A method, device and equipment for fast estimation of wind resistance coefficient and storage medium

Abstract

The embodiment of the application provides a kind of fast estimation method, device, equipment and storage medium of wind resistance coefficient, comprising: the target vehicle windward area, the test mass of the target vehicle and the target vehicle sliding resistance coefficient are obtained;The target vehicle windward area, the test mass of the target vehicle and the target vehicle sliding resistance coefficient are input into wind resistance coefficient prediction model, and the wind resistance coefficient of the target vehicle is obtained.To realize fast and high-precision estimation wind resistance coefficient.

Description

Technical Field

[0001] This application relates to the field of automotive technology, and more specifically to a method, apparatus, device, and storage medium for rapid estimation of drag coefficient. Background Technology

[0002] The drag coefficient (Cd) is a core parameter measuring the aerodynamic performance of an object moving in a fluid. It directly determines the energy efficiency, high-speed stability, and operational safety of a vehicle and is widely used in various fields such as automobiles, aerospace, drones, and virtual cycling. Especially in the field of new energy vehicles, the drag coefficient is one of the core indicators affecting the overall performance of a vehicle. For every 0.01 reduction in drag coefficient, the driving range can be increased by 1% to 2%. At a speed of 120 km / h, wind resistance accounts for as much as 70% of the total driving resistance. Low drag design has become a key path to enhance the core competitiveness of products, driving the iterative upgrade of drag coefficient estimation methods towards higher accuracy and higher efficiency.

[0003] Currently, there are two main methods for estimating drag coefficient: one is CAE (Computer Aided Engineering) simulation. CAE simulation relies on a 3D data model of the vehicle, but obtaining 3D data models of non-designed vehicle models is difficult, and flow field analysis is required, making implementation challenging. This results in low applicability of this method in the target scenario.

[0004] Another method is wind tunnel testing. This method achieves accurate measurements by simulating real airflow in a controlled environment. However, this method requires specialized equipment to build the simulated airflow environment, which is costly and has a long testing cycle, making it unsuitable for the rapid iteration needs of the automotive design phase.

[0005] Existing methods for estimating drag coefficients suffer from high costs, long cycles, and low practicality. There is an urgent need for a drag coefficient estimation method that can overcome existing technological bottlenecks and balance efficiency and accuracy. Summary of the Invention

[0006] In view of this, this application provides a method, apparatus, device and storage medium for rapid estimation of drag coefficient, so as to achieve rapid and high-precision estimation of drag coefficient.

[0007] In a first aspect, embodiments of this application provide a method for rapidly estimating the drag coefficient, including: obtaining the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle; The windward area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle are input into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

[0008] In one possible implementation of the first aspect, obtaining the frontal area of ​​the target vehicle includes: Acquire a projection image of the target vehicle on a target plane; wherein the target plane is a plane perpendicular to a first direction; the first direction is the driving direction of the target vehicle; Obtain the height and width information of the target vehicle; Based on the height information, width information, and projection contour information of the target vehicle on the target plane, the frontal area of ​​the target vehicle is determined.

[0009] In one possible implementation of the first aspect, acquiring the projection image of the target vehicle on the target plane includes: The front of the target vehicle is scanned along a second direction to obtain a projection image of the target vehicle on the target plane; the second direction is perpendicular to the first direction.

[0010] In one possible implementation of the first aspect, acquiring the projection image of the target vehicle on the target plane includes: The target vehicle or its model is scanned to obtain point cloud data of the target vehicle. The point cloud data of the target vehicle is orthographically projected along a first direction to obtain a projected image of the target vehicle on the target plane.

[0011] One possible implementation of the first aspect also includes: The frontal area, test mass, gliding drag coefficient, and drag coefficient of multiple historical vehicles are obtained; wherein the historical vehicles are of the same model as the target vehicle. A training sample set is formed based on the frontal area, test mass, gliding drag coefficient, and wind resistance coefficient of the aforementioned historical vehicles. The preset network model is trained based on the training sample set to obtain the drag coefficient prediction model.

[0012] One possible implementation of the first aspect also includes: Based on the frontal area, test mass, gliding drag coefficient, and wind resistance coefficient of the aforementioned historical vehicles, a validation sample set is formed; the validation sample set is different from the training sample set. The step of training the preset network model based on the training sample set to obtain the drag coefficient prediction model includes: The preset network model is trained based on the training sample set to obtain the initial drag coefficient prediction model; The initial drag coefficient prediction model is evaluated based on the validation sample set. When the initial drag coefficient prediction model passes the evaluation, the initial drag coefficient prediction model is determined as the drag coefficient prediction model.

[0013] One possible implementation of the first aspect also includes: If the initial drag coefficient prediction model fails the evaluation, the training sample set is updated, and the steps "training the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluating the initial drag coefficient prediction model based on the validation sample set" are re-executed based on the updated training sample set until the initial drag coefficient prediction model passes the evaluation.

[0014] Secondly, embodiments of this application provide a device for rapidly estimating the drag coefficient, comprising: The acquisition module is used to acquire the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle. The processing module is used to input the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

[0015] Thirdly, embodiments of this application provide an electronic device, including a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the electronic device is triggered to execute the method described in any of the first aspects above.

[0016] Fourthly, embodiments of this application provide a computer-readable storage medium including a stored program, wherein, when the program is executed, it controls the device where the computer-readable storage medium is located to perform the method described in any of the first aspects above.

[0017] The solution provided in this application embodiment obtains the frontal area, test mass, and drag coefficient of the target vehicle. These parameters are then input into a drag coefficient prediction model to obtain the drag coefficient of the target vehicle. In this way, the drag coefficient prediction model can be pre-trained in this application embodiment. Therefore, when the drag coefficient of a target vehicle needs to be obtained, the frontal area, test mass, and drag coefficient can be acquired, and the drag coefficient prediction model can be used to quickly estimate the drag coefficient. In other words, this application embodiment can use the drag coefficient prediction model to quickly predict the drag coefficient of a target vehicle, achieving rapid and high-precision estimation of the drag coefficient without the need for wind tunnel testing, thus reducing costs. Attached Figure Description

[0018] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating a rapid method for estimating drag coefficient provided in this application embodiment; Figure 2 A schematic diagram illustrating an example of a projection image of a target vehicle on a target plane, provided as an embodiment of this application; Figure 3 This application provides an example schematic diagram of scanning a target vehicle in a second direction according to an embodiment of the present application. Figure 4 This application provides another example schematic diagram of scanning a target vehicle in a second direction according to an embodiment of the present application; Figure 5 A flowchart illustrating another method for rapid estimation of drag coefficient provided in this application embodiment; Figure 6 A schematic diagram illustrating model training provided in an embodiment of this application; Figure 7 A schematic diagram of a rapid drag coefficient estimation device provided in an embodiment of this application; Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0020] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0021] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0022] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0023] In related technologies, there are two main methods for estimating the drag coefficient: one is CAE (Computer Aided Engineering) simulation. CAE simulation relies on a three-dimensional data model of the vehicle, but obtaining a three-dimensional data model of a non-designed vehicle model is difficult, and flow field analysis is required, which is quite challenging, resulting in low applicability of this method in the target scenario.

[0024] Another method is wind tunnel testing. This method achieves accurate measurements by simulating real airflow in a controlled environment. However, this method requires specialized equipment to build the simulated airflow environment, which is costly and has a long testing cycle, making it unsuitable for the rapid iteration needs of the automotive design phase.

[0025] Existing methods for estimating drag coefficients suffer from high costs, long cycles, and low practicality. There is an urgent need for a drag coefficient estimation method that can overcome existing technological bottlenecks and balance efficiency and accuracy.

[0026] To address the aforementioned problems, this application provides a rapid method for estimating the drag coefficient. The method includes: obtaining the frontal area, test mass, and skid drag coefficient of the target vehicle; inputting the frontal area, test mass, and skid drag coefficient of the target vehicle into a drag coefficient prediction model to obtain the drag coefficient of the target vehicle. In this way, the drag coefficient prediction model can be pre-trained in this embodiment. Therefore, when the drag coefficient of a target vehicle needs to be obtained, the frontal area, test mass, and skid drag coefficient of the target vehicle can be acquired, and the drag coefficient prediction model can be used to quickly estimate the drag coefficient of the target vehicle. In other words, this embodiment can use the drag coefficient prediction model to quickly predict the drag coefficient of the target vehicle, achieving rapid and high-precision estimation of the drag coefficient without wind tunnel testing, thus reducing costs. A detailed explanation follows.

[0027] See Figure 1 This is a flowchart illustrating a rapid method for estimating the drag coefficient provided in an embodiment of this application. Figure 1 As shown, the method includes: Step S101: Obtain the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle.

[0028] In this embodiment, based on the principle of determining vehicle resistance using the coasting method, the vehicle is only subject to rolling resistance Fr and air resistance Fa during coasting on a level road. Therefore, the total vehicle resistance can be expressed by the following formula 1: 1 in, Indicates the vehicle's weight. This represents the rolling resistance coefficient of the vehicle. The value represents the drag coefficient, and A represents the vehicle's frontal area. This represents air density, which at sea level and standard atmospheric pressure is approximately 1.225 kg / m³ (kilograms per cubic meter). This indicates the vehicle's speed.

[0029] As shown in Formula 1, the drag coefficient is mainly related to the vehicle's frontal area A, the vehicle's test mass m, and the rolling resistance coefficient. related.

[0030] Typically, total driving resistance can be expressed as the vehicle's velocity. Since it is a function of the vehicle's total resistance, the formula for the total driving resistance can be simplified to Formula 2: 2 a, b, and c are all sliding resistance coefficients. The values ​​of a, b, and c are constants and can be measured through sliding resistance tests. The rolling resistance coefficient in Formula 1 above... Related to the sliding resistance coefficients a, b, and c, and through the rolling resistance coefficient The regression equation shows that the frontal area A, the test mass m of the vehicle, and the drag coefficients a, b, and c are related to the drag coefficient.

[0031] Therefore, when it is necessary to determine the drag coefficient of a target vehicle, the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle can be obtained.

[0032] In some embodiments, the frontal area of ​​a target vehicle can be obtained using the laser point cloud-pixel method. That is, obtaining the frontal area of ​​the target vehicle includes: Obtain the projected image of the target vehicle on the target plane. Obtain the height and width information of the target vehicle. Based on the height, width, and projected outline information of the target vehicle on the target plane, determine the frontal area of ​​the target vehicle.

[0033] The target plane is a plane perpendicular to the first direction; the first direction is the direction of travel of the target vehicle.

[0034] In this embodiment, the frontal area of ​​the target vehicle can be the projected area of ​​its front view; therefore, a front view image of the target vehicle can be obtained first. That is, the projected image of the target vehicle on the target plane can be obtained, such as... Figure 2 As shown, the target plane is a plane perpendicular to the vehicle's direction of travel. The height H and width W of the target vehicle can be extracted from the vehicle's announcement information, for reference... Figure 2 As shown. Thus, the theoretical frontal area S of the target vehicle can be calculated based on its height H and width W. For example, the theoretical frontal area S of the target vehicle can be calculated using the following formula 3.

[0035] S = H × W 3 After obtaining the projected image of the target vehicle on the target plane and the theoretical frontal area S of the target vehicle, the total number of pixels P of the vehicle body region in the projected image can be calculated using image processing techniques based on the projected image of the target vehicle on the target plane. 车身 And the total number of pixels P, which is the theoretical total area of ​​the front of the target vehicle. 总 For example, refer to Figure 2 As shown, the black area represents the body area of ​​the target vehicle, from which statistics can be calculated. Figure 2 The total number of pixels in the black area of ​​the image shown is the total number of pixels P in the vehicle body area. 车身 , Figure 2 The total number of pixels within the box shown is P, which represents the theoretical total area of ​​the target vehicle's front side. 总 Based on the principle that "pixel ratio = area ratio", the frontal area of ​​the target vehicle can be calculated. Specifically, the frontal area of ​​the target vehicle can be calculated using Formula 4 below.

[0036] 4 In some embodiments, to obtain a more rapid and accurate projection image of the target vehicle on the target plane, the projection image of the target vehicle on the target plane can be obtained by laser scanning. That is, obtaining a projection image of the target vehicle on the target plane includes: scanning the front of the target vehicle along a second direction to obtain a projection image of the target vehicle on the target plane.

[0037] The second direction is perpendicular to the first direction. The second direction is parallel to the target plane.

[0038] That is, a projected image of the target vehicle on the target plane can be obtained through laser scanning. To obtain the projected image on the target plane more quickly, the laser scanning device can scan the target vehicle along a second direction, as shown in the reference... Figure 3As shown. For example, a laser scanning device can be installed at the front of the target vehicle, so that a projected image of the target vehicle on the target plane can be formed by receiving the reflected laser signal. In some embodiments, for faster implementation, a laser transmitting device can be installed at the front of the target vehicle, and a laser receiving device can be installed at the parking space of the target vehicle, as shown in the reference. Figure 4 As shown. In this way, the projected image of the target vehicle on the target plane can be directly acquired at the laser receiving device. This method allows scanning of the target vehicle only in the second direction to directly obtain its projected image on the target plane, eliminating the need to acquire scan data of the entire vehicle. This is simple, fast, and improves the efficiency of projected image acquisition.

[0039] In other embodiments, obtaining a projection image of the target vehicle on the target plane includes: scanning the target vehicle or a model of the target vehicle to obtain point cloud data of the target vehicle; and orthographically projecting the point cloud data of the target vehicle along a first direction to obtain a projection image of the target vehicle on the target plane.

[0040] In other words, to obtain a projected image of a target vehicle on a target plane, the target vehicle or its model can be scanned to acquire point cloud data of the target vehicle, i.e., to acquire point cloud data of the complete target vehicle model. That is, the point cloud data of the target vehicle is the complete point cloud data of that target vehicle. To obtain a frontal view image of the target vehicle, the acquired point cloud data can be orthographically projected along a first direction to obtain a projected image of the target vehicle on the target plane. Thus, by obtaining the complete point cloud data of the target vehicle and projecting it along the first direction to obtain a projected image of the target vehicle on the target plane, the obtained projected image of the target vehicle on the target plane can be more accurate, reducing the possibility of detail loss in the projected image of the target vehicle on the target plane.

[0041] In some embodiments, the test mass m of the target vehicle can be obtained through a weighing test.

[0042] In some embodiments, the gliding resistance coefficient α can be obtained by fitting free-gliding test data.

[0043] In some embodiments, the sliding resistance coefficient b can be obtained by fitting free-sliding test data.

[0044] In some embodiments, the sliding resistance coefficient c can be obtained by fitting free-sliding test data.

[0045] In other embodiments, after obtaining the drag coefficients a, b, and c from free-gliding test data of other vehicles of the same model as the target vehicle, these drag coefficients can be stored. Thus, when estimating the drag coefficient of the target vehicle, the drag coefficients a, b, and c can be retrieved from the storage device.

[0046] In other embodiments, the coasting drag coefficients a, b, and c can be obtained from a cloud server. That is, the cloud server can store the coasting drag coefficients a, b, and c corresponding to different vehicle models. When it is necessary to obtain the coasting drag coefficients a, b, and c of a target vehicle, the coasting drag coefficients a, b, and c corresponding to the target vehicle model can be obtained from the cloud server.

[0047] In other embodiments, at least one of the gliding drag coefficients a, b, and c can be obtained in a storage device or a cloud server. Other gliding drag coefficients are obtained by fitting free-gliding test data.

[0048] Of course, the sliding resistance coefficients a, b, c, or the vehicle test mass m can also be obtained through other means. This application embodiment does not limit this.

[0049] Step S102: Input the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

[0050] In this embodiment, to obtain the drag coefficient of the target vehicle more quickly and accurately, a drag coefficient prediction model can be pre-trained. The drag coefficient prediction model is used to predict the drag coefficient of a specific vehicle model. It should be understood that the vehicle model for which the drag coefficient prediction model can predict the drag coefficient is related to the training data of the drag coefficient prediction model. To obtain the drag coefficient of the target vehicle, the training data used when training the drag coefficient prediction model includes training data corresponding to the vehicle model of the target vehicle. After obtaining the frontal area, test mass, and sliding drag coefficient of the target vehicle, these parameters can be used as input values ​​to the drag coefficient prediction model. The drag coefficient prediction model can analyze and extract features from the frontal area, test mass, and sliding drag coefficient of the target vehicle, and output the drag coefficient of the target vehicle.

[0051] Thus, in this embodiment, a drag coefficient prediction model can be pre-trained. When the drag coefficient of a target vehicle needs to be obtained, the vehicle's frontal area, test mass, and skidding drag coefficient can be acquired. Using the drag coefficient prediction model, the drag coefficient of the target vehicle can be quickly estimated. In other words, in this embodiment, the drag coefficient prediction model can be used to quickly predict the drag coefficient of a target vehicle, achieving rapid and high-precision estimation without the need for wind tunnel testing, thereby reducing costs.

[0052] In some embodiments, a preset network model can be trained to obtain a drag coefficient prediction model. That is, a reference... Figure 5 As shown, the above method also includes: Step S103: Obtain the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of multiple historical vehicles.

[0053] Among them, the historical vehicles and the target vehicles are of the same model.

[0054] In this embodiment of the application, in order to obtain the drag coefficient prediction model, it is necessary to train a preset network model. Based on this, it is necessary to obtain the training dataset of the preset network model. At this time, the drag coefficient, frontal area, test mass, and skidding drag coefficient of historical vehicles of the same model as the target vehicle can be obtained.

[0055] In some embodiments, accurate data of multiple historical vehicles can be collected through the vehicle publication information of historical vehicles, including the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of the historical vehicles.

[0056] In other embodiments, the frontal area of ​​a historical vehicle can be obtained using the laser point cloud-pixel method, the test mass of a historical vehicle can be obtained through a weighing test, the sliding drag coefficients a, b, and c can be fitted through a sliding test, and the drag coefficient of a historical vehicle can be obtained through a wind tunnel test.

[0057] The specific methods for obtaining the windward area of ​​historical vehicles using the laser point cloud-pixel method, obtaining the test mass of historical vehicles through weighing tests, and fitting the sliding resistance coefficients a, b, and c through sliding tests can refer to the specific implementation methods of the windward area, test mass, and sliding resistance coefficients of the target vehicle mentioned above. These methods will not be elaborated further in this application.

[0058] Step S104: Based on the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of multiple historical vehicles, a training sample set is formed.

[0059] In this embodiment of the application, after obtaining the frontal area, test mass, gliding drag coefficient and wind resistance coefficient of multiple historical vehicles, a training sample set can be directly constructed using the frontal area, test mass, gliding drag coefficient and wind resistance coefficient of multiple historical vehicles.

[0060] In some embodiments, the frontal area, test mass, skidding drag coefficient, and drag coefficient of all historical vehicles collected can be directly used as the training sample set.

[0061] In other embodiments, the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of some historical vehicles can also be directly used as the training sample set.

[0062] Step S105: Train the preset network model based on the training sample set to obtain the wind resistance coefficient prediction model.

[0063] In other words, after obtaining the training sample set, the preset network model can be trained using the training sample set. When training the preset network model, the vehicle's frontal area, test mass, and gliding drag coefficient are used as inputs, and the drag coefficient is used as the output. At this time, the frontal area, test mass, and gliding drag coefficient of historical vehicles in the training sample set can be used as input values ​​to the preset network model. The preset network model can perform feature extraction and analysis on the input frontal area, test mass, and gliding drag coefficient of historical vehicles, output its predicted drag coefficient, and compare the difference between this drag coefficient and the corresponding drag coefficient in the training sample set. When the drag coefficient output by the preset network model differs significantly from the drag coefficient corresponding to the training sample set, it indicates that the output result of the preset network model is inaccurate. In this case, the parameters of the preset network model can be adjusted, and after adjusting the parameters, the preset network model can be retrained using the training sample set. This process is repeated until the difference between the drag coefficient output by the preset network model and the drag coefficient corresponding to the training sample set is small. For example, if the difference between the drag coefficient output by the preset network model and the drag coefficient corresponding to the training sample set is no more than 3%, then the output result of the preset network model can be considered relatively accurate, and the preset network model training is complete. The trained preset network model can then be used as the drag coefficient prediction model to obtain the drag coefficient prediction model.

[0064] In some embodiments, the preset network model can be a three-layer BP (Back Propagation) neural network model. In other embodiments, the preset network model can also be a model using the random forest algorithm; in this case, the algorithm flow of the preset network model can be referred to... Figure 6As shown, the process includes: inputting a training sample set into a preset network model; the preset network model performing random sampling and feature selection on the training sample set to construct N decision trees, for example, 100 decision trees. Each decision tree independently learns the input-output mapping relationship to predict the result. The average of the prediction results from all decision trees is taken as the predicted drag coefficient, which is then output by the output layer.

[0065] Of course, the preset network model can also be other neural network models, and this application embodiment does not limit this.

[0066] As one possible implementation, refer to Figure 5 As shown, the above method also includes: Step S106: Based on the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of multiple historical vehicles, a verification sample set is formed.

[0067] The validation sample set is different from the training sample set.

[0068] To ensure the stability of the drag coefficient prediction model, after training the prediction network model, the trained model can be validated using a validation set. This validation set can be formed based on the frontal area, test mass, coasting drag coefficient, and drag coefficient of multiple historical vehicles. The validation set differs from the training set. Specifically, from the frontal area, test mass, coasting drag coefficient, and drag coefficient of multiple historical vehicles, a portion of these parameters can be used to form the training set, while the remaining parameters form the validation set.

[0069] At this point, step S105 above trains the preset network model based on the training sample set to obtain the drag coefficient prediction model, which includes: S1051. Train the preset network model based on the training sample set to obtain the initial drag coefficient prediction model.

[0070] S1052. Evaluate the initial drag coefficient prediction model based on the validation sample set.

[0071] S1053. When the initial drag coefficient prediction model passes the evaluation, the initial drag coefficient prediction model shall be determined as the drag coefficient prediction model.

[0072] In other words, after constructing the validation and training sample sets, the preset network model can be trained using the training sample set, as detailed in step S105 above. After training the preset network model using the training sample set, the trained model can be used as the initial drag coefficient prediction model. At this point, the initial drag coefficient prediction model can be evaluated using the validation sample set. Specifically, the frontal area, test mass, and skidding drag coefficient of historical vehicles in the validation sample set can be used as input values ​​to the initial drag coefficient prediction model. The initial drag coefficient prediction model can extract and analyze features from the input values ​​and output the predicted drag coefficient. The drag coefficient output by the initial drag coefficient prediction model is compared with the corresponding drag coefficient in the validation sample set to check whether the difference between the two is within a preset range, such as checking whether the difference is within 3%. If the initial drag coefficient prediction model is found to be stable and its output is accurate, then the initial drag coefficient prediction model can be considered to have passed evaluation. Once the initial drag coefficient prediction model has passed evaluation, it can be officially recognized as the drag coefficient prediction model.

[0073] In other embodiments, if the initial drag coefficient prediction model fails the evaluation, the training sample set needs to be adjusted, and the preset network model needs to be retrained. (See reference...) Figure 5 As shown, the above method also includes: S1054. If the initial drag coefficient prediction model fails the evaluation, the training sample set is updated, and the steps "train the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluate the initial drag coefficient prediction model based on the validation sample set" are re-executed based on the updated training sample set until the initial drag coefficient prediction model passes the evaluation.

[0074] In other words, if the initial drag coefficient prediction model fails the evaluation, it indicates that the output of the initial drag coefficient prediction model is unstable and needs to be retrained. At this point, the training sample set can be updated. For example, samples with obvious errors in the original training sample set can be deleted, or new historical vehicle frontal area, test mass, skidding drag coefficient, and drag coefficient can be added to update the training sample set. The preset network model is then retrained based on the updated training sample set. This involves re-executing the steps "training the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluating the initial drag coefficient prediction model based on the validation sample set" until the initial drag coefficient prediction model passes the evaluation. In other words, after updating the training sample set, the preset network model can be retrained based on the updated training sample set to obtain the initial drag coefficient prediction model. After obtaining the initial drag coefficient prediction model, it can be evaluated using a validation sample set. If the evaluation fails, the training sample set needs to be updated again, and the steps "training the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluating the initial drag coefficient prediction model based on the validation sample set" need to be repeated until the initial drag coefficient prediction model passes the evaluation. The initial drag coefficient prediction model that passes the evaluation is then determined as the drag coefficient prediction model.

[0075] Thus, in this embodiment, by pre-obtaining a drag coefficient prediction model, the drag coefficient of the target vehicle can be predicted without wind tunnel testing. This avoids the high cost and long cycle of wind tunnel testing, reducing the cost and time required for drag coefficient estimation and meeting the need for rapid drag coefficient assessment in the early stages of development. Furthermore, when estimating the drag coefficient, the frontal area, test mass, and skidding drag coefficient of the target vehicle are obtained. Data acquisition is relatively easy, with high applicability and ease of use, further reducing the complexity of drag coefficient estimation and improving its efficiency.

[0076] refer to Figure 7 This is a schematic diagram of a device for rapidly estimating the drag coefficient provided in an embodiment of this application. Figure 7 As shown, the device includes: The acquisition module 701 is used to acquire the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle.

[0077] The processing module 702 is used to input the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

[0078] In some embodiments, the acquisition module 701 is specifically used to acquire a projection image of the target vehicle on a target plane. It acquires the height and width information of the target vehicle; based on the height, width, and projection contour information of the target vehicle on the target plane, it determines the frontal area of ​​the target vehicle.

[0079] The target plane is a plane perpendicular to the first direction; the first direction is the direction of travel of the target vehicle.

[0080] In some embodiments, the acquisition module 701 is specifically used to scan the front of the target vehicle along the second direction to obtain a projection image of the target vehicle on the target plane.

[0081] The second direction is perpendicular to the first direction.

[0082] In some embodiments, the acquisition module 701 is specifically used to scan the target vehicle or a model of the target vehicle to acquire point cloud data of the target vehicle; and to orthographically project the point cloud data of the target vehicle along a first direction to obtain a projected image of the target vehicle on the target plane.

[0083] In some embodiments, the processing module 702 is further configured to acquire the frontal area, test mass, skidding drag coefficient, and drag coefficient of multiple historical vehicles. A training sample set is formed based on the frontal area, test mass, skidding drag coefficient, and drag coefficient of the multiple historical vehicles. A preset network model is trained based on the training sample set to obtain a drag coefficient prediction model.

[0084] Among them, the historical vehicles and the target vehicles are of the same model.

[0085] In some embodiments, the processing module 702 is further configured to form a validation sample set based on the frontal area, test mass, skidding drag coefficient, and wind resistance coefficient of multiple historical vehicles. The validation sample set differs from the training sample set.

[0086] The processing module 702 is specifically used to train a preset network model based on a training sample set to obtain an initial drag coefficient prediction model. The initial drag coefficient prediction model is then evaluated based on a validation sample set. If the initial drag coefficient prediction model passes the evaluation, it is determined as the drag coefficient prediction model.

[0087] The processing module 702 is also used to update the training sample set when the initial drag coefficient prediction model fails the evaluation, and to re-execute the steps "training the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluating the initial drag coefficient prediction model based on the validation sample set" according to the updated training sample set, until the initial drag coefficient prediction model passes the evaluation.

[0088] Corresponding to the above embodiments, this application also provides an electronic device. Figure 8 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. The electronic device 800 may include a processor 801, a memory 802, and a communication unit 803. These components communicate through one or more buses. Those skilled in the art will understand that the structure of the server shown in the figure does not constitute a limitation on the embodiment of the present invention. It may be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0089] The communication unit 803 is used to establish a communication channel, enabling the storage device to communicate with other devices. It can receive user data sent by other devices or send user data to other devices.

[0090] The processor 801 serves as the control center of the storage device, connecting various parts of the electronic device via various interfaces and lines. It executes software programs and / or modules stored in the memory 802, and calls data stored in the memory to perform various functions of the electronic device and / or process data. The processor can be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 801 may consist only of a central processing unit (CPU). In this embodiment of the invention, the CPU may have a single processing core or include multiple processing cores.

[0091] The memory 802 is used to store the execution instructions of the processor 801. The memory 802 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0092] When the execution instructions in memory 802 are executed by processor 801, the electronic device 800 is able to perform some or all of the steps in the above embodiments.

[0093] In a specific implementation, the present invention also provides a computer storage medium, wherein the computer storage medium may store a program, which, when executed, may include some or all of the steps of the various embodiments of the rapid estimation method for the drag coefficient provided by the present invention. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0094] Those skilled in the art will clearly understand that the techniques in the embodiments of the present invention can be implemented using software plus necessary general-purpose hardware platforms. Based on this understanding, the technical solutions in the embodiments of the present invention, or the parts that contribute to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of the present invention.

[0095] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments and terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

Claims

1. A rapid method for estimating the drag coefficient, characterized in that, include: Obtain the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle; The windward area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle are input into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

2. The method according to claim 1, characterized in that, The acquisition of the frontal area of ​​the target vehicle includes: Acquire a projection image of the target vehicle on a target plane; wherein the target plane is a plane perpendicular to a first direction; the first direction is the driving direction of the target vehicle; Obtain the height and width information of the target vehicle; Based on the height information, width information, and projection contour information of the target vehicle on the target plane, the frontal area of ​​the target vehicle is determined.

3. The method according to claim 2, characterized in that, The process of acquiring the projection image of the target vehicle on the target plane includes: The front of the target vehicle is scanned along a second direction to obtain a projection image of the target vehicle on the target plane; the second direction is perpendicular to the first direction.

4. The method according to claim 2, characterized in that, The process of acquiring the projection image of the target vehicle on the target plane includes: The target vehicle or its model is scanned to obtain point cloud data of the target vehicle. The point cloud data of the target vehicle is orthographically projected along a first direction to obtain a projected image of the target vehicle on the target plane.

5. The method according to claim 1, characterized in that, Also includes: The frontal area, test mass, gliding drag coefficient, and drag coefficient of multiple historical vehicles are obtained; wherein the historical vehicles are of the same model as the target vehicle. A training sample set is formed based on the frontal area, test mass, gliding drag coefficient, and wind resistance coefficient of the aforementioned historical vehicles. The preset network model is trained based on the training sample set to obtain the drag coefficient prediction model.

6. The method according to claim 5, characterized in that, Also includes: Based on the frontal area, test mass, gliding drag coefficient, and wind resistance coefficient of the aforementioned historical vehicles, a verification sample set is formed. The verification sample set is different from the training sample set; The step of training the preset network model based on the training sample set to obtain the drag coefficient prediction model includes: The preset network model is trained based on the training sample set to obtain the initial drag coefficient prediction model; The initial drag coefficient prediction model is evaluated based on the validation sample set. When the initial drag coefficient prediction model passes the evaluation, the initial drag coefficient prediction model is determined as the drag coefficient prediction model.

7. The method according to claim 6, characterized in that, Also includes: If the initial drag coefficient prediction model fails the evaluation, the training sample set is updated, and the steps "train the preset network model based on the training sample set to obtain the initial drag coefficient prediction model" and "evaluate the initial drag coefficient prediction model based on the validation sample set" are re-executed based on the updated training sample set until the initial drag coefficient prediction model passes the evaluation.

8. A device for rapidly estimating the drag coefficient, characterized in that, include: The acquisition module is used to acquire the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the sliding drag coefficient of the target vehicle. The processing module is used to input the frontal area of ​​the target vehicle, the test mass of the target vehicle, and the skid drag coefficient of the target vehicle into the drag coefficient prediction model to obtain the drag coefficient of the target vehicle.

9. An electronic device, characterized in that, The device includes a memory for storing computer program instructions and a processor for executing the program instructions, wherein when the computer program instructions are executed by the processor, the device performs the method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method according to any one of claims 1 to 7.

Images