Column collision sliding table test boundary parameter optimization method

Abstract

The invention relates to the technical field of automobile safety performance development, in particular to a column collision sliding table test boundary parameter optimization method. Comprising the following steps: extracting related parts from a vehicle column collision passenger protection finite element collision model to obtain a simplified model; establishing a simulation column collision sliding table test finite element model; performing parameterization setting on key parameters in the model, and setting initial values and value ranges of the parameters; establishing a multi-objective optimization function and setting optimization constraint conditions; generating a DOE matrix of a plurality of groups of samples in a parameter value range, wherein each group of samples comprises a combination of all parameters; constructing a reduced order prediction model for predicting rib compression response under different parameter combinations; and with multi-objective function minimization as an objective, obtaining an optimal parameter combination meeting constraint conditions. According to the technical scheme, the calculation complexity can be reduced, the input boundary condition of the side column collision sliding table invasion equipment can be quickly determined, and the passenger injury condition under the whole vehicle column collision working condition can be highly reproduced.

Description

Technical Field

[0001] This invention relates to the field of automotive safety performance development technology, specifically to a method for optimizing boundary parameters in a pole impact slide test. Background Technology

[0002] In science and engineering, many problems are difficult to solve due to their numerical complexity or specific constraints. Taking high-dimensional models as an example, as the dimensionality increases, the computational complexity rises exponentially, putting enormous pressure on computing resources and making the solution process extremely time-consuming and difficult. Meanwhile, many practical problems require multi-parameter co-optimization, which not only demands the establishment of accurate optimization matrices but also poses stringent challenges to computational efficiency. How to quickly and efficiently complete multi-parameter optimization while ensuring computational accuracy has become a pressing problem to be solved.

[0003] In the field of automotive safety research and development, side pole impact, as an extremely stringent side collision test in mainstream domestic and international safety standards (such as Euro NCAP and C-NCAP), has always been a key area of ​​focus for automotive OEMs to invest in research and development. The complexity of the side pole impact test is reflected in many aspects. It is closely related to many factors such as the amount of intrusion, the intrusion speed, and the deformation of the inner door panel. These factors are intertwined and influence each other, jointly determining the injury situation of the occupants during the collision.

[0004] Currently, pole impact trolley testing, as an emerging testing method, shows great potential in optimizing occupant protection during pole impact scenarios. This method can effectively reduce costs and significantly improve R&D efficiency, providing a more economical and efficient approach for automotive safety R&D. However, existing side pole impact intrusion devices have significant shortcomings in key parameters. Specifically, key information such as input parameters and impact cylinder offset distance are still unclear, resulting in unsatisfactory simulation effects of dummy injuries when using pole impact trolley testing to reproduce full-vehicle pole impact scenarios. Because it cannot accurately simulate the injuries suffered by dummies in actual collisions, the test results are difficult to effectively guide the optimization and improvement of side pole impact trolley testing, failing to meet the needs of automotive safety R&D for accurate simulation and effective optimization.

[0005] In vehicle testing, to more realistically simulate side impact tests, the pole impact test bench needs to perform two important tasks: firstly, to simulate vehicle acceleration to reflect the impact of inertial loads on occupants; and secondly, to simulate door acceleration to reflect the injury caused by door intrusion. However, in actual pole impact tests, the side passenger compartment is often simplified, resulting in a significant difference between the constructed bench test environment and the full vehicle test environment. This is especially true regarding motion states and the mass of the test specimens, where consistency is difficult to maintain. This difference leads to a large deviation between the pole impact test results and the full vehicle test results. Therefore, ensuring that the final test results and dummy injury outcomes of the pole impact test closely match those of the full vehicle test has become a current challenge in pole impact tests. Summary of the Invention

[0006] The purpose of this invention is to propose a method for optimizing boundary parameters in a pole impact slide test. This technical solution can reduce computational complexity and quickly determine the input boundary conditions of the side pole impact slide intrusion device, thus highly reproducing the occupant injury situation under the pole impact condition of the whole vehicle.

[0007] To achieve the above objectives, this invention proposes a method for optimizing boundary parameters in a column-impact sliding stage test, comprising: Relevant components were extracted from the finite element collision model of the whole vehicle pole impact occupant protection to obtain a simplified model; A finite element model for simulating column-to-slide table test was established. The model includes the slide table, upper / middle / lower intrusion cylinders, and dummy. The boundary conditions and initial conditions of the model were set, including the slide table loading waveform, the intrusion behavior of the intrusion cylinders, and the initial position and attitude of the dummy. The key parameters in the model are parameterized, and the initial values ​​and ranges of the parameters are set. A multi-objective optimization function is established to represent the difference between the predicted values ​​of the damage parameters of the upper, middle, and lower ribs of the dummy chest output from the pole impact slide test and the experimental values ​​of the corresponding parameters in the whole vehicle pole impact test data, and optimization constraints are set. Several sets of DOE matrices are generated within the parameter value range. Each set of samples contains a combination of all key parameters. The parameters of each set of samples are substituted into the finite element model of the column-slide table test for calculation and analysis to obtain the response under each set of parameters, forming a DOE sample set. Based on the DOE sample set, a reduced-order prediction model was constructed to predict the rib compression response under different parameter combinations. With the goal of minimizing a multi-objective function, optimization algorithms are used to iteratively optimize the reduced-order model to obtain the optimal parameter combination that satisfies the constraints. The optimal parameter combination predicted by the reduced-order prediction model was substituted into the finite element model of the column-slide test to perform calculations and analysis, thereby verifying the accuracy of the prediction model.

[0008] Beneficial effects of the basic scheme: This method constructs a simplified model by extracting the core components of the whole vehicle pillar impact finite element model. At the same time, it uses the separation of variables method to decompose the high-dimensional optimization problem into multiple low-dimensional sub-problems, and introduces parameters, boundary conditions, etc. as additional coordinates into the model, which greatly reduces the mesh size and number of iterations of finite element calculation, and significantly reduces the computational complexity. On this basis, a reduced-order prediction model is constructed to replace the traditional time-consuming full model simulation, realize the rapid calculation of rib compression response under different parameter combinations, effectively shorten the parameter optimization cycle, and save the cost of high-performance computing resources.

[0009] This method focuses on the difference between the predicted values ​​of rib injury parameters at different locations on the dummy's chest and the values ​​from a full-vehicle pole impact test. It generates a sample set through Design of Experiments (DOE) and iteratively solves for the optimal parameter combination using an optimization algorithm. This allows for the accurate determination of key boundary parameters such as the sliding platform loading waveform, the intrusion behavior of the intruding cylinder, and the dummy's initial posture. The optimized sliding platform test model can highly reproduce the injury characteristics of the occupant's chest ribs under a full-vehicle pole impact condition, solving the problem of poor dummy injury simulation results caused by unclear boundary parameters in traditional sliding platform tests. This provides reliable test data support for the verification and optimization of occupant protection schemes in side pole impact tests of new energy vehicles.

[0010] To address the drawbacks of traditional vehicle pole impact physical tests, such as high cost, long cycle, and poor repeatability, this method optimizes the boundary parameters of the slide test through precise numerical simulation. This enables multiple rounds of parameter iteration and scheme verification in a virtual environment, reducing the number of unnecessary physical tests. At the same time, the optimized slide test can stably reproduce the target working condition, improving the consistency and reliability of test results and accelerating the research and development process of safety components such as the body structure and restraint system of new energy vehicles.

[0011] This method employs parametric modeling and multi-objective optimization design, which can flexibly adjust the parameter value range and optimization objective function for different vehicle models and different collision conditions. By using linearization and interpolation strategies, it effectively handles nonlinear problems, and the generated general solution can be quickly adapted to the post-processing and optimization needs of different parameter combinations. It is not only applicable to side pole impact slide test, but can also be extended to the boundary parameter optimization of slide tests for other collision forms, and has broad engineering application value.

[0012] As a feasible and preferred approach, relevant components are extracted from the finite element collision model of the whole vehicle pole impact occupant protection to obtain a simplified model, including the following: Related components include WorldSID 50 thSide-impact dummy, seat, seat belt, SAB airbag, and door trim assembly; detailed mesh of key areas such as the dummy's chest and ribs is preserved, and non-critical areas are geometrically cleaned and replaced with equivalent mass.

[0013] As a feasible preferred option, the key parameters in the model include the scaling factor of the slide loading waveform, the scaling factor of the upper / middle / lower intrusion cylinder waveform, and the Y-axis offset distance of the impact cylinder.

[0014] As a feasible and preferred option, the multi-objective optimization function formula is as follows:

[0015]

[0016]

[0017] In the formula, , , The root mean square errors of the upper ribs, middle ribs, and lower ribs in the dummy chest are represented by the root mean square errors in the simulation data and experimental data, respectively. , , These represent the compression amounts of the upper, middle, and lower ribs of the dummy at each time point ti measured in the experiment. , , Representing each time t in the simulation calculation i The corresponding compression amounts of the upper, middle, and lower ribs of the dummy's chest; n represents the number of data points; The optimization constraints are as follows:

[0018] The optimization objective is to ensure that the root mean square error of the predicted values ​​of the multi-objective optimized slide test and the vehicle test data is no greater than 5.

[0019] As a feasible and preferred approach, a reduced-order prediction model is constructed using the inherent generalized decomposition model reduction method, including the following: Create snapshots of each selected time and each result from the simulation results, and construct a snapshot matrix:

[0020] Calculate the basis and modes using the snapshot matrix, and then truncate:

[0021] Construct a reduced-order model using a simplified basis: .

[0022] As a feasible and preferred option, the prediction model formula is as follows:

[0023] in, It is a physical field variable. These are spatial coordinates. It is time. It is a parameter vector. Representation of spatial patterns, Indicates time mode; Indicates parameter mode; Indicates the number of patterns.

[0024] As a feasible and preferred approach, with the goal of minimizing a multi-objective function, an optimization algorithm is used to iteratively optimize the reduced-order model, including the following: Gradient descent or genetic algorithm is used for parameter optimization, and an alternating direction strategy is used to iteratively solve the unknown function during the optimization process.

[0025] As a feasible preferred option, it also includes substituting the optimized parameters into the finite element model of the column-slide table for calculation and verification, and comparing whether the root mean square error of the optimized model response curve and the test curve meets the target requirements.

[0026] As a feasible and preferred option, after substituting the optimal parameter combination into the finite element model of the slide table, the compression amount-time curves of the upper, middle and lower ribs of the dummy's chest are output. Calculate the root mean square error of the three curves and the corresponding curve of the whole vehicle pole impact test; Verify whether the constraints are met:

[0027] If the requirements are not met, optimize again until all errors meet the requirements. Attached Figure Description

[0028] Figure 1 This is a logical diagram of a method for optimizing boundary parameters in a column-to-slide table test.

[0029] Figure 2 This is a schematic diagram of a finite element model.

[0030] Figure 3 Illustration of key parameters of the model Figure 1 .

[0031] Figure 4 Illustration of key parameters of the model Figure 2 .

[0032] Figure 5 This is a schematic diagram of Latin hypercube sampling.

[0033] Figure 6 This is a diagram showing the comparison between predicted values ​​and experimentally measured values.

[0034] Figure 7 This is a bar chart of root mean square error.

[0035] Figure 8 This is a diagram showing the comparison between the predicted values ​​before and after optimization and the experimentally measured values. Detailed Implementation

[0036] To make the technical solution and advantages of this application clearer, the technical solution of the present invention will be further described in detail below with reference to the accompanying drawings. It is understood that the specific embodiments described herein are only some embodiments of the present invention, and are only used to explain this application, not to limit it. It should be noted that the technical features or combinations of technical features described in the following embodiments should not be considered isolated; they can be combined with each other to achieve better technical effects. The same reference numerals appearing in the accompanying drawings of the following embodiments represent the same features or components, and can be applied to different embodiments.

[0037] Furthermore, unless otherwise defined, the technical or scientific terms used in this invention description shall have the ordinary meaning understood by one of ordinary skill in the art to which this invention pertains.

[0038] The present invention will now be described in further detail with reference to the accompanying drawings.

[0039] Reference Figure 1 This disclosure provides a method for optimizing boundary parameters in a column-to-slide test, comprising the following steps.

[0040] Step S100: Model simplification processing, extracting relevant components from the whole vehicle pole impact occupant protection finite element collision model, including WorldSID 50. th The side-impact crash test dummy, seats, seat belts, SAB airbags, and door interior assemblies were simplified in their models. Key structural features were retained while non-critical details were removed to reduce computational load.

[0041] Specifically, preprocessing software such as HyperMesh is used for geometry cleanup and mesh generation. Detailed meshes of critical parts such as the dummy's chest and ribs are preserved, while meshes of non-critical parts are simplified. Components such as the seat and seatbelt are replaced with equivalent mass replacements to reduce the number of computational units.

[0042] Step S200: Establish a finite element model for simulating the column-to-slide test. Based on the simplified model, establish a finite element model for simulating the column-to-slide test. The model should include key components such as the slide, intrusion cylinder, and dummy, and set reasonable boundary and initial conditions.

[0043] Specifically, the sliding table model is used to simulate the collision motion field of the entire vehicle pillar, and the sliding table loading waveform is set. The intrusion cylinder model includes an upper intrusion cylinder, a middle intrusion cylinder, and a lower intrusion cylinder to simulate the intrusion behavior of the vehicle door. The dummy model adopts WorldSID 50. th The side-impact dummy is set with its initial position and attitude. Boundary conditions are set, including the displacement and velocity of the slide and intrusion cylinder. Initial conditions are set, including the initial velocities and accelerations of the dummy, slide, and intrusion cylinder.

[0044] Step S300, parameter settings, refer to Figure 3 and Figure 4 (Table 1 explains the parameters in the figure.) The key parameters in the model are parameterized, including the scaling factor of the sliding table loading waveform, the scaling factor of the intrusion cylinder waveform, and the offset distance of the impact cylinder.

[0045] Table 1

[0046] Specifically, the scaling factor of the slide table loading waveform controls the amplitude of the slide table acceleration waveform, with an initial value set to 1, and a range of 0.9-1.0 in this embodiment. The scaling factor of the intrusion cylinder waveform controls the intrusion speed and amplitude of the upper, middle, and lower intrusion cylinders respectively, with an initial value set to 1 for each, and a range of 0.8-1.0 in this embodiment. The offset distance of the impact cylinder controls the initial position of the impact cylinder, with an initial value set to 0, and a range of 0-30mm in this embodiment.

[0047] Step S400: Establish a multi-objective optimization function to represent the difference between the predicted values ​​of the damage parameters of the upper, middle, and lower ribs of the dummy chest output from the pole impact test and the experimental values ​​of the corresponding parameters in the whole vehicle pole impact test data.

[0048] The formula for the multi-objective optimization function is as follows:

[0049]

[0050]

[0051] In the formula, , , The root mean square errors of the upper ribs, middle ribs, and lower ribs in the dummy chest are represented by the root mean square errors in the simulation data and experimental data, respectively. l、 , These represent the compression amounts of the upper, middle, and lower ribs of the dummy at each time point ti measured in the experiment. , , _i_i represents the compression of the upper, middle, and lower ribs of the dummy at each time point _ti in the simulation calculation; _n_i represents the number of data points within the range of 0-70ms.

[0052] The optimization constraints are as follows: ; The optimization objective is to ensure that the root mean square error (RMSE) of the multi-objective optimized slide test prediction values ​​and vehicle test data values ​​is no greater than 5.

[0053] Step S500: Based on the results of the parameterization settings, determine the value range of each optimization parameter for subsequent DOE matrix creation and sampling.

[0054] The parameter value range is shown in Table 2.

[0055] Table 2

[0056] Step S600: Create the DOE matrix.

[0057] After the objective function, optimization constraints, and parameter value ranges are determined, sampling is performed, referring to... Figure 5 In this embodiment, Latin hypercube sampling is preferred. Latin hypercube sampling is a sampling method commonly used in high-dimensional numerical analysis and computational simulation. It is a hierarchical random sampling technique that aims to improve sampling efficiency and reduce sample redundancy by effectively filling high-dimensional space, thereby providing more accurate estimates within a given computational budget.

[0058] Specifically, the Latin hypercube sampling function in tools such as MATLAB or Python is used to generate DOE matrix samples. The parameters of each sample group are then used to modify the corresponding parameters in the column-to-slide table parameterized model to obtain the sample set for establishing the prediction model. In this embodiment, the number of samples is set to 20 groups, with each group containing values ​​for 5 optimized parameters. See Table 3.

[0059] Table 3

[0060] Step S700: Based on the sample set of the DOE matrix, a prediction model is established using a multi-objective optimization solution and an advanced model order reduction method. Model order reduction (MOR) technology is used to reduce the computational complexity of the mathematical model. In this embodiment, Proper Generalized Decomposition (PGD) is used to reduce the order of the model, improve optimization efficiency, and generate a low-dimensional prediction model.

[0061] The prediction model formula is:

[0062] in, These are physical field variables (such as displacement, velocity, etc.). These are spatial coordinates. It is time. It is a parameter vector. Representation of spatial patterns, Indicates time mode; Indicates parameter mode; Indicates the number of patterns.

[0063] The solution is obtained by iteratively adding new terms through PGD until convergence.

[0064] Specifically, snapshots of each selected time and each result (velocity, displacement, etc.) are created from the simulation results, and a snapshot matrix is ​​constructed:

[0065] Calculate the basis and modes using the snapshot matrix, and then truncate:

[0066] Construct a reduced-order model using a simplified basis:

[0067] Reference Figure 6 and Figure 7 The red (experimental measurement) and green (model prediction) curves show a high degree of overlap, indicating that the model can accurately capture the displacement trends of the upper, middle, and lower chest positions during rapid dynamic processes. The predicted values ​​show a high degree of agreement with the experimental measurements, and the RMSE index also verifies the reliability of the model.

[0068] Step S800: Real-time optimization is performed based on the prediction model. The objective function is minimized by adjusting the values ​​of the optimization parameters, while satisfying the optimization constraints.

[0069] The optimization process employs an alternating direction strategy to iteratively solve the unknown function.

[0070] Specifically, optimization algorithms (such as gradient descent, genetic algorithms, etc.) are used for parameter optimization. During the optimization process, the objective function value is calculated for the current parameters in each iteration, and the parameter values ​​are updated according to the optimization algorithm. Optimization stops when the objective function value satisfies the optimization constraints or the maximum number of iterations is reached.

[0071] Optimized parameter examples: 0.99 (large slide plate waveform scaling factor), 0.85 (upper intrusion cylinder waveform scaling factor), 0.85 (middle intrusion cylinder waveform scaling factor), 0.99 (lower intrusion cylinder waveform scaling factor), 0.63 (impact cylinder offset distance).

[0072] Step S900: Substitute the optimized parameters into the finite element model of the column-slide table for calculation and verification, and compare whether the root mean square error of the predicted curve and the experimental curve meets the target requirements.

[0073] Specifically, finite element analysis software such as LS-DYNA is used for simulation calculations.

[0074] Damage parameters of the upper, middle, and lower ribs of the dummy were extracted from the simulation calculations and compared with the data from the whole vehicle pole impact test.

[0075] Calculate the root mean square error and verify whether the constraints are met:

[0076] The root mean square error of both the predicted curve and the experimental curve meets the target requirements, proving that the optimization method is effective.

[0077] Reference Figure 8 As shown in Table 4, after optimizing the key parameters of the column-impact sliding table model, the simulated damage curves of the column-impact sliding table show a higher degree of fit with the experimental results, and the simulated damage values ​​are closer to the experimental values. These parameters can be used to guide the next stage of column-impact sliding table testing in this project.

[0078] Table 4

[0079] The above content is merely an embodiment of the present invention. Commonly known structures and characteristics of the solutions are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can improve and implement this solution based on the guidance provided in this application and their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for optimizing boundary parameters in a column-to-slide table test, characterized in that, include: Relevant components were extracted from the finite element collision model of the whole vehicle pole impact occupant protection to obtain a simplified model; A finite element model for simulating column-to-slide table test was established. The model includes the slide table, upper / middle / lower intrusion cylinders, and dummy. The boundary conditions and initial conditions of the model were set, including the slide table loading waveform, the intrusion behavior of the intrusion cylinders, and the initial position and attitude of the dummy. The key parameters in the model are parameterized, and the initial values ​​and ranges of the parameters are set. A multi-objective optimization function is established to represent the difference between the predicted values ​​of the damage parameters of the upper, middle, and lower ribs of the dummy chest output from the pole impact slide test and the experimental values ​​of the corresponding parameters in the whole vehicle pole impact test data, and optimization constraints are set. Within the parameter value range, generate several sets of DOE matrices, each set containing a combination of all key parameters; substitute the parameters of each set of samples into the finite element model of the column-slide table test for calculation and analysis, obtain the response under each set of parameters, and form a DOE sample set; Based on the DOE sample set, a reduced-order prediction model was constructed to predict the rib compression response under different parameter combinations. With the goal of minimizing a multi-objective function, optimization algorithms are used to iteratively optimize the reduced-order model to obtain the optimal parameter combination that satisfies the constraints. The optimal parameter combination predicted by the reduced-order prediction model was substituted into the finite element model of the column-slide test to perform calculations and analysis, thereby verifying the accuracy of the prediction model.

2. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, Relevant components were extracted from the finite element collision model of the whole vehicle pole impact occupant protection system to obtain a simplified model, including the following: Related components include WorldSID 50 th Side-impact dummy, seat, seat belt, SAB airbag, and door trim assembly; detailed mesh of key areas such as the dummy's chest and ribs is preserved, and non-critical areas are geometrically cleaned and replaced with equivalent mass.

3. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, Key parameters in the model include the scaling factor of the slide loading waveform, the scaling factor of the upper / middle / lower intrusion cylinder waveform, and the Y-axis offset distance of the impact cylinder.

4. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, The formula for the multi-objective optimization function is as follows: In the formula, , , The root mean square errors of the upper ribs, middle ribs, and lower ribs in the dummy chest are represented by the root mean square errors in the simulation data and experimental data, respectively. l、 , These represent the compression amounts of the upper, middle, and lower ribs of the dummy at each time point ti measured in the experiment. , , Representing each time t in the simulation calculation i The corresponding compression amounts of the upper, middle, and lower ribs of the dummy's chest; n represents the number of data points; The optimization constraints are as follows: The optimization objective is to ensure that the root mean square error of the predicted values ​​of the multi-objective optimized slide test and the vehicle test data is no greater than 5.

5. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, A reduced-order prediction model is constructed using the inherent generalized decomposition model reduction method, including the following: Create snapshots of each selected time and each result from the simulation results, and construct a snapshot matrix: Calculate the basis and modes using the snapshot matrix, and then truncate: Construct a reduced-order model using simplified bases: 。 6. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, The prediction model formula is: in, It is a physical field variable. These are spatial coordinates. It is time. It is a parameter vector. Representation of spatial patterns, Indicates time mode; Indicates parameter mode; Indicates the number of patterns.

7. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, With the goal of minimizing a multi-objective function, optimization algorithms are used to iteratively optimize the reduced-order model, including the following: Gradient descent or genetic algorithms are used to iteratively optimize the model parameters.

8. The method for optimizing boundary parameters in a column-to-slide test according to claim 1, characterized in that, It also includes substituting the optimized parameters into the finite element model of the column-slide table for calculation and verification, and comparing whether the root mean square error of the optimized model response curve and the experimental curve meets the target requirements.

9. The method for optimizing boundary parameters in a column-to-slide test according to claim 8, characterized in that, After substituting the optimal parameter combination into the finite element model of the slide table, the compression-time curves of the upper, middle and lower ribs of the dummy's chest are output. Calculate the root mean square error of the three curves and the corresponding curve of the whole vehicle pole impact test; Verify whether the constraints are met: If the requirements are not met, optimize again until all errors meet the requirements.

Images