Counterwall rigid channel selection method based on curve characteristic inflection points

Abstract

The invention relates to the technical field of automobile passive safety testing, discloses a method for selecting a barrier avoidance rigid channel based on curve characteristic inflection points, and aims to solve the problems of inaccurate result and tedious operation caused by dependence on manual subjective judgment in an existing method. The method comprises the following steps: preprocessing and normalizing an original force-displacement curve of a counterguard dynamic experiment, and filtering to generate a baseline; calculating the offset of the curve and the baseline, and superposing the weight to generate and filter the upper and lower envelope lines; determining a feature inflection point based on the baseline second derivative extreme point, segmenting an envelope line according to the inflection point, and fitting a straight line by a least square method; and introducing a linear offset parameter fine-tuning fitting line segment, correcting a calibration point force value, and then connecting to form a rigid channel. The method realizes channel selection automation, reduces subjective errors, improves channel precision and efficiency, and is suitable for a barrier wall rigidity evaluation scene in a vehicle collision test.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automobile passive safety testing, and particularly relates to a selection method of a barrier rigid channel based on a curve feature inflection point. BACKGROUND

[0002] The barrier rigid channel is a characteristic boundary for objectively describing the rigidity change range of the barrier in the collision process in vehicle collision testing, and is usually bounded by the upper and lower envelope lines of the force-displacement curve (F-S curve), forming a continuous channel area. The core of barrier rigid channel selection is to convert the scattered force-displacement curves into a channel with clear boundaries and quantifiable, so as to define the rigidity performance range of the barrier in the collision process, and to provide reliable support for the consistency of barrier performance in the vehicle passive safety test scene.

[0003] At present, the mainstream barrier rigid channel selection method in the industry highly depends on manual experience, and a large amount of data needs to be collected through barrier dynamic experiments. The barrier dynamic experiment is a dynamic process simulating the collision between a vehicle and a barrier (such as a rigid wall, a honeycomb aluminum barrier, a T-shaped tooling, etc.). The force, displacement, time and other parameters in the collision process are collected through sensors. The core is to reproduce different collision conditions (such as different speeds, angles, barrier types) to obtain the dynamic response data of the barrier in the collision process. After pre-processing the barrier dynamic experiment data, a plurality of force-displacement curves are generated, and then the engineers manually determine the barrier rigid channel by analyzing the curve features based on subjective judgment.

[0004] However, the traditional method based on subjective observation and analysis has significant limitations in the overall process, and cannot obtain accurate results.

[0005] In the process of defining the channel, multiple experimental curves need to be drawn at the same time, and then the drawing is made according to subjective judgment and gradually modified, which is very cumbersome. When the experimental curve data increases, multiple curves appearing on the same frame will interfere with subjective judgment. Originally, with the increase of experimental data, the accuracy of channel calibration should be improved, but in reality, adding curves multiple times will increase the difficulty of redrawing the drawing and affect human judgment, thus leading to inaccurate channel definition and strong subjectivity. The main reasons for the above problems and defects are as follows: (1) Engineers mainly rely on visual observation of the curve direction and distribution density to define the channel range through first impression. It is difficult to objectively quantify the curve similarity, resulting in different engineers drawing different conclusions under the same working conditions, and the results cannot be verified by mathematical indicators. (2) Each time a channel is defined, multiple curves need to be drawn by superimposing and repeatedly modified and adjusted. As the amount of experimental data increases, the number of curves in the same coordinate system increases, which not only increases the difficulty of drawing, but also interferes with visual judgment due to the intersection and overlap of curves, thus reducing the accuracy of channel calibration. (3) Barrier development involves multiple experimental working conditions, and the curve trends of different working conditions are quite different. When engineers process multiple sets of data consecutively, the visual impression of the preceding curves can interfere with the judgment of subsequent curves, leading to the accumulation of errors when analyzing across operating conditions.

[0006] Therefore, with the development of vehicle crash testing technology, the amount of experimental data continues to grow, and the complexity of operating conditions continues to increase. Traditional methods can no longer meet the development needs of high precision and high efficiency, and there is an urgent need for an automated channel selection scheme. Summary of the Invention

[0007] The present invention aims to provide a method for selecting a rigid barrier channel based on the inflection point of a curve feature, in order to solve the problems in existing vehicle collision tests where the selection of a rigid barrier channel relies on subjective human judgment, resulting in inaccurate results, cumbersome operation, difficulty in quantification, and interference from increased data volume.

[0008] To achieve the above objectives, the present invention adopts the following technical solution: A method for selecting a rigid barrier passage based on the inflection points of a curve feature includes: Data preprocessing steps: Screening the raw force-displacement curves obtained from the barrier dynamic experiment; Data normalization steps: For the filtered curves, normalize the data under the same working condition based on spline interpolation; Curve filtering steps: Process the standardized data using a filter; Baseline generation steps: Calculate the average value of the filtered curve as the baseline, and then filter the baseline. Offset calculation steps: Obtain the absolute value of the distance between each curve and the baseline at the same displacement point, and take the average value of all distances under the same displacement as the offset of that point; Envelope generation steps: Generate upper and lower envelopes by superimposing offset weights on the baselines respectively, and then filter the upper and lower envelopes; Inflection point identification steps: Determine feature inflection points based on the extreme points of the baseline second derivative; Piecewise fitting steps: Divide the upper and lower envelopes into several continuous line segments based on the feature inflection points, and use the least squares method to perform line fitting on each line segment to obtain the slope and intercept of each line segment. Channel calibration steps: Introduce a straight line offset parameter to fine-tune the fitted line segment, and determine the intersection of each line segment as the channel calibration point; if the force value of the current calibration point is less than the force value of the previous calibration point, then use the force value of the previous calibration point; keep the force value unchanged from the last calibration point to the end of the curve; Channel construction steps: Connect all calibration points to form the upper and lower boundaries of a continuous barrier rigid channel.

[0009] The principle and advantages of this scheme are as follows: In practical applications, data preprocessing eliminates abnormal interference, normalization enables mathematical processing of multiple sets of experimental data under the same working condition without correlation, achieving mathematical comparability of multiple sets of data, filtering removes noise to retain true characteristics, and inflection points are accurately identified based on the extreme points of the second derivative of the baseline. Channels are then generated through piecewise fitting and parameter fine-tuning using the least squares method. The calculation of the upper and lower envelope widths based on the baseline adaptively yields the appropriate channel width. This scheme can be optimized by selecting appropriate fine-tuning parameters, achieving process automation. Furthermore, in the post-processing of the calculation files, it can automatically extract the required feature values, calculate key channel performance, and generate output results in the form of data points and line graphs. Depending on different engineering needs, this scheme greatly improves the efficiency of barrier channel definition, reduces the subjectivity of traditional channel definition, and increases channel accuracy.

[0010] Preferably, as an improvement, the data preprocessing step includes filtering: removing redundant data points generated by collision rebound, non-uniform displacement coordinate data, and abnormal points with different force values ​​corresponding to the same displacement.

[0011] Technical benefits: It filters out invalid information, ensures that subsequent analysis is based on high-quality data, and reduces the interference of noise on channel feature extraction.

[0012] Preferably, as an improvement, the normalization process includes: calculating the average value of the displacements at the ends of all curves under the same working condition as a unified abscissa length, and using spline interpolation to map all curves onto uniformly distributed displacement coordinates to obtain a standardized data set.

[0013] Technical effect: Unifying the horizontal coordinate scale of multiple sets of curves enables the originally independent scattered data to have the same displacement benchmark, providing a mathematical basis for the calculation of the average value of multiple curves (baseline generation) and offset analysis (envelope generation), and ensuring that different experimental data can be directly compared and integrated.

[0014] Preferably, as an improvement, the filter adopts the channel frequency of the ISO standard. The channel frequency class includes CFC60, CFC180, and CFC1000. The cutoff frequency of CFC60 is fc=100Hz, the cutoff frequency of CFC180 is fc=300Hz, and the cutoff frequency of CFC1000 is fc=1650Hz. It also includes low-pass filters with custom cutoff frequencies of 50Hz and 30Hz.

[0015] Technical benefits: The use of multi-level ISO standard filters and custom frequency options allows for flexible adjustment of the filtering intensity based on the noise characteristics of experimental data. High-frequency noise (such as electromagnetic interference) can be filtered out by high-level CFC (such as CFC1000), while low-frequency noise (such as equipment vibration) can be processed by low custom frequency (such as 30Hz), thus preserving the true trend and characteristic inflection points of the curve to the greatest extent while denoising.

[0016] Preferably, as an improvement, the criteria for identifying the feature inflection point include: extreme points where the absolute value of the second derivative of the baseline is greater than a preset threshold, and the displacement difference between adjacent inflection points is not less than a preset proportion of the total displacement length.

[0017] Technical benefits: It helps avoid misjudgments caused by minor fluctuations and helps identify inflection points that are of practical significance for rigid channel division.

[0018] Preferably, as an improvement, the range of adjustment of the straight line offset parameter does not change the shape of the baseline.

[0019] Technical benefits: Facilitates ensuring the overall trend characteristics of the baseline.

[0020] Preferably, as an improvement, it also includes a channel verification step: extracting the curve after spline interpolation and collecting points of a preset proportion of test data as check points; calculating the minimum distance from each check point to the channel boundary and calculating the average value; adjusting the channel width parameter to make the average value approach the minimum and the original data not exceed the channel range and reach the preset threshold range.

[0021] Technical effect: By calculating checkpoints and fine-tuning parameters, the generated channels are ensured to cover all valid experimental data while minimizing the average distance from data points to channel boundaries, thus achieving an optimized balance in channel width and avoiding a decrease in reference value due to excessive width or narrowness.

[0022] Preferably, as an improvement, the channel verification step also corrects the channel slope, adjusting it to zero slope if a negative slope segment appears.

[0023] Technical effect: Adjusting the negative slope line segment to zero slope conforms to the physical law that the rigidity of the barrier will not actively weaken during the collision process, ensuring that the channel characteristics are consistent with the actual engineering, and avoiding application errors caused by the conflict between mathematical calculation results and physical common sense. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of a method for selecting a rigid barrier channel based on the inflection points of a curve feature; Figure 2 A schematic diagram illustrating the import of raw data for a segmented honeycomb aluminum channel case study of a barrier vehicle force-wall experiment. Figure 3 A schematic diagram of data normalization for a block-type honeycomb aluminum channel case study of a barrier vehicle force-wall experiment. Figure 4 A schematic diagram of normalized data filtering for a block-cell aluminum channel case in a barrier vehicle force-wall experiment; Figure 5 Schematic diagram of baseline and upper and lower envelope for the blocky honeycomb aluminum channel case of barrier vehicle force wall experiment; Figure 6 A schematic diagram illustrating the calculated characteristic inflection points of a segmented honeycomb aluminum channel case for a barrier vehicle force-wall experiment. Figure 7 Calculate the schematic diagram of the segmented upper and lower channel lines for the blocky vehicle force wall experiment case of the segmented honeycomb aluminum channel; Figure 8 Schematic diagram of the performance channel of the segmented honeycomb aluminum channel in the barrier vehicle force wall experiment. Figure 1 ; Figure 9 Schematic diagram of the performance channel of the segmented honeycomb aluminum channel in the barrier vehicle force wall experiment. Figure 2 ; Figure 10 Schematic diagram of the performance channel of the segmented honeycomb aluminum channel in the barrier vehicle force wall experiment. Figure 3 ; Figure 11 Schematic diagram of the performance channel of the segmented honeycomb aluminum channel in the barrier vehicle force wall experiment. Figure 4 ; Figure 12 Schematic diagram of the upper and lower channel lines in a case study of a barrier vehicle's test on a T-Shape tooling using a segmented honeycomb aluminum channel. Figure 1 ; Figure 13 Schematic diagram of the upper and lower channel lines in a case study of a barrier vehicle's test on a T-Shape tooling using a segmented honeycomb aluminum channel. Figure 2 ; Figure 14 Schematic diagram of the upper and lower channel lines in a case study of a barrier vehicle's test on a T-Shape tooling using a segmented honeycomb aluminum channel. Figure 3 ; Figure 15 This is a schematic diagram of the flow path for a segmented honeycomb aluminum channel used in an experiment involving a barrier vehicle and a T-Shape tooling. Detailed Implementation

[0025] The following detailed description illustrates the specific implementation method: The basic implementation examples are as follows: Figure 1 As shown, a method for selecting a rigid barrier channel based on the inflection point of a curve includes: Data preprocessing steps: The raw force-displacement curves obtained from the barrier dynamic experiment are screened. Raw data screening aims to remove meaningless data. During the barrier dynamic experiment, the vehicle begins to rebound in the later stages of the experiment, leading to a concentration and redundancy of data points. Simultaneously, in the initial data processing, time-domain asymmetry results in uneven displacement data points. Therefore, screening includes: removing redundant data points generated by collision rebound, data with uneven displacement coordinates, and outliers where the same displacement corresponds to different force values.

[0026] Data normalization steps: For the selected curves, data under the same working condition are normalized using spline interpolation to unify the abscissa scale of multiple sets of curves, ensuring that the originally independent scattered data have the same displacement reference. The normalization process includes: in multiple experimental data under the same working condition, the average displacement of the ends of all curves is used as the unified abscissa length. Sampling points are averaged on the displacement abscissa according to a preset multiple of the original data volume. Spline interpolation is used to interpolate the original data points onto a uniform displacement abscissa, and then uniform sampling is performed to ensure the data length remains consistent with the original data volume, resulting in a standardized data set. The standardized data set has a uniform abscissa, and multiple scattered data points share the same displacement abscissa, which is used as the basis for subsequent automated selection of upper and lower channel lines.

[0027] Curve filtering steps: The standardized data is processed using filters. The filters employ ISO standard channel frequencies, with channel frequency classes including CFC60, CFC180, and CFC1000. The cutoff frequency of CFC60 is fc=100Hz, CFC180 is fc=300Hz, and CFC1000 is fc=1650Hz. Custom low-pass filters with cutoff frequencies of 50Hz and 30Hz are also included.

[0028] Baseline generation steps: Calculate the average value of the filtered curves as the baseline, and then filter the baseline. The baseline is the benchmark for channel generation, and its smoothness directly affects the subsequent offset calculation and inflection point identification. Even if the original curve has been filtered, the superposition of small noises from multiple curves still has the probability of causing slight fluctuations in the baseline. Filtering further smooths the baseline, highlighting its core trend (such as the overall slope change), and provides a clear feature basis for subsequent inflection point identification.

[0029] Offset calculation steps: Obtain the absolute value of the distance between each curve and the baseline at the same displacement point, and take the average value of all distances at the same displacement as the offset of that point. The formula for calculating the absolute value of distance is:

[0030] in, Let i be the value of the curve at displacement point i. The value of the baseline at displacement point i.

[0031] Envelope generation steps: Introduce an upper offset weight and a lower offset weight The upper and lower envelopes are generated by superimposing offset weights on the baselines and offsetting them upward and downward by a*w distances, respectively.

[0032] The upper envelope is:

[0033] The lower envelope is:

[0034] in, Indicates the baseline. Indicates the offset. Indicates the upper offset weight. This indicates the weight of the lower offset.

[0035] The upper and lower envelopes are filtered. The envelope is the prototype of the channel boundary, generated by superimposing the baseline with the offset. The offset itself contains discrete information of the original curve (which may have a small amount of residual noise). Filtering can eliminate local fluctuations in the envelope, ensure its boundary is smooth and continuous, avoid the distortion of the line segment slope due to noise during subsequent piecewise fitting, and ultimately ensure the stability of the channel boundary.

[0036] Filtering is integrated into the aforementioned steps to achieve layered noise reduction: first, the original data is purified; then, the core baseline is optimized; and finally, the boundary shape (envelope) is smoothed. Each filtering step targets the noise characteristics of different data types, collectively ensuring the accuracy of the entire process from data input to channel output, avoiding noise interference with curve feature extraction and channel parameter calculation, and ultimately achieving an objective and reliable definition of a rigid channel.

[0037] Inflection point identification steps: Determine characteristic inflection points based on the extreme points of the second derivative of the baseline. Calculate the upper and lower extreme points of the second derivative of the baseline, and mark the coordinates of all extreme points on the baseline. These extreme points are the characteristic inflection points of the curve baseline. The identification criteria for characteristic inflection points include: extreme points where the absolute value of the second derivative of the baseline is greater than a preset threshold, and the displacement difference between adjacent inflection points is not less than a preset proportion of the total displacement length.

[0038] Piecewise fitting steps: Divide the upper and lower envelopes into several continuous line segments based on the feature inflection points, and use the least squares method to fit a straight line to each line segment to obtain the slope and intercept of each line segment.

[0039] Specifically, a feature inflection point selection method is introduced. Feature inflection points are manually selected from the automatically determined feature inflection points for final plotting. Under the same experimental conditions, only one inflection point selection is needed, named [P1, P2, P3…Pn]. Based on the selected inflection points, the upper and lower envelopes are divided into several data segments. For example, the upper envelope: ,[ The same applies to the lower envelope. Calculate the slope and intercept of each straight line segment using the least squares method from these multiple data segments.

[0040] Channel calibration steps: Introduce a straight line offset parameter to fine-tune the fitted line segment. In this embodiment, the vertical offset of the straight line is... and As a fine-tuning parameter, the straight line offset parameter is dynamically adjusted based on engineering requirements and verification results, shifting the straight line vertically. The adjustment range of the straight line offset parameter does not change the shape of the baseline, only slightly altering the channel width. The intersection point of multiple straight lines is determined as the channel calibration point. The calibration point is judged; if the force value of the current calibration point is less than the force value of the previous calibration point, the force value of the previous calibration point is used; the force value remains unchanged from the last calibration point to the end of the curve.

[0041] Channel construction steps: Connect all calibration points to form the upper and lower boundaries of a continuous barrier rigid channel.

[0042] The process also includes a channel verification step: extracting the curve after arbitrary spline interpolation and collecting points of a preset proportion of test data as checkpoints; in this scheme, the preset proportion is 1 / 20. The minimum distance from each checkpoint to the channel boundary (upper channel or lower channel) is calculated and averaged. The minimum distance is expressed as min(| |,| The channel performance is verified by adjusting the channel width parameter to minimize the average value while ensuring the original data does not exceed the channel range and reaches a preset threshold. The channel verification step also corrects the channel slope; if a negative slope segment appears, it is adjusted to zero slope. This ensures that the channel characteristics are consistent with engineering reality, avoiding application errors caused by conflicts between mathematical calculations and physical principles.

[0043] To better understand the solution, an example will be provided. Figures 2-7 This is a schematic diagram illustrating a segmented honeycomb aluminum channel definition for a barrier vehicle force-wall (LCW) experiment. Figures 8-11 This is a schematic diagram of the performance channel. Figures 12-14 This is a schematic diagram of the upper and lower channel lines for a segmented honeycomb aluminum channel definition case in the barrier vehicle's T-Shape tooling experiment (T-Shape). Figure 15 This is a schematic diagram of the flow path for a segmented honeycomb aluminum channel used in an experiment involving a barrier vehicle and a T-Shape tooling. Figure 15 In the optimization process, the definitions and functions of the flowcharts are shown in Table 1 below: Table 1

[0044] This application greatly improves the efficiency of barrier passage definition based on different engineering needs, reduces the subjectivity of traditional passage definition, and increases passage accuracy.

[0045] The above descriptions are merely embodiments of the present invention, and common knowledge such as specific technical solutions and / or characteristics are not described in detail here. It should be noted that those skilled in the art can make various modifications and improvements without departing from the technical solutions of the present invention, and these should also be considered within the scope of protection of the present invention. These modifications and improvements 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 selecting a rigid barrier channel based on the inflection point of a curve, characterized in that, include: Data preprocessing steps: Screening the raw force-displacement curves obtained from the barrier dynamic experiment; Data normalization steps: For the filtered curves, normalize the data under the same working condition based on spline interpolation; Curve filtering steps: Process the standardized data using a filter; Baseline generation steps: Calculate the average value of the filtered curve as the baseline, and then filter the baseline. Offset calculation steps: Obtain the absolute value of the distance between each curve and the baseline at the same displacement point, and take the average value of all distances under the same displacement as the offset of that point; Envelope generation steps: Generate upper and lower envelopes by superimposing offset weights on the baselines respectively, and then filter the upper and lower envelopes; Inflection point identification steps: Determine feature inflection points based on the extreme points of the baseline second derivative; Piecewise fitting steps: Divide the upper and lower envelopes into several continuous line segments based on the feature inflection points, and use the least squares method to perform line fitting on each line segment to obtain the slope and intercept of each line segment. Channel calibration steps: Introduce a straight line offset parameter to fine-tune the fitted line segment, and determine the intersection of each line segment as the channel calibration point; if the force value of the current calibration point is less than the force value of the previous calibration point, then use the force value of the previous calibration point; keep the force value unchanged from the last calibration point to the end of the curve; Channel construction steps: Connect all calibration points to form the upper and lower boundaries of a continuous barrier rigid channel.

2. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 1, is characterized in that... The data preprocessing step includes filtering: removing redundant data points generated by collision rebound, non-uniform displacement coordinate data, and abnormal points with different force values ​​corresponding to the same displacement.

3. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 1, is characterized in that... The normalization process includes: calculating the average value of the displacements at the ends of all curves under the same working condition as a unified horizontal coordinate length, and using spline interpolation to map all curves onto a uniformly distributed displacement coordinate system to obtain a standardized data set.

4. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 1, is characterized in that: The filters adopt the channel frequencies of the ISO standard. The channel frequency levels include CFC60, CFC180, and CFC1000. The cutoff frequency of CFC60 is fc=100Hz, the cutoff frequency of CFC180 is fc=300Hz, and the cutoff frequency of CFC1000 is fc=1650Hz. It also includes low-pass filters with custom cutoff frequencies of 50Hz and 30Hz.

5. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 1, is characterized in that... The criteria for identifying the feature inflection points include: extreme points where the absolute value of the second derivative of the baseline is greater than a preset threshold, and the displacement difference between adjacent inflection points is not less than a preset proportion of the total displacement length.

6. The method for selecting a rigid barrier channel based on the inflection point of a curve according to claim 1, characterized in that: The range of adjustment of the straight line offset parameter does not change the shape of the baseline.

7. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 1, is characterized in that... It also includes a channel verification step: extracting the curve after spline interpolation and collecting points of a preset proportion of test data as check points; Calculate the minimum distance from each checkpoint to the channel boundary and take the average value. Adjust the channel width parameter to minimize the average value while ensuring that the original data does not exceed the channel range and reaches the preset threshold range.

8. The method for selecting a rigid barrier channel based on the inflection point of a curve, as described in claim 7, is characterized in that: The channel verification step also corrects the channel slope; if a negative slope segment appears, it is adjusted to zero slope.

Images