Method and system for simulating carbon dioxide diffusion in cemented filling space with porosity evolution

Abstract

The invention discloses a porosity evolution cemented filling space carbon dioxide diffusion simulation method and system, and the simulation method comprises the following steps: calculating a proper physical similarity ratio, building a physical model having a certain similarity ratio with an engineering filling body, and carrying out a cemented filling carbon sequestration physical test; in the gradual solidification and carbon sequestration process of the cemented filling body, CT is used for scanning the cemented filling body in real time, a filling body digital model is established, then a pore voxel model is extracted, and spatial distribution characteristics and time sequence evolution characteristics of pores are analyzed; a voxel automatic encoder is used to invert the distribution characteristics of the pore structure in the physical similar model to an engineering scale, and sampling is carried out on an engineering site to verify the reliability of an inversion result; and CO2 flow diffusion simulation in the cemented filling space: establishing a simulation model, carrying out pretreatment and initial porosity setting, and carrying out CO2 diffusion simulation in the cemented filling space.

Description

Technical Field

[0001] The present invention relates to the field of geological storage of carbon dioxide, and in particular to a method and system for simulating carbon dioxide diffusion in a cemented filling space with porosity evolution. Background Art

[0002] Cemented filling technology is a green and emission-reducing mining method in coal mines. This technology mixes coal-based solid wastes such as crushed gangue and fly ash and fills them into the goaf of coal mines. It can not only reduce the emission of coal-based solid wastes, but also reduce the subsidence of the surface and protect the ecological environment of the mining area. It has been successfully applied in more than ten mining areas across the country.

[0003] Large-scale CO2 emissions will lead to the greenhouse effect. CO2 geological storage is an effective measure to reduce carbon emissions. Injecting CO2 into a gradually solidified cementing filling body can make CO2 react with substances such as calcium hydroxide and magnesium hydroxide in the cementing material, and then convert it into carbide, thereby sealing the CO2 in the underground cementing filling body. This method of geological storage through chemical reaction has high safety, avoids the re-escape of CO2, and effectively reduces the CO2 content in the atmosphere.

[0004] The mineralization and carbon fixation effect of cemented filling mining sites is closely related to the migration and diffusion of CO2 gas phase in the cemented filling space. The CO2 mineralization reaction of coal-based solid waste cemented filling bodies will significantly change the microstructure of the filling body, thereby affecting the porosity of the filling body, and ultimately affecting the range of CO2 gas phase migration and diffusion. The coal-based solid waste mineralization reaction, the evolution of the filling body porosity, and the migration and diffusion of CO2 in the filling space influence and restrict each other. Therefore, it is of great significance to analyze the spatiotemporal evolution characteristics of the porosity of the coal-based solid waste cemented filling body during the whole period of CO2 mineralization, and consider the CO2 migration and diffusion law of the cemented filling space with the spatiotemporal evolution of porosity for the carbon fixation effect.

[0005] In addition, the main methods for laboratory determination of the pore and fracture structure of cemented filling bodies include mercury injection, centrifuge, cast thin section, nuclear magnetic resonance and CT scanning. Among them, the digital core model obtained by CT scanning is one of the most intuitive methods to show the internal pore throat structure of the cemented filling body. Moreover, the cemented filling body is in a paste state during the solidification process, and it is not suitable to use other methods to test the internal voids. However, due to the heterogeneity of the material and structure inside the cemented filling body, the structure of the digital core model is very complex, which will bring about problems such as high simulation cost, long time consumption, and difficulty in convergence. Extracting and equating pores and throats from the digital core model and establishing a pore network model can simplify the difficulty of calculation while ensuring the calculation results. Summary of the invention

[0006] In response to the problems and needs raised above, this solution proposes a method for simulating the diffusion of carbon dioxide in cemented filling spaces with porosity evolution. Due to the adoption of the following technical features, it can achieve the above technical objectives and bring about many other technical effects.

[0007] The object of the present invention is to provide a method for simulating carbon dioxide diffusion in a cemented filling space with porosity evolution, comprising the following steps:

[0008] S10: Physical similarity simulation test of mineralized carbon fixation: Calculate the appropriate physical similarity ratio, establish a physical model with a certain similarity ratio with the engineering filling body, and conduct physical tests on cemented filling carbon fixation;

[0009] S20: Acquisition of spatiotemporal evolution of pore structure during the whole mineralization cycle: During the process of gradual solidification and carbon fixation of cemented filling bodies, CT is used to scan the cemented filling bodies in real time, a digital model of the filling bodies is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed;

[0010] S30: Engineering scale inversion of pore structure: Use voxel autoencoders to invert the distribution characteristics of pore structure in physically similar models to the engineering scale, and take samples at the engineering site to verify the reliability of the inversion results;

[0011] S40: Simulation of CO2 flow and diffusion in cemented filling space: Establish a simulation model, perform pretreatment and initial porosity setting, and then simulate the CO2 diffusion in the cemented filling space.

[0012] In one example of the present invention, in step S10, a suitable physical similarity ratio K is calculated by comprehensively considering the physical similarity model size L0, the scanning size limit L1 of the CT device and the engineering dimension L2 of the filling body, and the size relationship satisfies L2=KL0, L1≥L0.

[0013] In an example of the present invention, step S20 specifically includes the following steps:

[0014] Use high-precision CT equipment to scan the cemented filling body in the carbon fixation process in real time, and then perform noise reduction processing on the CT images and convert them into voxels;

[0015] The Otsu threshold segmentation method is used to maximize the inter-class variance of the pore area and the solid area to obtain the optimal threshold T, and then the pore area is extracted from the voxel image;

[0016] A three-dimensional pore voxel model of the filling body is constructed, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed based on the pore voxels.

[0017] In one example of the present invention, the inter-class variance of the pore region and the solid region The expression is:

[0018]

[0019] Among them, ω1 and ω2 are the weights of the two categories after segmentation, μ1 and μ2 are the means of the two categories, and T is the optimal threshold.

[0020] In an example of the present invention, in step S20, analyzing the spatial distribution characteristics and temporal evolution characteristics of the pores includes:

[0021] The position distribution function f of the pores in the physical similarity model is calculated based on the position, number, volume, shape and fractal dimension of the pores. A (x,y,z), quantity distribution function f N (N), volume distribution function f V (V), shape distribution function f Ψ (Ψ) and the pore distribution pattern function f D (D), constructing the spatial statistical model of pore distribution g(f A ,f N ,f V ,f Ψ ,f D );

[0022] In an example of the present invention, in step S20, analyzing the temporal evolution characteristics of the pores includes:

[0023] Analyze the pore distribution characteristics at each time point and obtain the position distribution time series function f At (x,y,z,t), quantity distribution time series function f Nt (N,t), volume distribution time series function f Vt (V, t), shape distribution time series function f Ψt (Ψ,t) and the pore distribution pattern time series function f Dt (D, t), connect the spatial statistical models of pore distribution at different time points to form the spatiotemporal evolution model of pore distribution g t (f At ,f Nt ,f Vt ,f Ψt ,f Dt ).

[0024] In an example of the present invention, step S30 includes the following steps:

[0025] First, input voxel information g input Mapped to low-dimensional information h, where the expression of status information h is: h = f encoder (g input )=σ(wencoder g input +b encoder ), where σ is the activation function, w encoder is the weight, b encoder is bias;

[0026] After capturing the key features of the pore structure, the low-dimensional features h are decoded back to voxel information g of the same dimension as the input output , where the expression of voxel information is: g output =f decoder (h)=σ(w decoder h+b decoder );

[0027] Minimize the loss function during voxel autoencoder training This can optimize the pore inversion results.

[0028] In one example of the present invention, in step S40, the CO2 diffusion simulation of the cementing filling space specifically includes the following steps:

[0029] The numerical model of the cemented filling engineering was established using COMSOL software. The initial porosity was set to be the same as that of the engineering, and the porosity that changes with time was set. Then, the CO2 flow and diffusion morphology and the spatial distribution law of CO2 concentration under different CO2 injection pressures, injection timings, injection durations and drilling arrangements were simulated.

[0030] Another object of the present invention is to provide a carbon dioxide diffusion simulation system for cemented filling space with porosity evolution, characterized by comprising:

[0031] A physical similarity simulation unit is configured to calculate a suitable physical similarity ratio, establish a physical model having a certain similarity ratio with the engineering filling body, and conduct a physical test of cementation filling carbon fixation;

[0032] The evolution law acquisition unit is configured to acquire the spatiotemporal evolution law of the pore structure during the whole period of mineralization: in the process of gradual solidification and carbon fixation of the cemented filling body, the cemented filling body is scanned in real time by CT, a digital model of the filling body is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed;

[0033] The pore structure inversion unit is configured for engineering scale inversion of pore structure: the distribution characteristics of pore structure in the physically similar model are inverted to the engineering scale using a voxel autoencoder, and the reliability of the inversion results is verified by sampling at the engineering site;

[0034] The flow diffusion simulation unit is configured to establish a simulation model, perform preprocessing and initial porosity setting, and then perform CO2 diffusion simulation in the cemented filling space.

[0035] In an example of the present invention, the evolution law acquisition unit includes:

[0036] An image processing module is configured to use a high-precision CT device to scan the cemented filling body in the carbon fixation process in real time, and after obtaining the CT image, perform noise reduction processing and convert it into voxels;

[0037] The pore region extraction module is configured to use the Otsu threshold segmentation method to maximize the inter-class variance of the pore region and the solid region, obtain the optimal threshold T, and then extract the pore region from the voxel image;

[0038] The pore feature acquisition module is configured to construct a three-dimensional pore voxel model of the filling body, and analyze the spatial distribution characteristics and temporal evolution characteristics of the pores based on the pore voxels.

[0039] The outstanding beneficial effects compared with the prior art are as follows:

[0040] The present invention proposes a method for simulating CO2 diffusion in cemented filling space based on the spatiotemporal evolution of porosity. A physical similarity model is established for the engineering-scale cemented filling body through a similarity ratio, the physical similarity model in the carbon fixation process is scanned in real time, a digital filling body model is constructed, the pore throat structure characteristics are extracted, and a pore network model is established. The model is imported into the software for CO2 diffusion simulation calculation. Compared with traditional methods, the technical effects of visualization of the spatiotemporal evolution of pore characteristics during carbon fixation and efficient simulation of CO2 diffusion in cemented filling space are achieved.

[0041] The best embodiment for carrying out the present invention will be described in more detail below with reference to the accompanying drawings so that the features and advantages of the present invention can be easily understood. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings of the embodiments of the present invention, wherein the drawings are only used to illustrate some embodiments of the present invention, but not to limit all embodiments of the present invention thereto.

[0043] Figure 1 A flow chart of a method for simulating carbon dioxide diffusion in a cemented filling space according to porosity evolution of an embodiment of the present invention;

[0044] Figure 2 A CT scanning schematic diagram of the cemented filling body during the carbon fixation process according to an embodiment of the present invention;

[0045] Figure 3 4 is a diagram of a pore voxel model of the present invention according to an embodiment of the present invention.

[0046] List of reference numerals:

[0047] Radiation source 1;

[0048] Physical similarity model 2;

[0049] Gas injection tube 3;

[0050] Stage 4;

[0051] Detector 5. DETAILED DESCRIPTION

[0052] In order to make the purpose, technical solution and advantages of the technical solution of the present invention clearer, the technical solution of the embodiment of the present invention will be clearly and completely described in conjunction with the drawings of specific embodiments of the present invention. The same figure marks in the drawings represent the same parts. It should be noted that the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the described embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] Unless otherwise defined, the technical terms or scientific terms used herein shall have the usual meanings understood by persons of ordinary skill in the field to which the present invention belongs. The words "first", "second" and similar words used in the patent application specification and claims of the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "one" do not necessarily indicate a quantity limitation. Words such as "include" or "comprise" mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. Words such as "connect" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0054] According to a first aspect of the present invention, a method for simulating carbon dioxide diffusion in a cemented filling space with porosity evolution is described as follows: Figure 1 As shown, the following steps are included:

[0055] S10: Physical similarity simulation test of mineralized carbon fixation: Calculate the appropriate physical similarity ratio, establish a physical model with a certain similarity ratio with the engineering filling body, and conduct physical tests on cemented filling carbon fixation;

[0056] S20: Acquisition of spatiotemporal evolution of pore structure during the whole mineralization cycle: During the process of gradual solidification and carbon fixation of cemented filling bodies, CT is used to scan the cemented filling bodies in real time, a digital model of the filling bodies is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed;

[0057] S30: Engineering scale inversion of pore structure: Use voxel autoencoders to invert the distribution characteristics of pore structure in physically similar models to the engineering scale, and take samples at the engineering site to verify the reliability of the inversion results;

[0058] S40: Simulation of CO2 flow and diffusion in cemented filling space: Establish a simulation model, perform pretreatment and initial porosity setting, and then simulate the CO2 diffusion in the cemented filling space.

[0059] This simulation method establishes a physical similarity model of the engineering-scale cemented filling body through similarity ratio, scans the physical similarity model in the carbon fixation process in real time, constructs a digital filling body model, extracts pore throat structure characteristics, establishes a pore network model, and imports it into the software for CO2 diffusion simulation calculation. Compared with traditional methods, it achieves the technical effect of visualization of the spatiotemporal evolution of pore characteristics during carbon fixation and efficient simulation of CO2 diffusion in the cemented filling space.

[0060] In one example of the present invention, in the step S10, a suitable physical similarity ratio K is calculated by comprehensively considering the physical similarity model size L0, the scanning size limit L1 of the CT equipment and the engineering scale L2 of the filling body, and the size relationship satisfies L2=KL0, L1≥L0. The physical similarity ratio includes but is not limited to the filling space size similarity ratio k1, the filling material similarity ratio k2, the filling pipeline size similarity ratio k3 and the CO2 pumping flow similarity ratio k4. When preparing the similarity model, the initial porosity is ensured to be similar; after the physical similarity model of the cemented filling is established, CO2 is pumped into it.

[0061] In an example of the present invention, step S20 specifically includes the following steps:

[0062] Use high-precision CT equipment to scan the cemented filling body in the carbon fixation process in real time, and then perform noise reduction processing and convert it into voxels after obtaining the CT image; for example, Figure 2 As shown, it includes a ray source 1, a physically similar model 2, a gas injection tube 3, a stage 4 and a detector 5; wherein the ray source 1, the physically similar model 2 and the detector 5 are arranged in sequence along the projection direction of the ray source 1, and the physically similar model 2 is placed on the stage 4, and the gas injection tube 3 is inserted into the physically similar model 2 for injecting CO2 therein.

[0063] The Otsu threshold segmentation method is used to maximize the inter-class variance of the pore area and the solid area to obtain the optimal threshold T, and then the pore area is extracted from the voxel image;

[0064] A three-dimensional pore voxel model of the filling body is constructed, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed based on the pore voxels.

[0065] In one example of the present invention, the inter-class variance of the pore region and the solid region The expression is:

[0066]

[0067] Among them, ω1 and ω2 are the weights of the two categories after segmentation, μ1 and μ2 are the means of the two categories, and T is the optimal threshold.

[0068] In an example of the present invention, in step S20, analyzing the spatial distribution characteristics and temporal evolution characteristics of the pores includes:

[0069] The position distribution function f of the pores in the physical similarity model is calculated based on the position, number, volume, shape and fractal dimension of the pores. A (x,y,z), quantity distribution function f N (N), volume distribution function f V (V), shape distribution function f Ψ (Ψ) and the pore distribution pattern function f D (D), constructing the spatial statistical model of pore distribution g(f A ,f N ,f V ,f Ψ ,f D );

[0070] The position expression of the pore is: A = (x i ,y i ,z i ), the expression of the number of pores is: The volume expression of the pore is: V = NV0, and the shape expression of the pore is: (sphericity) and the fractal dimension of pores are expressed as:

[0071] In an example of the present invention, in step S20, analyzing the temporal evolution characteristics of the pores includes:

[0072] Analyze the pore distribution characteristics at each time point and obtain the position distribution time series function f At (x,y,z,t), quantity distribution time series function f Nt(N, t), volume distribution time series function f Vt (V, t), shape distribution time series function f Ψt (Ψ,t) and the pore distribution pattern time series function f Dt (D, t), connect the spatial statistical models of pore distribution at different time points to form the spatiotemporal evolution model of pore distribution g t (f At ,f Nt ,f Vt ,f Ψt ,f Dt ).

[0073] In an example of the present invention, step S30 includes the following steps:

[0074] First, input voxel information g input Mapped to low-dimensional information h, where the expression of status information h is: h = f encoder (g input )=σ(w encoder g input +b encoder ), where σ is the activation function, w encoder is the weight, b encoder is bias;

[0075] After capturing the key features of the pore structure, the low-dimensional features h are decoded back to voxel information g of the same dimension as the input output , where the expression of voxel information is: g output =f decoder (h)=σ(w decoder h+b decoder );

[0076] Minimize the loss function during voxel autoencoder training This can optimize the pore inversion results.

[0077] In one example of the present invention, Figure 3 As shown, in step S40, the CO2 diffusion simulation of the cementing filling space specifically includes the following steps:

[0078] The numerical model of the cemented filling engineering was established using COMSOL software. The initial porosity was set to be the same as that of the engineering, and the porosity that changes with time was set. Then, the CO2 flow and diffusion morphology and the spatial distribution law of CO2 concentration under different CO2 injection pressures, injection timings, injection durations and drilling arrangements were simulated.

[0079] A carbon dioxide diffusion simulation system for cemented filling space with porosity evolution according to the second aspect of the present invention comprises:

[0080] A physical similarity simulation unit is configured to calculate a suitable physical similarity ratio, establish a physical model having a certain similarity ratio with the engineering filling body, and conduct a physical test of cementation filling carbon fixation;

[0081] The evolution law acquisition unit is configured to acquire the spatiotemporal evolution law of the pore structure during the whole period of mineralization: in the process of gradual solidification and carbon fixation of the cemented filling body, the cemented filling body is scanned in real time by CT, a digital model of the filling body is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed;

[0082] The pore structure inversion unit is configured for engineering scale inversion of pore structure: the distribution characteristics of pore structure in the physically similar model are inverted to the engineering scale using a voxel autoencoder, and the reliability of the inversion results is verified by sampling at the engineering site;

[0083] The flow diffusion simulation unit is configured to establish a simulation model, perform preprocessing and initial porosity setting, and then perform CO2 diffusion simulation in the cemented filling space.

[0084] The simulation system establishes a physical similarity model of the engineering-scale cemented filling body through similarity ratio, scans the physical similarity model in the carbon fixation process in real time, constructs a digital filling body model, extracts pore throat structure characteristics, establishes a pore network model, and imports it into the software for CO2 diffusion simulation calculations. Compared with traditional methods, it achieves the technical effect of visualizing the spatiotemporal evolution of pore characteristics during carbon fixation and efficient simulation of CO2 diffusion in cemented filling space.

[0085] In an example of the present invention, the evolution law acquisition unit includes:

[0086] An image processing module is configured to use a high-precision CT device to scan the cemented filling body in the carbon fixation process in real time, and after obtaining the CT image, perform noise reduction processing and convert it into voxels;

[0087] The pore region extraction module is configured to use the Otsu threshold segmentation method to maximize the inter-class variance of the pore region and the solid region, obtain the optimal threshold T, and then extract the pore region from the voxel image;

[0088] The pore feature acquisition module is configured to construct a three-dimensional pore voxel model of the filling body, and analyze the spatial distribution characteristics and temporal evolution characteristics of the pores based on the pore voxels.

[0089] The exemplary implementation of the method for simulating carbon dioxide diffusion in a cemented filling space with porosity evolution proposed in the present invention is described in detail above with reference to the preferred embodiments. However, those skilled in the art will appreciate that, without departing from the concept of the present invention, various modifications and variations may be made to the above-mentioned specific embodiments, and various technical features and structures proposed in the present invention may be combined in various ways without exceeding the protection scope of the present invention, which is determined by the appended claims.

Claims

1. A method for simulating carbon dioxide diffusion in cemented filling spaces with porosity evolution, characterized in that: The steps include: S10: Physical similarity simulation test of mineralized carbon fixation: Calculate the appropriate physical similarity ratio, establish a physical model with a certain similarity ratio with the engineering filling body, and conduct physical tests on cemented filling carbon fixation; S20: Acquisition of spatiotemporal evolution of pore structure during the whole mineralization cycle: During the process of gradual solidification and carbon fixation of cemented filling bodies, CT is used to scan the cemented filling bodies in real time, a digital model of the filling bodies is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed; S30: Engineering scale inversion of pore structure: Use voxel autoencoders to invert the distribution characteristics of pore structure in physically similar models to the engineering scale, and take samples at the engineering site to verify the reliability of the inversion results; S40: Simulation of CO2 flow and diffusion in cemented filling space: Establish a simulation model, perform pretreatment and initial porosity setting, and then simulate the CO2 diffusion in the cemented filling space.

2. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: In step S10, a suitable physical similarity ratio K is calculated by comprehensively considering the physical similarity model size L0, the scanning size limit L1 of the CT device and the engineering dimension L2 of the filling body, and the size relationship satisfies L2=KL0, L1≥L0.

3. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: The step S20 specifically includes the following steps: Use high-precision CT equipment to scan the cemented filling body in the carbon fixation process in real time, and then perform noise reduction processing on the CT images and convert them into voxels; The Otsu threshold segmentation method is used to maximize the inter-class variance of the pore area and the solid area to obtain the optimal threshold T, and then the pore area is extracted from the voxel image; A three-dimensional pore voxel model of the filling body is constructed, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed based on the pore voxels.

4. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 3, characterized in that: Between-class variance of pore and solid regions The expression is: Among them, ω1 and ω2 are the weights of the two categories after segmentation, μ1 and μ2 are the means of the two categories, and T is the optimal threshold.

5. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: In the step S20, analyzing the spatial distribution characteristics of the pores includes: The position distribution function f of the pores in the physical similarity model is calculated based on the position, number, volume, shape and fractal dimension of the pores. A (x,y,z), quantity distribution function f N (N), volume distribution function f V (V), shape distribution function f Ψ (Ψ) and the pore distribution pattern function f D (D), constructing the spatial statistical model of pore distribution g(f A ,f N ,f V ,f Ψ ,f D ).

6. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: In the step S20, analyzing the temporal evolution characteristics of the pores includes: Analyze the pore distribution characteristics at each time point and obtain the position distribution time series function f At (x,y,z,t), quantity distribution time series function f Nt (N,t), volume distribution time series function f Vt (V, t), shape distribution time series function f Ψt (Ψ,t) and the pore distribution pattern time series function f Dt (D, t), connect the spatial statistical models of pore distribution at different time points to form the spatiotemporal evolution model of pore distribution g t (f At ,f Nt ,f Vt ,f Ψt ,f Dt ).

7. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: The step S30 includes the following steps: First, input voxel information g input Mapped to low-dimensional information h, where the expression of status information h is: h = f encoder (g input )=σ(w encoder g input +b encoder ), where σ is the activation function, w encoder is the weight, b encoder is bias; After capturing the key features of the pore structure, the low-dimensional features h are decoded back to voxel information g of the same dimension as the input output , where the expression of voxel information is: g output =f decoder (h)=σ(w decoder h+b decoder ); Minimize the loss function during voxel autoencoder training This can optimize the pore inversion results.

8. The method for simulating carbon dioxide diffusion in cemented filling space with porosity evolution according to claim 1, characterized in that: In step S40, the CO2 diffusion simulation of the cementing filling space specifically includes the following steps: The numerical model of the cemented filling engineering was established using COMSOL software. The initial porosity was set to be the same as that of the engineering, and the porosity that changes with time was set. Then, the CO2 flow and diffusion morphology and the spatial distribution law of CO2 concentration under different CO2 injection pressures, injection timings, injection durations and drilling arrangements were simulated.

9. A carbon dioxide diffusion simulation system for cemented filling space with porosity evolution, characterized in that: include: A physical similarity simulation unit is configured to calculate a suitable physical similarity ratio, establish a physical model having a certain similarity ratio with the engineering filling body, and conduct a physical test of cementation filling carbon fixation; The evolution law acquisition unit is configured to acquire the spatiotemporal evolution law of the pore structure during the whole period of mineralization: in the process of gradual solidification and carbon fixation of the cemented filling body, the cemented filling body is scanned in real time by CT, a digital model of the filling body is established, and then a pore voxel model is extracted, and the spatial distribution characteristics and temporal evolution characteristics of the pores are analyzed; The pore structure inversion unit is configured for engineering scale inversion of pore structure: the distribution characteristics of pore structure in the physically similar model are inverted to the engineering scale using a voxel autoencoder, and the reliability of the inversion results is verified by sampling at the engineering site; The flow diffusion simulation unit is configured to establish a simulation model, perform preprocessing and initial porosity settings, and then perform CO2 diffusion simulation in the cemented filling space.

10. The carbon dioxide diffusion simulation system for cemented filling space with porosity evolution according to claim 9, characterized in that: The evolution law acquisition unit comprises: An image processing module is configured to use a high-precision CT device to scan the cemented filling body in the carbon fixation process in real time, and after obtaining the CT image, perform noise reduction processing and convert it into voxels; The pore region extraction module is configured to use the Otsu threshold segmentation method to maximize the inter-class variance of the pore region and the solid region, obtain the optimal threshold T, and then extract the pore region from the voxel image; The pore feature acquisition module is configured to construct a three-dimensional pore voxel model of the filling body, and analyze the spatial distribution characteristics and temporal evolution characteristics of the pores based on the pore voxels.

Images