Application of distributed fiber-optic sensing in mining pressure and overburden monitoring in three-dimensional similarity simulation experiments

Abstract

The manifestation of mining pressure and overburden deformation in mining fields is one of the critical issues that cannot be avoided in the safe and efficient extraction of coal. Precise monitoring and early warning of these factors are essential for disaster prevention and control. This study, based on three-dimensional similarity simulation experiments, integrates PPP-BOTDA distributed fiber-optic sensing technology with a self-developed traction displacement device to construct an internal deformation monitoring system for the model. The paper proposes a method for identifying roof pressure and quantifying overburden failure height using the average variation of the Brillouin frequency shift measured by optical fibers. Experimental results show that changes in the vertical Brillouin frequency shift curve are related to the degree of deformation of the overburden layers in the mining field. The more intense the overburden movement, the larger the change in the Brillouin frequency shift curve. This method successfully identified 16 instances of pressure occurrences during the advancement of the working face, with the step change in the average variation of the Brillouin frequency shift curve closely matching the location of the pressure. The fiber-optic monitoring of overburden failure deformation height was found to be 7%–15% higher than the surface observation results from the model, reflecting the high sensitivity of distributed fiber-optic sensing technology to small internal deformations in rock layers. The research results validate the accuracy and advancement of distributed fiber-optic sensing technology in monitoring mining pressure and overburden deformation in three-dimensional similarity simulation experiments. It overcomes the limitations of traditional monitoring methods that cannot obtain continuous deformation data from within, providing more options and security for safe mining and intelligent early warning systems.

1 Introduction

Coal still occupies a dominant position in China’s energy consumption. For example, in 2023, its consumption accounted for more than 55% of the total energy consumption (Jia-xuan et al., 2021). With the continuous increase in coal demand, mining capacity has been continually enhanced, and extraction intensity has been consistently intensified. This has led to frequent occurrences of roof accidents, sudden water inrushes from the roof and floor, rock burst disasters, and coal and gas outbursts, all of which are closely related to the manifestation of mining pressure and overburden movement in the mining field (Zhen-qi, 1988). In recent years, to effectively protect water resources in mining areas, coal mine underground reservoir technology has been widely applied. The dam structures of these reservoirs and the surrounding rock system are long-term subjected to the complex coupling environment of mining-induced stress and seepage, making the monitoring of overburden movement and deformation particularly important. Therefore, conducting research on the monitoring of overburden movement and deformation in mining fields is of significant importance for ensuring the safe extraction of coal resources, maintaining the stable operation of underground reservoirs, and protecting the ecological environment of mining areas (Zhi-jun and Wen-xiang, 2022).

Under high-intensity mining conditions, overburden fractures are characterized by large amplitude, wide spatial extent, and a propensity for abrupt changes, and the corresponding fracture and evolution laws have become a critical scientific issue that has attracted considerable attention and remains to be addressed in green and intelligent coal mining (Wang et al., 2022; Hui et al., 2025). Sun Binyang and co-workers, based on a self-developed optoelectronic multi-parameter distributed sensing system deployed in an extremely thick conglomerate mining site, investigated strata movement and deformation as well as the fracturing and caving behavior. Their results showed that the caving zone height was 44.73–48.00 m, the fracture zone development height was 256.29–276.54 m, and the post-mining re-compaction zone was located 150–204 m behind the working face (Bin-yang et al., 2024). Li Zhi and co-workers studied the strong roof pressure manifestations in the Tongxin mining area and revealed that the mining-induced overburden sequentially experiences slight bed separation, instantaneous subsidence, separation closure, intermittent stabilization, and instantaneous caving and compaction (Zhi et al., 2024). Chang Yunbo and co-workers, using in situ borehole monitoring and mine-pressure analysis, examined the overburden movement characteristics of a shallow-buried fully mechanized top-coal caving face beneath thick and hard bedrock. They found that key roof strata govern overburden movement and roof pressure manifestation; the overburden failure pattern develops upward in a saddle-shaped form; the maximum heights of the caving zone and fracture zone were 44 m and 58 m, respectively; and the thick, hard key stratum effectively inhibited further upward propagation of fractures (Yun-bo et al., 2024).

Current research on overburden deformation mainly relies on theoretical analysis (Xing et al., 2024), in situ measurements (Shang-shang and De-zhong, 2019), numerical simulations (Wang et al., 2003), and physical similarity simulations (Jie et al., 2022). Among these, three-dimensional similarity simulation experiments have become an important method for studying the movement patterns of overburden due to their ability to realistically replicate the spatial structure of the mining field (Bing et al., 2025). However, traditional methods for monitoring internal deformation in three-dimensional models still face several challenges: stress gauges, fiber Bragg gratings, and similar technologies can only obtain discrete point data, acoustic emission and photoelastic instruments are incapable of distributed measurement, and displacement monitoring relies on mechanical gauges or total stations, which cannot capture continuous internal deformation data (Ke et al., 2021). The processing of monitoring data also requires external equipment, resulting in a delayed cycle of “experiment - data monitoring - processing - analysis,” making it difficult to respond in real-time to dynamic overburden deformation (Gao-feng et al., 2028). Distributed fiber-optic sensing technology offers a new solution for dynamic monitoring of these issues (Jing et al., 2021). Based on the Brillouin optical time-domain analysis (BOTDA) principle, this technology utilizes the linear response characteristic of Brillouin frequency shift when optical fibers are subjected to strain or temperature disturbances, enabling long-distance, high-resolution distributed monitoring (Gong-yu et al., 2023). The main advantages of distributed fiber-optic sensing technology include: the ability of a single fiber to both sense and transmit, covering the entire strain field of the rock mass (Jing et al., 2020a); strong anti-interference ability, making it suitable for harsh environments with minimal impact on the physical properties of the model (Jing et al., 2018a); and high accuracy, with spatial resolution of 5 cm and sampling intervals of 1 cm, enabling precise capture of micro-deformations in rock layers (Gang et al., 2025).

Since the introduction of distributed fiber-optic sensing technology by Farahani M A et al., it has been widely applied in full-distribution measurements in spatial domains (Farahani and Gogolla, 1999). Horiguchi et al. proposed the BOTDA system, which uses the linear relationship between Brillouin frequency shift and strain and temperature to achieve distributed measurement of strain and temperature information along the sensing fiber (Horiguchi and Tateda, 1989). Existing studies have shown that distributed fiber-optic sensing technology can be used to monitor surface temperature, residual deformation, and cracks (Chun-xun et al., 2025). Chai Jing et al. introduced Brillouin distributed fiber-optic sensing technology into similarity material model experiments, enabling distributed strain measurement of internal deformation in the model, providing a novel testing method for similarity material simulation experiments (Jing et al., 2020b; Jing et al., 2018b). Qi et al. applied optical sensing monitoring technology to measure overburden deformation during field mining activities, studied distributed monitoring methods for overburden deformation, analyzed deformation characteristics induced by mining, and explored the delamination change mechanism (Qi et al., 2024).

This study addresses the issues of mining pressure manifestation and overburden movement during the mining process. Based on three-dimensional similarity simulation experiments, distributed fiber-optic sensing technology and internal displacement monitoring are applied to construct a monitoring system for overburden deformation during mining activities. The research aims to reveal the dynamic response patterns of overburden structure deformation, stress field evolution, and fracture expansion during the advancement of the working face. It also analyzes the precursory information of overburden instability and the intrinsic mechanisms of mining pressure manifestation. The study provides technical support for the construction and safe operation monitoring of underground reservoir dams and offers key technical guidance for safe and efficient mining.

2 Theoretical analysis

2.1 Distributed optical fiber sensing technology

In a Brillouin Optical Time Domain Analysis, hereinafter referred to as BOTDA, system, a short pulse of light is injected into one end of the optical fiber while a continuous probe beam is introduced into the opposite end. The strain variation at any point along the fiber can be determined by monitoring the frequency shift of the stimulated Brillouin scattering light (Lei et al., 2022; Peng-bai et al., 2022). The distributed fiber-optic sensing system is shown in Figure 1.

FIGURE 1

2.2 Principle of distributed optical fiber deformation monitoring

By detecting the power of the continuous light coupled out from one end of the optical fiber, the frequency difference corresponding to the maximum energy transfer in each segment of the fiber can be determined. This frequency shift is used to obtain distributed strain and temperature information (Xu-ping et al., 2024), as expressed by Equations 2.1, 2.2:

Where n_eff is the effective refractive index of the fiber core, v_A is the acoustic velocity, and λ_P is the vacuum wavelength of the pump light; Δv_B1 and Δv_B2 represent the variations in Brillouin frequency shift in the optical fiber;C_T1 and C_T2 denote the Brillouin temperature coefficients of different optical fibers, respectively;C_ε is the strain coefficient of the optical fiber under constant temperature conditions; Δ_ε and ΔT indicate the changes in strain and temperature, respectively.

2.3 Discrimination of working face pressure

The average variation degree D_i is adopted to characterize the difference between two Brillouin frequency shift curves, serving as an indicator for identifying intense overburden movement and predicting whether the mining face is under pressure (Kun-you et al., 2024)^. The calculation of D_i is given by the following Equation 2.3:

In the formula: D_i represents the average variation degree after the ith advancement of the mining face MHz; x_i (j) denotes the Brillouin frequency shift value at the location of optical fiber sampling point j after the ith advancement MHz; x_i−1 (j) refers to the Brillouin frequency shift value at the location of optical fiber sampling point j after the (i − 1)-th advancement MHz; i indicates the excavation step; j specifies the position index of the optical fiber sampling point; n is the total number of optical fiber sampling points, calculated as n = L/B, where L is the total length of the optical fiber cm and B is the sampling interval cm; m represents the total number of mining face advancements.

When Z optical fibers are deployed in the overburden strata of the mining face, the average variation degrees D_i1, D_i2, …, D_{i k}, …, D_{i Z} of the Brillouin frequency shift curves from these Z fibers after the ith excavation step are integrated to compute the overall average variation degree ∑D_i using Equation 2.4. This consolidated value D_i is then utilized to determine whether the mining face is under pressure.

In the formula: ∑D_i denotes the overall average variation degree of all optical fibers after the ith excavation step of the mining face MHz; i indicates the excavation step number; \ m \ represents the total number of mining face excavation steps; k is the index number of the optical fiber; Z specifies the total number of optical fibers deployed; D_i,_k refers to the average variation degree of the kth optical fiber after the ith excavation step MHz.

2.4 Determination of overburden failure height in the stope

After the lower coal seam has been extracted over a certain distance, the overlying strata above the stope move downward, undergoing deformation and failure. The resulting strata movement and deformation act on the vertically installed optical fibers within the model, causing fiber deformation. By detecting the strain and temperature along the optical fiber and eliminating the influence of temperature, the failure-induced deformation height H_d of the overburden strata can be determined. Specifically, when the frequency shift at a certain point of the optical fiber exceeds a predefined threshold K, the height corresponding to that point is identified as the failure deformation height of the overburden. The magnitude of the threshold K directly influences the measured failure height: a larger K results in a smaller measured failure height, and vice versa. The overburden strata above the stope are divided, from bottom to top, into the caved zone H₁, the fractured zone H₂, and the bending subsidence zone H₃, as illustrated in Figure 2. According to the empirical formulas for mining-induced overburden failure height, when the roof overburden above the coal seam is classified as moderately hard strata, the caving-zone height is given by H_m = 100ΣM/(4.7ΣM + 19) ± 2.2, where H_m is the caving-zone height in meters and M is the mining height of the coal seam in meters. The fracture-zone height is given by H_l = 100ΣM/(1.6ΣM + 3.9) ± 5.6 = 100∑M/(1.6∑M + 3.9) ± 5.6, where H_l is the fracture-zone height in meters. In the bending-subsidence zone, the rock strata are relatively less affected by coal extraction; moreover, due to pronounced differences in mechanical properties among strata, the subsidence magnitudes of different layers can vary substantially under mining-induced stress (Zhi-yong et al., 2024).

FIGURE 2

3 3D similarity simulation experiment

3.1 Geological conditions

The experiment is based on a coal mine as the engineering background. The main coal seam being mined is the No. 2 coal seam, with a thickness of 15 m. There are a total of 12 rock layers above the coal seam, and the total thickness of the coal seam and its overburden is 805.32 m. The characteristic parameters of the overburden are shown in Table 1.

3.2 Experimental equipment and design

The experiment was conducted using a variable-scale three-dimensional model frame independently designed and developed by Xi’an University of Science and Technology. The internal effective dimensions of the model frame are 3,600 mm in length, 2,000 mm in width, and 2,400 mm in height. The geometric similarity ratio of the model was set at 1:400, the unit weight similarity ratio at 1.6, and the stress similarity ratio at 640. The three-dimensional model frame and the physical similarity model are shown in Figure 3, and the mechanical parameters of each stratigraphic layer in the similarity model are listed in Table 2. The front face of the model was sealed with a transparent acrylic panel. The paving thicknesses of the coal-seam floor, the coal seam, and the overlying strata were 200 mm, 60 mm, and 1,740 mm, respectively. The coal seam was simulated using galvanized square steel tubes with a cross-section of 40 mm × 60 mm; during mining, coal extraction was realized by sequentially withdrawing the tubes one by one. The overburden strata above the model that were not constructed up to the ground surface were converted, based on proportional weight equivalence, into sandbag loads applied to the top of the model.

FIGURE 3

3.3 Monitoring system

The modeled working face had a width of 800 mm and an advance length of 2,400 mm, with a coal-seam height of 60 mm. Along the advance direction, 600 mm coal pillars were reserved on both the left and right sides. The working face was arranged adjacent to the front face of the model, and a 1,200 mm coal pillar was reserved in the dip direction.

A PPP-BOTDA distributed fiber-optic sensing system was adopted to monitor internal deformation of the model. The main hardware comprised a distributed optical-fiber measurement unit NBX-6055A and a control-system host. The supporting acquisition and analysis software of the interrogator was used to collect Brillouin scattering frequency data, with a maximum spatial resolution of 5 cm and a minimum sampling interval of 1 cm. In addition, to obtain the internal displacement of strata above the simulated working face, a self-developed traction-type multi-point displacement measurement device was employed to measure internal displacements at three different stratigraphic levels within the overburden.

Three vertical optical fibers, denoted as V11, V12, and V13, were installed in the overburden above the working face, each with a length of 1740 mm. Specifically, V11 was located 1,200 mm from the left boundary of the model; V12 was located 2,400 mm from the left boundary and 400 mm from the front face. Three sets of internal displacement measurement tubes, labeled 11#, 12#, and 13#, were installed in the strata above the working face with a spacing of 600 mm. These tubes monitored displacements at nine positions across three elevations, located 500 mm, 690 mm, and 1,190 mm above the coal-seam roof. The measurement points on the tubes were named from bottom to top as 111#, 112#, 113#, 121#, 122#, 123#, 131#, 132#, and 133#. Moreover, the internal displacement tubes 11#, 12#, and 13# were mapped to the corresponding external displacement measurement points on the model front surface, which were labeled as A1, A2, A3, B1, B2, B3, C1, C2, and C3, as shown in Figure 4.

FIGURE 4

4 Results and discussion

4.1 Determination of mining face pressure on the model front

The width of the model working face is 800 mm, and the height is 60 mm. Excavation progresses from left to right, with each excavation step being 40 mm, for a total of 60 steps, resulting in a total advancement of 2,400 mm. A 600 mm coal pillar is left on both sides of the model, and a 1,200 mm coal pillar is left in the dip direction.

When the working face advances to 560 mm, during the 14th excavation, the first fracture occurs in the first sub-critical layer. The sub-critical layer and its partial accompanying layers move toward the goaf, causing intense movement of the overburden. The working face experiences the first pressure occurrence, with a collapse height of 400 mm as shown in Figure 5a.

FIGURE 5

When the working face advances to 720 mm, during the 18th excavation, the first sub-critical layer fractures again, and the deformation of the overburden continues to develop upward. In the range of 540–740 mm, delamination occurs in the overburden, with the maximum delamination amount being approximately 1.5 mm and a length of 220 mm. The working face experiences the first cyclic pressure occurrence, and the height of the damaged deformation reaches 740 mm as shown in Figure 5b.

When the working face advances to 840 mm, during the 21st excavation, the second sub-critical layer fractures for the first time, and the range of overburden collapse expands. The working face experiences the third cyclic pressure occurrence. The second sub-critical layer is thick, and the entire critical layer is in two different states. In the height range of 540–640 mm, the rock layers form a hinged structure, while in the range of 640–1,020 mm, the rock layers are in a bent and sinking state. At this point, the maximum delamination occurs at a height of 1,020 mm, with a delamination amount of approximately 10 mm as shown in Figure 5c.

When the working face advances to 1,200 mm, during the 30th excavation, the main critical layer deforms, and the entire layer begins to slowly subside, although the subsidence is minimal. The number of fractures within the main critical layer increases significantly. The working face experiences the fifth cyclic pressure occurrence, and the overburden delamination caused by the collapsed rock layers develops to a height of 1,400 mm. The maximum delamination amount is approximately 5 mm as shown in Figure 5d.

When the working face advances to 1,320 mm, during the 33rd excavation, the fractures within the main critical layer continue to develop, and the overall bending deformation of the main critical layer increases. The subsidence reaches 15 mm. The working face experiences the sixth cyclic pressure occurrence, and the corresponding height of the damage deformation develops to 1,630 mm as shown in Figure 5e.

When the working face advances to 1,560 mm, the main critical layer undergoes its first fracture. The main critical layer and its accompanying layers move downward as a whole. The working face experiences the eighth cyclic pressure occurrence, and the corresponding collapse height on the front of the model develops to the surface. In the range of 200–1,600 mm on the surface, a subsidence state is observed as shown in Figure 5f.

When the working face advances to 1,840 mm, during the 46th excavation, the main critical layer fractures again, and the overburden experiences a large-scale collapse. The surface subsidence extends to the range of 200–1,800 mm, and the working face experiences the 10th cyclic pressure occurrence as shown in Figure 5g.

When the working face advances to 2,400 mm, during the 60th excavation, the mining of the working face is completed. At this point, the second sub-critical layer fractures again, and the overburden experiences intense movement. The working face experiences the 15th cyclic pressure occurrence, forming a subsidence basin on the surface approximately 1,450 mm long as shown in Figure 5h.

After the mining of the working face is completed, the phenomena on the front surface of the model are statistically analyzed. A total of 16 pressure occurrences were observed in the overburden, with the first pressure occurrence having a step distance of 560 mm, and the cyclic pressure step distances ranging from 80 to 160 mm, with an average of 120 mm. The height of the damage deformation gradually increased as the working face advanced. When the working face reached 1,200 mm, the fractures developed to the top of the model, and the pressure occurrence step distances are shown in Figure 6.

FIGURE 6

4.2 Discrimination of mining face pressure based on displacement measurement points

4.2.1 Analysis of displacement measurements

Figure 7 shows the variation curves of the internal and surface displacement measurement points. As shown in Figure 7A, when the working face advances to 840 mm, the displacement at the lower two displacement measurement points, 111# and 112#, of the 11# displacement measurement pipe changes for the first time. The displacements at 111# and 112# are 19.46 mm and 13.51 mm, respectively. Observation from the front of the model reveals that at this point, the working face experiences the third pressure occurrence, corresponding to the second cyclic pressure occurrence. The second sub-critical layer fractures for the first time, leading to a large-scale collapse of the overburden, which results in significant subsidence at the 111# and 112# displacement measurement points. The displacement at the external measurement points A1 and A2, corresponding to 111# and 112#, also changes for the first time, with displacements of 27.0 mm and 19.0 mm, respectively. At this time, the internal displacement measurement point 113# and its corresponding external measurement point A3 do not show any displacement change.

FIGURE 7

After the working face advances to 1,080 mm, the displacements at the internal displacement measurement points 111# and 112# and their corresponding external measurement points A1 and A2 begin to change slowly. When the working face advances to 1,200 mm, observation from the front of the model reveals the occurrence of the sixth pressure event, corresponding to the fifth cyclic pressure occurrence at the working face. The internal displacement measurement point 113# and its corresponding external measurement point A3 on the front surface of the model begin to change, with displacements of 3.92 mm and 5.0 mm, respectively. From the front of the model, it is observed that the main critical layer begins to deform, causing the subsidence of the two measurement points. After the working face advances to 1,440 mm, the displacements at the internal displacement measurement point 113# and the corresponding external measurement point A3 on the front surface of the model begin to change slowly. After the simulated working face advances to 1760 mm, the displacements at the three internal displacement measurement points and three external displacement measurement points eventually stabilize.

Figure 7B shows the displacement curves of the three internal displacement measurement points of the 12# displacement measurement pipe and their corresponding three external displacement measurement points on the front surface of the model. From Figure 7B, it can be seen that when the model working face advances to 1,560 mm, observation from the front of the model reveals the occurrence of the ninth pressure event, corresponding to the eighth cyclic pressure occurrence at the working face. The displacements at the three internal displacement measurement points of the 12# displacement pipe and the corresponding three external displacement measurement points all change. Among them, the displacements at 121# and 122# and their corresponding external measurement points B1 and B2 change significantly. The subsidence at 121# and 122# is 30.28 mm and 9.84 mm, respectively, while the subsidence at B1 and B2 is 38.0 mm and 15.0 mm, respectively. The displacement changes at the internal displacement measurement point 123# and its corresponding external displacement measurement point B3 are minimal, with changes of 0.6 mm and 5.0 mm, respectively. After the working face advances to 1,680 mm, observation from the front of the model reveals the occurrence of the 10th pressure event, corresponding to the ninth cyclic pressure occurrence at the working face. The internal displacement measurement points 121# and 122# and their corresponding external measurement points B1 and B2 begin to change slowly. After the working face advances to 2,280 mm, the displacements at the three internal displacement measurement points and their corresponding three external displacement measurement points begin to stabilize.

Figure 7C shows the displacement curves of the three internal displacement measurement points of the 13# displacement measurement pipe and their corresponding three external displacement measurement points on the front surface of the model. From Figure 7C, it can be seen that when the simulated working face advances to 1960 mm, observation from the front of the model reveals the occurrence of the 12th pressure event, corresponding to the 11th cyclic pressure occurrence at the working face. The two internal displacement measurement points, 131# and 132#, at the lower part of the 13# displacement pipe and their corresponding external displacement measurement points C1 and C2 on the front surface of the model begin to change. The displacements at 131# and 132# are 3.08 mm and 0.49 mm, respectively, while the displacements at C1 and C2 are 6.0 mm and 3.0 mm, respectively. When the simulated working face advances to 2080 mm, observation from the front of the model reveals the occurrence of the 13th pressure event, corresponding to the 12th cyclic pressure occurrence at the working face. The displacement at the internal displacement measurement point 131# and the corresponding external displacement measurement point C3 on the front surface of the model begin to change, with displacements of 2.17 mm and 4.0 mm, respectively.

4.2.2 Discriminant analysis of mining face pressure via displacement monitoring points

The movement of key strata above the mining face exhibits a direct correlation with face pressure, allowing for the discrimination of weighting events through displacement monitoring of these critical structural layers. Figure 8 illustrates the internal displacement variations within the three-dimensional model during face advancement. As the face approached each monitoring point, displacement increased sharply; with further advancement, the rate of displacement change gradually decelerated. Based on the comprehensive displacement monitoring results, the variation in total internal displacement at different advancement positions was obtained as shown in Figure 8D. When the face advanced to 840 mm, the total displacement change reached 32.97 mm, coinciding with the initial fracture of the second sub-key stratum and the ensuing face pressure. Subsequently, the total displacement change gradually decreased with continued face advancement until an advance of 1,560 mm, when the main key stratum fractured initially and fractures propagated to the model roof, causing a sudden increase in total displacement to 47.84 mm. When the face reached 2,080 mm, the total displacement change surged again, indicating ongoing settlement within the internal rock layers. By the completion of the mining operation, no abrupt changes were observed in the total displacement curve, suggesting the absence of significant caving deformation in the internal strata during this final stage. Comparative analysis with Figure 6 reveals that the advancement distances corresponding to 13 distinct weighting events align precisely with the positions of abrupt changes in the total displacement curve. This correlation demonstrates that during weighting events, the subsidence of specific displacement measurement points increased markedly, with their displacement curves exhibiting characteristic step-like abrupt changes. These findings collectively verify that the displacement monitoring system can accurately identify face pressure events, while also confirming that the movement patterns of internal displacement points maintain consistent correspondence with overburden caving and weighting behavior above the mining face.

FIGURE 8

4.3 Discrimination of mining face pressure using vertical optical fibers

Due to the 600 mm spacing between the vertical optical fibers V11, V12, and V13, the mining face reached positions directly beneath V11, V12, and V13 at advancement distances of 600 mm, 1,200 mm, and 1800 mm, respectively. Consequently, the characteristics of the Brillouin frequency shift curves evolved distinctively with increasing face advancement. When the face advanced to 840 mm, a weighting event occurred on the model front, and the BFS curve of fiber V11 exhibited pronounced changes as shown in Figure 9. The average variation degree of the BFS curves for fiber V11 across different advancement distances revealed a substantial shift at 840 mm, indicating intense movement of the overburden strata in the vicinity of V11.

FIGURE 9

At a face advancement of 1,320 mm, another weighting event was observed on the model front, accompanied by notable alterations in the BFS curves of both optical fibers V11 and V12 compared to previous stages as shown in Figure 10. The average variation degree of the BFS curves increased markedly at this advancement distance, indicating severe strata movement occurring near fibers V11 and V12. As mining operations continued, the average variation degrees for both V11 and V12 gradually decreased, suggesting that the surrounding overburden strata had progressively re-established a stable structural state.

FIGURE 10

During the advancement of the working face from 1,680 mm to 1,800 mm, the Brillouin frequency shift curves of fibers V11 and V12 remained largely consistent, as shown in Figure 11. When the working face reached 1,800 mm, the lower part of fiber V13 transitioned to a tensile state. The working face continued to advance to 1,840 mm, and pressure occurrence was observed on the front surface of the model. At this point, the Brillouin frequency shift curves of fibers V11, V12, and V13 all underwent significant changes. The average variation of the Brillouin frequency shift curve indicates that from 1,680 mm to 1,800 mm, the average variation of fibers V11 and V12 changed very little, suggesting that the rock layers near fibers V11 and V12 did not experience significant movement. When the working face reached 1800 mm, the average variation of fiber V13 increased, which was mainly due to the embedding process and the downward movement of the rock layers. As the working face continued to advance to 1,840 mm, the average variation curves of fibers V11, V12, and V13 all exhibited a step change, indicating that the rock layers near the embedded fibers underwent significant movement. The corresponding observations from the model front and displacement monitoring results confirmed that a pressure occurrence took place at this time.

FIGURE 11

During the advancement of the working face from 2,080 mm to 2,160 mm, the Brillouin frequency shift curves of fibers V11, V12, and V13 remained largely unchanged. When the working face reached 2,200 mm, the Brillouin frequency shift curves of fibers V12 and V13 underwent significant changes, as shown in Figure 12. The average variation of the Brillouin frequency shift curve indicates that fibers V12 and V13 experienced a step change when the working face reached 2,200 mm. By analyzing the average variation of fibers V12 and V13, it can be determined that a pressure occurrence took place at this point.

FIGURE 12

From the above analysis, it can be concluded that the step changes in the average variation of the Brillouin frequency shift curves mainly occur when a pressure event is observed at the front of the model, shortly after a pressure event is observed, or when the working face advances directly below the fiber. The first phenomenon occurs because when a pressure event happens, the overburden undergoes significant movement, causing the frequency shift curve of the fiber to change noticeably, thus increasing the average variation, and all the average variation curves exhibit a distinct step change at this point. The second reason is due to the large thickness of the three-dimensional model, where external deformation precedes internal deformation. The third reason for the step change in the average variation curve arises from experimental conditions. When the vertical fiber is installed, the lower part of the fiber is fixed with a cylindrical object. Therefore, when the working face is directly below the vertical fiber, the cylindrical object and the nearby rock layers immediately tend to move downward, causing the lower end of the vertical fiber to be in a tensile state, increasing the average variation. However, at this point, the working face has not experienced a pressure event.

Analysis of the average variation curves shows that the frequency shift curves of fibers in the delamination zone are the most active, while the frequency shift curves of fibers in the re-compaction zone gradually change less and eventually stabilize. Therefore, when determining a pressure event at the working face, it is necessary to use the existing judgment results as a reference. For fibers in the delamination zone, judgment can be made directly. However, if the fiber is directly above the working face, relying on just one fiber alone is insufficient to determine whether the increase in average variation is caused by the intense movement of the overburden. In this case, additional fibers need to be used to assist in the judgment.

4.4 Analysis of overburden deformation

4.4.1 Analysis of failure deformation height

The Brillouin frequency shift curve from the distributed fiber-optic monitoring system is related to the overburden failure deformation height. It is considered that when the Brillouin frequency shift value at a certain point on the fiber is greater than the threshold value K, the corresponding rock layer undergoes deformation; if it is less than the threshold value K, the corresponding rock layer does not experience failure or deformation. Through the characteristic analysis of the Brillouin frequency shift curves of fibers V11, V12, and V13, it was found that the frequency shift curves of the rock layers that have not collapsed, up to the top of the model, converge near the 0 value. The threshold value K is ultimately determined to be in the range of 10–20 MHz. During the advancement of the working face, the curve showing the change in internal deformation and failure height in the model is shown in Figure 13. The monitoring results at fibers V11, V12, and V13 indicate that the internal collapse deformation of the model increases in a stepwise manner as the working face advances. The initial failure height at the V11 measurement point is the lowest, mainly because the goaf range is small in the early stages of mining, and the collapse development of the overburden is limited. As the working face continues to advance, the collapse range of the overburden increases, and the failure deformation height also increases. When the working face reaches below V12 and V13, due to the increased collapse range and height of the overburden, the initial failure deformation height measured at V12 and V13 is higher than the initial failure deformation height at V11. As shown in Figure 13, the three curves increase in a stepwise manner, and the horizontal coordinates corresponding to the steps of the curves align with the occurrence of pressure at the working face, indicating that when the pressure occurs at the working face, the failure deformation height increases.

FIGURE 13

The comparison and analysis of the internal failure deformation height measured by vertical fibers and the corresponding front surface observation results of the model show that the fiber-measured internal failure deformation height is consistently higher than the model’s front surface failure deformation height. This difference is mainly due to the high sensitivity of distributed fiber-optic sensing technology to small deformations in the rock layers, which can monitor microscopic fractures and early-stage deformations that are not directly observable by the naked eye. In contrast, front surface observations of the model can only identify macro-scale damage deformations that have already formed. Therefore, the overburden failure deformation height obtained by the fiber is more comprehensive, reflecting the cumulative deformation, including micro-damage. It should be noted that the fiber monitoring results obtained using the 10–20 MHz threshold are relatively conservative when characterizing the height of fracture zones, with the values being slightly higher than the actual values. Nevertheless, this measurement still provides valuable reference information for determining the height of the water-conducting fracture zone within the overburden failure range.

Based on the curve of the internal failure–deformation height during face advance shown in Figure 13, the displacement increments of each monitoring point in the overburden were accumulated for each advance step to obtain the accumulated displacement variation curve of the overburden above the working face, as shown in Figure 14. The accumulated displacement variation ranges from 10 to 50 mm. After conversion using the geometric similarity ratio, the corresponding deformation and caving magnitude of the overburden in the prototype is estimated to be 4–20 m. The internal displacement evolution of the model is consistent with the caving behavior and periodic weighting characteristics of the overburden above the working face.

FIGURE 14

4.4.2 Analysis of caved zone height

The Brillouin frequency shift curves of vertical fibers V11, V12, and V13 during mining are shown in Figure 15, when a pressure occurrence takes place at the working face. From Figure 15, it can be seen that the length of the compressed section at the lower end of vertical fiber V11 is 510 mm, the compressed length at the lower end of vertical fiber V12 is 500 mm, and the compressed length at the lower end of vertical fiber V13 is 310 mm. Based on the principle of distributed fiber-optic sensing for identifying the height of the collapse zone, the collapse zone heights near fibers V11, V12, and V13 are determined to be 510 mm, 500 mm, and 310 mm, respectively. The collapse zone heights at V11 and V12 are quite similar, while the result at V13 is smaller, indicating that the rock layer at the V13 position has not collapsed as extensively. This phenomenon is consistent with the displacement monitoring results, which mutually verify the reliability of the monitoring data.

FIGURE 15

5 Discussion

This study aims to address the limitations of conventional monitoring methods in capturing the continuous internal deformation of overburden strata in a mining panel. By integrating PPP-BOTDA distributed fiber-optic sensing with a self-developed traction-type displacement device, we established a three-dimensional physical similarity simulation system, enabling dynamic, distributed, and high-precision monitoring of overburden deformation and mine-pressure manifestation during face advance. By proposing a quantitative approach based on the mean variation degree and a threshold K, we achieved accurate identification of 16 periodic weighting events. The fiber-optic and displacement monitoring results show good consistency, allowing reliable identification of periodic weighting, and the inferred internal displacement evolution of the overburden generally agrees with the strata caving process and periodic weighting behavior. The results are also consistent with those reported by Jing et al. (2020b), Jing et al. (2018b) and Liang Bing et al. Qi et al. (2024) on mining-induced overburden deformation. Compared with surface observations, fiber-optic monitoring improved the measurement accuracy of the failure–deformation height by 7%–15%, effectively revealing subtle deformation characteristics and confirming the reliability of the fiber-optic data. These findings further demonstrate the accuracy and advanced capability of distributed fiber-optic sensing for deformation monitoring in geotechnical and mining engineering.

Compared with existing understanding, our results exhibit certain discrepancies, providing new insights. Traditional studies generally consider that the overburden deformation observed on the surface of a two-dimensional similarity model can approximately represent the internal deformation response. Discrepancy 1: Unlike previous work, we employed PPP-BOTDA distributed fiber-optic sensing to continuously monitor the interior of a three-dimensional model. Benefiting from the three-dimensional model that better reproduces in situ spatial confinement, together with the high sensitivity and continuous sampling capability of distributed fiber-optic sensing to micro-strain responses, we obtained and identified richer information on the evolution of overburden movement, thereby more realistically depicting the spatiotemporal evolution of internal damage. The monitoring results indicate that, during face advance, the deformation and failure evolution of the internal overburden exhibits a certain lag relative to the deformation observed on the model surface. Discrepancy 2: When the working face advanced directly beneath the vertical fiber, no obvious periodic weighting phenomenon was observed on the model front face; however, the mean variation degree derived from the fiber-optic monitoring showed a step-like change. This suggests that the spatial transmission of overburden deformation and failure in a three-dimensional model involves a certain lag, and that internal strain accumulation and structural evolution often occur prior to surface-scale manifestations. The monitoring results further verify the strong early-warning potential of distributed fiber-optic sensing, which is highly sensitive to subcritical stages such as stress transfer, crack initiation, and bed separation development prior to key-stratum fracture, enabling earlier detection of the internal stress and deformation evolution before macroscopic failure becomes apparent.

The results are consistent with key-stratum theory. Both displacement and fiber-optic monitoring data indicate that the abrupt subsidence and stepwise increase in failure–deformation height are temporally consistent with the periodic fracturing heights of the sub-key stratum and the main key stratum, suggesting that mine-pressure manifestation is essentially the macroscopic response of instability of the overlying key-stratum structure. From a structural mechanics perspective, periodic weighting corresponds to a cyclic evolution of the key-stratum structure from a stable load-bearing beam or arch system to a process of fracture, rotation, and re-bearing. The accelerated-subsidence segment in the displacement curves reflects the fracture-induced instability of the key stratum, whereas the mean variation degree and threshold K obtained from fiber-optic monitoring provide a quantitative representation of the strain field during the instability process. The combined displacement and fiber-optic results effectively characterize the typical stages of overburden evolution, including an initial stage dominated by bending subsidence, an intermediate stage with rapid development of bed separation and fractures, a key-stratum fracturing stage featuring abrupt responses and a stepwise increase in failure–deformation height, and a post-fracture stage characterized by block compaction and gradual attenuation of the response. Based on these observations, the micro-strain signals captured by fiber-optic monitoring are linked more explicitly to the macroscopic structural instability process in mine-pressure mechanics, providing theoretical support for the mechanical interpretation of distributed fiber-optic sensing in studies of overburden movement and failure.

This study also has certain limitations. First, the physical and mechanical properties of the similarity materials differ from those of real rock masses, and groundwater softening as well as complex in situ stress distributions are not fully represented, which may affect the accuracy of predicting overburden failure patterns and heights. Second, the threshold K is determined in a partly empirical manner, and its general applicability needs further validation under different geological conditions and model scales. Building on these limitations, future work will focus on three-dimensional physical simulations incorporating seepage–stress coupling to better reproduce hydrogeological conditions, applying the PPP-BOTDA system in field investigations to optimize and calibrate the threshold K and the criteria for periodic weighting identification through inversion analysis, and integrating artificial intelligence with large-scale distributed fiber-optic data to develop machine-learning-based mine-pressure early-warning models, thereby providing stronger technical support for safe, green, and intelligent coal mining.

6 Conclusion

References

• 1

  Bin-yangS.LiangY.Ping-songZ.Rong-xinW. (2024). Experimental study on the optical-electrical sensing of overburden movement and bed separation evolution under extra-thick conglomerate strata. J. J. China Univ. Min. and Technol.53 (05), 977–992. 10.13247/j.cnki.jcumt.20240162

  • CrossRef
  • Google Scholar

• 2

  BingL.Jing-chaoX.Wei-jiS.Sheng-jieF. (2025). Fracture-induced mechanism and three-dimensional spatial characteristics of overburden under mining-induced breakage. J. Coal Sci. Technol.53 (01), 107–121.

  • Google Scholar

• 3

  Chun-xunT.Wan-qiuL.Wei-shanK.Da-congL. (2025). Design and performance test of BOTDA-Based sensor for roadbed cavity monitoring. J. Instrum. Tech. Sens. (05), 21–25.

  • Google Scholar

• 4

  FarahaniM. A.GogollaT. (1999). Spontaneous Raman scattering in optical fibers with modulated probe light for distributed temperature Raman remote sensing. J. J. Light. Technol.17 (8), 1379–1391. 10.1109/50.779159

  • CrossRef
  • Google Scholar

• 5

  GangC.Yu-miaoG.De-shengC.BinS.Tian-luC. (2025). Research and application of fiber optic monitoring for coal mine safety mining under the integrated framework of sensing-transmission-computing-control. J. Geol. J. China Univ., 1–10. 10.16108/j.issn1006-7493.2025054

  • CrossRef
  • Google Scholar

• 6

  Gao-fengS.Sheng-liY.Zhao-huiW. (2028). Mechanism analysis of coal wall failure based on Ritz method and three-dimensional physical simulation. J. J. China Coal Soc.43 (08), 2162–2172. 10.13225/j.cnki.jccs.2017.1301

  • CrossRef
  • Google Scholar

• 7

  Gong-yuH.Zhi-yuH.Zi-xiangL.Qing-ruZ.Dong-xingF.ChengC.et al (2023). Application status and prospect of distributed optical fiber and fiber Bragg grating sensing technology in coal mine safety monitoring. J. J. China Coal Soc.48 (S1), 96–110. 10.13225/j.cnki.jccs.2022.0527

  • CrossRef
  • Google Scholar

• 8

  HoriguchiT.TatedaM. (1989). BOTDA-nondestructive measurement of single-mode optical fiber attenuation characteristics using Brillouin interaction: theory. J. J. Light. Technol.7 (8), 1170–1176. 10.1109/50.32378

  • CrossRef
  • Google Scholar

• 9

  HuiX.YongW.Qian-huiL.Xiao-weiJ.Jing-jingG.ZhuT. (2025). Research on the collapse law and deformation-prediction of overburden in mining-induced slopes. J. Saf. Coal Mines56 (11), 88–99. 10.13347/j.cnki.mkaq.20250267

  • CrossRef
  • Google Scholar

• 10

  Jia-xuanT.JunL.Yue-qiangL.KunC.KunZ. (2021). Reflections on the development of China's coal industry during the 14th Five-Year Plan period. J. China Coal47 (10), 6–10. 10.19880/j.cnki.ccm.2021.10.002

  • CrossRef
  • Google Scholar

• 11

  JieZ.Yi-fengH.Nan-hongL.Jian-pingG.BinW.Jian-junW. (2022). Study on overburden movement and fracture evolution law under repeated mining in shallow coal seams. J. Saf. Coal Mines53 (03), 58–65. 10.13347/j.cnki.mkaq.2022.03.010

  • CrossRef
  • Google Scholar

• 12

  JingC.Wen-gangD.QiangY.Ding-dingZ. (2018a). Analysis of optical fiber sensing test method for physical model test. J. J. Xi’an Univ. Sci. Technol.38 (05), 728–736. 10.13800/j.cnki.xakjdxxb.2018.0505

  • CrossRef
  • Google Scholar

• 13

  JingC.Zi-wuX.RuiG.Ding-dingZ.QiangY.YiL. (2018b). Experimental study on distributed optical fiber detection of collapse morphology and evolution of overburden in mining area. J. J. China Univ. Min. Technol.47 (06), 1185–1192. 10.13347/j.cnki.mkaq.2020.04.010

  • CrossRef
  • Google Scholar

• 14

  JingC.Xu-baoZ.Ding-dingZ.Wen-gangD.Wu-linL.Yu-yuY. (2020a). BOTDA-FBG monitoring test on overburden movement during mining. J. Sci. Technol. Eng.20 (16), 6404–6410.

  • Google Scholar

• 15

  JingC.Shu-junL.Ding-dingZ. (2020b). Experimental study on deformation characteristics of overburden based on distributed optical fiber monitoring. J. Saf. Coal Mines51 (04), 47–51.

  • Google Scholar

• 16

  JingC.Yong-liangL.QiangY.Ding-dingZ.Wen-gangD.Zhi-chengH.et al (2021). Theory, technology and application of optical fiber sensing for deformation and failure of mine surrounding rock. J. Coal Sci. Technol.49 (01), 208–217. 10.13199/j.cnki.cst.2021.01.016

  • CrossRef
  • Google Scholar

• 17

  KeY.Wen-jieL.BiaoJ.Qing-heZ.ShuaiL.Zhai-nanZ. (2021). Three-dimensional physical simulation experiment on overburden movement in deep and thick hard roof fully-mechanized caving mining. J. Chin. J. Geotechnical Eng.43 (01), 85–93.

  • Google Scholar

• 18

  Kun-youZ.Lin-mingD.Jia-zhuoL.KeY.Yan-kunM.Guo-luY.et al (2024). Study on dynamic and static loading mechanism of extra-thick key strata on rock burst. J. J. Min. Saf. Eng.41 (05), 908–919. 10.13545/j.cnki.jmse.2023.0255

  • CrossRef
  • Google Scholar

• 19

  LeiZ.Wen-zheG.JingC.ZheM.Feng-qiQ. (2022). Distributed optical fiber monitoring of full-field deformation evolution process of overburden during mining. J. J. Min. Strata Control Eng.4 (01), 43–50. 10.13532/j.jmsce.cn10-1638/td.20211014.002

  • CrossRef
  • Google Scholar

• 20

  Peng-baiX.Xiao-longW.Kun-huaW.Yong-qiuZ.Jin-yunZ.Yong-kangD.et al (2022). Research on distributed optical fiber high temperature sensing based on BOTDA (Invited). J. Infrared Laser Eng.51 (10), 42–50.

  • Google Scholar

• 21

  QiL.Jia-baoN.Qing-haiL.Jin-haiZ.Jian-xiaoZ. (2024). Coupling and zonal characterization of optical fiber monitoring for fracture evolution in mining-induced overburden strata. J. J. China Coal Soc.49 (03), 1345–1357. 10.13225/j.cnki.jccs.2023.1282

  • CrossRef
  • Google Scholar

• 22

  Shang-shangZ.De-zhongK. (2019). Fracture characteristics of hard roof and overlying strata movement law in working face. J. Saf. Coal Mines50 (05), 257–262. 10.13347/j.cnki.mkaq.2019.05.062

  • CrossRef
  • Google Scholar

• 23

  WangC.TannantD. D.LillyP. A. (2003). Numerical analysis of the stability of heavily jointed rock slopes using PFC2D. J. Int. J. Rock Mech. Min. Sci.40 (3), 415–424. 10.1016/S1365-1609(03)00004-2

  • CrossRef
  • Google Scholar

• 24

  WangX.XieJ.ZhuW.XuJ. (2022). The field monitoring experiment of the high-level key stratum movement in coal mining based on collaborative DOFS and MPBX. J. Sci. Rep.1 (1), 665. 10.1038/s41598-021-04578-w

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 25

  XingF.Ming-luJ.TuoZ.Yun-pengL.Tian-haoL.Shi-weiC. (2024). Study on instability and rupture of overlying hard rock stratum in No. 8105 working face of Tongxin Coal Mine. J. J. Saf. Environ.24 (07), 2632–2640. 10.13637/j.issn.1009-6094.2023.1850

  • CrossRef
  • Google Scholar

• 26

  Xu-pingZ.Yi-xinZ.LiangW.Kuang-luY.BoL. (2024). Research status and future of distributed optical fiber sensing technology and its application. J. Acta Opt. Sin.44 (01), 11–73.

  • Google Scholar

• 27

  Yun-boC.Ji-shengX.Li-yongZ.Jin-kuiM.Xin-jianG.You-bingZ. (2024). Study on the movement and failure law of overburden in shallow-buried thick and hard bedrock fully mechanized caving face. J. Coal Eng.56 (06), 144–151.

  • Google Scholar

• 28

  Zhen-qiS. (1988). Practical mine pressure control. Xuzhou: China University of Mining and Technology Press, 1–2.

  • Google Scholar

• 29

  ZhiL.DongY.Yi-chaoJ. (2024). Research on evolution mechanism of overburden movement and ground pressure control during deep and thick coal seam mining. J. Min. Res. Dev.44 (12), 68–75. 10.13827/j.cnki.kyyk.2024.12.010

  • CrossRef
  • Google Scholar

• 30

  Zhi-junD.Wen-xiangZ. (2022). Study on overburden movement and pressure behavior in shallow coal seam mining face. J. Coal Technol.41 (11), 1–4. 10.13301/j.cnki.ct.2022.11.001

  • CrossRef
  • Google Scholar

• 31

  Zhi-yongZ.Yuan-yuanL.Yan-fengL. (2024). Fracture development in overlying strata and upward mining in closely spaced lower coal seam. J. Coal Eng.56 (01), 113–118.

  • Google Scholar