A physical simulation method and system for fracturing fracture imaging using non-uniformly distributed contact electrodes.

Abstract

This invention relates to the field of oilfield reservoir research technology, specifically a physical simulation method and system for fracturing imaging using non-uniformly distributed contact electrodes. The method includes: detecting resistance modes using a low-frequency excitation signal at the current time step during fracture propagation simulation to obtain the corresponding electrical signal; obtaining the conductive fluid distribution of the core fracturing specimen at the current time step using a conductive fluid distribution inversion model; fitting and comparing the electric field distribution result at the current time step with the conductive fluid distribution result; and outputting the fracture parameters and conductive fluid distribution results of the core fracturing specimen at the current time step based on the fitting and comparison results. This invention directly and in-situ senses the changes in electrical properties caused by natural fractures and their contained fluids, acquires electrical signals, reduces response delay, and effectively improves the accuracy and reliability of conductive fluid distribution inversion by combining prior knowledge and deep learning to construct a conductive fluid distribution inversion model.

Description

Technical Field

[0001] This invention relates to the field of oilfield reservoir research technology, and is a physical simulation method and system for fracturing fracture imaging that combines non-uniformly distributed contact electrodes. Background Technology

[0002] Fracture morphology is an important indicator of the effectiveness of hydraulic fracturing in oil and gas reservoirs. Obtaining information on fracture development quickly and accurately is crucial for evaluating the effectiveness of hydraulic fracturing in oil and gas reservoirs.

[0003] Currently, indoor experiments and numerical simulations are the main methods for obtaining crack development morphology, specifically:

[0004] Physical simulation experiments, which involve fracturing on small-scale rock cores, can directly reflect the initiation, propagation, and morphological characteristics of fractures, making them a key means of studying the formation mechanism of complex fracture networks and evaluating the effectiveness of fracturing schemes. However, existing physical simulation fracture monitoring technologies have significant limitations: 1. Acoustic emission technology locates fracture event points by capturing elastic waves generated by rock fracturing, but it is difficult to accurately depict the continuous fracture geometry, has insufficient spatial resolution, and has limited sensitivity to micro-cracks or non-brittle fractures, and is easily affected by background noise; 2. Although CT technology can provide high-resolution three-dimensional structural information, its slow scanning speed makes it difficult to capture the transient process of rapid dynamic fracture propagation during fracturing, resulting in missing or blurred morphological information in key propagation stages; 3. The high purchase and maintenance costs of imaging equipment such as CT and MRI, and their special working environment requirements limit their integrated application in conventional autoclave simulation devices.

[0005] Numerical simulation methods for predicting fracture morphology are highly dependent on the input constitutive model, fracture criteria, and parameters, which are often difficult to obtain or calibrate accurately, and still pose challenges for simulating complex geomechanical interactions.

[0006] Existing fracturing fracture imaging simulation inversion methods, such as:

[0007] Existing patent document CN114239431B discloses a simulation method, apparatus, and equipment for water-driven oil recovery in fractured reservoirs. Based on the characteristics of water displacement flow lines in fractured cores with different occurrences in fractured reservoirs, the matrix is ​​divided into several linear sub-regions, transforming the two-dimensional water-driven oil recovery problem into a one-dimensional linear sub-region coupled solution problem. The study area is divided into injection zone, fracture zone, and production zone, and corresponding linear flow models are created for each. Each linear flow model is solved sequentially to obtain analytical solutions for linear displacement in the matrix zone at the injection end, numerical solutions for linear conduction in the fractures, and analytical solutions for matrix displacement at the outlet end. The linear flow of the matrix and fracture systems is coupled, transforming the two-dimensional water-driven oil recovery problem into a one-dimensional linear sub-region coupled solution problem. A semi-analytical solution for the three-zone linear flow of water-driven oil recovery in fractured cores is obtained, thereby obtaining the pressure distribution and saturation distribution of fractured reservoirs, achieving accurate simulation of the water-driven oil recovery process in fractured reservoirs. This method does not involve constructing a physical simulation method for fracturing crack imaging that combines resistance mode detection, conductive fluid distribution inversion, and crack fitting with non-uniformly distributed contact electrodes. It cannot combine resistance mode detection, prior knowledge, and deep learning to dynamically analyze the geometry and conductivity distribution of non-planar and branched cracks.

[0008] Existing patent document 2, publication number CN116068662A, discloses a method and apparatus for random simulation of fractures based on fracture development intensity trends. The method includes: establishing a fracture development intensity distribution map; performing physical property simulation based on preset fracture simulation parameters to randomly generate fracture simulation centers and corresponding fracture probability parameters; extracting the fracture intensity corresponding to the fracture center from the fracture development intensity distribution map based on the fracture simulation center, and verifying the fracture probability parameters based on the fracture intensity; calculating the starting and ending coordinates of the fracture to which the fracture simulation center belongs, storing and drawing the fracture; considering the influence of geological factors, and using the fracture development intensity trend as a constraint to control random fracture simulation, the results of random fracture simulation are highly consistent with geological understanding, conforming to geological laws and possessing high reliability. However, this method does not involve constructing a physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes that integrates resistance mode detection, conductive fluid distribution inversion, and fracture fitting. It cannot combine resistance mode detection, prior knowledge, and deep learning to dynamically analyze the geometric morphology and conductivity distribution of non-planar and branch fractures. Summary of the Invention

[0009] This invention provides a physical simulation method and system for fracturing crack imaging that combines non-uniformly distributed contact electrodes, overcoming the shortcomings of the prior art. It can effectively solve the problems of response lag and high maintenance cost of existing crack development imaging technology.

[0010] One of the technical solutions of this invention is achieved through the following measures: a physical simulation method for fracturing fracture imaging combining non-uniformly distributed contact electrodes, comprising:

[0011] The fracture propagation of the core fracturing specimen held in the core clamping device was simulated, and the resistance mode was detected at the current time step using a low-frequency excitation signal during the fracture propagation simulation to obtain the corresponding electrical signal.

[0012] The effective porosity of the core fracturing specimen, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation are used as the prior knowledge input for the conductive fluid distribution inversion model. The electrical signal is used as the original data input for the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen at the current time step.

[0013] The conductive fluid distribution result of the previous time step is used as the initial field of the current time step to determine the fracture morphology distribution in the core at the current time step. The fracture morphology distribution in the core at the current time step is fitted and compared with the conductive fluid distribution result of the current time step. Based on the fitting and comparison results, the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen at the current time step are output.

[0014] The following are further optimizations and / or improvements to the above-mentioned technical solution:

[0015] The construction process of the above-mentioned conductive fluid distribution inversion model includes:

[0016] Several samples were obtained and divided into training sample set and test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen.

[0017] The initial model is trained using a training sample set. A loss function is introduced during training, and training ends when the value of the loss function stabilizes, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms, including dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module applies physical constraints. The decoder performs feature recovery on the encoded fused features.

[0018] The conductive fluid distribution inversion model was tested using a test sample set, the model parameters were optimized, and a conductive fluid distribution inversion model that meets the test evaluation requirements was output.

[0019] The above method uses the conductive fluid distribution results from the previous time step as the initial field for the current time step to determine the fracture morphology distribution within the core at the current time step. It then compares and fits the fracture morphology distribution with the conductive fluid distribution results at the current time step. Based on the fitting and comparison results, it outputs the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen at the current time step, including:

[0020] The conductive fluid distribution results from the previous time step are used as the initial field for the current time step to determine the fracture morphology distribution within the core at the current time step.

[0021] The distribution of fracture morphology in the core at the current time step is compared with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error, and it is determined whether the fitting error is greater than the set value.

[0022] In response, the initial field of the previous time step is randomly adjusted within the set adjustment range, and the core fracture morphology distribution of the current time step is redefined.

[0023] If no response is received, the output will show the distribution of conductive fluid and the distribution of fracture morphology within the core at the current time step.

[0024] The aforementioned core clamping device includes a clamping body and two end caps, which are respectively located at the front and rear ends of the clamping body. Each end cap has a fluid inlet / outlet. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure, and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes arranged at intervals. An insulating rubber block is arranged between adjacent resistance electrodes. Four stress loading modules are respectively arranged on the outside of the five modal sensing array measurement layers.

[0025] The above describes the simulation of crack propagation in a core fracturing specimen held in a core clamping device. During the crack propagation simulation, a low-frequency excitation signal is used to perform resistance mode detection at the current time step to obtain the corresponding electrical signal, including:

[0026] The core is processed into a core fracturing specimen, and a simulated wellbore is set in the center of the core fracturing specimen. The size of the core fracturing specimen is set according to the size of the core clamping device.

[0027] The core fracturing specimen is placed inside the modal sensing array component, which is then placed in the clamping assembly. Insulating silicone oil is filled between the clamping assembly and the modal sensing array component via an external pump. By applying pressure to the external stress loading module, stress is provided to the core fracturing specimen while simultaneously compacting the modal sensing array component, ensuring that the resistance electrodes on the modal sensing array component are in complete contact with the core fracturing specimen.

[0028] Conductive fluid is injected into the core fracturing specimen held in the core clamping device at a set flow rate or constant pressure. During the process of injecting the conductive fluid and driving the fracture propagation, the resistance mode is detected at the current time step using a low-frequency excitation signal to obtain the corresponding electrical signal.

[0029] The above-mentioned resistance mode detection process includes:

[0030] In the modal sensing array component, a low-frequency excitation signal is injected into the diagonal resistor electrode pair, the potential difference between the remaining adjacent resistor electrodes is measured, the corresponding voltage gradient is obtained, and the equivalent resistance value is calculated.

[0031] Repeat the above steps, applying current layer by layer in sequence, and determine the equivalent resistance of each measurement layer of the multimodal sensing array.

[0032] The second technical solution of the present invention is achieved through the following measures: a fracturing fracture imaging physical simulation system combined with non-uniformly distributed contact electrodes, including a simulation detection part and a dynamic inversion part;

[0033] The simulation and testing section includes a core clamping device and an excitation and testing device. The core clamping device holds the core fracturing specimen, and the core fracturing specimen clamped in the core clamping device is used to simulate crack propagation. During the crack propagation simulation, the excitation and testing device uses a low-frequency excitation signal to perform resistance mode detection at the current time step to obtain the corresponding electrical signal.

[0034] The dynamic inversion part includes:

[0035] The dynamic inversion unit takes the effective porosity of the core fracturing specimen, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation as the prior knowledge input of the conductive fluid distribution inversion model, and takes the electrical signal as the original data input of the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen at the current time step.

[0036] The fracture fitting unit uses the conductive fluid distribution result of the previous time step as the initial field of the current time step to determine the fracture morphology distribution in the core at the current time step. It then compares and fits the fracture morphology distribution in the core at the current time step with the conductive fluid distribution result. Based on the fitting comparison result, it outputs the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen at the current time step.

[0037] The following are further optimizations and / or improvements to the above-mentioned technical solution:

[0038] The aforementioned core clamping device includes a clamping body and two end caps, which are respectively located at the front and rear ends of the clamping body. Each end cap has a fluid inlet / outlet. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure, and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes arranged at intervals. An insulating rubber block is arranged between adjacent resistance electrodes. Four stress loading modules are respectively arranged on the outside of the five modal sensing array measurement layers.

[0039] The aforementioned excitation and detection device includes an excitation source module and a detection module;

[0040] The excitation source module is equipped with a frequency-controlled resistance excitation channel, which outputs a 10Hz–100kHz square wave current with a maximum current ≤20mA, and supports core axial and radial resistance gradient measurement.

[0041] The detection module includes a resistance gradient measurement module, which performs resistance mode detection.

[0042] The above also includes a model building unit for constructing the conductive fluid distribution inversion model, including:

[0043] The sample acquisition module acquires several samples and divides them into training sample set and test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen.

[0044] The training module trains the initial model using the training sample set. A loss function is introduced during training, and training ends when the value of the loss function stabilizes, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms, including dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module applies physical constraints. The decoder performs feature recovery on the encoded fused features.

[0045] The testing module uses a test sample set to test the conductive fluid distribution inversion model, optimizes the model parameters of the conductive fluid distribution inversion model, and outputs a conductive fluid distribution inversion model that meets the test evaluation requirements.

[0046] The aforementioned fracturing fitting unit includes:

[0047] The solution module uses the conductive fluid distribution results from the previous time step as the initial field for the current time step to determine the fracture morphology distribution within the core at the current time step.

[0048] The fitting module compares the distribution of fracture morphology in the core at the current time step with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error and determine whether the fitting error is greater than the set value.

[0049] If the fitting output module responds to "yes", it will randomly adjust the initial field of the previous time step within the set adjustment range and return to redetermine the fracture morphology distribution in the core at the current time step; otherwise, it will output the conductive fluid distribution result and the fracture morphology distribution in the core at the current time step.

[0050] The beneficial effects of this invention include:

[0051] (1) Millisecond-level dynamic capture, breaking through the inherent limitations of launch positioning ambiguity and CT scan delay, constructing a millisecond-level electric field signal dynamic response network: Through a non-uniformly distributed electrode array and a microsecond-level signal acquisition architecture, real-time tracking of crack propagation speed ≥1m / s is achieved, with a time resolution of 10ms / frame, fully recording transient behaviors such as crack initiation, branching, and turning.

[0052] (2) Intelligent analysis of complex cracks: Combining prior knowledge and deep learning, a conductive fluid distribution inversion model is constructed. This model is based on rock mechanics to construct prior constraints for crack propagation. Combined with deep learning network, the geometric shape and conductivity distribution of non-planar cracks and branch cracks are dynamically analyzed, which effectively improves the accuracy and reliability of conductive fluid distribution inversion. Attached Figure Description

[0053] Appendix Figure 1 This is a schematic diagram of the physical simulation method for imaging of hydraulic fracturing cracks provided in an embodiment of the present invention.

[0054] Appendix Figure 2 This is a schematic diagram of the core fracturing specimen provided in an embodiment of the present invention.

[0055] Appendix Figure 3 This is a schematic cross-sectional view of the modal sensing array component provided in an embodiment of the present invention.

[0056] Appendix Figure 4 This is a schematic diagram of the arrangement of resistive electrodes and the input and output of electrical signals in the modal sensing array measurement layer provided in an embodiment of the present invention.

[0057] Appendix Figure 5 A schematic diagram of the process for constructing a conductive fluid distribution inversion model according to an embodiment of the present invention.

[0058] Appendix Figure 6 This is a schematic diagram of the method for obtaining crack parameters and conductive fluid distribution results at the current time step, as provided in an embodiment of the present invention.

[0059] Appendix Figure 7 This is a schematic diagram of the physical simulation system for imaging of hydraulic fracturing cracks provided in an embodiment of the present invention.

[0060] The codes in the attached diagram are as follows: 1 is the simulated wellbore, 2 is the core fracturing specimen, 3 is the stress loading module, 4 is the resistance electrode, and 5 is the insulating rubber block. Detailed Implementation

[0061] The present invention is not limited to the following embodiments, and the specific implementation can be determined according to the technical solution of the present invention and the actual situation.

[0062] Those skilled in the art will understand that, unless specifically stated otherwise, in the embodiments of the present invention, a "module" or "unit" refers to a computer program or part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0063] In addition, in the embodiments of the present invention, "multiple" refers to two or more, and "first" and "second" are used to distinguish descriptions and should not be construed as implying relative importance.

[0064] Based on this, the technical solution of the present invention will be described and explained below with reference to several examples.

[0065] Example 1: As shown in the attached document Figure 1 As shown, this embodiment of the invention discloses a physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes, comprising:

[0066] Step S110: Simulate the crack propagation of the core fracturing specimen 2 held in the core clamping device, and use a low-frequency excitation signal to perform resistance mode detection at the current time step during the crack propagation simulation to obtain the corresponding electrical signal.

[0067] Step S120: The effective porosity of the core of the core fracturing specimen 2, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation are used as the prior knowledge input of the conductive fluid distribution inversion model. The electrical signal is used as the original data input of the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen 2 at the current time step.

[0068] Step S130: Use the conductive fluid distribution result of the previous time step as the initial field of the current time step, determine the fracture morphology distribution in the core of the current time step, and compare the fracture morphology distribution in the core of the current time step with the conductive fluid distribution result of the current time step. Based on the comparison result, output the fracture morphology distribution and conductive fluid distribution result of the core fracturing specimen 2 in the current time step.

[0069] In this embodiment, the process of obtaining the effective porosity of the core fracturing specimen 2 includes: drying the core fracturing specimen 2 in a 40℃ constant temperature oven for 24 hours and weighing the dry weight; then placing the core fracturing specimen 2 in a prepared KCl solution (concentration 0.01mol / L) and saturating it under vacuum for 24 hours and weighing the wet weight; calculating the saturated water volume from the difference between the wet and dry weights; and then calculating the effective porosity of the core from the solution density.

[0070] In this embodiment, the process of obtaining the initial electric field distribution of the core fracturing specimen 2 includes using a low-frequency excitation signal to perform resistance mode detection at the current time step, obtaining the corresponding electrical signal, calculating the initial electric field distribution, and inverting the development of core pores and microfractures. It should also be noted that the electrical signal is the conductivity, and the initial electric field distribution is calculated based on the conductivity.

[0071] In this embodiment, the flow field data pre-calculated using the Darcy equation is obtained by establishing a numerical model based on the Darcy equation and using numerical calculation methods to obtain the fluid distribution, pressure distribution, and composition changes at different locations at different time steps. This process can be implemented using finite element software.

[0072] This invention discloses a physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes. The method integrates resistance mode detection, conductive fluid distribution inversion, and fracture fitting. During fracture propagation simulation, a low-frequency excitation signal is used to perform resistance mode detection at the current time step, directly and in-situ sensing the changes in electrical properties caused by natural fractures and their contained fluids, obtaining the corresponding electrical signals. This method offers fast signal acquisition, reducing response delay, and has relatively relaxed material requirements and controllable costs. Furthermore, by combining prior knowledge and deep learning, a conductive fluid distribution inversion model is constructed. This model uses rock mechanics to establish prior constraints for fracture propagation and combines a deep learning network to dynamically analyze the geometry and conductivity distribution of non-planar and branch fractures, effectively improving the accuracy and reliability of conductive fluid distribution inversion.

[0073] Example 2: As shown in the attached document Figure 2 , 3 As shown, the embodiment of the present invention is a further optimization of the above embodiment. The core clamping device includes a clamping body and two end caps, which are respectively disposed at the front end and rear end of the clamping body. Both end caps are provided with fluid inlets / outlets. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block 5 to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes 4 arranged at intervals. An insulating rubber block 5 is disposed between adjacent resistance electrodes 4. Four stress loading modules 3 are respectively disposed on the outside of the five modal sensing array measurement layers.

[0074] It should be noted that the six insulating rubber blocks 5 in the sensing array measurement section are not exactly the same length. The stress loading module 3 can be used to simulate the stress on the rock core (formation stress, which can be 10MPa in this embodiment) while compacting the modal sensing array components, so that the resistance electrodes 4 on the modal sensing array components are in complete contact with the rock core fracturing specimen 2. Furthermore, the appendix... Figure 3 δ in H and δ h The numerical values ​​are the same, only the directions they represent are different.

[0075] The thickness of the aforementioned resistive electrode 4 can be 0.5 mm, and the material can be conductive rubber (silicone matrix + 20% carbon black). The outer shell of the clamping component is made of Hastelloy C276 material, with an embedded zirconia ceramic insulating layer, a pressure resistance of up to 20 MPa, and a temperature range covering -20 to 150℃.

[0076] Compared to traditional core clamping devices used in conventional resistivity imaging (low-resolution) methods, this embodiment combines low-frequency resistance measurement with a core clamping device featuring five modal sensing array measurement layers. Each modal sensing array measurement layer has 28 resistance electrodes 4, thus advancing from traditional "end-face measurement" or "single-point measurement" to "real-time spatial detection." This provides data support for subsequent real-time inversion calculation of the dynamic changes in the distribution of conductive fluid at different locations in the core fracturing specimen 2, making the conductive fluid distribution results more accurate.

[0077] Furthermore, based on the structure of the core clamping device described above, this embodiment simulates the crack propagation of the core fracturing specimen 2 clamped in the core clamping device. During the crack propagation simulation, a low-frequency excitation signal is used to perform resistance mode detection at the current time step to obtain the corresponding electrical signal, specifically including:

[0078] Step S210: The core is processed into a core fracturing specimen 2. A simulated wellbore 1 is set at the center of the core fracturing specimen 2. The size of the core fracturing specimen 2 is set according to the size of the core clamping device. In this embodiment, the core fracturing specimen 2 is a standard core of 10cm×10cm×10cm.

[0079] It should be noted that the simulated wellbore 1 is a metal cylinder with an inner insulating layer, and its size is selected according to the needs.

[0080] In step S220, the core fracturing specimen 2 is placed inside the modal sensing array component, which is then placed in the clamping assembly. Insulating silicone oil is filled between the clamping assembly and the modal sensing array component via an external pump. By externally pressurizing the stress loading module 3, stress is provided to the core fracturing specimen 2 (the stress simulates the formation stress experienced by the core underground, which can be set to 10MPa in this embodiment). At the same time, the modal sensing array component is compacted, so that the resistance electrode 4 on the modal sensing array component is in complete contact with the core fracturing specimen 2.

[0081] Step S230: Inject conductive fluid (i.e., conductive fracturing fluid, which can be a KCl solution with a concentration of 0.01 mol / L) into the core fracturing specimen 2 held in the core clamping device at a set displacement or constant pressure. During the process of injecting conductive fluid and driving fracture propagation, use a low-frequency excitation signal to perform resistance mode detection at the current time step to obtain the corresponding electrical signal.

[0082] In this embodiment, the discharge rate can be set to 30-100 ml / min, and the constant pressure can be set to 8 MPa.

[0083] In this embodiment, resistance mode detection utilizes electrical resistance tomography (EIT) to measure the electrical signal corresponding to the current time step in real time, as shown in the attached figure. Figure 4 As shown, it includes:

[0084] (1) Inject a low-frequency excitation signal into the diagonal resistor electrode 4 in the modal sensing array component, measure the potential difference between the remaining adjacent resistor electrodes 4, obtain the corresponding voltage gradient, and invert the corresponding conductivity, specifically including:

[0085] A low-frequency excitation signal is injected into the four pairs of diagonal resistor electrodes. The low-frequency excitation signal can be a square wave current from 10Hz to 100kHz, and the maximum current can be set to ≤20mA.

[0086] Measure the potential difference Δ between adjacent resistive electrodes 4 V i Determine the corresponding voltage gradient E i ;

[0087] E i =Δ V i / d

[0088] Determine the average voltage gradient E avg Calculate the total voltage drop between the diagonal resistor electrodes 4. V AB ;

[0089] V AB = E avg / I

[0090] in, d The spacing between adjacent resistive electrodes 4; I It is a low-frequency excitation signal (low-frequency AC signal);

[0091] According to Ohm's law, the equivalent resistance corresponding to the total voltage drop between the diagonal resistive electrodes 4 is obtained;

[0092] Conductivity is derived from the equivalent resistance.

[0093] (2) Repeat the above steps, apply current layer by layer in sequence, and determine the equivalent resistance of each multimodal sensing array measurement layer.

[0094] The traversal period can be set to 0.1ms, and the current range can be set to 0.1 to 15mA.

[0095] Furthermore, based on this embodiment, the process of obtaining the initial electric field distribution includes:

[0096] (1) The core is processed into a core fracturing specimen 2. A simulated wellbore 1 is set in the center of the core fracturing specimen 2. The size of the core fracturing specimen 2 is set according to the size of the core clamping device. In this embodiment, the core fracturing specimen 2 is a standard core of 10cm×10cm×10cm.

[0097] (2) The core fracturing specimen 2 is placed in the modal sensing array component, and the modal sensing array component is placed in the clamping assembly. Insulating silicone oil is filled between the clamping assembly and the modal sensing array component through an external pump. The stress loading module 3 is pressurized externally to provide stress to the core fracturing specimen 2 (the stress simulates the formation stress that the core bears underground, which can be set to 10MPa in this embodiment). At the same time, the modal sensing array component is compacted so that the resistance electrode 4 on the modal sensing array component is in complete contact with the core fracturing specimen 2.

[0098] (3) Inject conductive fluid (i.e. conductive fracturing fluid, which can be a KCl solution with a concentration of 0.01 mol / L) into the core fracturing specimen 2 held in the core clamping device at a set displacement or constant pressure. During the process of injecting conductive fluid and driving fracture propagation, use a low-frequency excitation signal to perform resistance mode detection at the current time step to obtain the corresponding electrical signal.

[0099] (4) By collecting the changes in electrical signals during the injection of conductive fluid, a quantitative calibration relationship between signal characteristics and conductive fluid distribution (especially crack morphology) is established to form an initial electric field distribution.

[0100] It should be noted that the signal characteristic here can be resistance.

[0101] Example 3: As shown in the attached document Figure 5 As shown, the embodiments of the present invention are further optimizations of the above embodiments, wherein the construction process of the conductive fluid distribution inversion model includes:

[0102] Step S310: Obtain several samples and divide them into training sample set and test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen 2, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen 2.

[0103] Step S320: Train the initial model using the training sample set. A loss function is introduced during training. When the value of the loss function is stable, the training ends, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms. It includes dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module performs physical constraints. The decoder performs feature recovery on the encoded fused features.

[0104] The above feature fusion can be performed using either concatenation or weighted methods. The physical constraint module embeds the mass conservation equation and Darcy's law as regularization constraints, as shown below:

[0105] mass conservation equation:

[0106]

[0107] Darcy's Law:

[0108]

[0109] Where φ is the effective porosity of the core, V is the permeability under fully saturated conditions, K is the permeability coefficient, and μ is the fluid viscosity. This represents the pressure gradient.

[0110] The aforementioned attention mechanism is used in the feature fusion module. It can automatically learn and assign different importance weights to different feature regions or different feature dimensions in the two input channels, highlighting the fusion features most relevant to saturation inversion and suppressing irrelevant or noisy information, thereby constructing a better tensor representation for subsequent encoding and decoding.

[0111] The loss function mentioned above can be selected as needed. In this embodiment, variance can be used as the loss function, which is usually used to measure the degree of dispersion between data points and the mean.

[0112] Step S330: Test the conductive fluid distribution inversion model using the test sample set, optimize the model parameters of the conductive fluid distribution inversion model, and output a conductive fluid distribution inversion model that meets the test evaluation requirements.

[0113] Example 4: As shown in the appendix Figure 6As shown, this embodiment of the invention is a further optimization of the above embodiment. The conductive fluid distribution result from the previous time step is used as the initial field for the current time step to determine the fracture morphology distribution within the core at the current time step. The fracture morphology distribution within the core at the current time step is then fitted and compared with the conductive fluid distribution result. Based on the fitting and comparison results, the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen 2 at the current time step are output, including:

[0114] Step S410: Use the conductive fluid distribution result of the previous time step as the initial field of the current time step to determine the fracture morphology distribution in the core at the current time step.

[0115] It should be noted that the above steps can be implemented using COMSOL software.

[0116] Step S420: The distribution of fracture morphology in the core at the current time step is compared with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error, and it is determined whether the fitting error is greater than a set value (in this embodiment, the set value can be 5%).

[0117] It should be noted that the above fitting and comparison process combines existing software based on fluid distribution models and rock physics models to calculate the theoretically measurable electrical signal corresponding to the fracture morphology distribution in the core at the current time step, and compares the theoretically calculated electrical signal with the actual electrical signal collected in the conductive fluid distribution results at the current time step.

[0118] In step S430, the initial field of the previous time step is randomly adjusted within the set adjustment range, and the process returns to step S410 to redetermine the distribution of fracture morphology in the core at the current time step.

[0119] In step S440, if the response is no, the conductive fluid distribution result and the fracture morphology distribution in the core at the current time step are output.

[0120] In this embodiment, the adjustment range can be [-1, 1]. The distribution of fracture morphology in the core in this embodiment includes the location, shape, and propagation process of the fractures.

[0121] Example 5: As shown in the attached document Figure 7 As shown, this embodiment of the invention discloses a fracturing fracture imaging physical simulation system combined with non-uniformly distributed contact electrodes, including a simulation detection part and a dynamic inversion part;

[0122] The simulation and testing section includes a core clamping device and an excitation and testing device. The core clamping device holds the core fracturing specimen 2. The core fracturing specimen 2 held in the core clamping device is used to simulate crack propagation. During the crack propagation simulation, the excitation and testing device uses a low-frequency excitation signal to perform resistance mode detection at the current time step to obtain the corresponding electrical signal.

[0123] The dynamic inversion part includes:

[0124] The dynamic inversion unit takes the effective porosity of the core of the core fracturing specimen 2, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation as the prior knowledge input of the conductive fluid distribution inversion model, and takes the electrical signal as the original data input of the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen 2 at the current time step.

[0125] The fracture fitting unit uses the conductive fluid distribution result of the previous time step as the initial field of the current time step to determine the fracture morphology distribution in the core at the current time step. It then compares and fits the fracture morphology distribution in the core at the current time step with the conductive fluid distribution result. Based on the fitting comparison result, it outputs the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen at the current time step.

[0126] In this embodiment, the dynamic inversion part can be implemented based on electronic devices, which include processors and memory. The memory stores computer programs corresponding to the corresponding units. The computer programs are loaded and executed by the processor to implement the corresponding method steps.

[0127] The aforementioned processor may be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof; it may implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention; it may also be a combination that implements computing functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc. The aforementioned memory may include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, portable hard drives, magnetic disks, or optical disks.

[0128] Among them, as attached Figure 2 , 3As shown, the core clamping device includes a clamping body and two end caps, which are respectively located at the front and rear ends of the clamping body. Each end cap has a fluid inlet / outlet. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure, and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block 5 to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes 4 arranged at intervals. An insulating rubber block 5 is arranged between adjacent resistance electrodes 4. Four stress loading modules 3 are respectively arranged on the outside of the five modal sensing array measurement layers.

[0129] It should be noted that the lengths of the six insulating rubber blocks 5 in the sensing array measurement section are not exactly the same. The stress loading module 3 can be used to simulate the stress (formation stress, which can be 10MPa in this embodiment) on the rock core while compacting the modal sensing array component, so that the resistance electrode 4 on the modal sensing array component is in complete contact with the rock core fracturing specimen 2.

[0130] The thickness of the aforementioned resistive electrode 4 can be 0.5 mm, and the material can be conductive rubber (silicone matrix + 20% carbon black). The outer shell of the clamping component is made of Hastelloy C276 material, with an embedded zirconia ceramic insulating layer, a pressure resistance of up to 20 MPa, and a temperature range covering -20 to 150℃.

[0131] The excitation and detection device includes an excitation source module and a detection module;

[0132] The excitation source module is equipped with a frequency-controlled resistance excitation channel, which outputs a 10Hz–100kHz square wave current with a maximum current ≤20mA, and supports core axial and radial resistance gradient measurement.

[0133] The detection module includes a resistance gradient measurement module, which performs resistance mode detection.

[0134] In practical use, the resistance electrode 4 in the core clamping device is connected to the excitation and detection device in advance through the high-pressure sealed feedthrough interface. The specific connection method is the same as that in the current injection method and the detection object in the method embodiment.

[0135] The fracturing fitting unit includes:

[0136] The solution module uses the conductive fluid distribution results from the previous time step as the initial field for the current time step to determine the fracture morphology distribution within the core at the current time step.

[0137] The fitting module compares the distribution of fracture morphology in the core at the current time step with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error and determine whether the fitting error is greater than the set value.

[0138] If the fitting output module responds to "yes", it will randomly adjust the initial field of the previous time step within the set adjustment range and return to redetermine the fracture morphology distribution in the core at the current time step; otherwise, it will output the conductive fluid distribution result and the fracture morphology distribution in the core at the current time step.

[0139] Example 6: This embodiment of the invention is a further optimization of the above embodiments, and also includes a model building unit for constructing a conductive fluid distribution inversion model, including:

[0140] The sample acquisition module acquires several samples and divides them into a training sample set and a test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen 2, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen 2.

[0141] The training module trains the initial model using the training sample set. A loss function is introduced during training, and training ends when the value of the loss function stabilizes, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms, including dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module applies physical constraints. The decoder performs feature recovery on the encoded fused features.

[0142] The testing module uses a test sample set to test the conductive fluid distribution inversion model, optimizes the model parameters of the conductive fluid distribution inversion model, and outputs a conductive fluid distribution inversion model that meets the test evaluation requirements.

[0143] The above content is only a specific embodiment of the present invention, which has strong adaptability and implementation effect. However, the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention. Therefore, equivalent changes made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes, characterized in that, include: The fracture propagation of the core fracturing specimen held in the core clamping device was simulated, and the resistance mode was detected at the current time step using a low-frequency excitation signal during the fracture propagation simulation to obtain the corresponding electrical signal. The effective porosity of the core fracturing specimen, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation are used as the prior knowledge input for the conductive fluid distribution inversion model. The electrical signal is used as the original data input for the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen at the current time step. The conductive fluid distribution of the core fracturing specimen in the previous time step is used as the initial field for the current time step. The fracture morphology distribution in the core at the current time step is determined. The fracture morphology distribution in the core at the current time step is fitted and compared with the conductive fluid distribution at the current time step. Based on the fitting and comparison results, the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen at the current time step are output.

2. The physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes according to claim 1, characterized in that, The process of constructing the conductive fluid distribution inversion model includes: Several samples were obtained and divided into training sample set and test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen. The initial model is trained using a training sample set. A loss function is introduced during training, and training ends when the value of the loss function stabilizes, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms, including dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module applies physical constraints. The decoder performs feature recovery on the encoded fused features. The conductive fluid distribution inversion model was tested using a test sample set, the model parameters were optimized, and a conductive fluid distribution inversion model that meets the test evaluation requirements was output.

3. The physical simulation method for fracturing fracture imaging based on non-uniformly distributed contact electrodes according to claim 1 or 2, characterized in that, The conductive fluid distribution of the core fracturing specimen in the previous time step is used as the initial field for the current time step. The fracture morphology distribution within the core in the current time step is determined. The fracture morphology distribution within the core in the current time step is then fitted and compared with the conductive fluid distribution results. Based on the fitting and comparison results, the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen in the current time step are output, including: The conductive fluid distribution of the core fracturing specimen in the previous time step is used as the initial field for the current time step to determine the fracture morphology distribution in the core at the current time step. The distribution of fracture morphology in the core at the current time step is compared with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error, and it is determined whether the fitting error is greater than the set value. In response, the initial field of the previous time step is randomly adjusted within the set adjustment range, and the core fracture morphology distribution of the current time step is redefined. If no response is received, the output will show the distribution of conductive fluid and the distribution of fracture morphology within the core at the current time step.

4. The physical simulation method for fracturing fracture imaging using non-uniformly distributed contact electrodes according to claim 1 or 2, characterized in that, The core clamping device includes a clamping body and two end caps, which are respectively located at the front and rear ends of the clamping body. Each end cap has a fluid inlet / outlet. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure, and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes arranged at intervals. An insulating rubber block is placed between adjacent resistance electrodes. Four stress loading modules are respectively arranged on the outside of the five modal sensing array measurement layers.

5. The fracturing fracture imaging physical simulation method combined with non-uniformly distributed contact electrodes according to claim 4, characterized in that, Fracture propagation simulation was performed on a core fracturing specimen held in a core clamping device. During the fracture propagation simulation, a low-frequency excitation signal was used to perform resistance mode detection at the current time step to obtain the corresponding electrical signals, including: The core is processed into a core fracturing specimen, and a simulated wellbore is set in the center of the core fracturing specimen. The size of the core fracturing specimen is set according to the size of the core clamping device. The core fracturing specimen is placed inside the modal sensing array component, which is then placed in the clamping assembly. Insulating silicone oil is filled between the clamping assembly and the modal sensing array component via an external pump. By applying pressure to the external stress loading module, stress is provided to the core fracturing specimen while simultaneously compacting the modal sensing array component, ensuring that the resistance electrodes on the modal sensing array component are in complete contact with the core fracturing specimen. Conductive fluid is injected into the core fracturing specimen held in the core clamping device at a set flow rate or constant pressure. During the process of injecting the conductive fluid and driving the fracture propagation, the resistance mode is detected at the current time step using a low-frequency excitation signal to obtain the corresponding electrical signal.

6. The physical simulation method for fracturing fracture imaging based on non-uniformly distributed contact electrodes according to claim 1, 2, or 5, characterized in that, The process of resistance mode detection includes: In the modal sensing array component, a low-frequency excitation signal is injected into the diagonal resistor electrode pair, the potential difference between the remaining adjacent resistor electrodes is measured, the corresponding voltage gradient is obtained, and the equivalent resistance value is calculated. Repeat the above steps, applying current layer by layer to determine the equivalent resistance of each measurement layer of the multimodal sensing array.

7. A fracturing fracture imaging physical simulation system combining non-uniformly distributed contact electrodes, employing the method described in any one of claims 1 to 6, characterized in that, It includes a simulation detection section and a dynamic inversion section; The simulation and testing section includes a core clamping device and an excitation and testing device. The core clamping device holds the core fracturing specimen, and the core fracturing specimen clamped in the core clamping device is used to simulate crack propagation. During the crack propagation simulation, the excitation and testing device uses a low-frequency excitation signal to perform resistance mode detection at the current time step to obtain the corresponding electrical signal. The dynamic inversion part includes: The dynamic inversion unit takes the effective porosity of the core fracturing specimen, the initial electric field distribution, and the flow field data pre-calculated by Darcy's equation as the prior knowledge input of the conductive fluid distribution inversion model, and takes the electrical signal as the original data input of the conductive fluid distribution inversion model to obtain the conductive fluid distribution result of the core fracturing specimen at the current time step. The fracture fitting unit uses the conductive fluid distribution of the core fracturing specimen in the previous time step as the initial field for the current time step, determines the fracture morphology distribution in the core in the current time step, and compares the fracture morphology distribution in the core in the current time step with the conductive fluid distribution in the current time step. Based on the comparison results, it outputs the fracture morphology distribution and conductive fluid distribution results of the core fracturing specimen in the current time step.

8. The fracturing fracture imaging physical simulation system combined with non-uniformly distributed contact electrodes according to claim 7, characterized in that, The core clamping device includes a clamping body and two end caps, which are respectively located at the front and rear ends of the clamping body. Each end cap has a fluid inlet / outlet. The clamping body includes clamping components of the same length and both having a hollow quadrangular prism structure, and a modal sensing array component. The modal sensing array component is fitted inside the clamping component. The modal sensing array component includes five modal sensing array measurement layers arranged sequentially from top to bottom. Each modal sensing array measurement layer includes four sensing array measurement segments with the same structure. Adjacent sensing array measurement segments are connected by an insulating rubber block to form a closed modal sensing array measurement layer. Each sensing array measurement segment includes seven resistance electrodes arranged at intervals. An insulating rubber block is arranged between adjacent resistance electrodes. Four stress loading modules are arranged on the outside of the five modal sensing array measurement layers. or / and, The excitation and detection device includes an excitation source module and a detection module; The excitation source module is equipped with a frequency-controlled resistance excitation channel, which outputs a 10Hz–100kHz square wave current with a maximum current ≤20mA, and supports core axial and radial resistance gradient measurement. The detection module includes a resistance gradient measurement module, which performs resistance mode detection.

9. The fracturing fracture imaging physical simulation system according to claim 7 or 8, characterized in that, It also includes a model building unit for constructing a conductive fluid distribution inversion model, including: The sample acquisition module acquires several samples and divides them into training sample set and test sample set according to the proportion. Each sample includes the identification information of the conductive fluid distribution result, the effective porosity of the core of the historical core fracturing specimen, the initial electric field distribution and the flow field data pre-calculated by Darcy's equation, and the electrical signal corresponding to a certain time step in the crack propagation simulation of the historical core fracturing specimen. The training module trains the initial model using the training sample set. A loss function is introduced during training, and training ends when the value of the loss function stabilizes, resulting in a conductive fluid distribution inversion model. The initial model is a dual-channel U-Net network with residual connections and attention mechanisms, including dual channels, a feature fusion module, an encoder, a physical constraint module, and a decoder. The dual channels include a parallel prior knowledge input channel and a raw data input channel, which are used to input prior knowledge and electrical signals, respectively. The feature fusion module performs feature fusion and tensor construction on the prior knowledge and electrical signals. The encoder performs 4-level downsampling on the input fused features and retains microscopic boundary features through residual connections. During the encoding process, the physical constraint module applies physical constraints. The decoder performs feature recovery on the encoded fused features. The testing module uses a test sample set to test the conductive fluid distribution inversion model, optimizes the model parameters of the conductive fluid distribution inversion model, and outputs a conductive fluid distribution inversion model that meets the test evaluation requirements.

10. The fracturing fracture imaging physical simulation system according to claim 7 or 8, characterized in that, The fracturing fitting unit includes: The solution module uses the conductive fluid distribution results of the core fracturing specimen in the previous time step as the initial field for the current time step to determine the fracture morphology distribution in the core at the current time step. The fitting module compares the distribution of fracture morphology in the core at the current time step with the distribution of conductive fluid at the current time step to obtain the corresponding fitting error and determine whether the fitting error is greater than the set value. If the fitting output module responds to "yes", it will randomly adjust the initial field of the previous time step within the set adjustment range and return to redetermine the fracture morphology distribution in the core at the current time step; otherwise, it will output the conductive fluid distribution result and the fracture morphology distribution in the core at the current time step.

Images