CCUS case comparison and selection method based on dynamic modulus length fusion cosine similarity
Through the dynamic mode-length fusion cosine similarity method and combined with spatial similarity calculation, the problem of mode-length difference and spatial distribution being ignored in the existing technology is solved, and a more accurate CCUS case comparison is achieved.
Patent Information
- Application Number
- CN202510444444.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-10
- Publication Date
- 2025-08-19
AI Technical Summary
When calculating the similarity of CCUS cases, the existing cosine similarity algorithm ignores the difference in modular length, resulting in vectors with the same direction but large difference in modular lengths being misjudged as highly similar, and does not consider the spatial distribution of CCUS cases, and lacks the capture of adaptability and nonlinear relationships.
The dynamic mode-length fusion cosine similarity method is adopted, and the mode-length influence coefficient is dynamically adjusted by assigning parameter weights and combining spatial similarity calculations, and calculating the dynamic mode-length fusion cosine similarity and spatial similarity of the case, and finally obtaining a more comprehensive case similarity.
A more accurate selection of similar CCUS cases is achieved, and the spatial distribution of CCUS cases is taken into account, which improves the accuracy of similarity measurement.
Smart Images
Figure CN120508832A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of CCUS technology, and in particular to a CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity. Background Art
[0002] Carbon dioxide capture, utilization, and storage (CCUS) is the process of separating CO2 from industrial processes, energy use, or the atmosphere, and using engineering techniques to reduce its emissions and generate ancillary benefits. CCUS is currently the only technological option for achieving low-carbon utilization of fossil energy and a key means of achieving large-scale greenhouse gas emissions reductions. It is a viable technical solution for deep decarbonization in difficult-to-abate industries such as steel, cement, and chemicals, and a key component of the technology portfolio for achieving carbon neutrality.
[0003] Carbon storage is a key link in CCUS, and the CCUS site selection issue is a key issue in carbon sequestration. Existing CCUS storage site selection research is based on existing cases to select storage indicators and construct an evaluation system. Therefore, the higher the similarity of the cases, the more reference value they have.
[0004] The cosine similarity algorithm is a common method for calculating case similarity, primarily for calculating textual similarity between cases. It works based on the vector space model and measures similarity by comparing the cosine of the angle between two vectors. In two-dimensional or three-dimensional space, vectors can be thought of as arrows emanating from the origin, and cosine similarity is the cosine of the angle between these two arrows. A larger cosine similarity indicates closer orientations between the two vectors, meaning higher similarity; a smaller cosine similarity indicates lower similarity.
[0005] The cosine similarity algorithm is used for CCUS case screening. On the one hand, it only measures vector directional similarity, ignoring differences in modulus length. Vectors with the same direction but significantly different modulus lengths are mistakenly identified as highly similar, while vectors with different directions but similar modulus lengths may actually be related but underestimated. Existing improved methods cannot dynamically adjust the weights of direction and modulus length based on data distribution, lacking adaptability. Simple weighting or penalty methods fail to capture nonlinear relationships. Furthermore, they calculate the textual similarity of cases and fail to consider the spatial similarity of CCUS cases across different spatial locations. Summary of the Invention
[0006] In view of the above problems, the present invention aims to provide a CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity.
[0007] The technical solutions of the present invention are as follows: A CCUS case comparison and selection method based on dynamic modulus fusion cosine similarity includes the following steps: S1: Collect basic data on existing CCUS cases; S2: selecting parameters that can affect CCUS from the basic data as case comparison parameters, and assigning weights to the comparison parameters; selecting parameters that can determine the spatial distribution of CCUS cases from the basic data as case distribution parameters; S3: Calculate the dynamic modulus fusion cosine similarity of each CCUS case based on the case comparison parameters and the weight assigned to each case comparison parameter, and calculate the spatial similarity of each CCUS case based on the case distribution parameters; S4: Calculating case similarity based on the dynamic modulus fusion cosine similarity and the spatial similarity; S5: According to the case similarity, combined with the case similarity output threshold, a similar case comparison result is obtained.
[0008] Preferably, in step S1, the basic data includes case number, case name, longitude, latitude, execution country, case location, closure area, reservoir thickness, reservoir porosity, reservoir permeability, cap rock thickness, cap rock permeability, formation water pressure, seismic peak acceleration, population density, and carbon source distance.
[0009] Preferably, in step S2, the case comparison parameters include trap area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness, and the case distribution parameters include longitude and latitude.
[0010] Preferably, in step S2, each comparison parameter is assigned a weight based on an entropy weight method, and the calculation formula is as follows: (1) Where: Indicates the weight of the comparison parameter; represents the information entropy of the data set; n represents the number of comparison parameters.
[0011] Preferably, the weights of the trapped area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness are 0.1052, 0.0876, 0.7060, 0.0529 and 0.0484 respectively.
[0012] Preferably, in step S3, the dynamic modulus length fusion cosine similarity is calculated by the following formula: (2) (3) Where: Represents the dynamic modulus fusion cosine similarity of vector A and vector B; A represents the vector composed of case A and selected parameter data; B represents the vector composed of case B and selected parameter data; Indicates the dynamic adjustment of the modulus length influence coefficient; Indicates the modulus of A vector; Indicates the modulus of the B vector; Indicates the standard deviation of the difference in vector modulus in the dataset; Represents the standard deviation of the traditional cosine similarity in the dataset.
[0013] Preferably, in step S3, the spatial similarity is calculated using the following formula: (4) Where: represents the case space similarity; represents the spatial distance between the case number i and the input case; is the Gaussian parameter; m is the number of cases in the data set.
[0014] Preferably, step S4 specifically includes the following sub-steps: weighting the dynamic modulus length fusion cosine similarity and the spatial similarity respectively, and then performing weighted calculation to obtain the case similarity.
[0015] Preferably, when assigning weights, geological experts are first used to score the dynamic modulus fusion cosine similarity and the spatial similarity respectively, and then the scores are introduced into the hierarchical analysis method to calculate the weights.
[0016] Preferably, in step S5, the case similarity output threshold is 0.5.
[0017] The beneficial effects of the present invention are: The present invention integrates the modulus information into the directional similarity calculation to achieve a more comprehensive similarity measurement. At the same time, the spatial distribution of CCUS cases is considered and spatial similarity is introduced, so that more similar CCUS cases can be selected. BRIEF DESCRIPTION OF THE DRAWINGS
[0018] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0019] Figure 1 Schematic diagram of the process of CCUS case comparison and selection method based on dynamic modulus fusion cosine similarity of the present invention. DETAILED DESCRIPTION
[0020] The present invention is further described below with reference to the accompanying drawings and examples. It should be noted that, in the absence of conflict, the embodiments in this application and the technical features in the embodiments can be combined with each other. It should be noted that, unless otherwise specified, all technical and scientific terms used in this application have the same meanings as those commonly understood by those of ordinary skill in the art to which this application belongs. The use of similar words such as "include" or "comprising" in the present invention means that the elements or objects appearing before the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects.
[0021] like Figure 1 As shown, the present invention provides a CCUS case comparison method based on dynamic modulus fusion cosine similarity, comprising the following steps: S1: Collect basic data on existing CCUS cases.
[0022] In a specific embodiment, the basic data includes case number, case name, longitude, latitude, execution country, case location, closure area, reservoir thickness, reservoir porosity, reservoir permeability, cap rock thickness, cap rock permeability, formation water pressure, seismic peak acceleration, population density, and carbon source distance.
[0023] S2: Selecting parameters that can affect CCUS from the basic data as case comparison parameters, and assigning weights to each comparison parameter; selecting parameters that can determine the spatial distribution of CCUS cases from the basic data as case distribution parameters.
[0024] In a specific embodiment, the case comparison parameters include trap area, reservoir thickness, reservoir porosity, reservoir permeability and caprock thickness, and the case distribution parameters include longitude and latitude.
[0025] In the above embodiment, the trapped area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness are the most influential CCUS assessment parameters in existing CCUS assessment case studies, and therefore are adopted as the case comparison parameters.
[0026] In a specific embodiment, each comparison parameter is assigned a weight based on an entropy weight method, and the calculation formula is as follows: (1) Where: Indicates the weight of the comparison parameter; represents the information entropy of the data set; n represents the number of comparison parameters.
[0027] Optionally, the weights of the trapped area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness are 0.1052, 0.0876, 0.7060, 0.0529 and 0.0484 respectively.
[0028] It should be noted that the entropy weight method in the above embodiment is only a preferred weight assignment method of the present invention, and other weight assignment methods in the prior art may also be applicable to the present invention.
[0029] S3: Calculate the dynamic modulus fusion cosine similarity of each CCUS case according to the case comparison parameters and the weight assigned to each case comparison parameter, and calculate the spatial similarity of each CCUS case according to the case distribution parameters.
[0030] In a specific embodiment, the dynamic modulus fusion cosine similarity is calculated by the following formula: (2) (3) Where: Represents the dynamic modulus fusion cosine similarity of vector A and vector B; A represents the vector composed of case A and selected parameter data; B represents the vector composed of case B and selected parameter data; Indicates the dynamic adjustment of the modulus length influence coefficient; Indicates the modulus of A vector; Indicates the modulus of the B vector; Indicates the standard deviation of the difference in vector modulus in the dataset; Represents the standard deviation of the traditional cosine similarity in the dataset.
[0031] In the above embodiment, when the present invention performs dynamic modulus fusion cosine similarity calculation, the original vector is mapped to a new space through dynamic nonlinear transformation and adaptive parameter mechanism so that its direction implicitly contains modulus information. The dynamic nonlinear transformation formula of the vector is as follows: (5) (6) Where: A', B' respectively represent the new vectors of A and B after dynamic nonlinear transformation; Indicates the modulus of the A' vector; Indicates the modulus of the B' vector Then calculate the dynamic modulus fusion cosine similarity: (7) Where: Indicates the modulus of the A' vector; Indicates the modulus of the B' vector; Substituting equations (5)-(6) into equation (7), we can obtain the dynamic modulus length fusion cosine similarity calculation formula shown in equation (2).
[0032] In the above embodiment, the dynamic adjustment of the modulus influence coefficient The range is between 0 and 1. If the module length difference fluctuates greatly ( higher value), then Increase, strengthen the influence of modulus length. If the direction similarity fluctuates significantly ( higher value), then Reduce, retain directional sensitivity.
[0033] The spatial similarity is calculated by the following formula: (4) Where: represents the case space similarity; represents the spatial distance between the case number i and the input case; is the Gaussian parameter; m is the number of cases in the data set.
[0034] In the above embodiment, when calculating the spatial similarity, the distance d between the selected case and all other cases is calculated based on the longitude and latitude of each CCUS case. Then, a Gaussian function is used to calculate the distance weight, and the weight is normalized to a value in the range [0, 1] to obtain the spatial similarity.
[0035] S4: Calculate the case similarity based on the dynamic modulus fusion cosine similarity and the spatial similarity.
[0036] In a specific embodiment, step S4 specifically includes the following sub-steps: weighting the dynamic modulus fusion cosine similarity and the spatial similarity respectively, and then performing weighted calculation to obtain the case similarity.
[0037] Optionally, when assigning weights, geological experts are first used to score the dynamic modulus fusion cosine similarity and the spatial similarity respectively, and then the scores are introduced into the hierarchical analysis method to calculate and obtain weights.
[0038] It should be noted that the weighting method in the above embodiment is only a preferred method of the present invention, and other weighting methods in the prior art may also be applicable to the present invention.
[0039] S5: According to the case similarity, combined with the case similarity output threshold, a similar case comparison result is obtained.
[0040] In a specific embodiment, the case similarity output threshold is 0.5. It should be noted that the case similarity output threshold of this embodiment is only a specific threshold preferred by the present invention. When using the present invention, other case similarity output thresholds can be used as needed. The larger the case similarity output threshold, the higher the output case similarity.
[0041] In a specific embodiment, CCUS cases were selected using the CCUS case comparison method based on dynamic modulus fusion and cosine similarity described in the present invention. In this example, 136 existing CCUS cases were collected based on literature and public data. The case selection parameters in this example included trap area, reservoir thickness, reservoir porosity, reservoir permeability, and caprock thickness, with weights of 0.1052, 0.0876, 0.7060, 0.0529, and 0.0484, respectively. The comparison resulted in 46 similar cases.
[0042] Furthermore, a CCUS case selection method using traditional cosine similarity, without considering spatial similarity, compared the 136 existing CCUS cases in the above embodiment and found 71 similar cases. This method considers module length information and combines the spatial similarity of the spatial distribution of each CCUS case, resulting in a more accurate output of similar cases.
[0043] The above description is merely a representative embodiment of the present invention and does not constitute any form of limitation to the present invention. Any technical personnel familiar with the present invention who, without departing from the scope of the technical solution of the present invention, makes some changes or modifications to the embodiments disclosed above using the technical contents disclosed above are equivalent embodiments of the present invention. However, any simple modifications, equivalent changes and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.
Claims
1. A CCUS case comparison method based on dynamic modulus fusion cosine similarity, characterized by: The following steps are involved: S1: Collect basic data on existing CCUS cases; S2: selecting parameters that can affect CCUS from the basic data as case comparison parameters, and assigning weights to the comparison parameters; selecting parameters that can determine the spatial distribution of CCUS cases from the basic data as case distribution parameters; S3: Calculate the dynamic modulus fusion cosine similarity of each CCUS case based on the case comparison parameters and the weight assigned to each case comparison parameter, and calculate the spatial similarity of each CCUS case based on the case distribution parameters; S4: Calculating case similarity based on the dynamic modulus fusion cosine similarity and the spatial similarity; S5: According to the case similarity, combined with the case similarity output threshold, a similar case comparison result is obtained.
2. The CCUS case comparison method based on dynamic modulus fusion and cosine similarity according to claim 1 is characterized in that: In step S1, the basic data includes case number, case name, longitude, latitude, execution country, case location, trap area, reservoir thickness, reservoir porosity, reservoir permeability, cap rock thickness, cap rock permeability, formation water pressure, seismic peak acceleration, population density, and carbon source distance.
3. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 2 is characterized in that: In step S2, the case comparison parameters include the trap area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness, and the case distribution parameters include longitude and latitude.
4. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 3 is characterized in that: In step S2, each comparison parameter is assigned a weight based on the entropy weight method, and the calculation formula is as follows: Where: W i Indicates the weight of the comparison parameter; e i represents the information entropy of the data set; n represents the number of comparison parameters.
5. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 4 is characterized in that: The weights of the trap area, reservoir thickness, reservoir porosity, reservoir permeability and cap rock thickness are 0.1052, 0.0876, 0.7060, 0.0529 and 0.0484 respectively.
6. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 1 is characterized in that: In step S3, the dynamic modulus fusion cosine similarity is calculated using the following formula: Where: C(A,B) represents the dynamic modulus fusion cosine similarity of vector A and vector B; A represents the vector composed of case A and selected parameter data; B represents the vector composed of case B and selected parameter data; α represents the dynamic adjustment modulus influence coefficient; ‖A‖ represents the modulus length of vector A; ‖B‖ represents the modulus length of vector B; std(lenth_diff) represents the standard deviation of the vector modulus length differences in the dataset; std(cos_sim) represents the standard deviation of the traditional cosine similarity in the dataset.
7. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 1 is characterized in that: In step S3, the spatial similarity is calculated using the following formula: Where: G represents the case space similarity; di represents the spatial distance between the case with case number i and the input case; σ is the Gaussian parameter; m represents the number of cases in the dataset.
8. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to any one of claims 1 to 7, characterized in that: Step S4 specifically includes the following sub-steps: weighting the dynamic modulus length fusion cosine similarity and the spatial similarity respectively, and then performing weighted calculation to obtain the case similarity.
9. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 8 is characterized in that: When assigning weights, geological experts are first used to score the dynamic modulus fusion cosine similarity and the spatial similarity respectively, and then the scores are introduced into the hierarchical analysis method to calculate the weights.
10. The CCUS case comparison and selection method based on dynamic modulus fusion and cosine similarity according to claim 1 is characterized in that: In step S5, the case similarity output threshold is 0.5.