Noise, vibration and harshness (NVH) simulation-based front wall welding assembly sound insulation performance optimization design method and device

Abstract

The invention discloses an NVH simulation-based sound insulation performance optimization design method and device for a front wall welding assembly, and the method comprises the steps: S1, selecting the thicknesses of a plurality of plates of the front wall welding assembly as design variables, and calculating the sound insulation of the front wall welding assembly corresponding to the design variables through a structure-sound coupling method; s2, generating thickness sample points of a plurality of plates according to the design variables, and establishing a Kriging approximation model of the thickness sample points and the sound reduction index; and S3, optimizing the Kriging approximation model by adopting a continuous quadratic programming optimization algorithm to obtain a thickness optimal solution of the plurality of plates. According to the optimization design method, for the problems existing in sound insulation performance optimization of the current automobile front wall welding assembly, efficient and reliable sound insulation performance optimization design of the front wall welding assembly is achieved by combining multidisciplinary simulation, a Kriging approximation model and a continuous quadratic programming optimization algorithm.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of automobile performance optimization, and in particular to a front-end welding assembly sound insulation performance optimization design method and device based on NVH simulation. BACKGROUND

[0002] With the development of the automobile industry, consumers' requirements for vehicle comfort are continuously increasing, and NVH (Noise, Vibration, and Harshness) performance has become one of the important indicators for measuring automobile quality. Interior noise directly affects the driving experience, and sound insulation performance is a key factor in controlling noise transmission.

[0003] During driving, noises such as engine, tire, and wind noise are transmitted to the interior of the vehicle through the vehicle body structure (such as the front-end panel, floor, and doors). Among them, the front-end welding assembly, as the main barrier to isolate engine compartment noise, has a significant impact on the interior noise level. If the sound insulation effect is poor, engine roar or road excitation noise will be transmitted into the cabin, reducing the comfort of the ride.

[0004] Currently, the optimization design of automobile sound insulation performance mainly relies on the following methods:

[0005] 1. Experimental test method: measure the sound insulation of different front-end welding assemblies through real vehicle or bench tests, but this method is low in efficiency, long in cycle, and difficult to quickly iterate and optimize.

[0006] 2. Empirical design method: adjust the parameters of the front-end welding assembly based on historical data or engineering experience, but it is difficult to realize the influence of each parameter on all performances and the ranking of each parameter, which is low in efficiency.

[0007] 3. Finite element simulation method: can be used to calculate the vibration characteristics of the structure, but the finite element simulation usually only considers the structural modal and does not fully consider the structural-acoustic coupling effect, so the simulation optimization precision is insufficient.

[0008] In recent years, the structural-acoustic coupling simulation optimization method (i.e. finite element-boundary element coupling) has been used to predict sound insulation performance. However, traditional optimization design methods (such as design of experiments + response surface method) are difficult to converge efficiently when there are many variables, which can easily lead to low optimization efficiency. In addition, the conventional approximation model (such as polynomial regression, radial basis function) has limited fitting ability for nonlinear sound insulation characteristics, and the reliability of the optimization result is low.

[0009] Therefore, in view of the deficiencies of the existing sound insulation performance optimization methods, it is an urgent technical problem to develop an efficient and reliable front-end welding assembly sound insulation performance optimization design method. SUMMARY

[0010] The application aims to provide a front-end welding assembly sound insulation performance optimization design method and device based on NVH simulation, so as to solve the problem of low efficiency and low reliability of the existing sound insulation performance optimization method.

[0011] To solve the above technical problems, in a first aspect, the application provides a front-end welding assembly sound insulation performance optimization design method based on NVH simulation, comprising the steps of:

[0012] S1: selecting the thicknesses of a plurality of plate parts of the front-end welding assembly as design variables, and calculating the sound insulation of the front-end welding assembly corresponding to the design variables by using a structure-sound coupling method;

[0013] S2: generating thickness sample points of the plurality of plate parts according to the design variables, and establishing a Kriging approximation model of the thickness sample points and the sound insulation;

[0014] S3: optimizing the Kriging approximation model by using a continuous quadratic programming optimization algorithm to obtain the optimal solution of the thicknesses of the plurality of plate parts.

[0015] Further, the calculation method of the sound insulation comprises:

[0016] S11: constructing a finite element model of the front-end welding assembly according to three-dimensional data of the front-end welding assembly;

[0017] S12: constructing an acoustic field boundary element model according to the actual use conditions of the front-end welding assembly;

[0018] S13: coupling the finite element model and the acoustic field boundary element model to calculate the sound insulation of the front-end welding assembly.

[0019] Further, the Kriging approximation model is:

[0020]

[0021] cov[z(X i ),z(X j )]=σ 2 R(X i ,X j )

[0022]

[0023] Wherein, y(X) is the sound insulation, f i (X) is the i-th known regression function; β i is the weight coefficient of the regression model; z(X) is the deviation generated by the approximation model, which is a random process subject to (0, σ 2 ) standard normal distribution; cov[z(X i ),z(X j) is the covariance of z(X); X i and X j are any two thickness sample points; X i = (x i,1 , x i,2 ,..., x i,m ), x i,m is the thickness of the m-th plate in the i-th thickness sample point; X j = (x j,1 , x j,2 ,..., x j,m ), x j,m is the thickness of the m-th plate in the i-th thickness sample point; R(X i , X j ) is the correlation function; i = 1, 2,..., n; j = 1, 2,..., n; n is the number of thickness sample points; α1 is the weight coefficient of the first kernel function ; θ is the long-scale parameter of the first kernel function; α2 is the weight coefficient of the second kernel function , Γ(v) is the gamma function, l is the long-scale parameter of the second kernel function, v is the smoothness control parameter, K v is the modified Bessel function of the second kernel function; α3 is the weight coefficient of the third kernel function .

[0024] Further, step S3 specifically comprises:

[0025] The negative log-likelihood function is used as the objective function of the Kriging approximation model, the weight coefficients α1, α2, α3 are adjusted to minimize the fitting error of the Kriging model on the thickness sample points, and the optimal solution of the thickness of the plurality of plates is obtained.

[0026] Further, the objective function of the Kriging approximation model is:

[0027]

[0028] Wherein, X is the input thickness sample point of the Kriging approximation model, y is the predicted value of the sound insulation amount output by the Kriging approximation model, and K(X, X, α1, α2, α3) is the covariance matrix dependent on the weight coefficients α1, α2, α3.

[0029] Further, minimizing the objective function specifically comprises:

[0030] Let the current iteration point be Then the objective function is twice Taylor expanded to obtain the quadratic approximation function of the objective function:

[0031]

[0032] wherein, is the gradient of the objective function, i.e. the partial derivative with respect to α1, α2, α3; H k is the second order derivative of the objective function with respect to the parameters, used to describe the curvature of the objective function;

[0033] minimizing the quadratic approximation function:

[0034]

[0035] iteratively solving to obtain a new iteration point and making the new iteration point P k+1 the previous iteration point P k = P k+1 , until the convergence condition is met, to obtain the optimal solution of the weight coefficients α1, α2, α3.

[0036] Further, the convergence condition is:

[0037]

[0038] or

[0039] ‖P k+1 -P k ‖<σ

[0040] wherein, ε is a pre-established error tolerance, and σ is a convergence tolerance.

[0041] Further, in the iterative solving process of minimizing the quadratic approximation function, an adaptive learning method is used to adaptively update the iteration point P k+1 according to the gradient change of each iteration step. The update method of the iteration point P k+1 is:

[0042]

[0043] G k = G k-1 + g k ⊙ g k

[0044]

[0045] wherein, g k is the current gradient; G k is the sum of the squares of the historical gradients; A is a non-zero constant; and η is a basic learning rate.

[0046] Further, the plurality of plate members of the front wall welding assembly include a front wall plate, a front windshield gutter rear connecting plate, a vacuum booster pump installation reinforcing plate, a front wall reinforcing member, a front wall beam, and a plate type radiator installation support.

[0047] In a second aspect, the application provides a device for optimizing the sound insulation performance of a front-end welding assembly based on NVH simulation, which comprises a processor and a memory; wherein the memory is used to store computer execution instructions, and when the device is running, the processor executes the computer execution instructions stored in the memory to enable the device to execute the optimization design method provided in the first aspect.

[0048] The application has the following beneficial effects:

[0049] 1. By using the structural-acoustic coupling method to calculate the sound insulation of the front-end welding assembly, the influence of different plate thicknesses on the sound insulation performance of the front windshield of the automobile can be accurately evaluated; the Kriging approximation model and the sequential quadratic programming optimization algorithm are used to solve the sound insulation optimization problem, the Kriging approximation model has high flexibility and can adapt to different spatial structure models, i.e., it can be applied to different vehicle models and different projects in the sound insulation performance optimization of the front-end welding assembly; the sequential quadratic programming optimization algorithm can improve the accuracy and generalization ability of the Kriging approximation model; this method can filter out the optimal solution or suboptimal solution from a large number of possible design schemes in a short time;

[0050] 2. By using the composite kernel function composed of different types of kernel functions, the complex relationship in the data can be effectively captured, and the accuracy and generalization ability of the Kriging approximation model can be improved; through the sequential quadratic programming optimization algorithm, the weight coefficients of the composite kernel function can be effectively optimized, thereby optimizing the Kriging approximation model and improving the performance of the Kriging approximation model;

[0051] 3. By using the adaptive learning method to dynamically adjust the step size according to the gradient change of each iteration step, the problems caused by excessively large or small step sizes can be avoided, making the sequential quadratic programming optimization algorithm more stable during the optimization process, especially when the objective function has a complex nonlinear relationship, the weight coefficient update can be more stable and the convergence speed can be accelerated. BRIEF DESCRIPTION OF DRAWINGS

[0052] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. In these drawings, the same reference numerals are used to represent the same or similar parts. The schematic embodiments of the present application and their descriptions are used to explain the present application, and do not constitute an improper limitation on the present application. In the drawings:

[0053] Figure 1 The method flowchart of an embodiment of the present application;

[0054] Figure 2 The coupling schematic diagram of the finite element model and the sound field boundary element model of an embodiment of the present application;

[0055] Figure 3 Error scatter plot for an embodiment of the application. DETAILED DESCRIPTION

[0056] A front-end welding assembly sound insulation performance optimization design method based on NVH simulation as shown in the figure, comprising the steps of: Figure 1

[0057] S1: selecting the thickness of a plurality of plate parts of the front-end welding assembly as design variables, and calculating the sound insulation of the front-end welding assembly corresponding to the design variables by using a structure-sound coupling method;

[0058] S2: generating thickness sample points of the plurality of plate parts according to the design variables, and establishing a Kriging approximation model of the thickness sample points and the sound insulation;

[0059] S3: using a continuous quadratic programming optimization algorithm to optimize the Kriging approximation model to obtain the optimal solution of the thickness of the plurality of plate parts.

[0060] The present application can accurately evaluate the influence of different plate thicknesses on the sound insulation performance of the front windshield of the automobile by using the structure-sound coupling method to calculate the sound insulation of the front-end welding assembly. The Kriging approximation model and the continuous quadratic programming optimization algorithm are used to solve the sound insulation optimization problem. The Kriging approximation model has high flexibility and can adapt to different spatial structure models, i.e. it can be applied to different models and different projects in the sound insulation performance optimization of the front-end welding assembly. The continuous quadratic programming optimization algorithm can improve the accuracy and generalization ability of the Kriging approximation model, and can filter out the optimal solution or suboptimal solution from a large number of possible design schemes in a short time.

[0061] In the above step S2, the thickness of a plurality of main plate parts constituting the front-end welding assembly can be selected as design variables by using Latin square experimental design. The optimal Latin square experimental design can control multiple thickness variables at the same time and reduce experimental errors.

[0062] According to an embodiment of the present application, the sound insulation calculation method comprises:

[0063] S11: constructing a finite element model of the front-end welding assembly according to three-dimensional data of the front-end welding assembly; the finite element model construction method specifically comprises: importing the three-dimensional model into a finite element analysis software (such as ANSYS), and performing meshing to obtain a structure mesh of the front-end welding assembly;

[0064] ​S12: constructing a sound field boundary element model according to the actual use condition of the front wall welding assembly; the sound field boundary element model construction method specifically comprises: meshing the boundary element of the sound wave incident surface of the front wall welding assembly into a hemisphere to form a hemispherical surface grid; then, applying the actually measured noise signal in the engine compartment as an acoustic excitation to the hemispherical surface grid, and applying a constraint at the boundary of the instrument panel and the position of the bolt hole according to the actual installation condition of the front wall welding assembly;

[0065] S13: coupling the finite element model and the sound field boundary element model, as shown in Figure 2 , calculating the sound insulation quantity of the front wall welding assembly. The contact surface of the structure grid and the sound field acoustic grid is defined as a coupling surface, the structure-sound coupling surface of the front wall welding assembly is obtained, and the sound insulation quantity is defined as the decibel number related to the incident sound energy on one side of the structure-sound coupling surface and the transmitted sound energy on the other side:

[0066]

[0067] In the formula: R represents the sound insulation quantity, E i and W i represent the incident sound energy and the incident sound power, E t and W t represent the transmitted sound energy and the transmitted sound power.

[0068] According to one embodiment of the present application, the Kriging approximation model is:

[0069]

[0070] cov[z(X i ),z(X j )]=σ 2 R(X i ,X j )

[0071]

[0072] Wherein: y(X) is the sound insulation quantity, f i (X) is the i-th known regression function; β i is the weight coefficient of the regression model; z(X) is the deviation generated by the approximation model, which is a random process subject to (0, σ 2 ) standard normal distribution; cov[z(X i ),z(X j )] is the covariance of z(X); X i and X j are any two thickness sample points; X i =(x i,1 ,x i,2 ,...,x i,m ), xi,m is the thickness of the m-th plate in the i-th thickness sample point; X j = (x j,1 ,x j,2 ,...,x j,m ), x j,m is the thickness of the m-th plate in the i-th thickness sample point; R(X i ,X j ) is a correlation function; i = 1, 2,..., n; j = 1, 2,..., n; n is the number of thickness sample points; a1 is a weight coefficient of a first kernel function ; θ is a long-scale parameter of the first kernel function; a2 is a weight coefficient of a second kernel function ; Γ(v) is a gamma function, l is a long-scale parameter of the second kernel function, and v is a smoothness control parameter; K v is a modified Bessel function of the second kernel function; a3 is a weight coefficient of a third kernel function .

[0073] The present application adopts a composite kernel function, which can capture a smooth relationship through the first kernel function, capture a more complex nonlinear relationship through the second kernel function, and especially cope with a mutation; the third kernel function captures a linear relationship, and through combination of different types of kernel functions, a complex relationship in the data can be captured, so as to achieve the purpose of improving the precision and generalization ability of the Kriging approximation model.

[0074] According to an embodiment of the present application, the step S3 specifically comprises:

[0075] The negative log-likelihood function is adopted as the objective function of the Kriging approximation model, the weight coefficients a1, a2, and a3 are adjusted to minimize the fitting error of the Kriging model on the thickness sample points, and the optimal solution of the thickness of the plurality of plates is obtained. By minimizing the negative log-likelihood function, the weight coefficients a1, a2, and a3 are adjusted to minimize the fitting error of the model on the training data, so as to ensure that the kernel function similarity structure is reasonable, thereby ensuring the precision and generalization ability of the Kriging approximation model.

[0076] In the optimization process, the weight coefficients a1, a2, and a3 can be set with constraint conditions: 1, the weight coefficients a1, a2, and a3 are non-negative, so as to avoid negative contribution; 2, the sum of the contributions of all kernel functions is 1, that is, a1+a2+a3=1, so as to ensure the relative balance of the weight coefficients.

[0077] According to an embodiment of the present application, the objective function of the Kriging approximation model is:

[0078]

[0079] wherein, X is an input thickness sample point of the Kriging approximation model, y is a predicted value of the sound insulation amount output by the Kriging approximation model, and K(X, X, a1, a2, a3) is a covariance matrix dependent on the weight coefficients a1, a2, and a3.

[0080] According to one embodiment of the present application, the minimization of the objective function specifically includes:

[0081] Let the current iteration point be P Then, the objective function is twice Taylor expanded to obtain a quadratic approximation function of the objective function:

[0082]

[0083] wherein, is the gradient of the objective function, i.e., the partial derivative of the objective function with respect to a1, a2, and a3; H k is the second derivative of the objective function with respect to the parameters, used to describe the curvature of the objective function;

[0084] The quadratic approximation function is minimized:

[0085]

[0086] The new iteration point P is obtained by iterative solving, and the new iteration point P k+1 is made to satisfy the convergence condition. k The previous iteration point P k+1 is updated, until the convergence condition is satisfied, and the optimal solution of the weight coefficients a1, a2, and a3 is obtained.

[0087] According to one embodiment of the present application, the convergence condition is:

[0088]

[0089] or

[0090] ‖P k+1 -P k ‖<σ

[0091] wherein, ε is a pre-established error tolerance, and σ is a convergence tolerance.

[0092] When the change of the objective function is small enough, or the change of the weight coefficients is small enough, the convergence is ended, at which time a1, a2, and a3 minimize the fitting error of the Kriging approximation model on the training data.

[0093] According to one embodiment of the present application, in the iterative solving process of minimizing the quadratic approximation function, an adaptive learning method is used to adaptively update the iteration point P k+1 according to the gradient change of each iteration step, and the iteration point Pk+1 The update method is as follows:

[0094]

[0095] G k =G k-1 +g k ⊙g k

[0096]

[0097] Among them, g k G represents the current gradient; k Let be the sum of squares of the historical gradients; A is a non-zero constant; η is the basic learning rate.

[0098] This embodiment employs an adaptive learning method to dynamically adjust the step size based on the gradient changes in each iteration step. This avoids problems caused by excessively large or small step sizes, making the continuous quadratic programming optimization algorithm more stable during the optimization process. In particular, when the objective function has complex nonlinear relationships, it can make the weight coefficient updates smoother and accelerate the convergence speed.

[0099] According to one embodiment of this application, the front bulkhead welded assembly includes multiple plates such as a front bulkhead plate, a front windshield drain channel rear connecting plate, a vacuum booster pump mounting reinforcement plate, a front bulkhead reinforcement, a front bulkhead crossbeam, and a plate radiator mounting bracket. When selecting the thickness of the main plates of the front bulkhead welded assembly as a design variable, upper and lower limits are set based on the selected initial thickness, as shown in Table 1.

[0100] Table 1 Range of Design Variables

[0101]

[0102] The optimal Latin square experimental design was adopted, and the design variables were divided into 51 layers within the upper and lower limits of each design variable, resulting in 51 sets of optimal Latin square experimental designs. Each set included 5 sample points with different thicknesses, as shown in Table 2.

[0103] Table 2 Experimental Design Scheme

[0104]

[0105] Based on the selected thickness sample points, the sound insulation of the front welded assembly corresponding to the design variables can be calculated using the structure-acoustic coupling method.

[0106] Through simulation experiments, the optimal solution and predicted values ​​of each response of the sound insulation optimization problem are obtained using the optimization algorithm disclosed in this invention. The predicted values ​​are then compared with the actual values, and the errors between the predicted and actual values ​​are all less than 5%. Figure 3As shown, the engineering requirements are met.

[0107] In a second aspect, the application provides a device for optimizing the sound insulation performance of a front-end welded assembly based on NVH simulation, comprising a processor and a memory; wherein the memory is used to store computer execution instructions; when the device is running, the processor executes the computer execution instructions stored in the memory, so that the device executes the optimization design method disclosed in the first aspect.

[0108] The method is not only suitable for optimizing the sound insulation performance of the front-end welded assembly, but also suitable for optimizing the weight, stiffness, modal and the like of the front-end welded assembly. By simultaneously optimizing multiple targets, it can be ensured that the final design scheme not only meets the requirements in sound insulation performance, but also performs well in other aspects, thereby achieving the best balance of overall performance.

[0109] Finally, it should be pointed out that the above embodiments are only used to illustrate the technical solutions of the application and are not limiting. Although the application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the application can be modified or replaced equivalently without departing from the purpose and scope of the application, and they should be covered in the scope of the claims of the application.

Claims

1. A method for optimizing the sound insulation performance of a front welded assembly based on NVH simulation, characterized in that, Including the following steps: S1: Select the thickness of multiple plates of the front welding assembly as design variables, and use the structure-sound coupling method to calculate the sound insulation of the front welding assembly corresponding to the design variables; S2: Generate multiple thickness sample points of the plates based on the design variables, and establish a Kriging approximate model of the thickness sample points and the sound insulation. S3: The Kriging approximation model is optimized using a continuous quadratic programming optimization algorithm to obtain the optimal solution for the thickness of multiple plates.

2. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 1, characterized in that, The method for calculating the sound insulation includes: S11: Construct a finite element model of the front welding assembly based on the three-dimensional data of the front welding assembly; S12: Construct a sound field boundary element model based on the actual usage conditions of the front welding assembly; S13: Couple the finite element model and the sound field boundary element model to calculate the sound insulation of the front welding assembly.

3. The sound insulation performance optimization design method for the front welded assembly based on NVH simulation according to claim 1 or 2, characterized in that, The correlation function used in the Kriging approximation model is: Where: X i and X j Let X be any two thickness sample points; i =(x i,1 ,x i,2 ,...,x i,m ), x i,m X represents the thickness of the m-th plate component in the i-th thickness sample point; j =(x j,1 ,x j,2 ,...,x j,m ), x j,m R(X) represents the thickness of the m-th plate component in the i-th thickness sample point; i ,X j ) represents the correlation function; i = 1, 2, ..., n; j = 1, 2, ..., n; n is the number of thickness sample points; α1 is the first kernel function. The weighting coefficients; θ is the long-scale parameter of the first kernel function; α2 is the weighting coefficient of the second kernel function. The weighting coefficients, Γ(v) are the gamma function. K is the long-scale parameter of the second kernel function, v is the smoothness control parameter, and K is the long-scale parameter of the second kernel function. v α3 is the modified Bessel function of the second kernel function; α3 is the third kernel function. The weighting coefficients.

4. The sound insulation performance optimization design method for the front welded assembly based on NVH simulation according to claim 3, characterized in that, Step S3 specifically includes: The negative log-likelihood function is used as the objective function of the Kriging approximation model. The optimization objective is to minimize the objective function. The weight coefficients α1, α2, and α3 are adjusted to minimize the fitting error of the Kriging model at the thickness sample points, thereby obtaining the optimal solution for the thickness of multiple plates.

5. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 4, characterized in that, The objective function of the Kriging approximation model is: Where X is the input thickness sample point of the Kriging approximation model, y is the predicted value of the sound insulation output by the Kriging approximation model, and K(X,X,α1,α2,α3) is the covariance matrix dependent on the weight coefficients α1,α2,α3.

6. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 5, characterized in that, Minimizing the objective function specifically includes: Let the current iteration point be Then, the objective function is expanded using a second Taylor series to obtain a second-order approximation function of the objective function: in, H is the gradient of the objective function, i.e., the partial derivatives with respect to α1, α2, α3; k The second derivative of the objective function with respect to the parameters is used to describe the curvature of the objective function; Minimize the quadratic approximation function: Iterative solution yields new iteration points And make the new iteration point P k+1 Update the previous iteration point P k =P k+1 This continues until the convergence condition is met, yielding the optimal solution for the weight coefficients α1, α2, and α3.

7. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 6, characterized in that, The convergence condition is: or ‖P k+1 -P k ‖<σ Where ε is the pre-set error tolerance and σ is the convergence tolerance.

8. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 7, characterized in that, In the iterative solution process of minimizing the quadratic approximation function, an adaptive learning method is used to adaptively update the iteration point P according to the gradient change of each iteration step. k+1 The iteration point P k+1 The update method is as follows: G k =G k-1 +g k ⊙g k Among them, g k G represents the current gradient; k Let be the sum of squares of the historical gradients; A is a non-zero constant; η is the basic learning rate.

9. The method for optimizing the sound insulation performance of the front welded assembly based on NVH simulation according to claim 1, characterized in that, The front bulkhead welded assembly includes multiple plates such as a front bulkhead plate, a front windshield drain channel rear connecting plate, a vacuum booster pump mounting reinforcement plate, a front bulkhead reinforcement, a front bulkhead crossbeam, and a plate radiator mounting bracket.

10. A device for optimizing the sound insulation performance of a front welded assembly based on NVH simulation, characterized in that, include: A processor and a memory; wherein the memory is used to store computer execution instructions, and when the device is running, the processor executes the computer execution instructions stored in the memory to cause the device to perform the optimization design method according to any one of claims 1-9.

Images