A method and system for determining a reservoir fracturing interwell interference prevention solution
By constructing an analysis model based on historical fracture boundary information and interference information, and combining machine learning and knowledge graph technology, prevention and control schemes between reservoir fracturing wells are generated and stored. This solves the problems of cumbersome analysis and poor data security in existing methods, and achieves efficient and accurate prevention and control scheme design.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA PETROLEUM & CHEMICAL CORP
- Filing Date
- 2024-11-29
- Publication Date
- 2026-05-29
AI Technical Summary
Existing methods for preventing interference between reservoir fracturing wells involve cumbersome analysis processes and lack determination of interference conditions, resulting in low matching between prevention schemes and actual conditions, low data processing efficiency, poor security, and inability to guarantee the accuracy of analysis results.
Based on historical fracture boundary information and disturbance information of the target reservoir, a disturbance information analysis model and a prevention and control data analysis model are constructed. Machine learning and knowledge graph technologies are used to generate accurate prevention and control plans, and data security is ensured through blockchain storage.
It enables accurate acquisition of reservoir fracturing well interference prevention and control schemes, shortens acquisition time, improves the convenience of the analysis process and the accuracy and efficiency of prevention and control scheme design, and ensures data security.
Smart Images
Figure CN122113553A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of fracturing well construction technology, and in particular relates to a method and system for determining the interference prevention scheme between reservoir fracturing wells. Background Technology
[0002] Shale reservoirs are characterized by self-generating and self-storing properties, low porosity, ultra-low permeability, and well-developed natural fractures. They generally lack natural production capacity and require hydraulic fracturing and horizontal well technology for economical development. However, due to the poor ability of the matrix to supply oil and gas to the fractures, it is difficult to achieve the expected production enhancement effect by fracturing a single main fracture. Therefore, special fracturing techniques (such as "fracture network fracturing" and "volume modification") are often required to create artificial fractures that branch or change direction, thereby connecting with natural fractures and increasing the complexity of the fractures. This improves the ability of the reservoir matrix to supply oil and gas to the artificial fractures, resulting in better production enhancement.
[0003] In the petroleum industry, "fracking" refers to a method of creating fractures in oil and gas reservoirs using hydraulic force during oil or gas production; it is also known as "hydraulic fracturing." Fracking is a way to artificially create fractures in the formation, improve the underground flow environment of oil, and increase oil well production. It plays an important role in improving bottomhole flow conditions, mitigating inter-layer flow, and improving reservoir activation.
[0004] In developing this invention, the inventors discovered that existing methods for preventing and controlling interference between reservoir fracturing wells are cumbersome in their analysis of actual interference conditions. Furthermore, they lack the ability to accurately determine these conditions, or rely solely on manual determination. Consequently, they cannot provide pre-defined prevention and control solutions, or the provided solutions have a low degree of compatibility with actual interference conditions, failing to guarantee the efficiency of subsequent prevention and control design. In addition, existing methods for preventing and controlling interference between reservoir fracturing wells also suffer from low data processing efficiency and poor data security, failing to ensure the accuracy of subsequent prevention and control analysis results. Summary of the Invention
[0005] To address the aforementioned problems, this invention provides a method for determining inter-well interference prevention and control schemes in reservoir fracturing, comprising: determining historical interference information that forms the historical fracture boundary information based on historical fracture boundary information of the target reservoir, and constructing an interference information analysis model; using historical interference prevention and control data for eliminating interference caused by the historical interference information, and combining the historical interference information, constructing a prevention and control data analysis model; obtaining the current fracture boundary information of the target reservoir, and using the interference information analysis model to obtain interference information analysis results, and then using the prevention and control data analysis model to determine the corresponding prevention and control data to form a prevention and control scheme.
[0006] Preferably, the step of acquiring the current fracture boundary information of the target reservoir includes: acquiring reservoir fracturing well images containing the fracture boundary information at different points on the target reservoir, and using the acquired reservoir fracturing well images as corresponding data to be transmitted; determining the time required for each point to complete data transmission based on the location and data transmission rate of each point, and forming a node queue for data transmission trajectories according to the order of data transmission completion; grouping the different points according to the type of the reservoir fracturing well images, and combining the node queues to obtain a set of data transmission trajectories for each point combination; using the set of data transmission trajectories as a population and the corresponding points as individuals, thereby using a genetic algorithm to generate a population matrix to obtain the optimal data transmission trajectory for transmitting reservoir fracturing well images.
[0007] Preferably, before constructing the interference information analysis model, the method further includes: collecting historical images of the target reservoir containing fractures, and dividing the acquired historical fracture images into blocks according to the display ratio; using Fourier transform to extract image blocks affected by noise in the block-processed historical fracture images; smoothing the extracted image blocks by Gaussian filtering, and then integrating the smoothed image blocks with the image blocks not affected by noise to obtain the historical fracture boundary information.
[0008] Preferably, the step of determining the historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir includes: screening fracture regions in the historical fracture boundary information that have fracture displacement and generate loads to balance fracture opening, and discretizing the fractures in the fracture regions to obtain multiple fracture boundary units; determining the boundary conditions corresponding to each fracture boundary unit based on the boundary distribution state when the internal pressure of each fracture in the fracture region is uniform; determining the stress characteristics at the midpoint of the fracture inside the corresponding fracture boundary unit by combining the displacement discontinuity of each fracture boundary unit; and using the stress characteristics as the corresponding historical interference information.
[0009] Preferably, the boundary conditions are expressed using the following expression:
[0010]
[0011] Where p represents the net pressure of the crack, s and n represent the horizontal and vertical axes respectively, j represents the serial number of the crack boundary element, N represents the total amount of the crack boundary segment elements, and σ represents the stress.
[0012] Preferably, the step of constructing a prevention and control data analysis model using historical interference prevention and control data to eliminate interference caused by historical interference information includes: normalizing the historical interference prevention and control data and screening qualified prevention and control data whose characterization ability meets the requirements for constructing the prevention and control data analysis model by calculating the variance coefficient of each normalized data; dividing the qualified prevention and control data into a training set and a test set, training a preset machine learning model using the training set to obtain the prevention and control data analysis model, and then using the test set to evaluate the analysis effect of the current prevention and control data analysis model; determining whether the current prevention and control data analysis model is the best prevention and control data analysis model based on the analysis effect evaluation result, and if it is not determined, continuing to train and determine the current prevention and control data analysis model until the best prevention and control data analysis model is obtained.
[0013] Preferably, the process of obtaining the optimal prevention and control data analysis model includes: dividing the training set into multiple sub-training sets, and using each sub-training set to sequentially train a preset machine learning model to obtain the optimal prevention and control data analysis model. After each sub-training set is trained, the accuracy of the analysis results of the current prevention and control data analysis model is obtained using the test set until a prevention and control data analysis model with an analysis result accuracy reaching a preset accuracy threshold is obtained, and this model is taken as the optimal prevention and control data analysis model.
[0014] Preferably, the step of forming a prevention and control plan includes: acquiring vectorized historical interference prevention and control data, and using the TransD algorithm to construct a topic text knowledge subgraph about the interference prevention and control plan; constructing a knowledge graph embedding model based on the topic text knowledge subgraph, and based on this, obtaining the entity table and relation table corresponding to the historical interference prevention and control data to construct a prevention and control graph library; retrieving current prevention and control data in the prevention and control graph library to obtain the corresponding prevention and control plan, wherein if the current prevention and control data is included in the prevention and control graph library, the prevention and control plan is directly formed based on the corresponding part in the prevention and control graph library; otherwise, a new prevention and control plan is formulated based on the current prevention and control data, and the prevention and control graph library is updated based on the new prevention and control plan.
[0015] Preferably, the method further includes: processing the construction parameters corresponding to the determined prevention and control scheme into several blocks, and storing the determined prevention and control scheme using a block-based storage method.
[0016] On the other hand, the present invention also provides a system for determining inter-well interference prevention and control schemes for reservoir fracturing. The system is used to execute a method for determining inter-well interference prevention and control schemes for reservoir fracturing. The system includes: a first model construction module, which is used to determine historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir, and construct an interference information analysis model; a second model construction module, which is used to construct a prevention and control data analysis model by combining historical interference prevention and control data used to eliminate interference caused by the historical interference information with the historical interference information; and a prevention and control scheme generation module, which is used to obtain the current fracture boundary information of the target reservoir, obtain interference information analysis results using the interference information analysis model, and then determine the corresponding prevention and control data using the prevention and control data analysis model to form a prevention and control scheme.
[0017] Compared with the prior art, one or more embodiments of the above solutions may have the following advantages or beneficial effects:
[0018] This invention provides a method and system for determining inter-well interference prevention and control schemes in reservoir fracturing. The method constructs an interference information analysis model based on historical fracture boundary information of the target reservoir and historical interference information that forms the historical fracture boundary information. Then, based on historical interference prevention and control data used to eliminate interference caused by historical interference information, and historical interference information, a prevention and control data analysis model is constructed. Finally, based on the constructed analysis models and using the current fracture boundary information of the target reservoir, the corresponding prevention and control data is determined to form a prevention and control scheme. This invention saves time in acquiring prevention and control schemes and makes the analysis process more convenient, achieving accurate acquisition of inter-well interference prevention and control schemes in reservoir fracturing, and ensuring the accuracy and efficiency of subsequent prevention and control scheme design.
[0019] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description
[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:
[0021] Figure 1 This is a step diagram of a method for determining a reservoir fracturing well interference prevention scheme according to an embodiment of this application.
[0022] Figure 2 This is a schematic diagram of the overall structure of a system for determining inter-well interference prevention schemes in reservoir fracturing, according to an embodiment of this application. Detailed Implementation
[0023] The embodiments of the present invention will be described in detail below with reference to the accompanying drawings and examples, so that the process of how the present invention uses technical means to solve technical problems and achieve technical effects can be fully understood and implemented accordingly. It should be noted that, as long as there is no conflict, the various embodiments and features in the various embodiments of the present invention can be combined with each other, and the resulting technical solutions are all within the protection scope of the present invention.
[0024] Furthermore, the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.
[0025] Existing methods for preventing inter-well interference during reservoir fracturing involve cumbersome analysis of actual interference conditions and lack the ability to accurately determine these conditions. They often rely solely on manual determination, resulting in inadequate pre-defined prevention solutions or low accuracy in matching actual interference conditions, thus compromising the efficiency of subsequent prevention scheme design. Furthermore, existing methods suffer from low data processing efficiency and poor data security, failing to guarantee the accuracy of subsequent prevention analysis results.
[0026] Therefore, to address the aforementioned problems, this invention proposes a method and system for determining inter-well interference prevention and control schemes in reservoir fracturing. This method constructs an interference information analysis model based on historical fracture boundary information of the target reservoir and historical interference information that forms the historical fracture boundary information. Then, based on historical interference prevention and control data used to eliminate interference caused by historical interference information, and historical interference information, a prevention and control data analysis model is constructed. Finally, based on the constructed analysis models and utilizing the current fracture boundary information of the target reservoir, the corresponding prevention and control data is determined to form a prevention and control scheme. This invention saves time in acquiring prevention and control schemes and makes the analysis process more convenient, achieving accurate acquisition of inter-well interference prevention and control schemes in reservoir fracturing, and ensuring the accuracy and efficiency of subsequent prevention and control scheme design.
[0027] Example 1
[0028] Figure 1 This is a step diagram illustrating a method for determining inter-well interference prevention schemes in reservoir fracturing, according to an embodiment of this application. See below for reference. Figure 1 This will explain each step of the method.
[0029] like Figure 1As shown, in step S110, historical interference information that forms the historical fracture boundary information is determined based on the historical fracture boundary information of the target reservoir, and an interference information analysis model is constructed. Specifically, in this embodiment, historical reservoir fractured well images containing fractures in the target reservoir are first acquired, and then historical fracture boundary information is extracted from them. Afterwards, the causes of the historical fracture boundary information are analyzed for the extracted historical fracture boundary information, thereby obtaining the historical interference information that forms the historical fracture boundary information. Accordingly, based on a first preset machine learning model, the historical reservoir fractured well images containing historical fracture boundary information are used as input information to the first preset machine learning model, and the corresponding historical interference information is used as output information to begin training the first preset machine learning model. After training is completed, the interference information analysis model is obtained.
[0030] Furthermore, in step S120, historical interference prevention data, used to eliminate interference caused by historical interference information, is combined with the historical interference information to construct a prevention data analysis model. Specifically, in this embodiment, after determining the corresponding historical interference information based on the historical fracture boundary information extracted from historical reservoir fractured well images containing fractures in the target reservoir, historical interference prevention data capable of eliminating interference caused by historical interference information is analyzed. Accordingly, based on the second preset machine learning model, the determined historical interference information is used as the input information of the second preset machine learning model, and the corresponding historical interference prevention data is used as the output information of the second preset machine learning model to begin training the second preset machine learning model. After training is completed, the prevention data analysis model is obtained.
[0031] In the process of constructing a prevention and control data analysis model by utilizing historical interference prevention and control data to eliminate interference from historical interference information, the following steps are taken: First, the historical interference prevention and control data is normalized, and qualified prevention and control data that meet the requirements for constructing a prevention and control data analysis model are selected by calculating the variance coefficient of each normalized data set. Next, the qualified prevention and control data is divided into training and testing sets. A pre-set machine learning model is trained using the training set to obtain the prevention and control data analysis model. Then, the analysis effect of the current prevention and control data analysis model is evaluated using the testing set. Finally, based on the evaluation results, it is determined whether the current prevention and control data analysis model is the optimal model. If not, the current model is trained and evaluated again until the optimal model is obtained.
[0032] Specifically, this embodiment extracts and analyzes historical interference prevention and control data, and normalizes the data to unify the format of different types of prevention and control data. After format unification, the normalized prevention and control data undergoes data cleaning and outlier / missing value handling. Next, by calculating the variance coefficient of the processed normalized prevention and control data, the correlation between each current prevention and control data and its corresponding historical interference information is determined, thereby selecting qualified prevention and control data whose correlation meets the requirements (i.e., whose characterization ability meets the requirements for constructing the prevention and control data analysis model). Thus, using the currently selected prevention and control data (i.e., qualified prevention and control data), the interference information analysis model and the prevention and control data analysis model are constructed, effectively improving the accuracy of the output analysis results of the constructed interference information analysis model and the prevention and control data analysis model, providing technical support for determining a prevention and control scheme that matches the current fracture boundary information of the target reservoir.
[0033] In one specific embodiment of this application, qualified prevention and control data whose characterization capabilities meet the requirements for constructing a prevention and control data analysis model are specifically data whose features or representations are sufficient to effectively capture key information or patterns of the data.
[0034] Next, in this embodiment, the selected qualified prevention and control data are divided into a training set and a test set according to a specified ratio. The qualified prevention and control data in the training set and its corresponding historical interference information are used to train a second preset machine learning model, while the qualified prevention and control data in the test set and its corresponding historical interference information are used to evaluate the analytical performance of the final obtained prevention and control data analysis model.
[0035] After obtaining the analysis effect evaluation results, this embodiment determines whether the current prevention and control data analysis model is the optimal prevention and control data analysis model based on the current analysis effect evaluation results. Specifically, if the analysis accuracy of the current prevention and control data analysis model meets the expected requirements, it is the optimal prevention and control data analysis model; otherwise, it needs to continue training and accuracy analysis until the optimal prevention and control data analysis model is obtained.
[0036] In the process of obtaining the best prevention and control data analysis model, the training set is divided into multiple sub-training sets. Each sub-training set is used to train a preset machine learning model in sequence to obtain the best prevention and control data analysis model. After training each sub-training set is completed, the accuracy of the analysis results of the current prevention and control data analysis model is obtained using the test set until the accuracy of the analysis results reaches the preset accuracy threshold, and this model is taken as the best prevention and control data analysis model.
[0037] Specifically, in this embodiment, the training set is divided into multiple sub-training sets. Then, any one sub-training set is used as the output information of the second preset machine learning model, and the corresponding historical interference information is used as the input information to obtain the corresponding prevention and control data analysis model. Next, the historical interference information corresponding to the test set is used as the input information of the current prevention and control data analysis model, and the current prevention and control data analysis model outputs the corresponding prevention and control data. The output prevention and control data is then analyzed and compared with the test set to obtain the accuracy of the analysis results of the current prevention and control data analysis model. Then, by comparing the accuracy of the current analysis results with a preset accuracy threshold, it is determined whether the analysis accuracy of the current prevention and control data analysis model meets the expected requirements. If the expected requirements are met, the current prevention and control data analysis model is taken as the optimal prevention and control data analysis model; if the expected requirements are not met, another sub-training set is randomly selected from the remaining sub-training sets to obtain the corresponding prevention and control data analysis model, and it is determined whether the analysis accuracy of the current prevention and control data analysis model meets the expected requirements, until the optimal prevention and control data analysis model is obtained. This embodiment divides the construction process of the prevention and control data analysis model into multiple stages, and evaluates the analysis effect of the prevention and control data analysis model for each stage, which effectively shortens the model training time, improves the construction efficiency of the prevention and control data analysis model, and ensures the accuracy of the interference prevention and control scheme output by the prevention and control data analysis model.
[0038] In one specific embodiment of this application, the second preset machine learning model has a convolutional neural network framework structure, and the training process includes input, convolution, pooling, full connection and output.
[0039] In one specific embodiment of this application, the historical interference information corresponding to the qualified prevention and control data is substituted into the second preset machine learning model, and the normalized value of the qualified prevention and control data is used as the label of the corresponding type of prevention and control data. Thus, the label is used as the output information of the second preset machine learning model to simplify the output historical interference prevention and control data, which effectively improves the model training efficiency and accuracy.
[0040] After the interference information analysis model and the prevention and control data analysis model are constructed, in step S130, the current fracture boundary information of the target reservoir is obtained, and the interference information analysis results are obtained using the interference information analysis model. Then, the corresponding prevention and control data are determined using the prevention and control data analysis model to form a prevention and control plan. Specifically, in this embodiment, images of fractured wells containing fractures in the target reservoir are obtained, and then fracture boundary information is extracted from them. Next, the images of fractured wells containing fracture boundary information are used as input information for the interference information analysis model to obtain the corresponding interference information analysis results. Then, the current interference information analysis results are used as input information for the prevention and control data analysis model, and the corresponding prevention and control data is output to determine the appropriate prevention and control plan for the target reservoir based on the prevention and control data.
[0041] In one specific embodiment of this application, current fracture boundary information of the target reservoir is obtained by deploying edge computing devices in the target reservoir. These edge computing devices include, but are not limited to, smart sensors, programmable logic controllers, edge smart routers, and ICT converged gateways.
[0042] In the step of acquiring the current fracture boundary information of the target reservoir, firstly, reservoir fracturing well images containing fracture boundary information are acquired at different points on the target reservoir, and these acquired images are used as corresponding data to be transmitted. Next, based on the location and data transmission rate of each point, the time required for each point to complete data transmission is determined. Based on this, a node queue for data transmission trajectories is formed according to the order in which data transmission is completed. Then, different points are grouped according to the type of reservoir fracturing well image, and combined with the node queues, a set of data transmission trajectories for each point combination is obtained. Finally, the set of data transmission trajectories is used as a population, and the corresponding points are used as individuals, thereby generating a population matrix using a genetic algorithm to obtain the optimal data transmission trajectory for transmitting reservoir fracturing well images.
[0043] Specifically, this embodiment uses sensors positioned at different points on the target reservoir to obtain images of fractured wells containing information about the current fracture boundaries of the target reservoir, and uses these images as the corresponding data to be transmitted. To avoid unnecessary energy consumption and prevent transmission delays caused by transmitting too much data at once, this embodiment optimizes the transmission trajectory of the data to be transmitted, ensuring that each piece of data is transmitted in a specific order (along the optimized transmission trajectory) from front to back or from back to front. During trajectory optimization, this embodiment calculates the time required for each point to complete data transmission based on its location, data transmission rate, and the corresponding data to be transmitted. After obtaining the data transmission time for each point, points that have completed data transmission are added sequentially to the node queue of the transmission trajectory according to the order of data transmission completion, forming a node queue for the data transmission trajectory. The remaining points form a candidate queue.
[0044] Next, this embodiment employs multiple sensor types to obtain various types of reservoir fracturing well images. The points are grouped according to the type of reservoir fracturing well image, and the data transmission trajectories involved in each point combination are obtained from the resulting node queue of data transmission trajectories, thus obtaining a set of data transmission trajectories for each point combination. Then, each data transmission trajectory set is represented as a population, and each point included in the corresponding point combination is treated as an individual. A population matrix is further generated using a genetic algorithm. Next, two groups of individuals are randomly selected from the population matrix, and a path segment is selected from each of the two groups. These segments are then swapped to form two new groups of individuals. The two new groups are used to update the paths of the two groups of individuals before the swap, until the update of each group of individuals in the population matrix is completed. Further, in the current population matrix, a group of individuals is randomly selected, and two path segments from that group are randomly selected and swapped to optimize the path. Next, starting from the end point of the path, each node is traversed. If a node can be unobstructedly connected to the starting point, then the node between the starting point and that node is a redundant node. After the redundant nodes are identified, these redundant nodes are deleted, and the fitness function of the path is calculated. Thus, the path is optimized through continuous iteration. Finally, the optimal data transmission trajectory for transmitting reservoir fracturing well images is obtained.
[0045] Before constructing the interference information analysis model, this embodiment also collects historical images of the target reservoir containing fractures and divides the acquired historical fracture images into blocks according to the display ratio; then, Fourier transform is used to extract the image blocks affected by noise in the block-processed historical fracture images; finally, Gaussian filtering is used to smooth the extracted image blocks, and then the smoothed image blocks are integrated with the image blocks that are not affected by noise to obtain the historical fracture boundary information.
[0046] Specifically, in this embodiment, before constructing the interference information analysis model, historical images of the target reservoir containing fractures are preprocessed to ensure that the output analysis results of the constructed interference information analysis model are more consistent with reality. During preprocessing, historical images of the target reservoir containing fractures are acquired, and then each acquired historical image is divided into blocks according to the display ratio. After block processing, Fourier transform is used to analyze and extract image blocks affected by noise interference in the fracture historical images. Next, Gaussian filtering is used to smooth the extracted image blocks. Then, the smoothed image blocks are integrated with the noise-free image blocks to form a new historical image with a block distribution consistent with the original historical image. This new historical image is then used to extract historical fracture boundary information.
[0047] In the step of determining the historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir, firstly, fracture regions with fracture displacement and loads that balance fracture opening are screened from the historical fracture boundary information, and the fractures in the fracture regions are discretized to obtain multiple fracture boundary units; then, based on the boundary distribution state when the internal pressure of each fracture in the fracture region is uniform, the boundary conditions corresponding to each fracture boundary unit are determined, and combined with the displacement discontinuity of each fracture boundary unit, the stress characteristics at the midpoint of the fracture inside the corresponding fracture boundary unit are determined, and the stress characteristics are used as the corresponding historical interference information.
[0048] Specifically, after obtaining historical fracture boundary information, this embodiment analyzes the fracture characteristics of rock blocks during historical artificial fracturing processes, and selects fracture regions from the historical fracture boundary information that possess fracture displacement and generate loads to balance fracture opening. Then, a local coordinate system is constructed with the center of the corresponding fracture region as the origin, and a global coordinate system is constructed with the center of the entire target reservoir as the origin, thereby discretizing each fracture within the fracture region into N boundary units in the horizontal direction. Next, based on the boundary distribution state when the internal pressure of each fracture in the fracture region is uniform, the boundary conditions for each fracture boundary unit are determined, and then the displacement discontinuity of each fracture boundary unit is obtained. Further, the boundary conditions are combined with the corresponding displacement discontinuities, and the stress characteristics at the midpoint of the fracture within the corresponding fracture boundary unit are obtained through analysis. Finally, the analyzed stress characteristics are used as the corresponding historical interference information, thereby achieving the purpose of determining the historical interference information that forms the historical fracture boundary information.
[0049] In this embodiment of the application, the boundary conditions are represented by the following expression:
[0050]
[0051]
[0052] Where p represents the net pressure of the crack, s and n represent the horizontal and vertical axes respectively, j represents the serial number of the crack boundary element, N represents the total amount of the crack boundary segment elements, and σ represents the stress.
[0053] In one specific embodiment of this application, in addition to collecting historical image information of the target reservoir, historical pressure information within the reservoir is also collected. Based on the historical pressure information, the fracture boundary information obtained from the historical image information is optimized and adjusted, thereby making the obtained fracture boundary information more consistent with reality.
[0054] Furthermore, in the step of formulating a prevention and control plan, vectorized historical interference prevention and control data is acquired, and the TransD algorithm is used to construct a topic text knowledge subgraph about the interference prevention and control plan. Then, based on the topic text knowledge subgraph, a knowledge graph embedding model is constructed. Based on this, the entity table and relation table corresponding to the historical interference prevention and control data are obtained to construct a prevention and control graph library. Finally, the current prevention and control data is retrieved from the prevention and control graph library to obtain the corresponding prevention and control plan. If the current prevention and control data is included in the prevention and control graph library, a prevention and control plan is directly formed based on the corresponding part in the prevention and control graph library. Otherwise, a new prevention and control plan is formulated based on the current prevention and control data, and the prevention and control graph library is updated according to the new prevention and control plan.
[0055] Specifically, this embodiment selects relevant prevention and control fields as the knowledge scope and acquires prevention and control data from relevant fields on the Internet to extract feature keywords from historical interference prevention and control data. After extraction, the extracted feature keywords are converted into word vectors, thus obtaining vectorized historical interference prevention and control data. Next, for the vectorized historical interference prevention and control data, based on the TransD algorithm, the preference vector is used to represent the vector from the origin in the space to the embedded prevention and control scheme. At the same time, the preference for a specified prevention and control scheme is measured by the Euclidean distance between preference and the scheme, constructing a topic text knowledge subgraph about interference prevention and control schemes. In this embodiment, the starting point or center point of the vector space is used as the origin in the space. Then, according to the relationships connected to entities in the entity extraction subgraph, a knowledge graph embedding model is constructed. Using the knowledge graph embedding model, entities and their connected relationships are represented as continuous low-dimensional vectors, thereby capturing the semantic association between them in the vector space. This allows for better capture of the semantic similarity between different entities and the semantic meaning of the relationships connected to entities, thereby enriching and improving the content of the prevention and control graph library. In constructing the knowledge graph embedding model, entity vectors are used as input to the CNN layer of the third machine learning model, and entity tables and relation tables are used as output. The knowledge graph embedding model is obtained by training the third machine learning model. Next, the Neo4j database is installed and configured, and the Neo4j service is started and data (e.g., processed and transformed prevention and control domain data, feature keywords, entity relationships, etc.) is imported. Combined with the prediction results of the corresponding knowledge graph embedding model, the prevention and control graph library is constructed. Finally, the current prevention and control data is retrieved from the prevention and control graph library. If the current prevention and control data is included in the library, the relevant part of the prevention and control graph is converted into a corresponding prevention and control plan and output. Otherwise, a new prevention and control plan is formulated for the current data, and the prevention and control graph library is updated according to the new plan. Therefore, this embodiment effectively solves the problems of severe information overload caused by big data, difficulty in obtaining implicit relationships constructed by knowledge graphs, and contextual association information of topic text information, resulting in a more reliable prevention and control plan.
[0056] Furthermore, in this embodiment, the construction parameters corresponding to the determined prevention and control plan are processed into several blocks, and the determined prevention and control plan is stored using a block-based storage method. Specifically, in this embodiment, the construction parameters corresponding to the determined prevention and control plan are first preprocessed into a unified format, and the unified format construction parameters are processed into blocks with the same data type and containing valid identification information. When joining the network, each node in the blockchain network generates a local public-private key pair as its own identifier in the network. When a node is waiting for its local role to become a candidate node, it broadcasts a leadership application to other nodes in the network and sends it. After the application is approved, the candidate node becomes the leader node, and other nodes become follower nodes. Then, the leader node broadcasts the block record information. After receiving the information, the follower nodes broadcast the received information to other follower nodes and record the number of repetitions. They use the information with the most repetitions to generate a block header and send a verification application to the leader node. After the verification is approved, the leader node sends an add command and enters a dormant period (and cannot apply to become a leader node again during the dormant period until the dormant period ends). After receiving the confirmation information, the follower nodes add the newly generated blocks to the blockchain and return the candidate identity, thereby realizing the block-based storage of the prevention and control plan. Therefore, this embodiment effectively improves data storage efficiency, prevents stored data from being maliciously tampered with, ensures data storage security, and thus provides a guarantee for the accuracy of subsequent analysis.
[0057] Example 2
[0058] Based on the method for determining inter-well interference prevention schemes in reservoir fracturing as described in Embodiment 1 above, this embodiment of the invention also provides a system for determining inter-well interference prevention schemes in reservoir fracturing. Figure 2 This is a schematic diagram of the overall structure of a system for determining inter-well interference prevention schemes in reservoir fracturing, according to an embodiment of this application. The structure and function of the system for determining inter-well interference prevention schemes in reservoir fracturing will be described in detail below with reference to embodiments of the present invention.
[0059] The system for determining inter-well interference prevention and control schemes in reservoir fracturing according to the present invention includes at least: a first model construction module 21, a second model construction module 22, and a prevention and control scheme generation module 23. Specifically, the first model construction module 21 is used to determine the historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir, and construct an interference information analysis model; the second model construction module 22 is used to construct a prevention and control data analysis model by combining historical interference information with historical interference prevention and control data used to eliminate interference caused by historical interference information; the prevention and control scheme generation module 23 is used to obtain the current fracture boundary information of the target reservoir, and obtain the interference information analysis results using the interference information analysis model constructed by the first model construction module 21, and then determine the corresponding prevention and control data using the prevention and control data analysis model constructed by the second model construction module 22 to form a prevention and control scheme.
[0060] This invention proposes a method and system for determining inter-well interference prevention and control schemes in reservoir fracturing. The method constructs an interference information analysis model based on historical fracture boundary information of the target reservoir and historical interference information that forms the historical fracture boundary information. Then, based on historical interference prevention and control data used to eliminate interference caused by historical interference information, and historical interference information, a prevention and control data analysis model is constructed. Finally, based on the constructed analysis models and using the current fracture boundary information of the target reservoir, the corresponding prevention and control data is determined to form a prevention and control scheme. This invention saves time in acquiring prevention and control schemes and makes the analysis process more convenient, achieving accurate acquisition of inter-well interference prevention and control schemes in reservoir fracturing, and ensuring the accuracy and efficiency of subsequent prevention and control scheme design.
[0061] The above description is merely a specific implementation example of the present invention, and the scope of protection of the present invention is not limited thereto. Any modifications or substitutions made to the present invention by those skilled in the art within the technical specifications described herein should be within the scope of protection of the present invention.
[0062] Of course, the present invention may have other various embodiments. Without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and modifications according to the present invention, but these corresponding changes and modifications should all fall within the protection scope of the claims of the present invention.
[0063] Those skilled in the art will understand that the modules or steps of the present invention described above can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using computer-executable program code, thereby storing them in a storage device for execution by a computing device, or fabricating them separately as individual integrated circuit modules, or fabricating multiple modules or steps as a single integrated circuit module. Thus, the present invention is not limited to any particular hardware and software combination.
[0064] While the embodiments disclosed in this invention are as described above, the content is merely for the purpose of facilitating understanding of the invention and is not intended to limit the invention. Any person skilled in the art to which this invention pertains may make any modifications and variations in form and detail of the implementation without departing from the spirit and scope disclosed herein; however, the scope of patent protection for this invention shall still be determined by the scope defined in the appended claims.
Claims
1. A method for determining inter-well interference prevention schemes in reservoir fracturing, characterized in that, include: Based on the historical fracture boundary information of the target reservoir, the historical interference information that formed the historical fracture boundary information is determined, and an interference information analysis model is constructed. Using historical interference prevention data to eliminate interference caused by the historical interference information, and combining the historical interference information, a prevention data analysis model is constructed. The current fracture boundary information of the target reservoir is obtained, and the interference information analysis model is used to obtain the interference information analysis results. Then, the corresponding prevention and control data is determined using the prevention and control data analysis model to form a prevention and control plan.
2. The method according to claim 1, characterized in that, The step of obtaining the current fracture boundary information of the target reservoir includes: Images of reservoir fractured wells containing fracture boundary information are acquired at different points on the target reservoir, and the acquired images of reservoir fractured wells are used as corresponding data to be transmitted. Based on the location and data transmission rate of each point, the time required for each point to complete data transmission is determined. Based on this, a node queue about the data transmission trajectory is formed according to the order in which data transmission is completed. According to the type of reservoir fracturing well image, the different points are grouped and combined with the node queue to obtain the data transmission trajectory set of each point combination; The set of data transmission trajectories is used as a population, and the corresponding points are used as individuals. A population matrix is generated using a genetic algorithm to obtain the optimal data transmission trajectory for transmitting reservoir fracturing well images.
3. The method according to claim 2, characterized in that, Before constructing the interference information analysis model, the method further includes: Collect historical images of the target reservoir containing fractures, and divide the acquired historical fracture images into blocks according to the display ratio; Fourier transform was used to extract image blocks that were disturbed by noise from the historical crack images after block processing. The extracted image blocks are smoothed by Gaussian filtering, and then the smoothed image blocks are integrated with the image blocks that are not affected by noise to obtain the historical crack boundary information.
4. The method according to any one of claims 1 to 3, characterized in that, The step of determining the historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir includes: Crack regions with crack slippage and loads that balance crack opening are selected from the historical crack boundary information, and the cracks in the crack regions are discretized to obtain multiple crack boundary units. Based on the boundary distribution state when the internal pressure of each crack in the crack region is uniform, the boundary conditions of each crack boundary element are determined. Combined with the displacement discontinuity of each crack boundary element, the stress characteristics at the midpoint of the crack inside the corresponding crack boundary element are determined, and the stress characteristics are used as the corresponding historical interference information.
5. The method according to claim 4, characterized in that, The boundary conditions are expressed using the following expression: Where p represents the net pressure of the crack, s and n represent the horizontal and vertical axes respectively, j represents the serial number of the crack boundary element, N represents the total amount of the crack boundary segment elements, and σ represents the stress.
6. The method according to any one of claims 1 to 5, characterized in that, The step of constructing a prevention and control data analysis model by utilizing historical interference prevention and control data used to eliminate interference caused by the historical interference information, and combining the historical interference information, includes: The historical interference prevention and control data are normalized, and qualified prevention and control data whose characterization ability meets the requirements for constructing a prevention and control data analysis model are screened by calculating the variance coefficient of each normalized data. The qualified prevention and control data are divided into a training set and a test set. A preset machine learning model is trained using the training set to obtain the prevention and control data analysis model. Then, the test set is used to evaluate the effectiveness of the current prevention and control data analysis model. Based on the results of the analysis, determine whether the current prevention and control data analysis model is the best prevention and control data analysis model. If it is not, continue to train and evaluate the current prevention and control data analysis model until the best prevention and control data analysis model is obtained.
7. The method according to claim 6, characterized in that, The process of obtaining the optimal prevention and control data analysis model includes: The training set is divided into multiple sub-training sets, and a preset machine learning model is trained sequentially using each sub-training set to obtain the optimal prevention and control data analysis model. After training each sub-training set is completed, the accuracy of the analysis results of the current prevention and control data analysis model is obtained using the test set until a prevention and control data analysis model with an analysis result accuracy reaching a preset accuracy threshold is obtained, and this model is taken as the optimal prevention and control data analysis model.
8. The method according to any one of claims 1 to 7, characterized in that, The steps involved in developing a prevention and control plan include: Obtain vectorized historical interference prevention data and use the TransD algorithm to construct a thematic text knowledge subgraph about interference prevention schemes; Based on the topic text knowledge subgraph, a knowledge graph embedding model is constructed, and based on this, the entity table and relation table corresponding to the historical interference prevention and control data are obtained to construct a prevention and control graph library. The current prevention and control data is retrieved from the prevention and control map library to obtain the corresponding prevention and control plan. If the current prevention and control data is included in the prevention and control map library, the prevention and control plan is directly formed based on the corresponding part in the prevention and control map library. Otherwise, a new prevention and control plan is formulated based on the current prevention and control data, and the prevention and control map library is updated based on the new prevention and control plan.
9. The method according to any one of claims 1 to 8, characterized in that, The method further includes: The construction parameters corresponding to the determined prevention and control plan are processed into several blocks, and the determined prevention and control plan is stored using a block-based storage method.
10. A system for determining inter-well interference prevention schemes in reservoir fracturing, characterized in that, The system is used to perform the method as described in any one of claims 1 to 9, the system comprising: The first model construction module is used to determine the historical interference information that forms the historical fracture boundary information based on the historical fracture boundary information of the target reservoir, and to construct an interference information analysis model. The second model construction module is used to construct a prevention and control data analysis model by combining historical interference prevention and control data, which is used to eliminate interference caused by the historical interference information, with the historical interference information. The prevention and control plan generation module is used to obtain the current fracture boundary information of the target reservoir, and use the interference information analysis model to obtain the interference information analysis results. Then, it uses the prevention and control data analysis model to determine the corresponding prevention and control data to form a prevention and control plan.