Nozzle surface pressure field distribution prediction method, system and equipment and storage medium

Abstract

The invention relates to the technical field of fluid mechanics, and discloses a nozzle surface pressure field distribution prediction method, system and device and a storage medium, and the method comprises the steps: obtaining a three-dimensional structure of a sample nozzle under different values of a geometric model critical dimension parameter, and combining with a fixed value of a boundary condition parameter of the sample nozzle, simulating each three-dimensional structure to obtain a surface pressure field distribution simulation result; based on the geometric feature data of each three-dimensional structure and the surface pressure field distribution simulation result, training the deep learning model to obtain a nozzle surface pressure field distribution prediction model; and determining a three-dimensional to-be-measured structure of the to-be-measured nozzle according to the actual values of the key dimension parameters of the geometric model of the to-be-measured nozzle, and obtaining a surface pressure field distribution prediction result by using the prediction model based on geometric feature to-be-measured data obtained by sampling the three-dimensional to-be-measured structure. The nozzle surface pressure field distribution can be quickly and accurately predicted, and the nozzle design efficiency and the performance optimization level are improved.

Description

Technical Field

[0001] This invention relates to the field of fluid mechanics, and in particular to a method, system, device, and storage medium for predicting the pressure field distribution on a nozzle surface. Background Technology

[0002] In the fields of fluid mechanics and engineering applications, accurately calculating the surface pressure field distribution of nozzles is crucial for optimizing nozzle design, improving fluid jetting efficiency, and enhancing the performance of related equipment. Traditional methods for calculating nozzle surface pressure field distribution often employ simulations with fixed geometric models and boundary conditions. For example, in many published patents and academic literature, researchers typically use commercial simulation software such as ANSYS to perform single-run simulations based on pre-defined nozzle geometry and fixed boundary condition parameters to obtain pressure field distribution results under specific geometric structures. Furthermore, when predicting pressure field distribution under different geometric structures, traditional methods mainly rely on rebuilding the geometric model, setting boundary conditions, and performing simulations again, or using empirical formulas combined with limited experimental data for estimation. This approach is extremely resource-intensive.

[0003] Existing technologies suffer from the following main defects and shortcomings: For predicting pressure field distribution under varying geometric structures, current methods are insufficient. Whether based on empirical formulas or extensive simulations, neither can quickly and accurately provide pressure field distribution results for different geometric structures. This fails to provide timely and effective references for engineering design, resulting in long nozzle design cycles, high costs, and an inability to adapt to the rapid iteration and optimization demands of modern engineering. Furthermore, traditional machine learning prediction is based on training the relationship between changes in boundary conditions and results under the same structure. Because structural changes alter the number of grid points, it's impossible to fix the number of input nodes in the neural network, thus hindering the training of machine learning models. All machine learning models require data with the same input dimension.

[0004] Therefore, there is an urgent need to provide a technical solution to address the above problems. Summary of the Invention

[0005] To address the aforementioned technical problems, this invention provides a method, system, device, and storage medium for predicting the pressure field distribution on the nozzle surface.

[0006] In a first aspect, the present invention provides a method for predicting the pressure field distribution on the nozzle surface, the technical solution of which is as follows: The three-dimensional sample structure of the sample nozzle under different values ​​of the key dimensional parameters of the geometric model is obtained, and the surface pressure field distribution simulation results of each three-dimensional sample structure are obtained by combining the fixed values ​​of the boundary condition parameters of the sample nozzle. Based on the geometric feature sample data obtained by sampling each three-dimensional sample structure, and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, the preset deep learning model is trained to obtain the nozzle surface pressure field distribution prediction model. Based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, the three-dimensional structure of the nozzle under test is determined. Based on the geometric feature test data obtained by sampling the three-dimensional structure under test, the surface pressure field distribution prediction result of the nozzle under test is obtained using the nozzle surface pressure field distribution prediction model.

[0007] The beneficial effects of the nozzle surface pressure field distribution prediction method of the present invention are as follows: The method of this invention solves the problem that traditional methods are difficult to adapt to changing geometric model structures, reduces the consumption of manpower and material resources, shortens the nozzle design cycle, reduces costs, and improves nozzle design efficiency. It can better meet the needs of rapid iteration and optimization in modern engineering and overcome the defect that traditional machine learning cannot be trained after structural changes.

[0008] Based on the above scheme, the nozzle surface pressure field distribution prediction method of the present invention can be further improved as follows.

[0009] In one alternative approach, the step of training a pre-defined deep learning model to obtain a nozzle surface pressure field distribution prediction model based on geometric feature sample data obtained by sampling each three-dimensional sample structure and simulation results of the surface pressure field distribution of each three-dimensional sample structure includes: The geometric feature sample data of any three-dimensional sample structure is input into the preset deep learning model used for nozzle surface pressure field distribution prediction to obtain the surface pressure field distribution prediction result of any three-dimensional sample structure. Based on the predicted surface pressure field distribution and the simulation results of the surface pressure field distribution of any three-dimensional sample structure, the loss value of any three-dimensional sample structure is determined until the loss value of each three-dimensional sample structure is obtained. Based on all the loss values, the parameters of the preset deep learning model are optimized to obtain an optimized deep learning model. The optimized deep learning model is then used as the preset deep learning model, and the step of inputting the geometric feature sample data of any three-dimensional sample structure into the preset deep learning model for nozzle surface pressure field distribution prediction is returned to be executed until the optimized deep learning model meets the preset iterative training conditions. At this point, the optimized deep learning model is determined as the nozzle surface pressure field distribution prediction model.

[0010] In one alternative approach, the key dimensional parameters of the geometric model include: the size of the oxygen-enriched gas inlet, the size of the fuel-enriched gas inlet, and the channel radius.

[0011] In one alternative approach, the boundary condition parameters include: inlet pressure, fluid flow rate, and outlet pressure.

[0012] In one alternative approach, the step of simulating each three-dimensional sample structure to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure includes: Using ANSYS software, each three-dimensional sample structure was simulated separately to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure.

[0013] In one alternative approach, the preset deep learning model is the RegDGCNN model.

[0014] In one alternative approach, it also includes: The predicted surface pressure field distribution of the nozzle under test is visualized and output.

[0015] Secondly, the present invention provides a nozzle surface pressure field distribution prediction system, the technical solution of which is as follows: It includes: a simulation module, a training module, and a prediction module; The simulation module is used to: obtain the three-dimensional sample structure of the sample nozzle under different values ​​of the key dimension parameters of the geometric model, and combine the fixed values ​​of the boundary condition parameters of the sample nozzle to simulate each three-dimensional sample structure and obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure. The training module is used to: train a preset deep learning model based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, so as to obtain a nozzle surface pressure field distribution prediction model. The prediction module is used to: determine the three-dimensional structure of the nozzle under test based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, and obtain the surface pressure field distribution prediction result of the nozzle under test based on the geometric feature test data obtained by sampling the three-dimensional structure under test and the nozzle surface pressure field distribution prediction model.

[0016] The beneficial effects of the nozzle surface pressure field distribution prediction system of the present invention are as follows: The system of this invention solves the problem that traditional methods are difficult to adapt to changing geometric model structures, reduces the consumption of manpower and material resources, shortens the nozzle design cycle, reduces costs, and improves nozzle design efficiency. It can better meet the needs of rapid iteration and optimization in modern engineering and overcome the defect of traditional machine learning that cannot be trained after structural changes.

[0017] Thirdly, the technical solution of an electronic device according to the present invention is as follows: It includes a memory, a processor, and a program stored in the memory and running on the processor, wherein the processor executes the program to implement the steps of the nozzle surface pressure field distribution prediction method of the present invention.

[0018] Fourthly, the technical solution of a computer-readable storage medium provided by the present invention is as follows: The computer-readable storage medium stores instructions that, when read, cause the computer-readable storage medium to perform the steps of the nozzle surface pressure field distribution prediction method of the present invention.

[0019] The above description is merely an overview of the technical solution of the present invention. In order to better understand the technical means of the present invention and to implement it in accordance with the contents of the specification, and in order to make the above and other objects, features and advantages of the present invention more apparent and understandable, specific embodiments of the present invention are described below. Attached Figure Description

[0020] The accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Furthermore, the same reference numerals denote the same parts throughout the drawings. In the drawings: Figure 1 This is a schematic flowchart of an embodiment of a nozzle surface pressure field distribution prediction method according to the present invention; Figure 2 A schematic diagram of an ANSYS parametric table; Figure 3 This is a schematic diagram of the Paraview interface; Figure 4 This is a schematic diagram of the predicted surface pressure field distribution. Figure 5 This is a comparison chart of the predicted surface pressure field distribution after repair and the simulated surface pressure field distribution. Figure 6 A visual representation of the training dataset; Figure 7 This is a schematic diagram of an embodiment of a nozzle surface pressure field distribution prediction system according to the present invention; Figure 8 This is a schematic diagram of an embodiment of an electronic device according to the present invention. Detailed Implementation

[0021] Exemplary embodiments of the invention will now be described in more detail with reference to the accompanying drawings. While exemplary embodiments of the invention are shown in the drawings, it should be understood that the invention can be implemented in various forms and should not be limited to the embodiments set forth herein.

[0022] Figure 1 This diagram illustrates a flowchart of an embodiment of a nozzle surface pressure field distribution prediction method provided by the present invention. This method can be executed by an electronic device such as a terminal device or a server. The terminal device can be any fixed or mobile terminal, such as a user equipment (UE), mobile device, user terminal, terminal, cellular phone, cordless phone, personal digital assistant (PDA), handheld device, computing device, vehicle-mounted device, or wearable device. The server can be a single server or a server cluster consisting of multiple servers. Any electronic device can implement the nozzle surface pressure field distribution prediction method by having its processor call computer-readable instructions stored in its memory. Figure 1 As shown, it includes the following steps: S1. Obtain the three-dimensional sample structure of the sample nozzle under different values ​​of the key dimension parameters of the geometric model, and combine the fixed values ​​of the boundary condition parameters of the sample nozzle to simulate each three-dimensional sample structure and obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure.

[0023] Here, the sample nozzle refers to the baseline nozzle model used to generate training data, whose geometry exhibits diverse variations through different combinations of key dimensional parameters. Key dimensional parameters of the geometric model are the core variables controlling the three-dimensional geometry of the nozzle. The three-dimensional sample structure refers to the three-dimensional solid model of the sample nozzle generated by adjusting the key dimensional parameters of the geometric model; each structure corresponds to a specific set of dimensional parameter combinations. Boundary condition parameters refer to the physical environment constraint variables set in the fluid simulation. The surface pressure field distribution simulation results refer to the dataset of the mapping relationship between the spatial position of the nozzle surface and the pressure value obtained after performing fluid dynamics simulation calculations on the three-dimensional sample structure.

[0024] S2. Based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, the preset deep learning model is trained to obtain the nozzle surface pressure field distribution prediction model.

[0025] Among them, geometric feature sample data refers to the set of coordinates extracted after sampling a fixed number of points (e.g., 10,000 points) on the surface of a 3D sample structure, with a uniform dimension of N×3 (N being the number of sampling points). The preset deep learning model refers to the initialized RegDGCNN neural network architecture, which is built based on a dynamic graph convolution mechanism and can process geometric input data in the form of point clouds. The nozzle surface pressure field distribution prediction model refers to the trained RegDGCNN model, which learns the mapping relationship between geometric feature data and pressure field distribution, and has the ability to directly output the pressure field from the input nozzle geometric data.

[0026] It should be noted that the steps for sampling the 3D sample structure to obtain geometric feature sample data include: importing the pressure field distribution data of the nozzle surface or body generated by ANSYS simulation (stored as a .vtu format file) into the processing program; and, based on the principle of spatial uniform distribution or curvature feature sensitivity, extracting a fixed number (e.g., 10,000) of coordinate points on the surface or body of the 3D structure. Each sampling point contains 3D spatial coordinates (x, y, z), ultimately generating a point cloud matrix with a unified dimension of N×4 (N is the number of sampling points, and 4 represents the x, y, z coordinates of the 3D space and the pressure data). This process standardizes the dimension of the input data by fixing the number of samples, ensuring that nozzle data of different geometric structures can be processed by the same deep learning model, while retaining key geometric feature information for the model to learn the pressure field distribution law. In addition, the nozzle surface pressure field distribution prediction model separates the input nodes of the deep neural network from the CFD mesh, which enables the model to predict pressure distribution or other physical field distributions under a wide range of different flow field domains.

[0027] S3. Based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, determine the three-dimensional structure of the nozzle under test, and based on the geometric feature test data obtained by sampling the three-dimensional structure under test, use the nozzle surface pressure field distribution prediction model to obtain the surface pressure field distribution prediction result of the nozzle under test.

[0028] The nozzle under test refers to the target nozzle whose pressure field distribution needs to be predicted in engineering design, and its geometry is determined by actual design requirements. The actual values ​​of the key dimensional parameters of the geometric model refer to the specific values ​​of the key dimensional parameters of the geometric model determined in the engineering application of the nozzle under test. The geometric feature test data refers to the standardized point cloud coordinate dataset generated after sampling the three-dimensional structure of the nozzle under test with a fixed number of points in the same manner as in the training phase. The predicted surface pressure field distribution of the nozzle under test refers to the set of pressure values ​​corresponding to the sampling point locations output after inputting the geometric feature test data into the trained nozzle surface pressure field distribution prediction model, which, after interpolation processing, forms a complete surface pressure field distribution map.

[0029] The technical solution of this embodiment solves the problem that traditional methods are difficult to adapt to changing geometric model structures, reduces the consumption of manpower and material resources, shortens the nozzle design cycle, reduces costs, and improves nozzle design efficiency. It can better meet the needs of rapid iteration and optimization in modern engineering and overcome the defect that traditional machine learning cannot be trained after structural changes.

[0030] In one alternative approach, S2 includes: S21. Input the geometric feature sample data of any three-dimensional sample structure into the preset deep learning model for predicting the pressure field distribution on the nozzle surface, and obtain the predicted result of the surface pressure field distribution of the any three-dimensional sample structure.

[0031] S22. Based on the predicted surface pressure field distribution and the simulation results of the surface pressure field distribution of any three-dimensional sample structure, determine the loss value of any three-dimensional sample structure until the loss value of each three-dimensional sample structure is obtained.

[0032] S23. Based on all the loss values, optimize the parameters of the preset deep learning model to obtain the optimized deep learning model. Use the optimized deep learning model as the preset deep learning model and return to execute S21 until the optimized deep learning model meets the preset iterative training conditions. Then, determine the optimized deep learning model as the nozzle surface pressure field distribution prediction model.

[0033] It should be noted that steps S21-S22 are performed separately for each 3D sample structure to obtain the loss value for each 3D sample structure. The default iterative training condition is: reaching the maximum number of iterations or loss convergence.

[0034] In the above-mentioned optional methods, by inputting geometric feature sample data into a preset deep learning model, calculating the loss value and optimizing the model parameters until the preset iterative training conditions are met, a high-precision nozzle surface pressure field distribution prediction model is obtained, which further improves the accuracy and reliability of the prediction results and provides a more powerful reference for nozzle design.

[0035] In one alternative approach, the key dimensional parameters of the geometric model include: the size of the oxygen-enriched gas inlet, the size of the fuel-enriched gas inlet, and the channel radius.

[0036] The oxygen-enriched gas inlet size refers to the cross-sectional area or diameter of the channel in the nozzle structure used to input high-oxygen-concentration gas, and its value determines the flow capacity of the oxygen-enriched gas into the mixing chamber. The fuel-enriched gas inlet size controls the flow distribution characteristics of the fuel-enriched gas before mixing. The channel radius refers to the cross-sectional radius of the main flow channel inside the nozzle, which directly affects the fluid velocity distribution and pressure loss characteristics.

[0037] There are fixed constraints between the size of the oxygen-enriched gas inlet, the size of the gas-enriched gas inlet, and the radius of the channel. For example, the size of the oxygen-enriched gas inlet must be smaller than the size of the gas-enriched gas inlet, and the size of the gas-enriched gas inlet must be smaller than the radius of the channel. These can be set according to the actual situation, and no restrictions are set here.

[0038] It should be noted that in this embodiment, the Latin hypercube sampling algorithm is used, and based on actual engineering requirements and the physical laws of nozzle operation, as well as the constraints between key dimension parameters, different values ​​of key dimension parameters of the sample nozzle in the geometric model are obtained to obtain multiple three-dimensional sample structures.

[0039] In the above-mentioned optional methods, by specifying key dimensional parameters of the geometric model, such as the size of the oxygen-enriched gas inlet, the size of the fuel-enriched gas inlet, and the channel radius, the pressure field distribution of key parts of the nozzle can be predicted in a targeted manner, which can better meet the needs of performance optimization of key parts of the nozzle in engineering design and improve the accuracy and effectiveness of nozzle design.

[0040] In one alternative approach, the boundary condition parameters include: inlet pressure, fluid flow rate, and outlet pressure.

[0041] In this context, inlet pressure refers to the pressure applied by the fluid at the nozzle inlet boundary, as set in ANSYS fluid simulation, and is expressed in Pa. Fluid flow rate refers to the volume or mass of fluid passing through the nozzle inlet per unit time, as set in ANSYS fluid simulation, and is expressed in cubic meters per second (m³). 3 / s or kg / s. The outlet pressure refers to the reverse pressure value applied by the environment at the nozzle outlet boundary as set in ANSYS fluid simulation, and the unit is Pa.

[0042] In the above-mentioned optional methods, by fixing the values ​​of the boundary condition parameters, the influence of boundary condition changes on the simulation results is eliminated, making the obtained simulation results more reflective of the influence of changes in key dimensional parameters of the geometric model on the pressure field distribution on the nozzle surface. This improves the efficiency of model training and the accuracy of prediction results, and provides a more reliable basis for nozzle geometry optimization.

[0043] In one alternative approach, the step of simulating each three-dimensional sample structure to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure includes: Using ANSYS software, each three-dimensional sample structure was simulated separately to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure.

[0044] ANSYS software refers to a commercial computing platform used to perform parametric fluid dynamics simulations. It solves the Navier-Stokes equations using its Fluent or CFX modules and outputs pressure field distribution data on the nozzle surface. In this embodiment, the key dimensional parameters and boundary condition parameters of the geometric model corresponding to the three-dimensional sample structure are written into an ANSYS parametric table (e.g., ...). Figure 2(As shown). After obtaining the simulation results of the surface pressure field distribution for each 3D sample structure, the simulation results are imported into Paraview software. Using Paraview's data conversion function, the result files are converted to .vtu format files, resulting in the pressure field distribution data file for each 3D sample structure (one 3D sample structure corresponds to one working condition). The .vtu file format facilitates subsequent deep learning model reading and processing, and can effectively store the geometric information and physical quantity data of the pressure field distribution. Since the result files for each group of nozzles are in different folders, one hundred sets of data need to be read manually one by one. However, a script can be run using the console to read them in batches. The results of all read data are shown below. Figure 3 As shown.

[0045] In the above-mentioned optional methods, ANSYS software is further used to simulate the three-dimensional sample structure. By utilizing its powerful fluid dynamics simulation capabilities, high-precision simulation results of surface pressure field distribution can be obtained, providing high-quality data support for the training of deep learning models, thereby improving the performance of the nozzle surface pressure field distribution prediction model and the reliability of the prediction results.

[0046] In one alternative approach, the preset deep learning model is the RegDGCNN model.

[0047] Specifically, the RegDGCNN model refers to a deep learning architecture based on dynamic graph convolutional neural networks, specifically designed for processing 3D point cloud data and regressing physical field distributions. The RegDGCNN model structure comprises the following core components: ① An input layer that receives the coordinates of a fixed number of 3D point clouds (e.g., 10,000 points); ② A transformation network that aligns the input point cloud coordinates using a spatial transformation matrix (T-Net); ③ A dynamic graph construction module that uses the K-nearest neighbor (KNN) algorithm to calculate the local neighborhood relationships for each point in real time, generating a dynamic relationship graph; ④ Multi-layer graph convolutional layers that use the core formula... Extract local geometric features, where, It is the first Layers Features It is a point The neighborhood group, and These are learnable weights and biases. It is an activation function (such as ReLU). For point With neighboring points The RegDGCNN model achieves rapid prediction of the surface pressure field of a variable nozzle structure by learning the mapping relationship between the geometric structure and the pressure field from the end-to-end. This overcomes the input dimensionality limitations caused by changes in grid topology in traditional methods. The model employs a multi-scale local feature aggregation layer to fuse multi-scale local features into a global description, and a fully connected layer to output pressure values ​​that correspond one-to-one with the input points.

[0048] Among the above optional methods, the RegDGCNN model is further adopted as the preset deep learning model. This model is built based on the dynamic graph convolution mechanism, which can effectively process geometric input data in the form of point clouds, better capture the complex mapping relationship between the geometric features of the nozzle surface and the pressure field distribution, improve the accuracy and generalization ability of the prediction model, and make the prediction results closer to the actual engineering situation.

[0049] In one alternative approach, it also includes: The predicted surface pressure field distribution of the nozzle under test is visualized and output.

[0050] In the above-mentioned optional methods, by visualizing the predicted surface pressure field distribution of the nozzle under test, designers can more intuitively understand the distribution of the surface pressure field of the nozzle, which facilitates the analysis and optimization of nozzle performance, further improving the efficiency and quality of nozzle design, and also enhancing the practicality of the invention in engineering applications and user experience.

[0051] It should be noted that the predicted surface pressure field distribution of the nozzle under test is as follows: Figure 4 As shown. To restore the predicted surface pressure field distribution to the three-dimensional structure under test, other values ​​need to be repaired and interpolated. This embodiment proposes using the KNN method for interpolation and repair, which can restore the pressure distribution very well. The comparison between the repaired results and the simulation results is as follows: Figure 5 As shown.

[0052] To better illustrate the technical solution of this embodiment, the following example is used for explanation: The Latin hypercube sampling method was used to obtain a table of nozzle structural parameters (.csv file). After importing it into ANSYS, batch simulations were performed and automatically saved. In this embodiment, the simulation results were used to calculate the surface pressure field distribution of the nozzle. Then, Paraview was used to read the .cas.h5 file, batch importing 100 sets of results, which were then converted into .vtu files. A RegDGCNN model was trained in Python, sampling 10,000 points for each nozzle, resulting in 100 sets of simulation data, with 75 sets used as training data. Figure 6(This is a schematic diagram of the training dataset; the key dimensional parameters of the geometric model for each 3D sample structure have different values.) There were 15 sets of test data and 10 sets of validation data, with a batch size of 5, and the training was conducted for 15 generations. After training, the model parameters were exported as a .pth file. During subsequent prediction, the parameter file was read again, yielding pressure data from 10,000 sampling points (predicted surface pressure field distribution). The pressure was then restored to the coordinate points, and interpolation was performed on other locations to obtain matching prediction results, which were almost identical to the simulation results. In this embodiment, pressure distribution estimation for a nozzle with a 300k hexahedral mesh was completed within 2 seconds on an NVIDIA GeForce GTX 1650, a significant improvement in efficiency compared to the 2 hours required for a standard CFD simulation using a 52-core Intel(R) Xeon(R) Platinum 8272CL CPU.

[0053] Figure 7 A schematic diagram of an embodiment of a nozzle surface pressure field distribution prediction system 200 provided by the present invention is shown. Figure 7 As shown, the system 200 includes: a simulation module 210, a training module 220, and a prediction module 230; The simulation module 210 is used to: obtain the three-dimensional sample structure of the sample nozzle under different values ​​of the key dimension parameters of the geometric model, and combine the fixed values ​​of the boundary condition parameters of the sample nozzle to simulate each three-dimensional sample structure and obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure. The training module 220 is used to: train a preset deep learning model based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, so as to obtain a nozzle surface pressure field distribution prediction model. The prediction module 230 is used to: determine the three-dimensional structure of the nozzle under test based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, and obtain the surface pressure field distribution prediction result of the nozzle under test based on the geometric feature test data obtained by sampling the three-dimensional structure under test and the nozzle surface pressure field distribution prediction model.

[0054] In one alternative embodiment, the training module 220 includes: a first training module, a second training module, a third training module, and an iterative training module; The first training module is used for: training a preset deep learning model based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, to obtain a nozzle surface pressure field distribution prediction model, including: The second training module is used to: input the geometric feature sample data of any three-dimensional sample structure into the preset deep learning model for predicting the pressure field distribution on the nozzle surface, and obtain the predicted result of the surface pressure field distribution of the any three-dimensional sample structure; The third training module is used to: determine the loss value of any three-dimensional sample structure based on the predicted surface pressure field distribution and the simulation results of the surface pressure field distribution of any three-dimensional sample structure, until the loss value of each three-dimensional sample structure is obtained. The iterative training module is used to: optimize the parameters of the preset deep learning model based on all loss values ​​to obtain an optimized deep learning model; use the optimized deep learning model as the preset deep learning model and return to call the first training module; until the optimized deep learning model meets the preset iterative training conditions, determine the optimized deep learning model as the nozzle surface pressure field distribution prediction model.

[0055] In one alternative approach, the key dimensional parameters of the geometric model include: the size of the oxygen-enriched gas inlet, the size of the fuel-enriched gas inlet, and the channel radius.

[0056] In one alternative approach, the boundary condition parameters include: inlet pressure, fluid flow rate, and outlet pressure.

[0057] In an alternative embodiment, the simulation module 210 is specifically used for: Using ANSYS software, each three-dimensional sample structure was simulated separately to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure.

[0058] In one alternative approach, the preset deep learning model is the RegDGCNN model.

[0059] In one alternative approach, it also includes: an output module; The output module is used to visualize and output the predicted surface pressure field distribution of the nozzle under test.

[0060] It should be noted that the beneficial effects of the nozzle surface pressure field distribution prediction system 200 provided in the above embodiments are the same as those of the nozzle surface pressure field distribution prediction method described above, and will not be repeated here. Furthermore, the system provided in the above embodiments is only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the system can be divided into different functional modules according to the actual situation to complete all or part of the functions described above. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and their specific implementation process is detailed in the method embodiments, and will not be repeated here.

[0061] The nozzle surface pressure field distribution prediction system 200 of the present invention can be a computer program (including program code) running on a computer device. For example, the nozzle surface pressure field distribution prediction system 200 of the present invention is an application software that can be used to execute the corresponding steps in the nozzle surface pressure field distribution prediction method of the present invention.

[0062] In some embodiments, the nozzle surface pressure field distribution prediction system 200 of the present invention can be implemented in a combination of hardware and software. As an example, the nozzle surface pressure field distribution prediction system 200 of the present invention can be a processor in the form of a hardware decoding processor, which is programmed to execute the nozzle surface pressure field distribution prediction method of the present invention. For example, the processor in the form of a hardware decoding processor can be one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), or other electronic components.

[0063] The modules described in the embodiments of this invention can be implemented in software or hardware. The names of the modules are not, in some cases, limiting the scope of the module itself.

[0064] An electronic device according to an embodiment of the present invention includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements any of the above-described nozzle surface pressure field distribution prediction methods. That is, an electronic device according to an embodiment of the present invention may include, but is not limited to: a processor and a memory; the memory is used to store the computer program; the processor is used to execute the nozzle surface pressure field distribution prediction method shown in any embodiment of the present invention by calling the computer program.

[0065] In one alternative embodiment, an electronic device is provided, such as Figure 8 As shown, Figure 8The illustrated electronic device 4000 includes a processor 4001 and a memory 4003. The processor 4001 and the memory 4003 are connected, for example, via a bus 4002. Optionally, the electronic device 4000 may further include a transceiver 4004, which can be used for data interaction between the electronic device and other electronic devices, such as sending and / or receiving data. It should be noted that in practical applications, the transceiver 4004 is not limited to one type, and the structure of the electronic device 4000 does not constitute a limitation on the embodiments of the present invention.

[0066] Processor 4001 may be a CPU (Central Processing Unit), a general-purpose processor, a DSP (Digital Signal Processor), an ASIC (Application Specific Integrated Circuit), an FPGA (Field Programmable Gate Array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. Processor 4001 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0067] Bus 4002 may include a path for transmitting information between the aforementioned components. Bus 4002 may be a PCI (Peripheral Component Interconnect) bus or an EISA (Extended Industry Standard Architecture) bus, etc. Bus 4002 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 8 The bus 4002 is represented by only one thick line, but this does not mean that there is only one bus or one type of bus.

[0068] The memory 4003 may be ROM (Read Only Memory) or other types of static storage devices capable of storing static information and instructions, RAM (Random Access Memory) or other types of dynamic storage devices capable of storing information and instructions, or EEPROM (Electrically Erasable Programmable Read Only Memory), CD-ROM (Compact Disc Read Only Memory) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but not limited thereto.

[0069] The memory 4003 stores application code (computer program) for executing the present invention, and its execution is controlled by the processor 4001. The processor 4001 executes the application code stored in the memory 4003 to implement the content shown in the foregoing method embodiments.

[0070] Among them, electronic devices can also be terminal devices. A terminal device can be any terminal device that can install applications and access web pages through applications, including at least one of smartphones, tablets, laptops, desktop computers, smart speakers, smartwatches, smart TVs, and smart in-vehicle devices.

[0071] It should be noted that, Figure 8 The electronic device shown is merely an example and should not be construed as limiting the functionality and scope of use of the embodiments of the present invention.

[0072] An embodiment of the present invention provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements any of the above-described methods for predicting the pressure field distribution on the nozzle surface.

[0073] Alternatively, the computer-readable storage medium may be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), magnetic tape, a floppy disk, and an optical data storage device, etc.

[0074] In an exemplary embodiment, a computer program product or computer program is also provided, which includes computer instructions stored in a computer-readable storage medium. A processor of an electronic device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the electronic device to perform the above-described nozzle surface pressure field distribution prediction method.

[0075] Computer program code for performing the operations of this invention can be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a local area network (LAN) or a wide area network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).

[0076] It should be understood that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of methods and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, may be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0077] The computer-readable storage medium provided in this invention can be, but is not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this invention, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0078] The aforementioned computer-readable storage medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the method shown in the above embodiments.

[0079] The above description is merely a preferred embodiment of the present invention and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of disclosure in this invention is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the above-disclosed concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this invention.

[0080] It should be noted that the terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and represent a limitation on a specific order or sequence. Where appropriate, the order of use for similar objects can be interchanged so that the embodiments of this application described herein can be implemented in an order other than that shown or described.

[0081] Those skilled in the art will recognize that this invention can be implemented as a system, method, or computer program product. Therefore, this invention can be specifically implemented in the following forms: it can be entirely hardware, entirely software (including firmware, resident software, microcode, etc.), or a combination of hardware and software, generally referred to herein as a "circuit," "module," or "system." Furthermore, in some embodiments, this invention can also be implemented as a computer program product contained in one or more computer-readable media, which includes computer-readable program code.

[0082] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for predicting the pressure field distribution on a nozzle surface, characterized in that, include: The three-dimensional sample structure of the sample nozzle under different values ​​of the key dimensional parameters of the geometric model is obtained, and the surface pressure field distribution simulation results of each three-dimensional sample structure are obtained by combining the fixed values ​​of the boundary condition parameters of the sample nozzle. Based on the geometric feature sample data obtained by sampling each three-dimensional sample structure, and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, the preset deep learning model is trained to obtain the nozzle surface pressure field distribution prediction model. Based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, the three-dimensional structure of the nozzle under test is determined. Based on the geometric feature test data obtained by sampling the three-dimensional structure under test, the surface pressure field distribution prediction result of the nozzle under test is obtained using the nozzle surface pressure field distribution prediction model.

2. The method for predicting the pressure field distribution on the nozzle surface according to claim 1, characterized in that, The steps for training a pre-defined deep learning model to obtain a nozzle surface pressure field distribution prediction model, based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, include: The geometric feature sample data of any three-dimensional sample structure is input into the preset deep learning model used for nozzle surface pressure field distribution prediction to obtain the surface pressure field distribution prediction result of any three-dimensional sample structure. Based on the predicted surface pressure field distribution and the simulation results of the surface pressure field distribution of any three-dimensional sample structure, the loss value of any three-dimensional sample structure is determined until the loss value of each three-dimensional sample structure is obtained. Based on all the loss values, the parameters of the preset deep learning model are optimized to obtain an optimized deep learning model. The optimized deep learning model is then used as the preset deep learning model, and the step of inputting the geometric feature sample data of any three-dimensional sample structure into the preset deep learning model for nozzle surface pressure field distribution prediction is returned to be executed until the optimized deep learning model meets the preset iterative training conditions. At this point, the optimized deep learning model is determined as the nozzle surface pressure field distribution prediction model.

3. The method for predicting the pressure field distribution on the nozzle surface according to claim 1, characterized in that, The key dimensional parameters of the geometric model include: the size of the oxygen-enriched gas inlet, the size of the fuel-enriched gas inlet, and the channel radius.

4. The method for predicting the pressure field distribution on the nozzle surface according to claim 3, characterized in that, The boundary condition parameters include: inlet pressure, fluid flow rate, and outlet pressure.

5. The method for predicting the pressure field distribution on the nozzle surface according to claim 1, characterized in that, The steps for simulating each three-dimensional sample structure to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure include: Using ANSYS software, each three-dimensional sample structure was simulated separately to obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure.

6. The method for predicting the pressure field distribution on the nozzle surface according to any one of claims 1 to 5, characterized in that, The preset deep learning model is the RegDGCNN model.

7. The method for predicting the pressure field distribution on the nozzle surface according to any one of claims 1 to 5, characterized in that, Also includes: The predicted surface pressure field distribution of the nozzle under test is visualized and output.

8. A nozzle surface pressure field distribution prediction system, characterized in that, include: Simulation module, training module, and prediction module; The simulation module is used to: obtain the three-dimensional sample structure of the sample nozzle under different values ​​of the key dimension parameters of the geometric model, and combine the fixed values ​​of the boundary condition parameters of the sample nozzle to simulate each three-dimensional sample structure and obtain the simulation results of the surface pressure field distribution of each three-dimensional sample structure. The training module is used to: train a preset deep learning model based on the geometric feature sample data obtained by sampling each three-dimensional sample structure and the simulation results of the surface pressure field distribution of each three-dimensional sample structure, so as to obtain a nozzle surface pressure field distribution prediction model. The prediction module is used to: determine the three-dimensional structure of the nozzle under test based on the actual values ​​of the key dimension parameters of the geometric model of the nozzle under test, and obtain the surface pressure field distribution prediction result of the nozzle under test based on the geometric feature test data obtained by sampling the three-dimensional structure under test and the nozzle surface pressure field distribution prediction model.

9. An electronic device, characterized in that, The electronic device includes a processor coupled to a memory storing at least one computer program, which is loaded and executed by the processor to enable the electronic device to implement the nozzle surface pressure field distribution prediction method as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one computer program, which is loaded and executed by a processor to enable the computer-readable storage medium to implement the nozzle surface pressure field distribution prediction method as described in any one of claims 1 to 7.

Images