Multi-source data fusion analysis method and analysis system for geological exploration

Abstract

The invention belongs to the technical field of data fusion and processing, and particularly relates to a multi-source data fusion analysis method and system for geological prospecting, and the method comprises the steps: 1, setting a plurality of detection points at the superficial part of a target geological prospecting region, marking a plurality of unknown points at the depth part of the target geological prospecting region, calibrating a position parameter of each detection point; collecting multi-source geological data of each detection point; 2, performing data fusion on the multi-source geological data of each detection point; 3, according to the position parameter of each detection point, constructing a multi-source composite covariance matrix of all the detection points in combination with a variation function; and 4, predicting multi-source geological parameters of the unknown point according to the multi-source composite covariance matrix. According to the method, the precision of geological parameter prediction is improved, the heterogeneity of the reservoir is effectively dealt with, and the stability and the physical reasonability of interpolation calculation are enhanced.

Description

Technical Field

[0001] The present invention belongs to the technical field of data fusion and processing, and particularly relates to a multi-source data fusion analysis method and analysis system for geological exploration. Background Art

[0002] Geological exploration is an indispensable link in resource development, environmental monitoring and engineering construction. The accurate prediction of underground geological parameters directly affects the reliability and economy of exploration. In traditional geological exploration technologies, the acquisition of geological parameters usually relies on limited drilling data, logging data and geophysical exploration data, and is estimated at unknown points through interpolation methods. However, due to the heterogeneity and multi-scale coupling effects of underground reservoirs, the analysis methods based on a single data source often cannot comprehensively reflect the formation characteristics, resulting in limited prediction accuracy. Therefore, in recent years, multi-source data fusion has become an important research direction in the field of geological exploration. By integrating various geophysical measurement parameters, higher-precision prediction of geological parameters can be achieved. However, there are still certain limitations in the existing technologies in aspects such as data fusion, variogram construction, covariance matrix calculation and interpolation prediction, resulting in limited applicability under complex reservoir conditions.

[0003] Traditional geological parameter interpolation methods mainly include inverse distance weighted interpolation (IDW), Kriging interpolation, geostatistical modeling (such as sequential Gaussian simulation, SGS), etc. These methods are all based on certain spatial variability assumptions and predict unknown points through statistical methods. For example, the Kriging interpolation method uses the variogram to calculate the spatial correlation of points at different positions, and thus assigns different weights in the interpolation process to make the predicted value optimal in a statistical sense. However, traditional Kriging interpolation usually only relies on a single parameter (such as porosity or permeability) for calculation and fails to fully utilize the multiple parameter information in logging data. In addition, in reservoirs with complex geological environments and significant heterogeneity, the prediction accuracy of the Kriging interpolation method is often limited and it is difficult to capture the characteristics of complex reservoirs.

[0004] In recent years, with the development of computer technology, machine learning and artificial intelligence methods have gradually been applied to the prediction of geological parameters. For example, methods such as support vector machine (SVM), random forest (RF), deep learning (such as convolutional neural network CNN, long short-term memory network LSTM) are used to build geological parameter prediction models. Compared with traditional methods, machine learning methods can extract features using a large amount of data and establish complex non-linear mapping relationships. However, the prediction accuracy of machine learning methods usually depends on a large amount of high-quality training data, and in actual geological exploration, it is often costly to obtain large-scale labeled data. In addition, the interpretability of machine learning models is poor, making it difficult to combine their prediction results with geological theories, thus affecting their practical applications in geological modeling. Summary of the Invention

[0005] In view of this, the main object of the present invention is to provide a multi-source data fusion analysis method and analysis system for geological exploration, which improves the accuracy of geological parameter prediction, effectively addresses the heterogeneity of reservoirs, enhances the stability and physical rationality of interpolation calculations, and improves the applicability in complex geological environments, and is widely applicable to fields such as oil and gas exploration, geothermal resource evaluation, and formation stability analysis.

[0006] The technical solution adopted by the present invention is as follows: A multi-source data fusion analysis method for geological exploration, the method comprising: Step 1: Set a plurality of detection points in the shallow part of the target geological exploration area, mark a plurality of unknown points in the deep part of the target geological exploration area, and calibrate the position parameters of each detection point; collect multi-source geological data of each detection point; Step 2: Perform data fusion on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters in the target geological exploration area; Step 3: According to the position parameters of each detection point, construct a multi-source composite covariance matrix of all detection points in combination with the variogram; each element in the multi-source composite covariance matrix is the multi-source composite covariance between different detection points; Step 4: Predict the multi-source geological parameters of the unknown points according to the multi-source composite covariance matrix.

[0007] Further, the multi-source geological parameters include: rock density , unit: g / cm³; solid particle density , unit: g / cm³; porosity ; horizontal permeability , unit: mD; water saturation ; total organic carbon content ; thermal conductivity , unit: W / m·K; rock Young's modulus , unit: GPa; fracture density , unit: pieces / m; effective pressure , unit: MPa; where represents the subscript index of the detection point or unknown point, taking positive integer values, representing the th detection point or unknown point, can be replaced by any subscript index letter, including: , or .

[0008] Further, the formula of the variogram is: ; Among them, and both represent the subscript index of the detection point, and the value is a positive integer, representing the th detection point or the th detection point respectively; represents the variation value between the th detection point and the th detection point; represents the nugget effect value of the th detection point; represents the nugget effect value of the th detection point; represents the vertical distance difference between the th detection point and the th detection point; represents the horizontal distance difference between the th detection point and the th detection point.

[0009] Furthermore, for the th detection point, the calculation formula for its nugget effect value is as follows: ; Among them, is the depth of the th detection point; is the initial nugget effect value. For the loose sediment layer the value range is ; for the shale layer the value range is ; for the carbonate rock layer the value range is ; for the igneous rock layer the value range is .

[0010] Furthermore, the multi-source composite covariance between the th detection point and the th detection point is calculated using the following formula: ; Among them, is the multi-source composite covariance between the th detection point and the th detection point; is the difference in the nugget effect value between the th detection point and the th detection point; represents the The elastic modulus at a detection point, with the unit of GPa; represents the elastic modulus at a detection point, with the unit of GPa; represents the maximum horizontal principal stress at a detection point, with the unit of MPa; represents the rock shear modulus at a detection point, with the unit of GPa; represents the shear strength at a detection point, with the unit of MPa; represents the vertical stress at a detection point, with the unit of MPa; represents the pore fluid pressure at a detection point, with the unit of MPa; represents the thickness of the oil and gas bearing layer at a detection point, with the unit of m; represents the thermal diffusivity at a detection point, with the unit of m² / s; represents the geostress adjustment coefficient between the detection point and the detection point;

[0011] Furthermore, for the thermal diffusivity at the detection point , it is calculated using the following formula: ; where represents the specific heat capacity at a detection point, with the unit of .

[0012] Furthermore, the geostress adjustment coefficient between the detection point and the detection point is calculated using the following formula: .

[0013] Furthermore, the multi-source geological parameters of unknown points are predicted based on the multi-source composite covariance between detection points through the following formula: ; where is the multi-source geological parameter of the predicted unknown point; is the standard value of the multi-source geological parameter of the detection point, which is a set value, and the standard values of different multi-source geological parameters are different; Multi-source geological parameters of detection points; is the weight of the th detection point; is the brittleness index at the th detection point; is the natural gamma value at the th detection point, with the unit of API; represents the shale index at the th detection point; is the spontaneous potential at the th detection point, with the unit of mV; is the true resistivity at the th detection point, with the unit of Ω·m; is the relative density of natural gas at the

[0014] Furthermore, through the following formula, the weight of the th detection point is calculated using the following formula, where ; wherein, is the Lagrange multiplier, ensuring that the sum of weights is 1; represents the depth difference between the th detection point and the unknown point; is the covariance matrix; represents the covariance between the th detection point and the th detection point, where

[0015] A multi-source data fusion analysis system for geological exploration, the system includes: a data acquisition part, which is used to set multiple detection points in the shallow part of the target geological exploration area, mark multiple unknown points in the deep part of the target geological exploration area, and calibrate the position parameters of each detection point; collect multi-source geological data of each detection point; a variogram construction part, which is used to perform data fusion on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters in the target geological exploration area; a composite covariance construction part, which is used to combine the variogram according to the position parameters of each detection point to construct a multi-source composite covariance matrix of all detection points; each element in the multi-source composite covariance matrix is the multi-source composite covariance between different detection points; an unknown point data prediction part, which is used to predict the multi-source geological parameters of the unknown point according to the multi-source composite covariance matrix.

[0016] Adopting the above technical solutions, the present invention has the following beneficial effects: Traditional geological parameter prediction methods usually rely on a single variable (such as porosity or permeability) for interpolation. However, the present invention constructs a more complete geological parameter prediction system by introducing multiple logging parameters, such as brittleness index, natural gamma value, shale index, true resistivity, relative density of natural gas, etc. Since the geological properties of the formation are usually affected by multiple physical factors, it is difficult to accurately describe the spatial variation of the underground reservoir relying only on a single variable. The multi-source data fusion method of the present invention incorporates the coupling relationship between different physical logging parameters into the prediction model by constructing a multi-source composite covariance matrix, enabling the final predicted value to comprehensively consider the lithology, permeability, and fluid distribution characteristics of the reservoir, thereby improving the prediction accuracy. The underground geological environment is highly heterogeneous. Especially in fault zones, fracture-developed areas, or multiphase reservoirs, the spatial distribution of geological parameters often has complex variation laws. The traditional Kriging interpolation method mainly relies on calculating the correlation based on spatial distance, but it cannot accurately describe the heterogeneous variation of the reservoir. When constructing the variogram in the present invention, a new parametric formula is adopted. This formula not only considers the variation of geological parameters in the horizontal and vertical directions but also combines the ratio relationships of multiple variables such as rock density, effective pressure, porosity, permeability, and water saturation, thus being able to more accurately characterize the spatial variability of the reservoir. In addition, by introducing an exponential decay model of the nugget effect into the variogram, the present invention can more effectively suppress the influence of local abnormal data, enabling the prediction model to still maintain high stability in the face of complex geological conditions and large measurement errors. Traditional interpolation methods usually adopt a fixed spatial weight allocation method. However, the present invention constructs an interpolation weight calculation model based on the multi-source composite covariance matrix, making the weight not only depend on the spatial distance but also be able to be dynamically optimized in combination with geomechanical parameters, logging data characteristics, and reservoir properties. In the interpolation weight calculation formula of the present invention, in addition to the traditional covariance matrix calculation, a ground stress adjustment coefficient is additionally introduced, enabling the prediction model to adaptively adjust the weight distribution in a complex stress environment. For example, in high-stress areas, the pore structure and permeability characteristics of the rock may be significantly affected by stress. The method of the present invention can correct the correlation between different detection points through the ground stress adjustment coefficient, making the interpolation prediction more accurate. In addition, a depth correction factor is also introduced in the weight calculation process of the present invention. This factor can dynamically adjust the influence degree of different detection points during the prediction process, making the interpolation model more accurate when predicting deep reservoirs. Brief Description of the Drawings

[0017] Figure 1 It is a schematic flow chart of a multi-source data fusion analysis method for geological exploration provided by an embodiment of the present invention; Figure 2Schematic diagram of the system structure of a multi-source data fusion analysis system for geological exploration provided by an embodiment of the present invention; Detailed implementation manners

[0018] All features disclosed in this specification, or all steps in the disclosed methods or processes, except for mutually exclusive features and / or steps, can be combined in any manner.

[0019] Any feature disclosed in this specification (including any additional claims, abstract) can be replaced by other equivalent or similar-purpose alternative features unless specifically stated. That is, unless specifically stated, each feature is only an example of a series of equivalent or similar features.

[0020] Example 1: Refer to Figure 1 , a multi-source data fusion analysis method for geological exploration, the method includes: Step 1: Set a plurality of detection points in the shallow part of the target geological exploration area, mark a plurality of unknown points in the deep part of the target geological exploration area, and calibrate the position parameters of each detection point; collect multi-source geological data of each detection point; The shallow part generally refers to the area near the surface or the relatively shallow area underground. The strata in this part are relatively easy to directly observe and collect data through traditional geophysical methods (such as surface measurement, shallow drilling or seismic exploration). The depth range of the shallow part does not have a fixed value, but is divided according to the specific exploration object. For example, in oil and gas exploration, the shallow part may refer to the strata within the range of 0 to 500 meters, while in groundwater exploration or environmental geological survey, the shallow part may only involve the range of 0 to 50 meters. The main purpose of arranging the detection points in the shallow part is to obtain basic data such as surface lithology, structural form, and preliminary physical properties, and provide constraints for deep speculation through this information. Since the geological information in the shallow part is relatively easy to obtain, in the method of the present invention, the density of the detection points in this part is usually relatively high to ensure a relatively complete distribution of geological parameters.

[0021] The deep part refers to the deeper underground area, usually beyond the direct detection range of traditional surface measurement means, and relies on geophysical inversion or well logging measurement for analysis. It is usually difficult and costly to obtain geological parameters of the deep part. Therefore, a large number of detection points cannot be arranged like the shallow part, but rely on a small amount of known well data and methods such as geophysical logging and seismic inversion for estimation. For example, in oil and gas exploration, the deep part may refer to the underground reservoir at a depth of 500 meters to several kilometers, while in geothermal resource exploration, the deep part may refer to the high-temperature geothermal reservoir area above 1000 meters. The method of the present invention does not directly set a large number of detection points in the deep part, but adopts the marking method of unknown points, that is, multiple underground target points to be predicted are set in the modeling stage, and through the method of multi-source data fusion, combined with the known information of the shallow part, mathematical modeling and interpolation prediction techniques are used to infer the geological parameters of these unknown points.

[0022] Setting multiple detection points in the shallow part and simultaneously marking multiple unknown points in the deep part to form a spatial distribution system that includes both known information and covers unknown areas is the basis for achieving high-precision prediction. In traditional geological exploration methods, usually a single data source is used for analysis. For example, the properties of underground media are inferred through density measurement or acoustic logging data. However, this method often has large errors when encountering areas with complex geological layers and drastic lithological changes, and it is difficult to comprehensively reflect the multi-scale characteristics of geological bodies. The present invention introduces a multi-source data acquisition strategy, enabling the detection points to cover different geophysical and geological parameters, including key geological parameters such as rock density, solid particle density, porosity, permeability, water saturation, total organic carbon content, thermal conductivity, Young's modulus, fracture density, and effective pressure. These parameters each have unique physical meanings when describing the underground structure. For example, permeability reflects the ability of fluid flow inside the rock, while Young's modulus determines the mechanical response of the rock. The combination of the two can more accurately characterize the mechanical behavior of the reservoir. Therefore, the method of the present invention lays the foundation for multi-parameter fusion at the data acquisition stage to avoid the limitations of single-parameter prediction.

[0023] During the calibration process of the detection point position parameters, the present invention fully considers the uniformity of spatial distribution and the geological characteristics of exploration targets. Traditional methods may use regular grids to arrange detection points. However, the actual distribution of geological bodies is often highly heterogeneous. Therefore, simple grid arrangement may lead to problems of mismatched sampling density. For example, in areas with developed fractures or areas where reservoir changes are drastic, if the distribution of detection points is too sparse, it may lead to the omission of key geological features, thereby affecting the prediction accuracy. To solve this problem, the present invention adopts a method for arranging detection points optimized based on geological information. In the shallow part, the distribution of detection points should ensure coverage of the main stratigraphic interfaces and lithologic transition zones. In the deep part, the marking of unknown points is optimized based on existing exploration data, so that the unknown points can be evenly distributed within the inferred key strata, rather than randomly or regularly distributed. In addition, to further improve data quality, the present invention considers the dynamic change characteristics of geological parameters during the data acquisition process. For example, in permeability measurement, the flow state of underground fluids will affect the measured values. Therefore, time series analysis needs to be combined during data acquisition to eliminate the influence of short-term disturbances, so that the finally obtained data can truly reflect the static characteristics of underground media.

[0024] Step 2: Perform data fusion on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters in the target geological exploration area; Specifically, during the implementation process of the present invention, the fusion of multi-source data is not just a simple averaging or linear weighting of different physical quantities, but a coupled modeling based on the physical relevance between different geological parameters. For example, permeability, porosity, and water saturation are closely related when describing the fluid migration ability of reservoirs, and Young's modulus, fracture density, and effective pressure determine the mechanical stability of rocks. Therefore, when constructing the variogram, the coupling effect between these parameters needs to be fully considered to ensure that the variogram after data fusion can accurately reflect the true change trend of geological parameters in space. Compared with traditional methods that only consider the variability of a single geological parameter, the method of the present invention adopts a multi-level variable association modeling strategy, that is, during the construction of the variogram, the ratio, interaction term, and non-linear transformation of different parameters are used to describe the mutual influence between multi-source data, so that the final variogram can not only describe the geometric distance relationship between detection points, but also reflect the physical correlation of geological parameters at different spatial scales.

[0025] In addition, during the geological exploration process, the nugget effect is a key factor affecting the accuracy of the variogram. The nugget effect refers to the anomaly of geological parameters caused by the drastic changes in the local geological environment or measurement errors within a small scale range. Such abnormal values may affect the stability of the overall variogram, thereby reducing the accuracy of subsequent interpolation prediction. The method of the present invention adaptively models the nugget effect, enabling the variogram to automatically adjust the correlation strength between different detection points, thus reducing the impact of local anomalies on the overall model. On this basis, the method of the present invention further optimizes the spatial scale parameter of the variogram, enabling the variogram to dynamically adjust its scope of action under different geological environments, thereby ensuring a high prediction accuracy under complex geological conditions.

[0026] More importantly, the spatial variation of geological parameters is usually affected by various environmental factors. For example, depth, temperature, pressure, etc. will change the physical properties of rocks. Therefore, in the process of constructing the variogram, the method of the present invention not only considers the horizontal and vertical distance relationships between detection points, but also introduces a correction factor related to the geological environment, enabling the variogram to adaptively adjust its parameters under different geological environments. The innovation of this method lies in that it does not rely on a fixed mathematical model, but optimizes the parameters according to the specific geological conditions of the detection points, enabling the variogram to more accurately reflect the true variation law of the geological body. This method of constructing the variogram based on data adaptive optimization makes the present invention significantly superior to traditional methods in terms of the prediction accuracy of geological parameters. Especially in complex geological environments such as fracture-developed areas, fault zones or high-temperature and high-pressure reservoirs, this method can effectively improve the accuracy of data interpolation and reduce the errors caused by the heterogeneity of the geological environment.

[0027] Step 3: According to the position parameters of each detection point, construct a multi-source composite covariance matrix for all detection points in combination with the variogram; each element in the multi-source composite covariance matrix is the multi-source composite covariance between different detection points; In specific implementation, the present invention first calculates the variation degree of geological parameters between different detection points by using the variogram, and then derives the spatial covariance between the detection points. The role of the variogram is to quantify the spatial autocorrelation of geological parameters, while the covariance matrix further establishes the numerical correlation between the detection points, enabling the mathematical description of the spatial data dependence relationship. Different from traditional methods, the covariance matrix of the present invention is not a simple Euclidean distance weight distribution, but constructs matrix elements in a multi-physical parameter coupling manner by combining the influences of multiple geological parameters such as in-situ stress, porosity, permeability, and thermal conductivity, so as to ensure the physical rationality of the covariance matrix. For example, in high-stress regions, due to the possible changes in the mechanical properties of rocks, the traditional covariance matrix may not accurately describe the true spatial correlation of geological parameters. However, the method of the present invention corrects the correlation calculation between different detection points by introducing parameters such as rock elastic modulus, in-situ stress adjustment factor, and pore fluid pressure, enabling the covariance matrix to adapt to the changes in complex geological environments.

[0028] In the process of constructing the covariance matrix by the method of the present invention, the anisotropic characteristics of geological bodies are also fully considered. Geological bodies are not homogeneous media, but anisotropic structures with directional differences. For example, sedimentary rock reservoirs often have different permeability distributions in the vertical and horizontal directions, and the fracture directions in fracture zones also affect the spatial correlation of geological parameters. Therefore, in the method of the present invention, the calculation of the covariance matrix is not only based on horizontal and vertical distances, but also combines the directional changes of geological parameters. For example, by introducing a directional adjustment factor, the covariance calculation can dynamically adapt to the changes in geological characteristics in different directions. The advantage of this method is that it avoids the errors of traditional covariance matrices under the isotropic assumption, enabling high prediction accuracy to be maintained under complex geological conditions.

[0029] In addition, during the calculation of the covariance matrix, the uncertainty of geological data is also an important issue. During actual geological exploration, the data of detection points may be affected by factors such as measurement errors, environmental noise, and data scarcity, resulting in a decrease in the stability of the covariance matrix. To overcome this problem, the method of the present invention combines geostatistical methods during the calculation of the covariance matrix, introducing a measurement error adjustment factor and a data weight correction strategy, enabling the covariance matrix to still maintain strong robustness under high data uncertainty. For example, during data acquisition, the data of some detection points may be affected by external interference, and their covariance calculation weights will be automatically reduced, thereby reducing the impact of error propagation on the overall covariance matrix. This adaptive data correction mechanism enables the covariance matrix to still provide a reliable prediction basis in complex geological environments.

[0030] It should be noted that the multi-source composite covariance matrix of the present invention is not only applicable to the calculation of the correlation between known detection points, but also provides key data support for the prediction of subsequent unknown points. After constructing the covariance matrix, this matrix is used to deduce the geological parameters of unknown points, so that the estimated values of unknown points can conform to the data distribution of known detection points to the greatest extent. Traditional interpolation methods usually rely on simple weight assignment, while the method of the present invention uses the covariance matrix to construct an optimal estimation model, so that the predicted values of unknown points are not only affected by the spatial distance, but also restricted by the coupling effect between multi-source data, thereby improving the prediction accuracy. For example, in the prediction of permeability, traditional methods may only perform interpolation based on spatial distance, while the method of the present invention incorporates parameters such as rock porosity, fracture density, and fluid pressure into the calculation through the covariance matrix, making the prediction results more consistent with the actual geological situation.

[0031] Step 4: Predict the multi-source geological parameters of unknown points according to the multi-source composite covariance matrix.

[0032] During the prediction process, the method of the present invention first uses the multi-source composite covariance matrix to establish the correlation weight between the detection points and the unknown points. This weight is not only based on the calculation of geometric spatial distance, but through the optimization of the variogram, so that the distribution of the weight can reflect the true spatial variability of geological parameters. For example, in the fracture development area, since the physical properties of the rock are closely related to the pore structure, traditional interpolation methods may only consider the horizontal distance for weight calculation, while the method of the present invention further combines multiple parameters such as elastic modulus, shear modulus, porosity, and permeability through the composite covariance matrix, making the weight assignment more in line with the true change trend of the geological structure. This weight calculation method based on the coupling of geological parameters enables the prediction results to more accurately describe the geological attributes of unknown points and avoids the problem of error accumulation caused by single-parameter interpolation in traditional methods.

[0033] When calculating the geological parameters of unknown points, the method of the present invention not only utilizes the data of known detection points, but also combines anisotropic adjustment, in-situ stress correction, and well logging data constraints, so that the final predicted value not only considers spatial correlation, but is also constrained by multiple physical fields. For example, in high-stress regions, due to the deformation behavior of rocks, the geological parameters may change non-linearly. Therefore, traditional methods relying solely on distance weighting may not accurately predict the permeability or porosity of the formation. The method of the present invention optimizes the covariance matrix and introduces a stress adjustment factor, enabling the prediction model to dynamically adapt to parameter changes in different geological environments, thereby maintaining a high prediction accuracy under complex geological conditions. In addition, during the prediction process, the method of the present invention further utilizes well logging data, such as natural gamma, acoustic travel time, resistivity and other parameters, as additional correction factors, so that the prediction results can be more reliably corrected in areas with lithological mutations. This prediction method combining well logging data enables the method of the present invention to still provide high-precision geological parameter estimation in the case of limited drilling data, effectively improving the stability and reliability of the prediction.

[0034] In the specific calculation process, the method of the present invention adopts the Kriging interpolation optimization algorithm, making the predicted value not only conform to the optimal estimation criterion, but also ensuring the minimization of the prediction error. Although the traditional Kriging interpolation method can provide a good spatial interpolation effect, it is often difficult to capture the detailed changes in complex geological environments in the case of highly heterogeneous geological parameters. The method of the present invention optimizes the calculation process of the covariance matrix and combines the dynamic adaptive adjustment of geological parameters, making the weight distribution of Kriging interpolation more reasonable and avoiding the problem of decreased prediction accuracy in data-sparse areas of traditional methods. For example, in the fault zone area, the mechanical properties of rocks may change suddenly, resulting in large errors in the interpolation method of a single physical quantity. The method of the present invention introduces parameters such as fracture density, elastic modulus, and stress state into the Kriging interpolation process through multi-source data fusion, enabling the prediction results to more accurately describe the geological characteristics of the fault zone.

[0035] In addition, to ensure the stability of the prediction results, the method of the present invention also introduces a data constraint optimization mechanism, that is, when calculating the geological parameters of unknown points, in addition to the weight distribution provided by the covariance matrix, the physical boundary conditions of the geological parameters are also considered. For example, in the prediction of reservoir permeability, if the predicted value in a certain area exceeds the reasonable physical range (such as far exceeding the maximum permeability of known detection points), the method of the present invention will automatically adjust the weight distribution and use other relevant parameters, such as porosity, total organic carbon content, etc., for correction, so as to ensure the rationality of the prediction results. This optimization mechanism based on physical constraints enables the method of the present invention to not only provide high-precision prediction results when the data is sufficient, but also maintain good prediction stability when the data is relatively scarce or the geological conditions are relatively complex, avoiding abnormal prediction situations caused by local data insufficiency.

[0036] Example 2: The multi-source geological parameters include: rock density , in g / cm³; solid particle density , in g / cm³; porosity ; horizontal permeability , in mD; water saturation ; total organic carbon content ; thermal conductivity , in W / m·K; rock Young's modulus , in GPa; fracture density , in fractures / m; effective pressure , in MPa; where represents the subscript index of the detection point or unknown point, taking positive integer values, representing the th detection point or unknown point, can be replaced by any subscript index letter, including: , or .

[0037] Specifically, the rock density , as a basic physical quantity describing the overall mass characteristics of the rock, is the basis for calculating parameters such as rock porosity, permeability, and thermal conductivity. The change in rock density often reflects the change in formation composition. For example, in carbonate reservoirs, the change in density may be affected by dolomitization or dissolution, while in sandstone reservoirs, it may be affected by cementation degree and mineral composition changes. Therefore, in geological exploration, rock density is not only used to judge formation lithology, but also can be an important parameter for well logging data interpretation and reservoir modeling. In contrast, the solid particle density characterizes the density of the rock matrix without including the influence of pore space. Usually, the ratio of rock density to solid particle density can be used to calculate porosity , which describes the proportion of the space available for fluid storage inside the rock.

[0038] Porosity is one of the key parameters determining reservoir quality, which directly affects the fluid storage capacity and seepage capacity. During the process of multi-source data fusion, porosity is not only used to calculate the reserves of the reservoir, but also one of the core variables in the calculation of the variogram. Since the spatial variation of porosity is usually closely related to the sedimentary environment, diagenesis, and fracture development degree, when constructing the multi-source composite covariance matrix, the method of the present invention fully considers the influence of porosity on the spatial variability of geological parameters, and through the collaborative analysis with parameters such as permeability and water saturation, optimizes the calculation accuracy of data fusion. Horizontal permeability further describes the fluid flow capacity of the reservoir. Its magnitude is not only related to porosity, but also affected by pore structure, pore-throat connectivity, and rock microstructure. The spatial distribution of permeability often has strong anisotropic characteristics. Therefore, during the calculation process of the variogram of the present invention, permeability acts together with parameters such as stress and fracture density to more realistically depict the fluid migration law of the reservoir.

[0039] Water saturation As an important indicator of reservoir fluid distribution, it characterizes the proportion of water in the fluid in the pore space of the reservoir. The change of water saturation is usually affected by reservoir pressure, lithology, and geological evolution process. During geological exploration, it not only affects the judgment of fluid characteristics of oil and gas reservoirs, but also is one of the key parameters in well logging data interpretation. In the process of data fusion of the method of the present invention, water saturation is analyzed collaboratively with parameters such as permeability, porosity, and effective pressure, so as to improve the sensitivity of the prediction model to reservoir fluid distribution.

[0040] Total organic carbon content is mainly used to characterize the hydrocarbon source rock abundance of shale gas reservoirs or reservoirs rich in organic matter. As an important indicator for shale gas exploration, the TOC content determines the hydrocarbon generation potential of the rock and affects the adsorbed gas content of the reservoir. In the process of multi-source data fusion of the method of the present invention, TOC is used as an important variable to describe the physical and chemical properties of the formation, and combined with thermal conductivity for comprehensive analysis. Thermal conductivity is an important factor affecting the distribution of formation temperature field. Especially in geothermal resource exploration or deep oil and gas exploration, the change of thermal conductivity can affect the formation temperature gradient, and further affect the fluid physical properties and pressure distribution of the reservoir. During the construction process of the variogram of the present invention, thermal conductivity acts together with parameters such as rock density and effective pressure to optimize the prediction accuracy of the reservoir temperature field.

[0041] Rock Young's modulus As an important parameter of rock mechanical properties, it determines the elastic stress response ability of rocks. The magnitude of Young's modulus is closely related to the composition, pore structure, and fracture development degree of rocks. During the calculation of the covariance matrix in the present invention, Young's modulus and fracture density act together to optimize the description of reservoir mechanical stability. Fracture density directly affects the seepage ability and stress distribution of the reservoir. Its role in the multi-source data fusion process is not limited to describing the fluid migration ability of the reservoir, but also used to optimize the anisotropy calculation of the covariance matrix, enabling the prediction model to more accurately describe the structural changes of the reservoir.

[0042] Effective pressure As the core variable affecting reservoir physical and chemical processes, it determines the pore compressibility, permeability change, and fracture closure state of rocks. In the multi-source data fusion process, the method of the present invention utilizes the coupling relationship between effective pressure and permeability, porosity, and stress state to optimize the accuracy of the reservoir parameter prediction model. Compared with the traditional single-parameter interpolation method, the method of the present invention enables a significant improvement in the prediction accuracy of geological parameters through the collaborative calculation of multiple physical parameters and can maintain strong adaptability in complex geological environments.

[0043] Example 3: The formula of the variogram is: ; where and both represent the subscript indices of the detection points, taking positive integer values, representing the th detection point or the th detection point respectively; represents the variation value between the th detection point and the th detection point; represents the nugget effect value of the th detection point; represents the nugget effect value of the th detection point; represents the vertical distance difference between the th detection point and the th detection point; represents the horizontal distance difference between the th detection point and the th detection point.

[0044] Specifically, in the formula is used to describe the variation value between the th detection point and the th detection point, thus providing a basis for the subsequent calculation of the multi-source composite covariance matrix. Here, and respectively represent the nugget effect values of the th detection point and the th detection point. These nugget effect values play a crucial role in geological exploration because geological parameters may be affected by local anomalies, measurement accuracy, or high-frequency disturbances caused by geological structures at very small scales, thus manifesting as an additional constant term that cannot be ignored in the variogram. By explicitly introducing and into the variogram, these local effects can be incorporated into the overall spatial variability analysis. The most distinctive part of the formula is the coupling of the exponential function term with the coefficient outside it. This coupling combines the multiple effects of geological parameters to reflect the spatial differences between different detection points in the vertical and horizontal directions, as well as the multiple coupling effects of lithology, mechanical, and fluid properties. Specifically, reflects the comprehensive spatial scale difference between the detection points in the vertical and horizontal directions. If the two are far apart in the vertical or horizontal direction, the value of this term will increase, thus having a significant impact on the negative exponential part of the exponential function; once the distance between the two points is close enough, the impact of this term on the overall variability value will be correspondingly weakened. At the same time, this part considers the ratio between the rock density and the solid particle density, aiming to capture the impact of the change in rock properties between different detection points on the variability; if the rock density of detection point is significantly different from the solid particle density of detection point , this means that there may be significant changes in the lithology or pore structure of the strata where the two are located, thus leading to more obvious differences in geological properties in space. In addition, this part further reflects the coupling relationship between mechanical and seepage properties. Specifically, is the effective pressure, is the porosity, represents the horizontal permeability, is the water saturation. When the effective pressure at point and the porosity at point are relatively large, while the permeability at point and the water saturation at point are relatively small, then the geological property difference between the two points will be amplified, thus introducing a larger value in the exponential term, resulting in a more significant decrease in the exponential function and an increase in the overall variability value. In other words, this structure clearly unifies mechanical parameters and fluid distribution parameters into the same variogram, providing a mathematical approach for the high integration of multi-source geological data. It should be noted that there is also a factor outside the exponential term, which means that after the exponential function reaches a certain value, the detection point The total organic carbon content and the detection point The thermal conductivity is taken into account as a multiplicative factor and is related to the detection point The Young's modulus of and the detection point The fracture density of are processed to obtain relative ratios. From a geophysical perspective, TOC characterizes the organic matter abundance and is crucial for the evaluation of shale gas or other unconventional oil and gas reservoirs; represents the rock thermal conductivity, reflecting the heat conduction ability of the formation; is the Young's modulus of the rock, determining the deformation behavior of the rock under elastic stress; is the fracture density, representing the degree of fracture development near the detection point. The more developed the fracture network, the more obvious the continuity and anisotropy of the geological parameters. The coupling of these parameters is finally reflected in a square - root form, that is , this operation numerically plays a regulatory role in amplifying or reducing the influence of these parameters on the variogram, so that when the TOC and thermal conductivity are relatively large, while the Young's modulus and fracture density are relatively small, more obvious differences may be presented between two points. In this way, the entire variogram is no longer simply determined by the distance factor or a single petrophysical parameter, but is a combined effect of multiple factors such as spatial position, mechanical properties, fluid properties, and geochemical properties, which is exactly the advantage of the present invention. Through this integrated variogram, the present invention can more accurately depict the complexity within the formation and show higher accuracy and reliability in the subsequent derivation of the covariance matrix and prediction of geological parameters at unknown points. This variogram with a nugget effect term and multi - source parameter coupling characteristics provides a new interpolation idea for the geological exploration field, can effectively overcome the limitations faced by traditional geostatistical methods in dealing with multi - physical parameter coupling, and still maintain strong adaptability and credibility in complex geological environments.

[0045] Example 4: For the th detection point, the calculation formula for the nugget effect value is as follows: ; where, is the depth of the th detection point; is the initial nugget effect value. For loose sedimentary layers the value range is ; for shale layers the value range is ; for carbonate rock layers the value range is ; for igneous rock layers the value range is .

[0046] Specifically, the present invention assumes that the greater the depth where the detection point is located, the more stable the geological environment is. Or rather, for areas in the shallow part that are easily affected by factors such as surface activities, local erosion, or weathering, the nugget effect should be relatively large. Therefore, by exponentially decaying the depth in the calculation of the nugget effect value, this gradually weakening random perturbation with depth can be reflected. At the same time, the present invention fully considers the characteristics of different lithologies in the initial nugget effect value. Due to differences in diagenetic conditions and rock structures, different stratigraphic types often have different local microscopic characteristics. For example, loose sedimentary layers usually have relatively weak particle binding forces and are easily affected by multiple factors such as the surface environment and sediment dynamics, resulting in large local measurement fluctuations during the data acquisition process. Therefore, a relatively smaller initial nugget effect value range needs to be selected to reflect the randomness brought by its high-frequency changes. For shale layers, since they may contain more organic matter and have a relatively dense structure, combined with the stress effect during the sedimentary evolution process, local fracture development zones or brittle regions are also likely to form, thus showing a certain degree of discreteness in the measurement data. Therefore, the selection of the initial nugget effect value is increased and combined with depth for corresponding exponential decay. For carbonate rock layers, due to their widespread dissolution and diverse pore structures, large differences in reservoir properties may occur in a small range. Therefore, the initial nugget effect value range is further increased, enabling the measurement uncertainty brought by greater microscopic fluctuations to be reflected in the final calculation. In igneous rock layers, due to the usually higher rock density and often experiencing relatively strong magmatic or metamorphic processes, local anomalies are more likely to be related to processes such as joint fractures or mineral replacement. The corresponding nugget effect needs to be taken in a higher range to reasonably depict the large measurement fluctuations that may occur in such extreme environments.

[0047] Example 5: The multi-source composite covariance between the th detection point and the th detection point is calculated using the following formula: ; where is the multi-source composite covariance between the th detection point and the th detection point; is the difference in the nugget effect value between the th detection point and the th detection point; represents the elastic modulus at the th detection point, with the unit of GPa; represents the elastic modulus at the th detection point, with the unit of GPa; represents the The maximum horizontal principal stress at the th test point, with the unit of MPa; th test point, with the unit of MPa; th test point, with the unit of MPa; th test point, with the unit of MPa; th test point, with the unit of m; th test point, with the unit of m² / s; th test point and the th test point, representing the in-situ stress adjustment coefficient;

[0048] Specifically, is the nugget effect value difference between the th test point and the th test point, used to correct the random component caused by local high-frequency disturbances or small-scale anomalies. In geostatistics, the nugget effect usually refers to the irregular fluctuations or local anomalies of measured values at very short distances. In the present invention, it is regarded as a basic correction quantity affecting the covariance: when both points are in a similar geological environment and the local disturbance intensity is approximately the same, its impact on the final result will be relatively small; however, if the formation types or local environments of the test points are very different, this difference will have a more significant correction on the covariance value. On this basis, represents the variation value between the th

[0049] Next, is the in-situ stress adjustment coefficient, which is used to describe the influence of stress field differences on the correlation of geological parameters. If the stress environment in a certain area is very different from that of the surrounding areas, then even if the spatial distance is relatively close, obvious deviations may occur in the mechanical properties of the reservoir and the fluid seepage performance; It is in this case that the covariance is weighted and corrected, so that the complexity of the stress field can be better incorporated into the model. Next, the in the numerator part reflects the coupling effect of the elastic modulus with the fracture density and the maximum horizontal principal stress. Here, is the fracture density at the th detection point, which largely determines the brittle characteristics of the rock and the fluid conduction ability; and are the elastic moduli of the th and the th detection points respectively, which describe the ability of the rock to resist deformation within the elastic range; is the maximum horizontal principal stress at the th detection point. If the rock at a certain detection point has a high elastic modulus, a large fracture density, and a relatively high maximum horizontal principal stress, then this point will show a stronger response to the changes in the geomechanical field. In the covariance calculation, the product of these factors will increase the correlation between the two points; on the contrary, if the rock is denser and has fewer fractures, and the stress level is lower, then the enhancement effect on the overall correlation will be weakened.

[0050] Comparing the above numerator part with the in the denominator, it can be seen that these denominator terms also represent several core variables that affect geomechanical coupling. is the rock shear modulus of the th detection point, which determines the deformation ability of the rock in the shear state; is the shear strength at the th detection point, which is used to measure the critical value that the rock can withstand before shear failure; represents the vertical stress at the th detection point, which covers the compaction effect of the overlying pressure of the rock formation in the vertical direction. When the parameter values in the denominator are larger, it means that the rock in this area can better resist stress actions or external force changes, making it difficult for the mechanical coupling effect from another detection point to be transmitted, thus showing a weakening effect on the covariance value. At the end of the formula, the correction term composed of exponential functions is used to describe the influence of depth difference, fluid pressure, and heat diffusion on the correlation between two points. Among them, is the th and the The difference in vertical depth between the detection points. If a detection point is buried deep underground while another is relatively shallow, there are often significant differences in thermal conditions and fluid pressure distribution between the two; represents the pore fluid pressure at the th detection point. When the pressure is high, the hydrodynamic processes within the rock pores and fractures are more active, having a significant impact on the attenuation or enhancement of geological properties; is the thickness of the oil and gas bearing layer at the th detection point. When this value is high, the fluid in the reservoir may be more abundantly distributed and heat convection is more obvious, thereby amplifying the regulatory effect of fluid pressure on geological parameters; is the thermal diffusivity at the th detection point. It characterizes the rock's ability to transfer heat. If the thermal diffusivity is large, the temperature gradient is more easily balanced within the region, which may weaken the temperature and fluid property differences caused by depth differences. Conversely, if the thermal diffusivity is small, the temperature change and fluid pressure difference generated by the same depth difference will be more obvious. The exponential form means that after these thermodynamic and hydrodynamic factors are combined with the depth difference, they will produce an exponential decay effect on the covariance: when

[0051] Example 6: For the thermal diffusivity at the th detection point, it is calculated using the following formula: ; where represents the specific heat capacity at the th detection point, with the unit of

[0052] Specifically, in this formula, is the thermal conductivity at the th detection point, usually with the unit of . A higher value means that the rock at that location has a stronger ability to conduct heat and is more likely to transfer temperature differences in the horizontal and vertical directions; represents the density of the rock at that location (the density without pore space or the bulk rock density, depending on the specific description in the foregoing of the present invention), with the unit of or , which measures the mass contained per unit volume of the rock; is the specific heat capacity, with the unit of , which represents the amount of heat required to raise the temperature of a unit mass of rock by 1 Kelvin without a phase change. The above three parameters together determine the comprehensive ability of the rock medium to absorb and diffuse heat during temperature changes: when the thermal conductivity is relatively large, heat energy can be quickly transmitted in the rock, making the temperature gradient tend to equilibrium faster; however, if the density of the rock and specific heat capacity are relatively high, it means that more heat needs to be absorbed or released per unit volume to achieve the same temperature change. In this case, the value of the thermal diffusivity will decrease as the denominator increases, resulting in relatively slow heat diffusion. Therefore, intuitively represents the ratio of "heat conduction ability" to "heat storage ability": if the thermal conductivity dominates, increases, and the rock medium will quickly balance the temperature field; if the density or specific heat capacity of the rock is relatively large, decreases, and heat is difficult to conduct to a farther area in a short time. Through the accurate estimation of by the method of the present invention, not only can the thermophysical characteristics at the detection point be incorporated into the correction of the covariance matrix or variogram during subsequent multi-source data fusion, but also the coupling relationship between the depth difference, pore fluid pressure, and thermodynamic parameters can be reflected through exponential terms, etc., so as to more realistically simulate the geological property changes of the deep reservoir under the interaction of temperature, pressure, and stress, making the overall geological exploration and data fusion model also maintain high adaptability and accuracy under high-temperature, high-pressure, or unconventional formation conditions.

[0053] Example 7: The in-situ stress adjustment coefficient between the th detection point and the th detection point is calculated using the following formula: .

[0054] Specifically, this embodiment gives a formula containing multiple mechanical parameters and geometric factors to quantify the difference in the in-situ stress environment between the th detection point and the th detection point. First, and respectively represent the maximum horizontal principal stress at the th detection point and the minimum horizontal principal stress at the th detection point. When the difference between the two is relatively large, it indicates that there is an obvious difference in the stress environment between the two points, which will thus affect the rock fracture mode and fracture propagation law; if the difference between the two is relatively small, it means that the in-situ stress levels of the two points are relatively close. Correspondingly, the adjustment amplitude of the in-situ stress adjustment coefficient to the covariance will be weakened. Next, and respectively represent the vertical stress at the and the th detection points, which together with the horizontal stress determine the complete stress state of the underground medium. When the denominator value is large, it indicates that the overlying pressure borne relatively is high, and the sensitivity of the rock mass to stress differences will be reduced, partially offsetting the actual influence of the stress environment differences between the two points.

[0055] Meanwhile, this part reflects the coupling between the shear modulus and the elastic modulus of the rock: are the rock shear moduli at the two points respectively, characterizing the deformation ability of the rock under shear action, while is the elastic modulus at the th detection point, used to characterize the resistance degree of the rock to elastic deformation. If the rock shear modulus or elastic modulus at a certain detection point changes drastically, it will increase or weaken the amplification effect of the in-situ stress adjustment coefficient on the covariance, and then reflect different feedbacks of the rock mass to stress actions. Finally, the exponential decay term is used to characterize the weakening effect of the difference in vertical depth between the two detection points on the in-situ stress coupling relationship: when is large, it means that they may be located in different tectonic horizons underground. Even if the thickness of the oil and gas-bearing layer is large, it is difficult to cross the physical environment differences brought about by the depth difference, resulting in obvious attenuation of this exponential term and causing the in-situ stress adjustment coefficient to tend to decrease; if the burial depths of the two points are close and the superposition degree of the oil and gas-bearing layer thickness is high, then this exponential value is relatively larger, making the amplification effect of the in-situ stress adjustment on the correlation between the two points more significant.

[0056] Example 8: According to the following formula, predict the multi-source geological parameters of an unknown point based on the multi-source composite covariance between detection points: ; where is the predicted multi-source geological parameter of the unknown point; is the standard value of the multi-source geological parameter of the detection point, which is a set value, and the standard values of different multi-source geological parameters are different; is the multi-source geological parameter of the detection point; is the weight of the th detection point; is the number of detection points; is the brittleness index at the th detection point; is the natural gamma value at the th detection point, with the unit of API; represents the shale index at the th detection point; is the spontaneous potential at the th detection point, with the unit of mV; is the true resistivity at the th detection point, with the unit of Ω·m; is the relative density of natural gas at the th detection point.

[0057] Specifically, the core structure of the formula is a weighted summation expression, where represents the geological parameter of the predicted unknown point, and the known detection point parameters after a series of corrections and weight adjustments, finally determine the estimated value of the unknown point. First, this summation term means that the predicted value is the weighted average of all detection point data, where represents the weight of the th detection point, which is calculated from the multi-source composite covariance matrix. The weight reflects the contribution of the detection point data in the prediction process, which is usually jointly determined by the spatial distance between the detection point and the unknown point, the variability of geological parameters, and the similarity of logging data. If the geological environment of a certain detection point is more similar to the unknown point, then its weight value will be relatively large, and the prediction result is more likely to be affected by the data of this point; on the contrary, if the geological conditions at a certain detection point are significantly different from the unknown point, then its weight is smaller and its contribution to the final predicted value is lower. The next correction term is used to eliminate the systematic deviation of logging data or rock physical parameters under different regions or different measurement standards. represents the standard value of the th detection point. Different types of geological parameters have different standard values. For example, for porosity, may be the regional average porosity, and for permeability, may be the average permeability under the same lithological background. By calculating the deviation between the actual data of the detection point and its standard value , the present invention can effectively correct the systematic error in the measurement data, so that the prediction model maintains stability and consistency under different regions or different measurement methods.

[0058] Based on the corrected geological parameters, the formula further introduces the coupling correction term of logging parameters, where is the brittleness index, which measures the likelihood of brittle fracture in reservoir rocks under external forces. The brittleness index is usually determined by mineral composition, fracture development degree, and stress environment. A higher brittleness index often means that the rock is more prone to fracture, which is crucial for the development of unconventional resources such as shale gas and coalbed methane. If the brittleness index at a certain detection point is high, then this point will have a more significant impact on the geological parameters of unknown points during the prediction process, because brittle fracture is often accompanied by reservoir transformation, resulting in large changes in key parameters such as permeability and porosity. is the natural gamma value (in API units), which is mainly used to reflect the content of radioactive elements in the formation and is usually positively correlated with the shale content. A higher natural gamma value usually means a higher shale content in the formation, while a low natural gamma value indicates the possibility of a sandstone or carbonate reservoir. Therefore, when predicting the geological parameters of unknown points, if the natural gamma value at a certain detection point is high, it may indicate a higher shale content in the reservoir at this point, thus affecting its key geological parameters such as permeability and water saturation. represents the standardized unit of the logging gamma value, which is used to normalize data between different logging instruments and regions, enabling direct comparison of logging data at different detection points. And is the pore fluid density, which represents the density of the fluid in the reservoir at this detection point. The greater the pore fluid density, usually means that the fluid components may contain a higher proportion of heavy hydrocarbon components, or are affected by higher pressure conditions. In the formula, the role of this term is to balance the influence of the brittleness index and the natural gamma value on the prediction, so that the reservoir fluid characteristics can appropriately participate in the prediction model and avoid relying solely on rock mineralogical parameters for judgment.

[0059] In addition, the formula also includes an additional correction factor , where is the shale index, which reflects the influence of the shale content in the formation on lithology and permeability and is usually used to distinguish between reservoirs and caps or low-permeability reservoirs. is the spontaneous potential, in mV units. It is mainly used to identify the salinity and permeability of formation water. A higher spontaneous potential is usually related to better reservoir permeability. If both the shale index and the spontaneous potential at the detection point are high, it may indicate a higher shale content in the reservoir, but at the same time, there may be better fluid migration ability. In the correction factor, is the true resistivity (in units), which is used to characterize the conductivity of the formation and is usually related to the water saturation of the reservoir. High resistivity generally indicates hydrocarbon enrichment, while low resistivity may mean that the formation is mainly filled with water. Finally, is the relative density of natural gas, which is used to measure whether the density of natural gas at the detection point is higher than the standard air density. If this value is large, it indicates that there may be richer natural gas resources in this area. The physical meaning of this correction factor is that by normalizing the shale index, spontaneous potential, true resistivity, and natural gas density, the physical and chemical properties of the reservoir can be more comprehensively described, enabling the prediction model to not only rely on spatial statistical correlations but also combine the specific characteristics of well logging data to optimize the accuracy of the prediction results. For example, if the shale index at a certain detection point is high, but the resistivity and natural gas density are low, it indicates that the reservoir may be more water-filled rather than rich in oil and gas, and the influence of this detection point will be weakened when predicting unknown points. Conversely, if the resistivity and natural gas density at a certain detection point are both high, it indicates that this point may be in an oil and gas enrichment area, so its contribution to the prediction of unknown points will be amplified.

[0060] Example 9: Through the following formula, the weight of the th detection point is calculated using the following formula, where the value of is 1, 2,..., N: where is the Lagrange multiplier to ensure that the sum of the weights is 1; represents the depth difference between the th detection point and the unknown point.

[0061] Specifically, in this formula, the calculation of the weight depends on a matrix equation, where the inverse matrix of the covariance matrix is used to solve the optimal weight allocation problem. The column vector on the left side of the matrix contains the weights of all detection points and an additional Lagrange multiplier , which is used to ensure that the sum of all weights is 1, thereby maintaining the unbiasedness of the interpolation process. On the right side of the matrix equation, the column vector consists of the depth differences between the detection points and the unknown point and the constant 1, to ensure that the interpolation result not only meets the requirements of statistical optimal estimation but also fully considers the geological feature changes of the reservoir in the depth direction. The covariance matrix is a square matrix composed of the multi-source composite covariance between each detection point, where represents the th detection point and the The covariance between detection points, which is calculated by the aforementioned composite covariance formula and includes various factors such as rock mechanics parameters, hydrodynamic parameters, stress field influence, and formation physical properties. Therefore, compared with the traditional method of calculating covariance only based on Euclidean distance or variogram, the method of the present invention further considers the coupling effect of multi-source geological data during weight calculation, enabling the prediction process to more accurately reflect the spatial variation trend within the reservoir.

[0062] The last column and the last row of the matrix are both composed of 1s, aiming to introduce Lagrange constraints so that the solution of the weights not only depends on the distribution of the covariance matrix but also ensures that the sum of all weights is 1, thereby ensuring that there are no systematic biases in the interpolation results. The advantage of this method is that even in the case of uneven data distribution or large spatial variability, it can still maintain the stability of the prediction and avoid abnormal interpolation phenomena that may occur in traditional methods. During the matrix solution process, the weights are obtained through an inverse operation, and its mathematical meaning is to allocate the influence degree of each detection point on the unknown point under the premise of minimizing the prediction error. If the covariance between a certain detection point and the unknown point is large, it indicates a strong correlation between their geological parameters, and the weight of this point will increase accordingly; conversely, if the covariance between a certain detection point and the unknown point is small, its weight will decrease, meaning that the contribution of this point to the prediction is small. In addition, the depth difference is used as a correction factor to further adjust the weight distribution, enabling the interpolation to consider not only the horizontal distribution but also the geological changes in the depth direction, which is particularly important in the case of obvious stratification or significant in-situ stress gradient in the formation.

[0063] Example 10, refer to Figure 2 : A multi-source data fusion analysis system for geological exploration, the system includes: a data acquisition part for setting a plurality of detection points in the shallow part of the target geological exploration area, marking a plurality of unknown points in the depth part of the target geological exploration area, calibrating the position parameters of each detection point; collecting multi-source geological data of each detection point; a variogram construction part for performing data fusion on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters in the target geological exploration area; a composite covariance construction part for combining the variogram according to the position parameters of each detection point to construct a multi-source composite covariance matrix of all detection points; each element in the multi-source composite covariance matrix is the multi-source composite covariance between different detection points; an unknown point data prediction part for predicting the multi-source geological parameters of the unknown points according to the multi-source composite covariance matrix.

[0064] Although the specific embodiments of the present invention have been described above, those skilled in the art should understand that these specific embodiments are merely illustrative. Without departing from the principles and essence of the present invention, those skilled in the art can make various omissions, substitutions, and changes to the details of the above methods and systems. For example, combining the above method steps so as to perform substantially the same function in a substantially the same way to achieve substantially the same result falls within the scope of the present invention. Therefore, the scope of the present invention is only defined by the appended claims.

Claims

1. A multi-source data fusion analysis method for geological exploration, characterized in that: The method comprises: Step 1: Set multiple detection points in the shallow part of the target geological exploration area, mark multiple unknown points in the deep part of the target geological exploration area, calibrate the position parameters of each detection point; collect multi-source geological data of each detection point; Step 2: Data fusion is performed on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters in the target geological exploration area; Step 3: Based on the location parameters of each detection point, the multi-source composite covariance matrix of all detection points is constructed in combination with the variation function; each element in the multi-source composite covariance matrix is ​​the multi-source composite covariance between different detection points; Step 4: Predict the multi-source geological parameters of unknown points based on the multi-source composite covariance matrix.

2. The multi-source data fusion analysis method for geological exploration according to claim 1, characterized in that: The multi-source geological parameters include: rock density , in g / cm³; solid particle density , unit is g / cm³; porosity ; Horizontal permeability , unit is mD; water saturation ;Total organic carbon content Thermal conductivity , unit is W / m·K; Young's modulus of rock , in GPa; crack density , unit is bar / m; effective pressure , the unit is MPa; among which, Indicates the subscript index of the detection point or unknown point, which is a positive integer. Test points or unknown points, Any subscript index letter may be used instead, including: , or .

3. The multi-source data fusion analysis method for geological exploration according to claim 2, characterized in that: The formula for the variogram is: ； in, and They all represent the subscript index of the detection point, which is a positive integer and represents the The first detection point or Testing points; Indicates The detection point and The variation value between the detection points; Indicates The nugget effect value of each detection point; Indicates The nugget effect value of each detection point; Indicates The detection point and The vertical distance difference between the detection points; Indicates The detection point and The horizontal distance difference between the detection points.

4. The multi-source data fusion analysis method for geological exploration according to claim 3, characterized in that: For The calculation formula of the nugget effect value is as follows: ； in, For the The depth of each detection point; is the initial nugget effect value, for loose sedimentary layer The value range is For shale formations The value range is For carbonate rock layers The value range is For igneous rock layers The value range is .

5. The multi-source data fusion analysis method for geological exploration according to claim 3, characterized in that: For the The detection point and The multi-source composite covariance between detection points is calculated using the following formula: ； in, For the The detection point and The multi-source composite covariance between the detection points; For the The detection point and The difference in the nugget effect between the detection points; Indicates The elastic modulus at each test point, in GPa; Indicates The elastic modulus at each test point, in GPa; Indicates The maximum horizontal principal stress at each test point, in MPa; Indicates The rock shear modulus at each detection point, in GPa; Indicates The shear strength at each test point, in MPa; Indicates The vertical stress at each test point is in MPa; Indicates The pore fluid pressure at each detection point, in MPa; Indicates The thickness of the oil and gas layer at each detection point, in meters; Indicates Thermal diffusion coefficient at each detection point, in m² / s; Indicates The detection point and The in-situ stress adjustment coefficient between the detection points; Represents the absolute value operator.

6. The multi-source data fusion analysis method for geological exploration according to claim 5, characterized in that: For Thermal diffusivity at each detection point , calculated using the following formula: ； in, Indicates The specific heat capacity at each detection point is in .

7. The multi-source data fusion analysis method for geological exploration according to claim 6, characterized in that: No. The detection point and The in-situ stress adjustment factor between the detection points Calculated using the following formula: 。 8. The multi-source data fusion analysis method for geological exploration according to claim 7, characterized in that: The multi-source geological parameters of unknown points are predicted based on the multi-source composite covariance between detection points through the following formula: ； in, Multi-source geological parameters for predicted unknown points; is the standard value of the multi-source geological parameter of the detection point, is the set value, and the standard values ​​of different multi-source geological parameters are different; Multi-source geological parameters for the detection points; For the The weight of each detection point; is the number of detection points; For the The brittleness index at each detection point; For the The natural gamma value at each detection point, in API; Indicates The mud index at each testing point; For the The natural potential at each detection point, in mV; For the The true resistivity at each detection point, in Ω·m; For the The relative density of natural gas at each detection point.

9. The multi-source data fusion analysis method for geological exploration according to claim 8, characterized in that: Through the following formula, The weight of the detection point Calculated using the following formula: The value of is 1,2,…,N: ； in, is the Lagrange multiplier, ensuring that the sum of weights is 1; Indicates The depth difference between the detected point and the unknown point; is the covariance matrix; Indicates The detection point and The covariance between the detection points, The value of is 1, 2,…, N.

10. A multi-source data fusion analysis system for geological exploration, characterized in that: The system comprises: a data acquisition part, which is used to set a plurality of detection points in the shallow part of the target geological exploration area, mark a plurality of unknown points in the deep part of the target geological exploration area, and calibrate the position parameters of each detection point; collect multi-source geological data of each detection point; a variogram construction part, which is used to perform data fusion on the multi-source geological data of each detection point to construct a variogram describing the spatial correlation of geological parameters of the marked geological exploration area; a composite covariance construction part, which is used to construct a multi-source composite covariance matrix of all detection points according to the position parameters of each detection point and in combination with the variogram; each element in the multi-source composite covariance matrix is ​​the multi-source composite covariance between different detection points; and an unknown point data prediction part, which is used to predict the multi-source geological parameters of the unknown point according to the multi-source composite covariance matrix.

Images