Grouting intelligent simulation method and system based on topological relationship mining
Through the intelligent simulation method based on topological relationship mining, the problems of low prediction accuracy and computational complexity in tunnel construction of traditional grouting simulation methods are solved, and efficient and accurate grouting simulation and slurry diffusion morphology prediction are achieved, which improves grouting effect prediction accuracy and construction safety.
Patent Information
- Application Number
- CN202410260862.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-03-07
- Publication Date
- 2024-11-22
- Estimated Expiration
- 2044-03-07
AI Technical Summary
Traditional grouting simulation methods have problems such as low prediction accuracy, high computational complexity, and unstable results in tunnel construction, making it difficult to effectively evaluate and optimize the grouting effect, which may lead to poor construction plans or cause disasters.
Using intelligent simulation method based on topological relationship mining, the grouting parameters and geological parameters are abstracted into topological graphs, and intelligent simulation is carried out through improved genetic algorithms and graph convolution neural network models to realize the constraint relationship optimization of grouting parameters and slurry diffusion. Form prediction.
Efficient and accurate grouting simulation is achieved, reducing dependence on grid division, providing more flexible and targeted parameter characterization, and improving the prediction accuracy of grouting effect and construction safety.
Smart Images

Figure CN118504066B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of geotechnical engineering technology, and in particular to a grouting intelligent simulation method and system based on topological relationship mining. Background Art
[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.
[0003] During tunnel construction, groundwater seepage, inrush and surge are quite common. Once such disasters occur, no matter how big or small, they will pose a serious threat to the environmental safety, construction quality assurance, economic benefits, and construction schedule of tunnel construction. Grouting, as an effective technology for blocking water channels and strengthening strata, plays a vital role in tunnel construction.
[0004] Grouting numerical simulation is a technology that uses numerical methods to simulate and analyze the grouting process. Through numerical simulation technology, the diffusion and blocking mechanism of slurry can be deeply understood and studied before grouting construction, which has important guiding significance for the design of grouting reinforcement and anti-seepage engineering and the prediction of grouting effect. In the early stage of tunnel construction, if this link cannot be carried out effectively, the construction plan may not be fully evaluated and optimized, resulting in poor final grouting effect, a large amount of trial and error costs, and even major disasters in serious cases.
[0005] Traditional grouting simulation methods, such as statistical models and numerical simulation methods, have certain limitations. The statistical model method relies on existing data and empirical formulas, has low prediction accuracy, and cannot take into account complex underground conditions and grouting processes; although the numerical simulation method can achieve a certain simulation accuracy, it is highly dependent on the accuracy of grid division, and fine grid division means extremely high computational complexity and extremely high time cost. According to statistics, model establishment and grid generation usually take up more than 80% of the entire project cycle, and the grid division effect is easily affected by subjective factors, and there is a certain instability in the results. In addition, the uncertainty of parameter settings, the limitations of physical model assumptions, model verification, and the limitations of verification data will have an uncertain impact on the results of grouting simulation. Summary of the invention
[0006] In order to solve the above problems, the present invention proposes a grouting intelligent simulation method and system based on topological relationship mining, which abstracts grouting parameters and geological parameters into topological nodes, and abstracts the parameters to be considered for grouting and the constraint relationship between the parameters into topological edges. Based on an improved genetic algorithm, the topological relationship is intelligently mined and a topological map is drawn. Finally, the topological structure is interpreted through a graph convolutional neural network model to realize intelligent grouting simulation.
[0007] In some embodiments, the following technical solutions are adopted:
[0008] A grouting intelligent simulation method based on topological relationship mining, comprising:
[0009] Based on historical grouting cases, the grouting parameter matrix, geological parameter matrix and parameter connection relationship matrix are extracted, and the grouting parameter topology map of historical grouting cases is drawn in combination with expert experience to form a topology map database;
[0010] Based on the topological map database, the grouting topological relationship intelligent constraint model based on the genetic algorithm is trained and optimized;
[0011] Obtain the grouting parameters and geological parameters of the grouting project to be simulated, use the trained and optimized grouting topology relationship intelligent constraint model to determine the constraint relationship of the grouting parameters and form an optimal grouting parameter topology map;
[0012] Based on the node and edge information of the grouting parameter topology graph, the node feature matrix and the connection relationship two-dimensional adjacency matrix are extracted; the extracted graph structure data is input into the graph convolutional neural network model for intelligent grouting simulation.
[0013] Also includes:
[0014] Evaluate the grouting results, optimize the grouting parameters in real time based on the evaluation results, adjust the parameters of the grouting topology relationship intelligent constraint model, and re-simulate the grouting based on the reconstructed topology structure data until the best grouting simulation effect is achieved;
[0015] The specific process of evaluating the grouting results is as follows: calculating the grouting simulation results and multiple statistical indicators set based on expert experience to evaluate the degree of difference, and performing weighted comprehensive evaluation on different statistical indicators according to actual needs and importance weights; if the weighted comprehensive evaluation result is lower than the preset comprehensive evaluation value, grouting parameters need to be optimized to guide construction.
[0016] In other embodiments, the following technical solutions are adopted:
[0017] A grouting intelligent simulation system based on topological relationship mining, comprising:
[0018] The data collection module is used to extract the grouting parameter matrix, geological parameter matrix and parameter connection relationship matrix based on historical grouting cases, and draw the grouting parameter topology map of historical grouting cases in combination with expert experience to form a topology map database;
[0019] A model optimization module, used for training and optimizing the grouting topology relationship intelligent constraint model based on the genetic algorithm based on the topology map database;
[0020] The topology drawing module is used to obtain the geological parameters and grouting parameters of the grouting project to be simulated, and use the trained and optimized grouting topology relationship intelligent constraint model to obtain the constraint relationship of the grouting parameters and form the optimal grouting parameter topology structure data;
[0021] The grouting simulation module is used to extract the node feature matrix and the connection relationship two-dimensional adjacency matrix based on the node and edge information of the grouting parameter topology map; and input the extracted data matrix into the graph convolutional neural network model for intelligent grouting simulation.
[0022] The grouting result evaluation module is used to evaluate the grouting results, optimize the grouting parameters according to the evaluation results, adjust the parameters of the grouting topological relationship intelligent constraint model, and re-simulate the grouting based on the reconstructed topological structure data until the best grouting simulation effect is achieved and the construction is guided;
[0023] The specific process of evaluating the grouting results is as follows: calculating the grouting simulation results and multiple statistical indicators set based on expert experience to evaluate the degree of difference, and performing weighted comprehensive evaluation on different statistical indicators according to actual needs and importance weights; if the weighted comprehensive evaluation result is lower than the preset comprehensive evaluation value, grouting parameter optimization is required.
[0024] In other embodiments, the following technical solutions are adopted:
[0025] A terminal device comprises a processor and a memory, wherein the processor is used to implement instructions; the memory is used to store a plurality of instructions, wherein the instructions are suitable for being loaded by the processor and executing the above-mentioned grouting intelligent simulation method based on topological relationship mining.
[0026] In other embodiments, the following technical solutions are adopted:
[0027] A computer-readable storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the above-mentioned grouting intelligent simulation method based on topological relationship mining.
[0028] Compared with the prior art, the present invention has the following beneficial effects:
[0029] (1) The present invention abstracts grouting parameters, geological parameters and parameter connection relationships into topological graph structure data, and realizes the intelligent description of grouting topological edges and topological graph structure information under different geological conditions, different constraint relationships and different grouting parameters through the grouting topological relationship intelligent constraint model; based on the acquired optimal topological graph, feature matrix extraction is performed, and the graph convolutional neural network model is used to realize efficient and accurate grouting simulation and prediction of slurry diffusion morphology, which can maximize the realization of grouting goals and provide new ideas for grouting intelligent simulation.
[0030] (2) The intelligent constraint model of grouting topological relationship proposed in the present invention improves the traditional genetic algorithm, encodes the grouting parameter nodes and route information into the gene sequence using binary coding, and introduces expert experience to constrain the parameters and association relationships, thereby realizing the determination and drawing of the optimal route and topological map in the topological node parameter space.
[0031] (3) The present invention realizes intelligent simulation and prediction of grouting based on learning and interpreting topological structures, which can effectively reduce the dependence of traditional grouting numerical simulation on model grid division, realize faster numerical simulation, more flexible and targeted parameter characterization, and more comprehensive information acquisition.
[0032] (4) The present invention proposes a grouting simulation parameter optimization feedback method based on topological relationships. Through comprehensive evaluation of the grouting simulation results, parameter tuning and real-time adjustment of constraint relationships are achieved based on the grouting evaluation model feedback mechanism, and the reconstruction of the topological graph structure is realized.
[0033] Other features and advantages of additional aspects of the present invention will be given in part in the following description, and in part will become obvious from the following description, or will be learned through the practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS
[0034] Figure 1 It is a flow chart of a grouting intelligent simulation method based on topological relationship mining in an embodiment of the present invention;
[0035] Figure 2 A schematic diagram of a grouting simulation topology diagram construction process in an embodiment of the present invention;
[0036] Figure 3 This is a diagram of an intelligent grouting simulation system based on topological relationship mining in an embodiment of the present invention. DETAILED DESCRIPTION
[0037] It should be noted that the following detailed descriptions are illustrative and are intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used in the present invention have the same meanings as those commonly understood by those skilled in the art to which the present application belongs.
[0038] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprise" and / or "include" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0039] Embodiment 1
[0040] In one or more embodiments, a grouting intelligent simulation method based on topological relationship mining is disclosed, combined with Figure 1 , specifically including the following process:
[0041] (1) Based on historical grouting cases, the grouting parameter matrix, geological parameter matrix, and parameter connection relationship matrix are extracted, and the grouting parameter topology map of the historical grouting cases is drawn in combination with expert experience to form a topology map database;
[0042] Specifically, this embodiment extracts the grouting parameter matrix (including grouting pressure, grouting speed, spatial distribution data, etc.) and geological parameter matrix (including formation type, lithology, permeability, porosity, aquifer conditions, fault zones, rock strength, etc.) corresponding to different cases and the parameter connection relationship matrix from the case-related materials and literature in the existing tunnel disaster grouting technology management case library, and draws the grouting parameter topology map of the case according to the actual constraint relationship between the parameters and expert experience, and constructs a topology map database based on this.
[0043] Specifically, the grouting parameter matrix is a matrix that describes the parameters used in the grouting process. Each row represents a parameter (feature) and each column represents a sample or time step. The elements in the matrix represent the values of the corresponding parameter (feature) in different samples or time steps. The parameter connection relationship matrix is a matrix that describes the connection relationship between parameters. It shows the dependency, mutual influence or connection strength between parameters. Usually, the rows and columns of the matrix correspond to parameters, and the elements in the matrix represent the connection relationship between the corresponding parameters.
[0044] In this embodiment, each node is represented as a grouting parameter or geological parameter, such as flow rate, pressure, slurry selection, etc. In the topological diagram, each node can be represented by a circle, square or other shape. Based on the correlation between the parameters, the nodes are connected to form a topological diagram; by interpreting the topological diagram, the grouting parameters and parameter relationships can be read. The optimal topological diagram can achieve the maximum effect or performance of the grouting process. Through reasonable parameter connection relationships, the grouting goals can be achieved to the maximum extent.
[0045] (2) Training and optimizing the grouting topology relationship intelligent constraint model based on genetic algorithm based on the topology map database;
[0046] In this embodiment, the grouting topological relationship intelligent constraint model is constructed based on the improved genetic algorithm; a large number of grouting parameter matrices, geological parameter matrices and parameter connection relationship matrices, as well as topological maps drawn based on these matrices are stored in the topological map database; these data are used as training parameter adjustment data sets to tune the improved genetic algorithm to obtain the optimized grouting topological relationship intelligent constraint model.
[0047] This embodiment can find the optimal route in the topological node parameter space based on the genetic algorithm, and at the same time, realize the drawing of the topological map. The improvement of the genetic algorithm is:
[0048] ① Before randomly generating individual populations, the grouting parameter nodes and route information are encoded into the gene sequence using binary coding according to the grouting topology parameters and constraint connection relationships. After the algorithm terminates, the node and route information of the grouting topology are decoded by inverse conversion according to the optimal grouting topology structure gene sequence, and the grouting topology structure is drawn according to the node and route information.
[0049] ② Based on expert experience, a population of individuals is randomly generated, and the node attribute values and the connection relationship between nodes are constrained by threshold setting to ensure that some individuals meet the expectations of experts.
[0050] This embodiment encodes the grouting parameter nodes and route information into the gene sequence using binary coding, and can automatically draw a grouting topology map based on the node and route information. By obtaining the optimal topology map, intelligent and rapid grouting simulation can be achieved, reducing the dependence of traditional grouting numerical simulation on model grid division; at the same time, expert experience is introduced to constrain parameters and association relationships to ensure the accuracy and effectiveness of topology map drawing, and realize the determination and drawing of the optimal route and topology map in the topological node parameter space.
[0051] (3) obtaining geological parameters and grouting parameters of the grouting project to be simulated, and using the trained and optimized grouting topological relationship intelligent constraint model to obtain the constraint relationship of the grouting parameters and form an optimal grouting parameter topological graph structure;
[0052] In this embodiment, existing data of the project for which grouting numerical simulation is to be carried out are analyzed, and the main geological parameter matrix and grouting parameter matrix are manually extracted.
[0053] The geological parameter matrix and grouting parameter matrix that need to be considered are input into the grouting topological relationship intelligent constraint model to realize the automatic determination of the grouting parameter constraint relationship (physical contact relationship, slurry flow relationship, diffusion influence relationship and connectivity relationship, etc.) and the automatic drawing of topological edges and topological graph structures.
[0054] Figure 2 A schematic diagram of the construction process of the grouting simulation topology map is given. Figure 2 The grouting parameter nodes or geological parameter nodes in the model represent the characteristic parameters of each node, and there is a parameter connection relationship between each node. By inputting the geological parameter matrix, the grouting parameter matrix and the parameter connection relationship matrix into the trained and optimized grouting topology relationship intelligent constraint model, the optimal grouting simulation topology model can be obtained.
[0055] In this embodiment, the process of generating the optimal grouting parameter topology map based on the grouting topology relationship intelligent constraint model of the improved genetic algorithm is as follows:
[0056] (3-1) Determine the number of individuals in the initial population, select grouting parameters to represent them as genes, and encode the grouting parameter nodes and route information into the gene sequence using binary coding according to the grouting topological parameters and constraint connection relationships. Randomly generate individual populations based on expert experience, and set constraint node attribute values and connection relationships between nodes through thresholds to ensure that some individuals meet expert expectations.
[0057] The encoding steps are as follows:
[0058] (3-1-1) Determine the coding scheme: Determine the binary coding length corresponding to each parameter and route based on the value range of the specific grouting parameters and route information.
[0059] (3-1-2) Establishing a coding mapping table: According to the coding scheme, establish a coding mapping table to map the value of each grouting parameter and route with the corresponding binary code.
[0060] (3-1-3) Encode parameter nodes and route information into gene sequences: For each gene sequence, according to the values of grouting parameter nodes and route information, use the coding mapping table to convert them into corresponding binary codes. The obtained binary codes are combined into a complete gene sequence according to certain rules.
[0061] (3-2) According to the performance indicators of the grouting topology, a fitness function is designed to evaluate the quality of each individual and calculate the fitness value of each individual in the population.
[0062] (3-3) According to the fitness value, a part of individuals are selected as parents for the next generation of reproduction. When selecting, a random number in the interval [0,1] is generated. If the random number is less than or equal to the cumulative probability of the individual and greater than the cumulative probability of individual 1, the individual is selected to enter the offspring population.
[0063] (3-4) Select appropriate crossover operators to exchange and combine genes, generate new candidate solutions for grouting topology, and determine the probability of crossover operation. The crossover probability is usually used to control the frequency of crossover.
[0064] (3-5) Perform mutation operations on the offspring individuals and introduce random disturbances to increase the diversity of the population. The genes of the offspring individuals can be randomly changed, such as mutation of gene values or mutation of gene positions, to generate new candidate solutions for the grouting topology.
[0065] (3-6) The parent and offspring individuals are merged to form a new population, and the next round of selection, crossover and mutation operations are carried out.
[0066] (3-7) According to the preset termination conditions, such as reaching the maximum number of iterations, fitness reaching a certain threshold, etc., determine whether the termination conditions are met. If not, return to (3-3) to continue optimization; if satisfied, select the individual with the highest fitness as the optimal grouting topology structure.
[0067] (3-8) Based on the optimal grouting topology structure gene sequence, the node and route information of the grouting topology is decoded by inverse conversion.
[0068] (3-9) Draw the grouting topology diagram based on the node and route information.
[0069] (4) Based on the node and edge information of the grouting parameter topology graph, the node feature matrix and the two-dimensional adjacency matrix of the connection relationship are extracted; the extracted data matrix is input into the graph convolutional neural network model for intelligent grouting simulation.
[0070] Based on the grouting parameter topological graph, topological parameter information and graph features are extracted. The topological parameter information is the topological nodes abstracted from the grouting parameters and geological parameters. The graph features characterize the density and connectivity of the nodes, which can indicate the degree of mutual influence between the parameters and the efficiency of information transmission. The centrality of the nodes can indicate the importance and controllability of the parameters, the correlation between the grouting parameters, the transmission path and the overall topological structure.
[0071] In this embodiment, based on the node and edge information of the topological graph, a node feature matrix and a two-dimensional adjacency matrix of connection relations are constructed (in the adjacency matrix, 1 indicates that there is a connection between nodes, and 0 indicates that there is no connection between nodes).
[0072] Specifically, the node feature matrix is constructed using One-hot encoding, which converts each value into a binary vector whose length is equal to the number of feature values. For each node feature, the corresponding binary vector is used as one or more columns of the node feature matrix.
[0073] The node feature matrix of this embodiment is a two-dimensional matrix, each row represents a node in the graph, and each column represents a feature of the node. The elements in the matrix represent the values of the corresponding features of the nodes. There is a graph structure of grouting simulation, which contains 5 parameter nodes, and each node has 3 features (such as node type / label, position, attributes, etc.).
[0074] The adjacency matrix of this embodiment is also a two-dimensional matrix, which is used to represent the connection relationship between nodes in the graph. The elements in the matrix represent the connection between nodes, which can be binary values (connected is 1, no connection is 0) or weighted values;
[0075] The construction method of the adjacency matrix is as follows: determine the node scale of the grouting parameters required by the topological graph, and initialize an adjacency matrix of size N×N (N is the number of nodes). Initially, all matrix elements are set to 0 (indicating that there is no edge connection); according to the connection relationship of the topological graph, add edges to the adjacency matrix one by one, and for each edge (u, v), where u and v represent the starting node and target node of the edge respectively, set the corresponding elements A[u][v] and A[v][u] in the adjacency matrix to 1 (indicating that there is a connection between the two nodes). Diagonal processing: If the topological graph allows connections between nodes and themselves (self-loops), the elements A[i][i] on the diagonal can be set to 1 (indicating that the node is connected to itself). If there are no self-loops, the elements on the diagonal can remain at 0.
[0076] The node feature matrix and adjacency matrix after feature extraction are input into the trained GCN graph convolutional neural network model as input to obtain the label, attributes or other relevant information of each node, usually represented by a vector or matrix related to the number of nodes and the task, to realize intelligent grouting simulation.
[0077] The training process of the GCN graph convolutional neural network model in this embodiment is as follows:
[0078] (4-1) Initialize model parameters: including GCN weights, activation functions, etc.
[0079] (4-2) Forward propagation: The input node feature matrix and adjacency matrix are passed into the GCN model, and the output is generated through multi-layer graph convolution operations and nonlinear activation functions.
[0080] (4-3) Loss calculation: Compare the model output with the true label or target, and calculate the loss function to measure the difference between the model's prediction and the true value.
[0081] (4-4) Back propagation: Through the back propagation algorithm, the gradient of the model parameters is calculated and updated according to the loss function, and the model parameters are optimized to reduce the loss function.
[0082] (4-5) Repeat steps (4-2)-(4-4) until convergence: Through multiple iterative training, the model parameters are continuously updated until the model converges or reaches the set value.
[0083] The intelligent grouting simulation method of this embodiment does not require modeling and meshing, which saves a lot of time, and can more accurately describe the grouting parameter settings, breaking through the limitations of physical model assumptions and inaccurate model verification.
[0084] (6) The model and simulation effects are evaluated according to the grouting results. If the best grouting effect cannot be obtained, the grouting parameters are optimized, and the intelligent constraint model of the grouting topological relationship is adjusted. Based on the reconstructed topological structure, the grouting simulation is performed step by step until the best grouting simulation effect is achieved.
[0085] Specifically, this embodiment compares and analyzes the output grouting results with the actual grouting measured data, and then evaluates the simulation effect, compares the differences between grouting parameters (such as grouting pressure, grouting flow, etc.) and grouting effects (such as penetration depth, slurry diffusion range, etc.), calculates statistical indicators between the simulation results and the measured data to evaluate the degree of difference, and conducts a comprehensive evaluation of different indicators based on actual needs and importance weights to obtain a comprehensive evaluation result. If the simulation result is lower than the comprehensive evaluation index, the grouting parameters are fine-tuned, and the whole process is repeated until the evaluation result meets the index requirements.
[0086] The final output of the grouting simulation results is combined with the actual engineering grouting effect for verification.
[0087] This embodiment abstracts grouting parameters and geological parameters into topological nodes, relies on the improved genetic algorithm in the grouting topological relationship intelligent constraint model to realize the intelligent construction of the topological graph structure, and uses the GCN graph convolutional neural network model to mine the relationship and dependency characteristics between parameter nodes, ultimately realizing efficient intelligent grouting simulation and prediction as well as intelligent optimization of grouting parameters and models.
[0088] Embodiment 2
[0089] In one or more embodiments, a grouting intelligent simulation system based on topological relationship mining is disclosed, combined with Figure 3 ,include:
[0090] (1) Data collection module, which is used to extract grouting parameter matrix, geological parameter matrix and parameter connection relationship matrix based on historical grouting cases, draw the grouting parameter topology map of historical grouting cases based on expert experience, and construct a topology map database.
[0091] The data collection module of this embodiment specifically includes:
[0092] Grouting case collection unit: used to automatically crawl and collect tunnel disaster grouting technology management cases and grouting case related information in the literature through crawler technology; collect and store grouting parameters and parameter connection relationship data of a large number of numerical simulation cases and store them in the numerical simulation database;
[0093] Grouting parameter semantic extraction unit: realizes the grouting parameters and connection relationship of crawling cases, automatic semantic extraction and grouting parameters and extraction of numerical simulation time synchronization files;
[0094] Topological data integration unit: used to unify the format of parameter data sets and convert them into parameter matrix and parameter connection relationship matrix formats and store them.
[0095] (2) a model training module, which is used to train and optimize the grouting topology relationship intelligent constraint model based on the genetic algorithm based on the topology map database;
[0096] In this embodiment, the grouting topology relationship intelligent constraint model draws the topological structure according to the improved genetic algorithm; by comparing the topological map drawn by the existing data with the corresponding case in the topological map database, the node validity and edge connection conditions are verified by comparison to see whether they are consistent with the actual situation. If not, the genetic algorithm is dynamically adjusted in real time.
[0097] (3) A topology drawing module is used to obtain the geological parameters and grouting parameters of the grouting project to be simulated, and to obtain the constraint relationship of the grouting parameters by using the trained and optimized grouting topology relationship intelligent constraint model to form an optimal grouting parameter topology map;
[0098] In this embodiment, by simulating the interface of the grouting parameters, the format of the input grouting parameters can be standardized and input into the grouting topological relationship intelligent constraint model.
[0099] (4) A grouting simulation module, which is used to extract the node feature matrix and the two-dimensional adjacency matrix of the connection relationship based on the node and edge information of the grouting parameter topology map; and input the extracted data matrix into the graph convolutional neural network model for intelligent grouting simulation;
[0100] Specifically, the grouting simulation module of this embodiment includes a coding unit and a grouting simulation unit;
[0101] Coding unit: extract topological parameters and graph features from the drawn topological graph structure, and construct a node feature matrix and a two-dimensional adjacency matrix of connection relationships based on the node and edge information of the topological graph (1 is used in the adjacency matrix to indicate that there is a connection between nodes, and 0 is used to indicate that there is no connection between nodes).
[0102] The construction of the above node feature matrix uses One-hot encoding to convert each value into a binary vector whose length is equal to the number of feature values. For each node feature, the corresponding binary vector is used as one or more columns of the node feature matrix.
[0103] The construction method of the above adjacency matrix is as follows: determine the node scale of the grouting parameters required by the topological graph, and initialize an adjacency matrix of size N×N (N is the number of nodes). Initially, all matrix elements are set to 0 (indicating that there is no edge connection); according to the connection relationship of the topological graph, add edges to the adjacency matrix one by one, and for each edge (u, v), where u and v represent the starting node and target node of the edge respectively, set the corresponding elements A[u][v] and A[v][u] in the adjacency matrix to 1 (indicating that there is a connection between the two nodes). Diagonal processing: If the topological graph allows connections between nodes and themselves (self-loops), the elements A[i][i] on the diagonal can be set to 1 (indicating that the node is connected to itself). If there are no self-loops, the elements on the diagonal can remain at 0.
[0104] Grouting simulation unit: The node feature matrix and adjacency matrix after feature extraction are input into the trained GCN graph convolutional neural network model for intelligent grouting simulation.
[0105] (5) A grouting result evaluation module, which is used to evaluate the grouting results, optimize the grouting parameters according to the evaluation results, adjust the parameters of the grouting topological relationship intelligent constraint model, and re-perform the grouting simulation based on the reconstructed topological graph structure until the best grouting simulation effect is achieved;
[0106] Among them, the process of evaluating the grouting results is specifically as follows: multiple statistical indicators between the grouting simulation results and the grouting measured data are calculated to evaluate the degree of difference, and different statistical indicators are weighted comprehensively evaluated according to actual needs and importance weights; if the weighted comprehensive evaluation result is lower than the preset comprehensive evaluation value, grouting parameter optimization is required.
[0107] (6) A result output module, which is used to visualize the simulation results; it includes a result visualization unit, a data query unit and a text output unit; wherein:
[0108] Result visualization unit: can realize visualization query of grouting topology diagram and operation steps;
[0109] Data query unit: can realize the query of historical operation data, database data, simulation results and other data;
[0110] Text output unit: The text data output module can realize the text output of system collection, processing, historical operation and other data.
[0111] The specific implementation methods of the above modules have been described in detail in Example 1 and will not be described in detail here.
[0112] Embodiment 3
[0113] In one or more embodiments, a terminal device is disclosed, including a server, the server including a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the grouting intelligent simulation method based on topological relationship mining in Example 1 when executing the program. For the sake of brevity, it will not be described here.
[0114] It should be understood that in this embodiment, the processor may be a central processing unit CPU, and the processor may also be other general-purpose processors, digital signal processors DSP, application-specific integrated circuits ASIC, off-the-shelf programmable gate arrays FPGA or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0115] The memory may include a read-only memory and a random access memory, and provide instructions and data to the processor. A portion of the memory may also include a non-volatile random access memory. For example, the memory may also store information about the device type.
[0116] In the implementation process, each step of the above method can be completed by an integrated logic circuit of hardware in a processor or an instruction in the form of software.
[0117] Embodiment 4
[0118] In one or more embodiments, a computer-readable storage medium is disclosed, in which a plurality of instructions are stored, wherein the instructions are suitable for being loaded by a processor of a terminal device and executing the grouting intelligent simulation method based on topological relationship mining described in Example 1.
[0119] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.
Claims
1. A grouting intelligent simulation method based on topological relationship mining, characterized in that: include: Based on historical grouting cases, the grouting parameter matrix, geological parameter matrix and parameter connection relationship matrix are extracted, and the grouting parameter topology map of historical grouting cases is drawn in combination with expert experience to form a topology structure database; Based on the topological graph structure database, the grouting topological relationship intelligent constraint model based on the genetic algorithm is trained and optimized; The geological parameters and grouting parameters of the grouting project to be simulated are obtained, and the constraint relationship of the grouting parameters is obtained by using the trained and optimized grouting topological relationship intelligent constraint model, and the optimal grouting parameter topological map is automatically formed. Specifically: the number of individuals in the initial population is determined, the grouting parameters are represented as genes, and the grouting parameter node information and parameter connection relationship information are encoded into the gene sequence using binary coding, and the individual population is randomly generated; after calculating the fitness, crossover and mutation iterations, the individual with the highest fitness is selected as the optimal grouting topological map structure; according to the optimal grouting topological map structure gene sequence, the node and route information of the grouting topological map is decoded by inverse conversion, and the grouting topological map is drawn according to the node and route information; Based on the node and edge information of the grouting parameter topology map, the node feature matrix and the connection relationship two-dimensional adjacency matrix are extracted; the extracted data matrix is input into the graph convolutional neural network model for intelligent grouting simulation.
2. The grouting intelligent simulation method based on topological relationship mining according to claim 1 is characterized in that: Also includes: Evaluate the grouting results, optimize the grouting parameters based on the evaluation results, adjust the parameters of the grouting topology relationship intelligent constraint model, and re-simulate the grouting based on the reconstructed topology structure until the best grouting simulation effect is achieved; The specific process of evaluating the grouting results is as follows: calculating the grouting simulation results and multiple statistical indicators set based on expert experience to evaluate the degree of difference, and performing weighted comprehensive evaluation on different statistical indicators according to actual needs and importance weights; if the weighted comprehensive evaluation result is lower than the preset comprehensive evaluation value, grouting parameter optimization is required.
3. The grouting intelligent simulation method based on topological relationship mining according to claim 1 is characterized in that: Also includes: Based on expert experience, a population of individuals is randomly generated, and the node attribute values and the connection relationship between nodes are constrained by threshold setting to ensure that the constraint relationship between some individuals meets the expectations of experts.
4. The grouting intelligent simulation method based on topological relationship mining according to claim 1 is characterized in that: The node feature matrix is constructed as follows: Use One-hot encoding to convert the features of each node into a binary vector. The length of the binary vector is equal to the number of feature values. The node feature matrix is a two-dimensional matrix, each row of the matrix represents a node, and the value of each row is a binary vector transformed from the node feature.
5. The grouting intelligent simulation method based on topological relationship mining according to claim 1 is characterized in that: The method for constructing the connection relationship two-dimensional adjacency matrix is: Initialize an adjacency matrix of size N×N, where N is the number of nodes in the topology graph; Initially, all matrix elements are set to 0; according to the connection relationship of the topological graph, add edges to the adjacency matrix one by one. For each edge (u, v), where u and v represent the starting node and target node of the edge respectively, set the corresponding elements A[u][v] and A[v][u] in the adjacency matrix to 1, indicating that there is a connection between the two nodes; If the topology graph allows connections between nodes and themselves, i.e. self-loops, set the diagonal elements A[i][i] to 1; If there are no self-loops, keep the diagonal elements at 0.
6. A grouting intelligent simulation system based on topological relationship mining, characterized in that: include: The data collection module is used to extract the grouting parameter matrix, geological parameter matrix and parameter connection relationship matrix based on historical grouting cases, and draw the grouting parameter topology map of historical grouting cases in combination with expert experience to form a topology structure database; A model optimization module, used for training and optimizing the grouting topology relationship intelligent constraint model based on the genetic algorithm based on the topology structure database; The topology drawing module is used to obtain the geological parameters and grouting parameters of the grouting project to be simulated, and obtain the constraint relationship of the grouting parameters by using the trained and optimized grouting topology relationship intelligent constraint model, and automatically form the optimal grouting parameter topology map. Specifically: determine the number of individuals in the initial population, express the grouting parameters as genes, encode the grouting parameter node information and parameter connection relationship information into the gene sequence using binary coding, and randomly generate individual populations; after calculating fitness, crossover and mutation iterations, select the individual with the highest fitness as the optimal grouting topology map structure; according to the optimal grouting topology map structure gene sequence, decode the node and route information of the grouting topology map by inverse conversion, and draw the grouting topology map according to the node and route information; The grouting simulation module is used to extract the node feature matrix and the connection relationship two-dimensional adjacency matrix based on the node and edge information of the grouting parameter topology map; and input the extracted data matrix into the graph convolutional neural network model for intelligent grouting simulation.
7. The grouting intelligent simulation system based on topological relationship mining according to claim 6 is characterized in that: Also includes: The grouting result evaluation module is used to evaluate the grouting results, optimize the grouting parameters according to the evaluation results, adjust the parameters of the grouting topological relationship intelligent constraint model, and re-perform the grouting simulation based on the reconstructed topological graph structure until the best grouting simulation effect is achieved; The specific process of evaluating the grouting results is as follows: calculating the grouting simulation results and multiple statistical indicators set based on expert experience to evaluate the degree of difference, and performing weighted comprehensive evaluation on different statistical indicators according to actual needs and importance weights; if the weighted comprehensive evaluation result is lower than the preset comprehensive evaluation value, grouting parameter optimization is required.
8. A terminal device, comprising a processor and a memory, wherein the processor is used to implement instructions; and the memory is used to store multiple instructions, characterized in that: The instructions are suitable for being loaded by a processor and executing the grouting intelligent simulation method based on topological relationship mining as described in any one of claims 1-5.
9. A computer-readable storage medium storing a plurality of instructions, characterized in that: The instructions are suitable for being loaded by a processor of a terminal device and executing the grouting intelligent simulation method based on topological relationship mining as described in any one of claims 1-5.
Citation Information
Patent Citations
Grid fault detection method and device based on graph neural network
CN115733730A
Line parameter identification method based on excavation of spatial topological characteristics of power grid
CN116484740A
