Vision-Assisted Intelligent Coal Flow System Transportation Monitoring Method

Abstract

This invention discloses a vision-assisted intelligent coal transport monitoring method, belonging to the field of coal transport monitoring technology. The method includes: acquiring surface texture images of the coal transport conveyor belt, extracting feature points and analyzing pixel displacement gradients, and establishing a transport strain field by combining linear and nonlinear strain models; locating the boundary contact area between the drive roller and the conveyor belt, and triggering a contact analysis signal through strain analysis; extracting the effective contact area based on the signal, analyzing contact efficiency and homogeneity, and determining the contact off-center loading area and travel deviation area; further extracting contact features and equipment features to establish a correlation matrix, tracing the causes of anomalies, and conducting a durability status assessment based on the causes to generate a dynamic adjustment strategy. This invention achieves monitoring and intelligent optimization of the conveyor belt contact state, improving the stability of the transport system and the lifespan of the equipment.

Description

Technical Field

[0001] This invention relates to the field of coal flow transportation monitoring technology, and specifically to a vision-assisted intelligent coal flow system transportation monitoring method. Background Technology

[0002] In coal transport systems, the contact condition between the conveyor belt and the drive roller directly affects transport efficiency and equipment lifespan. Uneven loading and insufficient contact between the conveyor belt and drive roller lead to poor contact, causing localized wear, slippage, or even breakage of the conveyor belt, resulting in production interruptions. Existing monitoring technologies have the following limitations:

[0003] Existing technologies fail to adequately identify the contact boundary between the belt and the roller, lack a clear distinction between the effective contact area and the non-contact area, make it difficult to assess contact quality through indicators such as contact efficiency and uniform distribution, and are unable to locate specific contact defect areas.

[0004] Existing technologies mostly focus on anomaly identification, failing to link the intrinsic relationship between roller status and contact characteristics, making it difficult to trace the cause of anomalies; moreover, they lack a closed-loop mechanism from cause analysis to dynamic adjustment, making it impossible to proactively improve the contact status through targeted strategies, resulting in high equipment maintenance costs and short service life.

[0005] Therefore, this invention provides a vision-assisted intelligent coal flow system transportation monitoring method. Summary of the Invention

[0006] The purpose of this invention is to provide a vision-assisted intelligent coal flow system transportation monitoring method to solve at least one of the aforementioned problems in the prior art.

[0007] A vision-assisted intelligent coal flow system transportation monitoring method includes the following steps:

[0008] Surface texture images of the coal transport conveyor belt are collected and pixel displacement gradient analysis is performed. The image displacement is converted into strain distribution to establish the transport strain field.

[0009] Locate the boundary contact area of ​​the transport strain field, extract the linear and nonlinear strain components of the boundary contact area and perform strain analysis to obtain strain analysis characteristics. If the strain analysis characteristics break through the analysis boundary, a contact analysis signal is triggered.

[0010] Based on the contact analysis signal, the effective contact area of ​​the boundary contact region is extracted and the contact efficiency is analyzed. If the contact efficiency is in the low efficiency range, the contact area is homogeneously analyzed from the axial and motion directions to determine the contact off-center load area and the stroke deviation area.

[0011] Contact characteristics and equipment characteristics of the contact off-center load zone and the travel deviation zone are extracted, an association matrix is ​​established, and the source analysis of the contact off-center load zone and the travel deviation zone is carried out. Based on the source analysis results, the durability status is assessed, and a dynamic adjustment strategy for the coal flow transportation system is established.

[0012] As a further technical solution of the present invention: the method for establishing the transportation strain field is as follows:

[0013] Obtain the actual displacement gradient and perform linear strain verification to determine whether the displacement gradient satisfies the displacement-linear strain relationship. If it does, directly extract the strain components of the linear strain.

[0014] If the condition is not met, the strain type will be distinguished, and the displacement gradient of different strain types will be extracted and converted into strain components.

[0015] The strain components of different strain types are labeled, and the labeled strain components are processed by a spatial interpolation algorithm to obtain continuous strain values ​​on the conveyor belt surface.

[0016] The strain values ​​are correlated with the spatial coordinates of the original conveyor belt image to construct the transport strain field.

[0017] As a further technical solution of the present invention, the displacement gradient is obtained as follows:

[0018] Collect surface texture images of the non-coal transport face of the coal transport conveyor belt and extract texture feature points from the images;

[0019] Calculate the pixel coordinate displacement of texture feature points between adjacent frames of the surface texture image;

[0020] Obtain the pixel coordinate displacement of all texture feature points, divide the conveyor belt surface into multiple grid cells, calculate the average displacement of texture feature points in each grid cell, and the displacement gradient of the grid cell.

[0021] Obtain the displacement gradient of all mesh elements and establish the displacement gradient matrix;

[0022] Construct a pixel-physical mapping relationship to transform the displacement gradient of pixels in the displacement gradient matrix into the actual displacement gradient.

[0023] As a further technical solution of the present invention, the strain analysis characteristics are obtained as follows:

[0024] Linear and nonlinear strain components of the boundary contact region are extracted for strain analysis to obtain strain state coefficients and contact width. The strain state coefficients and contact width are used as strain analysis features.

[0025] An analysis boundary is set. If the strain analysis characteristics exceed the analysis boundary, a contact analysis signal is triggered.

[0026] As a further technical solution of the present invention, the strain analysis is performed as follows:

[0027] Obtain the boundary of the initial contact region, construct the boundary contact region, and obtain the linear and nonlinear strain components of the boundary contact region in the transport strain field;

[0028] Obtain the transport strain field corresponding to the straight section of the coal transport conveyor belt, calculate the mean value of the linear shear strain in the linear strain component of the straight section, and the maximum value of the linear shear strain in the boundary contact area;

[0029] The shear strain ratio is obtained by calculating the deviation ratio between the maximum value and the mean value of the linear shear strain in the linear strain component of the segment.

[0030] The ratio of the nonlinear strain area to the total area of ​​the boundary contact region is obtained to get the boundary area ratio.

[0031] Calculate the rate of change of the boundary area ratio and the straight section area ratio, and sum the rate of change with the shear strain ratio to obtain the strain state coefficient.

[0032] As a further technical solution of the present invention, the homogeneity analysis is performed as follows:

[0033] If the effective contact efficiency is in the low efficiency range, a homogeneity analysis is triggered.

[0034] Axial homogeneity analysis of the contact area was performed to obtain the shear strain difference and axial deviation index to determine the contact off-center load zone;

[0035] The rate of change of contact area is obtained by homogeneous analysis of the direction of motion, and the travel deviation zone is determined.

[0036] As a further technical solution of the present invention: the effective contact efficiency is obtained by means of:

[0037] Obtain the mesh within the effective contact area, convert the pixel size of the mesh cell to the physical size based on the pixel-physical mapping relationship, and obtain the area of ​​a single mesh based on the physical size;

[0038] Obtain the number of grids within the effective contact area, and combine it with the area of ​​a single effective grid. Multiply the number of grids by the area of ​​a single grid to obtain the effective contact area.

[0039] Obtain the arc length corresponding to the wrap angle of the drive roller and the width of the conveyor belt, and multiply the arc length and the width of the conveyor belt to obtain the theoretical contact area.

[0040] The effective contact area is calculated by comparing the effective contact area with the theoretical contact area to obtain the effective contact efficiency.

[0041] As a further technical solution of the present invention, the method for performing the axial homogeneity analysis is as follows:

[0042] Divide the boundary contact area into N strips along the drive roller axis, calculate the effective contact area ratio of each strip, and determine the off-center strip and compliant strip based on the contact area ratio.

[0043] Obtain the maximum and average effective contact area ratio within the off-center strip group, and input the maximum and average values ​​into the axial deviation equation to obtain the axial deviation index.

[0044] Obtain the mean value of linear shear strain within the off-center load strip group to obtain the mean value of the off-center load group; obtain the mean value of linear shear strain of adjacent compliant strip groups within the off-center load strip group to obtain the mean value of the compliant group.

[0045] The shear strain difference is obtained by calculating the ratio of the difference between the mean of the off-center load group and the mean of the compliant group.

[0046] As a further technical solution of the present invention: the method for conducting the durability status assessment is as follows:

[0047] Obtain the source tracing results, extract the contact off-center load area caused by the tilt of the drum axis, and determine the peak contact pressure using Hertzian contact theory;

[0048] The peak contact pressure is input into the wear equation to obtain the wear amount. Based on the wear amount, a wear prediction model is constructed to predict the remaining wear resistance life of the conveyor belt in the contact off-center load area.

[0049] The friction force was obtained and a driving loss model was established to quantify the driving force loss rate of the driving roller.

[0050] As a further technical solution of the present invention: the method for obtaining the source tracing result is as follows:

[0051] Extract the axial deviation index, shear strain difference, and contact area change rate of the off-center load zone;

[0052] Axial deviation index, shear strain difference, and contact area change rate are used as contact characteristics;

[0053] The pressure difference between the two sides of the drive roller in the contact off-center loading area and the surface temperature of the drive roller are obtained, and the pressure difference and surface temperature are used as equipment characteristics.

[0054] Integrate contact features and equipment features to establish a feature correlation matrix;

[0055] Feature correlation analysis was performed on the feature correlation matrix to trace the causes of contact off-center load and determine the cause type.

[0056] The beneficial effects of this invention are:

[0057] 1. By acquiring surface texture images of the conveyor belt using an industrial camera, deformation-resistant feature points are extracted. Based on the displacement gradient threshold, linear and nonlinear strain models are dynamically switched, which improves the physical authenticity of strain components from the perspective of mechanical principles and provides a precise strain field data foundation for subsequent analysis.

[0058] 2. By locating the boundary contact area between the drive roller and the conveyor belt, and combining the strain characteristics of the straight section and the boundary contact area, the strain state coefficient is calculated. This is then correlated with the comparison between the width of the incompletely bonded belt and the critical width. This enables the quantitative judgment and signal triggering of contact anomalies, reducing the missed detections or misjudgments caused by the reliance on experience in traditional methods, and improving the timeliness and accuracy of contact status monitoring.

[0059] 3. Based on the contact analysis signal, the effective contact area is extracted, and the contact quality is evaluated through the effective contact efficiency. Combined with the homogeneity analysis of the axial and motion directions, the contact off-load area and the travel deviation area are located. This helps to solve the limitation of traditional monitoring that can only identify macroscopic anomalies and cannot quantify the distribution characteristics of abnormal areas, and provides specific spatial positioning and characteristic basis for tracing the cause of anomalies.

[0060] 4. By integrating contact characteristics and equipment characteristics to establish a correlation matrix, the causes of anomalies can be traced. Combined with Hertzian contact theory and driving loss model, durability status assessment is carried out, and targeted dynamic adjustment strategies are generated, forming a closed loop of identification, analysis, tracing, and optimization. This helps to improve the problem that traditional methods can only monitor anomalies but cannot achieve intelligent optimization, thereby improving the stability of the coal flow transportation system and the service life of equipment. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a flowchart of the vision-assisted intelligent coal flow system transportation monitoring method of the present invention;

[0063] Figure 2 This is a flowchart illustrating how to determine whether a displacement-linear strain relationship is satisfied, according to an embodiment of the present invention.

[0064] Figure 3 This is a schematic diagram of the modules of the intelligent coal flow system transportation monitoring system based on vision assistance of the present invention. Detailed Implementation

[0065] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0066] Example 1

[0067] like Figure 1 As shown in the figure, the vision-assisted intelligent coal flow system transportation monitoring method provided in this embodiment of the invention includes the following steps:

[0068] S1. Collect surface texture images of the coal transport conveyor belt and perform pixel displacement gradient analysis. Convert the image displacement into strain distribution and establish the transport strain field.

[0069] The method for acquiring surface texture images of the coal transport conveyor belt and performing pixel displacement gradient analysis is as follows:

[0070] Preferably, an industrial camera equipped with an infrared fill light is used to acquire surface texture images of the non-coal transport surface of the coal transport conveyor belt, and the surface texture images are preprocessed and texture feature points are extracted from the images.

[0071] Perform inter-frame matching on texture feature points, remove mismatched points, and calculate the pixel coordinate displacement of texture feature points between adjacent frames of the surface texture image.

[0072] It should be explained that, based on Scale Invariant Feature Transform (SIFT) or Fast Feature Extraction (ORB) algorithm, feature points with strong resistance to deformation, such as corners and edges of the conveyor belt surface, are preferentially selected. For texture feature points of two consecutive frames of surface texture images, mismatched points are eliminated by nearest neighbor matching algorithm.

[0073] Obtain the pixel coordinate displacement of all texture feature points, divide the conveyor belt surface into multiple grid units, and calculate the average displacement of texture feature points in each grid unit.

[0074] The displacement gradient of a mesh cell is calculated using the finite difference method based on the average displacement of texture features within the mesh cell. ;

[0075] Obtain the displacement gradient of all mesh elements and establish the displacement gradient matrix;

[0076] Construct a pixel-physical mapping relationship to transform the displacement gradient of pixels in the displacement gradient matrix into the actual displacement gradient;

[0077] Understandably, the pixel-physical mapping relationship is constructed as follows: a checkerboard calibration board of known size (e.g., 100mm × 100mm, with a grid size of 10mm × 10mm) is pasted onto the conveyor belt surface. Images of the calibration board are taken from three different angles, and the pixel coordinates of the corner points are extracted along with the known physical coordinates. The homography matrix H is solved using the least squares method to achieve coordinate system transformation, ensuring that the physical coordinates (x, y) and pixel coordinates (u, v) satisfy the following: ;

[0078] The homography matrix H contains rotation, translation, and scaling parameters, which are calculated using the findHomography function in the OpenCV machine vision library. Finally, the pixel values ​​in the displacement gradient matrix are multiplied by calibration coefficients to obtain the actual physical displacement gradient.

[0079] like Figure 2 As shown, linear strain verification is performed based on the actual displacement gradient to determine whether the displacement gradient satisfies the displacement-linear strain relationship. If it does, the strain components of the linear strain are directly extracted.

[0080] If the condition is not met, the strain type will be distinguished, and the displacement gradient of different strain types will be extracted and converted into strain components.

[0081] It should be further explained that, based on the engineering threshold of the small deformation mechanics assumption (taking the absolute value of each component of the displacement gradient ≤ 5%), to determine whether the current deformation is within the linear elastic range, when the displacement gradient... , When all components are ≤5%, it is assumed that the theoretical relationship between displacement and linear strain is satisfied, and direct conversion is performed;

[0082] Determining whether the displacement gradient satisfies the displacement-linear strain relationship uses a clear quantitative standard. The purpose is to reduce the significant errors that would result from forcibly applying a linear model when large deformations (displacement gradient > 5%) occur due to local wrinkles or material yielding in the conveyor belt.

[0083] By judging and verifying, linear transformation is applied to areas that meet the conditions, while nonlinear strain model based on the geometric nonlinear Green strain formula is switched to areas that exceed the threshold. The strain components of the nonlinear strain type are calculated, which improves the physical authenticity of the strain components from the perspective of mechanical principles. This provides a data foundation for subsequent tension inversion and deformation compensation, and reduces the deviation of the whole chain analysis caused by the mismatch between model assumptions and actual deformation.

[0084] The linear strain components include normal strain. , and shear strain + ;

[0085] It needs to be explained that, and The shear strain corresponds to the elongation or contraction rate in the x and y directions, respectively, reflecting the shear deformation. When the displacement gradient exceeds the threshold, the deformation gradient tensor F is first constructed, and F is composed of elements F ij Composition, F ij Represents the coordinates x after deformation i (e.g., in the x and y directions) relative to the initial coordinate x j (e.g., partial derivatives in the X and Y directions); then with respect to F ij Take the transpose (interchange rows and columns) and multiply it by itself to get... Then subtract the unit tensor I with diagonal elements of 1 and the rest of 0, and multiply the result by 1 / 2 to obtain the Green strain tensor.

[0086] The strain components of different strain types are labeled, and the processed strain components are processed by a spatial interpolation algorithm to obtain continuous strain values ​​on the conveyor belt surface.

[0087] The strain values ​​are correlated with the spatial coordinates of the original conveyor belt image to construct the transport strain field;

[0088] Those skilled in the art will understand that the processed strain components are discrete values ​​calculated based on grid cells. By using the Kriging interpolation method in spatial interpolation algorithms, the strain values ​​at the grid gaps can be estimated based on the known strain values ​​of adjacent grids, thus obtaining a continuous strain distribution covering the conveyor belt surface. Subsequently, with the help of the previously established pixel-physical mapping relationship, each continuous strain value is mapped to the spatial coordinates of the original image (i.e., the actual physical location on the conveyor belt surface, such as the x and y coordinates from the center of the roller), ultimately forming a transport strain field with a one-to-one correspondence between spatial location and strain magnitude. This ensures that the strain data not only has numerical values ​​but also clearly defines its specific distribution location on the conveyor belt surface, providing a complete strain data foundation with spatial attributes for subsequent steps such as locating the boundary contact area.

[0089] Understandably, the purpose of constructing a transportation strain field is:

[0090] Objective 1: To provide a strain distribution basis for spatial coordinate association for the subsequent positioning of the boundary contact area. By associating the strain value with the spatial coordinates of the original conveyor belt image, the contact area between the drive roller and the conveyor belt can be located based on the strain field.

[0091] Objective 2: To provide multi-type strain data support for the quantitative analysis of contact states. By distinguishing between linear and nonlinear strain components, it provides raw data for the subsequent extraction of strain characteristics of the boundary contact region (such as shear strain ratio and nonlinear strain area ratio).

[0092] Objective 3: To improve the accuracy of strain analysis at the mechanical principle level, by dynamically switching between linear and nonlinear strain models (based on displacement gradient threshold), the problem that traditional single models cannot adapt to scenarios such as local wrinkles and large deformations of belts is solved, providing reliable strain field data for contact anomaly identification and cause tracing.

[0093] S2. Locate the boundary contact area of ​​the transport strain field, extract the linear and nonlinear strain components of the boundary contact area and perform strain analysis to obtain strain analysis characteristics. If the strain analysis characteristics break through the analysis boundary, the contact analysis signal is triggered.

[0094] The method for locating the boundary contact area of ​​the transport strain field is as follows:

[0095] Obtain the physical position parameters of the drive roller of the coal transport conveyor belt, and locate the boundary of the initial contact area between the drive roller and the conveyor belt by combining the spatial coordinates of the transport strain field;

[0096] Those skilled in the art will understand that the physical position parameters of the drive roller, including the three-dimensional coordinates of the center, diameter, and axial width, are collected by equipment ledgers, laser scanning, or sensors to construct its spatial geometric model; simultaneously, coordinate calibration is carried out on the transport strain field, and the spatial coordinates of the strain data are aligned with the actual coordinate system on site by using the marking points on the conveyor belt surface or camera extrinsic parameters to establish a spatial position mapping of the strain value;

[0097] The gradient abruptness characteristics of the strain field are analyzed along the direction of conveyor belt movement (the strain is gentle in the straight section before contact, and the strain increases sharply due to compression or friction at the moment of contact). Combined with the starting position of the circumference of the roller geometric model, that is, the theoretical coordinates of the top of the roller, the actual boundary of the starting contact area between the two is located by matching the geometric position and the strain abruptness threshold.

[0098] The method for extracting the linear and nonlinear strain components of the boundary contact region and performing strain analysis to obtain strain analysis characteristics is as follows:

[0099] Based on the initial contact region boundary, a boundary contact region is constructed and the linear and nonlinear strain components of the boundary contact region in the transport strain field are obtained.

[0100] Among them, the linear strain component corresponds to the small deformation scenario, is calculated based on the linear strain formula, and includes the normal strain in the x direction, the normal strain in the y direction, and the shear strain; the nonlinear strain component corresponds to the large deformation scenario, and is calculated based on the geometric nonlinear Green's strain formula.

[0101] It should be further explained that when the conveyor belt enters the roller, it first gradually adheres to the roller in a line contact manner, and there is an incompletely adhered strip of about 5 to 10 mm at the edge. The width may increase as the tension of the conveyor belt changes, and the texture on the back of the incompletely adhered strip area is exposed.

[0102] The physical position parameters include the center position coordinates and diameter of the drive roller;

[0103] Obtain the transport strain field corresponding to the straight section of the coal transport conveyor belt, and calculate the mean value of the linear shear strain in the linear strain component of the straight section;

[0104] Calculate the maximum value of linear shear strain in the boundary contact area, and calculate the deviation ratio between the maximum value and the mean value of linear shear strain in the linear strain components to obtain the shear strain ratio.

[0105] Obtain the area of ​​nonlinear strain in the transport strain field corresponding to the straight section, and the area corresponding to the straight section in the transport strain field;

[0106] The area of ​​the straight section with nonlinear strain is calculated as a percentage of the area of ​​the corresponding straight section in the transport strain field, thus obtaining the area ratio of the straight section.

[0107] The ratio of the nonlinear strain area to the total area of ​​the boundary contact region is obtained to get the boundary area ratio.

[0108] Calculate the rate of change of the boundary area ratio and the straight section area ratio, and sum the rate of change with the shear strain ratio to obtain the strain state coefficient;

[0109] Get the contact width W when the conveyor belt does not fully adhere to the drive roller;

[0110] The strain state coefficient and contact width are used as characteristics for strain analysis.

[0111] It is understandable that the physical meaning of the strain state coefficient is to comprehensively quantify the degree of strain anomaly in the boundary contact area relative to the straight section of the conveyor belt. It is a core indicator reflecting the deviation of the strain state of the contact area from the normal (straight section) level. The shear-strain change ratio reflects the degree of local abrupt change in linear shear strain and also reflects the distribution difference of nonlinear strain (such as large deformations such as wrinkles and yielding) in the contact area. It is used to determine whether the strain state of the contact area exceeds the normal range and provides a quantitative basis for triggering the contact analysis signal.

[0112] Set analysis boundaries; if the strain analysis characteristics exceed the analysis boundaries, a contact analysis signal is triggered.

[0113] Preferably, if the contact width If the width W of the incompletely bonded strip is higher than the critical width, and the strain state coefficient is higher than or equal to the preset state coefficient threshold, then the contact analysis signal is triggered.

[0114] Where T is the tension of the conveyor belt, and k is the contact stiffness correlation coefficient. The coefficient of static friction between the conveyor belt and the drive roller;

[0115] It needs to be explained that when the strain analysis characteristics break through the analysis boundary, the contact analysis signal is triggered because the contact state between the conveyor belt and the drive roller has deviated from the normal range: if the width of the belt that is not fully in contact exceeds the critical value, it indicates that the belt is not in contact enough and may cause slippage; if the strain state coefficient exceeds the threshold, it reflects abnormal strain in the contact area (such as shear strain abrupt change, or excessive proportion of nonlinear strain), which may indicate the risk of local wear or deformation.

[0116] By using quantitative standard trigger signals, potential contact problems can be promptly isolated from the strain field and targeted analyses can be initiated (such as effective contact efficiency assessment and uniformity analysis). This reduces the lag and subjectivity of traditional experience-based judgments and provides trigger nodes for subsequent precise location of contact off-center loading areas, travel deviation areas, and source tracing optimization. It ensures that contact anomalies are captured in a timely manner and enter the full-chain analysis process, thereby ensuring the stable operation of the transportation system.

[0117] Among them, the critical bandwidth and the preset state coefficient threshold are the analysis boundaries;

[0118] T represents the real-time tension of the conveyor belt on the drive roller. and k are the friction coefficient and stiffness coefficient per unit length of the conveyor belt, respectively.

[0119] The technical solution of this embodiment is as follows: Surface texture images of the coal conveyor belt are acquired and pixel displacement gradient analysis is performed. The image displacement is converted into strain distribution to establish a transport strain field. The boundary contact area of ​​the transport strain field is located, and the linear and nonlinear strain components of the boundary contact area are extracted and strain analysis is performed to obtain strain analysis features. If the strain analysis features break through the analysis boundary, a contact analysis signal is triggered. This achieves quantitative judgment and signal triggering of contact anomalies, reducing the missed detections or misjudgments caused by the reliance on experience-based judgment in traditional methods, and improving the timeliness and accuracy of contact status monitoring.

[0120] Example 2

[0121] like Figure 1 As shown, the vision-assisted intelligent coal flow system transportation monitoring method further includes the following steps:

[0122] S3. Based on the contact analysis signal, extract the effective contact area of ​​the boundary contact region and perform contact efficiency analysis. If the contact efficiency is in the low efficiency range, perform homogeneous analysis on the contact area from the axial direction and the direction of movement to determine the contact off-center load area and the stroke deviation area.

[0123] If the contact analysis signal is triggered, the linear shear strain value of each grid cell within the boundary contact area is obtained;

[0124] Calculate the percentage of deviation between the linear shear strain value and the mean value of the linear shear strain for each grid cell to obtain a single variation ratio;

[0125] Obtain the strain type label of the mesh, and determine whether the boundary contact area is in the effective contact area based on the single change ratio and strain type label;

[0126] Understandably, each grid cell is first labeled with a strain type; then the single change ratio is calculated; finally, a preset criterion is used to determine whether the area where the grid cell is located is an effective contact area if the grid cell label is linear strain and the single change ratio is within the normal range, i.e., it does not exceed the preset threshold; if the label is nonlinear strain or the single change ratio exceeds the limit, it is determined to be an invalid contact area, thereby realizing the division of the validity of the boundary contact area.

[0127] Obtain the mesh within the effective contact area, convert the pixel size of the mesh cell to the physical size based on the pixel-physical mapping relationship, and obtain the area of ​​a single mesh based on the physical size;

[0128] Obtain the number of grids within the effective contact area, and combine it with the area of ​​a single effective grid. Multiply the number of grids by the area of ​​a single grid to obtain the effective contact area.

[0129] Obtain the arc length corresponding to the wrap angle of the drive roller and the width of the conveyor belt, and multiply the arc length and the width of the conveyor belt to obtain the theoretical contact area.

[0130] It should be noted that the range of the wrap angle is 0-180°;

[0131] The effective contact area is calculated by comparing the effective contact area with the theoretical contact area.

[0132] The effective contact efficiency is compared with the preset contact efficiency range. If the effective contact efficiency is in the inefficient range, a homogeneous analysis is triggered.

[0133] Understandably, the effective contact efficiency is calculated by the ratio of the effective contact area to the theoretical contact area, and directly reflects the overall contact quality between the conveyor belt and the drive roller: if the efficiency is in the low efficiency range, it indicates that there are defects such as local non-adhesion and uneven contact distribution.

[0134] Triggering homogeneous analysis, in order to further locate the spatial distribution characteristics of defects, the contact off-load area is determined by axial homogeneous analysis, and the stroke deviation area is determined by motion direction homogeneous analysis. The overall macroscopic judgment is transformed into the microscopic location of specific abnormal areas, providing a data foundation for subsequent extraction of contact features and correlation with equipment features for cause tracing. This avoids merely staying at the surface judgment of low efficiency, but deeply analyzes the root cause of inefficiency, thereby providing optimization objects for dynamic adjustment strategies and achieving targeted improvement of contact quality.

[0135] If a distribution homogeneity analysis is triggered, then a homogeneity analysis of the contact area in both the axial and motion directions will be performed.

[0136] The method for performing axial homogeneous analysis on the contact area is as follows:

[0137] Divide the boundary contact area into N equal strips along the drive roller axis, and calculate the effective contact area percentage of each strip. The off-center load strip and the compliant strip are determined based on the contact area ratio;

[0138] Preferably, when the effective contact area accounts for... When the value is less than 50, the strip is marked as an off-center load strip; otherwise, it is marked as a compliant strip.

[0139] Those skilled in the art will understand that the roller axis is perpendicular to the direction of conveyor belt movement, and the effective area ratio of each strip is calculated by the ratio of the effective area within a single strip to the total area of ​​the strip.

[0140] If adjacent eccentric load strips appear, they are merged to obtain an eccentric load strip group.

[0141] If adjacent compliance stripes appear, they are merged to obtain a compliance strip group.

[0142] Obtain the maximum effective contact area ratio within the off-center strip group. and minimum value and mean ;

[0143] Through the formula: Obtain the axial deviation index PI of the off-center load strip group;

[0144] Obtain the mean value of the linear shear strain within the eccentric load strip group to obtain the mean value of the eccentric load group;

[0145] Obtain the mean value of the linear shear strain of adjacent compliant strip groups in the off-center load strip group to obtain the mean value of the compliant group;

[0146] The shear strain difference is obtained by calculating the ratio of the difference between the mean of the off-center load group and the mean of the compliant group.

[0147] The contact off-center loading zone is determined based on the shear strain difference and axial deviation index.

[0148] For example, the contact area between the conveyor belt and the drive roller is divided into grid cells with a contact width of 10 along the axial direction. The grid cells are numbered 1-10, and the central grid cell has a contact width of 5, which is the theoretical center of symmetry.

[0149] Calculate the shear strain difference: The average shear strain of the mesh with a contact width of 3-7 in the normal region is equal to the contact width of 50. The average shear strain of the mesh contact width 1-2 is 90 mm at the contact width. The difference was (90-50) / 50=80%, which far exceeded the preset threshold of 30% for contact width, indicating that the shear strain in this area was abnormally high.

[0150] Calculate the axial deviation index: The center position of the grid contact width 1-2 is 60mm away from the axial symmetry center contact width, while the total axial width of the roller is 200mm. The deviation index is 60 / (200 / 2) = 60%, which exceeds the threshold contact width of 20%, indicating that the deviation from the symmetry center is significant.

[0151] Both exceeded the standard, therefore the axial section where grid 1-2 is located was determined to be the contact off-center load zone, that is, the area where the local shear strain is abnormal and off-center due to the concentration of axial pressure.

[0152] The method for analyzing the homogeneity of the contact area in terms of motion direction is as follows:

[0153] Extract the angle range of the wrap angle along the direction of movement of the drive roller wrap angle;

[0154] Divide the angle range into angle intervals according to a fixed degree, and calculate the rate of change of effective contact area within each angle interval;

[0155] For example, along the direction of the roller wrap angle from 0° to 180°, each 10° wrap angle is divided into an interval, and the effective contact area change rate (ΔS / Δθ) of each interval is calculated. Under normal working conditions, the change rate of the 0°-30° interval should be >0, because the conveyor belt gradually comes into contact with the roller, and the effective contact area increases.

[0156] The change rate in the 30°-150° range is ≤5%, and the effective contact area is stable. The change rate in the 150°-180° range decreases slightly, which is a normal attenuation before separation.

[0157] The travel deviation zone is determined based on the rate of change of contact area;

[0158] For example, suppose the conveyor belt is divided into 5 consecutive monitoring segments (numbered AE, each segment corresponding to 1 second of running time) along the direction of movement (length direction):

[0159] During normal operation, the average contact area of ​​each section is 1000 cm². 2 The rate of change of contact area (the ratio of the area difference between adjacent segments to the area of ​​the previous segment) remains stable within ±3% (e.g., segment B increases by 2% compared to segment A, and segment C decreases by 1% compared to segment B).

[0160] When the contact area of ​​segment D was monitored to drop sharply to 700 cm² 2 The E segment further decreased to 400cm 2The calculated change rate of segment D relative to segment C was (700-1000) / 1000=-30%, and the change rate of segment E relative to segment D was (400-700) / 700≈-42.9%, both of which far exceeded the preset threshold of ±8%.

[0161] Because the contact area in the direction of motion decreases sharply and the rate of change exceeds the limit, the area where sections D and E are located is determined to be the travel deviation area, that is, the abnormal area where the conveyor belt and the roller have unstable contact in the direction of motion.

[0162] It is understandable that the purpose of determining the travel deviation area and the contact off-center load area is:

[0163] Function 1: Provides spatial location basis for tracing the causes of anomalies. By identifying the contact off-center load area and the travel deviation area, the contact characteristics of the corresponding areas can be extracted in a targeted manner, providing specific analysis objects for subsequent establishment of a correlation matrix with equipment characteristics and tracing the causes, avoiding generalized analysis of overall contact anomalies;

[0164] Function 2: Quantifying the distribution characteristics and severity of anomalies. The shear strain difference in the contact off-center loading zone, the axial deviation index, and the contact area change rate in the travel deviation zone can transform the macroscopic judgment of inefficient contact into quantifiable microscopic indicators, clarifying the spatial distribution patterns of axial unilateral off-center loading and localized interruption of contact in the direction of movement, and providing data support for assessing the impact of anomalies on the transportation system;

[0165] Thirdly, it provides targeted optimization targets for dynamic adjustment strategies. After identifying two areas, precise adjustment plans can be formulated based on their causes: for the contact off-center load area, the axial distribution can be improved by fine-tuning the roller level with a servo motor; for the stroke deviation area, the local friction coefficient abnormality can be resolved by linking the drying system or spraying a coating, reducing blind optimization and improving the stability adjustment efficiency of the coal flow transportation system.

[0166] S4. Extract the contact characteristics and equipment characteristics of the contact off-center load area and the travel deviation area, establish an association matrix, and conduct source tracing analysis on the contact off-center load area and the travel deviation area. Based on the source tracing results, conduct a durability status assessment and establish a dynamic adjustment strategy for the coal flow transportation system.

[0167] The method for extracting contact features and equipment features from the contact off-center load area and the stroke deviation area is as follows:

[0168] Extract the axial deviation index, shear strain difference, and contact area change rate of the off-center load zone;

[0169] Axial deviation index, shear strain difference, and contact area change rate are used as contact characteristics;

[0170] The pressure difference between the two sides of the drive roller in the contact off-center loading area and the surface temperature of the drive roller are obtained, and the pressure difference and surface temperature are used as equipment characteristics.

[0171] Integrate contact features and equipment features to establish a feature correlation matrix;

[0172] The method for establishing the correlation matrix and conducting source analysis on the contact off-center load area and the travel deviation area is as follows:

[0173] Feature correlation analysis is performed on the feature correlation matrix to trace the causes of contact off-center load and determine the cause type;

[0174] The method for performing feature correlation analysis on the feature correlation matrix is ​​as follows:

[0175] Calculate the correlation coefficient between the contact off-center loading area and the pressure difference, as well as the correlation coefficient between the shear strain difference and the pressure difference;

[0176] Based on the correlation coefficient between the contact off-center load area and the pressure difference, as well as the correlation coefficient between the shear strain difference and the pressure difference, it is determined whether the contact off-center load area is formed due to the tilt of the drum axis.

[0177] It should be explained that if the relationship is linear, the correlation coefficient is the Pearson correlation coefficient, which is calculated using the Pearson correlation coefficient equation; if the relationship is nonlinear, the Spearman correlation coefficient is used.

[0178] For example, if the correlation coefficient between the contact off-center load area and the pressure difference between the bearings on both sides of the drum is >0.85, and the correlation coefficient R between the shear strain difference and the pressure difference is... 2 >0.9, meaning the shear strain difference increases linearly with the pressure difference, is determined to be caused by the tilt of the drum axis;

[0179] Calculate the correlation coefficient between the area change rate of the stroke deviation zone and the surface temperature of the drive drum, and extract the nonlinear strain abrupt value of the stroke deviation zone;

[0180] Calculate the abrupt change multiple between the nonlinear strain mutation value and the preset strain threshold, and combine the correlation coefficient between the change rate of the stroke deviation area and the surface temperature of the drive roller to determine whether the decrease in friction coefficient is due to local overheating of the conveyor belt.

[0181] For example, if the correlation coefficient between the rate of change of the area of ​​the travel deviation zone and the temperature of the roller surface is >0.7, and the abrupt change value of the nonlinear strain is >3 times that of the normal range, then it is determined that the friction coefficient is reduced due to local overheating of the conveyor belt.

[0182] The method for conducting durability status assessment based on source tracing results and establishing a dynamic adjustment strategy for the coal transport system is as follows:

[0183] The method for conducting durability assessment is as follows:

[0184] The contact off-center loading zone caused by the tilt of the drum axis is obtained, and the peak contact pressure is determined by Hertzian contact theory.

[0185] Preferably, the equations are set based on Hertzian contact theory: Obtain the peak pressure;

[0186] in, The total normal pressure exerted by the conveyor belt on the roller is given by the formula: get, For conveyor belt tension, For the wrap angle of the roller;

[0187] E is the equivalent elastic modulus, derived from the equation: It is derived that, , The elastic modulus and Poisson's ratio of the conveyor belt, , The elastic modulus and Poisson's ratio of the roller;

[0188] b is the half-width of the contact off-center load zone, determined by the formula: We obtain R as the drum radius and L as the contact length of the conveyor belt along the drum axis;

[0189] The peak contact pressure is input into the wear equation to obtain the wear amount;

[0190] Preferably, by formula: Obtain the wear amount S;

[0191] in, This represents the preset correction coefficient, which is used to compensate for the difference between the theoretical model and the actual working conditions. t is the contact time.

[0192] It should be further noted that the correction factor was set by professionals in the field based on experience;

[0193] A wear prediction model is constructed based on the amount of wear to predict the remaining wear life of the conveyor belt in the contact off-center loading area.

[0194] Those skilled in the art will understand that when constructing a wear prediction model, the focus is first on the contact off-center load area to calculate the current wear amount in the area; at the same time, the wear resistance limit is determined according to the conveyor belt design standards; the core of the model obtains the remaining life through the current wear amount and real-time wear rate, and by establishing a quantitative correlation between wear and life, and continuously inputting new pressure peak and t data, the model parameters are iteratively updated (such as pressure peak when off-center load intensifies). Sudden changes will synchronously correct the wear rate and remaining life, thereby realizing the dynamic extrapolation of the remaining wear resistance life of the conveyor belt in the contact off-center load area;

[0195] The methods for conducting situation assessments are as follows:

[0196] The travel deviation area caused by local overheating of the conveyor belt and the resulting decrease in friction coefficient is obtained. At the same time, the contact pressure and the real-time friction coefficient are collected. The contact pressure and the real-time friction coefficient are multiplied to obtain the friction force.

[0197] A driving loss model is established based on friction to quantify the driving force loss rate of the driving roller;

[0198] It should be explained that when establishing a driving loss model based on friction, the contact pressure and real-time friction coefficient in the stroke deviation area are collected first, and the two are multiplied to obtain the actual friction force. Then, the product of the theoretical ideal friction coefficient and the total normal pressure is calculated by combining the design parameters of the driving roller (such as the ideal friction coefficient and the total normal pressure) as the maximum driving force. Finally, the driving force loss rate is quantified by the ratio of the deviation between the theoretical maximum driving force and the actual friction force. The driving loss model is constructed in this way to reflect the degree of driving force loss caused by the decrease in friction coefficient and uneven distribution of contact pressure.

[0199] A dynamic adjustment strategy is established based on the remaining wear resistance life and driving force loss rate to optimize the control of the coal flow transportation system.

[0200] Preferably, the dynamic adjustment strategy is as follows: for the tilt of the roller axis, the height of the bearings at both ends of the roller is finely adjusted by the servo motor. The adjustment amount Δh = PI × 0.1 mm / unit exponent. After each adjustment of 1 mm, the effective contact efficiency in S3 is collected in real time until the efficiency rises to ≥85% and stabilizes for 3 minutes.

[0201] If the sudden drop in friction coefficient is caused by moisture, the hot air drying system will be activated. If it is caused by wear, a wear-resistant coating will be applied locally. The nonlinear strain recovery rate of the stroke deviation area will be monitored simultaneously. Adjustment will be stopped when the recovery rate is greater than 70%.

[0202] The technical solution of this embodiment is as follows: Based on the contact analysis signal, the effective contact area of ​​the boundary contact region is extracted and the contact efficiency is analyzed. If the contact efficiency is in the low efficiency range, the contact area is homogeneously analyzed from the axial direction and the direction of movement to determine the contact off-center load area and the travel deviation area. The contact characteristics and equipment characteristics of the contact off-center load area and the travel deviation area are extracted, an association matrix is ​​established, and the contact off-center load area and the travel deviation area are traced back to their source. Based on the traceability results, the durability situation is assessed, and a dynamic adjustment strategy for the coal flow transportation system is established. This is beneficial to improving the problem that traditional methods can only monitor anomalies but cannot achieve intelligent optimization, thereby improving the stability of the coal flow transportation system and the service life of the equipment.

[0203] Example 3

[0204] like Figure 3As shown, the vision-assisted intelligent coal flow system transportation monitoring system also includes the following modules:

[0205] Strain acquisition module: used to acquire surface texture images of the coal transport conveyor belt and perform pixel displacement gradient analysis, convert image displacement into strain distribution, and establish the transport strain field;

[0206] Contact Analysis Module: Used to locate the boundary contact area of ​​the transport strain field, extract the linear and nonlinear strain components of the boundary contact area and perform strain analysis to obtain strain analysis characteristics. If the strain analysis characteristics break through the analysis boundary, the contact analysis signal is triggered.

[0207] Homogeneous analysis module: Based on the contact analysis signal, extract the effective contact area of ​​the boundary contact region and perform contact efficiency analysis. If the contact efficiency is in the low efficiency range, perform homogeneous analysis on the contact area from the axial direction and the direction of movement to determine the contact off-center load area and the stroke deviation area.

[0208] Situation assessment module: used to extract contact features and equipment features of the contact off-load zone and the travel deviation zone, establish an association matrix and perform source analysis on the contact off-load zone and the travel deviation zone, perform durability situation assessment based on the source analysis results, and establish a dynamic adjustment strategy for the coal flow transportation system.

[0209] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A vision-assisted intelligent coal flow system transportation monitoring method, characterized in that, Includes the following steps: Surface texture images of the coal transport conveyor belt are collected and pixel displacement gradient analysis is performed. The image displacement is converted into strain distribution to establish the transport strain field. Locate the boundary contact area of ​​the transport strain field, extract the linear and nonlinear strain components of the boundary contact area and perform strain analysis to obtain strain analysis characteristics. If the strain analysis characteristics break through the analysis boundary, a contact analysis signal is triggered. Based on the contact analysis signal, the effective contact area of ​​the boundary contact region is extracted and the contact efficiency is analyzed. If the contact efficiency is in the low efficiency range, the contact area is homogeneously analyzed from the axial and motion directions to determine the contact off-center load area and the stroke deviation area. The contact characteristics and equipment characteristics of the contact off-center load area and the travel deviation area are extracted, an association matrix is ​​established, and the source analysis of the contact off-center load area and the travel deviation area is carried out. Based on the source analysis results, the durability status is assessed, and a dynamic adjustment strategy for the coal flow transportation system is established. The method for conducting the aforementioned durability status assessment is as follows: Obtain the source tracing results, extract the contact off-center load area caused by the tilt of the drum axis, and determine the peak contact pressure using Hertzian contact theory; The peak contact pressure is input into the wear equation to obtain the wear amount. Based on the wear amount, a wear prediction model is constructed to predict the remaining wear resistance life of the conveyor belt in the contact off-center load area. The frictional force was obtained and a driving loss model was established to quantify the driving force loss rate of the driving roller. The method for obtaining the source tracing results is as follows: Extract the axial deviation index, shear strain difference, and contact area change rate of the off-center load zone; Axial deviation index, shear strain difference, and contact area change rate are used as contact characteristics; The pressure difference between the two sides of the drive roller in the contact off-center loading area and the surface temperature of the drive roller are obtained, and the pressure difference and surface temperature are used as equipment characteristics. Integrate contact features and equipment features to establish a feature correlation matrix; Feature correlation analysis was performed on the feature correlation matrix to trace the causes of contact off-center load and determine the cause type.

2. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 1, characterized in that, The method for establishing the aforementioned transport strain field is as follows: Obtain the actual displacement gradient and perform linear strain verification to determine whether the displacement gradient satisfies the displacement-linear strain relationship. If it does, directly extract the strain components of the linear strain. If the condition is not met, the strain type will be distinguished, and the displacement gradient of different strain types will be extracted and converted into strain components. The strain components of different strain types are labeled, and the processed strain components are processed by a spatial interpolation algorithm to obtain continuous strain values ​​on the conveyor belt surface. The strain values ​​are correlated with the spatial coordinates of the original conveyor belt image to construct the transport strain field.

3. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 2, characterized in that, The displacement gradient is obtained as follows: Collect surface texture images of the non-coal transport face of the coal transport conveyor belt and extract texture feature points from the images; Calculate the pixel coordinate displacement of texture feature points between adjacent frames of the surface texture image; Obtain the pixel coordinate displacement of all texture feature points, divide the conveyor belt surface into multiple grid cells, calculate the average displacement of texture feature points in each grid cell, and the displacement gradient of the grid cell. Obtain the displacement gradient of all mesh elements and establish the displacement gradient matrix; Construct a pixel-physical mapping relationship to transform the displacement gradient of pixels in the displacement gradient matrix into the actual displacement gradient.

4. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 1, characterized in that, The strain analysis characteristics are obtained as follows: Linear and nonlinear strain components of the boundary contact region are extracted for strain analysis to obtain strain state coefficients and contact width. The strain state coefficients and contact width are used as strain analysis features. An analysis boundary is set. If the strain analysis characteristics exceed the analysis boundary, a contact analysis signal is triggered.

5. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 4, characterized in that, The strain analysis is performed as follows: Obtain the boundary of the initial contact region, construct the boundary contact region, and obtain the linear and nonlinear strain components of the boundary contact region in the transport strain field; Obtain the transport strain field corresponding to the straight section of the coal transport conveyor belt, calculate the mean value of the linear shear strain in the linear strain component of the straight section, and the maximum value of the linear shear strain in the boundary contact area; The shear strain ratio is obtained by calculating the deviation ratio between the maximum value and the mean value of the linear shear strain in the linear strain component of the segment. The ratio of the nonlinear strain area to the total area of ​​the boundary contact region is obtained to get the boundary area ratio. Calculate the rate of change of the boundary area ratio and the straight section area ratio, and sum the rate of change with the shear strain ratio to obtain the strain state coefficient.

6. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 1, characterized in that, The homogeneity analysis was performed as follows: If the effective contact efficiency is in the low efficiency range, a homogeneity analysis is triggered. Axial homogeneity analysis of the contact area was performed to obtain the shear strain difference and axial deviation index to determine the contact off-center load zone; The rate of change of contact area is obtained by homogeneous analysis of the direction of motion, and the travel deviation zone is determined.

7. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 6, characterized in that, The effective contact efficiency is obtained as follows: Obtain the mesh within the effective contact area, convert the pixel size of the mesh cell to the physical size based on the pixel-physical mapping relationship, and obtain the area of ​​a single mesh based on the physical size; Obtain the number of grids within the effective contact area, and combine it with the area of ​​a single effective grid. Multiply the number of grids by the area of ​​a single grid to obtain the effective contact area. Obtain the arc length corresponding to the wrap angle of the drive roller and the width of the conveyor belt, and multiply the arc length and the width of the conveyor belt to obtain the theoretical contact area. The effective contact area is calculated by comparing the effective contact area with the theoretical contact area to obtain the effective contact efficiency.

8. The vision-assisted intelligent coal flow system transportation monitoring method according to claim 6, characterized in that, The method for performing the axial homogeneity analysis is as follows: Divide the boundary contact area into N strips along the drive roller axis, calculate the effective contact area ratio of each strip, and determine the off-center strip and compliant strip based on the contact area ratio. Obtain the maximum and average effective contact area ratio within the off-center strip group, and input the maximum and average values ​​into the axial deviation equation to obtain the axial deviation index. Obtain the mean value of linear shear strain within the off-center load strip group to obtain the mean value of the off-center load group; obtain the mean value of linear shear strain of adjacent compliant strip groups within the off-center load strip group to obtain the mean value of the compliant group. The shear strain difference is obtained by calculating the ratio of the difference between the mean of the off-center load group and the mean of the compliant group.

Images