A deep-buried cavern random block instability probability cloud early warning method

Abstract

The application provides a deep-buried cavern random block instability probability cloud early warning method, comprising collecting rock mass structure data, constructing a three-dimensional geological model to identify potential random blocks, using Monte Carlo simulation to calculate instability probability, combining Internet of Things sensor real-time monitoring of deformation and stress data, fusing simulation and monitoring results to calculate comprehensive instability probability, and publishing graded risk early warning information based on probability threshold through a cloud platform. The application can realize accurate early warning of deep-buried cavern random block instability, and improve the safety and reliability of cavern construction and operation.

Description

Technical Field

[0001] This invention relates to the field of underground engineering safety monitoring and early warning technology, and more specifically, to a cloud-based early warning method for the instability probability of random blocks in deeply buried caverns. Background Technology

[0002] In underground engineering construction, the stability of deeply buried caverns has always been a critical issue for project safety. Traditional methods for cavern stability analysis mainly rely on empirical formulas and deterministic analysis. While these methods can provide a preliminary assessment of cavern stability to some extent, they have significant limitations. First, they cannot fully account for the complexity and randomness of rock mass structures, especially the instability risk of random blocks. Second, traditional methods struggle to reflect the dynamic changes of caverns in real time during construction and operation, failing to promptly detect minute changes that may lead to instability. Furthermore, existing early warning systems are mostly locally deployed, lacking sufficient data integration and information sharing capabilities, making remote monitoring and rapid response difficult.

[0003] In implementing the embodiments of the present invention, the prior art has at least the following problems or defects: existing methods cannot accurately assess the instability probability of random blocks in deeply buried caverns, and lack a mechanism for fusing real-time monitoring data with simulation results, resulting in untimely and inaccurate early warnings. At the same time, traditional early warning systems have insufficient information sharing and remote monitoring capabilities, making it difficult to meet the safety and efficiency requirements of modern underground engineering. Summary of the Invention

[0004] This invention provides a cloud-based early warning method for the probability of random block instability in deeply buried caverns, comprising: S1: Collect rock mass structure data of the deep-buried cavern, including structural surface feature data and rock mass quality parameter data; S2: Construct a three-dimensional geological model based on the rock mass structure data and identify potential random blocks; S3: Calculate the instability probability of the potential random block using the Monte Carlo simulation method to obtain the simulated instability probability; S4: Real-time monitoring of deformation and stress data of deeply buried caverns is achieved through IoT sensors, generating real-time monitoring data; S5: Combine the simulated instability probability and the real-time monitoring data to calculate the comprehensive instability probability; S6: Based on the comprehensive instability probability, conduct risk warning and release warning information through the cloud platform.

[0005] Furthermore, the acquisition of rock mass structure data in S1 specifically includes: S1.1: Obtain a structural surface statistics table through adit logging. The structural surface statistics table records the number and distribution location of JX shear fault zone, J joint, Gm extrusion surface, Fpd fault, JC extrusion fault zone and JK structural surface. S1.2: Scan the inner wall of the adit using a 3D laser scanner, setting up a scanning station every 3 meters to obtain structural surface attitude data, including strike, dip angle and dip direction; S1.3: Rock mechanics tests were conducted using an indoor testing machine to obtain uniaxial saturated compressive strength, rock mass integrity index Kv, and triaxial unloading mechanical parameters; S1.4: Combine the on-site measurement of rock wave velocity using a seismic wave velocity tester to calculate the rock mass integrity index Kv.

[0006] Furthermore, the construction of the three-dimensional geological model in S2 specifically includes: S2.1: Based on the scale of the structural surface, the structural surface is divided into five levels: Level I fault-type structural surface, Level II fracture-type structural surface, Level III non-penetrating structural surface, Level IV visible structural surface, and Level V hidden microstructural surface; S2.2: Using structural surface network simulation technology, based on the grouping results of structural surface attitude, the diameter, spacing and volume density parameters of structural surfaces are generated through probability distribution functions; S2.3: Process 3D laser scanning data using Cloudcompare and GIS software to determine the boundary conditions, spatial location, geometric shape, and volume of random blocks; S2.4: Based on the three-dimensional network simulation results of the structural plane, calculate the probability distribution function of the rock mass RQD value.

[0007] Furthermore, the calculation of instability probability using the Monte Carlo simulation method in S3 specifically includes: S3.1: Set the total number of Monte Carlo simulations N, where N ≥ 10000; S3.2: Each simulation randomly samples the geometric and mechanical parameters of the block. The geometric parameters include the block volume and the inclination angle of the structural surface, and the mechanical parameters include the cohesion c and the internal friction angle φ. S3.3: The safety factor Fs of the block is calculated using the limit equilibrium theory. When Fs < 1, it is determined to be unstable. S3.4: Probability of Instability The calculation formula is:

[0008] in, N represents the number of unstable blocks in the simulation, and N is the total number of simulations.

[0009] Furthermore, the real-time monitoring of deformation and stress data of the deeply buried cavern via IoT sensors in S4 specifically includes: S4.1: Arrange displacement sensors, stress sensors, and microseismic monitoring equipment arrays in the roof arch and sidewalls of the cavern; S4.2: The displacement sensor uses a multi-point displacement meter to measure the relative displacement of the rock mass with an accuracy of 0.1mm; S4.3: The stress sensor adopts a vibrating wire stress gauge to measure the stress change value of rock mass, with a range of 0-50MPa; S4.4: Microseismic monitoring equipment records the frequency and energy release of rock mass fracturing events; S4.5: The collected data is uploaded to the cloud platform database in real time via the wireless transmission module.

[0010] Furthermore, the calculation of the comprehensive instability probability by integrating the simulated instability probability and the real-time monitoring data in S5 specifically includes: S5.1: Establish a Bayesian update model to simulate instability probability. As a priori probability; S5.2: Define the conditional probability of monitoring data P(D|F), where D represents real-time monitoring data and F represents the block instability event; S5.3: Calculate the marginal probability P(D) of the monitoring data; S5.4: Overall Instability Probability The calculation formula is:

[0011] in, , This indicates a stable event in the block.

[0012] Furthermore, the risk warning based on the comprehensive instability probability in S6 specifically includes: S6.1: Set three levels of probability thresholds: low threshold 0.1, medium threshold 0.3, and high threshold 0.7; S6.2: When 0.1≤ When the value is less than 0.3, a blue alert is triggered, and an SMS reminder is sent via the cloud platform. S6.3: When 0.3 ≤ When the value is less than 0.7, a yellow alert is triggered, and the audible and visual alarm device is activated through the cloud platform. S6.4: When When the value is ≥0.7, a red alert is triggered, and an emergency support plan is automatically executed through the cloud platform.

[0013] Furthermore, it also includes: S7: Numerical simulation analysis of surrounding rock stability during deep-buried cavern excavation, specifically including: S7.1: A three-dimensional numerical model of the geological structure of the underground cavern area was established using Rhino software; S7.2: Invert the distribution of the geostress field in the underground cavern area using Flac3D software; S7.3: Simulates the distribution and variation of the surrounding rock stress field during excavation, and outputs the maximum and minimum principal stress values; S7.4: Analyze the deformation characteristics and evolution of the surrounding rock during the excavation process, and calculate the top settlement and sidewall convergence value; S7.5: Correct the block mechanical parameters in the Monte Carlo simulation based on the numerical simulation results.

[0014] Furthermore, it also includes: S8.1: The first surrounding rock score is calculated using the first surrounding rock classification method. The scoring indicators of the first surrounding rock classification method include rock strength, rock mass integrity, structural surface condition and groundwater conditions. S8.2: The second surrounding rock score is calculated using the second surrounding rock classification method, which is based on the rock mass quality index RQD, joint group number Jn, joint roughness coefficient Jr, joint alteration coefficient Ja, joint water reduction coefficient Jw and stress reduction factor SRF. S8.3: The third surrounding rock score is calculated using the third surrounding rock classification method. The scoring indicators of the third surrounding rock classification method include rock strength score, RQD score, joint spacing score, joint condition score, and groundwater score. S8.4: Dynamically adjust the values ​​of cohesion c and internal friction angle φ in the Monte Carlo simulation based on the surrounding rock quality classification results.

[0015] Furthermore, it also includes: S9: Enables data integration and visualization through a cloud platform, specifically including: S9.1: Establish a cloud database to store rock mass structure data, real-time monitoring data, simulated instability probability, and comprehensive instability probability; S9.2: Develop a data fusion algorithm to update the comprehensive instability probability calculation model in real time; S9.3: Construct a web-based visualization interface to dynamically display the warning level, the 3D model of the block instability risk area, and monitoring data curves; S9.4: Set multi-level user permissions to support remote access and push alert information to mobile terminals.

[0016] The embodiments of the present invention have at least the following beneficial effects: 1. By collecting rock mass structure data and constructing a three-dimensional geological model to identify potential random blocks, and combining the Monte Carlo simulation method to calculate the instability probability, the complexity and randomness of the rock mass structure can be fully considered, and the instability risk of random blocks in deep-buried caverns can be accurately assessed, solving the problem that traditional methods cannot accurately assess the instability probability of random blocks.

[0017] 2. By integrating simulated instability probability with deformation and stress data monitored in real time by IoT sensors, a comprehensive instability probability is calculated, enabling real-time monitoring and assessment of dynamic changes in deeply buried caverns. This improves the timeliness and accuracy of early warnings and overcomes the problem of delayed early warnings caused by the inability of existing early warning systems to capture minute changes in a timely manner.

[0018] 3. Based on the cloud platform, early warning information is released, realizing data integration and visualization, supporting remote access and mobile terminal early warning information push, enhancing information sharing and remote monitoring capabilities, solving the defects of insufficient information sharing and difficulty in remote monitoring in traditional early warning systems, and improving the efficiency and convenience of underground engineering safety management. Attached Figure Description

[0019] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent from the following detailed description taken in conjunction with the accompanying drawings. Several embodiments of the invention are illustrated in the drawings by way of example and not limitation, wherein: Figure 1 This is a flowchart illustrating a cloud-based early warning method for the probability of random block instability in a deep-buried cavern, as provided in an embodiment of the present invention. Detailed Implementation

[0020] The principles and spirit of the invention will now be described with reference to several exemplary embodiments. It should be understood that these embodiments are provided merely to enable those skilled in the art to better understand and implement the invention, and are not intended to limit the scope of the invention in any way. Rather, these embodiments are provided to make the invention more thorough and complete, and to fully convey the scope of the invention to those skilled in the art.

[0021] Those skilled in the art will recognize that embodiments of the present invention can be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention can be specifically implemented in the following forms: entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software.

[0022] It should be noted that the number of any elements in the accompanying drawings is for illustrative purposes only and not as a limitation, and any naming is for distinction only and has no limiting meaning.

[0023] The following is for reference. Figure 1, Figure 1 This is a flowchart illustrating a cloud-based early warning method for the probability of random block instability in deeply buried caverns, provided in an embodiment of the present invention. Figure 1 As shown, a cloud-based early warning method for the instability probability of random blocks in deeply buried caverns includes: S1: Collect rock mass structure data of the deep-buried cavern, including structural surface feature data and rock mass quality parameter data; S2: Construct a three-dimensional geological model based on the rock mass structure data and identify potential random blocks; S3: Calculate the instability probability of the potential random block using the Monte Carlo simulation method to obtain the simulated instability probability; S4: Real-time monitoring of deformation and stress data of deeply buried caverns is achieved through IoT sensors, generating real-time monitoring data; S5: Combine the simulated instability probability and the real-time monitoring data to calculate the comprehensive instability probability; S6: Based on the comprehensive instability probability, conduct risk warning and release warning information through the cloud platform.

[0024] In this embodiment, rock mass structure data refers to the structural surface features and rock mass quality parameters related to the rock mass of the deeply buried cavern; this data forms the basis for assessing cavern stability. A three-dimensional geological model is a digital representation of the geological structure used to simulate the rock mass structure and potentially unstable blocks surrounding the cavern. The Monte Carlo simulation method is a probabilistic statistical numerical simulation method used to assess the instability probability of random blocks. IoT sensors refer to various sensors installed within the cavern for real-time monitoring of deformation and stress changes. The cloud platform is a cloud-based data processing and information sharing platform used to store data, calculate the overall instability probability, and issue early warning information.

[0025] Specifically, structural surface characteristic data refers to the quantity, distribution, and attitude of various structural surfaces existing in the rock mass, such as joints and faults, as well as their strike, dip angle, and dip direction. This data is acquired through adit logging and 3D laser scanning. Rock mass quality parameter data includes uniaxial saturated compressive strength, rock mass integrity index Kv, and triaxial unloaded mechanical parameters, which are acquired through indoor testing machines and seismic wave velocity meters.

[0026] When constructing the three-dimensional geological model, structural surfaces are classified into five levels according to scale: Level I fault-type structural surfaces, Level II fracture-type structural surfaces, Level III non-penetrating structural surfaces, Level IV visible structural surfaces, and Level V hidden microstructural surfaces. Structural surface network simulation technology, based on the grouping results of structural surface attitudes, generates the diameter, spacing, and volume density parameters of structural surfaces through a probability distribution function. In the Monte Carlo simulation, a total number of simulations is set, and each simulation randomly samples the geometric parameters, volume, structural surface dip angle, mechanical parameters, cohesion, and internal friction angle of the block. The limit equilibrium theory is used to calculate the block safety factor; when the safety factor is less than 1, instability is determined. The instability probability is calculated by the ratio of the number of instable blocks in the simulation to the total number of simulations. In real-time monitoring, multi-point displacement gauges are used as displacement sensors to measure the relative displacement of the rock mass with an accuracy of 0.1 mm; vibrating wire stress gauges are used as stress sensors to measure the stress change value of the rock mass with a range of 0-50 MPa; microseismic monitoring equipment records the frequency of rock mass fracture events and energy release values. All monitoring data is uploaded to the cloud platform database in real time via a wireless transmission module.

[0027] Preferably, when constructing the three-dimensional geological model, the structural surfaces are first divided into five levels according to their scale. Level I fault-type structural surfaces refer to faults with large scale and significant displacement; Level II fracture-type structural surfaces refer to large but not fully continuous fractures; Level III non-continuous structural surfaces refer to small and non-continuous structural surfaces; Level IV visible structural surfaces refer to small structural surfaces visible on the rock mass surface; and Level V occult structural surfaces refer to tiny structural surfaces inside the rock mass that are difficult to observe directly. In the Monte Carlo simulation, the total number of simulations is set to over 10,000 to ensure the reliability of the simulation results.

[0028] In each simulation, geometric and mechanical parameters of the block are randomly sampled. Geometric parameters include block volume and structural surface inclination angle, while mechanical parameters include cohesion and internal friction angle. The safety factor of the block is calculated using limit equilibrium theory; a safety factor less than 1 indicates instability. The instability probability is calculated as the ratio of the number of instable blocks in the simulation to the total number of simulations. In real-time monitoring, multi-point displacement gauges are used to measure the relative displacement of the rock mass with an accuracy of 0.1 mm; vibrating wire stress gauges are used to measure stress changes in the rock mass with a range of 0-50 MPa; and microseismic monitoring equipment records the frequency of rock mass fracture events and energy release values. All monitoring data is uploaded to the cloud platform database in real-time via a wireless transmission module. When fusing simulated instability probabilities and real-time monitoring data, a Bayesian update model is established. The simulated instability probability is used as the prior probability, the conditional probability of the monitoring data is defined, and the marginal probability of the monitoring data is calculated. The overall instability probability is obtained by calculating the prior probability, conditional probability, and marginal probability.

[0029] In some embodiments, the acquisition of rock mass structure data in S1 specifically includes: S1.1: Obtain a structural surface statistics table through adit logging. The structural surface statistics table records the number and distribution location of JX shear fault zone, J joint, Gm extrusion surface, Fpd fault, JC extrusion fault zone and JK structural surface. S1.2: Scan the inner wall of the adit using a 3D laser scanner, setting up a scanning station every 3 meters to obtain structural surface attitude data, including strike, dip angle and dip direction; S1.3: Rock mechanics tests were conducted using an indoor testing machine to obtain uniaxial saturated compressive strength, rock mass integrity index Kv, and triaxial unloading mechanical parameters; S1.4: Combine the on-site measurement of rock wave velocity using a seismic wave velocity tester to calculate the rock mass integrity index Kv.

[0030] It should be noted that this invention further details the specific steps for collecting rock mass structure data, which forms the basis for constructing a three-dimensional geological model and calculating subsequent instability probabilities. Adit logging is a geological survey method that involves meticulously recording the rock mass structural surfaces within an adit to obtain the number and distribution of these surfaces. A three-dimensional laser scanner is a high-precision measuring device capable of rapidly acquiring three-dimensional data of the rock mass surface, thereby determining the attitude of the structural surfaces. An indoor testing machine is used to conduct rock mechanics tests and obtain the mechanical property parameters of the rock. A seismic wave velocity meter is used to measure the wave velocity of the rock mass in situ, and then calculate the rock mass integrity index Kv, which is an important parameter for assessing rock mass quality.

[0031] Specifically, the structural plane statistics table obtained from the adit logging records the number and distribution of structural planes, including the JX shear fault zone, J joint, Gm compression surface, Fpd fault, JC compression fault zone, and JK structural plane. These structural planes are various discontinuities present in the rock mass and have a significant impact on its stability. A 3D laser scanner was used to scan the inner wall of the adit, with a scanning station set every 3 meters, to acquire the attitude data of the structural planes, including strike, dip angle, and dip direction. This attitude data is an important parameter describing the spatial orientation of the structural planes and is crucial for the subsequent construction of 3D geological models.

[0032] The rock mechanics tests conducted by the indoor testing machine include uniaxial saturated compressive strength, rock mass integrity index Kv, and triaxial unloading mechanical parameters. Uniaxial saturated compressive strength is the maximum pressure that rock can withstand under saturated conditions; the rock mass integrity index Kv reflects the degree of rock mass integrity; and the triaxial unloading mechanical parameters describe the mechanical behavior of the rock mass under unloading conditions. By combining on-site measurement of rock mass wave velocity with a seismic wave velocity meter, the rock mass integrity index Kv is calculated using a specific formula. This index is an important indicator for evaluating rock mass quality, reflecting the integrity and continuity of the rock mass.

[0033] Preferably, during adit logging, the characteristics of each structural plane should be recorded in detail, including its type, quantity, distribution location, and relationship with other structural planes. When using a 3D laser scanner, ensure that the scanning station is set to cover the entire inner wall of the adit to obtain comprehensive structural plane attitude data. For rock mechanics tests conducted using indoor testing machines, standard testing methods should be strictly followed to ensure the accuracy and reliability of the test results. When measuring rock mass wave velocity, a suitable seismic wave velocity meter should be selected, and measuring points should be reasonably arranged according to site conditions to obtain accurate wave velocity data. Through these detailed operating procedures and parameter settings, the accuracy and completeness of the collected rock mass structural data can be ensured, providing reliable data support for subsequent 3D geological model construction and instability probability calculation.

[0034] In some embodiments, constructing the three-dimensional geological model in S2 specifically includes: S2.1: Based on the scale of the structural surface, the structural surface is divided into five levels: Level I fault-type structural surface, Level II fracture-type structural surface, Level III non-penetrating structural surface, Level IV visible structural surface, and Level V hidden microstructural surface; S2.2: Using structural surface network simulation technology, based on the grouping results of structural surface attitude, the diameter, spacing and volume density parameters of structural surfaces are generated through probability distribution functions; S2.3: Process 3D laser scanning data using Cloudcompare and GIS software to determine the boundary conditions, spatial location, geometric shape, and volume of random blocks; S2.4: Based on the three-dimensional network simulation results of the structural plane, calculate the probability distribution function of the rock mass RQD value.

[0035] The classification of structural planes is based on their geometric characteristics and mechanical properties. Structural plane network simulation technology is a probabilistic statistical method used to generate parameters such as the diameter, spacing, and volume density of structural planes. Cloudcompare and GIS software are specialized software for processing 3D laser scanning data, capable of accurately determining the boundary conditions and spatial locations of random blocks. The rock mass RQD value is one of the rock mass quality indicators, and the calculation of its probability distribution function helps to assess the overall quality of the rock mass.

[0036] Specifically, structural surfaces are classified into five levels based on their scale: Level I fault-type structural surfaces, Level II fracture-type structural surfaces, Level III non-penetrating structural surfaces, Level IV visible structural surfaces, and Level V hidden structural surfaces. This classification is based on the geometric dimensions, development degree, and impact on rock mass stability of the structural surfaces. Structural surface network simulation technology generates the diameter, spacing, and volume density parameters of structural surfaces based on their attitude grouping results using probability distribution functions. These parameters are crucial for constructing realistic 3D geological models because they directly affect the distribution and interrelationships of structural surfaces within the model. Processing 3D laser scanning data using Cloudcompare and GIS software allows for the determination of boundary conditions, spatial location, geometry, and volumetric size of random blocks. This information is critical for identifying potentially unstable blocks. The rock mass RQD value is a core quality index obtained through borehole core sampling; its probability distribution function reflects the overall quality and integrity of the rock mass.

[0037] Preferably, when constructing a three-dimensional geological model, structural surfaces are first classified and their parameters set in detail. For example, Class I fault-type structural surfaces typically have a large scale and significant displacement, greatly affecting rock mass stability; while Class V microstructural surfaces are smaller in scale and easily overlooked in routine surveys, but their cumulative effect can also impact stability. When using structural surface network simulation technology, reasonable probability distribution function parameters, such as the average diameter, spacing, and volume density of structural surfaces, need to be set based on actual geological survey data. These parameter settings should be based on field surveys and historical data to ensure the accuracy of the simulation results. When processing three-dimensional laser scanning data, Cloudcompare software is used for data preprocessing and preliminary analysis, while GIS software is used for further spatial analysis and modeling to accurately determine the boundaries and spatial locations of random blocks. When calculating the probability distribution function of the rock mass RQD value, sufficient borehole data needs to be collected, and the distribution law of the RQD value needs to be obtained through statistical analysis, thereby providing a more accurate rock mass quality assessment for the model. These detailed steps and parameter settings help improve the accuracy and reliability of the three-dimensional geological model, providing a solid foundation for subsequent instability probability calculation and early warning.

[0038] In some embodiments, the calculation of instability probability using the Monte Carlo simulation method in S3 specifically includes: S3.1: Set the total number of Monte Carlo simulations N, where N ≥ 10000; S3.2: Each simulation randomly samples the geometric and mechanical parameters of the block. The geometric parameters include the block volume and the inclination angle of the structural surface, and the mechanical parameters include the cohesion c and the internal friction angle φ. S3.3: The safety factor Fs of the block is calculated using the limit equilibrium theory. When Fs < 1, it is determined to be unstable. S3.4: Probability of Instability The calculation formula is:

[0039] in, N represents the number of unstable blocks in the simulation, and N is the total number of simulations.

[0040] Monte Carlo simulation is a probabilistic numerical simulation method that assesses the instability probability of random blocks through random sampling. This method fully considers the randomness of the block's geometric and mechanical parameters, thus providing a more accurate assessment of instability risk. In practice, the total number of Monte Carlo simulations is set. Each simulation simulates the geometric parameters of a randomly sampled block, such as volume and surface tilt angles, and mechanical parameters, such as cohesion and internal friction angles. The safety factor of the block is then calculated using limit equilibrium theory. When the safety factor is less than 1, the block is considered to be in an unstable state. Finally, the instability probability is obtained by statistically analyzing the ratio of the number of unstable blocks in the simulation to the total number of simulations.

[0041] Specifically, the implementation of the Monte Carlo simulation method includes: setting the total number of Monte Carlo simulations, N, which should typically be greater than or equal to 10,000 to ensure the statistical reliability of the simulation results. For each simulation, the geometric and mechanical parameters of the block are randomly sampled. Geometric parameters include the block's volume and the inclination angle of its structural surfaces, reflecting the block's physical shape and spatial orientation; mechanical parameters include cohesion c and the angle of internal friction φ, describing the mechanical properties of the block material. The safety factor Fs of the block is calculated using limit equilibrium theory. Limit equilibrium theory is a commonly used geotechnical analysis method that assesses the stability of a block by calculating its equilibrium conditions under limit states. When Fs is less than 1, it indicates that the block is in an unstable state. Finally, the instability probability is calculated. The number of unstable blocks in the simulation It is calculated as a ratio to the total number of simulations N.

[0042] Preferably, the Monte Carlo simulation process can be further refined. For example, when setting the total number of simulations N, an appropriate number can be selected based on the specific needs of the project and the availability of computing resources. Generally, the more simulations, the more reliable the results, but the higher the computational cost. When randomly sampling block parameters, the probability distribution function of the parameters can be set based on field surveys and experimental data. For example, the block volume can be assumed to be normally distributed, the dip angle of the structural surface can be assumed to be uniformly distributed, while the cohesion and internal friction angle can be set to specific probability distributions based on rock mechanics test results. When calculating the block safety factor Fs, an appropriate limit equilibrium theory formula can be selected based on the specific geological conditions. For example, for simple block sliding problems, the Mohr-Coulomb criterion can be used; for more complex three-dimensional problems, more advanced numerical methods are required. In statistical instability probability... Further analysis of the simulation results can be performed, such as calculating confidence intervals, to assess the uncertainty of the simulation results. These refined steps and parameter settings help improve the accuracy and reliability of instability probability calculations, providing stronger support for the safety assessment of deeply buried caverns.

[0043] In some embodiments, the real-time monitoring of deformation and stress data of the buried cavern via IoT sensors in step S4 specifically includes: S4.1: Arrange displacement sensors, stress sensors, and microseismic monitoring equipment arrays in the roof arch and sidewalls of the cavern; S4.2: The displacement sensor uses a multi-point displacement meter to measure the relative displacement of the rock mass with an accuracy of 0.1mm; S4.3: The stress sensor adopts a vibrating wire stress gauge to measure the stress change value of rock mass, with a range of 0-50MPa; S4.4: Microseismic monitoring equipment records the frequency and energy release of rock mass fracturing events; S4.5: The collected data is uploaded to the cloud platform database in real time via the wireless transmission module.

[0044] Specifically, the deployment of IoT sensors includes arrays of displacement sensors, stress sensors, and microseismic monitoring equipment placed in the cavern's ceiling and sidewalls. The displacement sensors utilize multi-point displacement gauges with an accuracy of 0.1 mm, precisely measuring the relative displacement of the rock mass. The stress sensors employ vibrating wire stress gauges with a range of 0-50 MPa, used to measure changes in rock mass stress. The microseismic monitoring equipment records the frequency and energy release of rock mass fracturing events. The placement and type of these sensors are determined based on the cavern's structural characteristics and stability assessment requirements. Displacement sensors primarily monitor rock mass displacement changes, stress sensors monitor the stress state of the rock mass, while microseismic monitoring equipment captures minute fracturing events within the rock mass. Through the coordinated operation of these sensors, a comprehensive understanding of the dynamic changes in the cavern's rock mass can be achieved.

[0045] Preferably, when deploying IoT sensors, the spacing and number of sensors should be rationally determined based on the specific dimensions and geological conditions of the cavern. For example, at critical locations such as the cavern's arch and sidewalls, the spacing of displacement sensors can be set to approximately 3 meters to ensure the detection of minute displacement changes in the rock mass. The placement of stress sensors should consider the characteristics of rock mass stress distribution, focusing on monitoring areas of stress concentration. Microseismic monitoring equipment can be arranged in an array according to the length and width of the cavern to more comprehensively record rock mass fracturing events. During data acquisition, the sensors wirelessly transmit data to a cloud platform database in real time. The cloud platform database is responsible for storing and managing this data and provides data interfaces for subsequent data processing and analysis. Through these detailed steps and parameter settings, the accuracy and reliability of real-time monitoring data can be ensured, providing strong data support for the stability assessment and early warning of deeply buried caverns.

[0046] In some embodiments, the calculation of the comprehensive instability probability by fusing the simulated instability probability and the real-time monitoring data in S5 specifically includes: S5.1: Establish a Bayesian update model to simulate instability probability. As a priori probability; S5.2: Define the conditional probability of monitoring data P(D|F), where D represents real-time monitoring data and F represents the block instability event; S5.3: Calculate the marginal probability P(D) of the monitoring data; S5.4: Overall Instability Probability The calculation formula is:

[0047] in, , This indicates a stable event in the block.

[0048] It should be noted that this invention calculates the comprehensive instability probability by fusing simulated instability probability and real-time monitoring data, thereby improving the accuracy and reliability of instability early warning for deeply buried caverns. This method combines theoretical simulation and actual monitoring data, enabling a more comprehensive reflection of the cavern's stability state. The simulated instability probability is obtained through Monte Carlo simulation, which is based on the geometric and mechanical parameters of the rock mass and calculates the probability of block instability through extensive random sampling. Real-time monitoring data is acquired through sensors deployed within the cavern, including deformation and stress data, which reflect the dynamic changes of the cavern during actual use. By fusing the two through a Bayesian update model, a more accurate comprehensive instability probability can be obtained, thus providing a scientific basis for early warning.

[0049] Specifically, when building the Bayesian update model, the instability probability will be simulated. As a prior probability, this is the initial instability probability obtained based on Monte Carlo simulations. The conditional probability of the monitoring data, P(D|F), is defined as follows: This indicates real-time monitoring data. This represents a block instability event, and the conditional probability reflects the probability of monitoring data occurring under conditions of block instability. Calculate the marginal probabilities of the monitoring data. This represents the total probability of the monitored data occurring, including the probabilities under both unstable and stable conditions. (Comprehensive Instability Probability) The calculation formula is ,in , This represents a stable event in the block. In this formula, It was obtained through Monte Carlo simulation, P(D|F) and This can be obtained by analyzing historical records and statistical patterns of monitoring data, and The total probability is calculated using the formula above.

[0050] Preferably, when constructing the Bayesian update model, it is necessary to collect a large amount of historical monitoring data, including data under known instability and stability conditions, to calculate the conditional probability P(D|F) and During the calculation process, statistical analysis methods, such as maximum likelihood estimation or Bayesian estimation, can be used to determine the specific values ​​of these conditional probabilities. For the processing of monitoring data, data smoothing and filtering techniques can be employed to remove noise and outliers, improving data reliability. When calculating the overall instability probability, it is necessary to ensure the accuracy and timeliness of all input parameters, especially real-time monitoring data, which should be updated in real time to reflect the current state of the cavern. In this way, dynamic assessment of the instability risk of deeply buried caverns can be achieved, providing strong support for timely early warning measures.

[0051] In some embodiments, the risk warning based on the comprehensive instability probability in S6 specifically includes: S6.1: Set three levels of probability thresholds: low threshold 0.1, medium threshold 0.3, and high threshold 0.7; S6.2: When 0.1≤ When the value is less than 0.3, a blue alert is triggered, and an SMS reminder is sent via the cloud platform. S6.3: When 0.3 ≤ When the value is less than 0.7, a yellow alert is triggered, and the audible and visual alarm device is activated through the cloud platform. S6.4: When When the value is ≥0.7, a red alert is triggered, and an emergency support plan is automatically executed through the cloud platform.

[0052] It should be noted that this invention triggers different levels of early warnings by setting three probability thresholds and sends corresponding warning information through a cloud platform to effectively manage the risk of random block instability in deeply buried caverns. This tiered early warning mechanism can take different levels of early warning measures based on different ranges of comprehensive instability probability, thereby ensuring safety while avoiding overreaction. The three probability thresholds are a low threshold of 0.1, a medium threshold of 0.3, and a high threshold of 0.7. These thresholds are set based on the assessment of instability risk, with the low threshold indicating low risk, the medium threshold indicating moderate risk, and the high threshold indicating high risk. Sending warning information through the cloud platform enables rapid information transmission and sharing, ensuring that relevant personnel can take timely measures.

[0053] Specifically, when the overall instability probability A blue alert is triggered when the probability of instability is between 0.1 and 0.3, indicating a low risk of instability but still requiring monitoring. At this point, an SMS alert is sent via the cloud platform to notify relevant personnel to pay attention to changes in the monitoring data. When the overall probability of instability... A yellow alert is triggered when the probability of instability is between 0.3 and 0.7, indicating a moderate risk of instability and requiring appropriate measures. At this time, an audible and visual alarm is activated via the cloud platform to alert on-site personnel to safety precautions and to take corresponding preventative measures. When the overall instability probability... A value greater than or equal to 0.7 triggers a red alert, indicating a high risk of instability and requiring immediate emergency measures. At this point, the emergency support plan is automatically executed via the cloud platform to prevent instability events. Here, the overall instability probability is considered. It is calculated using a Bayesian update model, which combines simulated instability probability with real-time monitoring data, enabling a more accurate reflection of current instability risk. SMS alerts, audible and visual alarms, and emergency support plans are all specific forms of early warning measures, and they can be designed and implemented according to actual engineering needs.

[0054] Preferably, blue alert SMS notifications can be sent via the cloud platform's SMS service interface. By inputting the relevant personnel's mobile phone number and the alert information, an alert SMS can be sent to the designated mobile phone. Audible and visual alarm devices can be installed in key locations within the cavern, such as areas with frequent personnel activity. When a yellow alert is triggered, a control signal is sent via the cloud platform, causing the alarm device to emit sound and light signals to attract the attention of on-site personnel. The automatic execution of the emergency support plan can be achieved by connecting the cloud platform with the support equipment within the cavern. When a red alert is triggered, the cloud platform automatically controls the support equipment to perform support operations according to the preset support plan, such as shotcreting and installing anchor bolts.

[0055] Calculating the overall instability probability At that time, first simulate the instability probability As a prior probability, the conditional probability is then calculated based on real-time monitoring data. and marginal probability Finally, the overall instability probability was calculated using Bayes' theorem. .in, This indicates the probability of monitoring data occurring in the event of a block instability event. This represents the total probability of the monitored data occurring; it can be expressed by... , and conditional probability under block-stable events The calculations yielded the results. This method enables dynamic assessment and early warning of the risk of instability in random blocks within deeply buried caverns.

[0056] In some embodiments, it also includes: S7: Numerical simulation analysis of surrounding rock stability during deep-buried cavern excavation, specifically including: S7.1: A three-dimensional numerical model of the geological structure of the underground cavern area was established using Rhino software; S7.2: Invert the distribution of the geostress field in the underground cavern area using Flac3D software; S7.3: Simulates the distribution and variation of the surrounding rock stress field during excavation, and outputs the maximum and minimum principal stress values; S7.4: Analyze the deformation characteristics and evolution of the surrounding rock during the excavation process, and calculate the top settlement and sidewall convergence value; S7.5: Correct the block mechanical parameters in the Monte Carlo simulation based on the numerical simulation results.

[0057] This invention improves the accuracy of instability probability calculation by establishing a three-dimensional numerical model, inverting the distribution of the geostress field, simulating the stress field and deformation characteristics of the surrounding rock during the excavation process, and correcting the block mechanical parameters in the Monte Carlo simulation based on the numerical simulation results.

[0058] Specifically, a three-dimensional numerical model of the geological structure of the underground cavern area was established using Rhino software. Rhino is a 3D modeling software capable of accurately constructing the geometry and geological structure of underground caverns. The distribution of the geostress field in the underground cavern area was inverted using Flac3D software. Flac3D is a finite difference software specifically designed for geotechnical engineering, capable of simulating the mechanical behavior of soil and rock masses and inverting the geostress field. The distribution and variation of the surrounding rock stress field during excavation were simulated, outputting the maximum and minimum principal stress values. These stress values ​​are important indicators for assessing the stability of the surrounding rock. The deformation characteristics and evolution of the surrounding rock during excavation were analyzed, and the roof settlement and sidewall convergence values ​​were calculated. These deformation parameters reflect the dynamic changes of the surrounding rock during excavation. Based on the numerical simulation results, the block mechanical parameters in the Monte Carlo simulation were corrected. This step ensures that the instability probability calculation is closer to the actual engineering situation. In the above steps, Rhino software is used to construct an accurate three-dimensional model, Flac3D software is used to simulate and invert the geostress field, and parameters such as maximum principal stress, minimum principal stress, top settlement, and sidewall convergence value are key indicators for evaluating the stability of the surrounding rock.

[0059] Preferably, when establishing a three-dimensional numerical model, it is first necessary to collect detailed geological data of the underground cavern, including the type of rock mass, the location and orientation of structural planes, etc. In Rhino software, a three-dimensional geometric model of the underground cavern is constructed based on this data, and corresponding geological structural features are added. In Flac3D software, the inversion calculation of the geostress field is performed by inputting geological parameters and boundary conditions. During the simulated excavation process, different excavation stages are set, and corresponding rock mass elements are gradually removed, while the stress and deformation data of each stage are recorded. The maximum and minimum principal stress values ​​can be directly output through the post-processing function of Flac3D software, while the roof settlement and sidewall convergence values ​​can be obtained by measuring the displacement changes at specific points. Based on these numerical simulation results, the block mechanical parameters in the Monte Carlo simulation are corrected, such as adjusting the block cohesion and internal friction angle, to better reflect the mechanical properties under actual engineering conditions. Through these refined steps, the stability of the surrounding rock during the excavation of deeply buried caverns can be simulated more accurately, providing a scientific basis for engineering design and construction.

[0060] In some embodiments, it also includes: S8.1: The first surrounding rock score is calculated using the first surrounding rock classification method. The scoring indicators of the first surrounding rock classification method include rock strength, rock mass integrity, structural surface condition and groundwater conditions. S8.2: The second surrounding rock score is calculated using the second surrounding rock classification method, which is based on the rock mass quality index RQD, joint group number Jn, joint roughness coefficient Jr, joint alteration coefficient Ja, joint water reduction coefficient Jw and stress reduction factor SRF. S8.3: The third surrounding rock score is calculated using the third surrounding rock classification method. The scoring indicators of the third surrounding rock classification method include rock strength score, RQD score, joint spacing score, joint condition score, and groundwater score. S8.4: Dynamically adjust the values ​​of cohesion c and internal friction angle φ in the Monte Carlo simulation based on the surrounding rock quality classification results.

[0061] The first surrounding rock classification method adopts the hydropower surrounding rock classification system, which comprehensively evaluates the quality of surrounding rock through multiple indicators. Rock strength is determined through indoor uniaxial compressive strength tests, classifying rocks into hard, medium-hard, and soft categories. Rock mass integrity is comprehensively evaluated using the rock mass integrity index Kv and RQD value, reflecting the degree of development of fractures and joints in the rock mass. Structural surface condition indicators assess the occurrence, spacing, continuity, and infill characteristics of structural surfaces. Groundwater condition indicators consider the influence of groundwater level, water pressure, and seepage state on rock mass stability. The first surrounding rock score is calculated based on preset weights for each indicator. The second surrounding rock classification method adopts the rock mass quality Q-system classification method, a comprehensive evaluation system established based on extensive engineering practice. The rock mass quality index RQD is obtained through borehole core statistics, reflecting the integrity of the rock mass. The number of joint groups is determined through field window measurements, characterizing the complexity of the rock mass structure. The joint roughness coefficient is determined through structural surface morphology description and comparison with standard maps.

[0062] The joint alteration coefficient is evaluated based on the properties and thickness of the infill material on the structural surface. The joint water reduction coefficient is determined based on the groundwater seepage conditions. The stress reduction factor considers the rock mass strength attenuation effect under high ground stress conditions. The second surrounding rock score is calculated by multiplying the above parameters. The third surrounding rock classification method adopts the Rock Mass RMR classification method, which is a surrounding rock quality assessment system based on multi-factor weighted scoring. The rock strength score is assigned according to a preset range based on the uniaxial compressive strength test results. The RQD score is determined according to the rock mass quality index values ​​according to the grading standard. The joint spacing score is obtained through on-site structural surface measurement and statistics. The joint condition score comprehensively considers the joint surface roughness, infill material properties, and joint surface weathering degree. The groundwater score is determined based on the seepage state and water volume within the cavern. The third surrounding rock score is obtained by summing all the score items.

[0063] Based on the scoring values ​​obtained from three surrounding rock classification methods, a mapping relationship between surrounding rock quality and mechanical parameters is established. A quantitative relationship model between the surrounding rock score and cohesion *c* and internal friction angle *φ* is established through multiple regression analysis. When the surrounding rock score changes, the mechanical parameter values ​​in the Monte Carlo simulation are dynamically updated according to preset parameter adjustment rules. For high-quality surrounding rock with higher scores, the benchmark values ​​of cohesion *c* and internal friction angle *φ* are correspondingly increased. For low-quality surrounding rock with lower scores, the mechanical parameter values ​​are decreased according to preset reduction coefficients. By adjusting the mechanical parameters in real time, the Monte Carlo simulation more accurately reflects the actual surrounding rock conditions, improving the reliability of instability probability prediction.

[0064] Specifically, the surrounding rock score is calculated using the hydroelectric surrounding rock classification method, which comprehensively considers multiple indicators such as rock strength, rock mass integrity, structural surface condition, and groundwater conditions. Rock strength refers to the maximum stress that rock can withstand under external forces; rock mass integrity reflects the degree of development of fractures and joints in the rock mass; structural surface condition includes the attitude, spacing, and infill of structural surfaces; and groundwater conditions involve groundwater level, pressure, and water quality.

[0065] The Q-value can be calculated using the Q-system classification method. The Q-value is a dimensionless parameter that comprehensively reflects the quality of the surrounding rock. The calculation formula is Q=(RQD / Jn)×(Jr / Ja)×(Jw / SRF). Where RQD represents the rock mass quality index, Jn represents the number of joints, Jr represents the joint roughness, Ja represents the joint shear strength, Jw represents the groundwater influence factor, and SRF represents the stress reduction factor.

[0066] The RMR value is calculated using the RMR classification method. The RMR value is another parameter that comprehensively reflects the quality of the surrounding rock. The calculation formula is as follows: RMR = Rock Strength Score + RQD Score + Joint Spacing Score + Joint Condition Score + Groundwater Score. These classification methods and parameter settings provide a scientific basis for dynamically adjusting the cohesion c and internal friction angle φ in Monte Carlo simulations.

[0067] Preferably, when dynamically adjusting the values ​​of cohesion *c* and internal friction angle *φ* in the Monte Carlo simulation based on the surrounding rock quality classification results, the adjustments can be refined by combining specific engineering examples. For instance, in the classification method for surrounding rock in hydropower projects, rocks can be divided into categories such as hard rock, relatively hard rock, and soft rock based on their strength, with each category corresponding to a different strength range. Rock mass integrity can be assessed using the RQD value; a higher RQD value indicates a more intact rock mass. The assessment of structural plane conditions requires consideration of factors such as the orientation, spacing, and infill of structural planes. For example, smaller joint spacing generally indicates poorer rock mass stability. The assessment of groundwater conditions requires consideration of factors such as groundwater level, pressure, and quality; the presence of groundwater may reduce rock mass stability.

[0068] In the Q-system classification method, the RQD value can be calculated by drilling and sampling and measuring the percentage of rock cores longer than 10 cm in the total borehole length; the number of joints Jn can be calculated by the number of joints per unit area; the joint roughness Jr can be evaluated by the roughness of the joint surface; the joint shear strength Ja can be obtained by laboratory tests; and the groundwater influence factor Jw and stress reduction factor SRF need to be evaluated according to specific engineering conditions.

[0069] In the RMR classification method, rock strength score is determined based on the uniaxial compressive strength of the rock; RQD score is determined based on the magnitude of the RQD value; joint spacing score is determined based on the magnitude of the joint spacing; joint condition score is determined based on factors such as joint infill material and roughness; and groundwater score is determined based on factors such as groundwater level, pressure, and quality. Through these specific parameter settings and calculation methods, the quality of the surrounding rock can be more accurately assessed, and the values ​​of cohesion c and internal friction angle φ in the Monte Carlo simulation can be dynamically adjusted accordingly, thereby improving the accuracy of instability probability calculation.

[0070] In some embodiments, it also includes: S9: Enables data integration and visualization through a cloud platform, specifically including: S9.1: Establish a cloud database to store rock mass structure data, real-time monitoring data, simulated instability probability, and comprehensive instability probability; S9.2: Develop a data fusion algorithm to update the comprehensive instability probability calculation model in real time; S9.3: Construct a web-based visualization interface to dynamically display the warning level, the 3D model of the block instability risk area, and monitoring data curves; S9.4: Set multi-level user permissions to support remote access and push alert information to mobile terminals.

[0071] It should be noted that this invention utilizes a cloud platform for data integration and visualization, aiming to improve the efficiency and accuracy of early warning systems for instability in deeply buried caverns. A cloud platform is a network service platform based on cloud computing technology, capable of data storage, processing, and sharing. Data integration refers to consolidating data from different sources onto a unified platform for comprehensive analysis and processing. Visualization presents complex data information intuitively through graphics, charts, and other formats, facilitating quick understanding and decision-making by users. Through data integration and visualization on the cloud platform, real-time monitoring and dynamic management of early warning systems for instability in deeply buried caverns can be achieved, improving the reliability and practicality of the early warning system.

[0072] Specifically, a cloud database is established to store rock mass structure data, real-time monitoring data, simulated instability probability, and comprehensive instability probability. The cloud database is a cloud-based data storage solution that provides high availability, high scalability, and high security data storage services. Rock mass structure data includes structural surface feature data and rock mass quality parameter data, which are collected using equipment such as adit logging, 3D laser scanners, indoor testing machines, and seismic wave velocity meters. Real-time monitoring data is collected through displacement sensors, stress sensors, and microseismic monitoring equipment arrays deployed on the tunnel roof arches and sidewalls. These sensors can monitor rock mass deformation and stress changes in real time. Simulated instability probability and comprehensive instability probability are calculated using Monte Carlo simulation and a Bayesian update model. A data fusion algorithm is developed to update the comprehensive instability probability calculation model in real time. This algorithm can fuse simulated instability probability and real-time monitoring data to calculate a more accurate comprehensive instability probability. A web-based visualization interface is constructed to dynamically display warning levels, 3D models of block instability risk areas, and monitoring data curves. Users can intuitively view warning information and monitoring data through the web interface. Setting up multi-level user permissions supports remote access and mobile terminal alert push notifications. Users at different levels can access corresponding data and functions according to their permissions, and can also receive alert information in a timely manner through mobile terminals, ensuring that users can take timely measures.

[0073] Preferably, the cloud database can be constructed using distributed storage technologies, such as the Hadoop Distributed File System (HDFS) or Amazon S3, to improve data storage efficiency and reliability. Data fusion algorithms can employ machine learning algorithms, such as Bayesian networks or neural networks. These algorithms can automatically adjust model parameters based on historical and real-time data, improving the accuracy of comprehensive instability probability calculations. The web visualization interface can be constructed using modern web development frameworks, such as React or Vue.js, combined with 3D visualization libraries, such as Three.js, to dynamically display the 3D model of the block instability risk area. Monitoring data curves can be plotted using chart libraries, such as ECharts or D3.js, allowing users to intuitively view the changing trends of the monitoring data. Multi-level user permission settings can be implemented using a Role-Based Access Control (RBAC) model, assigning different permissions based on user roles to ensure data security and confidentiality. Through these specific technical implementation steps, an efficient, reliable, and easy-to-use cloud platform for early warning of instability in deeply buried caverns can be built, providing strong support for the safety management of deeply buried caverns.

[0074] The above description is merely an explanation of some preferred embodiments of the present invention and the technical principles employed. Those skilled in the art should understand that the scope of the invention as described in the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-described inventive concept. For example, technical solutions formed by substituting the above-described features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A cloud-based early warning method for the instability probability of random blocks in deeply buried caverns, characterized in that, Includes the following steps: S1: Collect rock mass structure data of the deep-buried cavern, including structural surface feature data and rock mass quality parameter data; S2: Construct a three-dimensional geological model based on the rock mass structure data and identify potential random blocks; S3: Calculate the instability probability of the potential random block using the Monte Carlo simulation method to obtain the simulated instability probability; S4: Real-time monitoring of deformation and stress data of deeply buried caverns is achieved through IoT sensors, generating real-time monitoring data; S5: Combine the simulated instability probability and the real-time monitoring data to calculate the comprehensive instability probability; S6: Based on the comprehensive instability probability, conduct risk warning and release warning information through the cloud platform.

2. The method for early warning of probability instability of random blocks in deep-buried caverns according to claim 1, characterized in that, The rock mass structure data collected in S1 specifically includes: S1.1: Obtain a structural surface statistics table through adit logging. The structural surface statistics table records the number and distribution location of JX shear fault zone, J joint, Gm extrusion surface, Fpd fault, JC extrusion fault zone and JK structural surface. S1.2: Scan the inner wall of the adit using a 3D laser scanner, setting up a scanning station every 3 meters to obtain structural surface attitude data, including strike, dip angle and dip direction; S1.3: Rock mechanics tests were conducted using an indoor testing machine to obtain uniaxial saturated compressive strength, rock mass integrity index Kv, and triaxial unloading mechanical parameters; S1.4: Combine the on-site measurement of rock wave velocity using a seismic wave velocity tester to calculate the rock mass integrity index Kv.

3. The method for early warning of probability instability of random blocks in deeply buried caverns according to claim 1, characterized in that, The construction of the three-dimensional geological model in S2 specifically includes: S2.1: Based on the scale of the structural surface, the structural surface is divided into five levels: Level I fault-type structural surface, Level II fracture-type structural surface, Level III non-penetrating structural surface, Level IV visible structural surface, and Level V hidden microstructural surface; S2.2: Using structural surface network simulation technology, based on the grouping results of structural surface attitude, the diameter, spacing and volume density parameters of structural surfaces are generated through probability distribution functions; S2.3: Process 3D laser scanning data using Cloudcompare and GIS software to determine the boundary conditions, spatial location, geometric shape, and volume of random blocks; S2.4: Based on the three-dimensional network simulation results of the structural plane, calculate the probability distribution function of the rock mass RQD value.

4. The method for early warning of probability instability of random blocks in deep-buried caverns according to claim 1, characterized in that, The calculation of instability probability using the Monte Carlo simulation method in S3 specifically includes: S3.1: Set the total number of Monte Carlo simulations N, where N ≥ 10000; S3.2: Each simulation randomly samples the geometric and mechanical parameters of the block. The geometric parameters include the block volume and the inclination angle of the structural surface, and the mechanical parameters include the cohesion c and the internal friction angle φ. S3.3: The safety factor Fs of the block is calculated using the limit equilibrium theory. When Fs < 1, it is determined to be unstable. S3.4: Probability of Instability The calculation formula is: in, N represents the number of unstable blocks in the simulation, and N is the total number of simulations.

5. The method for early warning of probability instability of random blocks in deeply buried caverns according to claim 1, characterized in that, The S4 section specifically includes real-time monitoring of deformation and stress data of the deeply buried cavern via IoT sensors, including: S4.1: Arrange displacement sensors, stress sensors, and microseismic monitoring equipment arrays in the roof arch and sidewalls of the cavern; S4.2: The displacement sensor uses a multi-point displacement meter to measure the relative displacement of the rock mass; S4.3: The stress sensor uses a vibrating wire stress gauge to measure the stress change value of the rock mass; S4.4: Microseismic monitoring equipment records the frequency and energy release of rock mass fracturing events; S4.5: The collected data is uploaded to the cloud platform database in real time via the wireless transmission module.

6. The method for early warning of probability instability of random blocks in deep-buried caverns according to claim 1, characterized in that, The calculation of the comprehensive instability probability by integrating the simulated instability probability and the real-time monitoring data in S5 specifically includes: S5.1: Establish a Bayesian update model to simulate instability probability. As a priori probability; S5.2: Define the conditional probability of monitoring data P(D|F), where D represents real-time monitoring data and F represents the block instability event; S5.3: Calculate the marginal probability P(D) of the monitoring data; S5.4: Overall Instability Probability The calculation formula is: in, , This indicates a stable event in the block.

7. The method for early warning of probability instability of random blocks in deeply buried caverns according to claim 1, characterized in that, The risk warning based on the comprehensive instability probability in S6 specifically includes: S6.1: Set three levels of probability thresholds: low threshold 0.1, medium threshold 0.3, and high threshold 0.7; S6.2: When 0.1 ≤ When the value is less than 0.3, a blue alert is triggered, and an SMS reminder is sent via the cloud platform. S6.3: When 0.3 ≤ When the value is less than 0.7, a yellow alert is triggered, and the audible and visual alarm device is activated through the cloud platform. S6.4: When When the value is ≥0.7, a red alert is triggered, and an emergency support plan is automatically executed through the cloud platform.

8. The method for early warning of probability instability of random blocks in deep-buried caverns according to claim 1, characterized in that, Also includes: S7: Numerical simulation analysis of surrounding rock stability during deep-buried cavern excavation, specifically including: S7.1: A three-dimensional numerical model of the geological structure of the underground cavern area was established using Rhino software; S7.2: Invert the distribution of the geostress field in the underground cavern area using Flac3D software; S7.3: Simulates the distribution and variation of the surrounding rock stress field during excavation, and outputs the maximum and minimum principal stress values; S7.4: Analyze the deformation characteristics and evolution of the surrounding rock during the excavation process, and calculate the top settlement and sidewall convergence value; S7.5: Correct the block mechanical parameters in the Monte Carlo simulation based on the numerical simulation results.

9. The method for early warning of the probability of instability of random blocks in a deep-buried cavern according to claim 1, characterized in that, Also includes: S8: Adjust the instability probability calculation parameters based on the surrounding rock quality classification results, specifically including: S8.1: The first surrounding rock score is calculated using the first surrounding rock classification method. The scoring indicators of the first surrounding rock classification method include rock strength, rock mass integrity, structural surface condition and groundwater conditions. S8.2: The second surrounding rock score is calculated using the second surrounding rock classification method, which is based on the rock mass quality index RQD, joint group number Jn, joint roughness coefficient Jr, joint alteration coefficient Ja, joint water reduction coefficient Jw and stress reduction factor SRF. S8.3: The third surrounding rock score is calculated using the third surrounding rock classification method. The scoring indicators of the third surrounding rock classification method include rock strength score, RQD score, joint spacing score, joint condition score, and groundwater score. S8.4: Dynamically adjust the values ​​of cohesion c and internal friction angle φ in the Monte Carlo simulation based on the surrounding rock quality classification results.

10. The method for early warning of the probability of instability of random blocks in a deep-buried cavern according to claim 1, characterized in that, Also includes: S9: Enables data integration and visualization through a cloud platform, specifically including: S9.1: Establish a cloud database to store rock mass structure data, real-time monitoring data, simulated instability probability, and comprehensive instability probability; S9.2: Develop a data fusion algorithm to update the comprehensive instability probability calculation model in real time; S9.3: Construct a web-based visualization interface to dynamically display the warning level, the 3D model of the block instability risk area, and monitoring data curves; S9.4: Set multi-level user permissions to support remote access and push alert information to mobile terminals.

Images