Intelligent Monitoring Method and System for Geometric Dimensions and Deformations of Rotary Kilns

Abstract

The present invention discloses a method and system for intelligent monitoring of the geometric dimensions and deformation of a rotary kiln, which solves the problem of measuring the geometric dimensions and monitoring the deformation of a rotary kiln under dynamic operating conditions. The present invention includes: collecting point cloud data of the rotary kiln body using a three-dimensional laser scanning device; performing clustering processing on the original point cloud data using an automatic extraction algorithm for the rotary kiln tyre structure to obtain tyre point cloud data of the kiln head, middle gear, and kiln tail; fitting the diameter and center of the tyre using a high-precision fitting algorithm for the rotary kiln tyre parameters to obtain the fitting results and fitting errors of the diameter and center coordinates of the tyre; and implementing automatic calculation of the offset and axis of the rotary kiln using an automatic calculation method for the offset and axis of the rotary kiln, and realizing the deformation monitoring of the rotary kiln structure through the slope and axis offset.

Description

Technical Field

[0001] The invention relates to monitoring of geometrical dimensions and deformation of a rotary kiln. Background Art

[0002] The rotary kiln is a large-scale machine that operates under high temperature (about 1200℃) and rapidly changing mechanical stress conditions. It is the core of cement enterprises and the pillar of industrial production. The calcination of cement clinker in the building materials industry mainly depends on the operation of the rotary kiln. The operating status of the rotary kiln directly affects the output, economic benefits and quality of cement enterprises. When the kiln barrel is bent, it will seriously affect the operation of the rotary kiln and shorten the working life of the rotary kiln. Therefore, the rotary kiln needs to be regularly measured for geometric dimensions and monitored for deformation, so as to make adjustments and maintenance to ensure the safe operation and production of the kiln.

[0003] There are many methods for deformation monitoring of rotary kiln, such as geometric leveling method, side leveling method, fitting and solving space circle and three-dimensional laser scanning method, but they all require relatively complex measurement and tedious manual data processing process, and the degree of automation is not high. Summary of the invention

[0004] The present invention uses a three-dimensional laser scanner to achieve non-contact measurement to solve the problems of high temperature and severe vibration on the rotary kiln working platform; in the process of point cloud data processing, a robust, fast and accurate method for extracting and estimating the wheel belt of the rotary kiln is proposed to solve the problem of a large number of incomplete points in the scanned point cloud due to limited station space, environmental occlusion and mirror reflection of the wheel belt structure. The present invention can generate high-precision, efficient and robust rotary kiln geometric parameters and deformation, and realize the geometric dimension measurement and deformation monitoring analysis of the rotary kiln under dynamic operation.

[0005] In a first aspect, a method for intelligently monitoring the geometric dimensions and deformation of a rotary kiln is provided, comprising: extracting a wheel point cloud from a three-dimensional point cloud of a rotary kiln body; randomly downsampling the wheel point cloud; calculating the normal vector of each point in the downsampled wheel point cloud, performing SVD decomposition on the matrix composed of all normal vectors to obtain the axis direction vector of the wheel point cloud; rotating the axis direction vector to be parallel to the z-axis, and the wheel point cloud data is also rotated; projecting the rotated wheel point cloud onto an xoy plane; fitting a circle using a least squares algorithm for the projected two-dimensional plane points, estimating the coordinates of the center of the circle and the radius / diameter; back-projecting the center of the circle of the two-dimensional plane into a point on the axis of the cylinder, and obtaining the coordinates of the center point of the wheel and the initial values of the radius / diameter; after obtaining the coordinates of the center point of the wheel and the initial values of the radius / diameter, performing parameter optimization on them using a nonlinear least squares algorithm; calculating the wheel offset and the slope between the wheel bands by using the optimized wheel center point coordinates and radius / diameter, and judging whether the rotary kiln is deformed according to the wheel radius / diameter, the wheel offset and the slope between the wheel bands.

[0006] In a second aspect, there is provided an intelligent monitoring system for the geometric dimensions and deformation of a rotary kiln, including: a tyre point cloud data extraction module configured to extract tyre point clouds from the three-dimensional point cloud of the rotary kiln body; a tyre parameter estimation module configured to: perform random downsampling on the tyre point clouds, calculate the normal vector of each point in the downsampled tyre point clouds, perform SVD decomposition on the matrix composed of all the normal vectors to obtain the axis direction vector of the tyre point clouds, rotate the axis direction vector to be parallel to the z-axis, and rotate the tyre point cloud data accordingly, project the rotated tyre point clouds onto the xoy plane, for the two-dimensional plane points after projection, use the least squares algorithm to fit a circle, estimate the center coordinates and radius / diameter, back-project the center of the two-dimensional plane into a point on the cylinder axis, complete the acquisition of the initial values of the tyre center point coordinates, radius / diameter, and after obtaining the initial values of the tyre center point coordinates, radius / diameter, perform parameter optimization on them using the non-linear least squares algorithm; a rotary kiln deformation monitoring module configured to: calculate the tyre offset and the slope between tyres through the optimized tyre center point coordinates, radius / diameter, and determine whether the rotary kiln is deformed according to the tyre radius / diameter, tyre offset, and the slope between tyres.

[0007] In a third aspect, there is provided a computer, including: a processor; a memory including one or more computer program modules; wherein, the one or more computer program modules are stored in the memory and configured to be executed by the processor, and the one or more computer program modules include instructions for implementing the intelligent monitoring method for the geometric dimensions and deformation of the rotary kiln described above.

[0008] In a fourth aspect, there is provided a computer-readable storage medium for storing non-temporary computer-readable instructions, which can implement the intelligent monitoring method for the geometric dimensions and deformation of the rotary kiln when executed by a computer. Description of the Drawings

[0009] Figure 1a and Figure 1b is a schematic diagram of a rotary kiln.

[0010] Figure 2 is a flowchart of the intelligent monitoring method for the geometric dimensions and deformation of a rotary kiln provided by an embodiment of the present invention.

[0011] Figure 3 is a scanned point cloud diagram of a rotary kiln provided by an embodiment of the present invention.

[0012] Figure 4 is a point cloud diagram obtained by performing Euclidean clustering segmentation on the original point cloud data of voxel downsampling provided by an embodiment of the present invention.

[0013] Figure 5It is a point cloud map obtained by performing region growing on the roughly segmented point cloud data provided by an embodiment of the present invention.

[0014] Figure 6 It is a point cloud map of the three-gear tyre structure extracted by an embodiment of the present invention.

[0015] Figure 7 It is a flowchart of Euclidean clustering point cloud segmentation provided by an embodiment of the present invention.

[0016] Figure 8 It is a flowchart of region growing point cloud segmentation provided by an embodiment of the present invention.

[0017] Figure 9 It is a schematic diagram of the cylindrical space equation.

[0018] Figure 10 It is a schematic diagram of the interface of the intelligent monitoring system for the geometric dimensions and deformation of the rotary kiln provided by an embodiment of the present invention. Detailed implementation manners

[0019] The geometric dimensions of the rotary kiln include three aspects. First, the radius / diameter of the tyre is an important geometric dimension of the rotary kiln. By detecting the tyre diameter, it can be judged whether the tyre is expanded or worn. Another geometric dimension of concern is the tyre offset, which is the horizontal and vertical distance offset of the center point of the second (middle) gear tyre from the center line connecting the first and third gear tyres. The purpose is to detect whether the third gear tyres are collinear. If they are not collinear, it will lead to increased loss of the rotary kiln. The third geometric dimension is the slope between the tyres. By connecting the centers of each gear, the slope between each gear can be judged. If it differs greatly from the design value, it will also cause loss. Therefore, it is necessary to regularly monitor these three geometric dimensions and adjust the rotary kiln tyres according to the monitoring results to ensure normal operation.

[0020] Figure 1a 、 1b The rotary kiln is shown. The green line in the figure is the connection line between the centers of the first and third gear tyres; the red dotted lines are the connection lines between the centers of the first - second gears and the second - third gears; the red solid line is the distance from the center point of the second gear to the connection line between the first and third gear centers ( Figure 1a is the horizontal distance, Figure 1b is the vertical distance).

[0021] Figure 2 It shows a flowchart of a method for intelligent monitoring of the geometric dimensions and deformation of a rotary kiln. The method includes point cloud extraction of the rotary kiln tyres, high-precision parameter estimation of the rotary kiln tyres, and calculation of the axis offset and slope of the rotary kiln. The method shown in Figure 1 will be described in detail below.

[0022] Step 1: Use a terrestrial three-dimensional laser scanner to obtain the three-dimensional point cloud data of the rotary kiln shell. Observe the rotary kiln shell and its surrounding environment on-site, and select appropriate measurement points for scanning according to the coverage range of the three-dimensional laser scanner; align the scanner with the rotary kiln shell to obtain three-dimensional point cloud data in the horizontal direction of 0 - 360° and the vertical direction of 0 - 90°. The scanning point cloud map of the rotary kiln can be referred to Figure 3 .

[0023] Step 2: Use the automatic extraction algorithm for the rotary kiln tyre structure to process the point cloud data and extract the point cloud data of the tyre.

[0024] Step 2.1: Perform voxel downsampling preprocessing on the original point cloud data collected by the three-dimensional laser scanner to prepare for subsequent point cloud clustering.

[0025] Step 2.2: For the preprocessed point cloud data, judge the distance between the nearest neighbor points and the voxelization distance. If the distance between the nearest neighbor points is greater than twice the voxelization point spacing, then this point is considered to belong to a new cluster, thereby realizing rough segmentation by Euclidean clustering. Figure 4 is the point cloud map after Euclidean clustering segmentation. Figure 7 shows the Euclidean clustering point cloud segmentation process. After the segmentation is completed, clusters with less than 100 cluster points are considered as occlusions or noise points and are removed.

[0026] Step 2.3: Perform region growing on the roughly segmented point cloud data to achieve fine segmentation of the point cloud data and obtain N groups of clustered point cloud data. Figure 5 shows the point cloud after region growing segmentation. Figure 8 shows the region growing point cloud segmentation process.

[0027] Step 2.4: Traverse the N groups of clustered point clouds obtained by segmentation, and use cylinder fitting based on the RANSAC algorithm (the tyre is a cylindrical object). Calculate the sum of the squared residuals RSS of the clustered points to the cylinder according to the general expression of the cylinder equation and the geometric relationship of the cylinder. Compare the size of RSS with the threshold (usually taken as 100) to judge whether the clustered point cloud belongs to the tyre point cloud. If the RSS value is less than 100, then this part of the cluster is considered to belong to the tyre point cloud;

[0028]

[0029] In the formula, {x0, y0, z0, a, b, c, r0} are the seven parameters of the cylinder fitted by the RANSAC method, (x0, y0, z0) is a point on the axis of the cylinder, (a, b, c) is the axis direction vector, r0 is the radius of the fitted cylinder, (x i , y i , z i) is the clustering point. According to these 7 parameters, a cylindrical surface can be determined. In the parameters, the radius is the first geometric dimension of the rotary kiln - the radius / diameter of the tyre; a point on the axis and the axial vector can be used to calculate the tyre offset and the slope between tyres.

[0030] Step 2.5: Calculate the central point coordinates of the point cloud data belonging to the tyre clustering, and determine whether the central point distance between the central points of each pair of tyre clusterings is less than the threshold (usually taken as 10 cm). If it is satisfied, it is considered that these two types of clusterings belong to the same tyre, and the point cloud data of the same gear tyre is merged to obtain the Figure 6 point cloud data of the three gears of the tyre at the kiln head, middle gear, and kiln tail as shown.

[0031] The present invention realizes the automatic extraction of the tyre point cloud data through the method of Euclidean clustering combined with region growing, saving manpower and improving the efficiency of tyre extraction at the same time.

[0032] The following Figure 9 makes a brief introduction to the cylindrical space equation. As Figure 9 shown, P(x, y, z) is a point on the cylindrical surface, P0(x0, y0, z0) is a point on the central axis of the cylinder, the vector (a, b, c) is the central axis vector of the cylinder, and r0 is the radius of the cylinder. Through the formula of vector dot product, it can be obtained that

[0033]

[0034] According to the Pythagorean theorem, it can be obtained that:

[0035]

[0036] Combining the two equations can obtain the cylindrical space equation:

[0037]

[0038] Step 3: Use the high-precision fitting algorithm for the parameters of the rotary kiln tyre to fit the diameter and center of the tyre, and obtain the fitting results and fitting errors of the diameter and center of the tyre. This algorithm has the advantages of high precision, high robustness, high efficiency, etc.

[0039] Step 3.1: For the extracted tyre point cloud data, use the fast estimation algorithm of cylindrical parameters based on the projection of the cylindrical axis to estimate the cylindrical surface parameters. The specific operations include:

[0040] 1) Randomly downsample the tyre point cloud data, and set the number of output point clouds cloud_filter to N.

[0041] 2) Estimate the normal vector for each point in cloud_filter, generate the normal vector matrix, perform SVD decomposition, and obtain the axis direction vector of the tyre point cloud.

[0042]

[0043] Wherein, U is the left singular vector matrix, Σ is a non - negative real - valued diagonal matrix, and V T is the right singular matrix. The elements on the diagonal of Σ are the singular values of matrix A. The column vector of V corresponding to the minimum value among them is the axis - direction vector (a, b, c) of the point cloud cloud_filter.

[0044] 3) Rotate the axis - direction vector (a, b, c) to be parallel to the z - axis. Derive the rotation matrix T by calculating the quaternion between the axis (a, b, c) and the z - axis, and rotate the point cloud cloud_filter according to the rotation matrix to obtain the point cloud cloud_trans.

[0045] 4) Project the rotated point cloud onto the xoy plane to obtain the projected two - dimensional plane points Point_2D;

[0046] 5) For the projected two - dimensional plane points, use the following function model to fit a circle by the least - squares algorithm, and estimate the center coordinates and radius;

[0047]

[0048] Wherein, (x i , y i ) are known points, and x circle , y circle , r0 are the center coordinate and radius parameters of the fitting circle to be estimated.

[0049] 6) Transform the center coordinates of the two - dimensional plane through the inverse matrix of the rotation matrix T to obtain a point on the axis of the point cloud cloud_filter, and complete the acquisition of the initial values of the cylinder parameters.

[0050] Step 3.2: After obtaining the initial values of the cylinder parameters, it is necessary to optimize the parameters using the non - linear least - squares algorithm. The specific operations include:

[0051] 1) Construct a function model with the difference between the distance from a spatial point to the cylinder axis and the cylinder radius as the object;

[0052]

[0053] According to the square - difference formula The above formula can be written as:

[0054]

[0055] After merging and reorganizing, the following expression form can be obtained:

[0056]

[0057] In the formula, {x0, y0, z0, a, b, c, r0} are unknown parameters, and (x, y, z) are the coordinates of spatial points. Since there is a correlation among the unknown parameters: a 2 + b 2 + c 2 = 1, and (x0, y0, z0) can be any point on the axis. To simplify the solution process, the value of z0 on the axis is set to be unchanged, and the axial vector is represented by the spherical coordinate system. Let:

[0058]

[0059] Therefore, the function model of cylindrical fitting can be written as:

[0060]

[0061] Among them, (x i , y i , z i ) is the i-th known observation value in the point cloud data, and X T = [α, β, x0, y0, r0] are unknown parameters.

[0062] 2) Perform Taylor expansion on the function model to linearize it and obtain the error equation:

[0063]

[0064] 3) Perform iterative calculation on the error equation by combining the least squares algorithm with weight selection iteration. When the parameters meet the tolerance limit or the number of iterations reaches the upper limit, the iteration ends, and the optimization of the cylindrical parameters is completed to realize the geometric dimension measurement of the key structure of the rotary kiln;

[0065]

[0066] In the formula, w is the IGGⅢ weight function, v is the residual of the observation value obtained by the least squares calculation, and σ is the mean square error of unit weight. According to the literature, the constant values k0 = 2.5 and k1 = 3.5 are taken.

[0067] Step 4: The structural deformation of the rotary kiln will cause the change of the slope between the tyre rings and the offset of the axis of the tyre rings of the rotary kiln. The automatic calculation method of the offset and axis of the rotary kiln is adopted to realize the automatic calculation of the mid-range offset and slope of the rotary kiln.

[0068] Step 4.1: Project the point cloud data of each tyre ring onto the axis of the cylinder, and the center point coordinates of this tyre ring are the mean value of the coordinates of all projection points;

[0069] Step 4.2: Calculate the offset between the middle gear and the axial lines of the head and tail wheels of the kiln and the slope k between the three-gear tyre based on the center point coordinates of the tyre, the offset formula, and the slope formula, and realize the deformation monitoring of the rotary kiln structure through the slope and axial offset;

[0070]

[0071] Wherein, Δx, Δy, and Δz are the coordinate differences between the center points of two-gear tyres, and k is the slope of the rotary kiln to be determined.

[0072] The present invention determines the center point coordinates of the cylinder through the spatial geometric relationship, completes the calculation of the slope and offset, and simply and quickly realizes the automatic calculation of the deformation amount of the rotary kiln.

[0073] The tyre diameter size accuracy automatically solved by the present invention is better than 1.4 cm, and the axial offset accuracy is better than 4 mm, which can provide stable and reliable data for the adjustment process of the rotary kiln and ensure the safe operation and production of the kiln body.

[0074] The present invention also provides an embodiment of an intelligent monitoring system for the geometric dimensions and deformation monitoring of a rotary kiln. The system includes: a tyre point cloud data extraction module, a tyre parameter estimation module, and a rotary kiln deformation monitoring module.

[0075] The tyre point cloud data extraction module is configured to extract tyre point clouds from the three-dimensional point cloud of the rotary kiln body.

[0076] The tyre parameter estimation module is configured to perform random downsampling on the tyre point cloud, calculate the normal vector of each point in the downsampled tyre point cloud, perform SVD decomposition on the matrix composed of all normal vectors to obtain the axial direction vector of the tyre point cloud, rotate the axial direction vector to be parallel to the z-axis, and the tyre point cloud data rotates accordingly. Project the rotated tyre point cloud onto the xoy plane, use the least squares algorithm to fit a circle for the projected two-dimensional plane points, estimate the center coordinates and radius / diameter, back-project the center of the two-dimensional plane into a point on the cylinder axis, complete the acquisition of the initial values of the tyre center point coordinates, radius / diameter. After obtaining the initial values of the tyre center point coordinates, radius / diameter, use the nonlinear least squares algorithm to optimize the parameters.

[0077] Through the combined weighted iterative least squares algorithm, iterative calculation is performed on the linearized cylinder fitting function model. When the parameters meet the tolerance or the number of iterations reaches the upper limit, the iteration ends, and the optimization of the cylinder parameters is completed. The cylinder fitting function model is constructed based on the difference between the distance from a spatial point to the cylinder axis and the cylinder radius. The Taylor expansion of the cylinder fitting function model is used to achieve linearization.

[0078] The rotary kiln deformation monitoring module is configured to calculate the offset and slope between tyre rings based on the optimized centre point coordinates and radius / diameter of the tyre rings, and determine whether the rotary kiln is deformed according to the tyre ring radius / diameter, tyre ring offset and slope between tyre rings.

[0079] For a more detailed implementation method of the point cloud data extraction module, tyre ring parameter estimation module and rotary kiln deformation monitoring module, refer to steps 2, 3 and 4. Since this system is a product corresponding to the described method.

[0080] The above only introduces the most important modules of the described system. Figure 10 The interface of this system is shown. There is a button on the interface for importing the original point cloud data. After clicking the button, select the point cloud file to import. The parameters for RANSAC cutting of the tyre rings can be set on the interface. For example, the normal weight, the maximum number of iterations, the distance threshold (the distance threshold from the points of the point cloud to the fitting model, within this distance, the points of the point cloud are considered to be on the cylinder), and the kiln radius. Set the cutting parameters of the tyre rings in the input box on the interface, click the cutting button to automatically cut the tyre rings, and automatically save the point cloud data of the cut tyre rings to the local.

[0081] There is a button on the interface for importing the point cloud data of the cut tyre rings. Click the button to import the point cloud data of the cut tyre rings, automatically fit the tyre rings, and display the fitting results and errors of the tyre rings in the output box.

[0082] There is a calculation button on the interface. Click the calculation button to automatically calculate the offset and slope results of the tyre rings.

[0083] The present invention realizes the geometric dimension measurement and deformation monitoring and analysis of a rotary kiln in a dynamic operating state, with intelligent processing throughout the process, and generates high-precision, efficient and robust geometric parameters and deformation amounts, and has been successfully applied in many places across the country.

[0084] The present invention also provides an embodiment of a computer. The computer includes a processor and a memory. The memory is used to store non-transitory computer-readable instructions (such as one or more computer program modules). The processor is used to run the non-transitory computer-readable instructions. When the non-transitory computer-readable instructions are run by the processor, one or more steps in the intelligent monitoring method for the geometric dimensions and deformation monitoring of the rotary kiln described above can be executed. The memory and the processor can be interconnected through a bus system and / or other forms of connection mechanisms.

[0085] For example, the processor can be a central processing unit (CPU), a graphics processing unit (GPU), or other forms of processing units with data processing capabilities and / or program execution capabilities. For example, the central processing unit (CPU) can be of the X86 or ARM architecture, etc. The processor can be a general-purpose processor or a dedicated processor, and can control other components in the computer to execute the desired functions.

[0086] For example, the memory can include any combination of one or more computer program products, and the computer program products can include various forms of computer-readable storage media, such as volatile memory and / or non-volatile memory. Volatile memory can include, for example, random access memory (RAM) and / or cache memory, etc. Non-volatile memory can include, for example, read-only memory (ROM), hard disk, erasable programmable read-only memory (EPROM), compact disc read-only memory (CD-ROM), USB memory, flash memory, etc. One or more computer program modules can be stored on the computer-readable storage media, and the processor can run one or more computer program modules to implement various functions of the computer.

[0087] The present invention also provides an embodiment of a computer-readable storage medium for storing non-transitory computer-readable instructions, which can implement one or more steps in the above intelligent monitoring method for geometric dimensions and deformation monitoring of a rotary kiln when executed by a computer. That is, when the intelligent monitoring method and system for geometric dimensions and deformation monitoring of a rotary kiln provided by the embodiments of the present application are implemented in software and sold or used as an independent product, they can be stored in a computer-readable storage medium. For the relevant description of the storage medium, reference can be made to the corresponding description of the memory in the computer above, and details will not be repeated here.

Claims

1. An intelligent monitoring method for the geometric dimensions and deformation of a rotary kiln, characterized in that, Including: (1) Extraction of tyre point cloud Extract the tyre point cloud from the 3D point cloud of the rotary kiln body, which includes: Perform voxel downsampling on the original point cloud data collected by the 3D laser scanner; Perform Euclidean clustering algorithm on the downsampled point cloud data for rough segmentation; Apply the region growing algorithm on the basis of rough segmentation to achieve fine segmentation of the point cloud data; For each group of clustered point clouds after fine segmentation, a cylinder model fitting based on the RANSAC algorithm is adopted to obtain cylinder parameters, where the cylinder parameters include the axis direction vector , a point on the axis , and the radius ; Calculate each point according to the following formula RSS to the fitted cylinder: Determine the points with RSS value less than the set threshold as the tyre point cloud; Calculate the clustering center point coordinates of the tyre point cloud. If the distance between two clustering center points is less than the threshold, it is determined that they belong to the same tyre gear, and finally merged into three gears of tyre point cloud at the kiln head, middle gear, and kiln tail; (2) Obtaining initial values of tyre parameters Perform random downsampling on the identified tyre point cloud; Calculate the normal vector of each point in the downsampled tyre point cloud, and perform SVD decomposition on the matrix composed of all normal vectors to obtain the axis direction vector of the tyre point cloud; Rotate the axis direction vector to be parallel to the z axis, and synchronously rotate the tyre point cloud; project the rotated tyre point cloud onto the xoy plane, use the least squares algorithm to fit a circle, and estimate the center coordinates and radius / diameter; Back-project the center of the two-dimensional plane into a point on the cylinder axis to complete the acquisition of the initial value parameters of the tyre center point coordinates, radius / diameter; (3) Parameter optimization After obtaining the initial values of the tyre center point coordinates, radius / diameter, perform parameter optimization based on the non-linear least squares algorithm, including: Establish an objective function model: In the formula, represents the parameter of the cylinder, is the direction vector of the cylinder axis, is a point on the cylinder axis, is the radius of the cylinder, is the observation point in the point cloud data; The function model of cylindrical fitting in the spherical coordinate system is written as: The objective function model is linearized by Taylor expansion to obtain the error equation: Among them, is the i th known observation value in the point cloud data, is the cylinder parameter to be estimated, , is the spherical coordinate system representation of the axis direction vector; Combine the least squares algorithm with weight selection and iteration to perform iterative calculation on the error equation. When the parameters meet the limit error or the number of iterations reaches the upper limit, end the iteration and complete the optimization of the cylinder parameters; (4) Deformation judgment Calculate the tyre offset and the slope between tyres through the optimized tyre center point coordinates, radius / diameter; Judge whether the rotary kiln is deformed according to the tyre radius / diameter, tyre offset and the slope between tyres.

2. An intelligent monitoring system for the geometric dimensions and deformation of a rotary kiln, characterized in that, Including: Tyre point cloud data extraction module, which is configured to extract the tyre point cloud from the 3D point cloud of the rotary kiln body, which includes: perform voxel downsampling on the original point cloud data collected by the 3D laser scanner; Perform Euclidean clustering algorithm on the downsampled point cloud data for rough segmentation; Apply the region growing algorithm on the basis of rough segmentation to achieve fine segmentation of the point cloud data; For each group of clustered point clouds after fine segmentation, a cylinder model fitting based on the RANSAC algorithm is adopted to obtain cylinder parameters, where the cylinder parameters include an axis direction vector , a point on the axis , and a radius ; Calculate each point according to the following formula RSS to the fitted cylinder: Determine the points with RSS value less than the set threshold as the tyre point cloud; Calculate the clustering center point coordinates of the tyre point cloud. If the distance between two clustering center points is less than the threshold, it is determined that they belong to the same tyre gear, and finally merged into three gears of tyre point cloud at the kiln head, middle gear, and kiln tail; The tyre ring parameter estimation module is configured to: perform random downsampling on the identified tyre ring point cloud, calculate the normal vector of each point in the downsampled tyre ring point cloud, perform SVD decomposition on the matrix composed of all the normal vectors to obtain the axis direction vector of the tyre ring point cloud, rotate the axis direction vector to be parallel to the z axis, and synchronously rotate the tyre ring point cloud, project the rotated tyre ring point cloud onto the xoy plane, use the least squares algorithm to fit a circle, estimate the center coordinates and radius / diameter, back-project the center of the two-dimensional plane into a point on the cylinder axis, complete the acquisition of the initial values of the tyre ring center point coordinates, radius / diameter. After obtaining the initial values of the tyre ring center point coordinates, radius / diameter, perform parameter optimization based on the non-linear least squares algorithm, including: Establish an objective function model: In the formula, represents the parameter of the cylinder, is the direction vector of the cylinder axis, is a point on the cylinder axis, is the radius of the cylinder, is the observation point in the point cloud data; The function model of cylindrical fitting in the spherical coordinate system is written as: Linearize the objective function model through Taylor expansion to obtain the error equation: Among them, is the i th known observation value in the point cloud data, is the cylinder parameter to be estimated, , are the spherical coordinate representations of the axis direction vector; Combine the least squares algorithm with weight selection and iteration to perform iterative calculation on the error equation. When the parameters meet the limit error or the number of iterations reaches the upper limit, end the iteration and complete the optimization of the cylinder parameters; Rotary kiln deformation monitoring module, which is configured to: calculate the tyre offset and the slope between tyres through the optimized tyre center point coordinates, radius / diameter, and judge whether the rotary kiln is deformed according to the tyre radius / diameter, tyre offset and the slope between tyres.

3. A computer, characterized in that, Including: Processor; Memory, including one or more computer program modules; Wherein, the one or more computer program modules are stored in the memory and configured to be executed by the processor, and the one or more computer program modules include instructions for implementing the intelligent monitoring method for the geometric dimensions and deformation of the rotary kiln according to claim 1.

4. A computer-readable storage medium for storing non-transitory computer-readable instructions, characterized in that, When the non-transitory computer-readable instructions are executed by a computer, the intelligent monitoring method for the geometric dimensions and deformation of the rotary kiln according to claim 1 can be implemented.

Images