Tunnel gas emission dynamic prediction method and system based on multi-source monitoring

Abstract

The present application relates to the technical field of underground engineering, in particular to a tunnel gas emission dynamic prediction method and system based on multi-source monitoring, comprising: dividing gas emission monitoring units in the surrounding rock of the working face, and setting the arrangement points of gas pressure sensors, the collection frequency of gas pressure sensors and the gas concentration threshold; injecting gas historical monitoring records into the gas emission monitoring units, and based on the collection frequency and the gas concentration threshold, performing evolution calculation on the gas occurrence state of the surrounding rock and the corresponding area of the geological structure feature point for each geological structure feature point to obtain the initial amount of gas emission at the current time corresponding to each calculation unit. The present application can realize the visualization of gas migration behavior by outputting the coordinate trajectory of virtual tracer particles, help engineers understand and evaluate the dynamic process of gas movement, and facilitate subsequent analysis and processing.

Description

Technical Field

[0001] This invention relates to the field of underground engineering technology, specifically to a method and system for dynamic prediction of tunnel gas outburst based on multi-source monitoring. Background Technology

[0002] Tunnel gas refers to various gases that may be encountered during tunnel construction, the most common being methane (CH4) and carbon dioxide (CO2). When the gas concentration reaches a certain level and mixes with air, it may form an explosive gas mixture, which can easily cause an explosion when it encounters a source of ignition. Furthermore, high concentrations of gas may displace oxygen in the air, posing a risk of asphyxiation to workers. Gas leaks and their handling may also delay project progress and increase construction costs.

[0003] Currently, traditional methods typically rely on static monitoring and historical data, lacking real-time capabilities. This can lead to a failure to provide timely warnings when gas emission rates change, thereby increasing the risk of accidents. Furthermore, traditional methods often employ simple linear models or empirical formulas, which struggle to accurately capture the complex dynamic changes in gas emission rates, resulting in inaccurate predictions that cannot effectively support construction decisions. Since traditional models are generally based on fixed assumptions, they cannot flexibly adapt to changes in the construction environment, leading to poor applicability under different geological conditions and an inability to effectively handle complex and variable environments.

[0004] In addition, traditional methods often lack visualization tools for the gas migration process, making it difficult for engineers to intuitively understand the movement patterns and dynamic behavior of gas, thus affecting the effectiveness of analysis and treatment. Furthermore, traditional methods lack flexible emergency response capabilities when faced with emergencies, and may require a long time to adjust strategies, making it difficult to quickly adapt to changes on site. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic prediction of tunnel gas outburst based on multi-source monitoring, comprising:

[0006] Obtain geological survey data files and historical gas monitoring records for the tunnel construction section; the tunnel construction section is formed by combining the surrounding rock at the tunnel face and the excavated roadway.

[0007] The gas outburst monitoring units were divided in the surrounding rock of the working face, and the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold were set.

[0008] Extract the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file;

[0009] Historical gas monitoring records are injected into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, the evolution calculation of the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point is performed to obtain the initial gas emission amount at the current moment for each calculation unit.

[0010] At the current moment, virtual tracer particles are generated in each computational unit of the geological structural feature point; based on the current moment, the drift velocity and coordinates of all virtual tracer particles, the coordinates of all virtual tracer particles at the next moment are calculated; where the drift velocity of the virtual tracer particles refers to the initial gas outburst at the current moment corresponding to the location of the computational unit where the virtual tracer particle is located.

[0011] Preferably, the method further includes:

[0012] Determine whether the total gas emission in the tunnel construction section has reached the warning threshold;

[0013] If the total gas emission in the tunnel construction section does not reach the warning threshold, the collection frequency and gas concentration threshold are based on the initial gas emission at the current moment of the excavated roadway and each calculation unit.

[0014] The evolution calculation of the gas occurrence state of the excavated roadway and the area corresponding to each geological structural feature point is performed to obtain the initial gas outburst amount at the next moment for each calculation unit in each geological structural feature point.

[0015] Preferably, the method further includes:

[0016] Update the initial gas outburst amount for the next moment to the initial gas outburst amount for the current moment for each computational unit in each geological structural feature point, and take the next moment as the current moment, and return to execute the generation of virtual tracer particles in each computational unit of the geological structural feature point in the current moment.

[0017] Based on the current moment, the drift velocity and coordinates of all virtual tracer particles, calculate the coordinates of all virtual tracer particles at the next moment, until the total gas outburst in the tunnel construction section reaches the warning threshold.

[0018] If the total gas outburst in the tunnel construction section reaches the warning threshold, the coordinate trajectories of the virtual tracer particles at all times will be output.

[0019] Preferably, after extracting the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file, the method further includes:

[0020] The coordinates of each geological structural feature point in the surrounding rock of the tunnel face are encoded to obtain the feature identification code corresponding to each geological structural feature point.

[0021] Preferably, based on the acquisition frequency and gas concentration threshold, the evolution calculation of the gas occurrence state in the area corresponding to the surrounding rock and geological structure feature points is performed to obtain the initial gas outburst amount at the current moment for each calculation unit, including:

[0022] For each calculation unit in each geological structural feature point, a gas outburst prediction sub-model is constructed;

[0023] Obtain the geological stress direction of the calculation unit and the relative distance between the calculation unit and the tunnel axis;

[0024] Based on the acquisition frequency, gas concentration threshold, geological stress direction, and relative distance, the gas outburst prediction sub-model is trained and inferred to obtain the initial gas outburst amount at the current moment corresponding to the calculation unit.

[0025] Preferably, based on the current moment, the drift velocities and coordinates of all virtual tracer particles are used to calculate the coordinates of all virtual tracer particles at the next moment, including:

[0026] For each virtual tracer particle, the feature identifier code corresponding to the geological structure feature point is written into the data structure of the virtual tracer particle, and the initial gas outburst at the current moment corresponding to the location of the virtual tracer particle in the computing unit is determined as the drift velocity of the virtual tracer particle, thus obtaining the target tracer particle.

[0027] Calculate the time step between the next time step and the current time step;

[0028] Based on the time step, the drift velocity of the target tracer particle, and its coordinates, calculate the coordinates of the virtual tracer particle at the next moment.

[0029] Preferably, determining whether the total gas emission in the tunnel construction section has reached the warning threshold includes:

[0030] Based on the coordinate trajectories of virtual tracer particles at all times, calculate the particle aggregation density of each computational unit in each geological structural feature point.

[0031] If the particle aggregation density of each calculation unit in each geological structural feature point is greater than the set aggregation threshold, then the total gas outburst of the tunnel construction section is determined to have reached the warning critical value.

[0032] If the particle aggregation density of each calculation unit in each geological structural feature point is not greater than the set aggregation threshold, then it is determined that the total gas outburst of the tunnel construction section has not reached the warning threshold.

[0033] A dynamic prediction system for tunnel gas emission based on multi-source monitoring is applicable to the aforementioned dynamic prediction method for tunnel gas emission based on multi-source monitoring, including:

[0034] The data acquisition module is used to acquire geological survey data files and historical gas monitoring records for the tunnel construction section; the tunnel construction section is formed by combining the surrounding rock at the tunnel face and the excavated roadway.

[0035] The area division module is used to divide the gas outburst monitoring units in the surrounding rock of the working face, and to set the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold.

[0036] The information extraction module is used to extract the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file;

[0037] The gas monitoring module is used to inject historical gas monitoring records into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, the module performs evolution calculations on the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point, and obtains the initial gas emission amount at the current moment for each calculation unit.

[0038] The coordinate calculation module is used to generate virtual tracer particles in each calculation unit of the geological structure feature point at the current moment, and calculate the coordinates of all virtual tracer particles at the next moment based on the current moment, the drift velocity and coordinates of all virtual tracer particles; where the drift velocity of the virtual tracer particles refers to the initial gas outburst at the current moment corresponding to the location of the calculation unit where the virtual tracer particle is located.

[0039] Compared with the prior art, the beneficial effects of the present invention are:

[0040] This invention enables timely detection of potential gas accumulation risks and early warnings by real-time monitoring and dynamic prediction of gas emission, thereby reducing the occurrence of accidents such as gas explosions and improving the safety of tunnel construction. Furthermore, through the generation and drift calculation of virtual tracer particles, it is possible to predict the dynamic changes in gas emission, which helps to take preventive measures in advance. Moreover, by outputting the coordinate trajectory of virtual tracer particles, it is possible to visualize the gas migration behavior, helping engineers understand and evaluate the dynamic process of gas movement, which is convenient for subsequent analysis and processing.

[0041] This invention updates the initial gas outburst volume in real time and adjusts the prediction model according to the current situation, making the prediction results more timely and adaptable, and able to cope with complex construction environments. Furthermore, by fusing and analyzing multi-source data, it maximizes the use of existing geological and monitoring data, making the prediction process more efficient and reducing the decision-making risks caused by insufficient data. Attached Figure Description

[0042] Figure 1 This is a schematic flowchart of the overall method in one embodiment of the present invention;

[0043] Figure 2 This is a schematic diagram of the overall system architecture in one embodiment of the present invention.

[0044] In the diagram: 1. Data acquisition module; 2. Area division module; 3. Information extraction module; 4. Gas monitoring module; 5. Coordinate calculation module. Detailed Implementation

[0045] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0046] Example 1, please refer to Figure 1 This invention provides a technical solution: a method for dynamic prediction of tunnel gas outburst based on multi-source monitoring, comprising:

[0047] S1. Obtain geological survey data files and historical gas monitoring records for the tunnel construction section; the tunnel construction section is formed by combining the surrounding rock at the tunnel face and the excavated roadway.

[0048] S2. Divide the gas outburst monitoring units in the surrounding rock of the working face, and set the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold.

[0049] S3. Extract the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file;

[0050] S4. Inject the historical gas monitoring records into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, perform evolution calculations on the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point to obtain the initial gas emission amount at the current moment for each calculation unit.

[0051] S5. At the current moment, generate virtual tracer particles in each calculation unit of the geological structure feature point; calculate the coordinates of all virtual tracer particles at the next moment based on the current moment, the drift velocity and coordinates of all virtual tracer particles; where the drift velocity of the virtual tracer particles refers to the initial amount of gas outburst at the current moment corresponding to the location of the calculation unit where the virtual tracer particle is located.

[0052] It should be noted that, firstly, geological survey data files and historical gas monitoring records of the tunnel construction section need to be collected; these data include the characteristics of the rock surrounding the tunnel face and past changes in gas concentration; within the surrounding rock of the tunnel face, the area is divided into multiple gas emission monitoring units; these units can be divided based on factors such as geological structure and rock characteristics; then, the location of the gas pressure sensor to be installed in each monitoring unit, as well as the sensor's acquisition frequency (e.g., once per minute) and the gas concentration threshold, such as triggering an alarm when the concentration exceeds a certain value;

[0053] The coordinates of all important geological structural feature points of the surrounding rock at the tunnel face are extracted from the geological survey data files. These feature points may be important geological phenomena such as faults and folds, which will affect the gas occurrence state. Historical gas monitoring records are introduced into the gas emission monitoring unit. For each geological structural feature point, the gas occurrence state is calculated by combining the acquisition frequency and concentration threshold. The core is to construct or call a predictive model that can reflect the gas migration law in the pore-fracture medium. The model can be a physical and mathematical model based on Darcy's law, Fick's diffusion law, etc., or a machine learning model trained on historical monitoring data (such as an LSTM time series prediction model). During calculation, the key input parameters are the acquisition frequency (Δt), the gas concentration threshold (C_th), the geological stress direction vector of the unit (σ), and the relative distance from the unit to the working face (L). The model outputs the gas emission rate (Q_initial) of each calculation unit at the "current moment" by solving the governing equations under specific boundary conditions (such as the concentration threshold) or by performing forward inference. The unit is usually m³ / s or m³ / min, which is the initial amount of gas emission at the current moment.

[0054] In the computational unit of each feature point, virtual tracer particles are generated based on the initial gas emission at the current moment. These particles simulate the movement of gas, reflecting its distribution near the geological feature point. The coordinates of all particles at the next moment are calculated based on the drift velocity of the virtual tracer particles (determined by the current initial gas emission). The average seepage velocity (v = Q_initial / A) at the interface of the computational unit is calculated based on the area (A) and the initial gas emission (Q_initial). The direction of this velocity is determined by the direction of geological stress and the spatial structure of the tunnel (e.g., primarily pointing towards the free face of the tunnel). This average seepage velocity is then assigned to the virtual tracer particles generated within the computational unit as their drift velocity vector. The random motion of the particles themselves (such as Brownian motion) can be simulated by superimposing a random velocity component to more realistically reflect the diffusion behavior of gas molecules.

[0055] For example: Suppose that in a tunnel construction section, geological survey results show several important structural features, such as a fault and a fold; through monitoring, it was found that the gas concentration near the fault has exceeded the safety threshold multiple times in historical data, so multiple gas pressure sensors were deployed in this area; the collected geological survey data shows that the coordinates of the fault are (10,20) and the coordinates of the fold are (15,25); based on the geological features, the construction section is divided into three monitoring units: A, B, and C, where A is close to the fault, B is close to the fold, and C is the middle area;

[0056] Two sensors are placed in unit A, one in unit B, and one in unit C, with a sampling frequency set to once per minute. Historical monitoring data shows that the initial gas emission rate in unit A is 20 cubic meters, in unit B it is 10 cubic meters, and in unit C it is 5 cubic meters. At the current moment, based on these initial emission rates, a corresponding number of virtual tracer particles are generated, such as 20 particles in unit A, 10 in unit B, and 5 in unit C. If the particle drift speed in unit A is 2 meters per minute, then at the next moment, the coordinates of these 20 particles will expand outward by 2 meters, updating to (12, 20), (12, 20.1), and so on.

[0057] In an optional embodiment, the method further includes:

[0058] Determine whether the total gas emission in the tunnel construction section has reached the warning threshold;

[0059] If the total gas emission in the tunnel construction section does not reach the warning threshold, the collection frequency and gas concentration threshold are based on the initial gas emission at the current moment of the excavated roadway and each calculation unit.

[0060] The evolution calculation of the gas occurrence state of the excavated roadway and the area corresponding to each geological structural feature point is performed to obtain the initial gas outburst amount at the next moment for each calculation unit in each geological structural feature point.

[0061] It should be noted that, firstly, it is necessary to assess whether the total gas emission in the current tunnel construction section has reached the warning threshold. This warning threshold is set based on historical data, geological characteristics, and construction conditions. When the total gas emission exceeds this threshold, a safety alarm may be triggered, requiring immediate action. If the total gas emission has not reached the warning threshold, the construction team will continue routine monitoring and analyze the initial gas emission of the excavated tunnels and each calculation unit. This means that construction can continue within a certain range, but vigilance must still be maintained.

[0062] In subsequent analysis, the previously set acquisition frequency (e.g., once per minute) and gas concentration threshold (e.g., a specific concentration value) will be used. These parameters help monitor gas conditions in real time and respond accordingly. Based on the initial gas emission at the current moment, the evolution calculation of the gas occurrence state is performed on the excavated roadway and the area of ​​each geological structural feature point. This is to predict the gas emission at the next moment. This calculation usually takes into account factors such as surrounding geological conditions and pressure changes. Through evolution calculation, the initial gas emission of each geological structural feature point and its corresponding calculation unit at the next moment can be obtained. This data can help construction personnel assess future safety risks and adjust construction plans according to changes in gas levels.

[0063] For example: Suppose that in a tunnel construction section, monitoring shows that the current total gas emission is 15 cubic meters, while the warning threshold is 20 cubic meters; therefore, the current gas emission has not reached the warning threshold. Next, the construction team will perform the following operations: The team finds that the current gas emission is normal and has not yet triggered a safety alarm; data is collected at a frequency of once per minute, and the gas concentration threshold is set at 10%; in the excavated roadways, recent monitoring data shows that the initial gas emission in unit A (near the fault) is 5 cubic meters, and the initial gas emission in unit B (near the fold) is... The initial gas outburst was 3 cubic meters, and the initial gas outburst of Unit C (the intermediate area) was 7 cubic meters. Based on the current initial gas outburst and the surrounding geological conditions, evolution calculations were performed. Assuming that the calculations showed that: the initial gas outburst of Unit A at the next moment was 6 cubic meters; the initial gas outburst of Unit B at the next moment was 4 cubic meters; and the initial gas outburst of Unit C at the next moment was 8 cubic meters, the construction team recorded these data and decided whether to adjust the construction plan or take additional protective measures based on the newly calculated gas outburst to ensure construction safety.

[0064] In an optional embodiment, the method further includes:

[0065] Update the initial gas outburst amount for the next moment to the initial gas outburst amount for the current moment for each computational unit in each geological structural feature point, and take the next moment as the current moment, and return to execute the generation of virtual tracer particles in each computational unit of the geological structural feature point in the current moment.

[0066] Based on the current moment, the drift velocity and coordinates of all virtual tracer particles, calculate the coordinates of all virtual tracer particles at the next moment, until the total gas outburst in the tunnel construction section reaches the warning threshold.

[0067] If the total gas outburst in the tunnel construction section reaches the warning threshold, the coordinate trajectories of the virtual tracer particles at all times will be output.

[0068] It should be noted that in each computational unit, the initial gas emission rate at the current moment is updated to the value at the next moment. This step applies the latest gas data to the model for more accurate subsequent analysis. As time progresses, the gas emission rate at the next moment becomes the new current moment; this is a time-iterative process that ensures the model can be continuously updated and reflect real-time conditions. In the computational unit for each geological structural feature point, several virtual tracer particles are generated; each particle represents a gas molecule and can simulate the gas diffusion process. These particles will be used to track the movement trajectory of gas in the tunnel.

[0069] Based on the current moment and the drift velocity of all virtual tracer particles (which may be related to factors such as gas concentration, pressure, and temperature) and their initial coordinates, the new coordinates of all particles at the next moment are calculated; this process simulates the diffusion and movement of gas in the tunnel over time; the above process is repeated until the total gas outburst in the tunnel construction section reaches the warning threshold; if the total gas outburst exceeds this threshold at any moment, the simulation is stopped, and the coordinate trajectories of the virtual tracer particles at all moments are prepared for output; once the warning threshold is reached, the position trajectories of the virtual tracer particles throughout the simulation process are recorded and output; this data can help analyze the gas diffusion patterns and potential risk areas;

[0070] For example: Suppose there are three calculation units (A, B, and C) in a tunnel construction section, with initial gas emission rates of 5 cubic meters, 3 cubic meters, and 7 cubic meters for each unit, respectively, and the current total gas emission rate is 15 cubic meters. The initial gas emission rate at the current moment is updated to the value at the next moment (e.g., unit A is updated to 5 cubic meters, unit B to 3 cubic meters, and unit C to 7 cubic meters). In each of the three units A, B, and C, 10 virtual tracer particles are generated, for a total of 30 particles. Assume that the drift velocity of each particle is a different value, for example, the particle velocity in unit A is 0.5 m / s, in unit B it is 0.3 m / s, and in unit C it is 0.4 m / s.

[0071] After a period of time (e.g., 1 second), new coordinates are calculated based on velocity: particles in unit A move 0.5 meters along the tunnel direction; particles in unit B move 0.3 meters; particles in unit C move 0.4 meters. After each update, the gas emission amount at the new moment is calculated. If the updated total amount still does not reach the warning threshold (e.g., 20 cubic meters), new particles are generated and the drift calculation is repeated. Assuming that after several iterations, the total gas emission amount eventually reaches the warning threshold, the coordinate trajectories of all particles at each moment are recorded. The output includes trajectory data of all virtual tracer particles, which can be used to analyze the gas diffusion and impact range to formulate corresponding safety measures.

[0072] In an optional embodiment, after extracting the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file, the method further includes:

[0073] The coordinates of each geological structural feature point in the surrounding rock of the tunnel face are encoded to obtain the feature identification code corresponding to each geological structural feature point.

[0074] It should be noted that in the surrounding rock of the tunnel face, the first step is to identify key geological structural features. These features may include cracks, bedding, lithological variations, faults, etc., all of which are important factors affecting gas outbursts. Each identified feature is assigned a coordinate value based on its position in three-dimensional space. Typically, these coordinates are represented in the form of X, Y, and Z, corresponding to the horizontal and vertical positions, respectively. For ease of management and reference, the coordinates of each feature need to be converted into a unique feature identifier. This encoding process can be carried out in various ways; for example, a string can be generated based on the coordinate values, or a short code can be converted into the coordinates using certain rules. After generating the feature identifiers, they can be used to track and record the relationship between gas outbursts and each feature. In data analysis and simulation, the feature identifiers make the information of specific feature points easier to access and manage.

[0075] For example: Suppose that the following geological structural features are identified in the surrounding rock of a tunnel face: Feature 1: Crack (coordinates: X=10, Y=20, Z=5); Feature 2: Lithological variation (coordinates: X=15, Y=25, Z=10); Feature 3: Fault (coordinates: X=12, Y=22, Z=7); Next, these feature points are encoded: Feature 1 (crack): coordinates (X=10, Y=20, Z=5) can be converted into feature identifier code "F1", where "F" represents "Feature", and the following number indicates the sequence number of the feature point; Feature 2 (lithological variation): coordinates (X=15, Y=25, Z=10) is encoded as "F2"; Feature 3 (fault): coordinates (X=12, Y=22, Z=7) is encoded as "F3"; Finally, the feature points and their corresponding identifier codes are as follows: Crack—F1; Lithological variation—F2; Fault—F3.

[0076] In an optional embodiment, based on the acquisition frequency and gas concentration threshold, the evolution calculation of the gas occurrence state in the area corresponding to the surrounding rock and geological structure feature points is performed to obtain the initial gas outburst amount at the current moment for each calculation unit, including:

[0077] For each calculation unit in each geological structural feature point, a gas outburst prediction sub-model is constructed;

[0078] Obtain the geological stress direction of the calculation unit and the relative distance between the calculation unit and the tunnel axis;

[0079] Based on the acquisition frequency, gas concentration threshold, geological stress direction, and relative distance, the gas outburst prediction sub-model is trained and inferred to obtain the initial gas outburst amount at the current moment corresponding to the calculation unit.

[0080] It should be noted that for each geological structural feature point, it is first divided into multiple computational units; each computational unit can be regarded as a small region with relatively uniform geological features; for each computational unit, a gas emission prediction sub-model is established. The input feature vector of this model includes: the reciprocal of the acquisition frequency (characterizing the monitoring time scale), the gas concentration threshold, the component of the geological stress direction in the tunnel axis (cosθ), and the relative distance from the computational unit to the tunnel face; the output is the initial gas emission amount of the computational unit; the model is trained using a historical monitoring dataset {(feature 1, feature 2, ..., actual emission amount)} to minimize the prediction error; the goal of the sub-model is to predict the gas emission amount within the unit based on the input geological and environmental parameters; the geological stress direction refers to the direction of stress action caused by factors such as crustal movement and rock pressure in a specific area; the stress direction of each computational unit can be determined through measurement or simulation; understanding the stress direction is crucial for predicting gas emission because it affects the stability of the rock strata and the release of gas;

[0081] The position of each computational unit relative to the tunnel axis is also important, and this is usually expressed as a distance. This relative distance can help predict the migration speed and outburst volume of gas in the tunnel. During the model training phase, the collected frequency data (such as time series data recorded by monitoring equipment), gas concentration threshold, stress direction, and relative distance are used as inputs to train the gas outburst prediction sub-model. During the inference phase, the trained model is used in conjunction with the actual monitoring data at the current moment to perform inference to obtain the initial gas outburst volume corresponding to the computational unit at the current moment.

[0082] For example: Suppose that during tunnel construction, researchers identify a feature point with a crack and certain geological stress below it. The area around this feature point is divided into several computational units, each with different geological conditions. Computational unit A: Geological stress direction: southeast; relative distance on the tunnel axis: 10 meters; monitoring data (e.g., gas concentration fluctuations over the past hour, collected every minute). Computational unit B: Geological stress direction: northwest; relative distance on the tunnel axis: 20 meters; monitoring data: similar to computational unit A. For these two computational units, gas emission prediction sub-models are constructed respectively. The inputs to the models may include: the geological stress direction, relative distance, and historical gas concentration data for computational unit A; and the same input information for computational unit B. During training, the model learns the influence of different stress directions and distances on the gas emission rate, thus accurately predicting the initial gas emission rate.

[0083] In an optional embodiment, calculating the coordinates of all virtual tracer particles at the next moment, based on the current moment, the drift velocity, and coordinates of all virtual tracer particles, includes:

[0084] For each virtual tracer particle, the feature identifier code corresponding to the geological structure feature point is written into the data structure of the virtual tracer particle, and the initial gas outburst at the current moment corresponding to the location of the virtual tracer particle in the computing unit is determined as the drift velocity of the virtual tracer particle, thus obtaining the target tracer particle.

[0085] Calculate the time step between the next time step and the current time step;

[0086] Based on the time step, the drift velocity of the target tracer particle, and its coordinates, calculate the coordinates of the virtual tracer particle at the next moment.

[0087] It should be noted that each virtual tracer particle represents a simulation of a type of gas molecule or emission rate. To track the behavior of these particles, a data structure needs to be assigned to each particle, containing its associated attributes, such as the feature code of the geological structural feature point it corresponds to. For example, assuming a virtual tracer particle is assigned to a computational unit of a fracture, the feature code recorded in its data structure might be "F1," indicating that the particle is affected by the fracture. Based on the previously established gas emission prediction sub-model, each computational unit has a corresponding initial gas emission rate at the current moment. This initial rate can be regarded as the drift velocity of the virtual tracer particle, as it reflects the gas release rate within the computational unit. For example, if the initial gas emission rate of computational unit A is 50 cubic meters per hour, then the drift velocity of the virtual tracer particles in this computational unit is set to a value corresponding to this emission rate, such as 0.0139 meters per second.

[0088] The time step refers to the time interval from the current moment to the next moment, which is very important in simulations because it directly affects the updating of particle positions. Typically, the time step can be set according to the actual monitoring frequency or simulation requirements, such as 5 minutes (300 seconds) or 10 minutes (600 seconds). Once the drift velocity and time step of the target tracer particle are determined, the coordinates of the virtual tracer particle at the next moment can be calculated. The particle's position is updated based on the product of the drift velocity and the time step, building upon the current position. For example, if the current coordinates of the virtual tracer particle are (10, 20, 5), the drift velocity is 0.0139 m / s, and the time step is 300 seconds, then at the next moment, the new coordinates will be the original coordinates plus the drift distance (0.0139 m / s * 300 seconds). Thus, the new coordinates can be updated to (10 + 4.17, 20, 5), i.e., (14.17, 20, 5).

[0089] In an optional embodiment, determining whether the total gas emission in the tunnel construction section has reached the warning threshold includes:

[0090] Based on the coordinate trajectories of virtual tracer particles at all times, calculate the particle aggregation density of each computational unit in each geological structural feature point.

[0091] If the particle aggregation density of each calculation unit in each geological structural feature point is greater than the set aggregation threshold, then the total gas outburst of the tunnel construction section is determined to have reached the warning critical value.

[0092] If the particle aggregation density of each calculation unit in each geological structural feature point is not greater than the set aggregation threshold, then it is determined that the total gas outburst of the tunnel construction section has not reached the warning threshold.

[0093] It should be noted that particle aggregation density refers to the number of virtual tracer particles aggregated per unit volume or unit area within a computational unit at a specific geological structural feature point. By analyzing the coordinate trajectories of virtual tracer particles at all times, the number of particles in each computational unit can be counted, and the aggregation density can be calculated accordingly. For example, if the volume of a computational unit is 10 cubic meters, and 10 virtual tracer particles are detected in this computational unit, then the particle aggregation density of this computational unit is 1 particle / cubic meter.

[0094] Before determining the aggregation density, an aggregation threshold needs to be set. This threshold is determined based on historical data, laboratory studies, or safety standards. If the particle aggregation density within a calculation unit exceeds this threshold, the gas emission of that calculation unit is considered to pose a potential safety hazard. For all calculation units at each geological structural feature point, it is checked whether their particle aggregation density is greater than the set aggregation threshold. If all calculation units meet this condition, the total gas emission of the tunnel construction section can be considered to have reached the warning threshold; otherwise, it has not reached the warning threshold.

[0095] For example: Suppose the following conditions are monitored in a tunnel construction section: Geological structural feature point: G1; Calculation units: Unit A: volume 10 cubic meters, 8 virtual tracer particles were detected, with an aggregation density of 0.8 particles / cubic meter; Unit B: volume 15 cubic meters, 20 virtual tracer particles were detected, with an aggregation density of 1.33 particles / cubic meter; Unit C: volume 5 cubic meters, 4 virtual tracer particles were detected, with an aggregation density of 0.8 particles / cubic meter; Assume the set aggregation threshold is... 1 particle / cubic meter; Judgment process: Calculate the aggregation density of each unit: Unit A: 0.8 < 1 (not exceeding the threshold); Unit B: 1.33 > 1 (exceeding the threshold); Unit C: 0.8 < 1 (not exceeding the threshold); The aggregation densities of Unit A and Unit C did not exceed the set threshold, only Unit B exceeded the threshold; Therefore, the condition that "the particle aggregation density of each calculation unit is greater than the set aggregation threshold" is not met; The final conclusion is: The total amount of gas emission in the tunnel construction section has not reached the warning critical value.

[0096] Example 2, please refer to Figure 2 This invention provides a technical solution: a dynamic prediction system for tunnel gas emission based on multi-source monitoring, applicable to the aforementioned dynamic prediction method for tunnel gas emission based on multi-source monitoring, comprising:

[0097] Data acquisition module 1 is used to acquire geological survey data files and gas history monitoring records of the tunnel construction section; wherein, the tunnel construction section is formed by the combination of the surrounding rock at the tunnel face and the excavated roadway;

[0098] The area division module 2 is used to divide the gas outburst monitoring units in the surrounding rock of the working face, and to set the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold.

[0099] Information extraction module 3 is used to extract the coordinate information of all geological structural feature points in the surrounding rock of the working face from the geological survey data file;

[0100] Gas monitoring module 4 is used to inject historical gas monitoring records into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, the evolution calculation of the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point is performed to obtain the initial gas emission amount at the current moment for each calculation unit.

[0101] The coordinate calculation module 5 is used to generate virtual tracer particles in each calculation unit of the geological structure feature point at the current moment, and calculate the coordinates of all virtual tracer particles at the next moment based on the current moment, the drift velocity and coordinates of all virtual tracer particles; where the drift velocity of the virtual tracer particles refers to the initial amount of gas outburst at the current moment corresponding to the location of the calculation unit where the virtual tracer particle is located.

[0102] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings. However, the present invention is not limited thereto. Various changes can be made within the scope of knowledge possessed by those skilled in the art without departing from the spirit of the present invention.

Claims

1. A method for dynamic prediction of tunnel gas outburst based on multi-source monitoring, characterized in that, include: Obtain geological survey data files and historical gas monitoring records for the tunnel construction section; the tunnel construction section is formed by combining the surrounding rock at the tunnel face and the excavated roadway. The gas outburst monitoring units were divided in the surrounding rock of the working face, and the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold were set. Extract the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file; Historical gas monitoring records are injected into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, the evolution calculation of the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point is performed to obtain the initial gas emission amount at the current moment for each calculation unit. At the current moment, virtual tracer particles are generated in each computational unit of the geological structural feature point; based on the current moment, the drift velocity and coordinates of all virtual tracer particles, the coordinates of all virtual tracer particles at the next moment are calculated; where the drift velocity of the virtual tracer particles refers to the initial amount of gas outburst at the current moment corresponding to the location of the computational unit where the virtual tracer particle is located. The method further includes: Determine whether the total gas emission in the tunnel construction section has reached the warning threshold; If the total gas emission in the tunnel construction section does not reach the warning threshold, the collection frequency and gas concentration threshold are based on the initial gas emission at the current moment of the excavated roadway and each calculation unit. The evolution calculation of the gas occurrence state of the excavated roadway and the area corresponding to each geological structural feature point is performed to obtain the initial gas outburst amount at the next moment for each calculation unit in each geological structural feature point. The method further includes: Update the initial gas outburst amount for the next moment to the initial gas outburst amount for the current moment for each computational unit in each geological structural feature point, and take the next moment as the current moment, and return to execute the generation of virtual tracer particles in each computational unit of the geological structural feature point in the current moment. Based on the current moment, the drift velocity and coordinates of all virtual tracer particles, calculate the coordinates of all virtual tracer particles at the next moment, until the total gas outburst in the tunnel construction section reaches the warning threshold. If the total gas outburst in the tunnel construction section reaches the warning threshold, the coordinate trajectories of the virtual tracer particles at all times will be output. After extracting the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file, the following is also included: The coordinates of each geological structural feature point in the surrounding rock of the tunnel face are encoded to obtain the feature identification code corresponding to each geological structural feature point; Based on the acquisition frequency and gas concentration threshold, the evolution of gas occurrence status in the areas corresponding to the surrounding rock and geological structural feature points is calculated to obtain the initial gas outburst amount at the current moment for each calculation unit, including: For each calculation unit in each geological structural feature point, a gas outburst prediction sub-model is constructed; Obtain the geological stress direction of the calculation unit and the relative distance between the calculation unit and the tunnel axis; Based on the acquisition frequency, gas concentration threshold, geological stress direction and relative distance, the gas outburst prediction sub-model is trained and inferred to obtain the initial gas outburst amount at the current moment corresponding to the calculation unit. Based on the current drift velocity and coordinates of all virtual tracer particles, calculate the coordinates of all virtual tracer particles at the next moment, including: For each virtual tracer particle, the feature identifier code corresponding to the geological structure feature point is written into the data structure of the virtual tracer particle, and the initial amount of gas outburst at the current moment corresponding to the location of the virtual tracer particle in the computing unit is determined as the drift velocity of the virtual tracer particle, thus obtaining the target tracer particle. Calculate the time step between the next time step and the current time step; Based on the time step, the drift velocity of the target tracer particle, and its coordinates, calculate the coordinates of the virtual tracer particle at the next moment.

2. The method for dynamic prediction of tunnel gas outburst based on multi-source monitoring according to claim 1, characterized in that, Determining whether the total gas emission in a tunnel construction section has reached the warning threshold includes: Based on the coordinate trajectories of virtual tracer particles at all times, calculate the particle aggregation density of each computational unit in each geological structural feature point. If the particle aggregation density of each calculation unit in each geological structural feature point is greater than the set aggregation threshold, then the total gas outburst of the tunnel construction section is determined to have reached the warning critical value. If the particle aggregation density of each calculation unit in each geological structural feature point is not greater than the set aggregation threshold, then it is determined that the total gas outburst of the tunnel construction section has not reached the warning threshold.

3. A dynamic prediction system for tunnel gas emission based on multi-source monitoring, applicable to the dynamic prediction method for tunnel gas emission based on multi-source monitoring as described in claim 1 or 2, characterized in that, include: The data acquisition module is used to acquire geological survey data files and historical gas monitoring records for the tunnel construction section; the tunnel construction section is formed by combining the surrounding rock at the tunnel face and the excavated roadway. The area division module is used to divide the gas outburst monitoring units in the surrounding rock of the working face, and to set the placement points of the gas pressure sensors, the acquisition frequency of the gas pressure sensors, and the gas concentration threshold. The information extraction module is used to extract the coordinate information of all geological structural feature points in the surrounding rock of the tunnel face from the geological survey data file; The gas monitoring module is used to inject historical gas monitoring records into the gas emission monitoring unit. For each geological structural feature point, based on the acquisition frequency and gas concentration threshold, the module performs evolution calculations on the gas occurrence state of the surrounding rock and the area corresponding to the geological structural feature point, and obtains the initial gas emission amount at the current moment for each calculation unit. The coordinate calculation module is used to generate virtual tracer particles in each calculation unit of the geological structure feature point at the current moment, and calculate the coordinates of all virtual tracer particles at the next moment based on the current moment, the drift velocity and coordinates of all virtual tracer particles; where the drift velocity of the virtual tracer particles refers to the initial gas outburst at the current moment corresponding to the location of the calculation unit where the virtual tracer particle is located.

Images