A vehicle road noise analysis method based on physical tire model

Abstract

The present invention provides a whole vehicle road noise analysis method based on a physical tire model, and establishes a random vibration road noise analysis method that combines road noise sensitive frequency band extraction based on a virtual sample database with noise normal distribution fitting, thereby overcoming the problems of being unable to evaluate the noise calculation result curve deviation caused by tire parameter uncertainty when conducting whole vehicle virtual road noise simulation based on a physical tire model, and being unable to ensure the effectiveness of virtual road technology in predicting NVH performance results in the early stage of vehicle development.

Description

Technical Field

[0001] The present invention belongs to the field of automobile NVH simulation technology, and in particular relates to a vehicle road noise analysis method based on a physical tire model. Background Art

[0002] With the continuous development of the automotive industry, NVH performance has become an important indicator for users to evaluate the ride comfort and product satisfaction of automobiles. The NVH problem is the noise transmitted to the passenger compartment due to the tires being excited by the unevenness of the road surface. The body vibration caused by the unevenness of the road surface is inevitable. The vibration noise mainly generated by the body panels is in the low and medium frequency range, which is a frequency band that users can easily perceive. For new energy vehicles, since there is no engine excitation noise, the contribution of road noise to the noise level of the whole vehicle is more prominent.

[0003] The use of whole-vehicle virtual road noise simulation technology based on physical tire models can quickly troubleshoot typical noise problems of the design model and find optimization directions in the early data stage of development, reduce sample manufacturing and sample vehicle testing rounds, save R&D costs and improve development efficiency. However, due to the consistency of tire products, model errors and other reasons, the parameters inversely calculated when modeling physical tires are uncertain, and virtual road noise simulation technology has strong tire parameter sensitivity. Therefore, the uncertainty of tire model parameters will cause the deterministic whole-vehicle noise curve obtained through a one-time calculation using the established tire model to not match the actual situation in the local frequency band, and even calculate singular values. The simulation results cannot effectively describe the actual road noise response of the whole vehicle as a whole, which is also the reason that restricts the large-scale promotion and application of virtual road noise simulation technology. Summary of the invention

[0004] In view of this, in order to address the uncertainty problem of physical tire model parameters, a random vibration road noise analysis method combining road noise sensitive frequency band extraction based on a virtual sample database and noise normal distribution fitting is established, which has important guiding significance for the high reliability prediction of vehicle road noise NVH.

[0005] To achieve the above object, the technical solution of the present invention is achieved as follows:

[0006] A vehicle road noise analysis method based on a physical tire model comprises the following steps:

[0007] S1. Build a physical tire model;

[0008] S2. Build a vehicle road noise simulation environment;

[0009] S3, perturbation processing of the constructed physical tire model;

[0010] S4, extracting road noise sensitive frequency bands;

[0011] S5. Screening sensitive tire parameters to screen out key physical parameters corresponding to the tire model in sensitive frequency bands;

[0012] S6. Based on the sensitive frequency band and the sensitive tire parameters, a proxy model is established between the tire parameter design variables and the noise value at each frequency in the corresponding sensitive frequency band;

[0013] S7, randomly extracting multiple groups of samples related to sensitive tire parameters, and calculating the noise value at each frequency in the sensitive frequency band of each sample based on the proxy model;

[0014] S8. Using the average noise value of each group in the samples randomly collected in step S7 as a reference, perform normal distribution fitting on the noise distribution at each sensitive frequency to obtain the final simulation result of the sensitive frequency band.

[0015] Furthermore, the specific method of step S1 is as follows:

[0016] After the model is selected, the static, steady-state and dynamic mechanical properties of the tire are tested, and the test data is imported into the parameter identification software. The tire parameters are inversely calculated based on the comparison between the simulation and test curves to construct a real 3D nonlinear physical tire model.

[0017] Furthermore, the specific method of step S2 is as follows:

[0018] Establish a whole vehicle sound-solid coupling simulation model, linearize the 3D physical tire model and assemble it with the whole vehicle model, input the real NVH road surface PSD, define the simulation conditions, and calculate the full-band noise value of the measuring point.

[0019] Furthermore, the specific method of step S3 is as follows:

[0020] The key physical parameters in the initial tire model are numerically disturbed respectively, and the full-band noise value of the measuring point is obtained through vehicle simulation.

[0021] Furthermore, the specific method of step S4 is as follows:

[0022] The full-band noise difference curves corresponding to the tire model with various parameters disturbance and the initial tire model are obtained, and the sensitive frequency bands of the noise difference are obtained through cluster analysis method.

[0023] Furthermore, the specific method of step S5 is as follows:

[0024] The noise RMS value in the sensitive frequency band is defined as the output response, the output response contribution of each key physical parameter of the tire model is calculated, and several parameters with large contribution are selected as design variables.

[0025] Furthermore, before executing step S6, sampling work of the experimental design needs to be performed, and the specific method is as follows:

[0026] Define the design domain of sensitive tire parameters, extract the basic sample pool and the test sample pool based on the experimental design method, obtain the full-band noise value of each sample measurement point through vehicle simulation, and extract the noise value at each frequency in the corresponding sensitive frequency band.

[0027] Furthermore, after the proxy model is established in step S6, it is also necessary to select the proxy model with the highest prediction accuracy of the noise value of each sensitive frequency based on the test samples.

[0028] Furthermore, the process of performing random sampling in step S7 is as follows:

[0029] Thirty samples were randomly selected in a single group in the design domain of sensitive tire parameters, and a total of 100 groups of sampling tests were completed.

[0030] Furthermore, the noise mean value at each frequency of the non-sensitive frequency band corresponding to each basic sample is taken as the final simulation result of the non-sensitive frequency band.

[0031] Compared with the prior art, the vehicle road noise analysis method based on the physical tire model of the present invention has the following advantages:

[0032] The whole vehicle road noise analysis method based on the physical tire model described in the present invention is based on a constructed whole vehicle virtual road noise simulation environment. By extracting the road noise sensitive frequency band and fitting the noise normal distribution, a full-band noise simulation result curve containing probability statistical information is obtained. The method can effectively solve the problems of large deviation range of virtual road noise simulation results caused by tire parameter uncertainty and inability to ensure high reliability of simulation results, and provide a scientific basis for NVH optimization and tire selection of whole vehicle products. The present invention has low requirements on the user's engineering experience, saves human resource costs to a certain extent, and reduces R&D costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] The accompanying drawings constituting a part of the present invention are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the accompanying drawings:

[0034] Figure 1 This is a flow chart of the vehicle road noise analysis method based on the physical tire model. DETAILED DESCRIPTION

[0035] It should be noted that, in the absence of conflict, the embodiments of the present invention and the features in the embodiments may be combined with each other.

[0036] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside" and the like indicate positions or positional relationships based on the positions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as limiting the present invention. In addition, the terms "first", "second", and the like are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, features defined as "first", "second", and the like may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "multiple" means two or more.

[0037] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood by specific circumstances.

[0038] The present invention will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0039] like Figure 1 As shown, this embodiment takes a certain SUV road noise simulation as an example to further illustrate the technical solution of the present invention, which specifically includes the following steps:

[0040] Step 1: Carry out tests on the mass inertia, cross-sectional geometry, static and dynamic radial stiffness, longitudinal stiffness, lateral stiffness, static torsional stiffness, lateral slip, longitudinal slip, bump impact, etc. of a tire assembled on an SUV. Import the test data into the CDTire / PI tire modeling software. After entering the basic tire information to establish a cross-sectional geometry model, identify the tire parameters by comparing the errors and trends between the simulation curve and the test curve. Output the CDTire / 3D nonlinear physical tire model with the suffix CDT50.

[0041] Step 2: Establish finite element models of each subassembly of a SUV vehicle under the Optistruct solver of Hyperworks software. Switch to the NVHD module of the software, flexibly connect the subassemblies together through the CBUSH unit, and generate a sound cavity model based on the interior and body. After setting the rim inertia, tire pressure, wheel load, tire contact area and vehicle speed in the Tire manager interface, linearize the CDTire / 3D tire model. After entering the rough asphalt road elevation information in the PSD Catalog interface, calculate the self-power spectrum and cross-power spectrum of each wheel contact road surface. Define the sound pressure level at the driver's outer ear as the noise response. Determine the simulation status of the whole vehicle in the Analysis Manager interface, establish the analysis job, complete the construction of the whole vehicle road noise simulation environment, and obtain the initial full-band road noise simulation results.

[0042] Step 3: Increase the initial values ​​of physical parameters such as crown mass, bead mass, sidewall mass, rubber shear stiffness and damping coefficient, rubber diagonal stiffness and damping coefficient, rubber lateral stiffness and damping coefficient, rubber circumferential stiffness and damping coefficient, carcass stiffness and damping coefficient, cover layer stiffness and damping coefficient, metal belt layer stiffness and damping coefficient, X-direction bending stiffness and damping coefficient, Y-direction bending stiffness and damping coefficient, diagonal bending stiffness and damping coefficient, crown radial stiffness, crown lateral shear stiffness, and crown circumferential shear stiffness in the CDT50 tire model file by 20% to obtain the corresponding numerically perturbed CDTire / 3D model, import it into the whole vehicle road noise simulation environment to run and solve to obtain the corresponding road noise calculation results, and generate a disturbed tire model sample database.

[0043] Step 4: Calculate and obtain the full-band noise difference curve of the whole vehicle corresponding to each parameter perturbed CDTire / 3D tire model and the initial CDTire / 3D tire model, and normalize each difference curve. Use the fuzzy C-means clustering method to perform data clustering along the ordinate axis (sound pressure level axis) of each normalized difference curve, segment the curve based on the clustering result, and take the RMS value of the normalized sound pressure level in each segment to achieve data simplification. The clustered difference curves corresponding to all perturbed CDTire / 3D tire model samples are accumulated using formula (1):

[0044]

[0045] In the formula, With f tRespectively represent the sound pressure level values ​​of the clustered difference curve and the cumulative difference curve corresponding to the lth disturbed tire model sample at frequency t, and m is the number of samples. The fuzzy C-means clustering method is used again to cluster the cumulative difference curve data along the horizontal axis (frequency axis). Finally, the RMS value of the sound pressure level on the secondary cumulative curve is used as the threshold. If the RMS value of the sound pressure level of a certain frequency band obtained after the previous clustering process is not less than the threshold, the frequency band is extracted as the sensitive frequency band.

[0046] Step 5: Calculate the noise sensitivity of each tire physical parameter in step 3 based on equation (2):

[0047]

[0048] In the formula, f l is the noise sensitivity of the lth tire model parameter, is the mean of the RMS values ​​of the noise in each sensitive frequency band of the road noise simulation results of the lth disturbance CDTire / 3D model, is the mean of the RMS value of the noise in each sensitive frequency band of the road noise simulation results of the initial CDTire / 3D model. Several parameters with high noise sensitivity are selected as design variables.

[0049] Step 6: Define the upper and lower limits of each tire parameter design variable based on the actual tolerance, select the Latin hypercube experimental design method in the DOE module of the HyperStudy software to extract basic tire model samples and test tire model samples in the tire parameter design domain, and import them into the vehicle road noise simulation environment to run the solution to obtain the corresponding noise simulation curve, extract the noise value at each frequency in the sensitive frequency band, and generate the basic tire model sample database and the test tire model sample database.

[0050] Step 7: Based on the basic tire model sample database, the polynomial response surface fitting method, radial basis function and Kriging interpolation method are used to establish a proxy model between the tire parameter design variables and the noise value at each frequency in the corresponding vehicle sensitive frequency band. By predicting the noise value at the sensitive frequency of the test tire model sample, the fitting / interpolation method with the highest R value is selected based on formula (3) as the proxy model for predicting the noise value of each sensitive frequency.

[0051] R=(R 2 -RAAE-RMAE) / 3

[0052]

[0053] Where f(x) is the noise simulation value of the test tire model sample x at a certain sensitive frequency, is the corresponding proxy model prediction value, and It is the average value of noise simulation value of q test samples at a certain sensitive frequency.

[0054] Step 8: Based on the Monte Carlo simulation method, 30 samples are randomly selected in a single group in the tire parameter design domain, and a total of 100 groups of sampling tests are completed. The noise values ​​of all samples at all sensitive frequencies in each group of tests are calculated based on the proxy model established in step 7.

[0055] Step 9: Count the average noise value at each sensitive frequency in each group of experiments in step 8 and its occurrence frequency, make a frequency histogram, and use equations (4) and (5) to calculate the noise expectation μ and standard deviation σ of 100 groups of random sampling at each sensitive frequency.

[0056]

[0057]

[0058] In the formula, the superscript s is the sth sensitive frequency. The frequency histogram is fitted with a normal distribution using Origin software, and the noise probability density function h(y) and distribution function H(y) at each sensitive frequency are obtained as shown in formula (6) and formula (7), respectively.

[0059]

[0060]

[0061] Step 10: Calculate and obtain the noise mean value at the non-sensitive frequency of each basic tire model sample in step 6, and obtain the noise value at the 90% confidence level of each sensitive frequency based on formula (7) and table lookup. The two are combined to obtain the final vehicle road noise full-band simulation curve.

[0062] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A vehicle road noise analysis method based on a physical tire model, characterized in that: The following steps are involved: S1. Build a physical tire model; S2. Build a vehicle road noise simulation environment; S3, perturbation processing of the constructed physical tire model; S4, extracting road noise sensitive frequency bands; The full-band noise difference curves of the whole vehicle corresponding to the CDTire / 3D tire model with each parameter disturbance and the initial CDTire / 3D tire model are calculated and normalized. The fuzzy C-means clustering method is used to perform data clustering along the ordinate axis of each normalized difference curve. The curve is segmented based on the clustering result, and the RMS value of the normalized sound pressure level is taken in each segment to achieve data simplification. The clustered difference curves corresponding to all the disturbed CDTire / 3D tire model samples are accumulated by formula (1): In the formula, With f t Respectively represent the sound pressure level values ​​of the clustered difference curve and the cumulative difference curve corresponding to the first disturbed tire model sample at frequency t, and m is the number of samples; the fuzzy C-means clustering method is used again to cluster the cumulative difference curve data along the horizontal axis; finally, the RMS value of the sound pressure level on the secondary cumulative curve is used as the threshold; if the RMS value of the sound pressure level of a certain frequency band obtained after the previous clustering process is not less than the threshold, the frequency band is extracted as the sensitive frequency band; S5. Screening sensitive tire parameters to screen out key physical parameters corresponding to the tire model in sensitive frequency bands; The noise sensitivity of each tire physical parameter in step S5 is calculated based on formula (2): In the formula, S l is the noise sensitivity of the first tire model parameter, is the mean of the RMS values ​​of the noise in each sensitive frequency band of the road noise simulation results of the first disturbance CDTire / 3D model, The average of the noise RMS values ​​of the corresponding road noise simulation results of the initial CDTire / 3D model in each sensitive frequency band, and several parameters with high noise sensitivity are selected as design variables; Based on the actual tolerance, the upper and lower limits of each tire parameter design variable are defined. In the DOE module of HyperStudy software, the Latin hypercube experimental design method is selected to extract basic tire model samples and test tire model samples in the tire parameter design domain. These samples are then imported into the vehicle road noise simulation environment to run the solution to obtain the corresponding noise simulation curve, extract the noise value at each frequency in the sensitive frequency band, and generate a basic tire model sample database and a test tire model sample database. S6. Based on the sensitive frequency band and the sensitive tire parameters, a proxy model is established between the tire parameter design variables and the noise value at each frequency in the corresponding sensitive frequency band; S7, randomly extracting multiple groups of samples related to sensitive tire parameters, and calculating the noise value at each frequency in the sensitive frequency band of each sample based on the proxy model; S8, using the average noise value of each group in the samples randomly collected in step S7 as a reference, respectively performing normal distribution fitting on the noise distribution at each sensitive frequency to obtain the final simulation result of the sensitive frequency band; Count the average noise value and its occurrence frequency at each sensitive frequency in each group of experiments, make a frequency histogram, and count the noise expectation μ of 100 groups of random sampling at each sensitive frequency s With standard deviation σ s ; Origin software was used to fit the frequency histogram to a normal distribution and obtain the noise probability density function h at each sensitive frequency. s (y) and the distribution function H s (y) are shown in formula (6) and formula (7) respectively; The noise mean value at the non-sensitive frequency of each basic tire model sample is calculated, and the noise value at the 90% confidence level of each sensitive frequency is obtained based on formula (7) and table lookup. The two are combined to obtain the final full-band simulation curve of vehicle road noise.

2. The vehicle road noise analysis method based on a physical tire model according to claim 1, characterized in that: The specific method of step S1 is as follows: After the model is selected, the static, steady-state and dynamic mechanical performance tests of the tire are carried out, and the test data is imported into the parameter identification software. The tire parameters are inversely calculated based on the comparison between the simulation and test curves to construct a real 3D nonlinear physical tire model.

3. The vehicle road noise analysis method based on a physical tire model according to claim 1, characterized in that: The specific method of step S2 is as follows: Establish a whole vehicle sound-solid coupling simulation model, linearize the 3D physical tire model and assemble it with the whole vehicle model, input the real NVH road surface PSD, define the simulation conditions, and calculate the full-band noise value of the measuring point.

4. The vehicle road noise analysis method based on a physical tire model according to claim 1, characterized in that: The specific method of step S3 is as follows: The key physical parameters in the initial tire model are numerically disturbed respectively, and the full-band noise value of the measuring point is obtained through vehicle simulation.

5. The vehicle road noise analysis method based on a physical tire model according to claim 1, characterized in that: Before executing step S6, sampling work of the experimental design needs to be performed. The specific method is as follows: Define the design domain of sensitive tire parameters, extract the basic sample pool and test sample pool based on the experimental design method, obtain the full-band noise value of each sample measurement point through vehicle simulation, and extract the noise value at each frequency in the corresponding sensitive frequency band.

6. The vehicle road noise analysis method based on a physical tire model according to claim 5, characterized in that: After the proxy model is established in step S6, it is also necessary to select the proxy model with the highest prediction accuracy of the noise value of each sensitive frequency based on the test samples.

7. The vehicle road noise analysis method based on a physical tire model according to claim 6, characterized in that: The process of performing random sampling in step S7 is as follows: Thirty samples were randomly selected in a single group in the design domain of sensitive tire parameters, and a total of 100 groups of sampling tests were completed.

Images