A method and apparatus for predicting clastic reservoir flow unit type

Abstract

The application provides a method for predicting the flow unit type of clastic rock reservoirs, calculates a reservoir flow unit index FZI according to core measured data, and further obtains a cumulative probability distribution diagram, so that the reservoir flow unit is divided into three categories to obtain first logging data; the first logging data is taken as an independent variable, the three types of reservoir flow units are taken as dependent variables, abnormal values in the data are removed, the independent variable and the dependent variable data are subjected to crossplot analysis, and logging parameters for distinguishing the three types of flow units are screened out as second logging data; according to the second logging data, a training set and a test set are divided, the training set is used to train a random forest model; a confusion matrix is drawn to calculate the accuracy rate, the test set is used to evaluate the model, and the optimal random forest model is obtained to predict the actual clastic rock reservoir flow unit type. The application has the beneficial effect that the reservoir flow unit type of the uncored section is effectively predicted, the reservoir characteristics are clear, and a foundation is laid for the next exploration and development.

Description

Technical Field

[0001] This invention relates to the field of predicting flow unit types in clastic rock reservoirs, and more particularly to a method and apparatus for predicting flow unit types in clastic rock reservoirs. Background Technology

[0002] The study of flow units is of significant theoretical and practical value for further subdividing deep / ultra-deep reservoirs, predicting their distribution and properties, improving the accuracy of permeability interpretation and reservoir numerical simulation, and revealing the distribution patterns of remaining oil. Currently, there are three main categories of methods for delineating flow units: 1. Outcrop sedimentary interface research methods; 2. Flow unit delineation based primarily on geological research: a. Sedimentary facies method, b. Reservoir hierarchical analysis method, c. Heterogeneous composite index method; 3. Flow unit delineation based primarily on mathematical methods: a. Flow stratification index (FZI) method, b. Pore throat geometry (R35) method, c. Multi-parameter comprehensive method, d. Production dynamic parameter method, etc. Deep / ultra-deep reservoirs are highly heterogeneous, and the results of the above methods often only characterize the flow unit features of the cored section. Therefore, previous researchers have often used discriminant analysis and probabilistic statistical methods to predict the flow unit type of the uncored section, but the accuracy needs improvement. Summary of the Invention

[0003] To more accurately and quickly predict reservoir flow characteristics in uncored sections, this invention provides a method for predicting flow unit types in clastic reservoirs, mainly comprising the following steps:

[0004] S1: Collect core measured data and raw logging data. The core measured data includes porosity and permeability. Calculate the reservoir flow unit index FZI based on the core measured data. Obtain the cumulative probability distribution map based on the reservoir flow unit index FZI. Divide the reservoir flow units into 3 categories. Match the raw logging data and reservoir flow units according to the depth relationship to obtain the corresponding first logging data.

[0005] S2: Using the first logging data as the independent variable and the three types of reservoir flow units as the dependent variable, outliers in the data are removed, and the independent and dependent variable data are analyzed by cross-plot analysis to select logging parameters that can distinguish the three types of flow units as the second logging data.

[0006] S3: Based on the second well logging data, divide the data into a training set and a test set, and use the training set and test set to train and validate the established random forest model;

[0007] S4: Based on the training and validation results, a confusion matrix is ​​plotted to calculate the accuracy, and the model is evaluated to obtain the optimal random forest model, which is used to predict the flow unit type of actual clastic reservoirs.

[0008] Furthermore, in step S1, the formula for calculating the reservoir flow unit index FZI is as follows:

[0009]

[0010]

[0011]

[0012] In the formula: FZI is the flow unit index, RQI is the reservoir quality index, and k is the permeability. Porosity It is the ratio of pore volume to particle volume.

[0013] Furthermore, the second logging data includes neutron CNL, borehole caliber CAL, sonic transit time DT, natural gamma ray GR, density DEN, shallow lateral resistance RS, deep lateral resistance RD, and spontaneous potential SP.

[0014] Furthermore, in step S3, when building the random forest model, the number of decision trees is adjusted, the minimum number of samples for each leaf node is set, and OOB error prediction is enabled to calculate the accuracy of the random forest model.

[0015] Furthermore, the three types of reservoir flow units are as follows: Type I reservoirs have a flow stratification index of less than 0.1, an average porosity of 6.80%, and an average permeability of 0.252 mD, belonging to "low-porosity and ultra-low-permeability" reservoirs; Type II reservoirs have a flow stratification index of 0.1-0.5, an average porosity of 7.65%, and an average permeability of 7.135 mD, belonging to "low-porosity and low-permeability" reservoirs; and Type III reservoirs have a flow unit index > 0.5, an average porosity of 6.23%, and an average permeability of 14.832 mD, belonging to "low-porosity and medium-permeability" reservoirs.

[0016] A storage device that stores instructions and data for implementing a method for predicting the flow unit type of clastic rock reservoirs.

[0017] An apparatus for predicting the flow unit type of a clastic reservoir includes: a processor and a storage device; the processor loads and executes instructions and data in the storage device to implement a method for predicting the flow unit type of a clastic reservoir.

[0018] The beneficial effects of the technical solution provided by this invention are as follows: Based on statistical principles, this invention collects core thin section data and original well logging data to obtain core measured data, including porosity and permeability. The reservoir flow unit index (FZI) is calculated based on the core measured data, and the flow unit types are classified using the cumulative probability distribution map of the calculated FZI. Using well logging data as the independent variable and three types of reservoir flow units as the dependent variable, outliers are removed from the data. Correlation analysis is performed on the independent and dependent variables, and combined with machine learning methods, an optimal random forest model is obtained to predict the flow unit types of actual clastic reservoirs. This achieves the goal of quickly predicting the flow unit types of reservoirs in uncorked sections, clarifying reservoir characteristics, and laying the foundation for further exploration and development. Furthermore, compared to traditional methods for predicting reservoir flow in uncorked sections, this method is more accurate and faster. Attached Figure Description

[0019] The present invention will be further described below with reference to the accompanying drawings and embodiments. In the accompanying drawings:

[0020] Figure 1 This is a flowchart of a method for predicting the flow unit type of clastic reservoirs according to an embodiment of the present invention.

[0021] Figure 2 This is a cumulative probability distribution diagram in an embodiment of the present invention.

[0022] Figure 3 This is a comparison chart of the predicted values ​​and the true values ​​obtained using the training set and test set in an embodiment of the present invention.

[0023] Figure 4 This is a distribution diagram of flow units throughout the entire well section in an embodiment of the present invention.

[0024] Figure 5 This is a schematic diagram of the hardware device working in an embodiment of the present invention. Detailed Implementation

[0025] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0026] Embodiments of the present invention provide a method and apparatus for predicting the flow unit type of clastic rock reservoirs.

[0027] Please refer to Figure 1 , Figure 1 This is a flowchart of a method for predicting the flow unit type of clastic reservoirs according to an embodiment of the present invention, specifically including:

[0028] S1: Collect core measured data and raw logging data. Calculate the reservoir flow unit index (FZI) based on the core measured data (porosity, permeability), and classify the flow units into 3 categories according to the cumulative probability distribution map. Match the raw logging data and reservoir flow units according to depth relationships, and organize the corresponding logging data (i.e., the first logging data), as shown in Table 1:

[0029] Table 1

[0030]

[0031] S2: Using the first logging data as the independent variable and the three types of reservoir flow units as the dependent variable, outliers in the data are removed. For example, if a series of data are all around 2.23, and suddenly a data of 9999 appears, this 9999 is an abnormal value caused by logging. The independent variable and dependent variable data are analyzed by cross-plot analysis, and the logging parameters that can distinguish the three types of flow units are selected as the second logging data.

[0032] S3: Use MATLAB to build a random forest model. Based on the second well logging data, divide the data into training and test sets, import the data, and train the model.

[0033] S4: Draw a confusion matrix based on the training and validation results, calculate the accuracy, evaluate the model, and train the optimal model.

[0034] In step S1: Collect and organize core measurement data, including porosity and permeability. The flow unit index (FZI) is calculated using the following formula.

[0035]

[0036]

[0037]

[0038] In the formula: FZI is the flow unit index, dimensionless; RQI is the reservoir quality index, dimensionless; k is the permeability, mD. Porosity, %; It is the ratio of pore volume to particle volume.

[0039] Based on statistical principles, the calculated flow unit index FZI is used to create... Figure 2The cumulative probability distribution diagram shown is illustrated. It can be seen that there are multiple curves with different slopes in the diagram. The intersection of these curves is used as the boundary value for dividing the flow units, resulting in three flow units. The specific parameters of the three types of flow units are shown in Table 1. Type I reservoirs have a flow stratification index less than 0.1, an average porosity of 6.80%, and an average permeability of 0.252 mD, classifying them as "low-porosity, ultra-low-permeability" reservoirs. Type II reservoirs have a flow stratification index between 0.1 and 0.5, an average porosity of 7.65%, and an average permeability of 7.135 mD, classifying them as "low-porosity, low-permeability" reservoirs. Type III reservoirs have a flow unit index greater than 0.5, an average porosity of 6.23%, and an average permeability of 14.832 mD, classifying them as "low-porosity, medium-permeability" reservoirs. Type III reservoirs exhibit well-developed microfractures.

[0040] In step S2: Cross-plot analysis is performed on the independent and dependent variable data to select logging parameters that can better distinguish the three types of flow units as the second logging data. Specific parameters include, but are not limited to: neutron CNL, wellbore CAL, sonic transit time DT, natural gamma ray GR, density DEN, shallow lateral resistance RS, deep lateral resistance RD, spontaneous potential SP, etc.

[0041] In step S3: 1. First, clear the environment variables and import the data. 2. Divide the flow unit type and the corresponding second logging data into training and test sets: randomly shuffle the data, take the first sample as the training set, and the remaining samples as the test set; normalize the input and output data of the training and test sets so that their values ​​are between 0 and 1. 3. Establish the random forest model: adjust the number of decision trees, the minimum number of samples per leaf node, and enable OOB error prediction to calculate the accuracy of the random forest model. 4. Train the random forest model: open the neural network training window in MATLAB, train the neural network using the training set, and then use the trained neural network to predict on the training and test sets. Sort the data for plotting.

[0042] In step S4: the confusion matrix is ​​plotted, the accuracy is calculated, the random forest model is evaluated, and the optimal random forest model is trained. In this embodiment, the optimal model is one where the classification accuracy of both the training set and the test set is greater than 90%. Figure 3 In the training set, the accuracy rate was 93.33%, and the test set accuracy rate was 96.00%, indicating that the optimal random forest model had a good fit and high accuracy. Using the optimal random forest model for prediction yielded the following results: Figure 4 The distribution of flow units throughout the well section shown effectively predicts the reservoir flow characteristics in the section without coring.

[0043] Please see Figure 5 , Figure 5This is a schematic diagram of the hardware device in operation according to an embodiment of the present invention. The hardware device specifically includes: a device 401 for predicting the flow unit type of clastic rock reservoirs, a processor 402, and a storage device 403.

[0044] A device 401 for predicting the flow unit type of clastic reservoirs: The device 401 for predicting the flow unit type of clastic reservoirs implements the method for predicting the flow unit type of clastic reservoirs.

[0045] Processor 402: The processor 402 loads and executes the instructions and data in the storage device 403 to implement the method for predicting the flow unit type of clastic rock reservoirs.

[0046] Storage device 403: The storage device 403 stores instructions and data; the storage device 403 is used to implement the method for predicting the flow unit type of clastic rock reservoir.

[0047] The beneficial effects of this invention are as follows: Based on statistical principles, this invention collects core thin section data and raw well logging data to obtain core measured data, including porosity and permeability. The reservoir flow unit index (FZI) is calculated based on the core measured data, and the flow unit types are classified using the cumulative probability distribution map of the calculated FZI. Using well logging data as the independent variable and three types of reservoir flow units as the dependent variable, outliers are removed from the data. Correlation analysis is performed on the independent and dependent variables, and combined with machine learning methods, an optimal random forest model is obtained to predict the flow unit types of actual clastic reservoirs. This achieves the goal of quickly predicting the flow unit types of reservoirs in uncorked sections, clarifying reservoir characteristics, and laying the foundation for further exploration and development. Furthermore, compared to traditional methods for predicting reservoir flow in uncorked sections, this method is more accurate and faster.

[0048] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for predicting the flow unit type of clastic rock reservoirs, characterized in that: include: S1: Collect core measured data and raw logging data. The core measured data includes porosity and permeability. Calculate the reservoir flow unit index FZI based on the core measured data. Obtain the cumulative probability distribution map based on the reservoir flow unit index FZI. Divide the reservoir flow units into 3 categories. Match the raw logging data and reservoir flow units according to the depth relationship to obtain the corresponding first logging data. S2: Using the first logging data as the independent variable and the three types of reservoir flow units as the dependent variable, outliers in the data are removed, and the independent and dependent variable data are analyzed by cross-plot analysis to select logging parameters that can distinguish the three types of flow units as the second logging data. S3: Based on the second well logging data, divide the data into a training set and a test set, and use the training set and test set to train and validate the established random forest model; S4: Based on the training and validation results, a confusion matrix is ​​plotted to calculate the accuracy, and the model is evaluated to obtain the optimal random forest model, which is used to predict the flow unit type of actual clastic reservoirs.

2. The method for predicting the flow unit type of clastic reservoirs as described in claim 1, characterized in that: In step S1, the formula for calculating the reservoir flow unit index FZI is as follows: In the formula: FZI is the flow unit index, RQI is the reservoir quality index, and k is the permeability. Porosity It is the ratio of pore volume to particle volume.

3. The method for predicting the flow unit type of clastic reservoirs as described in claim 1, characterized in that: The second logging data includes neutron CNL, borehole caliber CAL, sonic transit time DT, natural gamma ray GR, density DEN, shallow lateral resistance RS, deep lateral resistance RD, and spontaneous potential SP.

4. The method for predicting the flow unit type of clastic reservoirs as described in claim 1, characterized in that: In step S3, when building the random forest model, the number of decision trees is adjusted, the minimum number of samples for each leaf node is set, and OOB error prediction is enabled to calculate the accuracy of the random forest model.

5. The method for predicting the flow unit type of clastic reservoirs as described in claim 1, characterized in that: The three types of reservoir flow units are as follows: Type I reservoirs have a flow stratification index of less than 0.1, an average porosity of 6.80%, and an average permeability of 0.252 mD, belonging to the "low porosity and ultra-low permeability" reservoirs; Type II reservoirs have a flow stratification index of 0.1-0.5, an average porosity of 7.65%, and an average permeability of 7.135 mD, classifying them as "low-porosity and low-permeability" reservoirs. Type III reservoirs have a flow stratification index >0.5, an average porosity of 6.23%, and an average permeability of 14.832 mD, classifying them as "low-porosity and medium-permeability" reservoirs.

6. A storage device, characterized in that: The storage device stores instructions and data for implementing the method for predicting the flow unit type of clastic reservoirs as described in any one of claims 1 to 5.

7. A device for predicting the flow unit type of clastic rock reservoirs, characterized in that: include: A processor and a storage device; the processor loads and executes instructions and data in the storage device to implement the method for predicting the flow unit type of clastic reservoirs as described in any one of claims 1 to 5.

Images