A shale favorable sedimentary facies belt prediction method, system, medium and processor

Abstract

The present application is suitable for the technical field of shale gas exploration and development, and provides a shale favorable sedimentary facies belt prediction method, system, medium and processor. The method comprises: performing step-by-step screening on multi-source geological data based on geological classification screening rules through the multi-source geological data to obtain a core evaluation index system; wherein the step-by-step screening comprises first-level screening based on lithological combination, sedimentary structure and paleontological marker to identify sedimentary facies types, second-level screening based on the correlation strength between each initial evaluation index and a shale gas content macroscopic characterization parameter, and third-level screening based on the adaptation degree of each index and key elements of tectonic preservation conditions; a sedimentary facies belt favorable quantitative characterization model is constructed according to the core evaluation index system; finally, the core evaluation index of the region to be measured is input into the model for prediction, and the shale favorable sedimentary facies belt grade and the uncertainty range of the prediction result are output. The accuracy and reliability of shale favorable sedimentary facies belt prediction are effectively improved.

Description

Technical Field

[0001] This invention relates to the field of shale gas exploration and development technology, and in particular to a method, system, medium and processor for predicting favorable sedimentary facies zones in shale. Background Technology

[0002] Predicting favorable sedimentary facies zones in shale is a crucial step in shale gas exploration and deployment. Its purpose is to identify the most favorable sedimentary environment combinations for the development of organic-rich shale and shale gas accumulation within the regional geological context. Existing prediction methods typically rely on comprehensive evaluation based on multi-source geological data. First, drilling, logging, seismic, and geochemical data for the target area are collected to extract multiple evaluation indicators reflecting the sedimentary environment, source rock quality, and structural preservation conditions. Then, based on geological experience or statistical analysis, weights are assigned to each indicator or thresholds are defined. Finally, through multi-factor overlay, the weighted sum of each indicator yields a favorable score, and the distribution range of favorable sedimentary facies zones is delineated based on the score. However, due to the significant complexity and heterogeneity of shale accumulation conditions, a non-linear coupling relationship exists between the evaluation indicators and shale gas-bearing capacity. Traditional linear weighting methods struggle to accurately characterize this complex mapping. Furthermore, structural preservation conditions act as a threshold constraint on shale gas enrichment; failure to meet preservation requirements for a single key indicator can lead to overall failure, resulting in low prediction accuracy and poor reliability, hindering effective guidance for shale gas exploration and deployment. Summary of the Invention

[0003] This invention provides a method, system, medium, and processor for predicting favorable sedimentary facies zones in shale, which addresses the problems of low prediction accuracy and poor reliability, making it difficult to effectively guide shale gas exploration and deployment.

[0004] The first aspect of this invention provides a method for predicting favorable sedimentary facies zones in shale, comprising: Acquire multi-source geological data for the target area, including initial evaluation indicators reflecting sedimentary environment characteristics, source rock quality characteristics, and tectonic preservation conditions. The multi-source geological data are screened step-by-step based on preset geological grading rules to obtain a core evaluation index system; wherein, the step-by-step screening includes: The first level of screening identifies sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, selecting areas that belong to the preset favorable sedimentary facies types. The second level of screening involves selecting indicators whose correlation with each initial evaluation index meets a preset threshold, based on the correlation strength between each index and the macroscopic characterization parameters of shale gas content. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. A quantitative characterization model for the favorable sedimentary facies zone is constructed based on the core evaluation index system. This model is used to calculate the comprehensive index of the favorable sedimentary facies zone in the area to be tested and to quantify the uncertainty of the prediction results. The core evaluation indicators of the area to be tested are input into the favorable sedimentary facies zone quantitative characterization model for prediction, and the favorable sedimentary facies zone level of the shale in the area to be tested and the uncertainty range of the prediction results are output.

[0005] Furthermore, the process of identifying sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, and then selecting areas belonging to a preset favorable sedimentary facies type, includes: Based on single-well facies analysis, according to the lithological assemblage, sedimentary structures and paleontological markers in the well logging data, the sedimentary facies types developed vertically in a single well are identified, and favorable intervals belonging to the preset favorable sedimentary facies types are delineated; the preset favorable sedimentary facies types include fine-grained sedimentary facies formed in deep-water anoxic environments, and the fine-grained sedimentary facies include platform facies, deep-water shelf facies and basin facies. Based on seismic facies analysis, according to seismic reflection structure, external geometry and seismic attribute characteristics, the planar distribution of sedimentary facies types is identified, and favorable areas belonging to the preset favorable sedimentary facies types are delineated. The favorable strata identified by a single well and the favorable areas delineated by seismic activity are cross-calibrated to determine the target spatial range that passes the first-level screening.

[0006] Furthermore, the process of selecting indicators whose correlation with each initial evaluation index meets a preset threshold based on the correlation strength between the initial evaluation index and the macroscopic characterization parameters of shale gas content includes: Obtain macroscopic characterization parameters of shale gas content from known wells, wherein the macroscopic characterization parameters of shale gas content include at least one of measured gas content, total hydrocarbon value measured by gas analysis, and shale gas indication thickness; Calculate the correlation coefficients between each initial evaluation index and the macroscopic characterization parameters of shale gas content; Initial evaluation indicators whose absolute value of the correlation coefficient is greater than or equal to the preset correlation threshold are identified as indicators that pass the second-level screening.

[0007] Furthermore, based on the degree of fit between each indicator and key elements of structural preservation conditions, indicators that can synergistically characterize the effectiveness of structural preservation are selected, including: To obtain key elements of structural preservation conditions, the key elements of structural preservation conditions include at least one of fault sealing index, distance from magmatic rock mass, formation pressure coefficient and effective thickness of the top plate; Set structural preservation effectiveness constraints for each key element; the effectiveness constraints are determined based on the regional structural evolution characteristics and the preservation effect of drilled wells. Each indicator obtained after the first and second level screening is matched with the constraints one by one; if the value of an indicator meets the constraints of all key elements, the indicator is determined to pass the third level screening; if the value of an indicator does not meet the constraints of any key element, the indicator is determined to fail the third level screening. All matching indicator combinations are identified as the indicator set that passes the third-level screening.

[0008] Furthermore, the construction of a favorable quantitative characterization model for sedimentary facies zones includes: Using the core evaluation index system as input features and the preset favorable zone level as output label, a neural network proxy model is constructed. The neural network surrogate model is trained using sample data from known wells, and the model parameters are optimized using the backpropagation algorithm to obtain a well-trained quantitative characterization model of sedimentary facies zones. The output layer of the neural network proxy model includes both a comprehensive advantage index output node and an uncertainty quantification output node.

[0009] Furthermore, the calculation of the comprehensive index of the favorable sedimentary facies zone of the area to be tested includes: Input the core evaluation indicators of the region to be tested into the trained neural network proxy model; The favorable index of sedimentary facies zone in the test area is obtained from the output node of the favorable index through forward propagation calculation of the model. The favorable index is a dimensionless value with a value range of 0 to 1. The larger the value, the more favorable the shale accumulation conditions.

[0010] Furthermore, the uncertainty of the quantitative prediction result includes: The Bayesian neural network method is used to assign the weight parameters of the neural network surrogate model to the prior distribution, and the posterior distribution of the weights is learned through variational inference. In the prediction phase, multiple samples are taken from the weighted posterior distribution, and multiple forward propagations are performed to obtain a set of prediction samples for the overall advantage index. The uncertainty range of the prediction results is calculated based on the statistical distribution of the predicted samples.

[0011] A second aspect of the present invention provides a system for predicting favorable sedimentary facies zones in shale, comprising: The multi-source geological data module is used to acquire multi-source geological data of the target area. The multi-source geological data includes initial evaluation indicators that reflect the characteristics of sedimentary environment, source rock quality, and tectonic preservation conditions. The core evaluation index system determination module is used to perform step-by-step screening of the multi-source geological data based on preset geological grading rules to obtain the core evaluation index system; wherein, the step-by-step screening includes: The first level of screening identifies sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, selecting areas that belong to the preset favorable sedimentary facies types. The second level of screening involves selecting indicators whose correlation with each initial evaluation index meets a preset threshold, based on the correlation strength between each index and the macroscopic characterization parameters of shale gas content. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. The sedimentary facies zone favorable quantitative characterization model construction module is used to construct a sedimentary facies zone favorable quantitative characterization model based on the core evaluation index system; the sedimentary facies zone favorable quantitative characterization model is used to calculate the comprehensive index of sedimentary facies zone favorableness in the area to be tested and to quantify the uncertainty of the prediction results. The prediction module is used to input the core evaluation indicators of the area to be tested into the favorable quantitative characterization model of the sedimentary facies zone for prediction, and output the favorable sedimentary facies zone level of the shale in the area to be tested and the uncertainty range of the prediction results.

[0012] A third aspect of the present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the shale favorable sedimentary facies zone prediction method as described above.

[0013] A fourth aspect of the present invention provides a processor for running a computer program, which, when running, executes the shale favorable sedimentary facies zone prediction method as described above.

[0014] As can be seen from the above technical solutions, the present invention has the following advantages: This invention acquires multi-source geological data of a target area and performs step-by-step screening of the multi-source geological data based on preset geological classification rules to obtain a core evaluation index system. The step-by-step screening includes a first-level screening based on lithological assemblage, sedimentary structures, and paleontological markers to identify sedimentary facies types; a second-level screening based on the correlation strength between each initial evaluation index and macroscopic characterization parameters of shale gas content; and a third-level screening based on the degree of fit between each index and key elements of structural preservation conditions. Based on the core evaluation index system, a quantitative characterization model for the favorable sedimentary facies zone is constructed. This model is used to calculate the comprehensive index of the favorable sedimentary facies zone in the area to be tested and quantify the uncertainty of the prediction results. Finally, the core evaluation indexes of the area to be tested are input into the model for prediction, outputting the favorable sedimentary facies zone level of shale and the uncertainty range of the prediction results. This invention effectively improves the representativeness and relevance of the core evaluation index system by systematically stripping away indicators that are irrelevant to or have a low contribution to shale gas accumulation through a three-level screening process. By constructing a characterization model that combines the functions of favorableness calculation and uncertainty quantification, it generates a comprehensive evaluation effect that takes into account both nonlinear accumulation patterns and threshold constraints, and provides quantifiable risk basis for exploration decisions, significantly improving the accuracy and reliability of predicting favorable sedimentary facies zones in shale. Attached Figure Description

[0015] Figure 1 This is a flowchart of the method for predicting favorable sedimentary facies zones in shale in this invention; Figure 2 This is a flowchart of the three-stage screening process for favorable sedimentary facies zones in shale in this invention; Figure 3 This is a flowchart of the advantage model construction and prediction process in this invention. Detailed Implementation

[0016] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “corresponding to,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0017] Example 1 The implementation method in this embodiment can be implemented in a system, on a server, or on a terminal; no specific limitation is made. The method in this application will be described below from the perspective of system implementation. Please refer to... Figure 1 The method provided in this application includes the following steps: S1. Obtain multi-source geological data for the target area. The multi-source geological data includes initial evaluation indicators that reflect the characteristics of the sedimentary environment, the quality characteristics of the source rocks, and the characteristics of the tectonic preservation conditions. Specifically, the initial evaluation indicators reflecting sedimentary environment characteristics refer to geological parameters used to identify the paleosedimentary environment type and its spatial distribution in the target area, including lithological assemblage, sedimentary structures, paleontological markers, sandstone content, mudstone color, bedding type, and seismic reflection structure. The initial evaluation indicators reflecting source rock quality characteristics refer to geological parameters used to evaluate the hydrocarbon generation potential and reservoir performance of organic-rich shale, including total organic carbon content, rock pyrolysis parameters, vitrinite reflectance, kerogen type, organic matter abundance, organic matter maturity, porosity, permeability, brittle mineral content, and clay mineral content. The initial evaluation indicators reflecting tectonic preservation conditions refer to geological parameters used to evaluate the effectiveness of shale gas reservoir preservation in the later stages, including fault development degree, fault sealing index, distance from magmatic bodies, formation pressure coefficient, roof lithology, roof thickness, burial depth, and tectonic deformation intensity. These initial evaluation indicators collectively constitute the basic dataset of multi-source geological data, used for subsequent step-by-step screening and comprehensive evaluation.

[0018] S2. Based on preset geological classification and screening rules, multi-source geological data are screened step by step to obtain a core evaluation index system; wherein, the step-by-step screening includes: The core evaluation index system refers to a set of key evaluation indicators that, after three levels of screening, are closely related to shale gas accumulation conditions and can synergistically characterize the sedimentary environment, source rock quality, and tectonic preservation effectiveness. These indicators are used to construct subsequent quantitative characterization models for favorable sedimentary facies zones. Please refer to [link to relevant documentation]. Figure 2 The specific implementation process of the three-level screening is as follows: S21. First-level screening: Based on lithological assemblage, sedimentary structure and paleontological markers, sedimentary facies types are identified to screen out areas belonging to the preset favorable sedimentary facies types; 1. Based on single-well facies analysis, according to the lithological assemblage, sedimentary structures and paleontological markers in the well logging data, the sedimentary facies types developed vertically in a single well are identified, and favorable intervals belonging to the preset favorable sedimentary facies types are delineated; the preset favorable sedimentary facies types include fine-grained sedimentary facies formed in deep-water anoxic environments, and the fine-grained sedimentary facies include platform facies, deep-water shelf facies and basin facies; Specifically, single-well facies analysis refers to identifying lithological variations based on logging data from a single well, utilizing characteristics such as logging curve morphology, amplitude, and smoothness. This is combined with logging response characteristics such as natural gamma ray, resistivity, and sonic transit time to determine sedimentary structures. Furthermore, based on paleontological markers reflected in the logging data, such as the stratigraphic locations of fossils like ostracods, conodonts, and graptolites, the analysis comprehensively identifies the sedimentary facies types developed vertically within the single well. Sedimentary facies types include platform facies, deep-water shelf facies, basin facies, platform margin slope facies, open platform facies, restricted platform facies, littoral facies, deltaic facies, and fluvial facies. Pre-defined favorable sedimentary facies types are sedimentary environment types most conducive to the development of organic-rich shale, determined based on regional shale gas exploration practices. These specifically include fine-grained sedimentary facies formed in deep-water anoxic environments, namely platform facies, deep-water shelf facies, and basin facies.

[0019] The process of identifying favorable facies intervals is as follows: First, sedimentary facies types are identified segment by segment along the vertical direction of a single well according to depth. The identified sedimentary facies types are compared with the preset favorable sedimentary facies types. When the sedimentary facies types continuously identified within a certain depth segment belong to the platform basin facies, deep-water shelf facies, or basin facies, the top and bottom depth ranges of that depth segment are recorded as candidate favorable facies intervals. After the identification of the entire well segment is completed, all candidate favorable facies intervals are merged, and thin segments with a thickness less than the minimum effective thickness are removed. Finally, a list of favorable facies intervals that have passed the first-level screening along the vertical direction of a single well is obtained. Each favorable facies interval includes top depth, bottom depth, and thickness information.

[0020] 2. Based on seismic facies analysis, according to seismic reflection structure, external geometry and seismic attribute characteristics, identify the planar distribution of sedimentary facies types and delineate favorable areas belonging to the pre-defined favorable sedimentary facies types; Seismic facies analysis refers to the comprehensive analysis of seismic data, including seismic reflection structures such as parallel, subparallel, divergent, progradational, and mound-like features, external seismic geometry such as sheet-like, wedge-like, lenticular, and mound-like features, and seismic attributes such as amplitude, frequency, and continuity, to identify the distribution range of sedimentary facies types on a plane.

[0021] The process of delineating favorable areas is as follows: First, the seismic data is interpreted and its attributes are extracted to identify seismic facies units with different reflection structures and geometric shapes on the seismic profile. Then, the correspondence between seismic facies and sedimentary facies types is established, and the seismic facies units are converted into a planar distribution map of sedimentary facies types. On the planar distribution map of sedimentary facies types, the boundaries of areas belonging to platform-basin facies, deep-water shelf facies, or basin facies are delineated to form multiple closed favorable area polygons. Each favorable area polygon contains its planar coordinate range and area information.

[0022] 3. Cross-calibrate the favorable strata identified by a single well with the favorable areas delineated by seismic analysis to determine the target spatial range that passes the first-level screening.

[0023] The favorable stratigraphic intervals identified by a single well are projected onto a plane to obtain the location of the favorable area at the well point. This location is then overlaid and compared with the favorable area polygon delineated by seismic analysis. If the well point location corresponding to the favorable stratigraphic interval of a single well is located inside the favorable area polygon delineated by seismic analysis, the two are considered to match, and this favorable area polygon is taken as the target spatial range for passing the first-level screening. If the well point location corresponding to the favorable stratigraphic interval of a single well is located outside the favorable area polygon delineated by seismic analysis, the two are considered to not match. In this case, the correspondence between seismic facies and sedimentary facies is recalibrated based on the favorable stratigraphic interval of the single well, and the seismic facies interpretation parameters are adjusted until the favorable area delineated by seismic analysis can cover the well point location. The above process is repeated until all single wells involved in the calibration match the seismically delineated area. The finally determined favorable area polygon is taken as the target spatial range for passing the first-level screening. This target spatial range is a three-dimensional spatial range, jointly defined by the favorable area polygon on the plane and the depth range of the favorable stratigraphic interval in the vertical direction.

[0024] S22. Second-level screening: Based on the correlation strength between each initial evaluation index and the macroscopic characterization parameters of shale gas content, indexes whose correlation meets the preset threshold are selected. 1. Obtain macroscopic characterization parameters of shale gas-bearing properties from known wells. These parameters include at least one of the following: measured gas content, total hydrocarbon value from gas analysis, and shale gas indication thickness. 2. Calculate the correlation coefficients between each initial evaluation index and the macroscopic characterization parameters of shale gas content; 3. Initial evaluation indicators whose absolute value of the correlation coefficient is greater than or equal to the preset correlation threshold are identified as indicators that pass the second-level screening.

[0025] Specifically, macroscopic characterization parameters of shale gas content refer to measured or calculated parameters used to directly or indirectly reflect the gas content of shale, including measured gas content, total hydrocarbon value, and shale gas indication thickness. Measured gas content refers to the gas content of shale samples directly measured through in-situ desorption experiments; total hydrocarbon value refers to the total hydrocarbon content detected by the gas logging instrument during drilling; and shale gas indication thickness refers to the thickness of continuous sections showing gas anomalies or core gas encounters during drilling. These parameters can be used individually or in combination. When multiple parameters are used in combination, each parameter needs to be normalized to eliminate dimensional effects before taking the arithmetic mean as the comprehensive characterization parameter.

[0026] The correlation coefficients between each initial evaluation index and the macroscopic characterization parameters of shale gas content were calculated using the Pearson correlation coefficient formula: in: This represents the correlation coefficient between the initial evaluation index and the macroscopic characterization parameters of shale gas content. Indicates the first The initial evaluation index values ​​for each sample point Indicates the first The values ​​of macroscopic characterization parameters of shale gas content at each sample point. This represents the average of the initial evaluation index for all sample points. This represents the average value of the macroscopic characterization parameters of shale gas content across all sample points. Indicates the number of sample points.

[0027] The preset correlation threshold refers to the critical correlation coefficient value set in advance for screening indicators. The value ranges from 0.3 to 0.7, and the specific value is determined according to the regional geological characteristics and the quality of sample data.

[0028] The specific process for determining initial evaluation indicators whose absolute value of the correlation coefficient is greater than or equal to the preset correlation threshold as indicators that pass the second-level screening is as follows: For each initial evaluation indicator, calculate its correlation coefficient with the macroscopic characterization parameters of shale gas content according to the above formula. and take the absolute value ;Will correlation threshold with preset Compare; if If the index is significantly correlated with the gas content of shale, it will be included in the list of indicators that have passed the second-level screening; if If an indicator is not significantly correlated with the gas content of shale, it is determined that the indicator fails the second-level screening and will not participate in subsequent evaluations. After completing the above judgment for all initial evaluation indicators, a list of indicators that have passed the second-level screening is obtained, and the original data of each indicator in the list is retained.

[0029] S23. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. 1. Obtain key elements of structural preservation conditions, which include at least one of the following: fault sealing index, distance from magmatic rock mass, formation pressure coefficient, and effective thickness of the top plate; Key elements of structural preservation conditions refer to geological parameters that control the effectiveness of shale gas reservoir preservation in the later stages. These include fault sealing index, distance from magmatic bodies, formation pressure coefficient, and effective roof thickness. The fault sealing index is a dimensionless parameter used to evaluate the lateral sealing capacity of faults, ranging from 0 to 1, with higher values ​​indicating better sealing. Distance from magmatic bodies refers to the planar distance between the evaluation point and the nearest magmatic body boundary, measured in kilometers. The formation pressure coefficient is the dimensionless ratio of measured formation pressure to hydrostatic pressure. The effective roof thickness refers to the thickness of the tight rock layer above the shale layer that effectively seals the shale gas. These key elements can be used individually or in combination. When multiple elements are used in combination, all elements must participate in the setting and matching verification of constraints.

[0030] 2. Set structural preservation effectiveness constraints for each key element; the effectiveness constraints are determined based on the regional structural evolution characteristics and the preservation effect of drilled wells. Structural preservation effectiveness constraints refer to the critical value ranges set for each key element to ensure effective shale gas preservation. The constraints are determined based on regional tectonic evolution characteristics and the preservation effect calibration of drilled wells. The process is as follows: First, collect the structural preservation effect evaluation results of drilled wells in the region. Based on the actual gas-bearing performance of the wells, classify them into two categories: effective preservation wells and ineffective preservation wells. Then, statistically analyze the numerical distribution range of each key element in effective and ineffective preservation wells, calculating the minimum, maximum, and average values ​​of each element in effective preservation wells and ineffective preservation wells, respectively. Finally, based on the differences in numerical distribution between the two types of wells, select the critical value that best distinguishes between effective and ineffective preservation as the constraint. For example, the constraint for fault sealing index is set to be greater than 0.7, the constraint for distance from magmatic bodies is set to be greater than 3 kilometers, the constraint for formation pressure coefficient is set to be greater than or equal to 1.2, and the constraint for effective roof thickness is set to be greater than or equal to 30 meters.

[0031] 3. Match each indicator obtained after the first and second level screening with the constraints one by one; if the value of an indicator meets the constraints of all key elements, the indicator is determined to pass the third level screening; if the value of an indicator does not meet the constraints of any key element, the indicator is determined to fail the third level screening. 4. All matching indicator combinations are identified as the indicator set that passes the third-level screening.

[0032] The process of matching each indicator obtained after the first and second levels of screening with the constraints is as follows: Obtain the list of indicators retained after the first and second levels of screening, where each indicator corresponds to a spatial location; for each indicator in the list, extract the corresponding key structural preservation condition values ​​from its spatial location, including the fault closure index. Distance from igneous rock mass Formation pressure coefficient and effective thickness of top plate Then, check whether the values ​​of each key element meet the corresponding constraints one by one. The judgment logic is as follows: like and and and If the indicator meets all the constraints of the key elements, then the indicator is marked as having passed the third level of screening. like or or or If any one of the above four conditions is not met, the indicator is determined to be unsatisfactory and is marked as failing the third-level screening.

[0033] After completing the above matching and verification for all indicators, a pass or fail flag is obtained for each indicator. An empty indicator set is created; the list of indicators retained after the first and second levels of screening is iterated through. For each indicator marked as passing the third level of screening, its identifier information along with its original data is added to the indicator set; indicators marked as failing the third level of screening are not added to the indicator set. After the iteration is complete, all indicators in the indicator set are those that have passed the third level of screening. This indicator set is an ordered set, and each element in the set contains the indicator name, indicator value, indicator source, spatial location information, and the flag information indicating that it has passed each level of screening. This indicator set, together with the indicators retained after the first and second levels of screening, constitutes the final core evaluation indicator system, used for the construction of subsequent sedimentary facies zone favorable quantitative characterization models.

[0034] S3. Construct a quantitative characterization model for the favorable sedimentary facies zone based on the core evaluation index system; the quantitative characterization model for the favorable sedimentary facies zone is used to calculate the comprehensive index of the favorable sedimentary facies zone in the area to be tested, and to quantify the uncertainty of the prediction results; In this embodiment, the architecture of the sedimentary facies zone favorable quantitative characterization model includes: using the core evaluation index system as input features and the preset favorable zone level as output label to construct a neural network surrogate model; using sample data from known wells to train the neural network surrogate model, and optimizing the model parameters through the backpropagation algorithm to obtain the trained sedimentary facies zone favorable quantitative characterization model; wherein, the output layer of the neural network surrogate model simultaneously includes a favorable comprehensive index output node and an uncertainty quantification output node.

[0035] Specifically, the pre-defined favorable area classifications are categorized labels based on the actual shale gas indications and production test data of known wells. Shale gas indication data from drilled wells within the study area are collected, including measured gas content, total hydrocarbon values, and test production. Wells are then classified into three categories: Category I favorable area wells (good gas content, test production reaching industrial standards), Category II favorable area wells (moderate gas content, good indications but not meeting industrial standards), and Category III unfavorable area wells (poor gas content or no indications). These classification labels serve as the target output for neural network training.

[0036] The neural network surrogate model employs a nonlinear mapping model constructed using a multi-layer feedforward neural network structure. The model consists of an input layer, hidden layers, and an output layer. The number of nodes in the input layer equals the number of indicators in the core evaluation index system. Two hidden layers are configured: the first hidden layer has twice the number of nodes as the input layer, and the second hidden layer has the same number of nodes as the first hidden layer. The activation function for the hidden layers is a linear rectified function. The output layer contains two nodes, corresponding to the comprehensive advantage index and the uncertainty quantification parameter, respectively. The output node for the comprehensive advantage index uses a sigmoid activation function to map the output value to the 0-1 range, while the output node for the uncertainty quantification parameter uses a linear activation function.

[0037] Please see Figure 3The specific implementation process of training a neural network surrogate model using sample data from known wells is as follows: First, the core evaluation index data of known wells are used as input samples, and the corresponding favorable zone level labels are used as target outputs to construct a training dataset. The input samples are normalized so that the numerical range of each index is unified to the interval between 0 and 1. The processed sample data is input into the neural network, and the predicted output is obtained through forward propagation. The loss function value between the predicted output and the true label is calculated. The loss function adopts a weighted combination of mean squared error and cross-entropy to simultaneously optimize the regression accuracy of the favorable index and the accuracy of the level classification. The gradient of the loss function with respect to the weights and biases of each layer is calculated through the backpropagation algorithm, and the network parameters are updated using the Adam optimizer. The training dataset is randomly divided into a training set and a validation set, and multiple rounds of iterative training are performed. Each iteration uses a small batch of samples to update the parameters until the loss function value of the validation set no longer decreases or reaches the preset maximum number of iterations. Training is stopped and the model parameters are saved to obtain a trained quantitative characterization model of the favorable sedimentary facies zone.

[0038] In this embodiment, the calculation of the comprehensive index of the favorable sedimentary facies zone of the area to be tested includes the following: 1. Input the core evaluation indicators of the region to be tested into the trained neural network proxy model; 2. The favorable index of sedimentary facies zone in the test area is obtained from the output node of the favorable index through forward propagation calculation of the model. The favorable index is a dimensionless value with a range of 0 to 1. The larger the value, the more favorable the shale accumulation conditions.

[0039] The forward propagation calculation process is as follows: The core evaluation index vector of the region to be measured is denoted as... ,in The number of indicators. First, calculate the output of the first hidden layer: in: For the first hidden layer The output value of each node, For the input layer The node is connected to the first hidden layer. The connection weight of each node For the first hidden layer The bias term of each node, This represents the number of nodes in the first hidden layer. It is a linear rectification function, defined as follows: Its function is to introduce nonlinearity.

[0040] Then calculate the output of the second hidden layer: in: Indicates the second hidden layer The output value of each node, Indicates the first hidden layer. The node is connected to the second hidden layer. The connection weight of each node Indicates the second hidden layer The bias term of each node, This represents the number of nodes in the second hidden layer.

[0041] Finally, calculate the overall advantage index of the output layer: in: The comprehensive index of the favorable conditions for sedimentary facies zones in the area to be tested. Indicates the second hidden layer The connection weights from each node to the output node This is the bias term for the output node. The Sigmoid function is defined as follows: Its function is to map the result of linear combination to a range of 0 to 1, making it easier to interpret as a probability or level of favorableness. This output value is the comprehensive index of favorableness of sedimentary facies zones in the area under test. The closer the value is to 1, the more favorable the shale accumulation conditions in the area, and the closer it is to 0, the worse the conditions.

[0042] In this embodiment, the uncertainty of the quantification prediction result includes the following: 1. The Bayesian neural network method is used to assign the weight parameters of the neural network surrogate model to the prior distribution, and the posterior distribution of the weights is learned through variational inference; 2. In the prediction phase, multiple samples are taken from the weighted posterior distribution, and multiple forward propagations are performed to obtain a set of prediction samples for the overall advantage index; 3. Calculate the range of uncertainty of the prediction results based on the statistical distribution of the prediction samples.

[0043] Specifically, assigning a prior distribution to the weight parameters of the neural network surrogate model is implemented by treating all trainable parameters in the network (including weights and biases) as random variables and assigning them a prior distribution. This embodiment uses a distribution with a mean of 0 and a variance of... The Gaussian distribution is used as a priori, that is, for any parameter Its prior probability density function is ,in The mean is variance is It follows a normal distribution. These are hyperparameters, set based on experience, for example, they can be taken as follows: .

[0044] The posterior distribution of the learned weights is inferred through variational inference, i.e., finding a variational distribution. To approximate the true posterior distribution ,in The training data is used. The variational distribution adopts an independent Gaussian distribution, meaning that each parameter... The variational distribution is ,here and These are the variational parameters that need to be learned; remember them. The goal of variational inference is to make the variational distribution as close as possible to the true posterior. Minimizing the KL divergence between the variational distribution and the true posterior is equivalent to maximizing the lower bound of evidence. in: The lower bound of evidence is represented by a variational parameter. The function; Traverse the training samples; Indicates variational distribution Below, sample The expected value of the log-likelihood measures how well the model fits the data. The KL divergence represents the difference between the variational distribution and the prior distribution, measuring the discrepancy between the two and serving a regularization function. During optimization, stochastic gradient descent is used to maximize... The variational distribution after training is obtained. That is, the approximation of the weighted posterior distribution.

[0045] In the prediction phase, the specific implementation of multiple samplings from the weighted posterior distribution is as follows: for the region to be tested, repeat the process. Next (in this embodiment, take) Prediction. Each time, from the trained variational distribution... Randomly sample a set of weight parameters; specifically, for each parameter... From the normal distribution A sample value is extracted from the network and combined to form a complete set of network weights. The core evaluation indicators of the region to be tested are input into the network, and the favorable comprehensive index is calculated according to the aforementioned forward propagation process to obtain the [number of] weights. Predicted value from the second sampling ,in Finally, a set of predicted samples was obtained. .

[0046] The specific implementation of calculating the uncertainty range of the prediction results based on the statistical distribution of the prediction samples is as follows: Calculate the mean of this set of predicted samples as the final composite favorability index point estimate: in The sample mean represents the central tendency of the overall favorability index. The sample standard deviation is calculated as a measure of uncertainty. in This represents the sample standard deviation; a larger value indicates greater volatility in the prediction results, i.e., higher uncertainty. The 95% confidence interval can then be calculated as follows: This interval indicates that the true value has a 95% probability of falling within this range. You can also output quantiles, such as the 5th quantile and the 95th quantile, denoted as [insert quantile here]. and The range of uncertainty is This range of uncertainty reflects the reliability of the model's predictions. The smaller the standard deviation or the narrower the range, the more stable and reliable the prediction results are, which helps in assessing risks in exploration decisions.

[0047] S4. Input the core evaluation indicators of the area to be tested into the quantitative characterization model of favorable sedimentary facies zones for prediction, and output the favorable sedimentary facies zone level of shale in the area to be tested and the uncertainty range of the prediction results.

[0048] First, the spatial location information of the area to be tested is obtained. Core evaluation index data, retained after first, second, and third-level screening, is extracted from the geological data of this area to form the input feature vector of the area. This input feature vector is standardized using the same normalization parameters as the training samples, ensuring data distribution consistency. The standardized input feature vector is then input into a trained sedimentary facies favorability quantitative characterization model. The model performs predictions based on the aforementioned Bayesian neural network forward propagation mechanism. During the prediction process, multiple samples are taken from the posterior distribution of the training weights. Each sample yields a set of network weight parameters, and a forward propagation calculation is performed to obtain a predicted value for the comprehensive favorability index. This process is repeated one hundred times to obtain one hundred comprehensive favorability index prediction samples. Based on this set of prediction samples, statistical characteristics are calculated: the sample mean is taken as the final comprehensive favorability index of the area to be tested; the sample standard deviation is taken as a measure of uncertainty in the prediction results; and a 95% confidence interval is calculated, i.e., the sample mean plus or minus 1.96 times the standard deviation, as the specific numerical expression of the uncertainty range. The favorable sedimentary facies zone level of the shale in the test area is determined based on the average of the comprehensive favorable index. The level classification criteria are consistent with the favorable area levels preset during the model training phase: when the average comprehensive favorable index is greater than or equal to 0.8, it is classified as a Class I favorable area, indicating excellent shale accumulation conditions; when the average comprehensive favorable index is greater than or equal to 0.6 and less than 0.8, it is classified as a Class II favorable area, indicating moderate shale accumulation conditions; when the average comprehensive favorable index is greater than or equal to 0.4 and less than 0.6, it is classified as a Class III favorable area, indicating poor shale accumulation conditions; and when the average comprehensive favorable index is less than 0.4, it is classified as an unfavorable area, indicating no shale gas accumulation conditions. The final output prediction results include the favorable area level, the average comprehensive favorable index, and the uncertainty range of the test area. The uncertainty range is presented in the form of a 95% confidence interval to guide the assessment of the reliability of the prediction results and decision-making risks during exploration deployment.

[0049] Example 2 An embodiment of the shale favorable sedimentary facies zone prediction system of the present invention includes the following: The multi-source geological data module is used to acquire multi-source geological data of the target area. The multi-source geological data includes initial evaluation indicators that reflect the characteristics of sedimentary environment, source rock quality, and tectonic preservation conditions. The core evaluation index system determination module is used to perform step-by-step screening of multi-source geological data based on preset geological classification screening rules to obtain the core evaluation index system; wherein, the step-by-step screening includes: The first level of screening identifies sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, selecting areas that belong to the preset favorable sedimentary facies types. The second level of screening involves selecting indicators whose correlation with each initial evaluation index meets a preset threshold, based on the correlation strength between each index and the macroscopic characterization parameters of shale gas content. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. The sedimentary facies zone favorable quantitative characterization model construction module is used to construct a sedimentary facies zone favorable quantitative characterization model based on the core evaluation index system; the sedimentary facies zone favorable quantitative characterization model is used to calculate the comprehensive index of sedimentary facies zone favorableness in the area to be tested and to quantify the uncertainty of the prediction results. The prediction module is used to input the core evaluation indicators of the area to be tested into the quantitative characterization model of favorable sedimentary facies zones for prediction, and output the favorable sedimentary facies zone level of the shale in the area to be tested and the uncertainty range of the prediction results.

[0050] For specific limitations regarding the system, please refer to the method limitations described above, which will not be repeated here. Each module in the above system can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in the computer device in hardware form, or stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0051] Example 3 The present invention provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the steps of the method described above.

[0052] Example 4 The present invention provides a processor for running a computer program, wherein the computer program performs the steps of the method described above when it runs.

[0053] It is understood that those skilled in the art can combine various implementation methods in the above embodiments under the guidance of the above examples to obtain technical solutions with multiple implementation methods.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting favorable sedimentary facies zones in shale, characterized in that, include: Acquire multi-source geological data for the target area, including initial evaluation indicators reflecting sedimentary environment characteristics, source rock quality characteristics, and tectonic preservation conditions. The multi-source geological data are screened step-by-step based on preset geological grading rules to obtain a core evaluation index system; wherein, the step-by-step screening includes: The first level of screening identifies sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, selecting areas that belong to the preset favorable sedimentary facies types. The second level of screening involves selecting indicators whose correlation with each initial evaluation index meets a preset threshold, based on the correlation strength between each index and the macroscopic characterization parameters of shale gas content. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. A quantitative characterization model for the favorable sedimentary facies zone is constructed based on the core evaluation index system. This model is used to calculate the comprehensive index of the favorable sedimentary facies zone in the area to be tested and to quantify the uncertainty of the prediction results. The core evaluation indicators of the area to be tested are input into the favorable sedimentary facies zone quantitative characterization model for prediction, and the favorable sedimentary facies zone level of the shale in the area to be tested and the uncertainty range of the prediction results are output.

2. The method for predicting favorable sedimentary facies zones in shale according to claim 1, characterized in that, The process of identifying sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, and then selecting areas belonging to preset favorable sedimentary facies types, includes: Based on single-well facies analysis, according to the lithological assemblage, sedimentary structures and paleontological markers in the well logging data, the sedimentary facies types developed vertically in a single well are identified, and favorable intervals belonging to the preset favorable sedimentary facies types are delineated; the preset favorable sedimentary facies types include fine-grained sedimentary facies formed in deep-water anoxic environments, and the fine-grained sedimentary facies include platform facies, deep-water shelf facies and basin facies. Based on seismic facies analysis, according to seismic reflection structure, external geometry and seismic attribute characteristics, the planar distribution of sedimentary facies types is identified, and favorable areas belonging to the preset favorable sedimentary facies types are delineated. The favorable strata identified by a single well and the favorable areas delineated by seismic activity are cross-calibrated to determine the target spatial range that passes the first-level screening.

3. The method for predicting favorable sedimentary facies zones in shale according to claim 1, characterized in that, The method of selecting indicators whose correlation with each initial evaluation index meets a preset threshold based on the correlation strength between each initial evaluation index and the macroscopic characterization parameters of shale gas content includes: Obtain macroscopic characterization parameters of shale gas content from known wells, wherein the macroscopic characterization parameters of shale gas content include at least one of measured gas content, total hydrocarbon value measured by gas analysis, and shale gas indication thickness; Calculate the correlation coefficients between each initial evaluation index and the macroscopic characterization parameters of shale gas content; Initial evaluation indicators whose absolute value of the correlation coefficient is greater than or equal to the preset correlation threshold are identified as indicators that pass the second-level screening.

4. The method for predicting favorable sedimentary facies zones in shale according to claim 1, characterized in that, Based on the degree of fit between each indicator and the key elements of structural preservation conditions, indicators that can synergistically characterize the effectiveness of structural preservation are selected, including: To obtain key elements of structural preservation conditions, the key elements of structural preservation conditions include at least one of fault sealing index, distance from magmatic rock mass, formation pressure coefficient and effective thickness of the top plate; Set structural preservation effectiveness constraints for each key element; the effectiveness constraints are determined based on the regional structural evolution characteristics and the preservation effect of drilled wells. Each indicator obtained after the first and second level screening is matched with the constraints one by one; if the value of an indicator meets the constraints of all key elements, the indicator is determined to pass the third level screening; if the value of an indicator does not meet the constraints of any key element, the indicator is determined to fail the third level screening. All matching indicator combinations are identified as the indicator set that passes the third-level screening.

5. The method for predicting favorable sedimentary facies zones in shale according to claim 1, characterized in that, The construction of a favorable quantitative characterization model for sedimentary facies includes: Using the core evaluation index system as input features and the preset favorable zone level as output label, a neural network proxy model is constructed. The neural network surrogate model is trained using sample data from known wells, and the model parameters are optimized using the backpropagation algorithm to obtain a well-trained quantitative characterization model of sedimentary facies zones. The output layer of the neural network proxy model includes both a comprehensive advantage index output node and an uncertainty quantification output node.

6. The method for predicting favorable sedimentary facies zones in shale according to claim 5, characterized in that, The calculation of the comprehensive index of the favorable sedimentary facies zone of the area to be tested includes: Input the core evaluation indicators of the region to be tested into the trained neural network proxy model; The favorable index of sedimentary facies zone in the test area is obtained from the output node of the favorable index through forward propagation calculation of the model. The favorable index is a dimensionless value with a value range of 0 to 1. The larger the value, the more favorable the shale accumulation conditions.

7. The method for predicting favorable sedimentary facies zones in shale according to claim 5, characterized in that, The uncertainty of the quantitative prediction results includes: The Bayesian neural network method is used to assign the weight parameters of the neural network surrogate model to the prior distribution, and the posterior distribution of the weights is learned through variational inference. In the prediction phase, multiple samples are taken from the weighted posterior distribution, and multiple forward propagations are performed to obtain a set of prediction samples for the overall advantage index. The uncertainty range of the prediction results is calculated based on the statistical distribution of the predicted samples.

8. A system for predicting favorable sedimentary facies zones in shale, characterized in that, The method for predicting favorable sedimentary facies zones in shale as described in any one of claims 1 to 7 includes: The multi-source geological data module is used to acquire multi-source geological data of the target area. The multi-source geological data includes initial evaluation indicators that reflect the characteristics of sedimentary environment, source rock quality, and tectonic preservation conditions. The core evaluation index system determination module is used to perform step-by-step screening of the multi-source geological data based on preset geological grading rules to obtain the core evaluation index system; wherein, the step-by-step screening includes: The first level of screening identifies sedimentary facies types based on lithological assemblage, sedimentary structures, and paleontological markers, selecting areas that belong to the preset favorable sedimentary facies types. The second level of screening involves selecting indicators whose correlation with each initial evaluation index meets a preset threshold, based on the correlation strength between each index and the macroscopic characterization parameters of shale gas content. The third level of screening involves selecting indicators that can synergistically characterize the effectiveness of structural preservation based on the degree of compatibility between each indicator and the key elements of structural preservation conditions. The sedimentary facies zone favorable quantitative characterization model construction module is used to construct a sedimentary facies zone favorable quantitative characterization model based on the core evaluation index system; the sedimentary facies zone favorable quantitative characterization model is used to calculate the comprehensive index of sedimentary facies zone favorableness in the area to be tested and to quantify the uncertainty of the prediction results. The prediction module is used to input the core evaluation indicators of the area to be tested into the favorable quantitative characterization model of the sedimentary facies zone for prediction, and output the favorable sedimentary facies zone level of the shale in the area to be tested and the uncertainty range of the prediction results.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the method for predicting favorable sedimentary facies zones of shale as described in any one of claims 1 to 7.

10. A processor, characterized in that, Used to run a computer program, which, when running, executes the method for predicting favorable sedimentary facies zones of shale as described in any one of claims 1 to 7.

Images