Method for accurately identifying favorable region of reservoir based on random forest bimodal output
Through the random forest dual-modal output model, the porosity and pore throat type prediction is used to use core and log data, which solves the high cost of core sampling dependence in the existing technology, and achieves a more efficient and accurate reservoir advantage Area identification.
Patent Information
- Application Number
- CN202411366996.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-29
- Publication Date
- 2025-01-10
AI Technical Summary
The existing porosity prediction methods rely on core sampling and experiments, resulting in large labor and material consumption, and lack of efficient multi-parameter comprehensive evaluation methods.
The random forest dual-modal output model is used to pre-process the core mercury injected experimental data and logging curve data, and the pore throat radius distribution frequency map is constructed, the pore throat type is classified, and the random forest model is trained to perform poreness and pore throat type. Forecasting to achieve accurate identification of the entire well and the entire district.
It improves the prediction efficiency and accuracy of porosity and pore throat structure types, replaces a single prediction model, has broadness and flexibility, and is suitable for different oilfield exploration scenarios.
Smart Images
Figure CN119294433A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of geology, and in particular to a method for accurately identifying favorable reservoir areas based on random forest bimodal output. Background Art
[0002] When evaluating low-porosity and low-permeability reservoirs, the most effective method is to use multiple parameters to conduct a comprehensive evaluation of the reservoir. Among them, porosity is one of the important factors in evaluating reservoir characteristics; pore throat structure has an important influence on reservoir performance and seepage capacity. Clarifying the microscopic morphology of the reservoir pore throat structure helps to identify the reservoir type with high-efficiency production potential, which has always been a hot spot and focus of oil and gas reservoir research.
[0003] At present, the porosity prediction methods mainly include three categories: rock physics experimental measurement, logging interpretation and well-seismic joint prediction. The methods for studying rock pore structure can be divided into two categories: direct observation method and indirect testing method. However, no matter which method is used, it depends on core sampling and corresponding experiments, which consumes a lot of manpower and material resources. With the rapid development of artificial intelligence, the use of machine learning and logging data has provided new ideas in parameter prediction and reservoir evaluation. Summary of the invention
[0004] The purpose of the present invention is to provide a method for accurately identifying favorable reservoir areas based on random forest dual-modal output in order to solve the problem that existing porosity prediction methods all rely on core sampling and corresponding experiments, which consumes a lot of manpower and material resources.
[0005] The above-mentioned purpose of the present application is achieved through the following technical solutions: S1: Obtain core mercury injection experimental data and well logging curve data and perform preprocessing; Core mercury injection test data include: porosity, mercury injection pressure, mercury injection saturation; S2: Calculate and draw the pore throat radius distribution frequency diagram based on mercury injection pressure and mercury injection saturation; According to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram, pore throat classification is carried out to determine the pore throat type; S3: Set the well logging data as the independent variable; set the porosity as the dependent variable y1; set the pore throat type as the dependent variable y2; S4: construct a random forest bimodal output model; train the random forest bimodal output model through well logging curve data, dependent variable y1 and dependent variable y2; S5: Obtain the well logging data to be predicted; predict the porosity and pore throat type of the well logging data to be predicted through the trained random forest bimodal output model, realize the prediction of porosity and pore throat type of the whole well and the whole area, and determine the target favorable area.
[0006] Optionally, the preprocessing step includes: Check whether the core mercury injection test data and logging curve data are missing. If so, remove the data; if not, retain them.
[0007] Optionally, step S2 includes: S21: Use the national standard formula to calculate the pore radius; S22: With pore radius as the horizontal coordinate and mercury saturation difference as the vertical coordinate, draw a pore throat radius distribution frequency diagram.
[0008] Optionally, step S3 includes: the pore throat types include: type I multimodal type, type II bimodal symmetrical type, type III single-modal concentrated type and type IV bimodal asymmetric type.
[0009] Optionally, step S4 includes: S41: Train the random forest bimodal output model, draw a scatter plot of the porosity predicted by the dependent variable y1 and the actual porosity value; construct a confusion matrix of the pore throat type predicted by the dependent variable y2 and the actual pore throat type; S42: Evaluate the accuracy of the random forest bimodal output model through scatter fit plots and confusion matrices; S43: Repeatedly train the random forest bimodal output model until the accuracy reaches a preset threshold without overfitting.
[0010] A system for accurately identifying favorable reservoir areas based on random forest dual-modal output, the system comprising: a data acquisition module, a data processing module and a display module; The data acquisition module is used to acquire core mercury injection experimental data and well logging curve data and perform preprocessing; The data processing module is used to calculate and draw a pore throat radius distribution frequency diagram according to the mercury injection pressure and mercury injection saturation; The display module is used to visually display the pore throat radius distribution frequency diagram; The data processing module is also used to classify pore throats according to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram to determine the pore throat type; The data processing module is also used to set the well logging curve data as an independent variable; set the porosity as a dependent variable y1; and set the pore throat type as a dependent variable y2; The data processing module is also used to construct a random forest dual-modal output model; the random forest dual-modal output model is trained through the well logging curve data, the dependent variable y1 and the dependent variable y2; The data processing module is also used to obtain the logging curve data to be predicted; predict the porosity and pore throat type of the logging curve data to be predicted through the trained random forest bimodal output model; and determine the target favorable area through the predicted values of porosity and pore throat type.
[0011] An electronic device includes a processor, a memory, a user interface and a network interface, wherein the memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory so that the electronic device executes a method for accurately identifying favorable areas of a reservoir based on random forest bimodal output.
[0012] A computer-readable storage medium stores instructions, and when the instructions are executed, a method for accurately identifying favorable reservoir areas based on random forest bimodal output is performed.
[0013] The beneficial effects of the technical solution provided by this application are: By constructing a random forest dual-modal output model, the prediction of porosity and pore throat type is realized. Compared with the current single prediction and single output machine learning prediction model, the improved random forest dual-modal output model can achieve multi-parameter dual-modal training and prediction, improving the prediction efficiency. The superposition of porosity and pore throat structure type can more accurately and effectively identify favorable areas; the present invention takes porosity and pore throat structure type as an example, which can be replaced by TOC and kerogen type for different scenarios such as source rock evaluation or geological and engineering dual evaluation. It has extensiveness and flexibility and can be applied to different oil field exploration. BRIEF DESCRIPTION OF THE DRAWINGS
[0014] The present application will be further described below with reference to the accompanying drawings and embodiments, in which: Figure 1 is a step diagram in an embodiment of the present application; Figure 2 These are four pore throat structure diagrams in the embodiments of this application; Figure 3 is a porosity prediction scatter point fitting diagram in the embodiment of the present application; Figure 4 is a confusion matrix diagram for pore throat type identification in an embodiment of the present application; Figure 5 It is the pore porosity and pore throat type in the embodiment of the present application that comprehensively identifies favorable areas; Figure 6 It is a module structure diagram in the embodiment of the present application; Figure 7 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION
[0015] In order to have a clearer understanding of the technical features, purposes and effects of the present application, the specific implementation methods of the present application are now described in detail with reference to the accompanying drawings.
[0016] The embodiments of the present application provide a method for accurately identifying favorable reservoir areas based on random forest bimodal output.
[0017] Please refer to Figure 1 , Figure 1 This is a step diagram of a method for accurately identifying favorable reservoir areas based on random forest dual-modal output in an embodiment of the present application, including: S1: Obtain core mercury injection experimental data and well logging curve data and perform preprocessing; Core mercury injection test data include: porosity, mercury injection pressure, mercury injection saturation; Specifically, conventional logging curve data include: natural gamma (API), resistivity (Ω•m), sonic time difference (us / ft), density (g / cm3), and neutron (%).
[0018] The pre-processing steps include: Check whether the core mercury injection test data and logging curve data are missing. If so, remove the data; if not, retain them.
[0019] S2: Calculate and draw the pore throat radius distribution frequency diagram based on mercury injection pressure and mercury injection saturation; According to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram, pore throat classification is carried out to determine the pore throat type; Step S2 includes: S21: Use the national standard formula to calculate the pore radius; S22: With pore radius as the horizontal coordinate and mercury saturation difference as the vertical coordinate, draw a pore throat radius distribution frequency diagram.
[0020] Specifically, the national standard formula is: pore radius = 0.735 / mercury intrusion pressure.
[0021] S3: Set the well logging data as the independent variable; set the porosity as the dependent variable y1; set the pore throat type as the dependent variable y2; Step S3 includes: pore throat types include: type I multimodal type, type II bimodal symmetrical type, type III single-modal concentrated type and type IV bimodal asymmetric type.
[0022] Specifically, the pore throat structure in the study area can be divided into four types according to the characteristics of the pore throat radius frequency distribution curve, see Figure 3:Type I multi-peak type, Type II bimodal symmetrical type, Type III single-peak concentrated type, Type IV bimodal asymmetric type. Type I multi-peak type usually has 3 or more pore throat radius frequency peaks, the curve shape shows strong irregularity, the peak distribution band is wide, and the structure is relatively complex; Type II usually has 2 pore throat radius frequency peaks, the difference between the main peak and the secondary peak is small, the curve shape shows weak irregularity, the peak distribution band is wide, and it has the best connectivity; Type III single-peak concentrated type has only 1 pore throat radius frequency peak, the curve shape is relatively regular, and the peak distribution band is narrow; Type IV bimodal asymmetric type usually has 2 pore throat radius frequency peaks, the curve shape shows strong irregularity, the peak distribution band is wide, and it has the worst connectivity.
[0023] S4: construct a random forest bimodal output model; train the random forest bimodal output model through well logging curve data, dependent variable y1 and dependent variable y2; Step S4 includes: S41: Train the random forest bimodal output model, draw a scatter plot of the porosity predicted by the dependent variable y1 and the actual porosity value; construct a confusion matrix of the pore throat type predicted by the dependent variable y2 and the actual pore throat type; S42: Evaluate the accuracy of the random forest bimodal output model through scatter fit plots and confusion matrices; S43: Repeatedly train the random forest bimodal output model until the accuracy reaches a preset threshold without overfitting.
[0024] Specifically, a random forest dual-modal output model was compiled in Matlab software. The model establishment process is as follows: (1) Environmental preparation: clear all variables and clear the command window; (2) Data import and preprocessing: set the numerical model and category model to correspond to the required data, then randomly shuffle the order of the data set, set the proportion of the training set to 80%, and normalize the input and output data of the training set and the test set to make their values between 0 and 1; (3) Set the parameters of the random forest dual-modal output model respectively: the number of decision trees of the random forest model, the maximum depth parameter, etc.; (4) Use the trained model to predict the test set; (5) Draw a scatter fit plot of the porosity predicted by the dependent variable y1 and the true value, and draw a confusion matrix of the pore throat type predicted by the dependent variable y2 and the true category, and perform performance evaluation (see Figure 4 ).
[0025] S5: Obtain the well logging data to be predicted; predict the porosity and pore throat type of the well logging data to be predicted through the trained random forest bimodal output model, realize the prediction of porosity and pore throat type of the whole well and the whole area, and determine the target favorable area.
[0026] Specifically, Figure 5As shown in the figure, the high porosity section with the dominant pore throat type is the target favorable area. The model accuracy is clarified by the porosity scatter fitting diagram and the pore throat type confusion matrix diagram. After the training of the first generation model is completed, the porosity and pore throat type of the whole well and the whole area are predicted through the logging curve data to determine the target favorable area.
[0027] Please refer to Figure 6 , Figure 6 This is a module structure diagram of a system for accurately identifying favorable reservoir areas based on random forest dual-modal output in an embodiment of the present application. The system includes: a data acquisition module, a data processing module, and a display module; The data acquisition module is used to acquire core mercury injection experimental data and well logging curve data and perform preprocessing; The data processing module is used to calculate and draw a pore throat radius distribution frequency diagram according to the mercury injection pressure and mercury injection saturation; The display module is used to visually display the pore throat radius distribution frequency diagram; The data processing module is also used to classify pore throats according to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram to determine the pore throat type; The data processing module is also used to set the well logging curve data as an independent variable; set the porosity as a dependent variable y1; and set the pore throat type as a dependent variable y2; The data processing module is also used to construct a random forest dual-modal output model; the random forest dual-modal output model is trained through the well logging curve data, the dependent variable y1 and the dependent variable y2; The data processing module is also used to obtain the logging curve data to be predicted; predict the porosity and pore throat type of the logging curve data to be predicted through the trained random forest bimodal output model; and determine the target favorable area through the predicted values of porosity and pore throat type.
[0028] The present application also discloses an electronic device. Figure 7 , Figure 7 The electronic device 500 may include: at least one processor 501 , at least one network interface 504 , a user interface 503 , a memory 505 , and at least one communication bus 502 .
[0029] The communication bus 502 is used to realize the connection and communication between these components.
[0030] The user interface 503 may include a display screen, and the optional user interface 503 may also include a standard wired interface or a wireless interface.
[0031] The network interface 504 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).
[0032] The present application also discloses a computer-readable storage medium storing a plurality of instructions suitable for loading by a processor to execute the above-mentioned method for accurately identifying favorable reservoir areas based on random forest dual-modal output.
[0033] The above are only exemplary embodiments of the present disclosure and cannot be used to limit the scope of the present disclosure. That is, any equivalent changes and modifications made according to the teachings of the present disclosure are still within the scope of the present disclosure. After considering the disclosure of the specification and the truth of practice, those skilled in the art will easily think of other embodiments of the present disclosure.
[0034] This application is intended to cover any variation, use or adaptation of the present disclosure, which follows the general principles of the present disclosure and includes common knowledge or customary technical means in the art not described in the present disclosure. The description and examples are to be regarded as exemplary only, and the scope and spirit of the present disclosure are defined by the claims.
Claims
1. A method for accurately identifying favorable reservoir areas based on random forest dual-modal output, characterized in that: The method comprises the following steps: S1: Obtain core mercury injection experimental data and well logging curve data and perform preprocessing; Core mercury injection test data include: porosity, mercury injection pressure, mercury injection saturation; S2: Calculate and draw the pore throat radius distribution frequency diagram based on mercury injection pressure and mercury injection saturation; According to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram, pore throat classification is carried out to determine the pore throat type; S3: Set the well logging data as the independent variable; set the porosity as the dependent variable y1; set the pore throat type as the dependent variable y2; S4: construct a random forest bimodal output model; train the random forest bimodal output model through well logging curve data, dependent variable y1 and dependent variable y2; S5: Obtain the well logging data to be predicted; predict the porosity and pore throat type of the well logging data to be predicted through the trained random forest bimodal output model, realize the prediction of porosity and pore throat type of the whole well and the whole area, and determine the target favorable area.
2. The method for accurately identifying favorable reservoir areas based on random forest dual-modal output according to claim 1, characterized in that: The pre-processing steps include: Check whether the core mercury injection test data and logging curve data are missing. If so, remove the data; if not, retain them.
3. The method for accurately identifying favorable reservoir areas based on random forest dual-modal output according to claim 1, characterized in that: Step S2 includes: S21: Use the national standard formula to calculate the pore radius; S22: With pore radius as the horizontal coordinate and mercury saturation difference as the vertical coordinate, draw a pore throat radius distribution frequency diagram.
4. The method for accurately identifying favorable reservoir areas based on random forest dual-modal output according to claim 1, characterized in that: Step S3 includes: pore throat types include: type I multimodal type, type II bimodal symmetrical type, type III single-modal concentrated type and type IV bimodal asymmetric type.
5. The method for accurately identifying favorable reservoir areas based on random forest dual-modal output according to claim 1, characterized in that: Step S4 includes: S41: Train the random forest bimodal output model, draw a scatter plot of the porosity predicted by the dependent variable y1 and the actual porosity value; construct a confusion matrix of the pore throat type predicted by the dependent variable y2 and the actual pore throat type; S42: Evaluate the accuracy of the random forest bimodal output model through scatter fit plots and confusion matrices; S43: Repeatedly train the random forest bimodal output model until the accuracy reaches a preset threshold without overfitting.
6. A system for accurately identifying favorable reservoir areas based on random forest dual-modal output, characterized in that: The system comprises: a data acquisition module, a data processing module and a display module; The data acquisition module is used to acquire core mercury injection experimental data and well logging curve data and perform preprocessing; The data processing module is used to calculate and draw a pore throat radius distribution frequency diagram according to the mercury injection pressure and mercury injection saturation; The display module is used to visually display the pore throat radius distribution frequency diagram; The data processing module is also used to classify pore throats according to the frequency distribution curve characteristics of the pore throat radius distribution frequency diagram to determine the pore throat type; The data processing module is also used to set the well logging curve data as an independent variable; set the porosity as a dependent variable y1; and set the pore throat type as a dependent variable y2; The data processing module is also used to construct a random forest bimodal output model; the random forest bimodal output model is trained through the well logging curve data, the dependent variable y1 and the dependent variable y2; The data processing module is also used to obtain the logging curve data to be predicted; predict the porosity and pore throat type of the logging curve data to be predicted through the trained random forest bimodal output model; and determine the target favorable area through the predicted values of porosity and pore throat type.
7. An electronic device, characterized in that: The electronic device comprises a processor (501), a memory (505), a user interface (503) and a network interface (504), wherein the memory (505) is used to store instructions, the user interface (503) and the network interface (504) are used to communicate with other devices, and the processor (501) is used to execute the instructions stored in the memory (505) so that the electronic device executes the method according to any one of claims 1 to 5.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores instructions, and when the instructions are executed by a computer, the method according to any one of claims 1 to 5 is executed.