Coal bed methane enrichment area seismic prediction method with multiple geological parameter constraints

Abstract

The application discloses a coalbed methane enrichment area prediction method based on multiple geological parameter constraints, determines a research area, analyzes geological features of the research area to determine lithology, physical property features and elastic parameters of rock strata, predicts geological elements of a coalbed methane enrichment area of the research area, and establishes a nonlinear mapping relationship between main geological parameters of the coalbed methane and enrichment characteristics of the coalbed methane, namely, a geological attribute discrimination mode of the coalbed methane enrichment area, by means of artificial neural network method and constraint of the geological parameters. The method for predicting the coalbed methane enrichment area under the constraint of multiple geological parameters simultaneously controls the convergence direction of the solution of the enrichment area by multiple mutually irrelevant geological parameters, reduces the prediction scale of the enrichment unit to obtain an optimal solution, avoids the uncertainty of single parameter prediction, improves the accuracy of the prediction result, and solves the technical problem that the method for predicting the coalbed gas content by single seismic attribute or parameter in the prior art has strong multiple solutions, thereby leading to inaccurate prediction results.

Description

Technical Field

[0001] This invention belongs to the field of coalbed methane seismic exploration and relates to coal seam seismic prediction, specifically a seismic prediction method for coalbed methane enrichment areas constrained by multiple geological parameters. Background Technology

[0002] Coalbed methane (CBM) is a self-generated and self-storing unconventional natural gas stored primarily in adsorbed form (with a small amount of free and dissolved gas). It contains more than 90% methane and is commonly known as "gas" in coal mines. It is a crucial clean energy source among unconventional oil and gas resources. The exploration and development of CBM is of great significance for increasing the effective supply of natural gas, ensuring national energy security, improving the safe development and comprehensive utilization of coal and natural gas resources, ensuring coal mine production safety, reducing greenhouse gas emissions, protecting the atmospheric environment, reducing smog, and developing a green economy.

[0003] Coalbed methane (CBM) is "hidden" in coal seams in various ways, which is quite different from oil, which exists in a liquid state in rocks, and natural gas, which exists in a free state in strata. Due to the significant differences in the formation mechanisms of CBM and conventional oil and gas, conventional oil and gas prediction is based on the Biot-Gassmann theory, which assumes that in porous media, fluids exist in a free state within the pores. When seismic waves propagate in fluid-containing porous media, the fluids disturb the seismic waves, and this characteristic can be used for oil and gas prediction. However, CBM is mostly adsorbed onto the surface of the coal matrix. When seismic waves propagate in the coal matrix, they exhibit mostly solid characteristics, and fluids have little ability to disturb the seismic waves. Currently, there is a lack of fundamental research in rock physics in this field. Furthermore, the spatial distribution of gas content and permeability in coal seams is highly heterogeneous, making direct application of oil and gas prediction methods for CBM prediction usually ineffective. Moreover, current conventional methods for predicting CBM enrichment areas often use single seismic attributes or parameters to predict CBM enrichment, resulting in significant errors and multiple solutions. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method for earthquake prediction in coalbed methane-rich areas constrained by multiple geological parameters, thereby solving the technical problem that existing methods for predicting coalbed methane content using a single seismic attribute or parameter have strong ambiguity, leading to inaccurate prediction results.

[0005] To solve the above-mentioned technical problems, the present invention adopts the following technical solution:

[0006] A method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints, characterized by the following steps:

[0007] Step 1: Determine the study area and analyze its geological characteristics to determine the lithology, physical properties, and elastic parameters of the rock strata.

[0008] The geological features mentioned include methane content and methane concentration;

[0009] Step 2: Predict the geological elements of the coalbed methane enrichment zone in the study area;

[0010] The geological elements mentioned include coal seam thickness, lithology of the roof and floor of the coal seam, distribution of sedimentary microfacies before and after coal accumulation, coal seam permeability, coal body structure distribution, and coal seam gas content.

[0011] Step 3: Using the coal seam thickness, lithology of the top and bottom plates of the coal seam, coal body structure distribution, coal seam permeability and coal seam gas content of the sample well points as the input layer, and the coalbed methane saturation of the sample well points as the output layer, the sample data is trained using the BP neural network method to obtain the geological attribute discrimination model of the coalbed methane enrichment area.

[0012] Step 4: Input the test data into the constructed geological attribute discrimination model of coalbed methane enrichment area for discrimination, and obtain the coalbed methane enrichment area type number value of each CDP point in the whole area.

[0013] The test data includes the coal seam thickness, roof and floor lithology, coal seam permeability, coal body structure, and coal seam gas content for the entire area;

[0014] Step 5: Use the coalbed methane enrichment zone type number value of CDP point to perform interpolation calculation to obtain the predicted planar map of enrichment zone category distribution in the whole area.

[0015] This invention includes the following technical features:

[0016] In step two, post-stack seismic inversion technology is used to predict the coal seam thickness and the lithology of the top and bottom plates of the coal seam.

[0017] In step two, seismic facies analysis is used to predict the distribution of sedimentary microfacies before and after coal accumulation.

[0018] In step two, the coal body structure distribution is extracted using the pre-stack seismic inversion method.

[0019] In step two, the permeability of coal seams is predicted using seismic anisotropy property analysis.

[0020] In step two, the PG attribute profile of the target layer is extracted based on AVO theory, and then the gas content of the coal seam is predicted.

[0021] Step four specifically includes the following steps:

[0022] Step 4.1: Determine the study area through drilling and logging analysis, dividing the study area into n enrichment zones. The five geological element parameters for each enrichment zone are as follows:

[0023] The five types of geological parameters mentioned are coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content;

[0024] in:

[0025] X i This represents the values ​​of the five seismic attribute parameters corresponding to the i-th enriched region in the earthquake data to be predicted;

[0026] Step 4.2 involves extracting X from the seismic data. i Substituting into equation (1), we obtain the neural network output value F corresponding to the earthquake data to be predicted;

[0027] F=(f(X 1 ),f(X 2 ),...,f(X i ),...,f(X n )) (1)

[0028] Where: f(X) i () represents the five types of geological elements corresponding to the i-th enrichment zone constructed after training with a neural network. The output value of the neural network.

[0029] In steps 4.3 and 4.2, the neural network output value F is f(X) i Substituting these values ​​into equation (2) yields y. i , then y i Substituting into equations (3) and (4), we obtain the enrichment zone type number I corresponding to the earthquake data to be predicted;

[0030] y i =(f(X) i )-l i ) 2 (2)

[0031] y min =min(y 1 ,y 2 ,…,y i (3)

[0032]

[0033] in:

[0034] l i This indicates the identifier for the corresponding enriched region.

[0035] y i The component to be represented is f(X) i The square of the difference between the enrichment region category identifier and the enrichment region category identifier;

[0036] I represent yi Take the minimum value i, which is the rich region type number value corresponding to the data to be predicted.

[0037] Compared with the prior art, the beneficial technical effects of this invention are:

[0038] This invention establishes a nonlinear mapping relationship between the main geological parameters of coalbed methane and its enrichment characteristics using an artificial neural network method, constrained by coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content. This is a geological attribute discrimination model for coalbed methane enrichment areas. This method for predicting coalbed methane enrichment areas under the joint constraint of multiple geological parameters controls the convergence direction of the enrichment area solution by simultaneously controlling multiple uncorrelated geological parameters. It reduces the prediction scale of enrichment units to obtain the optimal solution, effectively reducing ambiguity and exploration risk, and avoiding the uncertainty of single-parameter prediction. This improves the accuracy of the prediction results and solves the technical problem of inaccurate prediction results caused by strong ambiguity in existing methods for predicting coal seam gas content using single seismic attributes or parameters. Attached Figure Description

[0039] Figure 1 The logging response characteristics of various lithologies in coal reservoirs are shown in the diagram.

[0040] Figure 2 A diagram illustrating the physical characteristics of coal reservoir rocks;

[0041] Figure 3 This is a plan view of the coal seam thickness distribution.

[0042] Figure 4(a) shows the planar distribution of clay content in the roof of the coal seam;

[0043] Figure 4(b) shows the planar distribution of clay content in the bottom plate of the coal seam;

[0044] Figure 5(a) shows the planar distribution of sedimentary microfacies before coal accumulation;

[0045] Figure 5(b) shows the planar distribution of microfacies deposited after coal accumulation;

[0046] Figure 6 This is to simultaneously invert the density profile before stacking;

[0047] Figure 7 This is a planar distribution diagram of the coal seam structure.

[0048] Figure 8 This is a planar distribution diagram of crack density.

[0049] Figure 9 This is a cross-plot of fracture density versus permeability.

[0050] Figure 10 Planar distribution map of coal seam permeability;

[0051] Figure 11 A diagram showing the AVO response characteristics of the top and bottom of the coal seam;

[0052] Figure 12 This is a planar distribution map of the gas-bearing factors in coal seams.

[0053] Figure 13 This is a comprehensive evaluation map of coalbed methane enrichment areas.

[0054] The specific content of the present invention will be further explained in detail below with reference to the embodiments. Detailed Implementation

[0055] The following are specific embodiments of the present invention. It should be noted that the present invention is not limited to the following specific embodiments. All equivalent modifications made based on the technical solutions of this application fall within the protection scope of the present invention.

[0056] This invention presents a method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints. The method includes the following steps:

[0057] Step 1: Determine the study area and analyze its geological characteristics to determine the lithology, physical properties, and elastic parameters of the rock strata.

[0058] The geological features mentioned include methane content and methane concentration;

[0059] Step 2: Predict the geological elements of the coalbed methane enrichment zone in the study area;

[0060] The geological elements mentioned include coal seam thickness, lithology of the roof and floor of the coal seam, distribution of sedimentary microfacies before and after coal accumulation, coal seam permeability, coal body structure distribution, and coal seam gas content.

[0061] Step 3: Using the coal seam thickness, lithology of the top and bottom plates of the coal seam, coal body structure distribution, coal seam permeability and coal seam gas content of the sample well points as the input layer, and the coalbed methane saturation of the sample well points as the output layer, the sample data is trained using the BP neural network method to obtain the geological attribute discrimination model of the coalbed methane enrichment area.

[0062] Step 4: Input the test data into the constructed geological attribute discrimination model of coalbed methane enrichment area for discrimination, and obtain the coalbed methane enrichment area type number value of each CDP point in the whole area.

[0063] The test data includes the coal seam thickness, roof and floor lithology, coal seam permeability, coal body structure, and coal seam gas content for the entire area;

[0064] Step 5: Use the coalbed methane enrichment zone type number value of CDP point to perform interpolation calculation to obtain the predicted planar map of enrichment zone category distribution in the whole area.

[0065] In the above technical solution, by constraining coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content, an artificial neural network method is used to establish a nonlinear mapping relationship between the main geological parameters of coalbed methane and the enrichment characteristics of coalbed methane, i.e., a geological attribute discrimination model for coalbed methane enrichment areas. This method of predicting coalbed methane enrichment areas under the joint constraint of multiple geological parameters controls the convergence direction of the solution of the enrichment area by simultaneously controlling multiple uncorrelated geological parameters, reducing the prediction scale of enrichment units to obtain the optimal solution, effectively reducing multiple solutions and exploration risks, avoiding the uncertainty of single parameter prediction, improving the accuracy of prediction results, and solving the technical problem that the existing methods for predicting coal seam gas content by single seismic attributes or parameters have strong multiple solutions, resulting in inaccurate prediction results.

[0066] Specifically, in step two, post-stack seismic inversion technology is used to predict the coal seam thickness and the lithology of the top and bottom plates of the coal seam.

[0067] Specifically, in step two, seismic facies analysis is used to predict the distribution of sedimentary microfacies before and after coal accumulation.

[0068] Specifically, in step two, the coal body structure distribution is extracted using the pre-stack seismic inversion method.

[0069] Specifically, in step two, the permeability of coal seams is predicted using seismic anisotropy property analysis.

[0070] Specifically, in step two, the PG attribute profile of the target layer is extracted based on AVO theory, and then the gas content of the coal seam is predicted.

[0071] Specifically, step four includes the following steps:

[0072] Step four specifically includes the following steps:

[0073] Step 4.1: Determine the study area through drilling and logging analysis, dividing the study area into n enrichment zones. The five geological element parameters for each enrichment zone are as follows:

[0074] The five types of geological parameters mentioned are coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content;

[0075] in:

[0076] X i This represents the values ​​of the five seismic attribute parameters corresponding to the i-th enriched region in the earthquake data to be predicted;

[0077] Step 4.2 involves extracting X from the seismic data. i Substituting into equation (1), we obtain the neural network output value F corresponding to the earthquake data to be predicted;

[0078] F=(f(X 1 ),f(X 2 ),...,f(X i ),...,f(X n )) (1)

[0079] Where: f(X) i () represents the five types of geological elements corresponding to the i-th enrichment zone constructed after training with a neural network. The output value of the neural network.

[0080] In steps 4.3 and 4.2, the neural network output value F is f(X) i Substituting these values ​​into equation (2) yields y. i , then y i Substituting into equations (3) and (4), we obtain the enrichment zone type number I corresponding to the earthquake data to be predicted;

[0081] y i =(f(X) i )-l i ) 2 (2)

[0082] y min =min(y 1 ,y 2 ,…,y i (3)

[0083]

[0084] in:

[0085] l i This indicates the identifier for the corresponding enriched region.

[0086] y i The component to be represented is f(X) i The square of the difference between the enrichment region category identifier and the enrichment region category identifier;

[0087] I represent y i Take the minimum value i, which is the rich region type number value corresponding to the data to be predicted.

[0088] Example:

[0089] This embodiment presents a method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints. The method includes the following steps:

[0090] Step 1: Determine the study area and analyze its geological characteristics to determine the lithology, physical properties, and elastic parameters of the rock strata.

[0091] The geological features mentioned include methane content and methane concentration;

[0092] Lithology includes sandstone, mudstone, and coal, while physical properties include the porosity of sandstone and the permeability of coal seams. Elastic parameters include wave impedance, density, and Poisson's ratio.

[0093] In this embodiment, the study area is located in the central and eastern part of the Qinshui Coalfield, and structurally it is located on the eastern wing of the Qinshui Basin syncline. The geotectonic classification is the North China Fault Block (Level II), the Lvliang-Taihang Fault Block (Level III), and the Qinshui Block-Depression (Level IV). Low and gentle parallel folds are common, and linear structures are the main features. The main coal-bearing strata are the Upper Carboniferous Taiyuan Formation (C3t) and the Lower Permian Shanxi Formation (P1s).

[0094] Based on previous exploration data, the block has three main mineable coal seams: No. 3 Upper, No. 3, and No. 15. Previous gas testing results indicate that No. 3 coal seam has an average methane content of 16.58 m³ / t and an average methane concentration of 88.82%; while No. 15 coal seam has an average methane content of 21.17 m³ / t and an average methane concentration of 88.91%. Vertically, the coalbed methane content in the lower coal seams is higher than that in the upper coal seams. Furthermore, the roof of No. 3 Upper and No. 3 coal seams is mainly composed of siltstone or sandy mudstone and mudstone, while the floor is mainly composed of mudstone and sandy mudstone. From a capping perspective, this is conducive to the preservation of coalbed methane.

[0095] By utilizing the differences in the response curves of various rocks on well logging curves, the locations of coal seams, sandstone, and mudstone in the selected wells can be precisely identified, such as... Figure 1 As shown; through rock physical analysis, cross-plot analysis was performed on the natural gamma curves and wave impedance curves of eight wells in the exploration area to establish the fitting relationship between natural gamma and wave impedance for different target layers, such as... Figure 2 As shown, from Figure 2 As can be seen, the sandstone in this area exhibits high impedance characteristics, while the mudstone exhibits low impedance characteristics.

[0096] Step 2: Predict the geological elements of the coalbed methane enrichment zone in the study area;

[0097] The geological elements mentioned include coal seam thickness, lithology of the roof and floor of the coal seam, distribution of sedimentary microfacies before and after coal accumulation, coal seam permeability, coal body structure distribution, and coal seam gas content.

[0098] Step 3: Using the coal seam thickness, lithology of the top and bottom plates of the coal seam, coal body structure distribution, coal seam permeability and coal seam gas content of the sample well points as the input layer, and the coalbed methane saturation of the sample well points as the output layer, the sample data is trained using the BP neural network method to obtain the geological attribute discrimination model of the coalbed methane enrichment area.

[0099] In this embodiment, the coalbed methane enrichment characteristics revealed by well locations are selected as samples. Five types of parameters, namely coal body structure, coal seam thickness, anisotropic fracture density (coal seam permeability), roof and floor lithology and gas-bearing factors, are used as constraints for coalbed methane enrichment areas, as shown in Table 1. The BP neural network algorithm is used for training to construct the correspondence between seismic coalbed methane enrichment areas and seismic attribute parameters, that is, the geological attribute discrimination model of coalbed methane enrichment areas.

[0100] Table 1 Training Sample Table of Coalbed Methane Enrichment Characteristics and Geological Parameters at Well Points

[0101]

[0102] Step 4: Input the test data into the constructed geological attribute discrimination model of coalbed methane enrichment area for discrimination, and obtain the coalbed methane enrichment area type number value of each CDP point in the whole area.

[0103] The test data includes the coal seam thickness, roof and floor lithology, coal seam permeability, coal body structure, and coal seam gas content for the entire area;

[0104] Step 5: Use the coalbed methane enrichment zone type number value of CDP point to perform interpolation calculation to obtain the predicted planar map of enrichment zone category distribution in the whole area.

[0105] Based on the coalbed methane enrichment zone identification model and combined with the prediction results of geological elements across the region, a comprehensive evaluation of coalbed methane enrichment zones is conducted, taking into account the overall tectonic development. Figure 13 This is a comprehensive evaluation map of the coalbed methane enrichment area in Coalbed Methane (CBM) No. 3. White areas represent Class I enrichment areas, light gray areas represent Class II enrichment areas, and dark gray and black areas represent relatively non-enriched CBM areas. It is evident that the CBM enrichment areas in Coalbed Methane No. 3 are concentrated in the northern and eastern parts of the exploration area, with the western part of the south also being a relatively enriched area. Enrichment unit prediction is the foundation of comprehensive geological evaluation of CBM. Based on this, and combined with CBM geological development theory, enrichment units should be evaluated and classified to identify high-enrichment areas and guide the construction of CBM wellfields.

[0106] Specifically, in step two, post-stack seismic inversion technology is used to predict the coal seam thickness and the lithology of the top and bottom plates of the coal seam.

[0107] In this embodiment, within a 6ms time window upward from the bottom interface of coal seam No. 3, the CDP points on the extracted wave impedance data volume are below the coal seam wave impedance threshold value (7700g / cm). 3 The cumulative time value (in m / s) is used as the time thickness value of coal seam No. 3 at the CDP point. Then, the layer velocity of the coal seam is obtained by combining the sonic curves of each well in the coal seam No. 3. The product of the time thickness value and the layer velocity is used as the preliminary predicted thickness of coal seam No. 3. Finally, the actual thickness of the coal seam obtained from each well is used as the benchmark to obtain the final predicted planar map of coal seam No. 3 thickness, as shown below. Figure 3 As shown.

[0108] exist Figure 2 The analysis diagram of the coal reservoir's petrophysical characteristics shows the intersection of the natural gamma curves and acoustic impedance curves of eight wells. It is evident that the acoustic impedance parameters effectively distinguish between sandstone and mudstone within each layer, and the acoustic impedance and natural gamma curves exhibit a good fit. Therefore, we can also obtain the distribution characteristics of natural gamma curves in each layer through acoustic impedance inversion, and then calculate the planar distribution of clay content based on the predicted natural gamma curves. Figure 4 shows the predicted planar distribution of clay content in the roof and floor of Coal Mine No. 3. The grayscale reflects the variation in clay content; darker grayscale indicates higher clay content, and lighter grayscale indicates lower clay content. The inversion results are consistent with the geological background of the area. Areas with higher clay content in the roof of Coal Mine No. 3 are mainly concentrated in the northeast and central parts of the exploration area. These areas have good caprock sealing, which is conducive to coalbed methane accumulation. Similarly, areas with higher clay content in the floor of Coal Mine No. 3 are mainly concentrated in the northwest and northeast of the exploration area. These areas also have good underlying strata sealing, which is conducive to coalbed methane accumulation.

[0109] Specifically, in step two, seismic facies analysis is used to predict the distribution of sedimentary microfacies before and after coal accumulation.

[0110] Sequence stratigraphy can roughly determine the vertical development characteristics of different types of sandstone and mudstone strata. By studying the seismic facies within the target sequence, the sedimentary facies and lateral distribution range of the sandstone and mudstone bodies can be determined. Seismic facies analysis is based on the division of seismic sequences, using the differences in seismic parameter characteristics to divide the seismic sequences into different seismic facies zones, and then making inferences about lithofacies and sedimentary environments.

[0111] In this embodiment, the No. 3 coal seam is located in the Lower Permian Shanxi Formation (P1s), which is a coal-bearing sedimentary rock series dominated by fluvial deltaic facies. Figures 5(a) and 5(b) show the distribution of sedimentary microfacies before and after coal accumulation, respectively. It can be seen that the roof of the No. 3 coal seam mainly develops four types of sedimentary microfacies: deltaic plain channels (Type 1), interchannel bays (Type 3), natural levees (Type 2), and floodplains (Type 4). The floor of the No. 3 coal seam mainly develops four types of sedimentary microfacies: deltaic plain channels (Type 4), interchannel bays (Type 2), natural levees (Type 3), and floodplains (Type 1).

[0112] Specifically, in step two, the coal body structure distribution is extracted using the pre-stack seismic inversion method.

[0113] Because primary coal and fractured coal have similar physical properties, their logging responses show little difference. Compared to fractured coal, granular coal exhibits more severe structural damage and more developed pores; its density logging response is significantly lower. Granular coal also shows lower radioactivity per unit volume, lower strength, and is prone to pore enlargement, resulting in a lower natural gamma ray logging response. Granular coal is more fragmented, has poorer cementation, and lower sound velocity, leading to a significantly greater sonic transit time logging response than fractured coal. Mylonite has the most severely damaged primary structure and the most developed pores, making it more susceptible to drilling processes. Its pore mud content is higher than other types of tectonic coal, resulting in a low anomaly in apparent resistivity logging. It shows higher density and natural gamma ray logging responses than other types of tectonic coal, and a high anomaly in sonic transit time logging, slightly higher than granular coal. Tectonic coal exhibits many anomalous logging responses, primarily characterized by low resistivity, low density, and high sonic transit time. By simultaneously inverting density elastic parameters before stacking, the density elastic parameter attributes of the target coal seam are extracted, such as... Figure 6 As shown, the predicted location of the tectonic coal seam is as follows. Figure 7 As shown.

[0114] Specifically, in step two, the permeability of coal seams is predicted using seismic anisotropy property analysis.

[0115] In this embodiment, amplitude and time difference are used for crack prediction. The differences in amplitude and time difference of the target layer at different orientations are analyzed, and then crack density is predicted. Figure 8 As shown; the analysis of wells in the exploration area containing permeability test data extracts the predicted fracture density around the wellbore, and performs cross-plot analysis to fit the statistical relationship between permeability and fracture density, as shown. Figure 9 As shown, a planar distribution map of coal seam permeability is obtained by combining data from fracture density prediction. Figure 10 As shown.

[0116] Specifically, in step two, the PG attribute profile of the target layer is extracted based on AVO theory, and then the gas content of the coal seam is predicted.

[0117] The study of AVO response characteristics in coal seams revealed that in the study area, the reflection amplitudes of both the roof and floor reflection interfaces of the No. 3 coal seam's main reservoir significantly increased with increasing offset. Therefore, taking the sum of the intercept and gradient would enhance AVO anomalies associated with coalbed methane enrichment and weaken AVO inversion noise unrelated to coalbed methane enrichment. Taking the envelope of the sum of the intercept and gradient, the in-phase axis anomalies associated with the roof reflection interface and the floor reflection interface are merged and displayed as a single in-phase axis, as shown below. Figure 11 As shown. Based on the simplified equations of Hilterman and Shuey, the PG attribute is extracted. When the ratio of P-wave velocity to S-wave velocity is approximately equal to 2, P*G reflects the gas-bearing characteristics of the coal seam, such as... Figure 12 As shown, Figure 12 The image shows the planar distribution of gas-bearing factors in coal seam No. 3. Areas with lower gray levels (white) are gas-abnormal zones. Based on sedimentary patterns and using P+G attribute predictions, it is believed that coal seam No. 3 has high gas content in the central and southwestern strips of the exploration area, which are favorable zones for future drilling.

[0118] Specifically, step four includes the following steps:

[0119] Step 4.1: Determine the study area through drilling and logging analysis, dividing the study area into n enrichment zones. The five geological element parameters for each enrichment zone are as follows:

[0120] The five types of geological parameters mentioned are coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content;

[0121] in:

[0122] X i This represents the values ​​of the five seismic attribute parameters corresponding to the i-th enriched region in the earthquake data to be predicted;

[0123] Step 4.2 involves extracting X from the seismic data. i Substituting into equation (1), we obtain the neural network output value F corresponding to the earthquake data to be predicted;

[0124] F=(f(X 1 ),f(X 2 ),...,f(X i ),...,f(X n )) (1)

[0125] Where: f(X) i () represents the five types of geological elements corresponding to the i-th enrichment zone constructed after training with a neural network. The output value of the neural network.

[0126] In steps 4.3 and 4.2, the neural network output value F is f(X) i Substituting these values ​​into equation (2) yields y. i , then y i Substituting into equations (3) and (4), we obtain the enrichment zone type number I corresponding to the earthquake data to be predicted;

[0127] y i =(f(X) i )-l i ) 2 (2)

[0128] y min =min(y 1 ,y2 ,…,y i (3)

[0129]

[0130] in:

[0131] l i This indicates the identifier for the corresponding enriched region.

[0132] y i The component to be represented is f(X) i The square of the difference between the enrichment region category identifier and the enrichment region category identifier;

[0133] I represent y i Take the minimum value i, which is the rich region type number value corresponding to the data to be predicted.

Claims

1. A method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints, characterized in that, The method includes the following steps: Step 1: Determine the study area and analyze its geological characteristics to determine the lithology, physical properties, and elastic parameters of the rock strata. The geological features mentioned include methane content and methane concentration; Step 2: Predict the geological elements of the coalbed methane enrichment zone in the study area; The geological elements mentioned include coal seam thickness, lithology of the roof and floor of the coal seam, distribution of sedimentary microfacies before and after coal accumulation, coal seam permeability, coal body structure distribution, and coal seam gas content. Step 3: Using the coal seam thickness, lithology of the top and bottom plates of the coal seam, coal body structure distribution, coal seam permeability and coal seam gas content of the sample well points as the input layer, and the coalbed methane saturation of the sample well points as the output layer, the sample data is trained using the BP neural network method to obtain the geological attribute discrimination model of the coalbed methane enrichment area. Step 4: Input the test data into the constructed geological attribute discrimination model of coalbed methane enrichment area for discrimination, and obtain the coalbed methane enrichment area type number value of each CDP point in the whole area. The test data includes the coal seam thickness, roof and floor lithology, coal seam permeability, coal body structure, and coal seam gas content for the entire area; Step 4.1: Determine the study area through drilling and logging analysis, dividing the study area into n enrichment zones. The five geological element parameters for each enrichment zone are as follows: The five types of geological parameters mentioned are coal seam thickness, lithology of the roof and floor of the coal seam, coal body structure distribution, coal seam permeability, and coal seam gas content; in: X i This represents the values ​​of the five seismic attribute parameters corresponding to the i-th enriched region in the earthquake data to be predicted; Step 4.2 involves extracting X from the seismic data. i Substituting into equation (1), we obtain the neural network output value F corresponding to the earthquake data to be predicted; F=(f(X 1 ),f(X 2 ),...,f(X i ),...,f(X n )) (1) Where: f(X) i () represents the five types of geological elements corresponding to the i-th type of enrichment area constructed after training by a neural network. The output value of the neural network; In steps 4.3 and 4.2, the neural network output value F is f(X) i Substituting these values ​​into equation (2) yields y. i , then y i Substituting into equations (3) and (4), we obtain the enrichment zone type number I corresponding to the earthquake data to be predicted; y i =(f(X i )-l i ) 2 (2) and min =min(y 1 ,and 2 ,…,and i ) (3) in: l i This indicates the identifier for the corresponding enriched region. y i The component to be represented is f(X) i The square of the difference between the enrichment region category identifier and the enrichment region category identifier; I represent y i Take the minimum value i, which is the rich region type number value corresponding to the data to be predicted; Step 5: Use the coalbed methane enrichment zone type number value of CDP point to perform interpolation calculation to obtain the predicted planar map of enrichment zone category distribution in the whole area.

2. The method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints as described in claim 1, characterized in that, In step two, post-stack seismic inversion technology is used to predict the coal seam thickness and the lithology of the top and bottom plates of the coal seam.

3. The method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints as described in claim 1, characterized in that, In step two, seismic facies analysis is used to predict the distribution of sedimentary microfacies before and after coal accumulation.

4. The method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints as described in claim 1, characterized in that, In step two, the coal body structure distribution is extracted using the pre-stack seismic inversion method.

5. The method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints as described in claim 1, characterized in that, In step two, the permeability of coal seams is predicted using seismic anisotropy property analysis.

6. The method for predicting coalbed methane enrichment areas based on multiple geological parameter constraints as described in claim 1, characterized in that, In step two, the PG attribute profile of the target layer is extracted based on AVO theory, and then the gas content of the coal seam is predicted.

Images