High-precision part machining path planning method based on artificial intelligence optimization

Abstract

The invention discloses a high-precision part machining path planning method based on artificial intelligence optimization, and the method comprises the following steps: generating a first three-dimensional geometric model of a to-be-machined part, carrying out the curved surface division, and obtaining a manufacturing feature point according to a division result; determining a processing placement surface for the first three-dimensional geometric model according to the manufacturing feature points, and performing processing element positioning and shaping to obtain a second three-dimensional geometric model; processing path preliminary planning is carried out, optimization is carried out based on mixed # imgabs0 # equilibrium by using a neural network, and a second planning processing path is output; performing path optimization based on a Gaussian distribution function and a cubic # imgabs1 # spline curve to obtain a third planning processing path; and simulation verification is carried out, path adjustment and optimization are carried out by adopting a hierarchical optimization ant colony algorithm according to a verification result, and a final planning processing path is output. The precision, the efficiency, the flexibility and the adaptability in the high-precision part machining process are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of computer-aided manufacturing, and in particular to a high-precision parts machining path planning method based on artificial intelligence optimization. Background Art

[0002] With the continuous improvement of the precision, efficiency and intelligence level of parts processing in modern manufacturing, how to achieve efficient processing of high-precision parts in complex processing environments has become an important challenge in the manufacturing field. Traditional part processing path planning methods usually rely on empirical rules and manual intervention, and there are problems such as low path planning efficiency, difficulty in ensuring processing accuracy and long processing time. Especially when facing parts with complex shapes, traditional methods often have difficulty in handling their complex geometric features, resulting in difficulties in optimizing the processing path, and may even cause collisions and errors during the processing process.

[0003] With the rapid development of artificial intelligence technology, path planning methods based on artificial intelligence have gradually attracted widespread attention. In particular, by using neural networks, optimization algorithms, simulation verification and other means, it is possible to automatically generate and optimize processing paths according to the geometric characteristics of the processed parts and the processing environment, which can not only improve the processing accuracy, but also significantly improve the efficiency of path planning and reduce manual intervention. However, the existing path planning methods based on artificial intelligence still have certain limitations. For example, traditional neural network optimization methods cannot effectively handle small errors in high-precision processing, and smoothness and collision detection in the path optimization process often face great challenges. In addition, the real-time and adaptive capabilities of path planning have not been fully improved, resulting in potential path planning errors in the processing of some high-precision parts.

[0004] Therefore, how to propose an efficient, intelligent and accurate part processing path planning method by combining artificial intelligence algorithms, optimization technology and high-precision processing characteristics has become an important issue that needs to be urgently solved in the current manufacturing field. Summary of the invention

[0005] In order to overcome the shortcomings of the prior art, the present invention provides a high-precision parts machining path planning method based on artificial intelligence optimization, which is used to solve the technical problems that the existing machining path planning method cannot effectively handle small errors, the real-time performance of path planning and the lack of adaptability, thereby improving the accuracy, efficiency, flexibility and adaptability of high-precision parts machining.

[0006] To solve the above problems, the technical solution adopted by the present invention is as follows: A high-precision parts machining path planning method based on artificial intelligence optimization includes the following steps: Generate a first three-dimensional geometric model of the part to be processed and perform surface division, and obtain manufacturing feature points according to the division results; Determine the processing placement surface for the first three-dimensional geometric model according to the manufacturing feature points, perform positioning and shaping of the processing elements, and obtain the second three-dimensional geometric model; Perform preliminary planning of machining paths and use neural networks based on hybrid Balanced optimization is performed and the second planned processing path is output; Based on Gaussian distribution function, cubic The spline curve is used to optimize the path and obtain the third planning processing path; Simulation verification is carried out, and based on the verification results, the hierarchical optimization ant colony algorithm is used to adjust and optimize the path, and the final planned processing path is output.

[0007] As a preferred embodiment of the present invention, when generating the first three-dimensional geometric model, it includes: Get the geometry of the part to be processed and all its edges, and put them into the edge collection ; Get edge set The edges with arc type are put into the arc edge sub-collection middle; Get the arc center points of all edges in the arc edge subset and put them into the center point collection The center point number is ; Provides a temporary maximum circle center point set , the maximum circle center point set ,parameter ,parameter ; like , then execute the parallel judgment steps; like , then execute the first three-dimensional geometric model generation step.

[0008] As a preferred embodiment of the present invention, when performing surface division, it includes: Get the surface area in the first 3D geometric model and perform Spline surface configuration; according to The spline surface configuration obtains the first-order and second-order partial derivatives of the surface area; The regions where the deviations of the first-order and second-order partial derivatives are both smaller than a threshold are classified as a region set, thereby dividing the surface region into several surface sub-regions.

[0009] As a preferred embodiment of the present invention, when obtaining manufacturing feature points, it includes: Convex edge decomposition is performed based on several surface sub-regions, and the base surface is matched with the predefined manufacturing template to obtain the manufacturing feature points. , while giving the geometric surface mapping attributes; When determining the processing placement surface, include: Get manufacturing feature points The face set , get the face set from the first 3D geometric model Adjacent face sets ; If the face collection If there is a unique surface, the unique surface is used as the processing placement surface ; If the face collection If there are multiple faces, obtain the faces parallel to the base face among the multiple faces, and take the face with the largest distance from the base face as the processing placement face .

[0010] As a preferred embodiment of the present invention, when positioning the processing element, it includes: From manufacturing feature points Get the positioning surface , positioning surface , and align the positioning surface , positioning surface The type is identified and the corresponding processing element coordinate system is constructed according to the identification result; When processing element shaping, including: Traversing manufacturing feature points The parameter variable set , through the parameter variable set The name of the parameter variable determines the meaning of the parameter variable. According to the association relationship between the parameter variable and the attribute adjacency graph, the geometric surface associated with the parameter variable is obtained, and the value of the parameter variable is determined.

[0011] As a preferred embodiment of the present invention, when performing preliminary planning of the processing path, it includes: Constructing a road network 0-1 matrix for the processing path according to the second three-dimensional geometric model; Get manufacturing feature points The connectivity index is The algorithm obtains any two manufacturing feature points The matrix of the shortest path between them, obtaining all manufacturing feature points The betweenness is calculated and the betweenness index is analyzed to obtain the first planned processing path.

[0012] As a preferred embodiment of the present invention, when outputting the second planned processing path, it includes: Divide the first planned processing path into several equal parts according to the preset time interval , , The start and end time of the processing path; Get The position of the machining tool at the moment , The manufacturing feature point closest to the machining tool at the moment Location ; Get the waiting position of the machining tool when starting machining , the first manufacturing feature point on the first planned processing path Location ; Based on time and location and A rectangular coordinate system with as the origin.

[0013] As a preferred embodiment of the present invention, when outputting the second planned processing path, it also includes: Will Coordinates at the moment and As the training sample input of the neural network, Moment and The position difference corresponding to the moment uses the error function of the neural network The gain factor is obtained as output; The position change at the next moment is re-input into the neural network, and the gain coefficient updated in real time is output for adaptive optimization until the hybrid-based Balance and output the second planned processing path.

[0014] As a preferred implementation mode of the present invention, when performing path optimization, it includes: Define a Gaussian distribution function in the processing environment, perform random sampling within a preset range close to obstacles, and generate random sampling points , select random sampling points from the distance according to the Gaussian distribution function for , and the point located on the second planned processing path ; judge and Are the collision conditions the same? If so, the corresponding If they are not the same, the corresponding Keep it and get the processing path without collision points; Through three times The spline curve is used to smooth the planned processing path without collision points to obtain the third planned processing path; in, The preset distance.

[0015] As a preferred implementation mode of the present invention, when performing path adjustment and optimization, it includes: The manufacturing feature points on the third planned processing path according to The axis coordinate values ​​are divided into several sub-areas; Treat each sub-area as a small Use ant colony algorithm to optimize and obtain optimized data sets for each sub-area; The optimized data set of each sub-area is used as a new small-scale , and the ant colony algorithm is used again for optimization to obtain the final planned processing path; The ant colony algorithm includes: each ant starts from the starting point of the third planned processing path, moves along the third planned processing path, and gradually adjusts the optimized path according to the pheromone concentration and heuristic information.

[0016] Compared with the prior art, the present invention has the following beneficial effects: (1) Improve processing accuracy and efficiency The present invention realizes accurate planning of machining paths for high-precision parts through an optimization method based on artificial intelligence (including neural networks and ant colony algorithms). Compared with traditional path planning methods, the present invention can dynamically adjust the machining path according to the geometric characteristics of the parts and the machining environment, thereby effectively reducing errors and improving the dimensional accuracy, surface roughness and shape accuracy of the parts. The optimized path not only reduces the machining time, but also effectively avoids collisions and unnecessary machining in the path, significantly improving the machining efficiency.

[0017] (2) Intelligent path optimization and real-time adaptive adjustment The invention combines neural network and hybrid equalization technology to adjust the processing path according to real-time feedback, especially in complex processing environments, and can adaptively optimize the path according to the real-time position and state of the processing tool. This intelligent path planning method makes the processing process more efficient and flexible, and can adapt to the needs of different working conditions and part complexity.

[0018] (3) Applicability to various processing scenarios By introducing Gaussian distribution function and cubic spline curve optimization technology in machining path planning, the present invention can provide the optimal machining path in various complex machining scenarios (such as machining of high-precision and high-complexity parts, difficult-to-machine materials, etc.). Especially in the case of many obstacles or special machining requirements, it can effectively avoid collisions with obstacles, and ensure the continuity and smoothness of the machining path through smooth optimization processing, thereby improving machining quality and tool life.

[0019] (4) Simulation verification and comprehensive evaluation mechanism The present invention ensures the actual feasibility and processing effect of the planned path through tool path simulation and multi-dimensional verification mechanisms such as surface roughness, dimensional accuracy, and shape error. Advanced measurement technologies such as high-precision probes, three-coordinate measuring machines, and optical scanners are used to comprehensively evaluate the accuracy of parts after processing, providing a scientific basis for subsequent processing, thereby further optimizing the processing process and reducing the defect rate and rework rate.

[0020] (5) Optimized calculation method and path accuracy The present invention uses the ant colony algorithm to perform hierarchical optimization on the processing path, and improves the accuracy and optimization degree of path planning by intelligently optimizing the processing path of each small block. The dynamic adjustment of the pheromone concentration on the path ensures the effectiveness of the optimization process, so that the final planned path is not only accurate, but also takes into account various constraints in actual processing, such as the characteristics of the processing tool and the rationality of the processing sequence.

[0021] (6) Ability to adapt to complex geometric features By introducing technologies such as 3D geometric model construction based on arc recognition, surface division, manufacturing feature point acquisition, processing placement surface determination, processing element positioning, and processing element shaping, the present invention can process parts with complex shapes, especially those containing multiple different surface types (such as concave and convex shapes, ellipses, cylinders, etc.). When processing these parts, the most suitable processing path can be provided for each different geometric feature, and the processing error can be minimized through intelligent optimization.

[0022] (7) Applicable to various manufacturing environments This method is applicable to a variety of manufacturing environments and has strong versatility and scalability. In addition, combined with virtual simulation technology, simulation verification can be performed before actual processing to avoid potential problems in processing and improve the controllability and safety of the processing process.

[0023] The present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1It is a step diagram of the high-precision parts machining path planning method based on artificial intelligence optimization provided by the present invention; Figure 2 It is a schematic diagram of the hierarchical optimization ant colony algorithm provided by the present invention. DETAILED DESCRIPTION

[0025] The high-precision parts machining path planning method based on artificial intelligence optimization provided by the present invention is as follows: Figure 1 As shown, the following steps are included: Step S1: generating a first three-dimensional geometric model of a part to be processed, performing surface division on the first three-dimensional geometric model, and obtaining manufacturing feature points according to the division result; Step S2: determining a processing placement surface for the first three-dimensional geometric model according to the manufacturing feature points, positioning the processing elements, and shaping the processing elements to obtain a second three-dimensional geometric model; Step S3: Preliminary planning of the machining path is performed based on the second 3D geometric model and manufacturing feature points, and a neural network is used to Balanced optimization is performed and the second planned processing path is output; Step S4: The second planned processing path is based on the Gaussian distribution function, the cubic The spline curve is used to optimize the path and obtain the third planning processing path; Step S5: Simulate and verify the third planned processing path to obtain verification results of collision, smoothness, stress, surface roughness, dimensional accuracy, and shape error; Step S6: Based on the verification results, the hierarchical optimization ant colony algorithm is used to adjust and optimize the path, and the final planned processing path is output.

[0026] In the above step S1, when generating the first three-dimensional geometric model of the part to be processed, it includes: Get the geometry of the part to be processed, and based on the geometry, get all the edges of the part to be processed, and put them into the edge collection Among them, For the edge, is the number of sides; Get edge set All edge types in , and take out the edges with arc type, and put them into the arc edge subset middle; Get the arc edge subset The arc center points of all the edges in the circle are put into the center point set Among them, is the center point of the arc, is the center point of the circle; Provides a temporary maximum circle center point set and the maximum circle center point set , where, in the initial state, the temporary maximum circle center point set and the maximum circle center point set are all empty sets; Provide parameters ,parameter , where, in the initial state, the parameters ,parameter ; like , then execute the parallel judgment step, which includes: taking out the center point set In and , and order , determine the arc center point and and The vector composed of and Are they parallel? If so, set the arc center point Add to the temporary maximum circle center point set If not parallel, then ,until equal , execute the assignment step; like , then execute the first three-dimensional geometric model generation step; The assignment step includes: , , if the temporary maximum circle center point set The number of elements is less than the maximum circle center point set , then continue to compare and , and executing a parallel determination step or a first three-dimensional geometric model generation step according to the comparison result; If the temporary maximum circle center point set The number of elements is greater than or equal to the maximum circle center point set , then the temporary maximum circle center point set Assign to the maximum circle center point set Then, continue to compare and , and executing a parallel determination step or a first three-dimensional geometric model generation step according to the comparison result; The first three-dimensional geometric model generation step includes: Get the largest circle center point set Any two points in and As the center axis vector of the part to be processed ; Remove the edge set All the edges in the point set are obtained through the point set function, and all the points on all the edges are put into the point set middle; Get point collection Each point in the vector to the central axis The vertical distance of the workpiece is calculated, and the maximum value is taken as the radius of the workpiece to be processed; Get point collection Each point in The vectors formed are along the central axis vector Maximum projection distance , Minimum projection distance , and the maximum projection distance Minimum distance from projection The difference is taken as the height of the part to be processed; Increase the height and radius allowance, and create a sketch along the center axis vector Perform extrusion to generate the first three-dimensional geometric model.

[0027] In the above step S1, when the first three-dimensional geometric model is divided into surfaces, it includes: Get the surface area in the first 3D geometric model and perform The spline surface configuration is shown in Formula 1: (1); In the formula, For the surface area Spline surface, is the horizontal curve in the surface area, is the curve in the vertical direction of the surface area, For the configuration The control points behind the spline surface, For the configuration The weights corresponding to the control points behind the spline surface, For the configuration Spline surface is the spline in the horizontal direction, For the configuration Spline surface splines in the vertical direction, For the configuration The number of splines of the spline surface in the horizontal direction, For the configuration The number of splines of the spline surface in the vertical direction; according to The spline surface configuration obtains the first-order partial derivative and the second-order partial derivative of the surface area, as shown in Formula 2 and Formula 3: (2); In the formula, is the first-order partial derivative of the surface area, is the first-order partial derivative result of the surface area along the horizontal direction, is the first-order partial derivative result of the surface area along the vertical direction, is the horizontal curve in the surface area, is the curve in the vertical direction of the surface area; (3); In the formula, is the second-order partial derivative of the surface area, is the second-order partial derivative result of the surface area along the horizontal direction, is the second-order partial derivative result of the surface area along the vertical direction, It is the second-order partial derivative result of the surface area first along the vertical direction and then along the horizontal direction; It is the second-order partial derivative result of the surface area first along the horizontal direction and then along the vertical direction; The regions where the deviations of the first-order partial derivative and the second-order partial derivative are both less than the threshold are classified into a region set, thereby dividing the surface region into a number of surface sub-regions; Among them, several curved surface sub-areas include: a plane type curved surface area, a concave cylindrical type curved surface area, a concave elliptical type curved surface area, a concave saddle type curved surface area, a convex cylindrical type curved surface area, a convex elliptical type curved surface area, and a convex saddle type curved surface area.

[0028] In the above step S1, when obtaining the manufacturing feature points, it includes: Convex edge decomposition is performed based on several surface sub-regions, and the base surface is matched with the predefined manufacturing template to obtain the manufacturing feature points. , while giving the geometric surface mapping attributes; When determining the processing placement surface, include: Get manufacturing feature points The face set , get the face set from the first 3D geometric model Adjacent face sets ; If the face collection If there is a unique face, the unique face will be used as the manufacturing feature point Processing placement surface ; If the face collection If there are multiple faces, obtain the faces parallel to the base face among the multiple faces, and take the face with the largest distance from the base face among all parallel faces as the manufacturing feature point Processing placement surface .

[0029] In the above step S2, when positioning the processing element, it includes: From manufacturing feature points Get the positioning surface , positioning surface , and align the positioning surface , positioning surface Identify the type of The recognition results include: positioning surface , positioning surface All are the same cylindrical surface; positioning surface Flat surface, positioning surface Cylindrical surface; positioning surface , positioning surface All are flat; If the positioning surface , positioning surface They are all the same cylindrical surface: then get the cylindrical surface axis and the processing placement surface The intersection of As the origin of the machining element coordinate system; obtain the machining placement surface The normal vector , according to the normal vector Get the vector perpendicular to it , and obtain the perpendicular to the normal vector and vector Vector ; The vector ,vector and the normal vector As the machining element coordinate system axis; If the positioning surface Flat surface, positioning surface For a cylindrical surface: Get the positioning surface The axis and processing placement surface The intersection of As the origin of the machining element coordinate system; obtain the machining placement surface The normal vector and positioning surface The normal vector , and obtain the perpendicular to the normal vector and the normal vector Vector ; The normal vector ,vector and the normal vector As the machining element coordinate system axis; If the positioning surface , positioning surface All are planes: get the positioning surface , positioning surface With processing placement surface The intersection of As the origin of the machining element coordinate system; obtain the machining placement surface The normal vector , positioning surface The normal vector , positioning surface The normal vector ; The normal vector ,vector and the normal vector As the machining element coordinate system axis; When processing element shaping, including: Traversing manufacturing feature points The parameter variable set , through the parameter variable set The name of the parameter variable determines the meaning of the parameter variable. According to the association relationship between the parameter variable and the attribute adjacency graph, the geometric surface associated with the parameter variable is obtained, and the value of the parameter variable is determined.

[0030] In the above step S3, when outputting the second planned processing path, it includes: Constructing a road network 0-1 matrix for the processing path according to the second three-dimensional geometric model; Using mathematical software to obtain manufacturing feature points The connectivity index is The algorithm obtains any two manufacturing feature points The matrix of the shortest path between them, obtaining all manufacturing feature points Betweenness, and analyze the betweenness index to obtain the first planned processing path; Divide the first planned processing path into several equal parts according to the preset time interval , is the start time of the machining path, is the end time of the processing path; Get The position of the machining tool at the moment , The manufacturing feature point closest to the machining tool at the moment Location ; Get the waiting position of the machining tool when starting machining , the first manufacturing feature point on the first planned processing path Location ; Based on time and location and A rectangular coordinate system with the origin as the horizontal axis, the time as the horizontal axis, and the position as the vertical axis; Will Coordinates at the moment and As the training sample input of the neural network, Moment and The position difference corresponding to the moment uses the error function of the neural network The gain factor is obtained as output; The position change at the next moment is re-input into the neural network, and the gain coefficient updated in real time is output for adaptive optimization until the hybrid-based Balance, real-time planning is completed, and the second planned processing path is output; Among them, mathematical software includes: software.

[0031] In the above step S4, based on the Gaussian distribution function, the When the spline curve is used for path optimization, it includes: Define a Gaussian distribution function in the processing environment, perform random sampling within a preset range close to obstacles, and generate random sampling points , select random sampling points from the distance according to the Gaussian distribution function for , and the point located on the second planned processing path ; judge and Are the collision situations the same? If so, the corresponding Discard, if the collision situation is different, then the corresponding The collision-free point planning processing path within the range near the obstacle is finally obtained; Through three times The spline curve smoothes the planned processing path without collision points to obtain the third planned processing path, as shown in Formula 4: (4); In the formula, For the The equations of the segment control points, For three times The basis functions of the spline, Plan the processing path for the third; in, The preset distance.

[0032] In the above step S5, when performing simulation verification, it includes: Let the second 3D geometric model be , the geometric model of the tool is ; By using the bounding box method , Separate packaging boxes , To express; Run the tool based on the third planned machining path. If , it is considered that the third planned processing path will collide with the part to be processed; Get the curvature change rate of the third planned processing path, as shown in Formula 5: (5); In the formula, is the curvature change rate, is the tangent angle of the third planned machining path, The length of the third planned processing path; Determine whether the curvature change rate is less than a preset value. If so, it is considered that the smoothness of the third planned processing path meets the requirements; The tool path simulation software uses cutting force to simulate the force of the tool on the third planned machining path to identify whether there is an area where the tool stress exceeds the threshold; The third planned machining path is simulated by tool path simulation software, and the surface roughness, dimensional accuracy and shape error of the processed parts are evaluated by high-precision probe, three-dimensional measuring machine and optical scanner respectively; Among them, a high-precision probe is used to measure the surface of the processed part, and the surface roughness is evaluated by the average value of the absolute value of the surface profile deviation from the average line, as shown in Formula 6: (6); In the formula, is the height of each point on the surface of the machined part, is the average height of the surface of the machined part, is the number of sampling points for the part surface height data, is the average value of the absolute value of the surface profile deviation from the mean line; The machined parts are precisely measured using a three-dimensional coordinate measuring machine, and the dimensional accuracy is evaluated by comparing the dimensional error with the target size in the CAD model, as shown in Formula 7: (7); In the formula, is the size error, is the measured size of the processed part, is the target size; The processed parts are scanned in three dimensions using an optical scanner, and the shape error is evaluated by overlapping the shape error with the CAD model, as shown in Formula 8: (8); In the formula, is the surface point of the machined part, is the surface point of the CAD model.

[0033] In the above step S6, if Figure 2 As shown in the figure, when adjusting and optimizing the path, it includes: The manufacturing feature points on the third planned processing path according to The axis coordinate values ​​are divided into several sub-areas; Treat each sub-area as a small Use ant colony algorithm to optimize and obtain optimized data sets for each sub-area; The optimized data set of each sub-area is used as a new small-scale , and the ant colony algorithm is used again for optimization to obtain the final planned processing path; Among them, the ant colony algorithm includes: Each ant starts from the starting point of the third planned processing path, moves along the third planned processing path, and gradually adjusts the optimized path according to the pheromone concentration and heuristic information; the pheromone concentration on the path is updated, as shown in Formula 9: (9); In the formula, is the updated pheromone concentration on the path, is the current pheromone concentration on the path, is the amount of pheromone added to the path, is the pheromone volatility coefficient.

[0034] The above-mentioned embodiments are only preferred embodiments of the present invention and cannot be used to limit the scope of protection of the present invention. Any non-substantial changes and substitutions made by technicians in this field on the basis of the present invention shall fall within the scope of protection required by the present invention.

Claims

1. A high-precision parts machining path planning method based on artificial intelligence optimization, characterized in that: The following steps are involved: Generate a first three-dimensional geometric model of the part to be processed and perform surface division, and obtain manufacturing feature points according to the division results; Determine the processing placement surface for the first three-dimensional geometric model according to the manufacturing feature points, perform positioning and shaping of the processing elements, and obtain a second three-dimensional geometric model; Perform preliminary planning of machining paths and use neural networks based on hybrid Balanced optimization is performed and the second planned processing path is output; Based on Gaussian distribution function, cubic The spline curve is used to optimize the path and obtain the third planning processing path; Simulation verification is carried out, and based on the verification results, the hierarchical optimization ant colony algorithm is used to adjust and optimize the path, and the final planned processing path is output.

2. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 1 is characterized in that: When generating the first 3D geometric model, include: Get the geometry of the part to be processed and all its edges, and put them into the edge collection ; Get edge set The edges with arc type are put into the arc edge sub-collection middle; Get the arc edge subset The arc center points of all the edges in the circle are put into the center point set The center point number is ; Provides a temporary maximum circle center point set , the maximum circle center point set ,parameter ,parameter ; like , then execute the parallel judgment step; like , then execute the first three-dimensional geometric model generation step.

3. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 2 is characterized in that: When performing surface division, it includes: Get the surface area in the first 3D geometric model and perform Spline surface configuration; according to The spline surface configuration obtains the first-order and second-order partial derivatives of the surface area; The regions where the deviations of the first-order and second-order partial derivatives are both smaller than a threshold are classified as a region set, thereby dividing the surface region into several surface sub-regions.

4. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 3 is characterized in that: When obtaining manufacturing feature points, include: Convex edge decomposition is performed based on several surface sub-regions, and the base surface is matched with the predefined manufacturing template to obtain the manufacturing feature points. , while giving the geometric surface mapping attributes; When determining the processing placement surface, include: Get manufacturing feature points The face collection , get the face set from the first 3D geometric model Adjacent face sets ; If the face collection If there is a unique surface, the unique surface is used as the processing placement surface ; If the face collection If there are multiple faces, obtain the faces parallel to the base face among the multiple faces, and take the face with the largest distance from the base face as the processing placement face .

5. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 4 is characterized in that: When positioning the processing element, it includes: From manufacturing feature points Get the positioning surface , positioning surface , and align the positioning surface , positioning surface The type is identified and the corresponding processing element coordinate system is constructed according to the identification result; When processing element shaping, including: Traversing manufacturing feature points The parameter variable set , through the parameter variable set The name of the parameter variable determines the meaning of the parameter variable. According to the association relationship between the parameter variable and the attribute adjacency graph, the geometric surface associated with the parameter variable is obtained, and the value of the parameter variable is determined.

6. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 5 is characterized in that: When making preliminary planning for the machining path, include: Constructing a road network 0-1 matrix for the processing path according to the second three-dimensional geometric model; Get manufacturing feature points The connectivity index is The algorithm obtains any two manufacturing feature points The matrix of the shortest path between them, obtaining all manufacturing feature points The betweenness is calculated and the betweenness index is analyzed to obtain the first planned processing path.

7. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 6 is characterized in that: When outputting the second planned processing path, it includes: Divide the first planned processing path into several equal parts according to the preset time interval , , The start and end time of the processing path; Get The position of the machining tool at the moment , The manufacturing feature point closest to the machining tool at the moment Location ; Get the waiting position of the machining tool when starting machining , the first manufacturing feature point on the first planned processing path Location ; Based on time and location and A rectangular coordinate system with as the origin.

8. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 7 is characterized in that: When outputting the second planned processing path, it also includes: Will Coordinates at the moment and As the training sample input of the neural network, Moment and The position difference corresponding to the moment uses the error function of the neural network The gain factor is obtained as output; The position change at the next moment is re-input into the neural network, and the gain coefficient updated in real time is output for adaptive optimization until the hybrid-based Balance and output the second planned processing path.

9. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 8 is characterized in that: When performing path optimization, it includes: Define a Gaussian distribution function in the processing environment, perform random sampling within a preset range close to obstacles, and generate random sampling points , select random sampling points from the distance according to the Gaussian distribution function for , and the point located on the second planned processing path ; judge and Are the collision conditions the same? If so, the corresponding If they are not the same, the corresponding Keep it and get the processing path without collision points; Through three times The spline curve is used to smooth the planned processing path without collision points to obtain the third planned processing path; in, The preset distance.

10. The high-precision parts machining path planning method based on artificial intelligence optimization according to claim 9 is characterized in that: When adjusting and optimizing the path, include: The manufacturing feature points on the third planned processing path according to The axis coordinate values ​​are divided into several sub-areas; Treat each sub-area as a small Use ant colony algorithm to optimize and obtain optimized data sets for each sub-area; The optimized data set of each sub-area is used as a new small-scale , and the ant colony algorithm is used again for optimization to obtain the final planned processing path; The ant colony algorithm includes: each ant starts from the starting point of the third planned processing path, moves along the third planned processing path, and gradually adjusts the optimized path according to the pheromone concentration and heuristic information.

Images