Method and system for predicting effective leakage volume of shale oil
By constructing a shale oil effective oil leakage volume data set and using the random forest model for prediction, the limitations of the shale oil effective oil leakage volume prediction accuracy and calculation efficiency in the existing technology are solved, and more accurate shale oil development is achieved. Guidance and recovery enhancement.
Patent Information
- Application Number
- CN202310902165.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2023-07-21
- Publication Date
- 2025-01-21
AI Technical Summary
The prior art has limitations in predicting the effective oil discharge volume of shale oil, especially in complex shale formations with heterogeneous rocks and multiphase fluid flow, which are difficult to accurately predict.
By constructing a shale oil effective oil leakage volume data set, using machine learning models, especially the random forest method, combined with numerical simulation and production transient analysis, hyperparameters are optimized to establish a correlation model, and then predict the effective oil leakage volume.
The prediction accuracy of the effective oil drainage volume of shale oil is improved, which can guide the development of shale oil more accurately, optimize the fracturing plan and improve the recovery rate.
Smart Images
Figure CN119338035A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of unconventional oil and gas exploitation, and in particular to a method and system for predicting the effective leakage volume of shale oil. Background Art
[0002] With the increasing demand for oil and gas resources and the continuous reduction of conventional oil and gas resources, the exploration of shale oil resources is of great significance for the country to find alternative energy sources. Shale oil is an important part of my country's current resource development. It has huge resources, low porosity, poor permeability and other characteristics. Therefore, hydraulic fracturing is generally used on site to develop shale oil reservoirs. Affected by the heterogeneity of the shale reservoir itself, the reservoir will form an extremely complex fracture network system after fracturing. The effective oil leakage volume after fracturing is an important parameter for fracturing evaluation. However, field monitoring and production dynamic analysis show that most of the fracturing transformation volume does not contribute to the production of shale oil wells. The fracturing transformation volume is different from the actual effective oil leakage volume of shale oil. Only by accurately understanding the effective oil leakage volume of shale oil and its influencing factors and laws can we more effectively guide shale oil development.
[0003] The prediction of effective stimulation volume is a key step in shale oil production, as it determines the amount of fracturing fluid and proppant required for fracturing, which directly affects production and recovery. Currently, conventional effective drainage volume prediction methods, such as empirical correlation and numerical simulation, have certain limitations in accuracy and computational efficiency, and may not be applicable to complex shale formations with heterogeneous shales and multiphase fluid flow.
[0004] Although some researchers have used machine learning methods to predict the effective transformation volume of shale oil, there are still certain challenges and limitations in applying machine learning methods to effectively predict the transformation volume of shale oil. One of the main challenges is the limited availability of high-quality training data, as data collection in the oil and gas industry is usually time-consuming and expensive. In addition, data quality is often affected by various factors, such as measurement errors, missing values, and heterogeneity, which may affect the accuracy and reliability of machine learning models. Another challenge is the difficulty in feature selection and engineering, because the input features may have a complex relationship with the target output, and selecting the optimal set of input features is crucial to the performance of the machine learning model. Summary of the invention
[0005] In view of the above problems, the present invention is proposed to provide a shale oil effective drainage volume prediction method and system that overcomes the above problems or at least partially solves the above problems.
[0006] According to one aspect of the present invention, a method for predicting effective drainage volume of shale oil is provided, the prediction method comprising:
[0007] Step S1: constructing a shale oil effective drainage volume dataset;
[0008] Step S2: constructing a machine learning model and optimizing hyperparameters according to the shale oil effective drainage volume dataset to obtain a trained random forest prediction model;
[0009] Step S3: predicting the effective oil drainage volume according to the random forest prediction model.
[0010] Optionally, the step S1: constructing a shale oil effective drainage volume data set specifically includes:
[0011] Analyze the factors affecting the effective drainage volume of shale oil;
[0012] Determine the feature value range based on the actual distribution range of each factor in the target block, and generate samples by random sampling;
[0013] The effective oil drainage volume is calculated based on the samples.
[0014] Optionally, the step of determining the feature value range according to the actual distribution range of each factor in the target block and generating samples by random sampling specifically includes: determining the feature value range according to the actual distribution range of each factor in the target block, and generating samples by random sampling within the determined feature value range using the Latin hypercube sampling method.
[0015] Optionally, the calculating the effective oil leakage volume according to the sample specifically includes:
[0016] A numerical simulator is called to carry out numerical simulation on the sample, the production dynamics corresponding to the sample are obtained, and a production transient analysis method is used.
[0017] Optionally, the step S2: constructing a machine learning model according to the shale oil effective drainage volume data set specifically includes:
[0018] The processed sample data set is divided, random parameters are used as input, and effective oil drainage volume is used as output. The random forest method is used to establish a correlation model between geological parameters and engineering parameter variables and the effective oil drainage volume of shale oil.
[0019] Optionally, after establishing the correlation model between geological parameters and engineering parameter variables and the effective drainage volume of shale oil, the method further includes: using a Bayesian optimization algorithm with coupled cross-validation to optimize hyperparameters in the model using a k-fold cross-validation method.
[0020] Optionally, the step S3: predicting the effective oil drainage volume according to the random forest prediction model specifically includes:
[0021] The characteristic parameters of the target well were normalized in the same way as the training set, and the processed data were input into the random forest model to obtain the effective oil drainage volume.
[0022] Optionally, after obtaining the effective oil leakage volume, the method further includes:
[0023] The coefficient of determination R 2 Verify the prediction effect of the model, the calculation formula is as follows:
[0024]
[0025] In the formula, y in is the predicted value of the response data, y i is the true value of the response data, y is the mean of the response data, and n is the total number of samples.
[0026] Optionally, the analysis of factors affecting the effective drainage volume of shale oil specifically includes:
[0027] Porosity, permeability, reservoir pressure, number of fracturing stages, displacement, injection fluid volume, Young's modulus and Poisson's ratio.
[0028] Optionally, the ratio of the training set to the test set in the correlation model between the geological parameter and engineering parameter variables and the effective drainage volume of shale oil is 7:3.
[0029] The present invention also provides a shale oil effective drainage volume prediction system, which is applied to the above-mentioned shale oil effective drainage volume prediction method, and the prediction system comprises:
[0030] A data set construction module, used to construct a shale oil effective drainage volume data set;
[0031] A random forest prediction model building module is used to build a machine learning model and optimize hyperparameters according to the shale oil effective drainage volume data set to obtain a trained random forest prediction model;
[0032] The oil drainage volume prediction module is used to predict the effective oil drainage volume according to the random forest prediction model.
[0033] The present invention provides a shale oil effective oil leakage volume prediction method and system, the prediction method comprising: constructing a shale oil effective oil leakage volume data set; constructing a machine learning model and optimizing hyperparameters according to the shale oil effective oil leakage volume data set to obtain a trained random forest prediction model; predicting the effective oil leakage volume according to the random forest prediction model. The influencing factors of the shale oil effective oil leakage volume play an important role in the adjustment of shale oil development and the formulation of subsequent fracturing development plans for new wells, and can realize the prediction of the effective oil leakage volume of new wells.
[0034] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention, it can be implemented according to the contents of the specification. In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are listed below. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.
[0036] Figure 1 A flowchart of a specific embodiment of a method for predicting effective drainage volume of shale oil provided by an embodiment of the present invention;
[0037] Figure 2 It is an analysis of the main controlling factors of shale oil drainage volume in a specific implementation of the present invention;
[0038] Figure 3 It is a diagram showing the prediction effect of effective oil leakage volume according to a specific embodiment of the present invention. DETAILED DESCRIPTION
[0039] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art.
[0040] The terms "comprises" and "having" and any variations thereof in the description embodiments, claims and drawings of the present invention are intended to cover non-exclusive inclusions, for example, including a series of steps or units.
[0041] The technical solution of the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments.
[0042] A method for predicting the effective drainage volume of shale oil includes:
[0043] Step 1: Construction of shale oil effective drainage volume dataset;
[0044] Step 2: Based on step 1, machine learning model construction and hyperparameter optimization;
[0045] Step 3: Based on step 2, predict the effective oil leakage volume.
[0046] In step 1, the specific process of constructing the shale oil effective drainage volume data set includes the following steps: (1) Analyzing the factors affecting the shale oil effective drainage volume; (2) Determining the characteristic value range according to the actual distribution range of each factor in the target block, and using the Latin hypercube sampling method to randomly sample and generate a large number of samples within the determined characteristic value range; (3) Calling the numerical simulator to carry out numerical simulation on each sample to obtain the production dynamics corresponding to each sample, and using the production transient analysis method to calculate the effective drainage volume.
[0047] In step 2, the machine learning model is constructed and hyperparameters are optimized. The processed sample data set is divided, random parameters are used as input, and effective oil drainage volume is used as output. The correlation model between geological parameters and engineering parameter variables and effective oil drainage volume of shale oil is established using the random forest method. The hyperparameters in the model are optimized using the k-fold cross-validation method by coupling the Bayesian optimization algorithm with cross-validation to improve the model training and prediction effect.
[0048] In step 3, the effective oil drainage volume is predicted. After obtaining the trained random forest prediction model, the characteristic parameters of the target well are normalized in the same way as the training set, and the processed data are input into the random forest model to obtain the effective oil drainage volume. 2 ) Verify the prediction effect of the model, and the calculation formula is as follows:
[0049]
[0050] In the formula, y in is the predicted value of the response data, y i is the true value of the response data, y is the mean of the response data, and n is the total number of samples.
[0051] like Figure 1 As shown, Figure 1 The present invention is a flow chart of a specific embodiment of a method for calculating the effective drainage volume of shale oil.
[0052] In step 101, a dataset of effective shale oil drainage volume in the N1 well area is constructed. (1) Analysis of factors affecting the effective shale oil drainage volume: According to the actual geological conditions and existing construction data of the N1 well area, the factors affecting the drainage volume are determined to be porosity, permeability, reservoir pressure, number of fracturing stages, displacement, injection volume, Young's modulus and Poisson's ratio.
[0053] Input variable value range
[0054]
[0055] (2) The characteristic value range is determined according to the actual distribution range of each factor in the N1 well area, and a large number of samples are randomly sampled within the determined characteristic value range using the Latin hypercube sampling method.
[0056] After completing the calculation of 200 sets of model data, random parameters were used as input and effective oil drainage volume as output. The random forest method was used to establish a correlation model between geological parameters and engineering parameter variables and the effective oil drainage volume of shale oil. The ratio of training set to test set was 7:3.
[0057] Numerical simulation scheme generated by random sampling
[0058]
[0059] According to the contribution degree, the cumulative contribution of each parameter to the effective drainage volume of shale oil in the N1 well area was calculated, such as Figure 2 As shown in the figure, the injection volume, the number of stages and the displacement are the main controlling factors, and their contribution is greater than 15%. In terms of geological factors, Young's modulus and formation pressure are the main controlling factors.
[0060] (3) Call the numerical simulator to carry out numerical simulation on each sample to obtain the production dynamics corresponding to each sample, and use the production transient analysis method to calculate the effective oil leakage volume.
[0061] In step 102, based on step 101, a machine learning model is constructed and hyperparameters are optimized. First, the initial hyperparameters of the model and their search space are determined by manual tuning, and then the evaluation indicators under the initial hyperparameters are calculated and recorded by the cross-validation method, as shown in the model hyperparameter optimization result table.
[0062] Hyperparameter optimization space and results
[0063]
[0064] In step 103, based on step 102, the effective oil drainage volume is predicted. The trained random forest prediction model is obtained, and the data of the main controlling factors of shale oil drainage volume are input into the random forest model to obtain the effective oil drainage volume model of the N1 well area. The determination coefficient R is used 2 Verify the prediction effect of the model, the calculation formula is as follows:
[0065]
[0066] In the formula, y in is the predicted value of the response data, y i is the true value of the response data, y is the average value of the response data, and n is the total number of samples. The prediction effect of the effective drainage volume of shale oil in the N1 well area is as follows: Figure 3 shown.
[0067] Beneficial effect: The method for predicting the effective oil leakage volume of shale oil in the present invention summarizes the influencing factors of the effective oil leakage volume of shale oil, which plays an important role in the adjustment of shale oil development and the formulation of subsequent fracturing development plans for new wells, and can realize the prediction of the effective oil leakage volume of new wells.
[0068] The above specific implementation methods further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above are only specific implementation methods of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for predicting effective drainage volume of shale oil, characterized in that: The prediction method comprises: Step S1: constructing a shale oil effective drainage volume dataset; Step S2: constructing a machine learning model and optimizing hyperparameters according to the shale oil effective drainage volume dataset to obtain a trained random forest prediction model; Step S3: predicting the effective oil drainage volume according to the random forest prediction model.
2. A method for predicting effective shale oil drainage volume according to claim 1, characterized in that: The step S1: constructing a shale oil effective drainage volume data set specifically includes: Analyze the factors affecting the effective drainage volume of shale oil; Determine the feature value range based on the actual distribution range of each factor in the target block, and generate samples by random sampling; The effective oil drainage volume is calculated based on the samples.
3. The method for predicting effective drainage volume of shale oil according to claim 1, characterized in that: The step of determining the feature value range according to the actual distribution range of each factor in the target block and generating samples by random sampling specifically includes: determining the feature value range according to the actual distribution range of each factor in the target block, and generating samples by random sampling within the determined feature value range using the Latin hypercube sampling method.
4. A method for predicting effective shale oil drainage volume according to claim 1, characterized in that: The calculating of the effective oil leakage volume according to the sample specifically comprises: A numerical simulator is called to carry out numerical simulation on the sample, the production dynamics corresponding to the sample are obtained, and a production transient analysis method is used.
5. The method for predicting effective drainage volume of shale oil according to claim 1, characterized in that: The step S2: constructing a machine learning model according to the shale oil effective drainage volume data set specifically includes: The processed sample data set is divided, random parameters are used as input, and effective oil drainage volume is used as output. The random forest method is used to establish a correlation model between geological parameters and engineering parameter variables and the effective oil drainage volume of shale oil.
6. A method for predicting effective shale oil drainage volume according to claim 5, characterized in that: After establishing the correlation model between geological parameters and engineering parameter variables and the effective drainage volume of shale oil, the method also includes: using a Bayesian optimization algorithm with coupled cross-validation and utilizing a k-fold cross-validation method to optimize the hyperparameters in the model.
7. The method for predicting effective drainage volume of shale oil according to claim 1, characterized in that: The step S3: predicting the effective oil drainage volume according to the random forest prediction model specifically includes: The characteristic parameters of the target well were normalized in the same way as the training set, and the processed data were input into the random forest model to obtain the effective oil drainage volume.
8. A method for predicting effective shale oil drainage volume according to claim 7, characterized in that: The step of obtaining the effective oil drainage volume further includes: The coefficient of determination R 2 Verify the prediction effect of the model, the calculation formula is as follows: In the formula, y in is the predicted value of the response data, y i is the true value of the response data, y is the mean of the response data, and n is the total number of samples.
9. A method for predicting effective shale oil drainage volume according to claim 2, characterized in that: The analysis of the factors affecting the effective drainage volume of shale oil specifically includes: Porosity, permeability, reservoir pressure, number of fracturing stages, displacement, injection fluid volume, Young's modulus and Poisson's ratio.
10. A method for predicting effective shale oil drainage volume according to claim 5, characterized in that: The ratio of the training set to the test set in the correlation model between the geological parameter and engineering parameter variables and the effective drainage volume of shale oil is 7:
3.
11. A shale oil effective drainage volume prediction system, applied to a shale oil effective drainage volume prediction method according to any one of claims 1 to 10, characterized in that: The prediction system comprises: A data set construction module, used to construct a shale oil effective drainage volume data set; A random forest prediction model building module is used to build a machine learning model and optimize hyperparameters according to the shale oil effective drainage volume data set to obtain a trained random forest prediction model; The oil drainage volume prediction module is used to predict the effective oil drainage volume according to the random forest prediction model.