Source rock-reservoir-project multi-parameter comprehensive prediction method based on random forest

Abstract

The invention provides a hydrocarbon source rock-reservoir-engineering multi-parameter comprehensive prediction method based on a random forest, and relates to the field of geological exploration, the method comprises the following steps: collecting rock core experiment data, the rock core experiment data comprising free hydrocarbon content, movable porosity and brittleness index; logging information is obtained; constructing a data set through logging information and rock core experiment data; establishing a three-parameter prediction model based on a random forest algorithm; and through the trained three-parameter prediction model, completing three-parameter prediction from a single well to a research area and sweet spot identification in an oil and gas reservoir based on logging data. By improving a random forest algorithm, a three-parameter prediction model of various different input parameters and various different output parameters is established. Multi-parameter prediction from a single well to a research area is achieved, and whole-area dessert identification and evaluation are achieved according to a dessert classification evaluation table of the research area.

Description

Technical Field

[0001] The present application relates to the field of geological exploration, and in particular to a random forest-based comprehensive prediction method for source rock-reservoir-engineering multi-parameters. Background Art

[0002] With the increasing demand for oil and gas resources and the depletion of conventional oil and gas, oil and gas exploration has gradually moved into unconventional areas. Exploration practice has confirmed that China's continental shale oil resources are very rich, and a breakthrough has been achieved from quantitative change to qualitative change. However, shale reservoirs are highly dense, have high clay mineral content, low reservoir brittleness, and complex lithology, resulting in low shale oil recovery and little benefit.

[0003] The concept of "sweet spot" is widely used in unconventional oil and gas exploration and development. The selection of high-quality "sweet spots" is the basis for economic and effective development. Finding sweet spots is the key to shale oil exploration and evaluation and achieving economic and profitable development. Shale oil sweet spot prediction usually includes three aspects: source rock quality evaluation, reservoir quality evaluation and engineering quality evaluation. The evaluation process mainly relies on a large amount of experimental data, which is costly and inefficient.

[0004] At present, the use of machine learning and logging data has provided new ideas for sweet spot prediction. However, most of them are single-parameter predictions, involving multiple parameters and the need to establish multiple different models, which is relatively cumbersome. Although the existing BP multi-output neural network can predict multiple parameters, the premise is that the independent variables (logging parameters) corresponding to each parameter are the same. In areas with a long exploration history, the number of coring times and the long time span lead to complex and diverse samples (different experiments and different corings have different corresponding logging parameters). The traditional BP multi-output neural network model is difficult to adapt, and the model establishment process is cumbersome and inefficient. Summary of the invention

[0005] The purpose of the present invention is to provide a source rock-reservoir-engineering multi-parameter comprehensive prediction method based on random forest in order to solve the problem that it is cumbersome and inefficient to establish multiple models for different experiments and different corings and corresponding logging parameters.

[0006] The above-mentioned purpose of the present application is achieved through the following technical solutions: S1: Collect core test data, including free hydrocarbon content, movable porosity and brittleness index; obtain logging data corresponding to free hydrocarbon content , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; S2: construct a data set through well logging data and core experiment data; S3: Based on the random forest algorithm, a three-parameter prediction model is established; the three-parameter prediction model is trained through the data set; S4: Obtain the logging data to be predicted; through the trained three-parameter prediction model, complete the three-parameter prediction from the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the logging data.

[0007] Optionally, step S2 includes: The well logging data of the data set are defined as independent variables; the free hydrocarbon content, movable porosity and brittleness index of the data set are defined as dependent variables.

[0008] Optionally, step S2 further includes: The independent variables and dependent variables of the data set are normalized respectively; the data of the data set is divided into three groups of data sets, each of which contains a group of independent variables and their corresponding dependent variables.

[0009] Optionally, step S3 includes: Using the TreeBagger function, combined with the three sets of data in the dataset, three random forest regression models are trained respectively to obtain the trained three-parameter prediction model; each random forest regression model uses a different set of independent variables and their corresponding dependent variables.

[0010] Optionally, step S3 further includes: Set the number of trees in the random forest, the minimum number of samples required per leaf node, and the maximum depth of the trees.

[0011] A source rock-reservoir-engineering multi-parameter comprehensive prediction system based on random forest, the system comprising: a data acquisition module, a data processing module, and a display module; The data acquisition module is used to collect core test data, which includes: free hydrocarbon content, movable porosity and brittleness index; The data acquisition module is also used to obtain logging data corresponding to the free hydrocarbon content. , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; The data processing module is used to construct a data set through well logging data and core experiment data; The data processing module is also used to establish a three-parameter prediction model based on a random forest algorithm; and train the three-parameter prediction model through a data set; The data acquisition module is also used to acquire the well logging data to be predicted; The data processing module is also used to complete the three-parameter prediction of the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the well logging data through the trained three-parameter prediction model; The display module is used to visually display the dessert recognition result.

[0012] An electronic device comprises 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 random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method.

[0013] A computer-readable storage medium stores instructions, and when the instructions are executed, a random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method is executed.

[0014] The beneficial effects of the technical solution provided by this application are: At present, most of the machine learning prediction parameters are single prediction and single output. The improved random forest model can train different independent variables to predict multiple dependent variables, thereby improving prediction efficiency. By combining the three quality parameters of source rock, reservoir and engineering (free hydrocarbon content, movable porosity and brittleness index) for comprehensive prediction, it is possible to more comprehensively and accurately evaluate and identify sweet spots, and this technology is convenient and flexible. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] 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 is a three-parameter model prediction error diagram in an embodiment of the present application; Figure 3 It is a prediction result diagram in the embodiment of the present application; Figure 4 It is a schematic diagram of the structure of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION

[0016] 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.

[0017] The embodiments of the present application provide a random forest-based comprehensive prediction method for source rock-reservoir-engineering multi-parameters.

[0018] Please refer to Figure 1 , Figure 1 This is a step diagram of a random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method in an embodiment of the present application, including: S1: Collect core test data, including free hydrocarbon content, movable porosity and brittleness index; obtain logging data corresponding to free hydrocarbon content , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; The present application provides an example as follows: collecting core experimental data includes three aspects: Y-source rock quality (free hydrocarbon content-S1), Reservoir quality (mobile porosity - Φ) and Engineering quality (brittleness index - BI) and the corresponding logging data for each quality , , (Gamma-GR, Resistance-RT, Acoustic Time-Difference-AC, Density-DEN, Neutron-CNL), check whether the data parameters are missing. If a data parameter is missing, remove the sample data. If not, keep it and proceed to the next step.

[0019] S2: construct a data set through well logging data and core experiment data; Step S2 includes: The well logging data of the data set are defined as independent variables; the free hydrocarbon content, movable porosity and brittleness index of the data set are defined as dependent variables.

[0020] The independent variables and dependent variables of the data set are normalized respectively; the data of the data set is divided into three groups of data sets, each of which contains a group of independent variables and their corresponding dependent variables.

[0021] S3: Based on the random forest algorithm, a three-parameter prediction model is established; the three-parameter prediction model is trained through the data set; The present application provides an embodiment as follows, which draws a scatter point fitting diagram of the predicted value and the true value obtained through the data set, such as Figure 2 As shown, and calculated , train the model multiple times until the model reaches the preset accuracy.

[0022] Step S3 includes: Using the TreeBagger function, combined with the three sets of data in the dataset, three random forest regression models are trained respectively to obtain the trained three-parameter prediction model; each random forest regression model uses a different set of independent variables and their corresponding dependent variables.

[0023] Step S3 also includes: Set the number of trees in the random forest, the minimum number of samples required per leaf node, and the maximum depth of the trees.

[0024] S4: Obtain the logging data to be predicted; through the trained three-parameter prediction model, complete the three-parameter prediction from the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the logging data.

[0025] Step S4 includes: The prediction results of the core experimental data to be predicted are denormalized; the sweet spots in the oil and gas reservoir are identified through the denormalized prediction results.

[0026] This application provides an embodiment as follows: Figure 3 As shown in the figure, the core experimental data to be predicted is predicted by the trained three-parameter prediction model, and the prediction results are denormalized. Draw a scatter plot: Draw a scatter plot of the training set and the test set for each independent variable group, and show the relationship between the true value and the predicted value. Calculate and display Value: Calculate the performance of each model on the training and test sets. Value. Adjust the display range of the graphics. Set the horizontal and vertical coordinate ranges for all drawn graphics and modify the image details.

[0027] The present application provides an embodiment as follows: a three-parameter prediction model uses random forest regression (TreeBagger) for multi-parameter prediction, and trains and tests three sets of data (each set has five independent variables and one target variable). It divides the training set and the test set, standardizes the input data, and then trains three different regression models, and finally evaluates their prediction capabilities.

[0028] A source rock-reservoir-engineering multi-parameter comprehensive prediction system based on random forest, the system comprising: a data acquisition module, a data processing module, and a display module; The data acquisition module is used to collect core test data, which includes: free hydrocarbon content, movable porosity and brittleness index; The data acquisition module is also used to obtain logging data corresponding to the free hydrocarbon content. , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; The data processing module is used to construct a data set through well logging data and core experiment data; The data processing module is also used to establish a three-parameter prediction model based on a random forest algorithm; and train the three-parameter prediction model through a data set; The data acquisition module is also used to acquire the well logging data to be predicted; The data processing module is also used to complete the three-parameter prediction of the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the well logging data through the trained three-parameter prediction model; The display module is used to visually display the dessert recognition result.

[0029] The three-parameter prediction model includes the following modules: 1. Data import and preprocessing module: read data from Excel files and randomly divide them into training and test sets. Then, define training and test sets for three sets of independent variables and target variables respectively.

[0030] 2. Data normalization module: Use the mapminmax function to normalize the data to the range of [0,1] to ensure that data of different scales can be effectively processed.

[0031] 3. Regression model training module: Use the random forest algorithm to train three regression models, each of which uses five input variables to predict one target variable.

[0032] 4. Model simulation and prediction module: 5. Result visualization module: Draw a scatter plot of the true value and the predicted value, and calculate the R² value of the training set and the test set to evaluate the fitting effect of the model. Output the scatter plot and the R² value of the model in the training set and the test set.

[0033] 6. Model evaluation and output module: Display the R² value of the training set and test set in the command window as a quantitative indicator of the model effect.

[0034] The present application provides an embodiment as follows: a comprehensive prediction of multiple parameters of source rock, reservoir and engineering based on random forest. It can simultaneously predict the quality of source rock, reservoir and engineering, and achieve efficient and accurate sweet spot evaluation. By introducing the random forest algorithm, this method can handle complex data structures with multiple inputs and multiple outputs, avoiding the trouble of establishing multiple models separately in traditional methods, as well as the problem that different experimental data correspond to different logging parameters, greatly improving the adaptability and prediction accuracy of the model, and is particularly suitable for areas with long exploration history and frequent coring experiments.

[0035] The present application provides an embodiment as follows: Multi-parameter comprehensive prediction: It realizes simultaneous prediction and output of multiple parameters, which greatly improves efficiency compared with conventional machine learning prediction methods; Comprehensive sweet spot evaluation: This method integrates three types of parameters, namely source rock, geology and engineering, for comprehensive prediction, which can more comprehensively evaluate and identify sweet spots in oil and gas reservoirs and improve the accuracy of exploration; Adapt to complex coring experimental conditions: The model can adapt to the problem that different experimental data correspond to different logging parameters, make full use of new and old data, avoid waste of resources, and is of great significance for old well review and secondary exploration. Wide adaptability: This method has strong adaptability, and the selection of parameters should be based on the actual needs of exploration, and can be flexibly adjusted, providing reliable support for sweet spot evaluation under various complex geological environments, and can be applied to different oil field explorations (conventional oil and gas, tight sandstone, shale oil and gas, coalbed methane and other oil and gas reservoirs).

[0036] The present application provides an embodiment as follows. Table 1 is a dessert classification evaluation table divided according to the actual situation of the study area.

[0037] Table 1

[0038] The present application also discloses an electronic device. Figure 4 , Figure 4 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 .

[0039] The communication bus 502 is used to realize the connection and communication between these components.

[0040] 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.

[0041] The network interface 504 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0042] 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 random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method.

[0043] 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.

[0044] 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 random forest-based comprehensive prediction method for source rock-reservoir-engineering multi-parameters, characterized in that: The method comprises the following steps: S1: Collect core test data, including free hydrocarbon content, movable porosity and brittleness index; obtain logging data corresponding to free hydrocarbon content , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; S2: construct a data set through well logging data and core experiment data; S3: Based on the random forest algorithm, a three-parameter prediction model is established; the three-parameter prediction model is trained through the data set; S4: Obtain the logging data to be predicted; through the trained three-parameter prediction model, complete the three-parameter prediction from the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the logging data.

2. The random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method according to claim 1, characterized in that: Step S2 includes: The well logging data of the data set are defined as independent variables; the free hydrocarbon content, movable porosity and brittleness index of the data set are defined as dependent variables.

3. The random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method according to claim 2, characterized in that: Step S2 also includes: The independent variables and dependent variables of the data set are normalized respectively; the data of the data set is divided into three groups of data sets, each of which contains a group of independent variables and their corresponding dependent variables.

4. The random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method according to claim 3, characterized in that: Step S3 includes: Using the TreeBagger function, combined with the three sets of data in the dataset, three random forest regression models are trained respectively to obtain the trained three-parameter prediction model; each random forest regression model uses a different set of independent variables and their corresponding dependent variables.

5. The random forest-based source rock-reservoir-engineering multi-parameter comprehensive prediction method according to claim 4, characterized in that: Step S3 also includes: Set the number of trees in the random forest, the minimum number of samples required per leaf node, and the maximum depth of the trees.

6. A source rock-reservoir-engineering multi-parameter comprehensive prediction system based on random forest, 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 collect core test data, which includes: free hydrocarbon content, movable porosity and brittleness index; The data acquisition module is also used to obtain logging data corresponding to the free hydrocarbon content. , Well logging data corresponding to movable porosity And the logging data corresponding to the brittleness index ; The data processing module is used to construct a data set through well logging data and core experiment data; The data processing module is also used to establish a three-parameter prediction model based on a random forest algorithm; and train the three-parameter prediction model through a data set; The data acquisition module is also used to acquire the well logging data to be predicted; The data processing module is also used to complete the three-parameter prediction of the entire well section of a single well to the study area and the identification of sweet spots in the oil and gas reservoir based on the well logging data through the trained three-parameter prediction model; The display module is used to visually display the dessert recognition result.

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.

Images