A shale reservoir geologic engineering double sweet spot evaluation method based on phase control constraint and data driving

Abstract

The application discloses a shale reservoir geology engineering double-sweet-spot evaluation method based on phase control constraint and data driving, relates to the technical field of shale reservoir geology engineering evaluation, and comprises the following steps: integrating core, logging, seismic and fracturing operation response data through a multi-parameter collaborative calculation platform, and screening geology and engineering sweet-spot core parameters; constructing a phase control elastic parameter inversion network to realize elastic parameter space inversion, building a geology engineering double-attribute coupling model to establish cross-dimension parameter correlation; determining and calibrating main control parameters by using a sensitivity prediction algorithm, and generating a time-space matched parameter distribution field through multi-scale collaborative calculation; constructing an evaluation system based on the distribution field, and completing double-sweet-spot region definition through double-attribute threshold division and space superposition analysis. Through coupling model building, sensitivity analysis and collaborative calculation processes, the method realizes deep fusion of multi-source data and accurate coupling of double attributes, and improves the accuracy and reliability of sweet-spot evaluation.

Description

Technical Field

[0001] This invention relates to the field of shale reservoir geological engineering evaluation technology, and in particular to a double sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approaches. Background Technology

[0002] Shale reservoirs have become a crucial area for oil and gas exploration and development. The evaluation of their geological and engineering sweet spots directly impacts the scientific nature of well placement and the effectiveness of fracturing operations, making it a core technological step towards efficient reservoir development. Currently, shale reservoirs exhibit strong heterogeneity and significant cross-scale parameter differences, making it difficult for traditional evaluation methods to integrate multi-source heterogeneous data and accurately establish the correlation between geological reservoir performance and engineering stimulation potential. As exploration and development extend to deeper and more complex blocks, higher demands are placed on the accurate identification of sweet spots. There is an urgent need for an evaluation technology that can integrate multi-dimensional information from core samples, logging, seismic data, and operational responses, taking into account both geological patterns and engineering realities, to achieve continuous and detailed characterization across the entire reservoir area. This would address the insufficient accuracy of traditional methods under complex reservoir conditions.

[0003] Existing technologies suffer from two key shortcomings: First, they evaluate geological and engineering attributes in a fragmented manner, failing to fully consider the intrinsic relationship between the two. They rely solely on single-dimensional parameters or simple superposition methods to determine sweet spots, neglecting the synergistic effects of different attribute parameters, resulting in evaluation results that fail to reflect the true potential of the reservoir. Second, they lack the ability to integrate multi-source data, lacking a systematic parameter integration and optimization mechanism. This leads to vague identification of key factors influencing sweet spot formation and the absence of an effective parameter spatial matching and dynamic adjustment model, resulting in poor spatial continuity and reliability of evaluation results, and failing to provide accurate and comprehensive technical support for subsequent engineering construction. Summary of the Invention

[0004] To overcome the shortcomings and deficiencies of existing technologies, this invention provides a double sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approaches.

[0005] The technical solution adopted in this invention is a method for evaluating the geological and engineering double sweet spot of shale reservoirs based on phase control constraints and data-driven approaches, comprising the following steps: S1, integrating core mineral composition, well logging elastic parameters, seismic reflection characteristics, and fracturing operation response data through a multi-parameter collaborative calculation platform for shale reservoirs, and screening parameters associated with geological sweet spots, such as organic carbon content, porosity, and gas saturation, and associated with engineering sweet spots, such as brittleness index, Young's modulus, and geostress difference coefficient; S2, constructing a phase control elastic parameter inversion network, using seismic wavefield data as input, and combining the results of geological facies zone division to perform spatial calculation of elastic parameters. S3. Based on the inversion results, a geological engineering dual-attribute coupling model is built, and cross-dimensional correlations are established by integrating reservoir physical parameters and engineering fracturing parameters; S4. The sweet spot parameter sensitivity prediction algorithm is used to analyze the parameter influence of the dual-attribute coupling results and determine the calibration master control parameters; S5. The calibration master control parameters are calculated on a multi-scale collaborative basis through a shale reservoir multi-parameter collaborative calculation platform to generate a spatiotemporally matched parameter distribution field; S6. Based on the parameter distribution field, a shale reservoir geological engineering dual sweet spot evaluation system is constructed, and the dual sweet spot region is defined by dual-attribute threshold division and spatial overlay analysis.

[0006] Furthermore, the expression for the phased elastic parameter inversion network is: , , in, For inverted elasticity parameters, These are the weighting coefficients. This is a convolution operation; Seis contains seismic wavefield data. These are the parameters of the phase-controlled constraint kernel function. For Bayesian neural network operations, For well logging data, These are network feature parameters. Phase control adjustment coefficient, Phase Here, Geo represents the geological facies control function, and Geo represents the geological facies zone data. These are the phase control boundary parameters.

[0007] Furthermore, the expression for the geological engineering dual-attribute coupling model is as follows: in, The coupling coefficient is a two-attribute coupling coefficient. For coupling adjustment parameters, For geological dessert property functions, Porosity For gas saturation, Organic carbon content, For engineering dessert attribute functions, The brittleness index, For Young's modulus, This is the coefficient of geostress difference. For the Laplace operator.

[0008] Furthermore, the expression for the dessert parameter sensitivity prediction algorithm is as follows: in, For parameter sensitivity index, For the first A dessert rating parameter, Let be the partial derivative of the coupling coefficient with respect to this parameter. The mean of the parameters, The sensitivity attenuation coefficient, This represents the total number of parameters.

[0009] Furthermore, the collaborative calculation expression of the shale reservoir multi-parameter collaborative calculation platform is as follows: in, These are the parameter values ​​after collaborative computation. Principal component analysis operation, For Kriging interpolation operations, For spatial structure parameters, For random forest operations, For model training parameters, It is a quadratic coupling function. For time-dynamic parameters, For the calculated value of the geological dessert attribute, Calculate the value for the dessert attribute of the project.

[0010] Furthermore, the expression for the double sweet spot evaluation of the shale reservoir geological engineering is as follows: in, The rating index for both desserts is [not specified]. For activation function, For evaluation coefficients, The coupling coefficient is a two-attribute coupling coefficient. This is the parameter sensitivity index.

[0011] Further, S3 includes the following sub-steps: S31, spatially matching the elastic parameters output by the phase-controlled elastic parameter inversion network with the measured physical property parameters of the rock core, and extracting the basic parameters of the geological sweet spot, such as porosity, organic carbon content, and gas saturation, within the corresponding geological unit; S32, collecting the pumping pressure response data during fracturing operations and the brittleness index, Young's modulus, and geostress difference coefficient interpreted from well logging, and establishing a set of basic parameters for the engineering sweet spot; S33, mapping the geological sweet spot parameters and the engineering sweet spot parameters to the same three-dimensional grid system using spatial interpolation methods to form a parameter space alignment matrix; S34, substituting the aligned dual-attribute parameters into the calculation based on the coupled model structure to generate full-domain coverage dual-attribute coupling coefficient distribution data.

[0012] Further, step S4 includes the following sub-steps: S41, randomly extracting sample points from the dual-attribute coupling coefficient distribution data to construct a sample dataset including all sweet spot parameters and coupling coefficients; S42, dividing the sample dataset into a training set and a test set according to the proportion, and setting the number of algorithm iterations and convergence conditions; S43, calculating the partial derivatives of each parameter with respect to the coupling coefficient using the training set data, calculating the initial sensitivity value, and performing normalization processing in combination with the parameter mean; S44, verifying the stability of the sensitivity calculation results using the test set data, correcting abnormal sensitivity values ​​through an exponential decay function, and outputting the final parameter sensitivity ranking results.

[0013] Further, S5 includes the following sub-steps: S51, inputting the preceding calibration parameters from the parameter sensitivity ranking results into the shale reservoir multi-parameter collaborative calculation platform, and setting the spatial scale and temporal resolution of each parameter; S52, using a spatial interpolation algorithm to supplement the missing data of the calibration parameters, constructing a complete parameter spatial field, and combining it with time series data to form a spatiotemporal dynamic parameter matrix; S53, using the platform's built-in collaborative calculation module to perform cross-dimensional fusion calculations on the spatiotemporal dynamic parameter matrix, eliminating redundant information and conflicting data between parameters; S54, outputting the spatiotemporal distribution field of the calibration parameters after collaborative optimization, ensuring the spatial matching of parameter distribution with geological facies zones and engineering construction areas.

[0014] A dual-sweet spot evaluation method for shale reservoir geological engineering based on phase-controlled constraints and data-driven approaches is proposed. This method is implemented through different units, including: a phase-controlled elastic parameter inversion network construction and computation unit, which receives seismic wavefield data and geological facies data, completes the spatial inversion of elastic parameters, and transmits them to a geological engineering dual-attribute coupling processing unit; a geological engineering dual-attribute coupling processing unit, which receives the inverted elastic parameters and reservoir dual-attribute basic parameters, calculates the dual-attribute coupling coefficients through a coupling model, and sends them to a sweet spot parameter sensitivity analysis unit; and a sweet spot parameter sensitivity analysis unit, which runs a sensitivity prediction algorithm based on the coupling coefficient data. The system outputs the calibration master control parameter sorting results to the shale reservoir multi-parameter collaborative calculation unit; the shale reservoir multi-parameter collaborative calculation unit receives the calibration master control parameters, generates a spatiotemporally matched parameter distribution field through collaborative calculation, and transmits it to the double sweet spot evaluation index construction unit; the double sweet spot evaluation index construction unit sets dual-attribute evaluation thresholds based on the parameter distribution field, establishes an evaluation system, and sends it to the double sweet spot region definition and output unit; the double sweet spot region definition and output unit completes the double sweet spot region definition through threshold division and spatial overlay analysis, and outputs the evaluation results. Each unit performs bidirectional data interaction and collaborative work through a data bus.

[0015] Beneficial Effects: This invention proposes a dual-sweet spot evaluation method for shale reservoir geology and engineering based on phase control constraints and data-driven approaches. It integrates multi-source data and utilizes a dedicated computing platform to achieve comprehensive collaboration of core, logging, seismic, and construction response data, breaking down traditional data fragmentation and addressing the problem of insufficient multi-source information fusion. Through refined inversion of elastic parameters under phase control constraints, combined with a dual-attribute coupling mechanism between geology and engineering, a deep correlation is established between reservoir properties and engineering fracturing parameters, replacing traditional single-dimensional or simple superposition evaluation models and compensating for the shortcomings of fragmented evaluation of geological and engineering attributes. Simultaneously, parameter sensitivity analysis accurately identifies key controlling factors, and multi-scale collaborative calculation generates a spatiotemporally matched parameter distribution field, providing a scientific basis for the evaluation system construction and significantly improving the continuity and accuracy of sweet spot region definition. The supporting system units achieve bidirectional data interaction and collaborative work, ensuring efficient connection of the evaluation process and providing comprehensive and accurate technical support for well site deployment and fracturing design, thus contributing to the efficient development of shale reservoirs. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the overall process of the method of the present invention. Figure 2 This is a flowchart of method step S3 of the present invention; Figure 3 This is a flowchart of method step S4 of the present invention; Figure 4 This is a flowchart of step S5 of the method of the present invention; Figure 5This is a diagram showing the unit composition for implementing the method of the present invention. Detailed Implementation

[0017] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. The application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0018] like Figure 1 As shown, a shale reservoir geological engineering double sweet spot evaluation method based on phase control constraints and data-driven approach includes the following steps: S1, integrating core mineral composition, well logging elastic parameters, seismic reflection characteristics, and fracturing operation response data through a shale reservoir multi-parameter collaborative calculation platform to screen parameters associated with geological sweet spots, such as organic carbon content, porosity, and gas saturation, and associated with engineering sweet spots, such as brittleness index, Young's modulus, and geostress difference coefficient; S2, constructing a phase control elastic parameter inversion network, using seismic wavefield data as input, and combining geological facies zone division results to perform spatial inversion of elastic parameters; S 3. Based on the inversion results, a geological engineering dual-attribute coupling model is built, and a cross-dimensional correlation is established by integrating reservoir physical parameters and engineering fracturing parameters; S4. The sweet spot parameter sensitivity prediction algorithm is used to analyze the parameter influence of the dual-attribute coupling results and determine the calibration master control parameters; S5. The calibration master control parameters are calculated on a multi-scale collaborative basis through a shale reservoir multi-parameter collaborative calculation platform to generate a spatiotemporally matched parameter distribution field; S6. Based on the parameter distribution field, a shale reservoir geological engineering dual-sweet spot evaluation system is constructed, and the dual-sweet spot area is defined by dual-attribute threshold division and spatial overlay analysis.

[0019] Step S1 involves integrating multi-source data and screening core parameters using a multi-parameter collaborative computing platform for shale reservoirs. Specifically, this involves integrating four types of basic data: core mineral composition analysis data, well logging elastic parameter data, seismic reflection characteristic data, and fracturing operation response data. Core mineral composition data includes the content percentages of minerals such as quartz, clay, and feldspar; well logging elastic parameter data includes key indicators such as P-wave velocity, S-wave velocity, and density; seismic reflection characteristic data includes parameters such as reflection amplitude, frequency, and phase; and fracturing operation response data involves operational records such as pumping pressure, flow rate, and sand-fluid ratio. During data integration, the platform achieves compatibility and interoperability of data from different sources through data format standardization. A data quality control mechanism is used to remove invalid data with outliers and missing values ​​exceeding 15%, ensuring the integrity and reliability of the integrated data. Subsequently, based on the core requirements of shale reservoir geological engineering double sweet spot evaluation, three key parameters closely related to geological sweet spots were selected: organic carbon content, porosity, and gas saturation. The screening threshold for organic carbon content was set to be greater than 2.0%, the porosity screening range was 3.0% to 10.0%, and the gas saturation screening standard was greater than 60%. Simultaneously, three core parameters related to engineering sweet spots were selected: brittleness index, Young's modulus, and geostress difference coefficient. The screening threshold for brittleness index was set to be greater than 40, the Young's modulus screening range was 25 GPa to 50 GPa, and the geostress difference coefficient screening standard was less than 0.3. This step, through precise selection of core parameters, provides a targeted and reliable data foundation for subsequent inversion calculations, model construction, and evaluation analysis, effectively reducing the interference of redundant data on the evaluation results and improving the efficiency and accuracy of the overall evaluation process.

[0020] Step S2 involves constructing the phase-controlled elastic parameter inversion network and implementing the spatial inversion of elastic parameters. Specifically, based on the geological characteristics of the shale reservoirs in the study area, three main geological facies zones—deep-water shelf, shallow-water shelf, and tidal flat—are first identified. The spatial distribution and boundary characteristics of each facies zone are then clearly defined, serving as the basic conditions for phase-controlled constraints. Subsequently, a phase-controlled elastic parameter inversion network is constructed, comprising an input layer, a convolutional layer, a Bayesian neural network layer, a phase-controlled constraint layer, and an output layer. The input layer receives preprocessed seismic wavefield data with a sampling rate set to 2ms and a frequency range of 5Hz to 80Hz, covering the entire three-dimensional space of the study area. The convolutional layer has six convolutional kernels, each 3×3 in size, used to extract spatial feature information from the seismic wavefield data. The Bayesian neural network layer includes three hidden layers with 128, 64, and 32 neurons per layer, used to achieve nonlinear mapping between well logging data and seismic data. The phase-controlled constraint layer introduces geological facies zone boundary parameters and phase-controlled adjustment coefficients to spatially constrain the inversion process. During the inversion process, seismic wavefield data is input into the network and combined with the results of the divided geological facies zones. The elastic parameters are inverted and output through network iterative calculation. The number of iterations is set to 200, and the convergence error is controlled within 0.001. Finally, the three-dimensional spatial distribution data of elastic parameters such as P-wave velocity, S-wave velocity, and density are obtained. The spatial resolution of the inversion results reaches 20m×20m×5m, which can accurately reflect the differences in elastic parameters within different geological facies zones and provide high-precision parameter support for the subsequent construction of dual-attribute coupled models.

[0021] Step S3, based on the inversion results of Step S2, involves building a geological engineering dual-attribute coupling model and calculating the coupling coefficient. Specifically, the elastic parameters output from the phase-controlled elastic parameter inversion network, such as P-wave velocity, S-wave velocity, and density, are spatially matched with the geological sweet spot parameters measured in the rock core, such as porosity, organic carbon content, and gas saturation. The discrete rock core data is then interpolated into a three-dimensional grid system consistent with the elastic parameters using the Kriging interpolation method. The grid size is set to 20m × 20m × 5m to ensure accurate spatial correspondence of the parameters. Subsequently, pump pressure response data during fracturing operations are collected. Combined with the brittleness index, Young's modulus, and geostress difference coefficient obtained from well logging interpretation, a set of basic engineering sweet spot parameters is established. The pump pressure data is selected from the pressure values ​​during the stable phase of the operation, with a sampling interval of 10s. Finally, the geological sweet spot parameters and engineering sweet spot parameters are uniformly mapped to the same three-dimensional grid system using spatial interpolation, forming a parameter spatial alignment matrix with dimensions consistent with the three-dimensional grid system. Finally, a geological-engineering dual-attribute coupling model was constructed. This model introduces coupling adjustment parameters and achieves deep fusion of geological sweet spot attribute functions and engineering sweet spot attribute functions through nonlinear operations. The aligned dual-attribute parameters are substituted into the model to calculate the distribution data of dual-attribute coupling coefficients covering the entire domain. The coupling coefficient ranges from 0 to 1. The larger the value, the higher the matching degree between geological and engineering attributes. This step breaks the limitation of isolated analysis of geological and engineering attributes by establishing cross-dimensional parameter correlation, and provides comprehensive coupling parameter support for subsequent sensitivity analysis.

[0022] Step S4 utilizes a dessert parameter sensitivity prediction algorithm to analyze the parameter influence of the dual-attribute coupling results, in order to determine the calibration master control parameters. Specifically, 1000 sample points are randomly selected from the dual-attribute coupling coefficient distribution data. These sample points are evenly distributed across the entire three-dimensional space of the study area to ensure representativeness. Based on these sample points, a sample dataset is constructed, including all dessert evaluation parameters and their corresponding coupling coefficients, such as organic carbon content, porosity, gas saturation, brittleness index, Young's modulus, and geostress difference coefficient. The sample dataset is then divided into a training set and a test set in a 7:3 ratio. The training set is used for algorithm model training, and the test set is used for result verification. The algorithm iteration count is set to 150 times, with a convergence condition that the change in sensitivity index between two consecutive iterations is less than 0.0001. During training, the partial derivatives of each evaluation parameter with respect to the coupling coefficient are calculated using the training set data to obtain the initial sensitivity values. These values ​​are then normalized using the mean values ​​of each parameter to eliminate the influence of differences in parameter dimensions. In the verification phase, the stability of the sensitivity calculation results is verified using test set data. Abnormal sensitivity values ​​that deviate from the overall trend are corrected using an exponential decay function. Finally, the sensitivity ranking results of each parameter are output. Parameters with a sensitivity index greater than 0.6 are identified as calibration master parameters, which typically include three core parameters: organic carbon content, brittleness index, and porosity. This step, by accurately identifying and calibrating master parameters, can focus on the core factors affecting the formation of double sweet spots, providing a clear target direction for subsequent multi-parameter collaborative calculations.

[0023] Step S5 utilizes a multi-parameter collaborative calculation platform for shale reservoirs to perform multi-scale collaborative calculations on the calibration master control parameters determined in Step S4, generating a spatiotemporally matched parameter distribution field. Specifically, the calibration master control parameters (organic carbon content, brittleness index, and porosity) are first input into the multi-parameter collaborative calculation platform, with the spatial scale of each parameter set to 20m × 20m × 5m and the temporal resolution to one sedimentary cycle, ensuring that the spatiotemporal accuracy of the parameter calculations matches the research requirements. Subsequently, addressing the issue of missing key parameters in some areas, ordinary kriging interpolation is used to supplement the missing data. The interpolation search radius is set to 100m, and the minimum number of interpolation points is 8, constructing a complete parameter spatial field. Then, combined with the sedimentary evolution sequence of the shale reservoirs in the study area, parameter data from different periods are integrated to form a spatiotemporally dynamic parameter matrix. The matrix dimension is the number of grids in the study area × the number of time series × the number of parameters. Next, the platform's built-in collaborative computing module is invoked. This module integrates algorithms such as principal component analysis and random forest to perform cross-dimensional fusion calculations on the spatiotemporal dynamic parameter matrix. Principal component analysis extracts the main features of the parameters, eliminating redundant information between parameters. The random forest algorithm is used to handle conflicting data between parameters. The criteria for determining conflicting data is that the relative error of parameters from different sources is greater than 10%, and the processing method is a weighted average based on credibility weights. Finally, the spatiotemporal distribution field of the key parameters after collaborative optimization is output. This distribution field can accurately reflect the value characteristics of the calibration master control parameters in different spatiotemporal locations, and the spatial matching degree of the parameter distribution with the geological facies and engineering construction area reaches more than 90%, providing a high-quality parameter foundation for the subsequent construction of the double sweet spot evaluation system.

[0024] Step S6, based on the spatiotemporal distribution field of key parameters generated in Step S5, constructs a dual-sweet spot evaluation system for shale reservoir geology and engineering and defines the dual-sweet spot regions. Specifically, it first combines exploration and development practice data and industry standards for shale reservoirs in the study area to set dual-attribute evaluation thresholds for geological and engineering sweet spots. The geological sweet spot thresholds are set as follows: organic carbon content ≥ 2.5%, porosity ≥ 4.0%, and gas saturation ≥ 65%; the engineering sweet spot thresholds are set as follows: brittleness index ≥ 45, Young's modulus ≥ 30 GPa, and geostress difference coefficient ≤ 0.25. Subsequently, a dual-sweet spot evaluation system is constructed, comprising three stages: parameter standardization, weight allocation, and comprehensive scoring. Parameter standardization uses the extreme value method to transform each parameter to the 0-1 range. Weight allocation is determined using the analytic hierarchy process (AHP), with organic carbon content, brittleness index, and porosity having weights of 0.35, 0.35, and 0.30, respectively. The comprehensive scoring uses a weighted summation method to calculate the dual-sweet spot evaluation index. Regions with an evaluation index ≥ 0.7 are initially identified as candidate sweet spot regions. Finally, dual-attribute thresholding and spatial overlay analysis were conducted. First, geological sweet spots and engineering sweet spots were defined based on set thresholds. Then, spatial overlay operations were performed to obtain the overlapping area of ​​the geological and engineering sweet spots, i.e., the double sweet spot area. Boolean logic operations were used in the overlay analysis, and grid cells that simultaneously met the threshold conditions for both geological and engineering sweet spots were identified as double sweet spot cells. This step, through a scientific evaluation system and meticulous spatial analysis, accurately defined the spatial distribution range and boundary characteristics of the double sweet spot area, achieving a region division accuracy of 20m × 20m × 5m, providing precise target area guidance for engineering practices such as well site deployment and fracturing design.

[0025] Preferably, the expression for the phased elastic parameter inversion network is: , , in, For inverted elasticity parameters, These are the weighting coefficients. This is a convolution operation; Seis contains seismic wavefield data. These are the parameters of the phase-controlled constraint kernel function. For Bayesian neural network operations, For well logging data, These are network feature parameters. Phase control adjustment coefficient, Phase Here, Geo represents the geological facies control function, and Geo represents the geological facies zone data. These are the phase control boundary parameters.

[0026] Specifically, the phase-controlled elastic parameter inversion network expression is constructed based on the requirements of multi-source data fusion and geological phase control constraints. It combines the spatial feature extraction advantages of convolutional neural networks with the uncertainty quantification capabilities of Bayesian neural networks, while also incorporating the control effect of geological facies zones on the distribution of elastic parameters. It is clear that the elastic parameter response information included in seismic wavefield data needs to be extracted through convolution operations, hence the introduction of convolution operation terms. Well logging data has high vertical resolution and requires nonlinear mapping with seismic data through Bayesian neural networks, thus adding Bayesian neural network operation terms. Considering the differences in the variation patterns of elastic parameters within different geological facies zones, geological phase control function terms are added for constraint. Regarding parameter values, the weighting coefficients are determined based on the reliability of seismic and well logging data, and are set to 0.6 and 0.4 respectively; the phase control adjustment coefficient is set to 0.3 based on the accuracy and stability of geological facies division; the phase control constraint kernel function parameters are selected according to the frequency range of the seismic wavefield (5Hz to 80Hz); the network feature parameters are determined through iterative optimization using training set data, with the number of hidden layer neurons set to 128, 64, and 32 respectively; and the phase control boundary parameters are calculated based on the spatial boundary coordinates of the geological facies. This formula fully integrates the advantages of seismic, well logging, and geological facies data, and achieves spatial inversion of elastic parameters through multi-module collaborative computation. During implementation, the input data is first standardized in format, then substituted into the formula for iterative calculation. The number of iterations is set to 200, and the convergence error is controlled within 0.001. Finally, high-precision three-dimensional distribution data of elastic parameters are output, providing reliable parameter support for the subsequent construction of a dual-attribute coupled model.

[0027] Preferably, the expression for the geological engineering dual-attribute coupling model is: in, The coupling coefficient is a two-attribute coupling coefficient. For coupling adjustment parameters, For geological dessert property functions, Porosity For gas saturation, Organic carbon content, For engineering dessert attribute functions, The brittleness index, For Young's modulus, This is the coefficient of geostress difference. For the Laplace operator.

[0028] Specifically, the geological-engineering dual-attribute coupling model calculates the intrinsic correlation between geological sweet spots and engineering sweet spots. It is constructed based on the nonlinear coupling relationship between reservoir physical parameters and engineering fracturing parameters, combined with the Laplace operator's ability to characterize the spatial variation rate of parameters. First, the geological sweet spot attribute function is determined to consist of porosity, gas saturation, and organic carbon content, while the engineering sweet spot attribute function consists of brittleness index, Young's modulus, and geostress difference coefficient. The product of these two attributes reflects the synergistic effect of the two properties. The Laplace operator is introduced to calculate the spatial second derivative of the dual-attribute function, characterizing the drastic degree of spatial variation of the parameters and avoiding distortion of the coupling results due to abrupt parameter changes. A coupling adjustment parameter β is added to the denominator to prevent the denominator from being zero, and a coupling adjustment parameter α is introduced to the numerator to balance the weights of the dual-attribute functions. Regarding parameter values, the coupling adjustment parameter α was calibrated to 0.8 based on the geological characteristics and engineering practice data of the shale reservoirs in the study area, and β was set to 0.1. The weights of each parameter in the geological sweet spot attribute function were determined using the analytic hierarchy process (AHP), with organic carbon content, porosity, and gas saturation weights of 0.4, 0.3, and 0.3, respectively. Similarly, the weights of each parameter in the engineering sweet spot attribute function were determined using the AHP, with brittleness index, Young's modulus, and geostress difference coefficient weights of 0.4, 0.3, and 0.3, respectively. This formula achieves deep integration of geological and engineering attributes. In implementation, the elastic parameters obtained from phase control inversion are first spatially matched with the measured data, the basic parameters of the dual attributes are extracted and mapped to the same three-dimensional grid system, and then substituted into the formula to calculate the dual attribute coupling coefficient. The coupling coefficient ranges from 0 to 1; a larger value indicates a higher degree of matching between geological and engineering attributes, providing comprehensive coupling parameter support for subsequent parameter sensitivity analysis.

[0029] Preferably, the expression for the dessert parameter sensitivity prediction algorithm is: in, For parameter sensitivity index, For the first A dessert rating parameter, Let be the partial derivative of the coupling coefficient with respect to this parameter. The mean of the parameters, The sensitivity attenuation coefficient, This represents the total number of parameters.

[0030] Specifically, the dessert parameter sensitivity prediction algorithm is based on the core objective of parameter sensitivity analysis, combining the quantification ability of partial derivatives on the influence of parameters with the correction effect of exponential decay functions on outliers. By calculating the partial derivatives of the coupling coefficient with respect to each evaluation parameter, the marginal impact of a single parameter on the coupling result is quantified; the ratio of a parameter to its mean is introduced to eliminate dimensional differences between parameters, ensuring the comparability of the sensitivities of different types of parameters; and an exponential decay function is added to correct the sensitivity of parameters that deviate significantly from the mean, avoiding interference from extreme values ​​in the overall sensitivity ranking. Regarding parameter values, the sensitivity decay coefficient γ is calibrated to 0.5 based on the dispersion of the sample data; the total number of parameters n is determined to be 6 based on the number of core parameters selected, including organic carbon content, porosity, gas saturation, brittleness index, Young's modulus, and geostress difference coefficient; the parameter mean is calculated using the arithmetic mean of the sample data. This formula accurately identifies the calibration master control parameters that affect the formation of double sweet spots. In implementation, 1000 sample points are randomly selected from the dual-attribute coupling coefficient distribution data to construct a dataset, which is divided into training set and test set in a 7:3 ratio. The sensitivity index of each parameter is calculated by substituting it into the formula. The stability of the results is verified by the test set data. Finally, the parameter sensitivity ranking results are output. Parameters with a sensitivity index greater than 0.6 are identified as calibration master control parameters, providing a clear target direction for subsequent multi-parameter collaborative calculation.

[0031] Preferably, the collaborative calculation expression of the shale reservoir multi-parameter collaborative calculation platform is as follows: in, These are the parameter values ​​after collaborative computation. Principal component analysis operation, For Kriging interpolation operations, For spatial structure parameters, For random forest operations, For model training parameters, It is a quadratic coupling function. For time-dynamic parameters, For the calculated value of the geological dessert attribute, Calculate the value for the dessert attribute of the project.

[0032] Specifically, the collaborative computation expression of the shale reservoir multi-parameter collaborative computing platform is based on the fusion requirements of multi-scale and multi-source parameters. It integrates the spatial prediction capability of Kriging interpolation, the dimensionality reduction and redundancy removal capability of principal component analysis, and the conflict data processing capability of random forest, while introducing a quadratic coupling function to strengthen the construction of dual-attribute correlation. Kriging interpolation is used to spatially complete the inverted elastic parameters, generating a complete parameter space field; principal component analysis is used to reduce the dimensionality of the interpolated parameters, extracting key features and eliminating redundant information; the random forest algorithm is used to process conflict data between parameters, and a weighted average is performed based on confidence weights; finally, a quadratic coupling function is introduced to achieve a further fusion of geological and engineering dual attributes, improving the spatiotemporal matching of parameters. Regarding parameter values, spatial structure parameters are determined based on the geological structural characteristics of the study area and parameter spatial correlation analysis, with a search radius set to 100m; model training parameters are obtained through optimization of the training set data, with the number of decision trees in the random forest set to 100; temporal dynamic parameters are determined based on the sedimentary cycle period of the shale reservoir, with a time resolution set to one sedimentary cycle. This formula enables multi-scale collaborative optimization of calibration master parameters. During implementation, the calibration master parameters are first input into the platform, and the spatial scale is set to 20m×20m×5m. Missing data is supplemented by spatial interpolation to construct a spatiotemporal dynamic parameter matrix. Then, the matrix is ​​substituted into the formula for cross-dimensional fusion calculation to eliminate parameter redundancy and conflict. The spatiotemporal distribution field of key parameters after collaborative optimization is output, providing a high-quality parameter foundation for the construction of the double sweet spot evaluation system.

[0033] Preferably, the expression for the double sweet spot evaluation of shale reservoir geological engineering is: in, The rating index for both desserts is [not specified]. For activation function, For evaluation coefficients, The coupling coefficient is a two-attribute coupling coefficient. This is the parameter sensitivity index.

[0034] Specifically, the double-sweet spot evaluation expression for shale reservoir geology and engineering is constructed by combining the core objective of double-sweet spot evaluation with the comprehensive influence of collaborative calculation parameters, dual-attribute coupling coefficients, and parameter sensitivity indices, and introducing a Sigmoid activation function to normalize the evaluation results. The collaborative calculation parameters reflect the spatiotemporal distribution characteristics of the calibrated main control parameters, the dual-attribute coupling coefficients reflect the matching degree between geological and engineering attributes, and the parameter sensitivity index reflects the influence weight of each parameter. The effects of these three are integrated through linear combination. The Sigmoid activation function maps the comprehensive results to the 0-1 interval, facilitating threshold division and regional definition. Evaluation coefficients k1, k2, and k3 are used to balance the weights of the three and are calibrated using actual exploration and development data from the study area. Regarding parameter values, evaluation coefficients k1, k2, and k3 are calibrated to 0.4, 0.4, and 0.2, respectively, based on the geological and engineering characteristics and evaluation accuracy requirements of the shale reservoirs in the study area. The collaborative calculation parameters are the spatiotemporal distribution values ​​of the calibrated main control parameters output by the multi-parameter collaborative calculation platform; the dual-attribute coupling coefficients are the calculation results of the geological engineering dual-attribute coupling model; and the parameter sensitivity index is the output result of the sweet spot parameter sensitivity prediction algorithm. This formula enables quantitative evaluation of double sweet spots. In implementation, the dual-attribute evaluation threshold is first set based on the parameter distribution field. Then, the collaborative calculation parameters, dual-attribute coupling coefficient, and parameter sensitivity index are substituted into the formula to calculate the double sweet spot evaluation index. Areas with an evaluation index ≥ 0.7 are initially identified as sweet spot candidate areas. Combining dual-attribute threshold division and spatial overlay analysis, the double sweet spot area is finally defined, providing accurate target area guidance for engineering practices such as well site deployment and fracturing design.

[0035] Preferred, such as Figure 2 As shown, S3 includes the following sub-steps: S31, spatially matching the elastic parameters output by the phase-controlled elastic parameter inversion network with the measured physical property parameters of the rock core, and extracting the basic parameters of the geological sweet spot, such as porosity, organic carbon content, and gas saturation, within the corresponding geological unit; S32, collecting the pumping pressure response data during fracturing operations and the brittleness index, Young's modulus, and geostress difference coefficient interpreted from well logging, and establishing a set of basic parameters for the engineering sweet spot; S33, mapping the geological sweet spot parameters and the engineering sweet spot parameters to the same three-dimensional grid system through spatial interpolation methods to form a parameter space alignment matrix; S34, substituting the aligned dual-attribute parameters into the calculation based on the coupled model structure to generate full-domain coverage dual-attribute coupling coefficient distribution data.

[0036] Specifically, step S3 is implemented through four sub-steps to accurately build the geological engineering dual-attribute coupling model and calculate the coupling coefficient. In S31, the elastic parameters such as P-wave velocity, S-wave velocity, and density output from the phase-controlled elastic parameter inversion network are matched one-to-one with the geological sweet spot basic parameters such as porosity, organic carbon content, and gas saturation measured in the rock core, according to spatial coordinates. The matching accuracy is controlled within 5 meters to ensure accurate parameter correspondence within the same geological unit, thereby extracting complete geological sweet spot basic parameters for each geological unit. In S32, by collecting continuously recorded pumping pressure response data during fracturing operations and combining it with the brittleness index, Young's modulus, and geostress difference coefficient obtained from well logging interpretation, a set of engineering sweet spot basic parameters covering the entire study area is established according to the correspondence between construction blocks and geological units. The parameter recording interval is set to 10 seconds to ensure data continuity. S33 employs the Kriging interpolation method to map the discretely distributed geological sweet spot parameters and engineering sweet spot parameters into a unified 3D grid system with a resolution of 20m × 20m × 5m, forming a parameter spatial alignment matrix with consistent dimensions and spatial alignment, ensuring that the dual-attribute parameters are completely matched in spatial location. S34, based on a pre-defined geological-engineering dual-attribute coupling model structure, substitutes the spatially aligned dual-attribute parameters into the model one by one for calculation. Through 200 iterations, it generates full-domain coverage of dual-attribute coupling coefficient distribution data, with coupling coefficient values ​​ranging from 0 to 1. This step, implemented in stages, achieves effective fusion of dual-attribute parameters and accurate calculation of coupling coefficients, providing comprehensive and reliable basic data for subsequent parameter sensitivity analysis.

[0037] Preferred, such as Figure 3 As shown, step S4 includes the following sub-steps: S41, randomly extract sample points from the dual-attribute coupling coefficient distribution data to construct a sample dataset including all sweet spot parameters and coupling coefficients; S42, divide the sample dataset into a training set and a test set according to the proportion, and set the number of algorithm iterations and convergence conditions; S43, calculate the partial derivatives of each parameter with respect to the coupling coefficient using the training set data, calculate the initial sensitivity value, and perform normalization processing in combination with the parameter mean; S44, verify the stability of the sensitivity calculation results using the test set data, correct abnormal sensitivity values ​​through an exponential decay function, and output the final parameter sensitivity ranking results.

[0038] Specifically, step S4 involves four sub-steps to complete the sensitivity analysis and calibration of the main control parameters for the dessert parameters, ensuring the accuracy and reliability of the analysis results. S41: From the dual-attribute coupling coefficient distribution data of the full-domain 3D mesh, 1000 sample points are randomly sampled using a uniform sampling method. These sample points cover all geological facies zones and engineering construction areas. Based on these sample points, all dessert evaluation parameters and corresponding coupling coefficients, such as organic carbon content, porosity, gas saturation, brittleness index, Young's modulus, and geostress difference coefficient, are extracted to construct a complete sample dataset. S42: The sample dataset is divided into a training set and a test set in a 7:3 ratio. The training set is used for algorithm model training, and the test set is used for result verification. The algorithm iteration count is set to 150 times, with the convergence condition being that the change in sensitivity index between two adjacent iterations is less than 0.0001, ensuring sufficient and stable algorithm training. S43: The training set data is used to... By using the partial derivative method, the marginal impact of each evaluation parameter on the coupling coefficient is calculated to obtain the initial sensitivity value. Then, the arithmetic mean of each parameter is used for normalization to eliminate the influence of differences in the dimensions of different parameters, making the sensitivity values ​​comparable. S44 substitutes the test set data into the trained algorithm model to verify the stability of the sensitivity calculation results. For abnormal sensitivity values ​​that deviate from the overall trend and have a relative error greater than 10%, an exponential decay function is used for correction. Finally, the sensitivity ranking results of each parameter are output. Parameters with a sensitivity index greater than 0.6 are identified as calibration master parameters. This step, through step-by-step sample processing, model training, result verification and correction, accurately identifies the core factors affecting the formation of double sweet spots, providing a clear target for subsequent multi-parameter collaborative calculation.

[0039] Preferred, such as Figure 4 As shown, step S5 includes the following sub-steps: S51, inputting the preceding calibration parameters from the parameter sensitivity ranking results into the shale reservoir multi-parameter collaborative calculation platform, and setting the spatial scale and temporal resolution of each parameter; S52, using a spatial interpolation algorithm to supplement the missing data of the calibration parameters, constructing a complete parameter spatial field, and combining it with time series data to form a spatiotemporal dynamic parameter matrix; S53, performing cross-dimensional fusion calculations on the spatiotemporal dynamic parameter matrix through the platform's built-in collaborative calculation module to eliminate redundant information and conflicting data between parameters; S54, outputting the spatiotemporal distribution field of the calibration parameters after collaborative optimization, ensuring the spatial matching of parameter distribution with geological facies zones and engineering construction areas.

[0040] Specifically, step S5 involves four sub-steps to achieve multi-scale collaborative calculation of calibration master control parameters and generation of a spatiotemporally matched parameter distribution field. S51 inputs the determined calibration master control parameters (typically organic carbon content, brittleness index, and porosity) into the shale reservoir multi-parameter collaborative calculation platform. Based on the exploration accuracy requirements of the study area, the spatial scale of each parameter is set to 20m × 20m × 5m, and the time resolution is one sedimentary cycle, ensuring that the spatiotemporal accuracy of the parameter calculation matches the actual needs. S52 addresses the issue of missing key parameters in some areas due to data acquisition limitations by using a common kriging interpolation algorithm to supplement the missing data. The interpolation search radius is set to 100m, and the minimum number of interpolation points is 8, constructing a complete parameter spatial field. This field is then combined with the sedimentary evolution sequence of the shale reservoir in the study area to integrate parameter data from different sedimentary periods, forming a dimension of the number of grids in the study area × the number of time series × ... The system generates a spatiotemporal dynamic parameter matrix with a number of parameters. S53 calls the platform's built-in collaborative computing module, which integrates principal component analysis and random forest algorithms. Principal component analysis is used to extract the main features of the parameters and eliminate redundant information between them. Then, the random forest algorithm is used to process conflicting data with relative errors greater than 10% between parameters, and a weighted average is performed according to credibility weights to achieve optimized parameter fusion. S54 outputs the spatiotemporal distribution field of the key parameters after collaborative optimization. This distribution field is verified through spatial matching to ensure that the spatial matching degree between the parameter distribution and the geological facies and engineering construction area reaches over 90%. It can accurately reflect the value characteristics of the calibrated main control parameters at different spatiotemporal locations, providing a high-quality parameter foundation for the subsequent construction of a double-sweetness evaluation system.

[0041] like Figure 5 As shown, a double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approaches is proposed. This method is implemented through different units, including: The phase-controlled elastic parameter inversion network construction and computation unit receives seismic wavefield data and geological facies data, completes the spatial inversion of elastic parameters, and transmits them to the geological engineering dual-attribute coupling processing unit. The geological engineering dual-attribute coupling processing unit receives the inverted elastic parameters and the reservoir dual-attribute basic parameters, calculates and generates the dual-attribute coupling coefficient through the coupling model, and sends it to the sweet spot parameter sensitivity analysis unit. The sweet spot parameter sensitivity analysis unit runs a sensitivity prediction algorithm based on coupling coefficient data and outputs the calibration of the main control parameters to the shale reservoir multi-parameter collaborative calculation unit. The shale reservoir multi-parameter collaborative calculation unit receives the calibrated master control parameters, generates a spatiotemporally matched parameter distribution field through collaborative calculation, and transmits it to the double sweet spot evaluation index construction unit. The double sweet spot evaluation index construction unit sets the dual attribute evaluation thresholds based on the parameter distribution field, establishes the evaluation system, and sends it to the double sweet spot region definition and output unit; The double sweet spot region definition and output unit defines the double sweet spot region through threshold division and spatial overlay analysis, and outputs the evaluation results. Each unit interacts and collaborates through a data bus.

[0042] The formulas in this invention achieve fusion calculations of different scalar and vector parameters, constructing a unified computational framework through standardization, dimensional adaptation mechanisms, and operator mapping design. For scalar parameters (such as organic carbon content, porosity, brittleness index, etc., which only contain numerical values), the formulas eliminate dimensional differences by introducing weighting coefficients and mean normalization, making scalars with different physical meanings comparable at the numerical level. For vector parameters (such as seismic wavefield data, geological facies spatial distribution data, etc., which contain directional or spatial dimensional information), operators such as convolution and kriging interpolation are used to transform vector information into quantifiable scalar features. For example, spatial feature parameters are extracted from seismic wavefield vectors through convolution operations, and geological facies vectors are transformed into boundary-constrained scalars through phase control functions. Meanwhile, the coupling adjustment parameters and sensitivity attenuation coefficients in the formula can dynamically balance the numerical ranges after scalar and vector transformations. Operations such as the Laplacian operator and principal component analysis further realize the mathematical correlation of cross-dimensional parameters. For example, the scalar parameters of geological sweet spots and the vector-derived parameters of engineering sweet spots are correlated by multiplying the numerator of the coupling model, and the denominator is adapted to the dimensional differences between the two through spatial change rate calculation. In the end, all parameters are calculated collaboratively in a unified mathematical space to ensure the scientificity and rationality of the fusion results.

[0043] A dual-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approaches is proposed. This method integrates multi-dimensional data from core samples, well logging, seismic data, and construction response data through a dedicated computing platform system. This overcomes the limitations of traditional isolated data use, enabling efficient collaboration of parameters at different scales and of different types, providing comprehensive data support for the evaluation. By relying on phase control constraints to achieve spatial inversion of elastic parameters, combined with targeted parameter sensitivity analysis, the method accurately identifies key factors influencing sweet spot formation, significantly improving the reliability of parameter characterization. Simultaneously, a dedicated evaluation system and supporting systems are constructed. Through bidirectional data interaction and workflow integration among functional units, the efficiency and coherence of the entire evaluation process are ensured, forming an integrated technical chain from data input to result output.

[0044] This method addresses the problem of fragmented evaluation of geological and engineering attributes. By constructing a dual-attribute coupling mechanism, it deeply explores the intrinsic correlation between reservoir properties and engineering fracturing parameters, replacing the traditional single-dimensional or simple superposition evaluation mode, making the evaluation results more closely reflect the true potential of the reservoir. Addressing the issues of insufficient multi-source data fusion and ambiguous identification of key factors, it achieves cross-dimensional data fusion and optimization through multi-parameter collaborative calculation, eliminating parameter redundancy and conflicts. Combined with spatiotemporal parameter matching processing, it improves the spatial continuity and accuracy of the evaluation results, providing scientific and reliable technical support for well location deployment and fracturing design. This effectively solves the pain point of insufficient evaluation accuracy of traditional methods under complex reservoir conditions.

[0045] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various equivalent changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach, characterized in that, Includes the following steps: S1 integrates core mineral composition, well logging elastic parameters, seismic reflection characteristics, and fracturing operation response data through a multi-parameter collaborative calculation platform for shale reservoirs, and screens parameters related to geological sweet spots such as organic carbon content, porosity, gas saturation, and engineering sweet spots such as brittleness index, Young's modulus, and geostress difference coefficient. S2, construct a phase-controlled elastic parameter inversion network, take seismic wavefield data as input, and combine geological facies zone division results to perform spatial inversion of elastic parameters; S3, a geological engineering dual-attribute coupled model is built based on the inversion results, and cross-dimensional correlation is established by integrating the parameters filtered by S1; S4. The dessert parameter sensitivity prediction algorithm is used to analyze the parameter influence of the dual-attribute coupling results and determine the calibration master control parameter. S5 uses a multi-parameter collaborative calculation platform for shale reservoirs to perform multi-scale collaborative calculations on the calibrated master control parameters, generating a spatiotemporally matched parameter distribution field; S6, Based on the parameter distribution field, a double sweet spot evaluation system for shale reservoir geological engineering is constructed, and the double sweet spot region is defined by dual attribute threshold division and spatial overlay analysis. The expression for the phased elastic parameter inversion network is: , , in, For inverted elasticity parameters, These are the weighting coefficients. This is a convolution operation; Seis contains seismic wavefield data. These are the parameters of the phase-controlled constraint kernel function. For Bayesian neural network operations, For well logging data, These are network feature parameters. Phase control adjustment coefficient, Phase Here, Geo represents the geological facies control function, and Geo represents the geological facies zone data. These are phase control boundary parameters; The expression for the geological engineering dual-attribute coupling model is as follows: in, The coupling coefficient is a two-attribute coupling coefficient. For coupling adjustment parameters, For geological dessert property functions, Porosity For gas saturation, Organic carbon content, For engineering dessert attribute functions, The brittleness index, For Young's modulus, This is the coefficient of geostress difference. For the Laplace operator.

2. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 1, characterized in that, The expression for the dessert parameter sensitivity prediction algorithm is as follows: in, For parameter sensitivity index, For the first A dessert rating parameter, Let be the partial derivative of the coupling coefficient with respect to this parameter. The mean of the parameters, The sensitivity attenuation coefficient, This represents the total number of parameters.

3. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 2, characterized in that, The collaborative calculation expression of the shale reservoir multi-parameter collaborative calculation platform is as follows: in, These are the parameter values ​​after collaborative computation. Principal component analysis operation, For Kriging interpolation operations, For spatial structure parameters, For random forest operations, For model training parameters, It is a quadratic coupling function. For time-dynamic parameters, For the calculated value of the geological dessert attribute, Calculate the value for the dessert attribute of the project.

4. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 3, characterized in that, The expression for the double sweet spot evaluation of shale reservoir geological engineering is as follows: in, The rating index for both desserts is [not specified]. For activation function, For evaluation coefficients, The coupling coefficient is a two-attribute coupling coefficient. This is the parameter sensitivity index.

5. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 1, characterized in that, S3 includes the following steps: S31, Spatially match the elastic parameters output by the phase-controlled elastic parameter inversion network with the measured physical property parameters of the rock core, extract the porosity, organic carbon content and gas saturation within the geological unit, and establish a basic parameter set for the geological sweet spot; S32, collect pumping pressure response data and well logging interpretation of brittleness index, Young's modulus, and geostress difference coefficient during fracturing operations, and establish a basic parameter set for the engineering sweet spot; S33 uses spatial interpolation to map the geological sweet spot basic parameter set and the engineering sweet spot basic parameter set to the same three-dimensional mesh system, forming a parameter space alignment matrix; S34, based on the geological engineering dual-attribute coupling model structure, substitutes the aligned dual-attribute parameters into the calculation to generate dual-attribute coupling coefficient distribution data covering the entire domain.

6. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 5, characterized in that, S4 includes the following sub-steps: S41, randomly sample points from the dual-attribute coupling coefficient distribution data to construct a sample dataset including all dessert parameters and coupling coefficients; S42, divide the sample dataset into training and test sets according to the proportion, and set the number of algorithm iterations and convergence conditions; S43, calculate the partial derivatives of each parameter with respect to the coupling coefficient using the training set data, calculate the initial sensitivity value, and then normalize it by combining the parameter mean. S44 uses test set data to verify the stability of the sensitivity calculation results, corrects abnormal sensitivity values ​​through exponential decay function, and outputs the final parameter sensitivity ranking results.

7. The double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach as described in claim 6, characterized in that, S5 includes the following steps: S51, input the pre-calibration parameters from the parameter sensitivity ranking results into the shale reservoir multi-parameter collaborative calculation platform, and set the spatial scale and temporal resolution of each parameter; S52 uses a spatial interpolation algorithm to supplement the missing data of the calibration parameters, constructs a complete parameter space field, and combines it with time series data to form a spatiotemporal dynamic parameter matrix; S53 performs cross-dimensional fusion calculations on the spatiotemporal dynamic parameter matrix through the platform's built-in collaborative computing module, eliminating redundant information and conflicting data between parameters; S54 outputs a spatiotemporally matched parameter distribution field after collaborative optimization, ensuring the spatial matching of parameter distribution with geological facies zones and engineering construction areas.

8. A double-sweet spot evaluation method for shale reservoir geological engineering based on phase control constraints and data-driven approach, as described in any one of claims 1-7, characterized in that, This method is implemented through different units, including: The phase-controlled elastic parameter inversion network construction and computation unit receives seismic wavefield data and geological facies data, completes the spatial inversion of elastic parameters, and transmits them to the geological engineering dual-attribute coupling processing unit. The geological engineering dual-attribute coupling processing unit calculates and generates dual-attribute coupling coefficients through a coupling model and sends them to the sweet spot parameter sensitivity analysis unit. The sweet spot parameter sensitivity analysis unit runs a sensitivity prediction algorithm based on coupling coefficient data and outputs the calibration of the main control parameters to the shale reservoir multi-parameter collaborative calculation unit. The shale reservoir multi-parameter collaborative calculation unit receives the calibrated master control parameters, generates a spatiotemporally matched parameter distribution field through collaborative calculation, and transmits it to the double sweet spot evaluation index construction unit. The double sweet spot evaluation index construction unit sets the dual attribute evaluation thresholds based on the parameter distribution field, establishes the evaluation system, and sends it to the double sweet spot region definition and output unit; The double sweet spot region definition and output unit defines the double sweet spot region through threshold division and spatial overlay analysis, and outputs the evaluation results. Each unit interacts and collaborates through a data bus.

Images