Real-time simulation method for stress field of semi-flexible wall nozzle in wind tunnel

Through the model step reduction technology combined with finite element simulation and Modelica, real-time simulation of the stress field of the semi-flexible wall nozzle in the wind tunnel is realized, solving the problem of long calculation time in the existing technology, and ensuring the safety and operating efficiency of the nozzle section .

CN118838196BActive Publication Date: 2024-11-19INST OF HIGH SPEED AERODYNAMICS OF CHINA AERODYNAMICS RES & DEV CENT
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411320348.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-23
Publication Date
2024-11-19
Estimated Expiration
2044-09-23

AI Technical Summary

Technical Problem

The prior art is difficult to monitor and analyze the stress field of the semi-flexible wall nozzle in wind tunnel in real time, resulting in a long calculation time and the safety and operation efficiency of the nozzle section cannot be guaranteed.

Method used

The finite element simulation model is used and Modelica is combined to generate a 1D Modelica model through the model reduction method, integrating the semi-soft wall servo motor and mechanical motion model to form an overall system model, and using MWORKS software to simulate it to realize real-time calculation and display of stress fields. .

Benefits of technology

It significantly reduces simulation time and can display the stress field of the semi-flexible wall nozzle in real time, ensuring the safety and smooth progress of the test, and improving the efficiency and accuracy of wind tunnel tests.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118838196B_ABST
    Figure CN118838196B_ABST
Patent Text Reader

Abstract

The invention discloses a real-time simulation method for a wind tunnel semi-flexible wall nozzle stress field, which relates to the field of semi-flexible wall nozzle stress field simulation, including: obtaining a reduced-order model based on a finite element simulation model and a Modelica joint simulation, and encapsulating the reduced-order model into a 1D Modelica model; opening the reduced-order model in MWORKS software, docking a semi-flexible wall servo motor model and a semi-flexible wall mechanical motion model, and obtaining a semi-flexible wall 1D mechanism subsystem model; opening the reduced-order model in MWORKS software, docking the semi-flexible wall 1D mechanism subsystem model with a *.mo file in a semi-flexible wall nozzle 1D Modelica model, integrating them into a semi-flexible wall overall system model, so as to complete the simulation based on the semi-flexible wall overall system model; performing simulation in MWORKS software, and viewing the stress field calculation results in real time. The invention provides a real-time simulation method for a wind tunnel semi-flexible wall nozzle stress field, which does not need to pay attention to the solution principle and process of the model, and automatically performs a reduction solution to the existing equations when the model is solved, and the model has the advantage of being easy to use.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of semi-flexible wall nozzle stress field simulation, and more specifically, to a real-time simulation method for the stress field of a semi-flexible wall nozzle in a wind tunnel. Background Art

[0002] Wind tunnels are essential equipment for aerodynamic experiments. Through special pipeline design and power systems, the flow rate (Mach number) of the internal air is controlled to simulate the real aerodynamic environment. At present, in order to control the Mach number and obtain a uniform air flow field, each transonic and supersonic wind tunnel uses a nozzle section to accelerate the airflow from low subsonic speed to the speed of sound, and then reaches the required Mach number at the nozzle outlet.

[0003] The main forms of nozzles in general wind tunnels are solid block nozzles, full semi-flexible wall nozzles and semi-flexible wall nozzles. Semi-flexible wall nozzles are widely used due to their high operating efficiency, simple manufacturing and low processing difficulty. Semi-flexible wall nozzles use special steel plates as the profile, which are driven by the actuator in the nozzle to change the shape of the steel plate to form a profile curve with the required Mach number. As one of the core structural components of the entire wind tunnel, the nozzle will not only be driven by the actuator during the test, but also withstand various complex aerodynamic loads. Therefore, it is very necessary to conduct stress and deformation analysis on the semi-flexible wall nozzle to ensure the safe and reliable operation of the nozzle section.

[0004] At present, the finite element analysis method is mainly used for the stress and deformation analysis of semi-flexible wall nozzles. Due to the large size and complex structure of semi-flexible wall nozzles, there are problems such as a large number of grids and long calculation time when performing finite element analysis. Therefore, the structural strength and safety of the semi-flexible wall nozzle can only be analyzed by adding extreme working condition boundary conditions to the established finite element simulation model, and the stress field of the nozzle cannot be monitored in real time.

[0005] Based on these problems, scholars at home and abroad have found that model reduction can better solve the problem that finite element simulation cannot be simulated in real time with the system model under complex field conditions. Model reduction (ROM) is a model simulation method that expands a high-dimensional nonlinear complex model into a linear superposition model based on the order reduction method, and reconstructs and generates a new model by truncating the high-order quantity and retaining the low-order quantity. In practical applications, high-order models often have too much redundant information and complexity, while low-order models can better express the key characteristics and laws of the system. Feature extraction is performed through a large amount of data (which can be CAE calculation results or test data), and a proxy model method is used to reconstruct a highly approximate mapping. This mapping can give calculation results at the speed of real-time simulation while ensuring a certain accuracy. However, there is currently no research on how to use the reduced-order model to solve practical problems in the real-time simulation of the stress field of a semi-flexible wall nozzle in a wind tunnel. Summary of the invention

[0006] An object of the present invention is to solve at least the above problems and / or disadvantages and to provide at least the advantages which will be described hereinafter.

[0007] In order to achieve these purposes and other advantages of the present invention, a real-time simulation method for a wind tunnel semi-flexible wall nozzle stress field is provided, comprising:

[0008] S1. A reduced-order model is obtained based on the finite element simulation model and Modelica joint simulation, and the reduced-order model is encapsulated as a 1D Modelica model. The model directory structure exported from the 1D Modelica model includes a *.mo file, a Resourse folder, and an exte_file folder;

[0009] S2. In MWORKS software, the semi-flexible wall servo motor model and the semi-flexible wall mechanical motion model are connected to obtain the semi-flexible wall 1D mechanism subsystem model;

[0010] S3. Open the reduced-order model in MWORKS software, connect the semi-flexible wall 1D mechanism subsystem model with the *.mo file in the semi-flexible wall nozzle 1D Modelica model, and integrate them into a semi-flexible wall overall system model to complete the simulation based on the semi-flexible wall overall system model;

[0011] S4. After simulation in MWORKS software, the stress field calculation results can be viewed in real time.

[0012] Preferably, in S1, the process of obtaining the reduced-order model includes:

[0013] S10. Establish a finite element simulation model based on the semi-flexible wall nozzle assembly model;

[0014] S11. Based on the finite element simulation model, the simulation results are solved according to the maximum operating displacement range of the semi-flexible wall nozzle, and the simulation results are saved as a file with a suffix of OP2;

[0015] S12. Based on the OP2 file, the Multiviz software and MWORKS software are used to perform parsing and training to obtain a trained reduced-order model.

[0016] Preferably, in S10, the process of establishing the finite element simulation model includes:

[0017] S101. Import the semi-flexible wall nozzle assembly model established in the 3D software into the finite element analysis software Simcenter 3D;

[0018] S102, simplify the structural features that do not need to be involved in the analysis, and only retain the geometry of the effective force-bearing part;

[0019] S103, entering the simulation environment to construct, set working conditions, and solve finite element simulation models of various components;

[0020] S104. In the post-processing results, define the output displacement results, deformation results, stress results, and node force results to obtain the corresponding finite element simulation model.

[0021] Preferably, in S11, the process of solving the simulation result according to the maximum operating displacement range of the semi-flexible wall nozzle is:

[0022] S111. Carry out material definition related operations in FEM finite element environment;

[0023] S112, selecting a mesh type, setting a mesh size, and selecting a material for each component;

[0024] S113, constraining and setting loads for each component;

[0025] S114, select a finite element simulation model, perform simulation, define a solution name, and select a solution type;

[0026] The contents of constraint and load settings include:

[0027] The actuator mechanism of the actuator cylinder is simulated by a sliding pair, the joint brackets are constrained by a rotating pair, and the constraint type of the electric cylinder base is a fixed constraint;

[0028] In Load Type, select the Enforced Motion load type;

[0029] Select 1D element as the model object, enter the displacement in the x-direction, and add a gravity load to the model, selecting the gravity type.

[0030] Preferably, in S12, the process of training the reduced-order model is:

[0031] S121, import the OP2 file into Multiviz software to perform CAE data calculation;

[0032] S122. Based on the calculation results, MWORKS software is used to perform order reduction training to generate a reduced-order model.

[0033] Preferably, the CAE data is calculated as follows:

[0034] Based on the simulation results of S11, a simulation result sample set of semi-flexible wall displacement input is obtained based on the semi-flexible wall positive displacement finite element simulation and semi-flexible wall negative displacement finite element simulation, and on this basis, semi-flexible wall monitoring point data and semi-flexible wall physical field data are generated;

[0035] The physical quantity of the data in the working condition is extracted and the data is merged, and then the merged data is processed by dimension reduction to obtain a training data set after dimension reduction;

[0036] The content of the training data set includes: input data and DetaModel data.

[0037] Preferably, in S122, the scale reduction training is performed as follows:

[0038] Import the training data set into MWORKS software and classify the variables;

[0039] Select input data as input variable and DetaModel data as output variable;

[0040] Set the model solution parameters in the model processing interface and perform training. After the solution is completed, test it. When the result meets expectations, the model can be encapsulated as a reduced-order model. If the result does not meet expectations, it is necessary to iteratively modify the parameters to solve the model until the model meets expectations.

[0041] After the reduced-order training is completed, the model is packaged as a 1D Modelica model and exported.

[0042] Preferably, the method further comprises: loading a corresponding *.VTK file in MWORKS software for simulation, and viewing the stress field calculation results in real time.

[0043] The present invention has at least the following beneficial effects:

[0044] First, the real-time simulation method of the stress field of the semi-flexible wall nozzle of the wind tunnel of the present invention is modeled based on the Modelica language, and adopts the statement-process modeling method. It only needs to simplify the actual problem into a physical model appropriately, and then describe its specific physical phenomenon through mathematical equations, without paying attention to the solution principle and process of the model. When the model is solved, the existing equations are automatically solved by the specification. The model has the advantages of being easy to use and fast solving speed. That is, in practical application, the present invention involves multiple disciplines and fields such as mechanics, electrical, and control to solve multi-domain coupling problems in complex engineering system simulation, so Modelica is selected as the modeling language. Modelica is an open, object-oriented, equation-based language suitable for large-scale complex heterogeneous physical system modeling. It has the ability to cross different fields and conveniently realize the modeling of multiple fields of complex physical systems such as mechanics, electronics, electricity, hydraulics, heat, and control. It is widely used in the fields of aviation, aerospace, automobiles, ships, etc. Through the statement-process modeling method (i.e., describing physical behavior based on physical equations, without clarifying the cause-effect relationship or data flow of the equation system), the model solving efficiency is faster. Declarative-process modeling focuses on the description of the problem itself, shielding the details of the solution to the problem, which is more in line with the cognitive model of most problems in the objective world. In engineering physical system modeling, declarative modeling is more conducive to engineers focusing on design rather than solving.

[0045] Secondly, the real-time simulation method of the stress field of the semi-flexible wall nozzle of the wind tunnel of the present invention introduces state variables that change with time, and adopts the integral method, so that the variables do not need to solve nonlinear equations, and can be directly integrated to reduce the nonlinearity of the model. Specifically, for the Modelica language, its object-oriented modeling characteristics lead to the physical constitutive modeling of its components, and the constraints between the components are expressed in the connection mode of the connector. At this time, the coupling between the differential equation expressing the constitutive model and the algebraic equation of the connection transformation will form a general form of differential-algebraic equation. For differential-algebraic equations, derivatives are generally not explicitly expressed, and even the derivatives of some dependent variables may not appear in these equations. Therefore, the derivative of the dependent variable x is explicitly expressed in the form of independent transient variables time t and dependent variable x. At this time, the array x is regarded as a set of state variables. As long as the function f has sufficient continuity, the unique solution of the initial value problem can always be found. Therefore, the state variables of the differential-equation system can be selected before the model is solved, or the derivative of the dependent variable x is given (generally defined by the built-in der function of Modelica) to reduce the nonlinearity of the model, which is conducive to the solution of the system.

[0046] Thirdly, the real-time simulation method of the stress field of the semi-flexible wall nozzle in the wind tunnel of the present invention can display the stress field of the semi-flexible wall nozzle in real time during the wind tunnel test, solving the problem of long calculation time of the current finite element simulation method, facilitating direct and real-time observation of the stress condition of the semi-flexible wall nozzle, and ensuring the safe and smooth progress of the test.

[0047] Other advantages, objectives and features of the present invention will be embodied in part through the following description, and in part will be understood by those skilled in the art through study and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 It is a flow chart of the real-time simulation method for the stress field of a semi-flexible wall nozzle in a wind tunnel according to the present invention;

[0049] Figure 2 This is a structural diagram of the semi-flexible wall overall system model of the present invention. DETAILED DESCRIPTION

[0050] The present invention is further described in detail below in conjunction with the accompanying drawings so that those skilled in the art can implement the invention with reference to the description.

[0051] Since the semi-flexible wall nozzle in the wind tunnel is seriously loaded during the test, it is necessary to perform stress analysis on it to ensure safe operation. In view of the problems of slow calculation speed and inability to display simulation results in real time in current simulation methods, a real-time simulation method for stress field of semi-flexible wall nozzle in wind tunnel is provided. This method is based on finite element and Modelica joint simulation. The simulation calculation results are reduced in order by model order reduction to achieve real-time simulation of stress field of semi-flexible wall nozzle, which greatly reduces the simulation time and can obtain data results comparable to real-time simulation within its operating range, thereby solving the above problems.

[0052] A real-time simulation method for a wind tunnel semi-flexible wall nozzle stress field of the present invention comprises:

[0053] S1. Establish a finite element simulation model based on the semi-flexible wall nozzle assembly model;

[0054] The specific operation process of S1 includes:

[0055] Import the semi-flexible wall nozzle assembly model established in the 3D software into the finite element analysis software Simcenter 3D; by constructing idealized components, promote the models that need to be analyzed to the idealized components, simplify the structural features in the model that are little affected by the force, and retain the geometry of the effective force-bearing parts; enter the simulation environment, build the finite element simulation model of each component, and then set and solve the working conditions; in the post-processing results, define the output displacement results, deformation results, stress results, and node force results.

[0056] S2. Based on the finite element simulation model, the simulation results are solved according to the maximum operating displacement range of the semi-flexible wall nozzle, and the simulation results are saved as a file with the suffix OP2;

[0057] The specific operation process of S2 includes:

[0058] S21. Material definition: In the FEM finite element environment, perform operations related to material definition;

[0059] S22, mesh division; selecting mesh type, mesh size and material for each part;

[0060] S23, constraint and load settings; the actuator is simulated by sliding pairs, the joint brackets are constrained by rotating pairs, and the electric cylinder base constraint type is fixed constraint; in the load type, select the forced motion load type; select 1D unit as the model object, enter the displacement size in the x direction, and add gravity load to the model, and select the gravity type;

[0061] S24, solving calculations; selecting a finite element simulation model, performing simulation, defining a solution name, and selecting a solution type;

[0062] S25. Export the calculation result into a file with a suffix of OP2.

[0063] S3: Based on the OP2 file, the OP2 file is analyzed by Multiviz and MWORKS software to output the coordinates and point correspondence of the finite element grid points, the correspondence between the displacement and stress of the semi-flexible wall nozzle, and the *.VTK file for stress cloud display;

[0064] The specific operation process of S3 includes:

[0065] S31. Import the OP2 file into Multiviz software to perform CAE data calculation, that is, based on the simulation results, a simulation result sample set of the semi-flexible wall displacement input is obtained based on the semi-flexible wall positive displacement finite element simulation and the semi-flexible wall negative displacement finite element simulation (the positive displacement and negative displacement here refer to the maximum positive and negative displacement conditions, so as to include the displacement range of the semi-flexible wall, and the simulation result sample set is obtained by calculating the conditions of maximum positive and negative displacements, and the result sample includes all displacement conditions of the semi-flexible wall operation), and on this basis, the semi-flexible wall monitoring point data and the semi-flexible wall physical field data are generated; the physical quantity of the data in the condition is extracted and the data is merged, and then the merged data is processed by dimensionality reduction to obtain the training data after dimensionality reduction; the training data after dimensionality reduction includes: input data and DetaModel data;

[0066] S32. Based on the calculation results, MWORKS software is used to reduce the model order, that is, the data set to be trained is imported into MWORKS software, and the variables are classified; the input data is selected as the input variable, and the DetaModel data is selected as the output variable; the model solution parameters are set in the model processing interface, and training is performed; after the solution is completed, the test is performed, and when the result meets the expectation, the model can be encapsulated as a reduced-order model; if the result does not meet the expectation, the parameters need to be iteratively modified to solve the model until the model meets the expectation;

[0067] After the reduced-order model training is completed, the model is encapsulated as a 1D Modelica model and exported. The 1D Modelica model contains multi-domain joint simulation interfaces, reduced-order models, icons, etc., which can be integrated with the Modelica equation model established in the MWORKS software for simulation. The exported model directory structure consists of *.mo files, Resourse folders, and exte_file folders; among them, Resourse places data modes, and exte_file is used to place dynamic libraries and py files of reduced-order models; the physical field reduced-order model needs to place *.VTK files, icons, etc. in Resourse, among which *.VTK files are used to provide the spatial coordinates of the original geometric points, point connection relationships, and mesh topology information for three-dimensional visualization. It should be noted that in actual applications, *.mo files are used in MWORKS software for docking mechanism subsystems. The Resourse folder and exte_file folder contain training data. When the *.mo file is simulated in MWORKS, the data in these two folders are called to complete the corresponding data transmission.

[0068] S4: The data on the correspondence between the displacement and stress of the semi-flexible wall nozzle is reduced by the model order to retain the key information and features; the stress of the entire structural field of the semi-flexible wall nozzle is calculated by the displacement of the actuator.

[0069] The specific operation process of S4 includes:

[0070] Open the reduced-order model in MWORKS software, connect the semi-flexible wall servo motor model and the semi-flexible wall mechanical motion model, integrate the semi-flexible wall 1D mechanism subsystem model, and then connect the semi-flexible wall 1D mechanism subsystem model with the *.mo file in the semi-flexible wall nozzle 1DModelica model to form a semi-flexible wall overall system model; and perform simulation based on the semi-flexible wall overall system model.

[0071] S5. Load the corresponding *.VTK file in MWORKS software for simulation, and view the stress field calculation results in real time.

[0072] Example 1

[0073] See also Figure 1 , a real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel, specifically including the following:

[0074] S1: Based on the semi-flexible wall nozzle assembly model, a finite element simulation model is established. Specifically, S1 includes:

[0075] Import the semi-flexible wall nozzle assembly model established in the 3D software into the finite element analysis software Simcenter 3D; since the semi-flexible wall nozzle is an assembly with a complex structure and many detailed features, it is necessary to simplify the models of each component to improve the analysis efficiency. By constructing idealized components, the models that need to be involved in the analysis are promoted to idealized components, and the structural features in the model that are little affected by the force are simplified, cut, and deleted, retaining the geometry of the effective force-bearing part. Enter the simulation environment, build the finite element simulation model of each component, and then set and solve the working conditions. In the post-processing results, define the output displacement results, deformation results, stress results, and node force results.

[0076] S2: Based on the finite element simulation model, the simulation results are solved according to the maximum operating displacement range of the semi-flexible wall nozzle, and the simulation results are saved as a file with the suffix OP2; specifically, S2 includes:

[0077] S21: Material definition; in the FEM finite element environment, carry out related operations of material definition, that is, carry out related operations such as assigning model materials, creating mesh collectors, and creating mesh models; select the Create Mesh Collector command, create entity attributes, select materials in the entity attribute interface, and complete the definition of mesh collector material attributes; all parts except the flexible plate are made of 304L material, and the flexible plate is made of S03 material;

[0078] S22: Mesh division; select mesh type, mesh size and material for each part; that is, select hexahedron 10-node mesh type for all parts except the flexible plate, set the mesh size to 44.3, and select 304L as mesh collector; it should be noted that the selection of mesh collector must be consistent with the mesh collector corresponding to the material of the current model, and the material is assigned through the mesh collector; select hexahedron 10-node mesh type for the flexible plate, set the mesh size to 10, and select S03 as mesh collector;

[0079] S23: Constraint and load settings (i.e. Figure 1Set boundary conditions in ); The actuator is simulated by sliding pairs, the joint brackets are constrained by revolute pairs, and the electric cylinder base is selected as fixed constraint; in the load type, select the forced motion load type; select 1D unit as the model object, enter the displacement size in the x direction, and add gravity load to the model, and select the gravity type; the gravity acceleration is set to 9810mm / s 2 ;

[0080] S24: Solve calculation; select the finite element simulation model, perform simulation, define the name of the solution scheme, and select the solution type; that is, select the finite element simulation model, right-click, select New Simulation, define the name of the solution scheme, and select the SOL402 solution type;

[0081] S25: Export the calculation results into a file with the suffix OP2; that is, after the solution is completed, switch to the post-processing navigator interface, double-click the Solution result to view the simulation analysis results, and save the result file as a file with the suffix OP2.

[0082] S3: Based on the OP2 file, the finite element grid point coordinates and point correspondence, the correspondence between the semi-flexible wall nozzle displacement and stress, and the *.VTK file for stress cloud display are output. It should be noted that the finite element grid point coordinates and point correspondence and the correspondence between the displacement and stress will be used for training during model reduction, and the *.VTK file will be used for real-time display of stress field results through 3D visualization tools;

[0083] like Figure 1 As shown, S3 establishes a reduced-order model based on data + equations. It should be noted that the data here refers to the calculation result data of the finite element software, and the equation refers to modeling using the Modelica language. The two are combined to establish a reduced-order model. Specifically, S3 includes:

[0084] S31: Import the OP2 file into Multiviz software to perform CAE data calculation. Specifically, S31 includes:

[0085] Based on the simulation results, a simulation result sample set of semi-flexible wall displacement input was obtained based on the semi-flexible wall positive displacement finite element simulation and semi-flexible wall negative displacement finite element simulation. On this basis, semi-flexible wall monitoring point data and semi-flexible wall physical field data were generated. Among them, the mesh data volume of the semi-flexible wall nozzle is 130W, and the number of working conditions is 50, of which the number of positive displacement working conditions is 25 and the number of negative displacement working conditions is 25.

[0086] The physical quantity extraction and data merger of these 50 working conditions are performed, and then the merged data is subjected to dimensionality reduction processing to obtain the reduced-dimensional training data; the reduced-dimensional training data content is: input1, input2, input3, input4, dataModel_1, dataModel_2, …, dataModel_50, and the matrix dimension is 50×54; among them, input1, input2, input3, input4 correspond to the four inputs in the finite element simulation model.

[0087] S32: Based on the calculation results, MWORKS software is used to perform model reduction to establish a reduced-order model. Specifically, S32 includes:

[0088] Import the data set to be trained into the MWORKS software and classify the variables; select the input data as the input variable and the DetaModel data as the output variable; set the model solution parameters in the model processing interface and perform training; after the solution is completed, perform a test. When the result meets the expectations, the model can be encapsulated as a reduced-order model; if the result does not meet the expectations, it is necessary to iteratively modify the parameters to solve the model until the model meets the expectations;

[0089] After the reduced-order model training is completed, the model is encapsulated as a 1D Modelica model, which includes a multi-domain joint simulation interface, a reduced-order model, icons, etc. Similarly, in the MWORKS software, the model established using the Modelica language through a declarative-procedural modeling method is based on mathematical equations and includes a multi-domain joint simulation interface, icons, etc. Therefore, the reduced-order model can be integrated with the Modelica equation model for simulation.

[0090] The specific operations are as follows:

[0091] Use the "Model Reduction and Fusion Simulation" option in the model reduction toolbox in MWORKS.Sysplorer to reduce the model order; after opening the toolbox, create a new reduced-order model project; according to the characteristics of the step-by-step static solution of the finite element working condition, choose to establish a static model, enter a name and select the directory to save it;

[0092] After creating a new model, enter the tab page in the model reduction toolbox; it mainly includes three pages: data import, model processing, and result analysis; first import the data set to be trained, and after the data import interface is set up, classify the variables; select input1, input2, input3, input4 as input variables, and detaModel_1, dataModel_2, ..., dataModel_50, a total of 50-order data modal coefficients as output variables;

[0093] Set the model solution parameters in the model processing interface; select the model type as a multilayer feedforward neural network model, set the number of iterations to 1000, the learning rate to 0.2, the activation function to tanh, the loss function to MSELoss, the optimizer to AdamW, the number of hidden layers to 3, and the number of neurons to "10, 20, 10", and perform training;

[0094] When the solution is completed, enter the result analysis page. In the result analysis interface, select the verification set in the multi-layer feedforward neural network, click the test above, and select the working condition set and output to be compared in the new interface to compare the results; when the result meets the expectations, the model can be packaged as a field reduction model; if the result does not meet the expectations, it is necessary to iteratively modify the parameters to solve the model until the model meets the expectations;

[0095] After the training of the reduced-order model of the semi-flexible wall nozzle is completed, use the export function on the top of the model toolbox to encapsulate the model as a 1D Modelica model; fill in the name, save path and other information in the export interface; the model defaults to the reduced-order model generated by training in the current project directory;

[0096] The 1D Modelica model contains multi-domain joint simulation interfaces, reduced-order models, icons, etc., which can be integrated with the Modelica equation model established in the MWORKS software for simulation. The exported model directory structure consists of *.mo files, Resourse folders, and exte_file folders; Resourse places data modes, and exte_file is used to place dynamic libraries and py files of reduced-order models; physical field reduced-order models need to place *.VTK files, icons, etc. in Resourse, where *.VTK files are used to provide the spatial coordinates of the original geometric points, point connection relationships, and mesh topology information for 3D visualization.

[0097] S4: The data of the correspondence between the displacement and stress of the semi-flexible wall nozzle is reduced by the model order to retain the key information and features; the stress of the whole structure field of the semi-flexible wall nozzle is calculated by the displacement of the actuator. Specifically, S4 includes:

[0098] Using the reduced-order model in MWORKS software, the semi-flexible wall servo motor model and the semi-flexible wall mechanical motion model are connected, the semi-flexible wall 1D mechanism subsystem model is integrated, and then the semi-flexible wall 1D mechanism subsystem model is connected with the *.mo file in the semi-flexible wall nozzle 1D Modelica model to form a semi-flexible wall overall system model; and simulation is performed based on the semi-flexible wall overall system model;

[0099] The specific operation is as follows: after the package is completed, the reduced-order model is opened in MWORKS.Sysplorer. When the reduced-order model is integrated with the 1D model, it is only necessary to connect the input signal interface with the input and output interface of ROM_Basis by dragging and dropping;

[0100] The established semi-flexible wall servo motor model and semi-flexible wall mechanical motion model are connected, and the semi-flexible wall 1D mechanism subsystem model is integrated. Then, the semi-flexible wall 1D mechanism subsystem model is connected with the *.mo file in the semi-flexible wall nozzle 1D Modelica model to form a semi-flexible wall overall system model.

[0101] Semi-flexible wall overall system model (the integrated semi-flexible wall overall system model is established by connecting the mechanism model and the reduced-order model through an interface) Figure 2 shown.

[0102] In the "Modeling" interface of MWORKS.Sysplorer, click "Simulation Settings", set the simulation time to 200s, the simulation step to 0.1, the algorithm to "Dassl", and the accuracy to 0.0001;

[0103] S5. Load the corresponding *.VTK file in the 3D visualization tool MWORKS software for simulation, and view the stress field calculation results in real time. The specific operations are as follows:

[0104] After the simulation settings are determined, enter the "Simulation" interface of MWORKS.Sysplorer, click "3D Visualization", and you will automatically enter the 3D Visualization Toolbox interface. Select the semi-flexible wall nozzle data to be displayed as the data source, then load the corresponding "*.VTK" file, click "Simulation", and the system will be simulated after compilation. The stress field calculation results can be viewed in real time in the 3D Visualization Toolbox interface. From the 200s time taken to simulate the real physical world, under the conditions of a simulation step size of 0.1 and a simulation accuracy of 0.0001, the semi-flexible wall nozzle stress field calculation is completed, and the CPU calculation time only takes 1.242 seconds, indicating that the wind tunnel semi-flexible wall nozzle stress field real-time simulation method of the present invention can meet the real-time requirements.

[0105] The above solution is only an illustration of a preferred embodiment, but is not limited thereto. When implementing the present invention, appropriate replacement and / or modification can be performed according to user needs.

[0106] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and the embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the art, additional modifications can be easily realized. Therefore, without departing from the general concept defined by the claims and equivalent scope, the present invention is not limited to the specific details and the illustrations shown and described here.

Claims

1. A real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel, characterized in that: include: S1. A reduced-order model is obtained based on the finite element simulation model and Modelica joint simulation, and the reduced-order model is encapsulated as a 1D Modelica model. The model directory structure exported by the 1D Modelica model includes a *.mo file, a Resourse folder, and an exte_file folder; S2. In MWORKS software, the semi-flexible wall servo motor model and the semi-flexible wall mechanical motion model are connected to obtain a semi-flexible wall 1D mechanism subsystem model; S3. Open the packaged 1D Modelica model in MWORKS software, connect the semi-flexible wall 1D mechanism subsystem model with the *.mo file in the semi-flexible wall nozzle 1D Modelica model, and integrate them into a semi-flexible wall overall system model, so as to complete the simulation based on the semi-flexible wall overall system model; S4. After simulation in MWORKS software, the stress field calculation results can be viewed in real time.

2. The real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel according to claim 1, characterized in that: In S1, the process of obtaining the reduced-order model includes: S10. Establish a finite element simulation model based on the semi-flexible wall nozzle assembly model; S11. Based on the finite element simulation model, the simulation results are solved according to the maximum operating displacement range of the semi-flexible wall nozzle, and the simulation results are saved as a file with a suffix of OP2; S12. Based on the OP2 file, the Multiviz software and MWORKS software are used to perform parsing and training to obtain a trained reduced-order model.

3. The real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel according to claim 2, characterized in that: In S10, the process of establishing the finite element simulation model includes: S101. Import the semi-flexible wall nozzle assembly model established in the 3D software into the finite element analysis software Simcenter3D; S102, simplify the structural features that do not need to be involved in the analysis, and only retain the geometry of the effective force-bearing part; S103, entering the simulation environment to construct, set working conditions, and solve finite element simulation models of various components; S104. In the post-processing results, define the output displacement results, deformation results, stress results, and node force results to obtain the corresponding finite element simulation model.

4. The real-time simulation method for stress field of a wind tunnel semi-flexible wall nozzle according to claim 2, characterized in that: In S11, the process of solving the simulation results according to the maximum operating displacement range of the semi-flexible wall nozzle is: S111. Carry out material definition related operations in FEM finite element environment; S112, selecting a mesh type, setting a mesh size, and selecting a material for each component; S113, constraining and setting loads for each component; S114, select a finite element simulation model, perform simulation, define a solution name, and select a solution type; The contents of constraint and load settings include: The actuator mechanism of the actuator cylinder is simulated by a sliding pair, the joint brackets are constrained by a rotating pair, and the constraint type of the electric cylinder base is a fixed constraint; In Load Type, select the Enforced Motion load type; Select 1D element as the model object, enter the displacement in the x-direction, and add a gravity load to the model, selecting the gravity type.

5. The real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel according to claim 2, characterized in that: In S12, the process of training the reduced-order model is: S121, import the OP2 file into Multiviz software to perform CAE data calculation; S122. Based on the calculation results, MWORKS software is used to perform order reduction training to generate a reduced-order model.

6. The real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel according to claim 5, characterized in that: The CAE data is calculated as follows: Based on the simulation results of S11, a simulation result sample set of semi-flexible wall displacement input is obtained based on the semi-flexible wall positive displacement finite element simulation and semi-flexible wall negative displacement finite element simulation, and on this basis, semi-flexible wall monitoring point data and semi-flexible wall physical field data are generated; The physical quantity of the data in the working condition is extracted and the data is merged, and then the merged data is processed by MWORKS software to reduce the dimension and obtain the training data set after the dimension reduction; The content of the training data set includes: input data and DetaModel data.

7. The real-time simulation method for stress field of a wind tunnel semi-flexible wall nozzle according to claim 5, characterized in that: In S122, the method of scale reduction training is: Import the training data set into MWORKS software and classify the variables; Select input data as input variable and DetaModel data as output variable; Set the model solution parameters in the model processing interface and perform training. After the solution is completed, test it. When the result meets expectations, the model can be encapsulated as a reduced-order model. If the result does not meet expectations, it is necessary to iteratively modify the parameters to solve the model until the model meets expectations. After the reduced-order model training is completed, the model is packaged as a 1D Modelica model and exported.

8. The real-time simulation method for stress field of a semi-flexible wall nozzle in a wind tunnel according to claim 1, characterized in that: Also includes: Load the corresponding *.VTK file in MWORKS software for simulation and view the stress field calculation results in real time.

9. The real-time simulation method for stress field of a wind tunnel semi-flexible wall nozzle according to claim 1, characterized in that: 1DModelica model includes multi-domain joint simulation interface, reduced-order model, and icons; The data modality is placed in the Resourse folder, and the dynamic library and py file of the reduced-order model are placed in the exte_file folder; The physical field reduction model places *.VTK files and icons in Resourse, where the *.VTK files are used to provide the spatial coordinates of the original geometric points, point connection relationships, and mesh topology information to achieve three-dimensional visualization.