Automobile performance fatigue evaluation system based on finite element analysis

Abstract

The invention discloses an automobile performance fatigue evaluation system based on finite element analysis, which relates to the technical field of data processing, and is characterized in that an initial peripheral three-dimensional image model and an initial internal three-dimensional image model of a vehicle are generated according to an initial condition data set, and are overlapped and mapped to obtain an initial state three-dimensional image model; generating an initial state three-dimensional image model according to the real-time condition data set, and simultaneously dividing a plurality of initial finite element regions and real-time finite element regions with the same positions and space volumes in the initial state three-dimensional image model and the real-time state three-dimensional image model; according to the method, the real-time finite element areas are obtained, the displacement amount of each space pixel in all the real-time finite element areas under different time nodes is obtained, the real-time condition of the corresponding vehicle is judged according to the displacement amount of each space pixel in the real-time finite element areas under different time nodes, and the accuracy of automobile part condition evaluation is effectively improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and specifically to an automotive performance fatigue assessment system based on finite element analysis. Background Art

[0002] Automotive performance fatigue assessment refers to the process of evaluating and testing the performance stability and durability of an automobile when it is used for a long time or operates under extreme conditions. It aims to verify whether the automobile can maintain good performance, safety, and reliability during actual use.

[0003] The existing automotive performance fatigue assessment technologies have the following defects: Computer-Aided Engineering (CAE) analysis: Using computer simulation software to simulate and analyze the structure, dynamics, and fluid mechanics of an automobile, etc., to evaluate its performance and fatigue characteristics under different working conditions. However, analyzing and interpreting a large amount of test data requires professional knowledge and experience, and it is easy to produce misunderstandings or inaccurate conclusions.

[0004] Test bench testing: Conducting various simulation tests on an automobile on a specially designed test bench, including simulating different road conditions, vibrations, and acceleration conditions, etc. This method can control the test conditions more precisely, but it cannot fully simulate the real road conditions and is costly.

[0005] Therefore, how to improve the versatility of the evaluation results while ensuring the accuracy of automotive performance fatigue assessment is a difficult point in the existing technology. For this reason, an automotive performance fatigue assessment system based on finite element analysis is provided. Summary of the Invention

[0006] In order to solve the above technical problems, the purpose of the present invention is to provide an automotive performance fatigue assessment system based on finite element analysis.

[0007] In order to achieve the above purpose, the present invention provides the following technical solutions: An automotive performance fatigue assessment system based on finite element analysis, including a cloud computing platform, and the cloud computing platform is communicatively connected to a vehicle data acquisition module, a finite element analysis module, and a vehicle condition assessment module; The vehicle data acquisition module is used to collect various sensors arranged at various positions of the vehicle, and then collect the initial condition dataset and real-time condition dataset of the vehicle; The finite element analysis module is used to generate an initial peripheral three-dimensional image model and an initial internal three-dimensional image model of the vehicle according to the initial condition data set, overlap and map the two to obtain an initial state three-dimensional image model, and generate an initial state three-dimensional image model according to the real-time condition data set. A number of initial finite element regions and real-time finite element regions with the same position and spatial volume size are divided in the initial state three-dimensional image model and the real-time state three-dimensional image model, so as to obtain the displacement of each spatial pixel in all real-time finite element regions at different time nodes; The module is used to judge the real-time condition of the corresponding vehicle according to the displacement of each spatial pixel in the real-time finite element region at different time nodes.

[0008] Further, the acquisition process of the initial condition data set and the real-time condition data set includes: The vehicle data acquisition module installs a variety of sensors on each part of the vehicle and sets numbers for each part. The types of sensors include temperature sensors, cameras, and laser sensors; When the vehicle is in a driving state, the vehicle data acquisition module sets a unit data acquisition cycle. Whenever a unit data acquisition cycle starts, the vehicle data acquisition module generates a data acquisition instruction and sends it to each sensor. Then each sensor overwrites the data acquisition instruction received in the previous unit data acquisition cycle with the newly received data acquisition instruction; During the unit data acquisition cycle, the temperature sensor and the camera collect the real-time temperature value and real-time video data of the part where they are located, and the laser sensor continuously sends multiple laser signals to the part where it is located at the same time; When a unit data acquisition cycle ends, the temperature sensor generates a corresponding temperature change curve according to the temperature values collected during the unit data acquisition cycle, and the laser sensor generates multiple groups of laser reflection signal spectra according to the laser reflection signals collected during the unit data acquisition cycle; At the same time, the vehicle data acquisition module collects the initial video data and initial laser reflection signal spectra of each part in the vehicle in the initial state through sensors; The vehicle data acquisition module marks the data uploaded by each sensor with the corresponding number according to the number of the part where each sensor is located. Then the data with the same number in the initial state and the driving state are respectively integrated, so as to obtain the initial condition data set and the real-time condition data set of the corresponding part.

[0009] Further, the process of generating the initial peripheral three-dimensional image model of the vehicle according to the initial condition data set includes: The finite element analysis module extracts the initial video data and the initial laser reflection signal spectrum of each part from the initial state dataset, and divides each initial video data into N initial image data by frame, where N is a natural number greater than 0; Perform grayscale processing on each initial image data, obtain the pixel values of the grayscale pixels, and then calculate the gradient value h and the gradient direction angle θ of each grayscale pixel except the edge positions in the initial image data; Set a gradient value threshold, label the characteristic pixel points for the grayscale pixels with gradient values greater than or equal to the gradient value threshold, and do not perform any operation on the grayscale pixels with gradient values less than the gradient value threshold; Divide each initial image data into m image regions of equal size, and determine whether there are characteristic pixel points in each image region. If not, do not perform any operation; If there are, generate a feature vector with a horizontal angle of θ and a modulus of h according to the gradient direction angle θ and the gradient value h of the characteristic pixel points; Successively select the image regions distributed in a 3*3 square in the initial image data, establish a two-dimensional rectangular coordinate system, and map the feature vectors in the selected image regions onto the two-dimensional rectangular coordinate system; Obtain the vector sum vector of each feature vector, and map the vector sum vector proportionally into the initial image data. If the vector sum vector is located at the characteristic pixel points of the image regions without images, select the image regions distributed in a 4*4,..., t*t square to obtain the corresponding vector sum vector until the vector sum vector is located at the characteristic pixel points of the image regions with images, where t is a natural number greater than 4 and t is less than or equal to the number of image regions on any side of the initial image data; Label the image region where the vector sum vector is located as a characteristic region, and successively splice each initial image data according to the division order of each image region, and then splice each two-dimensional characteristic region to obtain the corresponding three-dimensional characteristic region; Furthermore, generate an initial three-dimensional image model of the appearance of the corresponding part from the initial real-time video data, and label each three-dimensional characteristic region on the initial three-dimensional image model of the periphery.

[0010] Furthermore, the calculation formula for the gradient value h is: The calculation formula for the gradient direction angle θ is: When Or = The gradient direction angle θ = 0; Where 、 、 And respectively represent the pixel values of the upper region, lower region, left region, and right region of the grayscale pixels.

[0011] Furthermore, the process of generating the initial internal three-dimensional image model of the vehicle according to the initial condition dataset includes: When the laser signal penetrates the part, whenever the laser signal passes through a component inside the part, a corresponding local laser reflection signal is generated. Then, the finite element analysis module obtains the position and thickness of the corresponding component inside the part based on the generation time interval and phase value of each phase on the laser reflection signal spectrum, and then establishes the initial internal three-dimensional image model of the corresponding part.

[0012] Furthermore, the establishment process of the initial state three-dimensional image model and the real-time state three-dimensional image model includes: Overlap and map the initial peripheral three-dimensional image model with the initial internal three-dimensional image model, and splice the pixels of the initial peripheral three-dimensional image model and the initial internal three-dimensional image model according to the three-dimensional feature regions on the initial peripheral three-dimensional image model; Then, match the positions of each pixel on the three-dimensional feature region with the pixels at the edge positions of the initial internal three-dimensional image model. If more than half of the pixel positions and pixel values are the same at the corresponding positions of the three-dimensional feature region and the initial internal three-dimensional image model, no operation is performed; If there are not more than half of the pixel positions and pixel values that are the same at the corresponding positions of the three-dimensional feature region and the initial internal three-dimensional image model, directly cover the pixels on the three-dimensional feature region with the pixels at the corresponding positions of the initial internal three-dimensional image model; Then, the initial state three-dimensional image models of each part are obtained. At the same time, according to the steps of generating the initial state three-dimensional image model, the real-time state three-dimensional image models of each part are generated according to the real-time condition dataset.

[0013] Furthermore, the process of obtaining the displacement amounts of each spatial pixel in the real-time finite element region at different time nodes includes: The finite element analysis module respectively divides Num initial finite element regions and real-time finite element regions with the same size in the initial state three-dimensional image model and the real-time state three-dimensional image model, and sets the same number for the initial finite element regions and real-time finite element regions at the same spatial position, where Num is a natural number greater than 0; Establish a three-dimensional space coordinate system, overlap and map the initial state three-dimensional image models and real-time state three-dimensional image models of each part in the three-dimensional space coordinate system, and then match the spatial pixels in the initial finite element regions and real-time finite element regions with the same number in the initial state three-dimensional image model and the real-time state three-dimensional image model; Set the positions and states of the corresponding spatial pixels in the initial finite element region according to the matching results to the initial spatial positions and initial states in the real-time finite element region; Since the components inside each part of the vehicle jitter regularly during driving, displacements and wear occur in the finite element regions corresponding to each component; According to the overlapping mapping results of the initial state three-dimensional image model and the real-time state three-dimensional image model, obtain the initial three-dimensional coordinates and real-time three-dimensional coordinates of each spatial pixel in each real-time finite element region at the initial spatial position and in the real-time state; Set K time nodes according to the time length of the unit data acquisition cycle, and then starting from the initial three-dimensional coordinates of each spatial pixel, obtain the displacement amounts of each spatial pixel at adjacent time nodes, and integrate the displacement amounts of all spatial pixels in each real-time finite element region between each time node during the unit data acquisition cycle to generate a pixel displacement data set, where K is a natural number greater than 0.

[0014] Further, the process of judging the real-time conditions of each part includes: Set a normal displacement threshold for each part respectively, and then compare the normal displacement threshold with the displacement amounts in the entire pixel displacement data set of the corresponding part. If the displacement amount is less than or equal to the normal displacement threshold, no operation is performed; If the displacement amount is greater than the normal displacement threshold, it is judged that the displacement amount between the corresponding time nodes is abnormal. If more than (K - 1) / 3 of the displacement amounts in the pixel displacement data set are abnormal, it is judged that the corresponding real-time finite element region is abnormal during the unit data acquisition cycle, otherwise it is judged that the corresponding real-time finite element region is normal; Count the number of abnormalities in the corresponding real-time finite element regions of each part during the unit data acquisition cycle; Set multiple abnormal quantity threshold intervals for each part respectively, and then judge the real-time state of the corresponding part according to the abnormal quantity threshold where the number of abnormalities is located Compared with the prior art, the beneficial effects of the present invention are: By simultaneously dividing several initial finite element regions and real-time finite element regions with the same positions and spatial volume sizes in the initial state three-dimensional image model and the real-time state three-dimensional image model, the present invention obtains the displacement amounts of each spatial pixel in all real-time finite element regions at different time nodes. According to the displacement amounts of each spatial pixel in the real-time finite element region at different time nodes, the real-time conditions of the corresponding vehicle are judged, thereby realizing the improvement of the versatility of the evaluation results while ensuring the accuracy of the fatigue evaluation of the vehicle performance. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] To more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required in the embodiments. Obviously, the drawings in the following description are only some embodiments described in the present invention.

[0016] Figure 1 This is the schematic diagram of the present invention. Detailed implementation manners

[0017] To make the objectives, technical solutions, and advantages of the present invention clearer, the following will describe the technical solutions of the present invention in detail. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other implementation manners obtained by those of ordinary skill in the art without creative efforts shall fall within the scope protected by the present invention.

[0018] As Figure 1 shown, an automotive performance fatigue evaluation system based on finite element analysis includes a cloud computing platform, which is communicatively connected to a vehicle data acquisition module, a finite element analysis module, and a vehicle condition evaluation module; The vehicle data acquisition module is used to collect a variety of sensors at various positions of the vehicle, and then collect the initial condition dataset and the real-time condition dataset of the vehicle; The finite element analysis module is used to generate an initial external three-dimensional image model and an initial internal three-dimensional image model of the vehicle according to the initial condition dataset, overlap and map the two to obtain an initial state three-dimensional image model, and generate an initial state three-dimensional image model according to the real-time condition dataset. A number of positions and initial finite element regions and real-time finite element regions with the same spatial volume size are divided in both the initial state three-dimensional image model and the real-time state three-dimensional image model, and then the displacement amounts of each spatial pixel in all real-time finite element regions at different time nodes are obtained; The module is used to judge the real-time condition of the corresponding vehicle according to the displacement amounts of each spatial pixel in the real-time finite element region at different time nodes.

[0019] Further, the working principle of the present invention is illustrated through the following embodiments: The vehicle data acquisition module installs a variety of sensors on each part of the vehicle, and numbers each part as a1, a2, a3,..., a n , where n is a natural number greater than 0, and the types of sensors include temperature sensors, cameras, and laser sensors; When the vehicle is in a driving state, the vehicle data acquisition module sets a unit data acquisition period. Whenever a unit data acquisition period starts, the vehicle data acquisition module generates a data acquisition instruction and sends it to each sensor. Then, each sensor overwrites the data acquisition instruction received in the previous unit data acquisition period with the newly received data acquisition instruction; During the unit data acquisition period, the temperature sensor and the camera collect the real-time temperature value and real-time video data of their respective part positions. The laser sensor continuously sends multiple laser signals to its respective part position at the same time. Since the vehicle shakes and undergoes local deformation during driving, multiple sets of laser reflection signals are generated when the laser signal hits the same part of the part; When a unit data acquisition period ends, the temperature sensor generates a corresponding temperature change curve based on the temperature values collected during the unit data acquisition period. The laser sensor generates multiple sets of laser reflection signal spectra based on the laser reflection signals collected during the unit data acquisition period. Then, each sensor sends the data it has collected to the vehicle data acquisition module; At the same time, the vehicle data acquisition module collects the initial video data and initial laser reflection signal spectra of each part in the vehicle in the initial state through sensors. It should be noted that the parts include the engine, tires, etc.; The vehicle data acquisition module labels the data uploaded by each sensor with the corresponding part number according to the part number of each sensor. Then, the data with the same number in the initial state and the driving state are respectively integrated, and the initial condition dataset and real-time condition dataset of the corresponding part are obtained; The vehicle data acquisition module sends the initial condition dataset and real-time condition dataset of each part to the finite element analysis module.

[0020] Furthermore, the finite element analysis module establishes an initial state image model for each part based on the initial condition set. The specific process includes: The finite element analysis module extracts the initial video data and initial laser reflection signal spectra of each part from the initial state dataset, and divides each initial video data into N initial image data by frame, where N is a natural number greater than 0; The initial image data is grayscale processed, and the pixel values of the grayscale pixels are obtained. Then, the gradient value h and gradient direction angle θ of each grayscale pixel except the edge position in the initial image data are calculated. The calculation formula for the gradient value h is: The calculation formula for the gradient direction angle θ is: When or = the gradient direction angle θ = 0; Among them 、 、 and respectively represent the pixel values of the upper neighborhood, lower neighborhood, left neighborhood, and right neighborhood of the grayscale pixel; Set the gradient value threshold, mark the feature pixel positions for the grayscale pixels whose gradient values are greater than or equal to the gradient value threshold, and do nothing for the grayscale pixels whose gradient values are less than the gradient value threshold; Divide each initial image data into m image regions of equal size, and determine whether there are feature pixel positions in each image region. If not, do nothing; If there are, generate a feature vector with a horizontal angle of θ and a modulus of h according to the gradient direction angle θ and gradient value h of the feature pixel position; Successively select the image regions distributed in a 3*3 square in the initial image data, establish a two-dimensional rectangular coordinate system, and map the feature vectors in the selected image regions onto the two-dimensional rectangular coordinate system; Obtain the vector sum vector of each feature vector, and map the vector sum vector proportionally into the initial image data. If the vector sum vector is located at the feature pixel position of the image region without an image, select the image regions distributed in a 4*4,..., t*t square to obtain the corresponding vector sum vector until the vector sum vector is located at the feature pixel position of the image region with an image, where t is a natural number greater than 4 and t is less than or equal to the number of image regions on any side of the initial image data; Mark the image region where the vector sum vector is located as the feature region, and successively splice each initial image data according to the division order of each image region, and then splice each two-dimensional feature region to obtain the corresponding three-dimensional feature region; Furthermore, generate the initial three-dimensional image model of the appearance of the corresponding part from the initial real-time video data, and mark each three-dimensional feature region on the initial three-dimensional image model of the periphery;

[0021] Furthermore, when the laser signal penetrates the part, whenever the laser signal passes through a component inside the part to generate a corresponding local laser reflection signal, the finite element analysis module obtains the position and thickness of the corresponding component inside the part according to the generation time interval and phase value of each phase on the laser reflection signal spectrum, and then establishes the initial three-dimensional image model of the inside of the corresponding part; Perform overlapping mapping on the initial three-dimensional image model of the periphery and the initial three-dimensional image model of the inside, and perform pixel splicing on the initial three-dimensional image model of the periphery and the initial three-dimensional image model of the inside according to the three-dimensional feature regions on the initial three-dimensional image model of the periphery; Since the initial peripheral three-dimensional image model and the initial internal three-dimensional image model correspond to the same part, there are corresponding parts between the initial laser reflection signal spectrum and the initial video data corresponding to the two; Furthermore, each pixel in the three-dimensional feature region is position-matched with the pixels at the edge position of the initial internal three-dimensional image model. If there are more than half of the pixel positions and pixel values that are the same at the corresponding positions between the three-dimensional feature region and the initial internal three-dimensional image model, no operation is performed; If there are not more than half of the pixel positions and pixel values that are the same at the corresponding positions between the three-dimensional feature region and the initial internal three-dimensional image model, the pixels in the three-dimensional feature region directly cover the pixels at the corresponding positions of the initial internal three-dimensional image model; Repeat the above operations to obtain the initial state three-dimensional image models of each part. At the same time, according to the steps of generating the initial state three-dimensional image models, generate the real-time state three-dimensional image models of each part based on the real-time condition data set.

[0022] Furthermore, the finite element analysis module respectively divides Num initial finite element regions and real-time finite element regions of the same size in the initial state three-dimensional image model and the real-time state three-dimensional image model, and sets the same numbers s1, s2,..., s for the initial finite element regions and real-time finite element regions at the same spatial positions Num , where Num is a natural number greater than 0; it should be noted that the finite element analysis module updates the real-time state three-dimensional image model in real time according to the real-time condition data set of each unit data acquisition cycle; Establish a three-dimensional space coordinate system, overlap and map the initial state three-dimensional image models and real-time state three-dimensional image models of each part in the three-dimensional space coordinate system, and then match the spatial pixels in the initial finite element regions and real-time finite element regions with the same numbers in the initial state three-dimensional image model and the real-time state three-dimensional image model; Set the position and state of the corresponding spatial pixels in the initial finite element region to the initial spatial position and initial state in the real-time finite element region according to the matching result; Since the components inside each part of the vehicle jitter regularly during driving, displacements and wear occur in the finite element regions corresponding to each component; Furthermore, perform stress analysis on each spatial pixel in the real-time finite element region. The stress analysis process includes: According to the overlap mapping results of the initial state three-dimensional image model and the real-time state three-dimensional image model, obtain the initial three-dimensional coordinates and real-time three-dimensional coordinates of each spatial pixel in the initial spatial position and real-time state of each real-time finite element region; Set K time nodes according to the time length of the unit data acquisition period, and then obtain the displacement p of each spatial pixel at adjacent time nodes starting from the initial three-dimensional coordinates. The calculation formula of the displacement p is , where represents the displacement between the i-th and j-th time nodes, , , , , , respectively represent the coordinate values of the spatial pixel at the i-th and j-th time nodes. K, j, and i are natural numbers greater than 0, and i, j are less than or equal to K.

[0023] Furthermore, the finite element analysis module integrates the displacements of all spatial pixels in each real-time finite element region between each time node within the unit data acquisition period to generate a pixel displacement data set, and sends all the pixel displacement data sets to the vehicle condition evaluation module; The vehicle condition evaluation module sets a normal displacement threshold for each part respectively, and then compares the normal displacement threshold with the displacements in all the pixel displacement data sets of the corresponding part. If the displacement is less than or equal to the normal displacement threshold, no operation is performed; If the displacement is greater than the normal displacement threshold, it is determined that the displacement between the corresponding time nodes is abnormal. If the number of abnormal displacements in the pixel displacement data set exceeds (K - 1) / 3, it is determined that the corresponding real-time finite element region within the unit data acquisition period is abnormal, otherwise it is determined that the corresponding real-time finite element region is normal; Count the number of abnormalities in the corresponding real-time finite element regions of each part within the unit data acquisition period; Set the first abnormal quantity threshold (0, β1), the second abnormal quantity threshold [β1, β2], and the third abnormal quantity threshold (β2, ∞) for each part respectively, where β1 is less than β2. If the number of existing abnormalities is within the third abnormal quantity threshold (β2, ∞), it is determined that the corresponding part cannot be used; If the number of existing abnormalities is within the second abnormal quantity threshold [β1, β2], it is determined that the corresponding part is slightly worn; If the number of existing abnormalities is within the first abnormal quantity threshold (0, β1), it is determined that the corresponding part is normal.

[0024] The above embodiments are only used to illustrate the technical method of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical method of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical method of the present invention.

Claims

1. An automobile performance fatigue assessment system based on finite element analysis, comprising a cloud computing platform, characterized in that: The cloud computing platform is communicatively connected to a vehicle data acquisition module, a finite element analysis module, and a vehicle condition assessment module; The vehicle data collection module is used to collect data from various sensors installed at various positions of the vehicle, thereby collecting the initial status data set and the real-time status data set of the vehicle; The finite element analysis module is used to generate an initial peripheral three-dimensional image model and an initial internal three-dimensional image model of the vehicle according to the initial state data set, overlap and map the two to obtain an initial state three-dimensional image model, and generate an initial state three-dimensional image model according to the real-time state data set, and divide the initial state three-dimensional image model and the real-time state three-dimensional image model into a plurality of initial finite element regions and real-time finite element regions with the same position and spatial volume, thereby obtaining the displacement of each spatial pixel in all real-time finite element regions at different time nodes; The module is used to determine the real-time status of the corresponding vehicle according to the displacement of each spatial pixel in the real-time finite element area at different time nodes.

2. The automobile performance fatigue evaluation system based on finite element analysis according to claim 1 is characterized in that: The collection process of the initial status data set and the real-time status data set includes: Set the unit data collection cycle. When the vehicle is in driving state, every time a unit data collection cycle starts and ends, the vehicle data collection module generates data to retrieve the temperature change curve and laser reflection signal spectrum generated by each sensor; At the same time, the vehicle data acquisition module collects the initial video data and initial laser reflection signal spectrum of each part in the vehicle in the initial state through sensors; Set a number for each part, and then mark the data uploaded by each sensor with a corresponding number according to the number of the part where each sensor is located. Then, integrate the data with the same number in the initial state and the driving state respectively, and then obtain the initial status data set and real-time status data set of the corresponding part.

3. The automobile performance fatigue evaluation system based on finite element analysis according to claim 2 is characterized in that: The process of the initial peripheral three-dimensional image model includes: Each initial video data is divided into N initial image data by frame, each initial image data is grayed, and the pixel value of the grayed pixel is obtained, and then the gradient value h and the gradient direction angle θ of each grayed pixel except the edge position in the initial image data are obtained; Set the gradient value threshold, mark the grayscale pixels with gradient values ​​greater than or equal to the gradient value threshold as feature pixels, and do nothing with the remaining grayscale pixels; Each initial image data is divided into m image regions of equal size, and it is determined whether there are characteristic pixel points in each image region. Based on the determination results, a characteristic vector with a horizontal angle of θ and a modulus of h is generated, where N and m are natural numbers greater than 0. The vector sum of the corresponding feature vectors is obtained by sequentially selecting image regions distributed in 3*3, 4*4, ..., t*t squares, until the feature vectors and the vector sum are located at feature pixel points with the image region, wherein t is a natural number greater than 4, and t is less than or equal to the number of image regions on any side of the initial image data; The vector and the image region where the vector is located are marked as feature regions, each initial image data is sequentially spliced, and then each two-dimensional feature region is spliced ​​to obtain a three-dimensional feature region; Then, the initial real-time video data generates an initial appearance three-dimensional image model of the corresponding part, and each three-dimensional feature area is marked on the initial peripheral three-dimensional image model.

4. The automobile performance fatigue evaluation system based on finite element analysis according to claim 3 is characterized in that: The calculation formula of the gradient value h is: , the calculation formula of the gradient direction angle θ is: ,when or = When , the gradient direction angle θ=0; , , as well as Respectively represent the pixel values ​​of the upper area, lower area, left area, and right area of ​​the grayscale pixel.

5. The automobile performance fatigue evaluation system based on finite element analysis according to claim 3 is characterized in that: According to the generation time interval and phase value of each phase on the spectrum of the laser reflection signal, the position and thickness of the corresponding component inside the part are obtained, and then the initial internal three-dimensional image model of the corresponding part is established.

6. The automobile performance fatigue evaluation system based on finite element analysis according to claim 5, characterized in that: The process of establishing the initial state three-dimensional image model and the real-time state three-dimensional image model includes: Overlapping and mapping the initial peripheral three-dimensional image model and the initial internal three-dimensional image model, and pixel-splicing the initial peripheral three-dimensional image model and the initial internal three-dimensional image model according to the three-dimensional feature area on the initial peripheral three-dimensional image model; Match each pixel on the 3D feature region with the pixel on the edge of the initial internal 3D image model. If more than half of the pixel positions and pixel values ​​at the corresponding positions of the 3D feature region and the initial internal 3D image model are the same, no operation is performed. If more than half of the pixel positions and pixel values ​​at the corresponding positions of the three-dimensional feature area and the initial internal three-dimensional image model do not have the same value, the pixels on the three-dimensional feature area are directly overlaid on the pixels at the corresponding positions of the initial internal three-dimensional image model to obtain the initial state three-dimensional image model of each part. At the same time, according to the steps of generating the initial state three-dimensional image model, the real-time state three-dimensional image model of each part is generated according to the real-time state data set.

7. The automobile performance fatigue evaluation system based on finite element analysis according to claim 6 is characterized in that: The process of obtaining the displacement of spatial pixels at different time nodes includes: Dividing Num initial finite element regions and real-time finite element regions of the same size in the initial state three-dimensional image model and the real-time state three-dimensional image model, and setting the same number for the initial finite element regions and the real-time finite element regions at the same spatial position, wherein Num is a natural number greater than 0; Establish a three-dimensional space coordinate system, overlap and map the initial state three-dimensional image model and the real-time state three-dimensional image model of each part in the three-dimensional space coordinate system, and then match the spatial pixels in the initial finite element area and the real-time finite element area with the same number in the two; According to the matching result, the position and state of the corresponding spatial pixel in the initial finite element area are set to the initial spatial position and initial state in the real-time finite element area; According to the overlapping mapping results of the initial state three-dimensional image model and the real-time state three-dimensional image model, the initial three-dimensional coordinates and real-time three-dimensional coordinates of each real-time finite element region at the initial spatial position and each spatial pixel in the real-time state are obtained; Set K time nodes, start from the initial three-dimensional coordinates of each spatial pixel, obtain the displacement of each spatial pixel at adjacent time nodes, integrate the displacement of all spatial pixels in each real-time finite element area between each time node within the unit data acquisition cycle to generate a pixel displacement data set, where K is a natural number greater than 0.

8. The automobile performance fatigue evaluation system based on finite element analysis according to claim 7 is characterized in that: The process of determining the real-time status of each component includes: A normal displacement threshold is set for each part, and then the normal displacement threshold is compared with the displacement in all pixel displacement data sets of the corresponding parts. The displacement between corresponding time nodes is judged to be abnormal according to the comparison result. If the displacement in the pixel displacement data set exceeds (K-1) / 3, the corresponding real-time finite element area in the unit data acquisition cycle is judged to be abnormal, otherwise the corresponding real-time finite element area is judged to be normal. The number of anomalies in the real-time finite element area corresponding to each part within the unit data collection period is counted, and multiple anomaly number threshold intervals are set for each part. Then, the real-time status of the corresponding part is determined according to the anomaly number threshold at which the number of anomalies exists.

Images