Method and apparatus for electromagnetic prospecting of gas-bearing coal reservoirs

Abstract

The application provides an electromagnetic exploration method and device for a gas-containing coal reservoir, and the method comprises the following steps: obtaining the frequency spectrum and spatial distribution characteristics of a self-emission simulation signal according to a pre-constructed coal reservoir self-emission forward model, and in combination with preset underground resistivity parameters, preset electric dipole depth parameters and preset angle parameters; obtaining a measured electromagnetic field at a surface measuring point, comparing the measured electromagnetic field with the frequency spectrum and spatial distribution characteristics of the self-emission simulation signal, so as to determine whether there is a coal bed methane in a preset range below the corresponding measuring point; in the case that it is determined that there is a coal bed methane in the preset range below the measuring point, the corresponding measured electromagnetic field is used in combination with an electromagnetic exploration inversion model to obtain a coal reservoir exploration result. The application corrects the influence of the coal reservoir self-emission on the natural source resistivity inversion, can obtain a correct resistivity through electromagnetic sounding in a coal bed methane enrichment area, and directly quantitatively evaluates the coal reservoir depth, structural direction and gas content parameters according to the obtained inversion result.

Description

Technical Field

[0001] This invention relates to the field of coalbed methane exploration technology, and in particular to an electromagnetic exploration method and apparatus for gas-bearing coal reservoirs. Background Technology

[0002] Coalbed methane, also known as coal seam gas, refers to a mixture of gases found in some coal seams, primarily composed of hydrocarbons such as methane (methane content typically exceeds 50%), along with other gases such as carbon dioxide and nitrogen. It is an important new type of oil and gas energy source. my country ranks third in the world in coalbed methane resources, with recoverable reserves at depths shallower than 2000 meters (36.81 × 10¹² m³) comparable to natural gas reserves (38 × 10¹² m³). Currently, my country heavily relies on imported natural gas, while coalbed methane shares the same main components as natural gas—both are high-calorific-value, low-pollution clean oil and gas energy sources—and will be an important supplement and replacement for natural gas in the future.

[0003] Natural source electromagnetic exploration (SEM) has the advantages of high efficiency, low cost, and minimal impact from topography. In the past fifteen years, it has attracted the attention of coalbed methane exploration researchers. SEM attempts to use classical algorithms to find underground water-rich and low-resistivity zones to indicate coalbed methane. It takes the absence of self-emission in underground rock strata as a basic assumption and focuses more on impedance and apparent resistivity rather than the anomalous characteristics of the original electromagnetic data. At present, there is a lack of detailed research on the characteristics of self-emission signals in actual underground coal reservoirs and their impact on impedance and apparent resistivity. Therefore, existing scholars only consider resistivity when conducting electromagnetic detection of coal reservoirs and do not introduce the self-emission of coal reservoirs.

[0004] However, forward modeling results that only consider the low resistivity of water-rich areas cannot match the amplitude anomalies of measured electromagnetic signals. Furthermore, coal and gas-bearing coal and rock will generate electromagnetic radiation due to piezoelectric effect, electrokinetic effect, Stepanov effect, etc., and within a certain range, the electromagnetic radiation will be significantly enhanced as the gas content increases (gas pressure increases), making it difficult to obtain correct inversion results due to interference from the self-emission of coal reservoirs. In addition, the coalbed methane depth inverted by only considering the low resistivity of water-rich areas has multiple solutions, and it is difficult to obtain further quantitative information on gas content in principle. Summary of the Invention

[0005] This invention provides an electromagnetic exploration method and apparatus for gas-bearing coal reservoirs, which solves the defects in the prior art where the self-emission of coal reservoirs leads to inaccurate coal reservoir exploration results. It corrects the influence of coal reservoir self-emission on the inversion of natural source resistivity and enhances the accuracy of coal reservoir exploration.

[0006] This invention provides an electromagnetic exploration method for gas-bearing coal reservoirs, comprising: obtaining the spectral characteristics and spatial distribution characteristics of a self-emission simulated signal based on a pre-constructed forward model of the coal reservoir self-emission, combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters; wherein the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include a simulated electromagnetic field signal; acquiring the measured electromagnetic field at a surface measuring point, comparing the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emission simulated signal, to determine whether, within a frequency range of a preset peak value determined based on the spectral characteristics of the self-emission simulated signal, a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field exists, and based on its existence, determining the perimeter of the corresponding surface measuring point. The presence of coalbed methane is determined. If coalbed methane is present around the corresponding surface measuring point, the horizontal magnetic field strength at the surface measuring point is determined to be greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal. If the horizontal magnetic field strength is greater than the vertical magnetic field strength, the coalbed methane center is determined to be located within a preset range below the corresponding surface measuring point. If coalbed methane is determined to exist within the preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with an electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is obtained based on simulated electromagnetic exploration training data and the corresponding electromagnetic exploration labels. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0007] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, when it is determined that coalbed methane exists within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with an electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. This includes: using the measured electromagnetic field at the surface measuring point where coalbed methane exists as the inversion input data of the electromagnetic exploration inversion model to obtain the coal reservoir exploration results output by the electromagnetic exploration inversion model; wherein the measured electromagnetic field is a spectrum curve composed of a mixture of natural source electromagnetic fields from the ionosphere and lightning and self-emitted source electromagnetic fields from the coal reservoir, and the measured electromagnetic field includes an amplitude spectrum curve and a phase spectrum curve.

[0008] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, before obtaining the coal reservoir exploration results by using the measured electromagnetic field at the corresponding surface measuring point and combining it with an electromagnetic exploration inversion model, the method further includes: determining a corresponding preset parameter range based on the geological conditions of the surface measuring point corresponding to the previously acquired electromagnetic field measurement data; randomly generating underground resistivity parameters, electric dipole depth parameters, and angle parameters based on the preset parameter range; obtaining corresponding electric dipole source simulated electromagnetic field training data based on the underground resistivity parameters, electric dipole depth parameters, and angle parameters, combined with the coal reservoir self-emission forward model; obtaining electric dipole source simulated impedance training data based on the electric dipole source simulated electromagnetic field training data; obtaining corresponding natural source simulated electromagnetic field training data based on the underground resistivity parameters and the natural source forward model; and obtaining natural source simulated impedance training data based on the natural source simulated electromagnetic field training data; and determining a corresponding intensity ratio parameter based on the electromagnetic field measurement data. The system includes preset ranges for the slope parameters of the natural source magnetic field and the phase of the natural source. Based on these preset ranges, training parameters for the intensity ratio of the characteristic frequency band, the slope parameters of the natural source magnetic field, and the magnetic field intensity ratio are obtained. Natural source phase training data is also obtained based on the preset range of the natural source phase. Hybrid impedance training data is obtained based on the simulated impedance training data of the electric dipole source, the simulated impedance training data of the natural source, and the magnetic field intensity ratio training coefficient. Simulated electromagnetic exploration training data is obtained based on the hybrid impedance training data and the natural source phase training data. Electromagnetic exploration labels are also obtained based on the training parameters for the characteristic frequency band intensity ratio, the slope parameters of the natural source magnetic field, the underground resistivity parameter, and the electric dipole depth and angle parameters. The simulated electromagnetic exploration training data is used as input data for training, and the electromagnetic exploration labels are used as training tags to train the electromagnetic exploration inversion model, resulting in the electromagnetic exploration inversion model.

[0009] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, the training of the electromagnetic exploration inversion model to be trained includes: inputting mixed impedance training data from the simulated electromagnetic exploration training data into a feature extraction layer for feature extraction to obtain mixed impedance features; inputting the mixed impedance features and natural source phase training data from the simulated electromagnetic exploration training data into a first electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain coal reservoir parameter inversion results; inputting the mixed impedance features and natural source phase training data from the simulated electromagnetic exploration training data into a second electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain subsurface resistivity distribution results; constructing a first loss function based on the coal reservoir parameter inversion results and the characteristic frequency band intensity ratio training parameters, natural source magnetic field slope training parameters, and electric dipole depth and angle parameters in the electromagnetic exploration tags; and constructing a second loss function based on the subsurface resistivity distribution results and the subsurface resistivity parameters in the electromagnetic exploration tags; obtaining a total loss function based on the first loss function and the second loss function, and converging based on the total loss function to end the training.

[0010] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, training parameters for intensity ratios in characteristic frequency bands, training parameters for slope of natural source magnetic fields, and training coefficients for magnetic field intensity ratios are obtained based on preset ranges for intensity ratio parameters and preset ranges for slope parameters of natural source magnetic fields. The method includes: obtaining training parameters for intensity ratios in characteristic frequency bands based on preset ranges for intensity ratio parameters; obtaining training parameters for slope of natural source magnetic fields based on preset ranges for slope parameters of natural source magnetic fields; obtaining training coefficients for systematic spectral changes based on a forward model simulation of the coal reservoir's self-emission; obtaining training coefficients for interference ratio changes based on Gaussian noise simulation; and obtaining training coefficients for magnetic field intensity ratios based on the training parameters for intensity ratios in characteristic frequency bands, training parameters for slope of natural source magnetic fields, training coefficients for systematic spectral changes, and training coefficients for interference ratio changes.

[0011] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, before using the measured electromagnetic field at the surface measuring point where coalbed methane exists as the inversion input data of the electromagnetic exploration inversion model, the method includes: performing static effect correction and topographic distortion correction on the measured electromagnetic field at the surface measuring point where coalbed methane exists.

[0012] According to the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention, before obtaining the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal, the method further includes: constructing a forward model of coal reservoir self-emission based on the electromagnetic field excited by electric dipoles in the underground layered medium and propagating to the surface.

[0013] This invention also provides an electromagnetic exploration device for gas-bearing coal reservoirs, comprising: a signal forward modeling module, which obtains the spectral characteristics and spatial distribution characteristics of a self-emission simulated signal based on a pre-constructed coal reservoir self-emission forward model and in conjunction with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters; wherein the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include a simulated electromagnetic field signal; and a coalbed methane judgment module, which acquires the measured electromagnetic field at a surface measuring point, compares the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emission simulated signal, and determines whether a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field exists within a frequency range of a preset peak value determined based on the spectral characteristics of the self-emission simulated signal, and, based on its existence, determines the corresponding location. The presence of coalbed methane around the surface measuring point; and in the case where coalbed methane is present around the corresponding surface measuring point, based on the spatial distribution characteristics of the self-emitted simulated signal, determining whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength, and based on this, determining that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point; the electromagnetic exploration inversion module, in the case where it is determined that coalbed methane exists within a preset range below the surface measuring point, uses the measured electromagnetic field at the corresponding surface measuring point and combines it with the electromagnetic exploration inversion model to obtain the coal reservoir exploration results; wherein, the electromagnetic exploration inversion model is obtained based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data; the electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of any of the above-described electromagnetic exploration methods for gas-bearing coal reservoirs.

[0015] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the electromagnetic exploration method for gas-bearing coal reservoirs as described above.

[0016] The present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of any of the above-described electromagnetic exploration methods for gas-bearing coal reservoirs.

[0017] The electromagnetic exploration method and apparatus for gas-bearing coal reservoirs provided by this invention obtains the spectral and spatial distribution characteristics of the self-emission simulated signal through a forward model of coal reservoir self-emission, facilitating comparison with the measured electromagnetic field. This allows for the determination of whether coalbed methane exists at the corresponding surface measuring point. Furthermore, it facilitates the inversion of electromagnetic field data measured at surface measuring points containing coalbed methane using an electromagnetic exploration inversion model. This corrects the influence of coal reservoir self-emission on the inversion of natural source resistivity (false lows, false highs, and depth offsets, etc.), enabling the accurate resistivity inversion when performing electromagnetic sounding in coalbed methane-rich areas. Based on the inversion results, the depth, structural orientation, and gas content parameters of the coal reservoir can be directly and quantitatively assessed. In addition, the inverted electric dipole depth and low-resistivity zone depth can be cross-referenced, enhancing the accuracy of coal reservoir depth prediction. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 This is one of the flowcharts of the electromagnetic exploration method for gas-bearing coal reservoirs provided by the present invention;

[0020] Figure 2 This is a schematic diagram of the coal reservoir self-emission forward model provided by the present invention;

[0021] Figure 3 This is a schematic diagram of the simulated surface-received self-emitted electric field spectrum curve of a coal reservoir and the measured electromagnetic field spectrum signal of V8, provided by the present invention.

[0022] Figure 4 The frequency location f of the V8 measured electromagnetic signal anomaly provided by this invention is... c Comparison of coalbed methane burial depth at a surface measuring point;

[0023] Figure 5 This invention provides a simulated horizontal spatial distribution map of the intensity of self-emitted electromagnetic field signals from coal reservoirs received on the surface.

[0024] Figure 6 This is a schematic diagram of the simulated impedance spectrum curve of natural surface reception and coal reservoir self-emission, and the measured impedance spectrum curve of V8 provided by the present invention.

[0025] Figure 7 The simulated k (f = 900) provided by this invention varies in the complex plane |Z xy Distribution map;

[0026] Figure 8 This is a schematic diagram of the electromagnetic exploration inversion model provided by the present invention;

[0027] Figure 9 The hybrid impedance Z provided by this invention xy Change simulation diagram;

[0028] Figure 10 This is a schematic diagram of the electromagnetic exploration device for gas-bearing coal reservoirs provided by the present invention;

[0029] Figure 11 This is a schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0031] Figure 1 A flowchart illustrating an electromagnetic exploration method for gas-bearing coal reservoirs is shown. The method includes:

[0032] S11. Based on the pre-constructed forward model of coal reservoir self-emission, and combined with preset underground resistivity parameters, preset electric dipole depth parameters and preset angle parameters, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal are obtained; among which, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include the simulated electromagnetic field signal.

[0033] S12, acquire the measured electromagnetic field at the surface measuring point, compare the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emitted simulated signal, and determine whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within the frequency range of the preset peak value determined based on the spectral characteristics of the self-emitted simulated signal. Based on its existence, determine that there is coalbed methane around the corresponding surface measuring point. In the case that there is coalbed methane around the corresponding surface measuring point, determine whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal. Based on its greater strength, determine that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point.

[0034] S13, if it is determined that there is coalbed methane within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, and combined with the electromagnetic exploration inversion model, the coal reservoir exploration results are obtained; wherein, the electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data; the electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0035] It should be noted that S1N in this specification does not represent the chronological order of electromagnetic exploration methods for gas-bearing coal reservoirs. The following details will explain this in conjunction with... Figures 2-9 The present invention describes an electromagnetic exploration method for gas-bearing coal reservoirs.

[0036] Step S11: Based on the pre-constructed forward model of coal reservoir self-emission, and combined with preset underground resistivity parameters, preset electric dipole depth parameters and preset angle parameters, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal are obtained; wherein, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include the simulated electromagnetic field signal.

[0037] In this embodiment, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal are obtained, including: the coal reservoir self-emission forward model can use the Hankel filtering algorithm, combined with preset underground resistivity parameters, preset electric dipole depth parameters and preset angle parameters, to obtain the simulated electromagnetic field signal; based on the simulated electromagnetic field signal, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal are obtained.

[0038] It should be noted that the simulated electromagnetic field signal includes the electromagnetic field of the horizontal electric dipole located in layer i at any layer j and the electromagnetic field of the vertical electric dipole located in layer i at any layer j. Specifically:

[0039] The electromagnetic field of a horizontal electric dipole located in the i-th layer at any layer j is expressed as:

[0040]

[0041]

[0042]

[0043]

[0044]

[0045] Where IdL represents the electric dipole moment intensity; The electric dipole moment azimuth angle is represented by z; z represents the detector's z-axis coordinate; z s The z-axis coordinate represents the position of the center of the electric dipole; z jThis represents the z-axis coordinate at the interface of the j-th layer; z0 = -∞, z1 = 0, z n+1 =∞; J n (rm) represents the nth-order Bessel function; m represents the spatial frequency with the reciprocal dimension of distance. In practical applications, m is an intermediate parameter for the integral of the Bessel function. ω represents the propagation coefficient, which in practical applications is... 2 με can be ignored; when j = i, δ ji =1, otherwise, δ ji =0; a j b j c j d j p j q j The coefficients can be derived recursively from the boundary conditions of each layer in practical applications.

[0046] Furthermore, the electromagnetic field of a vertical electric dipole located in the i-th layer at any layer j can be expressed as:

[0047]

[0048]

[0049]

[0050]

[0051] It should be noted that coal reservoir modeling primarily uses horizontal electric dipoles, but actual coal reservoirs have a certain degree of inclination, and under stress conditions, they will also generate a certain vertical electric dipole component. Therefore, the dip angle is taken into account when modeling the electric dipoles. Introduction Essentially, it involves superimposing the electromagnetic fields generated by a horizontal electric dipole and a vertical electric dipole in a certain proportion.

[0052] In addition, the preset angle parameters include the preset azimuth angle. and preset tilt angle

[0053] In one optional embodiment, before obtaining the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal, the method further includes: constructing a forward model of coal reservoir self-emission based on the electromagnetic field excited by electric dipoles in the underground layered medium and propagating to the surface.

[0054] In one alternative embodiment, V8 / AMT is used for natural source electromagnetic exploration in the coalbed methane-rich area of ​​Qinshui County. V8 / AMT can measure E... x E y H x Hy H z Five-component electromagnetic field 1-10 4 Hz spectral signal. The constructed coal reservoir self-emission forward model is as follows: Figure 2 As shown, the electric dipole is located in the middle thin layer. The simulation results for different depths of the electric dipole are as follows. Figure 3 As shown in a, 10 3 An electromagnetic signal anomaly was observed near Hz, consistent with the measured electromagnetic signal anomaly at V8 in the study area. Figure 3 b matches. Simultaneously, the simulated electromagnetic signal anomaly frequency position f... c The depth decreases with increasing electric dipole depth (e.g.) Figure 3 a) This is consistent with the measured data of V8 in the study area, corresponding to Figure 4 In the shallower coalbed methane area, the pentagram in survey line 1 is darker and the lines are thicker, while in the deeper coalbed methane area, the pentagram in survey line 2 is lighter and the lines are thinner. On the other hand, the simulated distribution of surface electromagnetic field signal intensity with the horizontal position of the surface survey points is shown in the following diagram. Figure 5 As shown, the distribution of the horizontal and vertical components of the electromagnetic field differs.

[0055] Step S12: Obtain the measured electromagnetic field at the surface measuring point, compare the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emitted simulated signal, and determine whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within the frequency range of the preset peak value determined based on the spectral characteristics of the self-emitted simulated signal. Based on its existence, determine that there is coalbed methane around the corresponding surface measuring point. In the case that there is coalbed methane around the corresponding surface measuring point, determine whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal. Based on its greater strength, determine that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point.

[0056] In this embodiment, the measured electromagnetic field is a spectrum curve composed of a mixture of natural source electromagnetic fields from the ionosphere and lightning, and a self-emitted source electromagnetic field from the coal reservoir. The measured electromagnetic field includes both amplitude and phase spectrum curves. It should be noted that the frequency range of the preset peak value is mainly affected by the coal reservoir depth, coalbed methane content, and underground resistivity, and can be determined based on the spectral characteristics and spatial distribution characteristics of the self-emitted simulated signal obtained from the simulated electromagnetic field signal. Furthermore, the preset range below can be determined based on the location directly below and around the surface measuring point within a certain range, and can be based on prior experience or experimental data; no further limitations are imposed here.

[0057] Furthermore, within the frequency range of a preset peak value determined based on the spectral characteristics of the self-emitted analog signal, it is determined whether there exists an analog electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field. This includes: determining the frequency range of the preset peak value based on the spectral characteristics of the self-emitted analog signal; and within the frequency range of the preset peak value, determining whether the analog electromagnetic field signal is greater than or equal to a preset proportion of the measured electromagnetic field. It should be noted that the preset proportion can be determined based on the actual coal reservoir depth and coalbed methane content.

[0058] For example, refer to Figure 3 By simulating electromagnetic field signals at a preset peak value of 10 3 A strong electromagnetic signal anomaly, exceeding the measured electromagnetic field by approximately 50%, was observed near the Hz level, indicating the presence of a coalbed methane center near the surface measuring point. Based on... Figure 5 It can be seen that for 10 3 Abnormal signals near Hz indicate that when the horizontal electromagnetic field is strong and the vertical electromagnetic field is weak during field detection, it means that the center of the coalbed methane is basically directly below the surface measuring point. Conversely, it means that the horizontal position of the surface measuring point is still offset from the center of the coalbed methane.

[0059] In one possible implementation, after determining whether there is a simulated electromagnetic field signal exceeding a preset ratio of the measured electromagnetic field within the frequency range of the preset peak value determined based on the spectral characteristics of the self-emitted simulated signal, the following steps are taken: if it is determined that there is coalbed methane around the surface measuring point, determine whether the amplitude of the horizontal electromagnetic field is greater than at least one times the amplitude of the vertical electromagnetic field. If it is greater, it indicates that the amplitude of the horizontal electromagnetic field is abnormally strong while the amplitude of the vertical electromagnetic field is abnormally weak, which corresponds to determining that the center of the coalbed methane is basically located directly below the surface measuring point.

[0060] In another possible implementation, after determining whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within the frequency range of the preset peak value determined based on the spectral characteristics of the self-emitted simulated signal, the method further includes: if coalbed methane is determined to exist around the surface measuring point, determining whether the horizontal electromagnetic field amplitude exceeds a first preset threshold and whether the vertical electromagnetic field amplitude exceeds a second preset threshold. If both exceed the thresholds, it indicates that the horizontal electromagnetic field amplitude is abnormally strong while the vertical electromagnetic field amplitude is abnormally weak, which corresponds to determining that the center of the coalbed methane is basically located directly below the surface measuring point. It should be noted that the first and second preset thresholds can be determined based on actual empirical data, and are not further limited here.

[0061] Step S13: If coalbed methane is found within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with the electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0062] In this embodiment, when it is determined that there is coalbed methane within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, and combined with the electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. This includes: using the measured electromagnetic field at the surface measuring point where coalbed methane exists as the inversion input data of the electromagnetic exploration inversion model, and obtaining the coal reservoir exploration results output by the electromagnetic exploration inversion model.

[0063] In an optional embodiment, before obtaining the coal reservoir exploration results by using the measured electromagnetic field at the corresponding surface measuring point and combining it with the electromagnetic exploration inversion model, the method further includes: training the electromagnetic exploration inversion model, provided that coalbed methane exists within a preset range below the surface measuring point.

[0064] Specifically, training the electromagnetic exploration inversion model includes:

[0065] S131. Based on the geological conditions of the corresponding surface measuring points corresponding to the previously acquired electromagnetic field measurement data, determine the corresponding preset parameter range, and randomly generate underground resistivity parameters, electric dipole depth parameters, and angle parameters according to the preset parameter range. The electromagnetic field measurement data includes hybrid electromagnetic field data that combines natural source electromagnetic data, electric dipole source electromagnetic data, and natural source phase data.

[0066] It should be added that the angle parameters include the azimuth angle. and tilt angle Furthermore, in order to facilitate the subsequent simulation of the natural source impedance using the natural source forward model and the electric dipole source impedance using the coal reservoir self-emission forward model, it is necessary to first generate the subsurface resistivity parameters, as well as the electric dipole depth and angle parameters.

[0067] The preset parameter range can be determined based on prior experience of the actual survey location, for example, by randomly generating 8 strata with variable thickness and resistivity as shown in Table 1 below. This is then mapped to 27 thinner strata of fixed thickness to facilitate inversion. Since the coalbed methane depth in Qinshui County ranges from 300m to 700m, the depths of the 27 thin strata are set as follows: the first 6 layers at 50m, the middle 16 layers at 25m, the subsequent 4 layers at 100m, and the final base layer at 99999m, to ensure a resolution shallower than 800m; electric dipole depth z s Azimuth inclination Random values ​​were taken from [200m, 800m], [-180°-180°], and [-80°-80°].

[0068] Table 1 Each layer Range of values

[0069]

[0070] S132. Based on the underground resistivity parameters, electric dipole depth parameters, and angle parameters, and combined with the coal reservoir self-emission forward model, the corresponding electric dipole source simulated electromagnetic field training data is obtained. Based on the electric dipole source simulated electromagnetic field training data, the electric dipole source simulated impedance training data is obtained.

[0071] In this embodiment, the simulated impedance training data of the electric dipole source is represented as follows:

[0072]

[0073] in, This represents the simulated impedance training data of the electric dipole source. This represents the electric field in the training data for simulating electromagnetic fields using an electric dipole source. This represents the magnetic field in the training data for simulating electromagnetic fields using an electric dipole source.

[0074] S133. Based on the underground resistivity parameters and the forward model of the natural source, the corresponding training data of the simulated electromagnetic field of the natural source is obtained, and the training data of the simulated impedance of the natural source is obtained based on the training data of the simulated electromagnetic field of the natural source.

[0075] It should be noted that before obtaining the corresponding natural source electromagnetic field training data, including: constructing a natural source forward model based on electromagnetic radiation from the ionosphere, lightning, and other sources at a distance above the Earth's surface.

[0076] In addition, the training data for the simulated impedance of natural sources are expressed as follows:

[0077]

[0078] in, This represents training data for simulated impedance from natural sources; This represents the electric field in the training data for simulating electromagnetic fields from natural sources. This represents the magnetic field in the training data for simulating electromagnetic fields from natural sources.

[0079] S134. Based on the measured electromagnetic field data, determine the preset range of the corresponding intensity ratio parameter, the preset range of the natural source magnetic field slope parameter, and the preset range of the natural source phase. Based on the preset range of the intensity ratio parameter and the preset range of the natural source magnetic field slope parameter, obtain the characteristic frequency band intensity ratio training parameter, the natural source magnetic field slope training parameter, and the magnetic field intensity ratio training coefficient. Based on the preset range of the natural source phase, obtain the natural source phase training data.

[0080] Specifically, based on the preset ranges of the intensity ratio parameter and the natural source magnetic field slope parameter, training parameters for the intensity ratio of the characteristic frequency band, training parameters for the slope of the natural source magnetic field, and training coefficients for the magnetic field intensity ratio are obtained, including: obtaining training parameters for the intensity ratio of the characteristic frequency band based on the preset range of the intensity ratio parameter; obtaining training parameters for the slope of the natural source magnetic field based on the preset range of the natural source magnetic field parameter; obtaining training coefficients for systematic spectral changes based on the forward model simulation of coal reservoir self-emission; obtaining training coefficients for interference ratio changes based on Gaussian noise simulation; and obtaining training coefficients for the magnetic field intensity ratio based on the training parameters for the intensity ratio of the characteristic frequency band, the training parameters for the slope of the natural source magnetic field, the training coefficients for systematic spectral changes, and the training coefficients for interference ratio changes.

[0081] It should be added that the characteristic frequency band intensity ratio training parameter is closely related to the coalbed methane content and is used to characterize the ratio of the amplitude intensity of the electric dipole source magnetic field to that of the natural source magnetic field; the natural source magnetic field slope training parameter is used to characterize the systematic spectral changes of the natural source magnetic field; the systematic spectral change training coefficient is used to characterize the systematic spectral changes of the electric dipole source magnetic field; and the interference ratio change training coefficient is used to characterize the ratio changes caused by unknown micro-perturbations of the signal source.

[0082] In addition, the training parameters for the slope of the natural source magnetic field are expressed as follows:

[0083]

[0084] Based on statistics of measured magnetic field signals, k2 is approximated by a double logarithmic linear relationship, and the phase... Take 10 2 -10 4 Hz mean, natural source phase data Depending on the polarization direction, there are The double logarithmic linear approximation is as follows:

[0085] log10(|H p |)=k2 log 10(f)+b

[0086] Where k2 represents the slope training parameter of the natural source magnetic field; f represents the frequency; and b represents the intercept. This represents the natural source phase data in the x and y directions, which can be any two perpendicular directions, to facilitate the determination of the azimuth angle. The included angle.

[0087] Optionally, the x-direction can be due north underground; or, the x-direction can be the direction of a mountain range. Correspondingly, the y-direction is perpendicular to the x-direction.

[0088] In an optional embodiment, the training coefficient for the ratio of the magnetic field strength of the natural source to that of the electric dipole source is expressed as:

[0089]

[0090] Where k(f) represents the training coefficient for the ratio of magnetic field strength between the natural source and the electric dipole source; This represents the magnetic field data in the simulated electromagnetic field data of an electric dipole source. k1 represents the magnetic field data in the simulated electromagnetic field data of the natural source; k2(f) represents the training parameter of the intensity ratio of the characteristic frequency band; k3(f) represents the training parameter of the slope of the magnetic field of the natural source; k4(f) represents the training coefficient of the systematic spectrum change; f represents the training coefficient of the interference ratio change; f represents the frequency.

[0091] It should be noted that when f = 1000Hz,

[0092] Furthermore, Z is obtained by superimposing k(f) at different times. xy The differences are significant, such as Figure 6 As shown. It can be seen that even when |k(f)| is small, the Z-axis under disturbance... xy It also deviates significantly from the impedance of a pure natural source plane wave in the 100Hz-6000Hz range, i.e., the natural source simulated impedance training data. Furthermore, the simultaneous existence of both overestimation and underestimation significantly impacts the natural source resistivity inversion results. When k(f) is large, the mixed impedance training data Z... xy Training data with natural simulated impedance Completely different, such data cannot be processed according to classical magnetotelluric theory. In the V8 measured data, both small and large k(f) exist at surface measuring points, and in most surface measuring points, k1 (i.e., |k(f)|) exceeds 0.1 (but does not completely mask the natural source signal). It is evident that it is indeed necessary to establish a natural source inversion method that considers the self-emission of coal reservoirs to supplement and correct classical magnetotelluric theory.

[0093] It is evident that constructing the magnetic field strength proportional training coefficient k(f) is crucial for building hybrid impedance data. After simplification, k(f) is composed of k1, k2, and... The decision, and k1, k2, The intensity ratio parameter and the natural source magnetic field slope parameter can be preset within a certain range, and values ​​can be randomly selected to generate n sets of k(f).

[0094] For example, let's simulate the impedance |Z| at f = 900 Hz using an underground resistivity of 50 Ω and an electric dipole depth of 550 m. xy |As k(f) changes, such as Figure 7 As shown. Based on the statistics of the measured data of V8, k1 (i.e., |k(f)|) roughly falls within Figure 7 Within the circle, considering the changes in resistivity and electric dipole depth, the range of values ​​can be slightly relaxed to [0 3]. In the measured signal of V8 in the study area, the magnetic field component is much more stable than the electric field, with fewer fluctuations. After removing power frequency interference and 10 3 After observing electromagnetic anomalies near Hz, taking a double logarithm yields an average linear correlation coefficient r of -0.96 for all data, indicating a very strong linear relationship. Therefore, simplifying |k2(f)| to k2 using a double logarithmic linear approximation is reasonable. In 10 2 -10 4 The values ​​are very stable between Hz, so simplifying them to the mean is reasonable. k2, It can take values ​​in the intervals [-0.6, 0], [-π, π], and [-π, π] respectively.

[0095] S135. Based on the simulated impedance training data of electric dipole source, the simulated impedance training data of natural source, and the magnetic field strength ratio training coefficient, mixed impedance training data is obtained.

[0096] It should be noted that the mixed impedance training data includes mixed data of first-direction impedance and mixed data of second-direction impedance, where the first direction represents the direction from the x-direction to the y-direction, and the second direction represents the direction from the y-direction to the x-direction.

[0097] Specifically, the first directional impedance mixed data is expressed as follows:

[0098]

[0099] Among them, Z xy This represents mixed impedance data in the first direction; E x H represents the electric field in the x-direction; y This represents the y-axis magnetic field data; This represents the electric field in simulated electromagnetic field data from natural sources; This represents the magnetic field in simulated electromagnetic field data from natural sources; This represents the electric field in the simulated electromagnetic field data of an electric dipole source. This represents the magnetic field in the simulated electromagnetic field data of the electric dipole source. This represents training data for simulated impedance from natural sources; represents the simulated impedance training data of the electric dipole source; k(f) represents the magnetic field strength proportional training coefficient.

[0100] Similarly, the second-direction impedance mixed data Z yx You can refer to the above Z. xy This will not be repeated here.

[0101] In the actual training process, the simulated impedance training data from the electric dipole source and the simulated impedance training data from the natural source were superimposed with training coefficients proportional to the magnetic field strength to obtain mixed impedance training data, which was used to generate 500,000 sets of mixed impedance training data. This process took 16 hours on a 6-core Intel(R) Core(TM) i7-8700K CPU in parallel. With better hardware, more datasets can be generated, a larger network can be used, and the range of parameter values ​​can be appropriately widened.

[0102] S136. Based on the mixed impedance training data and the natural source phase training data, simulated electromagnetic exploration training data is obtained. Based on the characteristic frequency band intensity ratio training parameters, the natural source magnetic field slope training parameters, the underground resistivity parameters, and the electric dipole depth and angle parameters, electromagnetic exploration tags are obtained.

[0103] S137, Using simulated electromagnetic exploration training data as input data for training and electromagnetic exploration labels as labels for training, the electromagnetic exploration inversion model to be trained is trained to obtain the electromagnetic exploration inversion model.

[0104] Specifically, training the electromagnetic exploration inversion model to be trained includes: inputting the mixed impedance training data from the simulated electromagnetic exploration training data into the feature extraction layer for feature extraction to obtain mixed impedance features; inputting the mixed impedance features and the natural source phase training data from the simulated electromagnetic exploration training data into the first electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain coal reservoir parameter inversion results; inputting the mixed impedance features and the natural source phase training data from the simulated electromagnetic exploration training data into the second electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain subsurface resistivity distribution results; constructing a first loss function based on the coal reservoir parameter inversion results and the characteristic frequency band intensity ratio training parameters, natural source magnetic field slope training parameters, and electric dipole depth and angle parameters in the electromagnetic exploration tags; and constructing a second loss function based on the subsurface resistivity distribution results and the subsurface resistivity parameters in the electromagnetic exploration tags; obtaining the total loss function based on the first and second loss functions, and converging based on the total loss function to end the training.

[0105] In this embodiment, the coal reservoir parameter inversion results include the inverted electric dipole depth z. s Inversion azimuth angle and inversion dip angle The results include the inversion intensity ratio parameter k1 and the inversion natural source magnetic field slope parameter k2, and the underground resistivity distribution results include resistivity parameters of 27 layers.

[0106] In an optional embodiment, the electromagnetic exploration inversion model can be improved based on the neural network to obtain a task-specific network. The last quarter convolutional layer and the fully connected layer are split into two task-specific networks to address the problem of outliers caused by the different sensitivities of the coal reservoir parameter inversion results and the underground resistivity distribution results to the input data. It should be noted that the neural network can be selected according to the actual design requirements, such as the residual neural network ResNet18.

[0107] like Figure 8 As shown, based on the data, input layer 1 can use a 4-component impedance spectrum curve (Z). xy Z yx (Real part, imaginary part) or 2-component impedance spectrum curve (|Z) xy |、|Z yx |), Input layer 2 is based on statistics from V8 data. Task 1 Output the inverted electric dipole depth z s Inversion azimuth angle Inversion dip angle Along with the inversion intensity ratio parameter k1 and the inversion natural source magnetic field slope parameter k2, there are a total of 5 parameters. Task 2 outputs 27 layers of resistivity parameters.

[0108] In one alternative embodiment, topographic distortion and static effects typically lead to an overall shift in resistivity inversion results. While this may not affect the final conclusions of some exploration tasks that only concern the relative magnitude of underground resistivity, for the inversion task presented in this paper, due to the coupling between electric dipole parameters and resistivity, topographic distortion and static effects can cause an overall deviation in the inverted electric dipole parameters. Therefore, before using the measured electromagnetic field at surface measuring points containing coalbed methane as the inversion input data for the electromagnetic exploration inversion model, the following steps are taken: static effect correction and topographic distortion correction are performed on the measured electromagnetic field at surface measuring points containing coalbed methane to address the inversion shift of electric dipole parameters caused by the deep coupling effect of resistivity-electric dipole.

[0109] It should be noted that, for reference Figure 9 In fixed Mixed impedance Z when ρ1 is changed xy As can be seen from the simulation diagram, if a certain interference causes the overall impedance of the measured data to be higher (or lower), but does not change the frequency position f of the abnormal measured data signal, c Therefore, the inverted coal reservoir exploration results may misidentify resistivity as too high (or too low), and may also misinterpret z. sMisidentification of depth (or shallowness) is due to the resistivity-dipole depth coupling (R-DDC) effect. Therefore, V8 data must first undergo static effect and topographic distortion correction before being input into a trained electromagnetic exploration inversion model for inversion. Ultimately, the depth of the low-resistivity layer and the electric dipole depth obtained from the inversion can be used to determine the depth of the gas-bearing coal reservoir. That is, the structural orientation of the coal reservoir (the direction of stress on the coal reservoir); and the electric dipole moment intensity can be calculated based on k1, and the gas content can be further estimated by combining the electromechanical coupling model.

[0110] It should be added that static effect correction includes impedance tensor correction, electromagnetic array profiling, transient electromagnetic approximation correction, and phase-assisted correction. For example, when using the impedance tensor correction method, regardless of the complexity of the underground structure, the impedance tensor Z is affected by the shallow penetration of high-frequency electromagnetic waves. h Theoretically, it should have a one-dimensional form, that is If the actual observed high-frequency impedance tensor Z′ h If the 1-dimensional characteristic is not satisfied, the correction coefficient matrix C can be calculated from the difference between theoretical and measured values. Furthermore, due to the consistent influence of static effects across all frequency bands (overall shift in the spectral curve), the correction matrix C can be used for all frequencies. The correction coefficient matrix C is expressed as:

[0111]

[0112] In addition, topographic distortion correction mainly adopts the ratio method, that is, first assuming that the underground resistivity is one-dimensionally uniform, and then obtaining the apparent resistivity ρ with topography. at (f,x) and apparent resistivity ρ without topography a0 The two are then compared to obtain the correction coefficient matrix d(f,x) for pure terrain; then the correction coefficient matrix d(f,x) is multiplied by the measured apparent resistivity ρ. aD (f,x) yields the corrected apparent resistivity ρ aN (f,x), and also corrects phase distortion (similarly, impedance can be corrected).

[0113] It should be added that the correction coefficient matrix d(f,x) is expressed as:

[0114]

[0115] Where d(f,x) represents the correction coefficient matrix; ρ at (f,x) represents the apparent resistivity with topographical view, i.e., the electromagnetic field data obtained at the corresponding surface measurement point; ρ a0 This represents the apparent resistivity without considering topography. It should be noted that ρ... at ρa0 d(f,x) is mainly obtained using 2D and 3D finite element forward modeling.

[0116] In addition, the corrected ρ aN (f,x) is represented as:

[0117]

[0118]

[0119] Where, ρ aN (f,x) represents the corrected apparent resistivity; ρ aD (f,x) represents the apparent resistivity in the electromagnetic field data at the corresponding ground surface measurement point; d(f,x) represents the correction coefficient matrix. Indicates the corrected phase; This indicates the phase in the electromagnetic field data obtained at the corresponding ground surface measuring point; This indicates a phase with terrain features.

[0120] In summary, this invention uses a forward model to obtain the spectral and spatial distribution characteristics of the self-emission simulated signal from a coal reservoir. This allows for comparison with measured electromagnetic fields, thereby determining whether coalbed methane exists at the corresponding surface measurement points. Furthermore, it facilitates the inversion of electromagnetic field data measured at surface measurement points containing coalbed methane using an electromagnetic exploration inversion model. This corrects the influence of coal reservoir self-emission on the inversion of natural source resistivity (false lows, false highs, and depth offsets, etc.). It enables the accurate resistivity inversion when performing electromagnetic sounding in coalbed methane-rich areas. Based on the inversion results, the depth, structural orientation, and gas content parameters of the coal reservoir can be directly and quantitatively assessed. Additionally, the inverted electric dipole depth and low-resistivity zone depth can be cross-referenced, enhancing the accuracy of coal reservoir depth prediction.

[0121] The electromagnetic exploration device for gas-bearing coal reservoirs provided by the present invention is described below. The electromagnetic exploration device for gas-bearing coal reservoirs described below can be referred to in correspondence with the electromagnetic exploration method for gas-bearing coal reservoirs described above.

[0122] Figure 10 A schematic diagram of an electromagnetic exploration device for gas-bearing coal reservoirs is shown. The device includes:

[0123] The signal forward modeling module 101, based on a pre-constructed coal reservoir self-emission forward modeling model and combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters, obtains the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal; wherein, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include the simulated electromagnetic field signal.

[0124] The coalbed methane determination module 102 acquires the measured electromagnetic field at the surface measuring point, compares the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emitted simulated signal, and determines whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within a frequency range of a preset peak value determined based on the spectral characteristics of the self-emitted simulated signal. Based on its existence, it determines that coalbed methane exists around the corresponding surface measuring point. Furthermore, if coalbed methane exists around the corresponding surface measuring point, it determines whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal, and based on its greater strength, determines that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point.

[0125] The electromagnetic exploration inversion module 103, when it is determined that there is coalbed methane within a preset range below the surface measuring point, uses the measured electromagnetic field at the corresponding surface measuring point and combines it with the electromagnetic exploration inversion model to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0126] In this embodiment, the signal forward modeling module 101 includes: a signal simulation unit, which can use the Hankel filtering algorithm and combine preset underground resistivity parameters, preset electric dipole depth parameters and preset angle parameters to obtain a simulated electromagnetic field signal; and a feature acquisition unit, which obtains the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal based on the simulated electromagnetic field signal.

[0127] In an optional embodiment, the device further includes: a first forward model building module, which, before obtaining the spectral characteristics and spatial distribution characteristics of the self-emission simulation signal, builds a forward model of the coal reservoir self-emission based on the electromagnetic field excited by electric dipoles in the underground layered medium and propagating to the surface.

[0128] The coalbed methane determination module 102 includes: a first determination unit, which acquires the measured electromagnetic field at a surface measuring point, compares the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of a self-emitted analog signal, and determines whether there is an analog electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within a frequency range determined based on the spectral characteristics of the self-emitted analog signal, and determines, based on its existence, that coalbed methane exists around the corresponding surface measuring point; and a second determination unit, which, when coalbed methane exists around the corresponding surface measuring point, determines, based on the spatial distribution characteristics of the self-emitted analog signal, whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength, and determines, based on its greater strength, that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point.

[0129] Furthermore, the first judgment unit includes: a range determination subunit, which determines the frequency range of the preset peak value based on the spectral characteristics of the self-emitted analog signal; and a first judgment subunit, which determines whether the analog electromagnetic field signal is greater than or equal to a preset proportion of the measured electromagnetic field within the frequency range of the preset peak value.

[0130] In one possible implementation, the second judgment unit further includes: a measurement point location determination subunit, which, within the frequency range of the preset peak value determined based on the spectral characteristics of the self-emitted analog signal, determines whether there is an analog electromagnetic field signal exceeding the preset ratio of the measured electromagnetic field, and, if it is determined that there is coalbed methane at the surface measurement point, determines whether the amplitude of the horizontal electromagnetic field is greater than at least one times the amplitude of the vertical electromagnetic field. If it is greater, it indicates that the amplitude of the horizontal electromagnetic field is abnormally strong while the amplitude of the vertical electromagnetic field is abnormally weak, and it can be determined that the center of the coalbed methane is basically located directly below the surface measurement point.

[0131] In another possible implementation, the second judgment unit further includes a measurement point location determination subunit. This subunit, within a frequency range determined based on the spectral characteristics of the self-emitted analog signal and a preset peak value, determines whether there exists an analog electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field. If coalbed methane is confirmed to exist around the surface measurement point, it then determines whether the horizontal electromagnetic field amplitude exceeds a first preset threshold and whether the vertical electromagnetic field amplitude exceeds a second preset threshold. If both exceed the thresholds, it indicates that the horizontal electromagnetic field amplitude is abnormally strong while the vertical electromagnetic field amplitude is abnormally weak, correspondingly determining that the center of the coalbed methane is basically located directly below the surface measurement point. It should be noted that the first and second preset thresholds can be determined based on actual empirical data and are not further limited here.

[0132] The electromagnetic exploration inversion module 103 includes: using the measured electromagnetic field at the surface measuring point where coalbed methane exists as the inversion input data of the electromagnetic exploration inversion model to obtain the coal reservoir exploration results output by the electromagnetic exploration inversion model; wherein, the measured electromagnetic field is a spectrum curve composed of the electromagnetic field from the ionosphere and the natural source electromagnetic field from lightning and the electromagnetic field emitted by the coal reservoir, and the measured electromagnetic field includes the amplitude spectrum curve and the phase spectrum curve.

[0133] In an optional embodiment, the device further includes a training module, which, when it is determined that there is coalbed methane within a preset range below the surface measuring point, trains the electromagnetic exploration inversion model before obtaining the coal reservoir exploration results by using the measured electromagnetic field at the corresponding surface measuring point and combining it with the electromagnetic exploration inversion model.

[0134] Specifically, the training module includes: a parameter determination unit, which determines the corresponding preset parameter range based on the geological conditions of the surface measuring points corresponding to the previously acquired electromagnetic field measurement data, and randomly generates underground resistivity parameters, electric dipole depth parameters, and angle parameters based on the preset parameter range; an electric dipole source data acquisition unit, which obtains corresponding electric dipole source simulated electromagnetic field training data based on the underground resistivity parameters, electric dipole depth parameters, and angle parameters, combined with the coal reservoir self-emission forward model, and obtains electric dipole source simulated impedance training data based on the electric dipole source simulated electromagnetic field training data; a natural source data acquisition unit, which obtains corresponding natural source simulated electromagnetic field training data based on the underground resistivity parameters, combined with the natural source forward model, and obtains natural source simulated impedance training data based on the natural source simulated electromagnetic field training data; and a range determination unit, which determines the preset range of the corresponding intensity ratio parameter, the preset range of the natural source magnetic field slope parameter, and the preset range of the natural source phase based on the electromagnetic field measurement data; data The acquisition unit obtains training parameters for the intensity ratio of the characteristic frequency band, the slope of the natural source magnetic field, and the magnetic field intensity ratio based on preset ranges for the intensity ratio parameters and the slope parameters of the natural source magnetic field. It also obtains natural source phase training data based on preset ranges for the natural source phase. The impedance data acquisition unit obtains mixed impedance training data based on the simulated impedance training data of the electric dipole source, the simulated impedance training data of the natural source, and the magnetic field intensity ratio training coefficients. The training data acquisition unit obtains simulated electromagnetic exploration training data based on the mixed impedance training data and the natural source phase training data. It also obtains electromagnetic exploration labels based on the intensity ratio training parameters of the characteristic frequency band, the slope parameters of the natural source magnetic field, the subsurface resistivity parameters, and the depth and angle parameters of the electric dipole. The training unit uses the simulated electromagnetic exploration training data as input data and the electromagnetic exploration labels as training labels to train the electromagnetic exploration inversion model, thus obtaining the electromagnetic exploration inversion model.

[0135] Further, the data acquisition unit includes: a first parameter acquisition subunit, which obtains training parameters for the intensity ratio of the characteristic frequency band based on a preset range of the intensity ratio parameter; a second parameter acquisition subunit, which obtains training parameters for the slope of the natural source magnetic field based on a preset range of the slope parameter of the natural source magnetic field; a first coefficient acquisition subunit, which obtains training coefficients for systematic spectral changes based on a forward model simulation of coal reservoir self-emission; a second coefficient acquisition subunit, which obtains training coefficients for interference ratio changes based on Gaussian noise simulation; and a second data acquisition subunit, which obtains training coefficients for magnetic field intensity ratio based on the training parameters for the intensity ratio of the characteristic frequency band, the training parameters for the slope of the natural source magnetic field, the training coefficients for systematic spectral changes, and the training coefficients for interference ratio changes.

[0136] Further, the training unit includes: a feature extraction subunit, which inputs the mixed impedance training data from the simulated electromagnetic exploration training data into the feature extraction layer for feature extraction to obtain mixed impedance features; a first inversion subunit, which inputs the mixed impedance features and the natural source phase training data from the simulated electromagnetic exploration training data into the first electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain coal reservoir parameter inversion results; a second inversion subunit, which inputs the mixed impedance features and the natural source phase training data from the simulated electromagnetic exploration training data into the second electromagnetic exploration inversion layer for electromagnetic exploration inversion to obtain subsurface resistivity distribution results; a loss function construction subunit, which constructs a first loss function based on the coal reservoir parameter inversion results and the characteristic frequency band intensity ratio training parameters, natural source magnetic field slope training parameters, and electric dipole depth and angle parameters in the electromagnetic exploration tags, and constructs a second loss function based on the subsurface resistivity distribution results and the subsurface resistivity parameters in the electromagnetic exploration tags; and a training subunit, which obtains a total loss function based on the first and second loss functions, and terminates training upon convergence of the total loss function.

[0137] In an optional embodiment, the electromagnetic exploration inversion model can be improved based on the neural network to obtain a task-specific network. The last quarter convolutional layer and the fully connected layer are split into two task-specific networks to address the problem of outliers caused by the different sensitivities of the coal reservoir parameter inversion results and the underground resistivity distribution results to the input data. It should be noted that the neural network can be selected according to the actual design requirements, such as the residual neural network ResNet18.

[0138] In an optional embodiment, the device further includes: a second forward model building module, which, before obtaining the corresponding natural source simulated electromagnetic exploration training data, builds a natural source forward model based on electromagnetic radiation from the ionosphere, lightning, and other sources at a distance above the Earth's surface.

[0139] In an optional embodiment, the electromagnetic exploration inversion module 103 further includes: a correction unit, which measures the electromagnetic field at the surface measuring point where coalbed methane exists before using it as the inversion input data of the electromagnetic exploration inversion model, and performs static effect correction and topographic distortion correction on the measured electromagnetic field at the surface measuring point where coalbed methane exists, so as to solve the problem of electric dipole parameter inversion offset caused by resistivity-electric dipole depth coupling effect.

[0140] In summary, this embodiment of the invention uses a signal forward modeling module to obtain the spectral and spatial distribution characteristics of the self-emission simulated signal based on the coal reservoir self-emission forward modeling model. This allows the coalbed methane judgment module to compare it with the measured electromagnetic field, thereby determining whether coalbed methane exists at the surface measuring point corresponding to the measured electromagnetic field. Furthermore, it facilitates the electromagnetic exploration inversion module to invert the electromagnetic field data measured at surface measuring points containing coalbed methane using the electromagnetic exploration inversion model. This corrects the influence of coal reservoir self-emission on the inversion of natural source resistivity (false lows, false highs, and depth offsets, etc.), enabling the correct resistivity to be obtained when performing electromagnetic sounding in coalbed methane-rich areas. Based on the inversion results, the depth, structural orientation, and gas content parameters of the coal reservoir can be directly and quantitatively assessed. Additionally, the inverted electric dipole depth and low-resistivity zone depth can be cross-referenced, enhancing the accuracy of coal reservoir depth prediction.

[0141] Figure 11 An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 11 As shown, the electronic device may include: a processor 111, a communication interface 112, a memory 113, and a communication bus 114. The processor 111, communication interface 112, and memory 113 communicate with each other via the communication bus 114. The processor 111 can call logical instructions in the memory 113 to execute an electromagnetic exploration method for gas-bearing coal reservoirs. This method includes: obtaining the spectral characteristics and spatial distribution characteristics of a self-emission simulated signal based on a pre-constructed forward model of the coal reservoir self-emission, combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters; wherein the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include a simulated electromagnetic field signal; acquiring the measured electromagnetic field at a surface measuring point, comparing the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emission simulated signal, and determining whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within a preset peak frequency range determined based on the spectral characteristics of the self-emission simulated signal, and based on... Based on the existence of coalbed methane, it is determined that coalbed methane exists around the corresponding surface measuring point. Furthermore, given the presence of coalbed methane around the corresponding surface measuring point, the spatial distribution characteristics of the self-emitted simulated signal are used to determine whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength. Based on this greater strength, it is determined that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point. If coalbed methane is determined to exist within the preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with the electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the corresponding electromagnetic exploration labels. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0142] Furthermore, the logical instructions in the aforementioned memory 113 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0143] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the electromagnetic exploration method for gas-bearing coal reservoirs provided by the above methods. The method includes: obtaining the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal based on a pre-constructed coal reservoir self-emission forward model and in combination with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters; wherein the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include a simulated electromagnetic field signal; acquiring the measured electromagnetic field at a surface measuring point, comparing the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emission simulated signal, and determining the frequency of the self-emission simulated signal based on the spectral characteristics of the self-emission simulated signal. Within the specified range, it is determined whether there is a simulated electromagnetic field signal exceeding the preset ratio of the measured electromagnetic field, and based on its existence, it is determined that coalbed methane exists around the corresponding surface measuring point; and if coalbed methane exists around the corresponding surface measuring point, it is determined whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal, and based on its greater than, it is determined that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point; if coalbed methane is determined to exist within the preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with the electromagnetic exploration inversion model, to obtain the coal reservoir exploration results; wherein, the electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data; the electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0144] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements an electromagnetic exploration method for gas-bearing coal reservoirs provided by the methods described above. This method includes: obtaining the spectral characteristics and spatial distribution characteristics of a self-emission simulated signal based on a pre-constructed forward model of the coal reservoir self-emission, combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters; wherein the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include a simulated electromagnetic field signal; acquiring the measured electromagnetic field at a surface measuring point, comparing the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emission simulated signal, to determine whether there is a frequency exceeding the preset peak value range determined based on the spectral characteristics of the self-emission simulated signal. The process involves measuring a simulated electromagnetic field signal at a preset ratio and determining the presence of coalbed methane around the corresponding surface measuring point. If coalbed methane is present around the surface measuring point, the spatial distribution characteristics of the self-emitted simulated signal are used to determine whether the horizontal magnetic field strength at the surface measuring point is greater than the vertical magnetic field strength. If it is greater, the coalbed methane center is determined to be located within a preset range below the corresponding surface measuring point. If coalbed methane is confirmed to exist within the preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with an electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is trained based on simulated electromagnetic exploration training data and the corresponding electromagnetic exploration labels. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

[0145] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0146] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0147] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; 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; and these 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.

Claims

1. An electromagnetic exploration method for gas-bearing coal reservoirs, characterized in that, include: Based on a pre-constructed forward model of coal reservoir self-emission, and combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal are obtained; wherein, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include simulated electromagnetic field signals. The measured electromagnetic field at the surface measuring point is acquired, and the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field are compared with those of the self-emitted simulated signal. Within a frequency range of a preset peak value determined based on the spectral characteristics of the self-emitted simulated signal, it is determined whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field. Based on its existence, it is determined that coalbed methane exists around the corresponding surface measuring point. In the case that coalbed methane exists around the corresponding surface measuring point, it is determined whether the horizontal magnetic field strength of the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal. Based on its greater strength, it is determined that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point. If coalbed methane is found within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with the electromagnetic exploration inversion model, to obtain the coal reservoir exploration results. The electromagnetic exploration inversion model is obtained by training based on simulated electromagnetic exploration training data and the electromagnetic exploration labels corresponding to the simulated electromagnetic exploration training data. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

2. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 1, characterized in that, If coalbed methane is determined to exist within a preset range below the surface measuring point, the measured electromagnetic field at the corresponding surface measuring point is used, combined with an electromagnetic exploration inversion model, to obtain coal reservoir exploration results, including: The measured electromagnetic field at the surface measuring point where coalbed methane exists is used as the inversion input data of the electromagnetic exploration inversion model to obtain the coal reservoir exploration results output by the electromagnetic exploration inversion model; wherein, the measured electromagnetic field is a spectrum curve composed of the electromagnetic field from natural sources such as the ionosphere and lightning and the electromagnetic field emitted by the coal reservoir itself, and the measured electromagnetic field includes an amplitude spectrum curve and a phase spectrum curve.

3. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 2, characterized in that, Before obtaining the coal reservoir exploration results by using the measured electromagnetic field at the corresponding surface measuring point and combining it with the electromagnetic exploration inversion model, after confirming the existence of coalbed methane within a preset range below the surface measuring point, the process also includes: Based on the geological conditions of the corresponding surface measuring points corresponding to the previously acquired electromagnetic field measurement data, the corresponding preset parameter range is determined, and the underground resistivity parameter, electric dipole depth parameter, and angle parameter are randomly generated according to the preset parameter range. Based on the underground resistivity parameters, electric dipole depth parameters, and angle parameters, and combined with the coal reservoir self-emission forward model, the corresponding electric dipole source simulated electromagnetic field training data is obtained, and based on the electric dipole source simulated electromagnetic field training data, the electric dipole source simulated impedance training data is obtained. Based on the underground resistivity parameters and the natural source forward model, training data for the simulated electromagnetic field of the corresponding natural source is obtained, and training data for the simulated impedance of the natural source is obtained based on the training data for the simulated electromagnetic field of the natural source. Based on the measured electromagnetic field data, the preset ranges for the corresponding intensity ratio parameter, the preset range for the natural source magnetic field slope parameter, and the preset range for the natural source phase are determined. Based on the preset range of the intensity ratio parameter and the preset range of the natural source magnetic field slope parameter, the characteristic frequency band intensity ratio training parameter, the natural source magnetic field slope training parameter, and the magnetic field intensity ratio training coefficient are obtained; and based on the preset range of the natural source phase, the natural source phase training data is obtained. Based on the simulated impedance training data of the electric dipole source, the simulated impedance training data of the natural source, and the magnetic field strength ratio training coefficient, hybrid impedance training data is obtained. Based on the hybrid impedance training data and the natural source phase training data, simulated electromagnetic exploration training data is obtained, and electromagnetic exploration tags are obtained based on the characteristic frequency band intensity ratio training parameters, the natural source magnetic field slope training parameters, the underground resistivity parameters, and the electric dipole depth and angle parameters. The simulated electromagnetic exploration training data is used as the input data for training, and the electromagnetic exploration labels are used as the labels for training. The electromagnetic exploration inversion model to be trained is then trained to obtain the electromagnetic exploration inversion model.

4. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 3, characterized in that, The training of the electromagnetic exploration inversion model to be trained includes: The mixed impedance training data in the simulated electromagnetic exploration training data is input into the feature extraction layer for feature extraction to obtain mixed impedance features; The hybrid impedance characteristics and the natural source phase training data in the simulated electromagnetic exploration training data are input into the first electromagnetic exploration inversion layer to perform electromagnetic exploration inversion and obtain the coal reservoir parameter inversion results. The hybrid impedance characteristics and the natural source phase training data in the simulated electromagnetic exploration training data are input into the second electromagnetic exploration inversion layer to perform electromagnetic exploration inversion and obtain the underground resistivity distribution results. Based on the coal reservoir parameter inversion results and the characteristic frequency band intensity ratio training parameters, natural source magnetic field slope training parameters, and electric dipole depth and angle parameters in the electromagnetic exploration tag, a first loss function is constructed, and based on the underground resistivity distribution results and the underground resistivity parameters in the electromagnetic exploration tag, a second loss function is constructed. The total loss function is obtained based on the first loss function and the second loss function, and the training ends when the total loss function converges.

5. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 3, characterized in that, Based on the preset range of the intensity ratio parameter and the preset range of the natural source magnetic field slope parameter, the characteristic frequency band intensity ratio training parameters, the natural source magnetic field slope training parameters, and the magnetic field intensity ratio training coefficients are obtained, including: Based on the preset range of the intensity ratio parameter, the intensity ratio training parameters of the feature frequency band are obtained; The natural source magnetic field slope training parameters are obtained based on the preset range of the natural source magnetic field slope parameters. Based on the self-emission forward model simulation of the coal reservoir, the training coefficients of the systematic spectral change were obtained; Based on Gaussian noise simulation, the training coefficients for the interference ratio variation are obtained. The magnetic field strength ratio training coefficient is obtained based on the characteristic frequency band intensity ratio training parameters, the natural source magnetic field slope training parameters, the systematic spectrum change training coefficient, and the interference ratio change training coefficient.

6. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 1, characterized in that, Before using the measured electromagnetic field at a surface measuring point containing coalbed methane as the inversion input data for the electromagnetic exploration inversion model, the following steps are included: Static effect correction and topographic distortion correction were performed on the measured electromagnetic field at the surface measuring point where coalbed methane exists.

7. The electromagnetic exploration method for gas-bearing coal reservoirs according to claim 1, characterized in that, Before obtaining the spectral characteristics and spatial distribution characteristics of the self-emitted analog signal, the method further includes: A forward model of coal reservoir self-emission is constructed based on the electromagnetic field excited by electric dipoles in the underground layered medium and propagated to the surface.

8. An electromagnetic exploration device for gas-bearing coal reservoirs, characterized in that, include: The signal forward modeling module, based on a pre-constructed coal reservoir self-emission forward modeling model and combined with preset underground resistivity parameters, preset electric dipole depth parameters, and preset angle parameters, obtains the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal; wherein, the spectral characteristics and spatial distribution characteristics of the self-emission simulated signal include the simulated electromagnetic field signal; The coalbed methane determination module acquires the measured electromagnetic field at a surface measuring point, compares the spectral characteristics and spatial distribution characteristics of the measured electromagnetic field with those of the self-emitted simulated signal, and determines whether there is a simulated electromagnetic field signal exceeding a preset proportion of the measured electromagnetic field within a frequency range of a preset peak value determined based on the spectral characteristics of the self-emitted simulated signal. Based on its existence, it determines that coalbed methane exists around the corresponding surface measuring point. Furthermore, if coalbed methane exists around the corresponding surface measuring point, it determines whether the horizontal magnetic field strength at the surface measuring point is greater than the vertical magnetic field strength based on the spatial distribution characteristics of the self-emitted simulated signal, and based on this, determines that the center of the coalbed methane is located within a preset range below the corresponding surface measuring point. The electromagnetic exploration inversion module, upon determining the presence of coalbed methane within a preset range below the surface measuring point, uses the measured electromagnetic field at the corresponding surface measuring point and combines it with the electromagnetic exploration inversion model to obtain coal reservoir exploration results. The electromagnetic exploration inversion model is trained based on simulated electromagnetic exploration training data and the corresponding electromagnetic exploration labels. The electromagnetic exploration inversion model is used to perform inversion based on the input measured electromagnetic field.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the electromagnetic exploration method for gas-bearing coal reservoirs as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the electromagnetic exploration method for gas-bearing coal reservoirs as described in any one of claims 1 to 7.

Images