Shale gas resource quantity prediction method and device

Abstract

The invention discloses a shale gas resource quantity prediction method and device. The method comprises the steps that the total amount of adsorbed gas contained in different types of lithofacies in shale with the preset mass is obtained; the total amount of free gas contained in different types of lithofacies in the shale with the preset mass is obtained; according to the total amount of the adsorbed gas and the total amount of the free gas contained in the different types of shale lithofacies, the total resource amount of the shale gas contained in each type of lithofacies in the preset mass of shale is calculated, and therefore the content of the adsorbed gas and the content of the free gas contained in all types of lithofacies and the total resource amount of the shale gas are comprehensively considered; therefore, the total resource quantity of the shale gas contained in the shale with the preset mass is calculated, the prediction result can be more accurate, and prediction of the resource quantity of the shale gas in the low-exploration-degree area in the south of China can be achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of shale gas exploration, and in particular to a shale gas resource prediction method and device. Background Art

[0002] Shale gas refers to unconventional natural gas that is found in organic-rich mudstone and its interlayers, and exists mainly in adsorption and free states. It is mainly composed of methane and is a clean and efficient energy resource and chemical raw material. It is mainly used for residential gas, urban heating, power generation, automobile fuel and chemical production, etc., and has a wide range of uses.

[0003] At present, southern my country is a key area for shale gas exploration, evaluation and profitable development, but there are still many areas with low exploration degree (such as the eastern Sichuan-western Hubei region). The exploration data of shale gas in these areas is incomplete, and the current geological exploration data and physical property data are incomplete, which causes certain troubles for the accurate calculation of shale gas volume and the determination of favorable gas-bearing zones, and seriously affects the exploration and exploitation efficiency of shale gas.

[0004] For example, the paper "Evaluation Methods and Evaluation Parameter Assignment of Shale Gas Resources - Taking the Wufeng Formation-Longmaxi Formation in the Central Yangtze Region as an Example" published in China Geology used the volume method to conduct a good study and exploration of the evaluation of shale gas resources in the Wufeng Formation-Longmaxi Formation in the Central Yangtze Region. The current evaluation method mainly draws on the United States, where the degree of shale gas exploration and development is relatively high. It mainly evaluates the low exploration characteristics through the high exploration characteristics of the existing exploration areas. Although it can also predict the shale gas resources in the low exploration areas, this method is more suitable for areas near the study area where a large amount of shale gas exploration has been carried out. It is not applicable to the study areas such as the eastern Sichuan-western Hubei region, where the geological differences between different blocks are large, the shale lithofacies types are complex, and the overall exploration degree is currently low. Therefore, it is necessary to calculate the shale gas resources in these low exploration degree areas (such as the eastern Sichuan-western Hubei region) to provide a basis for further exploration and development of shale gas in this area. Summary of the invention

[0005] The embodiment of the present invention provides a shale gas resource prediction method and device, which can make the prediction result more accurate by calculating the shale gas resource by comprehensively considering the adsorbed gas and free gas contents.

[0006] In order to achieve the above object, an embodiment of the present invention provides a shale gas resource prediction method, comprising the following steps:

[0007] The method includes systematically identifying centimeter-level lithofacies types of preset quality shale in different regions, obtaining the total amount of adsorbed gas contained in different types of lithofacies in shale by region; obtaining the total amount of free gas contained in different types of lithofacies in shale of preset quality; and calculating the total shale gas resources contained in each type of lithofacies in shale of preset quality according to the total amount of adsorbed gas and the total amount of free gas contained in the different types of shale lithofacies, thereby calculating the total shale gas resources contained in shale of preset quality by comprehensively considering the adsorbed gas content and free gas content contained in all types of lithofacies and the total shale gas resources, and then multiplying by the coefficient of reducing the actual gas production of shale gas by the nano-confinement effect, so as to calculate the total shale gas resources contained in shale of preset quality, which can make the prediction result more accurate and realize the prediction of shale gas resources in low-exploration areas in southern my country (such as eastern Sichuan-western Hubei region).

[0008] Further,

[0009] Further, the total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources;

[0010] The total shale gas resource contained in the shale of a preset mass is calculated according to the total amount of adsorbed gas and the total amount of free gas, including:

[0011] The total shale gas resources contained in the shale of the preset quality are calculated using Formula 1:

[0012] Q 总 =Q 吸 +Q 游 (one)

[0013] Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 Represents the total free gas resources.

[0014] Furthermore, the method further comprises: obtaining the amount of dissolved gas resources contained in shale of a preset quality;

[0015] The formula 1 is further: Q 总 =Q 吸 +Q 游 +Q 溶

[0016] The Q 溶 is the amount of dissolved gas resources.

[0017] Furthermore, the Q 吸 =0.01·A·h·ρ·q 吸 , where A represents the gas-bearing area of ​​shale, h represents the thickness of shale, ρ represents the density of shale, and q 吸 is the adsorption gas content;

[0018] wherein, q 吸 = VL·P / (PL + P), where VL represents the Langmuir volume, P represents the formation pressure, and PL represents the Langmuir pressure.

[0019] Further, the Q 游 = 0.01·A·h·Φg·Sg / Z, where Φg represents the porosity, Sg represents the gas saturation, and Z represents the natural gas compressibility factor.

[0020] Further, the total amount of adsorbed gas is the adsorbed gas content, and the total amount of free gas is the free gas content;

[0021] Calculating the total shale gas resource amount contained in a preset mass of shale according to the total amount of adsorbed gas and the total amount of free gas includes:

[0022] Using Equation 2 to calculate the total shale gas resource amount contained in the preset mass of shale:

[0023] Q 总 = 0.01·A·h·ρ·(q 吸 + q 游 ) / Z (Equation 2)

[0024] wherein, Q 总 represents the total shale gas resource amount, A represents the gas-bearing area of the shale, h represents the thickness of the shale, ρ represents the density of the shale, Z represents the natural gas compressibility factor, q 吸 represents the adsorbed gas content, and q 游 represents the free gas content.

[0025] Further, the adsorbed gas content and the free gas content are obtained by one of the desorption method, isothermal adsorption method, analogy method, statistical method, well logging interpretation and calculation method.

[0026] Further, Z ≤ 1.

[0027] Further, the nano-confinement effect hinders the generation of gas phase, thereby reducing the actual production of shale gas. The dynamic simulation of shale gas production in typical wells in the eastern and western parts of the study area shows that after considering the nano-confinement effect, the cumulative shale gas production of Well A in the east and Well B in the west are 87% and 93% respectively of that without considering the nano-confinement effect. Therefore, when calculating the shale gas production, a coefficient brought by the nano-confinement effect on the actual shale gas production needs to be added after the traditional calculation formula, that is, the coefficient in the eastern part of the study area is 0.87, and the coefficient in the western part of the study area is 0.93.

[0028] The embodiment of the present invention also provides a shale gas resource amount prediction device, including:

[0029] A first acquisition module is used to acquire the total amount of adsorbed gas contained in shale of a preset mass;

[0030] A second acquisition module is used to obtain the total amount of free gas contained in shale of a preset quality;

[0031] The calculation module is used to calculate the total shale gas resources contained in the shale of a preset mass according to the total amount of adsorbed gas and the total amount of free gas.

[0032] Further, the total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources;

[0033] The calculation module is specifically used to calculate the total shale gas resources contained in the shale of the preset quality using Formula 1:

[0034] Q 总 =Q 吸 +Q 游 (one)

[0035] Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 Represents the total free gas resources.

[0036] An embodiment of the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the shale gas resource prediction method described in any one of the above items.

[0037] In combination with the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solutions to be protected by the present invention are as follows:

[0038] First, the shale gas resource prediction method of the present invention includes obtaining the total amount of adsorbed gas contained in shale of a preset mass; obtaining the total amount of free gas contained in shale of a preset mass; and calculating the total shale gas resources contained in the shale of a preset mass based on the total amount of adsorbed gas and the total amount of free gas. By comprehensively considering the adsorbed gas content and the free gas content, the total shale gas resources are calculated, which can make the prediction result more accurate and realize the prediction of shale gas resources in low-exploration areas (such as the eastern Sichuan-western Hubei region).

[0039] Second, the significant technical progress brought about by the above-mentioned shale gas resource prediction method is mainly reflected in the following aspects:

[0040] 1) Improve the accuracy of resource assessment:

[0041] Through the comprehensive quantitative analysis of adsorbed gas, free gas and dissolved gas, the total shale gas resources can be assessed more accurately.

[0042] The use of Langmuir model parameters (such as Langmuir volume and Langmuir pressure) can better simulate shale gas adsorption behavior.

[0043] 2) Quantify the contribution of various gases:

[0044] Clearly distinguishing the quantitative contributions of adsorbed gas and free gas will help understand the physical and chemical characteristics of shale gas reservoirs and provide a scientific basis for the formulation of production strategies.

[0045] 3) Improve the reliability of resource prediction models:

[0046] Combining multiple methods of obtaining gas content (such as desorption method, isothermal adsorption method, analogy method, statistical method, well logging interpretation and calculation method) can improve the accuracy of data acquisition and the robustness of the prediction model.

[0047] 4) Optimize mining plan:

[0048] Through precise resource prediction, the accuracy of prediction of favorable exploration areas is greatly guaranteed, and it is possible to more effectively plan drilling locations, ensure the precise path of horizontal wells, design fracturing operations and optimize production management, thereby reducing costs and improving shale gas exploration and production efficiency, and promoting shale gas reserves and production. This is of great significance in alleviating the tense situation of natural gas supply in my country, reducing the external dependence of natural gas consumption, ensuring my country's energy security, and optimizing the energy structure.

[0049] 5) Scientific management of resources:

[0050] It will help the government and energy companies to scientifically manage and reasonably allocate resources, formulate long-term development plans, ensure the sustainable development of energy, and also drive infrastructure construction, improve infrastructure construction in shale gas producing areas, and promote the development of related supporting industries. In addition, it can also drive the development of related industries and fields such as equipment manufacturing, steel demand, cement manufacturing, and engineering construction, increase employment and tax revenue, and promote the sustainable development of the local economy and even the national economy.

[0051] 6) Environmental Impact Assessment:

[0052] Accurate resource estimation can better predict the impact of mining activities on the environment, help formulate effective environmental protection measures, and reduce the negative impact on the environment. In addition, the promotion and application of natural gas has become a vital means for the Chinese government to control pollution. As a clean and efficient energy resource, shale gas can reduce carbon dioxide emissions by 50% and nitrogen oxides by 75% compared with coal, and there is almost no emission of sulfur dioxide, carbon monoxide, particulate matter and mercury. It can effectively reduce greenhouse gas emissions and improve the ecological environment, which is of great strategic significance for promoting sustainable economic and social development.

[0053] 7) Improve exploration technology:

[0054] Incorporating advanced statistical and mathematical models into prediction methods has promoted the advancement of exploration technology, which is of great significance for the exploration of resources under complex geological conditions. It has provided theoretical guidance for increasing reserves and production of marine shale gas in the south, directly and indirectly promoted the establishment of multiple shale gas exploration areas in the Sichuan Basin, and played a positive role in promoting and facilitating the rapid growth of my country's oil and gas reserves.

[0055] 8) Economic benefits:

[0056] Accurate resource assessment is crucial for investment decisions. It can help China Petroleum & Chemical Corporation, China National Petroleum Corporation, China Geological Survey and other relevant energy companies and departments to evaluate potential economic benefits, guide the evaluation and selection of marine shale gas areas in southern my country, reduce exploration investment risks, and effectively improve economic benefits.

[0057] Through this comprehensive prediction method, we can have a more comprehensive understanding of the characteristics of shale gas reservoirs, provide more scientific decision-making support for the exploration and development of shale gas resources, and thus drive technological progress and economic benefits in the shale gas industry.

[0058] Third, the technical progress of the shale gas resource prediction method provided by the present invention is mainly reflected in the following aspects:

[0059] 1) Comprehensive assessment of shale gas resources: This method takes into account not only adsorbed gas and free gas, but also dissolved gas, thus providing a more comprehensive and accurate prediction of shale gas resources.

[0060] 2) Accurate resource calculation formula: By introducing specific physical parameters (such as the gas-bearing area, thickness, density, Langmuir volume and pressure of shale), the prediction is more dependent on quantifiable and measurable scientific parameters, thereby improving the accuracy of the prediction.

[0061] 3) Considering formation conditions: By introducing formation pressure and natural gas compression factor, this method incorporates the actual conditions of the formation into the calculation, making the prediction results more consistent with the actual geological conditions.

[0062] 4) Simplified calculation process: Through clear calculation formulas, the prediction process becomes simple and clear, making it easier for engineers and researchers to quickly calculate the shale gas resources.

[0063] 5) Strong adaptability: Each parameter in the method can be adjusted according to the specific conditions of different regions and different shale formations, making the prediction method highly adaptable and flexible.

[0064] 6) Helps in decision making: Accurate resource forecast is of great significance for the planning, investment decision and economic analysis of shale gas extraction projects.

[0065] 7) Scientific management of shale gas resources: Provide the government and relevant departments with a scientific method to evaluate and manage shale gas resources, which will help to formulate reasonable energy policies and mining strategies.

[0066] Through this approach, the development of shale gas resources can be more scientific and efficient, while minimizing resource waste and environmental impact. BRIEF DESCRIPTION OF THE DRAWINGS

[0067] The technical solutions and beneficial effects of the present invention will be made apparent by describing in detail the specific embodiments of the present invention in conjunction with the accompanying drawings.

[0068] Figure 1 is a dynamic simulation result of shale gas production of typical wells in the east and west of the study area of ​​the embodiment of the present invention;

[0069] Figure 2 is a flow chart of a shale gas resource prediction method according to an embodiment of the present invention;

[0070] Figure 3 It is a schematic diagram of the structure of a shale gas resource prediction device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0071] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0072] The scheme in the background technology paper has the following shortcomings due to insufficient adaptability when evaluating shale gas resources: First, the study did not carry out the key step of fine division of shale lithofacies, and then calculated the resources by type and region, which is not only not conducive to the accurate evaluation of shale gas resources, but also difficult to find the most favorable lithofacies for shale gas enrichment, and unable to ensure the precise path of horizontal wells, resulting in inaccurate exploration; second, the means adopted in the paper originated from the technical method of Standing et al. in 1942, which only uses a single plate method and is outdated; third, the paper did not consider the impact of nano-confinement effect on the actual production of shale gas, resulting in deviations in the calculation of shale gas resources.

[0073] The working principle of the shale gas resource prediction method provided by the present invention mainly involves the following key steps:

[0074] 1) Identification of shale lithofacies: First, a system is needed to identify different centimeter-level lithofacies types in shale. Lithofacies types refer to shale types with different physical and chemical properties (including color, particle size distribution and sorting degree, sedimentary structure and bedding structure, mineral composition and content, geochemical characteristics, etc.), which is critical to predicting its gas content.

[0075] 2) Quantification of adsorbed and free gas: After determining the lithofacies of the shale, the next step is to measure and calculate the total amount of adsorbed and free gas contained in these lithofacies. Adsorbed gas refers to the gas attached to the surface of rock pores, while free gas refers to the gas that flows freely in the rock pores.

[0076] 3) Calculation of shale gas resources: For each lithofacies, the total amount of adsorbed gas and the total amount of free gas are added together to obtain the total shale gas resources in that lithofacies type. This step involves complex calculations and requires specific formulas or models.

[0077] 4) Considering the nanoconfinement effect: This involves considering that the nanoconfinement effect hinders the generation of the gas phase, thereby reducing the actual production of shale gas. When calculating shale gas production, it is necessary to add the coefficient of the nanoconfinement effect on the actual shale gas production after the traditional calculation formula, that is, the coefficient in the eastern part of the study area is 0.87, and the coefficient in the western part of the study area is 0.93 ( Figure 1 ).

[0078] 5) Comprehensive calculation of total resources: Finally, the total shale gas resources contained in all types of lithofacies are added together to obtain the total shale gas resources contained in shale of a preset quality.

[0079] The key to the method provided by the present invention is to accurately identify and quantify the gas content in different lithofacies, and take into account the impact of the nano-confinement effect on reducing the actual production of shale gas, so as to accurately predict the total resources of shale gas and more accurately evaluate the potential value and development prospects of shale gas fields.

[0080] The following are four application embodiments provided by the present invention, which demonstrate the specific implementation scheme of the shale gas resource prediction method.

[0081] Example 1: Shale gas resource prediction based on core analysis

[0082] 1) Core collection and analysis:

[0083] Core samples are taken from the target shale formation.

[0084] Core analysis techniques (such as thin section analysis, X-ray diffraction, scanning electron microscopy, comprehensive analysis of well logging characteristics, etc.) are used to identify lithofacies types and characteristics.

[0085] 2) Laboratory testing:

[0086] The core samples were tested for porosity and permeability.

[0087] High pressure adsorption experiments were used to determine the amount of adsorbed gas in each lithofacies.

[0088] The free gas volume was estimated by pressure release tests under laboratory conditions.

[0089] 3) Data analysis and prediction:

[0090] The experimental data were input into a pre-established shale gas prediction model.

[0091] The model is used to calculate the shale gas resources of different lithofacies.

[0092] The estimated density of gas in the lithofacies is adjusted to account for nanoconfinement effects.

[0093] 4) Total resources:

[0094] The shale gas resources of all lithofacies are added up to obtain the total resources of the entire core sample.

[0095] The core data are extrapolated to the entire shale layer to estimate the shale gas resources in the entire area.

[0096] Example 2: Shale gas resource prediction based on well logging and geophysical data

[0097] 1) Well logging and geophysical data collection:

[0098] Well logging and geophysical data of shale formations are collected through seismic surveys and downhole logging techniques.

[0099] Well log and geophysical data are analyzed to identify the distribution and characteristics of different lithofacies.

[0100] 2) Identification and classification of lithofacies types:

[0101] Machine learning algorithms are used to process and interpret logging and geophysical data, and combined with sedimentary structures, shale components, mineral structure and content, organic matter abundance, elemental characteristics, and field emission scanning electron microscopy results to comprehensively identify different lithofacies types.

[0102] The lithology information is combined to classify and analyze the characteristics.

[0103] Identifying shale lithofacies is a complex process involving knowledge and technology from multiple fields such as geology, geophysics, geochemistry, and computer science. The following are the specific steps for implementing a technical solution:

[0104] 1. Data collection and preprocessing

[0105] Core collection: Collect core samples from the target shale layer to ensure representativeness and diversity of the samples.

[0106] Laboratory analysis: Systematic macro and micro analysis of core samples, including sedimentary structure, shale composition, mineral structure and content, organic matter abundance, elemental characteristics, field emission scanning electron microscopy analysis, etc.

[0107] Data preprocessing: Clean and standardize the collected data to ensure data quality.

[0108] 2. Identification of lithofacies types

[0109] Geological characteristics combined with image analysis technology: Based on the laboratory analysis results, high-resolution microscopes and image analysis software are used to observe and analyze thin sections and scanning electron microscopes of core samples, and different lithofacies types are identified and divided from macroscopic, microscopic, mineral, and chemical perspectives. For example, massive mud mixed shale, layered mud mixed shale, laminated mud mixed shale, laminated mud mixed shale, laminated mud mixed shale, massive silica mud mixed shale, layered silica mud mixed shale, laminated silica mud mixed shale, laminated silica mud mixed shale, massive silica mud mixed shale, layered silica mud mixed shale, laminated silica mud mixed shale, laminated silica mud mixed shale, and laminated silica mud mixed shale are comprehensively distinguished through sedimentary structure, shale components, mineral structure and content, organic matter abundance, element characteristics, and field emission scanning electron microscope results. For example, the sedimentary structure of massive ash mud mixed shale is mainly massive, and the shale components are mainly ash (>40%) and clay (>40%), and they are roughly uniform mixed structures in macro and microscopic terms, and the organic matter abundance is generally less than 2%, and the elements are characterized by enrichment of Ca, Al, Ti, and Fe; while the sedimentary structure of laminar silica mud mixed shale is mainly laminar, and siliceous-clay layer coupling and ash-clay layer coupling can be seen, and these layer couplings are superimposed and developed. The shale components are mainly siliceous (mainly biogenic silicon, >30%), ash (>30%), and clay (>30%), and the organic matter abundance is generally 2%-3%, and the elements are characterized by enrichment of Si, Ca, Al, V, and Cu. Introducing elemental characteristics and field emission scanning electron microscopy results, combining sedimentary structure, shale components, mineral structure and content, and organic matter abundance to comprehensively identify lithofacies types is a feature different from conventional shale lithofacies identification and classification.

[0110] Geophysical logging data analysis: Combined with geophysical logging data (such as gamma, density, acoustic wave time difference, etc.), the logging curve characteristics are used to further identify and distinguish the lithofacies types. Among them, in a certain study area, silica mud mixed shale generally has high gamma, low density, and high acoustic wave, lime mud mixed shale generally has low gamma, high density, and low acoustic wave; silica mud mixed shale generally has medium gamma, medium density, and medium acoustic wave.

[0111] 3. Machine Learning and Pattern Recognition

[0112] Feature extraction: Extract key features for lithofacies identification from geological, geophysical, and geochemical data.

[0113] Machine learning model training: The extracted features are used to train machine learning models, such as logistic regression models, support vector machines (SVM), random forests (RF), convolutional neural networks (CNN), etc., to achieve automated lithofacies type identification.

[0114] Model verification and optimization: Through cross-validation and actual application feedback, the model parameters are continuously optimized to improve the recognition accuracy.

[0115] 4. Comprehensive analysis and application

[0116] Integrated geological model: Combine the lithofacies types identified by machine learning with the geological model to construct a more detailed shale lithofacies distribution map.

[0117] Gas content prediction: Based on the gas content characteristics of different lithofacies types and combined with the quantitative analysis of adsorbed gas and free gas, the total shale gas resources are predicted.

[0118] Exploration and development decision support: Apply the results of lithofacies type identification to shale gas exploration and development decisions to improve exploration efficiency and development success rate.

[0119] Through the above technical solution, accurate identification of different centimeter-level lithofacies types in shale can be achieved, providing key geological information support for the prediction of shale gas resources.

[0120] Among them, the application of machine learning and pattern recognition technology in shale lithofacies type identification can adopt the following specific model architecture and training process:

[0121] Model Architecture

[0122] 1. Logistic regression is a commonly used classification algorithm, especially suitable for binary classification problems. In lithofacies type identification, if only two lithofacies types (for example, mud shale and siliceous shale) are considered for identification each time, logistic regression can be a suitable choice. Sample screening and multiple training can be performed according to the preset type proportion.

[0123] 2. Support Vector Machine (SVM)

[0124] Model principle: SVM is a supervised learning method used for classification and regression analysis. It distinguishes different data categories by finding an optimal hyperplane to maximize the interval between data points of different categories.

[0125] Applicability: It is suitable for classification problems with small datasets, especially when the data feature dimension is high.

[0126] 3. Random Forest (RF)

[0127] Model principle: RF is an integrated learning method consisting of multiple decision trees. Each decision tree is a classifier, and the final classification result is determined by voting or averaging.

[0128] Applicability: It is suitable for classification problems of large-scale data sets, can handle high-dimensional data, and is not prone to overfitting.

[0129] 4. Convolutional Neural Network (CNN)

[0130] Model principle: CNN is a deep learning model that is particularly suitable for processing image data. It extracts image features and classifies them through convolutional layers, pooling layers, and fully connected layers.

[0131] Applicability: Suitable for classification problems of image data, such as identifying lithofacies types through core thin section images.

[0132] Machine Learning and Training Process

[0133] Data preparation

[0134] Feature extraction: Extract key features from geological, geophysical and geochemical data, such as mineral content, porosity, permeability, well log characteristics, etc.

[0135] Data partitioning: Divide the dataset into training and test sets, usually in a ratio of 70% training set and 30% test set.

[0136] Model Training

[0137] Parameter selection: Choose appropriate parameters for each model, such as shale type for logistic regression model, number of layers for CNN, and number of neurons in each layer.

[0138] Training process: Use the training set data to train the model, and iteratively optimize the model parameters to achieve the best performance on the training set.

[0139] Model validation and optimization

[0140] Cross-validation: Use cross-validation methods (such as k-fold cross-validation) to evaluate the generalization ability of the model.

[0141] Performance evaluation: The classification performance of the model is evaluated through indicators such as accuracy, recall, and F1 score.

[0142] Model optimization: Adjust model parameters based on validation results or try different model architectures to improve model performance.

[0143] Model Application

[0144] Prediction and Identification: Use the trained model to identify the lithofacies type of new shale samples.

[0145] Result analysis: Analyze the identification results of the model and interpret and verify them in combination with geological knowledge.

[0146] Through the above-mentioned machine learning and training process, the automatic identification of shale lithofacies types can be effectively realized, and the accuracy and efficiency of shale gas resource prediction can be improved.

[0147] Among them, the training model is as follows:

[0148] The lithofacies type identification is performed based on the machine learning method and the logistic regression model is used for lithofacies type identification.

[0149] The basic formula of the logistic regression model is as follows:

[0150] P(y=1|x;θ)=(1+e -θTx ) －1

[0151] in:

[0152] P(y=1|x;θ) is the probability that a sample belongs to category 1 (eg, siliceous shale) given the feature vector x and the parameter θ.

[0153] x is the input feature vector, which contains various characteristics used to identify lithofacies, such as mineral content, porosity, permeability, etc.

[0154] θ is the parameter vector of the model, which needs to be learned through training data.

[0155] e is the base of natural logarithms.

[0156] Parameter meaning

[0157] The feature vector of the x input contains multiple eigenvalues, each of which represents an attribute of the shale sample. For example, x 1， Might represent mineral content, x2 might represent porosity, and so on.

[0158] θ: The parameter vector of the model, which contains multiple weight values, each of which corresponds to a feature. These weight values ​​determine the contribution of each feature to the classification result. Through training data, the model will learn the optimal weight value.

[0159] P(y=1|x;θ) is the probability that a sample belongs to category 1 given the feature vector x and parameter θ. This probability value is between 0 and 1 and can be used to determine which category a sample belongs to.

[0160] Training process

[0161] In practice, the training process of the logistic regression model involves using the training data to optimize the parameter θ. This is usually achieved by maximizing the likelihood function or minimizing the logarithmic loss function. Once the training is complete, the model can be used to identify the lithofacies type of new shale samples.

[0162] In practical applications, facies type identification may involve multi-classification problems. In this case, multi-classification logistic regression or other machine learning models suitable for multi-classification, such as random forest or convolutional neural network, can be used.

[0163] The logistic regression model assumes that the features are independent of each other, which may not be true in actual geological data. Therefore, the characteristics of the data need to be considered when selecting the model. The performance of the model needs to be evaluated and optimized through cross-validation and practical application.

[0164] Through the above-mentioned logistic regression model formula and training process, the basic identification of shale lithofacies types can be achieved. However, the model in actual application may be more complex and needs to be selected and optimized according to specific problems and data characteristics.

[0165] Convolutional neural network (CNN) is a deep learning model that is particularly suitable for processing image data. In lithofacies type identification, if CNN is used, thin-section images of core samples or other types of image data are usually processed. CNN extracts image features and classifies them through multiple layers of convolutional layers, activation functions, pooling layers, and fully connected layers.

[0166] The basic structure of CNN

[0167] The CNN structure in this embodiment includes the following parts:

[0168] Convolutional Layer: Extracts local features of the image through convolution operations.

[0169] Activation Function: Such as ReLU function, which introduces nonlinear factors to enhance the expressiveness of the model.

[0170] Pooling Layer: Usually the maximum pooling, which is used to reduce the dimension of the feature map, reduce the number of parameters and the amount of calculation, while retaining important features.

[0171] Fully Connected Layer: Summarizes the features extracted by the previous convolution and pooling layers and outputs the classification results.

[0172] CNN formula:

[0173]

[0174] in:

[0175] y i，j，k is an element in the output feature map.

[0176] x i+m，j+n，k′ is an element in the input feature map.

[0177] w m，n，k′ It is a weight of the convolution kernel (filter).

[0178] b k is the bias term.

[0179] M and N are the sizes of the convolution kernel.

[0180] k and k′ are the channel indices of the feature maps.

[0181] Training process

[0182] In the training process of CNN, the backpropagation algorithm and stochastic gradient descent (SGD) or other optimization algorithms are usually used to update the weights and bias terms in the network to minimize the loss function. The loss function can be cross entropy loss (for classification problems) or other types of loss functions.

[0183] Precautions

[0184] The training of CNN requires a large amount of labeled data, that is, image data with lithology type labels.

[0185] During the training process, data needs to be preprocessed, such as normalization and enhancement, to improve the generalization ability of the model.

[0186] The hyperparameters of the model (such as learning rate, batch size, convolution kernel size, number of layers, etc.) need to be adjusted and optimized through experiments.

[0187] Through the above basic structure and training process of CNN, automatic recognition of lithofacies types can be achieved.

[0188] Using both the Logistic Regression (LR) model and the Convolutional Neural Network (CNN) model for training and data recognition usually means that the two models are used to solve different aspects of the same problem or process different types of data. For example, LR may be used to process non-image feature data, such as geological parameters, while CNN is used to process image data, such as core slice images. In this case, the cross-validation process needs to be performed separately for each model while ensuring that the performance of the entire system is effectively evaluated.

[0189] Steps of Cross Validation

[0190] Data partitioning: First, divide the data set into multiple mutually exclusive subsets. Usually, k-fold cross-validation is used to divide the data into k folds of similar size.

[0191] Model training: For each model (LR and CNN), train on k-1 folds and then validate on the remaining fold. This process is repeated k times, using a different fold as the validation set each time.

[0192] Performance evaluation: For each model, calculate the performance indicators on the validation set, such as accuracy, recall, F1 score, etc.

[0193] Model integration: If the goal is to combine the prediction results of LR and CNN, the prediction results of the two models can be integrated (such as voting, weighted averaging, etc.) in each iteration of cross-validation, and the performance of the integrated model can be evaluated on the validation set.

[0194] Results Summary: Finally, the performance metrics of each model across all iterations are averaged to obtain an overall performance evaluation of the model.

[0195] The following points should be ensured:

[0196] Data consistency: Ensure that the data used for training and validation is consistent between the two models so that the performance of the ensemble model can be properly evaluated.

[0197] Model independence: The training of LR and CNN should be independent, and each model should have its own training and validation process.

[0198] Ensemble strategy: Choose a suitable ensemble strategy, such as hard voting (majority voting), soft voting (probability averaging), etc., to combine the prediction results of the two models.

[0199] Hyperparameter tuning: During the cross-validation process, the model's hyperparameters can be tuned simultaneously to find the best model configuration.

[0200] Through the above steps, the system that uses both the logistic regression model and the convolutional neural network can be effectively cross-validated to ensure that the generalization ability of the model and the overall performance of the system are accurately evaluated.

[0201] 3) Resource calculation model:

[0202] A comprehensive model was constructed that combined well log and geophysical data with known information on the gas content of the lithofacies.

[0203] Taking into account the nanoconfinement effect, the model is adjusted to more accurately predict the gas content in each lithofacies.

[0204] Nanoconfinement Effect refers to the phenomenon that the physical and chemical properties of substances change significantly in nanoscale pores or spaces. In shale gas reservoirs, gas mainly exists in the nanoscale pores of shale in adsorbed and free states. Due to the very small spatial size of these pores, the interaction between gas molecules and pore walls is significantly enhanced, resulting in gas adsorption, diffusion and flow behaviors that are very different from those in conventional porous media. The nanoconfinement effect hinders the generation of the gas phase, thereby reducing the actual production of shale gas. And sometimes it is dominated by nanopores. The model needs to be able to accurately describe the distribution and connectivity of nanopores, as well as their effects on gas adsorption and flow.

[0205] 2. Adsorption behavior: In nanopores, the adsorption behavior of gases usually follows laws that are different from conventional adsorption theory. For example, the Langmuir isotherm adsorption model may need to be modified or replaced with an adsorption model that is more suitable for the nanoscale, such as the Dubinin-Astakhov model or the Brunauer-Emmett-Teller (BET) model.

[0206] 3. Diffusion mechanism: In nanopores, the diffusion mechanism of gas may change from conventional Knudsen diffusion to surface diffusion or transition diffusion. The model needs to consider these different diffusion mechanisms and correctly describe their effects on gas migration.

[0207] 4. Flow characteristics: At the nanoscale, the flow of gas may change from Darcy flow to non-Darcy flow, or even show slip flow effect. The model needs to be able to describe these complex flow behaviors and accurately predict the output characteristics of the gas.

[0208] 5. Multiphase flow and phase changes: In nanopores, phase changes of gases (such as liquefaction or condensation) may occur more easily, which will affect the content and migration of gases. The model needs to take these phase changes into account and correctly simulate their effects on the gas content.

[0209] By considering the nanoconfinement effect in the model, the gas content of each facies in the shale gas reservoir can be predicted more accurately, thereby improving the accuracy and efficiency of shale gas resource evaluation and development. This usually requires a combination of experimental data, field tests and advanced numerical simulation techniques.

[0210] In this scheme, the coefficient of reducing the actual gas production of shale gas by the nanoconfinement effect is determined by the following steps:

[0211] 1. Field data collection: First, it is necessary to collect field production data of multiple wells in the study area, including gas production, pressure, temperature, etc. These data are the basis for determining the nanoconfinement effect coefficient.

[0212] 2. Experimental research: Conduct laboratory-scale experiments to simulate the nanopore structure of shale gas reservoirs and measure the gas adsorption, diffusion rate and flow characteristics under different conditions. These experiments can help understand the impact of nanoconfinement effects on gas migration.

[0213] 3. Numerical simulation: Use numerical simulation software for shale gas reservoirs to establish a geological model and fluid flow model for the study area. In the model, traditional gas production prediction methods are usually used, such as flow models based on Darcy's law. Through dynamic simulation of shale gas production of typical wells in the east and west of the study area, the shale gas production with and without nanoconfinement effect tends to be stable after 180 days. At this time, it is shown that after considering the nanoconfinement effect, the cumulative shale gas production of Well A in the east and Well B in the west is 87% and 93% of that without considering the nanoconfinement effect, respectively. Therefore, when calculating shale gas production, it is necessary to add the coefficient of the nanoconfinement effect to the actual shale gas production after the traditional calculation formula, that is, this patent obtains a coefficient of 0.87 in the east of the study area and a coefficient of 0.93 in the west of the study area based on typical drilling analysis.

[0214] 4. Model calibration: Compare the field production data with the numerical simulation results and adjust the model parameters so that the simulated gas production matches the actual gas production. In this process, it may be found that the traditional model cannot fully explain the actual gas production, especially when considering the nano-confinement effect.

[0215] 5. Introduce nanoconfinement effect coefficient: In order to explain the influence of nanoconfinement effect on gas production, a correction coefficient can be introduced into the numerical model. This coefficient reflects the reduction of gas flow capacity in nanopores. By adjusting this coefficient, the simulation results can be better consistent with the actual production data.

[0216] 6. Regional difference analysis: Due to differences in geological conditions, the impact of nanoconfinement effect on gas production may vary in different regions. Therefore, it is necessary to conduct a detailed analysis of the geological characteristics, pore structure, pressure system, etc. in the east and west of the study area to determine the nanoconfinement effect coefficients in different regions.

[0217] 7. Verification and adjustment: Use independent field data to verify the determined nanoconfinement effect coefficient. If the verification results show that the gas production predicted by the coefficient deviates significantly from the actual gas production, the coefficient may need to be further adjusted.

[0218] 8. Uncertainty analysis: Uncertainty analysis is performed to evaluate the impact of the uncertainty of the nanoconfinement effect coefficients on the model prediction results. This helps to understand the reliability and scope of application of the coefficients.

[0219] Through the above steps, the nanoconfinement effect coefficients in the eastern and western parts of the study area can be determined and applied to the gas production prediction of shale gas reservoirs to improve the accuracy of the prediction.

[0220] 4) Total resource estimation:

[0221] The model is used to estimate the shale gas resources throughout the shale formation.

[0222] The prediction results are corrected and verified by combining regional geological and engineering data.

[0223] Application Example 3: Regional Shale Gas Exploration Assessment

[0224] A national energy company plans to assess the shale gas potential of a certain area to decide whether to make further exploration investments.

[0225] 1) Data collection: First, shale samples are collected from the selected area to determine the adsorbed gas and free gas content, and relevant geological parameters (such as gas-bearing area, shale thickness, density, etc.) are collected.

[0226] 2) Laboratory analysis: Parameters such as Langmuir volume and pressure, formation pressure, porosity, gas saturation, and natural gas compressibility factor are measured in the laboratory.

[0227] 3) Resource calculation: Apply the formulas described in claims 4 and 5 to calculate the total resources of adsorbed gas and free gas, and then obtain the total resources of shale gas in the area.

[0228] 4) Evaluation report: Based on the calculation results, a detailed evaluation report is prepared, including resource forecast, development potential and economic feasibility analysis.

[0229] 5) Decision support: The report will provide decision support for energy companies on whether to conduct shale gas exploration and development in the region.

[0230] The calculation results obtained by the embodiments of the present application are more accurate. An intuitive comparison with the contents recorded in the comparative document (Shale gas resource evaluation method and evaluation parameter assignment - taking the Wufeng Formation-Longmaxi Formation in the Central Yangtze Region as an example) shows that the two are quite different. This also confirms that the technical solution of the present application can better solve the technical problems existing in the prior art and provide higher accuracy of the calculation results. Moreover, the prediction of favorable shale gas areas using the technical solutions of the comparative documents has a large deviation. In the same target area (the study area of ​​this application overlaps with the Central Yangtze area of ​​the comparative documents in eastern Sichuan and western Hubei), they did not identify the important favorable area of ​​Fengdu block. However, by implementing the solution of this application in actual projects, the shale gas resource volume was calculated based on the centimeter-level lithofacies fine division and type zoning. It was found that the main developed lithofacies in the area are laminar silica mud mixed shale, foliated silica mud mixed shale, and foliated silica mud mixed shale, and they all have high gas content. The favorable area of ​​Fengdu was accurately found, and the potential natural gas resource volume of the Wufeng Formation-Longmaxi Formation in the Fengdu block is 3100.58*10 8 m 3 , estimated resource volume 2300.31*10 8 m 3 The shale gas exploration and development potential is huge and has huge resources, which has been confirmed by field production practice. This obviously shows that the present application is a significant improvement over the prior art, or it can be determined that the present application has brought unexpected technical effects. This example can also prove that the present invention has a very high economic value and has been successful in business.

[0231] Application Example 4: Shale Gas Field Development Planning

[0232] A shale gas development company is preparing to make a detailed development plan for the proven shale gas fields.

[0233] 1) Parameter determination: Conduct comprehensive geological exploration of shale gas fields to determine the adsorbed gas and free gas content of shale, as well as other related parameters.

[0234] 2) Numerical simulation: By establishing geological and engineering numerical models, the development process of shale gas fields is simulated and the gas production under different development scenarios is calculated.

[0235] 3) Resource prediction: Based on Formula 2 described in claim 6 and combined with numerical simulation results, the total shale gas resources under different mining conditions are predicted.

[0236] 4) Development optimization: Based on the prediction results, optimize the development plan, including well network layout, fracturing design and production rhythm.

[0237] 5) Economic evaluation: Based on market conditions, evaluate the economic feasibility of shale gas development to ensure the project's return on investment.

[0238] Through these two embodiments, the shale gas resource prediction method can help different stakeholders to conduct scientific resource assessment and decision-making, and provide strong data support for shale gas exploration and development.

[0239] The prediction of shale gas resources is a complex task, which usually involves geological exploration, rock physics analysis, chemical analysis, and mathematical and statistical models. Two specific embodiments and corresponding implementation schemes are listed below.

[0240] Example 1 A quantitative assessment of shale gas resources in a specific area is performed.

[0241] 1) Geological survey:

[0242] Conduct geological exploration of the area, including core sampling, and analysis of surface and subsurface geological maps.

[0243] Determine the extent of the shale gas reservoir, i.e., the gas-bearing area (A) and the thickness of the shale (h).

[0244] 2) Physical and chemical analysis:

[0245] Shale samples are collected and physical measurements are performed in the laboratory to obtain the density (ρ) of the shale.

[0246] Porosity (Φg) and gas saturation (Sg) are obtained through core analysis or geophysical logging data.

[0247] The Langmuir volume (VL) and Langmuir pressure (PL) were obtained using laboratory adsorption / desorption experiments.

[0248] 3) Resource calculation:

[0249] Calculate the adsorption gas content (q 吸 ), based on the VL and PL measured in the laboratory, combined with the formation pressure (P), using the formula 'q 吸 =VL·P / (PL+P)`.

[0250] Gas injection and measurement: Place the processed core sample in an autoclave and inject a certain amount of shale gas (usually methane or a mixture of methane and other gases). Under the set temperature and pressure conditions, the gas and the core sample are fully in contact and reach a state of equilibrium. Use a gas analyzer to measure the volume, pressure, temperature and other parameters of the gas in the autoclave.

[0251] Calculation of compression factor (Z): Based on experimentally measured data, the gas state equation (Setzmann-Wagner equation) is used to calculate the compression factor of the actual gas, taking the formation temperature, pressure and gas composition as input parameters. The compression factor (Z) is a measure of the deviation between the actual gas state and the ideal gas state, and is usually defined by comparing the ratio of the actual gas volume to the ideal gas volume.

[0252] Lithofacies difference analysis: Repeat the above experimental process for shale samples of different lithofacies. Different types of shale lithofacies have different compression factor values, for example, Z = 0.75 for massive ash mixed shale, Z = 0.78 for layered ash mixed shale, Z = 0.83 for laminated ash mixed shale, Z = 0.85 for foliated ash mixed shale, Z = 0.82 for massive silica mixed shale, Z = 0.87 for layered silica mixed shale, Z = 0.91 for laminated silica mixed shale, and Z = 0.94 for laminated silica mixed shale.

[0253] Calculate the free gas content (q 游 ), which needs to be obtained through desorption method, isothermal adsorption method, etc.

[0254] Apply the formula 'Q 总 =0.01·A·h·ρ·(q 吸 +q 游 ) / Z' to calculate the total shale gas resources.

[0255] 4) Data verification and model correction:

[0256] The estimated resource volumes are verified through field testing and production data.

[0257] If necessary, model parameters are modified to improve prediction accuracy.

[0258] Example 2 uses statistical and simulation methods to predict shale gas resources.

[0259] 1) Data collection and preprocessing:

[0260] Collect historical geological, geophysical and production data for the area.

[0261] Perform data cleaning to ensure that the data quality is suitable for statistical analysis.

[0262] 2) Statistical analysis:

[0263] Statistical methods, such as regression analysis, are used to determine the relationship between different geological parameters and shale gas production.

[0264] Develop a statistical prediction model to estimate the resource volume in unknown areas.

[0265] 3) Model simulation:

[0266] The shale gas reservoirs are stochastically simulated using geostatistical methods such as Kriging or Monte Carlo simulation.

[0267] By simulating multiple scenarios, the impact of the uncertainty of each parameter on resource prediction is estimated.

[0268] 4) Resource estimation:

[0269] The statistical model and simulation results are combined to estimate the resources of adsorbed gas and free gas.

[0270] Calculate the dissolved gas resources (Q 溶 ), if supported by necessary data.

[0271] Apply the modified formula 'Q 总 =Q 吸 +Q 游 +Q 溶 'Conduct final resource estimate.

[0272] 5) Verification and optimization:

[0273] The model is validated a posteriori and compared with actual production data to check the accuracy of the prediction.

[0274] The model parameters are adjusted according to the validation results to achieve higher prediction accuracy.

[0275] Please refer to the drawings, in which the same component symbols represent the same components, and the principle of the present invention is illustrated by implementing it in an appropriate computing environment. The following description is based on the illustrated specific embodiment of the present invention, which should not be considered as limiting other specific embodiments of the present invention that are not described in detail herein.

[0276] See also Figure 2 In a shale gas resource prediction method according to an embodiment of the present invention, the following steps are included:

[0277] Step S11: obtaining the total amount of adsorbed gas contained in shale of a preset mass.

[0278] Step S12: Obtain the total amount of free gas contained in shale of a preset mass.

[0279] Step S13: Calculate the total shale gas resources contained in the shale of a preset mass according to the total amount of adsorbed gas and the total amount of free gas.

[0280] The total shale gas resources are the total amount of natural gas contained in shale of a given scale (quality) after it is extracted to the surface. Among them, the total shale gas resources can be approximately decomposed into the total amount of adsorbed gas and the total amount of free gas. Therefore, the total shale gas resources are obtained by calculating the sum of the total amount of adsorbed gas and the total amount of free gas.

[0281] Therefore, by comprehensively considering the adsorbed gas content and the free gas content, the total shale gas resources can be calculated, which can make the prediction results more accurate and realize the prediction of shale gas resources in low-exploration areas (such as the eastern Sichuan-western Hubei region).

[0282] In one embodiment, the total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources.

[0283] The total shale gas resources contained in the preset mass of shale are calculated according to the total amount of adsorbed gas and the total amount of free gas, specifically, the total shale gas resources contained in the preset mass of shale are calculated using the following formula 1:

[0284] Q 总 =Q 吸 +Q 游 (one)

[0285] Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 Represents the total free gas resources.

[0286] Among them, shale also contains dissolved gas, but the content of dissolved gas in shale is very small, so it can be ignored. In order to improve the accuracy of calculation, the dissolved gas resource can also be calculated. Further, the method also includes: obtaining the dissolved gas resource contained in the shale of preset quality; the formula 1 is further: Q 总 =Q 吸 +Q 游 +Q 溶 , where Q 溶 is the amount of dissolved gas resources.

[0287] Among them, the Q 吸 =0.01·A·h·ρ·q 吸 , where A represents the gas-bearing area of ​​shale, h represents the thickness of shale, ρ represents the density of shale, and q 吸 is the adsorption gas content; further, q 吸 =VL·P / (PL+P), VL represents the Langmuir volume, P represents the formation pressure, and PL represents the Langmuir pressure.

[0288] Among them, the Q 游=0.01·A·h·Φg·Sg / Z, where Φg represents porosity, Sg represents gas saturation, and Z represents natural gas compressibility factor.

[0289] Therefore, without considering the amount of dissolved gas resources, the total shale gas resources Q 总 for:

[0290] Q 总 =Q 吸 +Q 游 =0.01·A·h·(ρ·VL·P / (PL+P)+Φg·Sg / Z)

[0291] Where Z≤1. Adsorption gas content q 吸 It can be obtained by one of the following methods: desorption method, isothermal adsorption method, analogy method, statistical method, logging interpretation and calculation method. Among them, when the adsorbed gas content q is obtained by the field desorption method 吸 When the volume change of natural gas from underground to the surface due to the change of temperature and pressure conditions is included, the compression factor Z in the above formula is equal to 1; when other methods are used and the volume change caused by the change of temperature and pressure conditions is not taken into account, the adsorbed gas content obtained needs to be obtained through plates or other methods, and the compression factor Z is less than 1. Usually, the compression factor Z is less than 1.

[0292] When calculating the total shale gas resources in the eastern Sichuan-western Hubei region, the formation pressure P can be calculated by the empirical formula of the eastern Sichuan-western Hubei region, that is, P = 17.86 + 0.0014 (H-414.9), where H is the depth value corresponding to the formation pressure P. In the embodiment of the present invention, the depth value H is taken as 2132m, then the formation pressure P = 20.26394MPa. In addition, the actual measured gas saturation Sg and porosity Φg are 23.11% and 2.74%, respectively. According to isothermal adsorption simulation, VL = 2.17m 3 , PL=2.09MPa, ρ=2.73t / m 3 By analyzing the analyzed gas, lost gas and total gas content in the coring section of PY1 well, the compression factor Z=0.928 was determined.

[0293] Using the above-determined formation pressure P, gas saturation Sg, porosity Φg, Langmuir volume VL, Langmuir pressure PL, and shale density ρ, the total shale gas resources in the eastern Sichuan-western Hubei region are calculated as follows:

[0294] The shale gas of Wufeng Formation-Longmaxi Formation in East Sichuan-West Hubei is mainly enriched in the lower sequence Sq1, so the resource calculation is mainly carried out for the lower sequence Sq1. Since there are three synclines developed in the work area, and the effective high-quality shale is also mainly distributed in these three synclines, the three synclines are taken as three independent units for resource calculation. The three synclines are Sangzeping syncline, Wandi syncline and Daozhen syncline. According to the statistics of shale thickness and gas-bearing area in the transgressive system domain and high-stand system domain of Sangzeping syncline, Wandi syncline and Daozhen syncline, the specific resource calculation results of the three synclines are as follows:

[0295]

[0296] Table 2 Resources of Sangzeping Syncline

[0297]

[0298] Table 3 Bay area complex syncline resources

[0299]

[0300] Table 4 True syncline resources

[0301] According to the determined formation pressure P, gas saturation Sg, porosity Φg, Langmuir volume VL, Langmuir pressure PL, shale density ρ, and shale thickness h and gas-bearing area A in the above tables, the formula Q is used to calculate the gas saturation. 总 =0.01·A·h·(ρ·VL·P / (PL+P)+Φg·Sg / Z) The total resources of each syncline can be calculated separately. The total resources of the three synclines are added together to obtain the total shale gas resources Q of the Wufeng Formation-Longmaxi Formation in the eastern Sichuan-western Hubei region. 总 Up to 14175.36×10 8 m 3 .

[0302] Therefore, the shale gas resource prediction method of the present invention can predict the total shale gas resources in the low-exploration eastern Sichuan-western Hubei region.

[0303] In other embodiments, the total shale gas resources may also be calculated using the following formula 2. Specifically, in another embodiment, the total amount of adsorbed gas is the adsorbed gas content, and the total amount of free gas is the free gas content. Based on the total amount of adsorbed gas and the total amount of free gas, the total amount of shale gas resources contained in the shale of a preset mass is calculated, including:

[0304] The total shale gas resources contained in the shale of the preset mass are calculated using Formula 2:

[0305] Q 总= 0.01·A·h·ρ·(q 吸 + q 游 ) / Z (two)

[0306] Among them, Q 总 represents the total shale gas resource amount, A represents the gas-bearing area of the shale, h represents the thickness of the shale, ρ represents the density of the shale, Z represents the natural gas compression factor, and q 吸 represents the adsorbed gas content, and q 游 represents the free gas content. It should be noted that the parameters with the same symbol in the present invention have the same meaning. For example, the meanings of A, h, Z, q 吸 in Formula (two) are the same as those in the above embodiments, and will not be elaborated one by one here. Among them, q 吸 and q 游 can both be obtained by one of the desorption method, isothermal adsorption method, analogy method, statistical method, logging interpretation and calculation method.

[0307] The prediction method of the present invention takes into account the influence of underground temperature and pressure. When the adsorbed gas content and free gas content are obtained by the in-situ desorption method, the volume change caused by the change of temperature and pressure from underground to the surface has been included. At this time, the compression factor Z in the calculation formula is equal to 1, which further simplifies the calculation process and improves the calculation accuracy.

[0308] Through the prediction method provided by the present invention, it can be obtained that the total shale gas resource amount of the Wufeng Formation - Longmaxi Formation in the eastern Sichuan - western Hubei region with low exploration degree can reach 14175.36 × 10 8 m 3 .

[0309] Moreover, by applying the shale gas resource amount prediction method provided by the present invention to the shale gas resource exploration work, the acquisition methods of the basic calculation data are diverse and can be flexibly selected according to the actual situation on site. For example, methods such as the desorption method, isothermal adsorption method, analogy method, statistical method, logging interpretation, etc. can all be effectively used to predict the shale resource amount. Compared with other prediction methods that require large-scale destructive sampling, the present invention requires less capital investment and causes less damage to the environment.

[0310] Referring to Figure 3 , the embodiment of the present invention also provides a shale gas resource amount prediction device, including:

[0311] The first acquisition module 201 is used to acquire the total amount of adsorbed gas contained in the shale with a preset mass;

[0312] The second acquisition module 202 is used to acquire the total amount of free gas contained in the shale with a preset mass;

[0313] The calculation module 203 is used to calculate the total shale gas resources contained in the shale of a preset mass according to the total amount of adsorbed gas and the total amount of free gas.

[0314] The total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources; the calculation module 203 is specifically used to calculate the total amount of shale gas resources contained in the shale of the preset quality using formula 1:

[0315] Q 总 =Q 吸 +Q 游 (one)

[0316] Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 Represents the total free gas resources.

[0317] The calculation module 203 is specifically used to use the formula Q 吸 =0.01·A·h·ρ·q 吸 Calculate the total adsorbed gas resources Q 吸 , where A represents the gas-bearing area of ​​shale, h represents the thickness of shale, ρ represents the density of shale, and q 吸 is the adsorption gas content; where q 吸 =VL·P / (PL+P), VL represents the Langmuir volume, P represents the formation pressure, and PL represents the Langmuir pressure. The calculation module 203 is also used to use the formula Q 游 =0.01·A·h·Φg·Sg / Z to calculate the total free gas resources Q 游 , where Φg represents porosity, Sg represents gas saturation, and Z represents natural gas compressibility factor.

[0318] An embodiment of the present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the shale gas resource prediction method described in the above embodiments.

[0319] It should be noted that the embodiments of the present invention can be implemented by hardware, software, or a combination of software and hardware. The hardware part can be implemented using dedicated logic; the software part can be stored in a memory and executed by an appropriate instruction execution system, such as a microprocessor or dedicated design hardware. A person of ordinary skill in the art will understand that the above-mentioned devices and methods can be implemented using computer executable instructions and / or contained in a processor control code, such as a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. Such code is provided on a carrier medium such as a disk, CD or DVD-ROM, a programmable memory such as a read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The device and its modules of the present invention can be implemented by hardware circuits such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., can also be implemented by software executed by various types of processors, and can also be implemented by a combination of the above-mentioned hardware circuits and software, such as firmware.

[0320] Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A shale gas resource prediction method, characterized in that: The following steps are involved: Based on the system's identification of centimeter-level lithofacies types of shale, the total amount of adsorbed gas contained in different types of lithofacies in a preset mass of shale is obtained, and the total amount of free gas contained in different types of lithofacies in a preset mass of shale is obtained; According to the total amount of adsorbed gas and free gas contained in the different types of shale phases, the total shale gas resources contained in each type of phase in the preset mass of shale are calculated respectively, and the density changes of adsorbed gas and free gas in different types of phases in the nano-confined space caused by the nano-confined effect are considered and determined; the total shale gas resources contained in all types of phases are added together to calculate the total shale gas resources contained in the preset mass of shale.

2. The method according to claim 1, characterized in that The lithofacies type refers to the different physical and chemical properties in shale, including color, particle size distribution and degree of sorting, sedimentary structure and bedding structure, mineral composition and content, and geochemical characteristics.

3. The method according to claim 2, characterized in that Machine learning algorithms are used to process and interpret logging and geophysical data, and combined with sedimentary structures, shale components, mineral structure and content, organic matter abundance, elemental characteristics, and field emission scanning electron microscopy results to comprehensively identify different lithofacies types.

4. The method according to claim 3, characterized in that The system identifies centimeter-level shale lithofacies types including Data collection and preprocessing: Collect core samples from the target shale layer to ensure the representativeness and diversity of the samples, conduct systematic macroscopic and microscopic analysis of the core samples, including sedimentary structure, shale composition, mineral structure and content, organic matter abundance, element characteristics, and field emission scanning electron microscopy analysis, and clean and standardize the collected data to ensure data quality; Lithofacies type identification: Based on the laboratory analysis results, the thin sections and scanning electron microscopes of the core samples are observed and analyzed using high-resolution microscopes and image analysis software to identify and classify different lithofacies types from macroscopic, microscopic, mineral, and chemical perspectives; combined with geophysical logging data, the logging curve characteristics are used to further identify and distinguish lithofacies types; Machine Learning and Pattern Recognition: Key features for lithofacies identification are extracted from geological, geophysical and geochemical data. The extracted features are used to train machine learning models to achieve automated lithofacies type identification. Through cross-validation and practical application feedback, model parameters are optimized to improve identification accuracy.

5. The method according to claim 4, characterized in that Machine learning and pattern recognition include The logistic regression model combined with convolutional neural network was used for cross-validation to identify lithofacies types.

6. The method according to claim 5, characterized in that The basic formula of the logistic regression model is as follows: P(y=1∣x;θ)=(1+e -θTx ) －1 in: P(y=1|x; θ) is the probability that a sample belongs to category 1 given a feature vector x and parameter θ; x is the input feature vector, which contains various features used to identify lithofacies; θ is the parameter vector of the model, which is learned through training data; e is the base of the natural logarithm.

7. The method according to claim 6, characterized in that The convolutional neural network structure consists of the following parts: Convolution layer: extracts local features of the image through convolution operation; activation function: includes ReLU function, introduces nonlinear factors and enhances the expressiveness of the model; pooling layer: used to reduce the dimension of the feature map, reduce the number of parameters and the amount of calculation, while retaining important features; Fully connected layer: summarizes the features extracted by the previous convolution and pooling layers and outputs the classification results; The formula of convolutional neural network is: Where: y i，j，k is an element in the output feature map, x i+m，j+n，k′ is an element in the input feature map, w m，n，k′ is a weight of the convolution kernel, b k is the bias term, M and N are the sizes of the convolution kernels, and k and k′ are the channel indices of the feature maps; The convolutional neural network training process is: During the training process of CNN, the weights and bias terms in the network are updated using the back-propagation algorithm and stochastic gradient descent or other optimization algorithms to minimize the loss function, which is the cross entropy loss or other types of loss functions.

8. The method according to claim 7, characterized in that Cross validation involves the following steps First, divide the data set into multiple mutually exclusive subsets, usually using k-fold cross-validation to divide the data into k folds of similar size; For each model, we train it on k-1 folds and then validate it on the remaining fold. This process is repeated k times, each time using a different fold as the validation set; For each model, calculate the performance indicators on the validation set, including accuracy, recall, and F1 score; In each cross-validation iteration, the prediction results of the two models are integrated, and the performance of the integrated model is evaluated on the validation set; The performance metrics of each model across all iterations are averaged to obtain an overall performance evaluation of the model.

9. The method according to claim 8, characterized in that The total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources; The total shale gas resource contained in the shale of a preset mass is calculated according to the total amount of adsorbed gas and the total amount of free gas, including: The total shale gas resources contained in the shale of the preset quality are calculated using Formula 1: Q 总 =Q 吸 +Q 游 (one) Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 represents the total free gas resources; It also includes: obtaining the amount of solution gas resources contained in shale of a predetermined quality; The formula 1 is further: Q 总 =Q 吸 +Q 游 +Q 溶 The Q 溶 is the amount of dissolved gas resources; The Q 吸 =0.01·A·h·ρ·q 吸 , where A represents the gas-bearing area of ​​shale, h represents the thickness of shale, ρ represents the density of shale, and q 吸 is the adsorption gas content; Among them, q 吸 =VL·P / (PL+P), VL represents the Langmuir volume, P represents the formation pressure, and PL represents the Langmuir pressure; The Q 游 =0.01·A·h·Φg·Sg / Z, where Φg represents porosity, Sg represents gas saturation, and Z represents natural gas compressibility factor; The total amount of adsorbed gas is the adsorbed gas content, and the total amount of free gas is the free gas content; The total shale gas resource contained in the shale of a preset mass is calculated according to the total amount of adsorbed gas and the total amount of free gas, including: The total shale gas resources contained in the shale of the preset quality are calculated using Formula 2: Q 总 = 0.01·A·h·ρ·(q 吸 + q 游 ) / Z (two) Among them, Q 总 represents the total shale gas resources, A represents the gas-bearing area of ​​shale, h represents the thickness of shale, ρ represents the density of shale, Z represents the natural gas compression factor, q 吸 represents the adsorbed gas content, q 游 Represents the free gas content.

10. A shale gas resource prediction device based on the method according to any one of claims 1 to 9, characterized in that: include: A first acquisition module is used to acquire the total amount of adsorbed gas contained in shale of a preset mass; A second acquisition module is used to obtain the total amount of free gas contained in shale of a preset quality; A calculation module, used to calculate the total shale gas resources contained in shale of a preset mass according to the total amount of adsorbed gas and the total amount of free gas; The total amount of adsorbed gas is the total amount of adsorbed gas resources, and the total amount of free gas is the total amount of free gas resources; The calculation module is specifically used to calculate the total shale gas resources contained in the shale of the preset quality using Formula 1: Q 总 =Q 吸 +Q 游 (one) Where Q 总 represents the total shale gas resources, Q 吸 represents the total amount of adsorbed gas resources, Q 游 Represents the total free gas resources.

Images