Deposit lithology identification system and method for oil-gas exploration

Abstract

The invention provides a sedimentary body lithology identification system and method for oil-gas exploration, and the system comprises a data collection module which is used for collecting seismic data and logging data; the data processing module is used for preprocessing the seismic data and the logging data; the data fusion module is used for fusing the preprocessed seismic data, logging data and a geologic model to generate a comprehensive data set; the lithology identification module is used for analyzing the comprehensive data set by using a machine learning algorithm and identifying the lithology type of the sedimentary body; the result verification module is used for performing comparison verification on the identification result and actual core analysis data; and the display module is used for visualizing the recognition result. The system provided by the invention not only can identify the lithology of the sedimentary body, but also can generate a visual three-dimensional lithology distribution diagram through three-dimensional geological modeling and a visualization technology. The visualization results enable geologists and engineers to clearly understand the underground structure and lithologic distribution so as to make more accurate exploration decisions and development plans.

Description

Technical Field

[0001] The present invention relates to the technical field of oil and gas exploration, and in particular to a sedimentary body lithology identification system and method for oil and gas exploration. Background Art

[0002] Oil and gas exploration is a complex and costly project, the main goal of which is to find and evaluate underground oil and gas reservoirs. In the process of oil and gas exploration, accurate identification of the lithology of sedimentary bodies is crucial to determine the location and characteristics of oil and gas reservoirs. Lithology identification not only helps geologists understand the underground geological structure, but also provides important information about the physical properties of the reservoir and the distribution of fluids. However, existing lithology identification methods mainly rely on seismic data and core analysis, which have many limitations and challenges.

[0003] Seismic exploration technology obtains information about underground geological structures by sending seismic waves underground and recording their reflection and refraction signals. Although seismic data usually contain rich stratigraphic information and can reflect the underground sedimentary environment, structural characteristics and lithology changes, seismic exploration has the following limitations in lithology identification:

[0004] Limited resolution: The resolution of seismic data is limited by the frequency and propagation path of seismic waves. Especially in deep strata and complex geological conditions, seismic signals are easily interfered by noise and multiple waves, resulting in reduced resolution. Although high-frequency seismic waves can provide higher resolution, their penetration depth is limited; low-frequency seismic waves have a large penetration depth but low resolution. Therefore, when identifying small or complex structures, seismic exploration data may not provide a clear enough image.

[0005] Noise interference: Seismic data is easily interfered by various noises, including environmental noise, seismic instrument noise and multiple waves. These noises will affect the quality of seismic signals and thus the accuracy of lithology identification. Although denoising can be performed through digital filtering and other technologies, it is still impossible to completely eliminate the influence of noise, especially in a noisy environment, where the denoising effect is limited.

[0006] Complex data interpretation: The processing and interpretation of seismic data requires complex calculations and expertise. When seismic waves propagate in underground media, they are affected by various geological conditions, such as bedding, cracks, and lithology changes, which makes seismic reflection signals highly nonlinear and multi-solution. Accurate interpretation of seismic data requires geologists to have rich experience and expertise, and also requires the use of complex seismic data processing software, which increases the difficulty and uncertainty of lithology identification.

[0007] Well logging technology measures geological and physical parameters around the well wall by placing logging instruments into the well, such as acoustic wave velocity, resistivity, density, and natural gamma rays. Although logging data provides detailed information about the underground formation and can directly reflect the physical properties of the lithology, there are also the following limitations in its application:

[0008] High acquisition cost: The acquisition of logging data usually needs to be carried out during the drilling process, which not only increases the exploration cost, but also may affect the progress of drilling. Drilling costs are expensive, especially in deep wells or offshore drilling environments, where the cost of logging operations is even higher. In addition, logging equipment and operations also require professional technicians, further increasing exploration costs.

[0009] Limited data integrity: The integrity of logging data is easily affected by the drilling environment, such as wellbore collapse, drilling fluid invasion, etc. These factors will cause discontinuity or distortion of logging data, affecting the quality of data and the accuracy of interpretation results. Under complex geological conditions, it is more difficult to obtain high-quality logging data, which increases the challenge of lithology identification.

[0010] Limited spatial resolution: Well logging data mainly provides geological information near the wellbore, and its spatial resolution is limited by the logging spacing and instrument resolution. Although well logging data can provide detailed vertical profile information, it is difficult to reflect the lithology changes and spatial distribution characteristics outside the wellbore. Therefore, it is difficult to fully understand the spatial structure and lithology changes of the formation by relying solely on well logging data.

[0011] Limitations of core analysis: Core analysis is to take underground core samples and conduct laboratory analysis and testing to determine the detailed characteristics and physical properties of the lithology. Although core analysis can provide the most direct and accurate lithology information, it has the following limitations:

[0012] The process is complex and costly: The core analysis process is complex, including core sampling, preservation, transportation, laboratory analysis, and data interpretation. Each link requires professional technology and equipment, and the operation cost is high. In addition, core analysis can usually only be performed on limited drilling samples, which makes it difficult to cover a large exploration area.

[0013] Sampling and preservation issues: The sampling and preservation of core samples are prone to contamination and damage, which affects the accuracy of the analysis results. During the drilling process, core samples may be contaminated by drilling fluid, mud, etc., resulting in changes in chemical composition and physical properties. At the same time, the preservation and transportation of core samples also require strict control of environmental conditions, otherwise it may lead to sample degradation and data distortion.

[0014] Limited representativeness: Core analysis can only provide local geological information and is difficult to represent the lithology changes in the entire exploration area. Although the representativeness of the data can be improved by increasing the number of sampling points, this will significantly increase the cost and workload. Under complex geological conditions, the representativeness of local core samples is more difficult to ensure, increasing the uncertainty of lithology identification.

[0015] In summary, traditional lithology identification methods have many limitations in application, including resolution and noise problems in seismic exploration, acquisition cost and data integrity problems in logging technology, and process complexity and representativeness problems in core analysis. These limitations make traditional lithology identification methods face many challenges in actual oil and gas exploration, and new technologies and methods are urgently needed to improve the accuracy and efficiency of lithology identification. The sedimentary lithology identification system and method of the present invention are proposed to address these problems. By integrating multi-source data and applying advanced machine learning algorithms, it aims to overcome the limitations of traditional technologies and provide a more accurate, efficient and comprehensive lithology identification solution. Summary of the invention

[0016] Based on the existing technical problems, the present invention proposes a sedimentary lithology identification system for oil and gas exploration, which is characterized by comprising:

[0017] Data acquisition module, used to collect seismic data and well logging data;

[0018] A data processing module, used for preprocessing seismic data and well logging data;

[0019] Data fusion module, used to fuse pre-processed seismic data, well logging data and geological models to generate a comprehensive data set;

[0020] A lithology identification module is used to analyze the comprehensive data set using machine learning algorithms to identify the lithology type of sedimentary bodies;

[0021] The result verification module is used to compare and verify the identification results with the actual core analysis data;

[0022] The display module is used to visualize the recognition results.

[0023] 2. The system according to claim 1, characterized in that the data acquisition module comprises:

[0024] A seismic exploration unit, used for collecting seismic wave data through a seismic exploration instrument;

[0025] The logging unit is used to collect downhole measurement data through logging equipment.

[0026] 3. The system according to claim 1, wherein the data processing module comprises:

[0027] A denoising processing unit, used for denoising seismic data and well logging data;

[0028] A normalization processing unit, used for normalizing seismic data and well logging data;

[0029] The data interpolation unit is used to interpolate missing or incomplete data.

[0030] 4. The system according to claim 1, characterized in that the data fusion module comprises:

[0031] A geological modeling unit for generating a three-dimensional geological model;

[0032] A data mapping unit for mapping seismic data and well logging data into a three-dimensional geological model;

[0033] The data weighting unit is used to perform weighted processing on different data sources according to the importance of the data.

[0034] 5. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that the lithology identification module uses a convolutional neural network (CNN) algorithm to train and analyze the comprehensive data set, wherein the loss function of the convolutional neural network is:

[0035]

[0036] Among them, L is the loss value, N is the number of samples, and y i is the actual label, is the predicted value,

[0037] 6. The sedimentary lithology identification system for oil and gas exploration according to claim 5, characterized in that the optimization algorithm of the convolutional neural network is the Adam optimization algorithm, and its update formula is:

[0038]

[0039] Among them, θ t is the parameter of the tth iteration, α is the learning rate, and are the first and second order moment estimates, respectively.

[0040] 7. The sedimentary lithology identification system for oil and gas exploration according to claim 1 is characterized in that the evaluation indicators used in the result verification module include accuracy, precision, recall and F1-score, and their calculation formulas are:

[0041]

[0042] Among them, TP is the number of true positives, TN is the number of true negatives, FP is the number of false positives, and FN is the number of false negatives.

[0043] 8. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that the display module comprises:

[0044] 3D visualization unit: used to generate 3D lithology distribution map;

[0045] Chart generation unit: used to generate statistical charts of lithology identification results;

[0046] Interaction interface unit: used for interaction between users and the system.

[0047] 9. A sedimentary lithology identification method for oil and gas exploration, characterized in that it comprises the following steps:

[0048] Data acquisition: obtaining seismic data and well logging data through seismic exploration and well logging technology;

[0049] Data processing: denoising, normalizing and interpolating the collected data;

[0050] Data fusion: Fusion of pre-processed seismic data, well logging data and geological models to generate a comprehensive data set;

[0051] Lithology Identification: A convolutional neural network algorithm is used to analyze the comprehensive data set to identify the lithology type of the sedimentary body;

[0052] Result verification: Compare and verify the identification results with the actual core analysis data;

[0053] Result display: The identification results are presented in the form of three-dimensional lithology distribution maps and statistical charts.

[0054] 10. A sedimentary body lithology identification method for oil and gas exploration according to claim 9, characterized in that the training process of the convolutional neural network algorithm in the lithology identification step comprises the following steps:

[0055] Build a convolutional neural network model;

[0056] Input a comprehensive dataset for training;

[0057] Calculate the loss function value and use the Adam optimization algorithm to update the parameters;

[0058] Repeat the training until the loss function converges.

[0059] Beneficial effects:

[0060] Improve the accuracy and efficiency of lithology identification: This invention integrates multi-source data (seismic data, well logging data and geological models) and adopts advanced machine learning algorithms (such as convolutional neural networks) to achieve rapid and accurate identification of sedimentary lithology. Compared with traditional methods, the system shows higher recognition accuracy and stability in large-scale data processing and complex geological conditions, thereby significantly improving the success rate of oil and gas exploration.

[0061] Reduce exploration costs and time: Traditional lithology identification methods usually rely on time-consuming and costly core analysis and laboratory testing. The system of the present invention reduces the reliance on actual core samples through automated data processing and intelligent lithology identification, reducing labor costs and time investment in the exploration process. In addition, the automated process of preprocessing, data fusion and algorithm optimization further improves work efficiency and shortens the exploration cycle.

[0062] Providing more intuitive and comprehensive lithology distribution maps: The system of the present invention can not only identify the lithology of sedimentary bodies, but also generate intuitive three-dimensional lithology distribution maps through three-dimensional geological modeling and visualization technology. These visualization results enable geologists and engineers to have a clearer understanding of underground structures and lithology distribution, so as to make more accurate exploration decisions and development plans. At the same time, the application of statistical charts and interactive interfaces enhances the interpretability of results and user experience, further supporting the development and utilization of oil and gas reservoirs. BRIEF DESCRIPTION OF THE DRAWINGS

[0063] Figure 1 This is a system block diagram of a sedimentary lithology identification system for oil and gas exploration proposed by the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] 1. Data Collection

[0066] In the oil and gas exploration area, seismic exploration instruments (such as seismic detectors) and logging equipment (such as acoustic logging instruments and resistivity logging instruments) are used to collect seismic wave data and downhole measurement data. These data include but are not limited to parameters such as velocity, density, acoustic wave time difference, and resistivity. The specific steps are as follows:

[0067] Seismic exploration: Arrange an array of seismic detectors, stimulate seismic waves through artificial or mechanical sources, and record the propagation of seismic waves in underground media.

[0068] Logging operation: During the drilling process, logging instruments are used to record parameters such as the acoustic wave velocity and resistivity of the well wall.

[0069] 2. Data processing

[0070] Preprocess the collected seismic data and well logging data. The specific steps include:

[0071] De-noising: Use digital filtering techniques such as wavelet transform to denoise seismic data to remove environmental noise and instrument noise.

[0072] Normalization: Normalize the logging data to make the data values ​​within a unified range for easy subsequent analysis. Use the formula:

[0073]

[0074] Among them, x is the original data, x' is the normalized data, and x min and x max are the minimum and maximum values ​​of the data respectively.

[0075] Data interpolation: Interpolate missing or incomplete data using linear interpolation or spline interpolation to ensure data integrity and continuity.

[0076] 3. Data Fusion

[0077] The preprocessed seismic data and well logging data are integrated with the regional geological model.

[0078] Use geological modeling software to generate a three-dimensional geological model and map various data into the model to form a comprehensive data set. The specific steps include:

[0079] Generate geological model: Generate a three-dimensional geological model of the region based on geological exploration data and historical exploration information. Data mapping: Map seismic data and well logging data to the corresponding positions and attributes of the three-dimensional geological model to form a comprehensive data set.

[0080] Data weighting: Different data sources are weighted according to the importance of the data to generate a weighted comprehensive data set. The weighting formula is:

[0081]

[0082] Among them, D is a comprehensive dataset, D i is the i-th data source, w i is the weight of the i-th data source, and n is the number of data sources

[0083] 4. Lithology identification

[0084] A convolutional neural network (CNN) algorithm was used to train and analyze the comprehensive dataset to identify the lithological types of sedimentary bodies.

[0085] The specific steps include:

[0086] Model construction: Build a convolutional neural network model, including input layer, convolution layer, pooling layer, fully connected layer and output layer.

[0087] The model results are as follows:

[0088] Input layer: Input the preprocessed comprehensive dataset.

[0089] Convolution layer: perform feature extraction, the convolution kernel size is 3×3, the step size is 1, and the ReLU activation function is used. Pooling layer: perform feature dimensionality reduction, the pooling window size is 2×2, and the step size is 2.

[0090] Fully connected layer: expands and classifies the pooled features, and uses the Softmax activation function to output the lithology category

[0091] Model training: The comprehensive data set is input into the convolutional neural network model for training, and the cross entropy loss function and Adam optimization algorithm are used to update the parameters. The training process is as follows:

[0092]

[0093] Model prediction: Use the trained cross-section to predict the new comprehensive data set and output the lithology type of the sedimentary body.

[0094] 5. Result Verification

[0095] Compare and verify the lithology results identified by the machine learning algorithm with the actual core analysis results. By adjusting the algorithm parameters and training data, the model is continuously optimized to improve the accuracy of identification. The specific steps include: Data comparison: Compare the identification results with the actual core analysis data, and calculate the evaluation indicators such as accuracy, precision, recall and F1-score. The formula is as follows:

[0096]

[0097] Among them, TP is the number of true positives, TN is the number of true negatives, FP is the number of false positives, and FN is the number of false negatives. Model optimization: According to the evaluation index results, adjust the model parameters and training data, and repeat the training and verification process until the model performance meets the expected requirements.

[0098] 6. Results display

[0099] The recognition results are visualized through the display module. The specific steps include:

[0100] Generate three-dimensional lithology distribution map: Use three-dimensional geological modeling software to generate a three-dimensional lithology distribution map of the identification results.

[0101] Generate statistical charts: Use data analysis software to generate statistical charts of lithology identification results, such as bar charts, pie charts, etc.

[0102] Interactive operation: Through the interactive interface, users can query, browse and analyze the recognition results.

[0104] Example 1: Sedimentary lithology identification of an oil field block

[0105] In a certain oil field block, the sedimentary body lithology identification system of the present invention is applied. First, seismic wave data and well logging data are collected respectively through seismic exploration and well logging technology. Then, the seismic data is denoised by wavelet transform, the well logging data is normalized, and the missing data is linearly interpolated and supplemented. Next, a three-dimensional geological model is generated, and the seismic data and well logging data are mapped to the model to form a comprehensive data set. Finally, the convolutional neural network algorithm is used to analyze the comprehensive data set to identify the lithology type of the sedimentary body, and the identification results are compared and verified with the actual core analysis data to generate a three-dimensional lithology distribution map and statistical charts. Through this system, the efficiency and accuracy of oil and gas reservoir identification are greatly improved.

[0106] Example 2: Lithology identification of sedimentary bodies in an offshore oil field

[0107] In a certain offshore oil field, the sedimentary body lithology identification system of the present invention is applied. First, seabed seismic wave data and downhole measurement data are collected by offshore seismic exploration instruments and logging equipment. Then, Gaussian filtering is used to denoise the seabed seismic wave data, normalize the logging data, and perform spline interpolation to complete the missing data. Next, a three-dimensional seabed geological model is generated, and the seismic data and logging data are mapped to the model to form a comprehensive data set. Finally, a convolutional neural network algorithm is used to analyze the comprehensive data set to identify the lithology type of the seabed sedimentary body, and the identification results are compared and verified with the actual core analysis data to generate a three-dimensional lithology distribution map and statistical charts. Through this system, the efficiency and accuracy of offshore oil and gas exploration are improved.

[0108] Example 3: Sedimentary lithology identification in a complex geological area

[0109] In a complex geological area, the sedimentary body lithology identification system of the present invention is applied. First, seismic wave data and downhole measurement data under complex geological conditions are collected by seismic exploration instruments and logging equipment. Then, the double filtering algorithm is used to denoise the seismic data, normalize the logging data, and complete the missing data by Kriging interpolation. Next, a three-dimensional geological model under complex geological conditions is generated, and the seismic data and logging data are mapped to the model to form a comprehensive data set. Finally, the convolutional neural network algorithm is used to analyze the comprehensive data set to identify the sedimentary body lithology type under complex geological conditions, and the identification results are compared and verified with the actual core analysis data to generate a three-dimensional lithology distribution map and statistical charts. Through this system, the problem of oil and gas reservoir identification under complex geological conditions is effectively solved. .

[0110] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0111] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A sedimentary lithology identification system for oil and gas exploration, characterized in that: include: Data acquisition module: used to collect seismic data and well logging data; Data processing module: used to pre-process seismic data and well logging data; Data fusion module: used to fuse pre-processed seismic data, well logging data and geological models to generate a comprehensive data set; Lithology Identification Module: It is used to analyze the comprehensive data set using machine learning algorithms to identify the lithology type of the sedimentary body; Result verification module: used to compare and verify the identification results with the actual core analysis data; Display module: used to visualize the recognition results.

2. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The data acquisition module comprises: A seismic exploration unit, used for collecting seismic wave data through a seismic exploration instrument; The logging unit is used to collect downhole measurement data through logging equipment.

3. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The data processing module comprises: A denoising processing unit, used for denoising seismic data and well logging data; A normalization processing unit, used for normalizing seismic data and well logging data; The data interpolation unit is used to interpolate missing or incomplete data.

4. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The data fusion module comprises: A geological modeling unit for generating a three-dimensional geological model; A data mapping unit for mapping seismic data and well logging data into a three-dimensional geological model; The data weighting unit is used to perform weighted processing on different data sources according to the importance of the data.

5. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The lithology recognition module uses a convolutional neural network (CNN) algorithm to train and analyze the comprehensive data set, where the loss function of the convolutional neural network is: Among them, L is the loss value, N is the number of samples, and y i is the actual label, is the predicted value.

6. The sedimentary lithology identification system for oil and gas exploration according to claim 5, characterized in that: The optimization algorithm of the convolutional neural network is the Adam optimization algorithm, and its update formula is: Among them, θ t is the parameter of the tth iteration, α is the learning rate, and are the first and second order moment estimates, respectively.

7. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The evaluation indicators used in the result verification module include accuracy, precision, recall and F1-score, and their calculation formulas are: Among them, TP is the number of true positives, TN is the number of true negatives, FP is the number of false positives, and FN is the number of false negatives.

8. The sedimentary lithology identification system for oil and gas exploration according to claim 1, characterized in that: The display module comprises: 3D visualization unit: used to generate 3D lithology distribution map; Chart generation unit: used to generate statistical charts of lithology identification results; Interaction interface unit: used for interaction between users and the system.

9. A sedimentary lithology identification method for oil and gas exploration, characterized in that: The following steps are involved: Data acquisition: obtaining seismic data and well logging data through seismic exploration and well logging technology; Data processing: denoising, normalizing and interpolating the collected data; Data fusion: Fusion of pre-processed seismic data, well logging data and geological models to generate a comprehensive data set; Lithology Identification: A convolutional neural network algorithm is used to analyze the comprehensive data set to identify the lithology type of the sedimentary body; Result verification: Compare and verify the identification results with the actual core analysis data; Result display: The identification results are presented in the form of three-dimensional lithology distribution maps and statistical charts.

10. A sedimentary lithology identification method for oil and gas exploration according to claim 9, characterized in that: The training process of the convolutional neural network algorithm in the lithology identification step includes the following steps: Build a convolutional neural network model; Input a comprehensive dataset for training; Calculate the loss function value and use the Adam optimization algorithm to update the parameters; Repeat the training until the loss function converges.

Images