Canal channel culvert pipe diversion intelligent simulation method and system based on TransPipeMLP
Through the TransPipeMLP model, the problems of large amount of calculation and insufficient prediction accuracy in culvert diversion are solved, and efficient and accurate flow velocity field prediction is achieved, which is suitable for flow diversion design and risk assessment of complex hydraulic structures.
Patent Information
- Application Number
- CN202510538776.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-27
- Publication Date
- 2025-08-19
AI Technical Summary
The prior art has problems such as large calculation, long time and difficulty in simulating complex working conditions in culvert diversion. The traditional machine learning model lacks the accuracy of prediction of local flow characteristics.
Using the TransPipeMLP model, a conceptual model of culvert diversion is constructed, multiple sets of different working conditions are generated, hydraulic characteristics are analyzed, and a prediction model based on a multi-layer Transformer encoder and a fully connected network is established, and the data set is integrated for efficient flow velocity field prediction.
It significantly improves the intelligence level of the culvert pipe diversion solution, realizes high-precision and fast-responsive flow rate field prediction, and is suitable for flow diversion design and risk assessment of complex hydraulic structures.
Smart Images
Figure CN120509080A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of waterway engineering construction diversion, and in particular to a canal waterway culvert diversion intelligent simulation method and system based on TransPipeMLP. Background Art
[0002] Culvert diversion is a key hydraulic engineering technology that uses culverts within canal channels to divert, divert, and regulate water. It is widely used in waterway diversion, engineering diversion, and water environment management. Currently, this technology relies primarily on physical model testing and numerical simulation. While physical model testing can realistically reflect the diversion process, it is costly, time-consuming, and difficult to simulate complex conditions. Numerical simulation, based on the Navier-Stokes equations and employing discrete techniques such as the finite volume method or the finite element method, can accurately predict hydraulic characteristics such as velocity distribution within culverts, but it is computationally intensive and sensitive to boundary conditions.
[0003] In recent years, machine learning methods have been introduced into the field of hydraulic system modeling due to their rapid prediction advantages. However, the prediction accuracy of traditional machine learning models for local flow characteristics in culvert diversion still needs to be improved. Summary of the Invention
[0004] The purpose of the present invention is to propose an intelligent simulation method and system for canal waterway culvert diversion based on TransPipeMLP, which simulates the velocity field in the construction diversion pipeline and achieves high-precision and high-efficiency prediction of the culvert diversion process.
[0005] According to a first aspect of an embodiment of the present disclosure, a method for intelligent simulation of canal culvert diversion based on TransPipeMLP is provided, comprising the following steps:
[0006] Construct a culvert diversion conceptual model to simulate the construction of culvert diversion in canal waterways;
[0007] In the culvert diversion conceptual model, a random sampling method is used to generate multiple sets of different operating parameter combinations based on the upstream flow velocity;
[0008] Analyze the hydraulic characteristics of the culvert under different working conditions and obtain the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0009] Integrate the simulation results of multiple working conditions to construct a structured comprehensive data set, which includes culvert geometric parameters, boundary conditions and flow field characteristics;
[0010] Based on the comprehensive dataset, the TransPipeMLP culvert diversion prediction model was established to predict the flow velocity field at key locations inside and outside the culvert.
[0011] In one embodiment, a simulated annealing method is used to optimize the Latin hypercube sampling algorithm SA-LHS to randomly generate multiple groups of different operating parameters.
[0012] In one embodiment, high-precision hydrodynamic software is used to analyze the hydraulic characteristics of the culvert under different working conditions.
[0013] In one embodiment, the TransPipeMLP culvert diversion model includes an input layer, a multi-layer fully connected network, a multi-layer Transformer encoder, and an output mapping layer; the input layer receives upstream flow velocity and culvert geometric parameters, the multi-layer fully connected network performs nonlinear transformation and dimensionality enhancement on the upstream flow velocity, and then performs a reshape operation, the multi-layer Transformer encoder captures the global spatial dependency of the flow velocity field, and the output mapping layer flattens the encoder output features and maps them to the velocity field dimension through the MLP layer, and reconstructs it into a two-dimensional structure identical to the original matrix.
[0014] In one embodiment, the upstream flow rate is input into a multi-layer fully connected network to extract features, and 128-dimensional and 64-dimensional feature vectors are obtained:
[0015] y=f(Wx+b)
[0016] Where y is the output vector with dimension m; x is the input upstream velocity with dimension n; W is the weight matrix with dimension m×n; b is the bias vector with dimension m; and f is the activation function ReLU:
[0017] f(z)=max(0,z)
[0018] Where z is the linear output before activation.
[0019] In one embodiment, the multi-layer Transformer encoder includes four stacked Transformer modules, each layer of which includes a multi-head attention mechanism, layer normalization, and residual connections to enhance the global modeling capability of features; the attention of each head i is:
[0020]
[0021] Merge all headers:
[0022] MultiHead(Q,K,V)=Concat(head1,...,head h )W O
[0023] Where Q is the query matrix; K is the key matrix; V is the value matrix; d k is the key dimension of each head; h is the number of heads; W O is the output weight matrix.
[0024] In one embodiment, each layer in the Transformer module includes a residual connection and normalization as follows:
[0025] Output=LayerNorm(x+AttentionOutput).
[0026] According to a second aspect of an embodiment of the present disclosure, a canal culvert diversion intelligent simulation system based on TransPipeMLP is provided, comprising:
[0027] Construction diversion simulation module, which builds a culvert diversion conceptual model to simulate the construction diversion of canal waterways through culverts;
[0028] A working condition parameter acquisition module, in the culvert diversion conceptual model, generates multiple sets of different working condition parameter combinations based on upstream flow velocity using a random sampling method;
[0029] The hydraulic characteristics simulation module analyzes the hydraulic characteristics of the culvert under different working conditions and obtains the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0030] The dataset construction module integrates the simulation results of multiple working conditions to construct a structured comprehensive dataset, which includes culvert geometric parameters, boundary conditions and flow velocity field characteristics;
[0031] The prediction module establishes the TransPipeMLP culvert diversion prediction model based on the comprehensive data set to predict the flow velocity field at key locations inside and outside the culvert.
[0032] According to a third aspect of an embodiment of the present disclosure, an electronic device is provided, including a memory, a processor, and a computer program stored and running on the memory, wherein when the processor executes the program, the intelligent simulation method for canal waterway culvert diversion based on TransPipeMLP is implemented.
[0033] According to a fourth aspect of an embodiment of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the method for intelligent simulation of canal culvert diversion based on TransPipeMLP is implemented.
[0034] Compared with the existing technology, the above technical solution adopted by the present invention has the following advantages: on the one hand, by integrating the machine learning method of the attention mechanism, the present invention improves the simulation efficiency of the culvert diversion velocity distribution by an order of magnitude compared with the traditional three-dimensional hydrodynamic method while maintaining the prediction accuracy, effectively overcoming the problems of large computational complexity and long time consumption of the traditional method, and can meet the rapid response requirements of engineering design and real-time scheduling, and is particularly suitable for rapid evaluation of multiple working conditions; on the other hand, by introducing the attention mechanism, the modeling ability of complex nonlinear relationships is significantly enhanced, which not only supports large-scale working condition simulation and optimization design, but also exhibits excellent generalization performance under different boundary conditions, and can be widely used in various engineering scenarios. BRIEF DESCRIPTION OF THE DRAWINGS
[0035] The drawings in the specification, which constitute a part of this application, are used to provide further understanding of this application. The illustrative embodiments of this application and their descriptions are used to explain this application and do not constitute improper limitations on this application.
[0036] Figure 1 This is a flow chart of the intelligent simulation method for canal culvert diversion based on TransPipeMLP;
[0037] Figure 2 Schematic diagram of the verification of the TransPipeMLP culvert diversion prediction model in the embodiment. DETAILED DESCRIPTION
[0038] The present disclosure will be further described below with reference to the accompanying drawings and embodiments.
[0039] It should be noted that the following detailed descriptions are illustrative and intended to provide further explanation of the present application. Unless otherwise specified, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which the present application belongs.
[0040] 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, they indicate the presence of features, steps, operations, devices, components and / or combinations thereof.
[0041] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the methods and systems according to the various embodiments of the present disclosure. It should be noted that each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the module, program segment, or a part of code can include one or more executable instructions for implementing the logical functions specified in the various embodiments. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, or they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the flowchart and / or block diagram, and the combination of the boxes in the flowchart and / or block diagram, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or can be implemented using a combination of dedicated hardware and computer instructions.
[0042] Example 1:
[0043] This embodiment provides a canal culvert diversion intelligent simulation method based on TransPipeMLP, including the following steps:
[0044] S1. Construct a conceptual model of culvert diversion to simulate the construction of culvert diversion in canal waterways;
[0045] Specifically, a preliminary culvert diversion conceptual model was constructed based on the culvert geometric characteristics, water boundary conditions and bottom parameters of the actual project area. This model comprehensively considers factors such as the channel cross-sectional shape, culvert layout and water flow characteristics, providing a theoretical basis for subsequent numerical simulations.
[0046] S2. In the culvert diversion conceptual model, a random sampling method is used to generate multiple sets of different operating parameter combinations based on upstream flow velocity;
[0047] Specifically, the simulated annealing method is used to optimize the Latin hypercube sampling algorithm SA-LHS, and multiple sets of different operating parameters are randomly generated to cover different boundary conditions such as upstream flow velocities and water level changes.
[0048] S3. Analyze the hydraulic characteristics of the culvert under different operating conditions and obtain the velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0049] Specifically, high-precision hydrodynamic software is used to analyze the hydraulic characteristics of the culvert under different working conditions, and to obtain hydrodynamic parameters such as flow rate and pressure at key sections inside and outside the culvert (inlet section, elbow section, outlet section, etc.).
[0050] S4. Integrate the simulation results of multiple working conditions to construct a structured comprehensive data set, which includes culvert geometric parameters, boundary conditions and flow velocity field characteristics;
[0051] Specifically, the multi-condition simulation results extract key parameters from the numerical simulation results, including: boundary condition parameters (upstream flow velocity, water level, etc.), geometric parameters (culvert diameter, inclination, etc.), flow field data (velocity of each section, etc.), construct a structured data set and perform standardized processing to ensure data quality.
[0052] S5. Based on the comprehensive dataset, a TransPipeMLP culvert diversion prediction model is established to predict the flow velocity field at key locations inside and outside the culvert.
[0053] Specifically, the TransPipeMLP culvert diversion model includes an input layer, a multi-layer fully connected network, a multi-layer Transformer encoder, and an output mapping layer; the input layer receives upstream flow velocity and culvert geometric parameters, the multi-layer fully connected network performs nonlinear transformation and dimensionality enhancement on the upstream flow velocity, and then performs a reshape operation. The multi-layer Transformer encoder captures the global spatial dependencies of the flow velocity field, and the output mapping layer flattens the encoder output features and maps them to the velocity field dimension through the MLP layer, and reconstructs them into a two-dimensional structure identical to the original matrix. Preferably, the multi-layer Transformer encoder includes four stacked Transformer modules, each layer of which includes a multi-head attention mechanism, layer normalization, and residual connections to enhance the global modeling capabilities of features.
[0054] In the process of training the TransPipeMLP culvert diversion prediction model using a comprehensive dataset, the mean square error (MSE) is used as the loss function to evaluate the error between the model output and the actual value, and the model weights are gradually optimized through batch iterative method.
[0055] Example 2:
[0056] This embodiment provides a canal culvert diversion intelligent simulation system based on TransPipeMLP, including:
[0057] Construction diversion simulation module, which builds a culvert diversion conceptual model to simulate the construction diversion of canal waterways through culverts;
[0058] A working condition parameter acquisition module, in the culvert diversion conceptual model, generates multiple sets of different working condition parameter combinations based on upstream flow velocity using a random sampling method;
[0059] The hydraulic characteristics simulation module analyzes the hydraulic characteristics of the culvert under different working conditions and obtains the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0060] The dataset construction module integrates the simulation results of multiple working conditions to construct a structured comprehensive dataset, which includes culvert geometric parameters, boundary conditions and flow velocity field characteristics;
[0061] The prediction module establishes the TransPipeMLP culvert diversion prediction model based on the comprehensive data set to predict the flow velocity field at key locations inside and outside the culvert.
[0062] Example 3:
[0063] An electronic device includes a memory, a processor, and a computer program stored and running on the memory, wherein the processor, when executing the program, implements the above-mentioned TransPipeMLP-based canal culvert diversion intelligent simulation method, including:
[0064] Construct a culvert diversion conceptual model to simulate the construction of culvert diversion in canal waterways;
[0065] In the culvert diversion conceptual model, a random sampling method is used to generate multiple sets of different operating parameter combinations based on the upstream flow velocity;
[0066] Analyze the hydraulic characteristics of the culvert under different working conditions and obtain the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0067] Integrate the simulation results of multiple working conditions to construct a structured comprehensive data set, which includes culvert geometric parameters, boundary conditions and flow field characteristics;
[0068] Based on the comprehensive dataset, the TransPipeMLP culvert diversion prediction model was established to predict the flow velocity field at key locations inside and outside the culvert.
[0069] Example 4:
[0070] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned TransPipeMLP-based canal culvert diversion intelligent simulation method, comprising:
[0071] Construct a culvert diversion conceptual model to simulate the construction of culvert diversion in canal waterways;
[0072] In the culvert diversion conceptual model, a random sampling method is used to generate multiple sets of different operating parameter combinations based on the upstream flow velocity;
[0073] Analyze the hydraulic characteristics of the culvert under different working conditions and obtain the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert;
[0074] Integrate the simulation results of multiple working conditions to construct a structured comprehensive data set, which includes culvert geometric parameters, boundary conditions and flow field characteristics;
[0075] Based on the comprehensive dataset, the TransPipeMLP culvert diversion prediction model was established to predict the flow velocity field at key locations inside and outside the culvert.
[0076] Those skilled in the art will appreciate that the modules or steps of the present disclosure described above can be implemented using a general-purpose computer device. Alternatively, they can be implemented using program code executable by a computing device, which can then be stored in a storage device and executed by the computing device. Alternatively, they can be fabricated into separate integrated circuit modules, or multiple modules or steps can be fabricated into a single integrated circuit module for implementation. The present disclosure is not limited to any specific combination of hardware and software.
[0077] The above description is merely a preferred embodiment of the present application and is not intended to limit the present application. Various modifications and variations are possible for those skilled in the art. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present application shall be included within the scope of protection of the present application.
[0078] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Those skilled in the art should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.
Claims
1. The intelligent simulation method of canal culvert diversion based on TransPipeMLP is characterized by: The following steps are involved: Construct a culvert diversion conceptual model to simulate the construction of culvert diversion in canal waterways; In the culvert diversion conceptual model, a random sampling method is used to generate multiple sets of different operating parameter combinations based on the upstream flow velocity; Analyze the hydraulic characteristics of the culvert under different working conditions and obtain the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert; Integrate the simulation results of multiple working conditions to construct a structured comprehensive data set, which includes culvert geometric parameters, boundary conditions and flow field characteristics; Based on the comprehensive dataset, the TransPipeMLP culvert diversion prediction model was established to predict the flow velocity field at key locations inside and outside the culvert.
2. The canal culvert diversion intelligent simulation method based on TransPipeMLP according to claim 1 is characterized in that: The simulated annealing method is used to optimize the Latin hypercube sampling algorithm SA-LHS and randomly generate multiple groups of different working condition parameters.
3. The canal culvert diversion intelligent simulation method based on TransPipeMLP according to claim 1 is characterized in that: High-precision hydrodynamic software is used to analyze the hydraulic characteristics of the culvert under different working conditions.
4. The canal culvert diversion intelligent simulation method based on TransPipeMLP according to claim 1 is characterized in that: The TransPipeMLP culvert diversion model includes an input layer, a multi-layer fully connected network, a multi-layer Transformer encoder, and an output mapping layer. The input layer receives upstream flow velocity and culvert geometric parameters. The multi-layer fully connected network performs nonlinear transformation and dimensionality enhancement on the upstream flow velocity, followed by a reshape operation. The multi-layer Transformer encoder captures the global spatial dependencies of the flow velocity field. The output mapping layer flattens the encoder output features and maps them to the velocity field dimension through the MLP layer, reconstructing them into a two-dimensional structure identical to the original matrix.
5. The intelligent simulation method for canal waterway culvert diversion based on TransPipeMLP according to claim 4 is characterized in that: The upstream flow rate is input into the multi-layer fully connected network to extract features and obtain 128-dimensional and 64-dimensional feature vectors: y=f(Wx+b) Where y is the output vector with dimension m; x is the input upstream velocity with dimension n; W is the weight matrix with dimension m×n; b is the bias vector with dimension m; and f is the activation function ReLU: f(z)=max(0,z) Where z is the linear output before activation.
6. The canal waterway culvert diversion intelligent simulation method based on TransPipeMLP according to claim 4 is characterized in that: The multi-layer Transformer encoder consists of four stacked Transformer modules. Each layer includes a multi-head attention mechanism, layer normalization, and residual connections to enhance the global modeling capability of features. The attention of each head i is: Merge all headers: MultiHead(Q,K,V)=Concat(head1,...,head h )W O Where Q is the query matrix; K is the key matrix; V is the value matrix; d k is the key dimension of each head; h is the number of heads; W O is the output weight matrix.
7. The canal culvert diversion intelligent simulation method based on TransPipeMLP according to claim 6 is characterized in that: Each layer in the Transformer module includes a residual connection and normalization as follows: Output=LayerNorm(x+AttentionOutput).
8. The canal culvert diversion intelligent simulation system based on TransPipeMLP is characterized by: include: Construction diversion simulation module, which builds a culvert diversion conceptual model to simulate the construction diversion of canal waterways through culverts; An operating parameter acquisition module, in the culvert diversion conceptual model, generates multiple sets of different operating parameter combinations based on upstream flow velocity using a random sampling method; Hydraulic characteristics simulation module, which analyzes the hydraulic characteristics of the culvert under different working conditions and obtains the flow velocity distribution and spatial coordinate information of key locations inside and outside the culvert; The dataset construction module integrates the simulation results of multiple working conditions to construct a structured comprehensive dataset, which includes culvert geometric parameters, boundary conditions and flow velocity field characteristics; The prediction module establishes the TransPipeMLP culvert diversion prediction model based on the comprehensive data set to predict the flow velocity field at key locations inside and outside the culvert.
9. An electronic device comprising a memory, a processor, and a computer program stored and running on the memory, wherein when the processor executes the program, the method for intelligent simulation of canal culvert diversion based on TransPipeMLP according to any one of claims 1 to 7 is implemented.
10. A computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the method for intelligent simulation of canal culvert diversion based on TransPipeMLP according to any one of claims 1 to 7 is implemented.