gamma }}}=1$$
(1)

According to the similarity simulation theory, similarity constants were determined by analyzing actual geological conditions of roadway and the conditions of laboratory. The parameters of Cσ, Cγ, and Cl were 0.065, 0.065 and 1.00 respectively. A physical model was built by the artificial siltstone, artificial coal and artificial mudstone.

Data acquisition scheme

The MatchID-2D strain measurement system was adopted to acquire the full-field deformation data of monitored area (as showed in Fig. 2a, the resolution of this strain measurement system is 0.01 pixel. And a video camera was used to record the process of physical model. Besides, there were 39 points of strain data monitoring located in the physical model (as showed in Fig. 2b. Every monitoring point was placed one strain gauges which was perpendicular to the direction of strata inclination (the normal). The arrangement of equipment data collection equipment was shown in Fig. 3.

Fig. 2
Fig. 2The alternative text for this image may have been generated using AI.
Full size image

Data acquisition location of strain and deformation.

Fig. 3
Fig. 3The alternative text for this image may have been generated using AI.
Full size image

This is a figure. Schemes follow the same formatting.

Experiment schemes

20 kN/m3 has been chosen as the average unit weight of the rock according to our engineering experience, and 1 was chose as the lateral pressure coefficient by the empirical formula. Thus, the initial gravity stress of rock mass and the initial horizontal tectonic stress are both 15 MPa at the depth of 750 m. Because the stress similarity constant Cσ is 0.065, the physical model was applied 1.00 MPa at the top and the lateral sides to simulate the initial stress of rock mass.

The physical model experiment consisted of three stages, stage A, stage B and stage C. The stage A included five processes, A1, A2, A3, A4 and A5. During the A1, A2 A3, A4 and A5, 0.2 MPa was applied to the physical model at the top and the lateral sides respectively. The stage B included eights processes, B1, B2, B3, B4, B5, B6, B7 and B8. During the B1, B2, B3, B4, B5, B6, B7 and B8, 50 mm was excavated in the method of full-face excavation. The Stage C was one process. During the stage C, 0.4 MPa additional loading was applied to the top of physical model. The detailed loading scheme was showed in Fig. 4.

Fig. 4
Fig. 4The alternative text for this image may have been generated using AI.
Full size image

Loading scheme of the experiment.

Results and discussion

Deformation analysis

Floor heave occurred during the loading process of stage C. We paid more attention to some typical pictures of the process of floor heave. The von mises equivalent strain of surrounding rock in monitoring area was obtained by processing those typical pictures (as showed in Fig. 5). The color represents the size of the von mises equivalent strain and the small black arrow on the picture represents the direction of the displacement vector. In the following, VME strain is used to represent von Mises equivalent strain. As showed in Fig. 5a, the VME strain value of whole monitoring area is very small. As showed in Fig. 5b, The VME strain value of floor below the roadway began to increase. As showed in Fig. 5c, the VME strain of floor below the roadway increased, the maximum VME strain value of floor reaches 0.06. Then, the VME strain value of floor continue to increase, the maximum VME strain value of floor reaches 0.08 (As showed in Fig. 5d). After, the phenomenon of separation of floor stratum occurs, the area where the VME strain value increases is funnel-shaped, the maximum VME strain value of floor is 0.15(As showed in Fig. 5e). Last, the floor strata were broken and serious floor heave occurred, the VME strain value of floor strata obviously increased and the range of strain increase is enlarged, the maximum VME strain value of floor is near 0.19(As showed in Fig. 5f).

Fig. 5
Fig. 5The alternative text for this image may have been generated using AI.
Full size image

The von mises equivalent strain of surrounding rock in monitoring area.

After conducting the aforementioned analysis, it becomes evident that the failure mechanism of roadway floor heave lies in the fracture and uplift of the floor. The strain distribution of surrounding rock exhibits a funnel-shaped pattern, with a VME strain value for the floor approaching 0.19. It is worth noting that horizontal stress serves as the primary factor contributing to floor heave in thin layer soft rock roadways.

Strain analysis

The strain field of surrounding rock during the process of floor heave of roadway is obtained (as showed in Fig. 6). The area with positive value in the picture represents that the area is under tension, and the negative value area indicates that the area is under pressure.

Fig. 6
Fig. 6The alternative text for this image may have been generated using AI.
Full size image

The strain field of surrounding rock during the process of floor heave of roadway.

At first, there is a circular tensile strain zone at the lower left corner of the roadway, and the maximum tensile strain is 550 micro strain, which indicates that the zone is in a tensile stress state (as showed in Fig. 6a). The tensile strain in the lower left corner of the roadway increases to 800. At the same time, a button shaped tensile strain area also appears in the right bottom corner of the roadway, the maximum value of the tensile strain is about 450 micro strain. It indicates that the left bottom corner and the right bottom corner of the roadway are in a state of tensile stress (as showed in Fig. 6b).

The strain within the range of one tunnel diameter around the tunnel is still dominated by compressive strain. There is an arched tensile strain area on the roof of the tunnel, and the maximum tensile strain is about 1800 micro strains. It showed that the rock mass in these areas is in the tensile strain area. At the same time, there is a compressive strain circle around the tunnel, which indicates that the rock mass in these areas is in the state of compressive stress. Besides, there is a spherical tensile strain zone at the top right of the roadway, and the maximum tensile strain is 1500 micro strain (as showed in Fig. 6c). There is an arched tensile strain zone at the position of the floor rock under the roadway and the maximum tensile strain is 1600 micro strain. Besides, A circular compressive strain zone appears at a distance of 1 times the tunnel diameter below the tunnel, and the maximum value of compressive strain is about 1100 micro strain.

After conducting the aforementioned analysis, it can be deduced that the surrounding rock undergoes a state of tensile stress within a time range equivalent to the diameter of the roadway subsequent to its floor heave collapse process. Moreover, there is an alternating occurrence of tensile stress zones and compressive stress zones in close proximity to the roadway.

Numerical simulation

Modeling building

A numerical model was built by 3DEC(as showed in Fig. 7). The size of the numerical calculation model is 1.6 m×1.6 m×0.4 m, which is consistent with the size of the physical model experiment. The numerical model contains 2572 blocks and 168,068 grid units. The boundary conditions were same with the physical model test. The failure criterion was Mohr– coulomb.

Fig. 7
Fig. 7The alternative text for this image may have been generated using AI.
Full size image

Calculation model.

Numerical results

The displacement field of the surrounding rock of the tunnel is shown in Fig. 8a. The positions with significant displacement changes are located in the right foot area and the left shoulder area of the tunnel. The maximum displacement in the right foot area is 47 mm, and the maximum displacement in the left shoulder area is 43 mm. The deformation of the two sides of the tunnel has also increased. The displacement deformation of four key measurement points was obtained, among which the displacement deformation of 1 # measurement point was 31 mm, the displacement deformation of 2 # measurement point was 43 mm, the displacement deformation of 3 # measurement point was 20 mm, and the displacement deformation of 4 # measurement point was 21 mm. The failure mode of the surrounding rock of the tunnel is shown in Fig. 8b. The roof rock of the tunnel undergoes a certain degree of bending deformation downwards, and the floor rock of the right foot area of the tunnel undergoes a significant bending deformation upwards. The fracture occurs at the position with the maximum bending deformation, and separation from the lower rock layer results in separation.

Fig. 8
Fig. 8The alternative text for this image may have been generated using AI.
Full size image

Defomation of surrounding rock.

The displacement data of the observation points of the numerical simulation test and the physical model test were compared and analyzed (as shown in Fig. 9), and it was found that the displacement values of the 1 # and 3 # observation points differed by 3 mm, the displacement values of observation points 2 # and 4 # differ by 2 mm. The displacement data of the four key monitoring points in the numerical simulation experiment are in good agreement with the displacement data of the four key monitoring points in the model experiment.

Fig. 9
Fig. 9The alternative text for this image may have been generated using AI.
Full size image

Displacement of monitoring point.

Conclusions and progress

  1. (1)

    For the soft rock roadway located in deep-buried thin bedded rock stratum, horizontal stress is an important factor affecting the floor heave of roadway.Thus providing further theoretical basis and scientific basis for the control of surrounding rock in mine tunnels.

  2. (2)

    When the floor heave occurs in the roadway, the strain distribution of floor of roadway is funnel-shaped, the max VME strain value of the floor is near 0.19, the affected range of floor strata is about half of the roadway width.

  3. (3)

    The surrounding rock within one time of the tunnel diameter is in the state of tensile stress after the process of roadway floor heave failure, and the areas of tensile stress and compressive stress in the surrounding rock alternates around the roadway. The failure mechanism of floor heave of the roadway is the fracture and uplift of the floor of roadway.

  4. (4)

    The study of soft rock roadway under complex geological conditions remains inadequate. The next step will involve simulating and analyzing the behavior of soft rock roadways under various endowment states to effectively control the failure mechanism associated with floor fractures and uplift.