Safety monitoring and intelligent linkage system for coal seam gas extraction
By constructing an intelligent linkage system, the problems of crude area division and insufficient prediction accuracy in gas drainage monitoring have been solved, enabling accurate prediction and early warning of gas reserves and state characteristics, thus improving the safety and efficiency of gas drainage.
Patent Information
- Application Number
- CN202511980688.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-25
- Publication Date
- 2026-02-06
AI Technical Summary
Existing gas extraction monitoring technologies are difficult to achieve precise control under complex geological conditions. The extensive regional division leads to severe or excessive negative pressure attenuation in some areas, and the accuracy of gas state prediction is insufficient, making it impossible to accurately predict the desorption trend of adsorbed gas and the total gas reserves.
A safety monitoring and intelligent linkage system for gas extraction in this coal seam was constructed, including a pipeline division module, a status characteristic analysis module, a regional linkage adjustment module, an early warning threshold setting module, and an optimization report generation module. Through historical database, terrain similarity analysis, and adjacent area verification, combined with real-time free gas parameters, dynamic corrections were made, personalized graded early warning thresholds were set, and structured optimization reports were generated.
It has achieved accurate prediction of total gas reserves and gas state characteristics, reduced prediction errors, avoided misjudgment and lag problems, realized early warning and precise prevention and control, improved response speed and adjustment accuracy, and opened up a closed loop of the whole process of data collection, prediction analysis, linkage adjustment and on-site execution.
Smart Images
Figure CN121473902A_ABST
Abstract
Description
TECHNICAL FIELD
[0001] The present application relates to the technical field of gas extraction safety monitoring and intelligent linkage, in particular to a coal seam gas extraction safety monitoring and intelligent linkage system. BACKGROUND
[0002] Coal seam gas extraction is the core link of safe production in high-gas mines. Its core purpose is to reduce the gas concentration in the coal seam and the working space by extracting the free gas and adsorbed gas stored in the coal seam, prevent gas explosions and personnel poisoning accidents, and at the same time improve the utilization rate of gas resources, realize the dual goals of safe mining and resource recovery.
[0003] The existing gas extraction monitoring technology is widely used in various coal seam extraction scenes in coal mines. At present, the existing technology cannot meet the needs of precise control under complex geological conditions. First, the regional division is extensive, and the effective extraction range and topographic parameter differences are not combined. Only the physical boundary is divided into regions, resulting in serious extraction negative pressure decay in some regions and gas accumulation, while over-extraction in some regions causes energy waste. Second, the gas state prediction accuracy is insufficient, relying on single historical data or real-time parameters, without fusion of topographic similarity analysis and adjacent region verification, the prediction result error is large, and the desorption trend of adsorbed gas and total gas reserves cannot be accurately predicted. In order to reduce this situation, a coal seam gas extraction safety monitoring and intelligent linkage system is proposed. SUMMARY
[0004] The purpose of the present application is to provide a coal seam gas extraction safety monitoring and intelligent linkage system to solve the problems raised in the background art.
[0005] To achieve the above purpose, a coal seam gas extraction safety monitoring and intelligent linkage system is provided, which includes a pipeline division module, a state feature analysis module, a regional linkage adjustment module, a warning threshold setting module and an optimization report generation module. The pipeline division module is used to obtain the pipeline parameters of the coal seam, analyze the effective extraction range according to the pipeline parameters, and divide the pipeline regions of the coal seam according to the effective extraction range. The state feature analysis module is used to establish a reference database corresponding to historical gas extraction, and collect topographic parameters of the current pipeline region. The topographic parameters of the current pipeline region are input into the reference database to predict the total gas reserves prediction value and the predicted state feature of the gas state, and the adjacent pipeline regions are verified. The regional linkage adjustment module is used to obtain the free gas parameters of the current pipeline region, analyze the difference features of the free gas parameters combined with the predicted state features, adjust the predicted state features according to the difference features, and adjust the adjusted predicted state features combined with the adjacent pipeline regions. The pre-warning threshold setting module is configured to predict the remaining amount of adsorbed gas and the desorption trend of the current pipeline area by a gas conversion kinetics equation according to the adjusted predicted state feature combined with the free gas parameter, and set a pre-warning threshold for the pipeline area by combining the predicted state feature, the remaining amount of adsorbed gas and the desorption trend. The optimization report generation module is configured to simulate the feature nodes of the pipeline area triggering the pre-warning threshold, and generate a pipeline parameter optimization report according to the simulated feature nodes combined with the pre-warning threshold.
[0006] As a further improvement of the technical solution, in the pipeline division module, the pipeline parameters of the coal seam are obtained at the gas extraction management end through the connection of the gas extraction management end. According to the pipeline parameters, the effective extraction range of the coal seam is calculated; wherein the effective extraction range is a polygon area with the borehole as the center and the effective extraction radius as the boundary. According to the effective extraction range, the coal seam is divided into pipeline areas; wherein the physical area of a single pipeline area is not more than 1.2 times the total area of the effective extraction range.
[0007] As a further improvement of the technical solution, in the state feature analysis module, the historical gas extraction records are extracted in the gas extraction management end, including the topographic parameters, total gas reserves, state features, extraction efficiency and topographic parameters of the historical pipeline area. The historical gas extraction records are divided according to the corresponding historical pipeline area, and a reference database is established according to the records after the division.
[0008] As a further improvement of the technical solution, in the reference database, the topographic parameters of the current pipeline area are input into the reference database, and the Euclidean distance with the historical pipeline area is calculated by taking the input topographic parameters as the feature vector in the reference database, and then the first five areas with the smallest distance are selected as the similar areas. The topographic parameters of the similar areas and the current pipeline area are calculated for the difference coefficient one by one, and each parameter is given a weight, and the comprehensive topographic difference coefficient is obtained by weighted summation, and then the mean value is calculated by combining the total gas reserves of the similar areas, and the mean value calculation result is modified by combining the comprehensive topographic difference coefficient, so as to obtain the corresponding modified total gas reserve prediction value of the current pipeline area. At the same time, the state features of the most similar similar areas are selected, and the selected state features are modified by combining the topographic difference coefficient to obtain the corresponding predicted state features of the current pipeline area.
[0009] As a further improvement of the technical solution, the state characteristics are divided into 8 types of gas state characteristic types by cluster analysis, including high reserve low conversion complex type, balanced conversion medium type, rapid conversion high risk type, low reserve stable type, high concentration accumulation risk type, low permeability difficult desorption type, low reserve easy extraction type, and dynamic fluctuation complex type.
[0010] As a further improvement of the technical solution, according to the total gas reserve prediction value of the current pipeline area, the relevant adjacent pipeline area of the current coal seam is obtained, the terrain difference is calculated according to the corresponding terrain parameters of the adjacent pipeline area and the corresponding terrain parameters of the current pipeline area, the deviation threshold is set according to the terrain difference calculation result, then the deviation rate calculation is carried out on the total gas reserve prediction value of the current pipeline area and the adjacent pipeline area, and the deviation rate calculation result is compared with the deviation threshold; When the deviation rate is greater than the deviation threshold, the similar area is matched again and the terrain difference fusion weight is adjusted; When the deviation rate is less than the deviation threshold, the total gas reserve prediction value of the current management area is determined.
[0011] As a further improvement of the technical solution, in the regional linkage adjustment module, the free gas parameters of the current pipeline area are obtained, then the difference characteristics between the free gas parameters and the predicted state characteristics are obtained by combining the free gas parameters with the predicted state characteristics for difference characteristic analysis; An adjustment threshold is set based on the parameter type, then the difference characteristics are compared with the adjustment threshold; When the difference characteristics exceed the adjustment threshold, the predicted state adjustment is triggered, the gradient descent algorithm is used, the deviation value of the difference characteristic type is used as the optimization target, and the predicted state characteristics are iteratively corrected; When the difference characteristics do not exceed the adjustment threshold, the monitoring continues; The management area triggering the predicted state is matched with the adjacent pipeline area for predicted state characteristic matching, when the predicted state characteristics of the two areas are consistent, the predicted state characteristics are corrected synchronously; On the contrary, when the predicted state characteristics of the two areas are inconsistent, the correction is not synchronized.
[0012] As a further improvement of the technical solution, in the early warning threshold setting module, after the real-time desorption rate is calculated by the gas conversion kinetics equation based on the predicted state characteristics after adjustment and free gas, the adsorbed gas remaining amount is obtained by combining the extraction time; The initial adsorbed gas reserve is determined based on the predicted state characteristics after adjustment; The conversion kinetics equation coefficient is modified by substituting the real-time free gas pressure and temperature parameters; The desorption rate change curve of the next 24 hours is predicted, and the adsorbed gas remaining amount decay trend is output. The remaining amount of adsorbed gas and desorption tendency are combined to predict the state characteristics to set a dedicated early warning threshold for the pipeline area; The early warning threshold is divided into high risk characteristics, medium risk characteristics and low risk characteristics; Each level of early warning threshold is associated with a corresponding desorption rate threshold.
[0013] As a further improvement of the technical solution, in the optimization report generation module, the pipeline area triggering the early warning threshold is simulated by feature nodes, including gas concentration nodes, pressure nodes, desorption rate nodes, terrain parameter nodes and pipeline operation parameter nodes. The simulation of each feature node adopts digital twin technology to restore the gas extraction process of the pipeline area in proportion; According to the simulated feature nodes and the early warning threshold, a pipeline parameter optimization report is generated; the optimization report includes the final determination result of the gas state characteristics of the current pipeline area, the key data of the remaining amount of adsorbed gas and the desorption tendency, the specific value and triggering condition of the early warning threshold, the pipeline parameter adjustment suggestion, and the feasibility analysis of regional segmentation or new pipeline control area.
[0014] Compared with the prior art, the beneficial effects of the present application are: 1. In the coal seam gas extraction safety monitoring and intelligent linkage system, by constructing a historical reference database, fusing terrain similarity matching, comprehensive difference coefficient correction and adjacent area verification, the total gas reserves and gas state characteristics are accurately predicted; at the same time, combined with real-time free gas parameters, the prediction result is dynamically corrected by gradient descent algorithm, which effectively reduces the prediction error caused by single data source, and provides a reliable basis for subsequent early warning and optimization.
[0015] 2. In the coal seam gas extraction safety monitoring and intelligent linkage system, according to the adjusted gas state characteristics and risk level, a specialized hierarchical early warning threshold is set, and core indicators such as desorption rate are associated, which avoids the misjudgment and lag problem caused by traditional one-size-fits-all threshold, realizes advanced warning and accurate prevention and control, and greatly reduces the risk of gas overrun and accumulation.
[0016] 3. In the coal seam gas extraction safety monitoring and intelligent linkage system, the real-time parameters and predicted states are dynamically matched through the regional linkage adjustment module, and the coordinated correction is realized through the similarity degree judgment of adjacent area states, which avoids the imbalance of global control caused by local adjustment; the early warning and optimization link accurately predicts the remaining amount of adsorbed gas and desorption tendency by gas conversion kinetics equation, and completes feature node simulation by digital twin technology, generates structured and implementable optimization report, and breaks through the whole process closed loop of data collection-prediction analysis-linkage adjustment-site execution, reduces manual intervention, and improves response speed and adjustment accuracy. BRIEF DESCRIPTION OF DRAWINGS
[0017] Figure 1 This is a flowchart illustrating the safety monitoring and intelligent linkage system for coal seam gas extraction of this invention. Figure 2 This is a flowchart illustrating the pipeline division module of the present invention; Figure 3 This is a flowchart illustrating the state feature analysis module of the present invention; Figure 4 This is a flowchart illustrating the regional linkage adjustment module of the present invention. Figure 5 This is a flowchart illustrating the early warning threshold setting module of the present invention; Figure 6 This is a flowchart illustrating the optimized report generation module of the present invention. Detailed Implementation
[0018] 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.
[0019] Please see Figure 1 As shown, the purpose of this embodiment is to provide a safety monitoring and intelligent linkage system for coal seam gas extraction, including a pipeline division module, a status feature analysis module, a regional linkage adjustment module, an early warning threshold setting module, and an optimization report generation module; The pipeline division module is used to obtain the pipeline parameters of this coal seam, analyze the effective extraction range based on the pipeline parameters, and divide the coal seam into pipeline areas according to the effective extraction range. The coal seam is divided into pipeline areas that can be precisely controlled according to the extraction capacity boundary, which solves the problem of coarse extraction area division and control failure caused by negative pressure attenuation in traditional extraction areas. In the pipeline segmentation module, the pipeline parameters of this coal seam are obtained by connecting to the gas extraction management terminal. The effective extraction range for gas extraction in this coal seam is calculated based on pipeline parameters; the effective extraction range is a polygonal area centered on the borehole and bounded by the effective extraction radius. The coal seam is divided into pipeline zones according to the effective extraction range; the physical area of a single pipeline zone shall not exceed 1.2 times the total area of the effective extraction range.
[0020] Data acquisition is performed through the gas extraction management terminal to collect the parameters of the gas extraction pipeline in this coal seam (pipeline basic parameters, borehole configuration parameters, coal seam related parameters, such as coal seam permeability coefficient and coal seam porosity), and then the effective extraction radius is calculated based on the derivation formula of radial flow theory. Then, taking a single borehole as the center and the effective extraction radius as the radial boundary, and combining the effective extraction radius boundaries of adjacent boreholes, a polygonal region is drawn (to avoid overlapping or omission of extraction in adjacent regions and to ensure coverage integrity), and the boundary coordinates of the polygonal region are marked. The effective extraction polygon areas of all boreholes within this coal seam are integrated, and the areas are split or merged according to 1.2 times the area threshold to ensure that there is no significant attenuation of the extraction negative pressure in each area (attenuation ≤5%). Finally, a unique identifier is assigned to each divided pipeline area to associate the corresponding borehole group and pipeline branch, as shown in the following formula: ; in, To achieve an effective extraction radius, The coal seam permeability coefficient. To draw in negative pressure, For sampling time, For gas dynamic viscosity, The porosity of the coal seam. Gas density.
[0021] The state feature analysis module is used to establish a reference database corresponding to historical gas extraction, collect the terrain parameters of the current pipeline area, input the terrain parameters of the current pipeline area into the reference database, predict the total gas reserves and the predicted state characteristics of the gas in the pipeline area, and verify them with adjacent pipeline areas. By matching historical data with terrain, it can achieve accurate prediction of the total gas reserves and gas state characteristics of the pipeline area, and reduce prediction errors by verifying with adjacent areas. In the state characteristic analysis module, historical gas extraction records are extracted from the gas extraction management terminal, including the terrain parameters, total gas reserves, gas state characteristics, extraction efficiency, and terrain parameters of the historical pipeline area. Extraction is performed through the historical data query module of the gas extraction management terminal, selecting historical gas extraction records of this coal seam and mines with similar geological conditions over the past 5 years; Historical gas extraction records are divided according to the corresponding historical pipeline areas, and a reference database is established based on the divided records.
[0022] Using the unique identifier of the historical pipeline area as the association key, the filtered historical gas extraction records are bound to the corresponding historical pipeline areas. Each historical pipeline area is treated as an independent data unit, associated with all its corresponding extraction parameters (topography, gas, extraction efficiency, etc.). At the same time, if there are records of multiple extraction cycles for the same historical pipeline area, the data sub-units are subdivided according to the extraction cycle.
[0023] The reference database extracts the terrain parameters of the current pipeline area and inputs them into the reference database. In the reference database, the Euclidean distance with the historical pipeline area is calculated using the input terrain parameters as feature vectors. Then, the five areas with the smallest distance are selected as similar areas. Six core topographic parameters of the current pipeline area are extracted: coal seam dip angle, burial depth, coal thickness, permeability coefficient, porosity, and geological structure type. The current parameters are standardized using a normalization method (to eliminate dimensional differences) to construct a feature vector of topographic parameters for the current area. Then, the standardized terrain parameter feature vectors of all historical pipeline areas in the reference database are extracted, the Euclidean distance between the current feature vector and the feature vector of each historical area is calculated, and all Euclidean distances are sorted in ascending order. The top 5 historical areas with the smallest distances are selected as similar areas of the current area. ; in, It is the Euclidean distance between the current region feature vector and the i-th historical region feature vector (the smaller the distance, the higher the similarity). This is the standardized value of the k-th terrain parameter in the current region. Let k be the standardized value of the topographic parameter of the i-th historical region. The parameter dimensions are 1-6, corresponding to 6 core terrain parameters.
[0024] The difference coefficients of the terrain parameters between the similar area and the current pipeline area are calculated item by item, and each parameter is assigned a weight. The comprehensive terrain difference coefficient is obtained by weighted summation. Then, the average value is calculated by combining the total gas reserves of the similar area. The average value is then corrected by combining the comprehensive terrain difference coefficient, so as to obtain the corrected total gas reserves prediction value of the current pipeline area. The topographic parameter difference coefficient between the current region and each similar region is calculated item by item (for each of the 6 core parameters), and weights are assigned to each topographic parameter (based on the best weight allocation based on engineering experience, prioritizing key parameters that affect gas reserves), such as permeability coefficient (0.3), coal thickness (0.2), burial depth (0.2), coal seam dip angle (0.15), porosity (0.1), and geological structure type (0.05). Then, the comprehensive topographic difference coefficient between the current region and the group of similar regions is calculated by weighted summation. The historical total gas reserves of five similar areas are extracted and their average values are calculated. A comprehensive topographic difference coefficient is then used to correct the average value (the greater the difference, the larger the correction, ensuring the predicted value closely matches the actual topography of the current area). The corrected predicted total gas reserves for the current area are then output, as shown in the following formula:
[0025] in, This represents the coefficient of difference between the current region and a group of similar regions on the k-th topographic parameter. The k-th topographic parameter is the mean of the measured values of the five similar regions. This represents the measured value of the k-th topographic parameter in the current region. ; in, This is the comprehensive topographic difference coefficient between the current region and similar regions (the value ranges from [0,1], and the closer the value is to 0, the smaller the topographic difference). The weight of the k-th terrain parameter is... is the coefficient of variation for the k-th topographic parameter; ; in, This represents the predicted total gas reserves in the current pipeline area. The average of the historical total gas reserves in five similar regions. This is a correction factor; Simultaneously, the state features of the most similar regions are selected, and the selected state features are combined with the terrain difference coefficient for correction to obtain the predicted state features corresponding to the current pipeline area.
[0026] The historical gas state characteristics of the most similar region are selected, and the characteristics are corrected by combining the comprehensive topographic difference coefficient. If the difference coefficient is ≤0.1 (extremely similar topography), the characteristics of the most similar region are directly used. If 0.1 < difference coefficient ≤ 0.3 (moderate terrain similarity), fine-tune the feature dimension (e.g., the most similar area is low-permeability and difficult-to-desorb type, and the current air permeability coefficient is slightly high, correct it to balanced conversion to moderate type). If the difference coefficient is >0.3 (significant terrain differences), re-validate the matching results of similar regions; The final output is the predicted state characteristics of the current pipeline area, which includes the following 8 categories; Based on the gas state characteristics, eight types of gas state characteristics were identified through cluster analysis, including: high reserves and low conversion complex type, balanced conversion medium type, rapid conversion high risk type, low reserves and stable type, high concentration accumulation risk type, low permeability and difficult desorption type, low reserves and easy extraction type, and dynamic fluctuation complex type.
[0027] Based on the predicted total gas reserves of the current pipeline area, adjacent pipeline areas related to the current pipeline area are obtained in this coal seam. The terrain difference between the adjacent pipeline areas and the current pipeline area is calculated based on their terrain parameters. A deviation threshold is set based on the terrain difference calculation results. Then, the deviation rate between the predicted total gas reserves of the current pipeline area and the adjacent pipeline areas is calculated, and the deviation rate calculation result is compared with the deviation threshold. The steps are as follows: Based on the spatial boundary of the current pipeline area, pipeline areas with adjacent boundaries and connected extraction spaces are selected. Six core terrain parameters (consistent with the previous text) are extracted from the current area and each adjacent area, as well as the total gas reserves predicted for each adjacent area. The topographic parameters of the current region and each of its adjacent regions are standardized (normalization method). A weighted summation method based on the difference coefficients is used to calculate the comprehensive topographic difference coefficient between the current region and each of its adjacent regions. If multiple adjacent regions exist, the average of the comprehensive topographic difference coefficients between each adjacent region and the current region is taken as the final quantitative result of the topographic difference. The formula is as follows: ; in, This is the comprehensive terrain difference coefficient between the current region and a single adjacent region (values [0,1], the smaller the value, the more similar the terrain). This represents the measured value of the k-th topographic parameter in the current region. This represents the measured value of the k-th topographic parameter in the adjacent region; Then, establish the correspondence between the comprehensive terrain difference coefficient and the deviation threshold (engineering best fit rule). The smaller the terrain difference, the stricter the deviation threshold setting; the larger the terrain difference, the more relaxed the threshold should be. Substitute the mean of the comprehensive terrain difference coefficient into the value to calculate the deviation threshold for this verification. Next, the total gas reserves initially predicted in the current area and the total gas reserves confirmed in the adjacent areas are extracted to calculate the absolute deviation between the two, which is then converted into a deviation rate. The calculated reserve deviation rate is then compared with the set deviation threshold. When the deviation rate is greater than the deviation threshold, similar areas are rematched and the terrain difference fusion weight is adjusted (such as increasing the weight of key parameters such as permeability coefficient and burial depth), and the total gas reserve prediction value is recalculated. When the deviation rate is less than the deviation threshold, the predicted total gas reserves for the current management area are determined using the following formula: ; in, The threshold for reserve deviation, This represents the average comprehensive topographic difference coefficient between the current region and all adjacent regions. The base threshold (the maximum allowable rate of deviation in reserves when the terrain is perfectly similar). This is the threshold adjustment coefficient (the threshold increases by 2% for every 0.1 increase in terrain difference). ; in, The deviation rate of total gas reserves between the current region and adjacent regions. This represents the preliminary estimated total gas reserves for the current region. This represents the total confirmed gas reserves in the adjacent area. The regional linkage adjustment module is used to obtain the free gas parameters of the current pipeline area, analyze the difference characteristics of the free gas parameters in combination with the predicted state characteristics, adjust the predicted state characteristics according to the difference characteristics, and link the adjusted predicted state characteristics with adjacent pipeline areas for linkage adjustment; based on the real-time free gas parameters, the predicted state characteristics are verified and adjusted, and at the same time, through the linkage of adjacent areas, local correction and global coordination are achieved to avoid control failure caused by parameter deviation in a single area; In the regional linkage adjustment module, the free gas parameters of the current pipeline area are obtained, and then the free gas parameters are combined with the predicted state characteristics to perform difference feature analysis and obtain the difference features between the free gas parameters and the predicted state characteristics. The core free gas parameters of the current pipeline area are collected by intrinsically safe sensors, including gas concentration, gas pressure, gas flow rate, and concentration fluctuation range. From the preliminary predicted state characteristics of the current area, extract the corresponding free gas parameter benchmark threshold range (such as concentration benchmark value, pressure benchmark value, flow benchmark value), and calculate the relative deviation of the real-time parameters and the mean of the benchmark thresholds for each item (to avoid misjudgment caused by differences in parameter magnitudes) to obtain a set of difference features. The adjustment threshold is set based on the parameter type, and then the difference features are combined with the adjustment threshold for over-comparison. First, prioritize the parameters, with core safety parameters (concentration and concentration fluctuation range) having the highest priority, followed by auxiliary parameters (pressure and flow rate). Then, set thresholds based on the risk level (high / medium / low) of the predicted state characteristics. Higher risk characteristics correspond to stricter thresholds (e.g., a concentration deviation threshold of 0.08 for high risk, 0.12 for medium risk, and 0.15 for low risk), thus obtaining a set of adjustment thresholds for each differential characteristic. Each difference feature deviation value is compared with the corresponding adjustment threshold one by one. If any difference feature deviation value is greater than the corresponding adjustment threshold, the prediction state feature adjustment process is triggered. When the difference features exceed the adjustment threshold, the prediction state adjustment is triggered. The gradient descent algorithm is used to iteratively correct the prediction state features with the deviation value of the difference feature type as the optimization target. Construct an objective function (loss function) with the goal of minimizing the weighted sum of the bias values of each differential feature. Employ a gradient descent algorithm to update the core parameters of the predicted state features (such as concentration level and conversion rate level) with a preset learning rate. Iterate until the loss function value is ≤ the convergence threshold (e.g., 0.02) or the maximum number of iterations is reached (e.g., 10 times). Output the adjusted predicted state features, as shown in the following formula:
[0028] in, The loss function (optimization objective; the smaller the value, the smaller the bias). The core parameter vector for predicting state characteristics (parameters to be updated, such as the quantized values corresponding to concentration level and conversion rate level). The weights of each differential feature, This represents the relative deviation value for each difference characteristic; ; in, These are the predicted state feature parameters updated after the (t+1)th iteration. Let be the predicted state feature parameters at the t-th iteration. The learning rate (ranging from 0.05 to 0.1). For the loss function in The gradient at a given point (guides the direction of parameter updates, reducing the loss function); If the difference characteristics do not exceed the adjustment threshold, monitoring will continue. Specifically, the predicted state features of the management area that triggers the predicted state are matched with those of the adjacent pipeline areas. When the predicted state features of the two areas are consistent, the predicted state features are corrected synchronously. Conversely, if the predicted state characteristics of the two regions are inconsistent, they will be corrected asynchronously. Extract the predicted state features of the current region and adjacent regions that triggered the adjustment, and calculate the state similarity coefficient between them. If the state similarity coefficient is ≥0.8 (features are consistent), the correction result of the current region is synchronized to the adjacent regions, and the same correction is performed on the predicted state features of the adjacent regions. If the state similarity coefficient is <0.8 (features are inconsistent), only the predicted state features of the current region are corrected, while the adjacent regions retain their original features and are continuously monitored. The formula is as follows:
[0029] in, S represents the state similarity coefficient between the current region and its neighboring regions (value range [0,1], the closer S is to 1, the more consistent the features). The k-th parameter is the predicted state feature of the current region. The k-th parameter is used to predict the state characteristics of neighboring regions. The number of parameters for predicting state features.
[0030] The early warning threshold setting module is used to predict the remaining amount and desorption trend of adsorbed gas in the current pipeline area based on the adjusted predicted state characteristics and free gas parameters, through the gas conversion kinetic equation. It then sets an early warning threshold for the pipeline area by combining the remaining amount and desorption trend of adsorbed gas with the predicted state characteristics. Based on the dynamically corrected state characteristics and gas conversion law, it predicts the state of adsorbed gas (remaining amount, desorption trend) and sets graded early warning thresholds to achieve "advanced early warning and precise prevention and control". In the early warning threshold setting module, based on the adjusted predicted state characteristics and free gas, the real-time desorption rate is calculated through the gas conversion kinetic equation, and the remaining amount of adsorbed gas is obtained by combining the extraction time. The initial adsorbed gas reserves were determined based on the adjusted predicted state characteristics. Determine the proportion of adsorbed gas in the initial total gas reserves, and calculate the initial adsorbed gas reserves by combining the predicted total gas reserves, real-time free gas reserves, and cumulative extraction volume (initial adsorption volume = total reserves - initial free volume - cumulative extraction volume + adsorption → free conversion volume). Substitute the real-time free gas pressure and temperature parameters to correct the coefficients of the conversion kinetic equation; The pressure correction factor is calculated based on the ratio of real-time gas pressure to standard pressure. Based on the absolute temperature and standard temperature of the coal body, and combined with the desorption activation energy, the temperature correction coefficient is calculated; Multiply the original desorption rate constant by the pressure correction factor and the temperature correction factor to obtain the real-time desorption rate constant adapted to the current operating conditions; The desorption intensity and remaining reserves of adsorbed gas at the current moment are quantified, and then substituted into the modified gas conversion kinetic equation to calculate the real-time desorption rate. Based on the real-time desorption rate, the extraction time t is combined to perform integral calculation to obtain the remaining amount of adsorbed gas at the current moment (remaining amount = initial adsorption amount - cumulative desorption amount, where the cumulative desorption amount is the integral of the desorption rate over time). Predict the desorption rate change curve for the next 24 hours and output the decay trend of the remaining adsorbed gas. The next 24 hours are divided into several time steps (e.g., 1 hour / step, 24 steps in total). The desorption rate of each time step is calculated iteratively based on the real-time desorption rate constant and the current amount of adsorbed gas remaining. Then, the time-desorption rate change curve is plotted, and the decay trend of the amount of adsorbed gas remaining (e.g., linear decay, exponential decay) is output by curve fitting. By combining the remaining amount of adsorbed gas and the desorption trend with the predicted state characteristics, a specific early warning threshold is set for the pipeline area. The warning thresholds are divided into high-risk, medium-risk, and low-risk characteristics. Each warning threshold is associated with a corresponding desorption rate threshold.
[0031] Based on the adjusted predicted state characteristics, the risk level (high / medium / low) is determined. Threshold quantification involves setting corresponding gas concentration warning thresholds for different risk levels, such as high risk ≥70%, medium risk ≥50%, and low risk ≥30%. Associate desorption rate thresholds, and bind corresponding desorption rate thresholds to each concentration threshold level; Define the combined triggering conditions of the concentration threshold and desorption rate threshold in the current pipeline area.
[0032] The optimized report generation module simulates the feature nodes of the pipeline area that triggers the warning threshold, and generates a pipeline parameter optimization report based on the simulated feature nodes and the warning threshold. By simulating the core feature nodes of the area triggering the warning using digital twin technology, a practical pipeline parameter optimization report is generated, completing the prediction-warning-optimization closed loop. In the optimized report generation module, the pipeline area that triggers the warning threshold is simulated with characteristic nodes. The characteristic nodes include gas concentration nodes, pressure nodes, desorption rate nodes, terrain parameter nodes and pipeline operation parameter nodes. The simulation of each characteristic node adopts digital twin technology to restore the gas extraction process of the pipeline area on a proportional scale. Extract the boundary coordinates of the pipeline area that triggers the early warning threshold, synchronously associate them with the basic information of its adjacent areas, and summarize the full data of 5 types of feature nodes, including real-time and historical data of gas concentration / pressure / desorption rate (last 24 hours, sampling interval of 5 minutes), measured values of terrain parameters (inclination, burial depth, etc.), pipeline operation parameters (negative pressure, flow rate, pipe diameter, borehole layout, etc., early warning threshold data), and coal physical parameters (porosity, permeability coefficient, etc.). Through the fusion of multi-dimensional models, including geological and topographical models (based on BIM+GIS technology, 1:1 restoration of coal seam occurrence morphology and geological structure), equipment and pipeline models (3D modeling to restore the spatial layout and connection relationship of extraction pumping stations, pipelines and boreholes), and gas migration models (embedded with CFD numerical simulation modules to depict the gas adsorption-desorption-flow law). Substitute the full data and calibrate the model parameters by backtracking through historical data (gas concentration change curve in the last 24 hours); Simulation is performed in the logical order of terrain parameter node → pipeline operation parameter node → gas pressure node → desorption rate node → gas concentration node (the preceding node provides boundary conditions for the following node). Topographic parameter nodes simulate the constraining effect of coal seam dip angle and burial depth on gas migration (such as gas gravity accumulation caused by dip angle). Pipeline operation parameter nodes simulate the impact of negative pressure and flow rate changes on the gas flow velocity within the pipeline. The desorption rate node and gas concentration node, based on the calibrated CFD model, simulate the dynamic changes of gas throughout the entire process from coal seam desorption → borehole extraction → pipeline transportation under early warning conditions, and output the spatiotemporal distribution cloud map of the nodes, as shown in the following formula:
[0033] in, The porosity of the coal seam. For gas density, For time, Let the gas seepage velocity vector be... For the gas desorption rate source term, The gas extraction rate is summarized to characterize the gas mass conservation within the simulation area, supporting the dynamic simulation of concentration and pressure nodes.
[0034] Based on the simulated feature nodes and early warning thresholds, a pipeline parameter optimization report is generated. The optimization report includes the final determination results of the gas status characteristics of the current pipeline area, key data on the remaining amount of adsorbed gas and desorption trend, specific values and triggering conditions of the early warning threshold, pipeline parameter adjustment suggestions, and feasibility analysis of regional division or the construction of new control areas.
[0035] The report integrates the core content, including the final determination of gas status characteristics (corrected characteristic type + risk level), key data on the remaining amount of adsorbed gas and desorption trend (current remaining amount, decay curve for the next 12 hours, extreme value time points), details of warning thresholds (triggered concentration / desorption rate thresholds, trigger time, and description of associated risks), pipeline parameter adjustment suggestions (quantified optimization values, such as adjusting negative pressure from -10kPa to -18kPa and reducing borehole spacing from 8m to 6m), and feasibility analysis of regional segmentation and the establishment of new control areas (based on the simulated effective extraction range, determining whether it is necessary to split the region or add new control boundaries). Embedding simulated spatiotemporal distribution cloud maps of feature nodes, comparison curves before and after parameter adjustment, and regional layout diagrams enhances report readability.
[0036] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely preferred examples and are not intended to limit the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.
Claims
1. The safety monitoring and intelligent linkage system for coal seam gas extraction is characterized by: It includes a pipeline division module, a status characteristic analysis module, a regional linkage adjustment module, an early warning threshold setting module, and an optimization report generation module; The pipeline division module is used to obtain the pipeline parameters of this coal seam, analyze the effective extraction range based on the pipeline parameters, and divide the coal seam into pipeline areas according to the effective extraction range. The state feature analysis module is used to establish a reference database corresponding to historical gas extraction, collect the terrain parameters of the current pipeline area, input the terrain parameters of the current pipeline area into the reference database, predict and output the total gas reserves and the predicted state features of the gas state of the pipeline area, and verify them in conjunction with adjacent pipeline areas. The regional linkage adjustment module is used to obtain the free gas parameters of the current pipeline area, combine the free gas parameters with the predicted state characteristics to perform difference feature analysis, adjust the predicted state characteristics according to the difference features, and combine the adjusted predicted state characteristics with adjacent pipeline areas for linkage adjustment. The warning threshold setting module is used to predict the remaining amount of adsorbed gas and the desorption trend in the current pipeline area based on the adjusted predicted state characteristics and free gas parameters, through the gas conversion kinetic equation, and to set a warning threshold for the pipeline area by combining the remaining amount of adsorbed gas and the desorption trend with the predicted state characteristics. The optimization report generation module is used to simulate feature nodes in the pipeline area that triggers the warning threshold, and generate a pipeline parameter optimization report based on the simulated feature nodes and the warning threshold.
2. The coal seam gas extraction safety monitoring and intelligent linkage system according to claim 1, characterized in that: In the pipeline division module, the pipeline parameters of this coal seam are obtained by connecting to the gas extraction management terminal. The effective extraction range for gas extraction in this coal seam is calculated based on pipeline parameters; the effective extraction range is a polygonal area centered on the borehole and bounded by the effective extraction radius. The coal seam is divided into pipeline zones according to the effective extraction range; the physical area of a single pipeline zone shall not exceed 1.2 times the total area of the effective extraction range.
3. The safety monitoring and intelligent linkage system for coal seam gas extraction according to claim 2, characterized in that: In the state feature analysis module, historical gas extraction records are extracted from the gas extraction management terminal, including topographic parameters, total gas reserves, gas state characteristics, extraction efficiency, and topographic parameters of the historical pipeline area. Historical gas extraction records are divided according to the corresponding historical pipeline areas, and a reference database is established based on the divided records.
4. The coal seam gas extraction safety monitoring and intelligent linkage system according to claim 3, characterized in that: The reference database extracts the terrain parameters of the current pipeline area and inputs them into the reference database. In the reference database, the Euclidean distance with the historical pipeline area is calculated by using the input terrain parameters as feature vectors. Then, the five areas with the smallest distance are selected as similar areas. The difference coefficients of the terrain parameters between the similar area and the current pipeline area are calculated item by item, and each parameter is assigned a weight. The comprehensive terrain difference coefficient is obtained by weighted summation. Then, the average value is calculated by combining the total gas reserves of the similar area. The average value is then corrected by combining the comprehensive terrain difference coefficient, so as to obtain the corrected total gas reserves prediction value of the current pipeline area. Simultaneously, the state features of the most similar regions are selected, and the selected state features are combined with the terrain difference coefficient for correction to obtain the predicted state features corresponding to the current pipeline area.
5. The safety monitoring and intelligent linkage system for coal seam gas extraction according to claim 4, characterized in that: The aforementioned state characteristics were classified into eight types through cluster analysis, including high-reserve-low-conversion complex type, balanced-conversion medium type, rapid-conversion high-risk type, low-reserve-stable type, high-concentration-accumulation-risk type, low-permeability and difficult-to-desorb type, low-reserve-easy-to-extract type, and dynamic-fluctuation complex type.
6. The coal seam gas extraction safety monitoring and intelligent linkage system according to claim 4, characterized in that: Based on the predicted total gas reserves of the current pipeline area, the adjacent pipeline areas related to the current pipeline area are obtained in this coal seam. The terrain difference is calculated based on the terrain parameters corresponding to the adjacent pipeline areas and the terrain parameters corresponding to the current pipeline area. A deviation threshold is set based on the terrain difference calculation results. Then, the deviation rate is calculated between the predicted total gas reserves of the current pipeline area and the adjacent pipeline areas, and the deviation rate calculation results are compared with the deviation threshold. If the deviation rate is greater than the deviation threshold, similar regions are rematched and the terrain difference fusion weights are adjusted. If the deviation rate is less than the deviation threshold, the predicted total gas reserves for the current management area are determined.
7. The coal seam gas extraction safety monitoring and intelligent linkage system according to claim 1, characterized in that: In the regional linkage adjustment module, the free gas parameters of the current pipeline area are obtained, and then the free gas parameters are combined with the predicted state characteristics to perform difference feature analysis to obtain the difference features between the free gas parameters and the predicted state characteristics. The adjustment threshold is set based on the parameter type, and then the difference features are combined with the adjustment threshold for over-comparison. When the difference features exceed the adjustment threshold, the prediction state adjustment is triggered. The gradient descent algorithm is used to iteratively correct the prediction state features with the deviation value of the difference feature type as the optimization target. If the difference characteristics do not exceed the adjustment threshold, monitoring will continue. Specifically, the predicted state features of the management area that triggers the predicted state are matched with those of the adjacent pipeline areas. When the predicted state features of the two areas are consistent, the predicted state features are corrected synchronously. Conversely, if the predicted state characteristics of the two regions are inconsistent, they will be corrected asynchronously.
8. The safety monitoring and intelligent linkage system for coal seam gas extraction according to claim 1, characterized in that: In the aforementioned warning threshold setting module, based on the adjusted predicted state characteristics and free gas, the real-time desorption rate is calculated using the gas conversion kinetic equation, and the remaining amount of adsorbed gas is obtained by combining the extraction time. The initial adsorbed gas reserves were determined based on the adjusted predicted state characteristics. Substitute the real-time free gas pressure and temperature parameters to correct the coefficients of the conversion kinetic equation; Predict the desorption rate change curve for the next 24 hours and output the decay trend of the remaining adsorbed gas. By combining the remaining amount of adsorbed gas and the desorption trend with the predicted state characteristics, a specific early warning threshold is set for the pipeline area. The warning thresholds are divided into high-risk, medium-risk, and low-risk characteristics. Each warning threshold is associated with a corresponding desorption rate threshold.
9. The safety monitoring and intelligent linkage system for coal seam gas extraction according to claim 1, characterized in that: In the optimization report generation module, the pipeline area that triggers the warning threshold is simulated with feature nodes. The feature nodes include gas concentration nodes, pressure nodes, desorption rate nodes, terrain parameter nodes and pipeline operation parameter nodes. The simulation of each feature node adopts digital twin technology to restore the gas extraction process of the pipeline area on a proportional basis. Based on the simulated feature nodes and early warning thresholds, a pipeline parameter optimization report is generated. The optimization report includes the final determination results of the gas status characteristics of the current pipeline area, key data on the remaining amount of adsorbed gas and desorption trend, specific values and triggering conditions of the early warning threshold, pipeline parameter adjustment suggestions, and feasibility analysis of regional division or the construction of new control areas.