Model building method of shock absorber, whole vehicle simulation method, medium and electronic equipment

Abstract

The embodiment of the invention provides a model building method of a shock absorber, a whole vehicle simulation method, a medium and electronic equipment. The model building method comprises the following steps: obtaining bench test data of a shock absorber; preprocessing the bench test data to obtain target bench test data; importing the target bench test data into matlab software, distributing training data and test data according to a preset proportion, and training the neural network model based on the training data; obtaining a root mean square error and a correlation coefficient based on the test data; and converting the corresponding neural network model into a function prototype unit model under the condition of determining that the precision of the neural network model meets the requirement according to the root-mean-square error and the correlation coefficient. Therefore, according to the test data of the shock absorber rack test, the neural network model of the hydraulic recovery buffer shock absorber is obtained through matlab identification, and then the neural network model is converted into the functional prototype unit model, so that the road data test links are reduced, the load prediction efficiency can be improved, and the cost is reduced.

Description

Technical Field

[0001] This application relates to the field of dynamic simulation technology, and in particular to a method for establishing a model of a shock absorber, a method for simulating a whole vehicle, and media and electronic equipment. Background Technology

[0002] As a crucial component of automotive chassis suspension, shock absorbers play a vital role in isolating and attenuating vibrations from the road surface onto the vehicle body. Under extreme road impacts, when the damping force is insufficient to cushion the vehicle's movement, the shock absorber will reach its elastic limit at high speed. To absorb impact energy, a hydraulic damping structure is typically added to the end of a standard shock absorber. However, due to the complexity of shock absorber design theory, a reliable and stable design method for the critical valve system parameters within the shock absorber has yet to be found.

[0003] In related technologies, prototypes are made based on experience and then installed on vehicles for multiple rounds of road durability and bench frequency sweep tests. The tests are repeated and the parameters are modified, resulting in high development costs, long cycles, and low efficiency. Summary of the Invention

[0004] This application aims to at least partially address one of the technical problems in the related art. Therefore, the purpose of this application is to propose a method for establishing a shock absorber model, a vehicle simulation method, a medium, and electronic equipment. By constructing a shock absorber model using a neural network to replace the original physical shock absorber, simulation testing is completed, reducing the road data testing steps, improving the efficiency of load prediction, and lowering costs.

[0005] To achieve the above objectives, the first aspect of this application proposes a method for establishing a vibration damper model. The method includes: acquiring bench test data of the vibration damper; preprocessing the bench test data to obtain target bench test data; importing the target bench test data into MATLAB software, allocating training data and test data according to a preset ratio, and training a neural network model based on the training data; obtaining the root mean square error and correlation coefficient based on the test data; and, if the accuracy of the neural network model meets the requirements based on the root mean square error and correlation coefficient, converting the corresponding neural network model into a functional prototype unit model.

[0006] According to one embodiment of this application, the above-mentioned method for establishing a vibration damper model further includes: if the accuracy of the neural network model is not satisfactory based on the root mean square error and correlation coefficient, modifying the number of layers in the neural network model and retraining the neural network model based on the training data until the accuracy of the neural network model meets the requirements.

[0007] According to one embodiment of this application, the bench test data includes displacement data, velocity data, and damping force data. Importing the target bench test data into MATLAB software includes: using the displacement data and velocity data as input to a neural network model, and using the damping force data as output to the neural network model.

[0008] According to one embodiment of this application, the root mean square error and correlation coefficient are obtained using the following formula:

[0009]

[0010] Where RMSE represents the root mean square error, R 2 The correlation coefficient is represented by m, the number of samples is represented by y, and the measured value of the damping force is represented by y. This represents the neural network prediction value. This represents the sample mean.

[0011] According to one embodiment of this application, preprocessing of bench test data includes: performing at least one of channel sorting, detrending, deburring, and filtering high-frequency noise signals on the bench test data to achieve preprocessing.

[0012] According to one embodiment of this application, converting the corresponding neural network model into a functional prototype unit model includes: Export the corresponding neural network model to the Simulink platform, and set displacement and velocity as inputs and damping force as output; export the Simulink model as a functional prototype unit model, and select the version and co-simulation mode.

[0013] To achieve the above objectives, a second aspect of this application proposes a vehicle simulation method for a shock absorber. The simulation method includes the aforementioned model establishment method. The vehicle simulation method for the shock absorber further includes: establishing a vehicle dynamics model in dynamics simulation software based on vehicle design parameters; importing the exported functional prototype unit model into the vehicle dynamics model; connecting the functional prototype unit model with the suspension model; and performing simulation based on preset simulation conditions, time, and road surface to obtain load results.

[0014] According to one embodiment of this application, connecting a functional prototype unit model to a suspension model includes: connecting the output interface of the functional prototype unit to an actuator of the suspension model to output damping force, wherein the actuator is used to apply damping force; and connecting the input interface of the functional prototype unit to two sensors of the suspension model to obtain displacement and velocity, wherein the sensors are used to obtain displacement and velocity respectively.

[0015] To achieve the above objectives, a third aspect of this application provides a computer-readable storage medium storing a vehicle control program. When executed by a processor, the vehicle control program implements the aforementioned method for establishing a model of a shock absorber, or the aforementioned method for simulating a whole vehicle using a shock absorber.

[0016] To achieve the above objectives, a fourth aspect of this application provides an electronic device, including: a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned method for establishing a model of a shock absorber, or implements the aforementioned method for simulating a whole vehicle using a shock absorber.

[0017] According to the damper model establishment method, vehicle simulation method, medium, and electronic equipment embodiments of this application, based on the damper bench test data, a neural network model of the hydraulic restoring buffer damper is identified using MATLAB, and then imported into the vehicle multibody dynamics model to achieve vehicle-level simulation. This method reduces the road data testing step, improves the efficiency of load prediction, and reduces costs. Attached Figure Description

[0018] Figure 1 This is a structural diagram of a hydraulically recovering shock absorber. Figure 2 This is a flowchart of a method for modeling a vibration damper according to some embodiments of this application; Figure 3 This is a flowchart of a vehicle simulation method for a shock absorber according to some embodiments of this application; Figure 4 This is a simulation flowchart based on some embodiments of this application; Figure 5 This is a block diagram of an electronic device according to some embodiments of this application. Detailed Implementation

[0019] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0020] The following describes in detail, with reference to the accompanying drawings, the model building method for the shock absorber, the vehicle simulation method, the medium, and the electronic equipment of the present application embodiments.

[0021] The hydraulic recovery buffer damper in this application is a cylindrical double-cavity damping element that dissipates kinetic energy by utilizing the pressure loss generated when oil flows through the throttling orifice of the valve system.

[0022] Figure 1 This is a structural diagram of a hydraulic recovery buffer damper.

[0023] Specific reference Figure 1 Its dual-chamber structure consists of a piston cylinder, a floating piston, and a high-pressure air chamber. The floating piston divides the cylinder into an oil chamber and a high-pressure air chamber. The hydraulic recovery damper can automatically adjust the damping force according to the speed through the valve opening degree during the recovery stroke. When approaching the limit displacement, it generates a sharp increase in recovery resistance with the help of the buffer ring, which not only suppresses vehicle body vibration but also prevents impact limit, realizing the dual functions of nonlinear damping control throughout the stroke and end-of-stroke energy absorption.

[0024] In this application, MATLAB refers to the scientific computing and algorithm development environment launched by MathWorks, which has built-in toolboxes for matrix operations, numerical optimization, statistical analysis, and deep learning. In this solution, it is responsible for data preprocessing, neural network training and weight export, and one-click generation of Simulink modules, providing the algorithm kernel for subsequent modeling and simulation.

[0025] Simulink is a graphical modeling and simulation platform for MATLAB that supports drag-and-drop construction of multi-domain dynamic systems. By directly importing trained networks through the Neural Network module and binding input / output ports, time-domain, frequency-domain, and real-time simulations can be performed. It serves as a bridge connecting algorithms and system-level verification.

[0026] An FMU (Functional Mock-up Unit) is a binary package file that conforms to the FMI standard, containing model equations, solver interfaces, and variable descriptions. After exporting a Simulink model as an FMU, it can be directly called by multibody dynamics software, achieving "plug-and-play" co-simulation and completing a seamless migration from neural networks to vehicle load analysis.

[0027] Figure 2 This is a flowchart illustrating a method for modeling a vibration damper according to some embodiments of this application. (Refer to...) Figure 2 Methods for establishing a model of a vibration damper may include: S101, Obtain bench test data for the vibration damper.

[0028] Specifically, the vibration damper is first fixed to the test bench, and then a sinusoidal displacement signal is applied as the excitation signal. The amplitude of the excitation signal must include... Figure 1 The working range of the limiting structure in the test is used to test the performance of the vibration damper under extreme working conditions, and finally the bench test data is obtained to provide data support for subsequent modeling.

[0029] S102, preprocess the bench test data to obtain the target bench test data.

[0030] Specifically, the obtained bench test data of the vibration damper is checked and corrected to improve data quality and further obtain the target bench test data.

[0031] S103. Import the target bench test data into MATLAB software, allocate training data, validation data and test data according to the preset ratio, and train the neural network model based on the training data.

[0032] Specifically, first, the input and output data in the target bench test data are determined, and the relevant parameters of the neural network structure are set, such as setting the layer size to 10; then, the data is imported into the neural network fitting tool in MATLAB software, and then the training data, validation data and test data are allocated according to the preset ratio. Finally, the neural network model is trained based on the training data.

[0033] The preset ratio is used to divide the target bench test data into training data, validation data, and test data. For example, the preset ratio can be set to 70% for training data, 15% for validation data, and 15% for test data. The preset ratio is set to make efficient use of data resources, ensure the comprehensiveness of model training, the reliability of validation, and the objectivity of test results, thereby improving the model's generalization ability and stability.

[0034] S104, obtain the root mean square error and correlation coefficient based on the test data.

[0035] Specifically, after the neural network model has finished training, the root mean square error of the neural network model is obtained using test data. and correlation coefficient This provides support for evaluating the accuracy of neural network models.

[0036] S105, if the accuracy of the neural network model meets the requirements based on the root mean square error and correlation coefficient, the corresponding neural network model is converted into a functional prototype unit model.

[0037] Specifically, after obtaining the root mean square error and correlation coefficient Then, the root mean square error is... and correlation coefficient The accuracy of the neural network model is evaluated by comparing it with a preset accuracy standard. For example, the accuracy standard can be set as RMSE ≤ 2% and R 2 ≥0.98. If the accuracy of the neural network model meets the requirements, then the neural network model is further converted into a functional prototype unit model.

[0038] A sinusoidal displacement excitation covering the limiting range is applied to a test bench to obtain the full-stroke damping characteristics in one go. After detrending and deburring preprocessing, the model is divided into training, validation, and test sets according to a preset ratio. Network training is completed in MATLAB. The root mean square error and correlation coefficient are used as unified criteria to evaluate the model accuracy. Once the accuracy is qualified, it is converted into a functional prototype unit model. This solution shortens the modeling-validation-delivery process of the vibration damper, and the neural network model can be quickly embedded into the multibody dynamics simulation of the whole vehicle, reducing development costs and significantly improving the model's generalization ability and engineering practicality.

[0039] In some embodiments of this application, the above-described method for establishing a vibration damper model further includes: if the accuracy of the neural network model is not satisfactory based on the root mean square error and correlation coefficient, modifying the number of layers in the neural network model and retraining the neural network model based on the training data until the accuracy of the neural network model meets the requirements.

[0040] Specifically, if after the neural network model training is completed, the results calculated based on the root mean square error (RMSE) and correlation coefficient using test data and actual sample data... If the accuracy requirements of the preset neural network model are not met, the neural network structure parameters are modified, for example, the number of layers in the neural network model is changed, and the neural network model is retrained based on the training data until the accuracy of the neural network model meets the requirements.

[0041] If the RMSE calculated from the test data is equal to... If the preset accuracy requirements are not met simultaneously, the system will modify the neural network structure parameters, such as the number of neural network layers, while keeping the input / output layer structure unchanged. Then, it will retrain based on the original training data until the accuracy requirements are met, and the iteration will terminate to ensure that the final neural network model has the high accuracy and consistency required for vehicle-level simulation.

[0042] In some embodiments of this application, the bench test data includes displacement data, velocity data, and damping force data. Importing the target bench test data into MATLAB software includes: using displacement data and velocity data as input to a neural network model, and using damping force data as output to the neural network model.

[0043] Specifically, the bench test data includes displacement data, velocity data, and damping force data. Displacement data can be obtained from the applied sinusoidal displacement excitation signal, reflecting the displacement changes of the damper under different excitations; velocity data can be obtained from the frequency f of the excitation signal using the formula v. The calculated velocity characteristics of the shock absorber's motion are used to represent the damping force data. The damping force data can be obtained by synchronously measuring the tension and compression load sensors arranged at the piston rod end, which is used to represent the resistance of the shock absorber during operation.

[0044] Furthermore, when importing the target bench test data into MATLAB software, displacement data and velocity data are used as input data for training the neural network model to comprehensively reflect the motion state information of the shock absorber; damping force data are used as output data of the neural network model to train the model to specifically learn and predict the damping force variation law of the shock absorber.

[0045] By collecting displacement, velocity, and damping force data from shock absorber bench tests, a neural network model was constructed using MATLAB. Displacement and velocity were used as inputs, and damping force as the output for training. This method enables the model to accurately learn the dynamic performance of the shock absorber, effectively predict changes in damping force, and provide efficient and accurate data support for vehicle simulation, thereby improving simulation accuracy and efficiency.

[0046] In some embodiments of this application, the root mean square error and correlation coefficient are obtained using the following formulas:

[0047]

[0048] Where RMSE represents the root mean square error, R 2 The correlation coefficient is represented by m, the number of samples is represented by y, and the measured value of the damping force is represented by y. This represents the neural network prediction value. This represents the sample mean.

[0049] Specifically, firstly, based on the measured value of damping force y and the neural network prediction value... Calculate the sum of squared residuals This amplifies significant biases and suppresses sign cancellation; then the sum of squared residuals is... The square root is taken to restore the index to the same dimension as the damping force, obtaining the root mean square (RMS) form. Finally, it is divided by the product of the sample size m and the mean ȳ, m·ȳ, to complete dimensionless normalization, compressing the error to a percentage scale and obtaining the final root mean square error (RMSE). This is used to intuitively and uniformly evaluate the prediction accuracy of the vibration damper across its entire operating range. The closer the RMSE is to 0, the higher the accuracy of the neural network model's prediction results.

[0050] Furthermore, using the measured sample mean of damping force as a benchmark, the total sum of squares of fluctuations is calculated. This is used to characterize the inherent dispersion of the sample; subsequently, the sum of squared residuals is... Divide this by the model to obtain the proportion of volatility not explained by the model; finally, subtract this proportion from 1 to obtain the correlation coefficient. This achieves dimensionless normalization, thus allowing the neural network model of the vibration damper to intuitively reflect the interpretability and goodness of fit of the measured data using single-valued quantities within the range of 0–1. The correlation coefficient is among these values. The closer the value is to 1, the higher the accuracy of the neural network model's prediction results.

[0051] In some embodiments of this application, preprocessing of bench test data includes at least one of channel sorting, detrending, deburring, and filtering high-frequency noise signals to achieve preprocessing.

[0052] The process includes: channel sorting (first unifying naming and time reference, eliminating invalid channels to ensure one-to-one correspondence between displacement, velocity, and damping force); detrending (using zero-point offset and linear correction to eliminate sensor temperature drift and slow bench drift, bringing the baseline to zero); glitch removal (removing transient spikes in the data to prevent extreme value interference); and high-frequency noise filtering (using a low-pass Butterworth filter to cut off electromagnetic interference above the highest effective frequency of the signal, preserving the true frequency band of the vibration damper).

[0053] Preprocessing the data using one of the four methods mentioned above will yield clean, stable, and traceable high-quality test data, laying the foundation for building accurate, robust, and reproducible shock absorber neural network models, and directly determining the accuracy of vehicle simulation and development efficiency.

[0054] In some embodiments of this application, the corresponding neural network model is converted into a functional prototype unit model, including: exporting the corresponding neural network model to the Simulink platform, setting displacement and velocity as inputs and damping force as outputs; exporting the Simulink model as a functional prototype unit model, and selecting the version and co-simulation mode.

[0055] Specifically, after training the neural network model and determining that the accuracy meets the standard, the neural network model is exported to the Simulink platform, with the damper displacement and velocity set as two input signals and the damper damping force as the output signal. Further, the connected Simulink model is exported as a functional prototype unit model in FMU format, selecting the version and co-simulation mode during export. For example, version 2.0 can be selected.

[0056] By directly importing the accepted neural network model into Simulink, binding displacement and velocity as inputs and damping force as outputs, and then exporting it in FMU 2.0 co-simulation mode, a functional prototype unit compatible with the vehicle multibody dynamics platform can be quickly obtained, achieving seamless model replacement and significantly shortening integration time.

[0057] In summary, based on the test data from the damper bench test, a neural network model of the hydraulic restoring damper was obtained using MATLAB.

[0058] Figure 3This is a flowchart of a vehicle simulation method for a shock absorber according to some embodiments of this application.

[0059] Reference Figure 3 The whole-vehicle simulation method for shock absorbers also includes: S201, a vehicle dynamics model is established in the dynamics simulation software based on the vehicle design parameters.

[0060] Specifically, based on the vehicle design parameters, a suspension model was built in multibody dynamics software. An actuator was created between the upper and lower mounting points of the shock absorber, and its force attribute was set to the shock absorber damping force. At the same time, two sensors were defined at the same location to measure the relative displacement and relative velocity of the two points, respectively. This was used to replace the physical shock absorber and to provide a standard interface for subsequent co-simulation.

[0061] S202, import the exported functional prototype unit model into the vehicle dynamics model.

[0062] Specifically, a functional prototype unit model, namely the shock absorber FMU, constructed using a neural network is imported to replace the original shock absorber entity in the suspension for subsequent simulation experiments.

[0063] S203 connects the functional prototype unit model with the suspension model and performs simulation based on preset simulation conditions, time, and road surface to obtain load results.

[0064] Specifically, the suspension model is assembled with other vehicle models to form a complete vehicle model, and simulation conditions, time, and road surface are set. After the simulation is completed, the required load results are extracted.

[0065] Within the vehicle dynamics template, a suspension interface is quickly constructed using a single actuator and dual sensors. Then, a damper FMU directly replaces the physical component, completing the port connection. Simulations can then run under preset operating conditions and output load spectra. This eliminates the need for physical dampers, significantly reducing testing time and cost while ensuring simulation accuracy and data traceability.

[0066] In some embodiments of this application, connecting the functional prototype unit model to the suspension model includes: connecting the output interface of the functional prototype unit model to the actuator of the suspension model to output damping force, wherein the actuator is used to apply damping force; and connecting the input interface of the functional prototype unit model to two sensors of the suspension model to obtain displacement and velocity, wherein the sensors are used to obtain displacement and velocity respectively.

[0067] Specifically, the output interface of the shock absorber FMU is connected to the actuator in the suspension model, and the actuator applies the damping force calculated by the shock absorber FMU based on the current displacement and velocity conditions; at the same time, the input interface of the shock absorber FMU is connected to the two displacement and velocity sensors of the suspension model to obtain the displacement and velocity signals of the shock absorber under the current conditions in real time.

[0068] The FMU (Fluid Mutation Unit) replaces the physical damper to complete the whole vehicle simulation, reducing the road data testing process, improving the efficiency of load prediction, and reducing costs.

[0069] As a concrete example, such as Figure 4 As shown, the simulation process includes the following steps: S301, obtained bench test data for hydraulic recovery damper.

[0070] The vibration damper is fixed to the test bench, and the excitation signal is a sinusoidal displacement signal with an amplitude that includes... Figure 1 The working range of the limiting structure. The relationship between signal frequency f and velocity amplitude v is: v = 2πf. The velocity amplitude v is determined by the standards of each company, that is, the velocity corresponding to the damper's damping curve. Read the time-domain data of the three signals of damper displacement, velocity, and damping force.

[0071] S302, Preprocessing of experimental data.

[0072] The three signals obtained in the previous step are then examined and corrected, including channel smoothing, detrending, glitch removal, and filtering of high-frequency noise signals. This improves data quality and enhances the neural network's feature acquisition capabilities.

[0073] S303 defines the grid input and output.

[0074] S304, set the neural network structure parameters.

[0075] S305, grid training and prediction.

[0076] The processed data was imported into MATLAB software. The vibration damper displacement and velocity were placed in one array as input, and the vibration damper force was placed in another array as output. The built-in neural network fitting tool of the software was opened, and the input and output data were imported. The training data was set to 70%, the validation data to 15%, the test data to 15%, and the layer size to 10. Training then began.

[0077] S306, Accuracy evaluation. If satisfied, proceed to step S307; otherwise, return to step S304.

[0078] Among them, based on the root mean square error (RMSE) and correlation coefficient To evaluate, m represents the number of samples, and y represents the sample value. Represents the network prediction value. Represents the sample mean; the closer RMSE is to 0, the better. The closer the value is to 1, the higher the network accuracy. If the accuracy does not meet the requirements, return to step three, modify the number of neural network layers, and retrain the model until the accuracy meets the requirements.

[0079] S307, exported to a Simulink module.

[0080] S308, set the input / output channels.

[0081] The process involves exporting a neural network model that meets accuracy requirements to Simulink, setting the damper displacement and velocity as two input signals, and the damper damping force as the output signal.

[0082] S309, export the FMU model, and execute S312.

[0083] The process involves exporting the connected Simulink model as an FMU format model, selecting version 2.0 and co-simulation mode during the export.

[0084] S310, the multibody dynamics model of the whole vehicle.

[0085] S311 defines the input and output of the vibration damper model.

[0086] Based on the vehicle design parameters, a suspension model was built in multibody dynamics software. One actuator was defined as the damper damping force, and two sensors were defined as the damper displacement and velocity, respectively.

[0087] S312, Import FMU model.

[0088] In this process, the FMU model of the shock absorber is imported and replaced the original shock absorber model in the suspension.

[0089] S313, Interface connection.

[0090] Specifically, the FMU's output interface is connected to the suspension's actuator, and the suspension's sensors are connected to the FMU's input and output interfaces. S314, simulation time and road surface settings.

[0091] S315, Load result extraction.

[0092] It should be noted that for details not disclosed in the vehicle simulation method for the shock absorber in the embodiments of this application, please refer to the details disclosed in the model building method for the shock absorber in the embodiments of this application, which will not be repeated here.

[0093] This application also provides a computer-readable storage medium storing a vehicle control program, which, when executed by a processor, implements the aforementioned method for establishing a model of a shock absorber, or the aforementioned method for simulating a whole vehicle using a shock absorber.

[0094] Figure 5 This is a structural diagram of an electronic device according to some embodiments of this application.

[0095] Corresponding to the above embodiments, this application also provides an electronic device, referring to... Figure 5 The electronic device 500 includes: a memory 510, a processor 520, and a program stored in the memory 510 and executable on the processor 520. When the processor 520 executes the program, it implements the aforementioned method for establishing a model of the shock absorber, or implements the aforementioned method for simulating the whole vehicle of the shock absorber.

[0096] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0097] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0098] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0099] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

[0100] Any process or method described in the flowchart or otherwise herein is to be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0101] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0102] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0103] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0104] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0105] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

[0106] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this application can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this application can be achieved, and this is not limited herein.

[0107] The specific embodiments described above do not constitute a limitation on the scope of protection of this application. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for establishing a model of a vibration damper, characterized in that, The method includes: Obtain bench test data for the vibration damper; The bench test data is preprocessed to obtain the target bench test data; The target bench test data is imported into MATLAB software, and training data, validation data and test data are allocated according to a preset ratio. The neural network model is then trained based on the training data. The root mean square error and correlation coefficient are obtained based on the test data. If the accuracy of the neural network model meets the requirements based on the root mean square error and the correlation coefficient, the corresponding neural network model is converted into a functional prototype unit model.

2. The method for establishing a model of a vibration damper according to claim 1, characterized in that, The method further includes: If the accuracy of the neural network model does not meet the requirements based on the root mean square error and the correlation coefficient, the number of layers in the neural network model is modified, and the neural network model is retrained based on the training data until the accuracy of the neural network model meets the requirements.

3. The method for establishing a model of a vibration damper according to claim 1, characterized in that, The bench test data includes displacement data, velocity data, and damping force data. Importing the target bench test data into MATLAB software includes: The displacement and velocity data are used as inputs to the neural network model, and the damping force data is used as the output of the neural network model.

4. The method for establishing a model of a vibration damper according to claim 3, characterized in that, The root mean square error and the correlation coefficient are obtained using the following formulas: Wherein, RMSE represents the root mean square error, R 2 The correlation coefficient is represented by m, the number of samples is represented by y, and the measured value of the damping force is represented by y. This represents the neural network prediction value. This represents the sample mean.

5. The method for establishing a model of a vibration damper according to claim 1, characterized in that, The bench test data is preprocessed, including: The bench test data is preprocessed by performing at least one of the following: channel sorting, trend removal, glitch removal, and high-frequency noise filtering.

6. The method for establishing a model of a vibration damper according to claim 1, characterized in that, The corresponding neural network model is converted into a functional prototype unit model, including: The corresponding neural network model was exported to the Simulink platform, and displacement and velocity were set as inputs, and damping force as output. Export the Simulink model as the functional prototype unit model, and select the version and co-simulation mode.

7. A vehicle simulation method for a shock absorber, characterized in that, The simulation method includes the model building method according to any one of claims 1-6, and the method further includes: A vehicle dynamics model is established in dynamics simulation software based on the vehicle design parameters. Import the exported functional prototype unit model into the vehicle dynamics model; The functional prototype unit model is connected to the suspension model, and simulation is performed based on preset simulation conditions, time, and road surface to obtain load results.

8. The vehicle simulation method for the shock absorber according to claim 7, characterized in that, Connecting the functional prototype unit model to the suspension model includes: The output interface of the functional prototype unit model is connected to the actuator of the suspension model to output damping force, wherein the actuator is used to apply damping force; The input interface of the functional prototype unit model is connected to two sensors of the suspension model to obtain displacement and velocity, wherein the sensors are used to obtain the displacement and velocity, respectively.

9. A computer-readable storage medium, characterized in that, It stores the vehicle control program, which, when executed by the processor, implements the model building method for the shock absorber according to any one of claims 1-6, or the whole vehicle simulation method for the shock absorber according to claim 7 or 8.

10. An electronic device, characterized in that, include: The system includes a memory, a processor, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements the model building method for the shock absorber according to any one of claims 1-6, or the whole-vehicle simulation method for the shock absorber according to claim 7 or 8.

Images