A marine shale gas resource abundance prediction method, device and equipment

Abstract

This specification relates to the field of shale gas resource development technology, and particularly to a method, apparatus, and equipment for predicting the abundance of marine shale gas resources. The method includes: selecting multiple control parameters related to shale gas resource abundance from multiple geological parameters based on shale gas resource abundance data and corresponding geological data of a marine shale scale zone; classifying marine shale into various types based on a geological oil and gas mechanism model; determining one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance for each marine shale type; constructing a corresponding shale gas resource abundance prediction model using a multivariate nonlinear regression algorithm based on the primary and secondary control parameters for each marine shale type; and predicting resource abundance by applying the corresponding shale gas resource abundance prediction model according to the shale type and geological data of the target marine shale area.

Description

Technical Field

[0001] The embodiments in this specification relate to the field of shale gas resource development technology, specifically to a method, apparatus, and equipment for predicting the abundance of marine shale gas resources. Background Technology

[0002] Currently, oil and gas exploration and development has shifted from conventional to unconventional oil and gas. Shale gas, as a potentially huge unconventional resource, has recoverable reserves of approximately 80 to 100 trillion cubic meters, making it a crucial area for ensuring oil and gas resource succession and energy security. Developing shale gas can effectively improve domestic oil and gas self-sufficiency, reduce dependence on imports, and promote energy diversification.

[0003] In recent years, shale gas exploration and development have made significant progress, with annual production now accounting for approximately 10% of total natural gas production, playing an increasingly important role in the energy structure. Marine shale gas accounts for nearly 70% of recoverable resources, making it the mainstay of shale gas development. However, marine shale gas exploration is still in its early stages, with relatively limited research and understanding, and an incomplete resource potential assessment system. To achieve sustainable and stable shale gas development, it is urgent to establish a systematic and efficient resource assessment methodology.

[0004] Resource abundance is a key parameter for evaluating the enrichment level of shale gas, referring to the amount of natural gas resources contained in a unit area or unit volume of shale. This parameter is directly related to the calculation of total resources, and its accurate prediction has significant scientific and economic value for site selection evaluation, reserve assessment, and development decisions.

[0005] Scholars both at home and abroad have conducted numerous studies on the prediction of conventional oil and gas resources, resulting in a variety of methods such as linear regression, neural networks, random forests, and gradient boosting trees.

[0006] However, due to the fundamental differences between conventional and unconventional oil and gas in terms of accumulation mechanisms and main controlling factors, the above methods are difficult to apply directly to the prediction of shale gas resource abundance. Furthermore, they often fail to effectively characterize the complex nonlinear relationship between geological factors and resource abundance, resulting in limited prediction accuracy and restricting the scientific nature and effectiveness of shale gas exploration and deployment. Summary of the Invention

[0007] The purpose of the embodiments in this specification is to provide a method, apparatus, and equipment for predicting the abundance of marine shale gas resources, so as to overcome the problem of limited prediction accuracy in existing methods.

[0008] To solve the above-mentioned technical problems, the specific technical solutions of the embodiments in this specification are as follows: On the one hand, the embodiments of this specification provide a method for predicting the abundance of marine shale gas resources, including: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, several control parameters related to shale gas resource abundance were selected from multiple geological parameters. Based on the geological oil and gas mechanism model, marine shale is classified into multiple types; For each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources; Based on the primary and secondary control parameters of each marine shale type, a corresponding shale gas resource abundance prediction model is constructed using a multivariate nonlinear regression algorithm. Based on the shale type and geological data of the marine shale target area, the corresponding shale gas resource abundance prediction model is applied to predict resource abundance.

[0009] On another front, embodiments of this specification provide a device for predicting the abundance of marine shale gas resources, comprising: The filtering module is used to filter out multiple control parameters related to shale gas resource abundance from multiple geological parameters based on shale gas resource abundance data and corresponding geological data of marine shale scale areas. The classification module is used to classify marine shale into multiple types based on geological oil and gas mechanism models; The determination module is used to determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources for each type of marine shale. The module is used to construct a corresponding shale gas resource abundance prediction model based on the main and secondary control parameters of each marine shale type, using a multivariate nonlinear regression algorithm. The prediction module is used to predict resource abundance based on the shale type and geological data of the marine shale target area, applying the corresponding shale gas resource abundance prediction model.

[0010] In another aspect, a computer device is provided, including a memory for storing computer programs and a processor for executing the computer programs to implement the above-mentioned method for predicting the abundance of marine shale gas resources.

[0011] As can be seen from the technical solutions provided in the embodiments of this specification above, these embodiments can, based on shale gas resource abundance data and corresponding geological data from marine shale scale areas, select multiple control parameters related to shale gas resource abundance from multiple geological parameters; classify marine shale into various types based on a geological oil and gas mechanism model; for each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance; construct a corresponding shale gas resource abundance prediction model using a multivariate nonlinear regression algorithm based on the primary and secondary control parameters for each type of marine shale; and predict resource abundance using the corresponding shale gas resource abundance prediction model according to the shale type and geological data of the marine shale target area. Through correlation analysis based on measured data from marine shale scale areas, control parameters significantly related to resource abundance are selected from numerous geological parameters, significantly improving the objectivity, efficiency, and repeatability of initial parameter selection, while effectively reducing the dimensionality of subsequent modeling and avoiding interference from irrelevant or redundant parameters. Building upon this foundation, a geological hydrocarbon mechanism model was introduced as the classification basis, dividing marine shale into multiple genetic types. This breaks through the traditional coarse classification model dominated by a single lithology or lithology, achieving an essential and refined classification based on the main control mechanism of hydrocarbon accumulation. This provides a precise geological framework for subsequent classification modeling, enabling the model to specifically characterize the unique enrichment patterns of different shale types. Furthermore, based on the classification, the primary and secondary control parameters for each shale type were identified, clarifying the relative importance of different geological factors in different types. This endows the prediction model with a clear geological logic and differentiated modeling focus, transforming the model construction process from a black-box parameter stacking to a structured process with primary and secondary factors and mechanism-driven characteristics, greatly enhancing the geological interpretability of the model. In addition, a multivariate nonlinear regression algorithm was used to construct the prediction model, directly overcoming the limitations of traditional linear methods in characterizing complex geological relationships, and enabling a more accurate fit between the nonlinear response of control parameters and resource abundance. Finally, by identifying the shale type in the target area, the corresponding prediction model was directly invoked for accurate prediction, providing timely and reliable quantitative basis for exploration decisions. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments or prior art of this specification, the accompanying drawings used in the description of the embodiments or prior art will be briefly introduced below.

[0013] Figure 1 This is a flowchart of a method for predicting the abundance of marine shale gas resources provided in the embodiments of this specification; Figure 2 This is a statistical diagram illustrating the frequency distribution of marine shale gas resource abundance provided in the embodiments of this specification; Figure 3This is a schematic diagram showing the fitting effect of the main control parameter function model for marine shale gas resource abundance provided in the embodiments of this specification; Figure 4 This is a schematic diagram illustrating the applicability test of the marine shale gas resource abundance prediction model provided in the embodiments of this specification; Figure 5 This is a schematic diagram of the structural composition of a marine shale gas resource abundance prediction device provided in the embodiments of this specification; Figure 6 This is a schematic diagram of the structural composition of the computer device provided in the embodiments of this specification. Detailed Implementation

[0014] The technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of this specification.

[0015] It should be noted that the terms "first," "second," etc., used in this specification, claims, and the foregoing drawings 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 described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, apparatus, product, or device 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 devices.

[0016] In some embodiments, marine shale can be shale formations deposited in marine environments during geological history. These shale formations occur in deep-to-semi-deep-water environments far from land, where the water is calm and highly reducing, which is conducive to the long-term accumulation and preservation of fine-grained sediments and organic matter. Marine shale generally features high organic matter content, continuous distribution, large thickness, and relatively stable lateral direction, making it an excellent geological carrier for shale gas generation and enrichment.

[0017] In some embodiments, shale gas resource abundance can be a core parameter for evaluating the degree of shale gas enrichment, representing the total geological resources of natural gas contained in a unit area or unit volume of shale layer, expressed in units of 100 million cubic meters per square kilometer or 100 million cubic meters per cubic kilometer. This parameter directly characterizes the enrichment intensity of natural gas within the shale layer. Resource abundance can be used to further estimate the total resources in the evaluation area, thus providing a quantitative basis for exploration site selection, reserve assessment, and development decisions. Its prediction accuracy directly affects exploration effectiveness and economic feasibility.

[0018] In some embodiments, geological parameters may include at least: TOC, Ro, effective thickness, brittle mineral content, clay mineral content, and effective thickness. Total organic carbon (TOC) can be the percentage of total organic carbon by mass in the rock, representing the hydrocarbon generation potential of the source rock and the material basis of shale gas resources. A higher TOC indicates greater hydrocarbon generation potential. Vitrin reflectance (Ro) can be used to measure the degree of thermal evolution of organic matter, reflecting the hydrocarbon generation stage of the source rock. Its value directly affects the type and quantity of natural gas generated and can be used to determine whether the shale has entered a stage of significant gas generation. Effective thickness can be the cumulative thickness of the shale layer capable of producing industrial shale gas flow; it characterizes the spatial scale of shale gas resources, and a greater thickness indicates greater resource potential. Brittle mineral content can be the volume or mass percentage of brittle mineral components such as quartz, feldspar, and carbonate rocks in the shale. A higher brittle mineral content indicates that the shale is more likely to form complex network fractures during fracturing, thereby improving the reservoir's permeability. The clay mineral content can refer to the amount of clay minerals such as illite, montmorillonite, and kaolinite in shale. Clay minerals not only affect the mechanical properties of shale (such as plasticity and expansibility) but are also closely related to its gas adsorption capacity. Their type and content have a significant impact on reservoir stimulation effects and gas occurrence states.

[0019] This specification provides an embodiment of a method for predicting the abundance of marine shale gas resources, referring to... Figure 1 The specific implementation includes the following steps: S101: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, several control parameters related to shale gas resource abundance are selected from multiple geological parameters.

[0020] In some embodiments, step S101 may specifically include: calculating a first correlation between each geological parameter and the shale gas resource abundance based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area; and selecting multiple control parameters from multiple geological parameters based on the first correlation corresponding to each geological parameter using a preset correlation threshold.

[0021] The specific implementation of step S101 is as follows: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, select multiple control parameters that are significantly related to the shale gas resource abundance from multiple geological parameters.

[0022] In some embodiments, the marine shale calibration zone can be a representative marine shale geological unit selected in an area with a high degree of exploration and relatively clear geological understanding. The abundance of shale gas resources in this unit has been assessed by a reliable method and can serve as a benchmark for model construction and validation.

[0023] Shale gas resource abundance data can be the quantitative value of the total geological resources of natural gas contained in a unit area or unit volume of shale within the above-mentioned scale zone.

[0024] Geological data can be a collection of various measured or interpretive information related to the aforementioned scale area, reflecting its formation conditions and current status. Geological parameters can be variables extracted from geological data to quantitatively characterize specific geological properties, such as TOC, Ro, effective thickness, brittle mineral content, clay mineral content, and effective thickness mentioned above.

[0025] In some embodiments, the first correlation can be a statistical measure used to quantify the strength and direction of the linear or monotonic relationship between two variables.

[0026] In some embodiments, each geological parameter (such as a TOC value sequence) can be paired with a corresponding shale gas resource abundance value sequence, and the correlation coefficient can be used for calculation.

[0027] Preferably, the Pearson correlation coefficient can be calculated simultaneously to measure linear relationships, and the Spearman rank correlation coefficient can be calculated to capture monotonic nonlinear relationships, thereby comprehensively assessing the correlation between parameters and resource abundance. This calculation process can be automated using statistical analysis software or programming tools, outputting one or more correlation coefficient values ​​for each geological parameter.

[0028] By calculating specific correlation coefficients, the candidate parameters with the strongest correlation to resource abundance can be accurately and quickly identified from dozens of potential geological parameters. This overcomes the subjective bias and omissions that may result from screening based solely on geological experience, providing a reliable data foundation for subsequent modeling and significantly improving the efficiency and scientific rigor of the parameter screening stage.

[0029] In some embodiments, the preset relevance threshold can be a numerical limit set in advance based on statistical significance requirements (such as p-value <0.05 or <0.01) and actual geological significance, such as an absolute value of 0.5 or 0.6.

[0030] In some embodiments, the control parameters may be a subset of parameters that, through the threshold test described above, are considered to have a statistically significant and geologically important association with shale gas resource abundance.

[0031] In some embodiments, the first correlation (absolute value or combined with significance p-value) of each geological parameter can be compared with a preset correlation threshold. Parameters whose absolute correlation value is greater than or equal to the threshold and require statistical significance are determined to have a substantial impact on resource abundance and are thus selected as control parameters; otherwise, they are eliminated.

[0032] By setting thresholds, interfering parameters with weak or uncertain relationships to resource abundance can be automatically excluded, preventing their introduction into the prediction model. This ensures that the final model consists only of key driving factors, simplifying the model structure, reducing the risk of overfitting, and enhancing the model's physical interpretability and stability. This lays a solid foundation for building high-precision, highly generalizable prediction models in the future.

[0033] S102: Based on the geological oil and gas mechanism model, marine shale is classified into multiple types.

[0034] In some embodiments, the geological hydrocarbon mechanism model can be a conceptual or quantitative model based on the fundamental principles of hydrocarbon geology, used to describe and explain the inherent laws governing the entire process of hydrocarbon generation, migration, accumulation, and preservation. It can be used to link observed geological phenomena with the underlying dynamic processes, providing a theoretical framework for classification and prediction.

[0035] In some embodiments, the above-mentioned geological oil and gas mechanism model includes at least a shale diagenesis mechanism model, a paleoenvironmental control mechanism model, and an organic matter enrichment mechanism model.

[0036] Shale diagenesis model characterizes the physical and chemical changes (i.e., diagenesis) that shale undergoes after deposition under the influence of temperature, pressure, and pore fluids, and its control mechanism on shale mineral composition, pore structure, rock mechanical properties, and hydrocarbon generation potential.

[0037] The paleoenvironmental control mechanism model is used to characterize the paleogeography, paleoclimate, paleoceanography (such as water depth, salinity, redox conditions), and paleontology during shale deposition, and to elucidate how these factors control the type, rate, and input and early preservation of sediments.

[0038] Organic matter enrichment mechanism models characterize the main controlling factors and processes of organic matter enrichment and preservation during sedimentation and burial, involving the balance between productivity (organic matter supply), dilution (inorganic sediment input) and preservation conditions (redox state).

[0039] In some embodiments, step S102 may specifically include: classifying marine shale into multiple types based on shale diagenesis mechanism models, paleoenvironmental control mechanism models, and organic matter enrichment mechanism models.

[0040] In some embodiments, shale diagenesis mechanism models, paleoenvironmental control mechanism models, and organic matter enrichment mechanism models can be integrated and applied. Specifically, for a target marine shale system, the following are analyzed simultaneously: (a) the diagenetic environment and fluid history indicated by the diagenetic mineral assemblage and sequence; (b) the paleoenvironmental background reflected by sedimentary structures, fossils, and geochemical indicators; and (c) the enrichment mechanism reflected by the abundance, type, and distribution pattern of organic matter. For example, the redox state of paleowater bodies can be determined by geochemical indicators (such as the enrichment degree of Mo and U) (paleoenvironmental model), and combined with specific diagenetic minerals formed under these conditions (such as pyrite, diagenetic model), the organic matter preservation efficiency can be jointly inferred (enrichment model).

[0041] In some embodiments, based on the results of the aforementioned collaborative analysis, coupling indicators that reflect genetic differences are extracted as classification criteria. Specifically, marine shale can be classified into basic types such as open marine shale and confined marine shale based on the logical chain of correlation between paleowater energy and circulation (controlled by paleoenvironmental model) - main controlling factors of organic matter enrichment (reflected by enrichment model) - diagenetic products and rock mechanical response (determined by diagenetic model). The former may be characterized by better water circulation, productivity-driven organic matter enrichment, and strong siliceous diagenesis; the latter may be characterized by limited water, preservation conditions-driven enrichment, and more significant clay mineral or calcareous diagenesis. This classification process can be implemented through expert knowledge system rules or based on multi-index clustering analysis algorithms.

[0042] Unlike empirical classifications based solely on lithology or geochemical indicators, this classification, based on the intrinsic mechanisms controlling shale generation, storage, and conservation capabilities, directly links the classification results to the fundamental reasons for differences in resource endowment, making the classification more scientifically sound and predictively significant. Furthermore, shale types classified through mechanistic models share similar geological evolution histories and gas-controlling patterns. This ensures that subsequent prediction models can be tailored to each shale type (i.e., using different master control parameters or model structures), greatly improving the model's precision and prediction accuracy, overcoming the inherent weakness of one-size-fits-all models in adapting to geological heterogeneity. Moreover, explicitly embedding the mechanistic model into the classification process transforms the entire prediction method from a purely data black box into one deeply integrated with domain knowledge, significantly enhancing the credibility, interpretability, and theoretical extrapolation capabilities for new areas (data-sparse regions).

[0043] S103: For each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources.

[0044] In some embodiments, the master control parameter can be a geological variable that plays a primary and decisive role in controlling the spatial distribution and enrichment scale of shale gas resources under a specific geological type, and its changes can directly and significantly cause systematic changes in resource abundance.

[0045] In some embodiments, secondary control parameters may be parameters that have a significant impact on resource abundance, but whose role depends to some extent on or is limited by the basic geological background created by the primary control parameters, and play an optimization, regulation or local control role.

[0046] In some embodiments, the aforementioned marine shale includes at least open-sea marine shale and confined-sea marine shale.

[0047] In some embodiments, open marine shale can be shale deposited in marine environments (such as continental shelves or open platforms) with good connectivity between the water body and the open ocean and relatively smooth water circulation. Its organic matter input is dominated by surface water productivity, the sedimentary environment is relatively oxidative, and diagenesis is mainly characterized by siliceous cementation.

[0048] In some embodiments, confined marine shale can be shale deposited in marine environments (such as bays, lagoons, and confined platforms) where water flow is limited and the environment is semi-enclosed or closed. Its organic matter enrichment is often dominated by favorable preservation conditions (oxygen-deficient environment), the sedimentary environment is highly reducing, and clay minerals or early carbonate cementation are more developed during diagenesis.

[0049] In some embodiments, the aforementioned multiple control parameters include one or more source rock control parameters and one or more reservoir control parameters.

[0050] In some embodiments, source rock control parameters can be parameters that directly characterize the gas generation capacity and potential of shale, and may include TOC and Ro. TOC can characterize the material abundance of hydrocarbon generation, while Ro can characterize the hydrocarbon generation stage and gas generation amount reached by the thermal evolution of organic matter.

[0051] In some embodiments, reservoir control parameters can be parameters characterizing shale reservoir space and permeability, including effective thickness, brittle mineral content, and porosity. Effective thickness determines the macroscopic volume of the reservoir fluid, brittle mineral content affects reservoir modifiability, and porosity is directly related to the size of the reservoir space.

[0052] In some embodiments, step S103 may specifically include: for open marine shale, using source rock control parameters as primary control parameters and reservoir control parameters as secondary control parameters; for confined marine shale, using reservoir control parameters as primary control parameters and source rock control parameters as secondary control parameters.

[0053] In some embodiments, for each type of marine shale identified in step S102, the dominant shale gas enrichment mechanism can be analyzed. Specifically, for open-ocean marine shale, the key to resource formation lies in whether there is sufficient organic matter and sufficient maturity to generate large amounts of natural gas, i.e., source-controlled characteristics are more prominent. For confined-ocean marine shale, due to its superior organic matter preservation conditions (oxygen deficiency), the total organic matter abundance (TOC) is generally high. In this case, whether the generated natural gas can be effectively retained in the shale and form industrial enrichment depends more on whether the shale itself has sufficient effective reservoir space (thickness) and / or whether it can be modified to form an effective seepage network, i.e., reservoir-controlled characteristics are more significant.

[0054] By clearly distinguishing the decisive and regulatory roles of parameters in specific types from the perspectives of geological genesis and gas-controlling mechanisms, this approach overcomes the limitations of traditional modeling that treats all parameters equally or only ranks them by correlation strength. This provides a direct and reliable geological basis for subsequent prediction model construction, including weighted processing, hierarchical modeling, or setting different regularization constraints.

[0055] In some embodiments, based on the above-mentioned determination of causes and mechanisms, various control parameters can be logically assigned primary and secondary values. Specifically, for open marine shale: source rock control parameters that directly determine the scale and potential of hydrocarbon generation are determined as primary control parameters. Reservoir control parameters that affect gas storage and seepage are determined as secondary control parameters. For confined marine shale: reservoir control parameters that determine storage and reproducibility are determined as primary control parameters. Source rock control parameters are determined as secondary control parameters because, in this type, a higher TOC is a basic prerequisite, but its variation may have a smaller marginal impact on resource abundance than reservoir conditions.

[0056] By categorizing parameters into primary and secondary controlling factors, the established models can more accurately reflect the core enrichment contradictions of different types of shale. For example, for open marine models, the focus is on capturing the strong nonlinear relationship between TOC / Ro and abundance; for confined marine models, the focus is on characterizing the relationship between thickness, brittle minerals, and abundance. This targeted modeling strategy allows the models to not only fit better in known data areas but also exhibit greater geological consistency in their prediction logic when extrapolating to new areas with similar geological conditions, thereby improving the model's generalization performance and reliability.

[0057] S104: Based on the primary and secondary control parameters of each marine shale type, a multivariate nonlinear regression algorithm is used to construct the corresponding shale gas resource abundance prediction model.

[0058] In some embodiments, step S104 may specifically include: constructing a single-factor shale gas resource abundance prediction model for each control parameter based on shale gas resource abundance data and corresponding geological data of marine shale scale zones; calculating a fifth correlation between each control parameter and other control parameters, including: calculating a sixth correlation between each active parameter and the plurality of secondary control parameters; calculating a seventh correlation between each secondary control parameter and the plurality of primary control parameters; and, based on the fifth correlation of each control parameter, using a multivariate nonlinear regression algorithm to fuse the single-factor shale gas resource abundance prediction models corresponding to the primary and secondary control parameters for each marine shale type, to obtain a shale gas resource abundance prediction model corresponding to that marine shale type.

[0059] In some embodiments, the multivariate nonlinear regression algorithm can be a statistical modeling method that can simultaneously handle multiple independent variables (primary and secondary control parameters) and allow a nonlinear relationship between the independent variables and the dependent variable (shale gas resource abundance), describing this complex relationship by fitting the optimal mathematical function.

[0060] In some embodiments, a single-factor shale gas resource abundance prediction model may be a simplified prediction model that considers only the relationship between a geological parameter and resource abundance, used to characterize the independent influence pattern of that parameter.

[0061] In some embodiments, the fifth, sixth, and seventh correlations can be quantitative indicators of correlation between different combinations of variables, used to measure the degree and direction of mutual influence between parameters, the difference being that they target different sets of parameters.

[0062] In some embodiments, based on shale gas resource abundance data and corresponding geological data from marine shale grading zones, a univariate relationship model between each selected control parameter (regardless of whether it is a primary or secondary control) and resource abundance can be established. Specifically, for each control parameter, its numerical sequence is fitted to the corresponding resource abundance value sequence. Various basic nonlinear function forms can be attempted for fitting, and the function that best characterizes the relationship between the two can be selected as the single-factor shale gas resource abundance prediction model for that parameter through methods such as calculating goodness of fit and residual analysis. For example, it may be determined that TOC and abundance follow a power function relationship, while effective thickness and abundance follow another functional relationship.

[0063] By accurately characterizing the independent contribution pattern and mathematical form of each geological control parameter to resource abundance, optimized basic units are provided for subsequent integrated modeling. This avoids the inappropriate linear assumptions that may be introduced by directly using the original parameters in the multivariate model, ensuring that the influence of each parameter enters the final model in a nonlinear form that best conforms to geological laws.

[0064] In some embodiments, a fifth correlation degree can be calculated between each control parameter and other control parameters. Specifically, a sixth correlation degree can be calculated specifically between each primary control parameter and all secondary control parameters to reveal the potential influence or coupling path of the dominant factor on the secondary factors. Similarly, a seventh correlation degree can be calculated specifically between each secondary control parameter and all primary control parameters to assess the degree to which the secondary factors are constrained by or synergistic with the dominant factors. These calculations can be performed using correlation coefficient matrix analysis.

[0065] By revealing the specific correlation structure (sixth and seventh correlations) between the primary and secondary control parameters, this hierarchical dependency can be addressed in a targeted manner during model construction, such as by introducing interaction terms or employing specific weighting strategies in the regression. This effectively solves the problem of unstable model coefficient estimation or difficulty in interpretation that may result from neglecting structured collinearity among parameters in traditional multiple regression, thereby enhancing the robustness and geological reliability of the model.

[0066] In some embodiments, a multivariate nonlinear regression algorithm can be used to fuse the single-factor shale gas resource abundance prediction models with each parameter obtained above, as well as the results of multi-level correlation analysis (sixth and seventh correlation degrees) between them. Specifically, the outputs of each single-factor model (or their defined functional forms) can be used as new input features. At the same time, the correlation analysis results can be used to guide the algorithm settings (e.g., introducing interaction terms for strongly correlated parameter pairs, or adjusting regularization constraints according to the dominant position of the master parameter). Through regression fitting, a unified mathematical expression that integrates the nonlinear effects of all parameters and considers the hierarchical coupling relationship between parameters is finally obtained, which is the final shale gas resource abundance prediction model corresponding to this marine shale type.

[0067] By first establishing a nonlinear single-factor model and then fusing them to construct a comprehensive model, the final model accurately captures complex nonlinear responses while retaining a clear mathematical form. Its parameters possess clear geophysical meaning, achieving a combination of machine learning accuracy and the interpretability of traditional geological models. Furthermore, by systematically analyzing and incorporating the multi-level correlation structure between parameters, the model effectively avoids overfitting and multicollinearity. This makes the model's predictive behavior more stable and reliable when facing new area data outside the training data range, significantly enhancing the method's generalization value. Moreover, the entire construction process is not purely data-driven but deeply integrates geological understanding (the division of primary / secondary control parameters) and the analysis of the intrinsic connections within the geological system (correlation between parameters). This makes the final model not only mathematically optimal but also geologically logically sound, providing a more scientific and reliable quantitative tool for shale gas exploration decision-making.

[0068] In some embodiments, the above-mentioned single-factor shale gas resource abundance prediction model, which uses a multivariate nonlinear regression algorithm to fuse the primary and secondary control parameters corresponding to each marine shale type based on the fifth correlation of each control parameter, may further include: constructing a first prior constraint matrix to characterize the coupling relationship between the primary and secondary control parameters based on the sixth and seventh correlations; constructing a second prior constraint matrix to characterize the synergistic relationship between source and reservoir parameters if the sixth or seventh correlation is greater than a preset coupling threshold; and using a multivariate nonlinear regression algorithm to fuse the single-factor shale gas resource abundance prediction model corresponding to the primary and secondary control parameters corresponding to each marine shale type based on the first and second prior constraint matrices.

[0069] In some embodiments, the first prior constraint matrix can be used to characterize the coupling strength between the primary control parameters and the secondary control parameters. The construction process of this matrix includes: performing a Cartesian product operation on the set of primary control parameters and the set of secondary control parameters; for each parameter pair, using its corresponding sixth or seventh correlation degree as the matrix element value, forming a two-dimensional matrix with a dimension equal to the number of primary control parameters multiplied by the number of secondary control parameters. Each element in this matrix quantifies the degree of coupling between a specific primary control parameter and a specific secondary control parameter; the larger the absolute value of the element, the stronger the correlation between the two in terms of geological genesis. The first prior constraint matrix is ​​introduced as a regularization constraint term into the loss function of the multivariate nonlinear regression algorithm. In the loss function, the first prior constraint matrix imposes correlation constraints on the regression coefficients of the primary and secondary control parameters, ensuring that parameter pairs with high coupling strength maintain geological consistency in the direction of coefficient change during model training. Based on the loss function after introducing the first prior constraint matrix, a nonlinear regression fusion is performed on the single-factor shale gas resource abundance prediction model corresponding to the primary and secondary control parameters to obtain the shale gas resource abundance prediction model corresponding to this marine shale type.

[0070] By constructing a first prior constraint matrix and explicitly incorporating the coupling relationship into the loss function, the model training process is guided by geological coupling principles. When there is a strong coupling relationship between the primary and secondary control parameters, this constraint matrix can guide the regression coefficients to change in the same direction, avoiding the problems of opposite signs or excessive oscillations that may occur under pure data-driven approaches. This directly improves the stability and geological interpretability of the model parameter estimation, enabling the final prediction model to not only have high fitting accuracy but also whose internal parameter relationships conform to actual geological understanding, thus enhancing the model's generalization credibility in unknown regions.

[0071] In some embodiments, a coupling threshold can be used to determine whether there is a significant synergistic effect between the primary and secondary control parameters. The absolute values ​​of the sixth correlation degree between each primary control parameter and each secondary control parameter, and the absolute values ​​of the seventh correlation degree between each secondary control parameter and each primary control parameter, are compared with the coupling threshold. When any absolute value of the sixth or seventh correlation degree is greater than the coupling threshold, the parameter pair is marked as a high-synergistic parameter pair. Based on all marked high-synergistic parameter pairs, a second prior constraint matrix is ​​constructed. The second prior constraint matrix is ​​a sparse matrix, where non-zero elements correspond to high-synergistic parameter pairs, and element values ​​are the original values ​​of the sixth or seventh correlation degree corresponding to that parameter pair. This matrix is ​​used to characterize the significant synergistic effect between the source and store parameters, i.e., the special interaction relationships represented by parameter pairs whose coupling strength exceeds the threshold. Based on the second prior constraint matrix, interaction terms for high-synergistic parameter pairs are automatically generated in the multivariate nonlinear regression model. The mathematical form of the interaction term is the product of the output values ​​of the single-factor prediction model corresponding to the two parameters, or a power combination of the original values ​​of the two parameters. The generated interaction terms are incorporated as new features into the regression model and, together with the original single-factor models corresponding to the main and secondary control parameters, participate in the multivariate nonlinear regression fusion to obtain the shale gas resource abundance prediction model corresponding to this marine shale type.

[0072] By setting a coupling threshold, parameter pairs with significant synergistic effects are identified, and interaction terms are automatically generated, enabling the prediction model to explicitly characterize the synergistic contributions of source and reservoir parameters. The second prior constraint matrix complements the first: the first imposes generalized coupling constraints on all primary and secondary control parameters, ensuring overall geological consistency of the model parameters; the second focuses on special parameter pairs whose coupling strength exceeds the threshold, expanding the model's feature space by generating interaction terms, thus enabling it to capture strong synergistic effects. Specifically, the second prior constraint matrix first selects parameter pairs with significant synergistic effects, expanding the model's feature space by generating interaction terms; the first prior constraint matrix then applies coupling constraints to all parameter pairs, including interaction terms, during model training, ensuring that the estimation process of regression coefficients is consistent with geological laws. This two-layer mechanism deeply integrates geological mechanism knowledge into the prediction model at both the structural design and parameter estimation levels, achieving a qualitative leap from simple data fitting to mechanism-data fusion modeling.

[0073] S105: Based on the shale type and geological data of the marine shale target area, apply the corresponding shale gas resource abundance prediction model to predict resource abundance.

[0074] In some embodiments, step S105 may specifically include: predicting resource abundance by applying a corresponding shale gas resource abundance prediction model based on the shale type and geological data of the marine shale target area.

[0075] In some embodiments, the marine shale target area can be a specific geographical or geological unit (such as a new block, a tectonic zone, or a stratigraphic unit) in shale gas exploration and evaluation where the resource abundance needs to be predicted. Its exploration level is lower than that of the calibration zone on which the model is based, and direct drilling and testing data are limited.

[0076] In some embodiments, the shale type can be a marine shale category classified in step S102 based on geological hydrocarbon mechanism models (such as diagenesis, paleoenvironment, and organic matter enrichment mechanisms), such as open marine shale and confined marine shale. Each type corresponds to a different dominant enrichment mechanism.

[0077] In some embodiments, geological data may be actual or estimated values ​​of various geological parameters that conform to the requirements of the prediction model, obtained or interpreted from the target area through means such as seismic exploration, well logging, field outcrop surveys, analogy with adjacent wells, or laboratory testing. These parameters may include TOC, Ro, and effective thickness of the target area.

[0078] In some embodiments, the corresponding shale gas resource abundance prediction model can be a quantitative prediction model with a clear mathematical expression, specifically trained and verified by a multivariate nonlinear regression algorithm for each defined shale type in step S104, based on its primary and secondary control parameters.

[0079] In some embodiments, basic geological analysis can be performed on the target marine shale area, assessing its sedimentary environment, petrological characteristics, and geochemical indicators based on the same mechanistic model framework (diagenesis, paleoenvironment, and organic matter enrichment mechanism) used in step S102. For example, it can be analyzed whether the area exhibits open marine characteristics with open water and high silica content, or confined marine characteristics with limited water, well-developed clay minerals, and significant anoxic indicators. Based on this analysis, the target area can be classified into a predefined specific shale type.

[0080] In some embodiments, based on the shale type determined in the target area, a set of primary and secondary control parameters can be determined. Subsequently, the specific values ​​of these parameters can be accurately extracted or calculated from the geological data of the target area.

[0081] In some embodiments, the extracted geological parameters of the target area can be used as input variables and substituted into a shale gas resource abundance prediction model corresponding to the shale type of the target area. The mathematical calculation formula of the model is executed, and a quantitative prediction value is directly output, which is the predicted shale gas resource abundance of the marine shale target area.

[0082] By establishing a standardized process of type identification -> model matching -> data input -> result output, resource assessment is transformed from qualitative analogy heavily reliant on expert experience to quantitative calculation based on clear rules and mathematical models. This significantly reduces assessment bias caused by subjective human factors, making prediction results for different target areas comparable and repeatable. The above steps strictly adhere to the principles of classification modeling and classification application. Specific prediction models with similar geological origins are precisely matched to the target area, ensuring that the mathematical models (such as parameter weights and nonlinear forms) perfectly match the enrichment patterns of that type of shale. This overcomes the inherent adaptability problem when using a single general-purpose model to predict all types of shale, thus enabling more accurate and reliable quantitative predictions of the resource potential of target areas in the early stages of exploration when data is limited. Furthermore, by encapsulating complex mechanistic understanding and preliminary modeling results into a user-friendly prediction tool, decision-makers do not need to delve into the complex details of the underlying models; they only need to provide basic geological data of the target area and complete type identification to quickly obtain quantitative resource abundance predictions. This provides timely and crucial data support for quickly selecting favorable exploration target areas, optimizing exploration deployment, and conducting economic evaluations.

[0083] In some embodiments, prior to step S101, the method may further include: acquiring shale gas resource abundance data and corresponding geological data for a marine shale scale area; calculating the shale gas resource abundance distribution characteristics of the marine shale scale area based on the shale gas resource abundance data; determining the confidence interval of the shale gas resource abundance distribution in the marine shale scale area based on the shale gas resource abundance distribution characteristics; and filtering the shale gas resource abundance data and corresponding geological data of the marine shale scale area based on the confidence interval.

[0084] In some embodiments, shale gas resource abundance data and corresponding geological data for marine shale calibration zones can be obtained from raw datasets collected systematically from selected, representative, and highly explored reference areas (calibration zones).

[0085] Shale gas resource abundance data can be a set of dependent variable observations that reflect the degree of resource enrichment in a region.

[0086] The corresponding geological data can be a set of independent variable observations that match each resource abundance observation point, covering a variety of geological parameters such as TOC, Ro, and thickness.

[0087] In some embodiments, the distribution characteristics of shale gas resource abundance can be the characteristics obtained by statistically describing the resource abundance dataset mentioned above, which are used to quantify its overall behavior, including calculating the central tendency (such as median, mean) and dispersion (such as standard deviation, range, quantile) of the data, and can be visualized by drawing histograms, box plots, etc., to identify the data distribution pattern (such as whether it is a normal distribution or whether there is skewness).

[0088] In some embodiments, a confidence interval may be a data screening boundary set based on the distribution characteristics of resource abundance data to build a robust predictive model, rather than a strictly statistical confidence interval. Its purpose is to define the range of mainstream or typical resource abundance data used for modeling. It can be determined based on quantiles (e.g., the 5th to 95th percentile) or the mean ± N times the standard deviation, to exclude abnormally high or low values ​​that may be caused by measurement errors, local extreme geological conditions, or assessment uncertainties.

[0089] In some embodiments, the original evaluation results of all candidate marine shale grading zones can be systematically collected and organized to form an initial dataset containing the correspondence between resource abundance values ​​and geological parameter values. Statistical analysis can be performed on the collected shale gas resource abundance data series. Descriptive statistics (mean, median, standard deviation, skewness, kurtosis, etc.) can be calculated, and probability distribution maps can be plotted. By grasping the overall range and concentration areas of resource abundance, it is possible to intuitively reveal whether there are discrete points far from the main shale clusters.

[0090] In some embodiments, based on the distribution characteristics calculated above and combined with geological understanding and modeling objectives, a numerical range can be set as a confidence interval for screening. For example, the interval can be defined as the 5th percentile to the 95th percentile of the resource abundance data by removing the highest and lowest 5% extreme values; or robust statistical methods such as median ± 3 times absolute median difference can be used to determine the upper and lower limits of the interval.

[0091] In some embodiments, confidence intervals can be used as filters applied to the initial dataset. The resource abundance value of each sample is programmatically checked: if the value falls within the interval, the sample row (including its resource abundance and all geological parameters) is retained; if the value exceeds the interval, the sample row is removed from the modeling dataset. Ultimately, a filtered set of shale gas resource abundance data and corresponding geological data for marine shale scale zones is generated.

[0092] By eliminating statistical outliers or geologically atypical samples, the interference of noise and outliers on model parameter estimation is effectively reduced. This allows subsequent steps S101 (correlation analysis) and S104 (regression modeling) to focus more on revealing the main geological patterns, rather than being distorted by individual extreme cases, thus significantly improving the stability and generalization ability of the constructed prediction model. Furthermore, this screening process essentially ensures that the data used to train the model comes from geological conditions where resource abundance is within the normal or expected range. This avoids the model learning non-universal or even erroneous relationships due to the inclusion of samples with abnormally scarce or abundant resources (possibly caused by special, local factors). Therefore, the model built on a clean dataset has prediction logic that is closer to the general geological patterns of the region, and has stronger guiding significance for prediction in new areas. This approach of prioritizing data quality control provides clean input for steps S101 and all subsequent analyses. This avoids iteration failures, convergence difficulties, or abnormal results caused by data problems in subsequent complex calculations, thereby improving the overall smoothness of method execution, the reliability of results, and the engineering practicality of the entire technical solution.

[0093] In some embodiments, after the above preprocessing is completed, step S101 may specifically include: selecting multiple control parameters related to shale gas resource abundance from multiple geological parameters based on the shale gas resource abundance data and corresponding geological data of the screened marine shale scale area.

[0094] This means that all subsequent screening, analysis, and modeling operations will be based on this clean dataset, ensuring that the entire prediction process from the source is built on a high-quality, highly consistent data foundation.

[0095] In some embodiments, the calculation of the first correlation between various geological parameters and shale gas resource abundance in step S102 may further include: dividing the corresponding geological parameter data into multiple numerical intervals according to the numerical distribution of each geological parameter; calculating the Pearson correlation and Spearman rank correlation between the geological parameter and shale gas resource abundance within each numerical interval to obtain multiple segmented Pearson correlation coefficients and multiple segmented Spearman rank coefficients; determining the geological weight corresponding to each numerical interval; determining the eighth correlation corresponding to the geological parameter based on the geological weights corresponding to the multiple numerical intervals and the multiple segmented Pearson correlation coefficients; determining the ninth correlation corresponding to the geological parameter based on the geological weights corresponding to the multiple numerical intervals and the multiple segmented Spearman rank coefficients; analyzing the numerical variation patterns of the segmented coefficients and selecting a fusion strategy based on the identified patterns; and fusing the eighth and ninth correlations according to the selected fusion strategy to obtain the first correlation corresponding to the geological parameter.

[0096] In some embodiments, the numerical distribution can be the statistical characteristics of the range of values ​​of the geological parameter to be analyzed across all samples, which can be characterized by histograms, kernel density estimation, or calculation of quantiles (such as quartiles), reflecting the central tendency, dispersion, and possible multimodal shape of the parameter. The numerical interval can be a continuous sub-range of values, defined manually or automatically, based on the analysis of the above numerical distribution. The division can be based on equal width (equal division according to the value range), equal frequency (making the number of samples in each interval approximately equal), or key thresholds based on prior geological knowledge (e.g., dividing TOC into <1%, 1%-2%, and >2% to correspond to poor, medium, and excellent source rocks).

[0097] In some embodiments, the piecewise Pearson correlation coefficient and the piecewise Spearman correlation coefficient can be the Pearson correlation coefficient and Spearman rank correlation coefficient between the geological parameter and shale gas resource abundance, calculated independently using only sample data within each defined numerical interval. The eighth correlation and the ninth correlation can be two comprehensive scalar indices obtained by summarizing and aggregating the above two sets of piecewise correlation coefficients, representing new assessments of the overall correlation of the parameter from both linear and monotonic nonlinear dimensions.

[0098] In some embodiments, for each geological parameter to be analyzed, its numerical distribution across the entire calibration range sample is analyzed. Based on this distribution, all samples are divided into different numerical intervals according to a preset strategy (e.g., dividing into 3-5 intervals based on equal frequency). For example, the TOC data are sorted from smallest to largest, and the intervals are divided into low, medium, and high numerical intervals at one-third and two-thirds of the sample size. Each interval contains a set of (parameter value, resource abundance value) data pairs.

[0099] In some embodiments, for each defined numerical interval, the following two parallel computational sub-steps are performed: (a) Calculate the first segment Pearson correlation coefficient: Take the original parameter values ​​and corresponding resource abundance values ​​of all samples within the interval, and directly calculate the Pearson correlation coefficient. This coefficient reflects the strength and direction of the linear relationship between the two within a specific range of parameter values. (b) Calculate the second segment Spearman rank correlation coefficient: Calculate the Spearman rank correlation coefficient using sample data within the same interval. This coefficient reflects the strength and direction of the monotonic relationship (not necessarily linear) between the two within a range of parameter values. This operation is repeated for all numerical intervals, ultimately obtaining two sets: multiple segment Pearson correlation coefficients and multiple segment Spearman correlation coefficients, where the number of elements in each set is equal to the number of intervals.

[0100] In some embodiments, the geological weight can be an importance coefficient assigned to each interval divided according to the value of a geological parameter (such as low, medium, and high TOC ranges). This weight is based on the geological significance or exploration value represented by the range of parameter values. For example, for shale gas, high TOC ranges (e.g., >2%) represent high-quality source rocks, and their correlation with resource abundance is far more important than that of low TOC ranges; therefore, they should be assigned a higher geological weight.

[0101] In some embodiments, the numerical variation pattern can be a systematic change pattern exhibited by multiple piecewise Pearson correlation coefficients or multiple piecewise Spearman rank correlation coefficients as the geological parameter numerical range progresses (e.g., from low-value areas to high-value areas). This includes: monotonically increasing / decreasing patterns (correlation continuously strengthens or weakens as parameter values ​​increase), threshold abrupt change patterns (correlation undergoes a qualitative change around a certain critical value), saturation patterns (correlation initially increases and then tends to stabilize), and irregular fluctuation patterns, etc.

[0102] In some embodiments, a geological weight is assigned to each interval based on domain knowledge (such as literature, expert experience, and regional geological patterns). The allocation principle may include: the more critical the geological significance of the interval and the stronger its representativeness of resource enrichment, the higher the weight. The weights are normalized so that the sum of all weights is 1. A weighted average is calculated based on the piecewise Pearson correlation coefficient, the piecewise Spearman rank coefficient, and their corresponding geological weights for each numerical interval.

[0103] In some embodiments, based on numerical sequences of multiple segmented Pearson correlation coefficients and multiple segmented Spearman rank coefficients, the trend of coefficient sequences changing with intervals can be observed. Specifically, this can be achieved by fitting a simple trend line, calculating the difference in coefficients between adjacent intervals, or matching with a preset pattern template. If the pattern is identified as monotonically increasing / decreasing, it indicates that the parameter has a dose effect, and the larger value between the eighth and ninth correlation coefficients can be selected as the fusion strategy to capture the strongest correlation signal. If the pattern is identified as threshold abrupt change, it indicates that the parameter has a critical threshold, and the correlation coefficient type corresponding to the high-weight interval can be selected. If both patterns are consistent and stable, an arithmetic mean or weighted average can be selected. If the patterns conflict or fluctuate, it indicates a complex relationship, and the larger absolute value of the two can be selected, or an uncertainty assessment (such as calculating the standard deviation) can be introduced and the one with higher statistical significance can be selected. According to the selected specific fusion rule, the two eighth and ninth correlation coefficients are fused to obtain the first correlation of the geological parameter.

[0104] By introducing geological weights, the correlation analysis, which is purely mathematical and statistical, is transformed into a correlation assessment driven by geological significance. This ensures that the final selected control parameters are not only statistically relevant but also have a substantial and crucial geological impact on resource enrichment. It avoids selecting parameters that are statistically significant but geologically insignificant or dominated by abnormally high or low values, directly improving the geological rationality and exploration guidance value of subsequent prediction models. Furthermore, analyzing numerical variation patterns and selecting fusion strategies accordingly solves the problem of insufficient ability of traditional single correlation coefficients to describe complex nonlinear relationships, significantly reducing the risk of misjudging complex relationships or simplifying information.

[0105] In some embodiments, step S102 above, which involves selecting multiple control parameters from multiple geological parameters based on a first correlation degree corresponding to each geological parameter and using a preset correlation threshold, may further include: selecting multiple candidate parameters from multiple geological parameters based on a first correlation degree between each geological parameter and shale gas resource abundance and using a preset first correlation threshold; the multiple candidate parameters include multiple source rock candidate parameters and multiple reservoir candidate parameters; calculating a second correlation degree between each candidate parameter and other candidate parameters, including: calculating a third correlation degree between each source rock candidate parameter and the multiple reservoir candidate parameters; calculating a fourth correlation degree between each reservoir candidate parameter and the multiple source rock candidate parameters; and selecting multiple control parameters from multiple candidate parameters based on the second correlation degree corresponding to each candidate parameter and using a preset second correlation threshold.

[0106] In some embodiments, the first correlation can be a univariate measure of the strength of the correlation between each geological parameter and shale gas resource abundance. The preset first correlation threshold can be a minimum correlation standard set for initial screening, based on a statistical significance level (e.g., a correlation coefficient value corresponding to p < 0.05) or an absolute value set empirically (e.g., 0.3 or 0.4). A parameter is considered to have a preliminary significant association with resource abundance only when the absolute value of its correlation exceeds this threshold.

[0107] In some embodiments, candidate parameters can be a subset of geological parameters initially screened through a preset first correlation threshold, all of which show statistically significant correlations with resource abundance. These candidate parameters can be further classified into source rock candidate parameters and reservoir candidate parameters according to the geological processes they belong to.

[0108] In some embodiments, the second, third, and fourth correlations are all indicators used to measure the interrelationships between parameters. The second correlation refers to the correlation between any candidate parameter and other candidate parameters. The third correlation can be the correlation between each source rock candidate parameter and all reservoir candidate parameters, used to reveal the coupling relationship between hydrocarbon generation conditions and reservoir conditions. The fourth correlation can be the correlation between each reservoir candidate parameter and all source rock candidate parameters, providing another perspective to assess the mutual influence between these two types of parameters.

[0109] In some embodiments, the preset second correlation threshold can be a geological factor used to determine whether the current control parameter is local, specific, or weakly correlated with the dominant enrichment mechanism of the current type of shale. When the absolute value of the geothermal correlation corresponding to the control parameter is lower than this threshold (e.g., 0.3 or 0.2), it can be considered to be inconsistent with the current geological logic.

[0110] In some embodiments, for each of a plurality of geological parameters, a first correlation with the shale gas resource abundance data sequence is calculated. Then, the absolute value of the first correlation calculated for each parameter is compared with a preset first correlation threshold. All geological parameters that satisfy a first correlation absolute value greater than or equal to the preset first correlation threshold are included in a plurality of candidate parameter sets. This set can then be geologically divided into two subsets: a plurality of source rock candidate parameters and a plurality of reservoir candidate parameters.

[0111] In some embodiments, for the obtained complete set of multiple candidate parameters, a second correlation degree is calculated between each pair of parameters to form a correlation matrix. This matrix comprehensively characterizes the interdependencies among all candidate parameters. To gain a deeper understanding of the interaction between the hydrocarbon generation system and the reservoir system, calculations are performed: (a) calculating the third correlation degree between each candidate parameter of the source rock and all candidate parameters of the reservoir; (b) calculating the fourth correlation degree between each candidate parameter of the reservoir and all candidate parameters of the source rock. These two analyses can be used to quantify the correlation strength between different geological process control factors.

[0112] In some embodiments, based on all calculated second relevance scores (including third and fourth relevance scores), a preset second relevance score threshold is applied as a criterion to filter candidate parameters, i.e., control parameters with low second relevance scores are deleted.

[0113] If two control parameters exhibit a high second correlation, it means they are controlled by the same geological process or have a genetic connection. This correlation is not redundant noise, but evidence of the inherent consistency and synergy of the geological system. Retaining them helps to reproduce this synergy in the model, making the model closer to the operating mechanism of the real geological system. Conversely, if a parameter exhibits a low second correlation with most other control parameters, it means it represents a local, specific, or weakly related geological factor to the dominant enrichment mechanism of the current type of shale. It may be an outlier, and forcibly including it in the model will introduce interference signals, obscure the core geological laws, and reduce the model's mechanistic consistency and extrapolation stability.

[0114] High secondary correlations (especially across categories such as source rock and reservoir parameters) may reveal key source-reservoir coupling relationships. For example, Ro (thermal maturity) and brittle mineral content may be positively correlated, reflecting the co-evolution of organic matter hydrocarbon generation and diagenesis. Retaining such highly correlated parameter pairs and characterizing them through interaction terms or structures in subsequent multivariate nonlinear regressions enriches and deepens the model mechanism. Deleting parameters with weak correlations to other parameters is essentially a mechanistic purification. This ensures that every parameter ultimately used for modeling is connected to other parts of the system through a certain correlation network, collectively pointing to a unified geological narrative. Models constructed in this way will have more stable predictive behavior, be more geologically logical, and avoid unexplained anomalies in prediction results caused by unstable changes in individual isolated parameters.

[0115] Furthermore, the primary purpose of removing highly correlated parameters in traditional regression analysis is to address the mathematical problem of multicollinearity, preventing inaccurate coefficient estimation. This is a defensive, mathematically driven strategy. Removing low-correlation parameters, on the other hand, is an offensive, mechanism-driven strategy, aiming to proactively shape a parameter set with internal coordination and a clear mechanism. The parameter selection described above does not simply aim to reduce the number of variables, but rather to pursue the purity and consistency of the parameter set in a geological sense. Removing low-correlation parameters is a dimensionality reduction based on the knowledge standard of parameter integration within the system. The result is a more coherent and focused feature set of the geological story, laying the foundation for subsequently constructing a white-box model that is both accurate and interpretable.

[0116] In some embodiments, step S104 above, which involves constructing a single-factor shale gas resource abundance prediction model for each control parameter based on shale gas resource abundance data and corresponding geological data from marine shale scale areas, may specifically include: constructing multiple candidate shale gas resource abundance prediction models for each control parameter based on shale gas resource abundance data and corresponding geological data from marine shale scale areas. These candidate shale gas resource abundance prediction models include at least exponential, power-law, and logarithmic shale gas resource abundance prediction models; and selecting the candidate shale gas resource abundance prediction model with the highest fitting accuracy from among the multiple candidate shale gas resource abundance prediction models as the single-factor shale gas resource abundance prediction model corresponding to that control parameter.

[0117] The general form of an exponential shale gas resource abundance prediction model can be expressed as y = a The formula is exp(bx) + c or a similar variant, where y is the resource abundance, x is a geological parameter, and a, b, and c are undetermined coefficients. This model can be used to describe the relationship between resource abundance and geological parameters, where the abundance increases or decreases at an accelerated rate.

[0118] The general form of the power-law shale gas resource abundance prediction model is y = a x^b + c, where each variable has the same meaning as above. This model (power-law relationship) is extremely common in geological phenomena and is suitable for describing the relative relationship of resource abundance y changing with a constant elastic coefficient as parameter x changes.

[0119] The general form of a logarithmic shale gas resource abundance prediction model is y = a ln(x + d) + b, where ln is the natural logarithm and d is a possible constant offset. This model is suitable for describing the pattern that resource abundance increases with increasing geological parameters, but the growth rate gradually slows down and tends to saturate.

[0120] In some embodiments, fitting accuracy can be a quantitative indicator measuring the degree of fit between the candidate model and the observed data, and can be evaluated using statistics such as the coefficient of determination, adjusted coefficient of determination, sum of squared residuals, or the Akaike Information Criterion. A higher numerical value indicates a better model fit.

[0121] In some embodiments, for a certain control parameter currently being processed, the numerical sequence of the parameter in all samples and the corresponding numerical sequence of shale gas resource abundance are extracted from the shale gas resource abundance data and the corresponding geological data of the marine shale scale area to form a set of (x_i, y_i) data pairs, where i represents the sample index.

[0122] In some embodiments, three preset function forms—exponential, power, and logarithmic—can be used to perform nonlinear regression fitting on the above data pairs. Specifically, for each function form, optimization algorithms such as nonlinear least squares are used to automatically find a set of optimal model coefficients (a, b, c) that minimize the overall deviation between the function curve and all data points (x_i, y_i).

[0123] In some embodiments, the fitting accuracy of each instantiated candidate model can be calculated. For example, the coefficient of determination between each model's predicted value and the actual observed value can be calculated. Then, the fitting accuracy indices of all candidate models are compared and ranked. The model with the highest fitting accuracy is selected from all candidate shale gas resource abundance prediction models. For example, if the coefficient of determination for the power model is 0.85, the exponential model is 0.82, and the logarithmic model is 0.78, then the power model is selected. This selected model is determined as the final single-factor shale gas resource abundance prediction model for this control parameter.

[0124] By automatically selecting the best from a variety of representative nonlinear candidate models, this method can flexibly and accurately capture the diverse relationships that may exist between each control parameter and resource abundance, such as exponential, power-law, or logarithmic relationships. This overcomes the systematic fitting bias that may be caused by manually specifying a single function form (such as using only linear or power functions), and significantly improves the scientificity and accuracy of the single-factor relationship characterization. The subsequent multivariate nonlinear regression model (step S104) is essentially an integration of these single-factor models. The single-factor model with the highest fitting accuracy selected for each parameter means that its mathematical form best represents the independent contribution pattern of that parameter. Using this as the basis function input to the integrated model ensures that the starting point of the integrated model is the most realistic information expression of each parameter, laying a solid foundation for building a high-precision final prediction model.

[0125] The following is a specific embodiment of this specification: 1. Data from 21 known scale zones were collected (see Table 1). Statistical analysis of oil and gas resource abundance characteristics was performed on these scale zones (see...). Figure 2 It was found that over 80% of shale gas resources are concentrated in the range of 1–3 × 10⁻⁶. 8 m 3 / km 2 The distribution of high resource abundance is relatively small and uneven, ranging from 1.5 to 2 × 10⁻⁶. 8 m 3 / km 2 The interval frequency distribution ratio is the highest (accounting for 33.3%). Therefore, the abundance of marine shale gas resources is mainly distributed in the medium and low-abundance areas, while the high-abundance areas are few and unevenly distributed.

[0126] Table 1

[0127] 2. Statistical results on resource abundance reveal that high abundance is distributed in a limited and highly uneven manner. Substituting geological data from high resource abundance areas into the prediction model can interfere with the results, increasing the error. Furthermore, the calibration areas are generally mature exploration zones, characterized by high levels of exploration and good development outcomes. Therefore, the superior geological conditions in most calibration areas can lead to inaccurate prediction models, resulting in an overestimation of overall resource abundance and quantity. To improve the accuracy of the prediction model, the adverse effects of excessively high resource abundance are eliminated, and subsequent modeling parameters are selected to ensure the objectivity and generalization ability of the data and model.

[0128] 3. Preliminary screening of geological influencing factors based on data fusion model. The classic Pearson correlation analysis method in mathematical statistics calculates the covariance and standard deviation of two variables, proving highly effective in handling data with strong linear relationships. Spearman correlation analysis, based on rank, is more suitable for representing nonlinear relationships. Basic correlation analysis was performed on the collected geological data of the calibration area, including TOC, Ro, brittle mineral content, clay mineral content, and effective thickness (see Table 2). This analysis revealed that TOC, Ro, brittle mineral content, clay mineral content, and effective thickness all showed a certain correlation with shale gas resource abundance, with nonlinear characteristics being particularly pronounced.

[0129] Table 2

[0130] Through simple mathematical analysis, the correlation between geological factors such as TOC, brittle mineral content, clay mineral content, and effective thickness and shale gas resource abundance was obtained. Due to the complexity and uniqueness of geological processes, simple data models cannot measure the complex relationships between variables, nor can they determine the influence of other variables on the relationships between the variables under study, thus failing to truly express the complex geological relationships between variables. This study integrates multiple hydrocarbon system principles into the data model, upgrading the analytical methods from multiple dimensions. Hydrocarbon system theory clearly defines the geological processes and elements of hydrocarbon reservoirs as source rocks, reservoirs, caprocks, traps, and migration. Shale gas, as an unconventional hydrocarbon resource, differs from conventional hydrocarbon resources; due to its in-situ accumulation, it is mainly controlled by source rocks and reservoirs. In the theory of the entire oil and gas system, conventional and unconventional oil and gas can originate from the same source rock. After hydrocarbon generation in the source rock, hydrocarbons are expelled after adsorption. Oil and gas that have undergone long-distance migration are considered conventional, while those that have migrated near the source or remain in the source rock are considered unconventional. Shale gas belongs to the category of unexpelled hydrocarbons and is therefore primarily controlled by the source rock (reservoir). The theory of hydrocarbon molecular dynamics posits that shale gas has not undergone migration and is mainly controlled by viscosity, accumulating in the pores of shale, primarily in the form of adsorbed gas and free gas. Therefore, it is mainly controlled by the conditions of the source rock (reservoir). Source rock factors include TOC, Ro, and effective thickness, while reservoir factors include brittle mineral content, clay mineral content, and effective thickness. Due to the complexity of geological processes, it is necessary to consider the interrelationships between multiple variables, such as the impact of Ro on the relationship between TOC and shale gas resource abundance. Partial correlation analysis was conducted based on two influence types: source rock and reservoir. This study investigated the interactions between geological factors within either source rock or reservoir parameters (see Table 3). It was found that among source rock parameters, Ro (Ro content) was the controlling factor, having a relatively small impact on the relationship between TOC (Total Organic Carbon) and resource abundance, with only a slight decrease in correlation coefficient compared to Table 2. However, for different effective thicknesses, the correlation coefficient decreased significantly, and the overall correlation also decreased substantially, indicating that effective thickness has a significant impact on the relationship between TOC and resource abundance. Similarly, it was found that Ro had a relatively small impact on all source rock geological factors, and effective thickness was less affected by other factors but had a significant impact on them. Interactions also existed between brittle mineral content and clay mineral content. Comprehensive analysis suggests that all geological factors, individually, correlate well with marine shale gas resource abundance, with only Ro showing relatively poor correlation. Furthermore, all factors exhibit significant nonlinear characteristics. Considering the mutual influence of geological factors, the effective thickness has a significant impact on the relationship between all geological factors and resource abundance, while being less affected by other geological factors. This indicates that the effective thickness, which serves as both a source rock parameter and a reservoir parameter, has a substantial impact on resource abundance, and that there is a clear interaction among various geological factors.

[0131] Table 3

[0132] Based on the analysis of the correlation between geological factors, the interaction between source rocks and reservoirs is further analyzed. Marine shale gas is formed in situ, so the corresponding source rocks are also reservoirs. However, the definitions of source rocks and reservoirs are different, and the geological factors they contain are also different. Canonical correlation analysis was performed on the three source rock factors (TOC, Ro, and effective thickness) and the three reservoir factors (brittle mineral content, clay mineral content, and effective thickness), as shown in Table 4. Source rock factors are group 1, and reservoir factors are group 2. The correlation in the first dimension is very good, with correlation coefficients above 0.95. Wilke's statistic is used to test the statistical significance of the canonical correlation coefficient, which represents the proportion of variance explained by the subsequent dimensions. The closer to 0, the more significant the relationship. F is the degrees of freedom; the larger the value under a significant relationship, the better the correlation. It can be seen that the correlation in the first dimension is very good, proving a significant correlation between source rock factors and reservoir factors, that is, the better the quality of the source rock, the better the reservoir tends to be. Statistical analysis of source rock factors in set 1 and reservoir factors in set 2 revealed, as shown in Table 5, that Ro load is relatively small and has a relatively small impact on source rocks, while other geological factors have large loads and significant impacts.

[0133] Table 4

[0134] In summary, TOC, effective thickness, brittle mineral content, and clay mineral content were initially selected as the main controlling factors. Among them, TOC and effective thickness belong to source rock factors, while brittle mineral content, clay mineral content, and effective thickness belong to reservoir factors. Moreover, the mutual influence among geological factors cannot be ignored.

[0135] Table 5

[0136] 4. The dominant geological influencing factors were selected using a geological hydrocarbon mechanism model. Comparison of shale gas abundance in scaled regions of region A and other regions revealed that high-abundance scaled regions were all located in region A, while shale gas abundance in other regions was generally low to medium. This is due to the different environments in which marine shale was formed during sedimentation. Shale in region A was primarily formed in an open marine environment, while shale in other regions was primarily formed in a confined marine environment. Therefore, to ensure the accuracy of the prediction model, marine shale was further divided into open marine shale and confined marine shale.

[0137] From the perspective of paleoenvironmental models, although marine shale deposits are all in deep-water environments and accompanied by a large amount of buried organic matter, the paleoenvironments vary considerably. In open-ocean marine shale deposits, the paleoclimate was generally warm and humid. Warmth is more suitable for biological survival, while humidity facilitates terrestrial weathering, bringing nutrients and providing conditions for a surge in productivity. In terms of paleooceanic conditions, paleosalinity was moderate and its fluctuations were small, having minimal impact on organisms. Frequent upwelling brought nutrients to the bottom waters, and the periodic changes in marine transgression and regression also promoted biological growth. In confined-ocean marine shale deposits, the paleoclimate was mostly cold and dry with high evaporation, which facilitated water stratification and maintained a sulfidated environment at the bottom. The paleooceanic salinity was relatively high, resulting in poor benthic organism development and minimal biological disturbance, which helped maintain the bottom environment. Simultaneously, the large input of terrigenous debris due to weathering protected the organic matter from degradation during diagenesis. Therefore, open marine paleoenvironments are suitable for biological survival, providing the necessary environmental elements for biological flourishing. TOC (Total Organic Matter) directly reflects productivity, while effective thickness reflects the duration of biological flourishing. Clay minerals, as carriers of biological excrement and remains, indirectly reflect the development of life. Confined marine paleoenvironments are characterized by their closed nature. Although weathering brings a large amount of terrigenous debris, the overall environment is stratified and has a long-term sulfidation environment at the bottom. TOC, directly reflecting the strength of productivity, remains important, while effective thickness represents the duration of the sulfidation environment protecting organic matter, and brittle mineral content indicates the input of terrigenous debris.

[0138] From the perspective of organic matter enrichment models, two key conditions are required for organic matter enrichment: primary productivity and an oxygen-deficient environment. In open marine facies, frequent water exchange between the basin and the open ocean ensures timely replenishment of nutrients. The warm and humid environment fosters high primary productivity, while the deposited clay material significantly adsorbs organic matter, collectively promoting its burial. This is the organic matter enrichment model for open marine shale. In confined marine facies, obstructed water exchange leads to significant evaporation and distinct basin water stratification. Oxygen-deficient or even sulfided water slows down organic matter decomposition, providing an excellent environment for burial. Weathering also brings a large amount of detrital material, which, although diluting during the burial stage, acts as a rock framework during later diagenetic compaction, effectively protecting the internal pore structure of the shale. This is the organic matter enrichment model for confined marine shale. In summary, TOC, effective thickness, clay mineral content, and diagenesis of open marine shale are closely related, while TOC, effective thickness, brittle mineral content, and diagenesis of confined marine shale are closely related.

[0139] Considering the interrelationships among geological factors, no single geological factor can adequately characterize the abundance of marine shale resources. Therefore, a set of influencing factors is used to determine the dominant geological factors. Based on preliminary screening results from a data fusion model and classification using a geological hydrocarbon mechanism model, the abundance of marine shale resources in open marine environments is primarily controlled by primary productivity. The primary dominant geological influencing factor set is the source rock factor set, including TOC and effective thickness. The secondary dominant geological factors are the reservoir factor set, including clay mineral content and effective thickness. In confined marine environments, the abundance of marine shale resources is primarily controlled by anoxic or even sulfide-laden sedimentary environments. The primary dominant geological influencing factor is the reservoir set, including brittle mineral content and effective thickness. The secondary dominant geological factors are the source rock factor set, including TOC and effective thickness.

[0140] 5. For both open-ocean marine shale gas and confined-ocean marine shale gas resource abundance and all individual controlling geological factors, select the most suitable fundamental function model to characterize the relationship between the two. See [reference needed]. Figure 3 The results show that for characterizing the abundance of open marine shale gas resources, the TOC parameter performs best under the exponential model, while the effective thickness and clay mineral content are suitable for the power model. For characterizing the abundance of confined marine shale gas resources, the TOC parameter is suitable for the power model, while the effective thickness and brittle mineral content show the best fit under the logarithmic model.

[0141] 6. The prediction model is constructed in stages. First, the contribution of individual geological factors to the overall impact of the geological factor set on the abundance of marine shale resources is determined. Second, a prediction model for the geological factor set and resource abundance is established. Typical correlation analysis is performed on the source rock factor set, reservoir set, and resource abundance, respectively. The typical loads are shown in Table 6. The typical loads of each geological factor represent the degree of influence on its respective factor set. The factor set further influences the resource abundance, which is also the weight of individual geological factors in the factor set.

[0142] Table 6

[0143] Based on the above findings, and combining the single geological influence factor function model and its weight in the influence factor set, nonlinear multiple regression methods were used to establish prediction models for the abundance of marine shale gas resources in open and confined seas, with source rock clusters and reservoir clusters as independent variables. These models are shown in Table 7. Next, the applicability of the constructed prediction models will be tested. (See Table 7 for details.) Figure 3 The results show that the R-values ​​of the two prediction models are... 2 All values ​​were above 0.9, proving that the model's prediction performance was excellent and the predicted results were very close to the actual values.

[0144] Table 7

[0145] 7. Based on the established prediction models for the abundance of open-ocean and confined-ocean shale gas resources, and combined with the actual geological conditions of the target area, a prediction model is selected. Geological influencing factors such as TOC, effective thickness, and clay mineral content or brittle mineral content are substituted into the prediction model to obtain the shale gas resource abundance, and then the resource quantity of the target area is calculated to assess the shale gas resource potential. A practical application example is provided here. The basic geological influencing factors of the target area are shown in Table 8. The main controlling factors are substituted into the prediction model to calculate the shale gas resource quantity. Figure 4 See Table 9.

[0146] Table 8

[0147] Table 9

[0148] As can be seen from the above embodiments of the method for predicting the abundance of marine shale gas resources provided in this specification, this embodiment can, based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, select multiple control parameters related to shale gas resource abundance from multiple geological parameters; based on the geological oil and gas mechanism model, marine shale is divided into multiple types; for each type of marine shale, one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance are determined; based on the primary and secondary control parameters of each type of marine shale, a corresponding shale gas resource abundance prediction model is constructed using a multivariate nonlinear regression algorithm; according to the shale type and geological data of the marine shale target area, the corresponding shale gas resource abundance prediction model is applied to predict resource abundance. Through correlation analysis based on measured data of the marine shale scale area, control parameters significantly related to resource abundance are selected from numerous geological parameters, significantly improving the objectivity, efficiency, and repeatability of the initial parameter selection, while effectively reducing the dimensionality of subsequent modeling and avoiding interference from irrelevant or redundant parameters. Building upon this foundation, a geological hydrocarbon mechanism model was introduced as the classification basis, dividing marine shale into multiple genetic types. This breaks through the traditional coarse classification model dominated by a single lithology or lithology, achieving an essential and refined classification based on the main control mechanism of hydrocarbon accumulation. This provides a precise geological framework for subsequent classification modeling, enabling the model to specifically characterize the unique enrichment patterns of different shale types. Furthermore, based on the classification, the primary and secondary control parameters for each shale type were identified, clarifying the relative importance of different geological factors in different types. This endows the prediction model with a clear geological logic and differentiated modeling focus, transforming the model construction process from a black-box parameter stacking to a structured process with primary and secondary factors and mechanism-driven characteristics, greatly enhancing the geological interpretability of the model. In addition, a multivariate nonlinear regression algorithm was used to construct the prediction model, directly overcoming the limitations of traditional linear methods in characterizing complex geological relationships, and enabling a more accurate fit between the nonlinear response of control parameters and resource abundance. Finally, by identifying the shale type in the target area, the corresponding prediction model was directly invoked for accurate prediction, providing timely and reliable quantitative basis for exploration decisions.

[0149] Based on the above-mentioned method for predicting the abundance of marine shale gas resources, this specification also proposes embodiments of a marine shale gas resource abundance prediction device. For example... Figure 5 As shown, the marine shale gas resource abundance prediction device 500 may specifically include the following modules: The filtering module 501 is used to filter out multiple control parameters related to shale gas resource abundance from multiple geological parameters based on shale gas resource abundance data and corresponding geological data of marine shale scale areas. The classification module 502 is used to classify marine shale into multiple types based on the geological oil and gas mechanism model; The determination module 503 is used to determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources for each type of marine shale. Module 504 is used to construct a corresponding shale gas resource abundance prediction model based on the main and secondary control parameters of each marine shale type using a multivariate nonlinear regression algorithm. The prediction module 505 is used to predict resource abundance based on the shale type and geological data of the marine shale target area, and to apply the corresponding shale gas resource abundance prediction model.

[0150] In some embodiments, the marine shale gas resource abundance prediction device 500 can further be used to: acquire shale gas resource abundance data and corresponding geological data of a marine shale scale area; calculate the shale gas resource abundance distribution characteristics of the marine shale scale area based on the shale gas resource abundance data; determine the confidence interval of the shale gas resource abundance distribution of the marine shale scale area based on the shale gas resource abundance distribution characteristics; and filter the shale gas resource abundance data and corresponding geological data of the marine shale scale area based on the confidence interval. Based on this, the filtering module 501 can specifically be used to: filter multiple control parameters related to shale gas resource abundance from multiple geological parameters based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, including: filtering multiple control parameters related to shale gas resource abundance from multiple geological parameters based on the filtered shale gas resource abundance data and corresponding geological data of the marine shale scale area.

[0151] In some embodiments, the screening module 501 may also be used to: calculate the first correlation between each geological parameter and the shale gas resource abundance based on the shale gas resource abundance data and the corresponding geological data of the marine shale scale area; and select multiple control parameters from multiple geological parameters based on the first correlation corresponding to each geological parameter using a preset correlation threshold.

[0152] In some embodiments, the screening module 501 may further be used to: select multiple candidate parameters from multiple geological parameters based on a first correlation between various geological parameters and shale gas resource abundance using a preset first correlation threshold; the multiple candidate parameters include multiple source rock candidate parameters and multiple reservoir candidate parameters; calculate a second correlation between each candidate parameter and other candidate parameters, including: calculating a third correlation between each source rock candidate parameter and the multiple reservoir candidate parameters; calculating a fourth correlation between each reservoir candidate parameter and the multiple source rock candidate parameters; and select multiple control parameters from multiple candidate parameters based on the second correlation corresponding to each candidate parameter using a preset second correlation threshold.

[0153] In some embodiments, the above-mentioned geological oil and gas mechanism model includes at least a shale diagenesis mechanism model, a paleoenvironmental control mechanism model, and an organic matter enrichment mechanism model. Based on this, the above-mentioned classification module 502 can be specifically used to classify marine shale into multiple types based on the shale diagenesis mechanism model, the paleoenvironmental control mechanism model, and the organic matter enrichment mechanism model.

[0154] In some embodiments, the aforementioned marine shale includes at least open-sea marine shale and confined-sea marine shale; the aforementioned multiple control parameters include one or more source rock control parameters and one or more reservoir control parameters. Based on this, the determining module 503 can specifically be used to: for open-sea marine shale, use the source rock control parameters as primary control parameters and the reservoir control parameters as secondary control parameters; for confined-sea marine shale, use the reservoir control parameters as primary control parameters and the source rock control parameters as secondary control parameters.

[0155] In some embodiments, the aforementioned construction module 504 can be specifically used to: construct a single-factor shale gas resource abundance prediction model corresponding to each control parameter based on shale gas resource abundance data and corresponding geological data of marine shale scale areas; calculate the fifth correlation between each control parameter and other control parameters, including: calculating the sixth correlation between each active parameter and the plurality of secondary control parameters; calculating the seventh correlation between each secondary control parameter and the plurality of primary control parameters; and, based on the fifth correlation of each control parameter, using a multivariate nonlinear regression algorithm to fuse the single-factor shale gas resource abundance prediction models corresponding to the primary and secondary control parameters of each marine shale type, to obtain the shale gas resource abundance prediction model corresponding to that marine shale type.

[0156] In some embodiments, the above-mentioned construction module 504 can also be used to: construct multiple candidate shale gas resource abundance prediction models corresponding to each control parameter based on shale gas resource abundance data and corresponding geological data of marine shale scale areas; and select the candidate shale gas resource abundance prediction model with the highest fitting accuracy from the multiple candidate shale gas resource abundance prediction models as the single-factor shale gas resource abundance prediction model corresponding to the control parameter.

[0157] As can be seen from the marine shale gas resource abundance prediction device provided in the embodiments of this specification above, the embodiments of this specification can, based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, select multiple control parameters related to shale gas resource abundance from multiple geological parameters; based on the geological oil and gas mechanism model, marine shale is divided into multiple types; for each type of marine shale, one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance are determined; based on the primary and secondary control parameters of each type of marine shale, a corresponding shale gas resource abundance prediction model is constructed using a multivariate nonlinear regression algorithm; according to the shale type and geological data of the marine shale target area, the corresponding shale gas resource abundance prediction model is applied to predict resource abundance. Through correlation analysis based on measured data of the marine shale scale area, control parameters significantly related to resource abundance are selected from numerous geological parameters, significantly improving the objectivity, efficiency, and repeatability of the initial parameter selection, while effectively reducing the dimensionality of subsequent modeling and avoiding interference from irrelevant or redundant parameters. Building upon this foundation, a geological hydrocarbon mechanism model was introduced as the classification basis, dividing marine shale into multiple genetic types. This breaks through the traditional coarse classification model dominated by a single lithology or lithology, achieving an essential and refined classification based on the main control mechanism of hydrocarbon accumulation. This provides a precise geological framework for subsequent classification modeling, enabling the model to specifically characterize the unique enrichment patterns of different shale types. Furthermore, based on the classification, the primary and secondary control parameters for each shale type were identified, clarifying the relative importance of different geological factors in different types. This endows the prediction model with a clear geological logic and differentiated modeling focus, transforming the model construction process from a black-box parameter stacking to a structured process with primary and secondary factors and mechanism-driven characteristics, greatly enhancing the geological interpretability of the model. In addition, a multivariate nonlinear regression algorithm was used to construct the prediction model, directly overcoming the limitations of traditional linear methods in characterizing complex geological relationships, and enabling a more accurate fit between the nonlinear response of control parameters and resource abundance. Finally, by identifying the shale type in the target area, the corresponding prediction model was directly invoked for accurate prediction, providing timely and reliable quantitative basis for exploration decisions.

[0158] This specification also provides a computer device for predicting the abundance of marine shale gas resources, including a processor and a memory for storing processor-executable instructions. Specifically, the processor can perform the following tasks according to the instructions: based on shale gas resource abundance data and corresponding geological data of a marine shale scale area, select multiple control parameters related to shale gas resource abundance from multiple geological parameters; classify marine shale into various types based on a geological oil and gas mechanism model; for each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance; based on the primary and secondary control parameters of each marine shale type, construct a corresponding shale gas resource abundance prediction model using a multivariate nonlinear regression algorithm; and predict resource abundance by applying the corresponding shale gas resource abundance prediction model according to the shale type and geological data of the marine shale target area.

[0159] To execute the above instructions more accurately, please refer to... Figure 6 As shown in the embodiments of this specification, another specific computer device 600 is also provided, wherein the computer device 600 includes a network communication port 601, a processor 602 and a memory 603, and the above structures are connected by internal cables so that the various structures can perform specific data interaction.

[0160] The processor 602 can specifically be used to: based on shale gas resource abundance data and corresponding geological data of marine shale scale areas, select multiple control parameters related to shale gas resource abundance from multiple geological parameters; classify marine shale into multiple types based on a geological oil and gas mechanism model; for each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance; based on the primary and secondary control parameters of each type of marine shale, construct a corresponding shale gas resource abundance prediction model using a multivariate nonlinear regression algorithm; and predict resource abundance by applying the corresponding shale gas resource abundance prediction model according to the shale type and geological data of the marine shale target area.

[0161] The memory 603 can be used to store the corresponding instruction program.

[0162] In this embodiment, the network communication port 601 can be a virtual port bound to different communication protocols, thereby enabling the sending or receiving of different data. For example, the network communication port can be a port responsible for web data communication, a port responsible for FTP data communication, or a port responsible for email data communication. Furthermore, the network communication port can also be a physical communication interface or communication chip. For example, it can be a wireless mobile network communication chip, such as GSM or CDMA; it can also be a Wi-Fi chip; or it can be a Bluetooth chip.

[0163] In this embodiment, the processor 602 can be implemented in any suitable manner. For example, the processor can take the form of a microprocessor or processor and a computer-readable medium storing computer-readable program code (e.g., software or firmware) executable by the (micro)processor, logic gates, switches, application-specific integrated circuits (ASICs), programmable logic controllers, and embedded microcontrollers, etc. This specification is not limiting.

[0164] In this embodiment, the memory 603 includes volatile memory and non-volatile memory. The memory 603 can include multiple layers. In digital systems, anything that can store binary data can be a memory; in integrated circuits, a circuit with storage function but no physical form is also called a memory, such as RAM, FIFO, etc.; in a system, a storage device with a physical form is also called a memory, such as a memory stick, TF card, etc.

[0165] Furthermore, embodiments of this specification also provide a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described... Figure 1 The instructions for the method shown.

[0166] It should be understood that in the various embodiments of this specification, the sequence number of each process does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this specification.

[0167] It should also be understood that, in the embodiments of this specification, the terms and / or are merely descriptions of the relationships between related objects, indicating that three relationships may exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, and B existing alone. Additionally, the character " / " in this specification generally indicates that the preceding and following related objects have an "or" relationship.

[0168] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0169] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0170] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0171] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational tasks to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The task is a function specified in one or more boxes.

[0172] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for predicting the abundance of marine shale gas resources, characterized in that, include: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, several control parameters related to shale gas resource abundance were selected from multiple geological parameters. Based on the geological oil and gas mechanism model, marine shale is classified into multiple types; For each type of marine shale, determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources; Based on the primary and secondary control parameters of each marine shale type, a corresponding shale gas resource abundance prediction model is constructed using a multivariate nonlinear regression algorithm. Based on the shale type and geological data of the marine shale target area, the corresponding shale gas resource abundance prediction model is applied to predict resource abundance.

2. The method according to claim 1, characterized in that, The method further includes: Acquire shale gas resource abundance data and corresponding geological data for marine shale scale zones; Based on the shale gas resource abundance data, the distribution characteristics of shale gas resource abundance in the marine shale scale zone are calculated; Based on the shale gas resource abundance distribution characteristics, determine the confidence interval for the shale gas resource abundance distribution in the marine shale scale zone; Based on the confidence interval, the shale gas resource abundance data and corresponding geological data of the marine shale scale area are screened; The shale gas resource abundance data and corresponding geological data based on marine shale scale zones were used to select multiple control parameters related to shale gas resource abundance from various geological parameters, including: Based on the shale gas resource abundance data and corresponding geological data of the selected marine shale scale area, several control parameters related to shale gas resource abundance were selected from multiple geological parameters.

3. The method according to claim 1, characterized in that, The selection of multiple control parameters related to shale gas resource abundance from various geological parameters includes: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale area, the first correlation degree between each geological parameter and shale gas resource abundance is calculated; Based on the first correlation of each geological parameter, multiple control parameters are selected from multiple geological parameters using a preset correlation threshold.

4. The method according to claim 3, characterized in that, The first correlation coefficient corresponding to each geological parameter is used to filter out multiple control parameters from multiple geological parameters using a preset correlation threshold, including: Based on the first correlation between various geological parameters and shale gas resource abundance, multiple candidate parameters are selected from multiple geological parameters using a preset first correlation threshold; the multiple candidate parameters include multiple source rock candidate parameters and multiple reservoir candidate parameters; Calculate the second correlation between each candidate parameter and other candidate parameters, including: calculating the third correlation between each source rock candidate parameter and the plurality of reservoir candidate parameters; and calculating the fourth correlation between each reservoir candidate parameter and the plurality of source rock candidate parameters. Based on the second relevance corresponding to each candidate parameter, multiple control parameters are selected from multiple candidate parameters using a preset second relevance threshold.

5. The method according to claim 1, characterized in that, The geological oil and gas mechanism model includes at least a shale diagenesis mechanism model, a paleoenvironmental control mechanism model, and an organic matter enrichment mechanism model; The geological hydrocarbon mechanism model classifies marine shale into several types, including: Based on shale diagenesis mechanism models, paleoenvironmental control mechanism models, and organic matter enrichment mechanism models, marine shale is classified into multiple types.

6. The method according to claim 1, characterized in that, The marine shale includes at least open marine shale and confined marine shale; the multiple control parameters include one or more source rock control parameters and one or more reservoir control parameters; For each type of marine shale, one or more primary control parameters and one or more secondary control parameters affecting shale gas resource abundance are determined, including: For open marine shale, source rock control parameters are used as primary control parameters and reservoir control parameters are used as secondary control parameters. For marine shale with limited facies, reservoir control parameters are used as primary control parameters and source rock control parameters are used as secondary control parameters.

7. The method according to claim 1, characterized in that, The aforementioned shale gas resource abundance prediction model is constructed using a multivariate nonlinear regression algorithm based on the primary and secondary control parameters for each marine shale type, including: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale zone, a single-factor shale gas resource abundance prediction model corresponding to each control parameter is constructed. Calculate the fifth correlation between each control parameter and other control parameters, including: calculating the sixth correlation between each active parameter and the plurality of secondary control parameters; and calculating the seventh correlation between each secondary control parameter and the plurality of primary control parameters. Based on the fifth correlation of each control parameter, a multivariate nonlinear regression algorithm is used to fuse the single-factor shale gas resource abundance prediction models corresponding to the main control parameters and secondary control parameters of each marine shale type, so as to obtain the shale gas resource abundance prediction model corresponding to that marine shale type.

8. The method according to claim 7, characterized in that, Based on the shale gas resource abundance data and corresponding geological data of marine shale scale zones, a single-factor shale gas resource abundance prediction model is constructed for each control parameter, including: Based on the shale gas resource abundance data and corresponding geological data of the marine shale scale zone, multiple candidate shale gas resource abundance prediction models are constructed for each control parameter. Among the multiple candidate shale gas resource abundance prediction models, the candidate shale gas resource abundance prediction model with the highest fitting accuracy is selected as the single-factor shale gas resource abundance prediction model corresponding to this control parameter.

9. A device for predicting the abundance of marine shale gas resources, characterized in that, include: The filtering module is used to filter out multiple control parameters related to shale gas resource abundance from multiple geological parameters based on shale gas resource abundance data and corresponding geological data of marine shale scale areas. The classification module is used to classify marine shale into multiple types based on geological oil and gas mechanism models; The determination module is used to determine one or more primary control parameters and one or more secondary control parameters that affect the abundance of shale gas resources for each type of marine shale. The module is used to construct a corresponding shale gas resource abundance prediction model based on the main and secondary control parameters of each marine shale type, using a multivariate nonlinear regression algorithm. The prediction module is used to predict resource abundance based on the shale type and geological data of the marine shale target area, applying the corresponding shale gas resource abundance prediction model.

10. A computer device, characterized in that, include: Memory, used to store computer programs; A processor for executing the computer program to implement the method of any one of claims 1-8.

Images