Auxiliary analysis method and system for shale oil and gas resource development
By collecting and mining historical data in shale areas, building a neural network model and performing Monte Carlo simulation, the problem of geological uncertainty in shale oil and gas resource development is solved, the sensitivity to characteristic parameters and output prediction is achieved, and the development strategy is optimized.
Patent Information
- Application Number
- CN202410001948.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-01-02
- Publication Date
- 2025-07-04
AI Technical Summary
The existing technology cannot accurately identify the "desserts" of shale oil and gas resources, conduct effective potential risk analysis and efficient development, and geological uncertainty has become a key issue.
By collecting historical data from the shale area, data mining is carried out to obtain the sensitivity of characteristic parameters, neural network models are built for data mining, and output prediction is carried out in combination with Monte Carlo simulation, providing auxiliary analysis systems to evaluate the development potential of shale oil and gas wells.
The sensitivity assessment of characteristic parameters is achieved, the accuracy of yield prediction and the effectiveness of risk analysis is improved, and researchers can identify key parameters and optimize development strategies.
Smart Images
Figure CN120256847A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of shale oil and gas exploration and development, and particularly relates to an auxiliary analysis method and system for shale oil and gas resource development. Background Art
[0002] Shale oil and gas resources are abundant and are an important part of the global energy supply. Accurately estimating the original geological reserves, ultimate recovery rate, and single-well recovery factor is a key task in the exploration and development of shale oil and gas resources. Existing technical experience shows that it is impossible to directly achieve accurate identification of shale "sweet spots", potential risk analysis, and efficient development according to conventional oil and gas methods. With the development of horizontal drilling and multi-stage hydraulic fracturing technologies, the engineering uncertainty in shale oil and gas development has decreased, and geological uncertainty has become the key to the successful development of shale oil and gas. However, there are few technical studies on evaluating aspects such as recoverable volume and potential risk analysis from the perspective of data mining. Summary of the Invention
[0003] The purpose of the present invention is to solve the problems existing in the above-mentioned prior art, and provide an auxiliary analysis method and system for shale oil and gas resource development, so as to provide data reference for more accurately evaluating the development potential of shale oil and gas wells.
[0004] The present invention is achieved through the following technical solutions: An auxiliary analysis method for shale oil and gas resource development includes the following steps: collecting historical data of the shale area, including characteristic data and production data of different wells in the shale area; obtaining the sensitivity of characteristic parameters that have different degrees of influence on oil and gas production by performing data mining on the historical data.
[0005] Further, the characteristic data includes geological data and / or engineering data.
[0006] Further, the geological data includes at least one of formation pressure, formation temperature, porosity, permeability, total organic matter content, kerogen maturity, and water saturation.
[0007] Further, the engineering data includes at least one of the number of fracturing stages, horizontal well section length, fracture spacing, fracture half-length, and fracture conductivity.
[0008] Further, the production data is any one of single-well daily gas production data, cumulative production, and EUR.
[0009] Further, before performing data mining, data preprocessing is first performed, including at least one of data denoising, filling, and standardization.
[0010] Further, the steps of performing data mining include:
[0011] Construct a sample set, where each sample in the sample set includes feature data and corresponding production data;
[0012] Use the sample set to train a data mining model: the feature data is used as the input vector, and the production data is used as the label data;
[0013] Change the value of a single feature parameter in the feature vector to obtain a to-be-tested input vector, input it into the trained data mining model, output the corresponding production, calculate the absolute value of the change rate of the production after changing the single feature parameter, so as to obtain the sensitivity of the single feature parameter.
[0014] Furthermore, construct the data mining model in the following manner:
[0015] Use the sample set to train various types of neural network models, and then use the root mean square error as the evaluation index to evaluate the training effect of each neural network model; use the neural network model set with the best training effect as the data mining model.
[0016] Furthermore, it also includes performing Monte Carlo simulation on the production under the influence of a single geological feature parameter: obtain the geological data and engineering data of the to-be-tested single well, randomly generate a certain geological feature parameter according to the probability distribution of the statistically obtained geological data each time, and together with the remaining geological data and engineering data of the to-be-tested single well, form a random vector, and input the random vector into the trained data mining model to obtain the production prediction result of the to-be-tested single well.
[0017] Furthermore, statistically obtain the probability distribution of the geological data in the following manner: according to the grade intervals of the geological data, count the grades into which each geological feature parameter in the entire shale area falls, and calculate the percentile of each geological feature parameter.
[0018] Furthermore, divide the grade intervals of the geological data according to expert experience, including the value ranges of five grade intervals: extremely good, good, medium, not good, and poor.
[0019] The present invention also provides an auxiliary analysis system for shale oil and gas resource development, including a data management module and a data mining module; the data management module is used to store the historical data of the shale area, including the feature data and production data of different wells in the shale area; the data mining module is used to perform data mining on the historical data to obtain the sensitivity of the feature parameters that have different degrees of influence on the oil and gas production.
[0020] Further, it also includes a Monte Carlo module for performing Monte Carlo simulation on the production volume affected by a single geological feature parameter: obtaining the geological data and engineering data of the well to be measured, randomly generating a certain geological feature parameter according to the probability distribution of the statistically obtained geological data each time of simulation, and forming a random vector together with the remaining geological data and engineering data of the well to be measured, and inputting the random vector into the trained data mining model to obtain the production volume prediction result of the well to be measured.
[0021] Further, it also includes a visualization module for generating a distribution feature map according to the probability distribution of geological data, for generating a sensitivity ranking map according to the sensitivity of feature parameters, and for generating a production volume frequency distribution map under the influence of different geological feature parameters according to the Monte Carlo simulation result.
[0022] Further, it also includes a data preprocessing module for preprocessing the historical data in the shale area and inputting it into the data mining module.
[0023] Compared with the prior art, the beneficial effects of the present invention include:
[0024] 1. Through the sensitivity of the feature parameters mined by the present invention, researchers can judge which feature parameters should be focused on according to the strength of the sensitivity of the feature parameters, avoiding blind and indiscriminate treatment of various feature parameters.
[0025] 2. The present invention uses the absolute value of the change rate of the production volume after changing a single feature parameter as the sensitivity of a single feature parameter, realizing the association between the feature parameter and the production volume, enabling researchers to better evaluate the production volume.
[0026] 3. The present invention constructs a data mining model by integrating different types of neural network models, making full use of the advantages of different types of neural networks, thereby being able to improve the accuracy of the data mining model.
[0027] 4. The present invention performs Monte Carlo simulation based on the trained data mining model, and can more accurately simulate the production volume affected by different geological feature parameters.
[0028] 5. The present invention generates corresponding display charts through the visualization module according to the analysis results, enabling researchers to obtain data references more intuitively and quickly. Description of the Drawings
[0029] Figure 1 It is a structural diagram of an auxiliary analysis system for shale oil and gas resource development.
[0030] Figure 2 It is a distribution feature and evaluation diagram of shale geological factors;
[0031] Figure 3It is a tornado diagram for sensitivity analysis of shale oil and gas parameters;
[0032] Figure 4 It is a graph of the Monte Carlo simulation results of shale oil and gas EUR;
[0033] Figure 5 It is a display diagram of the visualization interface. Specific implementation manners
[0034] In the process of shale oil and gas development, engineering factors and geological factors are inseparable. Practice has shown that geological factors play a crucial role in the commercial development of shale oil and gas. In order to increase the success rate of shale oil and gas resource development, the present invention designs a new strategy for rapid shale formation opportunity screening by combining data mining and Monte Carlo methods, taking into account important parameters for shale reservoir development, such as rock physical properties, reservoir properties, drilling, completion and fracturing engineering operation parameters, etc.
[0035] Based on expert experience, each geological characteristic parameter is divided into five value ranges: excellent, good, medium, poor, and very poor. The data mining method is used to conduct sensitivity analysis on the influencing factors.
[0036] The Monte Carlo simulation method is used to conduct uncertainty analysis on the production of shale oil and gas wells under different geological risk factors, such as EUR (Estimated Ultimate Recovery, ultimate recovery). The final presented results are displayed by a visualization dashboard composed of a key parameter evaluation graph, a risk sensitivity analysis graph, and a set of EUR statistical graphs.
[0037] Reference Figure 1 As shown, for the convenience of on-site staff, an auxiliary analysis system for shale oil and gas resource development for implementing the above method is provided, including a preprocessing module, a data management module, a data mining module, a Monte Carlo module, a visualization algorithm, and an account management module.
[0038] The data preprocessing module is used to preprocess the historical data in the shale area and input it to the data mining module.
[0039] The data management module is used to store the historical data in the shale area, including the characteristic data and production data of different wells in the shale area; the data mining module is used to conduct data mining on the historical data to obtain the sensitivity of characteristic parameters that have different degrees of influence on oil and gas production.
[0040] A Monte Carlo module for performing Monte Carlo simulation on the production under the influence of a single geological feature parameter: obtaining the geological data and engineering data of the well to be measured, randomly generating a certain geological feature parameter according to the probability distribution of the statistically obtained geological data each time, and forming a random vector together with the remaining geological data and engineering data of the well to be measured, and inputting the random vector into the trained data mining model to obtain the production prediction result of the well to be measured.
[0041] A visualization module for generating a distribution characteristic map according to the probability distribution of geological data, generating a sensitivity ranking map according to the sensitivity of feature parameters, and generating a production frequency distribution map under the influence of different geological feature parameters according to the Monte Carlo simulation results.
[0042] Reference Figure 1 As shown, the user logs in to the web page to upload the historical data of the shale area, and the server receives the historical data and stores it in the data management module. After the original data set is processed by the preprocessing module, the data mining module, the Monte Carlo module, and the visualization module, the user can see the dashboard on the web page to intuitively display the geological risks and development opportunities of the study area.
[0043] More specifically, it includes the following steps:
[0044] (1) Collection and preprocessing of historical data in the shale area
[0045] Collect the historical data of the shale area, including the characteristic data and production data of different wells in the shale area. The characteristic data includes geological data and engineering data. The geological data is composed of different types of geological feature parameters, and the engineering data is composed of different types of engineering feature parameters.
[0046] The geological data includes formation pressure, formation temperature, porosity, permeability, total organic matter content, kerogen maturity, water saturation, etc.
[0047] The drilling and completion engineering data includes the number of fracturing stages, horizontal well section length, fracture spacing, fracture half-length, fracture conductivity, etc.
[0048] The production data includes single-well daily gas production data, cumulative production, EUR, etc. Any production data can be used as the label of the sample as needed for the training and testing of the data mining model.
[0049] Perform data denoising, filling, and standardization preprocessing operations on the collected samples to obtain a sample set.
[0050] (2) Statistical probability distribution of geological data
[0051] According to expert experience, the geological data is divided into five levels: excellent, good, medium, poor, and bad. According to the geological data level, the levels of the geological characteristic parameters of the entire shale area are counted, and the percentiles of the geological characteristic parameters are calculated. The percentiles include P10, P50, and P90 values. P50 is the median, which means that 50% of the numbers are less than the P50 value.
[0052] (3) Data Mining
[0053] Construct a sample set, each sample in the sample set includes feature data and corresponding yield data: use the k-fold crossover method to divide the preprocessed sample set into a training set and a test set.
[0054] Build a data mining model:
[0055] The sample set is used to train various types of neural network models, and then the root mean square error RMSE (Root Mean Square Error) is used as an evaluation indicator to evaluate the training effect of each neural network model. The smaller the value, the better the performance of the model.
[0056] The neural network model can select supervised regression models such as multivariate linear, neural network, naive Bayes, SVM, etc.
[0057] The neural network model set with the best training effect is used as the data mining model: the top three neural network models are selected and the Bagging method is used to establish an integrated learning data mining model. Other model fusion methods, such as Boosting, can also be used.
[0058] Change the value of a single feature parameter in the feature vector to obtain the input vector to be tested, and input it into the trained data mining model, output the corresponding output, calculate the absolute value of the rate of change of output after changing a single feature parameter (which can be either a geological feature parameter or an engineering feature parameter), and thus obtain the sensitivity of a single feature parameter.
[0059] For the trained data mining model, by changing a certain input feature factor by a small change range, the sensitivity of the EUR of shale oil and gas wells can be obtained. These factors are arranged in order from large to small and from top to bottom according to the change range of EUR of shale oil and gas wells, and a tornado diagram of sensitivity analysis of shale oil and gas parameters is obtained. It can be used for risk factor analysis. The higher the factor, the stronger its sensitivity is, and the researchers need to pay special attention to it.
[0060] (4) Monte Carlo simulation
[0061] Perform Monte Carlo simulation on the production volume affected by a single geological feature parameter: Obtain the geological data and engineering data of the well to be tested. Each time a simulation is performed, a certain geological feature parameter is randomly generated according to the probability distribution of the statistically obtained geological data, and together with the remaining geological data and engineering data of the well to be tested, a random vector is formed. The random vector is input into the trained data mining model to obtain the production volume prediction result of the well to be tested.
[0062] Repeat the simulation N times to obtain N EUR values of single wells, statistically analyze the frequency distribution characteristics of the N EUR values of single wells, and obtain the values of P10, P50, and P90.
[0063] Example 1
[0064] (1) The customer uploads the data of the research area
[0065] As shown in Table 1 and Figure 1 The user uploads the data table of the shale area through the network on the customer terminal (computer or mobile phone). Each row of the data table represents a sample. Suppose there are n columns. The first column is the sample serial number, the 2nd to n - 1th columns are the features of the sample, including various geological and engineering parameters that affect the shale oil and gas production volume, and the nth column is the label of the sample, which is the EUR of the shale oil and gas well. The data set adopts a unified csv format, and the first row is the standard header for convenient modular processing.
[0066] Table 1 Data table of the shale research area
[0067]
[0068] (2) Preprocess and statistically analyze the probability distribution of geological data
[0069] The original data uploaded by the customer is stored in the database of the server for the customer to view, download, and update at any time. The customer logs in to the web account for database management, data preprocessing, shale oil and gas geological risk analysis, and opportunity identification. Click the preprocessing button, and the server will perform operations such as denoising, filling, and standardization on the original data. The preprocessed data is also stored in the database of the data management module.
[0070] Generate a distribution characteristic diagram according to the probability distribution of geological data. Refer to Figure 2 As shown: Statistically analyze the distribution of sample features and visually present them on the evaluation axis of feature parameters based on expert experience. The blue box represents the distribution range of the parameters in the research area, and the blue triangle refers to the P50 value.
[0071] (3) Data mining and risk analysis
[0072] Use the preprocessed dataset to train and test the regression model in the data mining module. If the trained regression model meets the test requirements, the model can be used for risk sensitivity analysis and Monte Carlo simulation; otherwise, retrain the model. For the trained data mining model, make a small change in a certain characteristic parameter (the change range does not exceed 5%), calculate the absolute value of the production change rate after changing a single characteristic parameter, so as to obtain the sensitivity of a single characteristic parameter. In this embodiment, the sensitivity of the EUR of shale gas and oil wells changing accordingly is calculated.
[0073] Arrange these factors in descending order of the EUR change range of shale gas and oil wells, from top to bottom, to obtain a tornado diagram for sensitivity analysis of shale gas and oil parameters (see Figure 3 ). It can be used for risk factor analysis. The factors closer to the top indicate stronger sensitivity and require key attention from researchers.
[0074] Embodiment 2
[0075] This embodiment conducts a Monte Carlo simulation on the basis of Embodiment 1.
[0076] The original data uploaded by the customer is stored in the database of the server for the customer to view, download, and update at any time. The customer logs in to the web account for database management, data preprocessing, shale gas and oil geological risk analysis, and opportunity identification. Click the preprocessing button, and the server will perform operations such as denoising, filling in the blanks, and standardizing the original data. The preprocessed data is also stored in the database of the data management module.
[0077] Generate a distribution characteristic diagram according to the probability distribution of geological data, as shown in reference to Figure 2 : Statistically analyze the distribution of sample characteristics and visually present them on the characteristic parameter evaluation axis based on expert experience. The blue box represents the distribution range of parameters in the study area, and the blue triangle refers to the P50 value.
[0078] Use the preprocessed dataset to train and test the regression model in the data mining module. If the trained regression model meets the test requirements, the model can be used for risk sensitivity analysis and Monte Carlo simulation; otherwise, retrain the model. For the trained data mining model, make a small change in a certain characteristic parameter (the change range does not exceed 5%), calculate the absolute value of the production change rate after changing a single characteristic parameter, so as to obtain the sensitivity of a single characteristic parameter. In this embodiment, the sensitivity of the EUR of shale gas and oil wells changing accordingly is calculated.
[0079] Arrange these factors in descending order of the EUR change range of shale gas and oil wells, from top to bottom, to obtain a tornado diagram for sensitivity analysis of shale gas and oil parameters (see Figure 3) It can be used for risk factor analysis. The higher the factor, the stronger its sensitivity, and researchers need to pay key attention to it.
[0080] Obtain the geological data and engineering data of the single well to be measured. Each simulation randomly generates a certain geological characteristic parameter according to the probability distribution of the statistically obtained geological data, and together with the remaining geological data and engineering data of the single well to be measured, forms a random vector. Input the random vector into the trained data mining model to obtain the production prediction result of the single well to be measured.
[0081] Repeat the above random simulation N times, and N single well EUR results can be obtained. Draw a frequency distribution histogram and calculate the P10, P50, and P90 values. The Monte Carlo random results under different random geological characteristic parameters are as Figure 4 shown. Oil workers can know the confidence level of the single well EUR under this geological factor, which can be used for the opportunity identification of shale "sweet spots".
[0082] Example 3
[0083] This example is visualized, enabling researchers to obtain data references more intuitively and quickly.
[0084] Combine the Figure 2 , Figure 3 , Figure 4 obtained in Example 1 and Example 2 into a visual dashboard (see Figure 5 ), and display it on the web browser of the client. Oil research personnel can comprehensively observe and analyze the results.
[0085] On the left side of the visual dashboard is the distribution characteristic and evaluation diagram of geological characteristic parameters, which gives the distribution range and evaluation of key parameters in the form of parallel coordinates.
[0086] The upper right part of the visual dashboard is the tornado diagram of geological engineering parameter sensitivity analysis obtained by the data mining method, which is used to intuitively display the influence of each factor and analyze the risk factors affecting the EUR of shale oil and gas wells. The higher the sensitivity of the characteristic parameter, the greater the risk it brings.
[0087] The lower right part of the dashboard is the frequency distribution diagram of the EUR of shale oil and gas wells under different geological risk factors based on Monte Carlo simulation. For example, according to Figure 4 it is possible to know the probability production under a certain geological factor. For example, the production change range under the maturity factor is very large, so we give priority to considering maturity as one of the indicators for sweet spot selection.
[0088] In the description of the present invention, it should be noted that unless otherwise clearly specified and defined, the terms "connected" and "coupled" should be understood in a broad sense. For example, it may be a fixed connection, a detachable connection, or an integral connection; it may be a mechanical connection or an electrical connection; it may be directly connected or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.
[0089] In the description of the present invention, unless otherwise stated, the orientation or positional relationship indicated by the terms "upper", "lower", "left", "right", "inner", "outer", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the present invention.
[0090] The above technical solutions are only specific embodiments of the present invention. For those skilled in the art, based on the principles disclosed in the present invention, various types of improvements or deformations can be easily made, not limited to the technical solutions described in the above specific embodiments of the present invention. Therefore, the foregoing description is only preferred and does not have a limiting meaning.
Claims
1. An auxiliary analysis method for shale oil and gas resource development, characterized in that, The method includes the following steps: collecting historical data of the shale area, including characteristic data and production data of different wells in the shale area; obtaining the sensitivity of characteristic parameters that have different degrees of influence on oil and gas production by performing data mining on the historical data.
2. The auxiliary analysis method for shale oil and gas resource development according to claim 1, wherein The characteristic data includes geological data and / or engineering data.
3. The auxiliary analysis method for shale oil and gas resource development according to claim 2, wherein, The geological data includes at least one of formation pressure, formation temperature, porosity, permeability, total organic matter content, kerogen maturity, and water saturation.
4. The auxiliary analysis method for shale oil and gas resource development according to claim 2, wherein The engineering data includes at least one of the number of fracturing stages, horizontal well section length, fracture spacing, fracture half-length, and fracture conductivity.
5. The auxiliary analysis method for shale oil and gas resource development according to claim 2, characterized in that The production data is any one of the daily gas production data per well, cumulative production, and EUR.
6. The auxiliary analysis method for shale oil and gas resource development according to claim 2, wherein, Before performing data mining, data preprocessing is first performed, including at least one of data denoising, filling, and standardization.
7. The auxiliary analysis method for shale oil and gas resource development according to claim 1, wherein The steps of performing data mining include: Constructing a sample set, where each sample in the sample set includes characteristic data and corresponding production data; Training a data mining model using the sample set: using the characteristic data as the input vector and the production data as the label data; Changing the value of a single characteristic parameter in the characteristic vector to obtain a test input vector, inputting the test input vector into the trained data mining model, outputting the corresponding production, and calculating the absolute value of the change rate of the production after changing the single characteristic parameter, so as to obtain the sensitivity of the single characteristic parameter.
8. The auxiliary analysis method for shale oil and gas resource development according to claim 7, wherein The data mining model is constructed in the following manner: Training various types of neural network models using the sample set, and then using the root mean square error as an evaluation index to evaluate the training effect of each neural network model; taking the neural network model set with the best training effect as the data mining model.
9. The auxiliary analysis method for shale oil and gas resource development according to claim 7, characterized in that, It also includes performing Monte Carlo simulation on the production under the influence of a single geological characteristic parameter: obtaining the geological data and engineering data of the well to be tested, randomly generating a certain geological characteristic parameter according to the probability distribution of the geological data obtained by statistics each time, and forming a random vector together with the remaining geological data and engineering data of the well to be tested, and inputting the random vector into the trained data mining model to obtain the production prediction result of the well to be tested.
10. The auxiliary analysis method for shale oil and gas resource development according to claim 1 or 9, characterized in that, The probability distribution of the geological data is statistically obtained in the following manner: according to the grade intervals of the geological data, counting the grades into which each geological characteristic parameter in the entire shale area falls, and calculating the percentile of each geological characteristic parameter.
11. The auxiliary analysis method for shale oil and gas resource development according to claim 10, wherein, The grade intervals of the geological data are divided according to expert experience, including the value ranges of five grade intervals: extremely good, good, medium, not good, and poor.
12. An auxiliary analysis system for shale oil and gas resource development, characterized in that, It includes a data management module and a data mining module; the data management module is used to store the historical data of the shale area, including the characteristic data and production data of different wells in the shale area; the data mining module is used to perform data mining on the historical data to obtain the sensitivity of characteristic parameters that have different degrees of influence on oil and gas production.
13. The auxiliary analysis system for shale oil and gas resource development according to claim 12, characterized in that, It further includes a Monte Carlo module for performing Monte Carlo simulation on the production volume under the influence of a single geological feature parameter: obtaining the geological data and engineering data of the well to be tested, randomly generating a certain geological feature parameter according to the probability distribution of the statistically obtained geological data each time of simulation, and together with the remaining geological data and engineering data of the well to be tested to form a random vector, and inputting the random vector into the trained data mining model to obtain the production volume prediction result of the well to be tested.
14. The auxiliary analysis system for shale oil and gas resource development according to claim 13, wherein It further includes a visualization module for generating a distribution feature map according to the probability distribution of geological data, for generating a sensitivity ranking map according to the sensitivity of feature parameters, and for generating a production volume frequency distribution map under the influence of different geological feature parameters according to the Monte Carlo simulation results.
15. The auxiliary analysis system for shale oil and gas resource development according to claim 13, wherein, It further includes a data preprocessing module for preprocessing the historical data in the shale area and inputting it to the data mining module.