Carbon dioxide geological storage evaluation method and system based on multi-source data fusion and storage medium thereof

Abstract

The application discloses a kind of carbon dioxide geological storage evaluation methods based on multi-source data fusion, specifically for: obtaining target storage area data, space registration and scale uniformity are carried out, and multi-source reservoir data set is generated;The target storage area is divided into spatial evaluation unit and the reservoir attribute is extracted, and the spatial evaluation unit data set is generated;The storage space state, injection accessible state and storage safety state of each spatial evaluation unit are determined respectively;Identify the short board unit of storage that does not meet the preset storage condition;With the unit corresponding to injection well as starting point, the spatial connectivity relationship is established by expanding search along the transmissible connection, and those without alternative path are determined as blocked influence range;The theoretical unit storage capacity is calculated, and the effective storage potential and spatial distribution result are obtained by processing. Meanwhile, the application also provides an evaluation system and a storage medium. The application can identify local restriction area and its connectivity blocking influence, to avoid overestimating theoretical storage capacity.

Description

Technical Field

[0001] This invention relates to the field of carbon dioxide geological storage evaluation and data processing technology, and in particular to a method, system and storage medium for evaluating carbon dioxide geological storage based on multi-source data fusion. Background Technology

[0002] With the development of carbon dioxide capture, utilization, and storage (CCUS) technology, the suitability evaluation and storage potential calculation of carbon dioxide geological storage sites have gradually become important steps in the utilization of underground reservoirs. Existing technologies typically collect geological and engineering data on the target area, including porosity, permeability, reservoir thickness, formation pressure, fault distribution, caprock thickness, caprock fracture pressure, and injection well location. Data from different sources are then processed, normalized, or spatially interpolated, and the theoretical storage capacity is estimated based on parameters such as reservoir area, effective thickness, porosity, and storage efficiency. Some evaluation methods also comprehensively evaluate the reservoir's storage capacity, injection conditions, and storage safety by setting evaluation indicators and corresponding weights, thereby determining the suitability and storage potential of candidate storage areas. For geological data with spatial distribution attributes, geological parameters at different locations within the reservoir can be characterized through gridding or 3D geological modeling, providing fundamental data for carbon dioxide geological storage evaluation.

[0003] However, underground reservoirs typically exhibit significant spatial heterogeneity. Existing methods often rely on regional average porosity, average permeability, average reservoir thickness, or comprehensive evaluation results to calculate storage potential, making it difficult to further identify localized restricted areas caused by insufficient storage space, low permeability, abnormal formation pressure, or insufficient safety conditions due to adjacent faults or caprocks. Furthermore, for reservoirs whose geological conditions meet storage requirements but whose effective connection to the injection well is blocked by the aforementioned localized restricted areas, existing evaluation processes typically still include them in the available storage capacity, leading to a discrepancy between the theoretical storage capacity and the actual usable storage space under current injection conditions.

[0004] Therefore, how to identify local confined areas based on a unified spatial benchmark of multi-source geological data, and further consider the blocking effect of local confined areas on the spatial connectivity of other qualified reservoirs, so as to obtain effective storage potential that is consistent with the current injection conditions, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] Purpose of the invention: The purpose of this invention is to provide a method for evaluating carbon dioxide geological storage based on multi-source data fusion, which solves the problem that differences in coordinate system, sampling scale and data resolution between reservoir data from different sources and at different spatial scales affect the data basis for storage evaluation.

[0006] Meanwhile, the present invention also provides a carbon dioxide geological storage evaluation system based on multi-source data fusion to solve the above problems.

[0007] In addition, the present invention also provides a computer-readable storage medium that solves the above-mentioned problems.

[0008] Technical Solution: The present invention provides a method for evaluating carbon dioxide geological storage based on multi-source data fusion, comprising the following steps: Step S10: Obtain porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters composed of caprock thickness and caprock fracture pressure, and injection well location data of the target storage area; perform spatial registration and scale unification on the data to generate a multi-source reservoir dataset, wherein each reservoir data in the multi-source reservoir dataset is associated with a corresponding spatial location; Step S20: Divide the target storage area into multiple spatial evaluation units according to the multi-source reservoir dataset; extract the reservoir attributes corresponding to each spatial evaluation unit; and generate a spatial evaluation unit dataset recording the spatial number and adjacency relationship of each spatial evaluation unit; Step S30: Based on the spatial evaluation unit dataset, respectively... Step S40: Determine the storage space status, injection reachability status, and sealing safety status of each spatial evaluation unit; Step S50: Identify spatial evaluation units whose storage space status, injection reachability status, and sealing safety status do not meet the corresponding preset conditions as sealing bottleneck units; Step S51: Starting from the spatial evaluation unit corresponding to the injection well location data, perform an extended search among the spatial evaluation units based on the adjacency relationship in the spatial evaluation unit dataset and the injection reachability status to construct a reachable unit set; When the extended search reaches the sealing bottleneck unit, terminate the search along the extended direction; Determine the spatial evaluation unit located on the side of the sealing bottleneck unit away from the starting point, and whose storage space status and sealing safety status are both satisfied, as a blocked unit to be judged;

[0009] Determine whether each of the obstructed units to be determined has an alternative path that bypasses the sealing bottleneck unit and connects to the starting point. Determine the obstructed units to be determined without the alternative path as the obstruction influence range. Step S60: Calculate the theoretical unit sealing capacity of each spatial evaluation unit based on the spatial evaluation unit dataset and summarize it to obtain the theoretical sealing capacity. Subtract the theoretical unit sealing capacity corresponding to the sealing bottleneck unit and the spatial evaluation units within the obstruction influence range from the theoretical sealing capacity to obtain the effective sealing potential and generate the corresponding spatial distribution results.

[0010] A further limitation of this technical solution is that the specific method for step S30, which determines the storage space status, injection reachability status, and sealing security status of each spatial evaluation unit based on the spatial evaluation unit dataset, is as follows:

[0011] S31. Extract the porosity and reservoir thickness corresponding to each spatial evaluation unit from the spatial evaluation unit dataset. Determine the usable pore space of the corresponding spatial evaluation unit based on the porosity and reservoir thickness. Compare the usable pore space with a preset minimum storage space threshold. If the usable pore space is not lower than the preset minimum storage space threshold, mark it as a storage space satisfied state; otherwise, mark it as a storage space unsatisfied state, thus forming the storage space state. S32. Extract the permeability, formation pressure, and adjacency relationship corresponding to each spatial evaluation unit from the spatial evaluation unit dataset. Determine adjacent spatial evaluation units based on the adjacency relationship. When the permeability of adjacent spatial evaluation units is not lower than the preset minimum permeability and the absolute value of the difference between their formation pressures is within a preset allowable pressure difference range, determine the adjacent spatial evaluation units. The evaluation units are transitively connected; a spatial evaluation unit that forms a transitive connection with at least one adjacent spatial evaluation unit is marked as an injection reachable state, otherwise it is marked as an injection unreachable state, thus forming the injection reachable state; S33, extract the fault distance, caprock safety parameters, and formation pressure corresponding to each spatial evaluation unit from the spatial evaluation unit dataset, generate a fault risk marker based on the comparison result of the fault distance with the preset fault safety distance, generate a caprock risk marker based on the comparison result of the caprock safety parameters with the preset caprock safety threshold, and generate a pressure risk marker based on the comparison result of the formation pressure with the preset allowable pressure upper limit; when any risk marker represents a risk state, the corresponding spatial evaluation unit is determined to be in a state where the sealing safety is not met, otherwise it is determined to be in a state where the sealing safety is met, thus forming the sealing safety state.

[0012] Further, in step S40, when identifying a short-board unit, firstly, spatial evaluation units whose at least one of the three states does not meet the corresponding preset conditions are determined as candidate short-board units; then, spatially continuous candidate short-board units are merged according to the adjacency relationship of the candidate short-board units to obtain the short-board region; finally, all spatial evaluation units constituting the short-board region are determined as the short-board unit; and in step S50, the blocking boundary during the extended search is the outer boundary of the short-board region.

[0013] Further, the extended search in step S50 specifically involves: a graph search that visits adjacent spatial evaluation units layer by layer along the transitive connections, with the starting point as the root node, and the visited spatial evaluation units are written into the reachable unit set; the method for determining whether each of the blocked units to be determined has an alternative path includes: after removing all the sealed short-board units from the spatial evaluation units participating in the search, re-searching for a continuous path from the starting point to the blocked unit to be determined; if the continuous path can be found, it is determined that an alternative path exists; otherwise, it is determined that no alternative path exists.

[0014] Furthermore, the method for spatial registration and scale unification of the data in step S10 to generate a multi-source reservoir dataset is as follows:

[0015] S11. Obtain the data source identifier, spatial coordinates, and sampling accuracy corresponding to porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters, and injection well location data, respectively.

[0016] S12. Transform each of the spatial coordinates using a unified coordinate reference system, and perform grid matching on the transformed data according to a preset spatial resolution;

[0017] S13. Perform consistency comparison on similar data matched to the same grid position according to the data source identifier. When the difference between similar data exceeds a preset difference threshold, generate a difference marker and determine the fusion weight and fusion order according to the difference marker and the sampling precision.

[0018] S14. According to the fusion weight and the fusion order, perform weighted merging of similar data, and write the data that has completed grid matching and weighted merging into the multi-source reservoir dataset.

[0019] Furthermore, the specific steps for generating the spatial evaluation unit dataset in step 20 are as follows:

[0020] S21. Generate a three-dimensional grid based on the reservoir boundary of the target storage area and the preset spatial resolution, and determine each three-dimensional grid as a spatial evaluation unit;

[0021] S22. Based on the spatial location of each spatial evaluation unit, read the corresponding reservoir attributes from the multi-source reservoir dataset;

[0022] S23. For fault or caprock anomaly zones that span two or more spatial evaluation units, corresponding parameters are assigned according to the spatial overlap ratio between the fault and each of the spatial evaluation units. Specifically, for faults, the spatial evaluation units corresponding to the fault impact are determined based on the spatial overlap ratio, and the fault distance corresponding to each spatial evaluation unit is calculated in combination with the spatial positional relationship between the fault plane and each spatial evaluation unit. For caprock anomaly zones, the caprock safety parameters corresponding to each spatial evaluation unit are determined based on the spatial overlap ratio.

[0023] S24. The spatial number, adjacency relationship and corresponding reservoir attributes of each spatial evaluation unit are associated and stored to generate the spatial evaluation unit dataset.

[0024] Furthermore, the specific steps for generating the corresponding spatial distribution results in step S60 are as follows:

[0025] S61. Calculate the theoretical storage capacity of each spatial evaluation unit based on the porosity, reservoir thickness and spatial scale of each spatial evaluation unit in the spatial evaluation unit dataset, and summarize the theoretical storage capacity.

[0026] S62. Perform deduplication processing on the storage short-board unit and the space evaluation unit within the obstruction influence range, and subtract the theoretical unit storage capacity corresponding to the deduplicated space evaluation unit from the theoretical storage capacity to obtain the effective storage potential.

[0027] S63. Generate a capacity status marker according to whether each spatial evaluation unit belongs to the sealed short-board unit, whether it is located in the obstructed influence range, and whether the corresponding theoretical unit sealed capacity is retained. Associate the capacity status marker with the spatial number of each spatial evaluation unit to generate the spatial distribution result.

[0028] Furthermore, before performing step S30, the method further includes steps of obtaining feature importance through machine learning, adjusting state judgment conditions, and re-identifying candidate bottleneck units, specifically:

[0029] Porosity, permeability, reservoir thickness, formation pressure, fault distance, and caprock safety parameters are extracted from historical storage evaluation samples as input features, and the storage availability status corresponding to the historical samples is used as training labels.

[0030] Machine learning is used to train the correspondence between the input features and the training labels to obtain the feature importance corresponding to each input feature;

[0031] Based on the importance of the aforementioned features, at least one of the following can be adjusted: the preset minimum storage space threshold, the preset minimum permeability, the preset allowable pressure difference, the preset fault safety distance, the preset caprock safety threshold, and the preset allowable pressure upper limit.

[0032] Another technical solution provided by the present invention is: a carbon dioxide geological storage evaluation system based on multi-source data fusion, comprising: a memory for storing multi-source reservoir datasets and spatial evaluation unit datasets; and a processor coupled to the memory, the processor being configured to execute any of the above methods.

[0033] Another technical solution provided by the present invention is: a computer-readable storage medium storing a computer program thereon, characterized in that the computer program implements any of the above methods when executed by a processor.

[0034] Beneficial effects: The carbon dioxide geological sequestration evaluation method, system, and storage medium based on multi-source data fusion provided by this invention have the following effects:

[0035] (1) This invention obtains data on porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters, and injection well location of the target storage area. It performs spatial registration and scale unification on reservoir data from different sources and at different spatial scales, and divides the target storage area into spatial evaluation units according to a unified spatial resolution. This enables various geological data to be correlated according to their corresponding spatial locations, which can reduce the impact of differences in coordinate systems, sampling scales, and data resolutions of multi-source geological data on storage evaluation, and provides a unified data basis for subsequent reservoir status judgment and spatial connectivity analysis.

[0036] (2) The present invention evaluates each spatial evaluation unit from three aspects: storage space status, injection reachability status and storage safety status. It can implement the evaluation process to specific spatial locations, which is conducive to identifying local restricted areas formed by local low permeability, insufficient storage space, fault proximity, insufficient caprock conditions or abnormal formation pressure, thereby improving the ability of carbon dioxide geological storage evaluation results to characterize the heterogeneity of reservoir space.

[0037] (3) Based on the identification of the storage bottleneck unit, this invention establishes spatial connectivity relationships according to the injection well location, the adjacency relationship between spatial evaluation units, and the injection accessibility status. It also determines whether there is an alternative path that bypasses the storage bottleneck unit and can connect to the injection well for the spatial evaluation unit located on the other side of the storage bottleneck unit. For spatial evaluation units that meet the requirements of their own storage space conditions and storage safety conditions, but do not have an alternative path due to the blockage of the storage bottleneck unit, they are classified into the obstructed influence range. Thus, the storage potential evaluation is no longer limited to deducting the storage capacity corresponding to the storage bottleneck unit itself, but can also further consider the connectivity restrictions of the storage bottleneck unit on other usable reservoir spaces, so that the evaluation results can reflect the actual accessibility of the reservoir space under the current injection well layout conditions, and avoid the theoretical storage capacity being overestimated.

[0038] (4) The present invention calculates the theoretical unit storage capacity of each spatial evaluation unit. After obtaining the theoretical storage capacity, the theoretical storage capacity is corrected according to the storage short-board unit and the range of obstruction influence to obtain the effective storage potential. Combined with the spatial number of each spatial evaluation unit, the corresponding spatial distribution results are generated. The theoretical storage space can be distinguished from the actual usable storage space, so that the storage potential evaluation considers both reservoir capacity and space availability. This provides data basis for candidate storage area screening, injection well location evaluation and storage space utilization analysis.

[0039] Instruction manual illustrations

[0040] Figure 1 This is a flowchart of the carbon dioxide geological storage evaluation method based on multi-source data fusion proposed in this invention;

[0041] Figure 2 This is a flowchart of the multi-source reservoir data processing and spatial evaluation unit construction proposed in this invention;

[0042] Figure 3 This is a flowchart of the spatial evaluation unit state determination proposed in this invention;

[0043] Figure 4 This is a flowchart of the identification of the short-board unit proposed in this invention;

[0044] Figure 5 This is a flowchart illustrating the determination of the obstruction range proposed in this invention;

[0045] Figure 6 This is a schematic diagram of the identification of the short-board unit in this embodiment;

[0046] Figure 7 This is a schematic diagram illustrating the determination of the obstruction range in this embodiment. Detailed Implementation

[0047] The technical solution of the present invention will be described in detail below, but the scope of protection of the present invention is not limited to the embodiments described.

[0048] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. The described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0049] It should be noted that in this invention, the term "spatial evaluation unit" can also be expressed as an evaluation grid, reservoir block, or voxel, all of which mean the smallest evaluation space with defined spatial boundaries and used for independent evaluation; the term "transferable connection" refers to the connectivity between adjacent spatial evaluation units that meet the preset permeability conditions and formation pressure difference conditions; the term "storage bottleneck unit" refers to a spatial evaluation unit that does not meet the corresponding preset conditions in at least one of the three states, cannot be directly used as an effective storage space, or will restrict the actual utilization of other spatial evaluation units.

[0050] Example 1

[0051] like Figure 1 As shown, this embodiment provides a method for evaluating carbon dioxide geological sequestration based on multi-source data fusion, wherein the processor executes the following steps.

[0052] Step S10: Obtain porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters consisting of caprock thickness and caprock fracture pressure, and injection well location data of the target storage area. Perform spatial registration and scale unification on each data to generate a multi-source reservoir dataset.

[0053] In this step, the target storage area refers to the underground reservoir space for which the suitability and potential for carbon dioxide geological storage are to be evaluated. Porosity can be obtained from core experiments, well logging interpretation, or seismic attribute inversion; permeability can be obtained from core flow experiments, wellbore pressure recovery tests, or well logging interpretation; reservoir thickness can be obtained from wellbore interpretation data and seismic stratigraphic interpretation data; formation pressure can be obtained from downhole pressure measurement data or a calibrated pressure model. Fault distance refers to the spatial distance from the corresponding reservoir space location to the nearest identified fault, which can be calculated after obtaining the spatial coordinates of the fault plane from 3D seismic interpretation. Caprock safety parameters are composed of caprock thickness and caprock fracturing pressure. Caprock thickness is obtained through wellbore interpretation or 3D seismic interpretation, and caprock fracturing pressure is obtained through formation integrity testing, fracturing testing, or historical well data. Injection well location data includes at least the planar coordinates of candidate injection wells and the depth of the target injection interval.

[0054] Data source identifiers are used to distinguish the acquisition method of each data item, such as being labeled as core experimental data, well logging interpretation data, seismic inversion data, or spatial interpolation data. Sampling accuracy refers to the measurement or interpretation error corresponding to the data source, which can be obtained from the equipment's nominal accuracy, experimental repeated measurement results, or historical calibration results. Since different data may use different coordinate reference systems, different sampling intervals, and different spatial resolutions, the processor first transforms each spatial coordinate using a unified coordinate reference system, and then performs grid matching on the transformed data according to a preset spatial resolution. Data of the same type matched to the same grid position are compared for consistency according to the data source identifier. When the difference between data of the same type exceeds a preset difference threshold, a difference marker is generated, and the fusion weight and fusion order are determined based on the difference marker and sampling accuracy. Data of the same type are weighted and merged according to the fusion weight and fusion order, and the processed data is written into a multi-source reservoir dataset, so that each reservoir data is associated with its corresponding spatial location.

[0055] In one specific implementation, the preset spatial resolution in the horizontal direction is set to 50–200 m, and the spatial resolution in the vertical direction is set to 5–20 m; in this embodiment, the horizontal resolution is 100 m and the vertical resolution is 10 m throughout. After the above processing, a multi-source reservoir dataset is formed. Each reservoir attribute in the multi-source reservoir dataset is associated with a defined spatial location, thereby providing a data foundation for extracting multiple attributes according to the same spatial location.

[0056] Data source identifiers distinguish the acquisition method of each data item, such as core experimental data, well logging interpretation data, seismic inversion data, or spatial interpolation data. Sampling accuracy refers to the measurement or interpretation error corresponding to the data source, which can be obtained from the equipment's nominal accuracy, repeated experimental measurements, or historical calibration results. The preset difference threshold uses relative difference representation. For similar data such as porosity, permeability, reservoir thickness, and formation pressure, the preset difference threshold can be set to 10%–25%. A difference marker is generated when the relative difference between two data points from different data sources but corresponding to the same grid location exceeds 15%.

[0057] For data with discrepancies, data with higher sampling precision is prioritized. When multiple reliable data sources exist for the same location, the fusion weights are determined using a method negatively correlated with measurement error; that is, the smaller the measurement error, the higher the fusion weight. Then, data of the same type are normalized and weighted. The weighted merging can be expressed as:

[0058] In the formula, The corresponding reservoir attributes after fusion; U represents the number of data sources participating in the fusion at the same spatial location; The same reservoir properties provided for the u-th data source; Let be the fusion weight of the u-th data source. The sum of all fusion weights is 1. The fusion weight is determined by normalization based on the sampling accuracy of the corresponding data source.

[0059] Step S20: Based on the multi-source reservoir dataset, divide the target storage area into spatial evaluation units, extract the reservoir attributes corresponding to each spatial evaluation unit, and generate a spatial evaluation unit dataset, such as... Figure 2 As shown.

[0060] The reservoir boundary is determined based on the reservoir top surface, reservoir bottom surface, and planar distribution boundary obtained from 3D seismic interpretation. The processor establishes a regular 3D mesh based on the reservoir boundary and a preset spatial resolution, defining each 3D mesh as a spatial evaluation unit. A spatial evaluation unit is the smallest evaluation space with defined spatial boundaries, used for independently calculating storage capacity, injection accessibility, and storage security.

[0061] For each spatial evaluation unit, the processor extracts porosity, permeability, reservoir thickness, formation pressure, fault distance, and caprock safety parameters from the multi-source reservoir dataset that match its spatial location, and records the spatial relationship between these parameters and the injection well location data. For discrete data such as well points or core samples, when they do not cover all spatial evaluation units, ordinary kriging interpolation, inverse distance weighted interpolation, or existing geological modeling results can be used to map the data to the corresponding spatial evaluation unit. The interpolation method used and its corresponding source are recorded in the multi-source reservoir dataset.

[0062] When a fault crosses multiple spatial evaluation units, the processor determines the spatial overlap ratio between the fault plane and each spatial evaluation unit, determines the spatial evaluation unit corresponding to the fault's influence based on the spatial overlap ratio, and calculates the fault distance corresponding to each spatial evaluation unit based on the spatial positional relationship between the fault plane and each spatial evaluation unit. When a caprock anomaly zone covers multiple spatial evaluation units, the processor determines the caprock safety parameters corresponding to each spatial evaluation unit based on the spatial overlap ratio between the caprock anomaly zone and each spatial evaluation unit.

[0063] Adjacent spatial evaluation units are determined by sharing a common surface, meaning that a spatial evaluation unit in a 3D regular grid has a maximum of six directly adjacent spatial evaluation units: above, below, front, back, left, and right. The spatial evaluation unit dataset records at least the spatial number, 3D spatial coordinates, reservoir attributes, and adjacent spatial numbers. Subsequent state determination and connectivity analysis are all performed around the same spatial number to avoid interruptions in data correspondence between different processing steps.

[0064] Step S30: Determine the storage space status, injection reachability status, and sealing security status of each spatial evaluation unit based on the spatial evaluation unit dataset.

[0065] like Figure 3As shown, this step does not directly determine whether a spatial assessment unit has storage capacity based on a single evaluation score. Instead, it assesses the status from three aspects: storage space, injection accessibility, and storage safety. Storage space status reflects whether the spatial assessment unit has usable pore space that meets the requirements; injection accessibility status reflects whether the spatial assessment unit can form a transferable connection with adjacent spatial assessment units based on its permeability, formation pressure, and adjacency, meeting preset transfer conditions, thus characterizing the unit's local injection transfer capability; storage safety status reflects whether the spatial assessment unit is limited by fault, caprock capacity, or formation pressure conditions. Each spatial assessment unit ultimately generates three corresponding types of status data, linked by a spatial number. These three types of status serve as the data basis for subsequent identification of storage bottleneck units.

[0066] When determining the storage space status, the porosity and reservoir thickness of each spatial evaluation unit are extracted from the spatial evaluation unit dataset. The usable pore space of a unit is then determined based on the porosity and reservoir thickness. Therefore, usable pore space describes the volumetric pores within a spatial evaluation unit capable of participating in carbon dioxide storage, and is primarily determined by the unit's planar dimensions, effective reservoir thickness, and porosity. For a spatial evaluation unit with a horizontal dimension of 100m × 100m and a vertical effective reservoir thickness of 10m, a porosity of 0.18 corresponds to a usable pore space of 18000m³. The preset minimum storage space threshold can be set according to the scale of the spatial evaluation unit and the target evaluation accuracy. It is set to 5000-20000 m³ for a spatial evaluation scale of 50-200m horizontally and 5-20m vertically. In this embodiment, when the available pore space is less than 8000 m³, it means that the storage space that the spatial evaluation unit can contribute is insufficient under the current spatial evaluation scale, and it is determined to be a storage space unsatisfactory state. Otherwise, it is a storage space satisfied state.

[0067] When determining the reachable state of injection, the permeability, formation pressure, and adjacency relationship of each spatial evaluation unit are extracted from the spatial evaluation unit dataset. Adjacent spatial evaluation units are determined based on the adjacency relationship, and the permeability and formation pressure difference of adjacent spatial evaluation units are compared. When the permeability of adjacent spatial evaluation units is not lower than the preset minimum permeability and the absolute value of the formation pressure difference is within the preset allowable pressure difference range, the corresponding adjacent spatial evaluation units are determined to be transitively connected.

[0068] A preset minimum permeability is used to exclude spatial connections with significantly insufficient fluid transport capacity. This can be determined based on the reservoir type of the target storage area, core flow test results, well logging interpretation results, or wellbore pressure test results, and is set in conjunction with the spatial distribution range of permeability within the target storage area to distinguish between adjacent spatial evaluation units that meet and do not meet carbon dioxide transport conditions. A preset allowable pressure difference range is used to constrain the initial pressure difference between adjacent spatial evaluation units. In this embodiment, the upper limit of the absolute formation pressure difference between adjacent spatial evaluation units is set to 1.5 MPa, meaning that when adjacent spatial evaluation units meet the requirements... When the corresponding adjacent spatial evaluation unit is considered to meet the preset allowable pressure difference condition, it is considered that the spatial evaluation unit is in an injection reachable state; otherwise, it is marked as an injection unreachable state.

[0069] It should be noted that the injection reachability status here is used to characterize whether the spatial evaluation unit itself possesses local propagation conditions, and is not used to determine whether the spatial evaluation unit has formed a global connectivity relationship with the spatial evaluation unit corresponding to the injection well location data through a continuous path; the global connectivity relationship and the path obstruction caused by the sealed short-board unit are determined in step S50 by performing a spatial connectivity search starting from the spatial evaluation unit corresponding to the injection well location data. Therefore, for spatial evaluation units that possess local propagation conditions but cannot form a continuous path with the injection well due to obstruction by other sealed short-board units, they are not marked as injection unreachable in this step due to the path obstruction, and step S50 further determines whether they belong to the obstruction range.

[0070] When determining the safety status of landfill, the fault distance, caprock safety parameters, and formation pressure corresponding to each spatial evaluation unit are extracted from the spatial evaluation unit dataset. Fault risk markers are generated based on the comparison between the fault distance and a preset fault safety distance. The preset fault safety distance is used to determine whether a spatial evaluation unit is excessively close to an identified fault; in this embodiment, the preset fault safety distance is set to 300–1000 m. Caprock risk markers are generated based on the comparison between the caprock safety parameters and a preset caprock safety threshold. The preset caprock safety threshold is composed of caprock thickness and caprock fracture pressure margin conditions; the lower limit of caprock thickness is set to 20–50 m, and the pressure margin between the caprock fracture pressure and the current formation pressure is set to 3–8 MPa. Pressure risk markers are generated based on the comparison between the formation pressure and a preset allowable pressure upper limit, which is set to 70%–85% of the corresponding caprock fracture pressure. When any risk marker indicates a risk status, the corresponding spatial evaluation unit is determined to be in a state where landfill safety is not met; otherwise, it is determined to be in a state where landfill safety is met. In this embodiment, when the spatial evaluation unit is less than 500m from the fault, the caprock thickness is less than 30m, the margin between the caprock rupture pressure and the current formation pressure is less than 5MPa, or the current formation pressure reaches more than 80% of the corresponding caprock rupture pressure, the corresponding risk marker enters a risk state.

[0071] Step S40: Identify space evaluation units that do not meet the preset sealing conditions based on the storage space status, injection reachability status, and sealing security status, and obtain the sealing bottleneck units, such as... Figure 4 As shown.

[0072] The preset storage conditions are not a single parameter threshold, but a combination of conditions consisting of storage space status, injection accessibility status, and storage safety status. A spatial evaluation unit satisfies the preset storage conditions when it simultaneously meets the storage space satisfaction, injection accessibility, and storage safety requirements. If at least one of these conditions is not met, the spatial evaluation unit is identified as a candidate bottleneck unit. A bottleneck unit is used to represent a spatial evaluation unit that cannot be directly used as effective storage space, or that would restrict the actual utilization of other spatial evaluation units. This definition not only focuses on whether the bottleneck unit itself can store carbon dioxide, but also provides a basis for subsequent judgment on whether it cuts off the effective connectivity between the injection well and other reservoir spaces.

[0073] like Figure 6As shown, a candidate short-board unit refers to a spatial evaluation unit that does not meet the corresponding conditions in at least one of the three states. Since subsurface reservoir anomalies usually have a certain degree of spatial continuity, the processor further aggregates spatially continuous candidate short-board units based on adjacency relationships: coplanar adjacency is used as the merging condition; two candidate short-board units are considered continuous if they share a single grid surface; connected component search is used to classify all continuous candidate short-board units into the same sealed short-board region. All spatial evaluation units constituting the sealed short-board region are identified as sealed short-board units, and the spatial extent of the sealed short-board region is recorded. The outer boundary of this spatial extent is directly used as the blocking boundary. The blocking boundary is not artificially added independent data, but is composed of the outer boundaries of each sealed short-board unit in the sealed short-board region, and it is directly used for path search in step S50. Therefore, this process can convert discrete state judgment results into blocking regions with clear spatial extents, providing a basis for subsequent analysis of whether sealed short-board units further affect other qualified spatial evaluation units.

[0074] This embodiment also includes the steps of obtaining feature importance through machine learning, adjusting state judgment conditions, and re-identifying candidate bottleneck units.

[0075] Specifically, historical storage evaluation samples refer to data samples with reservoir attribute records and actual storage utilization results or verified evaluation results. Each historical sample corresponds to a defined porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock thickness, caprock fracture pressure, and corresponding storage availability status. Historical spatial samples that have reached the predetermined injection capacity, have not exceeded the preset pressure safety conditions during the monitoring period, and have not shown any identified leakage anomalies are marked as usable, with a training flag of 1; the remaining samples are marked as unusable, with a training flag of 0.

[0076] The machine learning uses a random forest classification model, which consists of multiple classification and regression trees. Each classification and regression tree receives the same type of input features and is trained using different bootstrap sampling training samples. Finally, the system outputs the sealed usable state based on the voting results and statistically assigns the importance of each feature to the degree of impurity reduction of the node.

[0077] In a feasible training method, at least 1000 labeled historical samples are obtained, with 70% used as the training set, 15% as the validation set, and 15% as the test set. A random forest is set up with 200 classification and regression trees, with a maximum depth of 12 for each tree, a minimum number of samples per leaf node, and the number of candidate features for each node is the square root of the number of input features. The node splitting index is Gini impurity, and training subsets corresponding to each classification and regression tree are established by random sampling with replacement.

[0078] During training, the processor sequentially performs data missingness handling, input feature normalization checks, random forest training, validation set parameter checks, and test set output validation. After training, it saves the tree node splitting parameters and the feature importance corresponding to each input feature. Random forests are tree models, therefore they do not involve the input layer, hidden layer, and output layer connection structures of neural networks.

[0079] In practical applications, the saved model and feature importance can be read and used to adjust the parameters of the current target archive area.

[0080] Regarding the specific relationship between feature importance and threshold, input features with a normalized feature importance greater than 0.20 can be identified as high-importance input features, and the corresponding conditions can be adjusted to be more stringent by 5% to 10%. When the feature importance is between 0.10 and 0.20, the original conditions are maintained. For safety-related parameters, existing safety conditions are not reduced based on low feature importance. The three states are redefined according to the adjusted conditions, and candidate weak link units are identified based on the redefined three states.

[0081] Step S50: Establish spatial connectivity based on injection well location data, spatial evaluation unit dataset, and injection reachability status; and determine the obstruction impact range based on the sealing bottleneck unit, such as... Figure 5 As shown.

[0082] In practical applications, this step is a crucial processing step that distinguishes this solution from evaluation methods that only deduct the capacity of the non-compliant space itself.

[0083] like Figure 7 As shown, spatial connectivity is based on the adjacency relationships between spatial evaluation units, while also introducing the concept of injection reachability. Only when two adjacent spatial evaluation units meet the specified permeability and formation pressure conditions—that is, when a transitive connection is formed—is an effective connection established between them for determining carbon dioxide injection propagation. Each spatial evaluation unit constitutes a node in the graph structure, and edges are established between adjacent spatial evaluation units that satisfy the transitive connection requirement.

[0084] The reachable unit set is established using a graph search method. Each spatial evaluation unit constitutes a node in the graph structure; edges are established between two spatial evaluation units that satisfy both adjacency and the propagation conditions corresponding to the injection reachable state. The processor uses the spatial evaluation unit corresponding to the injection well location data as the starting point and employs a breadth-first search to visit spatial evaluation units connected to the starting point layer by layer along the edges. Visited spatial evaluation units are written into the reachable unit set. When a bottleneck unit is encountered, it is set as an impassable node, and the search along that expansion direction is terminated.

[0085] Subsequently, spatial evaluation units located on the side of the sealing short-board unit away from the starting point, with storage space status and sealing safety status both satisfied, are identified as the units to be judged as obstructed units. For each unit to be judged as obstructed, the processor performs path search again, but deletes all sealing short-board units from the search graph: if a continuous path from the starting point to the unit to be judged as obstructed can still be found, then this path is the alternative path, and the corresponding unit to be judged as obstructed is not included in the obstruction influence range; if a continuous path cannot be obtained, then the corresponding unit to be judged as obstructed is included in the obstruction influence range. The obstruction influence range is not defined based on the distance or relative orientation to the injection well, but is determined by the actual spatial connectivity under the condition of the existence of the sealing short-board unit.

[0086] It should be noted that spatial evaluation units that have been identified as sealing-off bottleneck units will not be included in the obstruction impact range again, thus making the sealing-off bottleneck units and the obstruction impact range correspond to different sets of spatial evaluation units.

[0087] In one variant implementation, the extended search can also be implemented using depth-first search or shortest path search. The essence of both is to traverse connected nodes from the starting point outward along transitive connections. Those skilled in the art can choose according to the data scale and real-time requirements, and both are within the protection scope of this invention.

[0088] In another variant implementation, when multiple injection wells are arranged in the target storage area, a set of reachable units is constructed with the spatial evaluation unit corresponding to each injection well as the starting point. The union of the sets of reachable units is taken as the final set of reachable units, and then the scope of obstruction is determined. Accordingly, as long as the unit to be determined as obstructed can be connected to any injection well, it is considered that there is an alternative path.

[0089] Step S60: Calculate the theoretical storage capacity of each spatial evaluation unit based on the spatial evaluation unit dataset and summarize the theoretical storage capacity. Correct the theoretical storage capacity based on the storage bottleneck unit and the range of obstruction to obtain the effective storage potential and the corresponding spatial distribution results.

[0090] After obtaining the storage bottleneck unit and the extent of obstruction, the processor first calculates the theoretical pore space based on the area of ​​the spatial evaluation unit, the effective reservoir thickness, and the porosity: In the formula, The theoretical pore space corresponding to the spatial evaluation unit; The length of the spatial evaluation unit in the first horizontal direction is obtained from the preset spatial resolution; The length of the spatial evaluation unit in the second horizontal direction is obtained from the preset spatial resolution; The effective reservoir thickness within the spatial evaluation unit is determined by the reservoir thickness. The porosity corresponding to this spatial evaluation unit is extracted from the spatial evaluation unit dataset.

[0091] If it is necessary to express the theoretical unit storage capacity using mass units, it can be calculated based on the theoretical pore space, combined with the density under carbon dioxide reservoir conditions and a pre-determined storage efficiency coefficient. Specifically, the carbon dioxide density can be obtained from a property table or equation of state based on the target reservoir temperature and pressure, and the storage efficiency coefficient can be pre-set according to the target reservoir type.

[0092] When it is necessary to express the theoretical unit storage capacity in the form of output carbon dioxide mass, according to To perform the conversion, in the formula, This represents the theoretical unit storage capacity corresponding to the spatial evaluation unit; The aforementioned theoretical pore space is obtained based on the spatial scale, reservoir thickness, and porosity of the spatial evaluation unit; The density of carbon dioxide under the target reservoir temperature and formation pressure conditions is obtained based on the measured temperature of the target reservoir, formation pressure, and carbon dioxide physical property database. The preset storage efficiency coefficient is set according to the target reservoir type. For sandstone saline aquifers, the preset storage efficiency coefficient can be set to 0.01 to 0.06. The storage efficiency coefficient is a capacity conversion parameter and is not involved in determining the storage space status, injection reachability, or storage safety status.

[0093] After the processor aggregates the theoretical storage capacity of all theoretical units to obtain the theoretical storage capacity, it determines the storage bottleneck unit and the corresponding spatial evaluation unit to be deducted based on the spatial number of each spatial evaluation unit. It then deduplicates the spatial evaluation units to be deducted, ensuring that the theoretical unit storage capacity for each spatial evaluation unit is deducted only once. Subsequently, the deduplicated theoretical unit storage capacity corresponding to the spatial evaluation units to be deducted is subtracted from the theoretical storage capacity to obtain the effective storage potential. Through this process, the effective storage potential not only excludes reservoir spaces whose geological properties do not meet the requirements, but also excludes reservoir spaces that are practically inaccessible from the current injection well due to spatial obstruction.

[0094] The capacity status markers are set to three types: the first capacity status indicates that the theoretical unit's sealed capacity is retained; the second capacity status indicates that the corresponding spatial evaluation unit belongs to the sealed short-board unit and is deducted; the third capacity status indicates that the corresponding spatial evaluation unit belongs to the obstructed influence range and is deducted. The processor remaps the capacity status markers to the three-dimensional mesh according to the spatial number to generate the spatial distribution result, which can simultaneously represent the location of usable space, sealed short-board units, and obstructed influence range.

[0095] This invention can be applied to engineering scenarios such as suitability evaluation of carbon dioxide geological storage sites, storage potential calculation, candidate storage area screening, and injection well location evaluation. It can improve the consistency between storage potential evaluation results and actual reservoir utilization conditions, and has clear industrial applicability.

[0096] This embodiment also provides a carbon dioxide geological storage evaluation system, including a memory and a processor. The memory is used to store multi-source reservoir datasets and spatial evaluation unit datasets. The processor is coupled to the memory and configured to execute the method described in Embodiment 1.

[0097] This embodiment also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in Embodiment 1. The storage medium may be a read-only memory, a random access memory, a magnetic disk, or an optical disk, etc.

[0098] As described above, although the invention has been shown and described with reference to the specified preferred embodiments, it should not be construed as limiting the invention itself. Various changes in form and detail may be made without departing from the spirit and scope of the invention as defined in the appended claims.

Claims

1. A method for evaluating carbon dioxide geological sequestration based on multi-source data fusion, characterized in that, The process includes the following steps: Step S10: Obtain porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters (composed of caprock thickness and caprock fracture pressure), and injection well location data for the target storage area; perform spatial registration and scale unification on the data to generate a multi-source reservoir dataset, where each reservoir data point in the multi-source reservoir dataset is associated with a corresponding spatial location; Step S20: Divide the target storage area into multiple spatial evaluation units based on the multi-source reservoir dataset; extract the reservoir attributes corresponding to each spatial evaluation unit; and generate a spatial evaluation unit dataset that records the spatial number and adjacency relationship of each spatial evaluation unit; Step S30: Determine the storage space status, injection reachability status, and storage safety status of each spatial evaluation unit based on the spatial evaluation unit dataset. Step S40: Identify spatial evaluation units whose storage space status, injection reachability status, and sealing security status do not meet the corresponding preset conditions as sealing bottleneck units; Step S50: Starting from the spatial evaluation unit corresponding to the injection well location data, perform an extended search among the spatial evaluation units according to the adjacency relationship in the spatial evaluation unit dataset and the injection reachability status to construct a set of reachable units; When the extended search reaches the sealing short-board unit, the search along the extended direction is terminated; the space evaluation unit located on the side of the sealing short-board unit away from the starting point, and whose storage space state and sealing security state are both satisfied, is determined as the blocked unit to be judged; Determine whether each of the obstructed units to be determined has an alternative path that bypasses the sealing bottleneck unit and connects to the starting point. Determine the obstructed units to be determined without the alternative path as the obstruction influence range. Step S60: Calculate the theoretical unit sealing capacity of each spatial evaluation unit based on the spatial evaluation unit dataset and summarize it to obtain the theoretical sealing capacity. Subtract the theoretical unit sealing capacity corresponding to the sealing bottleneck unit and the spatial evaluation units within the obstruction influence range from the theoretical sealing capacity to obtain the effective sealing potential and generate the corresponding spatial distribution results.

2. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 1, characterized in that, The specific method for determining the storage space status, injection reachability status, and sealing security status of each spatial evaluation unit based on the spatial evaluation unit dataset in step S30 is as follows: S31. Extract the porosity and reservoir thickness corresponding to each spatial evaluation unit from the spatial evaluation unit dataset. Determine the usable pore space of the corresponding spatial evaluation unit based on the porosity and reservoir thickness. Compare the usable pore space with a preset minimum storage space threshold. If the usable pore space is not lower than the preset minimum storage space threshold, mark it as a storage space satisfied state; otherwise, mark it as a storage space unsatisfied state, thus forming the storage space state. S32. Extract the permeability, formation pressure, and adjacency relationship corresponding to each spatial evaluation unit from the spatial evaluation unit dataset. Determine adjacent spatial evaluation units based on the adjacency relationship. When the permeability of adjacent spatial evaluation units is not lower than the preset minimum permeability and... When the absolute value of the difference between the two formation pressures is within the preset allowable pressure difference range, the adjacent spatial evaluation units are determined to be transitively connected; spatial evaluation units that form the transitively connected with at least one adjacent spatial evaluation unit are marked as injection reachable, otherwise they are marked as injection unreachable, thus forming the injection reachable state; S33, extract the fault distance, caprock safety parameters and formation pressure corresponding to each spatial evaluation unit from the spatial evaluation unit dataset, generate a fault risk mark based on the comparison result of the fault distance and the preset fault safety distance, generate a caprock risk mark based on the comparison result of the caprock safety parameters and the preset caprock safety threshold, and generate a pressure risk mark based on the comparison result of the formation pressure and the preset allowable pressure upper limit; When any risk marker represents a risk state, the corresponding spatial evaluation unit is determined to be in a state where the storage security is not satisfied; otherwise, it is determined to be in a state where the storage security is satisfied, thus forming the storage security state.

3. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 1, characterized in that, In step S40, when a short-board unit is identified, firstly, spatial evaluation units whose at least one of the three states does not meet the corresponding preset conditions are determined as candidate short-board units; then, spatially continuous candidate short-board units are merged according to the adjacency relationship of the candidate short-board units to obtain the short-board region; finally, all spatial evaluation units constituting the short-board region are determined as the short-board units. In step S50, the blocking boundary during the extended search is the outer boundary of the sealed short plate region.

4. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 2, characterized in that, The extended search in step S50 specifically involves a graph search that uses the starting point as the root node and visits adjacent spatial evaluation units layer by layer along the transitive connections, with the visited spatial evaluation units written into the reachable unit set. The method for determining whether there is an alternative path for each of the blocked units to be determined includes: after removing all the blocked units from the spatial evaluation units participating in the search, searching again for a continuous path from the starting point to the blocked unit to be determined; if the continuous path can be found, it is determined that there is an alternative path; otherwise, it is determined that there is no alternative path.

5. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 1, characterized in that, The method for spatial registration and scale unification of the data in step S10 to generate a multi-source reservoir dataset is as follows: S11. Obtain the data source identifier, spatial coordinates, and sampling accuracy corresponding to porosity, permeability, reservoir thickness, formation pressure, fault distance, caprock safety parameters, and injection well location data, respectively. S12. Transform each of the spatial coordinates using a unified coordinate reference system, and perform grid matching on the transformed data according to a preset spatial resolution; S13. Perform consistency comparison on similar data matched to the same grid position according to the data source identifier. When the difference between similar data exceeds a preset difference threshold, generate a difference marker and determine the fusion weight and fusion order according to the difference marker and the sampling precision. S14. According to the fusion weight and the fusion order, perform weighted merging of similar data, and write the data that has completed grid matching and weighted merging into the multi-source reservoir dataset.

6. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 1, characterized in that, The specific steps for generating the spatial evaluation unit dataset in step 20 are as follows: S21. Generate a three-dimensional grid based on the reservoir boundary of the target storage area and the preset spatial resolution, and determine each three-dimensional grid as a spatial evaluation unit; S22. Based on the spatial location of each spatial evaluation unit, read the corresponding reservoir attributes from the multi-source reservoir dataset; S23. For fault or caprock anomaly zones that span two or more spatial evaluation units, corresponding parameters are assigned according to the spatial overlap ratio between the fault and each of the spatial evaluation units. Specifically, for faults, the spatial evaluation units corresponding to the fault impact are determined based on the spatial overlap ratio, and the fault distance corresponding to each spatial evaluation unit is calculated in combination with the spatial positional relationship between the fault plane and each spatial evaluation unit. For caprock anomaly zones, the caprock safety parameters corresponding to each spatial evaluation unit are determined based on the spatial overlap ratio. S24. The spatial number, adjacency relationship and corresponding reservoir attributes of each spatial evaluation unit are associated and stored to generate the spatial evaluation unit dataset.

7. The carbon dioxide geological sequestration evaluation method based on multi-source data fusion according to claim 1, characterized in that, The specific steps for generating the corresponding spatial distribution results in step S60 are as follows: S61. Calculate the theoretical storage capacity of each spatial evaluation unit based on the porosity, reservoir thickness and spatial scale of each spatial evaluation unit in the spatial evaluation unit dataset, and summarize the theoretical storage capacity. S62. Perform deduplication processing on the storage short-board unit and the space evaluation unit within the obstruction influence range, and subtract the theoretical unit storage capacity corresponding to the deduplicated space evaluation unit from the theoretical storage capacity to obtain the effective storage potential. S63. Generate a capacity status marker according to whether each spatial evaluation unit belongs to the sealed short-board unit, whether it is located in the obstructed influence range, and whether the corresponding theoretical unit sealed capacity is retained. Associate the capacity status marker with the spatial number of each spatial evaluation unit to generate the spatial distribution result.

8. The method for evaluating carbon dioxide geological sequestration based on multi-source data fusion according to any one of claims 1 to 7, characterized in that, Before performing step S30, the method further includes steps of obtaining feature importance through machine learning, adjusting state judgment conditions, and re-identifying candidate bottleneck units, specifically: Porosity, permeability, reservoir thickness, formation pressure, fault distance, and caprock safety parameters are extracted from historical storage evaluation samples as input features, and the storage availability status corresponding to the historical samples is used as training labels. Machine learning is used to train the correspondence between the input features and the training labels to obtain the feature importance corresponding to each input feature; Based on the importance of the aforementioned features, at least one of the following can be adjusted: the preset minimum storage space threshold, the preset minimum permeability, the preset allowable pressure difference, the preset fault safety distance, the preset caprock safety threshold, and the preset allowable pressure upper limit.

9. A carbon dioxide geological sequestration evaluation system based on multi-source data fusion, characterized in that, include: The memory is used to store multi-source reservoir datasets and spatial evaluation unit datasets. And a processor, coupled to the memory, the processor being configured to perform the method of any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method described in any one of claims 1 to 8.

Images