Multi-isotope analysis and associated gas collaborative coal mine methane emission tracing method

Abstract

The present application belongs to the field of information technology, and particularly relates to a coal mine methane emission tracing method based on multi-isotope analysis and associated gas coordination. By fusing multi-isotope characteristics and associated gas ratio analysis, a multi-dimensional tracing system for coal mine regional methane emission is constructed, which significantly improves the accuracy and reliability of methane source identification. Compared with the traditional tracing method relying on single isotope or concentration statistical method, the present application can effectively distinguish various emission types such as coal mine gas, coking plant, coal-fired and agricultural sources in a complex mixed source environment, and realize high-precision cause analysis of multi-source methane emission. The present application realizes the spatio-temporal matching of isotope characteristics and transport path, and can quantitatively identify the influence of high-altitude and external source input on ground concentration.

Description

Technical Field

[0001] This invention belongs to the field of information technology, and in particular relates to a method for tracing the source of methane emissions from coal mines using multi-isotope analysis and associated gas synergy. Background Technology

[0002] In areas with a high concentration of coal mines, the sources of methane in the atmosphere are complex. These include gases released during coal mining, ventilation, and transportation, as well as emissions from surrounding industrial and biological sources such as coking plants, coal-fired power plants, and landfills. There may even be long-distance transport from coal-producing areas thousands of kilometers away. Therefore, scientifically and accurately identifying different methane emission sources and their contribution ratios is crucial for improving the accuracy of coal mine emission reduction and methane emission inventory compilation.

[0003] Currently, commonly used methane source apportionment techniques are mainly based on carbon stable isotopes (δ¹³C-CH4), hydrogen stable isotopes (δD-CH4), and radioactive isotopes (δ¹³C-CH4). 14 C-CH4 analysis. Methane from different emission sources exhibits different isotopic characteristics, and its isotopic composition depends on the methane formation pathway. Microbial methane is typically produced by the anaerobic decomposition of organic matter, with a δ¹³C value of approximately -61.7 ± 6.2‰ and a δD value of approximately -250‰ to -300‰; while thermogenic methane is formed by the pyrolysis of organic matter, with a δ¹³C value of approximately -44.8 ± 10.7‰ and a δD value of approximately -200‰ to -150‰. Because fossil fuels lack radioactive carbon, their δ¹³C values ​​are relatively low. 14 The C value is approximately -1000‰; while the δ value of biogenic methane produced from the decomposition of modern organic matter is... 14 The C value is close to the δ value of contemporary atmospheric CO2. 14 The C value is approximately -5‰. Therefore, by jointly analyzing δ¹³C, δD, and δ... 14 C can effectively distinguish between thermogenic methane and biogenic methane, and further identify the source of methane.

[0004] In terms of industrial source identification, although methane emissions from coking plants and coal-fired power plants both originate from fossil sources, their stable isotope characteristics and associated gas features differ significantly. Coking plant emissions are often accompanied by high concentrations of CO, Methane emissions from coal-fired power plants, particularly aromatic hydrocarbons, typically have δ¹³C values ​​around -40‰ and moderate C₂H₆ / CH₄ ratios. Methane emissions from coal-fired power plants have extremely low C₂H₆ / CH₄ ratios (<1 ppm) but significantly higher CO₂ concentrations. In contrast, coal mine methane gas is usually rich in heavy hydrocarbons such as ethane (C₂H₆ / CH₄ ratio >10 ppm) while having lower CO concentrations. These differences provide important gas ratio and isotopic evidence for identifying industrial methane sources.

[0005] In identifying long-distance methane transport sources in coal mining areas, the temporal overlap between external and local emissions makes it difficult to determine the source solely based on meteorological observations or concentration changes. Studies have shown that combining isotopic characteristics with atmospheric transport models (such as HYSPLIT and FLEXPART) for reverse trajectory analysis can effectively identify potential source regions of gas masses. When observed δ¹³C and δD significantly deviate from local coal mine endmember values, and the gas mass trajectory points towards external coal mining areas, it can be inferred that the gas primarily originates from distant coal mine emissions. Furthermore, inversion of mixed gas samples using the Keeling diagram method or the Miller-Tans method allows for the quantitative estimation of the contribution proportions from different sources.

[0006] In recent years, multi-gas co-observation technology has gradually become an important direction for methane source apportionment. Associated gases such as CO2, CO, and N2O can be used as auxiliary indicators to distinguish between industrial and biological sources. Coal-fired power plants emit high concentrations of CO2 with relatively stable CH4 / CO2 ratios; coking plants' high-temperature pyrolysis processes are accompanied by significant CO emissions; livestock farms and landfills often show elevated N2O concentrations. By combining gas ratios (such as CH4 / CO2, CH4 / CO, CH4 / N2O) with isotopic characteristics, the accuracy of methane source identification can be improved.

[0007] In summary, while existing methane source apportionment methods have achieved qualitative and semi-quantitative differentiation based on stable isotopes, in areas with dense coal mine distribution, complex source composition, and significant long-distance transport, single isotope indicators are insufficient to accurately analyze multi-source mixing characteristics. There is an urgent need for a method that combines multiple isotopes (δ¹³C, δD, δ¹³C, δ¹³ ... 14 C) A collaborative source tracing method based on the analysis and associated gas ratio characteristics to achieve high-precision identification and quantitative analysis of methane emission sources in coal mining areas. Summary of the Invention

[0008] Based on the above analysis, this application provides a coal mine methane emission source tracing method based on multi-isotope analysis and associated gas synergy, so as to achieve high-precision identification and quantitative analysis of methane emission sources in coal mine areas, and solve the technical problems of insufficient differentiation ability of existing methane source tracing methods under multi-source mixed conditions, insufficient utilization of associated gas information, and low identification accuracy of remote transmission sources.

[0009] Therefore, the technical solution of this application discloses a method for tracing the source of methane emissions from coal mines through multi-isotope analysis and synergistic interaction with associated gases, including the following steps:

[0010] S1. Multi-isotope data acquisition: Atmospheric methane samples were collected from the area above and around the target region to obtain carbon isotope δ¹³C-CH4 and radioactive carbon isotope δ¹³C-CH4. 14 C-CH4 and the abundance ratio of the hydrogen isotope δD-CH4;

[0011] S2. Associated gas ratio analysis: The concentration of associated gases in atmospheric methane samples was measured to obtain the ratio characteristics of each associated gas to CH4;

[0012] S3. Construction of Collaborative Traceability Model: Integrating δ¹³C and δ... 14 The input of C, δD and gas ratio features is based on a multidimensional discriminant model constructed using multiple linear regression and Bayesian discriminant methods to identify methane sources. It outputs the posterior probability distribution of each potential source class and the corresponding isotope endmember values. Values ​​with dispersed posterior probability distributions, isotope endmember values ​​exceeding the confidence interval of the source class feature database, or abnormal residuals in multiple linear regression are identified as outliers.

[0013] S4. Air mass trajectory inversion and tracking: For the outliers output by the model described in S3, reverse trajectory simulation is carried out using the HYSPLIT or FLEXPART atmospheric transport model to track the transport path of the air mass in time and space.

[0014] S5. Quantitative analysis of methane source types by end-member inversion: The Keeling plot method and the Miller-Tans method are used to invert the isotopic end-member values ​​of methane in the mixed gas sample and calculate the relative contribution ratio of different methane sources.

[0015] S6. Results Verification and Emission Inventory Correction: The model calculation results are compared and verified with ground flux observations, satellite inversion and existing emission inventories, and the methane emission intensity in the coal mining area is corrected and optimized.

[0016] The method for establishing the source characteristic database mentioned in S3 is as follows: collect or investigate the isotopic characteristic value range and associated gas ratio characteristics of typical methane emission sources in the target area, and establish a regionalized source characteristic database.

[0017] Furthermore, after collecting the atmospheric methane sample as described in S1, the sample is also concentrated in a cold trap, separated by chromatography, and analyzed by isotope ratio mass spectrometry.

[0018] Furthermore, the associated gas in S2 is CO2, CO, And C2H6.

[0019] Furthermore, the specific construction method of the multidimensional discriminant model based on multiple linear regression and Bayesian discriminant methods described in S3 is as follows:

[0020] Based on the source class feature database, a multi-source hybrid model is established based on the assumptions of mass conservation and linear superposition. A multiple linear regression function is constructed, and a Bayesian discriminant method is introduced to probabilistically identify the solution results of the multiple linear regression function. This results in a multidimensional discriminant model that outputs the probability distribution of each potential source class and the corresponding isotope endmember values.

[0021] Beneficial Effects: This invention constructs a multi-dimensional source tracing system for methane emissions in coal mining areas by integrating multi-isotope characteristics and associated gas ratio analysis, significantly improving the accuracy and reliability of methane source identification. Compared to traditional source tracing methods relying on single isotope or concentration statistics, this invention can effectively distinguish multiple emission types, such as coal mine gas, coking plants, coal combustion, and agricultural sources, in complex mixed source environments, achieving high-precision analysis of the causes of multi-source methane emissions. Secondly, this invention utilizes the concentration and ratio constraints of associated gases (CO2, CO, C2H6, N2O, etc.) to enhance the physical and chemical constraints for isotope discrimination, improving the interpretability and applicability of the model. This synergistic constraint mechanism exhibits stronger robustness in multi-source mixed areas, avoiding misjudgments caused by interference from meteorological or chemical reactions of single indicators, and significantly improving the credibility of methane emission contribution inversion. Furthermore, this invention combines meteorological reanalysis and atmospheric trajectory simulation to achieve spatiotemporal matching of isotope characteristics and transport paths, enabling quantitative identification of the impact of upper-air and external inputs on ground concentrations. This innovative integrated approach fills the gap in existing technologies for identifying cross-layer transport and regional diffusion mechanisms, provides a scientific basis for the calibration and dynamic updating of methane emission inventories in coal mining areas, and also provides important technical support for carbon emission regulation and greenhouse gas emission reduction assessment. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of the technology of this invention. Detailed Implementation

[0024] To make the technical problems solved, technical solutions, and beneficial effects of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0025] like Figure 1 As shown, the embodiments of this application disclose a method for tracing the source of methane emissions from coal mines based on multi-isotope analysis technology and associated gases, including the following steps:

[0026] S1. Multi-isotope data acquisition: Atmospheric methane samples were collected from the area above and around the target region to obtain carbon isotope δ¹³C-CH4 and radioactive carbon isotope δ¹³C-CH4. 14C-CH4 and hydrogen isotope (δD-CH4) data, which are δ¹³C, δ 14 The abundance ratio of C to δD is expressed as parts per thousand (‰); among which, δ¹³C-CH4 is used to distinguish between thermogenic and biogenic methane, δ 14 C-CH4 is used to determine the source depth of biogenic methane (shallow organic matter decomposition or coal seam decomposition), while δD-CH4 serves as an important supplementary indicator to improve the accuracy and reliability of isotope discrimination. The sampled gas undergoes cold trap concentration, chromatographic separation, and isotope ratio mass spectrometry analysis to ensure the stability and representativeness of each isotope signal.

[0027] This step utilizes the carbon stable isotope (δ¹³C), hydrogen stable isotope (δD), and radioactive carbon isotope (δ¹³C) of methane. 14 C) Features: Construct a multidimensional isotopic fingerprint system covering biogenic, thermogenic, and fossil-derived sources. The genetic differences between microbial methane and pyrolytic methane are distinguished by δ¹³C and δD characteristics. δ¹³C is then used to further differentiate the genesis of methane from microbial and pyrolytic methane. 14 The C content further distinguishes between fossil and modern methane sources, enabling high-precision differentiation of multiple emission sources and overcoming the limitation of insufficient identification capability of a single isotope index in complex source regions.

[0028] S2. Associated gas ratio analysis: The concentration of associated gases in atmospheric methane samples was measured to obtain the ratio characteristics of each associated gas to CH4 (such as CH4 / CO2, CH4 / CO, C2H6 / CH4).

[0029] This step introduces the concentration and ratio characteristics of associated gases such as CO2, CO, C2H6, and N2O to construct a methane-associated gas collaborative identification model. Through joint analysis of key ratios such as CH4 / CO2, CH4 / CO, and C2H6 / CH4 with isotopic fingerprints, the separability of different source types, including coal mine gas, coking plant emissions, coal-fired power plant emissions, and wetland biological emissions, is enhanced, improving the reliability and interpretability of source apportionment results. By analyzing the variation patterns of these ratio characteristics, emission characteristic fingerprints of different industrial sources such as coal mines, coking plants, and coal-fired power plants are identified. A source type feature database is constructed based on the ratio distribution, providing a reference for subsequent multi-source methane identification.

[0030] S3. Construction of Collaborative Traceability Model: Integrating δ¹³C and δ... 14 The input of C, δD, and gas ratio features is based on a multidimensional discriminant model constructed using multiple linear regression and Bayesian discriminant methods. This model identifies methane sources and outputs the probability distribution of each potential source class and the corresponding isotope endmember values, enabling quantitative differentiation of thermogenic, biogenic, and mixed-source methane.

[0031] This step, based on a multidimensional discriminant model constructed using multiple linear regression and Bayesian discriminant methods, establishes linear isotopic constraints on multi-source mixed gases, quantitatively inverts the isotopic endmember values ​​of different emission sources and their contribution ratio to regional methane concentration, thus achieving quantitative analysis of multi-source methane emissions in mixed scenarios. Specifically:

[0032] (1) Establishment of source characteristic database: Based on measured data of the target study area and published literature, a database of potential methane emission source characteristics is established using typical isotope and associated gas characteristic parameters. For the target area, the isotope characteristic value ranges (δ¹³C, ... 14 Based on the characteristics of C, δD) and associated gas ratios (CH4 / CO2, CH4 / CO, C2H6 / CH4, etc.), a regionalized source class characteristic database is established. Different regions need to reconstruct their own source class characteristic databases according to the differences in local source class composition.

[0033] (2) Establishment of multi-source hybrid model: Based on the source class feature database, a multi-source hybrid model is established based on the assumptions of mass conservation and linear superposition. A multivariate linear regression function is constructed, and a Bayesian discriminant method is introduced to probabilistically identify the solution results of the multivariate linear regression function, so as to obtain a multidimensional discriminant model that outputs the probability distribution of each potential source class and the corresponding isotope endmember values.

[0034] The δ¹³C and δ obtained from S1 and S2 14 C, δD, and gas ratio features are collaboratively input into a multidimensional discriminant model.

[0035] It should be noted that when establishing the source class feature database, typical methane emission sources within the target area are collected. For example, if there is a coking plant in the target area, the coking plant is a typical methane emission source. This application first collects data from this typical methane emission source to establish a basic source class feature database. Then, by measuring atmospheric methane samples (i.e., mixed source methane, which does not only come from typical methane emission sources) over and around the target area, the data is input into the model. Based on the aforementioned pre-established source class feature database, a methane source identification model is established through multiple linear regression and Bayesian discriminant methods. The model outputs the probability distribution of each potential source class and the corresponding isotope endmember values, thereby achieving quantitative differentiation of thermogenic, biogenic, and mixed source methane.

[0036] During this process, there may be data that does not conform to the source characteristic database established based on typical methane emission sources in the target area, and the source cannot be determined. These are outliers, and the gas mass trajectory inversion and tracking step S4 needs to be performed.

[0037] In a further embodiment, the criteria for determining outliers include:

[0038] (1) Source class probability anomaly identification: The posterior probability distribution of each potential source class is dispersed, or the posterior probability of the most probable source class is lower than the set threshold, indicating that the sample is difficult to be reasonably explained by the known source class combination in the regional source class feature database.

[0039] (2) Isotope endmember deviation identification: The δ¹³C, δ¹⁴C or δD endmember values ​​significantly exceed the statistical confidence interval range of the corresponding source class in the source class feature database, indicating that there may be methane sources or foreign inputs that are not included in the database.

[0040] (3) Regression residual anomaly identification: If the fitting residual of the multiple linear regression model is significantly large, or the standardized residual exceeds the preset statistical threshold, it indicates that the assumptions of mass conservation and linear superposition do not hold true for this sample.

[0041] (4) Inconsistency in isotope-associated gas identification: The source type indicated by the isotope characteristics and the associated gas ratio is inconsistent, indicating that the sample may be affected by mixed transport or non-local emissions.

[0042] Samples that meet any of the above conditions are judged to have high uncertainty or significantly large fitting residuals and are defined as outlier samples. That is, outlier values ​​are those that show dispersion in the posterior probability distribution of each potential source class, or are isotope endmember values ​​that exceed the confidence interval of the source class feature database, or are abnormal in the multiple linear regression residuals.

[0043] S4. Air mass trajectory inversion and tracking: For the outliers in the model output described in S3, reverse trajectory simulation is carried out using the HYSPLIT or FLEXPART atmospheric transport model to track the transport path of the air mass in time and space; combined with the industrial activity type and isotopic characteristics of the trajectory endpoint region, potential source regions of remotely transported air masses are identified, and the contribution of external input to local methane concentration is quantitatively assessed.

[0044] This step combines meteorological reanalysis data with atmospheric transport trajectory models (such as HYSPLIT) to conduct reverse trajectory simulations of air masses, identifying external input pathways for methane concentration anomalies at observation points. By matching isotopic characteristics with the trajectory source region, the impact of upper-air or external transport on surface methane concentrations is quantitatively analyzed, refining the spatiotemporal attribution of regional emission inventories.

[0045] S5. Quantitative source classification by end-member inversion: The Keeling diagram method and the Miller-Tans method are used to invert the isotopic end-member values ​​of methane in the mixed gas sample and calculate the relative contribution ratio of different methane sources. Through this step, the contribution share of methane from multiple sources such as coal mining, coking plant emissions, biodegradation and long-distance transport can be accurately estimated.

[0046] S6. Results Validation and Emission Inventory Correction: The model calculation results are compared and validated with ground flux observations, satellite inversion, and existing emission inventories to correct and optimize the methane emission intensity in coal mining areas. This results in a high-precision, source-specific methane emission assessment, providing scientific support for regional emission reduction strategies and regulation.

[0047] Through the above technical means, this invention achieves multi-factor fusion source tracing of methane emissions in coal mining areas, enabling accurate identification and quantitative analysis of multi-source mixed methane in complex industrial source areas, providing technical support for methane emission inventory calibration and regional emission reduction decisions. The specific steps and technical effects of this application will be described in detail below through specific embodiments.

[0048] Example 1: Source tracing of methane emissions from coal mines through multi-isotope analysis and synergistic effects with associated gases

[0049] A typical coal mining area was selected as the research object. This area contains multiple operating coal mines, closed mines, and supporting coking plants. Within a 20 km radius, there are coal-fired power plants and natural wetlands. The methane emission sources are complex and exhibit typical multi-source mixing characteristics, making it suitable for verifying the applicability and effectiveness of the method of this invention in complex scenarios.

[0050] Step 1: Multi-isotope data acquisition

[0051] Within the target coal mine area, sampling points were set up at the following locations: downwind of the coal mine entrance, at the mine boundary, at the regional background point, at the upwind reference point, at the surrounding coking plant, coal-fired power plant, landfill, and natural wetland, for a total of 9 fixed sampling points.

[0052] (1) Sample collection:

[0053] Use 200 L high-pressure stainless steel cylinders to regularly collect near-source atmospheric samples from emission sources such as coal mines, coking plants, coal-fired power plants, and landfills. The sampling height is 1.5 meters above the ground, and 2-3 parallel samples are collected continuously at each sampling point.

[0054] (2) Sample pretreatment and testing:

[0055] The collected gas samples were processed through the following steps: 1) Methane was enriched using a liquid nitrogen cold trap; 2) Methane was separated using gas chromatography; 3) δ¹⁸O₂ mass spectrometry was used to determine the methane content. 13 C-CH4 and δD-CH4; 4) The enriched methane was converted into CO2, and the δD-CH4 was determined by accelerator mass spectrometry (AMS). 14 C-CH4.

[0056] (3) Data Example:

[0057] The observed values ​​at a certain sampling point are shown in Table 1:

[0058] Table 1. Observed values ​​at a certain sampling point

[0059] .

[0060] The CH4 concentration at the sampling point was 2.85 ppm, significantly higher than the background atmospheric methane concentration, indicating that the sampling point was affected by local methane emission sources. δ 13 The C-CH4 value is -36.2‰, significantly enriched in ¹³C, and more indicative of thermogenic or coal-associated methane compared to biogenic methane (typically less than -55‰). The δD-CH4 value is -185‰, consistent with the typical range for coalbed methane and coking-related methane, but significantly different from biogenic methane from wetlands (typically less than -250‰). 14 The C-CH4 value is -980‰, close to -1000‰, indicating that the methane in the sample does not contain modern carbon and belongs to fossil carbon sources, which can effectively exclude methane from modern biological sources or recent biomass decomposition.

[0061] Combining the characteristics of CH4 concentration increase and δ 13 C-CH4, δD-CH4 and δ 14 The C-CH4 multi-isotope fingerprint characteristics indicate that the methane at this sampling point mainly originates from thermally formed methane controlled by fossil carbon, and it is preliminarily identified as a coal mine gas or coking-related fugitive emission source.

[0062] Step 2: Associated Gas Ratio Analysis

[0063] Simultaneous determination of associated gas concentrations, including CO2, CO, and other gases, of the same gas sample. C2H6.

[0064] (1) Gas concentration measurement:

[0065] CO2 and CO were determined using a non-dispersive infrared analyzer; C2H6 was determined using a gas chromatography-flame ionization detector (GC-FID). The determination was performed using chemiluminescence analysis.

[0066] (2) Ratio calculation:

[0067] Calculate the ratio of each associated gas to methane:

[0068] ;

[0069] (3) Example results are shown in Table 2.

[0070] Table 2. Example Results

[0071] .

[0072] The CH4 / CO2 ratio is at 10. -3 The magnitude indicates a relatively high CO2 content in the sample, which may be related to coal seam desorption, mine ventilation dilution, or release of inorganic CO2 from coal-bearing strata. The CH4 / CO ratio is significantly higher than expected (10). 2 The concentration of CO (at a certain level) indicates a low CO content, which does not meet the emission characteristics of combustion or high-temperature oxidation, and can effectively eliminate emissions from vehicle exhaust, biomass combustion, and industrial combustion sources. CH4 / NO x The ratio is extremely high (on the order of 10³), indicating The extremely low concentration further indicates that the gas is not a combustion-type or high-temperature industrial emission source, but rather more consistent with the characteristics of non-combustion emissions from coal mine gas or coking processes. The high C2H6 / CH4 ratio (18 ppm) indicates the presence of significant thermogenic hydrocarbons in the methane, distinguishing it from biogenic methane and consistent with characteristics of coalbed methane or coking-related gas sources. In summary, this gas sample exhibits a high C2H6 / CH4 ratio and a relatively high CH4 / CO ratio. Ratio, low CO and low Based on the characteristic combination of the content, it was initially determined that the source was coal mine gas emissions or coking-related fugitive sources, rather than combustion-related or transportation-related emissions.

[0073] Step 3: Construction of Collaborative Traceability Model

[0074] (1) Establishment of source class feature database

[0075] Based on measured data from the study area and reported characteristic parameters of typical isotopes and associated gases in publicly available literature, a database of potential methane emission source categories was established. These potential methane sources include at least coal mine gas, coking plant emissions, coal-fired power plant emissions, and biogenic methane. The characteristic parameter ranges for each source category are shown in Table 3.

[0076] Table 3. Database of characteristics of potential methane sources

[0077] .

[0078] The aforementioned database is used to define "feature endmembers" of different source classes, serving as prior information for subsequent hybrid model calculations.

[0079] (2) Establishment of multi-source hybrid model

[0080] Under actual observation conditions, methane at monitoring points typically originates from the mixed contributions of multiple emission sources. Based on the assumptions of mass conservation and linear superposition, a multi-source linear mixing model is established.

[0081] 1) Definition of observation vector:

[0082] For a given observation point, construct an observation feature vector for methane source identification:

[0083] ;

[0084] in, These are the stable carbon isotopes, stable hydrogen isotopes, and radioactive carbon isotope values ​​of methane, respectively. Methane and CO2, CO, respectively The ratio of C2H6 to C2H6.

[0085] 2) Definition of source vector at the source class end:

[0086] For the k-th potential methane source, its characteristic endmember vector is represented as:

[0087] The terminator parameters are derived from the source class feature database established above.

[0088] 3) Linear mixed relationship:

[0089] Assuming the methane at the observation point is formed from a mixture of potential sources of type K, then:

[0090] ;

[0091] in, This represents the proportion of methane from the Kth methane source that contributes to the methane levels at that observation point. This contribution proportion satisfies the following physical constraints:

[0092] ;

[0093] (3) Initial solution of multiple linear regression

[0094] Since the source class endmember parameters are known but the contribution ratio is unknown, a multiple linear regression method with non-negativity constraints is used to initially solve for the contribution ratio. The objective function is constructed as follows:

[0095] ;

[0096] in, This step provides initial estimates of the contribution proportions of various sources. It is used to obtain physically feasible initial solutions, providing initial conditions for subsequent probabilistic discrimination.

[0097] (4) Construction of Bayesian discriminant model

[0098] To fully consider the observation error and the uncertainty of source class characteristics, a Bayesian discriminant model is introduced based on step (3) to probabilistically identify the methane source.

[0099] 1) Likelihood function construction:

[0100] Assuming the observation error follows a multivariate normal distribution, the likelihood function of the observation vector is expressed as:

[0101] ;

[0102] Where Σ is the observation error covariance matrix.

[0103] 2) Prior distribution specification

[0104] Considering the nonnegativity and normalization constraints of the contribution proportions of each source class, a Dirichlet prior distribution is introduced:

[0105] ;

[0106] in, Used to reflect the prior weight information of each source class.

[0107] 3) Solving for the posterior probability

[0108] Based on the likelihood function and prior distribution mentioned above, the posterior distribution is sampled using the Markov Chain Monte Carlo (MCMC) method to obtain the posterior probability of each potential methane source class.

[0109] Taking a certain observation point as an example, the probability distribution of its methane source is shown in Table 4.

[0110] Table 4. Probability distribution of methane sources

[0111] .

[0112] The probability results are used to characterize the uncertainty distribution of methane sources at the observation point and serve as the basis for subsequent source tracing analysis and emission attribution.

[0113] It should be noted that the discrimination method framework used in S3 of this application is general, but the model parameters depend on a regional source class feature database. Different regions differ due to variations in emission source types and characteristics, necessitating the reconstruction of the regional source class feature database and retraining of the discrimination model accordingly. Once the model is constructed for a specific region, subsequent source tracing work within that region can directly input the collected data without repeated modeling.

[0114] Step 4: Air mass trajectory inversion and tracking

[0115] For the high-probability coal mine gas sampling points identified in step S3, and the abnormal sampling points with multi-source mixing or high uncertainty in the source probability distribution, the atmospheric transport trajectory inversion method is further introduced to track and analyze the historical transport paths that methane may have experienced, so as to determine whether the methane at the observation point is affected by external transport and to quantitatively assess the contribution of external input.

[0116] (1) Trajectory inversion model and meteorological data construction:

[0117] In this example, the HYSPLIT (Hybrid Single-Particle Lagrangian Integrated Trajectory) atmospheric transport model is used to simulate the reverse trajectory of an air mass. The HYSPLIT model is based on the Lagrangian method and simulates the trajectory of air particles in a three-dimensional wind field through numerical integration.

[0118] The meteorological driving data used is global reanalysis data from GDAS (Global Data Assimilation System). The meteorological data includes, but is not limited to, three-dimensional wind speed field, pressure field, temperature field and boundary layer parameters, which are used to construct the spatiotemporal continuous meteorological field required for trajectory inversion.

[0119] (2) Setting parameters for reverse trajectory simulation:

[0120] Using the geographic coordinates of the observation point to be analyzed as the starting point of the trajectory, multiple starting heights are set for different atmospheric levels to characterize the transport characteristics of air masses at different altitudes. In this example, the starting heights are set to 100m, 300m, and 500m. The trajectory is terminated at the observation time, and trajectory integration is performed in the reverse direction of time to trace the transport path of the air mass over a historical 72-hour period. The trajectory calculation time step is set to 10 minutes to improve the temporal resolution and stability of the trajectory simulation.

[0121] (3) Equations of air mass motion and methods for calculating trajectory:

[0122] The reverse trajectory of an air mass is obtained by numerical integration of the air mass motion equation, and its basic form is as follows:

[0123] ;

[0124] Where X(t) represents the spatial position vector of the air mass at time t; Let X represent the time step of the trajectory integration; V(X, t) represents the three-dimensional wind speed vector obtained by interpolating meteorological reanalysis data at spatial location X and time t. By stepwise integration of the above equation over the reverse time series, a set of historical transport trajectories of the air mass under different altitude conditions is obtained.

[0125] (4) Trajectory spatial statistics and potential source region identification:

[0126] Spatial statistical analysis was performed on all the obtained reverse trajectory results. The specific method is as follows:

[0127] 1) Divide the study area into pre-defined spatial grid units;

[0128] 2) Assign the trajectory path points to the corresponding grid cells according to their spatial location;

[0129] 3) Statistical analysis of the residence time of each trajectory in different grid cells is used to characterize the intensity of air mass residence in different regions during historical transport.

[0130] Grid cells with longer residence times were identified as potential source impact areas to reflect the spatial location of air masses that may have been affected by emissions during historical transport.

[0131] (5) Calculation of the contribution index of external transport:

[0132] To quantitatively characterize the potential impact of exogenous transport on methane concentration at the observation point, an exogenous transport contribution index C is introduced. ext The calculation formula is as follows:

[0133] ;

[0134] Among them, RT ext RT indicates the residence time of air masses whose trajectory falls outside the study area or within potential external emission source areas; total This represents the total residence time of all reverse trajectories at the observation point. The exogenous transport contribution index is used to reflect the relative influence of exogenous regions on methane at the observation point during the historical transport of the gas mass.

[0135] (6) Joint identification of trajectory results and isotopic characteristics:

[0136] By combining the air mass trajectory inversion results with the isotopic end-member features obtained in step 3, a coupled identification mechanism of physical transport path and chemical fingerprint information is formed:

[0137] 1) When the reverse trajectory mainly points to the concentrated distribution area of ​​local coal mines, and the methane isotope characteristics of the observation point are consistent with the gas end-member characteristics of the local coal mines, it is determined that the methane at the observation point is mainly emitted by the local coal mines.

[0138] 2) When the reverse trajectory points to an external industrial area, a remote coal mining area or other potential emission area, and the isotopic characteristics of the observation point deviate significantly from the local coal mine gas end-member range, it is determined that there is a significant external source of methane transport at the observation point.

[0139] The results of this example show that the gas mass mainly originated from the coal mining area within the region and did not point to the distant coal mining area, thus ruling out the dominant influence of long-distance transport.

[0140] Step 5: Keeling plot and Miller-Tans endmember inversion

[0141] After completing the source class probability identification in step 3 and the gas mass trajectory inversion constraint in step 4, the Keeling diagram method and the Miller-Tans method are further introduced to invert the regional methane isotope end-member characteristics and quantitatively calculate the contribution ratio of different sources to the regional methane concentration.

[0142] (1) Keeling diagram construction and endmember isotope inversion:

[0143] The Keeling diagram method is based on the assumption that the background concentration is relatively stable in the mixed system and the source end-member isotope composition is approximately constant. It inverts the isotope end-member values ​​of the main emission sources by observing the linear relationship between methane concentration and its isotope composition.

[0144] Construct the Keeling relation by relating the methane carbon isotope composition δ¹³C at the observed sample points to the methane volume fraction [CH₄].

[0145] ;

[0146] Where δ¹³C represents the carbon isotopic composition of methane at the observation point; [CH4] represents the volume fraction of methane at the corresponding time or sample point; The slope of the linear regression; The intercept is the linear regression value, representing the isotopic endmember characteristic value of the dominant emission source in the mixed system.

[0147] By performing least-squares linear regression analysis on multiple observation points, the intercept value was obtained: b = -37.1

[0148] The intercept value b is highly consistent with the coal mine gas isotope end-member range constructed in step 3, indicating that methane emissions in the study area are mainly from coal mine gas.

[0149] (2) Miller–Tans method for verification:

[0150] The Miller-Tans method is suitable for situations involving multi-endmember mixing or significant changes in background concentration over time. Its basic formula is:

[0151] ;

[0152] in, Indicates the carbon isotopic composition of the source endmember; The background methane concentration and its isotopic composition are represented respectively.

[0153] Simplifying the formula, we get:

[0154] ;

[0155] The source-end-member isotope values ​​obtained by linear regression are consistent with the results obtained by the Keeling diagram method, further verifying the reliability of coal mine gas as the dominant source.

[0156] (3) Calculation method for the contribution ratio of each source:

[0157] The source class contribution coefficients of the multi-source hybrid model obtained in step 3 Based on this, combined with the total methane concentration C at the observation point or region total Calculate the absolute contribution of each source to the regional methane concentration:

[0158] ;

[0159] in, This represents the contribution of emission source k to the regional methane concentration; C represents the normalized contribution ratio of the corresponding source class in the multi-source hybrid model; total This represents the total methane concentration at the observation point or within the region. After normalizing the contributions of each source, the relative contribution ratios of different sources to the regional methane concentration are obtained.

[0160] (4) Regional methane source contribution results:

[0161] Taking the study area in this example as an example, combining the endmember inversion results of the Keeling plot, the contribution coefficients of the multi-source mixing model, and the physical constraints of trajectory inversion, the contribution ratio of methane sources in the region is obtained as follows:

[0162] 1) Sources of coal mine gas: approximately 70%;

[0163] 2) Sources from industrial processes such as coking plants: approximately 20%;

[0164] 3) Other sources (including bio-based and low-contribution industrial sources): approximately 10%.

[0165] Step 6: Result Verification and Emission Inventory Revision

[0166] The methane source identification results obtained from the above steps are compared and analyzed with ground flux observation data, satellite inversion results, and existing methane emission inventories to verify the methane emission intensity in the coal mining area. Based on the verification results, the emission amounts of various sources in the regional methane emission inventory are corrected and optimized to form a coal mining area methane emission assessment result with clear source identification and high accuracy. The method described in this example achieves the collaborative identification and quantitative analysis of multi-source methane emissions in coal mining areas. Compared with methods relying solely on a single isotope or concentration statistics, this invention can significantly improve the accuracy and stability of methane source identification in complex scenarios involving multi-source mixing and external transport influences, verifying the feasibility and practical value of the method.

[0167] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for tracing the source of methane emissions from coal mines through multi-isotope analysis and synergistic effects with associated gases, characterized in that, Includes the following steps: S1. Multi-isotope data acquisition: Atmospheric methane samples were collected from the area above and around the target region to obtain carbon isotope δ¹³C-CH4 and radioactive carbon isotope δ¹³C-CH4. 14 C-CH4 and the abundance ratio of the hydrogen isotope δD-CH4; S2. Associated gas ratio analysis: The concentration of associated gases in atmospheric methane samples was measured to obtain the ratio characteristics of each associated gas to CH4; S3. Construction of Collaborative Traceability Model: Integrating δ¹³C and δ... 14 The input of C, δD and gas ratio features is based on a multidimensional discriminant model constructed using multiple linear regression and Bayesian discriminant methods to identify methane sources. It outputs the posterior probability distribution of each potential source class and the corresponding isotope endmember values. Values ​​with dispersed posterior probability distributions, isotope endmember values ​​exceeding the confidence interval of the source class feature database, or abnormal residuals in multiple linear regression are identified as outliers. S4. Air mass trajectory inversion and tracking: For the outliers output by the model described in S3, reverse trajectory simulation is carried out using the HYSPLIT or FLEXPART atmospheric transport model to track the transport path of the air mass in time and space. S5. Quantitative analysis of methane source types by end-member inversion: The Keeling plot method and the Miller-Tans method are used to invert the isotopic end-member values ​​of methane in the mixed gas sample and calculate the relative contribution ratio of different methane sources. S6. Results Verification and Emission Inventory Correction: The model calculation results are compared and verified with ground flux observations, satellite inversion and existing emission inventories, and the methane emission intensity in the coal mining area is corrected and optimized. The method for establishing the source characteristic database mentioned in S3 is as follows: collect or investigate the isotopic characteristic value range and associated gas ratio characteristics of typical methane emission sources in the target area, and establish a regionalized source characteristic database.

2. The tracing method according to claim 1, characterized in that, After collecting atmospheric methane samples, S1 also includes cold trap concentration, chromatographic separation, and isotope ratio mass spectrometry analysis of the samples.

3. The tracing method according to claim 1, characterized in that, The associated gas in S2 is CO2, CO, And C2H6.

4. The tracing method according to claim 1, characterized in that, The specific construction method of the multidimensional discriminant model based on multiple linear regression and Bayesian discriminant method described in S3 is as follows: Based on the source class feature database, a multi-source hybrid model is established based on the assumptions of mass conservation and linear superposition. A multiple linear regression function is constructed, and a Bayesian discriminant method is introduced to probabilistically identify the solution results of the multiple linear regression function. This results in a multidimensional discriminant model that outputs the probability distribution of each potential source class and the corresponding isotope endmember values.

Images