Simulation benchmarking methods, simulation benchmarking equipment and storage media for low-speed collision conditions

By adding key information and parameter analysis of low-speed collision conditions to the high-speed collision simulation model, the problem of low-speed collision simulation benchmarking was solved, achieving an efficient and accurate simulation benchmarking process and reducing time and manpower costs.

CN115391990BActive Publication Date: 2026-11-17GUANGZHOU AUTOMOBILE GROUP CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202210898527.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2022-07-28
Publication Date
2026-11-17
Estimated Expiration
2042-07-28

AI Technical Summary

Technical Problem

Simulation benchmarking under low-speed collision conditions is quite difficult. The simulation benchmarking process is time-consuming and labor-intensive, and the accuracy is low. Existing technologies cannot effectively solve this problem.

Method used

The vehicle collision simulation model is benchmarked using test data from high-speed collision conditions to obtain a high-speed collision simulation model. Key information from low-speed collision conditions is added to this model, and key component parameters are extracted and analyzed. Sensitivity, correlation, and effectiveness analyses are performed to select effective benchmarking parameters. Finally, low-speed collision test data is used for simulation benchmarking.

Benefits of technology

It simplifies the construction efficiency of low-speed collision simulation models, improves the efficiency and accuracy of simulation benchmarking, and saves time and manpower costs.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN115391990B_ABST
    Figure CN115391990B_ABST
Patent Text Reader

Abstract

This invention discloses a simulation benchmarking method, simulation benchmarking device, and storage medium for low-speed collision conditions. The method includes: using high-speed collision test data to perform simulation benchmarking on a whole-vehicle collision simulation model to obtain a high-speed collision simulation model; processing the high-speed collision simulation model using target key information corresponding to the low-speed collision condition to obtain a low-speed collision simulation model; extracting key component parameters from the low-speed collision simulation model to obtain target key component parameters; performing sensitivity analysis on the target key component parameters to determine effective key component parameters; performing correlation analysis on the effective key component parameters to obtain original benchmarking parameters; performing validity analysis on the original benchmarking parameters to obtain effective benchmarking parameters; and using low-speed collision test data to perform simulation benchmarking on the effective benchmarking parameters corresponding to the low-speed collision simulation model to obtain the target collision simulation model. This method simplifies the simulation benchmarking process for low-speed collision conditions, improves efficiency and accuracy, and saves costs.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of automotive collision simulation benchmarking technology, and in particular to a simulation benchmarking method, simulation benchmarking equipment, and storage medium for low-speed collision conditions. Background Technology

[0002] With the rapid development of automotive technology, competition in the development of vehicle collision structures is becoming increasingly fierce. Currently, there is considerable experience in simulating high-speed collision conditions, especially under CNCCAP, CIASI, and national standards, and the accuracy of dynamic high-speed simulations has reached a high standard. However, simulation benchmarking under low-speed collision conditions is less common. This is mainly because, under low-speed collision conditions, sensor detection results are highly correlated with structural components, plastic parts, and their connections. Due to the complexity of sensor signal waves and the numerous influencing factors, simulation benchmarking under low-speed collision conditions is very difficult, requiring significant time and manpower, and the accuracy is relatively low. Summary of the Invention

[0003] This invention provides a simulation benchmarking method, simulation benchmarking equipment, and storage medium for low-speed collision conditions, which simplifies the simulation benchmarking process for low-speed collision conditions, saves time and labor costs, and improves the efficiency and accuracy of simulation benchmarking.

[0004] A simulation benchmarking method for low-speed collision conditions includes:

[0005] Using high-speed collision test data corresponding to high-speed collision conditions, the whole vehicle collision simulation model is simulated and benchmarked to obtain a high-speed collision simulation model.

[0006] The high-speed collision simulation model is processed using the target key information corresponding to the low-speed collision condition to obtain the low-speed collision simulation model.

[0007] Key component parameters are extracted from the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model.

[0008] Sensitivity analysis is performed on the target key component parameters corresponding to the low-speed collision simulation model to determine the effective key component parameters corresponding to the low-speed collision simulation model.

[0009] Correlation analysis was performed on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the original benchmark parameters corresponding to the low-speed collision simulation model.

[0010] The effectiveness of the original benchmark parameters corresponding to the low-speed collision simulation model is analyzed to obtain the effective benchmark parameters corresponding to the low-speed collision simulation model.

[0011] Using low-speed collision test data corresponding to low-speed collision conditions, the effective benchmark parameters corresponding to the low-speed collision simulation model are simulated and benchmarked to obtain the target collision simulation model.

[0012] Preferably, the step of using high-speed collision test data corresponding to high-speed collision conditions to perform simulation benchmarking on the whole vehicle collision simulation model to obtain a high-speed collision simulation model includes:

[0013] Acquire high-speed collision test data under high-speed collision conditions;

[0014] Using the high-speed collision test data corresponding to the high-speed collision condition, the vehicle collision simulation model is simulated and benchmarked to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition;

[0015] If the high-speed simulation accuracy is less than the first accuracy threshold, the high-speed collision test data is used to adjust the whole vehicle collision simulation model.

[0016] If the high-speed simulation accuracy is not less than the first accuracy threshold, then the whole vehicle collision simulation model is determined as the high-speed collision simulation model.

[0017] Preferably, the high-speed collision test data corresponding to the high-speed collision condition is used to perform simulation benchmarking on the whole vehicle collision simulation model to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition, including:

[0018] Obtain the simulation feature curves output by the whole vehicle collision simulation model and the test feature curves corresponding to the high-speed collision test data;

[0019] The simulation characteristic curve and the experimental characteristic curve are cross-correlation evaluated to obtain cross-correlation evaluation scores;

[0020] The simulated characteristic curve and the experimental characteristic curve are used to evaluate the corridor and obtain the corridor evaluation score;

[0021] Based on the cross-correlation evaluation score and the corridor evaluation score, the high-speed simulation accuracy corresponding to the high-speed collision condition is obtained.

[0022] Preferably, the step of extracting key component parameters from the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model includes:

[0023] Key component parameters are extracted from the low-speed collision simulation model to obtain the original key component parameters corresponding to the low-speed collision simulation model.

[0024] The original key component parameters corresponding to the low-speed collision simulation model are optimized and sampled to obtain the target key component parameters corresponding to the low-speed collision simulation model.

[0025] Preferably, the step of performing sensitivity analysis on the target key component parameters corresponding to the low-speed collision simulation model to determine the effective key component parameters corresponding to the low-speed collision simulation model includes:

[0026] Sensitivity analysis is performed on the target key component parameters corresponding to the low-speed collision simulation model to obtain the target sensitivity corresponding to the target key component parameters;

[0027] Target key component parameters whose target sensitivity is greater than a preset sensitivity are determined as valid key component parameters.

[0028] Preferably, the step of performing correlation analysis on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the original benchmark parameters corresponding to the low-speed collision simulation model includes:

[0029] Correlation analysis was performed on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the parameter correlations corresponding to the effective key component parameters.

[0030] The top N effective key component parameters with high parameter correlation are determined as the original benchmark parameters corresponding to the low-speed collision simulation model, where N≥1.

[0031] Preferably, the step of performing an effectiveness analysis on the original benchmark parameters corresponding to the low-speed collision simulation model to obtain the effective benchmark parameters corresponding to the low-speed collision simulation model includes:

[0032] Obtain the simulation result curve corresponding to the original benchmark parameters output by the low-speed collision simulation model. The simulation result curve includes an upper boundary curve and a lower boundary curve.

[0033] Obtain the test result curves corresponding to the original benchmark parameters;

[0034] If the test result curve is between the upper boundary curve and the lower boundary curve, then the original benchmarking parameter is determined as the effective benchmarking parameter corresponding to the low-speed collision simulation model.

[0035] Preferably, after performing an effectiveness analysis on the original benchmark parameters corresponding to the low-speed collision simulation model and obtaining the effective benchmark parameters corresponding to the low-speed collision simulation model, the simulation benchmarking method for the low-speed collision condition further includes:

[0036] The effective benchmarking parameters corresponding to the low-speed collision simulation model are inverted and identified to obtain the inversion and identification results corresponding to the effective benchmarking parameters.

[0037] If the inversion identification result corresponding to the effective benchmarking parameter meets the preset conditions, then the effective benchmarking parameter is determined as the target benchmarking parameter corresponding to the effective benchmarking parameter;

[0038] Using low-speed collision test data corresponding to low-speed collision conditions, the target benchmarking parameters corresponding to the low-speed collision simulation model are simulated and benchmarked to obtain the target collision simulation model.

[0039] A simulation benchmarking device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned simulation benchmarking method for low-speed collision conditions.

[0040] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described simulation benchmarking method for low-speed collision conditions.

[0041] The aforementioned simulation benchmarking method, equipment, and storage medium for low-speed collision scenarios utilize target key information corresponding to the low-speed collision scenario to process the high-speed collision simulation model completed through high-speed collision scenario simulation benchmarking, thereby obtaining the low-speed collision simulation model. This ensures that the low-speed collision simulation model is built upon the high-speed collision simulation model completed through benchmarking, simplifying the construction efficiency of the low-speed collision simulation model and improving the efficiency and accuracy of low-speed collision scenario simulation benchmarking. Furthermore, key component parameters are extracted from the low-speed collision simulation model to extract target key component parameters relevant to the low-speed collision scenario and exclude key component parameters irrelevant to the low-speed collision scenario. Next, sensitivity analysis, correlation analysis, and effectiveness analysis are performed on the target key component parameters in sequence to select effective benchmark parameters that meet the sensitivity, correlation, and effectiveness criteria from a large number of benchmark parameters. This ensures that the effective benchmark parameters have the greatest impact on the simulation results of the low-speed collision simulation model, but their number is small, which helps to simplify the subsequent simulation benchmarking process. Finally, low-speed collision test data is used to perform simulation benchmarking on the effective benchmark parameters corresponding to the low-speed collision simulation model. This can achieve rapid benchmarking while ensuring the high accuracy of the target collision simulation model, thus saving simulation benchmarking time and manpower costs. Attached Figure Description

[0042] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0043] Figure 1 This is a flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0044] Figure 2 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0045] Figure 3 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0046] Figure 4 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0047] Figure 5 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0048] Figure 6 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0049] Figure 7 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0050] Figure 8 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention;

[0051] Figure 9 This is another flowchart of a simulation benchmarking method for low-speed collision conditions in one embodiment of the present invention. Detailed Implementation

[0052] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0053] This invention provides a simulation benchmarking method for low-speed collision conditions. This simulation benchmarking method can be applied to vehicle controllers or other simulation benchmarking devices that can perform information processing. It can reduce the difficulty of simulation benchmarking for low-speed collision conditions, reduce the time and manpower costs of simulation benchmarking, and ensure the accuracy of simulation benchmarking.

[0054] In one embodiment, such as Figure 1 As shown, a simulation benchmarking method for low-speed collision conditions is provided. Taking the application of this method in a simulation benchmarking device as an example, the method includes the following steps:

[0055] S101: Using high-speed collision test data corresponding to high-speed collision conditions, the whole vehicle collision simulation model is simulated and benchmarked to obtain a high-speed collision simulation model.

[0056] S102: Using the key target information corresponding to the low-speed collision condition, process the high-speed collision simulation model to obtain the low-speed collision simulation model.

[0057] S103: Extract key component parameters from the low-speed collision simulation model and obtain the target key component parameters corresponding to the low-speed collision simulation model.

[0058] S104: Perform sensitivity analysis on the target key component parameters corresponding to the low-speed collision simulation model to determine the effective key component parameters corresponding to the low-speed collision simulation model.

[0059] S105: Perform correlation analysis on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the original benchmark parameters corresponding to the low-speed collision simulation model.

[0060] S106: Perform an effectiveness analysis on the original benchmark parameters corresponding to the low-speed collision simulation model to obtain the effective benchmark parameters corresponding to the low-speed collision simulation model.

[0061] S107: Using low-speed collision test data corresponding to low-speed collision conditions, simulate and benchmark the effective benchmark parameters corresponding to the low-speed collision simulation model to obtain the target collision simulation model.

[0062] The whole-vehicle collision simulation model is a pre-built simulation model used to simulate whole-vehicle collisions. The high-speed collision scenario refers to a collision occurring while the test vehicle is traveling at high speed. Correspondingly, the low-speed collision scenario refers to a collision occurring while the test vehicle is traveling at low speed. In this example, high speed refers to a vehicle speed exceeding a preset threshold (e.g., 60 km / h), while low speed refers to a vehicle speed not exceeding the preset threshold.

[0063] As an example, in step S101, since directly performing simulation benchmarking on the whole vehicle collision simulation model under low-speed collision conditions is complex and time-consuming, to simplify the process and overcome the shortcomings of direct simulation benchmarking under low-speed collision conditions, the simulation benchmarking equipment can first use high-speed collision test data corresponding to high-speed collision conditions to perform simulation benchmarking on the whole vehicle collision simulation model to obtain a high-speed collision simulation model. In this example, the high-speed collision simulation model refers to a high-precision high-speed collision test model under high-speed collision conditions.

[0064] In this example, to ensure the relevance of simulation benchmarking for high-speed and low-speed collision scenarios, it is necessary to ensure that the simulation benchmarking processes for both scenarios target the same critical collision area. That is, under high-speed collision scenarios, the vehicle collision simulation model corresponding to a specific critical collision area is used for benchmarking; under low-speed collision scenarios, the same vehicle collision simulation model corresponding to the same critical collision area is used for benchmarking. This ensures that the two collision scenarios target the same critical collision area, allowing low-speed collision scenario benchmarking to be performed after high-speed collision scenario benchmarking. Using the high-speed collision simulation model generated from high-speed collision scenario benchmarking simplifies the low-speed collision scenario benchmarking process, improves its efficiency, and reduces time and labor costs.

[0065] Among them, the target key information corresponding to the low-speed collision condition refers to the simulation benchmark information that affects the low-speed collision condition. This target key information can be determined based on the test results of the low-speed collision condition, and the target key information corresponding to the low-speed collision condition can be understood as the component information that differs from that of the high-speed collision condition.

[0066] As an example, in step S102, after obtaining a high-speed collision simulation model with a preset accuracy, the simulation benchmarking device can use the high-speed collision simulation model to build a low-speed collision simulation model under high-speed collision conditions. Generally, under high-speed collision conditions, it is not necessary to include non-metallic parts such as plastic parts in the high-speed collision simulation model. However, the simulation benchmarking of low-speed collision conditions is mainly used to verify or calibrate the airbag deployment time. At this time, non-metallic parts such as plastic parts become more critical. Therefore, when using the high-speed collision simulation model completed by simulation benchmarking to perform low-speed collision simulation benchmarking, the target key information corresponding to the low-speed collision condition containing non-metallic parts needs to be added to the high-speed collision simulation model to obtain the low-speed collision simulation model after adding the target key information.

[0067] In one embodiment, the target key information includes sensor information, suspension stiffness parameters, key component bolt information, and non-metallic component information.

[0068] As an example, the key target information for low-speed collision scenarios includes sensor information. This sensor information refers to the location and connection relationships of the sensors used during the low-speed collision process. In this example, the sensor information, including the location and connection relationships of the sensors related to the low-speed collision scenario in the test vehicle, needs to be written into the high-speed collision simulation model to improve the high-speed collision simulation model, ensure sensor information consistency, and help guarantee the accuracy of the low-speed collision simulation model for benchmarking.

[0069] As an example, key information for low-speed collision scenarios includes suspension stiffness parameters. Generally, suspension stiffness parameters are more sensitive under low-speed collision conditions. The stiffness parameters of the suspension in the high-speed collision simulation model need to be calibrated using the design stiffness parameters to ensure consistency. This addresses the balance between suspension and vehicle weight, and helps ensure the accuracy of the low-speed collision simulation model for benchmarking.

[0070] As an example, the key information for low-speed collision scenarios includes bolt information in critical locations. This information refers to modeling the bolts in critical locations using solid elements and adding information about failure and preload to the constructed bolt model. In this example, the bolt information is added to the high-speed collision simulation model to check and improve it. This ensures that the adjusted high-speed collision simulation model can accurately address bolt failure issues, thus helping to guarantee the accuracy of the low-speed collision simulation model for benchmarking.

[0071] As an example, the key information for low-speed collision scenarios includes information on non-metallic components, which includes, but is not limited to, information on various types of plastic parts. In this example, it is necessary to pre-calibrate the constitutive properties of non-metallic components related to low-speed collision scenarios in the material library. If the deviation is high, the non-metallic components will undergo significant deformation during the simulation benchmarking process of low-speed collision scenarios, thus affecting the accuracy of the simulation benchmarking. Therefore, it is necessary to first use more accurate non-metallic component information to check and improve the high-speed collision simulation model, so that the adjusted high-speed collision simulation model can achieve accurate simulation, which helps to ensure the accuracy of the low-speed collision simulation model for simulation benchmarking.

[0072] As an example, in step S103, after acquiring the low-speed collision simulation model, the simulation benchmarking equipment needs to extract key components with significant deformation under low-speed collision conditions from the low-speed collision simulation model and determine the thickness and material data of these key components as target key component parameters. During the collision simulation model building process, the thickness changes, material hardening, and residual stress that occur after sheet metal stamping are usually not considered. If all sheet metal hardening information (including but not limited to thickness, stress, and stress) is directly added to the collision simulation model, this process requires obtaining a large amount of sheet metal information after stamping, which will make the collision simulation model building difficult. To simplify the impact of sheet metal stamping on the collision simulation model and reduce the influence of material constitutive differences, it is necessary to extract the thickness and material data of key components with significant deformation under low-speed collision conditions from the low-speed collision simulation model and determine them as target key component parameters. In this example, the simulation benchmarking device can extract target key component parameters from the low-speed collision simulation model, including but not limited to the elastic modulus of the vehicle's front bumper foam, yield strength, anti-collision beam thickness, energy absorption box thickness, and front bumper thickness. These target key component parameters can be used for simulation benchmarking to achieve the purpose of rapid and accurate simulation benchmarking, which helps to improve the efficiency of simulation benchmarking.

[0073] As an example, in step S104, after obtaining the target key component parameters corresponding to the low-speed collision simulation model, the simulation benchmarking device can analyze the sensitivity of each target key component parameter to the simulation results of the low-speed collision simulation model, and select the target key component parameters with higher sensitivity as effective key component parameters. In this example, effective key component parameters with higher sensitivity are selected from the target key component parameters with a large initial range, so that the effective key component parameters can be used for subsequent simulation benchmarking operations. This can reduce the number of parameters for simulation benchmarking and achieve the effect of rapid benchmarking while ensuring high accuracy.

[0074] As an example, in step S105, after obtaining the effective key component parameters corresponding to the low-speed collision simulation model, the simulation benchmarking device can analyze the correlation between each effective key component parameter and the result indicators in the low-speed collision simulation model, and select the effective key component parameters with higher correlation as the original benchmarking parameters. In this example, the effective key component parameters with higher correlation are selected as the original benchmarking parameters so that the simulation benchmarking can be performed using the original benchmarking parameters, which can reduce the number of parameters for simulation benchmarking. Since the original benchmarking parameters have high correlation, a rapid benchmarking effect can be achieved while ensuring high accuracy.

[0075] As an example, in step S106, after obtaining the original benchmarking parameters corresponding to the low-speed collision simulation model, the simulation benchmarking device can obtain the simulation result curve and experimental result curve for each original benchmarking parameter. It then performs a validity analysis on the simulation result curve and experimental result curve, identifying the original benchmarking parameters with validity as valid benchmarking parameters. In this example, the simulation and experimental result curves of the original benchmarking parameters can be used to analyze the result error between simulation and experiment. If the result error is small, validity is determined, and the original benchmarking parameter with validity is identified as a valid benchmarking parameter. If the result error is large, validity is determined, and the original benchmarking parameter needs to be deleted. In this example, by performing a validity analysis on the original benchmarking parameters and selecting valid benchmarking parameters, the number of simulation benchmarking parameters can be reduced. Because the simulation and experiment of valid benchmarking parameters are valid, rapid benchmarking can be achieved while ensuring high accuracy.

[0076] As an example, in step S107, during the simulation benchmarking process under low-speed collision conditions, the test vehicle needs to be controlled to drive under low-speed collision conditions. Sensors mounted on the test vehicle are used to collect low-speed collision test data in real time, especially the low-speed collision test data corresponding to the effective benchmarking parameters. Then, using the low-speed collision test data collected under low-speed collision conditions, the effective benchmarking parameters corresponding to the low-speed collision simulation model are simulated and benchmarked to obtain the low-speed simulation accuracy corresponding to the low-speed collision conditions. This low-speed simulation accuracy reflects the degree of closeness between the simulation results output by the low-speed collision simulation model and the test results of the low-speed collision test data. If the low-speed simulation accuracy is less than a second accuracy threshold, the low-speed collision simulation model is adjusted using the low-speed collision test data; if the low-speed simulation accuracy is not less than the second accuracy threshold, the low-speed collision simulation model is determined as the target collision simulation model for completing the simulation benchmarking under low-speed collision conditions. Here, the second accuracy threshold refers to a pre-set threshold used to evaluate whether the preset accuracy standard required for the low-speed collision conditions has been met. For example, the second accuracy threshold can be set to 80% or 90%.

[0077] In this embodiment, target key information corresponding to the low-speed collision condition is used to process the high-speed collision simulation model completed by high-speed collision simulation benchmarking to obtain the low-speed collision simulation model. This ensures that the low-speed collision simulation model is built upon the high-speed collision simulation model completed by simulation benchmarking, which helps to simplify the construction efficiency of the low-speed collision simulation model and improve the simulation benchmarking efficiency and accuracy of the low-speed collision condition. Next, key component parameters are extracted from the low-speed collision simulation model to extract target key component parameters related to the low-speed collision condition and exclude key component parameters irrelevant to the low-speed collision condition. Then, target key components are sequentially... Sensitivity analysis, correlation analysis, and validity analysis are performed on key parameters to select effective benchmark parameters that meet these criteria from a large pool of benchmark parameters. This ensures that the effective benchmark parameters have the greatest impact on the simulation results of the low-speed collision simulation model, but their number is relatively small, which helps simplify the subsequent simulation benchmarking process. Finally, low-speed collision test data is used to perform simulation benchmarking on the effective benchmark parameters corresponding to the low-speed collision simulation model. This allows for rapid benchmarking while ensuring the high accuracy of the target collision simulation model, thus saving time and manpower costs in simulation benchmarking.

[0078] In one embodiment, such as Figure 2 As shown, step S101, which involves using high-speed collision test data corresponding to the high-speed collision condition to perform simulation benchmarking on the whole vehicle collision simulation model to obtain the high-speed collision simulation model, includes:

[0079] S201: Obtain high-speed collision test data under high-speed collision conditions;

[0080] S202: Using high-speed collision test data corresponding to high-speed collision conditions, the whole vehicle collision simulation model is simulated and benchmarked to obtain the high-speed simulation accuracy corresponding to high-speed collision conditions.

[0081] S203: If the high-speed simulation accuracy is less than the first accuracy threshold, the vehicle collision simulation model shall be adjusted using high-speed collision test data;

[0082] S204: If the high-speed simulation accuracy is not less than the first accuracy threshold, then the whole vehicle collision simulation model is determined as the high-speed collision simulation model.

[0083] Generally, before conducting simulation benchmarking under high-speed collision conditions, the simulation benchmarking equipment can receive user-inputted component information from the test vehicle. Using this component information, the equipment updates the pre-set component information of the whole-vehicle collision simulation model, ensuring consistency between the test vehicle's component information and the whole-vehicle collision simulation model during the simulation benchmarking process. This guarantees the feasibility of the simulation benchmarking. In essence, ensuring consistency between the test vehicle's component information and the whole-vehicle collision simulation model establishes a correlation between them, enabling the use of high-speed collision test data from the test vehicle under high-speed collision conditions to perform simulation benchmarking on the whole-vehicle collision simulation model, thus making the simulation benchmarking process feasible.

[0084] As an example, in step S201, during the simulation benchmarking process under high-speed collision conditions, the test vehicle needs to be controlled to drive under high-speed collision conditions. The high-speed collision test data collected in real time by sensors mounted on the test vehicle is also available. This high-speed collision test data refers to the test data collected by the test vehicle during its high-speed collision operation, including but not limited to the test vehicle's speed information, weight information, barrier height information, and test deformation.

[0085] As an example, in step S202, after acquiring the high-speed collision test data under high-speed collision conditions, the simulation benchmarking device can use the high-speed collision test data to perform simulation benchmarking on the whole vehicle collision simulation model to obtain the high-speed simulation accuracy corresponding to the high-speed collision conditions. This high-speed simulation accuracy can reflect the degree of closeness between the simulation results output by the whole vehicle collision simulation model and the high-speed collision test data.

[0086] In this example, the simulation benchmarking equipment first maps high-speed collision test data, such as speed, weight, and barrier height, onto a pre-adjusted vehicle collision simulation model. Then, based on the simulation results output by the vehicle collision simulation model and the test deformation data in the high-speed collision test data, the simulation benchmarking equipment determines the high-speed simulation accuracy of the vehicle collision simulation model. This accuracy is then used to assess whether the high-speed collision simulation benchmarking has been successfully completed.

[0087] The first accuracy threshold is a preset threshold used to evaluate whether the high-speed collision condition has been met. For example, the first accuracy threshold can be set to 85%.

[0088] As an example, in step S203, after obtaining the high-speed simulation accuracy corresponding to the high-speed collision condition, the simulation benchmarking device can compare the high-speed simulation accuracy with the preset first accuracy threshold. If the high-speed simulation accuracy is less than the first accuracy threshold, it means that the high-speed simulation accuracy of the whole vehicle collision simulation model has not reached the preset accuracy standard. At this time, the whole vehicle collision simulation model can be adjusted using high-speed collision test data to update the whole vehicle collision simulation model so that steps S201-S204 can be repeated until the high-speed simulation accuracy is not less than the first accuracy threshold. Only then can the simulation benchmarking operation of the high-speed collision condition be considered complete.

[0089] As an example, in step S204, after obtaining the high-speed simulation accuracy corresponding to the high-speed collision condition, the simulation benchmarking device can compare the high-speed simulation accuracy with a preset first accuracy threshold. If the high-speed simulation accuracy is not less than the first accuracy threshold, it indicates that the high-speed simulation accuracy of the whole vehicle collision simulation model has reached the preset accuracy standard. At this time, the whole vehicle collision simulation model can be determined as a high-speed collision simulation model. Here, the high-speed collision simulation model refers to the model after simulating and benchmarking the whole vehicle collision simulation model using high-speed collision test data under high-speed collision conditions.

[0090] In this embodiment, high-speed collision test data under high-speed collision conditions are used to perform simulation benchmarking on the whole vehicle collision simulation model to obtain the high-speed simulation accuracy corresponding to the high-speed collision conditions. Based on the comparison result between the high-speed simulation accuracy and the first accuracy threshold, it is determined whether the simulation benchmarking operation of the high-speed collision conditions has been completed, so as to ensure the accuracy of the high-speed collision simulation model and thus help improve the high-speed simulation accuracy of low-speed collision conditions.

[0091] In one embodiment, such as Figure 3 As shown, S202 involves using high-speed collision test data corresponding to high-speed collision conditions to perform simulation benchmarking on the whole vehicle collision simulation model, obtaining the high-speed simulation accuracy corresponding to high-speed collision conditions, including:

[0092] S301: Obtain the simulation feature curves output by the whole vehicle collision simulation model and the test feature curves corresponding to the high-speed collision test data;

[0093] S302: Perform cross-correlation evaluation on the simulation characteristic curve and the experimental characteristic curve, and obtain the cross-correlation evaluation score;

[0094] S303: Evaluate the corridor based on the simulation characteristic curve and the experimental characteristic curve, and obtain the corridor evaluation score;

[0095] S304: Obtain the high-speed simulation accuracy corresponding to the high-speed collision condition based on the cross-correlation evaluation score and the corridor evaluation score.

[0096] Since achieving the preset accuracy of the high-speed simulation model under high-speed collision conditions is a prerequisite for its applicability to simulation benchmarking under low-speed collision conditions, it is necessary to objectively evaluate the simulation benchmarking of the whole vehicle collision simulation model adjusted using high-speed collision test data. Specifically, based on pre-set indicators, key information of key collision areas, and market attention indicators, the high-speed simulation accuracy under high-speed collision conditions should be comprehensively evaluated.

[0097] As an example, in step S301, the simulation benchmarking device can obtain the simulation results output by the adjusted vehicle collision simulation model. Based on these simulation results, it can obtain the simulation characteristic curves corresponding to preset evaluation features. These simulation characteristic curves include, but are not limited to, the simulation characteristic curve corresponding to the evaluation feature of acceleration. In this example, the simulation characteristic curve refers to the curve related to a certain evaluation feature output by the vehicle collision simulation model after simulating high-speed collision test data. Accordingly, the simulation benchmarking device can obtain the high-speed collision test data corresponding to each moment and determine the test characteristic curve based on the high-speed collision test data corresponding to the same evaluation feature collected at different times. In this example, to ensure the comparability of the simulation characteristic curves and the test characteristic curves, it is necessary to obtain both the simulation characteristic curve and the test characteristic curve corresponding to the same evaluation feature (including but not limited to acceleration).

[0098] As an example, in step S302, the simulation benchmarking equipment can perform corridor evaluation on the simulation characteristic curve and the experimental characteristic curve according to the cross-correlation rating standard to obtain the cross-correlation rating score. The cross-correlation rating standard is an evaluation standard that calculates the correlation between two curves through specific analytical indicators. The cross-correlation rating score reflects the cross-correlation between the simulation characteristic curve and the experimental characteristic curve. In this example, the cross-correlation rating score can be a score between 0 and 1.

[0099] As an example, in step S303, the simulation benchmarking device can evaluate the simulation characteristic curve and the experimental characteristic curve according to the Corridor Rating standard to obtain a Corridor Rating score. The Corridor Rating standard is an evaluation standard that calculates the deviation between the two curves by defining or generating a corridor. The Corridor Rating score reflects the deviation between the simulation characteristic curve and the experimental characteristic curve; in this example, the Corridor Rating score can be a value between 0 and 1.

[0100] As an example, in step S304, after obtaining the cross-correlation evaluation score and corridor evaluation score corresponding to the simulation characteristic curve and the test characteristic curve, the simulation benchmarking equipment can comprehensively evaluate the cross-correlation evaluation score and corridor evaluation score to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition. Here, the high-speed simulation accuracy is the result of the comprehensive evaluation of the cross-correlation evaluation score and corridor evaluation score.

[0101] In this example, the simulation benchmarking device can employ, but is not limited to, a weighted algorithm to calculate the cross-correlation evaluation score and the corridor evaluation score using weighted methods to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition. For example, the simulation benchmarking device can use S = ω1K1 + ω2K2 to calculate the cross-correlation evaluation score and the corridor evaluation score using weighted methods to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition. Here, S represents the high-speed simulation accuracy, K1 and K2 refer to the cross-correlation evaluation score and the corridor evaluation score, respectively, ω1 and ω2 refer to the preset weights of the cross-correlation evaluation score and the corridor evaluation score, respectively, and ω1 + ω2 = 1. Generally, when both K1 and K2 are scores of 0-1, the closer the high-speed simulation accuracy S corresponding to the high-speed collision condition is to 1, the higher the high-speed simulation accuracy of the high-speed collision condition can be considered. For example, when the high-speed simulation accuracy S is not less than 85%, the simulation benchmarking of the high-speed collision condition can be considered complete. At this time, the resulting high-speed collision simulation model has high accuracy.

[0102] In this embodiment, cross-correlation evaluation and corridor evaluation are performed on the simulation characteristic curve and the experimental characteristic curve respectively to obtain the cross-correlation evaluation score and the corridor evaluation score. Then, the cross-correlation evaluation score and the corridor evaluation score are comprehensively evaluated to make the high-speed simulation accuracy corresponding to the high-speed collision condition have strong objectivity.

[0103] In one embodiment, such as Figure 4 As shown, step S302, which involves evaluating the cross-correlation between the simulation characteristic curve and the experimental characteristic curve and obtaining the cross-correlation evaluation score, includes:

[0104] S401: Calculate the correlation between the simulation characteristic curve and the experimental characteristic curve to obtain waveform correlation, peak correlation and phase correlation;

[0105] S402: Obtain the cross-correlation evaluation score based on waveform correlation, peak correlation, and phase correlation.

[0106] As an example, in step S401, during the cross-correlation analysis of the simulation characteristic curve and the experimental characteristic curve, the simulation benchmarking equipment needs to calculate the correlation of the waveform, peak value, and phase of the two curves respectively, and obtain the waveform correlation, peak value correlation, and phase correlation. Here, waveform correlation refers to the correlation between the waveforms of the simulation characteristic curve and the experimental characteristic curve. Peak value correlation refers to the correlation between the peak values ​​of the simulation characteristic curve and the experimental characteristic curve. Phase correlation refers to the correlation between the phases of the simulation characteristic curve and the experimental characteristic curve.

[0107] As an example, in step S402, after obtaining the waveform correlation, peak correlation, and phase correlation, the simulation benchmarking device can use, but is not limited to, a weighted algorithm to calculate the waveform correlation, peak correlation, and phase correlation to obtain the cross-correlation evaluation score. In this example, the simulation benchmarking device can use K1=ω 11 K 11 +ω 12 K 12 +ω 13 K 13 The waveform correlation, peak correlation, and phase correlation are calculated to obtain the cross-correlation evaluation score, where K1 is the cross-correlation evaluation score, and K... 11 K 12 and K 13 These refer to waveform correlation, peak correlation, and phase correlation, respectively, ω 11 ω 12 and ω 13 These refer to the preset weights for waveform correlation, peak correlation, and phase correlation, respectively.

[0108] In this embodiment, the correlation between the simulation characteristic curve and the experimental characteristic curve is calculated to obtain waveform correlation, peak correlation and phase correlation, etc. Then, the correlation of all indicators is comprehensively evaluated to obtain the cross-correlation evaluation score, so that the obtained cross-correlation evaluation score has strong objectivity.

[0109] In one embodiment, such as Figure 5 As shown, step S103 involves extracting key component parameters from the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model, including:

[0110] S501: Extract key component parameters from the low-speed collision simulation model to obtain the original key component parameters corresponding to the low-speed collision simulation model.

[0111] S502: Optimize and sample the original key component parameters corresponding to the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model.

[0112] As an example, in step S501, after acquiring the low-speed collision simulation model, the simulation benchmarking device can extract key component parameters from the low-speed collision simulation model to extract the original key component parameters that deform significantly under low-speed collision conditions. These original key component parameters include, but are not limited to, the elastic modulus of the front bumper foam, yield strength, anti-collision beam thickness, energy-absorbing box thickness, and front bumper thickness. Here, the original key component parameters refer to the key component parameters directly extracted from the low-speed collision simulation model; they are key component parameters that have not undergone optimized sampling.

[0113] As an example, in step S502, after obtaining the original key component parameters corresponding to the low-speed collision simulation model, the simulation benchmarking equipment can use an experimental design method to optimize and sample all the obtained original key component parameters to obtain the target key component parameters corresponding to the low-speed collision simulation model. The experimental design method is an experimental mathematical method that can reasonably arrange numerical experimental design points. It is a surrogate model sampling strategy for multidisciplinary design optimization parameter identification benchmarking, which determines the number of sample points selected by the surrogate model and the sampling form.

[0114] For example, a full factorial experimental design method can be used to optimize the sampling of original critical component parameters. Specifically, an approximate model combining full factorial experimental design and response surface methodology can be used for optimization sampling. The approximate model formula is as follows: Where y is the approximate response function; ξ0,ξ i ,ξ ii ,ξ ij The regression coefficient is determined by the full factorial experimental design; x i ,x j To design variable levels; Here, k represents the error rate, and k represents the number of design variable levels. The response surface here refers to the relationship between the response variable and the input variable. This relationship can usually be represented by a surface in space, hence the name response surface.

[0115] For example, the Latin hypercube experimental design method can be used to optimize the sampling of original critical component parameters. The Latin hypercube sampling point selection scheme is usually used for probabilistic analysis. Similar to Monte Carlo and space-filling point selection schemes, the number of trials needs to be determined. Latin hypercube sampling can be used to fit the response surface, but even if the Latin hypercube design has enough points to fit the response surface, it may still obtain poor prediction quality or approach singularities during the regression process.

[0116] For example, the orthogonal experimental design method, Latin hypercube design, can be used. Orthogonal experimental design evolved from Latin hypercube design, and its characteristics can be summarized as "uniform distribution and neat comparability." The entire process of orthogonal experimental design is determined by the orthogonal array, while non-standard orthogonal arrays do not possess orthogonality. Orthogonal experimental design only requires specifying the number of factors and levels to generate the corresponding orthogonal array. Subsequently, based on the orthogonal array and the given upper and lower limits of the factors, orthogonal experimental design is performed to generate a series of experimental points with orthogonality.

[0117] In this embodiment, the key component parameters are first extracted from the low-speed collision simulation model to extract the original key component parameters that deform significantly in the low-speed collision condition. Then, the original key component parameters are optimized and sampled. On the one hand, this ensures that the determined target key component parameters will undergo significant deformation in the low-speed collision condition. On the other hand, the target key component parameters that can be screened after optimization and sampling are more in line with the simulation design requirements. Furthermore, it ensures that the number of target key component parameters is small, which helps to reduce the number of parameters in subsequent processing and improves the simulation benchmarking efficiency.

[0118] In one embodiment, such as Figure 6 As shown, step S104 involves performing sensitivity analysis on the target key component parameters corresponding to the low-speed collision simulation model to determine the effective key component parameters corresponding to the low-speed collision simulation model, including:

[0119] S601: Perform sensitivity analysis on the target key component parameters corresponding to the low-speed collision simulation model, and obtain the target sensitivity corresponding to the target key component parameters;

[0120] S602: Target key component parameters with a target sensitivity greater than the preset sensitivity are determined as effective key component parameters.

[0121] As an example, in step S601, the simulation benchmarking device can analyze the sensitivity of all target key component parameters to the simulation results based on the simulation results of the low-speed collision simulation model, and obtain the target sensitivity corresponding to each target key component parameter. Here, the target sensitivity is the contribution of the target key component parameter to the target value in the simulation results. The target value refers to the expected value that needs to be achieved in the simulation results to reflect a certain indicator.

[0122] In this example, all target key component parameters are input into a low-speed collision simulation model for calculation. The magnitudes of individual variables (i.e., target key component parameters) in the low-speed collision simulation model are changed sequentially, and the contribution of changes in the magnitude of a particular individual variable (i.e., a target key component parameter) to a specific target value is analyzed. For example, when changing the single variable of acceleration, its impact on a specific target value is significant, indicating that it is highly sensitive to the target value. In this case, the obtained target sensitivity is high.

[0123] Among them, the preset sensitivity is a pre-set threshold used to evaluate whether the sensitivity meets the standard.

[0124] As an example, in step S602, after the simulation benchmarking device obtains the target sensitivity corresponding to each target key component parameter, it can compare the target sensitivity with the preset sensitivity. If the target sensitivity is greater than the preset sensitivity, it is determined that the sensitivity of the target key component parameter to the simulation result meets the standard. At this time, the target key component parameter can be determined as a valid key component parameter.

[0125] In this embodiment, based on the comparison results between the target sensitivity and the preset sensitivity, the target key component parameters with a large initial range can be selected, and some target key component parameters with low target sensitivity can be removed. Only some target key component parameters with high target sensitivity are retained as effective key component parameters, so that the effective key component parameters can be used for subsequent simulation benchmarking operations. This can reduce the number of parameters for simulation benchmarking. Since the effective key component parameters have high sensitivity, the effect of rapid benchmarking can be achieved while ensuring high accuracy.

[0126] In one embodiment, such as Figure 7 As shown, step S105 involves performing a correlation analysis on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the original benchmark parameters corresponding to the low-speed collision simulation model, including:

[0127] S701: Perform correlation analysis on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the parameter correlations corresponding to the effective key component parameters.

[0128] S702: The top N effective key component parameters with high parameter correlation are determined as the original benchmark parameters corresponding to the low-speed collision simulation model, where N≥1.

[0129] As an example, in step S701, the simulation benchmarking device can analyze the correlation between all effective key component parameters and the result indicators set in the simulation results based on the simulation results of the low-speed collision simulation model, in order to obtain the parameter correlation of the effective key component parameters. This parameter correlation reflects the correlation between the effective key component parameters and the result indicators in the simulation results.

[0130] Where N is the number of original benchmark parameters to be selected. As an example, the value of N can be a pre-set fixed value, or it can be determined by multiplying the total number of all valid key parameters by a preset ratio. For example, the preset ratio can be set to 30%, 35%, 40%, or other values.

[0131] As an example, in step S702, after the simulation benchmarking device obtains all valid key component parameters, it can obtain a pre-set screening quantity N, or determine the screening quantity N by multiplying the total number of valid key component parameters by a preset ratio; then, it sorts the parameter correlations corresponding to all valid key component parameters to obtain the correlation ranking results; from the correlation ranking results, it selects the top N valid key component parameters and determines them as the original benchmarking parameters corresponding to the low-speed collision simulation model. In this example, other valid key component parameters besides the top N valid key component parameters selected from the correlation ranking results will be deleted and will not participate in subsequent simulation benchmarking.

[0132] In this embodiment, the top N effective key component parameters with high parameter correlation are selected as the original benchmark parameters corresponding to the low-speed collision simulation model. This is to use the effective key component parameters with high parameter correlation as the original benchmark parameters so that the simulation benchmark can be performed using the original benchmark parameters, which can reduce the number of parameters for simulation benchmarking. Since the original benchmark parameters have high correlation, the effect of rapid benchmarking can be achieved while ensuring high accuracy.

[0133] In one embodiment, such as Figure 8 As shown, step S106 involves performing an effectiveness analysis on the original benchmark parameters corresponding to the low-speed collision simulation model to obtain the effective benchmark parameters corresponding to the low-speed collision simulation model, including:

[0134] S801: Obtain the simulation result curves corresponding to the original benchmark parameters output by the low-speed collision simulation model. The simulation result curves include the upper boundary curve and the lower boundary curve.

[0135] S802: Obtain the test result curves corresponding to the original benchmark parameters;

[0136] S803: If the test result curve is between the upper boundary curve and the lower boundary curve, then the original benchmark parameters will be determined as the effective benchmark parameters corresponding to the low-speed collision simulation model.

[0137] As an example, in step S801, the simulation benchmarking device can control the low-speed collision simulation model to perform simulation operations and obtain the simulation result curve corresponding to the original benchmarking parameters output by the low-speed collision simulation model. This simulation result curve is a curve output by the low-speed collision simulation model related to a certain original benchmarking parameter. The upper boundary curve is the curve formed by the larger parameter value of the original benchmarking parameter in the simulation result curve; correspondingly, the lower boundary curve is the curve formed by the smaller parameter value of the original benchmarking parameter in the simulation result curve.

[0138] As an example, in step S802, the simulation benchmarking device can acquire low-speed collision test data of the test vehicle under low-speed collision conditions. This low-speed collision test data refers to the test data collected during the test vehicle's operation under low-speed collision conditions. Then, from all the low-speed collision test data, the parameter values ​​of the original benchmarking parameters at different times are fitted or otherwise processed to form the test result curve.

[0139] As an example, in step S803, after obtaining the simulation result curve and the test result curve corresponding to the original benchmarking parameters, the simulation benchmarking device can compare the test result curve with the upper boundary curve and the lower boundary curve to determine whether the test result curve is within the area formed between the upper boundary curve and the lower boundary curve. If the test result curve is within the area formed between the upper boundary curve and the lower boundary curve, the original benchmarking parameter is determined to be a valid benchmarking parameter corresponding to the low-speed collision simulation model; otherwise, the original benchmarking parameter is determined to be a non-valid benchmarking parameter and needs to be removed.

[0140] Generally, due to the error between simulation and experiment, the simulation result curve may not cover the experimental result curve in all regions. If the experimental result curve is between the upper and lower boundary curves, the error between simulation and experiment is considered small, and the original benchmarking parameter can be determined as a valid benchmarking parameter. If the experimental result curve is not between the upper and lower boundary curves, the error between simulation and experiment is considered large, and the original benchmarking parameter needs to be deleted.

[0141] In this embodiment, it is determined whether the original benchmarking parameters can be identified as valid benchmarking parameters based on whether the experimental result curve is between the upper and lower boundary curves of the simulation result curve. This allows for the selection of valid benchmarking parameters from the original benchmarking parameters, enabling subsequent simulation benchmarking operations to be performed using the valid benchmarking parameters. This reduces the number of parameters required for simulation benchmarking. Since the error between the simulation and experiment of the valid benchmarking parameters is small, a rapid benchmarking effect can be achieved while ensuring high accuracy.

[0142] In one embodiment, such as Figure 9 As shown, after step S106, that is, after performing an effectiveness analysis on the original benchmark parameters corresponding to the low-speed collision simulation model and obtaining the effective benchmark parameters corresponding to the low-speed collision simulation model, the simulation benchmarking method for low-speed collision conditions further includes:

[0143] S901: Perform inversion identification on the effective benchmarking parameters corresponding to the low-speed collision simulation model, and obtain the inversion identification results corresponding to the effective benchmarking parameters;

[0144] S902: If the inversion identification result corresponding to the effective benchmarking parameter meets the preset conditions, then the effective benchmarking parameter will be determined as the target benchmarking parameter corresponding to the effective benchmarking parameter.

[0145] S903: Using low-speed collision test data corresponding to low-speed collision conditions, simulate and benchmark the target benchmark parameters corresponding to the low-speed collision simulation model to obtain the target collision simulation model.

[0146] As an example, in step S901, after obtaining the effective benchmarking parameters corresponding to the low-speed collision test model, the simulation benchmarking device can use a surrogate model method to invert and identify the effective benchmarking parameters corresponding to the low-speed collision simulation model, and obtain the inversion identification result corresponding to the effective benchmarking parameters output by the surrogate model method. This inversion identification result can reflect the degree of closeness between the simulation results and the experimental results of the effective benchmarking parameters.

[0147] For example, simulation benchmarking equipment can use the polynomial response surface methodology to invert and identify the simulation and experimental results of the effective benchmarking parameters corresponding to the low-speed collision test model, thereby obtaining the inversion identification results corresponding to the effective benchmarking parameters. The polynomial response surface methodology analyzes sampled points according to certain rules to find the explicit or implicit mapping relationship between the input and output domains, facilitating the simplification of the model within a certain error range. Therefore, the formula for the surface response surface in multidimensional space is: Among them, τ0, τ i τ ij The unknown coefficients, arranged in a certain order, form a column vector τ; i Let be the i-th component of the n-dimensional independent variable x. When the highest term only considers the quadratic term, the number of unknown coefficients N is... τ For N τ = (n+1)(n+2) / 2; If the inversion optimization method is used to obtain the column vector τ, it becomes the key to solving the response surface surrogate model: min:[y k -f(x k ) P ] 2 In the formula, k = 1, 2, ..., m, where m is the number of sample points; y k f(x) represents the output response value at the k-th sample point. k ) P These are the predicted values ​​for the polynomial response surface.

[0148] For example, the Kriging method can be used to invert and identify the simulation results and experimental results corresponding to the effective benchmark parameters of a low-speed collision test model, thereby obtaining the inversion identification results corresponding to the effective benchmark parameters. The Kriging method is similar to a semi-parametric interpolation technique, using relevant functions to simulate a high-dimensional space. It can effectively avoid noise information problems and does not require the establishment of a specific mathematical model. The Kriging model formula is: f(x) K= g(x) + q(x), f(x) K The surrogate model solver function is defined as follows: g(x) is a deterministic function, often expressed as a polynomial, and serves as a global approximation; q(x) represents fluctuations, i.e., a randomly distributed error function, and serves as a local approximation of the local bias. The covariance matrix of q(x) is as follows: E[q(x)] i ),q(x)]=σ 2 R k (c,x,x j ), i,j=1,2,…,n, where σ is the variance; R k (c,x,x j ) is a related function with c as a parameter, where Common kernel functions φ can be Gaussian or exponential functions, based on the sample points x. i After linearly weighting the output value with the weight vector ω, the method of superimposing the difference is used to predict the output value of the measurement point. Then the prediction model... The resulting prediction variance is: Based on the model's technical requirement of minimizing the predicted variance, the problem of solving the full coefficient ω needs to be transformed into solving an extremum problem. Using the Lagrange multiplier method, the weight function can be substituted into the fitting function to obtain the solution. In the formula, the coefficient β * γ * The weight function coefficients are obtained during the process of finding the optimal solution weight vector. g(x) is determined through regression analysis, just like the polynomial response surface. Ultimately, the computational work of this method mainly focuses on solving the vector φ(x).

[0149] For example, simulation benchmarking equipment can use the radial basis function (RBF) method to invert and identify the simulation and experimental results of the effective benchmarking parameters corresponding to the low-speed collision test model, thereby obtaining the inversion identification results corresponding to the effective benchmarking parameters. The RBF method can be used in ultra-high-dimensional parameter spaces and can mine hidden information in large-scale data domains, requiring less training time than traditional artificial neural network techniques. Its expression is: Where N is the number of sampling points, m is the number of effective calibration parameters, and R m It is the design space, x is the parameter vector, x j It is the j-th sampling point, and the response value is f(x). j ) R It is x j δ was calculated for the design point. j It is a factor of the weight coefficient matrix δ, and can be calculated using the following formula:

[0150]

[0151] in, These are radial basis functions, which have various representations, including: (1) linear: (2) Cubic: (3) Thin plate spline: (4) Multifaceted functions: (5) Inverse polyhedral function: (6) Gaussian function

[0152] Using a Gaussian function as the radial basis kernel: Where, χ j It is the width of the j-th basis function. In this example, it uses... Calculations show that d max It is the maximum Euclidean distance between sampling points.

[0153] As an example, in step S902, after obtaining the inversion identification result corresponding to each valid benchmarking parameter, the simulation benchmarking device can determine whether the inversion identification result meets the preset conditions. If the inversion identification result meets the preset conditions, the valid benchmarking parameter is determined as the target benchmarking parameter. In this example, a surrogate model method is used to perform inversion identification on the valid benchmarking parameters. The obtained inversion identification result can represent the degree of closeness between the simulation result and the experimental result of the valid benchmarking parameter. If the degree of closeness is greater than a preset threshold, the inversion identification result is considered to meet the preset conditions, and the valid benchmarking parameter can be determined as the target benchmarking parameter. Conversely, if the degree of closeness is not greater than the preset threshold, the inversion identification result is considered not to meet the preset conditions, and the valid benchmarking parameter needs to be deleted.

[0154] As an example, in step S903, after determining the target benchmarking parameters, the simulation benchmarking device can acquire low-speed collision test data corresponding to the low-speed collision condition, specifically low-speed collision test data related to the target benchmarking parameters. Then, using the low-speed collision test data related to the target benchmarking parameters, the simulation benchmarking device performs simulation benchmarking on the target benchmarking parameters corresponding to the low-speed collision simulation model to obtain the target collision simulation model. The simulation benchmarking process is similar to step S107 and will not be elaborated here. In this example, the simulation benchmarking device can compare the simulation results corresponding to the two target benchmarking parameters—deformation of key components or result curves—with the test results of the low-speed collision test data. If either the deformation of key components or the result curve does not meet its corresponding preset condition, the simulation benchmarking of the low-speed collision simulation model is considered incomplete, and the model parameters of the low-speed collision simulation model need to be adjusted. If both the deformation of key components or the result curves meet their corresponding preset conditions, the simulation benchmarking of the low-speed collision simulation model is considered complete, and the target collision simulation model is obtained.

[0155] In this embodiment, surrogate model methods, including but not limited to polynomial response surface methodology, Kriging method, and radial basis function method, can be used to invert and identify the simulation results and experimental results of the effective benchmark parameters corresponding to the low-speed collision test model. Through iterative calculation, the difference between the simulation results output by the low-speed collision simulation model and the experimental results is determined. If the difference reaches the expected value, the low-speed collision simulation model is considered to meet the preset conditions, and the simulation benchmarking of the low-speed collision simulation model is considered complete, thus obtaining the target collision simulation model. If the difference consistently fails to reach the expected value, it indicates a problem with the collision test simulation model or a defect in the response surface methodology used for the effective benchmark parameters, requiring re-model testing, parameter selection, or response surface selection. Understandably, selecting target benchmark parameters based on the inversion identification results of effective benchmark parameters, and then performing simulation benchmarking based on these target benchmark parameters, significantly reduces the risk of poor simulation benchmarking results due to insufficient experience, time, etc., lowers the threshold for simulation benchmarking under low-speed collision conditions, improves the efficiency of simulation benchmarking, and saves the time and manpower costs required for the simulation benchmarking process under low-speed collision conditions.

[0156] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0157] In one embodiment, a simulation benchmarking device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the simulation benchmarking method for low-speed collision conditions described in the above embodiments, for example... Figure 1 As shown in S101-S107, or Figures 2 to 9 As shown in the figure, to avoid repetition, it will not be repeated here.

[0158] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When executed by a processor, the computer program implements the simulation benchmarking method for low-speed collision conditions described in the above embodiment. The simulation benchmarking method for low-speed collision conditions, for example... Figure 1 As shown in S101-S107, or Figures 2 to 9 As shown in the figure, to avoid repetition, it will not be repeated here.

[0159] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and RAMbus dynamic RAM (RDRAM), etc.

[0160] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0161] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A simulation benchmarking method for low-speed collision conditions, characterized in that, include: Using high-speed collision test data corresponding to high-speed collision conditions, the whole vehicle collision simulation model is simulated and benchmarked to obtain a high-speed collision simulation model. The high-speed collision simulation model is processed using the target key information corresponding to the low-speed collision condition to obtain the low-speed collision simulation model; the target key information corresponding to the low-speed collision condition is component information that differs from that of the high-speed collision condition. Key component parameters are extracted from the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model. The target key component parameters include: the elastic modulus of the vehicle's front bumper foam, yield strength, anti-collision beam thickness, energy-absorbing box thickness, and front bumper thickness; Sensitivity analysis is performed on the target key component parameters corresponding to the low-speed collision simulation model to determine the effective key component parameters corresponding to the low-speed collision simulation model. Correlation analysis was performed on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the original benchmark parameters corresponding to the low-speed collision simulation model. Obtain the simulation result curve corresponding to the original benchmark parameters output by the low-speed collision simulation model, the simulation result curve including the upper boundary curve and the lower boundary curve; obtain the test result curve corresponding to the original benchmark parameters; if the test result curve is between the upper boundary curve and the lower boundary curve, then the original benchmark parameters are determined as the effective benchmark parameters corresponding to the low-speed collision simulation model. Using low-speed collision test data corresponding to low-speed collision conditions, the effective benchmark parameters corresponding to the low-speed collision simulation model are simulated and benchmarked to obtain the target collision simulation model.

2. The simulation benchmarking method for low-speed collision conditions as described in claim 1, characterized in that, The process involves using high-speed collision test data corresponding to high-speed collision conditions to perform simulation benchmarking on the whole vehicle collision simulation model, thereby obtaining a high-speed collision simulation model, including: Acquire high-speed collision test data under high-speed collision conditions; Using the high-speed collision test data corresponding to the high-speed collision condition, the vehicle collision simulation model is simulated and benchmarked to obtain the high-speed simulation accuracy corresponding to the high-speed collision condition; If the high-speed simulation accuracy is less than the first accuracy threshold, the high-speed collision test data is used to adjust the whole vehicle collision simulation model. If the high-speed simulation accuracy is not less than the first accuracy threshold, then the whole vehicle collision simulation model is determined as the high-speed collision simulation model.

3. The simulation benchmarking method for low-speed collision conditions as described in claim 2, characterized in that, Using the high-speed collision test data corresponding to the high-speed collision conditions, the vehicle collision simulation model is simulated and benchmarked to obtain the high-speed simulation accuracy corresponding to the high-speed collision conditions, including: Obtain the simulation feature curves output by the whole vehicle collision simulation model and the test feature curves corresponding to the high-speed collision test data; The simulation characteristic curve and the experimental characteristic curve are cross-correlation evaluated to obtain cross-correlation evaluation scores; The simulated characteristic curve and the experimental characteristic curve are used to evaluate the corridor and obtain the corridor evaluation score; Based on the cross-correlation evaluation score and the corridor evaluation score, the high-speed simulation accuracy corresponding to the high-speed collision condition is obtained.

4. The simulation benchmarking method for low-speed collision conditions as described in claim 1, characterized in that, The step of extracting key component parameters from the low-speed collision simulation model to obtain the target key component parameters corresponding to the low-speed collision simulation model includes: Key component parameters are extracted from the low-speed collision simulation model to obtain the original key component parameters corresponding to the low-speed collision simulation model. The original key component parameters corresponding to the low-speed collision simulation model are optimized and sampled to obtain the target key component parameters corresponding to the low-speed collision simulation model.

5. The simulation benchmarking method for low-speed collision conditions as described in claim 1, characterized in that, The sensitivity analysis of the target critical component parameters corresponding to the low-speed collision simulation model, to determine the effective critical component parameters corresponding to the low-speed collision simulation model, includes: Sensitivity analysis is performed on the target key component parameters corresponding to the low-speed collision simulation model to obtain the target sensitivity corresponding to the target key component parameters; Target key component parameters whose target sensitivity is greater than a preset sensitivity are determined as valid key component parameters.

6. The simulation benchmarking method for low-speed collision conditions as described in claim 1, characterized in that, The correlation analysis of the effective key component parameters corresponding to the low-speed collision simulation model, to obtain the original benchmark parameters corresponding to the low-speed collision simulation model, includes: Correlation analysis was performed on the effective key component parameters corresponding to the low-speed collision simulation model to obtain the parameter correlations corresponding to the effective key component parameters. The correlation between the parameters corresponding to the effective key component parameters is sorted to obtain the correlation sorting result; The top N effective key component parameters are selected from the correlation ranking results, and these top N effective key component parameters are determined as the original benchmark parameters corresponding to the low-speed collision simulation model, where N≥1.

7. The simulation benchmarking method for low-speed collision conditions as described in claim 1, characterized in that, After determining the original benchmarking parameters as the effective benchmarking parameters corresponding to the low-speed collision simulation model, the simulation benchmarking method for the low-speed collision condition further includes: The effective benchmarking parameters corresponding to the low-speed collision simulation model are inverted and identified to obtain the inversion and identification results corresponding to the effective benchmarking parameters. If the inversion identification result corresponding to the effective benchmarking parameter meets the preset conditions, then the effective benchmarking parameter is determined as the target benchmarking parameter corresponding to the effective benchmarking parameter; Using low-speed collision test data corresponding to low-speed collision conditions, the target benchmarking parameters corresponding to the low-speed collision simulation model are simulated and benchmarked to obtain the target collision simulation model.

8. A simulation benchmarking device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the simulation benchmarking method for low-speed collision conditions as described in any one of claims 1 to 7.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the simulation benchmarking method for low-speed collision conditions as described in any one of claims 1 to 7.

Citation Information

Patent Citations

  • Method, system and related device for testing computer program performance

    CN114201389A

  • Equivalent finite element simulation method for high and low temperature pendulum bobs of front and rear bumpers of automobile

    CN114239331A