High-resolution reservoir prediction method under seismic full-wave information mining

Abstract

The invention relates to the technical field of reservoir prediction, in particular to a high-resolution reservoir prediction method under seismic full-wave information mining. The method comprises the following steps: acquiring seismic full-wave information data; performing frequency domain conjoint analysis on the seismic full-wave information data to generate a multi-scale full-wave feature tensor; performing seismic full-wave signal multi-scale decoupling on the multi-scale full-wave feature tensor to generate seismic multi-wave field separation data; implicit reservoir feature mapping is carried out on the seismic multi-wave-field separation data to generate initial reservoir feature mapping data; performing nonlinear correlation analysis on the initial reservoir feature mapping data to generate high-order reservoir feature data; and performing inversion linkage optimization on the high-order reservoir feature data and the seismic multi-wave-field separation data to generate reservoir dynamic response optimization data. According to the method, through multi-wave-field information extraction, high-resolution reservoir detail display, dynamic reservoir change prediction and confidence evaluation, the accuracy and resolution of reservoir prediction are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of reservoir prediction, and in particular to a high-resolution reservoir prediction method under seismic full-wave information mining. Background Art

[0002] Initially, seismic exploration mainly relied on the traditional seismic reflection method, which inferred geological structures by analyzing the reflection of seismic waves at the interface of underground layers. However, this method has limited effect in complex geological environments, especially in the prediction of deep or complex reservoirs, which poses great challenges. With the advancement of science and technology, especially the improvement of computer technology and data processing capabilities, seismic full-wave information mining technology has gradually emerged. This technology can obtain richer underground information by analyzing the full waveform information of seismic waves, including multiple wave types such as reflection waves, transmission waves, and scattered waves. In particular, through holographic analysis of the wave field, key features such as the physical properties, porosity, and water saturation of the reservoir can be more accurately identified, thereby improving the resolution of reservoir prediction. In recent years, with the development of deep learning and big data technology, the combination of seismic full-wave information mining and machine learning has greatly improved the accuracy and efficiency of reservoir prediction. However, the current traditional seismic reservoir prediction methods often rely only on traditional seismic reflection wave information, ignoring other wave fields in the full wave information (such as transmission waves, scattered waves, etc.). At the same time, there is a problem of low resolution, making it difficult to identify details and accurately characterize reservoir characteristics, which in turn leads to low accuracy and resolution of reservoir prediction. Summary of the invention

[0003] Based on this, it is necessary to provide a high-resolution reservoir prediction method based on seismic full-wave information mining to solve at least one of the above technical problems.

[0004] To achieve the above object, a high-resolution reservoir prediction method based on seismic full-wave information mining is provided, the method comprising the following steps: Step S1: Acquire seismic full-wave information data; perform frequency domain joint analysis on the seismic full-wave information data to generate a multi-scale full-wave characteristic tensor; perform multi-scale decoupling of seismic full-wave signals on the multi-scale full-wave characteristic tensor to generate seismic multi-wave field separation data; Step S2: perform implicit reservoir characteristic mapping on the seismic multi-wavefield separation data to generate initial reservoir characteristic mapping data; perform nonlinear correlation analysis on the initial reservoir characteristic mapping data to generate high-order reservoir characteristic data; perform inversion linkage optimization on the high-order reservoir characteristic data and the seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; perform reservoir detail enhancement on the reservoir dynamic response optimization data to generate a high-resolution reservoir detail display diagram; Step S3: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; performing reservoir characteristic change prediction on the high-resolution reservoir detail display image based on the dynamic reservoir characteristic change data to generate a reservoir characteristic change prediction image; Step S4: classify the reservoir characteristics change prediction map to generate a reservoir classification map; calculate the confidence of the reservoir classification map to generate reservoir prediction confidence data; compare and output the reservoir prediction confidence data with a preset confidence threshold until a high-resolution reservoir prediction distribution map is generated.

[0005] By acquiring full-wave information and performing joint analysis in the frequency domain, the present invention can extract rich underground information from multiple scales and multiple wave fields, overcoming the limitation of traditional methods that only rely on a single waveform signal. Multi-scale decoupling and wave field separation technology further improve the resolution of the signal, so that the characteristics of different wave fields can be processed independently, providing clearer basic data for subsequent high-precision reservoir prediction. Through implicit reservoir feature mapping and nonlinear correlation analysis, the deep-level characteristic relationship of the reservoir can be revealed, and more accurate high-order reservoir feature data can be generated. The inversion linkage optimization method can establish a close connection between multi-wave field data, thereby generating a more realistic reservoir dynamic response. Reservoir detail enhancement technology improves the resolution of the reservoir structure, ensures high-precision detail display, and thus improves the accuracy of reservoir prediction. Dynamic reservoir feature change analysis can monitor the change trend of the reservoir in real time and identify potential change patterns. Based on this, the reservoir feature change prediction map provides a forward-looking prediction of the future evolution of the reservoir, helps to make more accurate development decisions, and effectively improves the dynamic management ability of the reservoir. Through reservoir classification and confidence calculation, the reservoir can be accurately divided into regions and risk assessed, further improving the accuracy of reservoir prediction. Confidence assessment ensures the reliability and credibility of the prediction results, avoids errors caused by uncertainty, and the high-resolution reservoir prediction distribution map finally generated provides a reliable decision-making basis for oil and gas exploration and development. Therefore, the present invention improves the accuracy and resolution of reservoir prediction through multi-wave field information extraction, high-resolution reservoir detail display, dynamic reservoir change prediction and confidence assessment.

[0006] Preferably, step S1 comprises the following steps: Step S11: using a multi-point distributed seismic signal acquisition array to acquire seismic full-wave information data; Step S12: preprocessing the seismic full-wave information data to generate standard seismic full-wave information data, wherein the data preprocessing includes data cleaning, data missing value filling and data standardization; Step S13: performing time domain analysis on the full-wave signal to generate full-wave signal time domain data; performing fast Fourier transform on the full-wave signal time domain data to generate full-wave signal frequency domain data; performing frequency domain joint analysis based on the full-wave signal time domain data and the full-wave signal frequency domain data to generate a multi-scale full-wave feature tensor; Step S14: Based on the preset seismic wave propagation characteristic coefficients, the multi-scale full-wave characteristic tensor is subjected to multi-scale decoupling of the seismic full-wave signal to generate seismic multi-wave field separation data.

[0007] The present invention can remove noise and invalid data by preprocessing the seismic full-wave information, including data cleaning, missing value filling and standardization, to ensure that the final data is more representative and usable, and provide high-quality basic data for subsequent analysis. The combination of time domain analysis and frequency domain analysis can comprehensively capture the characteristics of seismic signals from different perspectives (time domain and frequency domain). The multi-scale full-wave feature tensor generated by joint analysis in the frequency domain helps to understand the propagation characteristics of seismic waves from multiple scales and improve the depth and accuracy of signal analysis. By performing multi-scale decoupling based on the seismic wave propagation characteristic coefficient, different wave fields can be effectively separated and the characteristics of various types of seismic waves can be extracted. This is very important for accurately identifying and analyzing different types of seismic waves (such as body waves, surface waves, etc.), and helps to better understand the propagation of seismic waves in underground media. The time domain, frequency domain and multi-scale analysis in steps S13 and S14 provide a multi-dimensional perspective for the processing of seismic signals, enhance the data analysis capability, and help to extract meaningful information in a complex seismic environment. Through multi-angle analysis and decoupling technology, the analysis capability and accuracy of seismic signals are effectively improved, which has important practical application value for earthquake monitoring, early warning and disaster assessment.

[0008] Preferably, step S14 comprises the following steps: Step S141: performing seismic wave transmission velocity analysis on the multi-scale full-wave characteristic tensor to generate seismic wave transmission velocity data; performing seismic wave type classification on the multi-scale full-wave characteristic tensor based on the seismic wave transmission velocity data to generate seismic wave type classification data, wherein the seismic wave type classification data includes longitudinal wave data, shear wave data and surface wave data; Step S142: performing variational modal decomposition on the P-wave data, S-wave data and surface wave data to generate P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data; performing wave field separation on the P-wave data, S-wave data and surface wave data through the P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data to generate seismic full-wave wave field separation data; Step S143: performing wavefield characteristic coupling analysis on the seismic full-wave wavefield separation data to generate wavefield characteristic coupling data; performing multi-scale decoupling of the seismic full-wave signal on the multi-scale full-wave characteristic tensor using the wavefield characteristic coupling data to generate seismic multi-wavefield separation data.

[0009] The present invention can accurately distinguish longitudinal waves, transverse waves and surface waves through the analysis of seismic wave transmission velocity and classification of seismic wave types. This not only improves the level of structured processing of seismic wave data, but also provides strong support for subsequent wave field separation and decoupling, which helps to more clearly identify the type and propagation path of seismic waves. By processing longitudinal wave, transverse wave and surface wave data through variational mode decomposition (VMD) technology, the independent characteristics of each wave can be effectively extracted to avoid mutual interference between different wave fields, thereby achieving more accurate wave field separation. This step helps to deeply analyze the characteristics of each wave field, especially in a complex seismic wave propagation environment, and can accurately reveal the contribution of different waves. Characteristic coupling analysis of the data after wave field separation helps to reveal the potential relationship and interaction between different wave fields, so as to more comprehensively understand the behavior of seismic waves and their propagation mode in underground media. This analysis plays an important role in predicting the propagation trend of seismic waves and analyzing the coupling effect between wave fields. The multi-scale decoupling in step S143 not only helps to separate different wave fields in seismic waves, but also can further extract valuable seismic wave characteristic data. Through multi-scale decoupling, the resolution of the signal can be improved, providing more detailed support for the real-time monitoring, analysis and early warning of earthquake signals. Through the combined processing of the above steps, high-quality earthquake signal data can be generated, which not only improves the accuracy of earthquake monitoring, but also broadens the application potential of these data in earthquake early warning, geological exploration, structural safety assessment and other fields. By combining seismic wave classification, wave field separation and characteristic coupling analysis, the multi-scale decoupling process of seismic wave signals is optimized, which not only enhances the understanding of seismic wave propagation characteristics, but also provides a reliable data basis for subsequent earthquake disaster monitoring and prediction.

[0010] Preferably, step S2 comprises the following steps: Step S21: extracting implicit reservoir characteristics from the seismic multi-wavefield separation data to obtain seismic wave implicit reservoir characteristic data; performing spatial distribution mapping on the seismic multi-wavefield separation data according to the seismic wave implicit reservoir characteristic data to generate initial reservoir characteristic mapping data; Step S22: decoupling the initial reservoir characteristic mapping data from physical properties to generate reservoir physical property data; performing nonlinear correlation analysis on the reservoir physical property data to generate high-order reservoir characteristic data; performing inversion linkage optimization on the high-order reservoir characteristic data and seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; Step S23: performing frequency component decomposition on the reservoir dynamic response optimization data to generate a multi-frequency reservoir characteristic spectrum; performing reservoir detail enhancement on the multi-frequency reservoir characteristic spectrum to generate a high-resolution reservoir detail display map.

[0011] The present invention can deeply explore the propagation characteristics of seismic waves in underground media through implicit reservoir feature extraction, and the generated seismic wave implicit reservoir feature data provides a more accurate basis for reservoir feature analysis. Through spatial distribution mapping, these features are applied to the spatial analysis of reservoirs, which helps to determine the distribution of reservoirs in a wide area and provides a clear spatial basis for reservoir exploration and development. The decoupling of reservoir physical properties allows the influence of different physical properties (such as porosity, elastic modulus, etc.) to be analyzed separately, avoiding confusion between different physical properties. This decoupling analysis can more clearly display the physical properties of the reservoir and provide an accurate explanation for the relationship between seismic wave signals and reservoir physical properties. Nonlinear correlation analysis further enhances the grasp of complex physical properties and makes reservoir characteristic analysis more accurate. Through inversion linkage optimization, high-order reservoir feature data is combined with seismic wave data, which can effectively optimize the dynamic response model of the reservoir. This helps to better understand the dynamic behavior of the reservoir, such as oil and gas flow or water permeability characteristics, thereby improving the efficiency of oil and gas field development and water resource utilization. Frequency component decomposition allows the different frequency characteristics of the reservoir to be analyzed separately, revealing the reservoir details and behaviors corresponding to different frequency components. By enhancing the reservoir details, the key features of the reservoir can be highlighted, especially in complex reservoir structures, helping engineers to more clearly identify potential resource-rich areas or abnormal areas. The generated high-resolution reservoir detail display map can provide detailed reservoir information, making the spatial and physical characteristics of the reservoir better visualized. With this high-resolution map, geologists and engineers can more accurately evaluate, make decisions and optimize reservoirs, providing important data support for oil and gas exploration, oil production processes and resource assessment. Through in-depth analysis and optimization of seismic multi-wavefield separation data, the extraction accuracy of reservoir characteristics is improved, and by optimizing reservoir dynamic response and detail enhancement, a more efficient and accurate reservoir evaluation method is provided. These analyses are of great value in improving the analytical capabilities of seismic data and its application in oil and gas exploration and other fields.

[0012] Preferably, performing spatial distribution mapping on seismic multi-wavefield separation data according to seismic wave implicit reservoir characteristic data includes: Confirm the seismic acquisition points of the seismic multi-wave field separation data to obtain seismic acquisition point information data; confirm the geographic coordinates of the seismic acquisition point information data to obtain geographic space coordinate data; The seismic multi-wave field separation data and the geographic space coordinate data are mapped to a spatial coordinate grid to generate seismic multi-wave field grid mapping data; the propagation range of the seismic multi-wave field grid mapping data is evolved according to the seismic wave implicit reservoir characteristic data to generate seismic propagation evolution path data; Based on the seismic propagation evolution path data, the seismic multi-wave field grid mapping data is refined into grid units to generate seismic wave spatial propagation area data; the local implicit mode of the seismic wave implicit reservoir characteristic data is extracted to obtain the local implicit mode data of the seismic wave; The grid unit reservoir characteristic values ​​of the seismic wave spatial propagation area data are calculated according to the local implicit mode data of the seismic wave to obtain the grid unit reservoir characteristic values; the seismic multi-wave field grid mapping data are spatially jointly mapped based on the grid unit reservoir characteristic values ​​to generate initial reservoir characteristic mapping data.

[0013] The present invention ensures that all data can be accurately mapped to the actual geographical location by confirming the seismic acquisition points and obtaining the geographic coordinates. This not only provides accurate basic information for subsequent data analysis and spatial mapping, but also ensures the reliability and consistency of seismic data. By grid mapping the seismic multi-wave field separation data with the geographic spatial coordinate data, the seismic data can be finely divided into multiple grid units in space. This refined spatial grid mapping helps to provide higher resolution seismic data, so that the spatial distribution characteristics of the reservoir can be more finely displayed and analyzed in a local range. By evolving the propagation range of the grid mapping data according to the implicit reservoir characteristic data of the seismic wave, the propagation path of the seismic wave and its evolution process can be revealed. This not only improves the understanding of the propagation mechanism of seismic waves, but also helps to predict the propagation trend and influence range of seismic waves under different conditions. Local implicit pattern extraction can identify the unique propagation characteristics of seismic waves in different regions, thereby helping to deeply understand the local behavior of seismic waves. Combined with the calculation of the grid unit reservoir characteristic value with the local implicit pattern data, more accurate reservoir physical property characteristic data can be generated, providing a scientific basis for subsequent reservoir evaluation and optimization. Based on the reservoir characteristic values ​​of grid units, spatial framework joint mapping is performed, which can closely combine seismic multi-wave field grid data with reservoir characteristic values ​​to generate initial reservoir characteristic mapping data. This joint mapping can better reflect the relationship between seismic waves and reservoir characteristics, and provide a more intuitive and accurate model for the spatial distribution and physical property analysis of reservoirs. Through the above series of spatial distribution mapping and feature extraction processes, detailed and high-precision reservoir characteristic mapping data can be obtained. This helps to more accurately evaluate the spatial distribution, physical properties and dynamic response of the reservoir, thereby optimizing the exploration and development strategy of the reservoir and improving the utilization efficiency of oil and gas resources. By mapping the spatial distribution of seismic wave implicit reservoir characteristic data, high-resolution and high-precision reservoir characteristic mapping data are provided, which not only improves the understanding of the seismic wave propagation process, but also enhances the accuracy of reservoir evaluation and optimization, and has important practical application value for oil and gas exploration, resource evaluation and development.

[0014] Preferably, the inversion linkage optimization of the high-order reservoir characteristic data and the seismic multi-wavefield separation data includes: Conduct frequency domain analysis on seismic multi-wave field separation data to generate seismic full-wave frequency domain data; construct nonlinear wave equations for seismic full-wave data based on seismic full-wave frequency domain data to generate reservoir medium wave equations; Finite difference calculation is performed according to the reservoir medium wave equation to obtain reservoir medium wave differential data; the reservoir medium wave equation is numerically solved and simulated through the reservoir medium wave differential data to generate simulated seismic wave field data; By simulating seismic wave field data, high-order reservoir characteristic data are inverted and optimized to generate reservoir dynamic response optimization data.

[0015] The present invention can convert time domain signals into frequency domain data by performing frequency domain analysis on seismic multi-wave field separation data, revealing the propagation characteristics of different frequency components in the reservoir. This process helps to refine the analysis of reservoir wave characteristics and provide the necessary data basis for subsequent wave equation construction and reservoir optimization. By constructing the nonlinear wave equation of the reservoir medium based on the seismic full-wave frequency domain data, the propagation behavior of seismic waves in the reservoir can be accurately described. The nonlinear wave equation can simulate the process of interaction between complex underground media and seismic waves, and considering the nonlinear effect, the simulation results are more realistic and accurate. After the reservoir medium wave equation is calculated by finite difference, wave difference data can be obtained. This data is further used for numerical solution simulation to generate simulated seismic wave field data. Through this method, the propagation behavior of seismic waves in a complex reservoir environment can be accurately simulated, providing scientific simulation results for subsequent reservoir response analysis and optimization. By inverting and optimizing the high-order reservoir characteristic data through simulated seismic wave field data, the dynamic response of the reservoir can be obtained more accurately. The inversion optimization process can link the physical properties, structure and dynamic behavior of the reservoir. By matching the simulated wave field data with high-order reservoir characteristics, a more accurate reservoir response model can be provided. The generated reservoir dynamic response optimization data can help accurately evaluate the performance of the reservoir under different seismic waves, especially its response to fluctuations and its impact on oil and gas flow. The optimized response data can not only improve the prediction accuracy of reservoir dynamic characteristics, but also provide more accurate decision support for oil and gas exploration, reservoir development, etc. Through inversion linkage optimization, the dynamic response of the reservoir can be accurately obtained, further improving the efficiency of oil and gas field development. The dynamic response optimization of the reservoir can provide a more accurate basis for the planning and implementation of oil and gas development, especially in the exploitation process of oil and gas reservoirs, helping to predict and adjust production plans and improve resource utilization.

[0016] Preferably, step S23 includes the following steps: Step S231: performing wavelet packet decomposition on reservoir dynamic response optimization data to generate low-frequency component data and high-frequency component data; performing geological labeling on the low-frequency component data and the high-frequency component data to generate low-frequency component labeled data and high-frequency component labeled data; Step S232: reconstructing the frequency band characteristics of the reservoir dynamic response optimization data to generate a multi-band time domain image; performing spatial feature mapping on the multi-band time domain image to generate a multi-frequency reservoir characteristic map; Step S233: enhancing the texture features of the multi-frequency reservoir characteristic map based on the high-frequency component annotation data to generate map local texture feature enhancement data; Step S234: performing trend fitting on the multi-frequency reservoir characteristic map based on the low-frequency component annotation data to generate map trend fitting data; performing overall trend optimization on the multi-frequency reservoir characteristic map through the map trend fitting data to generate map overall trend optimization data; Step S235: Optimize and integrate the multi-frequency reservoir characteristic map based on the map local texture feature enhancement data and the map overall trend optimization data, thereby obtaining a high-resolution reservoir detail display map.

[0017] The present invention can effectively decompose the reservoir dynamic response optimization data into low-frequency and high-frequency components through wavelet packet decomposition, respectively capturing the overall trend and detailed characteristics of the reservoir response. The low-frequency component mainly reflects the large-scale behavior of the reservoir and can reveal the macroscopic dynamic response of the reservoir; the high-frequency component captures the high-frequency changes in the details of the reservoir and can provide more accurate clues for microstructural analysis. Through geological label annotation, these frequency components can be interpreted more accurately, helping to establish a direct relationship between physical phenomena and reservoir characteristics. The frequency band characteristics of the reservoir dynamic response optimization data are reconstructed to generate multi-band time domain images, so that the time domain characteristics of the reservoir are displayed in detail in multiple frequency bands. This image can clearly show the time response characteristics of the reservoir at different frequencies, helping to better understand the dynamic behavior of the reservoir. Through spatial feature mapping, the generated multi-frequency reservoir feature map provides important spatial information for subsequent analysis and processing, which helps to understand the various characteristics of the reservoir from a global perspective. The labeled data of the high-frequency component enhances the texture characteristics of the map, thereby improving the local details of the reservoir map. Texture feature enhancement can help reveal the microscopic features and heterogeneity in the reservoir, such as pore distribution, fracture structure, etc. This process effectively improves the resolution of reservoir features, making the analysis of complex reservoirs more accurate. The labeled data of low-frequency components helps to fit the trend of multi-frequency reservoir feature maps, and then optimize the overall trend. This process helps to identify the regular changes of reservoirs on a large scale, providing a basis for predicting the overall behavior of reservoirs. Overall trend optimization improves the grasp of the long-term evolution characteristics of reservoirs by eliminating noise and enhancing the reliability of signals. By combining the local texture feature enhancement data of the map and the overall trend optimization data of the map, the multi-frequency reservoir feature map is optimized and integrated to generate a high-resolution reservoir detail display map. These optimized and integrated high-resolution images show the rich details of the reservoir, including microstructure and macro changes, with extremely high accuracy and clarity.

[0018] Preferably, step S233 includes the following steps: Based on the high-frequency component annotation data, the high-frequency texture area of ​​the multi-frequency reservoir characteristic map is located to generate high-frequency texture area data; the high-frequency texture area data is separated into feature areas to generate high-frequency texture separation area data; Grayscale co-occurrence matrix calculation is performed on the high-frequency texture separation area data to obtain texture information data of the high-frequency texture area; local contrast enhancement is performed on the texture information data of the high-frequency texture area to generate texture enhancement data of the high-frequency texture area; texture component matching and alignment is performed on the high-frequency texture area data according to the texture enhancement data of the high-frequency texture area to generate high-frequency texture alignment area data; The multi-frequency reservoir characteristic map is updated with map details through high-frequency texture alignment regional data to generate map local texture feature enhancement data.

[0019] The present invention can accurately identify the details in the reservoir map, especially the high-frequency texture information, by locating the high-frequency texture area of ​​the multi-frequency reservoir characteristic map based on the high-frequency component annotation data, which is very important for revealing the subtle changes and details in the reservoir. Feature area separation of the high-frequency texture area data further refines the high-frequency texture information, so that different texture areas can be analyzed independently, improving the accuracy and pertinence of subsequent processing. Grayscale co-occurrence matrix calculation of the high-frequency texture separation area data helps to extract the spatial arrangement pattern of the texture. This process can reveal the texture features in the reservoir map, such as roughness, uniformity, contrast, etc., which are crucial for revealing the microstructure of the reservoir. The obtained high-frequency texture area texture information data provides a reliable basis for subsequent local contrast enhancement and texture enhancement. By performing local contrast enhancement on the high-frequency texture area texture information data, the details and features of the texture in the reservoir map can be highlighted, so that the microstructure of the reservoir is clearer. Local contrast enhancement helps to amplify the texture features, making the important details in the reservoir more prominent, and facilitating the identification of potential reservoir heterogeneity areas. Texture component matching and alignment are performed based on the texture enhancement data of the high-frequency texture area to ensure that texture information of different frequencies can be effectively aligned and fused in the same atlas. This process helps to eliminate the deviation between information of different frequencies and ensure the overall consistency and accuracy of the atlas. The multi-frequency reservoir characteristic atlas is updated in detail through the high-frequency texture alignment regional data to further improve the clarity and resolution of the atlas, so that the texture characteristics of the reservoir more accurately reflect the actual physical and structural characteristics of the reservoir. The local texture feature enhancement data of the atlas finally generated provides a more delicate and clear display of reservoir details, especially in the regional details of the high-frequency texture, which further improves the operability and visualization of the atlas. The enhanced atlas can effectively show the microscopic heterogeneity of the reservoir and provide more accurate data support for the analysis and evaluation of the reservoir.

[0020] Preferably, step S3 comprises the following steps: Step S31: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; dividing the dynamic reservoir characteristic change data into data sets to generate a model training set and a model test set; Step S32: using a convolutional neural network algorithm to perform model training on the model training set to generate a reservoir characteristic change prediction pre-model; performing model optimization iteration on the reservoir characteristic change prediction pre-model through a model test set to generate a reservoir characteristic change prediction model; Step S33: importing the dynamic reservoir characteristic change data into the reservoir characteristic change prediction model to predict the reservoir characteristic change, thereby generating reservoir characteristic change prediction data; Step S34: dynamically updating the high-resolution reservoir detail display map through the reservoir characteristic change prediction data to generate a reservoir characteristic change prediction map.

[0021] The present invention can identify and extract the change trend of the reservoir in the time dimension, such as the evolution of reservoir pressure, porosity, fluid distribution and other characteristics, by analyzing the dynamic reservoir characteristic changes on the high-resolution reservoir detail display map. This process helps to obtain the time series data of reservoir changes and provides data support for subsequent predictive modeling. By dividing the dynamic reservoir characteristic change data into data sets and generating training sets and test sets, the prediction model can be effectively trained and the performance can be verified, thereby ensuring the generalization ability and prediction accuracy of the model. Using the CNN algorithm to train the model training set can automatically learn the complex patterns of reservoir characteristic changes from the data, especially when processing high-dimensional data, CNN can effectively capture the local characteristics of space and time, thereby improving the accuracy of the prediction. By optimizing and iterating the model through the test set, the prediction ability of the model can be continuously improved, the overfitting or underfitting phenomenon can be reduced, and the generated reservoir characteristic change prediction model can accurately reflect the actual changes in the reservoir. After completing model training and optimization, the dynamic reservoir characteristic change data is imported into the reservoir characteristic change prediction model for prediction, which can accurately simulate the evolution trend of reservoir characteristics, such as predicting the porosity change, fluid flow and other characteristics of the reservoir in the future time period. This prediction capability is crucial for foreseeing reservoir changes in advance. Based on the reservoir characteristic change prediction data, by dynamically updating the high-resolution reservoir detail display map, the reservoir characteristic change prediction map is generated, which can clearly show the state changes of the reservoir at future moments. This process helps decision makers evaluate the best time for reservoir development, optimize production strategies, and predict mining risks. Through accurate prediction of reservoir characteristic changes, not only can the accuracy of reservoir evaluation be improved, but also data support can be provided for subsequent reservoir management and mining decisions. For example, predicting the pressure, temperature, and fluid changes of the reservoir can effectively help engineers adjust mining plans or design new reservoir development plans. The prediction map provides a visual way to help geologists and engineers intuitively understand the dynamic changes of the reservoir, improve the efficiency and scientificity of decision-making, and reduce uncertainty in the development process. By dynamically predicting changes in reservoir characteristics, potential reservoir problems (such as decreased fluid permeability or abnormal pressure) can be identified in advance, so that preventive measures can be taken to reduce risks and optimize production plans.

[0022] Preferably, step S4 comprises the following steps: Step S41: performing reservoir potential assessment on the reservoir characteristic change prediction map to generate reservoir potential assessment data; performing reservoir classification on the reservoir characteristic change prediction map according to the reservoir potential assessment data to generate a reservoir classification map; Step S42: performing actual reservoir observation on the reservoir classification map to generate actual reservoir observation result data; performing confidence calculation on the actual reservoir observation result data and the reservoir classification map to generate reservoir prediction confidence data; Step S43: Compare the reservoir prediction confidence data with the preset confidence threshold. When the reservoir prediction confidence data is greater than or equal to the preset confidence threshold, output the map based on the reservoir classification map to generate a high-resolution reservoir prediction distribution map. When the reservoir prediction confidence data is less than the preset confidence threshold, return to step S3 for feature re-extraction until the reservoir prediction confidence data is greater than or equal to the preset confidence threshold.

[0023] The present invention can quantify the potential of the reservoir by analyzing the prediction map of reservoir characteristic changes, and provide a basis for subsequent development decisions. By accurately evaluating the reservoir potential, it is possible to identify areas with efficient oil or gas production, thereby maximizing resource utilization. Based on the reservoir potential evaluation data, the reservoir is classified and a reservoir classification map is generated, and the reservoir can be divided into different categories (such as high-yield, low-yield or difficult-to-develop areas). This enables engineers to formulate development plans and optimize mining plans in a targeted manner. Through actual reservoir observation, field data can be obtained and compared with the reservoir classification map to verify the accuracy of the prediction results. The introduction of actual reservoir observation result data makes reservoir classification and prediction closer to the actual situation and reduces prediction errors. Combined with confidence calculation, by comparing the actual observation data with the reservoir classification map, the confidence of the reservoir prediction can be calculated. This process helps to identify which reservoir areas have more reliable prediction results, thereby providing a scientific basis for decision-making. By comparing with the preset confidence threshold, reservoir areas with high confidence can be screened out to ensure the accuracy and reliability of the prediction. If the reservoir prediction confidence data is higher than the set threshold, a high-resolution reservoir prediction distribution map will be generated to provide a clear direction for the adjustment of the mining strategy. When the confidence of the prediction result is insufficient, it will automatically return to step S3 for feature re-extraction and optimization until the confidence requirement is met. This feedback mechanism ensures the reliability and scientificity of the final output reservoir prediction distribution map. Through the calculation and feedback adjustment of the confidence, step S4 achieves adaptive optimization, thereby improving the accuracy of the prediction results. The generation of reservoir classification maps and high-resolution reservoir prediction distribution maps helps decision makers to more accurately identify the reservoir characteristics and development potential of different regions. Before mining, the dynamic change trend of the reservoir can be foreseen in advance through high-resolution reservoir prediction maps to help avoid risks. Reservoir classification maps and reservoir prediction distribution maps provide visual support for mining plans. Based on these maps, engineers can formulate more accurate mining plans, such as giving priority to the development of high-yield areas and adjusting development strategies to meet the challenges of low-yield areas or difficult-to-develop areas. This process reduces the errors caused by human intervention and makes the entire reservoir evaluation and development process more scientific and refined. Through the gradual optimization of confidence and feedback mechanism, the extraction and prediction of reservoir characteristics can be continuously improved, reducing the losses caused by wrong predictions during the development process. The final high-resolution reservoir prediction distribution map helps to allocate resources more accurately and avoid waste of resources or uneconomical mining areas. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1 A schematic diagram of the steps of a high-resolution reservoir prediction method based on seismic full-wave information mining; Figure 2 for Figure 1 Detailed implementation steps of step S2 in the flowchart; Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0025] The technical method of the present invention is described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are 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 technicians in this field without creative work are within the scope of protection of the present invention.

[0026] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0027] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0028] To achieve this, please refer to Figures 1 to 3 A high-resolution reservoir prediction method based on seismic full-wave information mining, the method comprising the following steps: Step S1: Acquire seismic full-wave information data; perform frequency domain joint analysis on the seismic full-wave information data to generate a multi-scale full-wave characteristic tensor; perform multi-scale decoupling of seismic full-wave signals on the multi-scale full-wave characteristic tensor to generate seismic multi-wave field separation data; Step S2: perform implicit reservoir characteristic mapping on the seismic multi-wavefield separation data to generate initial reservoir characteristic mapping data; perform nonlinear correlation analysis on the initial reservoir characteristic mapping data to generate high-order reservoir characteristic data; perform inversion linkage optimization on the high-order reservoir characteristic data and the seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; perform reservoir detail enhancement on the reservoir dynamic response optimization data to generate a high-resolution reservoir detail display diagram; Step S3: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; performing reservoir characteristic change prediction on the high-resolution reservoir detail display image based on the dynamic reservoir characteristic change data to generate a reservoir characteristic change prediction image; Step S4: classify the reservoir characteristics change prediction map to generate a reservoir classification map; calculate the confidence of the reservoir classification map to generate reservoir prediction confidence data; compare and output the reservoir prediction confidence data with a preset confidence threshold until a high-resolution reservoir prediction distribution map is generated.

[0029] By acquiring full-wave information and performing joint analysis in the frequency domain, the present invention can extract rich underground information from multiple scales and multiple wave fields, overcoming the limitation of traditional methods that only rely on a single waveform signal. Multi-scale decoupling and wave field separation technology further improve the resolution of the signal, so that the characteristics of different wave fields can be processed independently, providing clearer basic data for subsequent high-precision reservoir prediction. Through implicit reservoir feature mapping and nonlinear correlation analysis, the deep-level characteristic relationship of the reservoir can be revealed, and more accurate high-order reservoir feature data can be generated. The inversion linkage optimization method can establish a close connection between multi-wave field data, thereby generating a more realistic reservoir dynamic response. Reservoir detail enhancement technology improves the resolution of the reservoir structure, ensures high-precision detail display, and thus improves the accuracy of reservoir prediction. Dynamic reservoir feature change analysis can monitor the change trend of the reservoir in real time and identify potential change patterns. Based on this, the reservoir feature change prediction map provides a forward-looking prediction of the future evolution of the reservoir, helps to make more accurate development decisions, and effectively improves the dynamic management ability of the reservoir. Through reservoir classification and confidence calculation, the reservoir can be accurately divided into regions and risk assessed, further improving the accuracy of reservoir prediction. Confidence assessment ensures the reliability and credibility of the prediction results, avoids errors caused by uncertainty, and the high-resolution reservoir prediction distribution map finally generated provides a reliable decision-making basis for oil and gas exploration and development. Therefore, the present invention improves the accuracy and resolution of reservoir prediction through multi-wave field information extraction, high-resolution reservoir detail display, dynamic reservoir change prediction and confidence assessment.

[0030] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a high-resolution reservoir prediction method under seismic full-wave information mining of the present invention. In this example, the high-resolution reservoir prediction method under seismic full-wave information mining includes the following steps: Step S1: Acquire seismic full-wave information data; perform frequency domain joint analysis on the seismic full-wave information data to generate a multi-scale full-wave characteristic tensor; perform multi-scale decoupling of seismic full-wave signals on the multi-scale full-wave characteristic tensor to generate seismic multi-wave field separation data; In an embodiment of the present invention, full-wave information is obtained by using seismic detection equipment (such as a seismograph). The signals collected by these devices usually include different types of seismic waves such as P waves, S waves, and surface waves. The collected raw seismic data is stored in a database, and the preprocessing process includes denoising, time synchronization, calibration, etc. to ensure the quality of the data. Spectral analysis techniques such as fast Fourier transform (FFT) or wavelet transform are used to convert seismic wave data in the time domain into the frequency domain. Each type of seismic wave has different characteristics in frequency, and frequency domain analysis helps to separate the waveform characteristics in different frequency bands. The signal is processed by wavelet transform or multi-scale Fourier analysis method to obtain the fluctuation characteristics at different frequency and time scales. Each scale represents a different resolution of the signal, which helps to capture the propagation characteristics of seismic waves at different time and space scales. The data obtained by multi-scale analysis is integrated into a multi-dimensional feature tensor, and each element in the tensor represents the fluctuation intensity or other characteristics (such as amplitude, phase, etc.) at different frequencies and time scales. Decoupling algorithms (such as principal component analysis (PCA), independent component analysis (ICA), etc.) are used to process multi-scale full-wave feature tensors. The goal is to decompose complex seismic full-wave information into multiple independent signal components (multi-wavefield separation). Each component corresponds to a specific wavefield (for example, P wave, S wave, Rayleigh wave, etc.). Through the decoupling process, different types of wavefield signals are separated, and each wavefield signal is reconstructed for subsequent analysis or prediction. Multi-wavefield separation data can more accurately reflect the characteristics of different wavefields, which is convenient for subsequent seismic wavefield analysis, source location, and construction of seismic models.

[0031] Step S2: perform implicit reservoir characteristic mapping on the seismic multi-wavefield separation data to generate initial reservoir characteristic mapping data; perform nonlinear correlation analysis on the initial reservoir characteristic mapping data to generate high-order reservoir characteristic data; perform inversion linkage optimization on the high-order reservoir characteristic data and the seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; perform reservoir detail enhancement on the reservoir dynamic response optimization data to generate a high-resolution reservoir detail display diagram; In an embodiment of the present invention, multiple wavefield signals extracted from seismic multi-wavefield separation data are used as input. These wavefield signals contain preliminary information of the reservoir, including the propagation characteristics of seismic waves, reflection waves, refraction waves, etc. The multi-wavefield separation data is mapped to the characteristic space of the reservoir using an implicit mapping method (such as a deep learning algorithm based on a neural network, a support vector machine, etc.). Implicit mapping can mine the complex relationship between seismic data and reservoir physical properties (such as porosity, permeability, etc.). Through this process, initial reservoir characteristic mapping data is generated, which reflects the rough attributes and hierarchical structure of the reservoir. The initial reservoir characteristic mapping data is analyzed by a nonlinear model (such as a deep neural network, a convolutional neural network, a random forest, etc.). At this time, nonlinear analysis helps to extract complex patterns in the data that cannot be revealed by linear methods, especially the effects of seismic wave propagation paths, reflection intensity, etc. on reservoir properties. The relationship between different reservoir characteristics (such as rock types, reservoir geometry, etc.) and seismic wave characteristics is modeled to reveal nonlinear relationships. For example, there is a complex nonlinear mapping relationship between the reflection intensity and porosity or permeability in different regions of the reservoir. Through this process, high-order reservoir characteristic data are generated, which contain more detailed reservoir characteristics, such as interlayer changes, fracture characteristics and other detailed information. Based on the high-order reservoir characteristic data, an inversion model is constructed. Using seismic inversion technology (such as least squares inversion, Bayesian inversion, etc.), the seismic data and reservoir characteristics are linked and optimized through the inversion process, and the model parameters are iteratively adjusted. In the inversion process, the high-order reservoir characteristic data are linked with the seismic multi-wavefield separation data, and the error between the wavefield data and the reservoir response is minimized by optimizing the model. Through this linkage optimization, the dynamic response of the reservoir is accurately reflected, especially the response characteristics of the reservoir under different seismic events. Finally, through inversion and optimization, the dynamic response optimization data of the reservoir is obtained, which can accurately reflect the dynamic behavior of the reservoir under different conditions, such as stress changes, fluid penetration, etc. The reservoir dynamic response optimization data is enhanced in detail using data enhancement technology (such as super-resolution reconstruction, image processing algorithms, etc.). These technologies can improve the resolution of reservoir characteristics by supplementing missing information and amplifying small changes in the reservoir. Apply detail optimization algorithms to enhance the performance of important features in the reservoir, such as fracture networks, pore structures, permeability distribution, etc. This process involves image filtering, denoising, edge detection and other technologies to highlight the details in the reservoir. Finally, the enhanced reservoir details are presented in the form of high-resolution displays, which can be used for further analysis and decision support, helping geologists and engineers to more accurately understand the reservoir structure and characteristics.

[0032] Step S3: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; performing reservoir characteristic change prediction on the high-resolution reservoir detail display image based on the dynamic reservoir characteristic change data to generate a reservoir characteristic change prediction image; In an embodiment of the present invention, a set of high-resolution reservoir detail display images are obtained, and these images generally display the detailed characteristics of the reservoir, including the spatial distribution of properties such as porosity, permeability, and fracture network. If there are reservoir detail images at multiple time points, a time series data set can be constructed, and the display image at each moment reflects the current state of the reservoir. If there is no data at multiple time points, hypothetical time series data can be generated by simulation or other means. By using image analysis techniques such as image differentiation, edge detection, and image registration (for example, a phase-related image registration method), feature changes in the reservoir detail display image are detected. For example, by comparing the changes in images at different time points, the dynamic changes in features such as porosity, fractures, and permeability in the reservoir are extracted. The detected change features are quantified and recorded to generate dynamic reservoir feature change data. These data generally include the amplitude, direction, speed, etc. of reservoir attribute changes in the time series, and involve the change trend of the local area (such as the expansion of fractures or changes in porosity, etc.). According to the dynamic reservoir characteristic change data, select appropriate prediction methods, such as time series prediction based on machine learning, convolutional neural network (CNN) or long short-term memory network (LSTM) in deep learning, or mathematical models based on traditional numerical methods such as finite difference method. Use historical data and dynamic reservoir characteristic change data to train the model. The goal of this model is to learn the law of reservoir characteristic change over time and accurately predict the change of reservoir characteristics at future moments. The prediction model is optimized by means of cross-validation, error evaluation, etc. to ensure that the model can effectively capture the dynamic evolution of reservoir characteristics. The trained prediction model is applied to the high-resolution reservoir detail display map to predict the change of reservoir characteristics at future time points based on the reservoir state (image data) at the current moment and historical change data. This process can output predicted porosity changes, permeability changes, fracture expansion, etc. Generate reservoir characteristic change prediction maps based on the model prediction results. These prediction maps will show the state of the reservoir at future time points or under different environmental conditions, and reflect the change trends in different regions (such as areas of fracture expansion, areas of increased or decreased porosity, etc.). Visualize the prediction map so that expert teams such as geologists and engineers can conduct further analysis and decision-making. Reservoir changes can be displayed through color changes, 3D graphics or dynamic charts.

[0033] Step S4: classify the reservoir characteristics change prediction map to generate a reservoir classification map; calculate the confidence of the reservoir classification map to generate reservoir prediction confidence data; compare and output the reservoir prediction confidence data with a preset confidence threshold until a high-resolution reservoir prediction distribution map is generated.

[0034] In an embodiment of the present invention, a specific classification algorithm is selected based on the data of the reservoir characteristic change prediction map (such as changes in porosity, permeability, fracture density, etc.). Common classification methods include support vector machine (SVM), random forest (RF), decision tree, k-means clustering, deep learning (such as convolutional neural network CNN), etc. The reservoir characteristic change prediction map is converted into a data format for a specific input classification algorithm, such as a feature vector or pixel points of an image (depending on the prediction map, it may be two-dimensional data or three-dimensional data). The classification model is trained using known reservoir data (such as historical data, actual exploration results), and the model learns the relationship between reservoir characteristic changes and categories. For example, the reservoir is divided into different categories, such as "high permeability area", "low permeability area", "fracture area", etc. The trained classification model is used to classify the reservoir characteristic change prediction map to obtain reservoir category information for each area or pixel. These categories can represent different reservoir attributes, such as the distribution of porosity, permeability, fracture degree, etc. According to the classification results, a reservoir classification map is generated. This map represents different reservoir categories by color or different areas. It can be a two-dimensional image or a three-dimensional model, showing the spatial distribution of reservoirs of different categories. Confidence usually indicates the reliability of the classification result, that is, the probability of the model predicting the reservoir category. For example, the output probability of the classifier can be used to define the confidence of each predicted area. For deep learning models, the prediction probability of each pixel is usually obtained (for example, the probability of belonging to a certain category). For each predicted area or pixel, the confidence that it belongs to a certain category is calculated. For example, the maximum a posteriori probability (MAP) or Bayesian inference method is used to calculate the confidence score of each reservoir category. The calculated confidence value is assigned to each area in the reservoir classification map to obtain reservoir prediction confidence data. This data provides a confidence score for each reservoir area or pixel, indicating the reliability of the classification result. Based on the confidence calculation results, a confidence map is generated. This map can be a two-dimensional image, in which the value of each pixel represents the confidence of the area. Colors can represent the level of confidence, for example, dark colors represent high confidence and light colors represent low confidence. According to the actual application requirements, a confidence threshold is set. This threshold is used to screen reliable areas for reservoir prediction. The threshold can be determined by historical data or expert experience, and is usually set to 0.7, 0.8 or 0.9, indicating that the prediction result is considered reliable only when the confidence exceeds this value. The confidence value of each area or pixel is compared with the preset threshold. If the confidence is higher than the threshold, the reservoir prediction in the area is considered reliable; if it is lower than the threshold, the prediction result is considered unreliable or needs further confirmation. Based on the results of the comparison between the confidence and the threshold, a high-resolution reservoir prediction distribution map is generated. The figure shows all areas where the confidence exceeds the threshold, which represent reliable reservoir prediction results. Highlight high-confidence areas by marking with different colors or different transparencies.This map can help geological engineers, exploration teams or development decision makers to visually see reliable reservoir distribution and development areas.

[0035] Preferably, step S1 comprises the following steps: Step S11: using a multi-point distributed seismic signal acquisition array to acquire seismic full-wave information data; Step S12: preprocessing the seismic full-wave information data to generate standard seismic full-wave information data, wherein the data preprocessing includes data cleaning, data missing value filling and data standardization; Step S13: performing time domain analysis on the full-wave signal to generate full-wave signal time domain data; performing fast Fourier transform on the full-wave signal time domain data to generate full-wave signal frequency domain data; performing frequency domain joint analysis based on the full-wave signal time domain data and the full-wave signal frequency domain data to generate a multi-scale full-wave feature tensor; Step S14: Based on the preset seismic wave propagation characteristic coefficients, the multi-scale full-wave characteristic tensor is subjected to multi-scale decoupling of the seismic full-wave signal to generate seismic multi-wave field separation data.

[0036] In an embodiment of the present invention, by deploying a multi-point distributed seismic signal acquisition array in the target area, each acquisition point should be equipped with a highly sensitive seismic sensor that can capture the full-wave information generated during seismic activity in real time. The collected data includes information such as the amplitude, frequency, and phase of the seismic wave, covering different types of seismic waves from P waves, S waves to surface waves, and forming a full-wave information data set. The layout of the multi-point array should cover different geographical areas to ensure the comprehensive capture of different seismic wave propagation modes. Remove noise data and correct abnormal values ​​caused by sensor failure or data transmission errors. Use interpolation algorithms (such as linear interpolation, spline interpolation, etc.) to fill in missing data in seismic signals caused by equipment failure or transmission problems. Unify the data collected by different seismic stations, for example, process the data according to a standardized range (such as a mean of 0 and a variance of 1) to reduce the differences in the impact of different equipment or environments and ensure data comparability. Analyze the temporal change trend of the full-wave signal, remove irrelevant frequency components through time domain filtering technology, retain the main seismic wave information, and generate time domain data. Convert the time domain data into frequency domain data to reveal the frequency components in the signal. The FFT algorithm can effectively extract the amplitude and phase of each frequency component in the signal. The time domain data and frequency domain data are jointly analyzed to extract multi-scale information based on the characteristics of waveform changes. Wavelet transform or other multi-scale methods are used to analyze the signal characteristics at different time scales and frequency scales to generate a multi-scale full-wave characteristic tensor. This characteristic tensor is a key data structure for seismic signal identification, classification and further analysis. Using seismological theory or experimental data, different types of seismic wave propagation characteristic coefficients are determined, including wave velocity, propagation path, attenuation coefficient, etc. Based on the set seismic wave propagation characteristic coefficients, a decoupling algorithm (such as principal component analysis or matrix decomposition method) is used to decompose the multi-scale full-wave characteristic tensor to separate various seismic wave fields (such as P waves, S waves and surface waves) from the original signal. Through the decoupling results, independent signals of each wave field can be obtained, which reflect the characteristics of different types of seismic waves in the time and frequency domains. The final generated seismic multi-wave field separation data is used for subsequent seismic wave analysis and disaster prediction.

[0037] Preferably, step S14 comprises the following steps: Step S141: performing seismic wave transmission velocity analysis on the multi-scale full-wave characteristic tensor to generate seismic wave transmission velocity data; performing seismic wave type classification on the multi-scale full-wave characteristic tensor based on the seismic wave transmission velocity data to generate seismic wave type classification data, wherein the seismic wave type classification data includes longitudinal wave data, shear wave data and surface wave data; Step S142: performing variational modal decomposition on the P-wave data, S-wave data and surface wave data to generate P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data; performing wave field separation on the P-wave data, S-wave data and surface wave data through the P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data to generate seismic full-wave wave field separation data; Step S143: performing wavefield characteristic coupling analysis on the seismic full-wave wavefield separation data to generate wavefield characteristic coupling data; performing multi-scale decoupling of the seismic full-wave signal on the multi-scale full-wave characteristic tensor using the wavefield characteristic coupling data to generate seismic multi-wavefield separation data.

[0038] In an embodiment of the present invention, the seismic wave transmission velocity of different wave types is obtained by using existing seismological theories or experimental data according to the propagation characteristics of seismic waves. The specific transmission velocity can be estimated by the time delay of the seismic wave propagating in the crust. The seismic wave transmission velocity data is generated by calculating the propagation velocity of signals in different frequency bands in the multi-scale full-wave characteristic tensor. This data helps to distinguish different wave types, especially when there are significant differences in wave velocities (such as longitudinal waves, transverse waves, and surface waves). Based on the seismic wave transmission velocity data, a classification algorithm (such as support vector machine or K-means clustering) is applied to classify the seismic wave types of the multi-scale full-wave characteristic tensor. According to the propagation velocity of seismic waves in the medium, the signals can be divided into three main types: longitudinal waves (P waves), transverse waves (S waves), and surface waves, and seismic wave type classification data is generated, which contains the signals corresponding to each type of seismic wave, wherein the longitudinal wave data, transverse wave data, and surface wave data are independently marked for subsequent processing. For the longitudinal wave data, transverse wave data, and surface wave data, the variational mode decomposition (VMD) method is used for modal decomposition. VMD can decompose nonlinear and non-stationary signals into several intrinsic mode functions (IMFs), which represent the main oscillation components in the signal. VMD is performed on the longitudinal wave, shear wave and surface wave signals to obtain the longitudinal wave modal decomposition data, shear wave modal decomposition data and surface wave modal decomposition data respectively. These data capture multiple frequency and time domain features in different waveforms. Based on the modal decomposition data, the longitudinal wave, shear wave and surface wave signals are further separated by multi-channel wavefield separation technology (such as independent component analysis ICA or non-negative matrix factorization NMF). The goal of this process is to extract the independent contribution of each wave type and eliminate redundant or interfering signals. Through this method, each wave field in the seismic wave can be separated more clearly to generate the final seismic full-wave wavefield separation data. The coupling analysis of the seismic full-wave wavefield separation data aims to discover the relationship or common characteristics between different wavefields. Through cross-correlation analysis, mutual information analysis and other methods, the coupling characteristics between longitudinal waves, shear waves and surface waves are identified to generate wave field characteristic coupling data. This data contains the coupling information of different wave fields at multiple scales, helping to reveal the interaction between wave fields and their overall impact on seismic wave propagation. Using the generated wave field characteristic coupling data, combined with the previous seismic wave type classification data and modal decomposition data, the multi-scale full-wave characteristic tensor is further multi-scale decoupled through decoupling algorithms (such as frequency domain decoupling or wavelet packet decoupling). The goal of this process is to separate multiple independent wave field data from complex seismic wave signals to accurately reflect the propagation characteristics of each seismic wave at different scales. Through this decoupling step, the final seismic multi-wave field separation data is generated, which contains independent signals of all seismic wave types and can be used for further seismic analysis and prediction.

[0039] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes: Step S21: extracting implicit reservoir characteristics from the seismic multi-wavefield separation data to obtain seismic wave implicit reservoir characteristic data; performing spatial distribution mapping on the seismic multi-wavefield separation data according to the seismic wave implicit reservoir characteristic data to generate initial reservoir characteristic mapping data; Step S22: decoupling the initial reservoir characteristic mapping data from physical properties to generate reservoir physical property data; performing nonlinear correlation analysis on the reservoir physical property data to generate high-order reservoir characteristic data; performing inversion linkage optimization on the high-order reservoir characteristic data and seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; Step S23: performing frequency component decomposition on the reservoir dynamic response optimization data to generate a multi-frequency reservoir characteristic spectrum; performing reservoir detail enhancement on the multi-frequency reservoir characteristic spectrum to generate a high-resolution reservoir detail display map.

[0040] In an embodiment of the present invention, implicit features related to reservoir characteristics are extracted from seismic multi-wave field separation data by using deep learning or traditional seismic processing algorithms (such as reflection coefficient method, time-frequency analysis, etc.). These features include reflection intensity, reflection time, amplitude change, phase information, etc. Deep neural networks (such as convolutional neural networks CNN) or other machine learning models are used for automatic feature extraction to identify implicit features of reservoirs from seismic signals, such as lithology, porosity, aquifers, etc., and generate seismic wave implicit reservoir feature data, which is a preliminary identification of the reservoir and contains reservoir information reflected in the reflection and propagation process of seismic waves. Based on the seismic wave implicit reservoir feature data, the spatial distribution characteristics of the seismic signal are used for mapping, and the feature data is mapped into three-dimensional space using a geographic information system (GIS) or a spatial interpolation algorithm. Methods such as Kriging interpolation or inverse distance weighting (IDW) can be used to map the extracted implicit reservoir characteristics in space and generate initial reservoir characteristic mapping data, which can show the preliminary characteristic distribution of the reservoir in three-dimensional space, including properties such as layer thickness and porosity. Decouple the initial reservoir characteristic mapping data from physical properties. Usually, the characteristics of the reservoir are decoupled by physical parameters such as wave velocity, density, porosity, and permeability. Physical modeling methods or machine learning methods (such as principal component analysis PCA) can be used to extract information related to physical properties from seismic wave data. Common decoupling methods include identifying lithology changes, fluid state, porosity, etc. in the reservoir through inversion algorithms of wave velocity and density, and generating reservoir physical property attribute data, which can show the physical properties of the reservoir such as lithology, porosity, and permeability. Nonlinear correlation analysis is performed on reservoir physical property attribute data to explore the complex relationship between different physical properties. For example, the nonlinear relationship between porosity and permeability, or the association between lithology and fluid saturation. Nonlinear regression analysis, support vector machine (SVM), neural network (such as deep autoencoder) and other methods can be used for analysis to identify high-order nonlinear characteristics between different physical properties and generate high-order reservoir characteristic data, which can reflect deeper physical phenomena in the reservoir and reveal the complexity of the reservoir. Combine high-order reservoir characteristic data with seismic multi-wavefield separation data for inversion linkage optimization. Through seismic inversion technology, reservoir characteristic data and seismic wave reflection data are jointly optimized to improve the accuracy of the reservoir model. Inversion methods such as waveform inversion, least squares inversion, full waveform inversion, etc. can be used to optimize the dynamic response of the reservoir and generate reservoir dynamic response optimization data, which can accurately predict the response of the reservoir under different pressure and fluid conditions and help optimize oil and gas exploration and production strategies. The reservoir dynamic response optimization data is decomposed into frequency components. The signal is decomposed into different frequency components using methods such as Fourier transform, wavelet transform or EMD (empirical mode decomposition) to help identify the response characteristics of the reservoir at different frequencies.The decomposed data can reveal the high-frequency and low-frequency response characteristics of the reservoir, which helps to analyze the different characteristics of the reservoir, such as the physical elasticity and fluid state of the reservoir, and generate a multi-frequency reservoir characteristic map, which shows the characteristic distribution of the reservoir at different frequencies and provides more comprehensive reservoir information. Based on the multi-frequency reservoir characteristic map, the reservoir details are enhanced. Image processing techniques such as super-resolution reconstruction and detail enhancement filtering can be used to enhance the low-resolution details in the map. Using a high-resolution reconstruction algorithm, the microstructure and details of the reservoir are enhanced to a higher spatial resolution, so as to better display the detailed characteristics of the reservoir and generate a high-resolution reservoir detail display map, which can show the microstructural characteristics of the reservoir and help further analyze the performance of the reservoir, such as the recoverability, permeability and production potential of the reservoir.

[0041] Preferably, performing spatial distribution mapping on seismic multi-wavefield separation data according to seismic wave implicit reservoir characteristic data includes: Confirm the seismic acquisition points of the seismic multi-wave field separation data to obtain seismic acquisition point information data; confirm the geographic coordinates of the seismic acquisition point information data to obtain geographic space coordinate data; The seismic multi-wave field separation data and the geographic space coordinate data are mapped to a spatial coordinate grid to generate seismic multi-wave field grid mapping data; the propagation range of the seismic multi-wave field grid mapping data is evolved according to the seismic wave implicit reservoir characteristic data to generate seismic propagation evolution path data; Based on the seismic propagation evolution path data, the seismic multi-wave field grid mapping data is refined into grid units to generate seismic wave spatial propagation area data; the local implicit mode of the seismic wave implicit reservoir characteristic data is extracted to obtain the local implicit mode data of the seismic wave; The grid unit reservoir characteristic values ​​of the seismic wave spatial propagation area data are calculated according to the local implicit mode data of the seismic wave to obtain the grid unit reservoir characteristic values; the seismic multi-wave field grid mapping data are spatially jointly mapped based on the grid unit reservoir characteristic values ​​to generate initial reservoir characteristic mapping data.

[0042] In an embodiment of the present invention, the location of each collection point is confirmed by using a collection device such as a seismograph, a vibrometer, etc. according to the source of the seismic signal. The specific location of each collection point can be recorded using the GPS or other positioning system of the seismic signal collection device. The seismic signal is processed using seismic data processing software to determine the collection point coordinates of the signal and extract the collection point information data. The result is seismic collection point information data, including the collection point number, collection time, collection signal type, etc. The GPS information recorded in the seismic collection point information data is used to confirm the accurate geographic coordinates (longitude, latitude, elevation, etc.) of each collection point. The collection point data can be docked with the map coordinate system through the geographic information system (GIS) software to obtain the geographic spatial coordinate data of each collection point. The data includes the precise location of each collection point, providing a basis for subsequent spatial mapping. The seismic multi-wave field separation data is combined with the geographic spatial coordinate data, and the data is mapped to a predetermined grid system based on the geographic coordinate system (such as the UTM coordinate system or the longitude and latitude coordinate system). The seismic data is mapped to a uniform grid using a spatial interpolation algorithm (such as Kriging interpolation or inverse distance weighted method) to ensure that each seismic data point corresponds to a grid unit in space. The generated seismic multi-wavefield grid mapping data can accurately represent the multi-wavefield characteristics at different locations during the propagation of seismic waves. Based on the implicit reservoir characteristic data of seismic waves (such as wave velocity, reflectivity, absorption characteristics, etc.), the propagation range of the seismic multi-wavefield grid mapping data is evolved. The wave equation or the physics-based seismic wave propagation model is used to simulate the propagation process of seismic waves under different geological conditions to generate seismic propagation evolution path data, which describes the propagation path and its evolution process of seismic waves from the source to each acquisition point. According to the seismic propagation evolution path data, the seismic multi-wavefield grid mapping data is refined. Through the sensitivity analysis of the propagation path and wave velocity changes, the area affecting the propagation is divided into finer grids. The grid units can be locally encrypted according to the propagation velocity, wave type and reflection characteristics, thereby improving the resolution of the seismic wave propagation area. The result is the generation of seismic wave spatial propagation regional data, which shows the different areas of seismic wave propagation and reflects the complexity of propagation through refined grid units. Local implicit modes are extracted from seismic wave implicit reservoir characteristic data using deep learning or modal decomposition techniques (such as variational mode decomposition VMD). Modal decomposition is performed on seismic data of different frequency bands or different time windows to extract implicit modes related to seismic wave propagation, reflection and reservoir characteristics, and generate seismic wave local implicit mode data, which can describe the fluctuation mode and characteristics of the reservoir in different local areas. Using seismic wave local implicit mode data, reservoir characteristic values ​​are calculated for each refined grid unit of seismic wave spatial propagation regional data.By analyzing the implicit modes and wave propagation characteristics in different grid cells, the characteristic values ​​related to reservoir characteristics (such as porosity, permeability, lithology, etc.) are calculated to generate grid cell reservoir characteristic values, which helps to further describe the physical properties and reflection characteristics of the reservoir. The grid cell reservoir characteristic values ​​and seismic multi-wavefield grid mapping data are used to jointly map the two to generate a complete reservoir space framework. Spatial mapping algorithms (such as seismic data inversion, interactive mapping, etc.) are used to combine reservoir characteristic values ​​with seismic signal data to form a preliminary reservoir model and generate initial reservoir characteristic mapping data, which represents the distribution characteristics of the reservoir in space and contains joint information of seismic wave propagation and reservoir characteristics.

[0043] Preferably, the inversion linkage optimization of the high-order reservoir characteristic data and the seismic multi-wavefield separation data includes: Conduct frequency domain analysis on seismic multi-wave field separation data to generate seismic full-wave frequency domain data; construct nonlinear wave equations for seismic full-wave data based on seismic full-wave frequency domain data to generate reservoir medium wave equations; Finite difference calculation is performed according to the reservoir medium wave equation to obtain reservoir medium wave differential data; the reservoir medium wave equation is numerically solved and simulated through the reservoir medium wave differential data to generate simulated seismic wave field data; By simulating seismic wave field data, high-order reservoir characteristic data are inverted and optimized to generate reservoir dynamic response optimization data.

[0044] In an embodiment of the present invention, the seismic multi-wave field separation data is converted from the time domain to the frequency domain by Fourier transform or wavelet transform, and the frequency components of the seismic waves are extracted. The seismic data is divided into multiple frequency bands by using a frequency analysis method (such as fast Fourier transform FFT), so that the waveform characteristics at different frequencies can be analyzed more clearly, and seismic full-wave frequency domain data is generated. The data contains seismic wave information of different frequency components, which can help to deeply understand the propagation characteristics of seismic waves in the reservoir medium. According to the seismic full-wave frequency domain data, a nonlinear wave equation describing the propagation of seismic waves in the reservoir medium is constructed. These equations take into account factors such as the nonlinear effect of the wave, the complexity of the medium, and the change in wave velocity. The nonlinear wave equation can be analogous to the conventional wave equation (such as the Helmholtz equation, the wave propagation equation), and nonlinear terms (such as reflection and refraction terms that depend on the properties of the medium) are added. Through these equations, the propagation of seismic waves in complex reservoirs can be described more accurately, thereby affecting the inversion process of the wave field. Numerical calculations are performed by the finite difference method (FDM) using the reservoir medium wave equation. The finite difference method can discretize the continuous wave equation and convert it into a solvable difference equation. Through the grid reservoir model (for example, through an orthogonal grid or an unstructured grid), the equation is discretized in the time and space domain, and the fluctuation value of each grid point is calculated to generate reservoir medium fluctuation difference data. These data represent the numerical simulation results of the propagation of seismic waves in the reservoir medium, which can reflect the propagation characteristics, reflection, refraction and other phenomena of the wave. Based on the reservoir medium fluctuation difference data, the reservoir medium wave equation is numerically solved and simulated. Explicit or implicit numerical methods such as the Lax-Friedrichs method or the gradient descent method can be used. In the numerical simulation process, the physical properties of the reservoir (such as elastic modulus, density, porosity, etc.) are combined to simulate the propagation of seismic waves in the reservoir, and the heterogeneity and anisotropy of the reservoir are considered to generate simulated seismic wave field data. This data reflects the whole process of seismic wave propagation in the reservoir medium, including reflection, refraction and interaction of waves, which can help to better understand the dynamic response of the reservoir. The simulated seismic wave field data is used to invert and optimize the high-order reservoir characteristic data. During the inversion process, optimization algorithms such as least squares method, genetic algorithm, and back propagation can be used, with the goal of minimizing the error between the simulated wave field and the actual seismic wave field. According to the inversion results, the physical parameters of the reservoir, such as porosity, permeability, wave velocity, etc., are optimized to more accurately describe the actual dynamic response of the reservoir and generate reservoir dynamic response optimization data, which can describe in detail the dynamic changes of the reservoir under different wave field conditions, optimize the reservoir characteristic model, and provide support for oil and gas exploration and development.

[0045] Preferably, step S23 includes the following steps: Step S231: performing wavelet packet decomposition on reservoir dynamic response optimization data to generate low-frequency component data and high-frequency component data; performing geological labeling on the low-frequency component data and the high-frequency component data to generate low-frequency component labeled data and high-frequency component labeled data; Step S232: reconstructing the frequency band characteristics of the reservoir dynamic response optimization data to generate a multi-band time domain image; performing spatial feature mapping on the multi-band time domain image to generate a multi-frequency reservoir characteristic map; Step S233: enhancing the texture features of the multi-frequency reservoir characteristic map based on the high-frequency component annotation data to generate map local texture feature enhancement data; Step S234: performing trend fitting on the multi-frequency reservoir characteristic map based on the low-frequency component annotation data to generate map trend fitting data; performing overall trend optimization on the multi-frequency reservoir characteristic map through the map trend fitting data to generate map overall trend optimization data; Step S235: Optimize and integrate the multi-frequency reservoir characteristic map based on the map local texture feature enhancement data and the map overall trend optimization data, thereby obtaining a high-resolution reservoir detail display map.

[0046] In an embodiment of the present invention, the reservoir dynamic response optimization data is decomposed into multiple frequency bands, especially low-frequency components and high-frequency components, by using wavelet packet transformation (WPD). Wavelet packet transformation can capture the different frequency information of the signal more finely, and is suitable for analyzing non-stationary signals, such as reservoir dynamic response signals. During the decomposition process, a suitable wavelet basis function (such as Daubechies wavelet) can be selected and the number of decomposition layers can be set to balance the frequency resolution and time resolution according to the characteristics of the reservoir data. According to the geological characteristics of the reservoir (such as rock formation type, porosity, permeability, etc.), the low-frequency component data and the high-frequency component data are labeled. Manual labeling or automatic labeling methods based on machine learning, such as support vector machines (SVM) or convolutional neural networks (CNN) can be used. The low-frequency component labeled data and the high-frequency component labeled data will contain geological information of the reservoir, such as rock formation boundaries, porosity changes and other characteristics, which will help subsequent feature analysis and map construction. Based on the low-frequency component data and high-frequency component data after wavelet packet decomposition, the characteristics of each frequency band are recombined through inverse wavelet packet transform or frequency band reconstruction method to generate multi-band time domain images. These images can show the time domain characteristics of reservoir dynamic response under different frequency bands and reflect the changes of reservoir in time. Multi-band time domain images are spatially mapped. The spatial mapping process usually involves aligning time domain data with geographic spatial coordinates (such as geological coordinate systems based on exploration areas), and using spatial interpolation methods (such as Kriging interpolation, inverse distance weighted interpolation, etc.) to generate spatial distribution data. The generated multi-frequency reservoir characteristic map will contain reservoir characteristic distributions in different frequency bands and can reveal the spatial heterogeneity of the reservoir. Based on the high-frequency component annotation data, the multi-frequency reservoir characteristic map is enhanced by image processing methods. High-frequency components usually contain details and rapidly changing information, which helps to capture subtle changes in the reservoir. Texture enhancement is performed using methods such as Gabor filtering and Laplace pyramid to highlight small-scale structural features inside the reservoir, such as fractures and pores. The enhanced data can help identify and highlight the heterogeneity and complexity of the reservoir at a local scale. Based on the low-frequency component annotation data, the trend fitting technology is used to analyze the multi-frequency reservoir characteristic map. The low-frequency component usually reflects the overall trend and large-scale characteristics of the reservoir, so it can be used to fit the long-term trend or change law of the reservoir. Trend modeling is performed using methods such as polynomial regression, curve fitting, and local weighted regression to obtain the map trend fitting data, that is, the large-scale change pattern of the reservoir. According to the map trend fitting data, the overall trend of the multi-frequency reservoir characteristic map is optimized to remove unnecessary noise and fluctuations, highlight the global trend and macro structure of the reservoir, and generate the map overall trend optimization data, which shows the basic structure and trend of the reservoir and helps to deeply understand the main characteristics of the reservoir and its geological background.The multi-frequency reservoir characteristic map is optimized and integrated by combining the local texture feature enhancement data of the map and the overall trend optimization data of the map. By fusing local details and global trends, the complex characteristics of the reservoir can be presented more comprehensively. The optimization and integration are performed by using weighted average and fusion algorithms, so that the map can simultaneously show details and macro structures and retain the true physical properties of the reservoir. The high-resolution reservoir detail display map is generated by optimizing and integrating the data. The map shows the details and overall trends of the reservoir, has a high spatial resolution, and can accurately display the complex structure of the reservoir.

[0047] Preferably, step S233 includes the following steps: Based on the high-frequency component annotation data, the high-frequency texture area of ​​the multi-frequency reservoir characteristic map is located to generate high-frequency texture area data; the high-frequency texture area data is separated into feature areas to generate high-frequency texture separation area data; Grayscale co-occurrence matrix calculation is performed on the high-frequency texture separation area data to obtain texture information data of the high-frequency texture area; local contrast enhancement is performed on the texture information data of the high-frequency texture area to generate texture enhancement data of the high-frequency texture area; texture component matching and alignment is performed on the high-frequency texture area data according to the texture enhancement data of the high-frequency texture area to generate high-frequency texture alignment area data; The multi-frequency reservoir characteristic map is updated with map details through high-frequency texture alignment regional data to generate map local texture feature enhancement data.

[0048] In an embodiment of the present invention, by using information from high-frequency component annotation data, it is combined with a multi-frequency reservoir characteristic map, and texture area positioning is performed by image processing technology. High-frequency components usually reflect details and rapidly changing structures, mainly concentrated in small cracks, pores and other areas of the reservoir. Edge detection, texture analysis and other algorithms (such as Canny edge detection, Laplacian pyramid or Sobel operator) are used to analyze the multi-frequency map and mark the high-frequency texture area in the frequency map. These areas represent high-frequency changes or complex details in the reservoir. By locating these high-frequency areas, high-frequency texture area data are generated, which contain detailed information of the reservoir under high-frequency components. Based on the high-frequency texture area data, the complex texture area in the image is further separated. The separation process can extract different texture feature areas by means of threshold segmentation, regional growing algorithm and the like. Pixel clustering (such as K-means clustering) or morphological segmentation method (such as morphological opening and closing operation) is used to segment the high-frequency texture area into several feature areas, so as to analyze the reservoir characteristics of different areas more carefully, and generate high-frequency texture separation area data, which contains spatial information and feature identification of different texture feature areas. Gray Level Co-occurrence Matrix (GLCM) is calculated for the high-frequency texture separation area data. Gray Level Co-occurrence Matrix is ​​a method used to describe image texture, which can reveal the spatial relationship and texture characteristics between pixels in the image. Gray Level Co-occurrence Matrix of different directions and distances is calculated, and texture features such as contrast, uniformity, energy, entropy, etc. are extracted from it. These features can help to deeply understand the texture information of the high-frequency area of ​​the reservoir. The texture information data of the high-frequency texture area is obtained, which contains the detailed texture features of the high-frequency texture area of ​​the reservoir, such as roughness, contrast, etc. Based on the texture information data of the high-frequency texture area, local contrast enhancement is performed. Contrast enhancement can enhance the details in the image and highlight the tiny structural changes in the reservoir. The texture information data of the high-frequency texture area is enhanced by using a local contrast adjustment algorithm (such as local histogram equalization or CLAHE algorithm) to improve the recognizability of the reservoir texture details and generate texture enhancement data of the high-frequency texture area, which is characterized by a clearer and easier to analyze high-frequency texture area of ​​the reservoir, which is convenient for subsequent image processing and analysis. Based on the texture enhancement data of the high-frequency texture area, the texture areas in different frequency bands are matched and aligned. The purpose of texture component matching and alignment is to ensure that the texture features of the reservoir can be accurately aligned at different image scales. Image registration technology (such as feature point-based registration algorithm, optical flow method or mutual information method) is used to accurately align the texture areas in the high-frequency texture area data to ensure that all texture features correspond to the same spatial coordinates, and generate high-frequency texture alignment area data, which can accurately map the texture features of the reservoir and ensure the consistency of texture areas at different scales.The high-frequency texture alignment regional data is used to update the atlas details of the multi-frequency reservoir characteristic atlas. By updating the atlas details, the local details in the image can be enhanced while retaining the global trend, thereby improving the overall resolution and accuracy of the atlas. The updated high-frequency texture data is fused with the original atlas data using a texture enhancement algorithm (such as hierarchical weighted fusion or multi-scale fusion method) to optimize the atlas detail presentation and generate atlas local texture feature enhancement data, which will show the enhanced details of the fine structure in the reservoir and provide a more accurate reservoir representation.

[0049] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes: Step S31: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; dividing the dynamic reservoir characteristic change data into data sets to generate a model training set and a model test set; Step S32: using a convolutional neural network algorithm to perform model training on the model training set to generate a reservoir characteristic change prediction pre-model; performing model optimization iteration on the reservoir characteristic change prediction pre-model through a model test set to generate a reservoir characteristic change prediction model; Step S33: importing the dynamic reservoir characteristic change data into the reservoir characteristic change prediction model to predict the reservoir characteristic change, thereby generating reservoir characteristic change prediction data; Step S34: dynamically updating the high-resolution reservoir detail display map through the reservoir characteristic change prediction data to generate a reservoir characteristic change prediction map.

[0050] In the embodiment of the present invention, by using high-resolution reservoir detail display images (obtained from the previous step), these display images provide the fine structure and characteristics of the reservoir. Image processing and dynamic analysis algorithms (such as time series analysis, wavelet transform, edge detection, etc.) are used to analyze the characteristic changes in the reservoir image. By comparing the reservoir detail images at different times or states, the dynamic change information of the reservoir at each moment is extracted to generate dynamic reservoir characteristic change data, which reflects the characteristic changes of the reservoir at different time points or states, including the dynamic evolution of characteristics such as porosity, fractures, and permeability. The dynamic reservoir characteristic change data is divided to generate two data sets: a model training set and a model test set. The training set is used to train the prediction model, and the test set is used to evaluate the performance and generalization ability of the model. The data set is randomly divided into a training set and a test set using a common data partitioning method (such as the 80 / 20 or 70 / 30 rule). Ensure that the test set and the training set are representative in data distribution. Two data sets are obtained: a model training set (for model training) and a model test set (for model verification and optimization). Using the data in the model training set, a convolutional neural network (CNN) is selected as the main deep learning algorithm. CNN can extract multi-level feature information from image data, and is particularly suitable for spatial and local structural changes of reservoir characteristics. A CNN architecture for predicting specific reservoir changes is constructed, which usually includes multiple convolutional layers, pooling layers, fully connected layers, etc. Back propagation training is performed through the training set, and the network weights are adjusted to minimize the loss function (such as mean square error). A preliminary reservoir characteristic change prediction pre-model is obtained, which can predict the trend of reservoir characteristic changes to a certain extent. The trained pre-model is evaluated using the data in the model test set, and the model is optimized based on the test set results. The model hyperparameters (such as convolution kernel size, learning rate, number of layers, etc.) are tuned using cross-validation, grid search and other methods, and the model training iteration is performed using advanced optimization algorithms (such as Adam, RMSProp) to improve the accuracy and stability of the model. Through multiple iterative optimizations, the final reservoir characteristic change prediction model is generated, which can more accurately predict the trend of reservoir characteristic changes over time or state. The dynamic reservoir characteristic change data (obtained from step S31) is imported into the reservoir characteristic change prediction model to predict characteristic changes. Using the trained and optimized reservoir characteristic change prediction model, forward reasoning is performed on the new dynamic reservoir characteristic change data to predict changes in reservoir characteristics at future moments or under different environments, and reservoir characteristic change prediction data is generated, which reflects the changes in reservoir characteristics at the predicted moment or state, including porosity changes, fracture evolution and other characteristics. The reservoir characteristic change prediction data (obtained from step S33) is used to update the previous high-resolution reservoir detail display map. This step combines the predicted reservoir characteristic change data with the original display map to obtain a new image representation.Image fusion technology or pixel-level update methods (such as image deformation, color mapping, etc.) are used to superimpose the reservoir characteristic change prediction data onto the original reservoir display image, thereby obtaining an updated reservoir detail map and generating a reservoir characteristic change prediction map. This is a dynamically updated image that shows the evolution of reservoir characteristics at different time points or environmental conditions, providing a more accurate prediction of reservoir change trends.

[0051] Preferably, step S4 comprises the following steps: Step S41: performing reservoir potential assessment on the reservoir characteristic change prediction map to generate reservoir potential assessment data; performing reservoir classification on the reservoir characteristic change prediction map according to the reservoir potential assessment data to generate a reservoir classification map; Step S42: performing actual reservoir observation on the reservoir classification map to generate actual reservoir observation result data; performing confidence calculation on the actual reservoir observation result data and the reservoir classification map to generate reservoir prediction confidence data; Step S43: Compare the reservoir prediction confidence data with the preset confidence threshold. When the reservoir prediction confidence data is greater than or equal to the preset confidence threshold, output the map based on the reservoir classification map to generate a high-resolution reservoir prediction distribution map. When the reservoir prediction confidence data is less than the preset confidence threshold, return to step S3 for feature re-extraction until the reservoir prediction confidence data is greater than or equal to the preset confidence threshold.

[0052] In the embodiment of the present invention, the reservoir characteristic change prediction map generated by using step S34 shows the characteristic change of the reservoir at different time points or states. The reservoir evaluation algorithm (such as comprehensive analysis of reservoir evaluation index, permeability and porosity, etc.) is applied to evaluate the potential of the reservoir. The reservoir potential evaluation model can be based on geology, petrology and physical models, combined with historical data for evaluation, and generate reservoir potential evaluation data, which describes the development potential of the reservoir, including recoverable reserves, production capacity, reservoir quality and other factors. The reservoir characteristic change prediction map is classified using the reservoir potential evaluation data. Reservoir classification is based on reservoir attributes such as permeability, porosity, saturation, etc., and reservoirs are divided into different types. Clustering algorithms (such as K-means, hierarchical clustering, etc.) or supervised learning algorithms (such as decision trees, random forests, etc.) are used for reservoir classification. Reservoirs can be classified according to their different potential evaluation results (such as high potential, medium potential, low potential) to generate a reservoir classification map, which shows the distribution of different reservoir types and provides support for subsequent analysis and development. Use reservoir classification maps to perform actual reservoir observations. These observations include drilling data, field survey data, production data, etc., providing actual feedback on reservoir characteristics and performance. Combine actual observation methods (such as seismic exploration, downhole measurement, oil production testing, etc.) to obtain actual reservoir data, and perform correlation analysis based on the reservoir classification map to generate actual reservoir observation result data, which includes actual observed reservoir characteristics and performance data, such as actual porosity, permeability, pressure and production capacity, etc. Confidence calculations are performed using actual reservoir observation result data and reservoir classification maps to evaluate the accuracy and reliability of reservoir classification. Confidence calculations are performed using statistical methods (such as Bayesian inference, Monte Carlo simulation, etc.) or confidence scoring models in machine learning (such as probability output of decision trees, confidence scores of support vector machines, etc.). This step will combine actual observation data and model predictions to generate a confidence value and generate reservoir prediction confidence data, which represents the confidence level of reservoir classification and prediction. The reservoir prediction confidence data is compared with the preset confidence threshold. The confidence threshold is a standard set according to project requirements, indicating the degree of trust in the reservoir prediction. The reservoir prediction confidence data is compared with the threshold through simple conditional judgment (such as if statement). If the reservoir prediction confidence data is greater than or equal to the preset threshold, proceed to the next step, otherwise return to step S3 for feature re-extraction. If the confidence data meets the threshold requirement, a high-resolution reservoir prediction distribution map is generated based on the reservoir classification map. This figure shows the high-resolution prediction of the spatial distribution of reservoir characteristics, presenting information such as the exploitability and development potential of the reservoir. Use data visualization technology (such as heat map, contour map, etc.) to combine the reservoir classification map with the spatial information and prediction data of the reservoir for high-resolution rendering, generate a reservoir prediction distribution map, and generate a high-resolution reservoir prediction distribution map to provide an intuitive reference for subsequent development decisions.If the reservoir prediction confidence data does not reach the preset threshold, return to step S3 to re-extract features. By adjusting the feature extraction method or using more high-resolution observation data for re-extraction, the model is improved to increase the accuracy of the prediction. Based on the adjusted features and optimized prediction results, the confidence calculation is performed again until the preset confidence threshold is met.

[0053] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0054] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A high-resolution reservoir prediction method based on seismic full-wave information mining, characterized in that: The following steps are involved: Step S1: Acquire seismic full-wave information data; perform frequency domain joint analysis on the seismic full-wave information data to generate a multi-scale full-wave characteristic tensor; perform multi-scale decoupling of seismic full-wave signals on the multi-scale full-wave characteristic tensor to generate seismic multi-wave field separation data; Step S2: perform implicit reservoir characteristic mapping on the seismic multi-wavefield separation data to generate initial reservoir characteristic mapping data; perform nonlinear correlation analysis on the initial reservoir characteristic mapping data to generate high-order reservoir characteristic data; perform inversion linkage optimization on the high-order reservoir characteristic data and the seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; perform reservoir detail enhancement on the reservoir dynamic response optimization data to generate a high-resolution reservoir detail display diagram; Step S3: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; performing reservoir characteristic change prediction on the high-resolution reservoir detail display image based on the dynamic reservoir characteristic change data to generate a reservoir characteristic change prediction image; Step S4: classify the reservoir characteristics change prediction map to generate a reservoir classification map; calculate the confidence of the reservoir classification map to generate reservoir prediction confidence data; compare and output the reservoir prediction confidence data with a preset confidence threshold until a high-resolution reservoir prediction distribution map is generated.

2. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: using a multi-point distributed seismic signal acquisition array to acquire seismic full-wave information data; Step S12: preprocessing the seismic full-wave information data to generate standard seismic full-wave information data, wherein the data preprocessing includes data cleaning, data missing value filling and data standardization; Step S13: performing time domain analysis on the full-wave signal to generate full-wave signal time domain data; performing fast Fourier transform on the full-wave signal time domain data to generate full-wave signal frequency domain data; performing frequency domain joint analysis based on the full-wave signal time domain data and the full-wave signal frequency domain data to generate a multi-scale full-wave feature tensor; Step S14: performing multi-scale decoupling of seismic full-wave signals on the multi-scale full-wave characteristic tensor based on preset seismic wave propagation characteristic coefficients to generate seismic multi-wave field separation data.

3. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 2 is characterized in that: Step S14 includes the following steps: Step S141: performing seismic wave transmission velocity analysis on the multi-scale full-wave characteristic tensor to generate seismic wave transmission velocity data; performing seismic wave type classification on the multi-scale full-wave characteristic tensor based on the seismic wave transmission velocity data to generate seismic wave type classification data, wherein the seismic wave type classification data includes longitudinal wave data, shear wave data and surface wave data; Step S142: performing variational modal decomposition on the P-wave data, S-wave data and surface wave data to generate P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data; performing wave field separation on the P-wave data, S-wave data and surface wave data through the P-wave modal decomposition data, S-wave modal decomposition data and surface wave modal decomposition data to generate seismic full-wave wave field separation data; Step S143: performing wavefield characteristic coupling analysis on the seismic full-wave wavefield separation data to generate wavefield characteristic coupling data; performing multi-scale decoupling of the seismic full-wave signal on the multi-scale full-wave characteristic tensor using the wavefield characteristic coupling data to generate seismic multi-wavefield separation data.

4. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 1 is characterized in that: Step S2 includes the following steps: Step S21: extracting implicit reservoir characteristics from the seismic multi-wavefield separation data to obtain seismic wave implicit reservoir characteristic data; performing spatial distribution mapping on the seismic multi-wavefield separation data according to the seismic wave implicit reservoir characteristic data to generate initial reservoir characteristic mapping data; Step S22: decoupling the initial reservoir characteristic mapping data from physical properties to generate reservoir physical property data; performing nonlinear correlation analysis on the reservoir physical property data to generate high-order reservoir characteristic data; performing inversion linkage optimization on the high-order reservoir characteristic data and seismic multi-wavefield separation data to generate reservoir dynamic response optimization data; Step S23: performing frequency component decomposition on the reservoir dynamic response optimization data to generate a multi-frequency reservoir characteristic spectrum; performing reservoir detail enhancement on the multi-frequency reservoir characteristic spectrum to generate a high-resolution reservoir detail display map.

5. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 4 is characterized in that: The spatial distribution mapping of seismic multi-wave field separation data based on seismic wave implicit reservoir characteristic data includes: Confirm the seismic acquisition points of the seismic multi-wave field separation data to obtain seismic acquisition point information data; confirm the geographic coordinates of the seismic acquisition point information data to obtain geographic space coordinate data; The seismic multi-wave field separation data and the geographic space coordinate data are mapped to a spatial coordinate grid to generate seismic multi-wave field grid mapping data; the propagation range of the seismic multi-wave field grid mapping data is evolved according to the seismic wave implicit reservoir characteristic data to generate seismic propagation evolution path data; Based on the seismic propagation evolution path data, the seismic multi-wave field grid mapping data is refined into grid units to generate seismic wave spatial propagation area data; the local implicit mode of the seismic wave implicit reservoir characteristic data is extracted to obtain the local implicit mode data of the seismic wave; The grid unit reservoir characteristic values ​​of the seismic wave spatial propagation area data are calculated according to the local implicit mode data of the seismic wave to obtain the grid unit reservoir characteristic values; the seismic multi-wave field grid mapping data are spatially jointly mapped based on the grid unit reservoir characteristic values ​​to generate initial reservoir characteristic mapping data.

6. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 4 is characterized in that: The inversion linkage optimization of high-order reservoir characteristic data and seismic multi-wave field separation data includes: Conduct frequency domain analysis on seismic multi-wave field separation data to generate seismic full-wave frequency domain data; construct nonlinear wave equations for seismic full-wave data based on seismic full-wave frequency domain data to generate reservoir medium wave equations; Finite difference calculation is performed according to the reservoir medium wave equation to obtain reservoir medium wave differential data; the reservoir medium wave equation is numerically solved and simulated through the reservoir medium wave differential data to generate simulated seismic wave field data; By simulating seismic wave field data, high-order reservoir characteristic data are inverted and optimized to generate reservoir dynamic response optimization data.

7. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 4 is characterized in that: Step S23 includes the following steps: Step S231: performing wavelet packet decomposition on reservoir dynamic response optimization data to generate low-frequency component data and high-frequency component data; performing geological labeling on the low-frequency component data and the high-frequency component data to generate low-frequency component labeled data and high-frequency component labeled data; Step S232: reconstructing the frequency band characteristics of the reservoir dynamic response optimization data to generate a multi-band time domain image; performing spatial feature mapping on the multi-band time domain image to generate a multi-frequency reservoir characteristic map; Step S233: enhancing the texture features of the multi-frequency reservoir characteristic map based on the high-frequency component annotation data to generate map local texture feature enhancement data; Step S234: performing trend fitting on the multi-frequency reservoir characteristic map based on the low-frequency component annotation data to generate map trend fitting data; performing overall trend optimization on the multi-frequency reservoir characteristic map through the map trend fitting data to generate map overall trend optimization data; Step S235: Optimize and integrate the multi-frequency reservoir characteristic map according to the map local texture feature enhancement data and the map overall trend optimization data, thereby obtaining a high-resolution reservoir detail display map.

8. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 7 is characterized in that: Step S233 includes the following steps: Based on the high-frequency component annotation data, the high-frequency texture area of ​​the multi-frequency reservoir characteristic map is located to generate high-frequency texture area data; the high-frequency texture area data is separated into feature areas to generate high-frequency texture separation area data; Grayscale co-occurrence matrix calculation is performed on the high-frequency texture separation area data to obtain texture information data of the high-frequency texture area; local contrast enhancement is performed on the texture information data of the high-frequency texture area to generate texture enhancement data of the high-frequency texture area; texture component matching and alignment is performed on the high-frequency texture area data according to the texture enhancement data of the high-frequency texture area to generate high-frequency texture alignment area data; The multi-frequency reservoir characteristic map is updated with map details through high-frequency texture alignment regional data to generate map local texture feature enhancement data.

9. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 1, characterized in that: Step S3 includes the following steps: Step S31: performing dynamic reservoir characteristic change analysis on the high-resolution reservoir detail display image to generate dynamic reservoir characteristic change data; dividing the dynamic reservoir characteristic change data into data sets to generate a model training set and a model test set; Step S32: using a convolutional neural network algorithm to perform model training on the model training set to generate a reservoir characteristic change prediction pre-model; performing model optimization iteration on the reservoir characteristic change prediction pre-model through a model test set to generate a reservoir characteristic change prediction model; Step S33: importing the dynamic reservoir characteristic change data into the reservoir characteristic change prediction model to predict the reservoir characteristic change, thereby generating reservoir characteristic change prediction data; Step S34: dynamically updating the high-resolution reservoir detail display map through the reservoir characteristic change prediction data to generate a reservoir characteristic change prediction map.

10. The high-resolution reservoir prediction method based on seismic full-wave information mining according to claim 1, characterized in that: Step S4 includes the following steps: Step S41: performing reservoir potential assessment on the reservoir characteristic change prediction map to generate reservoir potential assessment data; performing reservoir classification on the reservoir characteristic change prediction map according to the reservoir potential assessment data to generate a reservoir classification map; Step S42: performing actual reservoir observation on the reservoir classification map to generate actual reservoir observation result data; performing confidence calculation on the actual reservoir observation result data and the reservoir classification map to generate reservoir prediction confidence data; Step S43: Compare the reservoir prediction confidence data with the preset confidence threshold. When the reservoir prediction confidence data is greater than or equal to the preset confidence threshold, output the map based on the reservoir classification map to generate a high-resolution reservoir prediction distribution map. When the reservoir prediction confidence data is less than the preset confidence threshold, return to step S3 for feature re-extraction until the reservoir prediction confidence data is greater than or equal to the preset confidence threshold.

Images