A method for optimizing sound field characteristics and controlling noise based on the arrangement of resonance cavities inside a vehicle

Through the transmission path analysis method and the optimization of the resonance cavity structure, the problem of difficulty in controlling low-frequency noise in traditional technologies is solved, and effective noise reduction is achieved on the transmission path of the main noise contribution in the car, improving the sound quality and driving comfort in the car.

CN115064145BActive Publication Date: 2024-11-19WUHAN UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202210151945.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-02-18
Publication Date
2024-11-19
Estimated Expiration
2042-02-18

AI Technical Summary

Technical Problem

Traditional acoustic packaging technology is difficult to effectively control low-frequency noise below 300Hz, affecting the sound quality and driving comfort in the car.

Method used

Transmission path analysis method (TPA) is used to extract the transmission path of the main noise contribution, and the resonance cavity structure is set on the main transmission path of the vehicle noise, and the resonance cavity geometric parameters are optimized using LMSVirtual Lab simulation analysis software and generalized regression neural network (GRNN) , establish a nonlinear functional relationship and reduce the sound pressure level at the peak frequency.

Benefits of technology

Effectively reduce the low-frequency noise level in the car, improve calculation accuracy, enhance noise control effect, and improve driving comfort.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115064145B_ABST
    Figure CN115064145B_ABST
Patent Text Reader

Abstract

The present invention provides a method for optimizing the sound field characteristics and controlling noise based on the arrangement of the resonance cavity in the vehicle. The transmission path of the main noise contribution is extracted by the transmission path analysis method, and several paths with significantly higher noise contribution than other transmission paths are selected, and the transmission paths of non-critical noise contribution are eliminated, so as to reduce the amount of calculation and improve the calculation accuracy. By arranging the resonance cavity on the main transmission path of the vehicle noise and optimizing the structure, the transmission function on the key path of noise transmission is improved, so as to achieve the purpose of noise reduction in the low frequency band.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention belongs to the technical field of automobile design, and in particular relates to a method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle. Background Art

[0002] The quality of vehicle sound directly affects the comfort of drivers and passengers, and has been used by various manufacturers as an important indicator of the comfort of automotive products. The evaluation of sound quality usually adopts the size of noise (i.e. loudness) and sound characteristics (sharpness, roughness, etc.). Traditional technical means such as acoustic packaging have a good control effect on medium and high frequency noise above 300Hz, but the control effect on low frequency noise below 300Hz in the car is not ideal. The present invention proposes to use the transfer path analysis (TPA) method to extract the transfer path of the main noise contribution (generally select several paths with significantly higher noise contribution than other transfer paths, eliminate the transfer paths of non-critical noise contribution, reduce the amount of calculation, and improve the calculation accuracy). By arranging the resonant cavity on the main noise transmission path in the car and optimizing the structure, the transfer function on the key path of noise transmission is improved to achieve the purpose of noise reduction in the low frequency band. Summary of the invention

[0003] In view of the problems existing in the prior art, the technical solutions adopted by the present invention to solve the problems existing in the prior art are as follows:

[0004] A method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle, characterized by comprising the following steps:

[0005] Step 1: Use the transfer path analysis (TPA) method to determine the noise contribution of the transfer path from the excitation source to the response point (the noise amount at the response point = the excitation force at the excitation point * the transfer function of the transfer path), and on this basis extract the transfer function characteristics of the main contribution transfer path;

[0006] Step 2: Design the in-vehicle acoustic cavity structure as a resonant cavity structure to achieve its noise reduction function, use LMS Virtual Lab simulation analysis software to analyze the acoustic mode of the in-vehicle resonant cavity, and thereby determine the key geometric parameter variables that affect the modal frequency of the resonant cavity and the key resonant cavity position and key geometric parameters that affect the in-vehicle sound field distribution, and establish a sample database for the key geometric parameters;

[0007] Step 3: Establish a nonlinear function correspondence between the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle obtained by the transfer path analysis (TPA) method in step 1 and the geometric parameters of the resonance cavity, and achieve the purpose of reducing the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle by optimizing the design of the resonance cavity structure and reasonably arranging the position.

[0008] The specific steps of extracting the noise transfer path contribution using the transfer path analysis (TPA) method in step 1 are as follows:

[0009] Step 1.1 According to the specific vehicle model and the test objectives, determine the number of excitation points, the number of reference points and their specific locations, and build the test environment;

[0010] Step 1.2 On a closed test road, shift gears and accelerate slowly, and use the LMS test software to measure the in-vehicle noise signals at the driver's right ear DR and the right rear passenger PL to obtain the in-vehicle vibration noise signals under typical working conditions;

[0011] Step 1.3 measures and saves the acceleration data signals of the three degrees of freedom of the excitation point, and obtains the working condition data of the excitation point position under typical working conditions. This step can be performed simultaneously in step 1.2;

[0012] Step 1.4 measures and saves the acceleration data signals of the three degrees of freedom of the reference point, and obtains the working condition data of the reference point position under typical working conditions. This step can be performed simultaneously in step 1.2;

[0013] Step 1.5 Test the transfer function between the excitation point-target point, excitation point-reference point, and excitation point-excitation point for analysis of the excitation force and contribution of each path;

[0014] Step 1.6 In the TPA analysis module, calculate the excitation force and analyze the contribution of each path. Based on the analysis results, you can know the most important contribution path and the excitation force at the starting point of each path;

[0015] Step 1.7 Use spectrum analysis and other methods to determine the frequency that causes the noise or vibration peak, and analyze the sound field characteristics of the vehicle interior noise target under typical working conditions;

[0016] Step 1.8 extracts and analyzes the sound field transfer function characteristics of the main noise contribution transfer path in the vehicle, and obtains the sound pressure level at the peak frequency of the main noise contribution transfer path according to the frequency characteristics of the sound field in the vehicle.

[0017] The sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle in step 1 can be obtained by using the Transfer Path Analysis in the test software LMS, using the transfer path analysis method (TPA) during the noise transfer path contribution test, and at the same time, a sample database is established for multiple groups of data.

[0018] The step 2 specifically includes the following steps:

[0019] Step 2.1, perform finite element modeling on the resonance cavity structure formed by enclosing the body panels inside the vehicle;

[0020] The modal analysis of the resonant cavity is: The final resonant frequency is Where Z is the acoustic impedance, which is a complex number, and the imaginary part is the acoustic impedance, real part is the acoustic impedance; ω is the resonant frequency, ρ is the air density, L is the length of the capillary, S is the cross-sectional area of ​​the capillary, V is the cavity volume, c is the speed of sound, and f n is the modal resonance frequency, j is the sign of the imaginary part;

[0021] Step 2.2, using the finite element method to analyze the influence of different resonance cavity geometric parameters on the resonance cavity modal frequency. The acoustic mode of the resonance cavity is determined by its geometric parameters, including the resonance cavity appearance parameters and internal structure parameters. The influence of these two parts of parameters on the acoustic cavity modal frequency is analyzed separately, so as to obtain the change relationship of the resonance cavity modal frequency caused by each variable acting alone, determine the key geometric parameter variables that affect the resonance cavity modal frequency, and establish a sample database for the key geometric parameters;

[0022] Step 2.3: Based on the original resonance cavity model, individually change the geometric parameters of the acoustic cavity shape and the seat position parameters, analyze the changes in the sound pressure level response in the vehicle, select the acoustic cavity models under several specific geometric parameter states, and obtain the sound pressure level response curve of the measuring point position in the vehicle, so as to obtain the influence of the geometric parameters of the resonance cavity on the sound field distribution, determine the key geometric parameters of the key resonance cavity positions that affect the sound field distribution in the vehicle, and establish a sample database for the key geometric parameters.

[0023] In step 3, the generalized regression neural network (GRNN) algorithm is used to self-learn the data in the existing samples through the neural network algorithm, identify the functional correspondence between the input and output data, and establish a nonlinear functional correspondence between the sound pressure level at the peak frequency of the transmission path of the main noise contribution in the vehicle and the geometric parameters of the resonance cavity, where the input quantity is X=(x1, x2,…, x n ) T , the corresponding output is Y=(y1,y2,…,y n ) T The number of neurons in the input layer is equal to the dimension of the input vector in the learning sample. Each neuron is a simple distribution unit that directly passes the input variable to the pattern layer.

[0024] The number of neurons in the pattern layer is equal to the number of learning samples n, and its transfer function is:

[0025]

[0026] p i is the mode layer transfer function, X is the network input variable; Xi is the learning sample corresponding to the i-th neuron, σ is the smoothing factor;

[0027] The output values ​​of all nodes in a type of summation layer are summed, and its transfer function is:

[0028]

[0029] S m Transfer function for the summation layer;

[0030] Another type of output layer for weighted summation is expressed as:

[0031]

[0032] S Nk is the transfer function of the weighted summation, y ik is the weight output between the i-th neuron in the pattern layer and the k-th molecule in the summation layer.

[0033] The number of neurons in the output layer is equal to the two-digit number k in the output direction of the learning sample. Each neuron divides the output of the summation layer. The output of neuron k corresponds to the estimated result The kth element of , that is:

[0034]

[0035] The regression value of y corresponding to x:

[0036]

[0037] x is a random variable (the observed value is x0), f(x, y) is the joint probability density function, x and y are two random variables;

[0038] Using Parzen nonparametric estimation, the unknown density function is:

[0039]

[0040] Where, X i , Y i are the sample observations of x and y, p is the dimension of x, n is the sample size, and σ is the smoothing factor (Gaussian function width):

[0041]

[0042] because The above formula can be rewritten as:

[0043]

[0044] is the total sample observation value Yi The weighted sum of

[0045] The specific steps of optimizing the geometric parameters of the resonant cavity in step 3 are as follows:

[0046] The mathematical model of “resonance cavity geometry parameters – sound pressure level at peak frequency of the main noise contribution transfer path in the vehicle” is regarded as a multi-input multi-output system, and the function correspondence is Y = F(X) (12), where Y = (y1, y2, …, y n ) T is the sound pressure level (dB) at the peak frequency of the main contribution to the vehicle interior noise transmission path, X = (x1, x2, ..., x n ) T is the resonant cavity geometric parameter variable. A generalized regression neural network (GRNN) controller is established using MATLAB to train the input and output sample data and establish a functional correspondence.

[0047] The function relationship of the optimization problem of the in-vehicle acoustic environment is Y = F (X). The optimization process is to solve the geometric parameters of the resonant cavity (input vector X = (x1, x2, ..., x n ) T ), so that the sound pressure level at the peak frequency (output vector Y = (y1, y2, ..., y n ) T ) reaches the minimum or extreme value, which is a global optimization problem. Genetic algorithm (GA) is used to solve the problem of optimizing the acoustic environment of the car's interior sound field.

[0048] The modeling idea of ​​optimizing the sound field environment is: first, use the GRNN network model to establish the functional correspondence between the input and output of the in-vehicle acoustic environment Y = F (X), and then use GA to search for the acoustic cavity geometry parameters X = (x1, x2, ..., x n ) T Input into the trained GRNN network model to predict the corresponding sound pressure level Y=(y1,y2,…,y n ) T , then Y=(y1,y2,…,y n ) T The minimum value (extreme value) of is converted into the fitness function of GA, and the genetic operation of survival of the fittest is performed according to its adaptability until Y=(y1,y2,…,y n ) T The optimal solution set is obtained, and the corresponding acoustic cavity geometry parameters X = (x1, x2, ..., x n )T .

[0049] The optimization of the sound pressure level at the peak frequency of the main noise contribution transfer path in the vehicle can be constructed into a multi-objective optimization problem. With F(X) as the objective function vector, the mathematical model can be described as:

[0050]

[0051] Where minF(X) is the objective function for achieving the best acoustic environment, that is, the sound pressure level (dB) at the peak frequency of the main noise contribution transmission path in the car reaches the minimum value, x in and x im are the upper and lower limits of the i-th design variable of the acoustic cavity geometry parameter, respectively. The Gamultiobj function based on the GA algorithm in the Matlab optimization toolbox is used to optimize the multi-objective problem and obtain multiple groups of non-inferior solutions. Finally, each group of non-inferior solutions is optimized and screened. The optimization objective function is:

[0052] f(x)=Var(minF(X))+τ mn E(minF(X)) (14)

[0053] In the formula: Var—represents the calculated variance, τ mn — represents the weight coefficient, E— represents the average value, and the geometric parameters of the acoustic cavity (resonance cavity) corresponding to the minimum sound pressure level at the peak frequency of the transmission path of the main noise contribution in the car after optimization are used as the resonant cavity design indicators to achieve the purpose of noise reduction in the car sound field.

[0054] The present invention has the following advantages:

[0055] 1. The structural parameters and layout of the resonance cavity are currently mostly determined based on simulation, which cannot meet the actual measurement requirements. This application mainly determines and improves the structural parameters and layout of the resonance cavity based on actual measurement data, which is more realistic than software simulation analysis;

[0056] 2. Select the noise signal of the main noise transmission path, which is conducive to simplifying the calculation while meeting the requirements of noise reduction;

[0057] 3. Using the transfer path analysis (TPA) method for measurement is beneficial to place the resonant cavity on the main noise transmission path and change the transfer function, so that it can not only reduce noise by changing the transfer function, but also achieve noise elimination by changing the structural parameters. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 This is a diagram of the test steps for the TPA experiment;

[0059] Figure 2 This is a diagram of the steps of modal analysis;

[0060] Figure 3 Optimize the flow chart for the in-car sound field;

[0061] Figure 4 The present invention is a flow chart for designing a resonance cavity in a vehicle. DETAILED DESCRIPTION

[0062] The technical solution of the present invention is further specifically described below through embodiments and in conjunction with the accompanying drawings. Taking the low-frequency noise of a vehicle as the research object, the transfer path analysis method (TPA) is used to determine the noise contribution of the transfer path from the excitation source to the response point (the noise amount at the response point = the excitation force at the excitation point * the transfer function of the transfer path), and on this basis, the transfer function characteristics of the main contribution transfer path are extracted. The method steps of extracting the noise transfer path contribution by the transfer path analysis method (TPA) are as follows:

[0063] Step 1.1 According to the specific vehicle model and the test objectives, determine the number of excitation points, the number of reference points and their specific locations, and build the test environment;

[0064] Step 1.2 On a closed test road, shift gears and accelerate slowly, and use the LMS test software to measure the in-vehicle noise signals at the driver's right ear DR and the right rear passenger PL to obtain the in-vehicle vibration noise signals under typical working conditions;

[0065] Step 1.3 measures and saves the acceleration data signals of the three degrees of freedom of the excitation point, and obtains the working condition data of the excitation point position under typical working conditions. This step can be performed simultaneously in step 1.2;

[0066] Step 1.4 measures and saves the acceleration data signals of the three degrees of freedom of the reference point, and obtains the working condition data of the reference point position under typical working conditions. This step can be performed simultaneously in step 1.2;

[0067] Step 1.5 Test the transfer function between the excitation point-target point, excitation point-reference point, and excitation point-excitation point for analysis of the excitation force and contribution of each path;

[0068] Step 1.6 In the TPA analysis module, calculate the excitation force and analyze the contribution of each path. Based on the analysis results, you can know the most important contribution path and the excitation force at the starting point of each path;

[0069] Step 1.7 Use spectrum analysis and other methods to determine the frequency that causes the noise or vibration peak, and analyze the sound field characteristics of the vehicle interior noise target under typical working conditions;

[0070] Step 1.8 extracts and analyzes the sound field transfer function characteristics of the main noise contribution transfer path in the vehicle, and obtains the sound pressure level at the peak frequency of the main noise contribution transfer path according to the frequency characteristics of the sound field in the vehicle.

[0071] TPA experimental test steps are as follows Figure 1 shown.

[0072] Table 1 List of experimental equipment required for TPA test

[0073]

[0074]

[0075] Step 2: Design the sound cavity structure in the car as a resonant cavity structure to achieve its noise reduction function. Use LMS Virtual Lab simulation analysis software to analyze the acoustic mode of the resonant cavity in the car and determine the influence of the geometric parameters of the resonant cavity on the acoustic characteristics of the sound field in the car, such as Figure 2 As shown:

[0076] Step 2.1, perform finite element modeling on the resonance cavity structure formed by enclosing the body panels inside the vehicle;

[0077] The modal analysis of the resonant cavity is: The final resonant frequency is Where Z is the acoustic impedance, which is a complex number, and the imaginary part is the acoustic impedance, real part is the acoustic impedance; ω is the resonant frequency, ρ is the air density, L is the length of the capillary, S is the cross-sectional area of ​​the capillary, V is the cavity volume, c is the speed of sound, and f n is the modal resonance frequency, j is the sign of the imaginary part;

[0078] Step 2.2, using software simulation to analyze the influence of different resonance cavity geometric parameters on the resonance cavity modal frequency. The acoustic mode of the resonance cavity is determined by its geometric parameters, including the resonance cavity appearance parameters and internal structure parameters. The influence of these two parts of parameters on the acoustic cavity modal frequency is analyzed separately, so as to obtain the change relationship of the resonance cavity modal frequency caused by each variable acting alone, determine the key geometric parameter variables that affect the resonance cavity modal frequency, and establish a sample database for the key geometric parameters;

[0079] Step 2.3: Based on the original resonance cavity model, individually change the geometric parameters of the acoustic cavity shape and the seat position parameters, analyze the changes in the sound pressure level response in the vehicle, select the acoustic cavity models under several specific geometric parameter states, and obtain the sound pressure level response curve of the measuring point position in the vehicle, so as to obtain the influence of the geometric parameters of the resonance cavity on the sound field distribution, determine the key geometric parameters of the key resonance cavity positions that affect the sound field distribution in the vehicle, and establish a sample database for the key geometric parameters.

[0080] Step 3. Since the in-vehicle sound field noise is affected by the main transfer path noise, improving the transfer function characteristics on the key lines can effectively and conveniently improve the in-vehicle sound field noise. Therefore, a nonlinear function correspondence is established between the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle obtained by the transfer path analysis method (TPA) and the geometric parameters of the resonance cavity. By optimizing the design of the resonance cavity structure and layout position, the purpose of reducing the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle is achieved. The sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle can be obtained by using the Transfer Path Analysis in the test software LMS and the transfer path analysis method (TPA) during the noise transfer path contribution test, and a sample database is established for multiple sets of data.

[0081] This patent proposes to use the generalized regression neural network (GRNN) algorithm to identify the functional correspondence between input and output data by self-learning the data in the existing samples through the neural network algorithm, and establish a nonlinear functional correspondence between the sound pressure level at the peak frequency of the transmission path of the main noise contribution in the car and the geometric parameters of the resonant cavity, where the input quantity is X = (x1, x2, ..., x n ) T , the corresponding output is Y=(y1,y2,…,y n ) T The number of neurons in the input layer is equal to the dimension of the input vector in the learning sample. Each neuron is a simple distribution unit that directly passes the input variable to the pattern layer.

[0082] The number of neurons in the pattern layer is equal to the number of learning samples n, and its transfer function is:

[0083]

[0084] p i is the mode layer transfer function, X is the network input variable; X i is the learning sample corresponding to the i-th neuron, and σ is the smoothing factor.

[0085] The output values ​​of all nodes in a type of summation layer are summed up, and the summation layer transfer function S m for:

[0086]

[0087] Another type of output layer for weighted summation is expressed as:

[0088]

[0089] S Nkis the transfer function of the weighted summation, y ik is the weight output between the i-th neuron in the pattern layer and the k-th molecule in the summation layer.

[0090] The number of neurons in the output layer is equal to the two-digit number k in the output direction of the learning sample. Each neuron divides the output of the summation layer. The output of neuron k corresponds to the estimated result The kth element of , that is:

[0091]

[0092] The regression value of y corresponding to x:

[0093]

[0094] x is a random variable (the observed value is x0), f(x, y) is the joint probability density function, x and y are two random variables;

[0095] Using Parzen nonparametric estimation, the unknown density function is:

[0096]

[0097] Where, X i , Y i are the sample observations of x and y, p is the dimension of x, n is the sample size, and σ is the smoothing factor (Gaussian function width):

[0098]

[0099] because The above formula can be rewritten as:

[0100]

[0101] is the total sample observation value Y i The weighted sum of

[0102] The mathematical model of “resonance cavity geometry parameters – sound pressure level at peak frequency of the main noise contribution transfer path in the vehicle” is regarded as a multi-input multi-output system, and the function corresponding relationship is Y = F (X) (12), Where Y=(y1,y2,…,y n ) T is the sound pressure level (dB) at the peak frequency of the main contribution to the vehicle interior noise transmission path, X = (x1, x2, ..., x n ) Tis the resonant cavity geometric parameter variable. A generalized regression neural network (GRNN) controller is established using MATLAB to train the input and output sample data and establish a functional correspondence.

[0103] The function relationship of the optimization problem of the in-vehicle acoustic environment is Y = F (X), and the optimization result can be described as: Under certain external excitation conditions, solve the geometric parameters of the resonant cavity (input vector X = (x1, x2, ..., x n ) T ), so that the sound pressure level at the peak frequency (output vector Y = (y1, y2, ..., y n ) T ) reaches the minimum or extreme value, which belongs to the problem of global optimization. Considering the optimization idea of ​​genetic algorithm (GA) to deal with the problem of optimizing the acoustic environment of the car's interior sound field.

[0104] The modeling idea of ​​optimizing the sound field environment is: first, use the GRNN network model to establish the functional correspondence between the input and output of the in-vehicle acoustic environment Y = F (X), and then use GA to search for the acoustic cavity geometry parameters X = (x1, x2, ..., x n ) T Input into the trained GRNN network model to predict the corresponding sound pressure level Y=(y1,y2,…,y n ) T , then Y=(y1,y2,…,y n ) T The minimum value (extreme value) of is converted into the fitness function of GA, and the genetic operation of survival of the fittest is performed according to its adaptability until Y=(y1,y2,…,y n ) T The optimal solution set is obtained, and the corresponding acoustic cavity geometry parameters X = (x1, x2, ..., x n ) T , the optimization process flow is as follows Figure 3 As shown:

[0105] The optimization of the sound pressure level at the peak frequency of the main noise contribution transfer path in the vehicle can be constructed into a multi-objective optimization problem. With F(X) as the objective function vector, the mathematical model can be described as:

[0106]

[0107] Where minF(X) is the objective function for achieving the best acoustic environment, that is, the sound pressure level (dB) at the peak frequency of the main noise contribution transmission path in the car reaches the minimum value, x in and x imare the upper and lower limits of the i-th design variable of the acoustic cavity geometry parameter, respectively. The Gamultiobj function based on the GA algorithm in the Matlab optimization toolbox is used to optimize the multi-objective problem and obtain multiple groups of non-inferior solutions. Finally, each group of non-inferior solutions is optimized and screened. The optimization objective function is:

[0108] f(x)=Var(minF(X))+τ mn E(minF(X)) (14)

[0109] In the formula: Var—represents the calculated variance, τ mn — represents the weight coefficient, E— represents the average value, and the geometric parameters of the acoustic cavity (resonance cavity) corresponding to the minimum sound pressure level at the peak frequency of the main noise contribution transmission path in the optimized vehicle are used as the resonant cavity design index to achieve the purpose of reducing the noise in the vehicle sound field. Its design circuit is as follows Figure 4 shown.

[0110] The protection scope of the present invention is not limited to the above-mentioned embodiments. Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the scope and spirit of the present invention. If these changes and modifications fall within the scope of the claims of the present invention and their equivalents, the present invention is intended to include these changes and modifications.

Claims

1. A method for optimizing sound field characteristics and controlling noise based on the arrangement of resonance cavities in a vehicle, characterized in that: The steps include: Step 1: Use the transfer path analysis method to determine the transfer path noise contribution from the excitation source to the response point, and on this basis extract the transfer function characteristics of the main contribution transfer path; Step 2: Design the in-vehicle acoustic cavity structure as a resonant cavity structure to achieve its noise reduction function, use simulation analysis software to analyze the acoustic mode of the in-vehicle resonant cavity, and thereby determine the key geometric parameter variables that affect the modal frequency of the resonant cavity and the key resonant cavity position and key geometric parameters that affect the in-vehicle sound field distribution, and establish a sample database for the key geometric parameters; Step 3: Establish a nonlinear function correspondence between the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle obtained by the transfer path analysis method in step 1 and the geometric parameters of the resonance cavity, and achieve the purpose of reducing the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle by optimizing the design of the resonance cavity structure and reasonably arranging the position.

2. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle according to claim 1, characterized in that: The specific steps of extracting the noise transfer path contribution using the transfer path analysis method in step 1 are as follows: Step 1.1 According to the specific vehicle model and the test objectives, determine the number of excitation points, the number of reference points and their specific locations, and build the test environment; Step 1.2 On a closed test road, shift gears and accelerate slowly, and use the LMS test software to measure the in-vehicle noise signals at the driver's right ear DR and the right rear passenger PL to obtain the in-vehicle vibration noise signals under typical working conditions; Step 1.3 measures and saves the acceleration data signals of the three degrees of freedom of the excitation point, and obtains the working condition data of the excitation point position under typical working conditions. This step can be performed simultaneously in step 1.2; Step 1.4 measures and saves the acceleration data signals of the three degrees of freedom of the reference point, and obtains the working condition data of the reference point position under typical working conditions. This step can be performed simultaneously in step 1.2; Step 1.5 Test the transfer function between the excitation point-target point, excitation point-reference point, and excitation point-excitation point for analysis of the excitation force and contribution of each path; Step 1.6 In the TPA analysis module, calculate the excitation force and analyze the contribution of each path. Based on the analysis results, you can know the most important contribution path and the excitation force at the starting point of each path; Step 1.7 Use spectrum analysis to determine the frequency that causes noise or vibration peaks, and analyze the sound field characteristics of the vehicle interior noise target under typical working conditions; Step 1.8 extracts and analyzes the sound field transfer function characteristics of the main noise contribution transfer path in the vehicle, and obtains the sound pressure level at the peak frequency of the main noise contribution transfer path based on the frequency characteristics of the sound field in the vehicle.

3. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle according to claim 1, characterized in that: In the step 1, the sound pressure level at the peak frequency of the transfer path of the main noise contribution in the vehicle is obtained by using the Transfer Path Analysis method in the test software LMS during the noise transfer path contribution test, and a sample database is established for multiple groups of data.

4. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle as claimed in claim 1, characterized in that: The step 2 specifically includes the following steps: Step 2.1, perform finite element modeling on the resonance cavity structure formed by enclosing the body panels inside the vehicle; The modal analysis of the resonant cavity is: The final resonant frequency is Where Z is the acoustic impedance, which is a complex number, and the imaginary part is the acoustic impedance, real part is the acoustic impedance; ω is the resonant frequency, ρ is the air density, L is the length of the capillary, S is the cross-sectional area of ​​the capillary, V is the cavity volume, c is the speed of sound, and f n is the modal resonance frequency, j is the sign of the imaginary part; Step 2.2, using the finite element method to analyze the influence of different resonance cavity geometric parameters on the resonance cavity modal frequency. The acoustic mode of the resonance cavity is determined by its geometric parameters, including the resonance cavity appearance parameters and internal structure parameters. The influence of these two parts of parameters on the acoustic cavity modal frequency is analyzed separately, so as to obtain the change relationship of the resonance cavity modal frequency caused by each variable acting alone, determine the key geometric parameter variables that affect the resonance cavity modal frequency, and establish a sample database for the key geometric parameters; Step 2.3: Based on the original resonance cavity model, individually change the geometric parameters of the acoustic cavity shape and the seat position parameters, analyze the changes in the sound pressure level response in the vehicle, select the acoustic cavity models under several specific geometric parameter states, and obtain the sound pressure level response curve of the measuring point position in the vehicle, so as to obtain the influence of the geometric parameters of the resonance cavity on the sound field distribution, determine the key geometric parameters of the key resonance cavity positions that affect the sound field distribution in the vehicle, and establish a sample database for the key geometric parameters.

5. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle according to claim 1, characterized in that: In step 3, the generalized regression neural network algorithm is used to self-learn the data in the existing samples through the neural network algorithm to identify the functional correspondence between the input and output data, and a nonlinear functional correspondence is established between the sound pressure level at the peak frequency of the transmission path of the main noise contribution in the vehicle and the geometric parameters of the resonance cavity, where the input quantity is X=(x1, x2,…, x n ) T , the corresponding output is Y=(y1,y2,…,y n ) T , the number of neurons in the input layer is equal to the dimension of the input vector in the learning sample, and each neuron is a simple distribution unit that directly passes the input variable to the pattern layer; The number of neurons in the pattern layer is equal to the number of learning samples n, and its transfer function is: p i is the mode layer transfer function, X is the network input variable; X i is the learning sample corresponding to the i-th neuron, σ is the smoothing factor; The output values ​​of all nodes in a type of summation layer are summed, and its transfer function is: S m Transfer function for the summation layer; Another type of output layer for weighted summation is expressed as: S Nk is the transfer function of the weighted summation, y ik is the weight output between the i-th neuron of the pattern layer and the k-th molecule of the summation layer; The number of neurons in the output layer is equal to the two-digit number k in the output direction of the learning sample. Each neuron divides the output of the summation layer. The output of neuron k corresponds to the estimated result The kth element of , that is: The regression value of y corresponding to x: x is a random variable, the observed value is x0, f(x, y) is the joint probability density function, x and y are two random variables; Using Parzen nonparametric estimation, the unknown density function is: Where, X i , Y i are the sample observations of x and y, p is the dimension of x, n is the sample size, and σ is the smoothing factor; because The above formula can be rewritten as: is the total sample observation value Y i The weighted sum of 6. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle as claimed in claim 1, characterized in that: The specific steps for optimizing the geometric parameters of the resonance cavity in step 3 are as follows: the mathematical model of "geometric parameters of the resonance cavity - sound pressure level at the peak frequency of the transmission path of the main noise contribution in the vehicle" is regarded as a multi-input multi-output system, and the function corresponding relationship is Y = F (X) (12), where Y = (y1, y2, ..., y n ) T is the sound pressure level at the peak frequency of the main contribution to the vehicle interior noise transmission path, X = (x1, x2, ..., x n ) T It is the geometric parameter variable of the resonant cavity. A generalized regression neural network controller is established using MATLAB to train the input and output sample data and establish a function correspondence.

7. The method for optimizing sound field characteristics and controlling noise based on the arrangement of resonance cavities in a vehicle as claimed in claim 6, characterized in that: The function relationship of the optimization problem of the in-vehicle acoustic environment is Y=F(X). The optimization process is to solve the geometric parameters of the resonant cavity under certain external excitation conditions. The input vector X=(x1,x2,…,x n ) T , so that the sound pressure level at the peak frequency, the output vector Y = (y1, y2, ..., y n ) T To reach the minimum or extreme value, a genetic algorithm is used to solve the problem of optimizing the acoustic environment of the car's interior sound field.

8. The method for optimizing sound field characteristics and controlling noise based on the arrangement of a resonance cavity in a vehicle as claimed in claim 7, characterized in that: A multi-objective optimization problem is constructed for optimizing the sound pressure level at the peak frequency of the main noise contribution transfer path in the vehicle, with F(X) as the objective function vector. The mathematical model is described as: Where minF(X) is the objective function for achieving the best acoustic environment, that is, the sound pressure level at the peak frequency of the main noise contribution transmission path in the car reaches the minimum value, x in and x im are the upper and lower limits of the i-th design variable of the acoustic cavity geometry parameter, respectively. The Gamultiobj function based on the GA algorithm in the Matlab optimization toolbox is used to optimize the multi-objective problem and obtain multiple groups of non-inferior solutions. Finally, each group of non-inferior solutions is optimized and screened. The optimization objective function is: f(x)=Var(min F(X))+τ mn E(min F(X)) (14) In the formula: Var—represents the calculated variance, τ mn — represents the weight coefficient, E— represents the average value, and the acoustic cavity geometric parameters corresponding to the minimum sound pressure level at the peak frequency of the transmission path of the main noise contribution in the car after optimization are used as the resonance cavity design indicators to achieve the purpose of noise reduction in the car sound field.