Stay wire connector machining method based on automation

Abstract

The invention discloses a stay wire connector processing method based on automation, and relates to the field of automation control, and the method comprises the steps: obtaining the point cloud data of a wedge-shaped connector physical model; real-time data in the machining process of the wedge-shaped connector is collected through a sensor; preprocessing the point cloud data; reconstructing a three-dimensional model of the wedge-shaped joint by using the processed point cloud data, and extracting geometric parameters of the wedge-shaped joint from the three-dimensional model; according to the collected real-time data, a physical model of the wedge-shaped connector machining process is established; according to the three-dimensional model, a tool machining motion path is generated through computer aided manufacturing software; a self-adaptive particle swarm optimization algorithm is adopted to optimize a tool machining motion path; and the optimized tool machining motion path is converted into a G code recognized by a numerical control machine tool, and the numerical control machine is controlled to machine the wedge-shaped connector. The machining precision of the wedge-shaped connector is improved by aiming at the low machining precision of the wedge-shaped connector in the prior art.

Description

Technical Field

[0001] The present application relates to the field of automation control, and in particular to a wire joint processing method based on automation. Background Art

[0002] In the field of construction engineering, steel bars are one of the most important load-bearing components in concrete structures, and the quality of their connections directly affects the overall performance and safety of the building. Wedge joints have been widely used in steel bar connections due to their advantages such as high bearing capacity, steel saving, and convenient construction. However, the processing quality of wedge joints has always been a difficult problem in engineering practice.

[0003] At present, the processing of steel bar wedge joints mainly adopts mechanical cutting methods, such as turning and milling. These methods have the following shortcomings: first, the processing equipment has poor versatility and is difficult to adapt to the processing needs of wedge joints of different specifications and shapes; second, the processing efficiency of mechanical cutting methods is limited. The above problems limit the application scope and reliability of wedge joints in steel bar connections.

[0004] In order to improve the processing level of steel bar wedge joints, a lot of research and practice have been carried out at home and abroad. Some scholars have proposed methods such as CNC milling and electrospark machining. By introducing CNC technology, the degree of automation and processing accuracy of wedge joints have been improved. Other scholars have optimized the tool structure and cutting parameters based on the complex geometric features of the wedge surface, and improved the surface quality of the wedge surface. Other scholars have explored the application of special processing methods such as laser processing and electroforming in the manufacture of wedge joints, broadening the range of processing technology options.

[0005] Although the above research has improved the processing problem of steel bar wedge joints to a certain extent, there are still the following shortcomings: First, the existing methods lack an accurate description of the geometric shape of the wedge joints, making it difficult to generate a highly matched CNC machining program; second, the existing methods ignore the complex dynamic characteristics of the tool and the workpiece during the cutting process, and the stability and reliability of the CNC machining process need to be improved; third, there is a lack of special CNC equipment and processes for steel bar wedge joints, and most of them rely on the transformation of general CNC machine tools, which increases the processing cost and the difficulty of process implementation. Summary of the invention

[0006] In response to the problem of low processing precision of wedge joints in the prior art, the present application provides an automated wire joint processing method, which reconstructs the three-dimensional model of the wedge joint and establishes a physical model of cutting joint processing including cutting force and vibration. On this basis, CAM software and optimization algorithms are used to generate optimized tool motion paths, thereby improving the processing precision of the wedge joint.

[0007] The purpose of this application is achieved through the following technical solutions.

[0008] The present application provides a wire joint processing method based on automation, comprising: scanning a physical model of a wedge joint using a three-dimensional scanner to obtain point cloud data of the physical model of the wedge joint; collecting real-time data during the processing of the wedge joint through a sensor, the real-time data including tool position, cutting force and vibration parameters; preprocessing the point cloud data, the preprocessing including filtering processing using a Gaussian filtering algorithm and denoising processing using a wavelet threshold denoising algorithm; reconstructing a three-dimensional model of the wedge joint using the processed point cloud data, and extracting geometric parameters of the wedge joint from the three-dimensional model, the geometric parameters including wedge angle, length and width; establishing a physical model of the wedge joint processing process based on the collected real-time data, the physical model including a cutting force model and a vibration model; generating a tool processing motion path based on the three-dimensional model using computer-aided manufacturing software; optimizing the tool processing motion path using an adaptive particle swarm optimization algorithm; wherein the adaptive particle swarm optimization algorithm optimizes the processing quality by adjusting the speed and position of the particles in combination with the established physical model; converting the optimized tool processing motion path into a G code recognized by a CNC machine tool, and controlling the CNC machine to process the wedge joint.

[0009] Furthermore, a Gaussian filtering algorithm is used for filtering processing, including: dividing the acquired point cloud data into different areas according to the geometric features of the wedge joint; adjusting the parameters of the Gaussian filter according to the curvature change of the divided area, the parameters including the size and standard deviation of the filter kernel; using the adjusted parameters, performing Gaussian filtering processing on each divided area to obtain filtered point cloud data; splicing the point cloud data of each area after filtering processing to obtain the point cloud data after filtering of the wedge joint.

[0010] Furthermore, a wavelet threshold denoising algorithm is used for denoising, including: performing wavelet decomposition on the spliced ​​point cloud data to obtain wavelet coefficients at different scales; dividing the decomposed wavelet coefficients into different regions according to the geometric characteristics of the wedge joint; calculating the variance of the wavelet coefficients of each region, comparing the variance with the preset noise threshold, and using the Bayes threshold to adjust the wavelet threshold of the corresponding region; using the adjusted threshold to perform threshold processing on the wavelet coefficients of each divided region; performing wavelet reconstruction on the denoised wavelet coefficients to obtain denoised point cloud data.

[0011] Furthermore, the geometric parameters of the wedge joint are extracted from the three-dimensional model, including: using an octree-based Poisson reconstruction algorithm to perform three-dimensional reconstruction on the preprocessed point cloud data, and obtaining a three-dimensional mesh model of the wedge joint by adjusting the division depth of the octree and the number of sampling points of each node; using a curvature-based mesh simplification algorithm to simplify the three-dimensional mesh model; using a Laplacian smoothing algorithm to smooth the simplified mesh model; and using a triangulation method and a least squares method to extract the wedge angle, length and width of the wedge joint from the simplified and smoothed three-dimensional model of the wedge joint.

[0012] Furthermore, a physical model of the wedge joint processing process is established, including: based on the collected tool position data, combined with the reconstructed three-dimensional model of the wedge joint, by establishing a transformation matrix between the tool coordinate system and the workpiece coordinate system, and using an analytical method to establish a contact geometry relationship model between the tool and the workpiece; the contact geometry relationship model describes the contact area and contact mode of the tool with the wedge joint during the processing; based on the collected cutting force data, combined with the contact geometry relationship model, by establishing microelements on the contact area and analyzing the force balance during the cutting process of a single tooth, a cutting force physical model is established using the Newton-Euler equation; the cutting force physical model includes tangential force, radial force and axial force; based on the collected vibration parameters, by modeling the forced vibration of the tool and the workpiece, a vibration physical model of the frequency characteristics of the tool and the workpiece is established using the transfer function method; the vibration physical model includes the relative displacement and velocity of the tool and the workpiece during the cutting process; the established cutting force physical model and vibration physical model are coupled to establish a dynamic equation between cutting force and vibration, and the physical model of the wedge joint processing process is obtained.

[0013] Furthermore, a contact geometry relationship model between the tool and the workpiece is established, including: establishing a tool coordinate system according to the geometric parameters and installation position of the tool, the origin of the tool coordinate system is located at the rotation center of the tool, and the direction of the coordinate axis is consistent with the rotation axis and radial direction of the tool; establishing a workpiece coordinate system according to the position and direction of the three-dimensional model of the wedge joint, the origin of the workpiece coordinate system is located at the geometric center of the wedge joint, and the direction of the coordinate axis is consistent with the length, width and height directions of the wedge joint; establishing a transformation matrix from the tool coordinate system to the workpiece coordinate system according to the relative position relationship between the tool coordinate system and the workpiece coordinate system, the transformation matrix includes a translation matrix and a rotation matrix; using an analytical method to calculate the spatial position of the tool in the workpiece coordinate system, obtain the contact point coordinates and normal vector between the tool and the workpiece, and establish a contact geometry relationship model between the tool and the workpiece.

[0014] Furthermore, a physical model of cutting force is established, including: establishing microelements along the cutting direction of the tool on the contact area according to the contact point position determined by the contact geometry relationship model, each microelement corresponding to a time step in the cutting process of a single tooth; establishing a microelement coordinate system, the origin of the microelement coordinate system is located at the geometric center of the microelement, and the direction of the coordinate axis is consistent with the cutting speed, radial feed and axial feed directions; establishing a cutting thickness and cutting width model on the microelement, analyzing the stress state in the cutting deformation zone based on the cutting mechanism, and calculating the tangential force, radial force and axial force on each microelement; accumulating the cutting forces on each microelement to obtain the change curve of the tangential force, radial force and axial force of the entire tool during the cutting process with time, and establishing a physical model of cutting force.

[0015] Furthermore, a vibration physical model of the frequency characteristics of the tool and the workpiece is established, including: simplifying the tool into a concentrated mass and torsional vibration system, establishing a vibration equation of the tool, and the vibration equation includes the mass, stiffness and damping coefficient of the tool; simplifying the workpiece into a uniform Euler beam model, and establishing a vibration equation of the workpiece in combination with the lateral vibration and torsional vibration of the workpiece under the action of the cutting force; setting the cutting force as an excitation term in the vibration equations of the tool and the workpiece, using the transfer function method to solve the vibration equation, and obtaining the relative displacement and velocity frequency response function of the tool and the workpiece during the cutting process; according to the frequency response function, calculating the relative displacement and velocity of the tool and the workpiece under different rotation speeds and feed conditions, and obtaining the vibration physical model of the frequency characteristics of the tool and the workpiece.

[0016] Furthermore, a physical model of the wedge joint processing process is obtained, including: setting the relative displacement and speed data of the tool and the workpiece output by the vibration physical model in the cutting force physical model, and establishing a cutting force dynamics equation, which reflects the influence of the vibration displacement and speed on the cutting force; setting the tangential force, radial force and axial force data output by the cutting force physical model in the vibration physical model as excitation terms of the vibration equation, and establishing a vibration dynamics equation, which reflects the influence of the cutting force on the vibration of the tool and the workpiece; coupling the cutting force dynamics equation and the vibration dynamics equation through mutual input and output to construct a coupled dynamics equation group, which includes cutting force and vibration and reflects the interaction between cutting force and vibration; using a numerical integration method to solve the coupled dynamics equation group, obtaining cutting force and vibration data at different times, and constructing a physical model of the wedge joint processing process.

[0017] Furthermore, a tool processing motion path is generated, including: importing the three-dimensional model of the wedge joint into the computer-aided manufacturing software, defining the tool type, tool geometric parameters and cutting parameters according to the processing requirements of the wedge joint; wherein the tool geometric parameters include tool diameter, helix angle and number of blades, and the cutting parameters include spindle speed, feed speed and cutting depth; in the computer-aided manufacturing software, the extracted wedge joint geometric parameters are used as input parameters, and the inner contour curve and outer contour curve of the wedge joint are generated by parametric programming to construct complete wedge joint processing contour data; according to the set tool parameters and the generated processing contour data, a tool position file generation algorithm is used to calculate the tool position during processing The spatial position coordinates on the contour and the tool axis direction vector are used to generate the initial tool motion trajectory containing the tool tip point coordinate sequence data and the tool axis vector sequence data; the cutting force data output by the cutting force physical model is used as the cutting force constraint condition to perform the cutting force constraint check on the initial tool motion trajectory, and the tool motion trajectory under the cutting force constraint is obtained by adjusting the feed speed and spindle speed on the initial tool motion trajectory; the vibration displacement and speed data output by the vibration physical model are used as vibration constraint conditions to perform the vibration constraint check on the corrected tool motion trajectory data, and the tool motion trajectory that meets the cutting force and vibration constraints is obtained by adjusting the cutting depth and adjacent trajectory spacing in the tool motion trajectory data.

[0018] Compared with the prior art, the advantages of this application are:

[0019] High-precision point cloud data of the physical model of the wedge joint is obtained through 3D scanning, and Gaussian filtering and wavelet threshold denoising algorithm are used to pre-process the point cloud data to eliminate noise and discrete points in the point cloud data, thereby improving the quality and accuracy of the point cloud data.

[0020] The octree Poisson reconstruction algorithm, mesh simplification algorithm and smoothing algorithm are used to reconstruct the 3D model of the wedge joint from the preprocessed high-precision point cloud data and extract key geometric parameters. Due to the use of high-precision point cloud data and advanced 3D reconstruction algorithms, the reconstructed 3D model of the wedge joint has a high degree of realism and accuracy, greatly improving the 3D model's ability to describe the actual geometric shape and size of the wedge joint.

[0021] The processing data is collected in real time by sensors, and a physical model of wedge joint processing is established based on the collected data, including a cutting force model and a vibration model. The cutting force model accurately describes the contact state between the tool and the workpiece and the cutting force acting on the tool through the contact area infinitesimal method and the Newton-Euler equation; the vibration model uses the transfer function method to model the vibration characteristics of the tool and the workpiece with high precision. Through the physical modeling method, the inherent physical mechanism of the processing process is revealed, so that the prediction and control of the processing process can be carried out under the guidance of the physical model.

[0022] When generating the tool motion path, the cutting force and vibration physical model are used as constraints. The initial tool path is corrected by optimizing parameters such as feed speed, spindle speed, cutting depth and adjacent path spacing, so that the generated tool motion path can achieve smooth, efficient and low-vibration cutting while meeting the physical constraints. The path optimization method based on the physical model overcomes the blindness of traditional empirical path planning.

[0023] The adaptive particle swarm algorithm is used to globally optimize the tool motion path, with the surface quality as the optimization target. By optimizing the continuity and smoothness of the tool trajectory, the impact of trajectory corners and vibration on the machining surface is minimized, thus obtaining the optimal tool motion path. The adaptive particle swarm algorithm can intelligently search and optimize complex, high-dimensional, nonlinear path planning problems, and realize the full-scale optimization of the tool path from macro to micro. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The present application will be further described in the form of exemplary embodiments, which will be described in detail by the accompanying drawings. These embodiments are not restrictive, and in these embodiments, the same number represents the same structure, wherein:

[0025] Figure 1 is an exemplary flow chart of an automated wire joint processing method according to the present application;

[0026] Figure 2 is a schematic diagram of the processing of a wedge joint according to the present application;

[0027] Figure 3 is an exemplary flow chart of preprocessing according to the present application;

[0028] Figure 4 is an exemplary flow chart of extracting geometric parameters according to the present application;

[0029] Figure 5 is an exemplary flow chart for constructing a physical model for wedge joint processing according to the present application. DETAILED DESCRIPTION

[0030] The method and system provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings.

[0031] Figure 1 is an exemplary flow chart of an automated wire joint processing method according to the present application, Figure 2The invention is a schematic diagram of the processing of a wedge joint according to the present application, comprising: scanning a physical model of the wedge joint with a three-dimensional scanner to obtain point cloud data of the physical model of the wedge joint; collecting real-time data of the processing of the wedge joint with a sensor, the real-time data including tool position, cutting force and vibration parameters; preprocessing the point cloud data, the preprocessing including filtering with a Gaussian filtering algorithm and denoising with a wavelet threshold denoising algorithm; reconstructing a three-dimensional model of the wedge joint with the processed point cloud data, and extracting geometric parameters of the wedge joint from the three-dimensional model, the geometric parameters including wedge angle, length and width; establishing a physical model of the processing process of the wedge joint according to the collected real-time data, the physical model including a cutting force model and a vibration model; generating a tool processing motion path according to the three-dimensional model using computer-aided manufacturing software; optimizing the tool processing motion path using an adaptive particle swarm optimization algorithm; wherein the adaptive particle swarm optimization algorithm optimizes the processing quality by adjusting the speed and position of the particles in combination with the established physical model; converting the optimized tool processing motion path into a G code recognized by a CNC machine tool, and controlling the CNC machine to process the wedge joint.

[0032] Use a 3D scanner to scan the physical model of the wedge joint and obtain the point cloud data of the physical model of the wedge joint; select a high-precision 3D laser scanner, such as Hexagon Metrology's LeicaAbsoluteScanner LAS-20-8, which has a measurement accuracy of up to ±25μm and a scanning frequency of up to 150,000 points / second, and can quickly obtain a large amount of measurement point data on the surface of the wedge joint. Fix the physical model of the wedge joint to be scanned on the workbench, adjust the position and angle of the 3D scanner so that its laser probe can cover the entire surface of the wedge joint. In order to eliminate the influence of surface reflection on the scanning quality, a thin layer of matte paint can be sprayed on the surface of the wedge joint. During the scanning process, the laser probe emits a beam of laser line to scan the surface of the wedge joint, and the 3D coordinates of each sampling point are calculated by measuring the time difference of the laser reflected back to the probe. By constantly changing the angle and position of the laser probe, the scanner can collect a large amount of point data on each surface of the wedge joint.

[0033] The real-time data of the wedge joint processing process is collected by sensors, and the real-time data includes tool position, cutting force and vibration parameters; a series of sensors are installed on the CNC machine tool to collect the processing process data in real time. Among them, the tool position data is obtained by the position sensors such as the grating ruler or electromagnetic ruler provided by the machine tool, which can accurately measure the displacement and speed of the tool in the three coordinate directions of X, Y and Z. The cutting force sensor based on strain gauges is installed on the tool holders, such as the Kistler 9257B three-axis cutting force sensor. The sensor uses three orthogonal quartz force sensors to simultaneously measure the cutting force components in three orthogonal directions (main cutting force direction, feed force direction and radial force direction) during the cutting process. By collecting cutting force data, the tool force state can be monitored in real time and the stability of the cutting process can be evaluated. An acceleration sensor, such as the PCB 356A16 three-axis acceleration sensor, is installed at the spindle end to measure the vibration signals of the tool and workpiece during the cutting process. The sensor has high sensitivity and wide frequency response, and can accurately collect high-frequency vibration signals. By analyzing the vibration signal in the time and frequency domains, the stability of the cutting process can be monitored and the impact of vibration on the quality of the machined surface can be evaluated.

[0034] Figure 3 This is an exemplary flowchart of the preprocessing according to the present application, in which the point cloud data is preprocessed, and the acquired point cloud data is divided into different regions according to the geometric features of the wedge joint. The wedge joint is usually composed of different characteristic regions such as the shoulder (region A), the wedge surface (region B) and the transition fillet (region C), and the curvature changes in different regions are different. In order to better retain the geometric features of the wedge joint, it is necessary to divide the point cloud into several sub-regions according to the normal vector and curvature change of the point cloud. The adaptive segmentation of the point cloud can be achieved by using methods such as the regional growing method and the curvature-based segmentation algorithm.

[0035] According to the change of curvature of the divided area, the parameters of Gaussian filtering are adjusted. The parameters include the size of the filter kernel (kernel size, ks) and the standard deviation (standard deviation, σ). Gaussian filtering is a commonly used point cloud smoothing algorithm. Its principle is to use a Gaussian kernel function to perform weighted averaging on each point of the point cloud to achieve the effect of denoising and smoothing. The two key parameters of the Gaussian kernel function, ks and σ, determine the degree and range of smoothing. For areas with large curvature changes, such as wedge surfaces (region B) and transition fillets (region C), a smaller ks (such as ks B =3,ks C =5) and σ (such as σ B =0.3,σ C=0.5) to avoid excessive smoothing and loss of geometric details; for areas with small curvature changes, such as shoulders (region A), a larger ks (such as ks A =10) and σ (such as σ A =1) to better remove noise.

[0036] Using the adjusted parameters, each divided region is subjected to Gaussian filtering to obtain the filtered point cloud data. For each divided sub-region (such as regionA, regionB, regionC), the corresponding Gaussian kernel parameter (ks A ,σ A ;ks B ,σ B ;ks C ,σ C ), for each point P in the region i Specifically, for each point P i , take this point as the center, select a point P that contains its neighborhood j The spherical area of ​​​​the point is calculated according to the Gaussian kernel function G(x) The weight w of all points in the neighborhood of the point ij , and then perform weighted averaging on the coordinates of these points to obtain the new filtered coordinates P of the point i '. Repeat this process until all points in each sub-region (regionA, region B, regionC) have completed filtering.

[0037] The point cloud data of each region (regionA', regionB', regionC') after filtering are spliced ​​to obtain the complete point cloud data M after the wedge joint is filtered. filtered Since different regions use different filtering parameters, it is necessary to perform transition processing on the boundary points of adjacent regions when stitching point clouds to avoid discontinuity or jaggedness. Point cloud registration algorithms, such as ICP (Iterative Closest Point) algorithm, can be used to finely register and smoothly transition the boundary points of adjacent regions (such as regionA' and regionB', regionB' and regionC'), and finally obtain a complete and smooth wedge joint point cloud model M. filtered .

[0038] For the spliced ​​point cloud data M filtered Perform wavelet decomposition to obtain wavelet coefficients at different scales. Select a suitable wavelet basis function ψ(x) (such as Daubechies wavelet db4) and the number of decomposition layers n (such as n = 4) to perform wavelet decomposition on the point cloud data M filteredPerform n-layer wavelet decomposition to obtain the wavelet coefficient matrix W={W1,W2,......,W n}, where W i Represents the wavelet coefficients obtained by decomposition at the i-th layer. Wavelet decomposition can separate the noise and features in the point cloud data. filtered Information is separated at different scales, which facilitates the subsequent threshold denoising process.

[0039] According to the geometric features of the wedge joint, the decomposed wavelet coefficient matrix W is divided into different regions. Similar to Gaussian filtering, according to the geometric features of the wedge joint such as the shoulder (region A), wedge surface (region B) and transition fillet (region C), each layer of wavelet coefficient W is divided into different regions. i Divide into corresponding sub-areas {W i,A ,W i,B ,W i,C}, where W i,A ,W i,B ,W i,C They represent the coefficients corresponding to regionA, B, and C in the i-th layer of wavelet coefficients respectively.

[0040] Calculate the variance Var(W) of the wavelet coefficients in each region i,A )、Var(W i,B )、Var(W i,C ), compare the variance with the preset noise threshold λ noise , use Bayes threshold to adjust the wavelet threshold λ of the corresponding area i,A ,λ i,B ,λ i,C Specifically, for each sub-region W of the i-th layer wavelet coefficients i,A ,W i,B ,W i,C , calculate the variance of its coefficient Var(W i,A )、Var(W i,B )、Var(W i,C ), the variance is compared with the preset noise threshold λ noise Compare, if the variance is greater than λ noise , then it is considered that the region contains more noise and requires a greater degree of threshold shrinkage; conversely, if the variance is less than or equal to λ noise , then it is considered that the region mainly contains feature information and only a small degree of threshold shrinkage is required. The Bayes criterion is used to adaptively adjust the wavelet threshold λ of each region. i,A ,λ i,B ,λ i,C , adapting it to the variance and noise level of the region.

[0041] Using the adjusted threshold λi,A ,λ i,B ,λ i,C , for each region W i,A ,W i,B ,W i,C The wavelet coefficients are thresholded to remove noise. For each wavelet coefficient w∈W i,A or W i,B , W i,C , compare its absolute value |w| with the threshold λ of the corresponding area i,A (or λ i,B , i,C ), if |w|≤λ i,A (or λ i,B , i,C ), then the coefficient is set to zero, w = 0; otherwise, the coefficient is retained. Through this soft threshold processing, the noise component in the wavelet coefficient can be effectively suppressed, while the large coefficient reflecting the characteristics of the wedge joint is retained.

[0042] The wavelet coefficient matrix after denoising is W'={W1',W2',......,W n '} Perform wavelet reconstruction to obtain the denoised point cloud data M denoised The wavelet reconstruction algorithm corresponding to the wavelet decomposition is used to reconstruct the processed wavelet coefficient matrix W' layer by layer, and finally the denoised point cloud data M is obtained. denoised . Compared with the original point cloud data M filtered In comparison, M denoised Has higher signal-to-noise ratio and feature fidelity.

[0043] Figure 4 This is an exemplary flow chart for extracting geometric parameters according to the present application. The three-dimensional model of the wedge joint is reconstructed using the processed point cloud data, and the geometric parameters of the wedge joint are extracted from the three-dimensional model. The geometric parameters include the wedge angle, length and width; an octree is constructed. The bounding box of the point cloud data M_denoised is used as the root node, and the three-dimensional space is recursively divided into an octree. The maximum division depth of the octree is set to d max (such as d max =8), the minimum size of an octree node is l min (such as l min =0.1mm). For each non-empty leaf node, record the number of point cloud sampling points n it contains. sample . Calculate the eigenvalues ​​of the octree nodes. For each leaf node O of the octree i , calculate the eigenvalue f of the node according to the normal vector of the point cloud sampling point it contains iThe principal direction can be extracted from the normal vector of the point cloud sampling point contained in the node by using methods such as PCA (principal component analysis) as the eigenvalue f of the node. i . Adjust the octree partition depth d and the number of sampling points n for each node sample According to the geometric characteristics of the wedge joint, different partitioning depths and sampling point numbers are used for octree nodes in different areas. For areas with more geometric details, such as wedge faces and transition fillets, a larger partitioning depth d is used. B =7,d C =6 and a larger number of sampling points n B =30,n C = 20 to improve the reconstruction accuracy; for areas with less geometric details, such as shoulders, a smaller division depth d is used A =5 and a smaller number of sampling points n A = 10 to reduce the amount of calculation. By adaptively adjusting the octree parameters, the efficiency of the algorithm can be improved while ensuring the reconstruction accuracy. Construct the Poisson equation.

[0044] According to the eigenvalue f of the octree i and the number of sampling points n sample , construct Poisson's equation Where u(x) is the reconstructed implicit function, and f(x) is the eigenvalue field of the octree node. Using the finite element method, the Poisson equation is discretized into a linear equation system Au=b, where A is the discrete Poisson operator matrix and b is the eigenvalue vector. Solve the Poisson equation. Using numerical algorithms such as the multigrid method, the discrete Poisson equation Au=b is solved to obtain the discrete solution u of the reconstructed implicit function u(x) on the octree node. i . Extract the isosurface. According to the discrete solution u of the reconstructed implicit function u(x) i , using the Marching Cubes algorithm, extract the isosurface of u(x)=C and generate the triangular mesh surface model Mesh of the wedge joint model Wherein, C is the threshold of the isosurface, which can be adjusted according to the density of the point cloud data and the required reconstruction accuracy (such as C = 0.5).

[0045] The reconstructed three-dimensional mesh model Mesh_model of the wedge joint is simplified and the mesh model Mesh is calculated. model Each vertex v i The mean curvature H i . Discrete differential geometry methods can be used, such as the Voronoi area-weighted mean curvature estimation formula: Among them A i For vertex v i The Voronoi area, α ij and β ijFor vertex v i Its adjacent vertex v j The two diagonal angles between them. According to the curvature distribution, the mesh model is partitioned. model Divided into regions R with large curvature changes h and the region R1 where the curvature changes less. According to the average curvature H i The distribution histogram of th (such as H th =0.2) for binary segmentation. Mesh simplification is performed on the region R1 with small curvature changes. The Edge Collapse algorithm is used to gradually simplify the mesh facets in the R1 region. Each time simplification is performed, the edge with the smallest folding cost is selected, its two endpoints are merged into a new point, and the relevant mesh topology is updated. This process is repeated until the preset simplification rate r1 (such as r1 = 50%) or the minimum number of facets N1 (such as N1 = 1000) is reached. For the region R1 with large curvature changes, h The simplified R1 region is kept unchanged to preserve the key geometric features of the wedge joint. h The regions are spliced ​​to obtain a simplified mesh model Mesh simplified .

[0046] The Laplacian smoothing algorithm is used to simplify the mesh model Mesh simplified Perform smoothing and calculate the mesh model Mesh simplified Each vertex v i The Laplacian coordinates L i The Laplacian coordinates represent the vertex v i The positional relationship relative to its adjacent vertices can be calculated using the following formula: where d i For vertex v i degree (i.e. the number of adjacent vertices), w ij For vertex v i With v j The weights between (Cotangent weights or Uniform weights can be used), For all vertex v i Adjacent vertices v j Sum. Construct the Laplacian matrix L and the mass matrix M. The Laplacian matrix L is a sparse matrix, where L ii =1, (If the vertex v i With v j adjacent), and the rest of the elements are 0. The mass matrix M is a diagonal matrix, where M ii Represents vertex vi The Voronoi area A i . Construct the Laplacian smoothing equation (M-λL)V'=MV, where V is the coordinate matrix of the mesh vertices, V' is the vertex coordinate matrix after smoothing, and λ is the smoothing factor (controls the smoothing intensity, usually 0.1~0.5). Solve the Laplacian smoothing equation to obtain the smoothed vertex coordinate matrix V'. Sparse linear equation solvers such as the conjugate gradient method can be used to efficiently solve the smoothing problem of large-scale mesh models. According to the smoothed vertex coordinate matrix V', update the mesh model Mesh simplified The vertex position of the smoothed mesh model Mesh smoothed .

[0047] S43, through the triangulation method, the wedge angle, length and width of the wedge joint are extracted from the simplified and smoothed three-dimensional model of the wedge joint using the least square method. According to the geometric characteristics of the wedge joint, the mesh model Mesh smoothed Extract two wedge-shaped surfaces F from a and F b The mesh model can be segmented into different feature regions according to the normal vector and curvature distribution of the mesh patch by using methods such as region growing or threshold segmentation, and two wedge-shaped surfaces F can be extracted. a and F b . For the two extracted wedge surfaces F a and F b The planes are fitted separately to obtain the parameterized representation of the wedge surface. The least squares plane fitting algorithm can be used to fit the vertex set {v i} to fit the plane equation ax+by+cz+d=0, where (a, b, c) is the plane normal vector and d is the distance from the plane to the origin. Let the coefficient vector in the plane equation be n=(a, b, c, d) and the vertex coordinate matrix be V. The least squares plane fitting problem can be expressed as: min||V n || 2 , constraint: ||n||=1. This problem can be solved by singular value decomposition (SVD) to obtain the plane parameter n a and n b The plane parameters of the two wedge surfaces obtained by fitting are

[0048] n a =(a a ,b a ,c a ,d a ) and n b =(a b ,b b ,c b ,db ), calculate the wedge angle α. The wedge angle can be obtained by the angle between the two plane normal vectors, that is, Among them, ||n α || and ||n b || respectively represent the plane normal vector n α and n b The modulus of , · represents the dot product operation of the vector, and arccos represents the inverse cosine function.

[0049] In the mesh model Mesh smoothed On the graph, select a series of measurement points {p i} (such as 20 points), measure these points to the two wedge surfaces F respectively a and F b The distance {d ai} and {d bi Then, using the principle of triangulation, calculate the wedge length l at each measuring point i Assume that the measurement point p i In F a The projection point on the surface is q ai , in F b The projection point on the surface is q bi , then the wedge length l i It can be calculated by the following formula: where α is the wedge angle. Finally, for all the wedge lengths {l i} to obtain the average length L of the wedge joint. Similarly, in the mesh model Mesh smoothed On the left, select a series of measurement line segments {s j} (such as 10 line segments), measure the distance between these line segments and the two wedge surfaces F a and F b The distance between the intersection points {w j Then, for all wedge widths {w j}Perform arithmetic averaging to obtain the average width W of the wedge joint.

[0050] In order to improve the accuracy and robustness of parameter measurement, the least square method can be used to optimize the above measurement results. The wedge angle α, length L and width W are taken as the parameters to be optimized, and the least square error function is constructed to minimize the residual square sum between the measured value and the theoretical value. Let the measured value of the wedge angle be {a k}, the length measurement is {L k}, the width measurement is {W k}, the corresponding theoretical values ​​are α, L, W, then the least squares error function can be expressed as: The least squares problem is solved by nonlinear least squares optimization methods such as the Levenberg-Marquardt algorithm to obtain the optimized wedge angle α*, length L and width W.

[0051] Figure 5 This is an exemplary flow chart for constructing a physical model for wedge joint processing according to the present application. Based on the collected real-time data, a physical model of the wedge joint processing process is established, and the physical model includes a cutting force model and a vibration model; based on the collected tool position data, combined with the reconstructed three-dimensional model of the wedge joint, by establishing a transformation matrix between the tool coordinate system and the workpiece coordinate system, an analytical method is used to establish a contact geometry relationship model between the tool and the workpiece; a tool coordinate system is established. Based on the geometric parameters and installation position of the tool, a tool local coordinate system {T} is established. The origin O of the tool coordinate system {T} T Located at the rotation center of the tool, that is, the tool tip. The coordinate axis direction is consistent with the rotation axis and radial direction of the tool, where Z T The axis is parallel to the rotary axis of the tool, X T The radial direction of the axis and the tool is consistent, Y T The axes satisfy the right-hand rule. The basis vectors of the tool coordinate system {T} are {X T ,Y T ,Z T}.

[0052] Establish the workpiece coordinate system. According to the position and orientation of the three-dimensional model of the wedge joint, establish the workpiece coordinate system {W}. The origin O of the workpiece coordinate system {W} W Located at the geometric center of the wedge joint, that is, the 3D model Mesh of the wedge joint smoothed The coordinate axis direction is consistent with the length, width and height of the wedge joint, where X W The axis is in the same direction as the length of the wedge joint, Y W The axis is aligned with the width of the wedge joint, Z W The height direction of the axis is consistent with that of the wedge joint. The basis vectors of the workpiece coordinate system {W} are {X W ,Y W ,Z W}.

[0053] Establish the transformation matrix from the tool coordinate system to the workpiece coordinate system. According to the relative position relationship between the tool coordinate system {T} and the workpiece coordinate system {W}, establish the transformation matrix W from the tool coordinate system to the workpiece coordinate system T The transformation matrix W T Contains the translation matrix W P T and the rotation matrix W R T Two parts, namely 0, 1]. Among them, the translation matrix WP T Indicates the origin of the tool coordinate system O T Coordinates in the workpiece coordinate system {W}, W P T =[p x ,p y ,p z ] T ; Rotation matrix W R T represents the posture of the tool coordinate system {T} relative to the workpiece coordinate system {W}, r ij Represents the projection of the basis vectors of the tool coordinate system in the workpiece coordinate system. Rotation matrix W R T It can be expressed by rotation transformation or Euler angles.

[0054] The analytical method is used to calculate the spatial position of the tool in the workpiece coordinate system. For any point Q on the tool, its coordinates in the tool coordinate system {T} are Q T =[x t ,y t ,z t ] T . Using the transformation matrix W T , the coordinates of point Q can be transformed from the tool coordinate system {T} to the workpiece coordinate system {W}, that is, By traversing the discrete points on the tool, the spatial position and posture of the tool in the workpiece coordinate system can be obtained.

[0055] Calculate the contact point coordinates and normal vectors between the tool and the workpiece. In the workpiece coordinate system {W}, use the analytical method to calculate the three-dimensional model Mesh of the tool and the wedge joint. smoothed The contact point coordinates and normal vector. For each discrete point W on the tool i Q , calculate the shortest distance d to the surface of the three-dimensional model of the wedge joint i , and record the corresponding nearest point W Pi If the shortest distance d i If it is less than a given threshold value δ (such as δ = 0.01 mm), then point W is considered Qi Contact with wedge joint surface, W Pi is the contact point coordinate. At the same time, calculate the contact point W Pi The unit normal vector W of the wedge joint surface at ni The calculation of the nearest point and normal vector can be accelerated by spatial indexing methods such as KD tree and octree.

[0056] The contact geometry model between the tool and the workpiece is established. According to the calculated contact point coordinate set {W Pi} and the corresponding normal vector set {Wni}, establish the contact geometry relationship model between the tool and the wedge joint workpiece. The contact geometry relationship model describes the contact area and contact mode between the tool and the wedge joint during the machining process, and can be used for subsequent cutting force calculation, tool trajectory optimization and machining parameter selection. Common contact geometry relationship models include point contact model, line contact model and surface contact model, which correspond to point contact, line contact and surface contact between the tool and the workpiece respectively. According to the distribution characteristics and number of contact points, select the appropriate contact geometry relationship model for modeling.

[0057] According to the collected cutting force data, combined with the contact geometry model, by establishing micro-elements on the contact area and analyzing the force balance during the cutting process of a single tooth, the cutting force physical model is established using the Newton-Euler equation, including: according to the contact point position determined by the contact geometry model, micro-elements are established along the tool cutting direction on the contact area. Each micro-element corresponds to a time step Δt in the cutting process of a single tooth. Assuming that within the time step Δt, the tool moves a distance Δl along the cutting direction, the length of the micro-element is Δl. The width Δw of the micro-element depends on the radial feed of the tool, and the height Δh depends on the axial feed.

[0058] Establish the micro-element coordinate system {M}. The origin O of the micro-element coordinate system {M} M Located at the geometric center of the microelement, the coordinate axis direction is consistent with the cutting speed, radial feed and axial feed direction. M The axis is consistent with the cutting speed direction v, Y M Axis and radial feed direction f r Consistent, Z M Axis and axial feed direction f a The basis vectors of the infinitesimal coordinate system {M} are {X M ,Y M ,Z M}.

[0059] The cutting thickness and cutting width model is established on the microelement. For the microelement (i, j), its cutting thickness h(i, j) and cutting width b(i, j) are: h(i, j) = f z ×sin(κ r )×sin(φ i -φ j ); Among them, f z is the feed per tooth, κ r is the principal deflection angle, φ i is the angular position of the i-th tooth, φ j is the angular position of the jth microelement, a pis the axial cutting depth. Based on the cutting mechanism, the stress state in the cutting deformation zone is analyzed, and the tangential force, radial force and axial force on each microelement are calculated. On the microelement (i, j), the stress balance equation in the cutting deformation zone is established: dF t (i,j)=τ s ×db(i,j)×dh(i,j); dF r (i,j)=σ n ×db(i,j)×dh(i,j);

[0060] dF a (i,j)=τ a ×db(i,j)×dh(i,j); where dF t (i,j),dF r (i,j),dF a (i, j) are the tangential force, radial force and axial force on the microelement (i, j), τ s is the shear stress, σ n is the normal stress, τ a is the axial shear stress, db(i,j) and dh(i,j) are the width and thickness of the element (i,j), respectively.

[0061] The cutting forces on each microelement are accumulated to obtain the tangential force, radial force and axial force of the entire tool during the cutting process. Assume that there are N teeth on the tool, and the cutting process of each tooth is divided into M microelements. Then, at time step k, the tangential force F on the tool is t (k), radial force F r (k) and axial force F a (k) is: Establish a physical model of cutting force. According to the above calculated curves of the change of tangential force, radial force and axial force over time, establish a physical model of cutting force. The physical model of cutting force describes the change law of the cutting force on the tool during the cutting process, which can be expressed as:

[0062] F(t)=[F t (t)F r (t)F a (t)] T ; Where F(t) is the cutting force vector, t is the time variable,

[0063] F t (t)F r (t)F a (t) are the time-varying functions of tangential force, radial force and axial force respectively.

[0064] According to the collected vibration parameters, the tool and workpiece are modeled by forced vibration, and the vibration physical model of the tool and workpiece frequency characteristics is established by using the transfer function method, including: simplifying the tool into a concentrated mass and torsional vibration system, and establishing the tool vibration equation. Assume that the mass of the tool is m t , the stiffness is k t , the damping coefficient is c t , under the action of cutting force F(t), the vibration displacement of the tool is x t (t), the vibration equation of the tool is:

[0065] The workpiece is simplified into a uniform Euler beam model, and the vibration equation of the workpiece is established by combining the lateral vibration and torsional vibration of the workpiece under the action of cutting force. Assuming that the length of the workpiece is L, the cross-sectional area is A, the density is ρ, the elastic modulus is E, and the moment of inertia is I, under the action of cutting force F(t), the lateral vibration displacement of the workpiece is y w (x, t), the torsional vibration angle is θ w (x, t), then the vibration equation of the workpiece is:

[0066]

[0067] Where x0 is the coordinate of the cutting force point, δ(*) is the Dirac function, M(t) is the cutting moment, G is the shear modulus, I p is the polar moment of inertia.

[0068] The cutting force is set as the excitation term in the vibration equation of the tool and the workpiece, and the vibration equation is solved using the transfer function method. Taking Laplace transform we get: (m t ×s 2 +c t ×s+k t )×X t (s) = F(s); where s is the complex frequency, X t (s) and F(s) are x t (t) and the Laplace transform of F(t). Then the frequency response function of the tool is:

[0069] Similarly, for the vibration equation of the workpiece, taking the Laplace transform we get:

[0070] Among them, y w (x,s) and θ w (x,s) are y w (x,t) and θ wLaplace transform of (x, t). Solving the above partial differential equation, we can get the frequency response function H of the workpiece w (x,s).

[0071] According to the frequency response function of the tool and the workpiece, the relative displacement and speed of the tool and the workpiece during the cutting process are calculated. Assume that the relative displacement of the tool and the workpiece is x r (t), relative speed is v r (t), then: x r (t) = x t (t)-y w (x0,t); In the frequency domain, the relative displacement and velocity can be expressed as: X r (s) = X t (s)-Y w (x0,s)=[H t (s)-H w (x0,s)]×F(s);

[0072] V r (s) = s × X r (s) = s × [H t (s)-H w (x0,s)]×F(s); According to the frequency response function, the relative displacement and speed of the tool and the workpiece under different speeds and feed conditions are calculated to obtain the vibration physical model of the tool and workpiece frequency characteristics. Assuming the tool speed is n and the feed per tooth is f(z), the cutting force F(t) can be expressed as: Among them, ω=2πn / 60 is the angular frequency, F k and φ k are the amplitude and phase of the kth harmonic respectively. Substituting F(t) into the frequency response function, we can get the frequency spectrum characteristics of relative displacement and velocity under different speeds and feed conditions: r (w)=[H t (jw)-H w (x0,jw)]×F(w);

[0073] V r (w) = jw × X r (w); where j is an imaginary unit. Based on the above spectrum characteristics, the frequency response curves of the relative displacement and velocity of the tool and the workpiece under different rotation speeds and feed conditions can be plotted to obtain the vibration physical model of the tool and workpiece frequency characteristics.

[0074] The established cutting force physical model and vibration physical model are coupled to establish the dynamic equation between cutting force and vibration, and obtain the physical model of the wedge joint processing process. This includes: setting the relative displacement x between the tool and the workpiece output by the vibration physical model in the cutting force physical model r (t) and speed v r (t) data, and establish the cutting force dynamics equation. Assume that the cutting thickness in the cutting force physical model is h(t), the cutting width is b(t), and the material shear stress is τ s , the material deformation coefficient is K c , then the cutting force dynamics equation is: F t (t) = K tc ×b(t)×h(t);F r (t) = K rc ×F t (t); F a (t) = K ac ×F t (t); where K rc and K ac are radial force coefficient and axial force coefficient respectively, φ c is the shear angle. Considering the influence of vibration on cutting thickness, we have:

[0075] h(t)=f z ×sin(κ r )+x r (t)×sin(κ r )+v r (t)×T×sin(κ r ), where f z is the feed per tooth, κ r is the main deflection angle, and T is the cutting cycle. Substituting h(t) into the cutting force dynamics equation, we can get the cutting force dynamics equation: F t (t) = K tc ×b(t)×[h(t)=f z ×sin(κ r )+x r (t)×sin(κ r )+v r (t)×T×sin(κ r )];

[0076] F r (t) = K rc ×F t (t); F a (t) = K ac ×F t (t); This equation reflects the influence of vibration displacement and speed on cutting force.

[0077] Set the tangential force F output by the cutting force physics model in the vibration physics model t (t), radial force F r (t) and axial force F a (t) data is used as the excitation term of the vibration equation to establish the vibration dynamics equation. For the vibration equation of the tool, the cutting force is used as the excitation term to obtain: For the vibration equation of the workpiece, taking the cutting force and cutting torque as excitation terms, we get:

[0078]

[0079] Among them, r t is the tool radius. This equation reflects the effect of cutting force on tool and workpiece vibration.

[0080] The cutting force dynamics equation and the vibration dynamics equation are coupled through mutual input and output to construct a coupled dynamics equation group. r (t) and speed v r (t) is replaced by the output of the vibration dynamics equation, and the cutting force F in the vibration dynamics equation is replaced by t (t), F r (t) and F a (t) is replaced by the output of the cutting force dynamics equation, and we get:

[0081]

[0082]

[0083]

[0084] The above equations constitute the coupled dynamic equations of cutting force and vibration, reflecting the interaction between cutting force and vibration. The coupled dynamic equations are solved by numerical integration method to obtain the cutting force and vibration data at different times. Commonly used numerical integration methods include Runge-Kutta method, Euler method and trapezoidal method. Taking the fourth-order Runge-Kutta method as an example, the state variable is

[0085] Then the coupled dynamic equations can be written in the form of state space equations: Where f(·) is the state equation. According to the fourth-order Runge-Kutta method, the update formula of the state variable in the time step Δt is: k1=f(z(t n ),t n );

[0086]

[0087] k4=f(z(t n )+k3×Δt,t n +Δt); Among them, t n is the nth time node, and Δt is the time step. Through iterative calculation, the state variables at different times can be obtained, and then the cutting force and vibration data can be obtained.

[0088] Based on the cutting force and vibration data obtained by numerical integration, a physical model of the wedge joint machining process is constructed. The physical model can be expressed in the form of time domain curves, frequency domain curves, three-dimensional waterfall charts, etc., which can intuitively reflect the changing laws and dynamic characteristics of cutting force and vibration during the machining process. At the same time, the physical model can also be used to analyze the influence of factors such as cutting parameters, tool geometry parameters and material properties on cutting force and vibration, optimize machining technology, and predict and control chatter behavior.

[0089] Generate a tool processing motion path by using computer-aided manufacturing software according to the three-dimensional model; including: importing the three-dimensional model of the wedge joint into the computer-aided manufacturing software, and defining the tool type, tool geometry parameters and cutting parameters according to the processing requirements of the wedge joint; wherein the tool geometry parameters include tool diameter, helix angle and number of blades, and the cutting parameters include spindle speed, feed speed and cutting depth;

[0090] Import the three-dimensional model of the wedge joint into the computer-aided manufacturing software, and define the tool type, tool geometry parameters and cutting parameters according to the processing requirements of the wedge joint. The tool type can be a ball end mill, a flat end mill or a taper mill, etc., determined according to the surface characteristics and processing accuracy requirements of the wedge joint. The tool geometry parameters include tool diameter D, helix angle β and number of blades Z. Select appropriate parameter values ​​according to the size and material properties of the wedge joint, for example, D = 10mm, β = 30°, Z = 4. The cutting parameters include spindle speed n, feed speed v f and cutting depth a p , set the parameter values ​​according to the processing efficiency and surface quality requirements of the wedge joint, for example, n = 5000r / min, v_f = 800mm / min, a_p = 0.2mm.

[0091] In the computer-aided manufacturing software, the extracted geometric parameters of the wedge joint are used as input parameters, and the inner and outer contour curves of the wedge joint are generated by parametric programming to construct the complete wedge joint processing contour data. The geometric parameters of the wedge joint are extracted, including the cross-sectional shape parameters of the wedge surface, the pitch circle radius R, the pitch circle angle θ and the helix angle γ. The inner contour curve C of the wedge joint is generated according to the mathematical model of the wedge surface using the parametric programming method. i (u) and outer contour curve C o(u), where u is the curve parameter. For example, for a spherical wedge joint, the inner contour curve can be expressed as: The outer contour curve can be expressed as: Among them, ΔR is the radius of the inner and outer contour curves. The generated inner and outer contour curve data are interpolated and fitted to obtain the complete wedge joint processing contour data.

[0092] According to the set tool parameters and the generated machining contour data, the tool location file generation algorithm is used to calculate the spatial position coordinates of the tool on the machining contour and the tool axis direction vector to generate the initial tool motion trajectory containing the tool tip point coordinate sequence data and the tool axis vector sequence data. According to the parameters such as tool diameter D, helix angle β and number of blades Z, the tool geometry model is established to calculate the tool envelope surface and tool tip point coordinates. The machining contour data is discretized into a series of machining contour points P i (x i ,y i ,z i ),i=1,2,.....,n, where n is the number of discrete points. For each machining contour point P i , calculate the intersection point Q of the tool envelope surface and the machining contour i (x qi ,y qi ,z qi ), as the tool tip coordinates on the tool motion trajectory. According to the coordinates of two adjacent tool tip points Q i and Q i+1 , calculate the tool axis direction vector V i (v xi ,v yi ,v zi ), as the tool axis vector on the tool motion trajectory. The calculated tool tip point coordinate sequence data {Q i} and tool axis vector sequence data {V i} are combined into the initial tool motion trajectory.

[0093] The cutting force data output by the cutting force physical model is used as the cutting force constraint condition to check the cutting force constraint of the initial tool motion trajectory. By adjusting the feed speed v on the initial tool motion trajectory, f and spindle speed n, the tool motion trajectory under cutting force constraints is obtained. According to the cutting force physical model, the cutting force F (F t ,F r ,F a ), and obtain the cutting force data series {F i}. Set cutting force constraints, such as the maximum tangential force F tmax , Maximum radial force F rmax and the maximum axial force F amaxFor each tool position on the initial tool motion trajectory, the corresponding cutting force data F i Compare with the cutting force constraints. If the cutting force data exceeds the constraints, adjust the feed rate v at that tool position. f and spindle speed n, so that the cutting force meets the constraints. The adjustment method can adopt iterative optimization algorithm, such as gradient descent method, Newton method, etc. f The tool motion trajectory under the cutting force constraint is obtained by interpolating and smoothing the spindle speed n.

[0094] The vibration displacement x output by the vibration physics model r and speed v r The data is used as the vibration constraint condition, and the vibration constraint check is performed on the corrected tool motion trajectory data. By adjusting the cutting depth a in the tool motion trajectory data p and the spacing between adjacent trajectories Δd, the tool motion trajectory that satisfies the cutting force and vibration constraints is obtained. According to the vibration physical model, the vibration displacement x at different tool positions is calculated. r and speed v r , and obtain the vibration data sequence {x ri} and {v ri}. Set vibration constraints, such as the maximum vibration displacement x rmax and maximum vibration velocity v rmax For each tool position on the tool motion trajectory under cutting force constraints, the corresponding vibration data x ri and v ri The vibration data is compared with the vibration constraints. If the vibration data exceeds the constraints, the cutting depth a is adjusted at that tool position. p and the adjacent track spacing Δd, so that the vibration meets the constraint conditions. The adjustment method can adopt an iterative optimization algorithm, such as a genetic algorithm, a particle swarm optimization algorithm, etc. p The interpolation and smoothing of the adjacent trajectory spacing Δd are performed to obtain the final tool motion trajectory that satisfies the cutting force and vibration constraints.

[0095] The adaptive particle swarm optimization algorithm is used to optimize the tool processing motion path; the adaptive particle swarm optimization algorithm adjusts the speed and position of the particles, combines the established physical model, and optimizes the processing quality; it includes: initializing the particle swarm, taking the tool motion trajectory data that meets the cutting force and vibration constraints as the initial particle position X i (j), where i is the particle number, j is the number of iterations, and initially j = 0. The feed speed v on the tool motion trajectory f , spindle speed n and cutting depth a p As the variable to be optimized, it constitutes the dimension of the particle. The initial velocity V of the randomly generated particlei (0), V i (0)∈[-V max ,V max ], where V max is the speed limit. Set the initial position of the particle X i (0) As the individual historical optimal position P of the particle i (0), and the particle position with the smallest fitness function value in the initial particle swarm is taken as the global optimal position G(0).

[0096] The fitness function is constructed, and the coupled dynamic model of the wedge joint machining process is established based on the machining surface roughness R a The objective function f(X i (j)), as the fitness function of the particle swarm algorithm. The objective function f(X i The expression of (j) is: minf(X i (j)) = R a (v f ,n,a p ); Constraints: v fmin ≤v f ≤v fmax ;

[0097] n min ≤n≤n max ; a pmin ≤a p ≤a pmax ; Among them, R a (v f ,n,a p ) is the roughness of the machined surface calculated by the coupled dynamics model, v fmin 、v fmax 、n min 、n max 、a pmin 、a pmax They are the value ranges of feed rate, spindle speed and cutting depth respectively.

[0098] Update particle speed and position. In the jth iteration, for the i-th particle, update its speed V according to the following formula i (j): Among them, ω(j) is the inertia weight, c1(j) is the individual learning factor, c2(j) is the social learning factor, r1 and r2 are random numbers between [0, 1]. According to the updated particle speed V i (j) Update the particle position X according to the following formula i (j):X i (j) = X i (j-1)+V i(j); Set the particle position X i (j) Substitute the objective function f(X i (j)) Calculate the fitness function value. Set the current position of particle i to X i (j) and its individual historical optimal position P i (j-1) for comparison, if f(X i (j))<f(P i (j-1)), then X i (j) as the new individual historical optimal position P i (j), otherwise keep P i (j) = P i (j-1). Set the current position of particle i to X i (j) is compared with the global optimal position G(j-1), if f(X i (j))<f(G(j-1)), then X i (j) is taken as the new global optimal position G(j), otherwise G(j)=G(j-1) is maintained.

[0099] Adaptive adjustment strategy, using the inertia weight strategy to adjust the weight ω(j) in the particle velocity update formula. The weight decreases as the number of iterations j increases. The expression is: Among them, ω max and ω min are the maximum and minimum values ​​of the inertia weight, j max The adaptive learning factor strategy is used to adjust the individual learning factor c1(j) and social learning factor c2(j) in the particle position update formula. The expression is: Among them, c 2max and c 2min are the maximum and minimum values ​​of the individual learning factor, c 2max and c 2min are the maximum and minimum values ​​of the social learning factor, respectively.

[0100] Termination condition judgment: the termination condition of the particle swarm algorithm is set to the fitness function value reaching the preset convergence accuracy ε or the number of iterations j reaching the preset maximum number of iterations j max When the termination condition is met, the tool motion trajectory parameter v corresponding to the current global optimal position G(j) is f , n and a p As the optimized tool motion trajectory data.

[0101] The optimized tool processing motion path is converted into G code recognized by the CNC machine tool to control the CNC machine to process the wedge joint. From the output of the adaptive particle swarm optimization algorithm, the optimized tool motion trajectory data is extracted, including the coordinate values ​​(X, Y, Z) of each point on the tool trajectory, the feed speed v f , spindle speed n and cutting depth a p Parameters such as. According to the model and control system of the CNC machine tool, select the corresponding G code format and instruction set, such as ISO standard G code, Siemens 840D system G code, etc. Convert the optimized tool motion trajectory data into G code instructions, mainly including the following instructions: G00 / G01: Linear rapid movement / linear cutting movement instruction, used to control the movement of the tool on the linear path. G02 / G03: Clockwise / counterclockwise arc interpolation instruction, used to control the movement of the tool on the arc path. G90 / G91: Absolute coordinate programming / incremental coordinate programming instruction, used to specify the coordinate mode of tool movement. G54-G59: Workpiece coordinate system selection instruction, used to select the coordinate system used in the processing process. G17 / G18 / G19: XY plane / ZX plane / YZ plane selection instruction, used to specify the plane of circular interpolation. M03 / M04 / M05: Spindle forward / reverse / stop instruction, used to control the rotation of the spindle. M06: Tool change instruction, used to control the tool replacement in the tool magazine. M08 / M09: Coolant on / off command, used to control the use of coolant. F: Feed rate command, used to specify the tool movement speed. S: Spindle speed command, used to specify the spindle rotation speed.

Claims

1. A method for processing a wire joint based on automation, wherein the wire joint is a wedge-shaped joint, characterized in that: include: Scanning the physical model of the wedge joint by using a three-dimensional scanner to obtain point cloud data of the physical model of the wedge joint; The sensors collect real-time data during the wedge joint processing process, including tool position, cutting force and vibration parameters; Preprocessing the point cloud data includes filtering using a Gaussian filtering algorithm and denoising using a wavelet threshold denoising algorithm; Reconstructing a three-dimensional model of the wedge joint using the processed point cloud data, and extracting geometric parameters of the wedge joint from the three-dimensional model, the geometric parameters including the wedge angle, length and width; According to the collected real-time data, a physical model of the wedge joint machining process is established, and the physical model includes a cutting force model and a vibration model; According to the three-dimensional model, the tool processing motion path is generated using computer-aided manufacturing software; Adaptive particle swarm optimization algorithm is used to optimize the tool processing motion path; the adaptive particle swarm optimization algorithm adjusts the speed and position of particles and combines the established physical model to optimize the processing quality; Convert the optimized tool processing motion path into G code recognized by CNC machine tools, and control the CNC machine to process the wedge joint; The Gaussian filtering algorithm is used for filtering, including: According to the geometric characteristics of the wedge joint, the acquired point cloud data is divided into different areas; According to the change of curvature of the divided area, the parameters of Gaussian filtering are adjusted, and the parameters include the size and standard deviation of the filter kernel; Using the adjusted parameters, Gaussian filtering is performed on each divided area to obtain filtered point cloud data; The point cloud data of each area after filtering are spliced ​​to obtain the point cloud data after wedge joint filtering.

2. The automated wire joint processing method according to claim 1, characterized in that: The wavelet threshold denoising algorithm is used for denoising, including: Perform wavelet decomposition on the spliced ​​point cloud data to obtain wavelet coefficients at different scales; According to the geometric characteristics of the wedge joint, the decomposed wavelet coefficients are divided into different regions; Calculate the variance of the wavelet coefficients in each region, compare the variance with the preset noise threshold, and use the Bayes threshold to adjust the wavelet threshold of the corresponding region; Using the adjusted threshold, threshold processing is performed on the wavelet coefficients of each divided area; The denoised wavelet coefficients are reconstructed by wavelet to obtain the denoised point cloud data.

3. The automated wire joint processing method according to claim 2, characterized in that: Extract the geometric parameters of the wedge joint from the 3D model, including: The octree-based Poisson reconstruction algorithm is used to reconstruct the preprocessed point cloud data into three dimensions. By adjusting the octree division depth and the number of sampling points at each node, the three-dimensional mesh model of the wedge joint is obtained. The curvature-based mesh simplification algorithm is used to simplify the three-dimensional mesh model; the Laplacian smoothing algorithm is used to smooth the simplified mesh model; Through triangulation method, the wedge angle, length and width of the wedge joint are extracted from the simplified and smoothed three-dimensional model of the wedge joint using the least square method.

4. The automated wire joint processing method according to claim 3, characterized in that: Establish a physical model of the wedge joint processing process, including: According to the collected tool position data and the reconstructed three-dimensional model of the wedge joint, the contact geometry model between the tool coordinate system and the workpiece coordinate system is established by using the analytical method; the contact geometry model describes the contact area and contact mode between the tool and the wedge joint during the machining process; Based on the collected cutting force data and the contact geometry model, the cutting force physical model is established using the Newton-Euler equation by establishing micro-elements on the contact area and analyzing the force balance of a single tooth during cutting. The cutting force physical model includes tangential force, radial force and axial force. According to the collected vibration parameters, the forced vibration model of the tool and the workpiece is modeled, and the vibration physical model of the frequency characteristics of the tool and the workpiece is established by using the transfer function method; the vibration physical model includes the relative displacement and speed of the tool and the workpiece during the cutting process; The established cutting force physical model and vibration physical model are coupled to establish the dynamic equation between cutting force and vibration, and obtain the physical model of the wedge joint machining process.

5. The automated wire joint processing method according to claim 4, characterized in that: Establish the contact geometry model between tool and workpiece, including: According to the geometric parameters and installation position of the tool, a tool coordinate system is established. The origin of the tool coordinate system is located at the rotation center of the tool, and the direction of the coordinate axis is consistent with the rotation axis and radial direction of the tool; According to the position and direction of the three-dimensional model of the wedge joint, a workpiece coordinate system is established. The origin of the workpiece coordinate system is located at the geometric center of the wedge joint, and the direction of the coordinate axis is consistent with the length, width and height of the wedge joint. According to the relative position relationship between the tool coordinate system and the workpiece coordinate system, a transformation matrix from the tool coordinate system to the workpiece coordinate system is established. The transformation matrix includes a translation matrix and a rotation matrix. The analytical method is used to calculate the spatial position of the tool in the workpiece coordinate system, the contact point coordinates and normal vector between the tool and the workpiece are obtained, and the contact geometric relationship model between the tool and the workpiece is established.

6. The automated wire joint processing method according to claim 5, characterized in that: Establish cutting force physical model, including: According to the contact point position determined by the contact geometric relationship model, micro-elements are established along the tool cutting direction on the contact area, and each micro-element corresponds to a time step in the cutting process of a single cutter tooth; Establish a microelement coordinate system, the origin of which is located at the geometric center of the microelement, and the direction of the coordinate axis is consistent with the cutting speed, radial feed and axial feed directions; The cutting thickness and cutting width models are established on the micro-element, the stress state in the cutting deformation zone is analyzed based on the cutting mechanism, and the tangential force, radial force and axial force on each micro-element are calculated; The cutting forces on each microelement are accumulated to obtain the time-varying curves of the tangential force, radial force and axial force of the entire tool during the cutting process, and a cutting force physical model is established.

7. The automated wire joint processing method according to claim 6, characterized in that: Establish a vibration physical model of the tool and workpiece frequency characteristics, including: The tool is simplified into a concentrated mass and torsional vibration system, and the vibration equation of the tool is established, which includes the mass, stiffness and damping coefficient of the tool; The workpiece is simplified into a uniform Euler beam model, and the vibration equation of the workpiece is established by combining the lateral vibration and torsional vibration of the workpiece under the action of cutting force. The cutting force is set as the excitation term in the vibration equation of the tool and the workpiece, and the vibration equation is solved by the transfer function method to obtain the relative displacement and velocity frequency response function of the tool and the workpiece during the cutting process. According to the frequency response function, the relative displacement and speed of the tool and the workpiece under different rotation speeds and feed conditions are calculated, and the vibration physical model of the frequency characteristics of the tool and the workpiece is obtained.

8. The automated wire joint processing method according to claim 7, characterized in that: The physical model of the wedge joint processing process is obtained, including: The relative displacement and speed data of the tool and the workpiece output by the vibration physical model are set in the cutting force physical model to establish the cutting force dynamics equation, which reflects the influence of the vibration displacement and speed on the cutting force. In the vibration physical model, the tangential force, radial force and axial force data output by the cutting force physical model are set as the excitation terms of the vibration equation, and the vibration dynamics equation is established. The vibration dynamics equation reflects the influence of the cutting force on the vibration of the tool and the workpiece. The cutting force dynamics equation and the vibration dynamics equation are coupled through mutual input and output to construct a coupled dynamics equation group, which includes cutting force and vibration and reflects the interaction between cutting force and vibration; The numerical integration method is used to solve the coupled dynamic equations, and the cutting force and vibration data at different times are obtained to construct the physical model of the wedge joint machining process.

9. The automated wire joint processing method according to any one of claims 2 to 8, characterized in that: Generate tool machining motion paths, including: The three-dimensional model of the wedge joint is imported into the computer-aided manufacturing software, and the tool type, tool geometry parameters and cutting parameters are defined according to the processing requirements of the wedge joint; the tool geometry parameters include tool diameter, helix angle and number of blades, and the cutting parameters include spindle speed, feed speed and cutting depth; In computer-aided manufacturing software, the extracted geometric parameters of the wedge joint are used as input parameters, and the inner contour curve and outer contour curve of the wedge joint are generated by parametric programming to construct complete wedge joint processing contour data; According to the set tool parameters and the generated machining contour data, the tool location file generation algorithm is used to calculate the spatial position coordinates of the tool on the machining contour and the tool axis direction vector to generate an initial tool motion trajectory containing tool tip point coordinate sequence data and tool axis vector sequence data; The cutting force data output by the cutting force physical model is used as the cutting force constraint condition to perform cutting force constraint verification on the initial tool motion trajectory. By adjusting the feed speed and spindle speed on the initial tool motion trajectory, the tool motion trajectory under the cutting force constraint is obtained. The vibration displacement and velocity data output by the vibration physics model are used as vibration constraint conditions, and the corrected tool motion trajectory data is checked for vibration constraints. By adjusting the cutting depth and adjacent trajectory spacing in the tool motion trajectory data, the tool motion trajectory that meets the cutting force and vibration constraints is obtained.

Images