Ejector optimization method and system based on improved orthogonal test method

Abstract

The invention relates to the technical field of electric digital data processing, provides an ejector optimization method and system based on an improved orthogonal test method, and combines the advantages of computational fluid mechanics simulation, neural network modeling and a multi-objective optimization algorithm. Firstly, flow field and sound field characteristics of the ejector are accurately simulated through a simulation model, and high-quality data support is provided for optimization. And secondly, decision variable identification and data enhancement are carried out by using an improved orthogonal test method, so that the training effect of the neural network model is improved. And finally, multi-objective optimization is carried out by adopting an NSGA-II genetic algorithm, so that the performance of the ejector is balanced among a plurality of objectives. The optimized ejector has remarkable advantages in the aspects of speed distribution, turbulent kinetic energy and acoustic power level. After optimization, the flow loss is reduced, the air flow expansion state is improved, the vacuum degree is improved, meanwhile, the flow noise is effectively reduced, and the ejector has the working efficiency and better performance.

Description

Technical Field

[0001] The present invention relates to the technical field related to electronic digital data processing, and in particular to an injector optimization method and system based on an improved orthogonal test method. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] In modern seawater desalination technology, the multi-effect distillation-thermal vapor compression (MED-TVC) system has been widely used in areas with water shortages due to its high energy efficiency and adaptability. The ejector, one of the core components of the system, injects low-pressure fluid through high-pressure working fluid to achieve energy transfer and fluid mixing. Its performance directly affects the overall efficiency and operating cost of the system.

[0004] Existing ejector designs mainly rely on empirical formulas and experimental methods, and it is often difficult to balance efficiency improvement, structural rationality, and flow stability during the optimization process. The flow inside the ejector involves complex fluid mechanics phenomena, such as shock waves, turbulence, and phase change, which can easily lead to energy loss and flow noise, affecting the long-term operational reliability of the equipment. In addition, the geometric structure of the ejector has an important influence on its performance, but existing optimization design methods often lack systematicity and it is difficult to accurately adjust multiple geometric parameters to achieve optimal performance.

[0005] Previous studies on injector geometry optimization have mainly focused on maximizing performance, usually using empirical formulas, CFD numerical simulations or experimental methods for optimization, and then using external equipment to reduce noise during the use of the injector, which increases costs. At the same time, noise has a certain impact on the working performance of the injector and affects the service life of the injector.

[0006] In recent years, the combination of computational fluid dynamics (CFD) and artificial intelligence technology has provided new ideas for injector optimization. Existing methods still have many problems: first, CFD simulations are computationally intensive and it is difficult to efficiently handle multi-variable and multi-objective optimization problems; second, artificial intelligence models such as neural network models are highly dependent on data quality and may be affected by the distribution of training data, resulting in limited generalization capabilities; in addition, although multi-objective optimization algorithms can balance between different objectives, their search efficiency depends on the initial population and fitness evaluation strategy, which may lead to local optimal solutions. Summary of the invention

[0007] In order to solve the above problems, the present invention proposes an ejector optimization method and system based on an improved orthogonal test method, comprehensively considers flow noise and performance optimization to build a model, combines decision variable identification, artificial neural network (ANN) agent model and non-dominated sorting genetic algorithm (NSGA-II), and realizes the optimal design of the ejector with the best performance and minimum flow noise.

[0008] In order to achieve the above object, the present invention adopts the following technical solution: One or more embodiments provide an injector optimization method based on an improved orthogonal test method, comprising the following steps: Construct the ejector flow field simulation model and the acoustic field simulation model; Based on the constructed models, the decision variables are identified through the improved orthogonal test method, and the objective function model is constructed through the trained neural network model with the goal of maximizing the ejection ratio and minimizing the average sound pressure level. The objective function model constructed based on the trained neural network model calculates the objective function value as the fitness value, and multi-objective optimization is performed on the geometric dimensions of the injector to obtain the optimized target dimensions.

[0009] One or more embodiments provide an injector optimization system based on an improved orthogonal test method, comprising: A model building module is configured to build an ejector flow field simulation model and an acoustic field simulation model; The objective function model building module is configured to identify decision variables through an improved orthogonal test method based on each constructed model, and to build an objective function model through a trained neural network model with the goal of maximizing the ejection ratio and minimizing the average sound pressure level; The target optimization module is configured to calculate the target function value as the fitness value based on the target function model constructed by the trained neural network model, and use the NSGA-II genetic algorithm to perform multi-objective optimization on the geometric size of the injector to obtain the optimized target size.

[0010] Compared with the prior art, the present invention has the following beneficial effects: The present invention replaces the traditional CFD optimization calculation with a neural network model, reduces repeated simulation calculations, and improves the optimization efficiency. The data enhancement technology based on the improved orthogonal test method is adopted to improve the generalization ability of the neural network model and make the optimization result more reliable. Combined with the NSGA-II genetic algorithm, the ejection ratio and the average sound pressure level of the ejector can be optimized at the same time, avoiding the limitations brought by single-objective optimization. The optimized ejector shows significant advantages in velocity distribution, turbulent kinetic energy and sound power level; after optimization, the flow loss is reduced, the airflow expansion state is improved, the vacuum degree is increased, and the flow noise is effectively reduced, so that the ejector has the working efficiency and better performance.

[0011] The advantages of the present invention and additional advantages will be described in detail in the following specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0012] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their description are used to explain the present invention but do not constitute a limitation of the present invention.

[0013] Figure 1 is a flow chart of the injector optimization method according to Embodiment 1 of the present invention; Figure 2 is a schematic diagram of geometric parameters of an injector according to Embodiment 1 of the present invention; Figure 3 is a schematic diagram of a training artificial neural network model according to Embodiment 1 of the present invention; Figure 4 It is a comparison diagram of the velocity, turbulent kinetic energy and sound power level cloud diagrams of the initial and optimized ejectors of Example 1 of the present invention. DETAILED DESCRIPTION

[0014] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0015] It should be noted that the following detailed descriptions are exemplary and are intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0016] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof. It should be noted that, in the absence of conflict, the various embodiments of the present invention and the features in the embodiments can be combined with each other. The embodiments will be described in detail below in conjunction with the accompanying drawings.

[0017] Example 1 In the technical solutions disclosed in one or more embodiments, Figures 1 to 4 As shown, the injector optimization method based on the improved orthogonal test method includes the following steps: Step 1, constructing an ejector flow field simulation model and an acoustic field simulation model; Step 2: Based on the constructed models, the decision variables are identified through the improved orthogonal test method, with the goal of maximizing the ejection ratio and minimizing the average sound pressure level, and the objective function model is constructed through the trained neural network model; Step 3: Calculate the objective function value as the fitness value based on the objective function model constructed based on the trained neural network model, perform multi-objective optimization on the geometric dimensions of the injector, and obtain the optimized target dimensions.

[0018] This method combines the advantages of computational fluid dynamics (CFD) simulation, neural network modeling and multi-objective optimization algorithm. First, the flow field and acoustic field characteristics of the ejector are accurately simulated by the simulation model to provide high-quality data support for optimization. Secondly, the improved orthogonal test method is used for decision variable identification and data enhancement, which improves the training effect of the neural network model. Finally, the NSGA-II genetic algorithm is used for multi-objective optimization to balance the performance of the ejector between multiple objectives.

[0019] In this embodiment, the traditional CFD optimization calculation is replaced by a neural network model, which reduces repeated simulation calculations and improves the optimization efficiency. The data enhancement technology based on the improved orthogonal test method is used to improve the generalization ability of the neural network model and make the optimization results more reliable. Combined with the NSGA-II genetic algorithm, the ejection ratio and the average sound pressure level of the ejector can be optimized at the same time, avoiding the limitations of single-objective optimization. The optimized ejector shows significant advantages in velocity distribution, turbulent kinetic energy and sound power level; after optimization, the flow loss is reduced, the airflow expansion state is improved, the vacuum degree is increased, and the flow noise is effectively reduced, so that the ejector has better working efficiency and performance.

[0020] A further technical solution, in step 2, an improved orthogonal test method, obtains multiple offset data corresponding to the current parameter data by dynamically adjusting the offset, expands the parameter data for data enhancement, and inputs the enhanced data into the ejector flow field simulation model and the sound field simulation model for orthogonal testing to obtain corresponding test results, which include the ejector's injection ratio and average sound pressure level.

[0021] A further technical solution is to expand the data set as a training set based on the identified decision variables through an improved orthogonal test method, use the parameter values ​​of the decision variables in the training set as input, and use the ejection ratio and average sound pressure level of the ejector model obtained under the corresponding parameter values ​​as output to train the constructed neural network model, and obtain an objective function model constructed by the neural network model; The test results of the orthogonal test method of this embodiment include the ejection ratio and the average sound pressure level of the ejector, which are used to evaluate the performance of the ejector. By adopting the improved orthogonal test method, the design space can be explored more comprehensively, the training data quality of the neural network model can be improved, and the accuracy and stability of the optimization process can be improved; the above steps are specifically described below.

[0022] like Figure 2 As shown, the ejector mainly includes eleven main geometric structures, including the nozzle inlet diameter (D nl ), nozzle tapering length (L nc ), nozzle throat diameter (D t ), nozzle expansion section length (L nd ), nozzle outlet diameter (D no ), nozzle outlet position (NXP), premixing chamber length (Lcs), equal area mixing chamber diameter (Dmix), equal area mixing chamber length (Lmix), diffusion chamber diameter (Ddiff) and diffusion chamber length (Ldiff).

[0023] In step 1, the ejector flow field simulation model and the acoustic field simulation model are constructed. First, the following assumptions are set: 1) Assume that the fluid inside the ejector is an ideal gas and follows the ideal gas equation; 2) The ejector wall is insulated; In step 1, the ejector flow field simulation model and the acoustic field simulation model are constructed as follows: Step 11, geometry establishment: using computer-aided engineering software to establish a two-dimensional geometric model of the injector fluid calculation domain; Specifically, the geometric model defines the structural shape of the injector and contains 11 geometric parameters required for optimization; the purpose of this step is to build an accurate physical model of the injector.

[0024] Step 12: Meshing: Meshing the constructed two-dimensional geometric model to discretize the geometric model; Furthermore, the mesh near the inner wall of the ejector is encrypted, while ensuring that the average mesh quality of the overall structure is greater than 0.9; the ability to analyze the boundary layer flow is enhanced by refining the mesh near the inner wall of the ejector; The purpose of meshing is to convert the continuous geometric model into a discrete grid suitable for numerical calculation, so that the CFD solver can perform calculations.

[0025] Step 13, CFD simulation: After completing the meshing, import the CFD solver for numerical calculation and build the ejector flow field simulation model; use the FW-H acoustic analogy method to predict the flow noise, build the sound field simulation model, and rely on the meshed flow field data to calculate the acoustic characteristics.

[0026] In the ejector flow field simulation model constructed by SST k-ω, the grid of the computational domain is used to discretize the Navier-Stokes equations during the simulation process, so that the fluid parameters (velocity, pressure, turbulence characteristics, etc.) can be calculated within each grid cell; The acoustic solution uses the FW-H acoustic analogy method to predict flow noise and relies on meshed flow field data to calculate acoustic characteristics. The time step, boundary conditions, and discrete format are solved and iterated based on the grid cells.

[0027] Among them, FW-H (Ffowcs Williams-Hawkings) is a method based on acoustic analogy theory and is used to calculate noise propagation in fluids.

[0028] SST k-ω (Shear Stress Transport k-ω) is a shear stress transport k-ω turbulence model that combines the advantages of the k-ε model and the k-ω model and is suitable for turbulence simulation near the wall.

[0029] In some embodiments, the ejector flow field simulation model includes a steady-state turbulence model and a transient turbulence model. The steady-state turbulence model uses SST kw to obtain flow velocity and turbulent kinetic energy distribution, and the transient turbulence model uses SAS scale adaptive simulation.

[0030] The steady-state turbulence model is constructed by the SST k-ω model, and the flow field inside the ejector is predicted by the SST k-ω model. The specific model is as follows: (1); (2); in, krepresents turbulent kinetic energy (TKE); ω represents specific dissipation rate, that is, the ratio of turbulent dissipation rate; represents the density of the fluid; Indicates time; It represents the generation term of turbulent kinetic energy and characterizes the energy source of turbulence; and is the amount of ω produced and dissipated; and It is a user-defined source item; and represents the amount of k generated and dissipated; and : represents the spatial coordinate component, and the subscripts i and j are used to represent different spatial directions, usually the x and y directions; : represents the velocity component, and the subscript i corresponds to different spatial directions (for example, Corresponding to the velocity in the x direction, Corresponding to the velocity in the y direction, Corresponding to the velocity in the z direction; describes the flow velocity of the fluid in the corresponding direction. The x direction, y direction and z direction are three orthogonal directions in the space coordinate system; : The dynamic viscosity of the fluid, which indicates the viscosity of the fluid; : Turbulent viscosity, used to describe the effect of turbulent motion on momentum transfer; and : Prandtl numbers for turbulent kinetic energy and specific dissipation rate, respectively, used to adjust the diffusion term in the turbulence model; Among them, turbulent viscosity The calculation formula is as follows; (3); Where S is the strain rate magnitude, is the low Reynolds number correction factor, is a constant.

[0031] represents the correction function, Represents the calculation correction function The intermediate variable is as follows: (4); (5); in, Represents the wall distance, which is used to describe the distance from the fluid point to the wall in the turbulence model; In some embodiments, the transient turbulence model in the ejector flow field simulation model is constructed by a SAS model (scale adaptive simulation model), and the specific model is as follows: The SAS model introduces an accurate transport equation based on the sea surface temperature k-ω model, as shown in equations (6) and (7). The SAS model formula is as follows: (6); (7); in, is an additional source term, and is obtained from formula (8): (8); in, : represents the model constant, which is used to adjust the relationship between the turbulent viscosity and the turbulent kinetic energy k and the specific dissipation rate ω in the turbulence model. Its value will affect the accuracy of turbulence calculation; : represents the model constant, which is used to adjust the generation term in the turbulence dissipation rate equation; : represents the model constant, which is used to adjust the dissipation term in the turbulence dissipation rate equation; : represents another turbulent Prandtl number used to adjust the diffusion term in the turbulent dissipation rate equation; : represents additional source terms, which are used to consider additional physical effects or corrections in the turbulence model; : represents a set coefficient, which is related to a specific correction or adjustment of the model; : Von Karman constant, used to describe the characteristics of turbulence near the wall, usually with a value of about 0.4; : The invariant of velocity gradient, used to describe the shear characteristics of the flow; : represents the characteristic length scale, which may be related to the geometric size or integral length scale of the flow; : represents the length scale associated with the turbulent kinetic energy k, which may be used to adjust the diffusion or dissipation characteristics in turbulence models; : represents a constant coefficient used to adjust the strength of the source term; : represents the Prandtl number related to the specific dissipation rate ω, which is used to adjust the diffusion term in the turbulence model; : represents the symbol of partial derivative, which means the partial derivative with respect to the corresponding variable.

[0032] The detailed definitions of other parameters and values ​​of constants in the above equations are as follows: (9); in, represents the strain rate tensor, which is the symmetric part of the velocity gradient tensor and is used to describe the velocity deformation of the fluid; represents the shear factor; U' represents the modulus of the velocity gradient, defined as the strain rate tensor The modulus of the second-order velocity derivative, U'', involves the curvature or rotation effect of the velocity field; represents the damping function, which is used to adjust the parameters of the model near the wall; tanh represents the hyperbolic tangent function, which is used for smooth transition; represents the parameters of the damping function; The max and min functions are used to limit range to ensure numerical stability; It represents the positive diffusion term to prevent numerical instability, and takes the larger of the gradient product and the minimum value. This formula ensures that the diffusion term is non-negative and avoids negative values ​​or singular points in the calculation; represents the component of the velocity vector, which represents the velocity of the fluid in the i-th spatial direction; Furthermore, the SAS model can use the standard RANS method in the steady-state region of the constructed geometric model and use a LES-like method for dynamic adjustment in the unsteady-state region, thus saving a lot of computing resources; The RANS (Reynolds-Averaged Navier-Stokes) method is based on the Reynolds-averaged Navier-Stokes equations and handles turbulent flows by time averaging. This method decomposes the instantaneous velocity field into the average velocity field and the turbulent pulsation part, and then averages the Navier-Stokes equations to obtain a set of average equations that can be solved; The LES-like (Large Eddy Simulation) method is a turbulence simulation method between the RANS method and the direct numerical simulation (DNS). The LES method retains the large-scale turbulence structure through filtering operations for direct simulation, while the small-scale turbulence structure is modeled through a sub-grid model. The LES-like method improves or simplifies the LES method in some aspects to adapt to specific engineering applications.

[0033] The ejector flow noise is another performance parameter that this study focuses on. The flow noise reflects the stability of the fluid dynamics inside the ejector to a certain extent. If the ejector makes abnormal noise during operation, it may be caused by unstable internal flow states, such as eddy currents, turbulence, or cavitation, which may affect the normal operation and service life of the ejector. If the ejector generates too much noise during operation, it will have an adverse effect on the surrounding environment and human health. Therefore, the pursuit of quieter ejectors can improve their performance stability and promote the wider application and development of ejectors.

[0034] In this embodiment, the ejector noise source and noise monitoring are realized by a numerical model, and the numerical model construction process is as follows: The noise index is detected on a circle with a radius of 1m and a monitoring point arranged every 15°. There are 13 monitoring points in total, and the noise sound pressure level index is reflected by the average value of the 13 points. The FW-H acoustic analogy method is used in the acoustic model to predict the far-field noise.

[0035] In some embodiments, in step 2, decision variables are identified based on the constructed models, and the key decision parameters and their value ranges are finally determined by calculating the contribution of each geometric parameter to the induced emission ratio (ER) and the average noise sound pressure level (SPLave). The decision variables are identified by an improved orthogonal test method, including the following steps: Step 21, obtain the geometric parameters of the ejector for analysis, dynamically adjust the offset, and select multiple different level values ​​for each parameter; The horizontal value refers to a parameter with multiple values ​​of different sizes. Preferably, the offset is adjusted dynamically. , the calculation formula is: (10); in, represents the intermediate value of the geometric parameter in the orthogonal experimental design, represents the maximum value of the geometric parameter in the orthogonal experimental design, Represents the minimum value of geometric parameters in orthogonal experimental design; are known parameters, and is the set parameter value; Indicates the deviation value, as the parameter value As the value of increases, the deviation value becomes smaller, thereby optimizing the orthogonal test parameter values ​​and finding the influencing parameters more accurately; Step 22, randomly combining the horizontal values ​​of the geometric parameters according to the given horizontal values ​​to obtain a generated experimental table and perform an orthogonal experiment; Step 23, conducting experiments using the constructed ejector flow field simulation model and acoustic field simulation model to obtain test results; Step 24: Perform a one-way ANOVA on the test results to obtain the influence ratio of each geometric parameter on the induced emission ratio (ER) and the average noise sound pressure level (SPLave), take the parameter whose influence ratio is greater than the set value as the decision variable, and determine the value range of the decision variable.

[0036] The method of performing one-way ANOVA on the test results to obtain the influence ratio of each geometric parameter on the induced emission ratio (ER) and the average noise sound pressure level (SPLave) includes the following steps: Step 241, treat each geometric parameter as an independent group, and set the original hypothesis ( ) and the alternative hypothesis ( ); First, the simulation data of the induced emission ratio (ER) and the average noise sound pressure level (SPLave) under different geometric parameter settings were collected and organized into a format suitable for one-way ANOVA, that is, each set of geometric parameters was treated as an independent group, and each group had corresponding ER and SPLave values; Specifically, in this embodiment, each group includes 11 geometric structures of the injector (such as Figure 2 In the middle, there are some differences in the parameters of the geometric structure between the groups; for example, the nozzle throat diameters between the two groups are different, or the nozzle inlet diameters are changed, etc. Specifically, establish the null hypothesis ( ) and the alternative hypothesis ( ), where the null hypothesis is that different geometric parameters have no significant effect on ER or SPLave, that is, the means of each group are equal; the alternative hypothesis is that at least one geometric parameter has a significant effect on ER or SPLave, that is, there are at least two groups with unequal means; significant effect is a relative concept. If there are multiple geometric parameters, after sorting them according to the size of the effect, the corresponding first n geometric parameters can be defined as having a significant effect, where n is the set value.

[0037] Step 242, calculating the within-group sum of squares (SSW) and the between-group sum of squares (SSB); Specifically, SSW measures the difference between each data point within a group and the mean of the group, and SSB measures the difference between the mean of each group and the overall mean; Step 243, calculate the within-group degrees of freedom (dfW) and between-group degrees of freedom (dfB); Calculate the degrees of freedom. The within-group degrees of freedom (dfW) is the total number of samples minus the number of groups, and the between-group degrees of freedom (dfB) is the number of groups minus one. Step 244, based on the calculated sums of squares and degrees of freedom, calculate the mean square error to obtain the within-group mean square error MSW and the between-group mean square error MSB; Calculate the mean square error (MS), that is, MSW is equal to SSW divided by dfW, and MSB is equal to SSB divided by dfB; Step 244, dividing the between-group mean square error MSB by the within-group mean square error MSW to obtain the statistic F; Step 245, compare the obtained statistical value with the set critical value to determine whether each group meets the null hypothesis, so as to screen out the group that meets the alternative hypothesis, that is, the group that has an impact on ER or SPLave; Specifically, the critical value can be set to 0.05, and the calculated F statistic is compared with the critical value. If the F statistic is greater than the critical value, the null hypothesis is rejected, indicating that at least one geometric parameter has a significant effect on ER or SPLave; Step 246: for the selected groups, extract the data of the same geometric parameters in each group, calculate the mean value of each geometric parameter, and obtain the influence ratio of each geometric parameter on the ejection ratio (ER) or the average noise sound pressure level (SPLave); The first few parameters with a larger proportion are taken as decision variables, and the approximate value range of each geometric parameter is determined according to the set values ​​of ER or average noise sound pressure level SPLave that meet the conditions.

[0038] Specifically, if multiple groups do not conform to the original hypothesis, the same parameters are taken out, such as taking out the nozzle inlet diameters of all groups to calculate the mean, and taking out the nozzle tapering section lengths of all groups to calculate the mean. By comparing the means, the influence value proportion of each parameter is obtained; the larger the mean, the greater the influence value proportion.

[0039] If certain geometric parameters are found to have a significant effect on ER or SPLave, further post hoc tests (such as Tukey test) can be performed to determine which specific parameters have significant differences, and a chart of the means of each group can be drawn to intuitively show the impact of different geometric parameters on ER and SPLave.

[0040] In this embodiment, the key factor identification uses the F value to quickly screen out the factors that have a significant impact on the response variable, and clarifies the contribution of each factor through the proportion to guide the priority allocation of resources; while reducing the number of experiments, it accurately locates the direction of process parameter optimization to improve efficiency or quality; and avoids blindly adjusting non-significant factors to reduce R&D or production costs.

[0041] Traditional data enhancement methods expand data based on experience according to existing values, that is, through manual value selection. This embodiment dynamically generates expanded data for each geometric parameter through a set deviation value calculation formula, and can adaptively adjust the disturbance amplitude according to the distribution characteristics of the original data, while avoiding excessive deviation from the real sample, effectively covering the potential feature space boundary; secondly, when selecting key parameters in the geometric parameters, combined with the one-way analysis of variance (One-Way ANOVA) method, the independent influence intensity of each geometric parameter on the target performance indicator can be quantitatively evaluated (such as sorting key parameters by F value), so as to give priority to the implementation of directional enhancement for highly sensitive parameters, so that the generated data is more focused on the optimization interval of the key parameter combination; Specifically, through decision variable identification, the decision variables determined include: the diameter D of the equal area mixing chamber mix , nozzle outlet position NXP, nozzle expansion section length L nd , Nozzle outlet diameter D no The length of the premixing chamber is L cs ; Furthermore, with the goal of maximizing the ejection ratio and minimizing the average sound pressure level, a method for constructing an objective function model through the trained neural network model includes the following steps: Step S1, obtaining data of decision variables of ejector design, expanding training data by improved orthogonal test method, constructing data including geometric parameter data of ejector decision variables, and corresponding data of simulation results of ejection ratio (ER) and average noise sound pressure level (SPLave); Specifically, the improved orthogonal test method includes the above steps 21 to 23, which will not be described in detail here; Step S2, taking the geometric parameter data in the training set as input and the induced emission ratio (ER) and the average noise sound pressure level (SPLave) as output, training the constructed neural network model; Step S3: Calculate the prediction loss according to the prediction output of the neural network model, adjust the parameters of the neural network model, and iterate the training until the prediction accuracy is met to obtain the trained model.

[0042] Step 3: Calculate the objective function value as the fitness value based on the objective function model constructed by the trained neural network model, and use the NSGA-II genetic algorithm to perform multi-objective optimization on the geometric dimensions of the injector to obtain the optimized target dimensions, including the following steps: Step 31, initializing the population: using Latin hypercube sampling (LHS) to randomly generate N initial individuals, each individual is a combination of geometric parameters of the ejector, and setting the population size P and the maximum number of iterations G; In this embodiment, Latin hypercube sampling (LHS) is used to replace traditional random initialization to improve the coverage of the population and reduce the deviation of the initial population; Step 32: Calculate the objective function value based on the trained neural network as the fitness value; Step 33, select individuals whose fitness values ​​are greater than the set threshold to form a new parent population; Step 34: Perform crossover and mutation operations on the individuals of the parent population to obtain a child population; Step 35: merge the offspring population and the parent population, and iterate (execute step 32) until the iteration cutoff condition is met to obtain the final Pareto optimal solution set; Step 36: Use the TOPSIS (Technique for Order Preference by Similarity to Ideal Solution) method to integrate the entropy weight method to select the optimal solution from the frontier solutions of the Pareto optimal solution set; including the following steps: Step 361, using entropy weight method to calculate target weight; In this embodiment, the objectives are twofold: maximizing the ejector's entrainment ratio and minimizing the average sound pressure level; The target weight calculation formula is: ; in, is the entropy value of the mth target; is the normalized value of individual i on target m; N is the number of individuals; Step 362: Based on the calculated target weight, the TOPSIS method is used to select the optimal solution; 1) Calculate the positive ideal solution A1 and the negative ideal solution A2; 2) Based on the target weight, calculate the distance of each solution to the positive ideal solution and the negative ideal solution; The distance from each solution to the positive ideal solution for: ; The distance of each solution to the negative ideal solution for: ; in, Represents the fitness value of the i-th individual under the m-th goal in the Pareto optimal solution set; 3) Calculate the score of each individual and select the individual with the highest score as the optimal solution; the calculated score is: ; In the above scheme, in the multi-objective optimization problem, the Pareto optimal solution set usually contains multiple possible solutions, each solution has different trade-offs in different objective dimensions, and the search efficiency of the traditional multi-objective optimization algorithm is limited by the initial population quality and fitness evaluation strategy, which may lead to convergence to the local optimal solution. In order to improve the global optimality of the solution and the decision-making quality of the optimization process, the scheme of this embodiment adopts the TOPSIS method to fuse the entropy weight method to screen the optimal solution. The entropy weight method is used to measure the effective information amount of each target, calculate the weight of each target according to the information entropy, and ensure that the subjective weighting is not relied on in the target trade-off process, but the weight is objectively assigned according to the discrete degree of the data, so as to enhance the objectivity of the optimization process. Subsequently, the TOPSIS method is used to calculate the proximity of each Pareto frontier solution to the ideal solution based on the method of constructing an ideal solution and a negative ideal solution, and select the optimal solution. The optimal solution screening is based on the principle of approaching the ideal solution, which reduces the risk of falling into the local optimum compared with the traditional multi-objective optimization algorithm; it avoids the computational cost of directly traversing all solutions, can quickly determine the optimal solution, and improves the convergence speed and solution efficiency of the algorithm.

[0043] In this embodiment, the traditional CFD simulation has a large amount of calculation, which affects the optimization efficiency. The objective function model is constructed through a neural network model, and the trained neural network is used to quickly predict the ejection ratio and average sound pressure level of the ejector, thereby replacing CFD simulation for fitness evaluation. This method reduces the consumption of computing resources, improves the efficiency and accuracy of fitness evaluation, and thus reduces the risk of the algorithm falling into a local optimal solution. The search space coverage is improved by an improved orthogonal test method. The improved orthogonal test method is used in the scheme to enhance the parameter data and increase the data distribution diversity of the decision variables. The data set is expanded by dynamically adjusting the offset to make the search space of the optimization algorithm more uniform, thereby reducing the probability of a local optimal solution. The scheme adopts NSGA-II (non-dominated sorting genetic algorithm). This algorithm selects individuals based on non-dominated sorting, so that the Pareto frontier can be explored more evenly during the optimization process, reducing the occurrence of local optimal problems.

[0044] In order to illustrate the effect of the above method of this embodiment, a comparative test was conducted to compare the performance of the initial injector and the injector optimized by the method of this embodiment.

[0045] The effectiveness of the optimization method was verified by comparing the flow field cloud map and spectrum map of the ejector before and after optimization. The optimized ejector showed a more uniform velocity gradient in the flow field distribution, reducing the acoustic waves generated by the collision, friction and shear between the fluid microclusters, thereby significantly reducing the flow noise. At the same time, the optimized ejector achieved full expansion under the design conditions, improved the vacuum degree, enhanced the entrainment capacity of the secondary flow, and further improved the overall performance of the MED-TVC system.

[0046] Spectral analysis shows that the noise energy of the optimized ejector in the main frequency band is significantly reduced, verifying the noise control effect. The comparison of the initial and optimized ejectors in terms of velocity, turbulent kinetic energy and sound power level is shown in Figure 4 .

[0047] like Figure 4 As shown in the small figure (a), Velocity (m / s): velocity (meters / second), represents the velocity distribution of the fluid. The initial length of the premixing chamber Lcs is 45.94mm, and the optimized length Lcs' is 50.82mm. The nozzle outlet position NXP is initially 0, and the optimized position NXP' is 6.27mm; the radius Rmix is ​​initially 11.09mm, and the radius Rmix' changes to 11.68mm after optimization; The optimized ejector has an increased length of premixing chamber, an increased diameter of equal-area mixing chamber, and a reduced nozzle expansion angle. These optimization measures reduce shock wave oscillations in the mixing chamber, reduce energy losses, and thus improve ejector efficiency.

[0048] In addition, at the nozzle outlet of the initial ejector, the primary flow is in an over-expanded state, while the optimized ejector achieves full expansion, improves the vacuum degree, and enhances the entrainment capacity of the secondary flow. This is also supported by the optimized ejector velocity profile, which increases the secondary flow rate, further increases the secondary flow rate, and improves the internal energy of the ejector.

[0049] The secondary flow is entrained and flows in layers with the primary flow in the premixing chamber. However, there is a significant velocity difference between the two fluids, resulting in significant local changes in velocity and density gradients. As a result, the momentum and energy exchange between the fluid microclusters becomes more frequent and intense, accompanied by rapid friction and shear effects, forming Figure 4 The distribution characteristics of turbulent kinetic energy (TKE) are shown in the small figure (b) in Figure 1. Specifically, TKE reaches its maximum value in the interface region where the two fluids interact. Figure 4 As shown in the (b) panel, both the premixing chamber and the diffuser chamber exhibit high TKE. In cavity flow fields, such as ejectors, an increase in TKE usually leads to higher noise levels. The high TKE area of ​​the optimized ejector is significantly reduced, thereby reducing the flow noise. Figure 4This is further supported by panel (c) in which the Acoustic Power Level (APL) distribution corresponds closely to the TKE distribution. In the high TKE region, the turbulent activity inside the fluid is particularly intense, resulting in an increase in the generation of acoustic waves from collisions, friction, and shear between fluid microclusters. These acoustic waves carry energy, which is quantified by the acoustic power level. Therefore, regions with higher TKE are generally associated with higher acoustic power levels, indicating a greater contribution from noise. Therefore, based on the above analysis of these contours, a detailed explanation is provided as to why the optimized ejector has lower noise levels and better entrainment performance.

[0050] It can be seen that the optimized ejector has significant advantages in velocity distribution, turbulent kinetic energy and sound power level. After optimization, flow loss is reduced, airflow expansion state is improved, vacuum degree is increased, and flow noise is effectively reduced, making the ejector have better working efficiency and performance.

[0051] Example 2 Based on Example 1, this embodiment provides an injector optimization system based on an improved orthogonal test method, including: A model building module is configured to build an ejector flow field simulation model and an acoustic field simulation model; The objective function model building module is configured to identify decision variables through an improved orthogonal test method based on each constructed model, and to build an objective function model through a trained neural network model with the goal of maximizing the ejection ratio and minimizing the average sound pressure level; The target optimization module is configured to calculate the target function value as the fitness value based on the target function model constructed by the trained neural network model, and use the NSGA-II genetic algorithm to perform multi-objective optimization on the geometric size of the injector to obtain the optimized target size.

[0052] Furthermore, the improved orthogonal test method obtains multiple offset data corresponding to the current parameter data by dynamically adjusting the offset, expands the parameter data for data enhancement, and inputs the enhanced data into the ejector flow field simulation model and the sound field simulation model for orthogonal testing to obtain corresponding test results, which include the ejector's entrainment ratio and average sound pressure level.

[0053] Furthermore, the objective function model building module includes a decision variable identification module, which is configured to identify decision variables by an improved orthogonal test method, including the following steps: The geometric parameters of the ejector are obtained for analysis, and a plurality of different level values ​​are selected for each parameter; According to the given horizontal values, the horizontal values ​​of the geometric parameters are randomly combined to obtain the generated experimental table and conduct orthogonal experiments; Experiments were conducted using the constructed ejector flow field simulation model and acoustic field simulation model to obtain test results; A one-way ANOVA was performed on the test results to obtain the influence ratio of each geometric parameter on the entrainment ratio and the average noise sound pressure level. The parameters with an influence ratio greater than the set value were taken as decision variables, and the value range of the decision variables was determined.

[0054] It should be noted here that the various modules in this embodiment correspond one-to-one to the various steps in Example 1, and the specific implementation process is the same, which will not be repeated here.

[0055] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. The injector optimization method based on the improved orthogonal test method is characterized by: The steps include: Construct the ejector flow field simulation model and the acoustic field simulation model; Based on the constructed models, the decision variables are identified through the improved orthogonal test method, and the objective function model is constructed through the trained neural network model with the goal of maximizing the ejection ratio and minimizing the average sound pressure level. The objective function model constructed based on the trained neural network model calculates the objective function value as the fitness value, and multi-objective optimization is performed on the geometric dimensions of the injector to obtain the optimized target dimensions.

2. The injector optimization method based on the improved orthogonal test method according to claim 1, characterized in that: The improved orthogonal test method obtains multiple offset data corresponding to the current parameter data by dynamically adjusting the offset, expands the parameter data for data enhancement, and inputs the enhanced data into the ejector flow field simulation model and the sound field simulation model for orthogonal testing to obtain the corresponding test results, which include the ejector's entrainment ratio and the average sound pressure level.

3. The injector optimization method based on the improved orthogonal test method according to claim 1, characterized in that: Based on the identified decision variables, the data set is expanded as the training set through the improved orthogonal test method. The parameter values ​​of the decision variables in the training set are used as input, and the injection ratio and average sound pressure level of the ejector model obtained under the corresponding parameter values ​​are used as output to train the constructed neural network model and obtain the objective function model constructed by the neural network model.

4. The injector optimization method based on the improved orthogonal test method according to claim 1, characterized in that: The decision variable identification is carried out by using the improved orthogonal test method, which includes the following steps: Obtain the geometric parameters of the ejector for analysis, dynamically adjust the offset, and select multiple different level values ​​for each parameter; According to the given horizontal values, the horizontal values ​​of the geometric parameters are randomly combined to obtain the generated experimental table and conduct orthogonal experiments; Experiments were conducted using the constructed ejector flow field simulation model and acoustic field simulation model to obtain test results; A one-way ANOVA was performed on the test results to obtain the influence ratio of each geometric parameter on the entrainment ratio and the average noise sound pressure level. The parameters with an influence ratio greater than the set value were taken as decision variables, and the value range of the decision variables was determined.

5. The injector optimization method based on the improved orthogonal test method according to claim 4, characterized in that: The method of performing one-way variance analysis on the test results to obtain the influence ratio of each geometric parameter on the induced emission ratio and the average noise sound pressure level includes the following steps: Treat each geometric parameter as an independent group, and set the null hypothesis and alternative hypothesis; the null hypothesis is that different geometric parameters have no significant effect on the entrainment ratio and the average noise sound pressure level, and the alternative hypothesis is that at least one geometric parameter has a significant effect on the entrainment ratio or the average noise sound pressure level; Calculate the within-group sum of squares and between-group sum of squares; Calculate the within-group and between-group degrees of freedom; Based on the calculated sums of squares and degrees of freedom, calculate the mean square error to obtain the within-group mean square error and between-group mean square error; Divide the mean square error between groups by the mean square error within groups to get the statistic F; Compare the obtained statistical value with the set critical value to determine whether each group meets the null hypothesis, so as to screen out the group that meets the alternative hypothesis; For the screened groups, the data of the same geometric parameters in each group are taken out, the mean of each geometric parameter is calculated, and the influence ratio of each geometric parameter on the induced emission ratio or the average noise sound pressure level is obtained.

6. The injector optimization method based on the improved orthogonal test method according to claim 1, characterized in that: The objective function model constructed based on the trained neural network model calculates the objective function value as the fitness value, and the NSGA-II genetic algorithm is used to perform multi-objective optimization on the geometric dimensions of the injector to obtain the optimized target dimensions, including the following steps: Latin hypercube sampling is used to randomly generate N initial individuals, each of which is a combination of geometric parameters of the ejector, and the population size P and the maximum number of iterations G are set; Calculate the objective function value based on the trained neural network as the fitness value; Select individuals whose fitness values ​​are greater than the set threshold to form a new parent population; Perform crossover and mutation operations on the individuals of the parent population to obtain the offspring population; The offspring population and the parent population are merged, and the iteration cycle is repeated until the iteration cutoff condition is met to obtain the final Pareto optimal solution set.

7. The injector optimization method based on the improved orthogonal test method according to claim 6, characterized in that: The TOPSIS method is integrated with the entropy weight method to select the optimal solution from the frontier solutions of the Pareto optimal solution set, including the following steps: The entropy weight method is used to calculate the target weight; Based on the calculated objective weights, the TOPSIS method is used to select the optimal solution.

8. The injector optimization system based on the improved orthogonal test method is characterized by: include: A model building module is configured to build an ejector flow field simulation model and an acoustic field simulation model; The objective function model building module is configured to identify decision variables through an improved orthogonal test method based on each constructed model, and to build an objective function model through a trained neural network model with the goal of maximizing the ejection ratio and minimizing the average sound pressure level; The target optimization module is configured to calculate the target function value as the fitness value based on the target function model constructed by the trained neural network model, and use the NSGA-II genetic algorithm to perform multi-objective optimization on the geometric size of the injector to obtain the optimized target size.

9. The injector optimization system based on the improved orthogonal test method according to claim 8, characterized in that: The improved orthogonal test method obtains multiple offset data corresponding to the current parameter data by dynamically adjusting the offset, expands the parameter data for data enhancement, and inputs the enhanced data into the ejector flow field simulation model and the sound field simulation model for orthogonal testing to obtain the corresponding test results, which include the ejector's entrainment ratio and the average sound pressure level.

10. The injector optimization system based on the improved orthogonal test method according to claim 8, characterized in that: The objective function model building module includes a decision variable identification module, which is configured to identify decision variables by an improved orthogonal test method, including the following steps: The geometric parameters of the ejector are obtained for analysis, and a plurality of different level values ​​are selected for each parameter; According to the given horizontal values, the horizontal values ​​of the geometric parameters are randomly combined to obtain the generated experimental table and conduct orthogonal experiments; Experiments were conducted using the constructed ejector flow field simulation model and acoustic field simulation model to obtain test results; A one-way ANOVA was performed on the test results to obtain the influence ratio of each geometric parameter on the entrainment ratio and the average noise sound pressure level. The parameters with an influence ratio greater than the set value were taken as decision variables, and the value range of the decision variables was determined.

Images