Model training method, parameter optimization method, test equipment and test system

Abstract

The invention discloses a model training method, a parameter optimization method, test equipment and a test system. The parameter optimization method comprises the following steps: inputting a second rim structure parameter into a rim lateral stiffness model determined by a rim lateral stiffness model training method, and determining a second predicted lateral stiffness corresponding to the second rim structure parameter; if the second predicted lateral stiffness does not meet the target stiffness condition, optimizing the second rim structure parameter, continuing to input the second rim structure parameter into a rim lateral stiffness model determined by the rim lateral stiffness model training method, and determining the second predicted lateral stiffness corresponding to the second rim structure parameter; and if the second predicted lateral stiffness meets the target stiffness condition, determining the second rim structure parameter as a target rim structure parameter. According to the method, complex simulation modeling testing does not need to be conducted on the vehicle rim based on the rim structure parameters, the mode of determining the target rim structure parameters is efficient and convenient, and manpower and material resources can be saved.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicle performance testing, and in particular to a model training method, a parameter optimization method, a testing device and a testing system. Background Art

[0002] The lateral stiffness of the rim is one of the important indicators that affect the NVH (Noise Vibration Harshness) level of the vehicle. The greater the lateral stiffness of the rim, the stronger the deformation resistance and vibration isolation performance of the vehicle rim, and the higher the NVH (Noise Vibration Harshness) level of the vehicle. Since different rim structural parameters correspond to different rim lateral stiffness, when designing a vehicle rim, it is usually necessary to pre-set a target value for the lateral stiffness of the rim. The target value is used as the target lateral stiffness of the rim. According to the pre-set target lateral stiffness of the rim, the rim structural parameters of the vehicle rim are designed so that the vehicle rim has the target lateral stiffness of the rim.

[0003] In the prior art, when designing a vehicle rim, it is usually necessary to establish a rim structure parameter model and perform simulation tests to determine the vehicle rim structure parameters that meet the target rim lateral stiffness. This method usually requires repeatedly establishing a rim structure parameter model and performing simulation tests on the rim structure parameters in the rim structure parameter model. If the simulation results determine that the rim structure parameters do not have the target rim lateral stiffness, the rim structure parameter model needs to be modified or rebuilt, and simulation tests need to be performed again until the simulation results determine that the rim structure parameters have the target rim lateral stiffness. Only then can the rim structure parameters that meet the target rim lateral stiffness be determined. This method requires repeated modeling and repeated simulation processes, which results in low efficiency due to a large workload and wastes a lot of human and material resources.

[0004] Therefore, how to more efficiently determine the rim structural parameters with the target rim lateral stiffness is a problem that needs to be solved urgently. Summary of the invention

[0005] The embodiments of the present invention provide a model training method, a parameter optimization method, a test device and a test system to solve the problem of how to efficiently determine the structural parameters of a rim with a target rim lateral stiffness.

[0006] A rim lateral stiffness model training method, comprising:

[0007] Performing a lateral stiffness test on a vehicle rim to obtain rim test data, wherein the rim test data includes a first rim structural parameter and a first test lateral stiffness corresponding to the first rim structural parameter;

[0008] Inputting the wheel rim test data into a neural network model for training, and determining a first predicted lateral stiffness corresponding to the first wheel rim structural parameter;

[0009] Determining a target loss function according to a first tested lateral stiffness and a first predicted lateral stiffness corresponding to the first rim structural parameter;

[0010] When the target loss function does not meet a preset convergence condition, updating the model parameters of the neural network model;

[0011] When the target loss function meets the preset convergence condition, the neural network model of the completed model parameters is updated and determined as the rim lateral stiffness model.

[0012] Preferably, the lateral stiffness test of the vehicle rim is performed to obtain the rim test data, including:

[0013] placing a vehicle rim corresponding to the first rim structural parameter on a test bench;

[0014] A lateral stiffness test is performed on the vehicle rim to determine a first test lateral stiffness corresponding to the first rim structural parameter.

[0015] Preferably, the performing a lateral stiffness test on the vehicle rim to determine a first test lateral stiffness corresponding to the first rim structural parameter comprises:

[0016] Performing a tapping test on a vehicle rim corresponding to the first rim structural parameter within a target frequency range to determine a resonance peak-to-peak frequency and an anti-resonance peak-to-peak frequency of each of the vehicle rims;

[0017] A first test lateral stiffness corresponding to the first rim structural parameter is determined according to the rim mass of the vehicle rim, the resonance peak-to-peak frequency, and the anti-resonance peak-to-peak frequency.

[0018] A method for optimizing wheel rim structural parameters, comprising:

[0019] Obtaining the second rim structural parameters;

[0020] Inputting the second rim structural parameter into the rim lateral stiffness model determined by the rim lateral stiffness model training method according to any one of claims 1 to 3, and determining a second predicted lateral stiffness corresponding to the second rim structural parameter;

[0021] If the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameters are optimized, the optimized second rim structural parameters are updated as the second rim structural parameters, and the second rim structural parameters are input into the rim lateral stiffness model determined by the rim lateral stiffness model training method according to any one of claims 1 to 3 to determine the second predicted lateral stiffness corresponding to the second rim structural parameters;

[0022] If the second predicted lateral stiffness meets the target stiffness condition, the second rim structural parameters are determined as target rim structural parameters.

[0023] Preferably, after determining the second predicted lateral stiffness corresponding to the second rim structural parameter, the rim structural parameter optimization method further comprises:

[0024] Determining a target prediction error according to the second predicted lateral stiffness and a preset target rim lateral stiffness;

[0025] If the target prediction error does not exceed a preset error range, determining that the second predicted lateral stiffness meets a target stiffness condition;

[0026] If the target prediction error exceeds a preset error range, it is determined that the second predicted lateral stiffness does not meet a target stiffness condition.

[0027] Preferably, the optimizing the second rim structural parameters and updating the optimized second rim structural parameters as the second rim structural parameters include:

[0028] Obtaining a target weight corresponding to each second rim structural parameter in the rim lateral stiffness model;

[0029] The first N second rim structural parameters with larger target weights are optimized, and the optimized second rim structural parameters are updated as the second rim structural parameters, wherein N≥2.

[0030] A testing device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned rim lateral stiffness model training method when executing the computer program, or implements the above-mentioned rim structural parameter optimization method when executing the computer program.

[0031] A test system comprises a test bench and the above-mentioned test equipment, and is used to implement the above-mentioned rim lateral stiffness model training method, or to implement the above-mentioned rim structural parameter optimization method.

[0032] Preferably, the test bench comprises a gantry, an elastic rope, a connecting assembly and an acceleration sensor, wherein both ends of the elastic rope are respectively fixed on the gantry for hanging a vehicle rim; the connecting assembly is equipped on the vehicle rim for performing a lateral stiffness test on the vehicle rim corresponding to a first rim structural parameter; the acceleration sensor is fixed on the connecting assembly for obtaining a first test lateral stiffness corresponding to the first rim structural parameter when performing a lateral stiffness test on the vehicle rim within a target frequency range.

[0033] Preferably, the connecting assembly includes a conical bolt and a conical connecting block, the conical bolt includes a conical body and a stud extending from a first side of the conical body, the second side of the conical body is clamped with the conical connecting block, and the stud is screwed onto the vehicle rim.

[0034] The above-mentioned model training method, test equipment and test system perform a lateral stiffness test on the vehicle rim to obtain rim test data, wherein the rim test data includes a first rim structural parameter and a first test lateral stiffness corresponding to the first rim structural parameter; the rim test data is input into a neural network model for training to determine a first predicted lateral stiffness corresponding to the first rim structural parameter; this method can obtain rim test data that is closer to the actual vehicle rim condition, so that the subsequent training of the neural network model based on the rim test data to obtain a more accurate rim lateral stiffness model. According to the first test lateral stiffness and the first predicted lateral stiffness corresponding to the first rim structural parameter, a target loss function is determined; when the target loss function meets the preset convergence condition, the neural network model of the updated model parameters is determined as the rim lateral stiffness model. This method is convenient for directly establishing a mapping relationship between the rim lateral stiffness and the rim structural parameters. When designing the rim structural parameters, the rim structural parameters that meet the target rim lateral stiffness can be directly obtained according to the rim lateral stiffness model without establishing a complex simulation model. This method can not only improve the efficiency of determining the rim structural parameters, which is convenient and fast, but also save manpower and material resources and reduce resource waste.

[0035] In the above parameter optimization method, test equipment and test system, the second predicted lateral stiffness corresponding to the second rim structural parameter is determined by the rim lateral stiffness model, and the second predicted lateral stiffness corresponding to the second rim structural parameter can be obtained without establishing a simulation model corresponding to the second rim structural parameter, which is highly efficient. When the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameter is optimized and updated, and after obtaining the updated second rim structural parameter, the lateral stiffness prediction is continued based on the updated second rim structural parameter through the rim lateral stiffness model, so as to facilitate the acquisition of the second rim structural parameter that meets the target rim lateral stiffness. When the second predicted lateral stiffness meets the target stiffness condition, the second rim structural parameter is determined as the target rim structural parameter. The method does not need to perform complex simulation modeling, simulation testing and model adjustment on the vehicle rim based on the rim structural parameter. The method is efficient and convenient in determining the target rim structural parameter, and can save manpower and material resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0037] Figure 1 is a flow chart of a rim lateral stiffness model training method in one embodiment of the present invention;

[0038] Figure 2 is another flow chart of a rim lateral stiffness model training method in one embodiment of the present invention;

[0039] Figure 3 is another flow chart of a rim lateral stiffness model training method in one embodiment of the present invention;

[0040] Figure 4 is a flow chart of a method for optimizing rim structural parameters in one embodiment of the present invention;

[0041] Figure 5 is another flow chart of a method for optimizing wheel rim structural parameters in one embodiment of the present invention;

[0042] Figure 6 is another flow chart of a method for optimizing wheel rim structural parameters in one embodiment of the present invention;

[0043] Figure 7 is a schematic diagram of a testing device in one embodiment of the present invention;

[0044] Figure 8is an origin transfer function curve generated during a knock test of a vehicle rim in one embodiment of the present invention;

[0045] Fig. 9 is a schematic diagram of a test bench in one embodiment of the present invention;

[0046] Fig.10 is a schematic diagram of a connection assembly in one embodiment of the present invention;

[0047] In the figure: 1, gantry; 2, elastic rope; 3, vehicle rim; 4, conical bolt; 41, conical body; 42, stud; 5, conical connecting block. DETAILED DESCRIPTION

[0048] 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 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.

[0049] The present invention provides a rim lateral stiffness model training method. The rim lateral stiffness model training method is applied in a rim lateral stiffness model training system, the rim lateral stiffness model training system is used to train a rim lateral stiffness model, and the rim lateral stiffness model is used to efficiently determine the rim structural parameters with a target rim lateral stiffness.

[0050] In one embodiment, if Figure 1 As shown, a rim lateral stiffness model training method is provided, and the method is applied in Figure 7 The test equipment in the example is used to illustrate, including the following steps:

[0051] S101: performing a lateral stiffness test on a vehicle rim to obtain rim test data, where the rim test data includes a first rim structural parameter and a first test lateral stiffness corresponding to the first rim structural parameter;

[0052] S102: inputting the rim test data into a neural network model for training, and determining a first predicted lateral stiffness corresponding to a first rim structural parameter;

[0053] S103: determining a target loss function according to a first tested lateral stiffness and a first predicted lateral stiffness corresponding to the first rim structural parameter;

[0054] S104: When the target loss function does not meet the preset convergence condition, updating the model parameters of the neural network model;

[0055] S105: When the target loss function satisfies a preset convergence condition, the neural network model of the completed model parameters is updated and determined as the rim lateral stiffness model.

[0056] The rim test data refers to the test data obtained by performing a lateral stiffness test on the vehicle rim, including the first rim structural parameters and the first test lateral stiffness corresponding to the first rim structural parameters. The first rim structural parameters are the rim structural parameters of the vehicle rim used to train the rim lateral stiffness model, including but not limited to the rim mass, wheel center thickness, wheel center diameter, wheel center root fillet, spoke inner thickness, wheel center weight reduction groove depth and spoke inner width. The first test lateral stiffness is the rim lateral stiffness of the vehicle rim obtained after performing a lateral stiffness test on the vehicle rim having the first rim structural parameters.

[0057] As an example, in step S101, the testing device obtains at least one rim structural parameter corresponding to a vehicle rim, takes the rim structural parameter corresponding to each vehicle rim as the first rim structural parameter corresponding to the vehicle rim, tests the vehicle rim having the first rim structural parameter, obtains the lateral stiffness of the vehicle rim, takes the lateral stiffness as the first test lateral stiffness of the corresponding vehicle rim, takes each first rim structural parameter and the first test lateral stiffness corresponding to it, and takes the data as rim test data. It can be understood that vehicle rims with different rim structural parameters have different lateral stiffnesses, and vehicle rims with different first rim structural parameters are tested to obtain rim test data including the first rim structural parameter and the first test lateral stiffness corresponding to the first rim structural parameter. This method can obtain rim test data that is closer to the actual vehicle rim condition, so that the rim lateral stiffness model obtained by subsequently training the neural network model based on the rim test data is more accurate.

[0058] The neural network model is an original model used to train the rim lateral stiffness model. The first predicted lateral stiffness is the lateral stiffness obtained by predicting the first rim structural parameter using the neural network model.

[0059] As an example, in step S102, the test equipment inputs the acquired rim test data corresponding to each vehicle rim into the neural network model for training, outputs the predicted value of the lateral stiffness corresponding to each first rim structural parameter, and uses the predicted value of the lateral stiffness as the first predicted lateral stiffness. In this example, the test equipment uses the first rim structural parameter in the rim test data as the input of the neural network model, uses the first test lateral stiffness in the rim test data as the output label of the neural network model, trains the neural network model, and obtains the first predicted lateral stiffness outputted in each training. In this example, a variety of neural network models such as BP (Back Propagation) neural network model and convolutional neural network model can be used for training to determine the first predicted lateral stiffness corresponding to the first rim structural parameter. For example, when the BP (Back Propagation) neural network model is used for training, M parameters including rim mass, wheel center thickness, wheel center diameter, wheel center root fillet, spoke inner thickness, wheel center weight reduction recess depth, and spoke inner width are set according to the first rim structural parameters, M inputs: X1, X2, ..., X i , …, X M , each input corresponds to a weight W1, W2, ..., W i , …, W M During training, the corresponding weight correction is applied to each parameter, and the bias value b is increased to obtain the model processing result The above model processing result net is passed through the activation function g(·) to obtain the output o=g(net) of the neural network model, which is the first predicted lateral stiffness. In this example, the rim test data is input into the neural network model for training to determine the first predicted lateral stiffness corresponding to the first rim structural parameter. This method uses the obtained rim test data that is closer to the actual vehicle rim condition for model training, which can make the subsequent training to obtain the rim lateral stiffness model more accurate.

[0060] Among them, the target loss function refers to the loss function generated during the model training process.

[0061] As an example, in step S103, when the test equipment is training the neural network model, the first predicted lateral stiffness is output, the first test lateral stiffness corresponding to the first rim structural parameter is used as the prediction label, the loss function between the first predicted lateral stiffness and the first test lateral stiffness is obtained, and the loss function is used as the target loss function. For example, when the BP (Back Propagation) neural network model is used for training, the first predicted lateral stiffness is obtained, and the target loss function between the first predicted lateral stiffness and the first test lateral stiffness is obtained. In this example, the target loss function is determined based on the first test lateral stiffness corresponding to the first rim structural parameter and the first predicted lateral stiffness, so as to facilitate the subsequent updating of the neural network model based on the target loss function, so that the neural network model converges faster and obtains the rim lateral stiffness model.

[0062] Among them, the preset convergence condition is a preset condition used to determine whether the neural network model converges according to the target loss function.

[0063] As an example, in step S104, the test device determines whether the target loss function meets the preset convergence conditions. When it is determined that the target loss function does not meet the preset convergence conditions, the model parameters of the neural network model continue to be updated based on the target loss function. In this example, the preset convergence conditions include but are not limited to the target loss function being within a preset range, the target loss function not exceeding a preset threshold, or the number of calculations (i.e., the number of training times) of the target loss function reaches a certain value. Among them, the preset range is used to judge the fluctuation of the target loss function, so as to stop training the neural network model when the target loss function fluctuates slightly within the preset range. The preset threshold is used to judge the size of the target loss function, so as to stop training the neural network model when the target loss function is lower than the preset threshold. In this example, after updating the model parameters of the neural network model, the test device continues to execute steps S102 and S103 until the target loss function meets the preset convergence conditions.

[0064] As an example, in step S105, when the test equipment determines that the target loss function meets the preset convergence condition, the neural network model that has been updated the most recently is determined as the rim lateral stiffness model. In this example, the rim lateral stiffness model is obtained through training, which is convenient for directly establishing the mapping relationship between the rim lateral stiffness and the rim structural parameters, and is convenient for directly determining the rim lateral stiffness that meets the target rim lateral stiffness according to the rim lateral stiffness model when designing the rim structural parameters, and determining the rim structural parameters corresponding to the input in the rim lateral stiffness model as the rim structural parameters that meet the target rim lateral stiffness requirements, without the need to establish a complex simulation model. This method can not only improve the efficiency of determining the rim structural parameters, which is convenient and fast, but also saves manpower and material resources and reduces resource waste.

[0065] In this embodiment, a lateral stiffness test is performed on a vehicle rim to obtain rim test data, wherein the rim test data includes a first rim structural parameter and a first test lateral stiffness corresponding to the first rim structural parameter; the rim test data is input into a neural network model for training to determine a first predicted lateral stiffness corresponding to the first rim structural parameter; this method can obtain rim test data that is relatively close to the actual vehicle rim condition, so that a subsequent rim lateral stiffness model obtained by training the neural network model based on the rim test data is relatively accurate. According to the first test lateral stiffness and the first predicted lateral stiffness corresponding to the first rim structural parameter, a target loss function is determined; when the target loss function meets the preset convergence condition, the neural network model of the updated model parameters is determined as the rim lateral stiffness model. This method is convenient for directly establishing a mapping relationship between the rim lateral stiffness and the rim structural parameters. When designing the rim structural parameters, the rim structural parameters that meet the target rim lateral stiffness can be directly obtained according to the rim lateral stiffness model without establishing a complex simulation model. This method can not only improve the efficiency of determining the rim structural parameters, which is convenient and fast, but also save manpower and material resources and reduce resource waste.

[0066] In one embodiment, if Figure 2 As shown, step S101, i.e., performing a lateral stiffness test on a vehicle rim to obtain rim test data, includes:

[0067] S201: placing a vehicle rim corresponding to the first rim structural parameter on a test bench;

[0068] S202: Performing a lateral stiffness test on a vehicle rim to determine a first test lateral stiffness corresponding to a first rim structural parameter.

[0069] Among them, the test bench refers to the equipment for testing the lateral stiffness of the vehicle rim.

[0070] As an example, in step S201, a vehicle rim corresponding to the first rim structural parameter is placed on an equipment stand to facilitate the testing equipment to perform a lateral stiffness test on the vehicle rim and obtain a first test lateral stiffness corresponding to the vehicle rim.

[0071] The lateral stiffness test refers to a test process of obtaining a first test lateral stiffness of a vehicle rim having different first rim structural parameters on a test bench.

[0072] As an example, in step S202, the test equipment performs a lateral stiffness test on a vehicle rim placed on a test bench and having a first rim structural parameter, and determines a first test lateral stiffness corresponding to the first rim structural parameter of the vehicle rim. The test equipment performs a lateral stiffness test on multiple vehicle rims having the first rim structural parameter, and obtains a first test lateral stiffness corresponding to the first rim structural parameter of each vehicle rim. In this example, a lateral stiffness test is performed on the vehicle rim to determine a first test lateral stiffness corresponding to the first rim structural parameter. The first test lateral stiffness is the lateral stiffness corresponding to the actual vehicle rim, so that the subsequent training of the neural network model based on the rim test data including the first rim structural parameter and the first test lateral stiffness is more in line with the actual situation, and the obtained rim lateral stiffness model is more accurate.

[0073] In this embodiment, a vehicle rim corresponding to the first rim structural parameter is placed on a test bench, and a lateral stiffness test is performed on the vehicle rim to determine the first test lateral stiffness corresponding to the first rim structural parameter. This makes the subsequent training of the neural network model based on the rim test data including the first rim structural parameter and the first test lateral stiffness more in line with the actual situation, and the obtained rim lateral stiffness model is more accurate.

[0074] In one embodiment, if Figure 3 As shown, step S202, i.e., performing a lateral stiffness test on the vehicle rim to determine a first test lateral stiffness corresponding to a first rim structural parameter, includes:

[0075] S301: performing a tapping test on a vehicle rim corresponding to the first rim structural parameter within a target frequency range to determine a resonance peak-to-peak frequency and an anti-resonance peak-to-peak frequency of each vehicle rim;

[0076] S302: Determine a first test lateral stiffness corresponding to a first rim structural parameter according to a rim mass, a resonance peak-to-peak frequency, and an anti-resonance peak-to-peak frequency of a vehicle rim.

[0077] The target frequency range refers to the frequency range in which the vehicle rim is struck when the lateral stiffness test is performed on the vehicle rim. The knock test refers to the process in which the vehicle rim is struck. The resonance peak frequency refers to the frequency corresponding to the resonance peak generated when the vehicle rim is subjected to the knock test within the target frequency range. The anti-resonance peak frequency refers to the frequency corresponding to the anti-resonance peak generated when the vehicle rim is subjected to the knock test within the target frequency range.

[0078] As an example, in step S301, the test equipment controls the transmission function test equipment to perform a tapping test on a vehicle rim having a first rim structural parameter within a target frequency range, obtains an origin transfer function curve generated by the vehicle rim during the tapping test, determines the resonance peak frequency and anti-resonance peak frequency of the vehicle rim according to the origin transfer function curve, performs a tapping test on each vehicle rim corresponding to the first rim structural parameter, and obtains the resonance peak frequency and anti-resonance peak frequency of each vehicle rim. The origin transfer function curve refers to a relationship curve between responses generated at the tapping point of the vehicle rim when the tapping point is subjected to tapping tests of different frequencies. Figure 8 The origin transfer function curve generated by the vehicle rim during the tapping test is shown. According to the origin transfer function curve, f1 is the resonance peak-to-peak frequency of the vehicle rim, and f2 is the anti-resonance peak-to-peak frequency of the vehicle rim. In this example, the vehicle rim is tapped within the target frequency range to improve the test efficiency, determine the resonance peak-to-peak frequency and anti-resonance peak-to-peak frequency of each vehicle rim, and determine the first test lateral stiffness corresponding to the first rim structural parameter of each vehicle rim according to the resonance peak-to-peak frequency and anti-resonance peak-to-peak frequency of each vehicle rim.

[0079] Among them, rim mass is the mass of the vehicle rim.

[0080] As an example, in step S302, after determining the resonance peak frequency and anti-resonance peak frequency of each vehicle rim, the test equipment determines the first test lateral stiffness corresponding to the first rim structural parameter of each vehicle rim according to the rim mass, resonance peak frequency and anti-resonance peak frequency of each vehicle rim. In this example, the test equipment processes the rim mass, resonance peak frequency and anti-resonance peak frequency of each vehicle rim using the following formula to obtain the first test lateral stiffness corresponding to the first rim structural parameter of each vehicle rim:

[0081]

[0082] Among them, M w is the rim mass of the vehicle rim, f1 is the resonance peak frequency of the vehicle rim, f2 is the anti-resonance peak frequency of the vehicle rim, K wheelis the first test lateral stiffness corresponding to the first rim structural parameter. As an example, f1 may be the first-order resonance peak frequency, and f2 may be the first-order anti-resonance peak frequency of the vehicle rim. In this example, the resonance peak frequency and anti-resonance peak frequency of each vehicle rim determined by the tapping test, as well as the rim mass of each vehicle rim, are used to determine the first test lateral stiffness corresponding to the first rim structural parameter of each vehicle rim, so that the first test lateral stiffness that meets the actual situation of the vehicle rim can be obtained, which facilitates the subsequent training based on the first test lateral stiffness to obtain a more accurate rim lateral stiffness model.

[0083] In the present embodiment, the resonance peak-to-peak frequency and the anti-resonance peak-to-peak frequency of each vehicle rim are determined through a tapping test, and the first test lateral stiffness corresponding to the first rim structural parameters of each vehicle rim is determined based on the resonance peak-to-peak frequency and the anti-resonance peak-to-peak frequency of each vehicle rim, as well as the rim mass of each vehicle rim. The first test lateral stiffness that meets the actual situation of the vehicle rim can be obtained, which facilitates subsequent training based on the first test lateral stiffness to obtain a more accurate rim lateral stiffness model.

[0084] The present invention provides a rim structural parameter optimization method. Specifically, the rim structural parameter optimization method is applied in a rim structural parameter optimization system, and the rim structural parameter optimization is used to optimize the rim structural parameters according to a rim lateral stiffness model, so as to efficiently determine the rim structural parameters with a target rim lateral stiffness.

[0085] In one embodiment, if Figure 4 As shown, a method for optimizing wheel rim structural parameters is provided, and the method is applied in Figure 7 The test equipment in the example is used to illustrate, including the following steps:

[0086] S401: Acquire the second rim structural parameters;

[0087] S402: inputting the second rim structural parameter into the rim lateral stiffness model determined by the rim lateral stiffness model training method to determine a second predicted lateral stiffness corresponding to the second rim structural parameter;

[0088] S403: If the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameter is optimized, the optimized second rim structural parameter is updated as the second rim structural parameter, and the second rim structural parameter is input into the rim lateral stiffness model determined by the rim lateral stiffness model training method to determine the second predicted lateral stiffness corresponding to the second rim structural parameter;

[0089] S404: If the second predicted lateral stiffness satisfies the target stiffness condition, the second rim structural parameter is determined as the target rim structural parameter.

[0090] The second rim structural parameter refers to a preset rim structural parameter to be optimized.

[0091] As an example, in step S401, the test equipment obtains the preset second rim structural parameters, so as to facilitate the subsequent optimization of the second rim structural parameters. Understandably, in order to obtain the target rim structural parameters that meet the target rim lateral stiffness, it is necessary to first set the initial rim structural parameters, which are the second rim structural parameters, so as to facilitate the subsequent prediction of the rim lateral stiffness of the second rim structural parameters through the rim lateral stiffness model, and adjust the second rim structural parameters according to the prediction results to obtain the target rim structural parameters that meet the target rim lateral stiffness.

[0092] The second predicted lateral stiffness refers to the lateral stiffness obtained by predicting the second rim structural parameters through the rim lateral stiffness model.

[0093] As an example, in step S402, after the test equipment obtains the rim lateral stiffness model and the second rim structural parameters determined by the rim lateral stiffness model training method, the second rim structural parameters are input into the rim lateral stiffness model for lateral stiffness prediction, and the second predicted lateral stiffness corresponding to the second rim structural parameters is output. In this example, the second predicted lateral stiffness corresponding to the second rim structural parameters is determined by the rim lateral stiffness model, and the second predicted lateral stiffness corresponding to the second rim structural parameters can be obtained without establishing a simulation model corresponding to the second rim structural parameters, which is highly efficient and convenient.

[0094] Among them, the target stiffness condition is used to determine whether the second predicted lateral stiffness predicted by the rim lateral stiffness model meets the preset condition of the target rim lateral stiffness. Among them, the target rim lateral stiffness refers to the target value of the lateral stiffness that the vehicle rim needs to achieve when designing the vehicle rim. It can be understood that in the design stage of the vehicle rim, it is necessary to determine the target value of the lateral stiffness of the designed vehicle rim, and use the target value as the target rim lateral stiffness. Based on the target rim lateral stiffness, the rim structural parameters that meet the target rim lateral stiffness are designed. In this example, when the difference between the second predicted lateral stiffness predicted by the rim lateral stiffness model and the target rim lateral stiffness is small, it indicates that the second predicted lateral stiffness meets the target stiffness condition, and when the difference between the second predicted lateral stiffness predicted by the rim lateral stiffness model and the target rim lateral stiffness is large, it indicates that the second predicted lateral stiffness does not meet the target stiffness condition.

[0095] As an example, in step S403, the test equipment determines whether the second predicted lateral stiffness obtained by the rim lateral stiffness model meets the target stiffness condition. When it is determined that the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameters are optimized, and the optimized second rim structural parameters are updated to the second rim structural parameters, and the second rim structural parameters are input into the rim lateral stiffness model determined by the rim lateral stiffness model training method to determine the second predicted lateral stiffness corresponding to the second rim structural parameters. In this example, when the test equipment determines that the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameters are optimized and updated to obtain the updated second rim structural parameters, and step S402 is continued until it is determined that the second predicted lateral stiffness meets the target stiffness condition. It is understandable that since the second predicted lateral stiffness does not meet the target stiffness condition, it indicates that the difference between the second predicted lateral stiffness and the target rim lateral stiffness is large. At this time, the second rim structural parameters corresponding to the second predicted lateral stiffness cannot be used for vehicle rim design, and need to be optimized and updated. Based on the updated second rim structural parameters, lateral stiffness prediction is performed through the rim lateral stiffness model, so that the predicted second predicted lateral stiffness meets the target stiffness condition, and the second rim structural parameters corresponding to the second predicted lateral stiffness are more in line with the design requirements of the target rim lateral stiffness.

[0096] In this example, if the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameters are optimized and updated, and after obtaining the updated second rim structural parameters, the lateral stiffness prediction is continued through the rim lateral stiffness model based on the updated second rim structural parameters, so as to facilitate obtaining the second rim structural parameters that meet the target rim lateral stiffness. This method does not need to establish a vehicle rim simulation model based on the rim structural parameters for simulation testing, thus avoiding the need to modify or rebuild the vehicle rim simulation model when the test results do not meet the target rim lateral stiffness during the simulation test of the simulation model, thereby reducing the waste of manpower and material resources, and improving the efficiency of determining the rim structural parameters with the target rim lateral stiffness, which is more convenient and quick.

[0097] The target rim structural parameter refers to the second rim structural parameter corresponding to the second predicted lateral stiffness that meets the target stiffness condition.

[0098] As an example, in step S404, when the second predicted lateral stiffness meets the target stiffness condition, the test equipment determines the second rim structural parameter corresponding to the second predicted lateral stiffness that meets the target stiffness condition as the target rim structural parameter. It can be understood that after the second rim structural parameter is input into the rim lateral stiffness model, the second predicted lateral stiffness is outputted, and when the second predicted lateral stiffness meets the target stiffness condition, it indicates that the second predicted lateral stiffness predicted by the rim lateral stiffness model for the second rim structural parameter is less different from the target rim lateral stiffness, and the second rim structural parameter corresponding to the second predicted lateral stiffness meets the design requirements of the rim parameter corresponding to the target rim lateral stiffness, and therefore, the second rim structural parameter corresponding to the second predicted lateral stiffness can be used as the target rim structural parameter. In this example, the lateral stiffness is predicted directly according to the second rim structural parameters through the rim lateral stiffness model to obtain the second predicted lateral stiffness. When the second predicted lateral stiffness meets the target stiffness conditions, the second rim structural parameters are determined as the target rim structural parameters. There is no need to perform complex simulation modeling, simulation testing and model adjustment. The method of determining the target rim structural parameters is efficient and convenient, and can save manpower and material resources.

[0099] In this embodiment, the second predicted lateral stiffness corresponding to the second rim structural parameter is determined by the rim lateral stiffness model, and the second predicted lateral stiffness corresponding to the second rim structural parameter can be obtained without establishing a simulation model corresponding to the second rim structural parameter, which is highly efficient. When the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameter is optimized and updated, and after obtaining the updated second rim structural parameter, the lateral stiffness prediction is continued based on the updated second rim structural parameter through the rim lateral stiffness model, so as to facilitate the acquisition of the second rim structural parameter that meets the target rim lateral stiffness. When the second predicted lateral stiffness meets the target stiffness condition, the second rim structural parameter is determined as the target rim structural parameter. This method does not require complex simulation modeling, simulation testing and model adjustment of the vehicle rim based on the rim structural parameter. This method is efficient and convenient in determining the target rim structural parameter, and can save manpower and material resources.

[0100] In one embodiment, if Figure 5 As shown, after step S402, that is, after determining the second predicted lateral stiffness corresponding to the second rim structural parameter, the rim structural parameter optimization method further includes:

[0101] S501: determining a target prediction error according to the second predicted lateral stiffness and a preset target rim lateral stiffness;

[0102] S502: If the target prediction error does not exceed the preset error range, determining that the second predicted lateral stiffness meets the target stiffness condition;

[0103] S503: If the target prediction error exceeds a preset error range, determining that the second predicted lateral stiffness does not meet the target stiffness condition.

[0104] The target prediction error refers to the difference between the second predicted lateral stiffness and the preset target rim lateral stiffness.

[0105] As an example, in step S501, after determining the second predicted lateral stiffness corresponding to the second rim structural parameter, the test device calculates the difference between the second predicted lateral stiffness and the preset target rim lateral stiffness, and determines the difference between the two as the target prediction error. For example, the test device can calculate the absolute value of the difference between the second predicted lateral stiffness and the target rim lateral stiffness to obtain an absolute error, and use the absolute error as the target prediction error. It can also obtain the ratio of the absolute value of the difference to the target rim lateral stiffness after calculating the absolute value of the difference between the second predicted lateral stiffness and the target rim lateral stiffness to obtain a relative error, and use the relative error as the target prediction error. In this example, the test device determines the target prediction error between the second predicted lateral stiffness and the preset target rim lateral stiffness, so as to judge whether the second predicted lateral stiffness meets the target stiffness condition according to the target prediction error.

[0106] Among them, the preset error range is the error range used to determine the size of the target prediction error.

[0107] As an example, in step S502, when the test device determines that the target prediction error does not exceed the preset error range, it determines that the second predicted lateral stiffness meets the target stiffness condition. Understandably, when the target prediction error does not exceed the preset error range, it indicates that the difference between the second predicted lateral stiffness and the preset target rim lateral stiffness is small, and the second predicted lateral stiffness meets the target stiffness condition.

[0108] As an example, in step S503, when the test device determines that the target prediction error exceeds the preset error range, it determines that the second predicted lateral stiffness does not meet the target stiffness condition. Understandably, when the target prediction error exceeds the preset error range, it indicates that the difference between the second predicted lateral stiffness and the preset target rim lateral stiffness is large, and the second predicted lateral stiffness does not meet the target stiffness condition.

[0109] In this embodiment, the testing equipment determines a target prediction error between the second predicted lateral stiffness and a preset target rim lateral stiffness, and judges whether the second predicted lateral stiffness meets the target stiffness condition based on whether the target prediction error does not exceed a preset error range, thereby facilitating determination of the target rim structural parameters when the second predicted lateral stiffness meets the target stiffness condition.

[0110] In one embodiment, if Figure 6As shown, step S403, namely optimizing the second rim structural parameters and updating the optimized second rim structural parameters as the second rim structural parameters, comprises:

[0111] S601: Obtaining a target weight corresponding to each second rim structural parameter in the rim lateral stiffness model;

[0112] S602: Optimize the first N second rim structural parameters with larger target weights, and update the optimized second rim structural parameters as the second rim structural parameters, where N≥2.

[0113] The target weight is the weight corresponding to each second rim structural parameter in the rim lateral stiffness model.

[0114] As an example, in step S601, when the test equipment determines that the second predicted lateral stiffness does not meet the target stiffness condition, the target weight corresponding to each second rim structural parameter in the rim lateral stiffness model is obtained. It can be understood that the second rim structural parameters include but are not limited to rim mass, wheel center thickness, wheel center diameter, wheel center root fillet, spoke inner thickness, wheel center weight reduction groove depth and spoke inner width. Therefore, the rim lateral stiffness model is a multi-input model. When performing the second predicted lateral stiffness and test, the size of the target weight corresponding to each second rim structural parameter determines the influence of each second rim structural parameter on the prediction result. The larger the target weight, the greater the influence of the second rim structural parameter corresponding to the target weight on the prediction result. When the second predicted lateral stiffness does not meet the target stiffness condition, it is necessary to give priority to adjusting the target weight with greater influence. Therefore, it is necessary to obtain the target weight corresponding to each second rim structural parameter in the rim lateral stiffness model. In this example, the target weight corresponding to each second rim structural parameter in the rim lateral stiffness model is obtained, so as to optimize the second rim structural parameter according to the target weight.

[0115] As an example, in step S602, after obtaining the target weight corresponding to each second rim structural parameter, the test device optimizes the first N second rim structural parameters with larger target weights, and updates the optimized second rim structural parameters to the second rim structural parameters, where N≥2. For example, if the test device determines that the target weights corresponding to the rim mass, wheel center thickness, and wheel center diameter among the second rim structural parameters are larger, the three rim structural parameters of rim mass, wheel center thickness, and wheel center diameter with larger target weights are optimized and updated to obtain the optimized second rim structural parameters, and the optimized second rim structural parameters are updated to the second rim structural parameters. In this example, the first N second rim structural parameters with larger target weights are optimized, and the optimized second rim structural parameters are updated to the second rim structural parameters. This method can relatively quickly determine the second rim structural parameters that need to be optimized, and facilitates the subsequent more efficient determination of the target rim structural parameters.

[0116] In this embodiment, the target weight corresponding to each second rim structural parameter in the rim lateral stiffness model is obtained, the first N second rim structural parameters with larger target weights are optimized, and the optimized second rim structural parameters are updated as the second rim structural parameters. This can relatively quickly determine the second rim structural parameters that need to be optimized, and facilitates the subsequent more efficient determination of the target rim structural parameters.

[0117] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0118] In one embodiment, a test device is provided. The test device may be a server, and its internal structure diagram may be as follows: Figure 7 As shown. The test device includes a processor, a memory, a network interface and a database connected through a system bus. Among them, the processor of the test device is used to provide computing and control capabilities. The memory of the test device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the test device is used to store data used or generated in the process of executing the rim lateral stiffness model training method, or, to store data used or generated in the process of executing the rim structure parameter optimization method. The network interface of the test device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a rim lateral stiffness model training method is implemented, or, a rim structure parameter optimization method is implemented.

[0119] In one embodiment, a testing device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the rim lateral stiffness model training method in the above embodiment is implemented, for example Figure 1 S101-S105 as shown, or Figures 2 to 3 As shown in , to avoid repetition, it will not be repeated here. Alternatively, when the processor executes the computer program, the wheel rim lateral stiffness model training method in the above embodiment is implemented, for example Figure 4 S401-S404 as shown, or Figure 5 to Figure 6 To avoid repetition, it will not be described here.

[0120] In one embodiment, a test system is provided, including a test bench and the above-mentioned test equipment, for implementing the rim lateral stiffness model training method in the above-mentioned embodiment, for example Figure 1 S101-S105 as shown, or Figures 2 to 3 As shown in , to avoid repetition, it will not be described here. Alternatively, the method for optimizing the rim structure parameters in the above embodiment is as follows: Figure 4 S401-S404 as shown, or Figure 5 to Figure 6 To avoid repetition, it will not be described here.

[0121] In one embodiment, the test bench includes a gantry, an elastic rope, a connecting assembly and an acceleration sensor, wherein two ends of the elastic rope are respectively fixed on the gantry for hanging a vehicle rim; the connecting assembly is equipped on the vehicle rim for performing a lateral stiffness test on the vehicle rim corresponding to a first rim structural parameter; the acceleration sensor is fixed on the connecting assembly for obtaining a first test lateral stiffness corresponding to the first rim structural parameter when performing a lateral stiffness test on the vehicle rim within a target frequency range.

[0122] As an example, Fig. 9 As shown, it is a schematic diagram of a test bench, which includes a gantry 1, an elastic rope 2, a connecting assembly (not shown in the figure) and an acceleration sensor (not shown in the figure), wherein both ends of the elastic rope 2 are respectively fixed on the gantry 1 for hanging the vehicle rim; the connecting assembly (not shown in the figure) is equipped on the vehicle rim 3 for performing a lateral stiffness test on the vehicle rim corresponding to the first rim structural parameter; the acceleration sensor (not shown in the figure) is fixed on the connecting assembly (not shown in the figure) for obtaining a first test lateral stiffness corresponding to the first rim structural parameter when performing a lateral stiffness test on the vehicle rim within a target frequency range.

[0123] The test bench provided in this embodiment equips the connecting assembly on the vehicle rim, and fixes the acceleration sensor on the connecting assembly. Compared with the method of gluing the connecting assembly to the vehicle rim for lateral stiffness testing, this embodiment can avoid the problem of reducing the lateral stiffness of the gluing point of the vehicle rim due to gluing, and facilitates the realization of more accurate lateral stiffness testing.

[0124] In one embodiment, the connecting assembly includes a conical bolt and a conical connecting block, wherein the conical bolt includes a conical body and a stud extending from a first side of the conical body, the second side of the conical body is clamped with the conical connecting block, and the stud is screwed onto the vehicle rim.

[0125] As an example, Fig.10 , which is a schematic diagram of a connection assembly, the conical bolt 4 includes a conical body 41 and a stud 42 extending from a first side of the conical body, the second side of the conical body 41 is engaged with the conical connection block 5, and the stud 42 is screwed to the vehicle rim (not shown) through a rim bolt mounting hole (not shown in the figure). In this example, the conical connection block 5 is a hexahedral mass block, which is engaged with the inner cavity of the conical body 41, and the inner cavity of the conical body 41 is consistent with the contour of the hexahedral mass block.

[0126] The connection assembly provided in this embodiment adopts the method of clamping the second side of the conical surface body with the conical surface connecting block, and the stud is screwed on the vehicle rim. Compared with the method of gluing the conical surface connecting block to the vehicle rim for lateral stiffness testing, it can avoid reducing the lateral stiffness of the rim at the bonding point between the vehicle rim and the conical surface connecting block due to gluing, thereby reducing the test error of the lateral stiffness test.

[0127] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A rim lateral stiffness model training method, characterized in that: include: Performing a lateral stiffness test on a vehicle rim to obtain rim test data, wherein the rim test data includes a first rim structural parameter and a first test lateral stiffness corresponding to the first rim structural parameter; Inputting the wheel rim test data into a neural network model for training, and determining a first predicted lateral stiffness corresponding to the first wheel rim structural parameter; Determining a target loss function according to a first tested lateral stiffness and a first predicted lateral stiffness corresponding to the first rim structural parameter; When the target loss function does not meet a preset convergence condition, updating the model parameters of the neural network model; When the target loss function meets the preset convergence condition, the neural network model of the completed model parameters is updated and determined as the rim lateral stiffness model.

2. The rim lateral stiffness model training method according to claim 1, characterized in that: The lateral stiffness test of the vehicle rim is performed to obtain the rim test data, including: placing a vehicle rim corresponding to the first rim structural parameter on a test bench; A lateral stiffness test is performed on the vehicle rim to determine a first test lateral stiffness corresponding to the first rim structural parameter.

3. The rim lateral stiffness model training method according to claim 2, characterized in that: The step of performing a lateral stiffness test on the vehicle rim to determine a first test lateral stiffness corresponding to the first rim structural parameter comprises: Performing a tapping test on a vehicle rim corresponding to the first rim structural parameter within a target frequency range to determine a resonance peak-to-peak frequency and an anti-resonance peak-to-peak frequency of each of the vehicle rims; A first test lateral stiffness corresponding to the first rim structural parameter is determined according to the rim mass of the vehicle rim, the resonance peak-to-peak frequency, and the anti-resonance peak-to-peak frequency.

4. A method for optimizing wheel rim structural parameters, characterized in that: include: Obtaining the second rim structural parameters; Inputting the second rim structural parameter into the rim lateral stiffness model determined by the rim lateral stiffness model training method according to any one of claims 1 to 3, and determining a second predicted lateral stiffness corresponding to the second rim structural parameter; If the second predicted lateral stiffness does not meet the target stiffness condition, the second rim structural parameters are optimized, the optimized second rim structural parameters are updated as the second rim structural parameters, and the second rim structural parameters are input into the rim lateral stiffness model determined by the rim lateral stiffness model training method according to any one of claims 1 to 3 to determine the second predicted lateral stiffness corresponding to the second rim structural parameters; If the second predicted lateral stiffness meets the target stiffness condition, the second rim structural parameters are determined as target rim structural parameters.

5. The method for optimizing the rim structural parameters according to claim 4, characterized in that: After determining the second predicted lateral stiffness corresponding to the second rim structural parameter, the rim structural parameter optimization method further includes: Determining a target prediction error according to the second predicted lateral stiffness and a preset target rim lateral stiffness; If the target prediction error does not exceed a preset error range, determining that the second predicted lateral stiffness meets a target stiffness condition; If the target prediction error exceeds a preset error range, it is determined that the second predicted lateral stiffness does not meet a target stiffness condition.

6. The method for optimizing the rim structural parameters according to claim 4, characterized in that: The step of optimizing the second rim structural parameters and updating the optimized second rim structural parameters as the second rim structural parameters comprises: Obtaining a target weight corresponding to each second rim structural parameter in the rim lateral stiffness model; The first N second rim structural parameters with larger target weights are optimized, and the optimized second rim structural parameters are updated as the second rim structural parameters, wherein N≥2.

7. A test device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the rim lateral stiffness model training method as described in any one of claims 1 to 3 is implemented, or, when the processor executes the computer program, the rim structural parameter optimization method as described in any one of claims 4 to 6 is implemented.

8. A testing system, characterized in that: It comprises a test bench and the test equipment as claimed in claim 7, and is used to implement the rim lateral stiffness model training method as claimed in any one of claims 1 to 3, or to implement the rim structure parameter optimization method as claimed in any one of claims 4 to 6.

9. The test system according to claim 8, wherein the test bench comprises a gantry, an elastic rope, a connection assembly and an acceleration sensor, wherein: The two ends of the elastic rope are respectively fixed on the gantry for hanging the vehicle rim; the connecting assembly is equipped on the vehicle rim for performing a lateral stiffness test on the vehicle rim corresponding to the first rim structural parameter; the acceleration sensor is fixed on the connecting assembly for obtaining a first test lateral stiffness corresponding to the first rim structural parameter when performing a lateral stiffness test on the vehicle rim within a target frequency range.

10. The test system according to claim 9, wherein the connection assembly comprises a conical bolt and a conical connection block, wherein: The conical bolt comprises a conical body and a stud extending from a first side of the conical body, a second side of the conical body is clamped with the conical connecting block, and the stud is screwed on the vehicle rim.

Images