A train shape parameter-based aerodynamic force prediction method and prediction system

Abstract

The application relates to the technical field of model construction, and discloses a train shape parameter-based aerodynamic force prediction method and a prediction system, which comprises the following steps: constructing an ideal train model, constructing an ideal train shape model based on the ideal train model and a coordinate system, constructing an actual train shape model according to a nose tip height, a streamline length, a train body height, a train body width, a train track gap and the coordinate system, constructing a final train shape model based on the ideal train shape model and the actual train shape model, obtaining an incoming flow wind speed and air density in a target train running process, and obtaining a head car resistance, a lateral force and a lift under different sideslip angles through the final train shape model, the incoming flow wind speed and the air density, calculating actual values of aerodynamic force coefficients of the target train through the head car resistance, the lateral force, the lift, a head car cross-sectional area, a lateral projection area and an overhead projection area, and solving the problem that existing analysis methods cannot simultaneously perform fitting analysis on multiple shape factors.

Description

Technical Field

[0001] The present invention relates to the technical field of model construction, and in particular to an aerodynamic force prediction method and prediction system based on train shape parameters. Background Art

[0002] Under strong crosswinds, the aerodynamic performance of high-speed trains deteriorates dramatically, leading not only to a rapid increase in drag, lateral force, and lift, but even, in severe cases, to overturning. Furthermore, the train's shape, including its cross-sectional area, body length, body height, and head shape, is a significant factor influencing the aerodynamic loads on trains subjected to crosswinds. However, traditional research methods for analyzing train crosswind performance often only consider single-variable conditions, and the mechanisms by which the coupling of different characteristic variables affects train aerodynamic loads remain unclear. Therefore, it is necessary to establish a train aerodynamic prediction method that can be applied under different characteristic parameters to rapidly determine the dynamics of train aerodynamic loads under unconventional conditions. To this end, a mapping relationship between different characteristic parameters and train aerodynamic loads is proposed. Summary of the Invention

[0003] The present invention provides an aerodynamic force prediction method and prediction system based on train shape parameters to solve the problem that existing analysis methods cannot perform fitting analysis on multiple shape factors at the same time.

[0004] In order to achieve the above object, the present invention is implemented through the following technical solutions:

[0005] In a first aspect, the present invention provides a method for predicting aerodynamic forces based on train profile parameters, comprising:

[0006] A coordinate system is constructed based on the target train's shape, and the target train's nose height, streamline length, body height, body width, track clearance, and uniform cross-section body length are obtained.

[0007] An ideal train model is constructed, and an ideal train shape model is constructed based on the ideal train model and the coordinate system.

[0008] An actual train shape model is constructed according to the nose height, the streamline length, the vehicle body height, the vehicle body width, the vehicle-rail gap, and the coordinate system.

[0009] A final train shape model is constructed based on the ideal train shape model and the actual train shape model.

[0010] The cross-sectional area, lateral projection area and top projection area of ​​the lead car of the target train are obtained according to the final train shape model.

[0011] The incoming wind speed and air density of the target train during its travel are obtained, and the leading vehicle resistance, lateral force, and lift at different sideslip angles are obtained using the incoming wind speed, the air density, and the final train shape model.

[0012] The actual value of the aerodynamic coefficient of the target train is calculated by using the lead vehicle resistance, the lateral force, the lift, the cross-sectional area of ​​the lead vehicle, the lateral projection area, and the top projection area.

[0013] Optionally, the ideal train shape model is:

[0014] |y| n +|z| n =c n ;

[0015] Where y represents the y-axis coordinate point of the ideal train shape model, z represents the z-axis coordinate point of the ideal train shape model, n represents a constant variable, and n decreases uniformly from 5 to 2, and c represents the change of the streamlined head section along the ideal train shape model according to the semi-elliptical profile.

[0016] Optionally, the actual train shape model is:

[0017]

[0018]

[0019] Among them, y max i Indicates the maximum y-axis coordinate point of the actual train shape model, z max i Indicates the maximum point of the z-axis coordinate of the actual train shape model, x i The ith coordinate point on the x-axis of the actual train shape model, w represents the body width, L1 represents the streamline length, h2 represents the body height, and h represents the body height. m represents the vehicle-rail clearance, and h1 represents the nose height.

[0020] Optionally, the final train shape model is:

[0021]

[0022]

[0023]

[0024] in, Indicates the coordinate points on the positive and negative directions of the y-axis of the final train shape model. Indicates the coordinate point on the positive direction of the z-axis of the final train shape model, Indicates the coordinate point on the negative direction of the z-axis of the final train shape model, yi The i-th coordinate point on the y-axis of the ideal train shape model, z i represents the i-th coordinate point on the z-axis of the ideal train shape model; c i It represents the change of the semi-elliptical profile of the streamlined head section along the ideal train shape model on the i-th plane.

[0025] Optionally, calculating the actual value of the aerodynamic coefficient of the target train by using the lead car resistance, the lateral force, the lift, and the cross-sectional area of ​​the lead car, the lateral projection area, and the top projection area includes:

[0026] The drag coefficient is calculated by the lead vehicle resistance and the cross-sectional area of ​​the lead vehicle. The calculation formula is as follows:

[0027]

[0028] The lateral force coefficient is calculated using the lateral force and the lateral projection area, and the calculation formula is as follows:

[0029]

[0030] The lift coefficient is calculated by the lift and the top-view projected area, and the calculation formula is as follows:

[0031]

[0032] Among them, C d represents the drag coefficient, C s represents the lateral force coefficient, C l represents the lift coefficient, A d Indicates the cross-sectional area of ​​the lead vehicle, A s Represents the lateral projection area, A l Indicates the top view projection area, F d Indicates the resistance of the leading vehicle, F s Indicates the lateral force, F l represents lift, U f represents the incoming wind speed, and ρ represents the air density.

[0033] Optionally, the sideslip angle ranges from 0° to 90°.

[0034] Optionally, the lead car resistance, the lateral force and the lift are obtained by numerically simulating the final train shape model in combination with the incoming wind speed and the air density.

[0035] Optionally, the cross-sectional area of ​​the lead car, the lateral projection area and the top projection area are obtained by numerically simulating the final train shape model.

[0036] Optionally, after calculating the actual value of the aerodynamic coefficient of the target train, the method further includes:

[0037] Based on the actual values ​​of the aerodynamic coefficients of the target train, the support vector machine regression algorithm is used to predict the changes in the aerodynamic coefficients of the train under different characteristic parameter changes.

[0038] In a second aspect, an embodiment of the present application provides an aerodynamic force prediction system based on train shape parameters, including a processor and a memory.

[0039] Memory, used to store computer programs.

[0040] The processor is configured to implement any one of the method steps described in the first aspect when executing a program stored in the memory.

[0041] Beneficial effects:

[0042] The present invention provides an aerodynamic force prediction method based on train shape parameters. Based on six design variables: nose height, streamline length, vehicle body height, vehicle body width, vehicle-rail clearance, and constant-section vehicle body length, the method automatically generates the train's aerodynamic shape. A multi-parameter coupled train aerodynamic load data sample library is established, and a method for predicting the train's aerodynamic force under the influence of different train shape parameters at different sideslip angles is finally obtained. This method eliminates the workload of calculating different train shape parameters one by one, can obtain train aerodynamic force under unconventional conditions, and has guiding significance for future research on high-speed train shape design. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Flowchart of a method for predicting aerodynamic forces based on train shape parameters according to a preferred embodiment of the present invention;

[0044] Figure 2 The appearance of the ideal train and prototype train of the preferred embodiment of the present invention;

[0045] Figure 3 Streamlined head profile curves of the ideal train and prototype train of the preferred embodiment of the present invention;

[0046] Figure 4 Comparison between the function train and the prototype train of the preferred embodiment of the present invention;

[0047] Figure 5 The spatial distribution of sampling points in a preferred embodiment of the present invention;

[0048] Figure 6 Correlation analysis between the target value and the predicted value of the drag coefficient in the preferred embodiment of the present invention;

[0049] Figure 7 is the drag coefficient prediction error of the preferred embodiment of the present invention;

[0050] Figure 8 Correlation analysis between target value and predicted value of lateral force coefficient of preferred embodiment of the present invention;

[0051] Figure 9 is the lateral force coefficient prediction error of the preferred embodiment of the present invention;

[0052] Figure 10 Correlation analysis between the target value and the predicted value of the lift coefficient in the preferred embodiment of the present invention;

[0053] Figure 11 is the lift coefficient prediction error of the preferred embodiment of the present invention. DETAILED DESCRIPTION

[0054] The following is a clear and complete description of the technical solutions of the present invention. It should be understood that the embodiments described are only a portion of the embodiments of the present invention, not all of them. All other embodiments derived by persons of ordinary skill in the art based on the embodiments of the present invention without inventive effort are intended to fall within the scope of protection of the present invention.

[0055] Unless otherwise defined, the technical or scientific terms used in the present invention shall have the usual meanings understood by persons of ordinary skill in the field to which the present invention belongs. The words "first", "second" and similar terms used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. Similarly, words such as "one" or "a" do not indicate a quantity limitation, but rather indicate the existence of at least one. Words such as "connected" or "connected" are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the object being described changes, the relative positional relationship also changes accordingly.

[0056] See Figure 1-11 The present application provides an aerodynamic force prediction method based on train shape parameters, including:

[0057] A coordinate system is constructed based on the target train's shape, and the target train's nose height, streamline length, body height, body width, track clearance, and uniform cross-section body length are obtained.

[0058] An ideal train model is constructed, and an ideal train shape model is constructed based on the ideal train model and the coordinate system.

[0059] An actual train shape model is constructed according to the nose height, the streamline length, the vehicle body height, the vehicle body width, the vehicle-rail gap, and the coordinate system.

[0060] A final train shape model is constructed based on the ideal train shape model and the actual train shape model.

[0061] The cross-sectional area, lateral projection area and top projection area of ​​the lead car of the target train are obtained according to the final train shape model.

[0062] The incoming wind speed and air density of the target train during its travel are obtained, and the leading vehicle resistance, lateral force, and lift at different sideslip angles are obtained using the incoming wind speed, the air density, and the final train shape model.

[0063] The actual value of the aerodynamic coefficient of the target train is calculated by using the lead vehicle resistance, the lateral force, the lift, the cross-sectional area of ​​the lead vehicle, the lateral projection area, and the top projection area.

[0064] The embodiment can be specifically divided into three steps:

[0065] (1) Establish a method for constructing the aerodynamic shape of a train with a multi-element cross design.

[0066] A simplified train model is used, consisting of a three-car formation consisting of a lead car, a middle car, and a tail car, and is based on a high-speed train model (see Figure 2 ), solid modeling is performed based on the train's shape parameters. Six design variables are selected here: nose height, streamline length, body height, body width, track gap, and uniform cross-section body length (corresponding to variables h1, L1, h2, w, h m , L2) to construct the train model. Table 1 gives the value range of each design variable, stipulating that the length of each vehicle (L) is equal and ranges from 20m to 30m.

[0067] Table 1: Design variable value range

[0068]

[0069] Since the cross-section of the straight section of the train is the same, Figure 3 Only the contour curve of the streamlined head is given. The length of the streamlined head can be controlled by the variable L1, and the end section of the streamlined head can be stretched to a length L2 to generate a train body with a uniform cross-section. First, based on the principle of generating an ideal train model, overall control is achieved to preliminarily generate the train's outer contour. The cross-sectional profile of the ideal train is expressed by equation (1), where c varies along the streamlined head section in a semi-elliptical profile; the major axis of the ellipse equation is 2L1, and the minor axis is w; approaching the nose of the train, n decreases uniformly from 5 to 2.

[0070] |y| n +|z| n =c n (1)

[0071] Then, based on the design parameters and outline of the prototype train, a series of coordinate points on the contour curves Line 1 and Line 2 are extracted to construct the contour function equations (2) and (3). Equation (2) represents the change in the maximum half-width of the streamlined head along the longitudinal direction. Since the train is symmetrical about the x-axis, the functional expressions of the positive and negative maximum half-widths are the same. Equation (3) represents the maximum height change of the streamlined head above the nose of the train. Since the bottom of the train can be approximated as a plane, the height change below the nose of the train is expressed by the constant h1–h m To express.

[0072]

[0073]

[0074] By using the above contour curve equation to transform the train's shape, we can roughly achieve the longitudinal development of the train's shape. Next, we need to control the cross-sectional profile of the train to achieve the overall change of the train. Since the train is a symmetrical model, only half of the train's cross-sectional profile needs to be modeled. The control equations are as follows (4), (5), and (6). and Respectively represent the final cross-section coordinates after adjustment. Finally, the train shape generated by the line function is compared with the prototype train. Figure 4 .

[0075]

[0076]

[0077]

[0078] (2) Establish a multi-parameter coupled train aerodynamic load data sample library.

[0079] The optimal Latin hypercube experimental design method is used to sample the design space. A total of 50 groups of sample points are designed. The spatial distribution of the normalized data sampling points is as follows: Figure 5 As shown in Figure 2. Since the leading car of a train is subject to the largest aerodynamic force and is the most dangerous in a strong wind environment, only the leading car is analyzed here. As shown in equations (7)–(9), the aerodynamic coefficient includes the drag coefficient C d , lateral force coefficient C s and lift coefficient C l . F d is the lead vehicle resistance, F s is the lateral force, F l is the lift. d , A s and A lThe cross-sectional area, side projection area and top projection area of ​​the train model head car are generated by the function control method in step (1). f The speed is 60m / s and the air density ρ is 1.225kg / m3. The present invention selects a sideslip angle of 30° for analysis. Table 2 shows the aerodynamic coefficient results of the lead vehicle under different parameter combinations.

[0080]

[0081]

[0082]

[0083] Table 2: Aerodynamic coefficients of the lead vehicle

[0084]

[0085]

[0086] (3) A method for predicting train aerodynamic forces under the influence of different shape parameters is proposed.

[0087] First, a support vector machine (SVM) regression algorithm was used to predict the train's drag coefficient, and the Gaussian radial basis function (Equation (10)) was selected as the kernel function of the regression model. A genetic algorithm was also used to quickly and efficiently select the optimal hyperparameter combination to improve the accuracy of the prediction model. The genetic algorithm population size and maximum number of generations were both 200. An elite strategy was applied, with the number of elite individuals set to 2. A tournament strategy was used as the selection function, with the number of elite individuals set to 2. The two-point function and the uniform function were set as the crossover function and mutation function, respectively, with the crossover probability and mutation probability being 0.85 and 0.01, respectively. The remaining parameters were set to their default values.

[0088] K(x i ,x)=exp(-g||x i -x|| 2 ) (10)

[0089] 80% of the sample data generated by the experimental design was randomly selected as the training set, and the remaining 20% ​​was used as the test set. An eight-fold cross-validation method was used to improve the robustness and generalization ability of the prediction model. Using MATLAB software, the optimal parameter combination for the drag coefficient prediction model was obtained: penalty factor C = 9.9723, kernel parameter g = 0.1965, and the final mean square error (MSE) of the training process was 7.23 × 10⁻4. Figure 6 is the regression curve between the predicted value and the target value of the drag coefficient. The correlation coefficients (R2) of the training set and the test set are 0.9984 and 0.9889 respectively. Figure 7As shown in the figure, the errors between the predicted values ​​and the target values ​​of the training set and the test set are both less than 5%, which means that the SVM regression model constructed above has good prediction performance and can well predict the train resistance coefficient.

[0090] Referring to the drag coefficient prediction model construction process described above, the same method was used to construct the side force coefficient and lift coefficient prediction models. Based on the same genetic algorithm parameter settings, the optimal parameter combination for the side force coefficient prediction model was determined to be: penalty factor C = 9.9644, kernel parameter g = 0.2136, and the final mean square error (MSE) of the training process was 7.11 × 10⁻¹⁴. The optimal parameter combination for the lift coefficient prediction model was: penalty factor C = 9.5752, kernel parameter g = 0.3538, and the final mean square error (MSE) of the training process was 1.97 × 10⁻¹³. Figure 8 and Figure 10 The regression curves between the predicted values ​​and target values ​​of the side force and lift coefficient are shown in Figure 2. The correlation coefficients (R2) of the training set are all higher than 0.99, and the correlation coefficients (R2) of the test set are all higher than 0.97. Figure 9 and Figure 11 It shows that the errors between the predicted values ​​and the target values ​​of the training and test sets of the lateral force and lift coefficient are both less than 5%, which shows that the constructed prediction model can well predict the train lateral force and lift coefficient.

[0091] An embodiment of the present application also provides an aerodynamic force prediction system based on train shape parameters, including a processor and a memory.

[0092] Memory, used to store computer programs.

[0093] The processor is configured to implement any one of the method steps described in the method for predicting aerodynamic forces based on train shape parameters when executing a program stored in the memory.

[0094] The above-mentioned aerodynamic force prediction system based on train shape parameters can implement various embodiments of the above-mentioned aerodynamic force prediction method based on train shape parameters and achieve the same beneficial effects, which will not be described in detail here.

[0095] Optionally, embodiments of the present application further provide a readable storage medium storing a program or instructions that, when executed by a processor, implement the steps of the machine vision recognition method described above. This readable storage medium can implement each embodiment of the model building method described above and achieve the same beneficial effects, and will not be described in detail here.

[0096] The above describes in detail the preferred embodiments of the present invention. It should be understood that those skilled in the art can make numerous modifications and variations based on the concepts of the present invention without inventive effort. Therefore, any technical solutions that can be derived by those skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A method for predicting aerodynamic forces based on train shape parameters, characterized in that: include: Construct a coordinate system based on the target train's shape and obtain the target train's nose height, streamline length, body height, body width, track clearance, and uniform cross-section body length; Constructing an ideal train model, and constructing an ideal train shape model based on the ideal train model and the coordinate system; constructing an actual train shape model based on the nose height, the streamline length, the vehicle body height, the vehicle body width, the vehicle-rail gap, and the coordinate system; Constructing a final train shape model based on the ideal train shape model and the actual train shape model; Obtaining the cross-sectional area, lateral projection area, and top projection area of ​​the lead car of the target train according to the final train shape model; Obtaining the incoming wind speed and air density of the target train during its travel, and obtaining the leading vehicle resistance, lateral force, and lift at different sideslip angles using the incoming wind speed, the air density, and the final train shape model; Calculate the actual value of the aerodynamic coefficient of the target train by using the lead car resistance, the lateral force, the lift, the cross-sectional area of ​​the lead car, the lateral projection area, and the top projection area; Wherein, the actual train shape model is: ； ； in, Indicates the maximum y-axis coordinate point of the actual train shape model, Indicates the maximum z-axis coordinate point of the actual train shape model, Represents the i-th coordinate point on the x-axis of the actual train shape model, Indicates the width of the vehicle body. represents the streamline length, Indicates vehicle height. Indicates the track clearance, Indicates the height of the nose tip; The final train shape model is: ； ； ； in, Indicates the coordinate points on the positive and negative directions of the y-axis of the final train shape model. Indicates the coordinate point on the positive direction of the z-axis of the final train shape model, Indicates the coordinate point on the negative direction of the z-axis of the final train shape model. represents the i-th coordinate point on the y-axis of the ideal train shape model, The i-th coordinate point on the z-axis of the ideal train shape model; It represents the change of the semi-elliptical profile of the streamlined head section along the ideal train shape model on the i-th plane.

2. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: The ideal train shape model is: ； Where y represents the y-axis coordinate point of the ideal train shape model, z represents the z-axis coordinate point of the ideal train shape model, n represents a constant variable, and n decreases uniformly from 5 to 2, and c represents the change of the streamlined head section along the ideal train shape model according to the semi-elliptical profile.

3. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: The calculating the actual value of the aerodynamic coefficient of the target train by using the lead car resistance, the lateral force, the lift, the cross-sectional area of ​​the lead car, the lateral projection area, and the top projection area includes: The drag coefficient is calculated by the lead vehicle resistance and the cross-sectional area of ​​the lead vehicle. The calculation formula is as follows: ； The lateral force coefficient is calculated using the lateral force and the lateral projection area, and the calculation formula is as follows: ； The lift coefficient is calculated by the lift and the top-view projected area, and the calculation formula is as follows: ； in, represents the drag coefficient, represents the lateral force coefficient, represents the lift coefficient, represents the cross-sectional area of ​​the lead vehicle, represents the lateral projection area, represents the top-view projected area, Indicates the resistance of the leading vehicle, represents the lateral force, represents lift, Indicates the incoming wind speed, Indicates the air density.

4. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: The sideslip angle ranges from 0° to 90°.

5. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: The leading vehicle resistance, the lateral force, and the lift are obtained by numerically simulating the final train shape model in combination with the incoming wind speed and the air density.

6. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: The cross-sectional area of ​​the lead car, the lateral projection area and the top projection area are obtained by numerically simulating the final train shape model.

7. The aerodynamic force prediction method based on train shape parameters according to claim 1, characterized in that: After calculating the actual value of the aerodynamic coefficient of the target train, the method further includes: Based on the actual values ​​of the aerodynamic coefficients of the target train, the support vector machine regression algorithm is used to predict the changes in the aerodynamic coefficients of the train under different characteristic parameter changes.

8. An aerodynamic force prediction system based on train shape parameters, characterized in that: Including processor and memory; Memory for storing computer programs; A processor, configured to implement the steps of any one of the methods described in claims 1-7 when executing a program stored in a memory.

Images