Fault sample generation method based on oil and gas field seismic data

Abstract

The invention discloses a fault sample generation method based on oil and gas field seismic data, and belongs to geophysical exploration and machine learning data engineering.The technical scheme includes the steps that seismic exploration data, well data and actual geological data of an actual work area are obtained; establishing an initial structure model based on seismic exploration data, well data and actual geological data; iteratively adjusting the initial structure model according to a forward modeling result to obtain a calibration structure model; performing process simulation based on the calibration structure model, and outputting a plurality of geologic bodies according to time steps; and converting the geologic body into a forward modeling parameter model, carrying out seismic forward modeling to generate a seismic data body, and carrying out data enhancement and sample amplification. The method has the advantages that compared with other sample generation methods, the generated samples better conform to the actual geological law, so that the sample quality is improved, and meanwhile, the number of the samples can be greatly increased by simulating the forward modeling process of the actual body.

Description

TECHNICAL FIELD

[0001] The present application relates to geophysical exploration and machine learning data engineering, in particular to a fault sample generation method based on oil and gas field seismic data. BACKGROUND

[0002] Fault interpretation is the basis of oil and gas field exploration and development work, and as the continuous advancement and deepening of exploration and development work, higher requirements are put forward for the accuracy and efficiency of fault interpretation. Among them, the efficiency improvement of artificial fault interpretation has become a key problem to be solved. Under this background, intelligent fault interpretation, as a hot development direction in recent years, can improve the efficiency of today's exploration and development work with its powerful data processing and analysis capability. The core advantage of intelligent fault interpretation is that it can use advanced artificial intelligence algorithms to efficiently and accurately analyze and interpret massive geological data, thereby greatly improving the accuracy and speed of fault interpretation. However, to achieve this goal, the scale and quality of sample data become the key factors restricting the performance improvement of intelligent fault interpretation. Specifically, the scale of the sample is directly related to the sufficiency and generalization ability of model training, and the quality of the sample directly determines the accuracy and reliability of the final interpretation result. Therefore, at the current stage, improving the quantity and quality of the sample has become an important task that needs to be solved in the work of intelligent fault interpretation. The existing intelligent fault recognition relies on large-scale labeled data, and the cost of real data labeling is high and lacks physical constraints consistent with the evolution of specific work areas, resulting in insufficient model generalization ability and weak interpretability. Therefore, a fault sample generation method based on oil and gas field seismic data is needed to meet the sample construction technology of geological rationality and approximate real acquisition statistical characteristics. SUMMARY

[0003] The purpose of the present application is to provide a fault sample generation method based on oil and gas field seismic data.

[0004] The present application is realized by the following measures: a fault sample generation method based on oil and gas field seismic data, characterized by comprising: Obtain seismic exploration data, well data and actual geological data of the actual work area; for the actual work area, require clear structural features, high degree of geological research, and high-quality seismic exploration data and well data.

[0005] The seismic exploration data is used to obtain the macroscopic morphology, fault distribution and structural style of the work area; The well data is used to obtain the rock physical parameters (such as core mechanical parameters, velocity, density, lithology, etc.) of the stratum; the actual geological data is used to prepare for establishing an initial structural model, and can include regional geological map, structural evolution history, ground stress field direction, denudation thickness, etc., to provide geological constraints for the model; An initial structural model was established based on seismic exploration data, well data, and actual geological data. The initial construction model is iteratively adjusted based on the forward simulation results to obtain the calibrated construction model; Based on the calibration construction model, process simulation is performed, and multiple geological bodies are output according to time steps; The geological body is converted into a forward modeling parameter model, and seismic forward modeling is performed to generate seismic data volumes, followed by data augmentation and sample amplification.

[0006] Furthermore, the initial construction model is established using discrete element method or a coupling method between discrete element method and finite element method.

[0007] Furthermore, establishing the initial construction model includes: Create a two-dimensional or three-dimensional geometric model, and define the model's dimensions and layers; Set boundary conditions and a gravity field.

[0008] Furthermore, the iterative calibration of the initial construction model is adjusted iteratively based on the forward simulation results as follows: The simulation results are quantitatively and qualitatively compared with the actual work area observations. During the iteration process, the parameters are adjusted until the forward simulation results of the initial construction model and the selected actual work area meet the quantitative and qualitative indicators.

[0009] Furthermore, when this method is used in extensional and strike-slip structural work areas, the boundary conditions include: Displacement boundary: Slow extrusion is applied to both sides of the calibrated construction model at a specific initial rate; Fixed Boundaries: The bottom of the calibration construction model is completely fixed, and the front and rear boundaries of the model restrict its normal displacement, allowing sliding along the strike to simulate the constraints of a real basin.

[0010] Furthermore, when this method is applied to extensional and strike-slip tectonic areas, the quantitative indicators are that the errors in fault dip angle and fault density are both <5%; the qualitative indicators are that the main fault, its conjugate secondary faults, and the rolling folds formed in the hanging wall due to the slip of the main fault are successfully simulated, and the spatial combination pattern matches the seismic profile with a degree of over 85%.

[0011] Furthermore, when this method is used to construct a work zone by extrusion, the boundary conditions include: Displacement boundary: A horizontal extrusion is applied to one side of the calibrated construction model at a specific initial constant rate; Bottom boundary: The bottom of the model is a slip surface, set as a free slip boundary, which allows for large-scale horizontal shortening.

[0012] Furthermore, when this method is used in a compressional structural work area, the quantitative indicators are that the fault spacing error is within ±10% and the fault propagation angle error is <6%; the qualitative indicators are that the imbricate thrust fault system, the double structure below it, and the triangular zone structure formed by the strong compressional folding at the fault front are successfully simulated, with an overall morphological similarity greater than 80%.

[0013] This embodiment provides a fault sample generation system based on seismic data from oil and gas fields, characterized in that it includes: The acquisition module acquires seismic exploration data, well data, and actual geological data for the actual work area; The construction module establishes an initial structural model based on seismic exploration data, well data, and actual geological data. The iteration module is used to iteratively adjust the initial construction model based on the forward simulation results to obtain the calibration construction model. The simulation module performs process simulation based on the calibration construction model and outputs multiple geological bodies according to time steps; Data augmentation and sample expansion module: Converts the geological body into a forward modeling parameter model, performs seismic forward modeling to generate seismic data volume, and performs data augmentation and sample expansion.

[0014] This embodiment provides an electronic device, characterized in that it includes a processor and a memory, wherein the processor is used to execute a program stored in the memory for a fault sample generation method based on oil and gas field seismic data, so as to implement the fault sample generation method based on oil and gas field seismic data.

[0015] The beneficial effects of the technical solution provided by the embodiments of the present invention are as follows: Based on the structural forward modeling method and structural dynamics, the present invention models actual complex geological bodies, and through a series of iterative adjustments, it can construct a structural model of a geological body with a high degree of similarity to the actual work area.

[0016] Compared with other sample generation methods, this invention generates samples that are more consistent with actual geological patterns, thereby improving sample quality. At the same time, by simulating the forward modeling process of actual bodies, it can greatly enrich the number of samples. Attached Figure Description

[0017] To more clearly illustrate the technical solution of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. Obviously, the drawings listed below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a flowchart of a method for generating fault samples based on seismic data from oil and gas fields, as described in an embodiment of the present invention. Detailed Implementation

[0019] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. Of course, the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0020] Example 1: See Figure 1 A method for generating fault samples based on seismic data from oil and gas fields, characterized by comprising: Obtain seismic exploration data, well data, and actual geological data for the actual work area; the actual work area should have clear structural features, a high level of geological research, and high-quality seismic exploration data and well data.

[0021] The seismic exploration data is used to obtain the macroscopic morphology, fault distribution, and structural style of the strata in the work area; The well data is used to obtain the rock physical parameters of the formation (such as core mechanical parameters, velocity, density, lithology, etc.); the actual geological data is used to prepare for the establishment of the initial structural model. The actual geological data may include regional geological maps, tectonic evolution history, geostress field direction, erosion thickness, etc., to provide geological constraints for the model. An initial structural model was established based on seismic exploration data, well data, and actual geological data. The initial construction model is iteratively adjusted based on the forward simulation results to obtain the calibrated construction model; Based on the calibration construction model, process simulation is performed, and multiple geological bodies are output according to time steps; The geological body is converted into a forward modeling parameter model, and seismic forward modeling is performed to generate seismic data volumes, followed by data augmentation and sample amplification.

[0022] The initial construction model is established using discrete element method or a coupling method between discrete element method and finite element method.

[0023] Establishing the initial construction model includes: Create a two-dimensional or three-dimensional geometric model, and define the model's dimensions and layers; Set boundary conditions and a gravity field.

[0024] The iterative calibration of the initial construction model is adjusted iteratively based on the forward simulation results as follows: The simulation results are quantitatively and qualitatively compared with the actual work area observations. During the iteration process, the parameters are adjusted until the forward simulation results of the initial construction model and the selected actual work area meet the quantitative and qualitative indicators.

[0025] Example 2: Based on Example 1, a Neogene hydrocarbon-rich area in a depression zone of an oilfield in the Bohai Bay Basin, belonging to an extensional strike-slip tectonic zone, was selected. The fault sample generation method based on oilfield seismic data was used, with the following specific steps: S1. Obtain seismic exploration data, well data, and actual geological data of the actual work area; collect 3D seismic exploration data (covering an area of ​​50 km²), well data (20 well locations), and actual geological data of the work area.

[0026] The logging curves (such as acoustic transit time AC and density DEN) of 20 wells were standardized, edited, and quality-controlled. Using seismic rock physics relationships such as the Gardner formula, lithological information was converted into wave impedance and velocity models, providing parameter constraints for the initial model. Well data included core mechanical parameters, including measured core friction coefficients (generally selected as 0.3-0.6) and compressive strength data. The core friction coefficient and compressive strength data were used to determine the range of micromechanical parameters for different lithological particles in the discrete element model.

[0027] Actual geological data, including tectonic evolution history, and comprehensive regional geological studies, divide the tectonic evolution since the Miocene into three main stages: the rifting period (~5 Ma), the depression period (~2 Ma), and the strike-slip alteration period (~1 Ma to present). This historical framework provides key temporal constraints for setting boundary conditions in numerical simulations.

[0028] S2. Establish an initial structural model based on seismic exploration data, well data, and actual geological data; Model Geometry and Layering: A 3D model measuring 10km x 5km x 3km was constructed. A 3000m deep basement served as the boundary, above which were interbedded Neogene sandstone and mudstone strata. Based on well logging data, 15 strata were incorporated into the model, with individual layer thicknesses randomly distributed between 10-30m. The sandstone layers were given a coarser grain (2mm), while the mudstone layers were given a finer grain (0.8mm). Weak Zone Implantation: Based on the fault plan interpreted from seismic data, a low-strength zone was implanted along the N60°E strike throughout the model (achieved by reducing the cohesion and friction coefficient between particles at this point), serving as the initial initiation point of the controlling fault.

[0029] Boundary condition settings include: Displacement boundary: Slow compression in the N45°E direction is applied to both sides of the model, with an initial rate of 0.5 mm / yr, to simulate the stress field of the region.

[0030] Fixed boundary: The bottom of the model is completely fixed, and the front and rear boundaries of the model restrict its normal displacement, allowing sliding along the strike to simulate the constraints of a real basin.

[0031] Gravitational field: gravitational acceleration applied g = 9.8 m / s².

[0032] Simulation run: Run the simulation using PFC3D software, with a total calculation time of 10 million years (Ma), and output the first preliminary results.

[0033] S3. Iteratively adjust the initial construction model based on the forward simulation results to obtain the calibration construction model; First round of quantitative and qualitative analysis: The three-dimensional fault model from the simulation results is converted into an earthquake-identifiable "fault resistive body" and compared with the coherence volume and ant volume attributes of actual seismic data.

[0034] Quantitative error: The average dip angle of the simulated fault is 55° < the actual 65° (error 15%); the fault density (fault length per unit area) is 0.8 km / km² < the actual 1.2 km / km².

[0035] Qualitative error: The simulated fault combination is too simple and lacks the rich secondary faults and "Y"-shaped conjugate fracture system observed in reality.

[0036] Parameter adjustment and iteration process: Iteration 1: Reduce the interparticle friction coefficient from 0.4 to 0.35, making the rock more susceptible to shear fracture.

[0037] Iteration 2: Increase the difference in mechanical properties between sandstone and mudstone (friction coefficient of sandstone 0.33, mudstone 0.38) to promote differential deformation and more complex fault combinations.

[0038] Iteration 3: Fine-tune the zone extrusion rate from 0.5 mm / yr to 0.6 mm / yr to accumulate greater strain in the same amount of time.

[0039] Iteration 4: Implant more randomly distributed low-intensity points in specific layers of the hanging wall of the main fault to guide the random emergence of secondary faults and increase fault density.

[0040] Iteration 5: Finely adjust the direction of the boundary force to make it form a smaller angle with the strike of the main fault, thereby promoting the generation of the strike-slip component.

[0041] Final calibration model: After five iterations, the simulation results showed a significant improvement in similarity to the actual work area.

[0042] Quantitative indicators: fault dip angle 63° (error <5%), fault density 1.15 km / km² (error <4%).

[0043] Qualitative indicators: The main fault, its conjugate secondary faults, and the rolling folds formed in the hanging wall due to the slippage of the main fault were successfully simulated, and their spatial combination pattern matched the seismic profile with a degree of consistency of over 85%. The final calibrated structural model was obtained.

[0044] S4. Based on the calibration construction model, perform process simulation and output multiple geological bodies according to time steps; Dynamic data output: After running the calibrated model, a total of 11 three-dimensional geological volumes (0 Ma to 10 Ma) are output at 1 Ma intervals. Each geological volume contains information on the location, velocity, stress state, and bond fracture (i.e., fault formation) of all particles.

[0045] Geological interpretation of key stages: During the rifting period (5 Ma): the main fault experienced intense activity, with vertical displacement reaching 800 m. The hanging wall strata rotated, forming a typical "rolling half-graben". Three sets of secondary faults developed in the local stress concentration zones on the footwall and hanging wall of the main fault.

[0046] Depression period (2 Ma): Regional stress field relaxed, and fault activity basically ceased. A new stratigraphic unit with a thickness of 200 m was deposited at the top of the model. These strata covered the earlier structures in the form of a drapery, which served to bury and preserve them.

[0047] Strike-slip alteration period (1 Ma to present): The regional stress direction has slightly shifted, and the early-formed normal faults have been reused, exhibiting significant strike-slip characteristics (strike-slip component accounts for 30%). At fault inflection points and intersections, complex and high-density "ribbon-like" fault fracture zones and fracture systems have formed, which are important oil and gas seepage channels in this area.

[0048] S5. Convert the geological body into a forward parameter model, perform seismic forward modeling to generate seismic data volumes, and perform data augmentation and sample expansion, specifically including: The 11 output geological bodies (selecting 10 key stages) are converted into forward modeling parameter models (velocity models), and velocity values ​​are assigned according to lithology.

[0049] Forward modeling was performed by solving the acoustic wave equations using the finite difference method. The dominant frequency of the Ricker wavelet was 30 Hz, the sampling rate was 2 ms, and the recording duration was 2 s. A centrally located shot and bilaterally received observation system was used to generate a 1000 (Inline) × 1000 (Crossline) × 500 (time sampling points) three-dimensional seismic data volume.

[0050] Data augmentation includes: (1) Add noise: Add random Gaussian noise with a signal-to-noise ratio of 10dB to simulate environmental noise and acquisition interference.

[0051] (2) Formation absorption effect: The formation filtering operator with Q value = 50 is used to filter the seismic traces to simulate the absorption and attenuation effect of the earth on high-frequency signals, so that the frequency band of the synthetic seismic record matches the actual seismic data.

[0052] (3) Wavelet variation: Slightly vary the dominant frequency (±5Hz) and phase of the wavelet in different data volumes to simulate the effect of different seismic acquisition and processing batches.

[0053] Sample amplification includes: (1) Sample cutting: For each enhanced seismic data volume, a 256×256 pixel slice is cut every 2 sampling points along the three directions of Inline, Crossline and time axis. 500 slices are cut in each direction, for a total of 10 stages × 3 directions × 500 slices = 15,000 slice images.

[0054] (2) Sample labeling: The position of each fault on each slice is accurately extracted to generate the corresponding binary fault label map (fault pixels are 1, non-fault pixels are 0). At the same time, the generated metadata also includes attribute labels such as fault displacement, fault dip, and dip angle corresponding to the slice.

[0055] The final result is a high-quality sample library containing 15,000 sets (seismic image slices, fault labels, attribute labels).

[0056] Example 3: Based on Example 1, the Kuqa foreland thrust belt in the Tarim Basin, which belongs to the compressional tectonic zone, was selected. The fault sample generation method based on oil and gas field seismic data was used, and the specific steps are as follows: S1. Acquire seismic exploration data, well data, and actual geological data for the actual work area. Specifically, high-fidelity processing of the 80 km² seismic exploration data is required, with a focus on pre-stack depth migration to improve the imaging quality of sub-salt structures. Attributes such as coherence, curvature, and ant-like structures are extracted for subsequent quantitative comparison with simulation results.

[0057] Well data includes core mechanical and rheological parameters. Core mechanical parameters include internal friction angle and cohesion obtained from triaxial compression experiments on mudstone and conglomerate cores, used to calibrate the particle contact model in the discrete element model. Creep experiments were conducted on gypsum-salt rock samples to obtain their steady-state creep rate, used to calibrate parameters of the viscoplastic constitutive model (viscosity ~1×10⁻⁶). 18 Pa·s).

[0058] Actual geological data, including tectonic evolution history, was used to verify that the direction of the maximum horizontal principal stress was N20°E, with a magnitude of 80-120 MPa, using wellbore collapse and fracturing data. This data served as the core basis for applying boundary forces. Combined with regional equilibrium profile analysis, the shortening and compression rate (0.8-1.2 mm / yr) during the Himalayan orogeny were determined, providing time-displacement constraints for the simulation.

[0059] S2. Establish an initial structural model based on seismic exploration data, well data, and actual geological data; Model geometry and layering: Construct a 2D cross-sectional model measuring 15km (length) x 8km (width) x 6km (depth) (expandable to 3D). The bottom of the model is the base slip surface at a depth of 4500m.

[0060] Bottom-up settings: Lower structural layer (below the basement-gypsum-salt layer): alternating layers of sandstone and mudstone, each layer 50-80m thick, set as discrete element particles (1-5mm in diameter) to simulate brittle deformation.

[0061] Slip layer: A 300 μm thick paste-salt layer with an initial viscosity set to 1 × 10⁻⁶. 19 Pa·s was used to simulate its flow characteristics.

[0062] Upper structural layer (above the gypsum-salt layer): a relatively thin layer of sandstone and mudstone, with a single layer thickness of 20-40m, also set as discrete element particles.

[0063] Boundary condition settings include: Displacement boundary: A horizontal compression in the N20°E direction is applied to the right side of the model, with an initial constant rate of 0.9 mm / yr.

[0064] Bottom boundary: The bottom of the model is a slip surface, set as a free slip boundary, which allows for large-scale horizontal shortening.

[0065] Gravitational field: Applying gravitational acceleration.

[0066] Simulation execution: A finite element-discrete element coupled method was employed. The flow of the gypsum-salt layer was simulated using FEM, while the fracturing of the upper and lower brittle strata was simulated using DEM. The tectonic deformation of the Himalayas during the middle 5 Ma period was simulated.

[0067] S3. Iteratively adjust the initial construction model based on the forward simulation results to obtain the calibration construction model; First round of quantitative and qualitative analysis: Quantitative deviations: The simulated reverse fault spacing is too large (800m, actual 400-600m); the fault propagation angle is too small (25°, actual 35°); the flow of gypsum-salt layer is not obvious, and salt arches have not been effectively formed.

[0068] Qualitative bias: lack of typical duplex structure and clear triangle zone morphology.

[0069] Parameter adjustment and iteration process: Iteration 1-2: Reduce the viscosity of the paste-salt layer from 1×10 19 Pa·s lowered to 5 × 10 18Pa·s enhances its fluidity, enabling it to better function as a slip layer and arch upwards.

[0070] Iteration 3-4: Optimize the particle size distribution of the brittle layer, increase the proportion of 1-3mm particles to 70% to improve the heterogeneity of the material and promote denser fracture (reduce the fault spacing).

[0071] Iteration 5: A phased variable speed loading strategy is adopted: slow extrusion at 0.6 mm / yr is used from 0 to 2 Ma to allow deformation to be fully transmitted and to initiate multiple faults; fast extrusion at 1.1 mm / yr is used from 2 to 5 Ma to accelerate the sliding of existing faults and connect them into a large structure.

[0072] Iterations 6-7: Fine-tuning the difference in friction coefficients between mudstone and conglomerate (mudstone decreased from 0.55 to 0.52, while conglomerate remained at 0.6), promoting selective slippage of the fault plane in the weaker layer (forming a fault plate).

[0073] Final calibration model: After seven iterations, the simulation results closely matched the actual seismic profile.

[0074] Quantitative indicators: Fault spacing 520m (error within ±8%); fault propagation angle 33° (error <6%). Qualitative indicators: Successfully simulated imbricate thrust fault system, its underlying dual structure (manifested as a group of closed fault blocks), and the triangular band structure formed by intense compression and folding at the fault front, with an overall morphological similarity of 82%. A calibrated structural model was finally obtained.

[0075] S4. Based on the calibration construction model, perform process simulation and output multiple geological bodies according to time steps; Dynamic data output: After running the calibrated model, high-resolution geological bodies are output at 11 time steps every 0.5 Ma, recording the displacement, strain and rupture history of each element.

[0076] Geological interpretation of key evolutionary stages: Initial thrust (1.5 Ma): Compressive stress activates the basement slip layer, the first main thrust fault breaks through the gypsum-salt layer and propagates upward, with a vertical displacement of 300 m, and the gypsum-salt layer begins to thicken locally.

[0077] Dual structural formation (3Ma): New thrust faults emerge sequentially on the footwall of the main fault, flowing upwards into the main detachment layer and downwards into the basement detachment layer, forming a typical "roof-floor" dual structural system, with a series of "steep at the top and gentle at the bottom" fault blocks inside.

[0078] Triangular zone formation (5 Ma): The thrust fault at the leading edge is blocked, and the stress is transformed into severe folding and uplift of the strata at the leading edge, forming a triangular zone with an uplift height of 1200 m and a dip angle of over 60°. The gypsum-salt layer undergoes significant plastic flow under tectonic loading, forming salt pillows.

[0079] S5. Convert the geological body into a forward parameter model, perform seismic forward modeling to generate seismic data volumes, and perform data augmentation and sample expansion, specifically including: Seismic forward modeling setup: Select geological bodies at 10 key evolution stages and transform them into three-dimensional elastic parameter models that include P-wave velocity, S-wave velocity, and density.

[0080] Forward modeling of elastic wave equations (such as the spectral method) is employed to more accurately simulate the anisotropy and complex wave field phenomena commonly found in compressional tectonic zones.

[0081] Wavelet: A zero-phase Rick wavelet with a dominant frequency of 20Hz and a bandwidth of 8-60Hz is used to simulate the wavelet characteristics of seismic strata in terrestrial formations.

[0082] Anisotropy: In fault zones and high-strain areas, transverse isotropic (VTI) parameters (Thomsen coefficients ε=0.15, δ=0.05) are introduced to simulate the effects of high-angle cracks and bedding.

[0083] Data augmentation includes: (1) Add noise: Add random noise and coherent noise with appropriate signal-to-noise ratio to simulate actual earthquake data.

[0084] (2) Velocity field perturbation: ±5% random perturbation is added to the velocity model used in forward modeling to simulate the errors in actual velocity analysis and the resulting construction artifacts.

[0085] (3) Offset error: The phenomenon of inaccurate offset positioning is simulated in part of the data volume, so that the maximum offset of the same phase axis reaches 5%.

[0086] Sample amplification includes: (1) Sample cutting: Vertical sections were cut along the traditional Inline / Crossline direction, and also along the structural dip (N20°E) and strike (N70°W) to display the structural features from the best perspective. Horizontal time slices were also cut at the same time. More than 2,000 image samples were generated for each processed seismic data volume.

[0087] (2) Sample labeling: Pixel-level labels are generated to label thrust faults, and special labels are also added for detachment, double tectonic boundaries, damage zones and triangular cores.

[0088] The final result is a sample library containing 20,000 sets (seismic images, fine structural labels).

[0089] Example 4: This embodiment provides a fault sample generation system based on seismic data from oil and gas fields, characterized in that it includes: The acquisition module acquires seismic exploration data, well data, and actual geological data for the actual work area; The construction module establishes an initial structural model based on seismic exploration data, well data, and actual geological data. The iteration module is used to iteratively adjust the initial construction model based on the forward simulation results to obtain the calibration construction model. The simulation module performs process simulation based on the calibration construction model and outputs multiple geological bodies according to time steps; Data augmentation and sample expansion module: Converts the geological body into a forward modeling parameter model, performs seismic forward modeling to generate seismic data volume, and performs data augmentation and sample expansion.

[0090] For the specific functions of each module, please refer to the relevant descriptions in the above method embodiments, which will not be repeated here.

[0091] Example 5: This embodiment provides an electronic device, characterized in that it includes a processor and a memory, wherein the processor is used to execute a program stored in the memory for a fault sample generation method based on oil and gas field seismic data, so as to implement the fault sample generation method based on oil and gas field seismic data in the above embodiment.

[0092] An electronic device includes at least one processor, memory, at least one network interface, and other user interfaces. The various components of the electronic device are coupled together via a bus system. It is understood that the bus system is used to enable communication between these components. In addition to a data bus, the bus system also includes a power bus, a control bus, and a status signal bus.

[0093] The user interface may include a display, keyboard, or clicking device (e.g., mouse, trackball, touchpad, or touchscreen). It is understood that the memory in this embodiment may be volatile memory or non-volatile memory, or may include both.

[0094] In this embodiment of the invention, the processor executes the method steps provided in each method embodiment by calling a program or instruction stored in the memory, specifically a program or instruction stored in an application program.

[0095] In some implementations, the memory stores elements such as executable units or data structures, or subsets thereof, or extended sets thereof: operating systems and applications.

[0096] The operating system includes various system programs, such as the framework layer, core library layer, and driver layer, used to implement various basic business functions and handle hardware-based tasks. The application programs include various applications, such as media players and browsers, used to implement various application functions. The program implementing the method of this invention can be included in the application programs.

[0097] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for generating fault samples based on seismic data from oil and gas fields, characterized in that, include: Acquire seismic exploration data, well data, and actual geological data for the actual work area; An initial structural model was established based on seismic exploration data, well data, and actual geological data. The initial construction model is iteratively adjusted based on the forward simulation results to obtain the calibrated construction model; Based on the calibration construction model, process simulation is performed, and multiple geological bodies are output according to time steps; The geological body is converted into a forward modeling parameter model, and seismic forward modeling is performed to generate seismic data volumes, followed by data augmentation and sample amplification.

2. The method for generating fault samples according to claim 1, characterized in that, The initial construction model is established using discrete element method or a coupling method between discrete element method and finite element method.

3. The method for generating fault samples according to claim 2, characterized in that, Establishing the initial construction model includes: Create a two-dimensional or three-dimensional geometric model, and define the model's dimensions and layers; Set boundary conditions and a gravity field.

4. The method for generating fault samples according to claim 3, characterized in that, The iterative calibration of the initial construction model is adjusted iteratively based on the forward simulation results as follows: The simulation results are quantitatively and qualitatively compared with the actual work area observations. During the iteration process, the parameters are adjusted until the forward simulation results of the initial construction model and the selected actual work area meet the quantitative and qualitative indicators.

5. The method for generating tomographic samples according to claim 3 or 4, characterized in that, When this method is used in extensional and strike-slip construction zones, the boundary conditions include: Displacement boundary: Slow extrusion is applied to both sides of the calibrated construction model at a specific initial rate; Fixed Boundaries: The bottom of the calibration construction model is completely fixed, and the front and rear boundaries of the model restrict its normal displacement, allowing sliding along the strike to simulate the constraints of a real basin.

6. The method for generating fault samples according to claim 5, characterized in that, The quantitative indicators are that the errors in fault dip angle and fault density are both <5%; The qualitative indicators are that the main fault, the secondary faults conjugate with it, and the rolling folds formed in the hanging wall due to the slip of the main fault are successfully simulated, and the spatial combination pattern matches the seismic profile with more than 85%.

7. The method for generating tomographic samples according to claim 3 or 4, characterized in that, When this method is used to construct a work zone by extrusion, the boundary conditions include: Displacement boundary: A horizontal extrusion is applied to one side of the calibrated construction model at a specific initial constant rate; Bottom boundary: The bottom of the model is a slip surface, set as a free slip boundary, which allows for large-scale horizontal shortening.

8. The method for generating fault samples according to claim 7, characterized in that, The quantitative indicator is: The fault spacing error is within ±10%; the fault propagation angle error is <6%. The qualitative indicators are that the imbricate thrust fault system, the dual structure at its lower part, and the triangular zone structure formed by the strong compression and folding at the fault front are successfully simulated, with an overall morphological similarity of more than 80%.

9. A fault sample generation system based on seismic data from oil and gas fields, characterized in that, include: The acquisition module acquires seismic exploration data, well data, and actual geological data for the actual work area; The construction module establishes an initial structural model based on seismic exploration data, well data, and actual geological data. The iteration module is used to iteratively adjust the initial construction model based on the forward simulation results to obtain the calibration construction model. The simulation module performs process simulation based on the calibration construction model and outputs multiple geological bodies according to time steps; Data augmentation and sample expansion module: Converts the geological body into a forward modeling parameter model, performs seismic forward modeling to generate seismic data volume, and performs data augmentation and sample expansion.

10. An electronic device, characterized in that, include: A processor and a memory, the processor being configured to execute a program stored in the memory for a method of generating fault samples based on seismic data from oil and gas fields, to implement the method of generating fault samples based on seismic data from oil and gas fields as described in any one of claims 1 to 8.

Images