Intelligent reservoir assessment method, device and equipment based on logging data and medium
By using an intelligent reservoir assessment method based on well logging data and generating continuous median pore throat radius curves using multiple models, the problem that micron-level median pore throat radius data in existing technologies cannot fully reflect the dynamic changes in the reservoir's micropore structure is solved, thus improving the accuracy and efficiency of reservoir assessment.
Patent Information
- Application Number
- CN202511049102.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-07-29
- Publication Date
- 2025-11-25
AI Technical Summary
Existing median pore throat radius data at the micrometer level are obtained through digital core experiments, which makes it difficult to fully reflect the dynamic changes in the micropore structure of the reservoir at different depths, resulting in low accuracy of reservoir evaluation results.
By acquiring well logging data and discrete median pore throat radius data, and using various base models such as extreme random trees, random forests, gradient boosting, CatBoost, and XGBoost, the target continuous median pore throat radius curve of the reservoir to be evaluated is generated. This curve is then fused with a meta-learner to generate more accurate evaluation results.
This method enables the indirect prediction of median pore throat radius using well logging data, reducing experimental costs, overcoming the limitation that median pore throat radius data is discrete data, and improving the accuracy of reservoir assessment.
Smart Images
Figure CN121007002A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the fields of machine learning and geological modeling technology, and in particular to an intelligent reservoir assessment method, apparatus, equipment and medium based on well logging data. Background Technology
[0002] Reservoir evaluation plays a crucial role in oil and gas exploration and development, directly impacting the efficiency and economic benefits of oil and gas fields. Accurate assessment of the reservoir's micropore structure is essential for understanding fluid flow characteristics, predicting oil and gas production, and optimizing development strategies. Among numerous micropore structure parameters, the micrometer-scale median pore throat radius is a key indicator. It directly reflects the size of the fluid flow channels, determines the fluid's transport capacity, and is strongly correlated with permeability. Therefore, the micrometer-scale median pore throat radius has significant guiding value in reservoir evaluation.
[0003] Existing median pore throat radius data at the micrometer level are mainly obtained through digital core experiments. However, the median pore throat radius data at the micrometer level obtained in this way cannot fully reflect the dynamic changes of the micropore structure of the reservoir at different depths, resulting in low representativeness in reservoir evaluation and consequently low accuracy of reservoir assessment results.
[0004] The above content is only used to help understand the technical solution of the present invention and does not represent an admission that the above content is prior art. Summary of the Invention
[0005] The main objective of this application is to provide an intelligent reservoir assessment method, apparatus, equipment, and medium based on well logging data, aiming to improve the accuracy of reservoir assessment results.
[0006] In a first aspect, embodiments of this application provide an intelligent reservoir assessment method based on well logging data, the method comprising:
[0007] Acquire characteristic data of the reservoir to be evaluated; wherein, the characteristic data includes well logging data and discrete median pore throat radius data;
[0008] The target information is input into at least one base model, and the geological features of the reservoir to be evaluated are extracted by each base model based on the target information to obtain the initial continuous median pore throat radius curves of the reservoir to be evaluated and each base model in the at least one base model; wherein, the target information is determined based on the feature data;
[0009] Based on the initial continuous median pore throat radius curves corresponding to each base model, the target continuous median pore throat radius curve of the reservoir to be evaluated is generated.
[0010] The reservoir to be evaluated is evaluated based on the target continuous median pore throat radius curve, and the evaluation result of the reservoir to be evaluated is obtained.
[0011] Secondly, embodiments of this application provide an intelligent reservoir assessment device based on well logging data, the device comprising:
[0012] The acquisition module is used to acquire characteristic data of the reservoir to be evaluated; wherein, the characteristic data includes well logging data and discrete median pore throat radius data;
[0013] The module is used to input target information into at least one base model, and extract geological features of the reservoir to be evaluated based on the target information through each base model in the at least one base model, thereby obtaining the initial continuous median pore throat radius curves of the reservoir to be evaluated and each base model in the at least one base model; wherein, the target information is determined based on the feature data;
[0014] The generation module is used to generate the target continuous median pore throat radius curve of the reservoir to be evaluated based on the initial continuous median pore throat radius curves corresponding to each base model.
[0015] The evaluation module evaluates the reservoir to be evaluated based on the target continuous median pore throat radius curve, and obtains the evaluation result of the reservoir to be evaluated.
[0016] Thirdly, embodiments of this application provide an intelligent reservoir assessment device based on well logging data. The device includes: a memory, a processor, and a computer processing program stored in the memory and executable on the processor. The computer processing program is configured to implement the intelligent reservoir assessment method based on well logging data as described in the first aspect.
[0017] Fourthly, embodiments of this application provide a storage medium storing a computer processing program, which, when executed by a processor, implements the intelligent reservoir evaluation method based on well logging data as described in the first aspect.
[0018] Fifthly, embodiments of this application provide a computer program product, which is stored in a storage medium and executed by at least one processor to implement the intelligent reservoir evaluation method based on well logging data as described in the first aspect.
[0019] This application proposes an intelligent reservoir assessment method, apparatus, equipment, and medium based on well logging data. The method involves acquiring characteristic data of the reservoir to be assessed, including well logging data and discrete median pore throat radius data. Target information is input into at least one base model. Each base model extracts geological features of the reservoir based on the target information, resulting in initial continuous median pore throat radius curves corresponding to the reservoir and each base model. The target information is determined based on the characteristic data. A target continuous median pore throat radius curve is generated for the reservoir based on the initial continuous median pore throat radius curves of each base model. The reservoir is then assessed based on the target continuous median pore throat radius curves to obtain the assessment result. Thus, by inputting the target information into the model, a continuous median pore throat radius curve for reservoir evaluation can be obtained. This not only enables the indirect prediction of the median pore throat radius using well logging data, thereby reducing experimental costs, but also overcomes the limitation that the median pore throat radius data is discrete data, generating a continuous median pore throat radius curve for the target, thereby providing a more accurate basis for reservoir evaluation and improving the accuracy of the evaluation results. Attached Figure Description
[0020] Figure 1 This is a flowchart illustrating the intelligent reservoir assessment method based on well logging data provided in the embodiments of this application;
[0021] Figure 2 This is a schematic diagram comparing a single model and multiple model stacking provided in an embodiment of this application;
[0022] Figure 3 , Figure 4 , Figure 5 , Figure 6 and Figure 7 These are schematic diagrams illustrating the different interpretations of dataset features by the various algorithms provided in the embodiments of this application;
[0023] Figure 8 This is a schematic diagram of the Stacking ensemble learning architecture based on Pycaret provided in the embodiments of this application;
[0024] Figure 9 This is a schematic diagram comparing the actual reservoir R50_pred curve with core data provided in the embodiments of this application;
[0025] Figure 10 This is a schematic diagram of the structure of the intelligent reservoir evaluation device based on well logging data provided in the embodiments of this application;
[0026] Figure 11 This is one of the structural schematic diagrams of the electronic device provided in the embodiments of this application;
[0027] Figure 12This is a second schematic diagram of the structure of the electronic device provided in the embodiments of this application;
[0028] The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0029] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.
[0030] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.
[0031] The intelligent reservoir evaluation method based on well logging data provided in this application will be described in detail below with reference to the accompanying drawings, through specific embodiments and application scenarios.
[0032] This application provides an intelligent reservoir assessment method based on well logging data, such as... Figure 1 As shown, the intelligent reservoir assessment method based on well logging data in this application embodiment may include the following steps:
[0033] Step 101: Obtain characteristic data of the reservoir to be evaluated; wherein, the characteristic data includes well logging data and discrete median pore throat radius data;
[0034] In some embodiments, in step 101 above, data for assessing reservoir quality are collected, including well logging data and discrete median pore throat radius data.
[0035] In this embodiment, it should be noted that the characteristic data may include well logging data and discrete median pore throat radius data, used to describe reservoir characteristics. Well logging data can be subsurface rock data obtained directly through drilling, such as neutron density, density, natural gamma ray, porosity, and clay content. Discrete median pore throat radius data can be statistical data describing the distribution of pore size, such as R50.
[0036] In some implementation methods, characteristic data of the reservoir to be evaluated are obtained through on-site measurements, laboratory analysis, or historical data recording.
[0037] In some implementation methods, the data of the reservoir to be evaluated, obtained through on-site measurement, laboratory analysis, or historical data recording, is preprocessed to obtain the characteristic data of the reservoir to be evaluated.
[0038] Step 102: Input the target information into at least one base model, and extract the geological features of the reservoir to be evaluated based on the target information using each base model in the at least one base model to obtain the initial continuous median pore throat radius curves of the reservoir to be evaluated and each base model in the at least one base model; wherein, the target information is determined based on the feature data;
[0039] In some embodiments, in step 102 above, the collected feature data is input into at least one base model to extract geological features. Based on the features extracted from the model, an initial continuous median pore throat radius curve is generated for each model.
[0040] In this embodiment, it should be noted that the base model can be an algorithmic model used to learn from data and make predictions. The continuous median pore throat radius curve can be a curve describing the continuous change in pore size.
[0041] In some implementations, feature data is input into multiple algorithm models, including Extreme Random Tree, Random Forest, Gradient Boosting, CatBoost, and XGBoost. Based on the model prediction results, a corresponding initial continuous median throat radius curve is generated.
[0042] In other implementations, feature data is input into an algorithm model, and an initial continuous median pore throat radius curve is generated based on the model's prediction results.
[0043] Step 103: Generate the target continuous median pore throat radius curve of the reservoir to be evaluated based on the initial continuous median pore throat radius curves corresponding to each base model.
[0044] In some embodiments, in step 103 above, the curves generated by each model are combined to generate the final target continuous median pore throat radius curve of the reservoir to be evaluated.
[0045] In some implementations, the best-performing model result is selected as the final target continuous median pore throat radius curve.
[0046] In other implementations, the results of multiple models are integrated to obtain the final target continuous median pore throat radius curve.
[0047] In other implementations, multiple model results are voted on using a voting mechanism to obtain the final target continuous median pore throat radius curve.
[0048] Step 104: Evaluate the reservoir to be evaluated based on the target continuous median pore throat radius curve to obtain the evaluation result of the reservoir to be evaluated.
[0049] In some embodiments, in step 104 above, the reservoir quality is assessed based on the generated target continuous median pore throat radius curve.
[0050] In some implementations, the shape, distribution, and characteristics of the target continuous median pore throat radius curve are analyzed to evaluate the reservoir.
[0051] In other implementations, the shape, distribution, and characteristics of the target continuous median pore throat radius curve are analyzed and evaluated in conjunction with geological and engineering standards.
[0052] For example, suppose an oilfield company needs to assess the productivity of a new reservoir: Well logging data and discrete median pore throat radius data for the reservoir are collected. Random forest and XGBoost models are used to train the data to extract the reservoir's geological features. Based on the model predictions, an initial continuous median pore throat radius curve is generated. By combining the results from each model, a target continuous median pore throat radius curve is generated, and the reservoir's productivity and quality are assessed accordingly.
[0053] In this embodiment, characteristic data of the reservoir to be evaluated is acquired, including well logging data and discrete median pore throat radius data. Target information is input into at least one base model. Each base model extracts geological features of the reservoir based on the target information, resulting in initial continuous median pore throat radius curves corresponding to the reservoir and each base model. The target information is determined based on the characteristic data. A target continuous median pore throat radius curve is generated for the reservoir. The reservoir is then evaluated based on the target continuous median pore throat radius curve, yielding the evaluation result. Thus, by inputting target information into the model, a continuous median pore throat radius curve for the evaluated reservoir can be obtained. This not only enables indirect prediction of the median pore throat radius using well logging data, reducing experimental costs, but also overcomes the limitation that median pore throat radius data is discrete, generating a target continuous median pore throat radius curve, thereby providing a more accurate basis for reservoir evaluation and improving the accuracy of the evaluation results.
[0054] In some implementations, acquiring the characteristic data of the reservoir to be evaluated includes:
[0055] Obtain the initial logging data and initial discrete median pore throat radius data of the reservoir to be evaluated;
[0056] The initial logging data and initial discrete median pore throat radius data are preprocessed to obtain the characteristic data of the reservoir to be evaluated; wherein, the preprocessing includes at least depth alignment of the initial logging data and initial discrete median pore throat radius data.
[0057] In this embodiment, initial logging data and initial discrete median pore throat radius data are collected for evaluating reservoir quality. The initial logging data and initial discrete median pore throat radius data are preprocessed, including depth alignment, to obtain characteristic data of the reservoir to be evaluated.
[0058] In this embodiment, it should be noted that depth alignment can be achieved by adjusting the initial logging data and the initial discrete median pore throat radius data to the same depth scale in order to facilitate feature extraction.
[0059] In some implementations, data (initial logging data and initial discrete median pore throat radius data) are collected through drilling sampling, logging, and specialized porosity measurement techniques (such as mercury intrusion porosimetry). The data undergoes the following processes: data cleaning to remove erroneous or incomplete data points; standardization to convert the data to a unified dimension, such as converting data from different units to the International System of Units (SI); and depth alignment to adjust data from different wells to the same depth scale for comparison and comprehensive analysis. Characteristic data are then extracted from the preprocessed data.
[0060] In other implementations, initial logging data and initial discrete median pore throat radius data are collected through drilling sampling, logging, and specialized porosity measurement techniques. Data preprocessing includes target well section selection and data cleaning; alignment of logging depth with digital core depth; data resampling; and Z-score normalization. This process yields feature data.
[0061] In this embodiment, the accuracy and consistency of the data are ensured by preprocessing the initial logging data and the initial discrete median pore throat radius data, providing high-quality input for subsequent analysis.
[0062] In some implementations, after acquiring the characteristic data of the reservoir to be evaluated and before inputting the target information into at least one base model, the method further includes:
[0063] Based on the characteristic data of the reservoir to be evaluated, a first data column is generated;
[0064] The target feature parameters in the feature data of the reservoir to be evaluated are enhanced to obtain new target feature parameters.
[0065] Based on the new target feature parameters, a second data column is generated;
[0066] The target information is determined based on the first data column and the second data column.
[0067] In this embodiment, based on the feature data of the reservoir to be evaluated, a first data column containing the feature data is generated. Key parameters in the feature data are then augmented to improve their discriminative ability in the model. Based on the augmented key parameters, a second data column is generated. Combining the first and second data columns, target information for model training and evaluation is determined.
[0068] In this embodiment, it should be noted that the target feature parameter can be a key parameter in the feature data, such as shallow or deep resistivity, conductivity, etc. Feature enhancement can be performed by amplifying features with high importance based on importance analysis, such as through square or polynomial calculations.
[0069] In some implementations, the first and second data columns can be used as the target information.
[0070] In some other implementations, the first data column can be used as the target information.
[0071] In some other implementations, the second data column can be used as the target information.
[0072] In this embodiment, by performing data augmentation on the target feature parameters, the model's ability to identify reservoir characteristics is improved, thereby enhancing the model's predictive accuracy. By using the first and second data columns to determine the target information input to the model, the focus of model training and evaluation can be ensured to be concentrated on the most relevant features, improving the efficiency and effectiveness of the evaluation.
[0073] In some implementations, the at least one base model includes at least one of the following:
[0074] The extreme random tree model is used to capture the global trend corresponding to the target information.
[0075] A random forest model, wherein the random forest model is used to resist noise interference based on the target information;
[0076] A gradient boosting regression tree model is used to enhance the local features of the target information.
[0077] The CatBoost model is used to process categorical parameters in the target information.
[0078] The XGBoost model is used to optimize the continuous extreme variables corresponding to the target information.
[0079] In this implementation, the Extra-Trees (ET) model, by introducing extreme randomness (e.g., random feature selection and random split point selection) when constructing the decision tree, is used to capture global trends in the data and improve the model's generalization ability. The Random Forest (RF) model, by constructing multiple decision trees and combining their predictions, is used to improve the model's stability and resistance to noise. The Gradient Boosting Regression (GBR) model strengthens the model's ability to capture local features by iteratively training the decision tree, focusing at each step on instances that the learner failed to predict correctly in the previous step. The CatBoost model, suitable for handling categorical parameters, is used for classification problems by combining multiple weak classifiers to improve classification performance. The XGBoost model, an implementation of the gradient boosting algorithm, is suitable for classification and regression problems; it trains the model by optimizing the objective function and is used to optimize the prediction of continuous extreme variables.
[0080] The choice of these models should be determined based on the specific needs of the problem and the characteristics of the data. For example, if the goal is to capture global trends, the Extreme Random Tree (ET) model is a good choice; if the goal is to improve the model's stability and robustness to noise, the Random Forest (RF) model is more suitable; if local features need to be enhanced, the Gradient Boosting Regression Tree (GBR) model is more effective; if dealing with classification problems, the CatBoost model is the best choice; if the goal is to optimize continuous variables, the XGBoost model is more applicable. It is also possible to integrate five models: Extreme Random Tree (ET) (for quickly capturing global trends), Random Forest (RF) (for robustness to noise), Gradient Boosting Regression Tree (GBR) (for enhancing local features), CatBoost (for handling categorical parameters), and XGBoost (for optimizing continuous variables).
[0081] In this embodiment, at least one model that can be included in the base model is given. By integrating at least one base model, the multidimensional parameter space can be divided and conquered to form complementary heterogeneous models. Since the data volume of micro parameters (such as discrete median pore throat radius data) is sparse, the different interpretations of the dataset features by various algorithms can be used during modeling to achieve the goal of comprehensively capturing geological features.
[0082] In some implementations, generating the target continuous median pore throat radius curve of the reservoir to be evaluated based on the initial continuous median pore throat radius curves corresponding to each base model includes:
[0083] The initial continuous median pore throat radius curves corresponding to each base model are input into the meta-learner. The meta-learner then fuses the initial continuous median pore throat radius curves corresponding to each base model to generate the target continuous median pore throat radius curve of the reservoir to be evaluated.
[0084] In this implementation, the prediction results (initial continuous median throat radius curves) obtained from various base models (such as extreme random trees, random forests, gradient boosting regression trees, etc.) are used as input data and fed into a meta-learner to train the meta-learner to learn the optimal combination of these results. The trained meta-learner is then applied to obtain the fused prediction result (target continuous median throat radius curve).
[0085] In this embodiment, it should be noted that the meta-learner can be a model used in ensemble learning to optimize the combination of base models.
[0086] In some implementations, at different reservoir levels, the prediction results (initial continuous median pore throat radius curves) obtained from each model for the same level of target information are selected, and the model with the highest fit is preferred. Thus, the prediction results obtained at different levels are derived from different models. Combining these prediction results constitutes the reservoir prediction result (target continuous median pore throat radius curve).
[0087] In other implementations, the predictions from each base model are used as input data and fed into a meta-learner. Algorithms (such as stacking, an ensemble technique that trains a meta-model by stacking the predictions from multiple models) are used to combine the predictions from the base models. The trained meta-learner is then applied to predict new data to obtain the fused pore throat radius curve.
[0088] In this implementation, the model's generalization ability is improved by fusing prediction results through a meta-learner. A target continuous median pore throat radius curve is generated, providing a detailed description of the reservoir and a basis for further evaluation and decision-making.
[0089] In some embodiments, evaluating the reservoir to be evaluated based on the target continuous median pore throat radius curve to obtain the evaluation result of the reservoir to be evaluated includes:
[0090] If the target continuous median pore throat radius curve shows that the pore size is uniformly distributed in the reservoir to be evaluated, the evaluation result of the reservoir to be evaluated is determined to be a high-quality reservoir.
[0091] When the target continuous median pore throat radius curve shows that the pore size is not uniformly distributed in the reservoir to be evaluated, the evaluation result of the reservoir to be evaluated is determined to be a low-quality reservoir or a potential reservoir.
[0092] In this embodiment, if the curve remains relatively smooth across the entire reservoir depth range without obvious peaks or valleys, this indicates a uniform pore size distribution, thus determining the reservoir to be of high quality. If the curve shows multiple peaks and valleys, particularly with significantly increased pore size in certain depth ranges and smaller pore size in others, this indicates a non-uniform pore size distribution, thus determining the reservoir to be of low quality or with high production potential in specific regions. Specifically, a reservoir with high production potential in specific regions may be a reservoir with larger pores in certain local areas, even if the overall pore distribution is uneven; a low-quality reservoir may be a reservoir with generally small or highly variable pores, which can lead to restricted fluid flow and affect overall production capacity.
[0093] In this embodiment, by analyzing the performance of the target continuous median pore throat radius curve, the characteristics of the reservoir can be assessed more accurately, providing a scientific basis for oilfield development and production decisions, and helping to optimize resource recovery and improve production efficiency.
[0094] This application also provides a complete process for an intelligent reservoir assessment method based on well logging data, which includes five steps:
[0095] Step 1: Data Preparation (Constructing a dataset)
[0096] Input data: Conventional logging data (neutron CNCF, density ZDEN, natural gamma ray GR, porosity POR, clay content Vsh, etc.), and discrete micron-level median pore throat radius data (R50) from digital core analysis.
[0097] Data preprocessing: target well section selection and data cleaning; alignment of logging depth with digital core depth; data resampling and Z-score normalization.
[0098] Step 2: Feature Engineering
[0099] Based on the importance and correlation analysis results, key feature parameters are selected for feature enhancement (feature enhancement can square or multiply a feature by a constant to create a new feature column (input column), which is then used in conjunction with the original data column (dataset) to train the model and obtain a continuous micrometer-level median pore throat radius curve).
[0100] Step 3: Multi-model integration and stacking
[0101] Five models are integrated, namely Extreme Random Tree (ET, which quickly captures global trends), Random Forest (RF, which resists noise interference), Gradient Boosting Regression Tree (GBR, which strengthens local features), CatBoost (which handles categorical parameters), and XGBoost (which optimizes continuous variables), to divide and conquer the multidimensional parameter space and form complementary heterogeneous models.
[0102] Because the amount of microscopic parameter data is sparse, different algorithms are used to interpret the characteristics of the dataset during modeling, so as to achieve the goal of comprehensively capturing geological features.
[0103] Reference Figure 2 , Figure 2 This is a comparison chart of single-model and multi-model stacking, showing the median pore radius curves at the micrometer level obtained through 10-fold cross-validation for small sample data (based on the dataset). The R-value of the single-model (RF) is also shown. 2 The R² value is 0.6, while multi-model ensemble stacking (ET, RF, GBR, CatBoost, and XGBoost) forms a complementary heterogeneous model, and the R² value of multi-model ensemble stacking is 0.6. 2 The value is 0.83. The results show that multi-model ensemble stacking performs excellently, with high explanatory power and prediction accuracy.
[0104] Reference Figure 3 , Figure 4 , Figure 5 , Figure 6 and Figure 7 Five feature importance plots illustrate the different interpretations of dataset features by various algorithms. The Feature Importance Plot represents the importance of features, while Variable Importance represents the importance of variables. Features include: POR (porosity), CAL (caliber), ZDEN (density), GR (natural gamma), ROP-AVG (drilling speed), MLR1C, MLR4C (shallow resistivity, deep resistivity), SH (shale content), CO2 (carbon dioxide content), TG (total hydrocarbons), DT24 (sonic transit time), CNCF (compensated neutrons), and C1 (methane content). The blocks of Lufeng, Kaiping, Enping, and Huizhou represent well logging data from Lufeng, Kaiping, Enping, and Huizhou, respectively.
[0105] Reference Figure 8 , Figure 8 A schematic diagram of the Stacking ensemble learning architecture based on Pycaret is presented. The data in the data sample library (dataset) is randomly divided into 10 mutually exclusive parts. Two parts are randomly selected as the validation set, and eight parts are used as the training set. Each algorithm used is trained and validated five times, resulting in five R... 2 For each algorithm, R for the same validation set 2 By comparing different models, the algorithm with the highest fit is selected for modeling a given data segment. This means that models for different well segments are constructed using different algorithms. By stacking these models, a final model for all the data is obtained.
[0106] Step 4: Model Training and Optimization
[0107] Loss function: A weighted combination of mean squared error (MSE) and mean absolute error (MAE);
[0108] Hyperparameter tuning: Pycaret framework automatically tunes;
[0109] Cross-validation: The sample size is too small, so K=10 is set when using K-fold cross-validation to prevent overfitting.
[0110] Step 5: Generation of Continuous Curves
[0111] The trained stacked model is applied to logging data across the entire well section, outputting a continuous micrometer-level median pore throat radius curve (R50_pred). (Refer to...) Figure 9 , Figure 9 A comparison chart is provided between the actual reservoir R50_pred curve (the green line representing the predicted value) and the core data (the red line representing the median pore throat radius in the micrometer range).
[0112] Furthermore, this application also provides a detailed implementation of the complete process of the above-mentioned intelligent reservoir evaluation method based on well logging data:
[0113] I. Data Collection:
[0114] We acquired conventional logging curves (POR, Vsh, GR, etc.) for the target well section and 400 sets of digital core micron-level median pore throat radius R50 data to form a small sample, which was then divided into training and test sets at a ratio of 9:1.
[0115] II. Example of Feature Engineering: Structural Features: ln (depth and shallow resistivity), ln (porosity);
[0116] III. Interpreting dataset features using the five models described above:
[0117] For example, the key parameters interpreted by the XGBoost model include: porosity (POR), drilling speed (ROP), borehole diameter (CAL), and natural gamma ray (GR).
[0118] IV. Stacking and fusion:
[0119] The meta-learner is trained using the predictions of the five base learners as input (using a simple averaging strategy (SMA) instead of a complex model to prevent overfitting).
[0120] V. Result Verification:
[0121] Using 10-fold cross-validation, test set R 2 The value reached 0.83, and the MAE was <0.1 μm, which was better than the single model (R<0.6).
[0122] Reference Figure 10One embodiment of this application provides an intelligent reservoir assessment device based on well logging data, the device comprising:
[0123] The acquisition module 1001 is used to acquire characteristic data of the reservoir to be evaluated; wherein, the characteristic data includes well logging data and discrete median pore throat radius data;
[0124] The module 1002 is used to input target information into at least one base model, and extract geological features of the reservoir to be evaluated based on the target information through each base model in the at least one base model, thereby obtaining the initial continuous median pore throat radius curves of the reservoir to be evaluated and each base model in the at least one base model; wherein, the target information is determined based on the feature data;
[0125] The generation module 1003 is used to generate the target continuous median pore throat radius curve of the reservoir to be evaluated based on the initial continuous median pore throat radius curves corresponding to each base model.
[0126] The evaluation module 1004 evaluates the reservoir to be evaluated based on the target continuous median pore throat radius curve, and obtains the evaluation result of the reservoir to be evaluated.
[0127] It should be noted that the present device embodiment and the above method embodiment are based on the same inventive concept. Therefore, the content of the above method embodiment is also applicable to the present device embodiment, and will not be repeated here.
[0128] Optionally, such as Figure 11 As shown, this application embodiment also provides an electronic device 1100, including a processor 1101 and a memory 1102. The memory 1102 stores a program or instructions that can run on the processor 1101. When the program or instructions are executed by the processor 1101, they implement the various steps of the above-described intelligent reservoir evaluation method embodiment based on well logging data and can achieve the same technical effect. To avoid repetition, they will not be described again here.
[0129] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.
[0130] Figure 12 A schematic diagram of the hardware structure of the electronic device used to implement the embodiments of this application.
[0131] The electronic device 1200 includes, but is not limited to, components such as: radio frequency unit 1201, network module 1202, audio output unit 1203, input unit 1204, sensor 1205, display unit 1206, user input unit 1207, interface unit 1208, memory 1209, and processor 1210.
[0132] Those skilled in the art will understand that the electronic device 1200 may also include a power supply (such as a battery) for supplying power to various components. The power supply may be logically connected to the processor 1210 through a power management system, thereby enabling functions such as managing charging, discharging, and power consumption through the power management system. Figure 12 The electronic device structure shown does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements, which will not be elaborated here.
[0133] The processor 1210 is used to: acquire characteristic data of the reservoir to be evaluated; wherein the characteristic data includes well logging data and discrete median pore throat radius data;
[0134] The target information is input into at least one base model, and the geological features of the reservoir to be evaluated are extracted by each base model based on the target information to obtain the initial continuous median pore throat radius curves of the reservoir to be evaluated and each base model in the at least one base model; wherein, the target information is determined based on the feature data;
[0135] Based on the initial continuous median pore throat radius curves corresponding to each base model, the target continuous median pore throat radius curve of the reservoir to be evaluated is generated.
[0136] The reservoir to be evaluated is evaluated based on the target continuous median pore throat radius curve, and the evaluation result of the reservoir to be evaluated is obtained.
[0137] In some embodiments, the processor 1210 is further configured to: acquire initial logging data and initial discrete median pore throat radius data of the reservoir to be evaluated;
[0138] The initial logging data and initial discrete median pore throat radius data are preprocessed to obtain the characteristic data of the reservoir to be evaluated; wherein, the preprocessing includes at least depth alignment of the initial logging data and initial discrete median pore throat radius data.
[0139] In some implementations, the processor 1210 is further configured to: generate a first data column based on the characteristic data of the reservoir to be evaluated;
[0140] The target feature parameters in the feature data of the reservoir to be evaluated are enhanced to obtain new target feature parameters.
[0141] Based on the new target feature parameters, a second data column is generated;
[0142] The target information is determined based on the first data column and the second data column.
[0143] In some implementations, the processor 1210 is further configured to: use a limit random tree model to capture the global trend corresponding to the target information;
[0144] A random forest model, wherein the random forest model is used to resist noise interference based on the target information;
[0145] A gradient boosting regression tree model is used to enhance the local features of the target information.
[0146] The CatBoost model is used to process categorical parameters in the target information.
[0147] The XGBoost model is used to optimize the continuous extreme variables corresponding to the target information.
[0148] In some embodiments, the processor 1210 is further configured to: input the initial continuous median pore throat radius curves corresponding to each base model into a meta-learner, and fuse the initial continuous median pore throat radius curves corresponding to each base model through the meta-learner to generate the target continuous median pore throat radius curve of the reservoir to be evaluated.
[0149] In some embodiments, the processor 1210 is further configured to: determine the evaluation result of the reservoir to be evaluated as a high-quality reservoir when the target continuous median pore throat radius curve shows that the pore size is uniformly distributed in the reservoir to be evaluated;
[0150] When the target continuous median pore throat radius curve shows that the pore size is not uniformly distributed in the reservoir to be evaluated, the evaluation result of the reservoir to be evaluated is determined to be a low-quality reservoir or a potential reservoir.
[0151] It should be understood that, in this embodiment, the input unit 1204 may include a graphics processing unit (GPU) 12041 and a microphone 12042. The GPU 12041 processes image data of still images or videos obtained by an image capture device (such as a camera) in video capture mode or image capture mode. The display unit 1206 may include a display panel 12061, which may be configured in the form of a liquid crystal display, an organic light-emitting diode, or the like. The user input unit 1207 includes a touch panel 12071 and at least one of other input devices 12072. The touch panel 12071 is also called a touch screen. The touch panel 12071 may include a touch detection device and a touch controller. Other input devices 12072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, power buttons, etc.), trackballs, mice, and joysticks, which will not be described in detail here.
[0152] The memory 1209 can be used to store software programs and various data. The memory 1209 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 1209 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 1209 in this embodiment includes, but is not limited to, these and any other suitable types of memory.
[0153] Processor 1210 may include one or more processing units; optionally, processor 1210 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 1210.
[0154] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described intelligent reservoir evaluation method embodiment based on well logging data and achieve the same technical effect. To avoid repetition, they will not be described again here.
[0155] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.
[0156] Furthermore, this application provides a computer program product stored in a storage medium. This program product is executed by at least one processor to implement the various processes of the above-described intelligent reservoir evaluation method embodiment based on well logging data, and can achieve the same technical effect. To avoid repetition, it will not be described again here.
[0157] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element. Furthermore, it should be noted that the scope of the methods and apparatus in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.
[0158] The sequence numbers of the above embodiments of the present invention are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments.
[0159] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of the present invention.
[0160] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.
Claims
1. A method of intelligent reservoir evaluation based on well logging data, characterized in that, The method comprises: obtaining characteristic data of a reservoir to be evaluated; wherein the characteristic data comprises well logging data and discrete median pore throat radius data; inputting target information into at least one base model respectively, extracting geological characteristics of the reservoir to be evaluated based on the target information by each base model in the at least one base model, and obtaining an initial continuous median pore throat radius curve corresponding to the reservoir to be evaluated and each base model in the at least one base model; wherein the target information is determined based on the characteristic data; generating a target continuous median pore throat radius curve of the reservoir to be evaluated according to the initial continuous median pore throat radius curve corresponding to each base model; evaluating the reservoir to be evaluated according to the target continuous median pore throat radius curve, and obtaining an evaluation result of the reservoir to be evaluated.
2. The method of claim 1, wherein, The method comprises: obtaining initial well logging data and initial discrete median pore throat radius data of the reservoir to be evaluated; preprocessing the initial well logging data and the initial discrete median pore throat radius data to obtain the characteristic data of the reservoir to be evaluated; wherein the preprocessing at least comprises depth alignment of the initial well logging data and the initial discrete median pore throat radius data.
3. The method of claim 1, wherein, After obtaining the characteristic data of the reservoir to be evaluated, before inputting the target information into the at least one base model respectively, the method further comprises: generating a first data column based on the characteristic data of the reservoir to be evaluated; performing feature enhancement on a target feature parameter in the characteristic data of the reservoir to be evaluated to obtain a new target feature parameter; generating a second data column based on the new target feature parameter; determining the target information based on the first data column and the second data column.
4. The method of claim 1, wherein, The at least one base model comprises at least one of: an extreme random tree model for capturing a global trend corresponding to the target information; a random forest model for resisting noise interference based on the target information; a gradient boosting regression tree model for strengthening local features of the target information; a CatBoost model for processing category type parameters in the target information; an XGBoost model for optimizing continuous extreme variables corresponding to the target information.
5. The method of claim 1, wherein, The method comprises: inputting the initial continuous median pore throat radius curve corresponding to each base model into a meta-learner, fusing the initial continuous median pore throat radius curve corresponding to each base model by the meta-learner, and generating the target continuous median pore throat radius curve of the reservoir to be evaluated.
6. The method of claim 1, wherein, The method comprises: in a case where the target continuous median pore throat radius curve displays that pore sizes are uniformly distributed in the reservoir to be evaluated, determining that the evaluation result of the reservoir to be evaluated is a high-quality reservoir. In a case that the target continuous median pore throat radius curve shows that the pore size is unevenly distributed in the reservoir to be evaluated, the evaluation result of the reservoir to be evaluated is determined as a low-quality reservoir or a potential reservoir.
7. An intelligent reservoir evaluation apparatus based on well logging data, characterized by, The device comprises: an acquisition module configured to acquire feature data of a reservoir to be evaluated, wherein the feature data comprises logging data and discrete median pore throat radius data; an obtaining module configured to input target information into at least one base model, and to obtain initial continuous median pore throat radius curves corresponding to the reservoir to be evaluated and each base model in the at least one base model by extracting geological features of the reservoir to be evaluated based on the target information by each base model in the at least one base model, wherein the target information is determined based on the feature data; a generating module configured to generate a target continuous median pore throat radius curve of the reservoir to be evaluated according to the initial continuous median pore throat radius curves corresponding to each base model; an evaluation module configured to evaluate the reservoir to be evaluated according to the target continuous median pore throat radius curve, and to obtain an evaluation result of the reservoir to be evaluated.
8. An intelligent reservoir evaluation device based on well logging data, characterized by, The device comprises a memory, a processor, and a computer processing program stored on the memory and executable on the processor, and the computer processing program is configured to implement the intelligent reservoir evaluation method based on logging data according to any one of claims 1 to 6.
9. A storage medium, characterized by The storage medium stores a computer processing program, and the computer processing program is executed by the processor to implement the intelligent reservoir evaluation method based on logging data according to any one of claims 1 to 6.
10. A computer program product, characterised in that, The instructions in the computer program product are executed by the processor of the electronic device, so that the electronic device performs the intelligent reservoir evaluation method based on logging data according to any one of claims 1 to 6.