Coal spontaneous combustion index gas traceability and analysis method based on isotope fingerprints

Abstract

The invention discloses a coal spontaneous combustion index gas traceability and analysis method based on isotope fingerprints, which comprises the following steps: drilling and sampling each coal seam of a coal mine to prepare desorbed gas, and measuring the abundance ratio of gas isotope to form an isotope fingerprint characteristic spectrum; drawing stable carbon-hydrogen-oxygen isotope fingerprint characteristic spectrums of the coal seam coal sample oxidation process at different temperatures through a temperature programming-gas chromatography-stable isotope mass spectrometer combined system; the method comprises the following steps: collecting underground non-spontaneous combustion area gas, measuring a normal background isotope value, and establishing a background database and stable isotope fingerprint characteristics of various gases to jointly construct a coal mine carbon-hydrogen-oxygen isotope fingerprint database; and applying a physical constraint optimization sub-source quantification mathematical model, and combining a Bayesian depth inversion method to realize rapid matching and high-precision traceability of stable isotope measured data and a fingerprint database. According to the method, tracing analysis of coal spontaneous combustion index gas abnormity in the shallow-buried short-distance multi-coal-seam mining process can be achieved, the coal spontaneous combustion disaster site is positioned, and safe production of a coal mine is guaranteed.

Description

Technical Field

[0001] This invention belongs to the field of coal mine safety technology, specifically relating to a method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting. Background Technology

[0002] In shallow, close-proximity mining of multiple coal seams, the surface, the current goaf, adjacent goaf, and overlying goaf are prone to collapse and interconnection, forming a continuous gas leakage channel between the surface, the complex goaf, and the working face. This results in numerous air leakage sources and complex leakage lines in the goaf under these mining conditions. Gases generated by spontaneous combustion of coal in the goaf and toxic and harmful gases accumulated in old goafs will flow between multiple goaf layers, and may even flow across to the working face, causing abnormally high gas concentrations in the goaf, upper corners, and other areas. This makes it difficult to determine the location and extent of the fire, and increases the difficulty of precise fire control.

[0003] Currently, most coal mines monitor the concentration of various gas components in the goaf and roadways through bundled tube monitoring or manual inspection, and combine this with the gas production characteristics of spontaneous combustion of coal to judge the safety status inside the goaf. This has played a role in preventing coal spontaneous combustion disasters to a certain extent. However, under the conditions of shallow-buried, close-range, multi-coal-seam mining, the concentration of coal spontaneous combustion indicator gases (such as CO and CO2) often exceeds the standard in the upper corner of the working face, while the temperature of the coal body in the goaf is not high. The source of the coal spontaneous combustion indicator gases is unclear, making it difficult to accurately locate the coal seam and specific area of ​​the fire source. There is a lack of systematic gas source identification and contribution measurement methods, which seriously affects the prediction, forecasting and prevention of coal spontaneous combustion disasters and poses a great threat to the safe mining of shallow-buried, close-range, multi-coal-seams. Summary of the Invention

[0004] The purpose of this invention is to provide a method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting. This method can trace and analyze the anomalies of coal spontaneous combustion indicator gases during shallow-buried, close-range multi-coal-seam mining, accurately locate the location of coal spontaneous combustion disasters, and ensure safe production in coal mines.

[0005] To achieve the above objectives, this invention provides a method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting, comprising the following steps:

[0006] S1. Collect coal samples from each coal seam in the coal mine to prepare desorption gas and coal oxidation index gas. The coal samples collected from each coal seam in the coal mine include those from exposed coal seams and those from unexposed coal seams.

[0007] S2. For unexposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics of the original occurrence index gas of the coal seam are analyzed by the desorbed gas prepared. For exposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics are analyzed by in-situ gas sampling, and the data is transmitted to the isotope fingerprint database.

[0008] S3. Determine the abundance of stable isotopes using a programmed temperature-gas chromatography-isotope ratio mass spectrometry system, and plot the temperature-fingerprint characteristic curve of the stable isotope of the index gas.

[0009] S4. By integrating the isotopic fingerprint features of the original gases in each coal seam, coal oxidation index gases, and normal gases in coal mines, a coal mine spontaneous combustion index gas isotopic fingerprint feature database is constructed.

[0010] S5. Optimize the mathematical model for isotope source quantification, and adopt the joint analysis method of carbon, hydrogen and oxygen isotope fingerprint feature database and Bayesian inversion to achieve molecular-level source tracing and accurate quantification of spontaneous combustion index gases in underground coal.

[0011] As a further aspect of the present invention: step S1 specifically includes the following steps:

[0012] S1.1 For coal seams that have not been exposed, geological coring boreholes shall be drilled using drilling rigs to obtain coal samples, and the principles of uniformity and representativeness shall be followed when sampling.

[0013] S1.2 For coal sample collection from exposed coal seams, first select the newly exposed area, remove the surface layer, and ensure that the coal sample taken is close to the properties of the original coal. The coal sample taken out should be sealed in a sealed container, and then the sampling record and numbering work should be carried out.

[0014] S1.3. The gas contained in the coal in the sealed container is collected by desorption method and placed in an aluminum foil gas sample bag and sealed.

[0015] S1.4 For in-situ gas sampling in exposed coal seams, select the newly exposed area, construct a gas sampling hole using a mine explosion-proof electric coal drill, and place thermocouples inside the hole to monitor the gas and coal body temperature in real time; purge nitrogen into the deepest part of the borehole; immediately after construction, seal the borehole with sealing material to ensure that the gas collected from the borehole is the original gas and nitrogen originally contained in the coal seam. After sealing, use an air pump to extract a certain amount of gas from the borehole and seal it in an aluminum foil gas sample bag.

[0016] As a further aspect of the present invention: the specific method for obtaining coal samples from the geological coring borehole in step S1.1 is as follows: the geological borehole is drilled to a distance of 2m from the coal seam, and compressed air is used to remove slag. After drilling to a distance of 0.5m from the coal seam, the coring drill bit is replaced. The coring drill bit should penetrate 0.5m into the coal seam before retracting the drill to obtain the coring sample. The extracted coal sample is quickly sealed in a sealing container, and then sampling records and numbering are performed.

[0017] As a further aspect of the present invention: step S2 specifically includes the following steps:

[0018] S2.1. Stable hydrogen isotopes of natively occurring gases are analyzed using a gas chromatography-isotope ratio mass spectrometry (GC-MS) system. Hydrogen-containing gases (such as CH4) need to be converted into H2 through high-temperature combustion or pyrolysis before being analyzed by mass spectrometry. Stable carbon isotopes of natively occurring gases are analyzed using isotope ratio mass spectrometry. Stable oxygen isotopes of natively occurring gases are first separated and purified, and then the purified gas is injected into the isotope ratio mass spectrometer for analysis. Liquid nitrogen is used at low temperature to separate and purify CO2 gas, and CO gas is separated and purified by a molecular sieve column of a gas chromatography sample separation system.

[0019] By comparing the isotope abundance ratio of the index gas collected downhole with the isotope abundance ratio of international standard materials, the δ values ​​of carbon, hydrogen, and oxygen isotopes of the gas are calculated, which is the difference in the isotope abundance ratio of the gas relative to the isotope abundance ratio of the standard material.

[0020] S2.2. After testing and analysis by isotope ratio mass spectrometry, a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of the coal seam's contained gas is formed and input into the isotope fingerprint database.

[0021] As a further aspect of the present invention: step S3 specifically includes the following steps:

[0022] S3.1. The gas production characteristics of each stage of coal spontaneous combustion were analyzed using a programmed heating device. Coal samples of different particle sizes were weighed using an electronic balance and mixed and dried in a vacuum drying oven.

[0023] S3.2 After drying, place the coal sample in a coal sample container and lay glass wool on the surface of the coal sample; seal the coal sample container, turn on the air generator, and calibrate the programmed heating device; set the initial temperature, heating rate, and compressed air inlet rate for testing;

[0024] S3.3 The programmed temperature rise device is run to raise the temperature and connected to the gas chromatograph to collect and detect the composition and content of the gas at different temperature stages using aluminum foil gas sample bags;

[0025] S3.4 After the programmed heating device cools to room temperature, refill with coal samples of the same particle size and weight, and introduce compressed nitrogen into the air inlet. Repeat the above steps to remeasure the gas composition and content at the outlet of each temperature point.

[0026] S3.5. Stable hydrogen isotopes of indicator gases at each stage of coal spontaneous combustion were tested and analyzed using a gas chromatography-isotope ratio mass spectrometry system. Hydrogen-containing gases (such as CH4) needed to be converted into H2 through high-temperature combustion or pyrolysis before entering the mass spectrometer to determine the isotope ratio. Stable carbon isotopes were analyzed using isotope ratio mass spectrometry. Stable oxygen isotope determination first involved separation and purification, followed by injection of the purified gas into the stable isotope mass spectrometer for analysis. Liquid nitrogen was used for low-temperature separation and purification of CO2 gas, and CO gas was separated and purified using a molecular sieve column of a gas chromatography sample separation system.

[0027] S3.6. Analyze the isotope values ​​of the indicator gases produced at different stages of coal oxidation using isotope ratio mass spectrometry, plot the "temperature-isotope value" characteristic curve, form a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of secondary indicator gases produced by coal oxidation, and input it into the isotope fingerprint database.

[0028] As a further aspect of the present invention: step S4 specifically includes the following steps:

[0029] S4.1 Collect gases from non-spontaneous combustion areas underground and measure gas isotope values ​​to establish a normal baseline database for coal mines, which will be used for subsequent data denoising.

[0030] S4.2 The isotope fingerprint database uses coal seam-temperature range as the primary key to construct the basic storage unit, ensuring that the characteristics of different oxidation stages of each coal seam are stored independently, providing temperature dimension analysis capability for fire source tracing;

[0031] S4.3. Isotope fingerprint database data preprocessing: Establish data filtering rules to reduce downhole operation noise interference and ensure the validity of input data.

[0032] As a further aspect of this invention: due to the adsorption / desorption effect of the coal seam, the isotope ratio of the index gas decreases during its transmission from the ignition source to the gas collection point, which can be described by an exponential decay model.

[0033] δ field =δ lab e -Lβ +ε

[0034]

[0035] Where, δ field The isotopic abundance ratio of the gas at the measured location downhole; δ lab The abundance ratio was calibrated for laboratory temperature-programmed tests; L is the straight-line distance from the ignition source to the monitoring point; β is the coal seam attenuation coefficient; ε is the measurement noise, which follows a zero-mean Gaussian distribution; σ noise The standard deviation of noise;

[0036] The coal seam attenuation coefficient β can be obtained by fitting downhole data using the least squares method.

[0037]

[0038] Where N represents the verification points arranged in the coal seam, and N≥30.

[0039] As a further aspect of the present invention: step S5 specifically includes the following steps:

[0040] S5.1 Considering gas generation capacity and transmission loss, physical constraints are applied to the source quantification mathematical model to make the source quantification results more consistent with the actual situation in coal mines. When the mixed index gas source has n end-members, the optimized stable isotope source calculation model is as follows:

[0041]

[0042] Where X is the content of one component gas in the gas mixture; δ mix δ represents the measured isotope value of a single component gas; δ is the isotope value of that single component gas in the index gas of each of the multiple coal seams; a, b, ... n are the contribution ratios of the gases from the n coal seams in the index gas, and f n The contribution ratio of a certain coal seam; Q is the gas production efficiency of a certain end-member coal body, mol / s;

[0043] S5.2. Use the isotope feature fingerprint database as the source tracing method to provide physical constraints and interpretability; use the Bayesian inversion method as the source tracing method to provide data probabilistic quantification and adaptability, and design a joint analysis method of isotope feature fingerprint database and Bayesian inversion.

[0044] Upgrade the static isotope fingerprint database to a temperature-related probability distribution:

[0045]

[0046] Where, μ k (T) represents the historical isotopic average value (‰) of coal seam k at temperature T; σ k (T) represents the standard deviation; T represents the temperature, which can be obtained from downhole sensors.

[0047] Establish a physical prior layer for contribution:

[0048]

[0049] Where f is the contribution ratio vector, satisfying ∑f k =1; α k For the prior concentration parameter, α k The larger the value, the higher the prior probability that the isotopic contribution of coal seam k is greater; h is the coal seam thickness, and I is the spontaneous combustion tendency of the coal seam.

[0050] A gas transport attenuation prior layer was established, and Hamiltonian Monte Carlo (HMC) sampling was used with ≥5000 iterations.

[0051]

[0052] Where, β k This is the coal seam gas transport attenuation coefficient; The prior mean; The prior standard deviation;

[0053] By utilizing graph neural networks to capture the spatiotemporal correlations between measurement points, the accuracy of cross-coal seam gas migration modeling is improved. Specifically, GNNs are used to capture cross-coal seam gas migration.

[0054]

[0055] in, The feature vector of node v at time t; Let W be the neighborhood of node v; W be the weight parameters of the neural network; e uv This refers to edge attributes;

[0056] The reliability of the output results of the source quantization unit is dynamically evaluated, and the formula for calculating the reliability of quantization based on multiple indicators is as follows:

[0057]

[0058] The formula for calculating the eigenvector v is as follows:

[0059]

[0060] Where Δδ is the deviation between the measured and predicted values; R is the range of the isotope fingerprint database; P consist SNR is the probability of consistency between the physical models;

[0061] The formula for calculating the signal-to-noise ratio (SNR) is:

[0062]

[0063] Where Var(signal) is the average power of the signal; Var(noise) is the average power of the noise;

[0064] The credibility-driven decision-making rules are as follows:

[0065]

[0066] When the calculated confidence level is greater than 0.8, the source tracing analysis results are reliable, and the gas source quantification unit can output the analysis results automatically. When the calculated confidence level is between 0.6 and 0.8, the source tracing analysis results are relatively reliable, but the source tracing analysis process needs to be manually reviewed to determine whether the source tracing results should be output or re-traced. When the calculated confidence level is less than 0.6, the source tracing analysis results have low confidence and require downhole monitoring verification. If necessary, sampling verification should be carried out, and the source quantification model should be retrained based on the verification results.

[0067] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0068] This invention prepares desorbed gas by drilling samples from various coal seams in a coal mine and determines the gas isotope abundance ratio using isotope ratio mass spectrometry to form isotope fingerprint characteristic maps. A temperature-programmed gas chromatography-isotope ratio mass spectrometry system is used to plot stable carbon, hydrogen, and oxygen isotope fingerprint characteristic maps of the oxidation process of coal samples from various coal seams at different temperatures. Gases from non-spontaneous combustion areas underground are collected, and normal background isotope values ​​are measured to establish a background database. This database, along with the stable isotope fingerprint characteristics of various gases, is used to construct a coal mine carbon, hydrogen, and oxygen isotope fingerprint database. Simultaneously, considering the adsorption / desorption effects of the coal seam, a coal seam attenuation coefficient is introduced to optimize the database data. After optimization, a source-specific quantification mathematical model with physical constraints and Bayesian deep inversion are used to achieve rapid matching and high-precision source tracing of carbon, hydrogen, and oxygen isotope fingerprints in coal spontaneous combustion indicator gases with the isotope fingerprint database. This effectively solves problems such as abnormal coal spontaneous combustion indicator gases in multiple coal seams, high false alarm rates, and unclear fire source location, improving the accuracy of coal spontaneous combustion disaster early warning and helping to ensure safe coal mine production. Attached Figure Description

[0069] Figure 1 This is a flowchart of the present invention.

[0070] Figure 2 This is a flowchart of the joint source quantification analysis of carbon, hydrogen, and oxygen isotope fingerprint feature database and Bayesian inversion in this invention. Detailed Implementation

[0071] The invention will now be further described with reference to the accompanying drawings.

[0072] like Figure 1 As shown, a method for tracing and analyzing the source of spontaneous combustion indicators of coal based on isotope fingerprinting includes the following steps:

[0073] S1. Collect coal samples from each coal seam in the coal mine to prepare desorption gas and coal oxidation index gas. The coal samples collected from each coal seam in the coal mine include those from exposed coal seams and those from unexposed coal seams.

[0074] Specifically, the following steps are included:

[0075] S1.1 For coal seams that have not been exposed, geological coring boreholes shall be drilled using drilling rigs to obtain coal samples, and the principles of uniformity and representativeness shall be followed when sampling.

[0076] Specifically, in the horizontal direction, sampling is conducted in zones, and sampling points are evenly distributed within the same area, taking into account site conditions. In the vertical direction, within the same coal seam, square grid lines are drawn on the coal seam cross-section, and a piece of coal is taken from each intersection of the grid lines to obtain uniform coal samples as much as possible from different fault cutting surfaces, mining area division surfaces, and geological unit division surfaces. At the same time, sampling locations should be selected from typical geological profiles of coal-bearing strata and all adverse geological anomalies (coal seam outcrops, burned areas, and faults, etc.) should be avoided.

[0077] The specific method for obtaining coal samples from the geological coring borehole in step S1.1 is as follows: When the geological borehole is drilled to a distance of 2m from the normal distance of the coal seam, compressed air is used to remove slag. After drilling to a distance of 0.5m from the coal seam, the coring drill bit is replaced. The coring drill bit should penetrate 0.5m into the coal seam before retracting the drill to obtain the coring sample. The extracted coal sample is quickly sealed in a sealing container, and then the sampling record and numbering work is carried out (indicating the sampling location, sample weight, sampling personnel, sampling time, layer number, etc.).

[0078] S1.2 For coal sample collection from exposed coal seams, first select the newly exposed area, and use a pickaxe or shovel to peel off the surface layer (peeling thickness ≥300mm) to ensure that the coal sample taken out is close to the properties of the original coal. The coal sample taken out should be sealed in a sealed container, and then the sampling record and numbering work should be carried out.

[0079] S1.3. The gas contained in the coal in the sealed container is collected by desorption method and placed in an aluminum foil gas sample bag and sealed.

[0080] S1.4 For in-situ gas sampling in exposed coal seams, select the newly exposed area, construct a gas sampling hole using a mine explosion-proof electric coal drill, and place thermocouples inside the hole to monitor the gas and coal body temperature in real time. Introduce nitrogen at the deepest point of the borehole to purge the small amount of CO and CO2 gas generated by drill bit friction. Immediately after construction, seal the borehole with sealing material to a depth of ≥50cm to ensure that the gas collected from the borehole is the original gas and nitrogen gas originally contained in the coal seam. Embed steel pipes in yellow mud for laying gas collection pipelines and thermocouples. After sealing, use an air pump to extract a certain amount of gas from the borehole and seal it in an aluminum foil gas sample bag.

[0081] S2. For unexposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics of the original occurrence index gas of the coal seam are analyzed by the desorbed gas prepared. For exposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics are analyzed by in-situ gas sampling, and the data is transmitted to the isotope fingerprint database.

[0082] Specifically, the following steps are included:

[0083] S2.1 The stable hydrogen isotopes of the native gas were tested using gas chromatography-isotope ratio mass spectrometry (GC-MS / MS), with the method being GC-MS-high-temperature pyrolysis-MS online analysis. The stable carbon isotopes of the native gas were analyzed using isotope ratio mass spectrometry (SMS), with the method conforming to the national standard GB / T 18340-2001. The stable oxygen isotopes of the native gas were first separated and purified, and then the purified gas was injected into the stable isotope mass spectrometer for analysis. Liquid nitrogen was used for low-temperature separation and purification of CO2 gas, and CO gas was separated and purified using a molecular sieve column in the GC sample separation system.

[0084] By comparing the isotopic abundance ratios of the index gas collected downhole with those of international standard materials, the carbon, hydrogen, and oxygen isotope δ values ​​of the gas are calculated, which are the differences in the isotopic abundance ratios of the gas relative to the isotopic abundance ratios of the standard materials. The formulas for calculating the carbon, hydrogen, and oxygen isotope δ values ​​are as follows:

[0085]

[0086] Wherein, the subscripts sample and VPDB represent the isotopic abundance ratios of carbon, oxygen and hydrogen in the gas sample and the stable isotopic molecular abundance ratios of carbon, oxygen and hydrogen in the international standard sample (ViennaPee Dee Belemnite), respectively.

[0087] S2.2. After testing and analysis by isotope ratio mass spectrometry, a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of the coal seam's contained gas is formed and input into the isotope fingerprint database.

[0088] S3. Determine the abundance of stable isotopes using a temperature-gas chromatography-isotope ratio mass spectrometry system and plot the "temperature-isotope value" characteristic curve of the stable isotope of the index gas.

[0089] Specifically, the following steps are included:

[0090] S3.1. The gas production characteristics of each stage of coal spontaneous combustion were analyzed using a programmed heating device. Coal samples of different particle sizes were weighed using an electronic balance and mixed and placed in a vacuum drying oven for 24 hours. The temperature was set to a fixed temperature of 30℃.

[0091] S3.2 After drying, place the coal sample in the coal sample container and lay glass wool on the surface of the coal sample with a thickness of 4-5 mm to prevent impurities from clogging the pipe; seal the coal sample container, turn on the air generator, and calibrate the programmed temperature rise device; set the initial temperature to 25℃, run it at a temperature rise rate of 1℃ / min, and introduce compressed air into the air inlet at a rate of 150ml / min, and maintain this state for the test;

[0092] S3.3 The programmed temperature rise device is used to raise the temperature and is connected to a gas chromatograph to collect and detect the composition and content of the gas at different temperature stages (30℃, 50℃, 60℃, 70℃, 80℃, 90℃, 100℃, 120℃, 140℃, 160℃, 180℃, 200℃, 220℃) using aluminum foil gas sample bags.

[0093] S3.4 After the programmed heating device cools to room temperature, refill with coal samples of the same particle size and weight, and introduce compressed nitrogen at a rate of 150 ml / min through the inlet. Repeat the above steps to remeasure the gas composition and content at the outlet of each temperature point.

[0094] S3.5. Stable hydrogen isotopes of indicator gases at each stage of coal spontaneous combustion were tested and analyzed using a gas chromatography-isotope ratio mass spectrometry (GC-MS / MS) system. The testing method was GC-MS-high-temperature pyrolysis-MS-online analysis. Stable carbon isotopes were analyzed using a stable isotope ratio mass spectrometer, and the testing method was in accordance with the national standard GB / T18340-200. Stable oxygen isotope determination involved separation and purification, followed by injection of the purified gas into a stable isotope mass spectrometer for analysis. Liquid nitrogen was used for low-temperature separation and purification of CO2 gas, and CO gas was separated and purified using a molecular sieve column in a GC sample separation system.

[0095] S3.6. Analyze the isotope values ​​of the indicator gases produced at different stages of coal oxidation using isotope ratio mass spectrometry, plot the "temperature-isotope value" characteristic curve, form a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of secondary indicator gases produced by coal oxidation, and input it into the isotope fingerprint database.

[0096] S4. By integrating the isotopic fingerprint features of the original gases in each coal seam, coal oxidation index gases, and normal gases in coal mines, a coal mine spontaneous combustion index gas isotopic fingerprint feature database is constructed.

[0097] Specifically, the following steps are included:

[0098] S4.1 Collect gas from non-spontaneous combustion areas underground (such as intake airways and unmined coal seams) and measure gas isotope values ​​to establish a normal background database for coal mines, which will be used for subsequent data noise reduction.

[0099] S4.2 The isotopic fingerprint database uses coal seam-temperature range as the primary key to construct the basic storage unit, ensuring that the characteristics of different oxidation stages of each coal seam are stored independently, providing temperature dimension analysis capability for fire source tracing. The primary key constraint formula for the isotopic fingerprint database is:

[0100] Primary Key = (s, T)

[0101] Where s represents the coal seam, a dimensionless quantity; T represents the temperature range of the coal sample from the coal seam, in °C;

[0102] S4.3. Isotope fingerprint database data preprocessing: Establish data filtering rules to reduce noise interference from downhole operations (such as vehicle exhaust) and ensure the validity of input data. When the measured isotope δ in downhole gas... meas At that time, the noise filtering rule is:

[0103]

[0104] Where ν represents the type of noise source (e.g., vehicle exhaust); Δ C Noise characteristic range; p monitor p represents the monitoring location; d represents the sampling location; eff The noise radius is [value].

[0105] Furthermore, due to the adsorption / desorption effects of coal seams, the isotope ratios of the indicator gas decrease during its transmission from the ignition source to the gas collection point, which can be described by an exponential decay model.

[0106] δ field =δ lab e -Lβ +ε

[0107]

[0108] Where, δ field The isotopic abundance ratio of the gas at the measured location downhole; δ lab The abundance ratio was calibrated for laboratory temperature-programmed tests; L is the straight-line distance from the ignition source to the monitoring point; β is the coal seam attenuation coefficient; ε is the measurement noise, which follows a zero-mean Gaussian distribution; σ noise The standard deviation of noise;

[0109] The attenuation coefficient β of coal seam i can be obtained by fitting downhole data using the least squares method:

[0110]

[0111] Where N represents the verification points arranged in the coal seam, and N≥30.

[0112] S5, such as Figure 2 As shown, an optimized isotope source quantification mathematical model was developed, and a joint analysis method combining carbon, hydrogen, and oxygen isotope fingerprint feature databases and Bayesian inversion was adopted to achieve molecular-level source tracing and accurate quantification of spontaneous combustion index gases in underground coal.

[0113] Specifically, the following steps are included:

[0114] S5.1. Considering gas generation capacity and transmission loss, physical constraints are applied to the isotope source quantification mathematical model to make the source quantification results more consistent with the actual situation in coal mines. When the mixed index gas source has n endmembers, the optimized stable isotope source quantification mathematical model is as follows:

[0115]

[0116] Where X is the content of one component gas in the gas mixture; δ mix δ represents the measured isotope value of a single component gas; δ is the isotope value of that single component gas in the index gas of each of the multiple coal seams; a, b, ... n are the contribution ratios of the gases from the n coal seams in the index gas, and f n The contribution ratio of a certain coal seam; Q is the gas production efficiency of a certain end-member coal body, mol / s;

[0117] S5.2. Use the isotope feature fingerprint database as the source tracing method to provide physical constraints and interpretability; use the Bayesian inversion method as the source tracing method to provide data probabilistic quantification and adaptability, and design a joint analysis method of isotope feature fingerprint database and Bayesian inversion.

[0118] Upgrade the static isotope fingerprint database to a temperature-related probability distribution:

[0119]

[0120] Where, μ k (T) represents the historical isotopic average value (‰) of coal seam k at temperature T; σ k (T) represents the standard deviation; T represents the temperature, which can be obtained from downhole sensors.

[0121] Establish a physical prior layer for contribution:

[0122]

[0123] Where f is the contribution ratio vector, satisfying ∑f k =1; α k For the prior concentration parameter, α k The larger the value, the higher the prior probability that the isotopic contribution of coal seam k is greater; h is the coal seam thickness, and I is the spontaneous combustion tendency of the coal seam.

[0124] A gas transport attenuation prior layer was established, and Hamiltonian Monte Carlo (HMC) sampling was used with ≥5000 iterations.

[0125]

[0126] Where, β k This is the coal seam gas transport attenuation coefficient; The prior mean; The prior standard deviation;

[0127] By utilizing graph neural networks to capture the spatiotemporal correlations between measurement points, the accuracy of cross-coal seam gas migration modeling is improved. Specifically, GNNs are used to capture cross-coal seam gas migration.

[0128]

[0129] in, The feature vector of node v at time t; Let W be the neighborhood of node v; W be the weight parameters of the neural network; e uv This refers to edge attributes;

[0130] The reliability of the output results of the source quantization unit is dynamically evaluated, and the formula for calculating the reliability of quantization based on multiple indicators is as follows:

[0131]

[0132] The formula for calculating the eigenvector v is as follows:

[0133]

[0134] Where Δδ is the deviation between the measured and predicted values; R is the range of the isotope fingerprint database; P consist SNR is the probability of consistency between the physical models;

[0135] The formula for calculating the signal-to-noise ratio (SNR) is:

[0136]

[0137] Where Var(signal) is the average power of the signal; Var(noise) is the average power of the noise;

[0138] The credibility-driven decision-making rules are as follows:

[0139]

[0140] When the calculated confidence level is greater than 0.8, the source tracing analysis results are reliable, and the gas source quantification unit can output the analysis results automatically. When the calculated confidence level is between 0.6 and 0.8, the source tracing analysis results are relatively reliable, but the source tracing analysis process needs to be manually reviewed to determine whether the source tracing results should be output or re-traced. When the calculated confidence level is less than 0.6, the source tracing analysis results have low confidence and require downhole monitoring verification. If necessary, sampling verification should be carried out, and the source quantification model should be retrained based on the verification results.

Claims

1. A method for tracing and analyzing the source of spontaneous combustion indicator gases in coal based on isotope fingerprinting, characterized in that, Includes the following steps: S1. Collect coal samples from each coal seam in the coal mine to prepare desorption gas and coal oxidation index gas. The coal samples collected from each coal seam in the coal mine include those from exposed coal seams and those from unexposed coal seams. S2. For unexposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics of the original occurrence index gas of the coal seam are analyzed by the desorbed gas prepared. For exposed coal seams, the stable carbon, hydrogen and oxygen isotope fingerprint characteristics are analyzed by in-situ gas sampling, and the data is transmitted to the isotope fingerprint database. S3. Determine the abundance of stable isotopes by programmed temperature rise-gas chromatography-stable isotope mass spectrometry, and plot the temperature-fingerprint characteristic curve of the index gas stable isotope. S4. By integrating the isotopic fingerprint features of the original gases in each coal seam, coal oxidation index gases, and normal gases in coal mines, a coal mine spontaneous combustion index gas isotopic fingerprint feature database is constructed. S5. Optimize the mathematical model for isotope source quantification, and adopt the joint analysis method of carbon, hydrogen and oxygen isotope fingerprint feature database and Bayesian inversion to achieve molecular-level source tracing and accurate quantification of spontaneous combustion index gases in underground coal.

2. The method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting according to claim 1, characterized in that, Step S1 specifically includes the following steps: S1.1 For coal seams that have not been exposed, geological coring boreholes shall be drilled using drilling rigs to obtain coal samples, and the principles of uniformity and representativeness shall be followed when sampling. S1.2 For coal sample collection from exposed coal seams, first select the newly exposed area, remove the surface layer, and ensure that the coal sample taken is close to the properties of the original coal. The coal sample taken out should be sealed in a sealed container, and then the sampling record and numbering work should be carried out. S1.

3. The gas contained in the coal in the sealed container is collected by desorption method and placed in an aluminum foil gas sample bag and sealed. S1.4 For in-situ gas sampling in exposed coal seams, select the newly exposed area, construct a gas sampling hole using a mine explosion-proof electric coal drill, and place thermocouples inside the hole to monitor the gas and coal body temperature in real time; purge nitrogen into the deepest part of the borehole; immediately after construction, seal the borehole with sealing material to ensure that the gas collected from the borehole is the original gas and nitrogen originally contained in the coal seam. After sealing, use an air pump to extract a certain amount of gas from the borehole and seal it in an aluminum foil gas sample bag.

3. The method for tracing and analyzing the source of coal spontaneous combustion index gases based on isotope fingerprinting according to claim 2, characterized in that, The specific method for obtaining coal samples from the geological coring borehole in step S1.1 is as follows: when the geological borehole is drilled to a distance of 2m from the coal seam, compressed air is used to remove slag. After drilling to a distance of 0.5m from the coal seam, the coring drill bit is replaced. The coring drill bit should penetrate 0.5m into the coal seam before retracting the drill to obtain the coring sample. The extracted coal sample is quickly sealed in a sealing container, and then the sampling record and numbering work is carried out.

4. The method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting according to claim 1, characterized in that, Step S2 specifically includes the following steps: S2.1 The stable hydrogen isotopes of the native gas were tested and analyzed using a gas chromatography-isotope ratio mass spectrometry system; the stable carbon isotopes of the native gas were analyzed using an isotope ratio mass spectrometer; the stable oxygen isotope determination of the native gas first involved separation and purification, and then the purified gas was injected into a stable isotope mass spectrometer for analysis. Liquid nitrogen was used at low temperature to separate and purify CO2 gas, and CO gas was separated and purified using a molecular sieve column of a gas chromatography sample separation system. By comparing the isotope abundance ratio of the index gas collected downhole with the isotope abundance ratio of international standard materials, the δ values ​​of carbon, hydrogen, and oxygen isotopes of the gas are calculated, which is the difference in the isotope abundance ratio of the gas relative to the isotope abundance ratio of the standard material. S2.

2. After testing and analysis by isotope ratio mass spectrometry, a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of the coal seam's contained gas is formed and input into the isotope fingerprint database.

5. The method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting according to claim 1, characterized in that, Step S3 specifically includes the following steps: S3.

1. The gas production characteristics of each stage of coal spontaneous combustion were analyzed using a programmed heating device. Coal samples of different particle sizes were weighed using an electronic balance and mixed and dried in a vacuum drying oven. S3.2 After drying, place the coal sample in a coal sample container and lay glass wool on the surface of the coal sample; seal the coal sample container, turn on the air generator, and calibrate the programmed heating device; set the initial temperature, heating rate, and compressed air inlet rate for testing; S3.3 The programmed temperature rise device is run to raise the temperature and connected to the gas chromatograph to collect and detect the composition and content of the gas at different temperature stages using aluminum foil gas sample bags; S3.4 After the programmed heating device cools to room temperature, refill with coal samples of the same particle size and weight, and introduce compressed nitrogen into the air inlet. Repeat the above steps to remeasure the gas composition and content at the outlet of each temperature point. S3.

5. Stable hydrogen isotopes of indicator gases at each stage of coal spontaneous combustion were tested and analyzed using a gas chromatography-isotope ratio mass spectrometry system; stable carbon isotopes were analyzed using an isotope ratio mass spectrometer; stable oxygen isotope determination first involved separation and purification, followed by injection of the purified gas into a stable isotope mass spectrometer for analysis. Liquid nitrogen was used for low-temperature separation and purification of CO2 gas, while CO gas was separated and purified using a molecular sieve column in a gas chromatography sample separation system. S3.

6. Analyze the isotope values ​​of the indicator gases produced at different stages of coal oxidation using isotope ratio mass spectrometry, plot the "temperature-isotope value" characteristic curve, form a stable carbon, hydrogen and oxygen isotope characteristic fingerprint spectrum of secondary indicator gases produced by coal oxidation, and input it into the isotope fingerprint database.

6. The method for tracing and analyzing coal spontaneous combustion indicator gases based on isotope fingerprinting according to claim 1, characterized in that, Step S4 specifically includes the following steps: S4.1 Collect gases from non-spontaneous combustion areas underground and measure gas isotope values ​​to establish a normal baseline database for coal mines, which will be used for subsequent data denoising. S4.2 The isotope fingerprint database uses coal seam-temperature range as the primary key to construct the basic storage unit, ensuring that the characteristics of different oxidation stages of each coal seam are stored independently, providing temperature dimension analysis capability for fire source tracing; S4.

3. Isotope fingerprint database data preprocessing: Establish data filtering rules to reduce downhole operation noise interference and ensure the validity of input data.

7. The method for tracing and analyzing the source of coal spontaneous combustion index gases based on isotope fingerprinting according to claim 6, characterized in that, Due to the adsorption / desorption effects of coal seams, the isotope ratios of the indicator gas decrease during its transfer from the ignition source to the gas collection point, which can be described by an exponential decay model. d field =δlab e -Lβ +e Where, δ field The isotopic abundance ratio of the gas at the measured location downhole; δ lab The abundance ratio was calibrated for laboratory temperature-programmed tests; L is the straight-line distance from the ignition source to the monitoring point; β is the coal seam attenuation coefficient; ε is the measurement noise, which follows a zero-mean Gaussian distribution; σ noise The standard deviation of noise; The coal seam attenuation coefficient β can be obtained by fitting downhole data using the least squares method. Where N represents the verification points arranged in the coal seam, and N≥30.

8. The method for tracing and analyzing the source of coal spontaneous combustion index gases based on isotope fingerprinting according to claim 1, characterized in that, Step S5 specifically includes the following steps: S5.

1. Considering gas generation capacity and transmission loss, physical constraints are applied to the isotope source quantification mathematical model to make the source quantification results more consistent with the actual situation in coal mines. When the mixed index gas source has n endmembers, the optimized stable isotope source quantification mathematical model is as follows: Where X is the content of one component gas in the gas mixture; δ mix δ represents the measured isotope value of a single component gas; δ is the isotope value of that single component gas in the index gas of each of the multiple coal seams; a, b, ... n are the contribution ratios of the gases from the n coal seams in the index gas, and f n The contribution ratio of a certain coal seam; Q is the gas production efficiency of a certain end-member coal body, mol / s; S5.

2. Use the isotope feature fingerprint database as the source tracing method to provide physical constraints and interpretability; use the Bayesian inversion method as the source tracing method to provide data probabilistic quantification and adaptability, and design a joint analysis method of isotope feature fingerprint database and Bayesian inversion. Upgrade the static isotope fingerprint database to a temperature-related probability distribution: Where, μ k (T) represents the historical isotopic average value (‰) of coal seam k at temperature T; σ k (T) represents the standard deviation; T represents the temperature, which can be obtained from downhole sensors. Establish a physical prior layer for contribution: Where f is the contribution ratio vector, satisfying ∑f k =1; α k Let α be the prior concentration parameter. k The larger the value, the higher the prior probability that the isotopic contribution of coal seam k is greater; h is the coal seam thickness, and I is the spontaneous combustion tendency of the coal seam. A gas transport attenuation prior layer was established, and Hamiltonian Monte Carlo (HMC) sampling was used with ≥5000 iterations. Where, β k This is the coal seam gas transport attenuation coefficient; The prior mean; The prior standard deviation; By utilizing graph neural networks to capture the spatiotemporal correlations between measurement points, the accuracy of cross-coal seam gas migration modeling is improved. Specifically, GNNs are used to capture cross-coal seam gas migration. in, The feature vector of node v at time t; N(v) is the neighborhood of node v; W is the weight parameter of the neural network; e uv This refers to edge attributes; The reliability of the output results of the source quantization unit is dynamically evaluated, and the formula for calculating the reliability of quantization based on multiple indicators is as follows: The formula for calculating the eigenvector v is as follows: Where Δδ is the deviation between the measured and predicted values; R is the range of the isotope fingerprint database; P consist SNR is the probability of consistency between the physical models; The formula for calculating the signal-to-noise ratio (SNR) is: Where Var(signal) is the average power of the signal; Var(noise) is the average power of the noise; The credibility-driven decision-making rules are as follows: When the calculated confidence level is greater than 0.8, the source tracing analysis results are reliable, and the gas source quantification unit can output the analysis results automatically. When the calculated confidence level is between 0.6 and 0.8, the source tracing analysis results are relatively reliable, but the source tracing analysis process needs to be manually reviewed to determine whether the source tracing results should be output or re-traced. When the calculated confidence level is less than 0.6, the source tracing analysis results have low confidence, and downhole monitoring verification is required. If necessary, sampling verification should be carried out, and the source quantification model should be retrained based on the verification results.

Images