Design method of a spinning draft air duct structure

Abstract

The present application relates to a kind of design methods of spinning draft air duct structure, comprising the following steps: 1), input structure parameter and fluid parameter, construct parameterization spinning draft air duct fluid dynamics simulation model, generate flow field state original data;2), the original data is preprocessed, constructs simulation model input parameter and the training data set of output result;3), using support vector regression model to establish the mapping relationship of input parameter and output result;4), based on genetic algorithm to establish multi-objective optimization model, generate offspring population;5), select the pareto front solution of new generation population to carry out fluid dynamics verification, and high error sample is added to training data set, and support vector regression model is retrained;6), repeat step 4-5), until multi-objective optimization model reaches convergence condition;7), output optimal air duct structure design scheme.The present application realizes the multi-objective optimization of air duct structure, and improves airflow uniformity.

Description

TECHNICAL FIELD

[0001] The application relates to a design method of an air duct, in particular to a design method of a spinning draft air duct structure and belongs to the technical field of digital design. BACKGROUND

[0002] The air flow uniformity of the spunbond draft air duct directly affects the quality of fiber forming. The traditional air duct design relies on experience and trial and error, and problems such as uneven air flow distribution and large pressure loss exist. The existing technology lacks a systematic optimization method, and it is difficult to balance fluid performance and structural constraints. In addition, the nonlinear coupling effect of air duct geometric parameters and fluid response is significant, and intelligent algorithms are needed to achieve efficient optimization.

[0003] Therefore, in order to solve the above problems, it is necessary to provide an innovative design method of a spinning draft air duct structure to overcome the defects in the prior art. SUMMARY

[0004] The purpose of the present application is to provide a design method of a spinning draft air duct structure, which solves the problems of uneven air flow distribution and high pressure loss in traditional air duct design through the cooperative optimization of a fluid dynamics simulation model (CFD) and a genetic algorithm.

[0005] To achieve the above purpose, the technical scheme adopted by the present application is as follows: a design method of a spinning draft air duct structure, comprising the following process steps:

[0006] 1) input the structure parameters and fluid parameters, build a parameterized spinning draft air duct fluid dynamics simulation model, and generate original data of the flow field state;

[0007] 2) pre-process the original data of step 1) to build a training data set of the simulation model input parameters and output results;

[0008] 3) use a support vector regression model to establish a mapping relationship between the input parameters and output results obtained in step 2);

[0009] 4) establish a multi-objective optimization model based on a genetic algorithm to generate a sub-population;

[0010] 5) select the Pareto frontier solution of the new generation population for fluid dynamics verification, and add high error samples to the training data set to retrain the support vector regression model of step 3);

[0011] 6) repeat steps 4-5) until the multi-objective optimization model reaches the convergence condition;

[0012] 7) output the optimal air duct structure design scheme, which includes the optimized geometric parameters, fluid parameters, performance indicators and results that meet the engineering constraint conditions.

[0013] The design method of the spinning draft air duct structure of the present application further is that: in the step 1), the input structure parameters include air duct cross section shape, number of guide vanes, guide vane plane position, inlet / outlet diameter ratio, bending section curvature radius;

[0014] The fluid parameters include inlet air speed, turbulence intensity, fluid density and viscosity;

[0015] The simulation model is constructed as a three-dimensional steady-state turbulence model, the Realizable k-ε model is selected for the turbulence model, the grid division is realized through ICEM CFD, the grid number is greater than 500,000, and the boundary conditions are the minimum distance of the guide vanes, the total volume of the air duct, and the inlet / outlet ratio;

[0016] The original data of the flow field state include air flow uniformity standard deviation σ, pressure loss coefficient ΔP and energy consumption E; wherein, the formula of the air flow uniformity standard deviation is:

[0017]

[0018] In the formula, N is the total number of air duct cross section monitoring points, Ti is the wind speed of the i th monitoring point, is the average wind speed;

[0019] E=ΔP×Q, Q is the volume flow.

[0020] The design method of the spinning draft air duct structure of the present application further is that: in the step 2), the pretreatment includes data cleaning, normalization and feature selection; first, the 3σ criterion is used to identify and eliminate abnormal samples caused by backflow or numerical instability; then, the Min-Max normalization method is used to map the input parameters of different dimensions to the [0, 1] interval; finally, based on the Pearson correlation coefficient |r|>0.7, the feature variables strongly related to air flow uniformity, pressure loss and energy consumption are screened, and redundant parameters are eliminated.

[0021] The design method of the spinning draft air duct structure of the present application further is that: in the step 2), the input of the simulation model includes geometric and fluid parameters, and the output is wind speed distribution, pressure loss and air flow uniformity, which together constitute the training data set.

[0022] The design method of the spinning draft air duct structure of the present application further is that: in the step 3), the kernel function of the support vector regression is selected as the radial basis function, and its expression is:

[0023]

[0024] Wherein, the input vector a m , a nThe wind channel geometric parameters and fluid parameters are composed, and specifically include guide vane position coordinates, inlet / outlet diameter ratio, curved section curvature radius, inlet wind speed, and turbulence intensity; nuclear bandwidth For adjusting the sensitivity of the model to the local air flow disturbance of the guide vane.

[0025] The design method of the spinning draft wind channel structure of the present application is further: the step 4) is specifically: by setting the design variable range of the number, position, curvature radius and inlet / outlet diameter ratio of the guide vane, combining initialization population, fitness function and genetic operation, the construction of the optimization model is realized;

[0026] Among them, Latin hypercube sampling is used to generate the initial population, which ensures uniform coverage of the parameter space;

[0027] The fitness function is defined as the weighted comprehensive target of the standard deviation σ of air flow uniformity, pressure loss ΔP and energy consumption E, and the specific calculation formula is:

[0028]

[0029] In the formula: α, β, γ are weight coefficients;

[0030] Genetic operation includes tournament selection, uniform crossover and Gaussian mutation, and the offspring population is generated by continuous iteration.

[0031] The design method of the spinning draft wind channel structure of the present application is further: the step 5) is specifically: the fluid dynamics verification is carried out on each generation of Pareto front solution, if the relative error of the predicted value and the actual value of the support vector regression model is > 10%, the sample is marked as high error sample and added to the training set; after every 5 iterations, the support vector regression model is updated by incremental learning strategy.

[0032] The design method of the spinning draft wind channel structure of the present application is further: the incremental learning strategy is: without retraining all samples, the new high error sample is added to the training set; only the hyperparameters C and γ of the support vector regression model are optimized for the new sample, the Bayesian optimization method is adopted, and the adjustment range is limited within ± 10%.

[0033] The design method of the spinning draft wind channel structure of the present application is further: the convergence condition of the step 6) is specifically: the change rate of the fitness function is < 1% for 3 generations in succession, or the total iteration number reaches 50 generations, the optimization is terminated.

[0034] The design method of the spinning draft wind channel structure of the present application is further: the engineering constraint condition of the step 7) includes: the minimum spacing of the guide vane is ≥ 50 mm;

[0035] The total volume of the wind channel is ≤ 0.5 m 3 ;

[0036] Inlet / outlet ratio ≤ 3:1.

[0037] Compared with the prior art, the present application has the following beneficial effects:

[0038] 1. The design method of the spinning draft air duct structure of the present application realizes multi-objective optimization of the air duct structure through the cooperation of the fluid dynamics simulation model and the genetic algorithm, solves the problems of uneven airflow distribution and high pressure loss in traditional air duct design, can be widely applied to spinning draft air duct design, and significantly improves the flow field energy efficiency.

[0039] 2. The design method of the spinning draft air duct structure of the present application adopts a support vector regression model, which can map input parameters to a high-dimensional feature space to enhance the model's fitting ability for complex nonlinear relationships and improve prediction accuracy.

[0040] 3. The design method of the spinning draft air duct structure of the present application is not only suitable for the draft air duct design of non-woven equipment, but also can be extended to multi-objective optimization scenarios of other fluid conveying systems. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 is a flow chart of the design method of the spinning draft air duct structure of the present application.

[0042] Figure 2 is a structural schematic diagram of the spinning draft air duct structure of the present application.

[0043] Figure 3 is a flow chart of the optimization model of steps 3) to 6) of the present application. DETAILED DESCRIPTION

[0044] Please refer to the drawings in the specification Figure 1 The present application is a design method of a spinning draft air duct structure, which comprises the following process steps:

[0045] 1), input structure parameters and fluid parameters, build a parameterized spinning draft air duct fluid dynamics (CFD) simulation model, and generate flow field state original data of different draft air duct structure parameters.

[0046] The input structure parameters include air duct cross-sectional shape, number of guide vanes, guide vane plane position, inlet / outlet diameter ratio, and bending segment curvature radius. Based on the ANSYS Fluent simulation platform, the above geometric parameters are bound to the modeling script as variables to directly drive the automatic generation of air duct shape and guide vane arrangement.

[0047] The fluid parameters include inlet air speed, turbulence intensity, fluid density and viscosity, which are used to set boundary conditions to ensure that the simulation results are consistent with the actual engineering working conditions.

[0048] The constructed simulation model is a three-dimensional steady-state turbulent flow model. The Realizable k-ε model is selected as the turbulent flow model, which has higher accuracy in rotating flow and strong shear flow field. Grid division is realized by ICEM CFD, and the number of grids is more than 500,000 to ensure the convergence and accuracy of calculation. The boundary conditions are the minimum distance of the guide vane, the total volume of the air duct, the inlet / outlet ratio.

[0049] The original data of the flow field state, that is, the simulation output, includes the standard deviation σ of air flow uniformity, the pressure loss coefficient ΔP and the energy consumption E. The formula of the standard deviation σ of air flow uniformity is:

[0050]

[0051] In the formula, N is the total number of monitoring points of the air duct cross section, Ti is the wind speed of the ith monitoring point, and is the average wind speed.

[0052] E=ΔP×Q, Q is the volume flow.

[0053] 2), the original data of step 1) are preprocessed to construct a training data set of simulation model input parameters and output results.

[0054] The preprocessing includes data cleaning, normalization and feature selection. First, the 3σ criterion is used to identify and eliminate abnormal samples (such as pressure sudden change points caused by backflow or numerical instability) caused by backflow or numerical instability. Then, the Min-Max normalization method is used to map the input parameters of different dimensions to the interval [0, 1]. Finally, based on the Pearson correlation coefficient |r|>0.7, the feature variables that are strongly related to air flow uniformity, pressure loss and energy consumption are selected, and redundant parameters (such as the influence of wall roughness on low Reynolds number conditions) are removed.

[0055] The input of the above simulation model includes geometric and fluid parameters, and the output is the wind speed distribution, pressure loss and air flow uniformity, which together constitute a training data set for the establishment and optimization of the subsequent proxy model.

[0056] Please continue to refer to the description of the accompanying drawings Figure 3 As shown in FIG. 3, step 2), a support vector regression (SVR) model is used to establish a mapping relationship between the input parameters and the output results obtained in step 2).

[0057] Specifically, the SVR model takes geometric and fluid parameters as input and outputs performance indicators such as air flow uniformity, pressure loss and energy consumption, thereby realizing the mapping between them.

[0058] The kernel function of the support vector regression model is selected as a radial basis function (RBF), which aims to map the input parameters to a high-dimensional feature space to enhance the fitting ability of the model to complex nonlinear relationships and improve the prediction accuracy, and the expression is:

[0059]

[0060] wherein the input vector am, an is composed of the geometric parameters and fluid parameters of the air duct, specifically including the guide vane position coordinates, the inlet / outlet diameter ratio, the curvature radius of the curved section, the inlet air speed, and the turbulence intensity; the kernel bandwidth is used to adjust the sensitivity of the model to the local airflow disturbance of the guide vane.

[0061] The hyperparameters (penalty coefficient C, kernel function bandwidth ) of the SVR model are optimized through grid search combined with 5-fold cross-validation, and the search range is C ∈ [1, 100], ∈ [0.01, 1], the final preferred C = 50, = 0.1; the validation accuracy of the model is evaluated by mean square error (MSE < 0.05) and determination coefficient (R2 > 0.92).

[0062] 4), a multi-objective optimization model is established based on genetic algorithm to generate a child population, and the specific process is as follows: by setting the design variable range of the number of guide vanes, position, curvature radius and inlet / outlet diameter ratio, combining with the initialization population, fitness function and genetic operation, the construction of the optimization model is realized. This step uses the proxy model and training data obtained in steps 1) to 3) as support to explore the design variable space and realize multi-objective optimization of the air duct structure.

[0063] wherein the initial population is generated by Latin hypercube sampling to ensure uniform coverage of the parameter space.

[0064] The fitness function is defined as the weighted comprehensive objective of the airflow uniformity standard deviation σ, pressure loss ΔP and energy consumption E, and the specific calculation formula is:

[0065]

[0066] wherein α, β, γ are weight coefficients for balancing the importance of different optimization objectives.

[0067] The genetic operation includes tournament selection, uniform crossover and Gaussian mutation, and the child population is generated by continuous iteration. The tournament strategy is used to select the top 20% of individuals as parents. Uniform crossover (probability 0.8), randomly exchange the guide vane position codes of the parent individuals. Gaussian mutation (probability 0.05), apply ±10% perturbation to the curvature radius.

[0068] 5), the Pareto front solution of the new generation is selected for fluid dynamics verification, and high error samples are added to the training data set to retrain the support vector regression model of step 3).

[0069] Specifically, the fluid dynamics verification is performed on each generation of Pareto front solution, and if the relative error between the predicted value and the actual value of the support vector regression model is >10%, the sample is marked as a high error sample and added to the training set; every 5 iterations, the support vector regression model is updated through incremental learning strategy.

[0070] The incremental learning strategy is: without retraining all samples, the new high error samples are added to the training set; only for the new samples, the hyperparameters C and γ of the support vector regression model are optimized, the Bayesian optimization method is adopted, and the adjustment range is limited within ±10% to improve the local prediction accuracy while maintaining the overall stability of the model.

[0071] 6), repeat steps 4-5), until the multi-objective optimization model reaches the convergence condition. The convergence condition is specifically: the change rate of the fitness function is <1% for 3 consecutive generations, or the total number of iterations reaches 50 generations.

[0072] 7), output the optimal air duct structure design scheme, as shown in the accompanying drawings Figure 2 The scheme includes the optimized geometric parameters, fluid parameters, performance indicators and results that meet the engineering constraint conditions. The engineering constraint conditions include: the minimum distance of the guide vane is ≥50 mm;

[0073] The total volume of the air duct is ≤0.5 m 3 ;

[0074] The inlet / outlet ratio is ≤3:1.

[0075] The specific implementation process of the present application will be described in detail in combination with specific embodiments.

[0076] The geometric parameters of the air duct model are: the air duct cross-sectional shape (tapered inlet diameter 200 mm, outlet diameter 100 mm), the number of guide vanes is 3, the plane position coordinates are (x1, y1)=(0.2L, 0.5H), (x2, y2)=(0.5L, 0.3H), (x3, y3)=(0.8L, 0.7H), where L=2 m is the length of the air duct, H=200 mm is the cross-sectional height; the bending section curvature radius R=1.5D (D=200 mm). The fluid parameters are set as inlet wind speed 20 m / s, turbulence intensity 10%, and fluid medium air (density ρ=1.225 kg / m 3, viscosity μ = 1.8 × 10−5 Pa⋅s). The k-ε turbulence model was adopted. High-precision meshes were generated by meshing. The boundary conditions were set as mass flow inlet (flow rate 20 m / s) and static pressure outlet (0 Pa gauge pressure). The simulation outputted the wind speed distribution data of 100 monitoring points in the air duct cross section, and the standard deviation σ of air flow uniformity was calculated:

[0077]

[0078] where N = 100 is the total number of monitoring points, T i is the wind speed at the i-th point, and = 18.5 m / s is the average wind speed. The total pressure drop ΔP = 420 Pa, and the energy consumption E = ΔP × Q, where the volume flow rate Q = 0.0314 m 3 / s, corresponding to a fan power of 4.8 kW.

[0079] Data cleaning, normalization, and feature selection were performed on the simulation data. The 3σ criterion was used for data cleaning to remove abnormal samples (such as monitoring points in the backflow area) with a wind speed standard deviation > 5 m / s. The Min-Max method was used for normalization to map parameters such as guide vane coordinates (x, y) and curvature radius R to the interval [0, 1]:

[0080]

[0081] where X min and X max are the minimum and maximum values of the parameters, respectively. Feature selection was based on the Pearson correlation coefficient r to select parameters strongly correlated with the optimization objectives (σ, ΔP, E) (such as inlet wind speed r = 0.85 and guide vane number r = 0.78) and remove weakly correlated parameters (such as wall roughness |r| < 0.3).

[0082] The RBF kernel function was used to construct the SVR model, and the kernel function expression is:

[0083]

[0084] The hyperparameters (penalty coefficient C and kernel bandwidth ) were optimized by grid search, with search ranges C ∈ {1, 10, 50, 100} and γ ∈ {0.01, 0.1, 1}. A set of hyperparameters was obtained through optimization. The training set and test set were divided according to the ratio 8:2, and the 5-fold cross-validation results showed that the mean square error MSE = 0.042 and the determination coefficient R2 = 0.93, with a prediction relative error < 8%.

[0085] The initial population is generated by Latin hypercube sampling with 100 individuals, parameter ranges include number of vanes n ∈ [1, 5], curvature radius R ∈ [100 mm, 400 mm], inlet / outlet diameter ratio ∈ [1:1, 1:3]. The fitness function is defined as:

[0086] .

[0087] Genetic operations include:

[0088] Selection: Tournament selection chooses the top 20% of individuals by fitness;

[0089] Crossover: Uniform crossover (probability 0.8), offspring parameters are generated by linear combination of parents:

[0090]

[0091] Mutation: Gaussian mutation (probability 0.05), perturbation is applied to parameters .

[0092] Pareto front solutions are extracted by multi-objective optimization strategies, non-dominated sorting and crowding calculation ensure solution set diversity and convergence.

[0093] CFD validation is performed on each generation of Pareto front solutions, if the SVR prediction error > 10%, the sample is added to the training set. Every 5 generations update the SVR model through incremental learning, the hyperparameter adjustment step is ±5%. The convergence condition is that the change rate of the fitness function is <1% for 3 consecutive generations, and the final iteration is performed until the convergence condition is met (F decreases from 1.82 to 1.80).

[0094] The optimal solution selects 4 vanes, position coordinates are (0.3L, 0.6H), (0.6L, 0.4H), (0.7L, 0.8H), (0.9L, 0.2H), inlet / outlet diameter ratio 1:2, curvature radius R = 300 mm.

[0095] Engineering constraint verification: minimum distance between vanes 60 mm > 50 mm, duct volume 0.48 m 3 <0.5 m 3 , inlet / outlet ratio 1:2 < 3:1.

[0096] The above detailed implementation is only the preferred embodiment of the present invention, and does not limit the present invention, any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present invention shall be included in the protection scope of the present invention.

Claims

1. A design method for a spinning drawing air duct structure, characterized in that: The process includes the following steps: 1) Input structural and fluid parameters to construct a parameterized spinning and drawing duct fluid dynamics simulation model and generate raw flow field state data; 2) Preprocess the raw data from step 1) to construct a training dataset for the input parameters and output results of the simulation model; 3) Use a support vector regression model to establish a mapping relationship between the input parameters and output results obtained in step 2); 4) Establish a multi-objective optimization model based on genetic algorithms to generate offspring populations; 5) Select the Pareto front solution of the new generation population for hydrodynamic verification, and add high error samples to the training dataset to retrain the support vector regression model in step 3); 6) Repeat steps 4-5) until the multi-objective optimization model reaches the convergence condition; 7) Output the optimal duct structure design scheme, which includes optimized geometric parameters, fluid parameters, performance indicators and results that meet engineering constraints.

2. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: In step 1), the input structural parameters include the cross-sectional shape of the air duct, the number of guide vanes, the planar position of the guide vanes, the inlet / outlet diameter ratio, and the radius of curvature of the curved section. Fluid parameters include inlet wind speed, turbulence intensity, fluid density, and viscosity; The simulation model is constructed as a three-dimensional steady-state turbulence model. The turbulence model selected is the Realizable k-ε model. The mesh generation is achieved by ICEM CFD, with a mesh size of >500,000. The boundary conditions are the minimum spacing of the guide vanes, the total volume of the air duct, and the inlet / outlet ratio. The raw flow field data includes the standard deviation of airflow uniformity σ, the pressure loss coefficient ΔP, and the energy consumption E; the formula for the standard deviation of airflow uniformity is: In the formula: N is the total number of monitoring points on the duct cross-section, and Ti is the wind speed at the i-th monitoring point. Average wind speed; E = ΔP × Q, where Q is the volumetric flow rate.

3. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: In step 2), the preprocessing includes data cleaning, normalization, and feature selection. First, the 3σ criterion is used to identify and remove abnormal samples caused by backflow or numerical instability. Then, the Min-Max normalization method is used to map input parameters of different dimensions to the [0,1] interval. Finally, based on the Pearson correlation coefficient |r|>0.7, feature variables that are strongly correlated with airflow uniformity, pressure loss, and energy consumption are screened, and redundant parameters are removed.

4. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: In step 2), the inputs to the simulation model include geometric and fluid parameters, and the outputs are wind speed distribution, pressure loss, and airflow uniformity, which together constitute the training dataset.

5. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: In step 3), the kernel function for support vector regression is the radial basis function, whose expression is: Where, input vector a m a n It consists of duct geometry and fluid parameters, specifically including the guide vane position coordinates, inlet / outlet diameter ratio, radius of curvature of the bend section, inlet wind speed, and turbulence intensity; core bandwidth. Used to adjust the model's sensitivity to local airflow disturbances caused by the deflector.

6. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: Step 4) specifically involves: by setting the design variable ranges for the number, position, radius of curvature, and inlet / outlet diameter ratio of the guide vanes, and combining the initialization population, fitness function, and genetic operations, the optimization model is constructed. Among them, Latin hypercube sampling was used to generate the initial population to ensure uniform coverage of the parameter space; The fitness function is defined as a weighted composite objective of the standard deviation of airflow uniformity σ, pressure loss ΔP, and energy consumption E, and its specific calculation formula is as follows: In the formula: α, β, γ are weighting coefficients; Genetic operations include tournament selection, uniform crossover, and Gaussian mutation, which generate offspring populations through continuous iteration.

7. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: Step 5) specifically involves: performing hydrodynamic verification on the Pareto front solution for each generation; if the relative error between the predicted value and the actual value of the support vector regression model is >10%, then the sample is marked as a high-error sample and added to the training set; after every 5 iterations, the support vector regression model is updated through an incremental learning strategy.

8. The design method of the spinning drawing air duct structure as described in claim 7, characterized in that: The incremental learning strategy is as follows: without retraining all samples, newly added high-error samples are incorporated into the training set; the hyperparameters C and γ of the support vector regression model are optimized only for the newly added samples, using the Bayesian optimization method, and the adjustment range is limited to ±10%.

9. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: The convergence condition for step 6) is specifically: the rate of change of the fitness function is less than 1% for three consecutive generations, or the optimization is terminated when the total number of iterations reaches 50 generations.

10. The design method of the spinning drawing air duct structure as described in claim 1, characterized in that: The engineering constraints in step 7) include: minimum spacing between guide vanes ≥ 50 mm; Total duct volume ≤ 0.5 m³ 3 ; The import / export ratio is ≤3:1.

Images