Tire wear degree estimation method based on in-tire strain

Abstract

The invention discloses a tire wear degree estimation method based on in-tire strain, which comprises the following steps: modeling through finite element analysis software to obtain three-dimensional tire models with different wear degrees, extracting strain data of the three-dimensional tire models under different working conditions (air pressure, load and speed), and drawing a curve graph to extract features; and carrying out correlation analysis on the extracted features and tire wear, selecting features with relatively high correlation with the tire wear as regression analysis features, taking the air pressure, the load, the speed and the regression analysis features as input quantities, and taking the tire wear degree as an output quantity to train and construct a tire wear degree estimation model. And estimating the wear degree of the tire by using the established model, and verifying the validity of the model by using actual test data. According to the method, the tire wear degree estimation result is more accurate, and the method can better adapt to tire wear degree estimation under complex working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of tire wear degree estimation, and in particular to a tire wear degree estimation method based on tire internal strain. Background Art

[0002] As the only part of a car that contacts the road, the performance of the tire is related to the safety, stability and comfort of driving. Tire wear is an important indicator of tire performance, which will change the tread morphology, stiffness and contact characteristics, thereby reducing the drainage and grip performance of the tire, affecting the safety of car driving. Estimation of the degree of tire wear can remind the driver to replace the tire in time after the tire reaches a certain degree of wear, thereby avoiding excessive wear of the tire, improving the vehicle's grip, and thus improving driving safety and stability.

[0003] With the rapid development of sensor technology, more and more domestic and foreign scholars use sensors to obtain data information related to tire wear, and combine them with corresponding algorithms to estimate the degree of tire wear. According to different measurement principles, there are mainly types such as acceleration sensors, strain sensors and optical sensors.

[0004] In the invention patent application "A tire wear monitoring method and system" with application number 202110174140.0, the three-axis acceleration sensor and tire pressure sensor attached to the inner wall of the tire collect the three-axis acceleration signal and tire pressure signal of the tire during operation at different wear levels, analyze and extract characteristic parameters, and build a wear prediction model with the wear level as the target and the characteristic parameters as the independent variable; and collect the three-axis acceleration signal and tire pressure signal of the tire during operation in real time, extract the corresponding characteristic parameters, and use the wear prediction model to estimate the real-time wear level of the tire. However, the number of features extracted is relatively small, namely the trough area in the tire pressure, speed and z-axis acceleration characteristic waveform, which reduces the accuracy of the tire wear estimation.

[0005] In the invention patent application "Tire Wear Estimation System, Tire Wear Estimation Program, and Tire Wear Estimation Method" with application number 202080073649.X, the strain sensor installed inside the pneumatic tire obtains the strain signal generated during the operation of the tire, and estimates the wear state of the tread of the pneumatic tire based on the strain signal output at the moment when the contact area of ​​the tread corresponding to the strain sensor is pushed out from the road surface and the reference value of the strain signal. However, the strain signal extracted has only one time series waveform and a relatively single feature type, which leads to insufficient adaptability to the estimation of tire wear under complex working conditions.

[0006] It can be seen that the current method of estimating the degree of tire wear based on sensor-collected data has the problems of a small number of extracted features and a single type, making it difficult to obtain a large amount of data related to tire wear, resulting in low accuracy in estimating the degree of tire wear and poor adaptability to complex tire operating conditions. Summary of the invention

[0007] The present invention aims to solve the problems of low accuracy and poor adaptability of the current method for estimating the degree of tire wear based on sensor collected data, and provides a method for estimating the degree of tire wear based on tire internal strain.

[0008] To solve the above problems, the present invention is achieved through the following technical solutions:

[0009] A method for estimating tire wear based on tire internal strain comprises the following steps:

[0010] Step 1: Use finite element analysis software to model the tire and obtain three-dimensional tire models with different degrees of wear;

[0011] Step 2: Using the established three-dimensional tire models with different wear degrees, different air pressures, different loads and different speeds are set for simulation;

[0012] Step 3: For each working condition, extract the strain data in the axial, circumferential and radial directions within a circle of the inner liner centerline of the three-dimensional tire model, and draw the corresponding strain curve graph;

[0013] Step 4: For each working condition, the circumferential strain curve graph and the radial strain curve graph are selected for feature extraction among the axial strain curve graph, the circumferential strain curve graph and the radial strain curve graph;

[0014] The features extracted based on the circumferential strain curve graph include the circumferential strain small peak height feature, the circumferential strain small peak width feature, the circumferential strain small peak area feature, the circumferential strain large peak height feature, the circumferential strain large peak width feature, and the circumferential strain large peak area feature;

[0015] The features extracted based on the radial strain curve graph include radial strain small valley height feature, radial strain small valley width feature, radial strain small valley area feature, radial strain large valley height feature, radial strain large valley width feature, and radial strain large valley area feature;

[0016] Step 5: First, calculate the correlation coefficient between each extracted feature and tire wear under each working condition, then calculate the average value of the correlation coefficient between the same feature and tire wear under all working conditions, and finally select the feature with the larger absolute value of the average value of the correlation coefficient as the regression analysis feature;

[0017] Step 6: Taking the regression analysis features of the three-dimensional tire model and the working conditions of the tire, i.e., air pressure, load and speed, as input quantities, and the wear degree of the three-dimensional tire model as output quantity, the support vector regression model is machine-learned to obtain a tire wear degree estimation model;

[0018] Step 7: Use the strain sensor installed inside the tire to be tested to collect the strain data of the tire to be tested after one rotation, and draw the corresponding strain curve to extract the regression analysis characteristics of the tire to be tested; at the same time, use the air pressure sensor and load sensor installed inside the tire to be tested, and the speed sensor installed in the vehicle to collect the air pressure, load and speed of the tire to be tested respectively; then send the regression analysis characteristics, air pressure, load and speed of the tire to be tested into the tire wear degree estimation model to estimate the wear degree of the tire to be tested.

[0019] The wear degree in the above step 1 includes zero wear degree, the air pressure condition in the above step 2 needs to include the rated air pressure condition, and the load condition needs to include the rated load condition.

[0020] The speed in step 2 above is the speed at which the tire reaches a steady free rolling state.

[0021] In the above step 4, the specific process of extracting features from the circumferential strain curve graph with the horizontal axis being the sampling point and the vertical axis being the circumferential strain is as follows:

[0022] Calculate the mean circumferential strain of all sampling points on the circumferential strain curve, and draw a circumferential strain mean line on the circumferential strain curve;

[0023] The vertical coordinate difference between the circumferential strain mean line and the maximum peak point of the circumferential strain curve is taken as the circumferential strain small peak height feature;

[0024] The horizontal coordinate difference between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the small peak width feature of the circumferential strain;

[0025] The peak area of ​​the circumferential strain curve between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the circumferential strain small peak area feature;

[0026] The difference between the maximum peak value and the minimum valley value of the circumferential strain curve is taken as the large peak height feature of the circumferential strain;

[0027] The horizontal coordinate difference between the two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak width feature of the circumferential strain;

[0028] The peak area of ​​the circumferential strain curve between two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak area feature of the circumferential strain.

[0029] In the above step 4, the specific process of extracting features from the radial strain curve graph with the horizontal axis being the sampling point and the vertical axis being the radial strain is as follows:

[0030] Calculate the radial strain mean of all sampling points on the radial strain curve, and draw a radial strain mean line on the radial strain curve;

[0031] The difference between the radial strain mean line and the minimum valley point of the radial strain curve is taken as the radial strain valley height feature;

[0032] The horizontal coordinate difference between the two intersection points of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain small valley width feature;

[0033] The valley area of ​​the radial strain curve between the two intersection points on the left and right of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain small valley area feature;

[0034] The difference between the maximum peak value and the minimum valley value of the radial strain curve is taken as the radial strain valley height feature;

[0035] The horizontal coordinate difference between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley width feature;

[0036] The valley area of ​​the radial strain curve between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley area feature.

[0037] In the above step 5, the correlation coefficient is the Pearson correlation coefficient.

[0038] In the above step 6, in step 6, it is necessary to expand the data set consisting of the regression analysis characteristics of the three-dimensional tire model, the tire working conditions, i.e., air pressure, load and speed, and the wear degree of the three-dimensional tire model to increase the data volume, and then perform machine learning on the support vector regression model to improve the accuracy of the regression prediction results and reduce errors.

[0039] Compared with the prior art, the present invention first obtains a large number of tire models of different working conditions and extracts the required strain data through finite element software modeling, thereby greatly reducing the time cost and financial cost of actual experiments; then, based on the strain data obtained by finite element software modeling, multiple features on the circumferential and radial strain curves are extracted, and the features are selected according to correlation analysis; then, a tire wear degree estimation model is constructed through a machine learning method to describe the complex nonlinear relationship between different features and tire wear degrees; when predicting and estimating, the air pressure, load, speed of the tire to be tested and features with a high correlation with tire wear are used as input quantities, and the established model is used to accurately estimate the tire wear degree, thereby solving the problems of low tire wear degree estimation accuracy and inability to better adapt to complex tire working conditions due to a small number of input features. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The flowchart of a tire wear degree estimation method based on tire internal strain is shown in FIG.

[0041] Figure 2 Two-dimensional tire cross-section models with four different degrees of wear: (a) 0 mm wear, (b) 2 mm wear, (c) 4 mm wear, and (d) 6 mm wear.

[0042] Figure 3 A circular path is created on the centerline of the tire inner liner.

[0043] Figure 4 are the features extracted from the circumferential strain curve.

[0044] Figure 5 are the features extracted from the radial strain curve.

[0045] Figure 6 Results of the tire wear estimation training set for the SVR algorithm.

[0046] Figure 7 Results of the tire wear estimation test set for the SVR algorithm.

[0047] Figure 8 The results of tire wear estimation using experimental test data. DETAILED DESCRIPTION

[0048] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in combination with specific examples and with reference to the accompanying drawings.

[0049] A tire wear estimation method based on tire internal strain, such as Figure 1 As shown, the specific steps are as follows:

[0050] Step 1: Use finite element analysis software to model the tire and obtain n three-dimensional tire models with different wear degrees, where n is a set value, n≥4.

[0051] Finite Element Analysis (FEA) uses mathematical approximation to simulate real physical systems. By using simple and interacting elements, i.e. units, a finite number of unknowns can be used to approximate real systems with infinite unknowns. Finite element analysis replaces complex problems with simpler ones and then solves them. It regards the solution domain as consisting of many small interconnected subdomains called finite elements, assumes a suitable approximate solution for each unit, and then derives the total satisfaction conditions for solving this domain to obtain the solution to the problem. This solution is not an exact solution, but an approximate solution, because the actual problem is replaced by a simpler problem. Since most practical problems are difficult to obtain exact solutions, and finite elements not only have high calculation accuracy, but can also adapt to various complex shapes, they have become an effective means of engineering analysis.

[0052] Step 1.1, use finite element software to establish a two-dimensional tire cross-section model with 0 mm wear (non-wear), and then translate its tread part inward by the same or different distances each time to obtain n two-dimensional tire cross-section models with different wear degrees.

[0053] In this embodiment, Abaqus finite element software is used to establish a two-dimensional tire cross-section model with 0mm wear, and its tread part is modified. The specific method is to trace the tread edge line, translate the edge line inward to obtain a new tread edge contour and remove the excess tread part. This method can ensure that the tread is smooth and natural after wear. In this embodiment, n=4. The tread part is translated inward 2mm each time, and the translation is performed three times, and finally four two-dimensional tire cross-section models with different wear degrees of 0mm, 2mm, 4mm, and 6mm are obtained.

[0054] Step 1.2: Mesh the two-dimensional tire cross-section models with n different wear degrees and check the mesh quality, then submit the analysis job.

[0055] In this embodiment, meshing is performed on two-dimensional tire cross-section models with four different degrees of wear. The number of meshes has a great impact on the calculation time and results of tire analysis. Although more meshes can improve the accuracy of the results, it takes more calculation time; while fewer meshes save calculation time, it may lead to reduced accuracy of the results. Therefore, it is necessary to introduce an appropriate number of meshes to reduce the time required for analysis and calculation while taking into account the accuracy of the results. After the model is meshed, it is necessary to check the quality of the mesh to confirm that no erroneous elements are displayed. Then submit the job for analysis, and view the model results through the visualization module after the analysis is completed. The results of the two-dimensional tire cross-section model with four different degrees of wear are shown below. Figure 2 shown.

[0056] Step 1.3: Generate n three-dimensional tire models with different degrees of wear by rotating the two-dimensional tire cross-section model around the central axis, and generate a road surface in contact with the three-dimensional tire model.

[0057] In this embodiment, four two-dimensional tire cross-section models with different wear degrees are rotated around the central axis to obtain four three-dimensional tire models with different wear degrees, and the road surface type in contact with the three-dimensional tire model is set to a rigid plane.

[0058] Step 2: Using the established three-dimensional tire models with n different degrees of wear, set m different air pressures, k different loads, and t different speeds for simulation, and finally obtain tire simulation data with n degrees of wear under m×k×t conditions. Among them, m, k, and t are all set values, and m, k, and t ≥ 4.

[0059] Step 2.1, set m different air pressures for n three-dimensional tire models with different wear degrees.

[0060] The air pressure is simulated by applying a pressure load on the inner surface of the tire. In this embodiment, m=4. According to the actual tire pressure range of the automobile, the air pressure is set to 0.211Mpa, 0.241Mpa, 0.271Mpa, and 0.301Mpa, among which the rated air pressure is 0.241Mpa.

[0061] Step 2.2: Set k different loads for n three-dimensional tire models with different wear degrees.

[0062] When applying the load, the road surface is moved toward the tire to establish contact between the road surface and the tire. Once the contact is established, the vertical load will be applied to the road surface, thereby completing the load simulation. In this embodiment, k=4. The load is set to 1100N, 2200N, 3300N, and 4400N, of which the rated load is 3300N.

[0063] Step 2.3, set t different speeds for n three-dimensional tire models with different wear degrees.

[0064] The speed is the speed of the tire when it reaches a steady free rolling state. The steady free rolling state of the tire is the state when the torque applied to the tire is zero, which is different from the braking state and driving state of the tire. In this embodiment, t=4. The speed is set to 20km / h, 40km / h, 60km / h, and 80km / h.

[0065] For the same linear motion speed ν, when the tire rotates at different angular speeds ω, it corresponds to three motion conditions: braking, free rolling, and driving. Usually, the combination of the angular velocity ω0 and the linear motion speed ν during free rolling is unknown in advance, and the steady-state rolling analysis requires the free rolling angular velocity ω0 and the linear motion speed ν to be specified at the same time, so an indirect method is needed to obtain the free rolling angular velocity ω0. The specific method steps are as follows:

[0066] Step 2.3.1: First, perform tire inflation and static loading analysis to obtain the tire's inflation radius R1 and static loading radius R2.

[0067] Step 2.3.2: Divide the linear motion velocity ν by the inflated radius R1 of the tire to obtain the angular velocity ω1. At this time, the tire is in a braking state.

[0068] Step 2.3.3: Divide the linear motion velocity ν by the static loading radius R2 of the tire to obtain the angular velocity ω2. At this time, the tire is in the driving state.

[0069] Step 2.3.4: Assume that the tire free rolling radius is R0, which is smaller than the inflation radius R1 but larger than the static loading radius R2. Therefore, the free rolling angular velocity ω0 is within the range determined by the following formula:

[0070]

[0071] By analyzing the angular velocity between ω1 and ω2, the relationship between torque and angular velocity is obtained. The angular velocity corresponding to zero torque is the free rolling angular velocity ω0.

[0072] Step 3: For each working condition, extract the strain data within a circle (one cycle) on the center line of the inner liner of the three-dimensional tire model (extract strain data in the axial, circumferential and radial directions in total), and draw the corresponding strain curve graph.

[0073] For each working condition, the generation process of the strain curve diagram is as follows:

[0074] Step 3.1: Create a circular path along the center line of the tire inner liner on the 3D tire model. Figure 3 shown.

[0075] Since the trajectory of the circular path is the same as the trajectory of the strain sensor installed on the center line of the inner liner of the actual tire as the tire moves around, the extraction of tire strain data under this path can be used to simulate the process of strain sensor collecting tire strain data. The method for creating a circular path is to select the point on the center line of the inner liner of the tire that is farthest from the road surface, that is, the vertex, as the starting point, and select points on the center line of the inner liner of the tire in a counterclockwise direction. The last selected end point coincides with the starting point to form a complete circular path.

[0076] Step 3.2: Convert the rectangular coordinate system into a cylindrical coordinate system, and extract the axial, circumferential and radial strain data of the tire in the cylindrical coordinate system.

[0077] Because the tire is a cylinder, the rectangular coordinate system needs to be converted into a cylindrical coordinate system to extract the axial, circumferential and radial strains of the tire in the cylindrical coordinate system. The axial, circumferential and radial directions together constitute the three orthogonal directions of the cylindrical coordinates. The axial direction is along the central axis of rotation of the cylinder, the circumferential direction is along the circumference of the cylinder section, and the radial direction is along the radius of the cylinder section. The conversion relationship between the rectangular coordinate system and the cylindrical coordinate system is as follows:

[0078]

[0079] Where r∈[0,+∞), z∈R. x, y, z are three coordinate variables of the rectangular coordinate system, r, z is the three coordinate variables of the cylindrical coordinate system.

[0080] Step 3.3: extract the strain data of the three-dimensional tire model under different working conditions along the center line of the inner liner under the circumferential path created in step 3.1, and draw strain curves in the axial, circumferential and radial directions.

[0081] The tire carcass is simulated using shell elements in the finite element analysis software. Since the tire is a typical nonlinear mechanical system, the output is the true strain LE under geometric nonlinear conditions. According to the mapping relationship, the direction of the shell element 1 is the z direction in the cylindrical coordinate system, that is, the axial direction, and the axial strain is LE11; the direction of the shell element 2 is the z direction in the cylindrical coordinate system, that is, the axial direction, and the axial strain is LE11; direction, that is, the circumferential direction, then the circumferential strain is LE22; the shell element 3 direction is the r direction in the cylindrical coordinate system, that is, the radial direction, then the radial strain is LE33.

[0082] Step 4: Compare and analyze the strain curves in the axial, circumferential and radial directions, and select the strain curves in the circumferential and radial directions for feature extraction.

[0083] By analyzing the curves of axial strain, circumferential strain and radial strain of the tire under different working conditions, it is found that with the change of working conditions, the axial strain curve does not show good regularity, while the circumferential strain curve and radial strain curve show good regularity, so the circumferential strain curve and radial strain curve are selected to extract features.

[0084] Step 4.1, extract features from the circumferential strain curve (the abscissa of the circumferential strain curve represents the sampling point, and the ordinate represents the circumferential strain), such as Figure 4 As shown, the specific process is as follows:

[0085] Calculate the mean circumferential strain of all sampling points on the circumferential strain curve, and draw a circumferential strain mean line on the circumferential strain curve;

[0086] The vertical coordinate difference between the mean line of circumferential strain and the maximum peak point of the circumferential strain curve is taken as the small peak height feature of circumferential strain, which is represented by LE22-h;

[0087] The horizontal coordinate difference between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the small peak width feature of the circumferential strain, which is represented by LE22-w;

[0088] The peak area of ​​the circumferential strain curve between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the circumferential strain small peak area feature, which is represented by LE22-s;

[0089] The difference between the maximum peak value and the minimum valley value of the circumferential strain curve is taken as the large peak height feature of the circumferential strain and is represented by LE22-H.

[0090] The horizontal coordinate difference between the two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak width feature of the circumferential strain, which is represented by LE22-W;

[0091] The peak area of ​​the circumferential strain curve between two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak area feature of the circumferential strain and is represented by LE22-S.

[0092] Step 4.2: Extract features from the radial strain curve (the abscissa of the radial strain curve represents the sampling point, and the ordinate represents the radial strain), such as Figure 5 As shown, the specific process is as follows:

[0093] Calculate the radial strain mean of all sampling points on the radial strain curve, and draw a radial strain mean line on the radial strain curve;

[0094] The difference between the radial strain mean line and the minimum valley point of the radial strain curve is taken as the radial strain valley height feature, which is represented by LE33-h.

[0095] The horizontal coordinate difference between the two intersection points of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain valley width feature, which is represented by LE33-w.

[0096] The valley area of ​​the radial strain curve between the two intersection points of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain small valley area feature and is represented by LE33-s.

[0097] The ordinate difference between the maximum peak point and the minimum valley point of the radial strain curve is taken as the radial strain valley height feature, which is represented by LE33-H.

[0098] The horizontal coordinate difference between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley width feature, which is represented by LE33-W;

[0099] The valley area of ​​the radial strain curve between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley area feature and is represented by LE33-S.

[0100] Step 5: First, calculate the correlation coefficient between each extracted feature and tire wear under each working condition, then calculate the average value of the correlation coefficient between the same feature and tire wear under all working conditions, and finally select the feature with the larger absolute value of the average value of the correlation coefficient as the regression analysis feature for subsequent regression analysis.

[0101] In statistics, the Pearson correlation coefficient is widely used to measure the degree of correlation between two variables, and its value is between -1 and 1. If the correlation coefficient is greater than 0, it indicates that the two variables are positively correlated, that is, the larger the value of one variable, the larger the value of the other variable; if the correlation coefficient is less than 0, it indicates that the two variables are negatively correlated, that is, the larger the value of one variable, the smaller the value of the other variable. The closer the absolute value of the correlation coefficient is to 1, the stronger the correlation between the two variables, and the closer the absolute value of the correlation coefficient is to 0, the weaker the correlation between the two variables. To this end, the present invention uses Pearson correlation analysis, assuming that variable X is the eigenvalue extracted in step 4 above, and variable Y is the degree of tire wear, and the Pearson correlation coefficient between the two variables can be calculated by the following formula:

[0102]

[0103] In the formula, N represents the number of variable values. In this embodiment, there are four tire wear degrees, so N=4.

[0104] First, under each working condition, the correlation coefficient between each feature and tire wear is calculated, and m×k×t groups of correlation coefficients can be obtained between each feature and tire wear. Then, the correlation coefficients between the same feature and tire wear under all working conditions are averaged, that is, the average value of the m×k×t groups of correlation coefficients is calculated to determine the strength of the correlation between the feature and tire wear. Finally, after the average value of the correlation coefficients is calculated, the features whose absolute value of the average value is greater than the preset value (such as 0.8) are selected for regression analysis.

[0105] In this embodiment, tire wear is used as a variable to perform correlation analysis with the extracted features. It is necessary to take the tire wear degree under the same working conditions (same air pressure, load and speed) as a group. Then, after the correlation analysis between each feature and tire wear, 64 sets of correlation coefficients can be obtained. The average value of the 64 sets of correlation coefficients obtained is calculated, and the strength of the correlation between the feature and tire wear is determined according to the size of the average value. After the average value of the correlation coefficients is calculated, the overall correlation coefficient between each feature and tire wear is obtained, as shown in the following table:

[0106] feature Correlation coefficient LE22-h 0.2579 LE22-w 0.8902 LE22-s 0.9377 LE22-H 0.4968 LE22-W 0.2053 LE22-S 0.6676 LE33-h -0.8213 LE33-w 0.3648 LE33-s -0.8992 LE33-H -0.8369 LE33-W 0.4355 LE33-S -0.9096

[0107] As can be seen from the table, there are six features with absolute values ​​of correlation coefficients greater than 0.8, namely LE22-w, LE22-s, LE33-h, LE33-s, LE33-H, and LE33-S. Therefore, these six features are selected for subsequent regression analysis.

[0108] Step 6: Using the regression analysis features of the three-dimensional tire model and the tire operating conditions, i.e., air pressure, load, and speed, as input, and the wear degree of the three-dimensional tire model as output, perform machine learning on the support vector regression model to obtain a tire wear degree estimation model.

[0109] Step 6.1. The present invention utilizes the data (air pressure, load, speed, regression analysis characteristics and wear degree) obtained by the above finite element simulation to form a data set.

[0110] The data set is composed of tire wear, air pressure, load, speed and six features with high correlation with wear, where the first column of data is tire wear, and the second to tenth columns of data are air pressure, load and speed, as well as six features with high correlation with wear (regression analysis features).

[0111] Step 6.2: Use data augmentation to increase the data volume.

[0112] In order to improve the accuracy of regression prediction results and reduce errors, data expansion is used to increase the amount of data. Since the extracted data is a one-dimensional array data, linear interpolation is used. The linear interpolation method makes numerical estimates based on the two adjacent data to the left and right of the point to be interpolated in the one-dimensional data sequence. It does not calculate the average value of the data size of the two points, but allocates the proportion according to the distance to the two points. Taking two adjacent points of a feature of the data set as an example, the principle of linear interpolation is introduced as follows:

[0113] Assuming that the coordinates of two adjacent points are (x0, y0) and (x1, y1), they form a straight line, and the coordinates of an unknown point on the straight line are (x, y), then the following formula can be obtained:

[0114]

[0115] Where k is the slope of the line. If the value of x is known and the value of y is required, transform formula (4) to obtain the formula for solving the value of y as follows:

[0116]

[0117] Given the value of y, the method for finding the value of x is the same as above, except that x and y are swapped.

[0118] Under the premise of not changing the characteristic variation trend, linear interpolation is used to perform uniform interpolation to expand the data set to u (u≥2) times of the original data set. In this embodiment, u=4, that is, the original 256 rows of data are expanded to 1024 rows of data.

[0119] Step 6.3: Use the six features with high correlation with wear in the data set after data expansion, as well as tire pressure, load and speed as input, and the tire wear degree as output to perform machine learning on the support vector regression (SVR) model to build a tire wear degree estimation model.

[0120] Support vector regression is a machine learning algorithm that originates from support vector machine (SVM) and is an important research direction of SVM. The core idea of ​​SVR is to find a hyperplane to minimize the interval between samples, which is used to solve regression problems. Compared with traditional machine learning regression algorithms, SVR has obvious advantages in preventing over-training, operation speed and result accuracy.

[0121] The support vector regression (SVR) algorithm is used to estimate the degree of tire wear. Suppose the sample set A = {(x1, y1), (x2, y2), …, (x p ,y p )}, where x irepresents the input vector, which in this embodiment is air pressure, load, speed and six features with high correlation with wear (regression analysis features); i represents the output vector, which is the tire wear degree in this embodiment; p is the number of samples. Assume that the regression model is:

[0122] f(x i )=w T x i +b (6)

[0123] Where w and b are the model parameters to be solved.

[0124] The optimization goal of SVR is:

[0125]

[0126] In the formula, ‖w‖ is the Euclidean norm, which represents the distance between two points (the norm of the vector).

[0127] Points within the boundary satisfy the following conditions:

[0128] |y i -(w T x i +b)|≤ε (8)

[0129] Where ε is the allowed prediction value f(x i ) and the actual value y i The difference between the two, the prediction result error is less than or equal to ε when the prediction value is correct, that is, the prediction result error value is within the interval [-ε,ε], that is, the sample data point must fall within the interval [f(x i )-ε,f(x i )+ε]. In actual situations, it is often impossible to directly find a suitable ε so that most of the data are within the interval. In order to make more data fall into the interval, it is necessary to introduce a slack variable to allow some data to be out of the interval, but a penalty term will be generated. Introducing the slack variable ξ, It is worth noting that for any sample data x i , when it is inside or on the edge of the interval, ξ=0; when it is above the interval, ξ>0, When it is below the interval, ξ=0, After introducing the slack variables, the regression prediction problem of support vector regression is transformed into a minimum optimization problem, and the following formula is obtained:

[0130]

[0131] Wherein, C is the penalty coefficient, which is set to 1000 in this embodiment.

[0132] Introducing the Lagrange multiplier λ≥0, μ ≥ 0, Transform the minimum optimization problem into a dual problem and obtain the following formula:

[0133]

[0134] In the formula, K(x i ,x j ) is a kernel function. In this embodiment, a radial basis kernel function is used, and its formula is:

[0135] K(x i ,x j ) = exp(-γ||x i -x j || 2 ) (11)

[0136] Wherein γ is a radial basis kernel function parameter, which is set to 1 in this embodiment.

[0137] Solving formula (10), we can obtain the tire wear estimation model expression of support vector regression:

[0138]

[0139] This embodiment uses the root mean square error (RMSE) to evaluate the prediction accuracy. The smaller the RMSE value, the more accurate the prediction of the model. The formula is as follows:

[0140]

[0141] In the formula, represents the estimated value of tire wear in the i-th group of data, y i Represents the actual value of the tire wear degree in the i-th group of data.

[0142] The data set is divided and 75% of the data is randomly selected as the training set to train the model parameters; the remaining 25% of the data is used as the test set to evaluate the performance of the model. The results of the tire wear estimation training set using the SVR algorithm are as follows: Figure 6 The results of the test set are shown in Figure 7 shown.

[0143] Different air pressure, load and speed conditions were applied to four real tires with wear degrees of 0mm, 2mm, 4mm and 6mm, including air pressure of 0.21Mpa, 0.24Mpa and 0.27Mpa, load of 1100N, 2200N and 3300N, and speed of 20km / h, 40km / h and 60km / h. The strain sensor installed inside the real tire collected data under the above conditions (the same data type as that extracted by finite element simulation), and the data was sent to the constructed tire wear estimation model to estimate the tire wear degree and verify the effectiveness of the model. The results of tire wear degree estimation using experimental test data are shown in the figure. Figure 8 shown.

[0144] Step 7, using the strain sensor installed inside the tire to be tested to collect the strain data of the tire to be tested after one rotation, and draw the corresponding strain curve diagram (circumferential strain curve diagram and radial strain curve diagram) to extract the regression analysis characteristics of the tire to be tested; at the same time, using the air pressure sensor and load sensor installed inside the tire to be tested to collect the air pressure and load of the tire to be tested, and using the speed sensor installed in the vehicle to obtain the speed signal; then the regression analysis characteristics, air pressure, load and speed of the tire to be tested are sent to the tire wear degree estimation model to estimate the wear degree of the tire to be tested.

[0145] The strain sensor, the air pressure sensor, and the load sensor are simultaneously installed at the position of the center line of the inner liner of the tire to be tested, and the installation method is to attach the sensor to the inner wall of the tire. It should be noted that when attaching the sensor to the inner wall of the tire, it is necessary to ensure that the sensor is firmly attached to prevent the sensor from falling off during the movement of the tire. The strain sensor is used to collect strain data in the axial, circumferential and radial directions of the tire to be tested for one rotation (one cycle). The circumferential and radial strain data are selected and the corresponding strain curves are drawn, and the regression analysis features determined in step 5 are extracted using the same method as step 4. The air pressure sensor is used to collect the air pressure of the tire to be tested in the current cycle. The load sensor is used to collect the load of the tire to be tested in the current cycle. The speed sensor installed in the vehicle collects the speed of the tire to be tested in the current cycle. The regression analysis features, air pressure, load and speed of the tire to be tested in the current cycle collected are sent to the tire wear degree estimation model, and the tire wear degree estimation model outputs the predicted wear degree of the tire to be tested.

[0146] The invention discloses a tire wear degree estimation method based on tire internal strain. The method uses finite element analysis software to perform modeling, obtains three-dimensional tire models with different wear degrees, and sets different air pressure, load and speed conditions for the three-dimensional tire models with different wear degrees. The axial, circumferential and radial strain data on the center line of the inner liner of the tire model under different working conditions are extracted and a curve graph is drawn. After analyzing the curve graph, the circumferential and radial strain curves are selected to extract features. The extracted features are subjected to correlation analysis with tire wear, and features with greater correlation with tire wear are selected for subsequent regression analysis. A tire wear degree estimation model is constructed using a machine learning method, and air pressure, load, speed and selected features with greater correlation with tire wear are used as input quantities, and tire wear degree is used as output quantity. The established model is used to estimate the tire wear degree, and the validity of the model is verified using actual test data. The invention uses the extracted multiple features and the air pressure, load and speed of the tire as model inputs, so that the tire wear degree estimation result is more accurate and can better adapt to the tire wear degree estimation under complex working conditions.

[0147] It should be noted that although the embodiments of the present invention described above are illustrative, they are not intended to limit the present invention, and therefore the present invention is not limited to the above specific embodiments. Without departing from the principles of the present invention, any other embodiments obtained by those skilled in the art under the guidance of the present invention are deemed to be within the protection of the present invention.

Claims

1. A tire wear degree estimation method based on tire internal strain, characterized in that: The steps include: Step 1: Use finite element analysis software to model the tire and obtain three-dimensional tire models with different degrees of wear; Step 2: Using the established three-dimensional tire models with different wear degrees, different air pressures, different loads and different speeds are set for simulation; Step 3: For each working condition, extract the strain data in the axial, circumferential and radial directions within a circle of the inner liner centerline of the three-dimensional tire model, and draw the corresponding strain curve graph; Step 4: For each working condition, the circumferential strain curve graph and the radial strain curve graph are selected for feature extraction among the axial strain curve graph, the circumferential strain curve graph and the radial strain curve graph; The features extracted based on the circumferential strain curve graph include the circumferential strain small peak height feature, the circumferential strain small peak width feature, the circumferential strain small peak area feature, the circumferential strain large peak height feature, the circumferential strain large peak width feature, and the circumferential strain large peak area feature; The features extracted based on the radial strain curve graph include radial strain small valley height feature, radial strain small valley width feature, radial strain small valley area feature, radial strain large valley height feature, radial strain large valley width feature, and radial strain large valley area feature; Step 5: First, calculate the correlation coefficient between each extracted feature and tire wear under each working condition, then calculate the average value of the correlation coefficient between the same feature and tire wear under all working conditions, and finally select the feature with the larger absolute value of the average value of the correlation coefficient as the regression analysis feature; Step 6: Taking the regression analysis features of the three-dimensional tire model and the working conditions of the tire, i.e., air pressure, load and speed, as input quantities, and the wear degree of the three-dimensional tire model as output quantity, the support vector regression model is machine-learned to obtain a tire wear degree estimation model; Step 7: Use the strain sensor installed inside the tire to be tested to collect strain data of the tire to be tested after one rotation, and draw the corresponding strain curve to extract the regression analysis characteristics of the tire to be tested; at the same time, use the air pressure sensor and load sensor installed inside the tire to be tested, and the speed sensor installed in the vehicle to collect the air pressure, load and speed of the tire to be tested respectively; The regression analysis characteristics, air pressure, load and speed of the tire to be tested are then sent to the tire wear degree estimation model to estimate the wear degree of the tire to be tested.

2. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: The wear degree in step 1 includes the zero wear degree, the air pressure condition in step 2 needs to include the rated air pressure condition, and the load condition needs to include the rated load condition.

3. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: The speed in step 2 is the speed at which the tire moves when it reaches a steady free-rolling state.

4. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: In step 4, the specific process of extracting features from the circumferential strain curve graph with the horizontal axis being the sampling point and the vertical axis being the circumferential strain is as follows: Calculate the mean circumferential strain of all sampling points on the circumferential strain curve, and draw a circumferential strain mean line on the circumferential strain curve; The vertical coordinate difference between the circumferential strain mean line and the maximum peak point of the circumferential strain curve is taken as the circumferential strain small peak height feature; The horizontal coordinate difference between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the small peak width feature of the circumferential strain; The peak area of ​​the circumferential strain curve between the two intersection points of the circumferential strain mean line and the circumferential strain curve below the maximum peak point is taken as the circumferential strain small peak area feature; The difference between the maximum peak value and the minimum valley value of the circumferential strain curve is taken as the large peak height feature of the circumferential strain; The horizontal coordinate difference between the two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak width feature of the circumferential strain; The peak area of ​​the circumferential strain curve between two small valley points of the circumferential strain curve below the circumferential strain mean line is taken as the large peak area feature of the circumferential strain.

5. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: In step 4, the specific process of extracting features from the radial strain curve graph with the horizontal axis being the sampling point and the vertical axis being the radial strain is as follows: Calculate the radial strain mean of all sampling points on the radial strain curve, and draw a radial strain mean line on the radial strain curve; The difference between the radial strain mean line and the minimum valley point of the radial strain curve is taken as the radial strain valley height feature; The horizontal coordinate difference between the two intersection points of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain small valley width feature; The valley area of ​​the radial strain curve between the two intersection points on the left and right of the radial strain mean line and the radial strain curve above the minimum valley point is taken as the radial strain small valley area feature; The difference between the maximum peak value and the minimum valley value of the radial strain curve is taken as the radial strain valley height feature; The horizontal coordinate difference between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley width feature; The valley area of ​​the radial strain curve between the two large peak points of the radial strain curve above the radial strain mean line is taken as the radial strain large valley area feature.

6. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: In step 5, the correlation coefficient is the Pearson correlation coefficient.

7. The tire wear degree estimation method based on tire internal strain according to claim 1, characterized in that: In step 6, it is necessary to expand the data set consisting of the regression analysis characteristics of the three-dimensional tire model, the tire working conditions, i.e., air pressure, load and speed, and the wear degree of the three-dimensional tire model to increase the data volume, and then perform machine learning on the support vector regression model.

Images