Geological digital twinborn model construction method and system in oil exploration

Abstract

The invention relates to the technical field of oil exploration, and discloses a geological digital twinborn model construction method and system in oil exploration. The method comprises the steps of receiving a multi-source heterogeneous geological data flow of a target exploration area, performing multi-scale alignment and feature fusion through a geological feature decoupling model, and generating a three-dimensional geological attribute field distribution map and the like. Constructing an adaptive grid division model to generate a dynamic flow field simulation instruction set; simulating dynamic response under reservoir exploitation disturbance based on a multi-physics field coupling model and optimizing simulation instruction set parameters; and iteratively updating parameters through a quantum computing optimization framework, and outputting a model construction sequence to an exploration decision platform. The system comprises a data receiving module, a feature processing module, a grid instruction generation module, a simulation optimization module and an output module. Multi-source data are integrated, the geological process is accurately simulated, the exploration decision is optimized, and the oil exploration efficiency and exploitation benefits are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of oil exploration, and particularly to a method and system for constructing a geological digital twin model in oil exploration. Background Art

[0002] In the field of oil exploration, accurately grasping the underground geological structure and reservoir characteristics is the key to efficient exploitation of oil and gas resources. With the continuous growth of global energy demand, oil exploration is gradually expanding into complex geological areas, which poses many challenges to traditional exploration technologies.

[0003] Traditional oil exploration methods mainly rely on single data sources or simple data processing methods. For example, in the early days, only seismic reflection wave data was relied on to infer the underground geological structure, but the information obtained by this method is limited and it is difficult to accurately describe complex geological structures. Although core analysis can provide information on the physical properties of rocks, the number of core samples is limited and it is impossible to comprehensively reflect the characteristics of the entire reservoir. Moreover, these data are often isolated and not effectively integrated and analyzed, making it difficult for exploration personnel to grasp the geological situation as a whole.

[0004] With the development of technology, various multi-source data acquisition means have emerged continuously, such as data obtained from geophysical exploration, geological drilling, logging, etc., and the amount of data is increasing day by day. However, these data are heterogeneous, and there are significant differences in the formats, precisions, time scales, etc. of different types of data, making data fusion and analysis extremely difficult. For example, seismic reflection wave data reflects the macroscopic characteristics of the underground geological structure, while the core porosity distribution matrix focuses on the microscopic rock characteristics. How to effectively integrate and analyze these data of different scales and properties has become an important issue restricting the accuracy of oil exploration.

[0005] In addition, in terms of reservoir simulation, traditional models are often too simplified and cannot accurately simulate the migration laws of multiphase fluids under complex geological conditions. The fluid migration in the reservoir is affected by various factors, such as formation pressure gradient, rock permeability anisotropy, fluid viscosity, etc. Traditional models are difficult to comprehensively consider these factors, resulting in a large deviation between the simulation results and the actual situation. In terms of well location optimization and deployment, there is a lack of scientific and effective methods, and decisions are often made based on experience, which not only increases the exploration cost but also may miss the best exploitation area.

[0006] During the oil extraction process, the reservoir will be disturbed by the extraction activities, resulting in changes in its physical properties and fluid distribution. However, existing technologies are difficult to monitor these dynamic changes in real time and accurately predict them, and cannot provide timely and effective support for extraction decisions. Facing these problems, there is an urgent need for a technology that can integrate multi-source heterogeneous geological data, accurately simulate geological processes, and optimize exploration decisions. The geological digital twin model construction method and system emerge as the times require to meet the urgent needs of the modern oil exploration industry for high-precision and high-efficiency exploration technologies. Summary of the Invention

[0007] The purpose of the present invention is to provide a method and system for constructing a geological digital twin model in oil exploration to solve the problems raised in the above background technology.

[0008] To achieve the above purpose, the present invention provides the following technical solution: A method for constructing a geological digital twin model in oil exploration, the method includes: Receiving multi-source heterogeneous geological data streams of the target exploration area, the data streams include seismic reflection wave data, core porosity distribution matrix, and formation pressure gradient time series data; Based on a preset geological feature decoupling model, performing multi-scale alignment and feature fusion on the multi-source heterogeneous data to generate a three-dimensional geological attribute field distribution map and a reservoir heterogeneity tensor; According to the attribute field distribution map, constructing an adaptive grid division model to generate a dynamic flow field simulation instruction set, the instruction set includes multi-phase fluid migration path topology and well location optimization deployment strategy; Based on a preset multi-physical field coupling model, simulating the dynamic response of the target reservoir under extraction disturbance, and optimizing the multi-parameter collaborative constraint conditions in the simulation instruction set; Iteratively updating the collaborative constraint conditions through a quantum computing optimization framework, and outputting a geological digital twin model construction sequence to the exploration decision-making platform.

[0009] Preferably, the construction steps of the geological feature decoupling model include: Collecting historical exploration data sets of multiple blocks, and constructing a three-dimensional feature training set including formation fracture patterns, sedimentary facies spatial distributions, and rock mechanics parameters; Performing non-linear dimensionality reduction on the three-dimensional feature training set through a spectral clustering algorithm, and extracting independent representations of dominant geological factors and secondary disturbances; Combining with the geological mechanics constitutive equation, constructing partial differential constraint conditions for dynamic association between factors; Embedding the partial differential constraint conditions into a Bayesian inference network to generate the geological feature decoupling model that supports online update.

[0010] Preferably, the adaptive grid division model includes: Dynamically divide the density levels of unstructured grids according to the spatial gradient distribution of the three-dimensional geological attribute field distribution map; Calculate the adaptive refinement weight of grid cells based on the anisotropic tensor of reservoir permeability and the fluid viscosity coefficient; Non-linearly map the refinement weight and density level through the hyperbolic sine function to generate a multi-resolution nested grid division scheme; Trigger the initial boundary condition loading protocol of the fluid-solid coupling simulation according to the scheme.

[0011] Preferably, the steps for constructing the multi-physical field coupling model include: Collect the pressure-stress response correlation matrix in historical production data and construct a multi-field coupling causal diagram dataset; Extract the interaction strength parameters of the seepage field and the stress field through the random forest regression algorithm; Combine with an implicit solver to construct a multi-physical field dynamic balance equation and quantify the relaxation time constant of energy transfer between fields; Input the relaxation time constant and the real-time production disturbance factor into the cellular automaton network to generate the multi-physical field coupling model.

[0012] Preferably, the method further includes: Identify the sensitive areas of reservoir dynamic response according to the simulation results of the multi-physical field coupling model; Configure a real-time data assimilation strategy for the area in the dynamic flow field simulation instruction set; Based on the assimilation strategy, automatically generate a multi-scale observation data fusion scheme, including seismic wave velocity field correction instructions and well-to-well interpolation parameter update strategies.

[0013] Preferably, the calculation of the adaptive refinement weight of grid cells includes: Obtain the formation fracture density distribution map and the pore pressure diffusion coefficient matrix, and construct a grid quality evaluation hypercube; Calculate the singular value contribution degree of each dimension feature through the tensor decomposition algorithm; Perform a Hadamard product operation on the contribution degree and the hypercube to obtain a comprehensive refinement weight; Among them, the calculation formula of the comprehensive refinement weight is:

[0014] In the formula, represents the comprehensive refinement weight value, represents the singular value of the -th dimension feature, represents the geological attribute field function, represents the spatial coordinate component, represents the non - linear amplification exponent, represents the total number of feature dimensions, is the Hadamard product operator, representing the product of corresponding elements.

[0015] Preferably, the embedding of the partial differential constraint conditions includes: Conduct Lie group symmetry analysis on the non - linear dimensionality reduction result, and screen the correlation patterns that conform to the geomechanics conservation law; Generate the characteristic evolution path that satisfies the constitutive constraint through Markov chain Monte Carlo sampling; Use the path data to regularize the training of the hyperparameters of the Bayesian inference network to ensure that the model output conforms to the reservoir dynamic evolution law.

[0016] Preferably, the execution of the real - time data assimilation strategy includes: Establish the spatio - temporal covariance matrix of multi - source observation data and define the data fusion confidence score; Update the posterior distribution of the reservoir state estimate through the ensemble Kalman filter algorithm; Use the posterior distribution as the initial condition to input the next simulation time step to complete the closed - loop correction of dynamic data.

[0017] Preferably, the execution of the quantum computing optimization framework includes: Define the Hamiltonian function for multi - objective optimization, including the dual constraints of reservoir recovery rate improvement and energy consumption reduction; Solve the ground - state configuration of the Hamiltonian through the quantum annealing algorithm to obtain the optimal set of collaborative constraint parameters; Among them, the Hamiltonian function is:

[0018] In the formula, represents the Hamiltonian, represents the parameter and the coupling strength of represents the binary decision variable, represents the external field action coefficient.

[0019] Preferably, the present invention further includes a geological digital twin model construction system for oil exploration, and the features include: Data receiving module: used to receive multi - source heterogeneous geological data streams of the target exploration area, and the data streams include seismic reflection wave data, core porosity distribution matrix, and formation pressure gradient time - series data; Feature processing module: Based on the preset geological feature decoupling model, perform multi - scale alignment and feature fusion on the multi - source heterogeneous data to generate a three - dimensional geological attribute field distribution map and a reservoir heterogeneity tensor; Grid instruction generation module: According to the property field distribution map, construct an adaptive grid division model, and generate a dynamic flow field simulation instruction set, where the instruction set includes the topology of the multiphase fluid migration path and the well position optimization deployment strategy; Simulation optimization module: Based on a preset multi-physical field coupling model, simulate the dynamic response of the target reservoir under mining disturbances, and optimize the multi-parameter collaborative constraint conditions in the simulation instruction set; Output module: Iteratively update the collaborative constraint conditions through a quantum computing optimization framework, and output the geological digital twin model construction sequence to the exploration decision-making platform.

[0020] Compared with the prior art, the beneficial effects of the present invention are: The method and system for constructing a geological digital twin model in oil exploration proposed by the present invention significantly optimize the oil exploration process from multiple key levels, bringing many positive impacts to the development of the industry.

[0021] In terms of data processing and analysis, the invention can receive multi-source heterogeneous geological data streams in the target exploration area, including seismic reflection wave data, core porosity distribution matrix, and formation pressure gradient time series data, etc. Through a preset geological feature decoupling model, multi-scale alignment and feature fusion of these complex data are carried out to generate a three-dimensional geological property field distribution map and a reservoir heterogeneity tensor. This process breaks the limitations of traditional single data utilization, organically integrates different types and scales of data, provides more comprehensive and accurate geological information for exploration personnel, enables them to understand the underground geological structure and reservoir characteristics more deeply, and thus lays a solid foundation for subsequent exploration decision-making.

[0022] At the level of model construction and simulation, based on the generated property field distribution map, an adaptive grid division model is constructed, and then a dynamic flow field simulation instruction set including the topology of the multiphase fluid migration path and the well position optimization deployment strategy is generated. This adaptive grid division method fully considers the spatial variation of geological properties and can more accurately simulate the migration of multiphase fluids under complex geological conditions. At the same time, combined with a preset multi-physical field coupling model, the dynamic response of the target reservoir under mining disturbances is simulated, and the multi-parameter collaborative constraint conditions in the simulation instruction set are continuously optimized. This makes the simulation of reservoir dynamic changes closer to reality, effectively improves the prediction accuracy, and provides strong support for reasonably planning the mining plan.

[0023] In terms of optimizing decisions, the collaborative constraint conditions are iteratively updated through a quantum computing optimization framework, and the construction sequence of the geological digital twin model is output to the exploration decision-making platform. This optimization framework can quickly find the optimal set of collaborative constraint parameters under multi-objective constraints, such as achieving a balance between improving reservoir recovery rate and reducing energy consumption. With the help of these accurate models and optimized decision-making sequences, exploration personnel can plan well positions more scientifically, avoid blind exploitation, and effectively reduce exploration costs. During the exploitation process, according to the real-time monitoring data and model simulation results, the exploitation strategy is adjusted in a timely manner to improve the exploitation efficiency of oil and gas resources, increase the production volume, and thus enhance the economic benefits of the entire oil exploration and development project.

[0024] In terms of real-time monitoring and feedback adjustment, the sensitive areas of reservoir dynamic response are identified based on the simulation results of the multi-physics field coupling model, and a real-time data assimilation strategy is configured for them to automatically generate a multi-scale observation data fusion scheme. This enables the system to track the dynamic changes of the reservoir in real time, integrate the newly acquired data into the model in a timely manner, continuously update the model parameters, and ensure the accuracy and timeliness of the model. This closed-loop data processing and model update mechanism can timely detect problems occurring during the exploitation process and provide a scientific basis for adjusting the exploitation plan to ensure the safe and efficient progress of the exploitation activities. In short, this invention comprehensively improves the technical level and efficiency of oil exploration from data integration, model construction, decision optimization to real-time monitoring and feedback, and is of great significance for promoting the intelligent and efficient development of the oil exploration industry. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 It is the working principle diagram of the method for constructing a geological digital twin model in the oil exploration described in the present invention; Figure 2 It is the working principle diagram of the adaptive mesh generation model; Figure 3 It is the step diagram of the relevant expansion of the multi-physics field coupling model; Figure 4 It is the flowchart of the calculation of the adaptive refinement weight of grid cells. DETAILED DESCRIPTION OF THE INVENTION

[0026] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0027] Please refer to Figures 1 - 4The present invention provides a method and system for constructing a geological digital twin model in petroleum exploration to achieve accurate simulation and analysis of the geological conditions in the target exploration area and provide strong support for exploration decision-making. The specific implementation steps are as follows: Obtain multi-source heterogeneous geological data streams in the target exploration area, including seismic reflection wave data, core porosity distribution matrix, and formation pressure gradient time series data. Seismic reflection wave data is collected through seismic exploration technology, which records the seismic wave information reflected from different underground formation interfaces and can be used to infer the structure and construction of the formation; the core porosity distribution matrix is ​​obtained by analyzing and testing the actual collected core samples, reflecting the distribution of rock pores, which is crucial for evaluating the reservoir performance of the reservoir; the formation pressure gradient time series data records the changes in formation pressure over time, which helps to understand the flow law of underground fluids.

[0028] The pre-set geological feature decoupling model is used to process the collected multi-source heterogeneous data. Through multi-scale alignment, data from different sources and scales are unified in time and space to make the data comparable. Then feature fusion is performed to integrate the feature information contained in various geological data, and finally a three-dimensional geological attribute field distribution map and reservoir heterogeneity tensor are generated. The three-dimensional geological attribute field distribution map intuitively shows the distribution of various attributes of underground geological bodies in three-dimensional space, such as lithology, porosity, etc.; the reservoir heterogeneity tensor quantitatively describes the heterogeneity of the reservoir, providing an important basis for subsequent reservoir evaluation and development.

[0029] Based on the generated three-dimensional geological attribute field distribution map, an adaptive grid division model is constructed. This model divides the grid reasonably according to the spatial variation characteristics of the geological attribute field, making the grid more refined in areas with drastic geological changes and relatively sparse in areas with gentle changes, thereby improving the accuracy and efficiency of the simulation. The model generates a dynamic flow field simulation instruction set, which includes the multiphase fluid migration path topology and well location optimization deployment strategy. The multiphase fluid migration path topology describes the flow paths and relationships of various fluids such as oil, gas, and water in underground reservoirs; the well location optimization deployment strategy determines the optimal drilling location and mining plan based on geological conditions and fluid migration laws to improve the recovery rate.

[0030] The preset multi-physics coupling model is used to simulate the dynamic response of the target reservoir under mining disturbance. Considering that the mining process will cause changes in physical fields such as formation pressure and stress, the multi-physics coupling model can comprehensively consider the interaction between these factors and accurately simulate the dynamic changes of the reservoir. During the simulation process, the multi-parameter collaborative constraints in the simulation instruction set are optimized to ensure that the simulation results are more in line with the actual situation.

[0031] With the help of the quantum computing optimization framework, the collaborative constraint conditions are iteratively updated. Quantum computing has powerful parallel computing capabilities and can search for better solutions in a short time. Through continuous iteration, the optimal collaborative constraint conditions are found, and finally the construction sequence of the geological digital twin model is output to the exploration decision-making platform. The exploration decision-making platform can formulate a more scientific and reasonable exploration and development plan based on these construction sequences, improving exploration efficiency and economic benefits.

[0032] The following further illustrates the implementation of the present invention in combination with Examples 1 to 6. Example 1:

[0033] In the process of constructing the geological feature decoupling model, a large amount of data collection work needs to be carried out first. Historical exploration data sets are collected from multiple different blocks. These data sets contain rich geological information. Among them, the formation fracture pattern records information such as the distribution, trend, and density of fractures in the formation. These fractures have an important impact on the migration of underground fluids and the connectivity of reservoirs; the spatial distribution of sedimentary facies describes the spatial distribution law of rock facies formed under different sedimentary environments. Different sedimentary facies have different reservoir properties; rock mechanical parameters reflect the physical and mechanical properties of rocks, such as elastic modulus, Poisson's ratio, etc. These parameters are crucial for analyzing the stability and deformation characteristics of the formation. Integrate this information to construct a three-dimensional feature training set containing the formation fracture pattern, spatial distribution of sedimentary facies, and rock mechanical parameters.

[0034] Use the spectral clustering algorithm to perform non-linear dimensionality reduction on the three-dimensional feature training set. The spectral clustering algorithm is a clustering algorithm based on graph theory, which can discover complex clustering structures in high-dimensional data. Through this algorithm, high-dimensional geological feature data can be mapped into a low-dimensional space while retaining the main feature information of the data. During the dimensionality reduction process, the independent representations of the dominant geological factors and secondary perturbations are extracted. The dominant geological factors represent the main trends and laws in geological features, while the secondary perturbations reflect some local and minor change factors.

[0035] Combined with the geomechanics constitutive equation, partial differential constraint conditions for the dynamic association between factors are constructed. The geomechanics constitutive equation describes the deformation and failure laws of rocks under the action of forces. By combining it with the geological factors obtained through dimensionality reduction, a dynamic association relationship between the factors can be established. These partial differential constraint conditions reflect the physical laws in geological processes and play a key role in the accuracy of subsequent models.

[0036] Embed partial differential constraint conditions into the Bayesian inference network to generate a decoupled geological feature model that supports online update. The Bayesian inference network is a machine learning model based on probabilistic inference, which can continuously update the model's parameters and prediction results according to new observed data. By embedding partial differential constraint conditions into it, when the model performs feature decoupling, it can not only consider the statistical features of the data but also follow the physical laws of geomechanics. In practical applications, as new geological data is continuously collected, the model can be updated online to improve the decoupling accuracy of geological features and better adapt to different exploration situations. Example 2:

[0037] When constructing an adaptive mesh generation model, dynamically divide the density level of unstructured meshes according to the spatial gradient distribution of the three-dimensional geological attribute field map. The spatial gradient of the geological attribute field reflects the rate of change of the geological attribute in space. In areas with large gradients, such as near formation boundaries and faults, the geological attribute changes violently, and denser meshes need to be divided to accurately describe its changes; while in areas with small gradients, the geological attribute changes relatively gently, and sparser meshes can be used to reduce the computational amount.

[0038] Based on the anisotropic tensor of reservoir permeability and the fluid viscosity coefficient, calculate the adaptive refinement weight of grid cells. Obtain the formation fracture density distribution map and the pore pressure diffusion coefficient matrix, and construct a hypercube for grid quality assessment. The formation fracture density distribution map reflects the distribution of fractures in the formation. Fractures will affect the seepage of fluids. The greater the fracture density, the more complex the fluid seepage, and the higher the requirement for the fineness of the grid; the pore pressure diffusion coefficient matrix describes the diffusion ability of pore pressure in the medium. The larger its value, the wider the influence range of pressure changes on the surrounding area, and finer meshes are also needed to simulate.

[0039] Calculate the singular value contribution degree of each dimension feature through the tensor decomposition algorithm. The tensor decomposition algorithm can decompose a high-order tensor into a combination of multiple low-order tensors. By calculating the singular value contribution degree, the importance of each dimension feature to the overall grid quality can be determined.

[0040] Perform the Hadamard product operation on the contribution degree and the hypercube to obtain the comprehensive refinement weight, and its calculation formula is:

[0041] In this formula, represents the comprehensive refinement weight value, which comprehensively considers the influence degree of each dimension feature on grid refinement. The larger the value, the more the grid cell needs to be refined. represents the singular value of the th dimension feature. The larger the singular value, the greater the contribution of this dimension feature to the overall feature. represents the geological attribute field function, which describes the spatial distribution of geological attributes; represents the spatial coordinate components, which are used to determine the characteristic changes in different directions; represents the non - linear amplification index, which is used to adjust the influence degree of different characteristic changes on the weight. By reasonably setting the value, the weight calculation can be made more in line with the actual geological situation; represents the total number of characteristic dimensions, covering all geological characteristic dimensions involved in the calculation; is the Hadamard product operator, which represents the product of corresponding elements. Through this operation method, the contribution degree of characteristics in each dimension is combined with the change rate of the geological attribute field function to obtain the comprehensive refined weight of each grid cell.

[0042] The refined weight and the density level are non - linearly mapped through the hyperbolic sine function to generate a multi - resolution nested grid division scheme. The hyperbolic sine function can perform non - linear transformation on data, making the relationship between the refined weight and the density level more in line with the geological actual situation. The multi - resolution nested grid division scheme generated in this way can effectively reduce the computational amount while ensuring the simulation accuracy.

[0043] According to this scheme, trigger the initial boundary condition loading protocol for the fluid - solid coupling simulation. The fluid - solid coupling simulation takes into account the interaction between the fluid and the solid. In oil exploration, there are complex interaction relationships between the reservoir rock and the fluid in it. By loading appropriate initial boundary conditions, this interaction can be accurately simulated, providing a reliable basis for the subsequent dynamic flow field simulation. Example 3:

[0044] When constructing a multi - physical - field coupling model, first collect the pressure - stress response correlation matrix in the historical production data to construct a multi - field coupling causal graph data set. The historical production data contains the change information of formation pressure and stress under different production conditions. The pressure - stress response correlation matrix records the mutual relationship between the pressure change and the stress change. By analyzing these data, the causal connections between multi - physical fields can be found. Constructing a multi - field coupling causal graph data set represents these causal relationships in a graphical way, providing an intuitive data structure for subsequent extraction of interaction intensity parameters.

[0045] Extract the interaction strength parameters of the seepage field and the stress field through the random forest regression algorithm. The random forest regression algorithm is a machine learning algorithm based on decision trees. By constructing multiple decision trees and making comprehensive predictions, it can effectively handle high-dimensional data and complex non-linear relationships. In the problem of multi-physical field coupling, there are complex interactions between the seepage field and the stress field, and the random forest regression algorithm can extract parameters reflecting the intensity of this interaction from a large amount of historical data.

[0046] Combine an implicit solver to construct a multi-physical field dynamic equilibrium equation and quantify the relaxation time constant of the energy transfer between fields. The implicit solver has the advantages of good stability and high accuracy in numerical calculations and can effectively solve the multi-physical field dynamic equilibrium equation. When constructing the equation, considering the energy transfer between the seepage field and the stress field, a relaxation time constant is introduced to quantify the speed and efficiency of this energy transfer. The relaxation time constant reflects the response speed of one physical field to the change of another physical field. By accurately determining this constant, the coupling process between multi-physical fields can be simulated more precisely.

[0047] Input the relaxation time constant and the real-time mining disturbance factor into the cellular automaton network to generate a multi-physical field coupling model. The cellular automaton network is a discrete dynamic system model composed of multiple cells, and each cell updates its state according to the states of its surrounding cells and certain rules. In the multi-physical field coupling model, taking the relaxation time constant and the real-time mining disturbance factor as inputs, the cellular automaton network can simulate the dynamic changes of multi-physical fields in space and time, thus generating a multi-physical field coupling model that can accurately reflect the actual situation. Example 4:

[0048] Identify the sensitive areas of reservoir dynamic response according to the simulation results of the multi-physical field coupling model. During the simulated mining process, the multi-physical field coupling model will output the changes in physical quantities such as pressure and stress in each area of the reservoir. By analyzing these simulation results, find the areas where the physical quantity changes significantly and is more sensitive to mining disturbances. These sensitive areas may be the key channels for fluid flow in the reservoir or the areas prone to formation deformation or damage. Paying attention to and studying these areas is crucial for optimizing the mining plan.

[0049] Configure a real-time data assimilation strategy for sensitive areas in the dynamic flow field simulation instruction set. The purpose of the real-time data assimilation strategy is to combine the real-time collected observation data with the simulation model to improve the accuracy of the simulation. First, establish the spatio-temporal covariance matrix of multi-source observation data and define the data fusion confidence score. The spatio-temporal covariance matrix describes the correlation between observation data at different times and spaces. By calculating the covariance matrix, the reliability of the observation data and the degree of influence between them can be understood. According to the covariance matrix, define the data fusion confidence score, which is used to measure the importance of each observation data in the data fusion process.

[0050] Update the posterior distribution of the reservoir state estimate through the Ensemble Kalman Filter algorithm. The Ensemble Kalman Filter algorithm is a commonly used data assimilation method that continuously updates the estimate of the reservoir state by combining the observation data with the model prediction results. In this process, using the established spatio-temporal covariance matrix and the data fusion confidence score, the observation data is weighted and then incorporated into the model prediction to obtain the posterior distribution of the reservoir state estimate.

[0051] Use the posterior distribution as the initial condition to input the next simulation time step to complete the closed-loop correction of dynamic data. In this way, in each simulation time step, the latest observation data can be used to correct the simulation results, enabling the simulation model to better track the actual dynamic changes of the reservoir and improving the accuracy and reliability of the simulation. At the same time, according to the results of the closed-loop correction of dynamic data, the production plan can be adjusted in a timely manner, optimizing the topology of the multiphase fluid migration path and the well location optimization deployment strategy, and improving the efficiency and safety of oil production. Example 5:

[0052] In the process of embedding the partial differential constraint conditions into the Bayesian inference network, first perform Lie group symmetry analysis on the nonlinear dimensionality reduction results. Lie group symmetry analysis is a method for studying the symmetry of mathematical and physical systems. In the geological feature decoupling model, by performing Lie group symmetry analysis on the nonlinear dimensionality reduction results, the correlation patterns that conform to the geomechanics conservation laws can be found. The geomechanics conservation laws include mass conservation, momentum conservation, energy conservation, etc., and these laws play an important constraining role in geological processes. By analyzing the symmetry, it can be found which correlations between geological factors conform to these conservation laws, thereby screening out effective correlation patterns.

[0053] Generate the characteristic evolution path that satisfies the constitutive constraints through Markov chain Monte Carlo sampling. Markov chain Monte Carlo sampling is a stochastic sampling method based on Markov chains, which can perform efficient sampling in high-dimensional spaces. In this embodiment, using the Markov chain Monte Carlo sampling method, according to the geomechanics constitutive equation and the selected correlation patterns, generate the characteristic evolution path that satisfies the constitutive constraints. These characteristic evolution paths reflect the change process of geological features in time and space and are the key to constructing an accurate geological feature decoupling model.

[0054] Use the path data to perform regularization training on the hyperparameters of the Bayesian inference network. The hyperparameters of the Bayesian inference network have an important impact on the performance of the model. Through regularization training, overfitting of the model can be avoided and the generalization ability of the model can be improved. During the training process, use the generated characteristic evolution path data as input to adjust and optimize the hyperparameters of the Bayesian inference network to ensure that the model output conforms to the reservoir dynamic evolution law. For example, when adjusting the hyperparameters, methods such as cross-validation can be used to select the optimal combination of hyperparameters so that the model can perform well on different data sets. In this way, the partial differential constraint conditions are effectively embedded in the Bayesian inference network, enabling the geological feature decoupling model to better reflect the actual geological situation and providing reliable support for subsequent geological analysis and exploration decisions. Example 6:

[0055] When executing the quantum computing optimization framework, first define the Hamiltonian function for multi-objective optimization, which includes the dual constraints of reservoir recovery rate improvement and energy consumption reduction. The Hamiltonian function is an important concept in quantum computing, which describes the energy state of a quantum system. In the present invention, in order to achieve multi-objective optimization in oil exploration, the improvement of reservoir recovery rate and the reduction of energy consumption are used as two important optimization objectives to construct the Hamiltonian function.

[0056] The expression of the Hamiltonian function is: . In this formula, represents the Hamiltonian, which represents the energy of the quantum system. In this application scenario, the smaller the value of, the closer it is to the optimal combination of collaborative constraint conditions, that is, it can better balance the two objectives of reservoir recovery rate improvement and energy consumption reduction; represents the parameter and the coupling strength between, which reflects the interaction relationship between different parameters. Different geological parameters may be correlated in terms of affecting reservoir recovery rate and energy consumption. The larger the value of, the stronger the coupling effect between the parameter and , and their influence needs to be considered more comprehensively during the optimization process; represents a binary decision variable, which is used to represent the value states of different parameters. In actual optimization, by adjusting the values of these binary decision variables, the optimal parameter combination can be found; represents the external field action coefficient, which takes into account the influence of external factors on the system. In oil exploration, external factors may include extraction technology, geological environment, etc. These factors will affect the reservoir recovery rate and energy consumption. The value of is determined according to specific external conditions.

[0057] The ground state configuration of the Hamiltonian is solved by the quantum annealing algorithm to obtain the optimal set of cooperative constraint parameters. The quantum annealing algorithm is an optimization algorithm based on the principles of quantum mechanics, which can find the global optimal solution in a complex energy landscape. In this embodiment, the defined Hamiltonian function is solved using the quantum annealing algorithm. By continuously adjusting the state of the quantum system, it gradually approaches the ground state configuration. In the ground state configuration, the corresponding binary decision variable The value of is the optimal set of cooperative constraint parameters. These parameter sets can be used to optimize the multi-parameter cooperative constraint conditions in the simulation instruction set, thereby improving the accuracy and practicality of the geological digital twin model and providing a more scientific basis for oil exploration decisions. For example, in practical applications, the obtained optimal set of cooperative constraint parameters can be applied to multi-physics field coupling models and adaptive mesh generation models to further optimize the simulation results of the models, improve the prediction accuracy of reservoir dynamic responses, and provide more reliable support for the formulation of oil extraction plans.

[0058] It should be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or device.

[0059] Although the embodiments of the present invention have been shown and described, it will be understood by those of ordinary skill in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for constructing a geological digital twin model in petroleum exploration, characterized in that: include: Receiving a multi-source heterogeneous geological data stream of a target exploration area, wherein the data stream includes seismic reflection wave data, a core porosity distribution matrix, and formation pressure gradient time series data; Based on a preset geological feature decoupling model, multi-scale alignment and feature fusion are performed on the multi-source heterogeneous data to generate a three-dimensional geological attribute field distribution map and a reservoir heterogeneity tensor; According to the attribute field distribution map, an adaptive grid partitioning model is constructed to generate a dynamic flow field simulation instruction set, wherein the instruction set includes a multiphase fluid migration path topology and a well location optimization deployment strategy; Based on a preset multi-physics field coupling model, the dynamic response of the target reservoir under mining disturbance is simulated, and the multi-parameter collaborative constraint conditions in the simulation instruction set are optimized; The collaborative constraints are iteratively updated through a quantum computing optimization framework, and the geological digital twin model construction sequence is output to the exploration decision-making platform.

2. The method for constructing a geological digital twin model according to claim 1, characterized in that: The steps of constructing the geological characteristic decoupling model include: Collect historical exploration data sets from multiple blocks and construct a three-dimensional feature training set including formation fracture patterns, spatial distribution of sedimentary phases, and rock mechanical parameters; Performing nonlinear dimensionality reduction on the three-dimensional feature training set by using a spectral clustering algorithm to extract independent representations of dominant geological factors and secondary disturbances; Combined with the geomechanical constitutive equation, partial differential constraints for dynamic correlation between factors are constructed; The partial differential constraint condition is embedded in a Bayesian inference network to generate the geological feature decoupling model supporting online updating.

3. The method for constructing a geological digital twin model according to claim 1, characterized in that: The adaptive meshing model includes: Dynamically dividing the density level of the unstructured grid according to the spatial gradient distribution of the three-dimensional geological attribute field distribution map; Based on the reservoir permeability anisotropy tensor and fluid viscosity coefficient, the grid cell adaptive refinement weight is calculated; The refinement weight and the density level are nonlinearly mapped through a hyperbolic sine function to generate a multi-resolution nested grid division scheme; The initial boundary condition loading protocol for fluid-structure interaction simulation is triggered according to the scheme.

4. The method for constructing a geological digital twin model according to claim 1, characterized in that: The steps of constructing the multi-physics field coupling model include: Collect the pressure-stress response correlation matrix from historical mining data and construct a multi-field coupling causal diagram data set; The interaction intensity parameters between the seepage field and the stress field are extracted by using the random forest regression algorithm; Combined with implicit solvers, the multi-physics dynamic equilibrium equations are constructed to quantify the relaxation time constants of energy transfer between fields; The relaxation time constant and the real-time mining disturbance factor are input into a cellular automaton network to generate the multi-physics coupling model.

5. The method for constructing a geological digital twin model according to claim 4, characterized in that: Also includes: According to the simulation results of the multi-physics field coupling model, identifying the reservoir dynamic response sensitive area; configuring a real-time data assimilation strategy for the region in the dynamic flow field simulation instruction set; Based on the assimilation strategy, a multi-scale observation data fusion scheme is automatically generated, including seismic wave velocity field correction instructions and inter-well interpolation parameter update strategy.

6. The method for constructing a geological digital twin model according to claim 3, characterized in that: The calculation of the grid unit adaptive refinement weight includes: Obtain formation fracture density distribution map and pore pressure diffusion coefficient matrix, and construct grid quality assessment hypercube; The singular value contribution of each dimensional feature is calculated through the tensor decomposition algorithm; Performing a Hadamard product operation on the contribution degree and the hypercube to obtain a comprehensive refinement weight; The calculation formula of the comprehensive refinement weight is: ; In the formula, represents the comprehensive refinement weight value, Indicates The singular values ​​of the dimensional features, represents the geological attribute field function, represents the spatial coordinate component, represents the nonlinear amplification exponent, represents the total number of feature dimensions, is the Hadamard product operator, which means the product of corresponding elements.

7. The method for constructing a geological digital twin model according to claim 2, characterized in that: The embedding of the partial differential constraint condition includes: Conduct Lie group symmetry analysis on the independent representations of the dominant geological factors and secondary disturbances to screen out association patterns that conform to the conservation laws of geomechanics; Generate feature evolution paths that satisfy constitutive constraints through Markov chain Monte Carlo sampling; The path data is used to conduct regularized training on the hyperparameters of the Bayesian inference network to ensure that the model output conforms to the dynamic evolution law of the reservoir.

8. The method for constructing a geological digital twin model according to claim 5, characterized in that: The execution of the real-time data assimilation strategy includes: Establish the spatiotemporal covariance matrix of multi-source observation data and define the data fusion confidence score; Updating the posterior distribution of reservoir state estimates by integrating the Kalman filter algorithm; The posterior distribution is input as the initial condition into the next simulation time step to complete the dynamic data closed-loop correction.

9. The method for constructing a geological digital twin model according to claim 1, characterized in that: The implementation of the quantum computing optimization framework includes: Define the Hamiltonian function for multi-objective optimization, including the dual constraints of increasing reservoir recovery and reducing energy consumption; Solving the ground state configuration of the Hamiltonian by a quantum annealing algorithm to obtain an optimal set of cooperative constraint parameters; Wherein, the Hamiltonian function is: ; In the formula, represents the Hamiltonian, Representation parameters and The coupling strength, represents a binary decision variable, Represents the external field effect coefficient.

10. A geological digital twin model construction system for petroleum exploration, characterized in that: include: Data receiving module: used to receive multi-source heterogeneous geological data streams in the target exploration area, the data streams including seismic reflection wave data, core porosity distribution matrix and formation pressure gradient time series data; Feature processing module: Based on the preset geological feature decoupling model, multi-scale alignment and feature fusion are performed on the multi-source heterogeneous data to generate a three-dimensional geological attribute field distribution map and a reservoir heterogeneity tensor; Grid instruction generation module: construct an adaptive grid partitioning model according to the attribute field distribution map, and generate a dynamic flow field simulation instruction set, wherein the instruction set includes a multiphase fluid migration path topology and a well location optimization deployment strategy; Simulation optimization module: based on a preset multi-physics field coupling model, simulates the dynamic response of the target reservoir under mining disturbance, and optimizes the multi-parameter collaborative constraints in the simulation instruction set; Output module: Iteratively update the collaborative constraints through the quantum computing optimization framework, and output the geological digital twin model construction sequence to the exploration decision-making platform.

Images