Non-uniform corner grid optimization method for reservoir numerical simulation based on attribute distribution

Through the non-uniform corner mesh optimization method based on attribute distribution, geological body information is identified and grid adjustment is performed, the problem of low grid optimization efficiency in reservoir numerical simulation is solved, and the grid scale is reduced and calculation accuracy is achieved. promote.

CN111324947BActive Publication Date: 2024-11-08CHINA PETROLEUM & CHEMICAL CORP +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202010073765.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-01-21
Publication Date
2024-11-08
Estimated Expiration
2040-01-21

AI Technical Summary

Technical Problem

The prior art cannot effectively optimize the non-uniform corner grid of fine geological models in numerical simulation of reservoirs, resulting in large calculation volume, low calculation efficiency and small loss of geological information, affecting the simulation accuracy.

Method used

Using the non-uniform corner mesh optimization method based on attribute distribution, we identify the geological information by defining the mesh grade, parameter threshold and inclusion relationship, conducting overall roughening and local re-sampling, and generating suitable numerical simulations with attribute resampling. multi-level grid structure.

Benefits of technology

The model grid scale is effectively reduced, geological characteristics are retained, computational efficiency and accuracy are improved, and convergence and computing efficiency of numerical simulation are ensured.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111324947B_ABST
    Figure CN111324947B_ABST
Patent Text Reader

Abstract

The present invention provides a non-uniform corner grid optimization method for reservoir numerical simulation based on attribute distribution, comprising: step 1, defining multiple parameters as a basis for judging the adjustment of the non-uniform grid; step 2, identifying the geological body information in the fine geological model, and overall coarsening the other parts of the model to obtain a preliminary optimized grid model; step 3, checking the physical property parameters of the fine grid contained in each coarsened grid in the optimized grid model, if the physical property parameters change more drastically, the coarse grid is adjusted twice to form a local refined grid; step 4, resampling the reservoir physical property parameters; step 5, saving the optimization results, and generating a non-uniform optimized grid model. The non-uniform corner grid optimization method for reservoir numerical simulation based on attribute distribution effectively guarantees the convergence of numerical simulation operations, and provides a good application bridge for the continuation of the results of the fine configuration model in oil and gas numerical simulation.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of oilfield development, and in particular to a non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution. Background Art

[0002] In order to accurately depict the reservoir structure, three-dimensional geological models usually use relatively fine grids for description, with a scale of millions, tens of millions, or even hundreds of millions of nodes. However, due to the large amount of computation required for reservoir numerical simulation models, the computing power and efficiency of the simulator are limited by software and hardware conditions and cannot withstand excessively large grid sizes. Therefore, in the process of applying geological models to numerical simulations, it is necessary to coarsen the fine geological models to ensure the application of numerical models. Conventional use of uniform coarsening to obtain numerical simulation models will result in the loss of fine geological information and errors in physical parameters, affecting the accuracy of numerical simulation research.

[0003] In the Chinese patent application with application number: 201610700948.7, a method for optimizing the vertical grid size of an ultra-high water-cut reservoir model is involved. The method for optimizing the vertical grid size of an ultra-high water-cut reservoir model includes: step 1: obtaining a reasonable grid size under different physical properties; step 2: obtaining a reservoir thickness influencing factor; step 3: obtaining a viscosity influencing factor; step 4: obtaining a water content influencing factor; and step 5: calculating a reasonable vertical grid size. In the process of determining the grid scale, this method only considers the influence of various influencing factors (reservoir thickness, viscosity, water content) on model optimization in the vertical direction, and does not consider the actual attribute numerical connection and the spatial distribution of fine geological bodies. Therefore, the influence of fine geological research results is ignored in the optimization process.

[0004] In the Chinese patent application with application number: 201510029000.9, a non-uniform grid wave equation finite difference template optimization design method is involved, wherein the method includes the following steps: gridding the simulation area of ​​the actual geological model according to the time sampling interval and the space sampling interval; according to the given maximum allowable error and wave number range, the corresponding operator length is calculated for different sound wave velocities; based on the least squares optimized time-space domain finite difference method and the corresponding operator length, the optimized finite difference coefficient of the grid is obtained; the obtained optimized finite difference coefficient is brought into the wave equation in the difference format to perform forward simulation of the wave equation. This method is mainly used in the field of seismic wave forward numerical simulation technology, and is not suitable for fine geological model grid optimization. Moreover, this method is only suitable for optimizing cube grids into rectangular and cuboid grids, and is not suitable for optimizing corner grids.

[0005] To this end, we invented a new non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution to solve the above technical problems. Summary of the invention

[0006] The purpose of the present invention is to provide a non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution, which can reduce the model scale and ensure the numerical simulation operation requirements.

[0007] The purpose of the present invention can be achieved through the following technical measures: a non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution, the non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution comprises: step 1, defining the order of the coarsening grid in the I and J directions, parameter thresholds, and the coarse and fine grid inclusion relationship as the judgment basis for the adjustment of the non-uniform grid; step 2, identifying the geological body information in the fine geological model, and overall coarsening the other parts of the model to obtain a preliminary optimized grid model; step 3, checking the physical property parameters of the fine grid contained in each coarsening grid in the optimized grid model, if the physical property parameters change more drastically, the coarse grid is adjusted twice to form a local refined grid; step 4, resampling the reservoir physical property parameters; step 5, saving the optimization results, and generating a non-uniform optimized grid model.

[0008] The purpose of the present invention can also be achieved by the following technical measures:

[0009] Step 1 includes:

[0010] (5) Set the maximum grid level in the I and I directions as the optimization constraint parameter;

[0011] (6) Set the maximum grid level in the I and J directions of this optimization;

[0012] (7) Setting the spatial scale inclusion relationship of each level of grid from fine to coarse in the I and J directions, that is, the number of fine grids of the next level contained in each level of coarse grid;

[0013] (8) Setting the grid attribute subdivision optimization threshold

[0014] In step 2, the geological body information in the fine geological model is identified, including wells, faults, pinch-out lines, and configuration boundaries, which are retained as the original fine grid, and the rest of the model is overall coarsened to obtain a preliminary optimized grid model.

[0015] Step 2 includes:

[0016] Step 2a, marking the grid positions of geological objects in the model, including faults, wells, pinch-out lines, and configuration boundaries, as the starting positions for grid optimization;

[0017] Step 2b, according to the size of the grid level, the entire study area is gradually coarsened from the initial grid level to the surrounding areas, and the coarsest level is optimized after reaching the highest level;

[0018] Step 2c: optimize the entire area and generate a mesh model after initial optimization.

[0019] In step 2b, the grids in the rows and columns where the well is located and its surrounding adjacent grids are set as the initial finest grids; the grids in the rows or columns on the direction of the main fault and pinch-out line are set as the initial finest grids.

[0020] In step 2c, starting from the set initial finest level grid, optimize step by step from the finest level to the coarsest level in all directions. When the set coarsest level is reached, keep optimizing at the coarsest level; optimize all fine grids in the study area according to the set grid level to obtain a preliminary optimized grid model.

[0021] Step 3 includes:

[0022] Step 3a: Calculate the volume weighted deviation between the optimized coarse mesh attribute parameters and the original fine mesh attribute parameters and mesh subdivision threshold Make comparisons;

[0023] Step 3b, if the coarse grid attribute parameter deviation value Greater than the set threshold Then reduce the grid order of the coarse grid;

[0024] Step 3c, recalculate the attribute parameter deviation value of the downgraded grid until

[0025] Step 3d, complete the adjustment of the entire area to form the final optimized grid model.

[0026] In step 3a, the changes in the reservoir physical parameters of the fine grids contained in the coarse grid are checked in turn, and the variable deviations of all coarse grids are calculated. Mesh subdivision criteria Make a comparison.

[0027] In step 3b, if there are more than 1 / n coarse grids in a row or column, the variable deviation is greater than the set threshold (where n is determined by the accuracy requirements of reservoir numerical simulation and actual application needs), then the grid level of the coarse grid in the row or column is reduced by one level and the variable deviation is calculated until Record the grid level after the current coarse grid is adjusted and mark it.

[0028] In step 3c, starting from the marked coarse grid, steps 3a and 3b are repeated, that is, the current optimized grid model is readjusted according to the optimization rule and the set grid level until the variable deviation of the coarse grid is Less than threshold

[0029] In step 4, the reservoir physical property parameters are resampled according to the adjusted non-uniform grid structure to ensure the rationality and smoothness of various reservoir parameters.

[0030] The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution in the present invention, for a fine geological model, constructs a multi-level grid structure, and performs reasonable coarsening, merging and refining adjustments on homogeneous areas while effectively retaining effective geological structural features such as faults, pinch-outs, and configuration interfaces, as well as attribute heterogeneity, so as to reduce the grid scale of the model and meet the needs of numerical simulation operations.

[0031] The excellent effect of the present invention is that, drawing on the idea of ​​the adaptive grid method, an improved non-uniform grid optimization method is proposed. The method adopts a two-level optimization method. First, fine grids are retained in the key research areas of reservoir geological characteristics, and other areas are coarsened; secondly, areas with strong physical property heterogeneity are detected and refined twice to form fine grids. This method can not only effectively retain the inheritance and application of fine geological results in numerical simulation research, but also reduce the number of grids as much as possible while ensuring the calculation accuracy, thereby ensuring the computational efficiency of numerical simulation applications. In the two-level optimization process, a uniform gradual transition between adjacent coarse and fine grids is also achieved, which effectively ensures the convergence of numerical simulation operations. Through the implementation of this grid optimization algorithm, a good application bridge is provided for the continuation of the results of fine configuration models in oil and gas numerical simulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Figure 1 It is a flow chart of a specific embodiment of the method for optimizing non-uniform corner point grids for reservoir numerical simulation based on attribute distribution of the present invention;

[0033] Figure 2 A schematic diagram of a conceptual model of a grid with uniform transition between thick and thin after optimization in a specific embodiment of the present invention;

[0034] Figure 3 It is a schematic diagram of a grid model in which a red frame portion is required to be adjusted in a specific embodiment of the present invention;

[0035] Figure 4 is a schematic diagram of a grid model after local adjustment in a specific embodiment of the present invention;

[0036] Figure 5 A comparison diagram of attribute rendering effects before and after non-uniform grid optimization in a specific embodiment of the present invention;

[0037] Figure 6 This is a property distribution diagram before and after grid optimization in a specific embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to make the above and other purposes, features and advantages of the present invention more obvious and easy to understand, preferred embodiments are given below and described in detail with reference to the accompanying drawings.

[0039] like Figure 1 As shown, Figure 1 The present invention is a flow chart of the non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution.

[0040] Step 101, respectively define the order of the coarsening grid in the I and J directions, parameter thresholds, coarse and fine grid inclusion relations and other parameters as a basis for determining the uneven grid adjustment;

[0041] The specific steps include:

[0042] (1) Set the maximum grid level in the I and I directions as the optimization constraint parameter;

[0043] (2) Set the maximum grid level in the I and J directions of this optimization;

[0044] (3) Setting the spatial scale inclusion relationship of each level of grid from fine to coarse in the I and J directions, that is, the number of the next level of fine grids contained in each level of coarse grid;

[0045] (4) Setting the grid attribute subdivision optimization threshold

[0046] Step 102, identifying geological body information (such as wells, faults, pinch-out lines, configuration boundaries, etc.) in the fine geological model, retaining it as the original fine grid, and overall coarsening the rest of the model to obtain a preliminary optimized grid model;

[0047] The specific steps include:

[0048] (1) Mark the grid position of geological objects (such as faults, wells, pinch-out lines, configuration boundaries, etc.) in the model as the starting position for grid optimization;

[0049] (2) According to the size of the grid level, the entire study area is gradually coarsened from the initial grid level to the surrounding areas. After reaching the highest level, the coarsest level is optimized; the grids in the rows and columns where the well is located and its surrounding adjacent grids are set as the initial finest level grids (grid level is 1); the grids in the rows or columns along the main faults and pinch-out lines are set as the initial finest level grids.

[0050] (3) Optimize the entire area and generate the mesh model after the initial optimization; starting from the initial finest level mesh, optimize from the finest level to the coarsest level in all directions. When the coarsest level is reached, keep optimizing at the coarsest level. Optimize all fine meshes in the study area according to the set mesh level to obtain a preliminary optimized mesh model.

[0051] Step 103, checking the physical property parameters of the fine grids contained in each coarse grid in the optimized grid model, if the physical property parameters change drastically, the coarse grid is adjusted twice to form a local fine grid;

[0052] The specific steps include:

[0053] (1) Calculate the volume weighted deviation between the optimized coarse mesh attribute parameters and the original fine mesh attribute parameters and mesh subdivision threshold Compare; check the changes in the reservoir physical parameters of the fine grids contained in the coarse grid in turn, and calculate the variable deviation of all coarse grids Mesh subdivision criteria Make a comparison.

[0054] (2) If the coarse grid attribute parameter deviation value Greater than the set threshold The grid level of the coarse grid is reduced; if there are more than 1 / n coarse grids in a row or column, the variable deviation is greater than the set threshold (where n is determined by the accuracy requirements of reservoir numerical simulation and actual application needs), then the grid level of the coarse grid in the row or column is reduced by one level and the variable deviation is calculated until Record the grid level after the current coarse grid is adjusted and mark it;

[0055] (3) Recalculate the attribute parameter deviation value of the downgraded grid until Starting from the marked coarse grid, repeat steps (1) and (2), that is, readjust the current optimized grid model according to the optimization rule and the set grid level until the variable deviation of the coarse grid is Less than threshold

[0056] (4) Complete the adjustment of the entire area to form the final optimized grid model;

[0057] Step 104, resampling the reservoir physical property parameters according to the adjusted non-uniform grid structure to ensure the rationality and smoothness of various reservoir parameters;

[0058] Step 105, saving the optimization results and generating a non-uniform optimized grid model.

[0059] In a specific embodiment of the present invention, taking a regional grid model as an example, the number of grids in the grid model before optimization is 202*202*1, and the number of grids after optimization is 170*156*1, and the number of grids is reduced by 35.01%. Using the attribute weighted analysis method, the grid attribute distribution before and after optimization is comprehensively analyzed, the number of grids is effectively reduced, and the grid attribute distribution remains consistent. At the same time, the grids in the area with drastic attribute changes are finer, and the grids in the area with small attribute changes are coarser. The grid model is evenly transitioned from coarse to fine, and there is no grid nesting. The following steps are included:

[0060] Step 1: Set the maximum grid level in the I and J directions to 5, set the grid level in the I and J directions for this optimization to 3, and set the grid after optimization to 3 scales from coarse to fine. Set the grid attribute subdivision optimization threshold

[0061] Step 2, record the grid positions of wells, faults, pinch-out lines, etc., and start from the above positions as the starting position of grid optimization. The grid level is level 1 (the finest grid). Through attribute deviation calculation, combined with the set grid level, the grid model after initial optimization is generated ( Figure 2 ), the model contains three different levels of grids in the I and J directions, and the grids transition evenly from fine to coarse.

[0062] Step 3: Calculate the volume weighted deviation of the attribute parameters between the optimized coarse grid and the original fine grid. By distinguishing between parameter deviation and threshold, if the variable deviation of the coarse grid is greater than the set threshold Then the grid level of the coarse grid of the row or column is reduced by one level and the variable deviation is calculated until Record the grid level after the current coarse grid is adjusted, mark it, and make adjustments based on the grid in the marked area ( Figure 3 ) related non-compliant grid areas until the grid attribute deviation value is less than the set threshold Save the locally adjusted mesh model, that is, the final optimized mesh model ( Figure 4 ).

[0063] Step 4: The local attribute distribution of the actual mesh model generated by the above algorithm ( Figure 5 ), and the model attribute distribution before and after optimization is compared by grid volume weighted calculation ( Figure 6 ).

Claims

1. A non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution, characterized in that: The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution includes: Step 1, define the parameters such as the order of the coarsening grid in the I and J directions, the parameter threshold, and the coarse and fine grid inclusion relationship as the basis for judging the uneven grid adjustment; Step 2: Identify the geological body information in the fine geological model and perform overall coarsening on other parts of the model to obtain a preliminary optimized grid model; Step 3: Check the physical parameters of the fine grid contained in each coarse grid in the optimized grid model. If the physical parameters change drastically, the coarse grid is adjusted twice to form a local fine grid, including: Step 3a: Calculate the volume weighted deviation between the optimized coarse mesh attribute parameters and the original fine mesh attribute parameters Step 3b, if the coarse grid attribute parameter deviation value Greater than the set threshold Then reduce the grid order of the coarse grid; Step 3c, recalculate the attribute parameter deviation value of the downgraded grid until Step 3d, complete the adjustment of the entire area to form the final optimized grid model; Step 4, resampling the reservoir physical property parameters; Step 5: Save the optimization results and generate an uneven optimized grid model.

2. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1 is characterized in that: Step 1 includes: (1) Set the maximum grid levels in the I and J directions as optimization constraint parameters; (2) Set the maximum grid level in the I and J directions of this optimization; (3) Setting the spatial scale inclusion relationship of each level of grid from fine to coarse in the I and J directions, that is, the number of the next level of fine grids contained in each level of coarse grid; (4) Setting the grid attribute subdivision optimization threshold 3. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1 is characterized in that: In step 2, the geological body information in the fine geological model is identified, including wells, faults, pinch-out lines, and configuration boundaries, which are retained as the original fine grid, and the rest of the model is overall coarsened to obtain a preliminary optimized grid model.

4. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 3 is characterized in that: Step 2 includes: Step 2a, marking the grid positions of geological objects in the model, including faults, wells, pinch-out lines, and configuration boundaries, as the starting positions for grid optimization; Step 2b, according to the size of the grid level, the entire study area is gradually coarsened from the initial grid level to the surrounding areas, and the coarsest level is optimized after reaching the highest level; Step 2c: optimize the entire area and generate a mesh model after initial optimization.

5. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 4 is characterized in that: In step 2b, the grids in the rows and columns where the well is located and its surrounding adjacent grids are set as the initial finest grids; the grids in the rows or columns on the direction of the main fault and pinch-out line are set as the initial finest grids.

6. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 4 is characterized in that: In step 2c, starting from the set initial finest level grid, optimize step by step from the finest level to the coarsest level in all directions. When the set coarsest level is reached, keep optimizing at the coarsest level; optimize all fine grids in the study area according to the set grid level to obtain a preliminary optimized grid model.

7. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1 is characterized in that: In step 3a, the changes in the reservoir physical parameters of the fine grids contained in the coarse grid are checked in turn, and the variable deviations of all coarse grids are calculated. Mesh subdivision criteria Make a comparison.

8. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1 is characterized in that: In step 3b, if there are more than 1 / n coarse grids in a row or column, the variable deviation is greater than the set threshold Where n is determined according to the accuracy requirements of reservoir numerical simulation and actual application needs. The variable deviation is calculated by reducing the grid level of the coarse grid in the row or column by one level. until Record the grid level after the current coarse grid is adjusted and mark it.

9. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1, characterized in that: In step 3c, starting from the marked coarse grid, steps 3a and 3b are repeated, that is, the current optimized grid model is readjusted according to the optimization rule and the set grid level until the variable deviation of the coarse grid is Less than threshold 10. The non-uniform corner point grid optimization method for reservoir numerical simulation based on attribute distribution according to claim 1, characterized in that: In step 4, the reservoir physical property parameters are resampled according to the adjusted non-uniform grid structure to ensure the rationality and smoothness of various reservoir parameters.