Numerical control machining path control system based on artificial intelligence

Abstract

The invention discloses a numerical control machining path control system based on artificial intelligence, particularly relates to the field of numerical control machining, is used for solving the problems of continuity and stability of a curvature mutation area in cutter path planning, and accurately captures global and local geometric characteristics of a curved surface through a curvature constraint feature matrix and a curvature adaptive convolutional network. A machining track meeting the differential geometric continuity is generated, and the problem of path breakage of a curvature sudden change area is effectively avoided; meanwhile, energy scale criteria and curvature manifold constraints are introduced, the machining track is dynamically repaired, geometric repair and physical stability are balanced, cutting force sudden change and stress concentration are reduced, and therefore the machining reliability is enhanced; in addition, through finite element simulation of the digital twin platform and synchronous updating of servo system parameters, dynamic closed-loop matching of a control instruction and a machining state is achieved, and the control efficiency and the surface quality are further optimized. And therefore, high-precision, high-stability and high-efficiency collaborative optimization is realized in complex curved surface processing.

Description

Technical Field

[0001] The present invention relates to the field of numerical control machining, and more particularly, to a numerical control machining path control system based on artificial intelligence. Background Art

[0002] In the free-form surface machining of fields such as aerospace components and precision molds, the surface of the workpiece often presents the sudden curvature characteristics of non-uniform rational B-spline surfaces (NURBS). Such geometric forms require that the tool path must strictly follow the differential topological properties of the surface to ensure the continuity of the cutting trajectory in the normal vector direction and the tangent vector direction. Existing path planning models based on deep learning (such as PointNet++) rely on discrete point cloud data to construct the tool contact point sequence. Its core defect is that the point cloud feature extraction network captures geometric information in a local neighborhood aggregation manner, and it is difficult to effectively represent the global differential geometric characteristics of implicit surfaces (such as principal curvature distribution, Gaussian curvature change gradient). When the machining object exceeds the surface type covered by the training set, the path coordinate sequence output by the model shows topological breaks near the curvature extreme points, and the dynamic contact relationship between the tool envelope surface and the workpiece surface is misaligned.

[0003] The essence of this problem stems from the cognitive limitations of artificial intelligence models on the laws of geometric deformation. Existing methods simplify surface machining to a coordinate regression task of discrete points, ignoring the internal relationship between the tool movement trajectory and the continuous surface manifold. When the tool moves to the curvature mutation area, the local geometric features extracted by the point cloud network cannot reflect the overall topological constraints of the surface, resulting in connection errors in the generated path between adjacent cutting rows. Such errors show two failure modes in physical machining: one is the residual uncut material, forming surface protrusions, forcing additional subsequent finishing processes; the other is overcutting, causing subsurface damage to the workpiece, triggering stress concentration and a decrease in fatigue life. Especially in sensitive scenarios such as thin-walled parts and heterogeneous material composite structures, such defects will directly lead to functional failure of the workpiece, forcing the machining process to be interrupted and accompanied by the scrapping of high-value materials.

[0004] To solve the above problems, a technical solution is provided now. Summary of the Invention

[0005] To overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides a numerical control machining path control system based on artificial intelligence. Through a curvature constraint feature matrix and a curvature adaptive convolutional network, it accurately captures the global and local geometric characteristics of the surface, generates a machining trajectory that satisfies differential geometric continuity, and effectively avoids path breakage problems in regions with sudden curvature changes. At the same time, an energy scale criterion and a curvature manifold constraint are introduced to dynamically repair the machining trajectory, balance geometric repair and physical stability, reduce sudden changes in cutting force and stress concentration, thereby enhancing machining reliability. In addition, through the finite element simulation of the digital twin platform and the synchronous update of servo system parameters, dynamic closed-loop matching between control instructions and machining states is achieved, further optimizing control efficiency and surface quality. Furthermore, high-precision, high-stability, and high-efficiency collaborative optimization are achieved in complex surface machining, providing technical support for high-quality manufacturing in fields such as aerospace and precision molds to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] Surface feature extraction module: Based on the surface geometric properties of the workpiece model, obtain the principal curvature direction field and Gaussian curvature distribution, and fuse the curvature change gradient parameter and topological connectivity parameter to generate a curvature constraint feature matrix;

[0008] Trajectory generation network module: Convert the curvature constraint feature matrix into a heat map channel, and after splicing it with the original point cloud data in the spatial dimension, input it into a pre-trained convolutional network to generate an initial machining trajectory;

[0009] Trajectory effect analysis module: Based on the residual height difference between adjacent cutting rows of the initial machining trajectory, reverse map the cutting force mutation gradient, and combine the curvature continuity deviation to calculate the local surface strain energy density to construct an energy scale criterion;

[0010] Trajectory dynamic repair module: When the energy scale criterion breaks through the critical value of material plastic deformation, generate a transition spline curve according to the curvature constraint feature matrix, and repair the geometric parameters and physical stability of the path through the relationship between the curvature decay rate and energy dissipation;

[0011] Simulation optimization control module: Import the repaired path into the digital twin platform for finite element simulation to verify the cutting stress distribution and dynamic cutting force fluctuation range; Screen out the optimized trajectory that meets the preset stress constraints, and synchronously update the trajectory parameters and the acceleration feedforward coefficient of the servo system.

[0012] In a preferred embodiment, the surface feature extraction module includes the following:

[0013] Extract the implicit representation of the surface from the workpiece computer-aided design model and parameterize it as a function dependent on two parametric coordinates; calculate the first-order and second-order change rates of the surface in the two parametric coordinate directions, as well as the normal direction vector of the surface, based on the implicit representation of the surface; calculate the first set of basic geometric quantities and the second set of basic geometric quantities using the tangent vector and the normal direction vector of the surface, construct the curvature feature matrix, and determine the principal curvatures and principal curvature directions in the curvature constraint feature matrix through eigenvalue analysis, that is, the maximum bending degree, the minimum bending degree, and their corresponding main bending directions.

[0014] In a preferred embodiment, the surface feature extraction module further includes the following:

[0015] Calculate the Gaussian curvature by multiplying the maximum bending degree by the minimum bending degree, and generate the Gaussian curvature distribution on the parametric grid; calculate the change rates of the Gaussian curvature in the two parametric coordinate directions to form the curvature change gradient parameters; calculate the topological property values and the number of connected components according to the topological structure of the surface to determine the topological connectivity parameters; for each sampling point on the surface, construct a feature vector containing the main bending direction, the Gaussian curvature, the curvature change gradient parameters, and the topological connectivity parameters, and arrange all the feature vectors according to the parametric grid to form the curvature constraint feature matrix.

[0016] In a preferred embodiment, the trajectory generation network module includes the following:

[0017] Normalize each feature dimension of the curvature constraint feature matrix to generate the heatmap channels; map the heatmap channels to the coordinate positions of the original point cloud data through bilinear interpolation to generate the interpolated feature matrix, and splice it with the original point cloud data in the spatial dimension to form the comprehensive input data; input the comprehensive input data into the pre-trained convolutional network and train it by minimizing the geometric loss function to generate the initial machining trajectory that satisfies the normal vector orthogonality and tangent vector continuity constraints.

[0018] In a preferred embodiment, the geometric loss function is constructed as follows:

[0019] Construct a geometric loss function that includes the normal vector orthogonality constraint and the tangent vector continuity constraint, and ensure that the path satisfies differential geometric continuity by calculating the dot product of the path tangent direction and the surface normal vector and the deviation of the cosine value of the angle between the tangent directions of adjacent path points from 1.

[0020] In a preferred embodiment, the trajectory effect analysis module includes the following:

[0021] Calculate the residual height difference between adjacent cutting rows in the surface normal direction based on the initial machining trajectory, and inversely map the cutting force mutation gradient through the ratio of the residual height difference to the average spacing between adjacent cutting rows; calculate the Euclidean norm of the curvature feature vector between adjacent points based on the curvature constraint feature matrix and the initial machining trajectory, and then calculate the global curvature continuity deviation; calculate the local surface strain energy density through the sum of the weighted term of the square of the cutting force mutation gradient and the weighted term of the square of the global curvature continuity deviation; extract the maximum value in the local surface strain energy density sequence as the energy scale criterion.

[0022] In a preferred embodiment, the trajectory dynamic repair module includes the following:

[0023] Judge whether the energy scale criterion exceeds the critical value of material plastic deformation, and identify the over-standard area by comparing the energy scale criterion with the critical value of material plastic deformation; generate a transition spline curve based on the curvature constraint feature matrix, extract the maximum bending degree, minimum bending degree, their corresponding main bending directions, and Gaussian curvature distribution from the curvature constraint feature matrix, and use cubic B-spline interpolation technology to construct the transition spline curve, and ensure that the boundary conditions and curvature continuity requirements are met; dynamically adapt the geometric parameters and physical stability of the repair path by calculating the curvature decay rate and energy dissipation function of the transition spline curve, and adjust the control points to minimize the difference between the local surface strain energy density and the energy dissipation function.

[0024] In a preferred embodiment, the simulation optimization control module includes the following:

[0025] Discretize the repaired machining trajectory into three-dimensional coordinate points and import them into the digital twin platform to establish a tool path model; analyze the cutting stress distribution and dynamic cutting force fluctuation range on the workpiece surface through finite element simulation, calculate the maximum principal stress and minimum principal stress on the workpiece surface, as well as the peak and valley values of the dynamic cutting force; screen and optimize the trajectory according to the preset stress constraint conditions, check whether the maximum principal stress exceeds the preset threshold and whether the cutting force fluctuation range is within the allowable range, and if not, adjust the trajectory and repeat the simulation; extract the speed and acceleration parameters of the tool movement from the optimized machining trajectory, and calculate the acceleration feedforward coefficient; update the speed, acceleration, and acceleration feedforward coefficient to the control instruction stream of the servo system to achieve the dynamic matching of the control instruction and the machining state.

[0026] The technical effects and advantages of the numerical control machining path control system based on artificial intelligence of the present invention:

[0027] By means of the curvature constraint feature matrix and the curvature adaptive convolution network, the global and local geometric characteristics of the curved surface are accurately captured, and a machining trajectory that meets differential geometric continuity is generated, effectively avoiding the path breakage problem in the regions of sudden curvature changes. At the same time, an energy scale criterion and a curvature manifold constraint are introduced to dynamically repair the machining trajectory, balance geometric repair and physical stability, reduce sudden changes in cutting force and stress concentration, thereby enhancing machining reliability. In addition, through the finite element simulation of the digital twin platform and the synchronous update of the servo system parameters, the dynamic closed-loop matching of control instructions and machining states is achieved, further optimizing the control efficiency and surface quality. Furthermore, the collaborative optimization of high precision, high stability and high efficiency is realized in the machining of complex curved surfaces, providing technical support for high-quality manufacturing in fields such as aerospace and precision molds. Brief Description of the Drawings

[0028] Figure 1 This is a schematic structural diagram of the numerically controlled machining path control system based on artificial intelligence according to the present invention.

[0029] Figure 2 This is a schematic flow diagram of the curved surface feature extraction module of the numerically controlled machining path control system based on artificial intelligence according to the present invention. Detailed Embodiments

[0030] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0031] Embodiment 1: Figure 1 The numerically controlled machining path control system based on artificial intelligence according to the present invention is provided, including:

[0032] Curved surface feature extraction module: Based on the curved surface geometric attributes of the workpiece model, the principal curvature direction field and the Gaussian curvature distribution are obtained, and a curvature constraint feature matrix is generated by fusing the curvature change gradient parameter and the topological connectivity parameter;

[0033] Trajectory generation network module: The curvature constraint feature matrix is converted into a heat map channel, and after being stitched with the original point cloud data in the spatial dimension, it is input into a pre-trained convolution network to generate an initial machining trajectory;

[0034] Trajectory effect analysis module: Based on the residual height difference between adjacent cutting rows of the initial machining trajectory, the cutting force mutation gradient is inversely mapped, and the local surface strain energy density is calculated by combining the deviation of curvature continuity, and an energy scale criterion is constructed;

[0035] Trajectory Dynamic Repair Module: When the energy scale criterion breaks through the critical value of material plastic deformation, a transitional spline curve is generated based on the curvature constraint characteristic matrix, and the geometric parameters and physical stability of the path are repaired through the relationship between the curvature decay rate and energy dissipation;

[0036] Simulation Optimization Control Module: Import the repaired path into the digital twin platform for finite element simulation to verify the cutting stress distribution and the dynamic cutting force fluctuation range; Screen the optimized trajectories that meet the preset stress constraints, and synchronously update the trajectory parameters and the acceleration feedforward coefficient of the servo system.

[0037] In the free-form surface machining of aerospace components, precision molds and other fields, the workpiece surface usually presents as a non-uniform rational B-spline surface (NURBS), and its curvature characteristics may change abruptly in local areas. To ensure the continuity of the tool path in the normal vector direction and the tangent vector direction, it is necessary to extract accurate geometric constraint information based on the implicit surface differential geometric properties of the workpiece CAD model. Aiming at the problem that the existing path planning model is difficult to effectively represent the global differential geometric characteristics of the curvature mutation area, the surface feature extraction module provides a geometric constraint benchmark for the curvature adaptive convolution network of the trajectory generation network module by generating a curvature constraint characteristic matrix, avoiding path breakage and machining defects.

[0038] As Figure 2 shown, the surface feature extraction module includes the following:

[0039] S1.1, Extract the implicit surface differential geometric properties:

[0040] Obtain the implicit representation form of the surface from the computer-aided design model of the workpiece, and parameterize it as a function depending on two parameter coordinates, where these two parameter coordinates represent two directions on the surface respectively. Calculate the rate of change of the surface in the first parameter coordinate direction to obtain the tangent vector in this direction; Similarly, calculate the rate of change of the surface in the second parameter coordinate direction to obtain the tangent vector in the other direction. Then, calculate the second-order rate of change of the surface in the first parameter coordinate direction, the mixed rate of change of the two parameter coordinate directions, and the second-order rate of change of the surface in the second parameter coordinate direction. These results are used to analyze the bending characteristics of the surface subsequently. Finally, calculate the normal direction vector of each point on the surface by performing a vector cross product operation on the two tangent vectors and normalizing their lengths, representing the vertical direction of the surface in the local area.

[0041] By calculating the first-order and second-order rates of change of the surface in the two parameter coordinate directions, and the resulting normal direction vector, the geometric characteristics of the surface in the local area can be comprehensively captured. This information provides basic data for subsequent analysis of the bending degree and direction of the surface, ensuring that the tool path planning can adapt to the shape changes of the surface in the micro area, thereby improving the machining accuracy and path continuity.

[0042] S1.2, Calculate the principal curvatures and principal curvature directions:

[0043] Based on the two tangent vectors obtained in the first step, calculate the first set of basic geometric quantities of the surface. Specifically, calculate the dot product of the first tangent vector with itself, the dot product of the first tangent vector with the second tangent vector, and the dot product of the second tangent vector with itself. These dot product results form a metric tensor that describes the local distance characteristics of the surface. Then, using the second derivative rate and the normal direction vector in the first step, calculate the second set of basic geometric quantities, namely, calculate the second derivative rate in the first parameter coordinate direction, the mixed derivative rate in the two parameter coordinate directions, and the dot product of the second derivative rate in the second parameter coordinate direction with the normal direction vector. These results reflect the bending change of the surface in the normal direction. Then, construct a curvature characteristic matrix by multiplying the inverse of the metric tensor by the matrix composed of the second set of basic geometric quantities, and perform eigenvalue analysis on this matrix to obtain the principal curvatures and principal curvature directions at each point of the surface. The principal curvatures include the maximum bending degree and the minimum bending degree, as well as the corresponding two main bending directions, that is, the principal curvature directions. Finally, continuously organize these local directions into a global field to obtain the principal curvature direction field. For example, the following calculation process can be used to obtain:

[0044] Calculate the first fundamental form:

[0045] E = S u ·S u , F = S u ·S v , G = S v ·S v

[0046] where E, F, and G respectively represent the components of the surface metric tensor, the first tangent vector S u and the second tangent vector S v represent the tangent vectors of the surface in the u and v directions.

[0047] Calculate the second fundamental form:

[0048] L = S uu ·N, M = S uv ·N, N = S vv ·N

[0049] where L, M, and N reflect the curvature change of the surface in the normal direction.

[0050] Construct the curvature characteristic matrix:

[0051]

[0052] Perform eigen - decomposition on B to obtain eigenvalues and eigenvectors. Among them, the eigenvalues correspond to the principal curvatures of the surface, that is, the maximum and minimum bending degrees, and the eigenvectors correspond to the principal curvature directions, that is, the two main bending directions corresponding to the maximum and minimum bending degrees.

[0053] The maximum bending degree, the minimum bending degree, and their corresponding main bending directions are the core indicators for describing the local geometric characteristics of the surface, which can accurately reflect the bending degree of the surface in different directions. This information provides a key basis for tool - path planning, enabling the tool attitude to be adjusted according to the main bending direction of the surface, thereby avoiding path interruption or machining defects and improving the quality of the machined surface.

[0054] S1.3, Calculate the Gaussian curvature distribution:

[0055] By multiplying the maximum bending degree obtained in the second step by the minimum bending degree, a scalar measure is defined, called Gaussian curvature, which is used to represent the overall characteristics of the local geometric properties of the surface. Then, on the grid formed by the surface parameter coordinates, the Gaussian curvature at each grid point is taken to generate a distribution set containing the Gaussian curvature values of all sampling points.

[0056] As a comprehensive measure of the local geometric properties of the surface, Gaussian curvature can reflect the overall topological properties of the surface. By analyzing the Gaussian curvature distribution, special regions on the surface can be identified, such as saddle - shaped regions, convex regions, or concave regions. This information is crucial for tool - path planning, helping to avoid or specially handle these regions during the planning process, thereby ensuring the stability and consistency of the machining path.

[0057] S1.4, Generate the curvature change gradient parameter:

[0058] Based on the Gaussian curvature obtained in the third step, calculate the rate of change of the Gaussian curvature in the first parameter - coordinate direction and the rate of change in the second parameter - coordinate direction respectively to form the gradient vector of the Gaussian curvature. Specifically, by taking a small step in the parameter coordinates, calculate the difference in Gaussian curvature between adjacent grid points, and use this to approximately represent the rate of change of the Gaussian curvature in the two directions. Then, take the square root of the sum of the squares of the rate of change in the two directions to define a scalar parameter, called the curvature change gradient parameter, which is used to represent the severity of the Gaussian curvature change. For example, the following calculation process can be used to obtain it:

[0059] Calculate the gradient of the Gaussian curvature:

[0060]

[0061] where K represents the Gaussian curvature.

[0062] Use the finite - difference method for approximate calculation:

[0063]

[0064] where Δu and Δv are parameter step sizes. Define the curvature change gradient parameter:

[0065]

[0066] indicating the severity of the curvature change.

[0067] The curvature change gradient parameter can effectively capture the regions where the Gaussian curvature on the surface undergoes sudden changes, and these regions are often the difficulties in path planning. By quantifying the severity of the curvature change, it can provide a basis for tool path planning, enabling the planning algorithm to adopt special strategies in the regions of curvature mutation to avoid the breakage of the tool path due to curvature discontinuity, thereby improving the reliability and surface quality of machining.

[0068] S1.5, Extract the topological connectivity parameter:

[0069] According to the topological structure of the surface, calculate a numerical value of topological characteristics, that is, by subtracting the number of edges from the number of vertices of the surface and then adding the number of faces, a characteristic number reflecting the overall topological properties of the surface is obtained. Based on this characteristic number, judge the topological type of the surface, and further calculate the total number of mutually independent parts in the surface as the topological connectivity parameter. The topological connectivity parameter reflects the overall connectivity and structural complexity of the surface, and is of great significance for understanding the global characteristics of the surface. In tool path planning, considering the connectivity of the surface can ensure that the path covers all machining areas and avoid path incoherence caused by surface separation, thereby improving the integrity and machining efficiency of the path planning.

[0070] S1.6. Fusion to generate the curvature constraint feature matrix:

[0071] For each sampling point on the surface parameter coordinate grid, construct a feature vector containing various geometric information. Specifically, this feature vector includes the two principal curvature directions calculated by the surface feature extraction module, the Gaussian curvature calculated in step S1.3, the curvature change gradient parameter calculated in step S1.4, and the topological connectivity parameter calculated in step S1.5. Then, arrange the feature vectors of all sampling points in the row and column order of the parameter grid to form a multi-dimensional feature matrix, called the curvature constraint feature matrix, whose dimension is jointly determined by the number of rows and columns of the grid and the number of components of the feature vector.

[0072] Reasons and benefits: The curvature constraint feature matrix encodes the curvature continuity and geometric mutation features of the surface in the form of a multi-dimensional tensor by integrating the local geometric characteristics of the surface (such as the main bending direction and Gaussian curvature) and global characteristics (such as topological connectivity). This comprehensive geometric description provides a comprehensive constraint benchmark for subsequent tool path planning, ensuring that the path can maintain continuity in the curvature mutation area and adapt to the overall structure of the surface, thereby reducing machining defects and improving machining accuracy.

[0073] The curvature constraint feature matrix encodes the global curvature continuity and local geometric mutation features of the surface in tensor form, providing a geometric constraint benchmark for the trajectory generation network module. Based on the curvature constraint feature matrix and combined with the original point cloud data, the trajectory generation network module generates an initial machining trajectory that satisfies the normal vector orthogonality and tangent vector continuity constraints, laying a foundation for the inverse mapping and optimization of the trajectory effect analysis module.

[0074] The trajectory generation network module includes the following:

[0075] S2.1, Conversion of the curvature constraint feature matrix:

[0076] First, perform conversion processing on the curvature constraint feature matrix generated by the surface feature extraction module. The curvature constraint feature matrix is a multi-dimensional tensor that contains the global curvature continuity and local geometric mutation features of the surface. During the conversion process, perform normalization processing on each feature dimension of the curvature constraint feature matrix. Specifically, for each feature dimension, first calculate the minimum and maximum values in all the data of this dimension, and then perform the following operation on each eigenvalue: subtract the minimum value from the eigenvalue, and then divide the obtained result by the difference between the maximum value and the minimum value, so as to map the eigenvalue to the interval from 0 to 1. After completion of the normalization, the curvature constraint feature matrix is converted into multiple heatmap channels, each channel corresponding to a feature dimension, reflecting the relative intensity distribution of the surface geometric characteristics.

[0077] The purpose of converting the curvature constraint feature matrix into heatmap channels is to present the complex geometric features of the surface in a standardized form, facilitating subsequent processing and learning by the network model. The normalization processing ensures that the data of different feature dimensions have a unified numerical range, avoids the instability of model training caused by differences in eigenvalue ranges, and at the same time retains the relative change information of the surface curvature distribution.

[0078] S2.2, Stitching with the original point cloud data:

[0079] The heat map channel is stitched and fused with the original point cloud data. The original point cloud data contains the discrete point coordinate information on the surface. In the stitching process, first, through the bilinear interpolation method, the eigenvalue in the heat map channel is mapped to the coordinate position of the original point cloud data. Specifically, for each point coordinate in the point cloud data, find its corresponding position in the heat map channel, and calculate the interpolated eigenvalue of this point using the weighted average of the four nearest neighbor eigenvalues. The weights are based on the distance between the point coordinate and the positions of the neighboring eigenvalues. After interpolation, an interpolated feature matrix with the same number of points as the original point cloud data is generated. Then, the three-dimensional coordinates of the original point cloud data and the interpolated feature matrix are merged in the spatial dimension to form a comprehensive input data, which includes the point cloud coordinates and the corresponding curvature features.

[0080] By stitching the original point cloud data and the interpolated feature matrix of the heat map channel, it is possible to provide comprehensive input information for the subsequent network model. This information fusion method includes both the discrete geometric positions and the continuous geometric characteristics of the surface, enabling the network model to comprehensively consider the local point positions and the global curvature distribution of the surface during the learning process. Bilinear interpolation ensures the continuity and smoothness of the features in the point cloud space, thereby improving the adaptability of the generated tool path to the geometric characteristics of the surface and ensuring the accuracy and integrity of the path planning.

[0081] S2.3, Design of the geometric loss function

[0082] Construct a geometric loss function to guide the curvature adaptive convolutional network to generate the tool path coordinate sequence, ensuring that the path satisfies the normal vector orthogonality and tangent vector continuity constraints. The geometric loss function consists of two parts: the normal vector orthogonality constraint and the tangent vector continuity constraint. The calculation method of the normal vector orthogonality constraint is that for each point on the path, calculate the dot product of the path tangent direction and the surface normal vector, and sum the squared values of the dot products of all points. The goal is to make the dot product close to zero, thereby ensuring that the path tangent direction is perpendicular to the surface normal vector. The calculation method of the tangent vector continuity constraint is that for adjacent point pairs on the path, calculate the deviation between the cosine value of the included angle of the two tangent directions and 1, and sum the deviation values of all adjacent point pairs. The goal is to make the change of the tangent direction smooth. The final geometric loss function is the weighted sum of the normal vector orthogonality constraint and the tangent vector continuity constraint, and the weight coefficient is used to adjust the relative importance of the two parts of the constraints.

[0083] The construction of the geometric loss function directly targets the differential geometric continuity requirements of the tool path. Through the normal vector orthogonality constraint, it ensures that the tool posture is aligned with the surface normal, and through the tangent vector continuity constraint, it ensures the smooth transition of the path in the curvature mutation region. This loss function can effectively avoid path breakage or discontinuity phenomena and improve the accuracy and stability of path planning.

[0084] S2.4, Network training and path generation:

[0085] Input the comprehensive input data into the curvature pre-trained convolutional network, and train the network by minimizing the geometric loss function to generate the tool path coordinate sequence. The training process includes the following steps: The convolutional network generates a preliminary path coordinate sequence according to the input data, calculates the value of the geometric loss function, adjusts the network parameters through backpropagation, and repeats this process until the value of the geometric loss function converges. After training, the network outputs the initial machining trajectory that satisfies the normal vector orthogonality and tangent vector continuity constraints as the path reference for the subsequent steps.

[0086] Generating the initial machining trajectory through network training can automatically learn the geometric characteristics of the surface and generate a path that meets the requirements of differential geometric continuity. Compared with the traditional method of manually designing paths, it greatly improves the efficiency and consistency. The guidance of the geometric loss function ensures the high quality of the path and reduces the complexity of subsequent optimization and adjustment.

[0087] The pre-trained convolutional network consists of multiple convolutional modules and fully connected layers, and its core feature lies in the dynamic generation of convolutional kernel weights. Specifically, a weight generation network is constructed, which takes the interpolation feature matrix as the input, calculates through a multi-layer neural network, and outputs the convolutional kernel weights applicable to the convolutional module. These dynamically generated convolutional kernel weights are adjusted according to the local curvature characteristics of the surface, enabling the convolutional operation to adapt to the geometric characteristics of the curvature mutation region. The network gradually generates a preliminary feature representation of the tool path through layer-by-layer convolution and feature extraction.

[0088] The trajectory generation network module is based on the curvature constraint feature matrix provided by the surface feature extraction module. By converting it into a heat map channel and splicing it with the original point cloud data, it is input into the convolutional network, and uses the dynamic convolutional kernel and geometric loss function to generate the initial machining trajectory. The initial machining trajectory maintains differential geometric continuity in the curvature mutation region, providing a high-quality path reference for the inverse mapping and optimization of the subsequent trajectory effect analysis module, ensuring the stability of the machining process and the surface quality.

[0089] In free-form surface machining, tool path planning needs to ensure that the machining trajectory meets the differential geometric continuity of the surface while maintaining physical stability during the machining process. The trajectory generation network module generates the initial machining trajectory through the convolutional network based on the curvature constraint feature matrix generated by the surface feature extraction module, and this trajectory satisfies the normal vector orthogonality and tangent vector continuity constraints. However, in the curvature mutation region, the initial machining trajectory may cause sudden changes in cutting force, affecting the machining stability. The trajectory effect analysis module analyzes the residual height difference between adjacent cutting rows and the deviation of curvature continuity of the initial machining trajectory, constructs an energy scale criterion, and provides a basis for path repair of the trajectory dynamic repair module to ensure physical stability during the machining process.

[0090] The trajectory effect analysis module includes the following:

[0091] S3.1, Calculate the residual height difference between adjacent cutting rows:

[0092] Based on the initial machining trajectory generated by the trajectory generation network module, first divide it into cutting rows. Each row represents a continuous cutting path. Then, in each pair of adjacent cutting rows, select point pairs at the same parametric positions and calculate the residual height difference of these point pairs in the surface normal direction. The specific calculation process is as follows: for each pair of points, first determine the normal vector of the surface at one of the points, then calculate the position vector difference between the two points, and then obtain the residual height difference of the two points in the normal direction through the absolute value of the dot product of the normal vector and the position vector difference. This residual height difference represents the residual amount of the material in the normal direction after machining and reflects the geometric deviation of the machining trajectory. For each pair of adjacent cutting rows, calculate the average value of the residual height differences of all corresponding point pairs to obtain the average residual height difference of this pair of adjacent cutting rows.

[0093] The purpose of calculating the residual height difference between adjacent cutting rows is to quantify the geometric inconsistency of the machining trajectory. Especially in areas with large curvature changes, the residual height difference may lead to unevenness on the machined surface. By analyzing the residual height difference, areas that may cause physical instability in the machining trajectory can be identified, providing basic data for analyzing the change of cutting force. This method ensures effective control of the geometric accuracy of the machining trajectory.

[0094] S3.2, Reverse map the cutting force mutation gradient:

[0095] Based on the calculation results of the residual height difference between adjacent cutting rows, further calculate the cutting force mutation gradient. The cutting force mutation gradient is defined as the ratio of the average residual height difference between adjacent cutting rows to the average distance between adjacent cutting rows. The specific calculation process is as follows: first, average the Euclidean distances between corresponding point pairs in each pair of adjacent cutting rows to obtain the average distance between adjacent cutting rows. Then, divide the average residual height difference between adjacent cutting rows by the average distance to obtain the cutting force mutation gradient. This gradient represents the change rate of the cutting force between adjacent cutting rows and reflects the degree of dynamic fluctuation of the force during the machining process.

[0096] The calculation of the cutting force mutation gradient aims to quantify the physical instability during the machining process. Especially in areas with large residual height differences, the cutting force may change rapidly, resulting in an unstable machining process. By reverse mapping the cutting force mutation gradient, the geometric deviation can be associated with the physical and mechanical effects, providing a basis for energy analysis. This method enhances the understanding of the physical characteristics of the machining process, helps prevent machining defects and improve machining stability.

[0097] S3.3, Calculate the deviation of curvature continuity:

[0098] Based on the curvature constraint feature matrix generated by the surface feature extraction module and the initial machining trajectory of the trajectory generation network module, calculate the curvature continuity deviation. The specific calculation process is as follows: For each point on the initial machining trajectory, extract the corresponding curvature feature vector from the curvature constraint feature matrix, which reflects the local curvature attribute at that point. Then, calculate the difference between the curvature feature vectors of adjacent points to obtain the change vector of the curvature feature. Next, calculate the Euclidean norm of this change vector as the curvature continuity deviation between adjacent points. Finally, take the average of the curvature continuity deviations for all adjacent point pairs to obtain the global curvature continuity deviation, which represents the overall curvature consistency of the machining trajectory.

[0099] The purpose of calculating the curvature continuity deviation is to evaluate the geometric consistency of the machining trajectory in the regions of abrupt curvature change. Especially in the regions where the curvature changes sharply, a large deviation value may lead to an uneven machining trajectory. By quantifying the curvature continuity deviation, it can provide a basis for geometric discontinuity in energy analysis. This method ensures that the geometric continuity of the machining trajectory is effectively controlled, which helps to improve the consistency and quality of the machined surface.

[0100] S3.4, Calculate the local surface strain energy density:

[0101] First, standardize the cutting force mutation gradient and the global curvature continuity deviation so that their values are between 0 and 1. Based on the transformed cutting force mutation gradient and the global curvature continuity deviation, calculate the local surface strain energy density. The local surface strain energy density is defined as the sum of the weighted term of the square of the cutting force mutation gradient and the weighted term of the square of the global curvature continuity deviation. The specific calculation process is as follows: First, calculate the square of the cutting force mutation gradient, which represents the energy accumulation of local physical instability; then, calculate the square of the global curvature continuity deviation, which represents the influence of geometric mutation on machining stability; finally, add the two parts of energy through the weight coefficient to obtain the local surface strain energy density. The weight coefficient is calibrated according to the material properties and machining conditions to adjust the proportion of the energy contribution of different factors. For example, the following calculation method is adopted:

[0102] Define the local surface strain energy density E k , Coupling the influence of cutting force mutation and curvature continuity deviation:

[0103]

[0104] α, β: Weight coefficients, which respectively adjust the energy contribution of the cutting force mutation gradient and the curvature continuity deviation, and need to be calibrated according to the material properties and machining conditions.

[0105] The square term of the cutting force mutation gradient represents the energy accumulation of local physical instability.

[0106] The squared term of the global curvature continuity deviation represents the influence of geometric mutations on machining stability.

[0107] The calculation of the local surface strain energy density couples geometric deviations and physical and mechanical effects, comprehensively evaluating the stability during the machining process. By mapping the cutting force mutation and the curvature continuity deviation to a unified energy scale, the stability of the machining trajectory can be analyzed more comprehensively.

[0108] S3.5, Construct the energy scale criterion:

[0109] Based on the calculation results of the local surface strain energy density, construct the energy scale criterion. The energy scale criterion is defined as the maximum value in the sequence of local surface strain energy density, reflecting the energy level of the most unstable region during the machining process. It is used to evaluate whether the machining trajectory may cause plastic deformation of the material or other physical instabilities, providing a decision basis for subsequent path repair.

[0110] The purpose of constructing the energy scale criterion is to provide a clear judgment standard for path repair by quantifying the maximum instability during the machining process. By comparing the energy scale criterion with the critical value of the material's plastic deformation, it can be determined whether the machining trajectory needs to be adjusted, thereby preventing machining defects and improving machining quality.

[0111] Based on the initial machining trajectory of the trajectory generation network module and the curvature constraint feature matrix of the surface feature extraction module, the trajectory effect analysis module calculates the residual height difference between adjacent cutting rows, the reverse mapping cutting force mutation gradient, the curvature continuity deviation, and the local surface strain energy density, and finally constructs the energy scale criterion. This criterion provides a clear stability evaluation basis for the path repair of the trajectory dynamic repair module, ensuring effective control of geometric continuity and physical stability during the machining process.

[0112] The trajectory effect analysis module constructs the energy scale criterion by analyzing the residual height difference between adjacent cutting rows and the curvature continuity deviation of the initial machining trajectory, which is used to evaluate the stability of the machining trajectory. When the energy scale criterion exceeds the critical value of the material's plastic deformation, it indicates that the machining trajectory may cause irreversible deformation of the material and path repair is required. Based on this, the trajectory dynamic repair module generates a transitional spline curve according to the curvature constraint feature matrix for the area where the energy scale exceeds the standard, and optimizes the geometric parameters and physical stability of the repair path through the dynamic adaptation of the curvature decay rate and the energy dissipation relationship, providing a high-quality repaired path for the finite element simulation and optimization of the simulation optimization control module.

[0113] The trajectory dynamic repair module includes the following:

[0114] S4.1, Determine whether the energy scale breaks through the critical value of the material's plastic deformation:

[0115] The trajectory dynamic repair module first determines whether there are areas at risk of instability in the machining trajectory based on the energy scale criterion generated by the trajectory effect analysis module and the pre-determined critical value of material plastic deformation. The determination process is as follows: Numerically compare the energy scale criterion generated by the trajectory effect analysis module with the critical value of material plastic deformation. When the energy scale criterion exceeds the critical value of material plastic deformation, the corresponding area is determined as an over-standard area and subsequent path repair is required. The critical value of material plastic deformation is determined through mechanical property experiments of the workpiece material and represents the energy limit level at which the material undergoes irreversible deformation during machining. The marking of the over-standard area provides a clear scope basis for path repair.

[0116] By comparing the energy scale criterion with the critical value of material plastic deformation, it is possible to accurately identify the areas in the machining trajectory that may cause material plastic deformation. This judgment method sets clear starting conditions for path repair, ensuring that the repair operation is carried out for specific problems, avoiding unnecessary processing of areas that do not require adjustment, thereby improving the utilization efficiency of computing resources and the stability of the machining process.

[0117] S4.2, Generate a transition spline curve according to the curvature constraint feature matrix constraint:

[0118] After determining the over-standard area, the trajectory dynamic repair module generates a transition spline curve according to the curvature constraint feature matrix to repair the geometric defects of the machining trajectory. The generation process is as follows: First, extract the principal curvature direction field and Gaussian curvature distribution of the over-standard area from the curvature constraint feature matrix as the geometric constraint conditions for the transition spline curve. Then, between adjacent cutting rows in the over-standard area, interpolation is performed along the principal curvature direction field, that is, the main bending direction corresponding to the maximum and minimum bending degrees, to construct a transition spline curve connecting the starting points of adjacent cutting rows. The construction method uses cubic B-spline interpolation technology to ensure that the curve has at least second-order continuity geometrically and meets the boundary conditions and curvature continuity requirements. The boundary conditions stipulate that the starting point and ending point of the transition spline curve coincide with the starting point and ending point of the adjacent cutting rows in the over-standard area respectively; the curvature continuity requirement is that the curvature change of the transition spline curve matches the local Gaussian curvature distribution of the surface, and the matching error is controlled within the tolerance range specified by the machining accuracy.

[0119] Generating a transition spline curve according to the curvature constraint feature matrix constraint ensures that the repair path is consistent with the geometric characteristics of the free-form surface, avoiding machining defects caused by path mutations. Using cubic B-spline interpolation technology ensures the smoothness and continuity of the curve, reducing vibration and stress concentration problems caused by path discontinuity during machining. At the same time, by matching the local Gaussian curvature distribution of the surface, the repair path is highly compatible with the surface characteristics, improving the accuracy and quality of the machining surface.

[0120] S4.3, Dynamically adapt the geometric parameters and physical stability of the repair path:

[0121] After generating the transition spline curve, dynamically adapt to optimize the geometric parameters and physical stability of the repair path. The adaptation process is as follows: First, calculate the rate of change of the curvature of the transition spline curve with respect to the parameter, which is defined as the curvature decay rate and represents the speed of change of the curve curvature. Then, introduce an energy dissipation function to describe the energy release situation of the repair path at different positions. The specific calculation method is to multiply the square of the curvature decay rate by the dissipation coefficient, and the dissipation coefficient is determined through experimental calibration of material properties and processing conditions. Then, by adjusting the control points of the transition spline curve, minimize the difference between the local surface strain energy density and the energy dissipation function. The optimization goal is to make the energy dissipation distribution of the repair path match the strain energy density of the out-of-specification area, reduce the impact of energy mutation, and at the same time keep the boundary conditions and curvature continuity requirements unchanged.

[0122] By dynamically adapting the geometric parameters and physical stability of the repair path, it is possible to optimize the physical characteristics during the processing while maintaining geometric continuity. Adjust the curvature decay rate and the energy dissipation relationship to ensure that the repair path effectively releases energy in the out-of-specification area and reduce the risks of stress concentration and material deformation during the processing. This method realizes the dual optimization of geometric repair and physical stability, improves the reliability and surface quality of the processing, and provides a stable path basis for subsequent processing.

[0123] The trajectory dynamic repair module identifies the out-of-specification area by judging whether the energy scale breaks through the critical value of material plastic deformation, generates a transition spline curve according to the curvature constraint feature matrix constraint, and optimizes the geometric parameters and physical stability of the repair path through dynamic adaptation, and finally generates the repaired processing trajectory. The repaired processing trajectory realizes the dual guarantee of geometric continuity and physical stability in the curvature mutation area, and provides a high-quality input basis for subsequent finite element simulation and optimization.

[0124] The trajectory dynamic repair module generates a transition spline curve for the area where the energy scale exceeds the standard and outputs the repaired processing trajectory, which is optimized in terms of geometric continuity and physical stability. The simulation optimization control module takes over the repaired processing trajectory of the trajectory dynamic repair module, conducts finite element simulation verification through the digital twin platform, and further optimizes the trajectory to meet the requirements of processing stability and surface quality.

[0125] The simulation optimization control module includes the following:

[0126] S5.1, Import the repaired processing trajectory into the digital twin platform

[0127] The repaired machining trajectory first needs to be discretized. Specifically, the continuous parametric curve output by the trajectory dynamic repair module is decomposed into a series of discrete three-dimensional coordinate points. These three-dimensional coordinate points are generated by uniform sampling or adaptive sampling according to curvature changes, ensuring that the point set can completely characterize the geometric features of the machining trajectory. The discretized three-dimensional coordinate points are used as geometric inputs and imported into the digital twin platform to establish a digital model of the tool path. The digital twin platform then generates a virtual representation of the machining trajectory based on these coordinate points, laying the foundation for subsequent simulation analysis.

[0128] Reasons and benefits: Decomposing the continuous machining trajectory into discrete three-dimensional coordinate points facilitates geometric modeling and simulation calculations on the digital twin platform. The discretization process can transform complex curve trajectories into point set data that can be processed by a computer, ensuring that the input data for simulation analysis is accurate and easy to operate.

[0129] S5.2. Conduct finite element simulation to verify the cutting stress distribution and the dynamic cutting force fluctuation range:

[0130] In the digital twin platform, use finite element analysis tools to establish mesh models of the workpiece and the tool. The specific process includes dividing the geometric shape of the workpiece into multiple small elements to form a mesh structure, and at the same time performing a similar mesh division on the tool model. Based on the discretized machining trajectory, simulate the behavior of the tool moving along the trajectory and gradually removing materials during the machining process. During this process, calculate the maximum principal stress and the minimum principal stress of each mesh element on the workpiece surface to generate a cutting stress distribution map, which is used to characterize the stress state on the workpiece surface during the machining process. In addition, through the cutting force physical model, calculate the change of the dynamic cutting force received by the tool over time during the machining process, record the maximum value and the minimum value of the cutting force, and then determine the fluctuation range of the dynamic cutting force.

[0131] Through finite element simulation, the material removal behavior during the machining process can be accurately simulated, and the cutting stress distribution on the workpiece surface and the change of the dynamic cutting force borne by the tool can be analyzed. It can identify stress concentration areas or problem points with excessive cutting force fluctuations that may be caused by the machining trajectory. The simulation verification predicts the performance of the machining trajectory before actual machining, helps avoid potential machining defects, and ensures the stability of the machining process and the compliance of the workpiece surface quality.

[0132] S5.3. Screen the optimized trajectory that meets the preset stress constraints:

[0133] Based on the cutting stress distribution map and the dynamic cutting force fluctuation range generated by finite element simulation, constraint condition checking is carried out. Specifically, it is checked whether the maximum principal stress of some mesh elements in the cutting stress distribution map exceeds the preset threshold, and at the same time, it is evaluated whether the dynamic cutting force fluctuation range is within the predefined allowable range. If it is found that the maximum principal stress exceeds the threshold or the cutting force fluctuation range exceeds the allowable range, local adjustment is made to the repaired machining trajectory, such as modifying the curvature of the trajectory or the tool feed rate, to generate a new machining trajectory. Subsequently, the above finite element simulation process is repeated until both the cutting stress distribution and the dynamic cutting force fluctuation range meet the preset constraint conditions, and finally the optimized machining trajectory is output.

[0134] Screen the optimized trajectories that meet the preset stress constraints to ensure that the maximum principal stress on the workpiece surface during the machining process does not exceed the yield strength of the material, and avoid material deformation or damage caused by excessive stress. At the same time, by controlling the dynamic cutting force fluctuation range within the allowable range, the vibration and instability during the machining process are reduced, thereby improving the machining accuracy and the quality of the workpiece surface. This iterative optimization method can improve the machining trajectory in a virtual environment and reduce the risks in actual machining.

[0135] S5.4, synchronously update the optimized trajectory parameters with the acceleration feedforward coefficient of the servo system:

[0136] Extract the speed and acceleration parameters of the tool movement from the optimized machining trajectory. Specifically, by analyzing the distance and time interval between each discrete point in the machining trajectory, the speed of the tool in each section is calculated, and further the acceleration is calculated through the change of speed over time. According to the dynamic response model of the servo system, the acceleration feedforward coefficient is determined, which is calculated by the ratio of acceleration to speed. Subsequently, the extracted speed, acceleration, and the calculated acceleration feedforward coefficient are updated to the control instruction stream of the servo system to achieve dynamic matching between the control instruction and the motion state of the machining trajectory.

[0137] Synchronously updating the speed, acceleration parameters and the acceleration feedforward coefficient of the optimized trajectory can improve the control performance of the servo system, reduce the tracking error during the tool movement, and improve the execution accuracy of the machining trajectory. By dynamically matching the control instruction with the machining state, it is ensured that the servo system can respond to the changes of the tool path in real time and optimize the stability of the machining process. This closed-loop coordination method enhances the response speed of the machining equipment and provides a guarantee for high-precision machining.

[0138] The simulation optimization control module verifies the cutting stress distribution and the dynamic cutting force fluctuation range of the repaired machining trajectory through the finite element simulation of the digital twin platform to ensure that it meets the preset constraint conditions. On this basis, the optimized machining trajectory is selected, and the speed and acceleration parameters of the trajectory are synchronously updated with the acceleration feedforward coefficient of the servo system to achieve the dynamic closed-loop matching of the control instruction flow and the physical machining state. Through the comprehensive processing of simulation verification and parameter optimization, the stability in the free-form surface machining process is ensured, and at the same time, the surface quality of the workpiece is improved, providing reliable technical support for high-precision machining.

[0139] The above formulas are all dimensionless and take their numerical values for calculation. The formula is obtained by collecting a large amount of data for software simulation to get a formula that is closest to the actual situation. The preset parameters in the formula are set by those skilled in the art according to the actual situation.

[0140] It should be noted that the system of the present invention can be deployed on the device itself to achieve embedded applications, or can also run on a PC or other terminals with a user interface, so as to meet various hardware environments and usage requirements.

[0141] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, without departing from the spirit and scope of the present invention, the described embodiments can be modified in various different ways. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

[0142] It should be noted that in this article, if there are relational terms such as first and second, they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "including a..." does not exclude the existence of another identical element in the process, method, article or device including the element.

[0143] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A numerical control machining path control system based on artificial intelligence, characterized in that, Including the steps: Surface feature extraction module: Based on the surface geometric properties of the workpiece model, obtain the principal curvature direction field and Gaussian curvature distribution, and fuse the curvature change gradient parameter and topological connectivity parameter to generate a curvature constraint feature matrix; Trajectory generation network module: Convert the curvature constraint feature matrix into a heatmap channel, and after splicing it with the original point cloud data in the spatial dimension, input it into a pre-trained convolutional network to generate an initial machining trajectory; Trajectory effect analysis module: Based on the residual height difference between adjacent cutting rows of the initial machining trajectory, inversely map the cutting force mutation gradient, and combine the curvature continuity deviation to calculate the local surface strain energy density, and construct an energy scale criterion; Trajectory dynamic repair module: When the energy scale criterion breaks through the critical value of material plastic deformation, generate a transition spline curve based on the curvature constraint feature matrix, and repair the geometric parameters and physical stability of the path through the relationship between the curvature decay rate and energy dissipation; Simulation optimization control module: Import the repaired path into the digital twin platform for finite element simulation to verify the cutting stress distribution and dynamic cutting force fluctuation range; Screen the optimized trajectory that meets the preset stress constraints, and synchronously update the trajectory parameters and the acceleration feedforward coefficient of the servo system.

2. The numerical control machining path control system based on artificial intelligence according to claim 1, wherein, The surface feature extraction module includes the following: Extract the implicit representation of the surface from the computer-aided design model of the workpiece, and parameterize it as a function depending on two parameter coordinates; Calculate the first-order and second-order change rates of the surface in the two parameter coordinate directions, and the normal direction vector of the surface based on the surface implicit representation; Calculate the first set of basic geometric quantities and the second set of basic geometric quantities using the tangent vector and normal direction vector of the surface, construct a curvature feature matrix, and determine the principal curvature and principal curvature direction in the curvature constraint feature matrix through feature analysis, that is, the maximum bending degree, the minimum bending degree, and their corresponding main bending directions.

3. The numerical control machining path control system based on artificial intelligence according to claim 2, characterized in that, The surface feature extraction module also includes the following: Calculate the Gaussian curvature by multiplying the maximum bending degree and the minimum bending degree, and generate a Gaussian curvature distribution on the parameter grid; Calculate the change rates of the Gaussian curvature in the two parameter coordinate directions to form a curvature change gradient parameter; Calculate the topological characteristic value and the number of connected components according to the topological structure of the surface to determine the topological connectivity parameter; For each sampling point on the surface, construct a feature vector containing the main bending direction, Gaussian curvature, curvature change gradient parameter, and topological connectivity parameter, and arrange all feature vectors according to the parameter grid to form a curvature constraint feature matrix.

4. The numerically controlled machining path control system based on artificial intelligence according to claim 3, wherein, The trajectory generation network module includes the following: Normalize each feature dimension of the curvature constraint feature matrix to generate a heatmap channel; Map the heatmap channel to the coordinate position of the original point cloud data through bilinear interpolation to generate an interpolation feature matrix, and splice it with the original point cloud data in the spatial dimension to form comprehensive input data; Input the comprehensive input data into a pre-trained convolutional network, and train it by minimizing the geometric loss function to generate an initial machining trajectory that meets the constraints of normal vector orthogonality and tangent vector continuity.

5. The numerically controlled machining path control system based on artificial intelligence according to claim 4, characterized in that, The geometric loss function is constructed as follows: Construct a geometric loss function that includes normal vector orthogonality constraints and tangent vector continuity constraints. By calculating the dot product of the path tangent direction and the surface normal vector, as well as the deviation of the cosine value of the included angle between adjacent path point tangent directions from 1, ensure that the path satisfies differential geometric continuity.

6. The numerical control machining path control system based on artificial intelligence according to claim 4, wherein The trajectory effect analysis module includes the following: Calculate the residual height difference between adjacent cutting rows in the surface normal direction based on the initial machining trajectory, and inversely map the cutting force mutation gradient through the ratio of the residual height difference to the average spacing between adjacent cutting rows; calculate the Euclidean norm of the curvature eigenvector between adjacent points based on the curvature constraint feature matrix and the initial machining trajectory, and then calculate the global curvature continuity deviation; calculate the local surface strain energy density through the sum of the weighted term of the square of the cutting force mutation gradient and the weighted term of the square of the global curvature continuity deviation; extract the maximum value in the local surface strain energy density sequence as the energy scale criterion.

7. The numerical control machining path control system based on artificial intelligence according to claim 6, wherein The trajectory dynamic repair module includes the following: Judge whether the energy scale criterion exceeds the critical value of material plastic deformation, and identify the over-standard area by comparing the energy scale criterion with the critical value of material plastic deformation; generate a transition spline curve based on the curvature constraint feature matrix, extract the maximum bending degree, minimum bending degree, their corresponding main bending directions, and the Gaussian curvature distribution from the curvature constraint feature matrix, and use cubic B-spline interpolation technology to construct the transition spline curve, ensuring that the boundary conditions and curvature continuity requirements are met; dynamically adapt the geometric parameters and physical stability of the repair path by calculating the curvature decay rate and energy dissipation function of the transition spline curve, and adjusting the control points to minimize the difference between the local surface strain energy density and the energy dissipation function.

8. The numerically controlled machining path control system based on artificial intelligence according to claim 7, characterized in that, The simulation optimization control module includes the following: Discretize the repaired machining trajectory into three-dimensional coordinate points and import them into the digital twin platform to establish a tool path model; analyze the cutting stress distribution and dynamic cutting force fluctuation range on the workpiece surface through finite element simulation, calculate the maximum principal stress and minimum principal stress on the workpiece surface, as well as the peak and valley values of the dynamic cutting force; screen and optimize the trajectory according to the preset stress constraint conditions, check whether the maximum principal stress exceeds the preset threshold and whether the cutting force fluctuation range is within the allowable range, and if not, adjust the trajectory and repeat the simulation; extract the speed and acceleration parameters of the tool movement from the optimized machining trajectory, and calculate the acceleration feedforward coefficient; update the speed, acceleration, and acceleration feedforward coefficient to the control instruction stream of the servo system to achieve dynamic matching of the control instruction and the machining state.

Images