Vehicle wind resistance coefficient data processing method and device, terminal equipment and storage medium

Abstract

The invention discloses a data processing method and device for a whole vehicle wind resistance coefficient, terminal equipment and a storage medium. The method comprises the steps of obtaining a plurality of shape parameter design variable samples of a vehicle; designing a variable sample based on each shape parameter, and training to obtain a wind resistance coefficient prediction model; carrying out global optimization on the wind resistance coefficient prediction model by taking a prediction result of the wind resistance coefficient prediction model as a fitness function value based on a particle swarm algorithm to obtain an optimal appearance parameter design variable; and obtaining an optimal shape design scheme of the vehicle based on the optimal shape parameter design variables. According to the method, the wind resistance coefficient is predicted through the prediction model and the particle algorithm, automatic calculation of the vehicle wind resistance coefficient is achieved, the accuracy problem caused by artificial subjective analysis is avoided, and the analysis accuracy of the whole vehicle wind resistance coefficient is greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of automobile wind resistance optimization, and in particular to a data processing method, device, terminal equipment and storage medium for the wind resistance coefficient of a whole vehicle. Background Art

[0002] Accurately predicting and optimizing the aerodynamic performance of the whole vehicle, especially the size of the whole vehicle's wind resistance coefficient, during the automobile design and development stage is of great significance for achieving efficient and low-energy operation of the vehicle. At present, the main methods for predicting and optimizing the wind resistance coefficient include CFD numerical simulation and wind tunnel testing. Wind tunnel testing repeatedly changes the appearance of the car to find the best solution to reduce the wind resistance of the whole vehicle. The test preparation cycle is long, the cost is high, and it is limited by factors such as wind tunnel flow boundary conditions and brackets. The CFD numerical simulation analysis of the vehicle's external flow field can only determine whether a certain area has an impact on the wind resistance performance of the whole vehicle. Then, based on the experience of the analyst, each design variable is selected for analysis and calculation one by one. The calculation amount is large and the cycle is long. It is highly dependent on the experience of the analyst, and the design iteration speed is limited. At the same time, it is impossible to consider the interaction between the design variables, and it is difficult to achieve the comprehensive optimization of multiple design variables of the whole vehicle. Therefore, the existing vehicle wind resistance coefficient has the problem of low analysis accuracy.

[0003] Therefore, a data processing strategy for the vehicle's drag coefficient is urgently needed to solve the problem of low analysis accuracy of the vehicle's drag coefficient. Summary of the invention

[0004] The embodiments of the present invention provide a data processing method, device, terminal equipment and storage medium for the wind resistance coefficient of a whole vehicle, so as to solve the problem of low analysis accuracy of the wind resistance coefficient of the whole vehicle.

[0005] In order to solve the above problem, an embodiment of the present invention provides a method for processing data of a vehicle drag coefficient, comprising:

[0006] Obtaining several design variable samples of vehicle appearance parameters;

[0007] Based on each of the shape parameter design variable samples, a drag coefficient prediction model is obtained by training;

[0008] Based on the particle swarm algorithm, the prediction result of the drag coefficient prediction model is used as the adaptation function value to perform global optimization on the drag coefficient prediction model and obtain the optimal shape parameter design variables.

[0009] Based on the optimal appearance parameter design variables, an optimal appearance design solution for the vehicle is obtained.

[0010] As an improvement of the above solution, the step of obtaining several design variable samples of vehicle appearance parameters includes:

[0011] Obtain samples of several key design variables;

[0012] Based on each of the key design variable samples, construct a grid parameterized model corresponding to each of the key design variable samples;

[0013] Perform CFD simulation on each grid parameterized model to obtain a predicted drag coefficient corresponding to each of the key design variable samples;

[0014] Based on each key design variable sample annotated with the predicted drag coefficient, a shape parameter design variable sample is obtained.

[0015] As an improvement of the above solution, the step of obtaining several key design variable samples includes:

[0016] Obtaining several initial design variables;

[0017] Performing a single sensitivity CFD simulation calculation on each of the initial design variables to obtain a CFD simulation result corresponding to each of the initial design variables;

[0018] Each CFD simulation result is screened by using the Pearson correlation coefficient method and the mutual information value method;

[0019] Based on the successfully screened initial design variables, a sample of key design variables is generated.

[0020] As an improvement of the above scheme, the key design variable samples are generated based on the successfully screened initial design variables, including:

[0021] The successfully screened initial design variables are used as sample design factors;

[0022] The Hammersley sampling method is used to extract samples from the design space corresponding to the sample design factors, and the sample points obtained are used as the key design variable samples; the ratio between the number of sample points and the number of sample design factors is a preset value.

[0023] As an improvement of the above scheme, the screening of each CFD simulation result by the Pearson correlation coefficient method and the mutual information value method includes:

[0024] Based on the Pearson correlation coefficient method, the Pearson coefficient of each CFD simulation result is calculated;

[0025] Based on the mutual information value method, the mutual information value of each CFD simulation result is calculated;

[0026] Determine the Pearson coefficient and mutual information value of each CFD simulation result;

[0027] If the Pearson coefficient of the CFD simulation result is within the Pearson coefficient threshold range, and the mutual information value is within the mutual information value threshold range, the screening is successful;

[0028] Otherwise, the screening fails.

[0029] As an improvement of the above solution, the generation of a drag coefficient prediction model based on each of the shape parameter design variable samples includes:

[0030] According to a preset ratio, all the shape parameter design variable samples are divided into a training set and a test set;

[0031] Based on the training set and the test set, the iterative training of the XGBOOST model is repeatedly performed, and when the calculated goodness of fit error is within the goodness of fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped, and the current initial model is output as the drag coefficient prediction model.

[0032] As an improvement of the above solution, the iterative training of the XGBOOST model includes:

[0033] Based on the current model parameters, the training set is input into the XGBOOST model for training to obtain an initial model;

[0034] Inputting the test set into the initial model to predict the drag coefficient, and calculating the fit error and the root mean square error based on the prediction result;

[0035] Judge the fit error and root mean square error;

[0036] If the fit error is within the fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped and the current initial model is output as the drag coefficient prediction model;

[0037] Otherwise, based on the genetic algorithm, the current model parameters are updated, the updated model parameters are used as the parameters of the XGBOOST model in the next XGBOOST iterative training, and the iterative training of the XGBOOST model is re-executed.

[0038] Accordingly, an embodiment of the present invention further provides a data processing device for the wind resistance coefficient of a whole vehicle, comprising: a data acquisition module, a data training module, a data optimization module and a result generation module;

[0039] The data acquisition module is used to acquire several design variable samples of the vehicle's appearance parameters;

[0040] The data prediction module is used to train and obtain a drag coefficient prediction model based on each of the shape parameter design variable samples;

[0041] The data optimization module is used to perform global optimization on the drag coefficient prediction model based on a particle swarm algorithm and take the prediction result of the drag coefficient prediction model as the adaptation function value to obtain the optimal shape parameter design variables;

[0042] The result generation module is used to obtain the optimal appearance design solution of the vehicle based on the optimal appearance parameter design variables.

[0043] Correspondingly, an embodiment of the present invention further provides a computer terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, a data processing method for the whole vehicle drag coefficient as described in the present invention is implemented.

[0044] Correspondingly, an embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a data processing method for the vehicle drag coefficient as described in the present invention.

[0045] As can be seen from the above, the present invention has the following beneficial effects:

[0046] The present invention provides a data processing method for the drag coefficient of a whole vehicle, by obtaining several appearance parameter design variable samples of the vehicle; based on each of the appearance parameter design variable samples, a drag coefficient prediction model is trained; based on a particle swarm algorithm, the prediction result of the drag coefficient prediction model is used as an adaptation function value, the drag coefficient prediction model is globally optimized to obtain the optimal appearance parameter design variable; based on the optimal appearance parameter design variable, the optimal appearance design scheme of the vehicle is obtained. The present invention trains the drag coefficient prediction model through the appearance can'shu design variable samples, and sets the prediction structure of the drag coefficient prediction model as the adaptation function value through the particle algorithm, thereby performing global optimization of the drag coefficient prediction model, obtaining the optimal appearance parameter design variable, and then generating the optimal appearance design scheme. The present invention predicts the drag coefficient through the prediction model and the particle algorithm, realizes the automatic calculation of the vehicle drag coefficient, avoids the accuracy problem caused by human subjective analysis, and greatly improves the analysis accuracy of the whole vehicle drag coefficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 It is a schematic flow chart of a method for processing data of a vehicle drag coefficient provided by an embodiment of the present invention;

[0048] Figure 2 It is a schematic diagram of the structure of a data processing device for the wind resistance coefficient of a whole vehicle provided by an embodiment of the present invention;

[0049] Figure 3 This is a schematic diagram of the structure of a terminal device provided by an embodiment of the present invention;

[0050] Figure 4 It is a flow chart of optimizing XGBoost model hyperparameters using a genetic algorithm according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] Embodiment 1

[0053] See also Figure 1 , Figure 1 FIG. 1 is a flow chart of a method for processing data of a vehicle drag coefficient provided by an embodiment of the present invention. Figure 1 As shown, this embodiment includes steps 101 to 104, and each step is specifically as follows:

[0054] Step 101: Obtain several design variable samples of vehicle appearance parameters.

[0055] In this embodiment, the step of obtaining several design variable samples of vehicle appearance parameters includes:

[0056] Obtain samples of several key design variables;

[0057] Based on each of the key design variable samples, construct a grid parameterized model corresponding to each of the key design variable samples;

[0058] Perform CFD simulation on each grid parameterized model to obtain a predicted drag coefficient corresponding to each of the key design variable samples;

[0059] Based on each key design variable sample annotated with the predicted drag coefficient, a shape parameter design variable sample is obtained.

[0060] In a specific embodiment, a full-detail mesh model of the initial vehicle external flow field CFD simulation is constructed, the body shape parameters are selected as design variables and the range of variation of the design variables is determined, and in the mesh deformation software, a corresponding deformation control body is established based on the initial mesh model, and the mesh is deformed by arbitrarily adjusting the position of the control body nodes to achieve control over the definitions of each design variable, thereby constructing a mesh parameterized model.

[0061] In this embodiment, the step of obtaining several key design variable samples includes:

[0062] Obtaining several initial design variables;

[0063] Performing a single sensitivity CFD simulation calculation on each of the initial design variables to obtain a CFD simulation result corresponding to each of the initial design variables;

[0064] Each CFD simulation result is screened by using the Pearson correlation coefficient method and the mutual information value method;

[0065] Based on the successfully screened initial design variables, a sample of key design variables is generated.

[0066] In this embodiment, the generation of key design variable samples based on the successfully screened initial design variables includes:

[0067] The successfully screened initial design variables are used as sample design factors;

[0068] The Hammersley sampling method is used to extract samples from the design space corresponding to the sample design factors, and the sample points obtained are used as the key design variable samples; the ratio between the number of sample points and the number of sample design factors is a preset value.

[0069] For better explanation, the initial design variables are used as sample design factors to determine the variation range of each sample design factor, and then the design space is determined through the variation range. According to the number K of sample design factors, the Hammersley sampling method is used to extract N sample points from the design space. The number of sample points N is not less than 5 times the number K of design factors (i.e., the preset value described in the present invention). The corresponding grid model is obtained according to the sample horizontal deformation corresponding to the generated N sample points. CFD simulation calculations are performed on all grid models to obtain the drag coefficient values ​​corresponding to all sample points to mark the key design variable samples as the shape parameter design variable samples.

[0070] In a specific embodiment, the variable values ​​of each parameter in the design variable sample for the shape parameters are normalized, and for a certain set of parameters, the data are transformed to the interval [0, 1].

[0071] In this embodiment, the screening of each CFD simulation result by the Pearson correlation coefficient method and the mutual information value method includes:

[0072] Based on the Pearson correlation coefficient method, the Pearson coefficient of each CFD simulation result is calculated;

[0073] Based on the mutual information value method, the mutual information value of each CFD simulation result is calculated;

[0074] Determine the Pearson coefficient and mutual information value of each CFD simulation result;

[0075] If the Pearson coefficient of the CFD simulation result is within the Pearson coefficient threshold range, and the mutual information value is within the mutual information value threshold range, the screening is successful;

[0076] Otherwise, the screening fails.

[0077] In a specific embodiment, the threshold range of the Pearson coefficient is a numerical range greater than 0.95, and the threshold range of the information value is a numerical range greater than 0.4.

[0078] Step 102: Based on each of the shape parameter design variable samples, a drag coefficient prediction model is trained.

[0079] In this embodiment, generating a drag coefficient prediction model based on each of the shape parameter design variable samples includes:

[0080] According to a preset ratio, all the shape parameter design variable samples are divided into a training set and a test set;

[0081] Based on the training set and the test set, the iterative training of the XGBOOST model is repeatedly performed, and when the calculated goodness of fit error is within the goodness of fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped, and the current initial model is output as the drag coefficient prediction model.

[0082] In a specific embodiment, the data samples are divided into a training sample set and a test sample data set in a ratio of 7:3 (i.e., the preset ratio described in the present invention). The sample points of the training set are used to establish a vehicle drag coefficient prediction model, and the sample points of the test set are used to compare and verify the accuracy of the prediction model.

[0083] In this embodiment, the iterative training of the XGBOOST model includes:

[0084] Based on the current model parameters, the training set is input into the XGBOOST model for training to obtain an initial model;

[0085] Inputting the test set into the initial model to predict the drag coefficient, and calculating the fit error and the root mean square error based on the prediction result;

[0086] Judge the fit error and root mean square error;

[0087] If the fit error is within the fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped and the current initial model is output as the drag coefficient prediction model;

[0088] Otherwise, based on the genetic algorithm, the current model parameters are updated, the updated model parameters are used as the parameters of the XGBOOST model in the next XGBOOST iterative training, and the iterative training of the XGBOOST model is re-executed.

[0089] In a specific embodiment, a genetic algorithm is used to optimize the hyperparameters of the XGBOOST model to obtain the optimal values ​​of the number of basic classifiers, maximum tree depth, learning rate, and minimum leaf node weight in the XGBoost model parameters, so as to realize automatic optimization of model parameters and improve model accuracy. The workflow diagram is as follows: Figure 4 As shown, the specific steps include:

[0090] 1) Select the number of classifiers, maximum tree depth, learning rate and minimum leaf node weight in the XGBoost model parameters as optimization parameters;

[0091] 2) Set the population size, generally 100-1000, and set the maximum number of iterations, generally 100-500, use binary coding to encode the parameters, and randomly generate a set of initial values, which constitute the initial population of the genetic algorithm;

[0092] 3) Using the mean prediction error (MAE) of the XGBOOST model as the evaluation index, calculate the objective function value of the individual, obtain the objective function value corresponding to each individual and store it in the result list;

[0093]

[0094] 4) Calculate the fitness function value of the individual, and ensure that the result of adding the constant 0 to each objective function value is greater than 0, and store the result in the fitness value list;

[0095] 5) Select the individual with the smallest fitness function value and its corresponding fitness value, and record the results;

[0096] 6) Perform the selection operation, divide each fitness value by its sum to obtain the probability of each individual being selected in the roulette algorithm, and use the roulette algorithm to select individuals with strong fitness for reproduction;

[0097] 7) Cross-breeding to update the population, by randomly selecting the crossover point position and based on the crossover probability, which is generally taken as 0.4-0.99, select individuals for gene exchange, and use the crossover gene combination to update the corresponding position in the original population to generate a new population;

[0098] 8) Increase the diversity of the population by introducing random mutations, and randomly choose whether to mutate the gene site. If the generated random number is less than the given mutation probability, which is generally 0.001-0.1, then perform the mutation to generate a new population;

[0099] 9) Calculate the total number of generations and determine whether the maximum number of iterations has been reached. If so, the loop terminates and the individual with the best fitness is selected. After decoding this individual, the optimal parameter combination of the XGBoost model is obtained.

[0100] Step 103: Based on the particle swarm algorithm, the drag coefficient prediction model is globally optimized with the prediction result of the drag coefficient prediction model as the adaptation function value to obtain the optimal shape parameter design variables.

[0101] In this embodiment, a particle swarm algorithm (PSO) is used to perform global optimization on the established drag coefficient prediction model. The specific steps are as follows:

[0102] 1) Set the relevant parameters of the particle swarm algorithm (speed update parameters c1, c2, inertia weight ω, number of iterations maxgen, population size sizepop, individual maximum value popmax and minimum value popmin, maximum speed Vmax and minimum speed Vmin);

[0103] 2) Population initialization, randomly initialize particle positions and particle velocities;

[0104] 3) The drag coefficient predicted by the XGBoost model is used as the fitness function value, the fitness function value of each particle is calculated, the neural network is used to evaluate all population individuals, and the fitness value is calculated based on the evaluation results;

[0105] 4) Find the initial optimal extreme value, calculate the optimal fitness value and the corresponding particle position in the group, that is, the global optimal position and the individual optimal position;

[0106] 5) Speed ​​update: In each iteration, the speed of the particle is updated according to the following formula. When the speed is greater than the maximum speed, the maximum speed is taken; when the speed is less than the minimum speed, the minimum speed is taken:

[0107]

[0108] Among them, V j (t+1) is the velocity of particle j in the t+1th iteration; ω is the inertia weight, which controls the influence of the velocity of the particle in the previous iteration; c1 and c2 represent the degree to which the particle approaches the global optimal position and the individual optimal position respectively; r1 and r2 are random numbers in the range of [0,1] respectively; G is the global optimal position; P j is the individual optimal position of particle j; is the position of particle j in the tth iteration.

[0109] 6) Position update: The current position of the particle is updated using the following formula:

[0110]

[0111] 7) Update the individual optimal and global optimal positions, and use the new position to calculate the fitness function. If the current solution of particle j is better than its historical best solution, update its individual optimal position; if the current solution of particle j is better than the global optimal solution, update the global optimal position;

[0112] 8) Result collection: record the fitness function value of the current group after each iteration. These values ​​can be plotted to monitor the effect of the optimization process.

[0113] 9) The algorithm terminates and determines whether the number of iterations reaches the maximum number of iterations. If so, the loop terminates and outputs detailed information on the optimal drag coefficient and its position based on the position of the group extreme value at the end of the iteration.

[0114] Step 104: Based on the optimal appearance parameter design variables, an optimal appearance design solution for the vehicle is obtained.

[0115] In this embodiment, the optimal shape parameter design variables obtained by optimization are subjected to denormalization processing: the parameters obtained by optimization are subjected to denormalization processing, and according to the formula used for normalization, the corresponding denormalization formula is:

[0116] x=x ‘ (max(x)-min(x))+min(x).

[0117] It is understandable that when using XGBoost-type machine learning algorithms to build a prediction model for the wind resistance coefficient of the whole vehicle, only sample data needs to be calculated and collected, and the rest of the data can be predicted, which can provide guidance and data support for the optimization of the wind resistance of the whole vehicle in the early stage of styling design; by using the Pearson correlation coefficient method and the mutual information value method to comprehensively analyze the correlation of the design variables to the wind resistance coefficient of the whole vehicle, the design variables can be screened, and the characteristic parameters with high sensitivity to the target variables can be selected to improve the learning effect and calculation efficiency of the model; in view of the problem that there are many hyperparameters and it is difficult to converge in the XGBoost model, the genetic algorithm is used to optimize the hyperparameters, which can reduce the risk of overfitting while ensuring the performance of the model, making the model more stable; by performing global optimization on multiple predicted values ​​of the prediction model, the interactive influence between the design variables can be considered to obtain a comprehensive optimization solution for multiple design variables of the whole vehicle.

[0118] See also Figure 2 , Figure 22 is a schematic diagram of the structure of a data processing device for a whole vehicle drag coefficient provided by an embodiment of the present invention, comprising: a data acquisition module 201, a data training module 202, a data optimization module 203 and a result generation module 204;

[0119] The data acquisition module is used to acquire several design variable samples of the vehicle's appearance parameters;

[0120] The data prediction module is used to train and obtain a drag coefficient prediction model based on each of the shape parameter design variable samples;

[0121] The data optimization module is used to perform global optimization on the drag coefficient prediction model based on a particle swarm algorithm and take the prediction result of the drag coefficient prediction model as the adaptation function value to obtain the optimal shape parameter design variables;

[0122] The result generation module is used to obtain the optimal appearance design solution of the vehicle based on the optimal appearance parameter design variables.

[0123] It can be understood that the above-mentioned system item embodiment corresponds to the method item embodiment of the present invention, which can implement the data processing method of the whole vehicle drag coefficient provided by any one of the above-mentioned method item embodiments of the present invention.

[0124] This embodiment obtains several appearance parameter design variable samples of the vehicle; based on each of the appearance parameter design variable samples, a drag coefficient prediction model is trained; based on the particle swarm algorithm, the prediction result of the drag coefficient prediction model is used as the adaptation function value, the drag coefficient prediction model is globally optimized to obtain the optimal appearance parameter design variable; based on the optimal appearance parameter design variable, the optimal appearance design scheme of the vehicle is obtained. The present invention trains the drag coefficient prediction model through the appearance can'shu design variable samples, and sets the prediction structure of the drag coefficient prediction model as the adaptation function value through the particle algorithm, thereby performing global optimization of the drag coefficient prediction model, obtaining the optimal appearance parameter design variable, and then generating the optimal appearance design scheme. The present invention predicts the drag coefficient through the prediction model and the particle algorithm, realizes the automatic calculation of the vehicle drag coefficient, avoids the accuracy problem caused by human subjective analysis, and greatly improves the analysis accuracy of the whole vehicle drag coefficient.

[0125] Embodiment 2

[0126] See also Figure 3 , Figure 3 It is a schematic diagram of the structure of a terminal device provided in one embodiment of the present invention.

[0127] A terminal device of this embodiment includes: a processor 301, a memory 302, and a computer program stored in the memory 302 and executable on the processor 301. When the processor 301 executes the computer program, the steps of the above-mentioned data processing method for the drag coefficient of each vehicle in the embodiment are implemented, for example Figure 1 Alternatively, when the processor executes the computer program, the functions of each module in the above-mentioned device embodiments are realized, for example: Figure 2 All modules of the data processing device for the vehicle drag coefficient are shown.

[0128] In addition, an embodiment of the present invention further provides a computer-readable storage medium, which includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the data processing method for the vehicle drag coefficient as described in any of the above embodiments.

[0129] Those skilled in the art will understand that the schematic diagram is merely an example of a terminal device and does not constitute a limitation on the terminal device. The terminal device may include more or fewer components than shown in the diagram, or a combination of certain components, or different components. For example, the terminal device may also include input and output devices, network access devices, buses, etc.

[0130] The processor 301 may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor, etc. The processor 301 is the control center of the terminal device, and uses various interfaces and lines to connect various parts of the entire terminal device.

[0131] The memory 302 can be used to store the computer program and / or module. The processor 301 implements various functions of the terminal device by running or executing the computer program and / or module stored in the memory and calling the data stored in the memory 302. The memory 302 can mainly include a program storage area and a data storage area, wherein the program storage area can store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area can store data created according to the use of the mobile phone (such as audio data, a phone book, etc.), etc. In addition, the memory can include a high-speed random access memory, and can also include a non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), at least one disk storage device, a flash memory device, or other volatile solid-state storage devices.

[0132] Wherein, if the module / unit integrated in the terminal device is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Wherein, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium, etc.

[0133] It should be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the accompanying drawings of the device embodiments provided by the present invention, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines. A person of ordinary skill in the art may understand and implement it without paying any creative effort.

[0134] The above is a preferred embodiment of the present invention. It should be pointed out that a person skilled in the art can make several improvements and modifications without departing from the principle of the present invention. These improvements and modifications are also considered to be within the scope of protection of the present invention.

Claims

1. A data processing method for vehicle drag coefficient, characterized in that: include: Obtaining several design variable samples of vehicle appearance parameters; Based on each of the shape parameter design variable samples, a drag coefficient prediction model is obtained by training; Based on the particle swarm algorithm, the prediction result of the drag coefficient prediction model is used as the adaptation function value to perform global optimization on the drag coefficient prediction model and obtain the optimal shape parameter design variables. Based on the optimal appearance parameter design variables, an optimal appearance design solution for the vehicle is obtained.

2. The data processing method of vehicle drag coefficient according to claim 1, characterized in that: The step of obtaining a plurality of design variable samples of vehicle appearance parameters includes: Obtain samples of several key design variables; Based on each of the key design variable samples, construct a grid parameterized model corresponding to each of the key design variable samples; Perform CFD simulation on each grid parameterized model to obtain a predicted drag coefficient corresponding to each of the key design variable samples; Based on each key design variable sample annotated with the predicted drag coefficient, a shape parameter design variable sample is obtained.

3. The data processing method of vehicle drag coefficient according to claim 2 is characterized in that: The obtaining of several key design variable samples includes: Obtaining several initial design variables; Performing a single sensitivity CFD simulation calculation on each of the initial design variables to obtain a CFD simulation result corresponding to each of the initial design variables; Each CFD simulation result is screened by using the Pearson correlation coefficient method and the mutual information value method; Based on the successfully screened initial design variables, a sample of key design variables is generated.

4. The data processing method of vehicle drag coefficient according to claim 3 is characterized in that: The method generates a sample of key design variables based on the successfully screened initial design variables, including: The successfully screened initial design variables are used as sample design factors; The Hammersley sampling method is used to extract samples from the design space corresponding to the sample design factors, and the sample points obtained are used as the key design variable samples; the ratio between the number of sample points and the number of sample design factors is a preset value.

5. The data processing method of vehicle drag coefficient according to claim 3 is characterized in that: The screening of each CFD simulation result by the Pearson correlation coefficient method and the mutual information value method includes: Based on the Pearson correlation coefficient method, the Pearson coefficient of each CFD simulation result is calculated; Based on the mutual information value method, the mutual information value of each CFD simulation result is calculated; Determine the Pearson coefficient and mutual information value of each CFD simulation result; If the Pearson coefficient of the CFD simulation result is within the Pearson coefficient threshold range, and the mutual information value is within the mutual information value threshold range, the screening is successful; Otherwise, the screening fails.

6. The data processing method of vehicle drag coefficient according to claim 5, characterized in that: The generating of a drag coefficient prediction model based on each of the shape parameter design variable samples comprises: According to a preset ratio, all the shape parameter design variable samples are divided into a training set and a test set; Based on the training set and the test set, the iterative training of the XGBOOST model is repeatedly performed, and when the calculated goodness of fit error is within the goodness of fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped, and the current initial model is output as the drag coefficient prediction model.

7. The data processing method of vehicle drag coefficient according to claim 6, characterized in that: The iterative training of the XGBOOST model includes: Based on the current model parameters, the training set is input into the XGBOOST model for training to obtain an initial model; Inputting the test set into the initial model to predict the drag coefficient, and calculating the fit error and the root mean square error based on the prediction result; Judge the fit error and root mean square error; If the fit error is within the fit error range and the root mean square error is within the root mean square error range, the iterative training is stopped and the current initial model is output as the drag coefficient prediction model; Otherwise, based on the genetic algorithm, the current model parameters are updated, the updated model parameters are used as the parameters of the XGBOOST model in the next XGBOOST iterative training, and the iterative training of the XGBOOST model is re-executed.

8. A data processing device for vehicle drag coefficient, characterized in that: include: Data acquisition module, data training module, data optimization module and result generation module; The data acquisition module is used to acquire several design variable samples of the vehicle's appearance parameters; The data prediction module is used to train and obtain a drag coefficient prediction model based on each of the shape parameter design variable samples; The data optimization module is used to perform global optimization on the drag coefficient prediction model based on a particle swarm algorithm and take the prediction result of the drag coefficient prediction model as the adaptation function value to obtain the optimal shape parameter design variables; The result generation module is used to obtain the optimal appearance design solution of the vehicle based on the optimal appearance parameter design variables.

9. A computer terminal device, characterized in that: It comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, a data processing method for the wind resistance coefficient of a whole vehicle as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute a data processing method for the wind resistance coefficient of a whole vehicle as described in any one of claims 1 to 7.

Images