Reservoir prediction method and device based on deposition and earthquake dual forward modeling combination

Abstract

The invention provides a reservoir prediction method and device based on deposition and earthquake dual forward modeling combination, and the method comprises the steps: carrying out the deposition numerical simulation of a research region, and obtaining the deposition numerical simulation result of each grid in the research region; constructing a three-dimensional geologic model based on a deposition numerical simulation result, and performing three-dimensional seismic forward modeling based on the three-dimensional geologic model to obtain seismic forward modeling data of each grid; dividing the research area into a plurality of subareas based on the deposition numerical simulation result and the seismic forward modeling data of each grid; determining the optimal attribute of each partition based on the seismic forward modeling data in each partition, and determining the optimal attribute data of each partition from the seismic data of each partition based on the optimal attribute of each partition; and for each partition, carrying out fusion processing on the preferred attribute data in the partition to obtain fusion attribute data of the partition, and carrying out reservoir prediction based on the fusion attribute data of each partition. According to the invention, the reliability and precision of reservoir prediction can be improved.

Description

Technical Field

[0001] The invention relates to the technical field of reservoir prediction, and in particular to a reservoir prediction method and device based on the combination of sedimentation and seismic dual forward modeling. Background Art

[0002] Based on the existing geological data and the understanding of sedimentological principles and sedimentary processes, the key link in the field of oil and gas exploration is to predict the relevant characteristics of underground reservoirs that have not yet been fully explored by using a variety of technical means (such as geostatistical methods, seismic interpretation technology, numerical simulation technology, etc.). Seismic attributes refer to various parameters related to the geometry, kinematics and dynamic characteristics of seismic waves from seismic information. Among them, amplitude and frequency attributes can reflect the changes in lithology and oil content of the formation and are widely used in reservoir prediction.

[0003] However, the existing reservoir prediction methods directly optimize attributes on the scale of the entire study area when performing attribute priority, resulting in the accuracy of the prediction results needing to be improved. Summary of the invention

[0004] The embodiment of the present invention provides a reservoir prediction method and device based on the combination of sedimentation and seismic dual forward modeling, which improves the accuracy of reservoir prediction by partitioning the study area and fusing its attributes.

[0005] In a first aspect, an embodiment of the present invention provides a reservoir prediction method based on a combination of sedimentation and seismic dual forward modeling, comprising: performing sedimentation numerical simulation on a study area to obtain sedimentation numerical simulation results for each grid in the study area; constructing a three-dimensional geological model based on the sedimentation numerical simulation results, performing three-dimensional seismic forward modeling based on the three-dimensional geological model, and obtaining seismic forward modeling data for each grid; dividing the study area into a plurality of partitions based on the sedimentation numerical simulation results and the seismic forward modeling data for each grid; determining preferred attributes for each partition based on the seismic forward modeling data within each partition, and determining preferred attribute data for each partition from the seismic data of each partition based on the preferred attributes of each partition; for each partition, fusing the preferred attribute data within the partition to obtain fused attribute data for the partition, and performing reservoir prediction based on the fused attribute data for each partition.

[0006] In some embodiments, the diffusion equation in the deposition numerical simulation incorporates hydrodynamic constraints.

[0007] In some embodiments, the study area is divided into a plurality of partitions based on the sedimentation numerical simulation results and seismic forward modeling data of each grid, including: clustering the grids according to the sedimentation numerical simulation results and seismic forward modeling data of each grid according to a density-based clustering algorithm to obtain clustering results; and determining the plurality of partitions based on the clustering results.

[0008] As a possible implementation method, the preferred attributes of each partition are determined based on the seismic forward modeling data in each partition, including: for each partition, based on the seismic forward modeling data in the partition, performing a sensitivity analysis on multiple seismic attributes in the partition, and based on the sensitivity analysis results, determining the preferred attributes of the partition from the multiple seismic attributes.

[0009] In a second aspect, an embodiment of the present invention provides a reservoir prediction device based on a combination of sedimentation and seismic dual forward modeling, characterized in that it includes: a numerical simulation module, which is used to perform sedimentation numerical simulation on a study area and obtain sedimentation numerical simulation results of each grid in the study area; a seismic forward modeling module, which is used to construct a three-dimensional geological model based on the sedimentation numerical simulation results, and perform three-dimensional seismic forward modeling based on the three-dimensional geological model to obtain seismic forward modeling data of each grid; a partitioning module, which is used to divide the study area into multiple partitions based on the sedimentation numerical simulation results and seismic forward modeling data of each grid; a determination module, which is used to determine the preferred attributes of each partition based on the seismic forward modeling data in each partition, and determine the preferred attribute data of each partition from the seismic data of each partition based on the preferred attributes of each partition; a fusion module, which is used to fuse the preferred attribute data in each partition for each partition, obtain the fused attribute data of the partition, and perform reservoir prediction based on the fused attribute data of each partition.

[0010] In some embodiments, the diffusion equation in the deposition numerical simulation incorporates hydrodynamic constraints.

[0011] In some embodiments, the partitioning module is specifically used to: cluster the grids according to the sedimentation numerical simulation results and seismic forward modeling data of each grid according to a density-based clustering algorithm to obtain clustering results; and determine the multiple partitions based on the clustering results.

[0012] In a third aspect, an embodiment of the present invention provides an electronic device, comprising a processor and a memory storing a computer program, wherein when the processor executes the program, the reservoir prediction method and device based on the combination of sedimentation and seismic dual forward modeling described in the first aspect are implemented.

[0013] In a fourth aspect, an embodiment of the present invention provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the reservoir prediction method and apparatus based on the combination of sedimentation and seismic dual forward modeling as described in the first aspect.

[0014] In a fifth aspect, an embodiment of the present invention provides a computer program product, including a computer program, which, when executed by a processor, implements the reservoir prediction method and apparatus based on the combination of sedimentation and seismic dual forward modeling as described in the first aspect.

[0015] According to the reservoir prediction method and device based on the combination of sedimentation and seismic dual forward modeling of the embodiment of the present invention, the sedimentation numerical simulation is performed on the study area to obtain the sedimentation numerical simulation results of each grid in the study area; based on the sedimentation numerical simulation results, a three-dimensional geological model is constructed, and based on the three-dimensional geological model, a three-dimensional seismic forward modeling is performed to obtain the seismic forward modeling data of each grid; based on the sedimentation numerical simulation results and seismic forward modeling data of each grid, the study area is divided into multiple partitions; based on the seismic forward modeling data in each partition, the preferred attributes of each partition are determined, and based on the preferred attributes of each partition, the preferred attribute data of each partition is determined from the seismic data of each partition; for each partition, the preferred attribute data in the partition is fused to obtain the fused attribute data of the partition, and the reservoir prediction is performed based on the fused attribute data of each partition. On the one hand, the present invention constrains the three-dimensional address forward modeling based on the sedimentation numerical simulation results, which can improve the rationality of the forward modeling results. On the other hand, the present invention can improve the accuracy of the attribute priority process and the accuracy of the fused attribute data by partitioning the study area and performing attribute optimization and attribute fusion for each partition, thereby improving the reliability and accuracy of reservoir prediction. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0017] Figure 1 A schematic flow chart of a reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling provided in an embodiment of the present invention; Figure 2 A schematic diagram of the structure of a reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling provided in an embodiment of the present invention; Figure 3 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] It should be noted that based on the existing geological data and the understanding of sedimentological principles and sedimentary processes, the use of various technical means (such as geostatistical methods, seismic interpretation technology, numerical simulation technology, etc.) to predict the relevant characteristics of underground reservoirs that have not yet been fully explored is a key link in the field of oil and gas exploration. Seismic attributes refer to various parameters related to the geometry, kinematics and dynamic characteristics of seismic waves from seismic information. Among them, amplitude and frequency attributes can reflect the changes in lithology and oil content of the formation and are widely used in reservoir prediction.

[0020] In terms of reservoir prediction technology, some methods have been proposed, such as reservoir identification and reservoir prediction through seismic inversion, and reservoir identification and reservoir prediction based on seismic multi-attribute fusion methods. However, most of them rely on geostatistical interpolation algorithms, or simply rely on seismic attribute analysis to predict reservoirs. When there are not enough control points, the accuracy of geostatistical methods in predicting reservoir boundaries and overall morphology will be greatly reduced; and relying solely on seismic attribute analysis can often only obtain relatively superficial geological information, and it is difficult to go deep into the sedimentary process and fine lithology changes inside the reservoir.

[0021] In the prior art, CN202310012792.3 introduces an experimental research method based on the integration of physical simulation and sedimentation numerical simulation of the source-sink system. By collecting geological data of the study area, preparing physical simulation experimental materials, and designing experimental bottom shapes, multiple physical sedimentation simulation experiments are carried out, and the data are recorded after adjusting to the geological conditions of the study area; then a grid model is constructed and numerical simulation parameters are set to carry out numerical sedimentation simulation experiments. After meeting the limited conditions, a single river channel physical sedimentation simulation experiment is carried out for a specific area and compared with the numerical simulation results. If the comparison results are consistent, the data is retained and the sand body is analyzed. Then, the hydrodynamic characteristics are analyzed to predict the sand body distribution through numerical simulation to find favorable areas. Finally, a favorable position is anchored to carry out smaller-scale physical simulations and integrate multiple sets of experimental results to complete the study of the sand body in the study area. However, this method is difficult to adapt to the special needs of different types of reservoirs. The integrated experimental research method includes multiple links of physical simulation and numerical simulation. Different reservoirs have great differences in sedimentary environment, physical properties and fluid characteristics. Therefore, it is necessary to adjust the physical model and numerical simulation model to varying degrees. In addition, physical simulation requires the establishment of a physical sand filling model. Sand filling models of different types and scales are not universal. Secondly, this method involves a variety of experimental equipment, a large amount of material preparation and repeated experiments. Whether it is the purchase cost of experimental materials or the use and maintenance costs of equipment, they are relatively high.

[0022] Prior art CN202310623431.2 introduces a thin reservoir prediction method based on frequency division phase control. This method uses geophysical methods as the basic research means, and establishes a variable γ prestack deterministic inversion based on a big data waveform-driven low-frequency model for thin reservoirs, and performs phase-controlled prestack geostatistical inversion combination to characterize the sand body. However, this method mainly relies on the analysis of seismic data, especially low-frequency models and variable γ prestack deterministic inversion, which has certain advantages for the identification of thin reservoirs, but in complex sedimentary environments or multi-type reservoirs, it may not be able to fully reflect the complexity and dynamic changes of the reservoir body. Especially in highly heterogeneous or drastically changing reservoir systems, the applicability of this method is relatively limited. Secondly, this method has a weak reflection of dynamic characteristics such as the evolution process of the reservoir body and changes in the sedimentary environment.

[0023] In order to solve the above problems, an embodiment of the present invention provides a reservoir prediction method and device based on the combination of sedimentation and seismic dual forward modeling.

[0024] Figure 1 A schematic diagram of a reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling provided by an embodiment of the present invention. Figure 1 As shown, the method may include the following steps.

[0025] Step 101, perform numerical simulation of sedimentation in the study area to obtain numerical simulation results of sedimentation for each grid in the study area.

[0026] Sedimentation numerical simulation technology is a technical means to quantitatively simulate and analyze the sedimentation process by using computer technology and mathematical models. Sedimentation numerical simulation technology is based on the principles of sedimentology, and comprehensively considers the supply of material sources (such as the type, quantity, and supply rate of sediments), transportation media (such as the velocity, direction, and intensity of water flow and wind), sedimentary environment (topography, hydrodynamic conditions, water depth, and other factors in different environments such as oceans, lakes, and rivers), and tectonic movement, etc. Many factors that affect sedimentation, by constructing corresponding mathematical and physical equations to describe the complete dynamic process of sediments from erosion, transportation to final deposition, and then using the powerful computing power of computers to solve these equations, and then present the distribution of sediments in different geological periods and different regions, thickness changes, and the distribution of sedimentary phases in a visual and quantitative form.

[0027] Numerical simulation of sedimentation can quantitatively describe the sedimentation mechanism and process in nature, and can also record the spatiotemporal evolution of sedimentary strata and the spatiotemporal distribution of reservoirs, providing a new methodological approach for revealing the sedimentation mechanism, quantitatively describing the distribution of underground reservoirs, and clarifying their spatial distribution range.

[0028] Generally, the numerical simulation methods of sedimentation can be divided into five categories according to theoretical methods: diffusion equation model, geometric rule model, fuzzy logic model, cellular automaton model and hydrodynamic equation model. The numerical simulation results obtained based on the diffusion equation model are non-unique and need to be compared with the work area; the model based on geometric rules lacks the constraints of mechanical equations, and the accuracy of its results needs to be further verified; the data required for the cellular automaton model is very fine and the scope of application is limited; the fuzzy logic model is difficult to use and requires strong operating experience. Although the hydrodynamic equation model has a small simulation speed and time and space scale, it can more accurately describe the sedimentation process. It is necessary to select a suitable model for numerical simulation according to the specific situation.

[0029] In some embodiments, the study area may be first gridded, and based on the paleo-geomorphological information of the study area, the study area may be numerically simulated by using an existing sedimentary numerical simulation method to obtain sedimentary numerical simulation results for each grid in the current study area. The sedimentary numerical simulation results may include sedimentary facies, and may also include physical parameters such as sedimentary thickness, porosity, and density.

[0030] In order to describe the sedimentation process more accurately, the embodiment of the present invention proposes a new type of sedimentation numerical simulation scheme, which is a hydrodynamic constrained sedimentation numerical simulation method based on the diffusion equation. It utilizes the advantages of the two models to more realistically reflect the coordinated evolution process of sedimentation and hydrodynamics, and provides an effective solution for the study of sedimentation processes in complex oil and gas reservoirs.

[0031] Specifically, the hydrodynamic equation is used to provide the water flow velocity field and generate constraints on the diffusion equation during the calculation process. The hydrodynamic equation is shown in equation (1). The influence of the hydrodynamic equation is introduced into the diffusion equation in the form of constraints, and the diffusion equation with hydrodynamic constraints is obtained, as shown in equation (2).

[0032] (1); (2); in, , is the total water depth, is the sediment thickness, is the water surface height; is the water pressure gradient, which is related to the change in fluid density; is the acceleration due to gravity; is the sediment concentration (or sediment thickness); is the water flow velocity field; is the diffusion coefficient, which is related to sediment properties and water flow conditions; It is the sediment supply item; It is the transportation and erosion loss term, which is related to the transportation capacity and flow rate of the water flow and can be calculated based on empirical formulas.

[0033] As a possible implementation method, the model implementation process based on the introduction of the hydrodynamic constrained diffusion equation can include the following steps: Step S1, obtaining hydrodynamic constraint variables.

[0034] Calculate the velocity field of the water flow in the study area using hydrodynamic equations (such as the shallow water equations or the Navier-Stokes equations) . Update the velocity field in combination with the topography (sediment distribution affects water depth); interpolate or discretize the velocity field and provide it to the diffusion equation.

[0035] According to water flow speed , calculate the hydrodynamics' sediment transport capacity and erosion effect, as shown in the following formula (3).

[0036] (3); in, is the handling efficiency coefficient; is the critical flow velocity (calculated by Shields parameter or empirical formula).

[0037] Step S2, calculation of diffusion equation.

[0038] At each time step, the change in sediment concentration or sediment thickness is calculated based on the diffusion equation with hydrodynamic constraints: Discretize; add hydrodynamic convection terms The impact of supply and transportation loss Update the sediment distribution. Solve the diffusion equation using numerical methods (such as finite difference or finite volume method) to obtain the time evolution of sediment distribution.

[0039] Step S3: two-way feedback and iteration.

[0040] (1) Feedback of sediment on hydrodynamics: Using sediment distribution results Update total water depth ; Water depth changes affect the water velocity field , and feedback is given to the hydrodynamic equation; (2) Iterative calculation: In each time step, the diffusion equation and the hydrodynamic equation are calculated alternately until convergence; the updated hydrodynamic field is used for the diffusion calculation of the next time step.

[0041] In some embodiments, the numerical calculation process of the deposition numerical simulation based on the above model may include the following steps: Step S4: grid division and discretization.

[0042] The study area is gridded based on a two-dimensional regular grid or a three-dimensional grid. The grid size can be determined based on the actual scale of the study area. Spatial discretization: finite difference or finite volume method is used for diffusion and convection terms respectively; Time discretization: explicit method is used for fast calculation or implicit method is used to improve stability.

[0043] Need to ensure the time step When the following conditions are met, the time step is adjusted adaptively, and the calculation accuracy and efficiency are adjusted according to the dynamic changes of the simulation results.

[0044] (4); in, is the grid size.

[0045] Step S5, implementation of the deposition numerical simulation algorithm.

[0046] Setting the initial sediment distribution and the initial velocity field ; In each time step, solve the hydrodynamic equation to obtain the velocity field and water depth ; Use the updated velocity field to solve the diffusion equation and calculate the deposition thickness ; Update the effects of sediment (such as feedback from topographic changes on flow velocity) and iterate until convergence within the current time step.

[0047] S6, visual display.

[0048] Use heat maps or contour maps to dynamically display the diffusion and accumulation of sediments; generate snapshots after each time step to generate dynamic videos or animations of sediment evolution. Use vector field maps to display velocity field distribution (arrows indicate flow direction and magnitude); superimpose on sediment distribution maps to intuitively display the interaction between water flow and sediments, and obtain the sedimentation numerical simulation results for each grid.

[0049] Step 102: construct a three-dimensional geological model based on the sedimentary numerical simulation results, perform three-dimensional seismic forward modeling based on the three-dimensional geological model, and obtain seismic forward modeling data for each grid.

[0050] Three-dimensional seismic forward modeling is a technology that uses wave theory (usually based on elastic wave or acoustic wave equations) based on known underground geological structure models (including information such as lithology, thickness, burial depth and structural morphology of the strata) and corresponding rock physical parameters (such as velocity, density, etc.) to simulate the entire physical process of seismic wave propagation, reflection, refraction and scattering in underground media through computer numerical calculation methods, thereby obtaining theoretical seismic records (that is, synthetic seismic data) under a set observation system.

[0051] Before multi-attribute fusion, the forward model can help build an idealized geological model and clarify the relationship between seismic attributes (such as amplitude, frequency, phase, etc.) and geological body characteristics (such as lithology, porosity, fractures, etc.). Through forward modeling, the response of these attributes under different geological models can be obtained, thereby providing more accurate attribute information and geological background for multi-attribute fusion and improving the accuracy of subsequent analysis.

[0052] Traditional three-dimensional seismic forward modeling usually relies on simplified geological assumptions, while the embodiments of the present invention accurately reproduce the complexity and dynamic changes of the sedimentary environment through sedimentary numerical simulation, which can better simulate the distribution and morphology of the reservoir and reflect the spatiotemporal characteristics of the geological process. This method reduces the reliance on assumptions in traditional forward modeling, builds a geological model based on sedimentological principles and actual geological data, and improves the geological rationality of the forward modeling results. By introducing sedimentary model guidance, the present invention can optimize the correlation between seismic attributes and sedimentary characteristics, provide more accurate reservoir responses, and thus improve the accuracy of multi-attribute fusion. In addition, the method has shown stronger capabilities in the description of complex oil and gas reservoirs (such as carbonate rocks, heterogeneous sandstones, etc.), can more accurately reveal the geological characteristics of oil and gas reservoirs, and provide more reliable seismic data support for oil and gas exploration and development.

[0053] In some embodiments, a three-dimensional geological model constructed based on numerical simulation of sedimentation may include: stratigraphic structure, such as stratifying different reservoir bodies in the sedimentation model and simulating the geological connections between layers; it may also include assignment of physical parameters, such as mapping physical properties such as lithology, porosity, density, and elastic modulus to the three-dimensional geological model to ensure that the geological body represented by each cell in the model has reasonable physical properties.

[0054] In order to conduct seismic forward modeling, after establishing the sedimentary numerical model and the three-dimensional geological model, it is necessary to define a seismic wave propagation model that interacts with the reservoir. The key parameters of the seismic model propagation model include: wave velocity (P-wave and S-wave velocity), which can calculate the wave velocity of different reservoirs based on the physical properties of the rock (such as density, elastic modulus, etc.); attenuation and scattering, which considers the attenuation effect, frequency dependence, scattering effect, etc. of seismic waves when propagating in the reservoir, especially in heterogeneous or complex reservoir areas; reflection coefficient, which calculates the reflection coefficient through the changes in physical properties in the geological model (such as lithology, porosity, etc.) to simulate the reflection behavior of seismic waves encountering different reservoirs.

[0055] In some embodiments, the specific process of performing three-dimensional seismic forward modeling based on the three-dimensional geological model and obtaining seismic forward modeling data for each grid may include the following steps.

[0056] Step S7, defining earthquake source definition.

[0057] Define the type, location, and excitation method of the earthquake source (such as source location, frequency, amplitude, etc.). You can choose ground excitation or underground source.

[0058] Step S8, setting the source and receiving point.

[0059] Set the location of the source and receiving points according to the actual seismic exploration plan. Usually the receiving points can be placed on the surface or underground of the model to simulate different seismic data according to the needs.

[0060] Step S9, simulating the propagation process.

[0061] The propagation process of seismic waves in a three-dimensional geological model is calculated through numerical simulation methods such as finite difference and finite element, taking into account the effects of wave reflection, refraction, attenuation, and multiple reflections.

[0062] Step S10, recording seismic forward modeling data.

[0063] The simulated seismic forward modeling data (such as amplitude, frequency, phase, velocity, etc.) are recorded and output in the form of seismic measurement data, such as reflection wave data (CDP data).

[0064] Step 103: Divide the study area into multiple subareas based on the sedimentation numerical simulation results and seismic forward modeling data of each grid.

[0065] For complex sedimentary reservoirs and large-scale sedimentary reservoirs, the results of linear fusion of seismic attributes are difficult to depict the spatial changes of reservoir sand bodies and cannot meet the requirements of fine reservoir characterization. In order to solve the above problems, the present invention divides the study area into zones so that the grids in each zone have similar geological characteristics. When performing attribute priority and attribute fusion in subsequent steps, each zone can be performed separately, thereby improving the accuracy of reservoir prediction.

[0066] In some embodiments, the similarity of adjacent grids can be determined based on the sedimentation numerical simulation results and seismic forward modeling results of each grid; multiple adjacent grids with similarities greater than or equal to a threshold are divided into a partition to obtain multiple partitions after division. The similarity can be calculated based on the sedimentation numerical simulation results and seismic forward modeling results of each adjacent grid, the simulation result difference of the adjacent grids is calculated, and the reciprocal of the difference is determined as the similarity of the adjacent grids.

[0067] In some embodiments, each grid may be partitioned by clustering. As an example, each grid may be clustered according to the sedimentation numerical simulation results and seismic forward modeling data of each grid using a density-based clustering algorithm to obtain a clustering result; and multiple partitions may be determined based on the clustering result.

[0068] Among them, in the clustering process, according to the sedimentation numerical simulation results of each grid, the simulated changes in physical parameters such as stratum velocity and density can be obtained, and the gradient changes and extreme value characteristics of the above physical parameters can also be obtained. These physical parameters are closely related to the type and boundary of the sedimentary area. In the clustering process, according to the seismic forward modeling data of each grid, the strength distribution of the simulated seismic wave reflection amplitude, the continuity of the reflection phase axis, the phase change, etc., as well as the propagation time characteristics of the seismic wave can also be obtained. Different sedimentary areas will have different effects on the propagation time of seismic waves, which can be used as an important basis for zoning. At the same time, the frequency-related characteristics in seismic attributes will also be different in different sedimentary areas.

[0069] The density-based clustering algorithm DBSCAN (Density-Based Spatial Clustering of Applications with Noise) can better discover the clustering of non-spherical distribution and irregular sedimentary areas, especially for situations with complex geological conditions and irregular boundaries of sedimentary areas. Its core idea is: if the density of data points in an area exceeds a certain threshold (that is, a set of densely connected data points is formed), then these points belong to the same cluster; and those data points in low-density areas are regarded as noise points.

[0070] As an example, according to the density-based clustering algorithm, the process of clustering each grid according to the sedimentation numerical simulation results and seismic forward modeling data of each grid includes: extracting features according to the sedimentation numerical simulation results and seismic forward modeling data of each grid, removing highly correlated and redundant features through statistical means such as correlation analysis, and retaining the most representative key features that have a significant indicative effect on the sedimentation area division; standardizing these screened features to ensure that features of different magnitudes and properties can be treated equally when participating in subsequent clustering calculations and are uniformly incorporated into the analysis framework of the DBSCAN algorithm; when using the DBSCAN algorithm, the determination of the two key parameters of neighborhood radius and minimum number of points is crucial. For the specific application scenario of zoning division, it is necessary to combine the prior knowledge in the geological field and the preliminary understanding of the current study area to make preliminary settings of the parameters. For example, if the approximate scale of the sedimentation area in the study area and the density of the data point distribution are known, the appropriate range of neighborhood radius and minimum number of points can be estimated accordingly. Subsequently, through multiple tests, using a part of the verification samples of the known sedimentary area division (these samples can come from previous geological exploration results or data verified by high-precision field measurements), the degree of consistency between the partitions divided by the DBSCAN algorithm under different parameter values ​​and the actual situation is compared, so as to continuously optimize the parameter values ​​so that the algorithm can accurately identify the boundaries of each partition and the distribution of data points inside. Finally, each cluster area obtained represents a partition after division.

[0071] Step 104, based on the seismic forward modeling data in each partition, determine the preferred attribute of each partition, and based on the preferred attribute of each partition, determine the preferred attribute data of each partition from the seismic data of each partition.

[0072] Different sediment thicknesses have different sensitivities to different seismic attributes. Generally speaking, for seismic attributes such as amplitude, changes in the thickness of thinner sediment layers may cause more obvious changes, because the absorption and reflection of seismic wave energy by thin sediment layers are relatively more susceptible to slight changes in their own thickness. A slight increase or decrease in the thickness of thin sediment layers may cause noticeable differences in the amplitude of the reflected seismic waves. For frequency-related seismic attributes, the sensitivity will be more prominent when the thickness of relatively thicker sediments changes. Thicker sediment layers often mean that the paths and processes of seismic wave propagation in them are more complicated. Once the thickness changes, the filtering and transformation of the frequency components of seismic waves will change accordingly, which will lead to more significant changes in frequency-related attributes.

[0073] In order to accurately obtain the preferred attributes, the embodiment of the present invention performs attribute optimization for each partition respectively to obtain the preferred attributes corresponding to each partition, so that the preferred attributes of each partition can more accurately characterize the characteristics of the partition, thereby improving the accuracy of reservoir prediction.

[0074] In some embodiments, the process of determining the preferred attributes of each partition based on the seismic forward modeling data in each partition may include: for each partition, based on the seismic forward modeling data in the partition, determining multiple seismic attribute data (for example, seismic attribute data such as phase and reflection coefficient) corresponding to each grid in the partition; for each seismic attribute, determining the degree of change of the seismic attribute data corresponding to the seismic attribute in the partition; and determining multiple seismic attributes whose degree of change is greater than or equal to a threshold as preferred attributes in the partition.

[0075] In other embodiments, the process of determining the preferred attributes of each partition based on the seismic forward modeling data within each partition may include: for each partition, based on the seismic forward modeling data within the partition, performing a sensitivity analysis on multiple seismic attributes within the partition, and based on the sensitivity analysis results, determining the preferred attributes of the partition from the multiple seismic attributes.

[0076] As an example, for each partition, multiple seismic attribute data of each grid can be determined based on the seismic forward modeling data in the partition; since the physical properties (such as porosity, lithology, density, etc.) between each grid are different, a sensitivity analysis can be performed between each seismic attribute and the physical property, that is, the sensitivity of each seismic attribute to the change of the physical property. For seismic attributes with greater sensitivity, it has greater distinguishing power, so seismic attributes with greater sensitivity can be determined as preferred attributes. For example, for each partition, the correlation between each seismic attribute and each physical property in the partition can be calculated, and multiple seismic attributes with correlation greater than or equal to the threshold are determined as the preferred attributes of the partition.

[0077] Among them, based on the preferred attributes of each partition, the preferred attribute data of each partition is determined from the seismic data of each partition, which means that the attribute values ​​corresponding to the preferred attributes of each partition are determined from the real seismic data of the study area, wherein the preferred attribute data are the attribute values ​​corresponding to each preferred attribute.

[0078] Step 105: for each partition, the preferred attribute data in the partition are fused to obtain fused attribute data of the partition, and reservoir prediction is performed based on the fused attribute data of each partition.

[0079] In some embodiments, the process of fusing the preferred attribute data in a partition may be to fusing the attribute values ​​corresponding to the different priority attributes at different positions in the partition to obtain the fused attribute data corresponding to the different positions in the partition. The fusion process may be a simple superposition process, a multivariate linear regression process, or a linear fusion algorithm such as RGB frequency division fusion.

[0080] In other embodiments, the preferred attribute data in the partitions may be fused based on a deep feedforward neural network. For example, the preferred attribute data in each partition is preprocessed, and the attribute values ​​of each preferred attribute are normalized; for each partition, the processed preferred attribute data is input into the deep feedforward neural network to obtain the fused attribute data of the partition.

[0081] Among them, the deep feedforward neural network can extract features for the attribute values ​​of each priority attribute respectively, and then fuse the features of multiple priority attributes to obtain fused attribute features, thereby realizing nonlinear fusion of attributes. The deep feedforward neural network can be trained based on the preferred attribute data samples and the sedimentary facies layer labels determined based on the logging data. The fused attribute data output by the model here can be the sedimentary facies layers at different positions in each partition.

[0082] In some embodiments, reservoir prediction is performed based on the fused attribute data of each partition, which is equivalent to comprehensively describing the three-dimensional spatial morphology of reservoir distribution in the study area based on the fused attribute data of each partition.

[0083] In some embodiments, vertically, the configuration mode (such as river channel, wedge, etc.) is matched on the seismic section; horizontally, the boundary of the reservoir is determined by integrating the distribution law of attribute data in the horizontal slice, and its distribution in different sedimentary layers and different partitions is depicted to construct a three-dimensional spatial model of the reservoir, clearly showing the spatial relationship between the reservoir and the surrounding reservoirs, geological structures, etc. Considering the characteristics of different types of reservoirs (such as single pore type, pore-fracture composite type, etc.), the appropriate modeling algorithm is used to accurately construct its complex spatial morphology, so that the depicted reservoir model is as close to the actual geological situation as possible.

[0084] As an example, the configuration patterns (such as river channels, sand bars, etc.) are matched on the seismic profile, and the lateral distribution boundaries of the reservoir are accurately identified by combining the fused attribute data; a three-dimensional geological model of the reservoir is constructed based on the fusion results of the preferred attributes and sedimentary zoning; the distribution morphology of the reservoir in different sedimentary layers and its relationship with the geological structure are depicted; the reservoir model is verified using drilling and production data, and reservoir parameters (such as porosity and permeability) are dynamically adjusted; the three-dimensional spatial morphology of the reservoir and its physical properties are dynamically displayed, and profiles, isosurfaces and lateral distribution maps are generated to intuitively present the prediction results.

[0085] According to the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling of the embodiment of the present invention, the study area is subjected to sedimentation numerical simulation to obtain the sedimentation numerical simulation results of each grid in the study area; based on the sedimentation numerical simulation results, a three-dimensional geological model is constructed, and based on the three-dimensional geological model, a three-dimensional seismic forward modeling is performed to obtain the seismic forward modeling data of each grid; based on the sedimentation numerical simulation results and the seismic forward modeling data of each grid, the study area is divided into multiple partitions; based on the seismic forward modeling data in each partition, the preferred attributes of each partition are determined, and based on the preferred attributes of each partition, the preferred attribute data of each partition is determined from the seismic data of each partition; for each partition, the preferred attribute data in the partition is fused to obtain the fused attribute data of the partition, and the reservoir prediction is performed based on the fused attribute data of each partition. On the one hand, the present invention constrains the three-dimensional address forward modeling based on the sedimentation numerical simulation results, which can improve the rationality of the forward modeling results. On the other hand, the present invention can improve the accuracy of the attribute priority process and the accuracy of the fused attribute data by partitioning the study area and performing attribute optimization and attribute fusion for each partition, thereby improving the reliability and accuracy of reservoir prediction.

[0086] In order to implement the above embodiment, the present invention also provides a reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling.

[0087] Figure 2 The schematic diagram of the structure of the reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling provided by the embodiment of the present invention. Figure 2 As shown, the device may include: a numerical simulation module 210 , a seismic forward modeling module 220 , a partitioning module 230 , a determination module 240 and a fusion module 250 .

[0088] The numerical simulation module 210 is used to perform numerical simulation of sedimentation in the study area and obtain numerical simulation results of sedimentation for each grid in the study area; the seismic forward modeling module 220 is used to construct a three-dimensional geological model based on the numerical simulation results of sedimentation, perform three-dimensional seismic forward modeling based on the three-dimensional geological model, and obtain seismic forward modeling data for each grid; the partitioning module 230 is used to divide the study area into multiple partitions based on the numerical simulation results of sedimentation and the seismic forward modeling data of each grid; the determination module 240 is used to determine the preferred attributes of each partition based on the seismic forward modeling data in each partition, and determine the preferred attribute data of each partition from the seismic data of each partition based on the preferred attributes of each partition; the fusion module 250 is used to fuse the preferred attribute data in each partition to obtain the fused attribute data of the partition, and perform reservoir prediction based on the fused attribute data of each partition.

[0089] In some embodiments, hydrodynamic constraints are incorporated into the diffusion equations in the numerical simulation of deposition.

[0090] In some embodiments, the partitioning module 230 is specifically used to: cluster each grid according to the sedimentation numerical simulation results and seismic forward modeling data of each grid according to a density-based clustering algorithm to obtain a clustering result; and determine multiple partitions based on the clustering result.

[0091] In some embodiments, the determination module 240 is specifically used to: for each partition, based on the seismic forward modeling data in the partition, perform sensitivity analysis on multiple seismic attributes in the partition, and based on the sensitivity analysis results, determine the preferred attribute of the partition from the multiple seismic attributes.

[0092] According to the reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling of the embodiment of the present invention, the numerical simulation module is used to perform sedimentation numerical simulation on the study area to obtain sedimentation numerical simulation results of each grid in the study area; the seismic forward modeling module is used to construct a three-dimensional geological model based on the sedimentation numerical simulation results, and perform three-dimensional seismic forward modeling simulation based on the three-dimensional geological model to obtain seismic forward modeling data of each grid; the partitioning module is used to divide the study area into multiple partitions based on the sedimentation numerical simulation results and the seismic forward modeling data of each grid; the determination module is used to determine the preferred attributes of each partition based on the seismic forward modeling data in each partition, and determine the preferred attribute data of each partition from the seismic data of each partition based on the preferred attributes of each partition; the fusion module is used to fuse the preferred attribute data in each partition to obtain fused attribute data of the partition, and perform reservoir prediction based on the fused attribute data of each partition. On the one hand, the present invention constrains three-dimensional address forward modeling based on the results of sedimentary numerical simulation, which can improve the rationality of the forward modeling results. On the other hand, the present invention divides the study area into zones and performs attribute optimization and attribute fusion for each zone, thereby improving the accuracy of the attribute priority process and the accuracy of the fused attribute data, thereby improving the reliability and accuracy of reservoir prediction.

[0093] It should be noted that the above explanations and descriptions in the embodiment of the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling are also applicable to the reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling in the embodiment of the present invention, and will not be repeated here.

[0094] Figure 3 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 3 As shown, the electronic device may include: a processor 310, a communication interface 320, a memory 330 and a communication bus 340, wherein the processor 310, the communication interface 320 and the memory 330 communicate with each other through the communication bus 340. The processor 310 may call the computer program in the memory 330 to implement the steps of the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling provided in the above embodiments.

[0095] For example, the method includes: performing numerical simulation of sedimentation in the study area to obtain numerical simulation results of sedimentation for each grid in the study area; constructing a three-dimensional geological model based on the numerical simulation results of sedimentation, performing three-dimensional seismic forward modeling based on the three-dimensional geological model, and obtaining seismic forward modeling data for each grid; dividing the study area into multiple partitions based on the numerical simulation results of sedimentation and the seismic forward modeling data of each grid; determining the preferred attributes of each partition based on the seismic forward modeling data within each partition, and determining the preferred attribute data of each partition from the seismic data of each partition based on the preferred attributes of each partition; for each partition, fusing the preferred attribute data within the partition to obtain fused attribute data of the partition, and performing reservoir prediction based on the fused attribute data of each partition.

[0096] In addition, the logic instructions in the above-mentioned memory 330 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0097] On the other hand, an embodiment of the present invention further provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the steps of the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling provided in the above embodiments.

[0098] On the other hand, an embodiment of the present invention further provides a non-transitory computer-readable storage medium, which stores a computer program, and the computer program is used to enable a processor to execute the steps of the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling provided in the above embodiments.

[0099] The non-transitory computer-readable storage medium can be any available medium or data storage device that can be accessed by a computer, including but not limited to magnetic storage (such as floppy disks, hard disks, magnetic tapes, magneto-optical disks (MO), etc.), optical storage (such as CDs, DVDs, BDs, HVDs, etc.), and semiconductor storage (such as ROM, EPROM, EEPROM, non-volatile memory (NAND FLASH), solid-state drives (SSDs)), etc.

[0100] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0101] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0102] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling, characterized in that: include: Performing numerical simulation of sedimentation in the study area to obtain numerical simulation results of sedimentation for each grid in the study area; Based on the sedimentary numerical simulation results, a three-dimensional geological model is constructed, and based on the three-dimensional geological model, a three-dimensional seismic forward modeling simulation is performed to obtain seismic forward modeling data for each of the grids; Based on the sedimentation numerical simulation results and seismic forward modeling data of each grid, the study area is divided into a plurality of subareas; Determine the preferred attribute of each of the partitions based on the seismic forward modeling data in each of the partitions, and determine the preferred attribute data of each partition from the seismic data of each of the partitions based on the preferred attribute of each of the partitions; For each of the partitions, the preferred attribute data in the partition is fused to obtain fused attribute data of the partition, and reservoir prediction is performed based on the fused attribute data of each of the partitions.

2. The method according to claim 1, characterized in that The diffusion equation in the sedimentation numerical simulation incorporates hydrodynamic constraints.

3. The method according to claim 1, characterized in that Based on the sedimentary numerical simulation results and seismic forward modeling data of each grid, the study area is divided into a plurality of subareas, including: According to a density-based clustering algorithm, clustering is performed on each grid according to the sedimentation numerical simulation result and seismic forward modeling data of each grid to obtain a clustering result; Based on the clustering result, the plurality of partitions are determined.

4. The method according to claim 1, characterized in that Determining the preferred attributes of each of the sub-areas based on the seismic forward modeling data in each of the sub-areas comprises: For each of the partitions, a sensitivity analysis is performed on a plurality of seismic attributes in the partition based on the seismic forward modeling data in the partition, and a preferred attribute of the partition is determined from the plurality of seismic attributes based on the sensitivity analysis result.

5. A reservoir prediction device based on the combination of sedimentation and seismic dual forward modeling, characterized in that: include: A numerical simulation module is used to perform a numerical simulation of sedimentation in a study area and obtain a numerical simulation result of sedimentation for each grid in the study area; A seismic forward modeling module, used to construct a three-dimensional geological model based on the sedimentary numerical simulation results, perform three-dimensional seismic forward modeling based on the three-dimensional geological model, and obtain seismic forward modeling data for each of the grids; A partitioning module, used for dividing the study area into a plurality of partitions based on the sedimentation numerical simulation results and seismic forward modeling data of each grid; A determination module, for determining the preferred attribute of each of the partitions based on the seismic forward modeling data in each of the partitions, and determining the preferred attribute data of each partition from the seismic data of each partition based on the preferred attribute of each of the partitions; The fusion module is used to fuse the preferred attribute data in each partition to obtain the fused attribute data of the partition, and perform reservoir prediction based on the fused attribute data of each partition.

6. The device according to claim 5, characterized in that The diffusion equation in the sedimentation numerical simulation incorporates hydrodynamic constraints.

7. The device according to claim 5, characterized in that The partition module is specifically used for: According to a density-based clustering algorithm, clustering is performed on each grid according to the sedimentation numerical simulation result and seismic forward modeling data of each grid to obtain a clustering result; Based on the clustering result, the plurality of partitions are determined.

8. An electronic device comprising a processor and a memory storing a computer program, characterized in that: When the processor executes the computer program, the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling according to any one of claims 1 to 4 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling as claimed in any one of claims 1 to 4 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the reservoir prediction method based on the combination of sedimentation and seismic dual forward modeling according to any one of claims 1 to 4 is implemented.

Images