Method for automatically determining mineral inversion logging response parameters

Abstract

The invention relates to the technical field of geophysical logging and mineral content analysis, and particularly discloses a method for automatically determining mineral inversion logging response parameters. The method comprises the steps of collection and preprocessing of logging and core data, construction of a logging response parameter matrix, setting of the value range of logging response parameters, construction of an actual calculation matrix, mineral volume content inversion, mineral weight content calculation, parameter optimization and result storage and analysis. According to the method, the parameter matrix required by mineral content inversion can be accurately determined according to specific stratum conditions, so that the mineral content inversion precision is improved, and the workload of interpretation personnel is remarkably reduced. In addition, the method is suitable for well logging interpretation requirements in different regions and under complex geological conditions, and a reliable basis is provided for complex reservoir mineral component inversion and reservoir feature evaluation.

Description

Technical Field

[0001] The invention relates to the technical field of geophysical logging and mineral content analysis, in particular to a method for automatically determining mineral inversion logging response parameters. Background Art

[0002] In geophysical logging, the quantitative interpretation of formation mineral components using logging data has become an important means of evaluating reservoir characteristics. Traditional logging interpretation methods are based on fixed interpretation models and calculate different reservoir parameters separately. These methods lack sufficient adaptability when dealing with complex and changeable formation conditions, and it is difficult to fully integrate a variety of logging information, so that when faced with complex reservoirs with multiple components, the interpretation results are often inaccurate.

[0003] To overcome the above limitations, interpreters began to use the optimal logging interpretation method based on mathematical optimization theory. This method integrates multiple logging curves (such as density, acoustic time difference, neutron porosity, natural gamma and resistivity logging curves) into the same optimization framework, establishes an objective function and adds constraints for iterative solution. Compared with traditional methods, the optimal logging interpretation method not only significantly improves the adaptability and flexibility of complex reservoirs, but also improves the accuracy of interpretation results.

[0004] In the process of optimization inversion, determining the logging response parameter values ​​of each mineral is a critical and challenging link. At present, this process usually relies on the experience of interpreters or refers to the typical logging response parameter values ​​of related minerals in other regions. However, due to the differences in geological genesis and diagenesis in different regions, even for the same mineral, the logging response parameter values ​​in different regions are different. If fixed parameters or empirical parameters are continued to be used, deviations may occur in the inversion results. Summary of the invention

[0005] The object of the present invention is to provide a method for automatically determining mineral inversion logging response parameters to solve the problems raised in the above background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions: Terminology explanation: Damped least squares method: A numerical method commonly used in the optimization process, mainly used to deal with problems with nonlinear objective functions. The damping coefficient is introduced to stabilize the solution process to prevent convergence difficulties caused by the selection of initial values ​​or drastic changes in the objective function gradient.

[0007] XRD data: Mineral information obtained through X-ray diffraction technology, which is usually used in the field of well logging to determine the mineral weight content in the formation.

[0008] Optuna parameter optimization framework: Optuna is an automated hyperparameter optimization framework that aims to automatically find the optimal parameter combination through efficient algorithms (such as Bayesian optimization, random search, etc.). It can be widely used in parameter optimization of various algorithms.

[0009] MAE: Mean Absolute Error, which calculates the average of the absolute values ​​of the corresponding position differences between two sets of values ​​to evaluate the size of the deviation between the two.

[0010] MSE: Mean Squared Error, which calculates the average of the squares of the corresponding position differences between two sets of values. It is used to evaluate the size of the deviation between the two and is more sensitive to larger differences.

[0011] A method for automatically determining mineral inversion logging response parameters, the method comprising: Step 1: Collection and preprocessing of logging and core data: Obtain logging curve data and remove invalid values, and collect XRD data as the basis for inversion effect verification, wherein the logging curve data includes density, acoustic wave time difference, neutron, gamma, element logging, etc.; Step 2: Constructing the logging response parameter matrix: Based on the formation characteristics and the collected logging curves and XRD data, comprehensively establish the logging response parameter matrix; Step 3: Set the value range of the logging response parameter: according to the theoretical value and the empirical data of the region, set the value range of each logging response parameter in the matrix; Step 4: Construct the actual calculation matrix: select specific values ​​according to the value range set in step 3 and fill them into the matrix to construct a complete matrix for calculation; Step 5: Mineral volume content inversion: Mineral volume content inversion is performed based on the damped least squares method to solve the mineral volume content within the treatment depth range; Step 6: Calculation of mineral weight content: According to the mineral density logging response parameters, the mineral volume content is converted into weight content percentage; Step 7: Parameter optimization: Use the Optuna parameter optimization framework to optimize the logging response parameters, with the goal of minimizing the error between the calculated mineral weight content and the XRD measured value. Through multiple rounds of iterations, until the set number of iterations is reached, the optimal logging response parameter matrix is ​​obtained; Step 8: Save and analyze results: Save the optimized logging response parameter matrix and mineral weight content inversion results, draw a comparison chart between the inversion results and the XRD data, and generate an error table to analyze the mineral inversion effect after parameter optimization.

[0012] As a further technical solution of the present invention, step 1, collection and preprocessing of logging and core data includes the following steps: S11. Well logging data collection: Well logging data includes well logging curve data, which includes density, acoustic wave time difference, neutron, gamma, element logging, etc. S12. Preprocessing of logging data: Process invalid values ​​in the logging data and delete rows containing -99999 to avoid calculation errors in subsequent steps; S13. Scope of data processing: Set the depth range for data processing, specify the maximum and minimum processing depth, and perform calculations only within this depth range; S14. XRD data collection: Collect XRD data of a well, extract data within the depth range set in S13, and obtain the weight content data of the main minerals in the formation within this range.

[0013] As a further technical solution of the present invention, step 2 of constructing the logging response parameter matrix includes the following steps: S21. Determine the calculation curve and the minerals to be calculated: According to the logging curve data, XRD data and logging processing requirements (logging data preprocessing, specifying the data processing range) collected in step 1, determine the logging curve used in subsequent calculations, as well as the types and quantities of minerals to be calculated; S22. Construct mineral logging response parameter matrix: According to the logging curves and mineral types determined in S21, a mineral logging response parameter matrix is ​​constructed. Each row of the matrix corresponds to a logging curve, and the last row is a balance equation to ensure that the sum of the mineral and pore volume contents is 1. Each column of the matrix corresponds to a mineral, and the last column is the pore fluid logging response value, which is used to calculate the pore volume. Each element in the matrix represents the theoretical logging response value of a certain mineral in a certain logging curve.

[0014] As a further technical solution of the present invention, step 4 of constructing the actual calculation matrix includes the following steps: S41. Obtain parameter value from the value range: According to the value range set in step 3, select a specific value for each logging response parameter that needs to be adjusted according to the specified method.

[0015] S42, fill in the corresponding position of the matrix: Fill the selected specific parameter values ​​into the corresponding positions of the logging response parameter matrix constructed in step 3 one by one to ensure that the matrix is ​​complete and meets the subsequent calculation requirements.

[0016] As a further technical solution of the present invention, step 5, mineral volume content inversion, comprises the following steps: S51. Use damped least squares method to invert mineral volume content: The damped least squares method solves the inversion result through the matrix equation (Formula 1), and the calculation formula is as follows: Formula 1: ; in Obtain the volume content vector of each mineral for a certain depth, is the logging response parameter matrix, for The transpose of is the damping coefficient, which is used to enhance the stability of the matrix equation. is the identity matrix, is the logging value vector of a certain depth; S52, Volume content constraints: The calculation results are constrained to be reasonable, and values ​​less than 0 are replaced with 0, and values ​​greater than 1 are replaced with 1.

[0017] As a further technical solution of the present invention, step six of calculating the mineral weight content comprises the following steps: S61. Calculate the mineral weight content: Calculate the weight content of each mineral at each depth point using the following formula: Formula 2: ; in For the depth Previous The volume content of minerals, For the Density logging response parameters of minerals, For the depth Previous The weight content of each mineral, is the number of mineral species, For the range from 1 to For every value of is the number of depth points calculated, For the range from 1 to Every value of ; S62. Calculate the total mass of minerals: Calculate the total mass of minerals at each depth point using the formula: Formula 3: ; in For Depth Total mass of the above minerals; S63. Calculate the mineral weight percentage: Calculate the weight percentage of each mineral using the formula: Formula 4: ; in For the depth Previous The weight percentage of a mineral.

[0018] As a further technical solution of the present invention, step seven parameter optimization includes the following steps: S71. Initial Exploration: First, a series of parameter combinations are randomly generated, and the mineral weight content percentage is calculated based on these parameters , while calculating The total mean square error between the measured value and the XRD value is used as the objective function to record the parameter combination of each attempt to provide initial data for subsequent optimization; The total mean square error is calculated as follows: Formula 5: ; in For the Minerals at depth XRD measured weight percentage on is the total number of corresponding XRD depth points, is the total mean square error, first calculate Minerals in The mean square error of the depth points, and then all The mean square errors of various minerals are accumulated to obtain the total mean square error; S72. Data accumulation and classification: After a preset number of attempts, the results are divided into parameter combinations with better effects and parameter combinations with worse effects based on historical data; S73, Intelligent Sampling: Bayesian optimization tends to select new parameters from the parameter range with better effects for trial, while trying to avoid the parameter range with poorer effects. To prevent falling into the local optimal solution, the preset probability threshold is retained during the optimization process to try the parameter range that has not been explored; S74, Iterative Optimization: Through multiple rounds of iterations, the parameter values ​​used to construct the response parameter matrix are continuously updated and optimized, and the parameter search range is gradually narrowed. Finally, after reaching the specified number of iterations, the value that can make the objective function The smallest optimal parameter combination.

[0019] As a further technical solution of the present invention, step eight, result preservation and analysis, comprises the following steps: S81. Save optimization results: Save the final optimized logging response parameter matrix to ensure that the result file format is standardized and easy to reuse, save the mineral weight content inversion results, and provide data support for calculation errors; S82. Draw a comparison chart of the results: Use drawing tools to draw a comparison chart between the mineral weight content inversion results and the XRD measured data, which can intuitively show the matching degree between the inversion effect and the actual measured value; S83, generate error table: Calculate the error indicators between the weight content of each mineral and the XRD measured value, including the mean absolute error (MAE) and the mean square error (MSE), and organize the calculation results into an error table and save them for subsequent analysis and evaluation; The MAE calculation formula is as follows: Formula 6: ; in For the The mean absolute error of the minerals; The MSE calculation formula is as follows: Formula 7: ; in For the The mean square error of the minerals.

[0020] Compared with the prior art, the present invention has the following beneficial effects: (1) Eliminate human errors. The algorithm realizes the automated parameter selection process, avoids the parameter selection errors caused by differences in personal experience, and improves the objectivity and reliability of parameter determination.

[0021] (2) Improve adaptability and versatility. This tuning method can be used in different regions and different lithologies, and can perform targeted parameter adjustments based on specific logging response characteristics.

[0022] (3) Save time and resources. Automated parameter adjustment significantly reduces the time cost and manpower input for determining the logging response parameter matrix, making the inversion work more efficient.

[0023] (4) Improve the inversion accuracy. To minimize the total mean square error ( ) is the objective function. The optimized logging response parameters make the inversion results closer to the XRD measured values, thereby improving the accuracy of the inversion. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention.

[0025] Figure 1 The present invention is a flow chart of a method for automatically determining mineral inversion logging response parameters.

[0026] Figure 2 This is a comparison chart between the inversion results of the standard well mineral weight content and the XRD measured values.

[0027] Figure 3 To verify the comparison between the well mineral weight content inversion results and the XRD measured values. DETAILED DESCRIPTION

[0028] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0029] See also Figure 1 The embodiment of the present invention provides a method for automatically determining mineral inversion logging response parameters, the method comprising: Step 1: Collection and preprocessing of logging and core data: Obtain logging curve data and remove invalid values, and collect XRD data as the basis for inversion effect verification, wherein the logging curve data includes density, acoustic wave time difference, neutron, gamma, element logging, etc.; Step 2: Constructing the logging response parameter matrix: Based on the formation characteristics and the collected logging curves and XRD data, comprehensively establish the logging response parameter matrix; Step 3: Set the value range of the logging response parameter: according to the theoretical value and the empirical data of the region, set the value range of each logging response parameter in the matrix; Step 4: Construct the actual calculation matrix: select specific values ​​according to the value range set in step 3 and fill them into the matrix to construct a complete matrix for calculation; Step 5: Mineral volume content inversion: Mineral volume content inversion is performed based on the damped least squares method to solve the mineral volume content within the treatment depth range; Step 6: Calculation of mineral weight content: According to the mineral density logging response parameters, the mineral volume content is converted into weight content percentage; Step 7: Parameter optimization: Use the Optuna parameter optimization framework to optimize the logging response parameters, with the goal of minimizing the error between the calculated mineral weight content and the XRD measured value. Through multiple rounds of iterations, until the set number of iterations is reached, the optimal logging response parameter matrix is ​​obtained; Step 8: Save and analyze results: Save the optimized logging response parameter matrix and mineral weight content inversion results, draw a comparison chart between the inversion results and the XRD data, and generate an error table to analyze the mineral inversion effect after parameter optimization.

[0030] In this embodiment, the collection and preprocessing of logging and core data in step 1 includes the following steps: S11. Well logging data collection: Well logging data includes well logging curve data, which includes density, acoustic wave time difference, neutron, gamma, element logging, etc. Collect LAS format data of a well, usually including conventional logging curves, such as acoustic transit time (AC, in μs / ft), natural gamma (GR, in API), density (DEN, in g / cm 3 ), neutron porosity (CNL, in %), deep resistivity (RLLD, in OHMM) and shallow resistivity (RLLS, in OHMM), etc. If more types of minerals need to be calculated, element logging curves can be added. Such as sodium (Na, in w / w), magnesium (Mg, in w / w), aluminum (Al, in w / w), silicon (Si, in w / w), potassium (K, in w / w), calcium (Ca, in w / w), titanium (Ti, in w / w), and iron (Fe, in w / w). This well will be used as a standard well, mainly to construct a logging response parameter matrix suitable for this area.

[0031] S12. Preprocessing of logging data: Process invalid values ​​in the logging data and delete rows containing -99999 to avoid calculation errors in subsequent steps; For example, in the logging curve data of a well at 5005.5-5005.8m (see Table 1), invalid data such as -99999 appears at 5005.7-5005.8m, and the data in this depth range needs to be deleted.

[0032] Table 1 - Example of invalid values ​​of well logging data ; S13. Scope of data processing: Set the depth range for data processing, specify the maximum and minimum values ​​for the processing depth, and perform calculations only within this depth range; for example, for a certain well, the processing depth range is specified to be 4641-5000m, and calculations are only performed within this depth range.

[0033] S14. XRD data collection: Collect XRD data of a well, extract data within the depth range set in S13, and obtain the weight content data of the main minerals in the formation within this range; For example, for a certain well, the XRD data of the well is collected, the XRD data within the specified depth range is extracted, and the weight content data of the main minerals in the formation (clay minerals, quartz, feldspar, dolomite, calcite, amorphous, biotite) are obtained.

[0034] In this embodiment, step 2 of constructing the logging response parameter matrix includes the following steps: S21. Determine the calculation curve and the minerals to be calculated: According to the logging curve data, XRD data and logging processing requirements (logging data preprocessing, specifying the data processing range) collected in step 1, determine the logging curve used in subsequent calculations, as well as the types and quantities of minerals to be calculated; For example, the minerals that need to be calculated in a certain well are clay minerals, quartz, feldspar, dolomite, calcite, amorphous, and biotite, and the logging curves include density, neutron porosity, sonic time difference, Ca, Si, Mg, and Fe curves.

[0035] S22. Construct mineral logging response parameter matrix: According to the logging curves and mineral types determined in S21, a mineral logging response parameter matrix is ​​constructed. Each row of the matrix corresponds to a logging curve, and the last row is a balance equation to ensure that the sum of the mineral and pore volume contents is 1. Each column of the matrix corresponds to a mineral, and the last column is the pore fluid logging response value, which is used to calculate the pore volume. Each element in the matrix represents the theoretical logging response value of a certain mineral in a certain logging curve.

[0036] Taking a certain well as an example, Table 2 is the logging response parameter matrix. The unassigned part is the logging response parameter that needs to be adjusted. DEN_XXX represents the mineral density logging response parameter, among which DEN_clay is the clay mineral density logging response parameter, DEN_quartz is the quartz density logging response parameter, DEN_feldspar is the feldspar density logging response parameter, DEN_dolomite is the dolomite density logging response parameter, DEN_calcite is the calcite density logging response parameter, DEN_amorphous is the amorphous density logging response parameter, and DEN_biotite is the biotite density logging response parameter. CNL_XXX represents the mineral neutron porosity logging response parameter, among which CNL_clay is the clay mineral neutron porosity logging response parameter, CNL_quartz is the quartz neutron porosity logging response parameter, and CNL_feldspar is the feldspar neutron Porosity logging response parameters, CNL_dolomite is the neutron porosity logging response parameter of dolomite, CNL_calcite is the neutron porosity logging response parameter of calcite, CNL_amorphous is the amorphous neutron porosity logging response parameter, CNL_biotite is the biotite neutron porosity logging response parameter, AC_XXX represents the mineral acoustic transit time logging response parameter, among which AC_clay is the clay mineral acoustic transit time logging response parameter, AC_quartz is the quartz acoustic transit time logging response parameter, AC_feldspar is the feldspar acoustic transit time logging response parameter, AC_dolomite is the dolomite acoustic transit time logging response parameter, AC_calcite is the calcite acoustic transit time logging response parameter, AC_amorphous is the amorphous acoustic transit time logging response parameter, and AC_biotite is the biotite acoustic transit time logging response parameter. In addition, the element logging response parameters are fixed values, as shown in Table 2.

[0037] Table 2 - Logging response parameter matrix ; In this embodiment, step 3 of setting the value range of the logging response parameter includes the following steps: According to the theoretical value and the empirical data of the region, the value range of each logging response parameter in the matrix is ​​set. During the optimization process, the logging response parameter will be taken within this range.

[0038] Taking a certain well as an example, the value range of its logging response parameters is shown in Table 3.

[0039] Table 3- Range of mineral logging response parameters for a certain well ; In this embodiment, step 4 of constructing the actual calculation matrix includes the following steps: S41. Obtain parameter value from the value range: According to the range of values ​​set in step 3, select specific values ​​for each logging response parameter that needs to be adjusted according to the specified method. The specific value selection method is detailed in step 7; S42, fill in the corresponding position of the matrix: Fill the selected specific parameter values ​​into the corresponding positions of the logging response parameter matrix constructed in step 3 one by one to ensure that the matrix is ​​complete and meets the subsequent calculation requirements.

[0040] In this embodiment, step 5, mineral volume content inversion includes the following steps: S51. Use damped least squares method to invert mineral volume content: The damped least squares method solves the inversion result through the matrix equation (Formula 1), and the calculation formula is as follows: Formula 1: ; in Obtain the volume content vector of each mineral for a certain depth, is the logging response parameter matrix, for The transpose of is the damping coefficient, which is used to enhance the stability of the matrix equation. is the identity matrix, is the logging value vector of a certain depth; S52, Volume content constraints: The calculation results are constrained to be reasonable, and values ​​less than 0 are replaced with 0, and values ​​greater than 1 are replaced with 1.

[0041] In this embodiment, step 6, calculating the mineral weight content, comprises the following steps: S61. Calculate the mineral weight content: Calculate the weight content of each mineral at each depth point using the following formula: Formula 2: ; in For the depth Previous The volume content of minerals, For the Density logging response parameters of minerals, For the depth Previous The weight content of each mineral, is the number of mineral species, For the range from 1 to For every value of is the number of depth points calculated, For the range from 1 to Every value of ; S62. Calculate the total mass of minerals: Calculate the total mass of minerals at each depth point using the formula: Formula 3: ; in For Depth Total mass of the above minerals; S63. Calculate the mineral weight percentage: Calculate the weight percentage of each mineral using the formula: Formula 4: ; in For the depth Previous The weight percentage of a mineral.

[0042] In this embodiment, step seven parameter optimization includes the following steps: S71. Initial Exploration: First, a series of parameter combinations are randomly generated, and the mineral weight content percentage is calculated based on these parameters , while calculating The total mean square error between the measured value and the XRD value is used as the objective function to record the parameter combination of each attempt to provide initial data for subsequent optimization; The total mean square error is calculated as follows: Formula 5: ; in For the Minerals at depth XRD measured weight percentage on is the total number of corresponding XRD depth points, is the total mean square error, first calculate Minerals in The mean square error of the depth points, and then all The mean square errors of various minerals are accumulated to obtain the total mean square error; S72. Data accumulation and classification: After a preset number of attempts, the results are divided into the best parameter combinations based on historical data ( Smaller, that is, smaller than the preset threshold ) and the less effective parameter combination ( Larger, that is, greater than the preset threshold ), this step provides a basis for subsequent sampling; S73, Intelligent Sampling: Bayesian optimization tends to select new parameters from the parameter range with better effects for trial, while trying to avoid the parameter range with poorer effects. To prevent falling into the local optimal solution, the preset probability threshold is retained during the optimization process to try the parameter range that has not been explored. This method can improve search efficiency, quickly find high-quality parameter combinations, and avoid falling into the local optimal solution.

[0043] S74, Iterative Optimization: Through multiple rounds of iterations, the parameter values ​​used to construct the response parameter matrix are continuously updated and optimized, and the parameter search range is gradually narrowed. Finally, after reaching the specified number of iterations, the value that can make the objective function The smallest optimal parameter combination.

[0044] In this embodiment, step eight, result preservation and analysis includes the following steps: S81. Save optimization results: Save the final optimized logging response parameter matrix to ensure that the result file format is standardized and easy to reuse, save the mineral weight content inversion results, and provide data support for calculation errors; S82. Draw a comparison chart of the results: Use drawing tools to draw a comparison chart between the mineral weight content inversion results and the XRD measured data, which can intuitively show the matching degree between the inversion effect and the actual measured value; S83, generate error table: Calculate the error indicators between the weight content of each mineral and the XRD measured value, including the mean absolute error (MAE) and the mean square error (MSE), and organize the calculation results into an error table and save them for subsequent analysis and evaluation; The MAE calculation formula is as follows: Formula 6: ; in For the The mean absolute error of the minerals; The MSE calculation formula is as follows: Formula 7: ; in For the The mean square error of the minerals.

[0045] In order to facilitate those skilled in the art to better understand the technical solution of the present invention, specific embodiments of the present invention are given as follows: The present invention is applied to a certain well as a standard well. First, the logging and core data are collected, the logging response matrix is ​​constructed and the parameter adjustment range is set, the damped least squares method is used to invert the mineral content, and the Optuna framework is used to adjust the parameters, and finally a suitable logging response parameter matrix is ​​obtained. Figure 1 Flow chart of this method.

[0046] Finally, the optimized logging response parameters were obtained (see Table 4), and a comparison chart between the standard well mineral weight content inversion results and the XRD measured values ​​was drawn (see Figure 2 ), which is used to evaluate the optimization effect. At the same time, the error table of the standard well (see Table 5) is generated to quantify the error between the mineral weight content inversion results and the XRD measured values, and to evaluate the rationality of parameter optimization.

[0047] Table 4 - Optimized logging response parameters ; Table 5 - Standard Well Error Table ; The optimal logging response parameter matrix was used for a well section with similar lithology in another verification well in the area, and a comparison chart of the mineral weight content inversion results and the XRD measured values ​​was drawn (see Figure 3 ), and an error table is generated (see Table 6). It can be seen that the inversion results are in good agreement with the measured XRD data, which verifies the applicability of the parameters.

[0048] Table 6-Verification Well Error Table ; It should be noted that, in this article, the term "comprises" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0049] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for automatically determining mineral inversion logging response parameters, characterized in that: The method comprises: Step 1: Collection and preprocessing of logging and core data: Obtain logging curve data and remove invalid values, and collect XRD data as the basis for inversion effect verification, wherein the logging curve data includes density, acoustic wave time difference, neutron, gamma and element logging; Step 2: Constructing the logging response parameter matrix: Based on the formation characteristics and the collected logging curves and XRD data, comprehensively establish the logging response parameter matrix; Step 3: Set the value range of logging response parameters: according to theoretical values ​​and regional empirical data, set the value range of each logging response parameter in the matrix; Step 4: Construct the actual calculation matrix: select specific values ​​according to the value range set in step 3 and fill them into the matrix to construct a complete matrix for calculation; Step 5: Mineral volume content inversion: Mineral volume content inversion is performed based on the damped least squares method to solve the mineral volume content within the treatment depth range; Step 6: Calculation of mineral weight content: According to the mineral density logging response parameters, the mineral volume content is converted into weight content percentage; Step 7: Parameter optimization: Use the Optuna parameter optimization framework to optimize the logging response parameters, with the goal of minimizing the error between the calculated mineral weight content and the XRD measured value. Through multiple rounds of iterations, until the set number of iterations is reached, the optimal logging response parameter matrix is ​​obtained; Step 8: Save and analyze results: Save the optimized logging response parameter matrix and mineral weight content inversion results, draw a comparison chart between the inversion results and the XRD data, and generate an error table to analyze the mineral inversion effect after parameter optimization.

2. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 1 The collection and preprocessing of logging and core data includes the following steps: S11. Well logging data collection: Well logging data includes well logging curve data, which includes density, acoustic wave time difference, neutron, gamma and element logging; S12. Preprocessing of logging data: Process invalid values ​​in the logging data and delete rows containing -99999 to avoid calculation errors in subsequent steps; S13. Scope of data processing: Set the depth range for data processing, specify the maximum and minimum values ​​for processing depth, and perform calculations only within this depth range; S14, XRD data collection: Collect XRD data and extract data within the depth range set in S13 to obtain the weight content data of the main minerals in the formation within this range.

3. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 2: Constructing the logging response parameter matrix includes the following steps: S21. Determine the calculation curve and the minerals to be calculated: According to the well logging curve data, XRD data and well logging processing requirements collected in step 1, determine the well logging curve used in subsequent calculations, as well as the types and quantities of minerals to be calculated; S22. Construct mineral logging response parameter matrix: According to the logging curves and mineral types determined in S21, a mineral logging response parameter matrix is ​​constructed. Each row of the matrix corresponds to a logging curve, and the last row is a balance equation to ensure that the sum of the mineral and pore volume contents is 1. Each column of the matrix corresponds to a mineral, and the last column is the pore fluid logging response value, which is used to calculate the pore volume. Each element in the matrix represents the theoretical logging response value of a certain mineral in a certain logging curve.

4. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 4: Constructing the actual calculation matrix includes the following steps: S41. Obtain parameter value from the value range: According to the value range set in step 3, select a specific value for each logging response parameter to be adjusted according to the specified method; S42, fill in the corresponding position of the matrix: Fill the selected specific parameter values ​​into the corresponding positions of the logging response parameter matrix constructed in step 3 one by one to ensure that the matrix is ​​complete and meets the subsequent calculation requirements.

5. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 5: Mineral volume content inversion includes the following steps: S51. Use damped least squares method to invert mineral volume content: The damped least squares method solves the inversion result through the matrix equation, and the calculation formula is as follows: Formula 1: ; in Obtain the volume content vector of each mineral for a certain depth, is the logging response parameter matrix, for The transpose of is the damping coefficient, which is used to enhance the stability of the matrix equation. is the identity matrix, is the logging value vector of a certain depth; S52, Volume content constraints: The calculation results are constrained to be reasonable, and values ​​less than 0 are replaced with 0, and values ​​greater than 1 are replaced with 1.

6. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 6 The mineral weight content calculation includes the following steps: S61. Calculate the mineral weight content: Calculate the weight content of each mineral at each depth point using the following formula: Formula 2: ; in For the depth Previous The volume content of minerals, For the Density logging response parameters of minerals, For the depth Previous The weight content of each mineral, is the number of mineral species, For the range from 1 to For every value of is the number of depth points calculated, For the range from 1 to Every value of ; S62. Calculate the total mass of minerals: Calculate the total mass of minerals at each depth point using the formula: Formula 3: ; in For Depth Total mass of the above minerals; S63. Calculate the mineral weight percentage: Calculate the weight percentage of each mineral using the formula: Formula 4: ; in For the depth Previous The weight percentage of a mineral.

7. A method for automatically determining mineral inversion logging response parameters according to claim 1, characterized in that: Step 7 Parameter optimization includes the following steps: S71, Initial Exploration: First, a series of parameter combinations are randomly generated, and the mineral weight content percentage is calculated based on these parameters , while calculating The total mean square error between the measured value and the XRD value is used as the objective function to record the parameter combination of each attempt to provide initial data for subsequent optimization; The total mean square error is calculated as follows: Formula 5: ; in For the Minerals at depth XRD measured weight percentage on is the total number of corresponding XRD depth points, is the total mean square error, first calculate Minerals in The mean square error of the depth points, and then all The mean square errors of various minerals are accumulated to obtain the total mean square error; S72. Data accumulation and classification: After a preset number of attempts, the results are divided into parameter combinations with better effects and parameter combinations with worse effects based on historical data; S73, Intelligent Sampling: The probability of Bayesian optimization selecting new parameters from the parameter range with better effects to try is greater than the probability of selecting from the parameter range with worse effects. In order to prevent falling into the local optimal solution, the preset probability threshold is retained during the optimization process to try the parameter range that has not been explored; S74, Iterative Optimization: Through multiple rounds of iterations, the parameter values ​​used to construct the response parameter matrix are continuously updated and optimized, and the parameter search range is gradually narrowed. Finally, after reaching the specified number of iterations, the value that can make the objective function The smallest optimal parameter combination.

8. A method for automatically determining mineral inversion logging response parameters according to claim 7, characterized in that: Step 8: Result preservation and analysis includes the following steps: S81. Save optimization results: Save the final optimized logging response parameter matrix to ensure that the result file format is standardized and easy to reuse, save the mineral weight content inversion results, and provide data support for calculation errors; S82. Draw a comparison chart of the results: Use drawing tools to draw a comparison chart between the mineral weight content inversion results and the XRD measured data, which can intuitively show the matching degree between the inversion effect and the actual measured value; S83, generate error table: Calculate the error indicators between the weight content of each mineral and the XRD measured value, including the mean absolute error (MAE) and the mean square error (MSE), and organize the calculation results into an error table and save them for subsequent analysis and evaluation; The MAE calculation formula is as follows: Formula 6: ; in For the The mean absolute error of the minerals; The MSE calculation formula is as follows: Formula 7: ; in For the The mean square error of the minerals.

Images