Icing wind tunnel test section flow field multidimension simulation method, device, equipment and medium

Abstract

The application discloses an icing wind tunnel test section flow field multidimensional simulation method, device, equipment and medium, the method comprises the steps that S1, a geometric model is established according to the actual structure of test section, S2, grid division is carried out to the geometric model;S3, control equation is determined and physical model is selected, S4, the boundary condition of test section calculation domain is defined;S5, using the performance analysis or test data of aeroengine, constructs aeroengine performance proxy model and CFD solver coupling;S6, data is transmitted between CFD solver and proxy model in the way of time step iteration, and coupling calculation is carried out to realize multidimensional simulation.The application introduces aeroengine performance proxy model into icing wind tunnel flow field simulation, develops a kind of high-precision, high-efficiency numerical simulation method, and then realizes the prediction and analysis to aeroengine icing weather condition test simulation result.

Description

Multi-dimensional simulation methods, devices, equipment and media for flow field in icing wind tunnel test section Technical Field

[0001] This application relates to the field of icing wind tunnel testing technology, and in particular to a method, apparatus, equipment and medium for multi-dimensional simulation of flow field in icing wind tunnel test sections. Background Technology

[0002] Existing methods for analyzing the flow field in wind tunnel test sections for aero-engine icing include experimental measurement and numerical simulation. These methods have the following shortcomings: First, experimental measurement methods are limited and costly. Wind tunnel tests are constrained by cabin size and measurement capabilities, resulting in limited experimental data that cannot fully reflect the characteristics of the three-dimensional gas-liquid two-phase flow field. Second, while numerical simulation can provide more flow field information, current methods often combine three independent modules—airflow, droplet motion, and icing state—lacking bidirectional coupling with the aero-engine performance model, making it difficult to balance computational accuracy and efficiency.

[0003] In summary, existing experimental measurement methods and numerical simulation methods each have their shortcomings and cannot meet the needs for rapid, accurate, and comprehensive evaluation of the flow field in the wind tunnel test section of an aero-engine icing system. Summary of the Invention

[0004] This application provides a multi-dimensional simulation method for the flow field of an icing wind tunnel test section, enabling rapid, accurate, and comprehensive evaluation of the flow field in an aero-engine icing wind tunnel test section.

[0005] This application is achieved through the following solution:

[0006] Multi-dimensional simulation of the flow field in the icing wind tunnel test section, including the following steps:

[0007] S1. Based on the actual structure of the test section, a three-dimensional geometric model was established using three-dimensional modeling software, including the rear compartment of the test chamber, corner section, engine intake duct, free jet nozzle, test bench, exhaust diffuser, aero-engine and drive shaft.

[0008] S2. After completing the geometric modeling of the test section, the computational domain of the test section is meshed. During the meshing process, boundary markers are introduced for subsequent physical model settings and surrogate model coupling.

[0009] S3. Determine the governing equations and select the physical model, including the gas phase flow governing equations, droplet dynamics model, droplet transformation strategy, and interphase coupling between droplets and gas flow.

[0010] S4. Define the boundary conditions of the test section computational domain, including the test section inlet boundary conditions, the test section outlet boundary conditions, the boundary conditions at the spray device outlet, the wall boundary conditions, and the aerodynamic interface between the aero-engine and the rear compartment of the test cabin.

[0011] S5. Using aero-engine performance calculation or test data, establish an aero-engine performance proxy model through machine learning, encapsulate the trained proxy model into a general executable file, and couple it with the CFD solver through code.

[0012] S6. By adopting a time-step iterative approach to transfer information between the CFD solver and the surrogate model, multi-dimensional coupled calculations are performed to complete all time-step calculations and realize multi-dimensional simulation of the flow field in the icing wind tunnel test section.

[0013] Furthermore, in step S1, when establishing the geometric model of the aero-engine, a simplified outline model is used to reduce the amount of computation, and an interface for interaction with the proxy model is reserved.

[0014] Furthermore, in step S2, during the mesh generation process, a hybrid mesh type is selected. For key areas such as nozzle outlet, engine inlet and outlet, and exhaust diffuser inlet, locally densified or unstructured meshes are used to capture the flow details of the airflow field and cloud field. The mesh quality is guaranteed by checking the indicators of cell distortion, orthogonality, and volume distribution.

[0015] Furthermore, in step S3, the gas phase flow control equations adopt the incompressible or weakly compressible Navier-Stokes equations as the control equations for the gas phase, including the mass conservation equation, momentum equation, and energy equation, and an appropriate solver is selected according to the incoming Mach number; the turbulence simulation adopts the Reynolds-averaged flow (RANS) method or the large eddy simulation (LES) method; the energy equations consider heat transfer and latent heat exchange during phase change;

[0016] When setting up the droplet dynamics model, a Lagrangian description is used for the droplets. The changes in mass, position, velocity, and temperature of each droplet or droplet group over time are tracked. The equation for calculating droplet acceleration is:

[0017] ;

[0018] Where, m p v is the mass of the droplet. p For droplet velocity, For droplet acceleration, The drag force acting on the droplet, For lift, For gravity, The pressure gradient force is the drag force, which is the product of the drag coefficient and the velocity difference. The drag coefficient can be determined using the Schiller-Naumann model or the Oseen model based on the Reynolds number and droplet shape. Collision aggregation uses the O'Rourke model, and breakup uses the Kelvin-Helmholtz and Rayleigh-Taylor instability criteria. Wall collision is determined by the elastic coefficient and the critical Weber number to determine whether it is rebound, wetting, or breakup.

[0019] The droplet conversion strategy includes: when the droplet diameter is less than a threshold or the local droplet volume fraction is lower than a threshold, the Lagrange particle model is continued; when the local volume fraction is higher than the threshold, in order to avoid excessive single-particle tracking leading to a surge in computation, the droplet swarm is regarded as a continuous medium, and the swarm behavior is described by Smoothed Particle Hydrodynamics (SPH) or Discrete Element Method (DPM). The swarm model describes the momentum and energy exchange between the droplet swarm and the gas phase through the control volume fraction equation and the swarm dynamics equation.

[0020] The interphase coupling between droplets and gas flow is specifically as follows: the reaction of droplet motion on the gas phase is added to the gas phase control equation through source terms, including momentum source terms and energy source terms, to express the droplet's drag and heat transfer effects on the gas flow; the CFD solver achieves two-phase coupling by cyclically updating the interphase source terms and droplet trajectory.

[0021] Furthermore, in step S4, the inlet boundary conditions of the test section include the incoming flow velocity, pressure, temperature, relative humidity, and turbulence intensity;

[0022] The test section outlet boundary conditions include set pressure outlet or full pressure outlet boundary conditions.

[0023] The boundary conditions at the outlet of the spray device are set using an atomizing nozzle model. The droplet size distribution and initial velocity are calculated based on the nozzle structure and the water and air supply pressures. The droplet size is usually represented by a log-normal distribution or a Rosner-Marcano distribution. The liquid water content (LWC) and median volume diameter (MVD) are also set.

[0024] The wall boundary conditions include setting the outer surface of the drive shaft as a rotating wall and setting other walls in the test section as non-slip solid walls, taking into account the influence of wall temperature, roughness and ice layer growth on surface friction;

[0025] The aerodynamic interface between the aero-engine and the rear compartment of the test module includes the engine inlet section and the outlet section. The aerodynamic interface parameters are dynamically updated by constructing an aero-engine performance proxy model and coupling it with the test section CFD model.

[0026] Furthermore, step S5 specifically includes the following steps:

[0027] S51. Experimental Design and Data Sampling: Based on the space filling strategy, sample points are selected within the entire working envelope of the engine. The sample input variables include flight altitude, flight Mach number, total intake temperature, speed or fuel flow rate, and the output variables include engine inlet and outlet flow rates, temperature and pressure. The collected samples are from aero-engine performance calculations or experimental data.

[0028] S52. Feature processing and model selection: Normalize the input and output data and select a suitable surrogate model structure, including multinomial response surface surrogate model, radial basis function surrogate model, kriging surrogate model or deep neural network surrogate model.

[0029] S53. Model Training and Validation: Train the surrogate model using cross-validation and validate it with data not used in the training; evaluate model performance based on prediction error and correlation coefficient, and optimize the model by increasing the sample size or adjusting the model structure.

[0030] S54. Proxy Model Embedding and Coupling: The trained proxy model is encapsulated into a general executable file and coupled with the CFD solver through code. During the model coupling calculation process, the data exchange rule between models is: data exchange is performed before the start of each iteration step.

[0031] Furthermore, step S6 specifically includes the following steps:

[0032] S61. Initialization Phase: Import geometry and mesh, set initial temperature field, pressure field, velocity field and cloud field; initialize engine operating conditions and call the proxy model to calculate the engine inlet and outlet flow rates, temperature and pressure at set speed or fuel flow rate;

[0033] S62, CFD solution stage: Under given boundary conditions, solve the gas phase and droplet motion equations, calculate the instantaneous distribution of the airflow field and cloud field, update the interphase coupling source terms, use pressure-based coupling algorithms (SIMPLE, PISO, etc.) to solve the flow field, and use a set time step to ensure numerical stability.

[0034] S63, Proxy Model Calculation Stage: Extract the total temperature and pressure of the engine inlet section and the static pressure of the exhaust environment, and input them into the proxy model to calculate the inlet flow rate, outlet flow rate, temperature and pressure of the engine at a set speed or fuel flow rate; the output of the proxy model is used to update the boundary conditions of the CFD calculation of the test section.

[0035] S64. Iteration and Convergence Judgment: Compare the CFD calculation results and the surrogate model output in the current time step to see if they meet the convergence condition. If they do not meet the condition, return the surrogate model output to the CFD solver and continue iterating. If they meet the condition, proceed to the next time step.

[0036] S65, Ice Layer Growth: If it is necessary to simulate the growth of ice layer on the wall, the ice layer thickness is calculated based on the wall water collection rate, heat transfer rate, freezing rate, etc., and the wall shape is updated at each time period; the mesh is adjusted accordingly by embedding a mesh dynamic deformation or local reconstruction module, and the ice layer change is ignored by using a fixed geometry during the iteration process or updated once at a set time scale;

[0037] S66. Post-processing and data output: After completing the calculation of all time steps, perform statistical and visual analysis on the flow field data, and output velocity vector map, pressure cloud map, temperature cloud map, droplet concentration distribution, particle size distribution, total pressure / total temperature non-uniformity, liquid water content distribution, etc.

[0038] This application also provides a multi-dimensional simulation system for the flow field of an icing wind tunnel test section, including:

[0039] The geometric model building module is used to create a three-dimensional geometric model based on the actual structure of the test section using three-dimensional modeling software. The model includes the rear compartment of the test chamber, corner section, engine air intake duct, free jet nozzle, test bench, exhaust diffuser, aero-engine and drive shaft.

[0040] The mesh generation and preprocessing module is used to mesh the computational domain of the test section after completing the geometric modeling of the test section. During the mesh generation process, boundary markers are introduced for subsequent physical model settings and surrogate model coupling.

[0041] The module for defining governing equations and selecting physical models is used to determine governing equations and select physical models, including gas phase flow governing equations, droplet dynamics models, droplet transformation strategies, and interphase coupling between droplets and gas flow.

[0042] The boundary condition setting module is used to define the boundary conditions of the test section computational domain, including the test section inlet boundary conditions, the test section outlet boundary conditions, the boundary conditions at the spray device outlet, the wall boundary conditions, and the aerodynamic interface between the aero-engine and the rear compartment of the test cabin.

[0043] The module for constructing and coupling the performance proxy model of aero-engines is used to build a performance proxy model of aero-engines using aero-engine performance calculation or test data and machine learning methods. The trained proxy model is encapsulated into a general executable file and coupled with the CFD solver through code.

[0044] The multi-dimensional coupled calculation module is used to transfer information between the CFD solver and the surrogate model in a time-step iterative manner, perform multi-dimensional coupled calculations, complete all time-step calculations, and realize multi-dimensional simulation of the flow field in the icing wind tunnel test section.

[0045] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement the steps of the multi-dimensional simulation method for the flow field of the icing wind tunnel test section.

[0046] This application also provides a storage medium that includes a stored program that, when the program is executed, controls the device containing the storage medium to perform the steps of the multi-dimensional simulation method for the flow field of the icing wind tunnel test section.

[0047] Compared with the prior art, this application has the following advantages:

[0048] This application proposes a multi-dimensional simulation method, apparatus, equipment, and medium for the flow field of an icing wind tunnel test section. The proposed multi-dimensional simulation method addresses the problem that existing experimental testing methods and numerical simulation methods cannot quickly, accurately, and comprehensively assess whether the flow field of the test section meets the requirements of aero-engine icing tests. On one hand, based on the Eulerian-Lagrange framework and the droplet transformation sub-model, a three-dimensional gas-liquid two-phase flow model coupled with an aero-engine performance proxy model is constructed, enabling predictive analysis of the simulation results of engine icing meteorological conditions. On the other hand, this method compensates for the limited experimental test data, providing multi-dimensional data such as velocity, pressure, temperature, liquid water content, and droplet size distribution of the test section, providing reliable support for the accurate evaluation of the flow field in the test section.

[0049] On the other hand, this application addresses the core challenges of insufficient computational accuracy due to fixed boundary conditions and low computational efficiency in full 3D simulation. Current numerical simulations suffer from two drawbacks: first, they ignore the coupling effect between the aero-engine operating conditions and the flow field in the test section, setting the engine inlet and outlet as fixed boundary conditions, leading to distorted simulation results; second, adopting full 3D simulation of the test section and engine introduces significant computational difficulty and workload, making it unsuitable for engineering applications requiring rapid iteration. This application proposes a coupled computational strategy of an aero-engine performance proxy model and test section CFD. Using the aero-engine performance proxy model, before the start of each iteration step, the total temperature and pressure of the engine inlet section and the static pressure of the exhaust environment are read from the test section CFD calculation results. This predicts the engine's inlet flow rate, outlet flow rate, temperature, and pressure at a set speed or fuel flow rate, and updates the test section CFD calculation boundary conditions in real time. This strategy effectively avoids the resource bottleneck of full 3D simulation and the model errors introduced by simplified boundaries while balancing computational efficiency, achieving synergistic optimization of computational accuracy and efficiency. The iterative solution of the coupled model can accurately map the complex coupling relationship between the flow and engine performance in the test section. Its high-confidence simulation results provide a reliable basis for optimizing the aerodynamic layout of the test section and for precise control of the environmental parameters of the engine icing test.

[0050] In addition to the purposes, features, and advantages described above, this application has other purposes, features, and advantages. A further detailed description of this application will be provided below with reference to the figures. Attached Figure Description

[0051] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0052] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort, wherein:

[0053] Figure 1 is a flowchart illustrating the multi-dimensional simulation method of the flow field in the icing wind tunnel test section according to a preferred embodiment of this application;

[0054] Figure 2 is a schematic diagram of the flow field calculation area in the ice wind tunnel test section;

[0055] Figure 3 is a schematic diagram of the geometric model of the test section;

[0056] Figure 4 is a schematic diagram of a tetrahedral unstructured mesh;

[0057] Figure 5 is a schematic diagram of data transfer between CFD calculation, changes in the hood cross-section conditions, and the aero-engine performance proxy model.

[0058] Figure 6 is a schematic diagram of the construction process of the coupling algorithm;

[0059] Figure 7 is a schematic diagram of serial coupling and parallel coupling schemes;

[0060] Figure 8 is a schematic diagram of the multi-dimensional simulation system module for the flow field of the icing wind tunnel test section according to a preferred embodiment of this application;

[0061] Figure 9 is a schematic block diagram of an electronic device according to a preferred embodiment of this application;

[0062] Figure 10 is an internal structural diagram of a computer device according to a preferred embodiment of this application.

[0063] In the diagram: 1. Drive shaft; 2. Spray device; 3. Rear compartment of the test chamber; 4. Aero engine; 5. Exhaust diffuser; 6. Test bench; 7. Free jet nozzle; 8. Engine intake duct; 9. Corner section. Detailed Implementation

[0064] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.

[0065] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.

[0066] Terminology Explanation:

[0067] Icing wind tunnel: An icing wind tunnel is a test facility specifically designed to simulate the operation of aircraft and engines in high-altitude icing environments. It generates low-temperature, high-humidity airflow within a sealed chamber, and injects liquid water droplets into the airflow through atomizing nozzles to create a cloud-like field. During testing, parameters such as incoming flow velocity, temperature, pressure, liquid water content, and median droplet diameter are typically controlled to approximate real atmospheric conditions.

[0068] Non-uniformity: refers to the degree of uniformity of the distribution of a physical quantity on a cross section, commonly expressed by the formula δ=(x max -x min ) / x avg This indicates that one of the design goals of the icing wind tunnel test section is to ensure that the non-uniformity of the flow field at the engine inlet and outlet is below a certain threshold, in order to simulate real flight conditions and ensure test accuracy.

[0069] Proxy models are approximate models trained using machine learning, response surface methodology, and other methods based on high-fidelity simulation results or experimental data. Proxy models can quickly predict the output of complex systems with extremely low computational cost. For example, aero-engine performance proxy models predict engine inlet and outlet flow rates, temperatures, and pressures based on inputs such as total intake pressure, total intake temperature, exhaust static pressure, and engine speed. This can replace complex engine performance calculations, saving simulation time.

[0070] CFD, or Computational Fluid Dynamics, is a computer-aided engineering technique that uses numerical methods and computer simulations to analyze and predict systems involving fluid flow, heat conduction, and related physical phenomena. Its basic principle lies in discretizing the continuous fluid motion control equations (such as the Navier-Stokes equations) in the spatial and temporal domains, thereby constructing corresponding numerical models. By iteratively solving these discretized algebraic equations, the quantitative distribution and evolution of key physical parameters such as velocity, pressure, and temperature of the flow field under specific boundary and initial conditions can be obtained.

[0071] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or a multi-dimensional simulation device for the flow field of an icing wind tunnel test section capable of performing the above functions. The following description uses a multi-dimensional simulation device for the flow field of an icing wind tunnel test section as an example to illustrate this embodiment and the subsequent embodiments.

[0072] As shown in Figure 1, a preferred embodiment of this application provides a multi-dimensional simulation method for the flow field of an icing wind tunnel test section, including the following steps:

[0073] S1. Based on the actual structure of the test section (see Figure 2), a geometric model of the test section is established using 3D modeling software (see Figure 3) to ensure the smoothness of the geometric components and the rationality of the structure. The actual structure of the test section includes the drive shaft 1, spray device 2, rear compartment of the test chamber 3, aero engine 4, exhaust diffuser 5, test bench 6, free jet nozzle 7, engine intake duct 8, and corner section 9.

[0074] S2. After completing the geometric modeling of the test section, the computational domain of the test section is meshed. During the meshing process, boundary markers are introduced for subsequent physical model settings and proxy model coupling. Figure 4 shows the tetrahedral unstructured mesh generated using meshing software.

[0075] S3. Determine the governing equations and select the physical model, including the gas phase flow governing equations, droplet dynamics model, droplet transformation strategy, and interphase coupling between droplets and gas flow.

[0076] S4. Define the boundary conditions of the test section computational domain, including the test section inlet boundary conditions, the test section outlet boundary conditions, the boundary conditions at the spray device outlet, the wall boundary conditions, and the aerodynamic interface between the aero-engine and the rear compartment of the test cabin.

[0077] S5. Using aero-engine performance calculation or test data, establish an aero-engine performance proxy model through machine learning, encapsulate the trained proxy model into a general executable file, and couple it with the CFD solver through code.

[0078] S6. By adopting a time-step iterative approach to transfer information between the CFD solver and the surrogate model, multi-dimensional coupled calculations are performed to complete all time-step calculations and realize multi-dimensional simulation of the flow field in the icing wind tunnel test section.

[0079] This embodiment proposes a multi-dimensional simulation method for the flow field of an icing wind tunnel test section. This method solves the problem that existing experimental testing methods and numerical simulation methods cannot quickly, accurately, and comprehensively assess whether the flow field of the test section meets the requirements of aero-engine icing tests. On the one hand, based on the Eulerian-Lagrange framework and the droplet transformation sub-model, a three-dimensional gas-liquid two-phase flow model coupled with an aero-engine performance proxy model is constructed, realizing the predictive analysis of the simulation results of engine icing meteorological conditions. On the other hand, this method compensates for the limitation of experimental testing data, providing multi-dimensional data such as velocity, pressure, temperature, liquid water content, and droplet size distribution of the test section, providing reliable support for the accurate evaluation of the flow field of the test section.

[0080] On the other hand, this embodiment addresses the core challenges of insufficient computational accuracy due to fixed boundary conditions and low computational efficiency in full 3D simulation. Current numerical simulations suffer from two drawbacks: first, they ignore the coupling effect between the aero-engine operating conditions and the flow field in the test section, setting the engine inlet and outlet as fixed boundary conditions, leading to distorted simulation results; second, adopting full 3D simulation of both the test section and the engine introduces significant computational difficulty and workload, making it unsuitable for engineering applications requiring rapid iteration. This embodiment proposes a coupled computational strategy of an aero-engine performance proxy model and test section CFD. Using the aero-engine performance proxy model, before the start of each iteration step, the total temperature and pressure of the engine inlet section and the static pressure of the exhaust environment are read from the test section CFD calculation results. This predicts the engine's inlet flow rate, outlet flow rate, temperature, and pressure at a set speed or fuel flow rate, and updates the test section CFD calculation boundary conditions in real time. This strategy effectively avoids the resource bottleneck of full 3D simulation and the model errors introduced by simplified boundaries while maintaining computational efficiency, achieving synergistic optimization of computational accuracy and efficiency. The iterative solution of the coupled model can accurately map the complex coupling relationship between the flow and engine performance in the test section. Its high-confidence simulation results provide a reliable basis for optimizing the aerodynamic layout of the test section and for precise control of the environmental parameters of the engine icing test.

[0081] Preferably, in step S1, when establishing the geometric model of the aero-engine, a simplified outline model is used to reduce the amount of computation, and an interface for interaction with the proxy model is reserved.

[0082] Preferably, in step S2, during the mesh generation process, a hybrid mesh type is selected. Specifically, for key areas such as nozzle outlet, engine inlet and outlet, and exhaust diffuser inlet, locally densified or unstructured meshes are used to capture the flow details of the airflow field and cloud field. The mesh quality is ensured by checking the indicators of cell distortion, orthogonality, and volume distribution.

[0083] Preferably, in step S3, the gas phase flow control equations adopt the incompressible or weakly compressible Navier-Stokes equations as the control equations for the gas phase, including the mass conservation equation, momentum equation, and energy equation, and an appropriate solver is selected according to the incoming Mach number; the turbulence simulation adopts the Reynolds-averaged flow (RANS) method or the large eddy simulation (LES) method; the energy equations consider heat transfer and latent heat exchange during phase change;

[0084] When setting up the droplet dynamics model, a Lagrangian description is used for the droplets. The changes in mass, position, velocity, and temperature of each droplet or droplet group over time are tracked. The equation for calculating droplet acceleration is:

[0085] ;

[0086] Where, mp v is the mass of the droplet. p For droplet velocity, For droplet acceleration, The drag force acting on the droplet, For lift, For gravity, The pressure gradient force is the drag force, which is the product of the drag coefficient and the velocity difference. The drag coefficient can be determined using the Schiller-Naumann model or the Oseen model based on the Reynolds number and droplet shape. Collision aggregation uses the O'Rourke model, and breakup uses the Kelvin-Helmholtz and Rayleigh-Taylor instability criteria. Wall collision is determined by the elastic coefficient and the critical Weber number to determine whether it is rebound, wetting, or breakup.

[0087] The droplet transformation strategy includes: when the droplet diameter is less than a threshold or the local droplet volume fraction is below a threshold, the Lagrange particle model continues to be used; when the local volume fraction is higher than the threshold, to avoid excessive single-particle tracking leading to a surge in computational load, the droplet swarm is treated as a continuous medium, and the swarm behavior is described using Smoothed Particle Hydrodynamics (SPH) or Discrete Element Method (DPM). The swarm model describes the momentum and energy exchange between the droplet swarm and the gas phase through the control volume fraction equation and the swarm dynamics equation.

[0088] The interphase coupling between droplets and gas flow is specifically as follows: the reaction of droplet motion on the gas phase is added to the gas phase control equation through source terms, including momentum source terms and energy source terms, to express the droplet's drag and heat transfer effects on the gas flow; the CFD solver achieves two-phase coupling by cyclically updating the interphase source terms and droplet trajectory.

[0089] Preferably, in step S4, the boundary conditions of the test section inlet (section A) include the incoming flow velocity, pressure, temperature, relative humidity, and turbulence intensity.

[0090] The boundary conditions at the test section outlet (sections E and F) include set pressure outlet or full pressure outlet boundary conditions.

[0091] The boundary conditions at the outlet of the spray device (section B) are set using an atomizing nozzle model. The droplet size distribution and initial velocity are calculated based on the nozzle structure and the water and air supply pressures. The droplet size is usually represented by a log-normal distribution or a Rosner-Marcano distribution. The liquid water content (LWC) and median volume diameter (MVD) are also set.

[0092] The wall boundary conditions include setting the outer surface of the drive shaft as a rotating wall and setting other walls in the test section as non-slip solid walls, taking into account the influence of wall temperature, roughness and ice layer growth on surface friction;

[0093] The aero-engine and the rear compartment of the test module include the engine inlet section (C section) and the outlet section (D section). The aero-engine performance proxy model and the test section CFD model are coupled for calculation to realize the dynamic update of the aero-engine interface parameters.

[0094] Preferably, the operating state of the aero-engine has a significant impact on the flow field of the test section, but directly conducting a full three-dimensional coupled simulation of the test section and the engine would bring extremely high computational difficulty and workload. Therefore, this embodiment adopts a surrogate model method, using aero-engine performance calculation or test data to establish an aero-engine performance surrogate model through machine learning. As shown in Figure 5, step S5 specifically includes the following steps:

[0095] S51. Experimental Design and Data Sampling: Based on the space filling strategy, sample points are selected within the entire working envelope of the engine. The sample input variables include flight altitude, flight Mach number, total intake temperature, speed or fuel flow rate, and the output variables include engine inlet and outlet flow rates, temperature and pressure. The collected samples are from aero-engine performance calculations or experimental data.

[0096] S52. Feature Processing and Model Selection: Normalize the input and output data, and select a suitable surrogate model structure, including multinomial response surface surrogate models, radial basis function (RBF) surrogate models, Kriging surrogate models, or deep neural network surrogate models; for highly nonlinear relationships, hybrid surrogate models or ensemble learning methods can be used to improve prediction accuracy; among which:

[0097] The Polynomial Response Surface Model (PRSM) approximates the mapping between input variables and system response by constructing a low-order polynomial function. A typical form is a second-order polynomial containing linear, square, and cross terms. This model offers significant advantages in computational efficiency and strong parameter interpretation, making it particularly suitable for rapid analysis of low-dimensional linear or weakly nonlinear problems, such as parameter sensitivity studies in the early stages of engineering design. However, its limitations lie in its insufficient fitting ability for high-dimensional complex nonlinear systems and its susceptibility to overfitting due to higher-order terms, leading to increased extrapolation prediction errors. Therefore, PRSM is mostly used in optimization scenarios with dimensions less than 10 or as an initialization tool for other high-precision models. The commonly used form is a second-order polynomial.

[0098] ;

[0099] In the formula, , , , y is the polynomial coefficient, x is the dependent variable, n is the polynomial order, and i and j range from 1 to n.

[0100] The Radial Basis Function (RBF) surrogate model is based on radially symmetric functions of spatial distance. Its strong nonlinear fitting ability and adaptability to high-dimensional data make it outstanding in reconstructing the response surface of complex systems, especially when data distribution is irregular or there are local abrupt changes. RBF can flexibly adapt by adjusting the basis function centers and shape parameters. However, the model is sensitive to noisy data and is prone to overfitting, leading to a decline in generalization performance. Furthermore, the basis function center selection strategy (such as random sampling or cluster optimization) directly affects prediction accuracy. This model is widely used in multidisciplinary optimization, real-time control, and other fields, and is particularly valuable in scenarios requiring fast interpolation and high dimensionality (such as robot path planning and fluid dynamics parameter inversion). Interpolation models are constructed based on radially symmetric functions (such as Gaussian functions and multiple quadratic functions).

[0101] ;

[0102] in, For radial basis functions, Here, y(x) represents the weighting coefficients, y(x) represents the objective function, x represents the independent variable, N represents the interpolation order, and i ranges from 1 to n. i This is the interpolation point.

[0103] The Kriging model integrates a global trend function with a local Gaussian stochastic process. Its core advantage lies in its ability to quantify prediction uncertainty (e.g., 95% confidence intervals) and capture spatial correlations through covariance functions (e.g., Matern kernels), thereby accurately characterizing the response features of nonlinear and non-stationary systems. The Kriging model maintains high accuracy even with small sample sizes, making it particularly suitable for replacing expensive simulation models (e.g., CFD) and for high-fidelity optimization design. However, its computational complexity increases cubically with sample size (O(N)). 3 Furthermore, the optimization of the covariance function hyperparameters has a significant impact on model performance, and often requires the combination of maximum likelihood estimation or Bayesian methods for parameter tuning.

[0104] ;

[0105] in, Let y(x) be a zero-mean Gaussian process, y(x) be the objective function, and f(x) be the global trend function.

[0106] S53. Model Training and Validation: Train the surrogate model using cross-validation and validate it with data not used in the training; evaluate model performance based on prediction error and correlation coefficient, and optimize the model by increasing the sample size or adjusting the model structure.

[0107] S54. Proxy Model Embedding and Coupling: The trained proxy model is encapsulated into a general executable file and coupled with the CFD solver through code. During the model coupling computation process, the data exchange rule between models is: data exchange occurs before the start of each iteration step. The model coupling computation process consists of three main stages:

[0108] 1. The program code initializes its data and initializes the coupled modules.

[0109] 2. The coupling module establishes connections between code relationships and neighborhood searches. During iteration, each piece of code calculates its own part of the problem and exchanges data at specific times.

[0110] 3. End the computation by disconnecting the code and stopping all code and coupled modules.

[0111] During the iteration process, the number of iterations a transient problem involves corresponds to the number of data exchanges, and the coupling module cannot control the simulation process of each coupled program. Therefore, when one side of the coupled program sends data, the other side should be prepared to receive it.

[0112] For transient solutions in co-simulation, a bidirectional coupling method is adopted, where different simulation models both send and receive data, and each model performs data exchange before the start of each iteration step, i.e., pre-iteration exchange. Figure 6 illustrates the construction process of the solution exchange between the two coupled codes. ①~⑨ represent the sequence numbers of each data processing step, where code A exchanges data before solving, and code B exchanges data after solving. For code A, the box is located at the end of the time step, while the box for code B is always located at the beginning of the time step. The first box in each code represents the initial transmission. From left to right, the first send operation x is found in code A. A transmission line is started here, connected to the r end of code B. Then, the first receive operation is found in code B, which is also the first box r. This generates the first transmission of the code. Now, the next send operation continues, starting at the second box in code B, and then connecting it to the first box in code A, and so on. In this way, the final algorithm described in Figure 6 is obtained, which can be identified as a sequential coupling algorithm. However, note that not all combinations of initial transmissions will produce a reasonable coupling algorithm.

[0113] In most coupled simulations, data transfer is bidirectional, and the type of coupling algorithm depends on the order of solving and transfer, as well as the chosen initial data transfer method. The accuracy and convergence speed of the solution computation are also highly dependent on the choice of code coupling method. Swapping one piece of code after iteration and another before iteration generates a wide variety of algorithms. However, only a small number of these codes are actually recommended for use. To add more algorithms, a virtual "idle program" can be added after the code swap, where the two pieces of code are not solved synchronously. One piece of code is always one step ahead of the other in time. Generally speaking, serial code coupling is slower but converges more easily; while parallel code coupling is faster but presents significant challenges to convergence. For unidirectional and bidirectional transfer, as the names suggest, in some coupled applications, only unidirectional transmission is used, where one piece of code only sends data to another, and the other only receives data. This reduces the possible types of coupling algorithms; in fact, there is only one basic coupling algorithm. The following describes all possible code coupling methods based on parameter-based pre-solution transfer:

[0114] First, let's take the first coupled code segment in Figure 7 as an example, where "serial coupling" indicates serial coupling. In this code, data is first sent from code B to code A, which is process ① in the figure; then code A begins iterative calculation to obtain the calculation result, corresponding to process ②; the calculation result is sent to code B through process ③, and B then undergoes iterative calculation through process ④ to obtain the calculation result; through the transfer process ⑤, the data is passed back to code A, and this completes one loop. This loop continues until the required calculation time is reached. For the "parallel coupling" process, let's take the first parallel coupled code segment as an example. In this code, data is sent from code A to code B, and code B also sends data to code A simultaneously. Then, both code segments begin their respective solution processes. At the next time point, data transfer and exchange begin simultaneously again, continuing until the calculation ends.

[0115] During the data transmission and exchange between the 3D CFD calculation program and the engine proxy model calculation program, convergence control and parameter range limitations are required. To avoid unrealistic values ​​during data transmission, this embodiment employs the Relaxion Factor convergence control method. The Relaxion Factor method is used to under-relax or over-relax the quantity values ​​sent to the receiver:

[0116] ;

[0117] In the formula, This is the new value used to start the next iteration after being controlled by the relaxation factor; 'a' is the relaxation factor. The variable values ​​newly calculated by the solver in the current iteration step. This is the value of the variable from the previous iteration.

[0118] The relaxation factor, or relaxation control factor, is a commonly used convergence control method in multidimensional co-simulation. It can control the reasonable and accurate transfer and exchange of physical quantity data at the coupling boundary. Under-relaxation (less than 1) may lead to more stable solutions for some problems, such as when there are many numerical spikes in the data transfer and exchange process during coupling. However, for some steady-state coupling situations, the over-relaxation factor method may be more suitable, as it can improve the convergence of some slow-converging problems.

[0119] Preferably, the multi-dimensional simulation in this embodiment uses a time-step iterative approach to transfer information between the CFD solver and the surrogate model. Therefore, step S6 specifically includes the following steps:

[0120] S61. Initialization Phase: Import geometry and mesh, set initial temperature field, pressure field, velocity field and cloud field; initialize engine operating conditions and call the proxy model to calculate the engine inlet and outlet flow rates, temperature and pressure at set speed or fuel flow rate;

[0121] S62, CFD solution stage: Under given boundary conditions, solve the gas phase and droplet motion equations, calculate the instantaneous distribution of the airflow field and cloud field, update the interphase coupling source terms, use pressure-based coupling algorithms (SIMPLE, PISO, etc.) to solve the flow field, and use a set time step to ensure numerical stability.

[0122] S63, Proxy Model Calculation Stage: Extract the total temperature and pressure and exhaust static pressure of the engine inlet section (C section), and input them into the proxy model to calculate the inlet (C section) flow rate and outlet (D section) flow rate, temperature and pressure of the engine at a set speed or fuel flow rate; the proxy model output is used to update the boundary conditions of the CFD calculation of the test section;

[0123] S64. Iteration and Convergence Judgment: Compare the CFD calculation results and the surrogate model output in the current time step to see if they meet the convergence condition. If they do not meet the condition, return the surrogate model output to the CFD solver and continue iterating. If they meet the condition, proceed to the next time step.

[0124] S65, Ice layer growth: If it is necessary to simulate the growth of the ice layer on the wall, the ice layer thickness is calculated based on the wall water collection rate, heat transfer rate, freezing rate, etc., and the wall shape is updated at each time period; the mesh is adjusted accordingly by embedding a mesh dynamic deformation or local reconstruction module. Since this embodiment mainly focuses on the distribution of airflow parameters and cloud parameters, a fixed geometry is used to ignore the ice layer change or to update it once at a set time scale during the iteration process.

[0125] S66. Post-processing and data output: After completing the calculation of all time steps, perform statistical and visual analysis on the flow field data, and output velocity vector map, pressure cloud map, temperature cloud map, droplet concentration distribution, particle size distribution, total pressure / total temperature non-uniformity, liquid water content distribution, etc.

[0126] In summary, the above embodiments provide a multi-dimensional simulation method for the flow field of an icing wind tunnel test section considering real-time calculation of aero-engine performance. Based on geometric modeling and mesh generation of the test section, this method constructs a three-dimensional gas-liquid two-phase flow model. The momentum, heat, and mass exchange between air and droplets is described using an Eulerian-Lagrange framework and a droplet transformation sub-model. An aero-engine performance proxy model is introduced to calculate the engine inlet and outlet flow rates, temperature, and pressure in real time. The CFD solver and the proxy model exchange data before each iteration step, updating boundary conditions in real time to achieve high-fidelity bidirectional coupling simulation of the test section flow field and engine performance. The obtained simulation results include multi-dimensional data such as test section velocity, pressure, temperature, liquid water content, and droplet size distribution, which can be used for aerodynamic layout optimization of the test section and precise control of engine icing test environment parameters. Furthermore, the above embodiments provide a coupling calculation strategy between the aero-engine performance proxy model and the test section CFD. This strategy utilizes a pre-trained aero-engine performance proxy model to read parameters such as total intake pressure, total intake temperature, and exhaust static pressure from the CFD calculation results of the test section before the start of each iteration step. It then predicts the engine's inlet and outlet flow rates, temperature, and pressure at a set speed or fuel flow rate, updating the boundary conditions of the test section's CFD calculations in real time. This strategy iterates cyclically at each time step until convergence, ensuring the accuracy and computational efficiency of the coupling between engine performance and the flow field in the test section, and providing a universal data exchange framework for multidisciplinary collaborative simulation.

[0127] As can be seen, this application addresses the problem that existing testing methods and numerical simulation techniques cannot quickly, accurately, and comprehensively assess whether the flow field of the test section meets the requirements of aero-engine icing tests. It provides a high-precision simulation technology that can realistically simulate the gas-liquid two-phase flow field in the test section and its coupling with engine performance. The high-precision simulation technology introduces an aero-engine performance proxy model and a droplet dynamics model into the three-dimensional CFD simulation to achieve efficient coupling and prediction of complex gas-liquid two-phase flow and engine performance.

[0128] As shown in Figure 8, another preferred embodiment of this application also provides a multi-dimensional simulation system for the flow field of an icing wind tunnel test section, including:

[0129] The geometric model building module is used to create a three-dimensional geometric model based on the actual structure of the test section using three-dimensional modeling software. The model includes the rear compartment of the test chamber, corner section, engine air intake duct, free jet nozzle, test bench, exhaust diffuser, aero-engine and drive shaft.

[0130] The mesh generation and preprocessing module is used to mesh the computational domain of the test section after completing the geometric modeling of the test section. During the mesh generation process, boundary markers are introduced for subsequent physical model settings and surrogate model coupling.

[0131] The module for defining governing equations and selecting physical models is used to determine governing equations and select physical models, including gas phase flow governing equations, droplet dynamics models, droplet transformation strategies, and interphase coupling between droplets and gas flow.

[0132] The boundary condition setting module is used to define the boundary conditions of the test section computational domain, including the test section inlet boundary conditions, the test section outlet boundary conditions, the boundary conditions at the spray device outlet, the wall boundary conditions, and the aerodynamic interface between the aero-engine and the rear compartment of the test cabin.

[0133] The module for establishing and coupling the performance proxy model of aero-engines is used to establish a performance proxy model of aero-engines using aero-engine performance calculation or test data and machine learning methods. The trained proxy model is packaged into a general executable file and coupled with the CFD solver through code.

[0134] The multi-dimensional coupled calculation module is used to transfer information between the CFD solver and the surrogate model in a time-step iterative manner, perform multi-dimensional coupled calculations, complete all time-step calculations, and realize multi-dimensional simulation of the flow field in the icing wind tunnel test section.

[0135] The multi-dimensional simulation system for the flow field of an icing wind tunnel test section provided in this embodiment adopts the multi-dimensional simulation method for the flow field of an icing wind tunnel test section in the above embodiments, solving the problem that existing test methods and numerical simulation methods cannot quickly, accurately, and comprehensively evaluate whether the flow field of the test section meets the requirements of aero-engine icing tests. Compared with the prior art, the beneficial effects of the multi-dimensional simulation system for the flow field of an icing wind tunnel test section provided in this application are the same as those of the multi-dimensional simulation method for the flow field of an icing wind tunnel test section provided in the above embodiments, and other technical features of the multi-dimensional simulation system for the flow field of an icing wind tunnel test section are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0136] As shown in Figure 9, a preferred embodiment of this application also provides an electronic device, including 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 the steps of the multi-dimensional simulation method of the flow field of the icing wind tunnel test section in the above embodiment.

[0137] This application provides an electronic device that employs the multi-dimensional simulation method of the icing wind tunnel test section flow field in the above embodiments, solving the problem that existing experimental testing methods and numerical simulation methods cannot quickly, accurately, and comprehensively evaluate whether the flow field of the test section meets the requirements of aero-engine icing tests. Compared with the prior art, the beneficial effects of the electronic device provided in this application are the same as those of the multi-dimensional simulation method of the icing wind tunnel test section flow field provided in the above embodiments, and other technical features of the electronic device are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0138] As shown in Figure 10, a preferred embodiment of this application also provides a computer device, which can be a terminal or a liveness detection server, and its internal structure diagram is shown in Figure 10. The computer device includes a processor, memory, and a network interface connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores an operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The network interface is used to communicate with other external computer devices via a network connection. When the computer program is executed by the processor, it implements the steps of the above-described multi-dimensional simulation method for the flow field of the icing wind tunnel test section.

[0139] Those skilled in the art will understand that the structure shown in Figure 10 is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or may combine certain components, or may have different component arrangements.

[0140] The computer equipment provided in this application employs the multi-dimensional simulation method of the icing wind tunnel test section flow field in the above embodiments, solving the problem that existing experimental testing methods and numerical simulation methods cannot quickly, accurately, and comprehensively assess whether the flow field of the test section meets the requirements of aero-engine icing tests. Compared with the prior art, the beneficial effects of the computer equipment provided in this application are the same as those of the multi-dimensional simulation method of the icing wind tunnel test section flow field provided in the above embodiments, and other technical features in the electronic equipment are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.

[0141] A preferred embodiment of this application also provides a storage medium, the storage medium including a stored program, which, when the program is executed, controls the device where the storage medium is located to perform the steps of the multi-dimensional simulation method of the flow field in the icing wind tunnel test section in the above embodiments.

[0142] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0143] If the functions described in this embodiment are implemented as software functional units and sold or used as independent products, they can be stored in one or more computing device-readable storage media. Based on this understanding, the parts of this application's embodiments that contribute to the prior art or the technical solutions can be embodied in the form of software products. These software products are stored in a storage medium and include several instructions to cause a computing device (which may be a personal computer, server, mobile computing device, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage media include: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and various other media capable of storing program code.

[0144] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language C++ and the embedded programming language C.

[0145] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0146] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0147] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0148] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the multi-dimensional simulation method for the flow field of the icing wind tunnel test section as described above.

[0149] The computer program product provided in this application solves the problem that existing experimental testing methods and numerical simulation methods cannot quickly, accurately, and comprehensively evaluate whether the flow field of the test section meets the requirements of aero-engine icing tests. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as those of the multi-dimensional simulation method of the flow field of the icing wind tunnel test section provided in the above embodiments, and will not be elaborated here.

[0150] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0151] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A multi-dimensional simulation method for the flow field in an icing wind tunnel test section, characterized in that, The steps include: S1. Establishing a 3D geometric model of the test section using 3D modeling software, including the rear compartment of the test chamber, corner sections, engine intake ducts, free jet nozzles, test bench, exhaust diffuser, aero-engine, and drive shaft; S2. After completing the geometric modeling of the test section, meshing the computational domain of the test section, introducing boundary markers during meshing for subsequent physical model settings and surrogate model coupling; S3. Determining the governing equations and selecting physical models, including gas phase flow governing equations, droplet dynamics model, droplet conversion strategy, and interphase coupling between droplets and airflow; S4. Defining the boundary conditions of the computational domain of the test section, including the test section inlet boundary conditions, test section outlet boundary conditions, boundary conditions at the spray device outlet, and wall boundary stripes. S5. Using engine performance calculations or test data, establish an aero-engine performance proxy model through machine learning, encapsulate the trained proxy model into a general executable file, and couple it with the CFD solver through code; S6. Use a time-step iterative approach to transfer information between the CFD solver and the proxy model, perform multi-dimensional coupled calculations, complete all time-step calculations, and realize multi-dimensional simulation of the flow field in the icing wind tunnel test section. The specific steps include: S61. Initialization stage: import geometry and mesh, set initial temperature field, pressure field, velocity field and cloud field; initialize engine operating conditions and call the proxy model to calculate the inlet and outlet flow rates, temperature and pressure of the engine at a set speed or fuel flow rate; S62, CFD Solution Stage: Under given boundary conditions, solve the gas phase and droplet motion equations, calculate the instantaneous distribution of the airflow field and cloud field, and update the interphase coupling source terms. A pressure-based coupling algorithm is used to solve the flow field, and a set time step is used to ensure numerical stability. S63, Surrogate Model Calculation Stage: Extract the total temperature and pressure at the engine inlet section and the static pressure of the exhaust environment. Input these into the surrogate model to calculate the engine's inlet flow rate, outlet flow rate, temperature, and pressure at a set speed or fuel flow rate. The surrogate model output is used to update the boundary conditions for the CFD calculation in the test section. S64, Iteration and Convergence Judgment: Compare the CFD calculation results within the current time step with the surrogate model output to see if the convergence condition is met. If not, then... The proxy model output is returned to the CFD solver for further iteration; if satisfied, the process proceeds to the next time step; S65, Ice layer growth: If it is necessary to simulate the growth of the ice layer on the wall, the ice layer thickness is calculated based on the wall water collection rate, heat transfer rate, and freezing rate at each time step, and the wall shape is updated; the mesh is adjusted accordingly by embedding a dynamic mesh deformation or local reconstruction module, and the ice layer change is ignored by using a fixed geometry during the iteration process or updated once at a set time scale; S66, Post-processing and data output: After completing the calculation of all time steps, the flow field data is statistically and visually analyzed, and the velocity vector map, pressure cloud map, temperature cloud map, droplet concentration distribution, particle size distribution, total pressure / total temperature non-uniformity, and liquid water content distribution are output.

2. The multi-dimensional simulation method for the flow field of an icing wind tunnel test section according to claim 1, characterized in that, In step S1, when establishing the geometric model of the aero-engine, a simplified outline model is used to reduce the amount of computation, and an interface for interaction with the proxy model is reserved.

3. The multi-dimensional simulation method for the flow field of an icing wind tunnel test section according to claim 1, characterized in that, In step S2, during the mesh generation process, a hybrid mesh type is selected. Specifically, for key areas including the nozzle outlet, engine inlet and outlet, and exhaust diffuser inlet, locally densified or unstructured meshes are used to capture the flow details of the airflow field and cloud field. The mesh quality is ensured by checking the indicators of cell distortion, orthogonality, and volume distribution.

4. The multi-dimensional simulation method for the flow field of an icing wind tunnel test section according to claim 1, characterized in that, In step S3, the gas phase flow control equations adopt the incompressible or weakly compressible Navier-Stokes equations as the control equations for the gas phase, including the mass conservation equation, momentum equation and energy equation, and an appropriate solver is selected according to the incoming Mach number; the turbulence simulation adopts the Reynolds-averaged method or the large eddy simulation method; the energy equations consider heat transfer and latent heat exchange during phase change. When setting up the droplet dynamics model, the droplets are described using the Lagrangian method. The changes in mass, position, velocity and temperature of each droplet or droplet group are tracked over time. The main forces acting on the droplets include drag force, gravity, buoyancy, pressure gradient force, lift force and virtual mass force. The equation for calculating droplet acceleration is: ; where m p v is the mass of the droplet. p For droplet velocity, For droplet acceleration, The drag force acting on the droplet. For lift, For gravity, The pressure gradient force is the drag force, which is the product of the drag coefficient and the velocity difference. The drag coefficient can be determined using the Schiller-Naumann model or the Oseen model based on the Reynolds number and droplet shape. Collision aggregation uses the O'Rourke model, and breakup uses the Kelvin-Helmholtz and Rayleigh-Taylor instability criteria. Wall collisions are judged as rebound, wetting, or breakup based on the elastic coefficient and the critical Weber number. The droplet transformation strategy includes: when the droplet diameter is less than a threshold or the local droplet volume fraction is less than a threshold... The Lagrange particle model is continued. When the local volume fraction is higher than the threshold, the droplet swarm is treated as a continuous medium, and the swarm behavior is described by smooth particle hydrodynamics or discrete element method. The swarm model describes the momentum and energy exchange between the droplet swarm and the gas phase through the control volume fraction equation and the swarm dynamics equation. The interphase coupling between the droplets and the gas flow is as follows: the reaction of the droplet motion on the gas phase is added to the gas phase control equation through source terms, including momentum source terms and energy source terms, to express the drag and heat transfer effect of the droplets on the gas flow. The CFD solver realizes the two-phase coupling by cyclically updating the interphase source terms and droplet trajectory.

5. The multi-dimensional simulation method for the flow field of an icing wind tunnel test section according to claim 1, characterized in that, In step S4, the inlet boundary conditions of the test section include incoming flow velocity, pressure, temperature, relative humidity, and turbulence intensity; the outlet boundary conditions of the test section include setting pressure outlet or full pressure outlet boundary conditions; the boundary conditions at the outlet of the spray device are set using an atomizing nozzle model, and the droplet size distribution and initial velocity are calculated based on the nozzle structure and water and air supply pressures. The droplet size is commonly represented by a log-normal distribution or a Rosner-Marcano distribution, and the liquid water content and median volume diameter are set; the wall boundary conditions include setting the outer surface of the drive shaft as a rotating wall and setting other walls of the test section as non-slip solid walls, considering the influence of wall temperature, roughness, and ice growth on surface friction; the aerodynamic interface between the aero-engine and the rear compartment of the test cabin includes the engine inlet section and the outlet section, and the dynamic updating of the aerodynamic interface parameters is achieved by constructing an aero-engine performance proxy model and coupling calculation with the test section CFD model.

6. The multi-dimensional simulation method for the flow field of an icing wind tunnel test section according to claim 1, characterized in that, Step S5 specifically includes the following steps: S51, Experimental Design and Data Sampling: Based on a space-filling strategy, sample points are selected within the entire working envelope of the engine. Input variables include flight altitude, flight Mach number, total intake temperature, engine speed, or fuel flow rate. Output variables include engine inlet and outlet flow rates, temperature, and pressure. The collected samples come from engine performance calculations or experimental data. S52, Feature Processing and Model Selection: Input and output data are normalized, and suitable surrogate model structures are selected, including multinomial response surface surrogate models, radial basis function surrogate models, Kriging surrogate models, and deep neural network surrogate models. S53, Model Training and Validation: The surrogate model is trained using cross-validation and validated using data not used in training. Model performance is evaluated based on prediction error and correlation coefficient indicators, and the model is optimized by increasing the sample size or adjusting the model structure. S54, Surrogate Model Embedding and Coupling: The trained surrogate model is packaged into a general executable file and coupled with the CFD solver through code. During the model coupling calculation process, the data exchange rule between models is: data exchange occurs before the start of each iteration step.

7. A multi-dimensional simulation system for the flow field of an icing wind tunnel test section, characterized in that, include: The geometric model construction module is used to create a 3D geometric model of the test section using 3D modeling software, including the rear compartment of the test chamber, corner sections, engine intake ducts, free jet nozzles, test bench, exhaust diffuser, aero-engine, and drive shaft. The mesh generation and preprocessing module is used to mesh the computational domain of the test section after geometric modeling. Boundary markers are introduced during mesh generation for subsequent physical model settings and surrogate model coupling. The governing equation definition and physical model selection module is used to determine the governing equations and select physical models, including gas phase flow governing equations, droplet dynamics models, droplet transformation strategies, and interphase coupling between droplets and airflow. Boundary condition settings are also included. The module defines the boundary conditions of the test section computational domain, including the test section inlet boundary conditions, test section outlet boundary conditions, boundary conditions at the spray device outlet, wall boundary conditions, and the aerodynamic interface between the aero-engine and the test compartment rear compartment; the aero-engine performance proxy model construction and coupling module is used to build an aero-engine performance proxy model using aero-engine performance calculation or test data through machine learning, encapsulate the trained proxy model into a general executable file, and couple it with the CFD solver through code; the multi-dimensional coupling calculation module is used to transfer information between the CFD solver and the proxy model in a time-step iterative manner, perform multi-dimensional coupling calculations, complete the calculations for all time steps, and realize... The current icing wind tunnel test section flow field multi-dimensional simulation is used for: Initialization stage: importing geometry and mesh, setting initial temperature, pressure, velocity, and cloud fields; initializing engine operating conditions and calling the surrogate model to calculate the engine's inlet and outlet flow rates, temperature, and pressure at a set speed or fuel flow rate; CFD solution stage: solving the gas phase and droplet motion equations under given boundary conditions, calculating the instantaneous distribution of the airflow and cloud fields, updating the interphase coupling source terms, using a pressure-based coupling algorithm to solve the flow field, and using a set time step to ensure numerical stability; surrogate model calculation stage: extracting the total temperature and pressure at the engine inlet section and the exhaust environment static pressure, inputting them into the surrogate model to calculate the engine's operating conditions at a set speed. The system calculates the inlet and outlet flow rates, temperature, and pressure under different speeds or fuel flow rates. The surrogate model output is used to update the boundary conditions for the CFD calculations in the test section. Iteration and convergence judgment: The system compares the CFD calculation results with the surrogate model output within the current time step to see if the convergence conditions are met. If not, the surrogate model output is returned to the CFD solver for further iteration. If the conditions are met, the system proceeds to the next time step. Ice layer growth: To simulate the growth of the ice layer on the wall, the ice layer thickness is calculated based on the wall's water collection rate, heat transfer rate, and freezing rate at each time interval, and the wall shape is updated. The mesh is adjusted accordingly by embedding a dynamic mesh deformation or local reconstruction module. During iteration, a fixed geometry is used to ignore ice layer changes, or the system updates once at a set time scale.Post-processing and data output: After completing all time-step calculations, statistical and visual analysis is performed on the flow field data, outputting velocity vector map, pressure contour map, temperature contour map, droplet concentration distribution, particle size distribution, total pressure / total temperature non-uniformity, and liquid water content distribution.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the multi-dimensional simulation method for the flow field of the icing wind tunnel test section as described in any one of claims 1 to 6.

9. A storage medium comprising a stored program, characterized in that, When the program is running, it controls the device containing the storage medium to perform the steps of the multi-dimensional simulation method for the flow field of the icing wind tunnel test section as described in any one of claims 1 to 6.

Images