A three-dimensional phase-controlled sand body modeling and connectivity evaluation method and system

Abstract

This invention establishes a method and system for three-dimensional facies-controlled sand body modeling and connectivity evaluation. The method is based on a dimensionality-reduction modeling architecture: a planar sedimentary facies distribution map is generated through probability-guided sequential indicator simulation; under the constraint of planar facies zones, a one-dimensional vertical sequence of interbedded sand and mud layers is independently generated point-by-point using probabilistic sampling; the planar facies map and the vertical sequence are stitched together in three-dimensional space; then, based on graph theory undirected graph algorithms, the connectivity of the three-dimensional sand body is identified using vertical depth interval overlap and non-interlayer occlusion as dual criteria; finally, connectivity uncertainty is quantified through Monte Carlo iteration. Compared with traditional full-3D grid sequential Gaussian simulation, this method is applicable to clastic reservoirs with gradual lateral variation and dominant vertical accretion (lateral variation coefficient of planar facies zones less than 0.3). It eliminates the computationally expensive three-dimensional spatial interpolation and complex structural modeling, forming a simplified engineering path of "trading computational power for uncertainty assessment," which can be used for reservoir connectivity evaluation and development scheme optimization.

Description

Technical Field

[0001] This invention relates to the field of oil and gas geological reservoir modeling and evaluation technology, and in particular to a method and system for three-dimensional phase-controlled sand body modeling and connectivity evaluation. Background Technology

[0002] In the demonstration of oil and gas reservoir development plans, the three-dimensional connectivity pattern of reservoir sand bodies directly affects the design of injection and production well locations and well network density. Therefore, an accurate three-dimensional sand body model is an unavoidable basic task.

[0003] The following are some existing 3D modeling methods and their shortcomings:

[0004] Existing commercial reservoir modeling software widely employs a full 3D mesh approach, which involves creating a fine 3D mesh within the modeling area and simultaneously simulating lithology and physical properties at each node, constrained by sedimentary facies models. This type of method typically generates millions to tens of millions of mesh nodes, and a single complete modeling operation often takes hours or even tens of hours. During the feasibility study phase, geological engineers often require multiple rounds of parameter adjustments and model updates, making computational cost a major bottleneck to work efficiency. In data-sparse exploration areas, the constraint capability of the full 3D mesh approach on parameter distribution in distant well areas weakens rapidly with increasing distance from the well, making it difficult to fully reflect the geological significance.

[0005] In addition, existing modeling methods also include objective methods and process methods.

[0006] Object-based simulation directly places sand bodies (such as channels and sandbars) with predetermined geometries into the modeling space, resulting in good geological consistency in spatial distribution. However, the computational cost of each object placement operation increases quadratically with the mesh size, especially in fine meshes (≥10 planar nodes). 5 The computational efficiency drops significantly in scenarios involving multi-parameter joint optimization, especially in Monte Carlo loops where frequent model updates are required.

[0007] Process-based simulation, which uses sedimentary dynamics as its modeling basis, can generate sand body morphologies that conform to sedimentary mechanisms. However, its computational cost is much greater than that of statistical methods, and it requires a large number of paleogeomorphological and paleohydraulic parameters that are difficult to obtain directly, which greatly limits its applicability in practical engineering applications.

[0008] Regarding connectivity evaluation, the existing techniques and their shortcomings are as follows: First, simple connectivity assessments based on permeability thresholds only consider lithological matching between adjacent grid cells while ignoring overlap in vertical depth intervals, easily misclassifying actually disconnected sandstone cells as connected. Second, direct connectivity evaluation through full-scale numerical fluid simulations (such as the finite difference simulator Eclipse) offers high computational accuracy but is extremely time-consuming, making it difficult to support the need for extensive repetitive modeling in Monte Carlo uncertainty analysis. Neither of these methods can simultaneously achieve both evaluation accuracy and computational efficiency. Summary of the Invention

[0009] To address the aforementioned problems, the present invention aims to propose a three-dimensional phase-controlled sand body modeling and connectivity evaluation method and system, which employs a dimensionality reduction modeling architecture to solve the computational complexity problem and improve the accuracy of connectivity identification.

[0010] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, this application provides a method for three-dimensional phase-controlled sand body modeling and connectivity evaluation, including: Step (1): Obtain prior geological data for the area to be modeled; Step (2): Based on prior geological data, a planar sedimentary facies distribution map of the area to be modeled is generated by probability-guided sequential indices. Step (3): Based on prior geological data and planar sedimentary facies distribution map, a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map is generated one by one using a probability sampling method; Step (4): Based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point, perform three-dimensional spatial stitching to obtain a three-dimensional grid model; Step (5): Based on the graph theory undirected graph algorithm, identify the connectivity of the three-dimensional sand body mesh of the three-dimensional mesh model; Step (6): Output the uncertainty analysis results of connectivity using the Monte Carlo method.

[0011] In one implementation, the prior geological data in step (1) includes: a prior probability distribution map of sedimentary facies in the modeling area; statistical parameters of sandstone volume proportions and single-layer thickness distribution parameters for each facies type; and drilling and seismic interpretation constraint data.

[0012] In one implementation, in step (2), the simulated value of each grid point in the planar sedimentary facies distribution map is calculated based on a heuristic linear fusion method of conditional probability. The factors included in the calculation are: the prior probability of each grid point and the neighborhood probability corresponding to each grid point.

[0013] In one implementation, the neighborhood probability of each grid point is calculated using inverse distance weighted interpolation.

[0014] In one implementation, in step (3), when there are constraints from drilling and seismic interpretation data at the corresponding location of a grid point, a one-dimensional vertical sequence corresponding to that grid point is generated based on the drilling and seismic interpretation data.

[0015] In one implementation, in step (3), when there is no constraint from drilling and seismic interpretation data at the corresponding location of the grid point, a probability distribution model corresponding to the sedimentary facies type is selected for sampling based on the sedimentary facies type corresponding to the grid point.

[0016] In one implementation, the probability distribution model includes: normal distribution, log-normal distribution, and uniform distribution.

[0017] In one implementation, in step (5), a graph theory undirected graph algorithm calculates whether there are connected edges between two three-dimensional sand body meshes based on spatial proximity conditions, compatibility conditions, vertical overlap conditions, and non-layer occlusion conditions.

[0018] In one implementation, in step (6), based on a set number of independent samples, the mean, standard deviation, and set percentile values ​​of the connectivity rate are calculated according to the sampling results, and the uncertainty analysis results of the connectivity rate are output.

[0019] Secondly, a three-dimensional phase-controlled sand body modeling and connectivity evaluation system is provided, including: The data management module is used to acquire prior geological data for the area to be modeled. The planar facies map simulation module is used to simulate and generate a planar sedimentary facies distribution map of the area to be modeled based on prior geological data and through probability-guided sequential instructions. The vertical profile generation module is used to generate a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map based on prior geological data and planar sedimentary facies distribution map using a probability sampling method. The 3D integration module is used to perform 3D spatial stitching based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point to obtain a 3D mesh model; The graph theory connectivity analysis module is used to identify the connectivity of 3D sand body meshes in 3D mesh models based on graph theory undirected graph algorithms. The Monte Carlo analysis module is used to output uncertainty analysis results of connectivity using the Monte Carlo method.

[0020] By adopting the above technical solutions, this invention uses a dimensionality reduction modeling architecture to solve the problem of computational complexity and improve the accuracy of connectivity recognition. Attached Figure Description

[0021] Figure 1 A flowchart of the method provided in an embodiment of the present invention; Figure 2 A flowchart illustrating the method for providing detailed calculation examples in embodiments of the present invention. Detailed Implementation

[0022] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the described embodiments of the present invention are within the scope of protection of the present invention.

[0023] This invention provides a method for three-dimensional phase-controlled sand body modeling and connectivity evaluation, comprising: Step (1): Obtain prior geological data for the area to be modeled; Step (2): Based on prior geological data, a planar sedimentary facies distribution map of the area to be modeled is generated by probability-guided sequential indices. Step (3): Based on prior geological data and planar sedimentary facies distribution map, a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map is generated one by one using a probability sampling method; Step (4): Based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point, perform three-dimensional spatial stitching to obtain a three-dimensional grid model; Step (5): Based on the graph theory undirected graph algorithm, identify the connectivity of the three-dimensional sand body mesh of the three-dimensional mesh model; Step (6): Output the uncertainty analysis results of connectivity using the Monte Carlo method.

[0024] The following is based on the appendix to this application. Figure 1 The methods of this application are described in one or more more detailed embodiments.

[0025] Detailed Implementation Examples Figure 1 This is a flowchart of a detailed implementation method. The method integrates sequential indicator simulation, one-dimensional vertical probabilistic stratification, three-dimensional volume integration, graph theory connectivity identification, and Monte Carlo uncertainty analysis into a unified process. This method is applicable to clastic reservoirs with gradual lateral variation and dominated by vertical accretion. Typical applicable scenarios include thick lacustrine delta fronts, marine retrograde sandstones, and vertically stacked sand bodies in peat bogs—depositional environments characterized primarily by vertical sand-mud interlocking and with a lateral variation coefficient of less than 0.3 for planar facies zones. For scenarios with strong lateral channel cutting (such as frequent course changes in braided rivers), significant lateral accretion (such as point-bar migration systems in meandering rivers), or strict regional isochronous stratigraphic constraints, this method cannot fully reflect lateral heterogeneity; simulation methods based on full three-dimensional meshes should be prioritized in these cases.

[0026] The methodology includes: Step (1): Obtain prior geological data for the area to be modeled; Specifically, in the data preparation stage, the prior probability distribution map of sedimentary facies in the modeling area, statistical parameters of sandstone volume proportions of each facies type, single-layer thickness distribution parameters, and hard constraint data from drilling and seismic interpretation are obtained.

[0027] A more detailed example is provided, using a 10×15 grid, 500m total thickness lacustrine delta front vertical accretionary reservoir as an example. The sedimentary background of this example is a shallow-water deltaic distributary channel and interdistributary bay alternating system during the lacustrine basin regression stage, with a lateral variation coefficient of approximately 0.22 for the planar facies zones. The flowchart of the example can be found in [reference needed]. Figure 2 .

[0028] More specifically, in the above example, step (1) defines four sedimentary facies types: channel facies (code 1), levee facies (code 2), crevasse-splay facies (code 3), and floodplain facies (code 0). The channel facies has a sandstone volume ratio of 75% with a standard deviation of 5% and a range of 50%–90%. The thickness distribution follows a normal distribution with a mean of 8m and a standard deviation of 2.5m. The vertical distribution pattern is uniform or blocky. The levee facies has a sandstone ratio of 40% with a normal distribution (mean of 5m and standard deviation of 1.5m). The crevasse-splay facies has a sandstone ratio of 55% with a log-normal distribution (mean of 4m and log-standard deviation of 0.3). The floodplain facies has a sandstone ratio of 10% with a uniform distribution (range [0.5m, 2m]). The prior probability map uses a background probability of 0.25, with a high probability of 0.75 set for the expected channel development zone, and a small random perturbation is superimposed to avoid overly regular geometric shapes. This example does not use hard-constrained drilling data to demonstrate the simulation effect of pure prior probability-driven approaches. For boundary validation, the lateral variation coefficient of the planar facies zone is defined as the variation coefficient of the sedimentary facies indicator value (channel facies = 1, non-channel facies = 0) on the planar grid points. ,in For plane standard deviation, This is the mean. In this example, the value is approximately 0.22, which meets the applicable conditions. .

[0029] Step (2): Based on prior geological data, a planar sedimentary facies distribution map of the area to be modeled is generated by probability-guided sequential indices. Specifically, planar facies simulation is the starting point of the method. Using prior probability distributions and hard-constraint data as conditions, a probability-guided sequential indicator simulation method is employed. Monte Carlo sampling is performed by traversing grid points via random paths and calculating conditional probabilities point-by-point, generating a planar sedimentary facies distribution map encompassing various sedimentary facies types, including channel facies, natural levee facies, breach fan facies, and floodplain facies. Conditional probabilities are calculated as a weighted fusion of prior probabilities and information from neighboring simulated grid points, with weighting coefficients set by the user based on data richness. This step forms the spatial framework foundation for all subsequent modeling work.

[0030] Correspondingly, the execution process of step (2) in the detailed example is as follows: Perform a stochastic simulation of planar sedimentary facies. For each grid point to be simulated... The conditional probability is calculated using a heuristic linear fusion formula:

[0031] in Let be the prior probability of that position. For the prior probability weighting coefficients, this example takes... This formula sacrifices the optimality of the variance of the standard indicator Kriging estimate (BLUE property) in exchange for the computational simplicity of eliminating the need to solve the Kriging equations, reducing the simulation time per grid point to the microsecond level. This is a key design trade-off for supporting large-scale Monte Carlo cycles. Neighborhood probability Inverse distance weighted interpolation is used for calculation:

[0032] in From the current grid point to the th The distance to each simulated neighboring grid point To prevent decimals from being divided by zero, This is the indicator value for nearby points (1 for the target phase and 0 for the non-target phase). The search radius is the total number of simulated neighboring points within the search radius, which is 3 grid cells in this example. After generating the random access order, the conditional probability is calculated point by point and Monte Carlo sampling is performed. Finally, a planar sedimentary facies distribution map covering a 10×15 grid is generated, in which channel facies account for approximately 28% of the total area, floodplain facies account for approximately 55%, and natural levee facies and breach fan facies account for approximately 17% in total. The distribution pattern is basically consistent with the preset prior probability map.

[0033] Step (3): Based on prior geological data and planar sedimentary facies distribution map, a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map is generated one by one using a probability sampling method; Specifically, in the vertical stratification stage, a one-dimensional sequence generation is performed for each planar grid point: if the grid point has drilling hard constraint data, the measured lithology sequence is directly read as its one-dimensional sequence; if there is no drilling hard constraint, the sedimentary facies type of the grid point is read, and the global sandstone proportion and thickness probability distribution model of that facies type is directly retrieved. To maintain lateral synergy among connected domains, grid points belonging to the same continuous sedimentary facies patch share a basic pseudo-random seed. Then, the sandstone layer thickness is iteratively sampled with the target sandstone volume fraction as a constraint, and the non-sandstone layer thickness is allocated according to the selected vertical distribution pattern. After staggered combination, total thickness conservation normalization is performed to generate the one-dimensional sand-mud interbedded sequence of the grid point. This step avoids spatial white noise by conditionalizing hard constraints and sharing random seeds, and skips the traditional three-dimensional parametric field kriging interpolation, significantly reducing the spatial computation.

[0034] Correspondingly, in the detailed example, step (3) is specifically described as follows: One-dimensional vertical sand-mud sequence is generated for each grid point. First, a hard constraint condition is determined: if hard drilling data exists at the grid point location, the well logging interpretation lithology sequence of the corresponding well is directly extracted as the vertical sequence for that grid point, skipping subsequent random sampling; if no hard constraint exists, one of the following three probability distribution models is selected based on the sedimentary facies type of the grid point in the planar facies diagram (normal distribution model is suitable for channel facies and natural levee facies, log-normal distribution model is suitable for crevasse fan facies, and uniform distribution model is suitable for floodplain facies): normal distribution:

[0035] Log-normal distribution:

[0036] Uniform distribution: ,

[0037] Calculate the total thickness of the target sandstone based on its volume fraction. (in The total thickness of the reservoir. (Target sandstone volume fraction), iteratively sample sandstone layer thickness until the cumulative thickness meets the target value. The thickness of non-sandstone layers is weighted according to a vertical distribution pattern: top-concentrated pattern. Bottom-focused mode Centralized model Uniform distribution pattern ,in The numbers represent the non-sandstone layers from top to bottom. This represents the total number of non-sandstone layers. The final thickness of each non-sandstone layer is... ,in This represents the total thickness of non-sandstone layers. After completing the sequence cross-combination, the total thickness is conserved and normalized, and adjacent thin layers of the same type are merged (layers with a thickness of less than 0.1m are merged into adjacent layers of the same type).

[0038] Step (4): Based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point, perform three-dimensional spatial stitching to obtain a three-dimensional grid model; Specifically, 3D integration stitches together the planar coordinates of the phase diagram with the one-dimensional sequence depth coordinates of each grid point to form a complete 3D geological body, creating an independent columnar aggregate without an isochronous framework.

[0039] Corresponding to the detailed example, the planar coordinates of the phase diagram are combined with the cumulative depth coordinates of the one-dimensional profiles of each grid point. Through data stitching, a regular three-dimensional columnar system without isochronous stratigraphic framework is constructed. The three-dimensional geological body generated in this example has an irregular layered structure of 10×15× (the number of layers at each grid point varies), with a total number of nodes of approximately 6000–9000, depending on the number of sand and mud layers at each grid point.

[0040] Step (5): Based on the graph theory undirected graph algorithm, identify the connectivity of the three-dimensional sand body mesh of the three-dimensional mesh model; Specifically, connectivity identification employs a graph-based undirected graph algorithm: unique identifiers are generated for sandstone layers that meet thickness thresholds; undirected connecting edges are established for sand body unit pairs that simultaneously meet four conditions: spatial proximity, phase compatibility, vertical depth overlap, and non-interlayer occlusion; then, breadth-first search is used to identify all connected components, assigning sand body numbers to each component and statistically analyzing indicators such as volume and span. The vertical depth overlap condition differs from simple connectivity judgments based on lithology thresholds; the non-interlayer occlusion condition further eliminates misjudgments due to lateral blockage by transverse interlayers. Together, these two conditions constitute the core geophysical constraints for graph-based connectivity evaluation.

[0041] For the example, construct an undirected graph. Vertex set For all samples that meet the thickness threshold (in this example, we take...), Sandstone layer unit under condition m), edge set A sand body element is a pair of elements that simultaneously satisfy all four connectivity criteria. The unique identifier for a sand body element is defined as follows: ,in row index ( (Total number of rows in the grid) For column indexes ( (Total number of columns in the grid) The grid point is assigned the top-to-bottom sequence number of the sand layer, and the absolute top depth of each unit is also recorded. and bottom depth .

[0042] For two sand body elements to establish a connected edge, four conditions must be met simultaneously.

[0043] Spatial proximity condition is ,in and The coordinates of the two grid points are the row and column coordinates in the plane.

[0044] The compatibility conditions stipulate that connectivity is allowed between channel facies and between floodplain facies, and the rules for other facies combinations can be extended based on sedimentological principles.

[0045] The vertical overlap condition is:

[0046] in and These represent the top and bottom depths of the two sand body units, respectively. m is the overlap threshold.

[0047] Non-interlayer occlusion requirements: Extract the complete layer sequence of two grid points within the vertical overlap region, and check layer by layer whether there is a continuous thickness exceeding the interlayer thickness threshold. (This example uses a default value of 0.5m, corresponding to a mudstone layer capillary displacement pressure that is significantly greater than the minimum thickness of sandstone) The mudstone layer segment simultaneously covers the overlapping interval; if it exists, it is considered as a layer-separated lateral blockage, and no connecting edge is established.

[0048] Taking a specific grid point pair as an example: At grid point (5,5), a sandstone layer has a top depth of 120m and a bottom depth of 128m; at grid point (5,6), a sandstone layer has a top depth of 122m and a bottom depth of 130m. The vertical overlap interval is [122m, 128m]. Checking the stratigraphic sequence of grid point (5,5) within this interval reveals: 120–123m is sandstone, 123–125m is mudstone (2m thick), and 125–128m is sandstone. Checking grid point (5,6) within the same interval reveals: 122–130m is continuous sandstone. Since the mudstone layer only appears at grid point (5,5), it does not constitute continuous cover at the grid point boundaries and does not trigger layer shading; therefore, a connecting edge is established. If the intervals from 123 to 126m of grid point (5,5) are all mudstone (thickness 3m > 0.5m), and the corresponding interval of grid point (5,6) is also mudstone, then it is determined to be a layer occlusion, and no connected edge is established. It should be noted that in this method, non-sandstone layers (mudstone layers) are not considered nodes in the connected graph. Therefore, sandstone layers vertically separated by mudstone layers do not share depth intervals because they are independent sandstone units. The non-layer occlusion condition further solves the problem of determining lateral layer blockage between adjacent grid points. A breadth-first search algorithm is used to traverse and identify connected components in the constructed adjacency graph, and the overall connectivity rate is:

[0049] in The number of sand body units that are connected to at least one neighboring sand body unit. This represents the total number of sand body elements that meet the thickness threshold condition across the entire region. In this example, approximately 1800 qualified sand body elements were extracted within a 10×15 grid, and three main connected sand body groups were identified. The overall connectivity rate was 71.3%, and the largest connected sand body group contained approximately 1100 elements, extending laterally to cover approximately 68% of the modeling area.

[0050] Step (6): Output the uncertainty analysis results of connectivity using the Monte Carlo method.

[0051] Specifically, N Monte Carlo samplings are performed on key parameters such as sandstone proportion and vertical distribution pattern. Each time, the complete modeling process is completed again with the sampling parameters. The P10, P50, P90 and other quantiles are calculated for the N sets of connectivity results, and a connectivity uncertainty report is output.

[0052] In the detailed examples, an unconditional full refresh strategy is adopted: the planar phase diagram and vertical stratigraphic sequence are independently regenerated in each loop to ensure that each sample is statistically independent, thus correctly reflecting the inherent variability of stochastic modeling. The random sampling model for the key parameters is: the target sandstone proportion follows... (truncation) The vertical distribution pattern is randomly selected from four patterns with equal probability. 1000 independent simulations are performed, with each simulation re-executed using the sampling parameters, following steps S1 to S5. The mean, standard deviation, and quantiles of the 1000 connectivity results are calculated.

[0053]

[0054] in, To simulate the total number of times, For the first The connectivity obtained from the simulation. This is the inverse function (i.e., the quantile function) of the cumulative distribution function of connectivity results. The results of 1000 Monte Carlo simulations in this example are: mean connectivity 68.7%, standard deviation 8.2%, P10 57.4%, P50 69.1%, P90 79.3%, and the P10–P90 interval is approximately 22 percentage points wide. This indicates that within the uncertainty range of the sandstone proportion, there is a moderate degree of uncertainty in the reservoir connectivity of this block. The P50 result can be used as a basic parameter for optimizing development schemes.

[0055] Based on the above embodiments and detailed calculation examples, a comparison of theoretical complexity shows that traditional Sequential Gaussian Simulation (SGSIM) requires a three-dimensional neighborhood search for each grid node and the solution of the covariance matrix during execution (computational complexity is O(n log n)). The "2D+1D" dimensionality reduction architecture and inverse distance weighted fusion formula used in this example not only eliminate the need for three-dimensional search, but also reduce the complexity of core probability calculation to a minimum. The computational efficiency is significantly lower than that of traditional methods. Therefore, for any given grid size, the time consumption and memory allocation for a single simulation are far lower than those for a full 3D simulation. Furthermore, due to the linear nature of the algorithm, it avoids the computational bottleneck of traditional methods when processing industrial-grade grids with millions of nodes, such as 100×100×50 grids. Building upon this, graph theory connectivity identification introduces physical threshold constraints. In the comparative verification between this example and the geological conceptual model, it effectively eliminates misjudgments of lateral blockage by interlayers and maintains effective connectivity of sand bodies through a continuous phase random seed sharing mechanism, achieving a good balance between computational efficiency and geological evaluation accuracy.

[0056] exist Figure 2 A three-dimensional phase-controlled sand body modeling and connectivity evaluation system includes: The data management module is used to acquire prior geological data for the area to be modeled. The planar facies map simulation module is used to simulate and generate a planar sedimentary facies distribution map of the area to be modeled based on prior geological data and through probability-guided sequential instructions. The vertical profile generation module is used to generate a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map based on prior geological data and planar sedimentary facies distribution map using a probability sampling method. The 3D integration module is used to perform 3D spatial stitching based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point to obtain a 3D mesh model; The graph theory connectivity analysis module is used to identify the connectivity of 3D sand body meshes in 3D mesh models based on graph theory undirected graph algorithms. The Monte Carlo analysis module is used to output uncertainty analysis results of connectivity using the Monte Carlo method. In the embodiments provided by this invention, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0057] The integrated units implemented as software functional units described above can be stored in a computer-readable storage medium. These software functional units, stored in a storage medium, include several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) or processor to execute some steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0058] The above description is merely 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 scope of protection of the present invention.

Claims

1. A method for three-dimensional phase-controlled sand body modeling and connectivity evaluation, characterized in that, include: Step (1): Obtain prior geological data for the area to be modeled; Step (2): Based on prior geological data, a planar sedimentary facies distribution map of the area to be modeled is generated by probability-guided sequential indices. Step (3): Based on prior geological data and planar sedimentary facies distribution map, a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map is generated one by one using a probability sampling method; Step (4): Based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point, perform three-dimensional spatial stitching to obtain a three-dimensional grid model; Step (5): Based on the graph theory undirected graph algorithm, identify the connectivity of the three-dimensional sand body mesh of the three-dimensional mesh model; Step (6): Output the uncertainty analysis results of connectivity using the Monte Carlo method.

2. The method according to claim 1, characterized in that, The prior geological data in step (1) includes: a prior probability distribution map of sedimentary facies in the modeling area; statistical parameters of sandstone volume ratio and single-layer thickness distribution parameters for each facies type; and drilling and seismic interpretation constraint data.

3. The method according to claim 2, characterized in that, In step (2), the simulated value of each grid point in the planar sedimentary facies distribution map is calculated based on a heuristic linear fusion method of conditional probability. The factors included in the calculation are: the prior probability of each grid point and the neighborhood probability corresponding to each grid point.

4. The method according to claim 3, characterized in that, in, The neighborhood probability of each grid point is calculated using inverse distance weighted interpolation.

5. The method according to claim 2, characterized in that, In step (3), when there are constraints from drilling and seismic interpretation data at the corresponding location of a grid point, a one-dimensional vertical sequence corresponding to that grid point is generated based on the drilling and seismic interpretation data.

6. The method according to claim 5, characterized in that, In step (3), when there are no constraints from drilling and seismic interpretation data at the corresponding location of the grid point, a probability distribution model corresponding to the sedimentary facies type is selected for sampling based on the sedimentary facies type corresponding to the grid point.

7. The method according to claim 6, characterized in that, Probability distribution models include: normal distribution, log-normal distribution, and uniform distribution.

8. The method according to claim 2, characterized in that, In step (5), the graph theory undirected graph algorithm calculates whether there are connected edges between two three-dimensional sand body meshes based on spatial proximity conditions, compatibility conditions, vertical overlap conditions, and non-layer occlusion conditions.

9. The method according to claim 2, characterized in that, In step (6), based on a set number of independent samples, the mean, standard deviation, and set percentile values ​​of the connectivity rate are calculated according to the sampling results, and the uncertainty analysis results of the connectivity rate are output.

10. A three-dimensional phase-controlled sand body modeling and connectivity evaluation system, characterized in that, include: The data management module is used to acquire prior geological data for the area to be modeled. The planar facies map simulation module is used to simulate and generate a planar sedimentary facies distribution map of the area to be modeled based on prior geological data and through probability-guided sequential instructions. The vertical profile generation module is used to generate a one-dimensional vertical sequence corresponding to each grid point of the planar sedimentary facies distribution map based on prior geological data and planar sedimentary facies distribution map using a probability sampling method. The 3D integration module is used to perform 3D spatial stitching based on the planar sedimentary facies distribution map and the one-dimensional vertical sequence corresponding to each grid point to obtain a 3D mesh model; The graph theory connectivity analysis module is used to identify the connectivity of 3D sand body meshes in 3D mesh models based on graph theory undirected graph algorithms. The Monte Carlo analysis module is used to output uncertainty analysis results of connectivity using the Monte Carlo method.

Images