Method for optimizing dynamic performance of two-stage vibration isolation rubber mount

Abstract

The invention discloses a method for optimizing the dynamic performance of a two-stage vibration isolation rubber mount, which comprises the step of optimizing the dynamic performance of a vibration reduction rubber mount on an automobile power assembly, and is realized by the following steps of: establishing a finite element analysis model of the two-stage vibration isolation rubber mount; constructing an orthogonal test scheme comprising a plurality of parameters and a plurality of levels; calculating static stiffness and dynamic performance of the test scheme model by utilizing finite element software; predicting static stiffness and dynamic performance of the test scheme model by using a neural network model optimized by a genetic algorithm; the peak frequency of the static stiffness and the peak frequency of the dynamic stiffness serve as optimization targets, and structural parameters of the rubber mount are optimized through a genetic algorithm; and the optimization work of the secondary vibration isolation rubber suspension is completed. The optimized rubber mount can better meet the vibration isolation requirement in practical application, has the advantages of being high in calculation efficiency, high in adaptability and accurate in result, and provides a new technical approach for design and optimization of the rubber mount.

Description

Technical Field

[0001] The present invention relates to the performance optimization of vibration damping rubber mounts on automotive powertrains, and particularly to the dynamic characteristic optimization of a two-stage vibration isolation rubber mount. Background Art

[0002] As an important vibration isolation component, rubber mounts are widely used in the connection between the powertrain and the chassis, and their performance is directly related to the NVH (noise, vibration, and harshness) characteristics of the vehicle. Therefore, the dynamic performance optimization of rubber mounts is of great significance for improving ride comfort and reducing noise pollution. Traditional rubber mount designs mainly rely on experience and trial-and-error methods, lacking a systematic optimization strategy. With the development of finite element analysis technology, the static performance calculation of rubber mounts has been relatively mature. However, for the dynamic performance of rubber mounts, especially the high-frequency dynamic stiffness performance, traditional design methods often fail to meet the strict requirements of modern electric vehicles for NVH performance.

[0003] In existing research, the dynamic stiffness calculation of rubber mounts usually adopts constitutive models combined with finite element analysis or dynamic models. Although these methods can provide certain calculation results, they have problems such as low calculation efficiency and difficulty in dealing with nonlinear and complex working conditions. Existing research has shown that the dynamic stiffness of two-stage mounts is related to many factors (Research on High-Frequency Dynamic Models of Rubber Mounts with Second-Stage Isolation, Jiagi Zhang and Xiao-ang Liu Hebei University of Technology, Citation: Zhang, J. and Liu, Research on High-frequency Dynamic Models of Rubber Mounts with Second-stage Isolation, SAE Technical Paper 2022-01-0617, 2022, doi:10.4271 / 2022-01-0617.), and it is difficult to give an accurate calculation formula. In addition, existing rubber mount designs often do not comprehensively consider the mutual influence of static stiffness and high-frequency dynamic characteristics, resulting in the mounts designed being difficult to achieve the ideal vibration isolation effect in actual applications. Summary of the Invention

[0004] To solve the above problems, the present invention proposes a method for optimizing the dynamic performance of a two-stage vibration isolation rubber mount based on a neural network cooperative genetic algorithm. This method optimizes the structural parameters of the rubber mount by establishing a finite element analysis model of the rubber mount, combining orthogonal experimental design, and using a neural network cooperative genetic algorithm to achieve the dual optimization goals of static stiffness and peak frequency of dynamic stiffness. This method not only improves the vibration isolation performance of the rubber mount, but also has the advantages of high computational efficiency, strong adaptability, and accurate results, providing a new technical approach for the design and optimization of rubber mounts. This method has higher computational efficiency than traditional methods and comprehensively considers static stiffness and high-frequency dynamic characteristics, thus meeting the higher vibration isolation requirements in practical applications of the mount.

[0005] The present invention is achieved at least by one of the following technical solutions.

[0006] A method for optimizing the dynamic performance of a two-stage vibration isolation rubber mount, comprising the following steps:

[0007] (1) Establish a finite element analysis model of the two-stage vibration isolation rubber mount;

[0008] (2) According to the parameters of the two-stage vibration isolation rubber mount, set an orthogonal experimental scheme to evaluate the influence of each parameter on the performance;

[0009] (3) Use finite element software to calculate the static stiffness and dynamic performance of the orthogonal experimental scheme;

[0010] (4) Use a neural network model optimized by a genetic algorithm to predict the static stiffness and dynamic performance of the orthogonal experimental scheme;

[0011] (5) Use a genetic algorithm to optimize the structural parameters of the rubber mount with the static stiffness and peak frequency of dynamic stiffness as the optimization goals.

[0012] Furthermore, the orthogonal experimental scheme includes the following steps:

[0013] (1) Determine the key parameters affecting the performance of the rubber mount;

[0014] (2) Set different levels for each parameter;

[0015] (3) Construct an orthogonal table, which includes the combinations of the key parameters and the levels of the key parameters;

[0016] (4) Conduct finite element calculations according to the orthogonal table to obtain the static stiffness and peak frequency data of dynamic stiffness under each combination; In the optimization design process, the structural parameters of the rubber mount are used as design variables, and the static stiffness and peak frequency of dynamic stiffness are used as optimization goals.

[0017] Furthermore, the construction of the neural network model includes the following steps:

[0018] (S1) Set up a multi-layer neural network, which includes an input layer, a hidden layer, and an output layer. The number of nodes in the input layer is determined according to the number of parameters. The number of nodes in the hidden layer is determined by comparing the training effects according to an empirical formula in sequence. The number of nodes in the output layer is determined according to the number of calculation targets;

[0019] (S2) Select an activation function and a loss function;

[0020] (S3) Use the stiffness information of the secondary vibration isolation rubber mount calculated by the finite element software as a data set, divide the data set into a training set and a test set, and perform normalization processing on the training set and the test set to improve the generalization ability of the model;

[0021] (S4) Use a genetic algorithm to optimize the initial parameters of the neural network, and the initial parameters include weights and thresholds;

[0022] (S5) Through the selection, crossover, and mutation operations of the genetic algorithm, iteratively generate a new population until the stop condition is met or the number of iterations is reached;

[0023] (S6) According to the fitness function, that is, the error between the predicted value and the experimental value of the neural network, select the optimal neural network parameters.

[0024] Further, the parameter settings of the genetic algorithm include: a. The population size, the number of iterations, the crossover probability, and the mutation probability are all determined in advance according to the experimental design; b. The fitness function is the sum of the absolute values of the errors between the predicted value of the neural network and the calculated value in the finite element.

[0025] Further, in step (5), the function for optimizing the structural parameters of the rubber mount uses the difference from the target value.

[0026] Further, the preprocessing of the data set of the neural network model includes the following steps: a. Check the correctness of the finite element model, that is, use the results calculated by the finite element as the data set of the neural network, where the results calculated by the finite element are the results with an error of no more than 10% from the suspension performance parameters in actual production; b. Shuffle the samples in the orthogonal experiment to improve the generalization ability of the model.

[0027] Furthermore, the secondary vibration isolation rubber mount includes a metal inner core, an outer tube, an intermediate frame, and a rubber bracket. The rubber bracket includes an inner rubber layer, an outer rubber layer, and multiple connecting arms that are cross-distributed along the circumferential direction of the inner rubber layer and the outer rubber layer. The inner ends of the connecting arms of the inner rubber layer are integrally connected to the inner rubber layer, the outer ends of the connecting arms of the inner rubber layer are integrally connected to the inner side of the intermediate frame, the inner ends of the connecting arms of the outer rubber layer are integrally connected to the outer side of the intermediate frame, the outer ends of the connecting arms of the outer rubber layer are integrally connected to the inner side of the outer tube. The metal inner core is fixedly connected to the output shaft of the drive motor, and the outer tube is fixed to the subframe.

[0028] Furthermore, the intermediate frame is an annular structure located inside the rubber bracket for connecting the connecting arms of the inner rubber layer and the outer rubber layer. The intermediate frame is made of metal or nylon. The number of connecting arms of the inner rubber layer and the outer rubber layer is four, which is an even number and is evenly distributed in a cross shape.

[0029] Furthermore, the connecting arms of the inner rubber layer are vulcanized and fixed to the metal inner core and the inner side of the intermediate frame. The connecting arms of the outer rubber layer are fixed to the outer side of the intermediate frame and vulcanized and fixed to the inner side of the outer tube. The outer tube is made of nylon. Taking the axial direction of the output shaft of the drive motor as the x-axis, the up and down directions in the radial direction as the z-axis, and the left and right directions as the y-axis, the rubber connecting arms are all arranged at a 45-degree angle with the x-direction.

[0030] A computer device of the present invention includes: a memory, a processor, and a computer program stored on the memory. When the computer program is executed on the processor, the dynamic performance optimization method of the secondary vibration isolation rubber mount is realized.

[0031] Compared with the existing technology, the beneficial effects of the present invention are as follows:

[0032] The present invention adopts a method based on the cooperation of a neural network and a genetic algorithm. Compared with the traditional design method that relies on experience and trial and error, the calculation efficiency is greatly improved, and the design cycle is reduced. Through the orthogonal test scheme, the present invention systematically evaluates the influence of multiple parameters at multiple levels on the performance of the rubber mount, improves the experimental efficiency, and reduces the number of experiments.

[0033] The present invention comprehensively considers the mutual influence of static stiffness and high-frequency dynamic characteristics. The optimized rubber mount can better meet the vibration isolation requirements in practical applications. The neural network model can adapt to the non-linearity and complex working conditions of the rubber mount, providing a new technical approach and being applicable to various different design and working condition requirements. Using the neural network model optimized by the genetic algorithm to predict static stiffness and dynamic performance improves the prediction accuracy and ensures the effectiveness of the optimization results. Description of the Drawings

[0034] Figure 1 is a flowchart of a method for optimizing the dynamic performance of a secondary vibration isolation rubber mount in Example 1;

[0035] Figure 2 is a model diagram of the secondary vibration isolation rubber mount in Example 2;

[0036] Figure 3 is a diagram of the static stiffness result based on neural network prediction in Example 3;

[0037] Figure 4 is a diagram of the peak frequency result of the dynamic stiffness based on neural network prediction in Example 4;

[0038] Figure 5 is a diagram of the static stiffness result of the neural network prediction optimized by the genetic algorithm in Example 5;

[0039] Figure 6 is a diagram of the peak frequency result of the dynamic stiffness of the neural network prediction optimized by the genetic algorithm in Example 6;

[0040] Figure 7 is a Pareto front diagram of multi-objective optimization in Example 7;

[0041] Figure 8 is a structural diagram of a partial cross-section of the secondary vibration isolation rubber mount in Example 8. Detailed Embodiment

[0042] To make the objectives, technical solutions and advantages of the present invention clearer and more definite, the following describes the present invention in further detail with reference to the accompanying drawings and by way of examples.

[0043] As Figure 2 shown, a method for optimizing the dynamic performance of a secondary vibration isolation rubber mount includes the following steps:

[0044] (1) Establish a finite element model of the secondary vibration isolation rubber mount in a finite element analysis software,

[0045] The secondary vibration isolation rubber mount in this embodiment includes a metal inner core, an outer tube, an intermediate skeleton and a rubber bracket. The rubber bracket includes an inner rubber layer, an outer rubber layer and a plurality of connecting arms distributed in a cross shape along the circumferential direction. The inner end of the inner rubber connecting arm is integrally connected to the inner rubber layer, the outer end of the inner rubber connecting arm is integrally connected to the inner side of the intermediate skeleton, the inner end of the outer rubber connecting arm is integrally connected to the outer side of the intermediate skeleton, the outer end of the outer rubber connecting arm is integrally connected to the inner side of the outer tube, the metal inner core is fixedly connected to the output shaft of the driving motor, and the outer tube is fixed to the subframe.

[0046] The middle skeleton is an annular structure located inside the rubber bracket, used to connect the inner and outer rubber connecting arms; the middle skeleton is made of metal or nylon; the number of the inner rubber and outer rubber connecting arms is four, which is an even number, and they are evenly distributed in a cross shape.

[0047] The inner rubber connecting arm is vulcanized and fixed to the metal inner core, and the inner rubber connecting arm is vulcanized and fixed to the inner side of the middle skeleton. The outer rubber connecting arm is fixed to the outer side of the middle skeleton, and the outer rubber connecting arm is vulcanized and fixed to the inner side of the outer tube. The material of the outer tube is nylon. Taking the axial direction of the output shaft of the driving motor as the x-axis, the vertical direction in the radial direction as the z-axis, and the left and right as the y-axis, all the rubber connecting arms are arranged at an angle of 45 degrees with the x-direction.

[0048] The established finite element model of the secondary vibration isolation rubber mount includes the inner rubber, the middle skeleton, the outer rubber, the outer tube, and the rigid surface, where the rigid surface is extracted from the outer surface of the outer tube. In the finite element software, rigid constraints need to be applied to the rigid surface, and an interference fit with the outer tube needs to be added. As an example, this example takes -0.75mm.

[0049] In the finite element analysis, since the stiffness of the metal inner core and the metal middle skeleton is much greater than that of the inner rubber, the outer rubber, and the nylon outer tube, the metal inner core is set as a rigid body. According to the cross-point dynamic stiffness reading method, the loading point of the metal inner core is set as the load and displacement input point, and the center point of the surface extracted from the outer wall of the outer tube is set as the output point. The rubber bracket mesh uses C3D8H hexahedral hybrid elements, with a mesh unit length of 1.5mm. The outer tube uses C3D8 hexahedral elements, with a mesh unit length of 3mm. The middle skeleton uses C3D8 hexahedral elements, with a mesh unit length of 3mm. Set the reference point of the metal inner core to be rigidly connected to the inner surface of the rubber (the part where the rubber and the metal inner core are vulcanized), and bind the unit surface of the inner wall of the outer tube and the outer surface of the rubber. Apply a preload of -1000N to the reference point of the metal inner core, and constrain the 6 degrees of freedom of the surface extracted from the outer wall of the outer tube. In the analysis step setting, select the steady-state dynamic analysis, set the excitation frequency range to 50 - 3000Hz, and the displacement amplitude to ±0.05mm;

[0050] (2) Select the parameters of the secondary vibration isolation rubber mount, set up an orthogonal test plan with four factors and five levels, and evaluate the influence of each parameter on the performance.

[0051] The orthogonal test plan is a systematic method used to evaluate the influence of multiple parameters at multiple levels on the performance of the rubber mount, and this plan can reduce the number of required experiments and improve the experimental efficiency. The orthogonal test plan with four factors and five levels includes the following steps:

[0052] 1) The four key parameters affecting the performance of the rubber mount are determined as R m , t m , d, ω r , where R m is the radius of the center plane of the middle skeleton (hereinafter referred to as the skeleton radius), t m is the radial thickness of the middle skeleton (hereinafter referred to as the skeleton thickness), ω r is the circumferential width of the rubber connecting arm (hereinafter referred to as the rubber width), and d is the axial thickness of the rubber connecting arm (hereinafter referred to as the rubber thickness). These four dimensional parameters can determine a unique mount structure.

[0053] Taking the U direction of the secondary vibration isolation rubber mount as the x-axis and the W direction as the y-axis, where U and W are the x-direction and y-direction of the vehicle coordinate system respectively, and taking the installation center point of the mount as the origin, a rectangular coordinate system as shown in the figure is established. The mount cross-section is symmetric about the origin, and a 1 / 4 structure is selected for parametric design. In order to more intuitively define the constraint conditions between the structure parameters, the following intermediate variables are established respectively, as Figure 8 shown: eight coordinate parameters A(x1, y1), B(x2, y2), C(x3, y3), D(x4, y4) and two auxiliary angle parameters α, β, R in is the radius of the center plane of the skeleton, and R1 is the radius of the inner-layer rubber of the skeleton. The following relational expressions can be obtained:

[0054]

[0055] 2) Four different levels are set for each parameter respectively; the levels of the skeleton radius are 25, 27.5, 30, 32.5, 35, the levels of the skeleton thickness are 1, 2, 3, 4, 5, the rubber width is 18, 21, 24, 27, 30, and the rubber thickness is 16, 17, 18, 19, 20.

[0056] 3) Construct an orthogonal table, which, as shown in Table 1, contains the combinations of the parameters and the levels of the parameters;

[0057] Table 1 List of levels of each factor

[0058]

[0059]

[0060] 4) Use the finite element method to calculate each sample in the orthogonal experiment to obtain the static stiffness, dynamic stiffness peak frequency data and the results of the orthogonal experiment under each combination.

[0061] Table 2 Results of the orthogonal experiment of the embodiment

[0062]

[0063]

[0064] Table 3 Static Stiffness Response Table of Examples

[0065] Horizontal <![CDATA[R m > <![CDATA[t m > d <![CDATA[ω r > 1 279.572 243.774 196.762 241.267 2 277.402 254.323 237.525 254.079 3 271.965 276.425 272.118 273.863 4 273.171 288.602 316.246 290.984 5 272.002 310.987 351.460 312.918 Range 7.57 67.213 154.698 71.651

[0066] Table 4 Dynamic Stiffness Peak Frequency Response Table of Examples

[0067] Horizontal <![CDATA[R m > <![CDATA[t m > d <![CDATA[ω r > 1 476.387 361.162 304.302 329.284 2 382.185 354.539 329.795 331.799 3 352.437 392.546 367.124 365.966 4 304.027 353.516 390.663 375.598 5 273.641 326.914 396.792 387.029 Range 202.746 34.247 92.490 58.746

[0068] 3) In combination with finite element software, first use the Python language to batch build the mathematical models of each example in the orthogonal test scheme inside the software, then divide the mesh according to the quarter model, generate the mesh inp file of the entire digital model, modify the material properties and boundary settings in the inp file, complete the finite element model, calculate the static stiffness of the orthogonal test scheme and the dynamic stiffness curve within the required range, and use the generated odb file to extract the frequency at which the static stiffness and the maximum dynamic stiffness within the required range appear, that is, the dynamic stiffness peak frequency. Mainly control the dynamic stiffness in the entire frequency domain by controlling the magnitude of the peak frequency.

[0069] (4) During the optimization design process, take the structural parameters of the rubber mount as the design variables, and the static stiffness and the dynamic stiffness peak frequency as the optimization objectives. Train the neural network model and the neural network model optimized by the genetic algorithm respectively to predict the static stiffness and the dynamic stiffness peak frequency of the samples in the orthogonal test model. The comparison results are Figures 3 - 6 , Figure 3 and Figure 5 are the results of the static stiffness and the dynamic stiffness predicted by the neural network respectively. Figure 4 and Figure 6 are the results of the static stiffness and the dynamic stiffness predicted after optimizing the network parameters. Compared with before optimization, the prediction results are more accurate.

[0070] (5) Save the neural network model optimized by the genetic algorithm. In combination with the NSGA-II genetic algorithm, take the static stiffness of 400 N / mm and the dynamic stiffness peak frequency of 300 Hz as the optimization objectives, establish a difference objective function, optimize the structural parameters of the rubber mount, and draw the Pareto front diagram according to the optimization results as shown in Figure 7, appropriate rubber mount parameters are selected. Since the peak dynamic stiffness frequency has a greater impact on the high-frequency performance of the mount compared to the static stiffness, data with a smaller peak dynamic stiffness is preferentially considered. Considering the dimensional errors in actual production, in this example, a frame radius of 30 mm, a frame thickness of 3 mm, a rubber thickness of 26 mm, and a rubber width of 17 mm are selected. The calculated static stiffness is 394.34 N / mm, and the peak dynamic stiffness frequency is 314.4 Hz. The percentage errors from the target values are 1.42% and 4.8% respectively, completing the optimization of the rubber mount.

[0071] As an embodiment, the steps for constructing the neural network model include:

[0072] (s1) Set up a multi-layer neural network. The number of nodes in the input layer is selected as 4 structural parameters, the number of nodes in the hidden layer is selected as 11 according to an empirical formula, and the number of nodes in the output layer is selected as 2, namely the static stiffness and the peak dynamic stiffness frequency;

[0073] (s2) Select the activation function Sigmoid and the loss function MSE;

[0074] (s3) Use the results of finite element calculations as the data set. The data set is divided into a training set and a test set in a ratio of 4:1, and the training set and the test set are normalized to improve the generalization ability of the model;

[0075] (s4) Use the genetic algorithm to optimize the initial parameters of the neural network. The initial parameters include the weights and thresholds connecting each neuron in the neural network, that is, the biases in the activation function;

[0076] (s5) Select and evaluate individuals in the initial parameter population based on the fitness function. The fitness function is usually associated with the performance of the neural network on the training data. Exchange some genes (i.e., weights or thresholds) of the selected parent individuals to generate offspring. To increase the population diversity and prevent premature convergence to a local optimal solution, randomly change some genes (i.e., weights or thresholds) on the offspring individuals, and iteratively generate a new population until the stop condition is met or a predetermined number of iterations is reached;

[0077] (s6) Select the optimal neural network parameters according to the fitness function, that is, the error between the predicted value and the test value.

[0078] The parameter settings of the genetic algorithm include: a. The population size is set to 30, the number of iterations is set to 100, the crossover probability is set to 0.6, and the mutation probability is set to 0.1; b. The fitness function is the sum of the absolute values of the errors between the predicted values of the neural network and the calculated values using the finite element model.

[0079] The function with optimized structural parameters uses the difference from the target value, i.e., the expected value in engineering practice. Here, the static stiffness target is 400 N / mm, and the peak frequency of the dynamic stiffness is 350 Hz.

[0080] The data preprocessing of the neural network model includes: a. Checking the correctness of the finite element model, that is, whether there are unpredictable large deviations in the results calculated by the finite element method, and using the stiffness information of the samples in the orthogonal experiment as the data set of the neural network; b. Shuffling the samples in the orthogonal experiment to improve the generalization ability of the model.

[0081] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principle and practical application of the present invention, so that those skilled in the art can understand and utilize the present invention well.

Claims

1. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount, characterized in that, It includes the following steps: (1) Establish a finite element analysis model of the secondary vibration isolation rubber mount; (2) According to the parameters of the secondary vibration isolation rubber mount, set an orthogonal test scheme to evaluate the influence of each parameter on the performance; (3) Use finite element software to calculate the static stiffness and dynamic performance of the orthogonal test scheme; (4) Use a neural network model optimized by a genetic algorithm to predict the static stiffness and dynamic performance of the orthogonal test scheme; (5) Use a genetic algorithm to optimize the structural parameters of the rubber mount with the static stiffness and the peak frequency of the dynamic stiffness as the optimization objectives.

2. The dynamic performance optimization method of a secondary vibration isolation rubber mount according to claim 1, wherein, The orthogonal test scheme includes the following steps: (1) Determine the key parameters that affect the performance of the rubber mount; (2) Set different levels for each parameter; (3) Construct an orthogonal table, which contains the combinations of the key parameters and the levels of the key parameters; (4) Conduct finite element calculations according to the orthogonal table to obtain the static stiffness and the peak frequency data of the dynamic stiffness under each combination; In the optimization design process, use the structural parameters of the rubber mount as the design variables and the static stiffness and the peak frequency of the dynamic stiffness as the optimization objectives.

3. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 1, characterized in that, The construction of the neural network model includes the following steps: (s1) Set up a multi-layer neural network, which includes an input layer, a hidden layer, and an output layer. The number of nodes in the input layer is determined according to the number of parameters, the number of nodes in the hidden layer is determined by comparing the training effects in turn according to the empirical formula, and the number of nodes in the output layer is determined according to the number of calculation objectives; (s2) Select an activation function and a loss function; (s3) Use the stiffness information of the secondary vibration isolation rubber mount calculated by finite element software as the data set, divide the data set into a training set and a test set, and perform normalization processing on the training set and the test set to improve the generalization ability of the model; (s4) Use a genetic algorithm to optimize the initial parameters of the neural network, and the initial parameters include weights and thresholds; (s5) Through the selection, crossover, and mutation operations of the genetic algorithm, iteratively generate a new population until the stop condition is met or the number of iterations is reached; (s6) According to the fitness function, that is, the error between the predicted value and the experimental value of the neural network, select the optimal neural network parameters.

4. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 3, characterized in that, The parameter settings of the genetic algorithm include: a. The population size, the number of iterations, the crossover probability, and the mutation probability are all determined in advance according to the experimental design; b. The fitness function is the sum of the absolute values of the errors between the predicted values based on the neural network and the calculated values in the finite element.

5. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 1, characterized in that In step (5), the function for optimizing the structural parameters of the rubber mount uses the difference from the target value.

6. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 1, characterized in that, The preprocessing of the data set of the neural network model includes the following steps: a. Check the correctness of the finite element model, that is, use the results calculated by finite element as the data set of the neural network, where the results calculated by finite element are the results with an error of no more than 10% from the suspension performance parameters in actual production; b. Shuffle the samples in the orthogonal experiment to improve the generalization ability of the model.

7. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 1, characterized in that: The secondary vibration isolation rubber mount includes a metal inner core, an outer tube, an intermediate skeleton and a rubber bracket. The rubber bracket includes an inner-layer rubber, an outer-layer rubber and a plurality of connecting arms distributed in a cross shape along the circumferential direction of the inner-layer rubber and the outer-layer rubber. The inner ends of the inner-layer rubber connecting arms are integrally connected with the inner-layer rubber, the outer ends of the inner-layer rubber connecting arms are integrally connected with the inner side of the intermediate skeleton, the inner ends of the outer-layer rubber connecting arms are integrally connected with the outer side of the intermediate skeleton, the outer ends of the outer-layer rubber connecting arms are integrally connected with the inner side of the outer tube, the metal inner core is fixedly connected with the output shaft of the driving motor, and the outer tube is fixed to the subframe.

8. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 7, characterized in that: The intermediate skeleton is an annular structure located inside the rubber bracket for connecting the inner-layer rubber and the outer-layer rubber connecting arms; the intermediate skeleton is made of metal or nylon; the number of the connecting arms of the inner-layer rubber and the outer-layer rubber is four, which are evenly distributed in a cross shape.

9. A method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to claim 7, characterized in that: The inner-layer rubber connecting arms are vulcanized and fixed to the metal inner core, and the inner-layer rubber connecting arms are vulcanized and fixed to the inner side of the intermediate skeleton. The outer-layer rubber connecting arms are fixed to the outer side of the intermediate skeleton, and the outer-layer rubber connecting arms are vulcanized and fixed to the inner side of the outer tube. The outer tube is made of nylon. Taking the axial direction of the output shaft of the driving motor as the x-axis, the up and down directions in the radial direction as the z-axis, and the left and right directions as the y-axis, the rubber connecting arms are all arranged at a 45-degree angle with the x direction.

10. A computer device, characterized in that, It includes: a memory, a processor and a computer program stored on the memory. When the computer program is executed on the processor, it realizes a method for optimizing the dynamic performance of a secondary vibration isolation rubber mount according to any one of claims 1 to 9.

Images